tracks30000.tsv\n\nhead -200000 artist_tags.tsv > artist_tags200K.tsv \n\n\n"},{"alpha_fraction":0.7053459286689758,"alphanum_fraction":0.7182297110557556,"avg_line_length":43.11070251464844,"blob_id":"a9d168831822b1242bfbc717004f4fb39d178621","content_id":"f6e44ad26ed136b9bcb9d8cb905bf8ad1b561740","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":11953,"license_type":"no_license","max_line_length":496,"num_lines":271,"path":"/docs/CreateMahoutRecommender.md","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"# Mahout for Newbies: How to Create a Recommender #\nA short tutorial on actions you have to take to create your first Mahout recommender.\n\n## Introduction ##\n\n[Mahout](http://mahout.apache.org/) is a library containing scalable implementation of many machine learning algorithms, notably the collaborative filtering and other algorithms used in recommender systems.\n\nThere are many tutorials on creating simple or more complex recommenders but for Mahout newbies it's quite difficult to cope with all the technical details you have to do to make Mahout work in your way.\n\nI'm not an Eclipse fan, so all instructions use only the shell console. I was running the examples on Ubuntu 10.04, but it should be working on most of the linux distros and even on windows with cygwin.\n\nThis page is aiming to be a step-by-step description for Mahout/Maven beginners. I'm also a Mahout/Maven beginner, so any corrections and comments are more then welcome.\n\n## Part 1: Installation ##\nThere's a good [installation guide in the Mahout wiki](https://cwiki.apache.org/confluence/display/MAHOUT/BuildingMahout).\n\nI'm running the 0.4 Mahout release, so all stuff in this article is primarily for that, but should be working also on trunk and other releases.\n\nAs described in the guide, after JDK and Maven you should do something like:\n```\ncd core\nmvn compile\nmvn install\n```\n\nA few things to note here:\n * If you installation fails, try the fix proposed in the tutorial:\n```\nrm -rf ~/.m2/\nmvn clean install\n```\n * If this doesn't help, try again\n```\nmvn install\n```\nand look to the `core/target/surefire-reports/`. If you can find the cause of the problem there, you're a lucky person. If not, you can seek help on the `user@mahout.apache.org` mailing list as [I successfully did.](http://comments.gmane.org/gmane.comp.apache.mahout.user/5881)\n * Alternatively, you can try running the install without tests\n```\nmvn -Dmaven.test.skip=true install\n```\n\n## Part 2: Creating a simple recommender ##\n\nNow that you have mahout successfully installed, you can try some built-in examples, as [described in the Mahout wiki](https://cwiki.apache.org/confluence/display/MAHOUT/RecommendationExamples). Firstly you have to download some dataset, and then pass it as the `-i` parameter. It's a kind of a detective work but you can find some hints in the source code of the examples. I've successfully run the recommender on [the Book Crossing dataset](http://www.informatik.uni-freiburg.de/~cziegler/BX/)\n\nAnd now to creating your own recommender on your own data. Editing the examples isn't a good idea, because the code gets quite messy and all the paths are terribly long. Instead of that follow the oncoming steps.\n\n### Create a Maven project ###\nFirstly, you have to create a maven project. Go to the Mahout directory (the parent of the core directory) and run something like:\n```\nmvn archetype:create -DarchetypeGroupId=org.apache.maven.archetypes -DgroupId=com.unresyst -DartifactId=mahoutrec\n```\nThis creates an empty project called `mahoutrec` with the package namespace `com.unresyst`. Now change to the `mahoutrec` directory. You can try out the new project by running:\n```\nmvn compile\nmvn exec:java -Dexec.mainClass=\"com.unresyst.App\"\n```\nThis should print the Hello world message. If not try passing the `-e` option for the second command and try to realize what's going wrong.\n\nMore info on creating a Maven project can be found in [the official guide](http://maven.apache.org/guides/getting-started/)\n\n### Set the project dependencies ###\nNow that you have your maven project, let's set it up for your Mahout instalation.\n\nOpen the `pom.xml` file from your project directory (`mahoutrec`) in your favourite text editor. And now:\n * add the Mahout dependencies to your `pom.xml`. The following `` elements must be under the `` parent element.\n```\n \n org.apache.mahout \n mahout-core \n 0.4 \n \n \n org.apache.mahout \n mahout-math \n 0.4 \n \n \n org.apache.mahout \n mahout-math \n 0.4 \n test-jar \n test \n \n \n org.apache.mahout \n mahout-utils \n 0.4 \n \n```\n\n> If you are using other than the 0.4 version replace the 0.4 by the number of your version.\n * Set the relative path to the parent project `pom.xml` by adding the `` element to the `` element:\n```\n ../pom.xml \n```\nYour whole `pom.xml` should look like that:\n```\n\n\n 4.0.0 \n \n mahout \n org.apache.mahout \n 0.4 \n ../pom.xml \n \n com.unresyst \n mahoutrec \n 1.0-SNAPSHOT \n mahoutrec \n http://maven.apache.org \n \n \n junit \n junit \n 3.8.1 \n test \n \n \n org.apache.mahout \n mahout-core \n 0.4 \n \n \n org.apache.mahout \n mahout-math \n 0.4 \n \n \n org.apache.mahout \n mahout-math \n 0.4 \n test-jar \n test \n \n \n org.apache.mahout \n mahout-utils \n 0.4 \n \n \n \n UTF-8 \n \n \n```\n\nNow try compiling your project by running\n```\nmvn compile\n```\nYou should get a Build successful message.\n\n### Create a simple recommender ###\nAnd now to the promised creation of your own recommender running on your own data. Firstly we'll create some dummy data to feed the recommender. Create a `datasets` directory in your project directory and create the `dummy-bool.csv` file with the following contents:\n```\n#userId,itemId\n1,3\n1,4\n2,44\n2,46\n3,3\n3,5\n3,6\n4,3\n4,5\n4,11\n4,44\n5,1\n5,2\n5,4\n```\nThe first line is a comment, all the others are boolean preferences for items expressed by the users. Boolean preferences are useful when you don't have any explicit user-item ratings in your data. A boolean preference can mean e.g. that the user has viewed a profile page of the item. The example file states that the user id `1` has viewed the item id `3`, and item id `4`, user `2` item `44` and `46`, and so on.\n\nNow, let's create the recommender. Create a file UnresystBoolRecommend.java in the well hidden subddirectory `src/main/java/com/unresyst/` of your project directory. Place the following content to the file\n```\npackage com.unresyst;\n\nimport java.io.File;\nimport java.io.FileNotFoundException;\nimport java.util.List;\nimport java.io.IOException;\n\nimport org.apache.commons.cli2.OptionException; \nimport org.apache.mahout.cf.taste.common.TasteException;\nimport org.apache.mahout.cf.taste.impl.model.file.FileDataModel;\nimport org.apache.mahout.cf.taste.impl.recommender.CachingRecommender;\nimport org.apache.mahout.cf.taste.impl.recommender.slopeone.SlopeOneRecommender;\nimport org.apache.mahout.cf.taste.model.DataModel;\nimport org.apache.mahout.cf.taste.recommender.RecommendedItem;\nimport org.apache.mahout.cf.taste.impl.common.LongPrimitiveIterator;\n\npublic class UnresystBoolRecommend {\n \n public static void main(String... args) throws FileNotFoundException, TasteException, IOException, OptionException {\n \n // create data source (model) - from the csv file \n File ratingsFile = new File(\"datasets/dummy-bool.csv\"); \n DataModel model = new FileDataModel(ratingsFile);\n \n // create a simple recommender on our data\n CachingRecommender cachingRecommender = new CachingRecommender(new SlopeOneRecommender(model));\n\n // for all users\n for (LongPrimitiveIterator it = model.getUserIDs(); it.hasNext();){\n long userId = it.nextLong();\n \n // get the recommendations for the user\n List recommendations = cachingRecommender.recommend(userId, 10);\n \n // if empty write something\n if (recommendations.size() == 0){\n System.out.print(\"User \");\n System.out.print(userId);\n System.out.println(\": no recommendations\");\n }\n \n // print the list of recommendations for each \n for (RecommendedItem recommendedItem : recommendations) {\n System.out.print(\"User \");\n System.out.print(userId);\n System.out.print(\": \");\n System.out.println(recommendedItem);\n }\n } \n }\n}\n```\nSome notes to that:\n * If Java isn't your tongue language, be aware of the package declaration on the first line. If you omit that the Maven runner will never find your class.\n * In the main function we first create a data model from our csv file, then we instantiate a simple Slope One recommender filled with the data. Then we iterate over all users in the dataset and we print recommendations for them.\n * Sorry for my poor java.\n\nTo run the recommender on our dummy data, do:\n```\nmvn compile\n```\nThis should end up by BUILD SUCCESSFUL. Then run the main function:\n```\nmvn exec:java -Dexec.mainClass=\"com.unresyst.UnresystBoolRecommend\" \n```\n\nIn the middle of the Maven spam you should be able to see:\n```\nUser 1: RecommendedItem[item:5, value:1.0]\nUser 2: RecommendedItem[item:5, value:1.0]\nUser 2: RecommendedItem[item:3, value:1.0]\nUser 3: no recommendations\nUser 4: no recommendations\nUser 5: RecommendedItem[item:5, value:1.0]\nUser 5: RecommendedItem[item:3, value:1.0]\n```\nThese are recommendations for our users. Our recommender recommends the item `5` to the user `1`, with certainty 1.0 (this is a constant value for boolean recommenders), items `5` and `3` to user 2 and so on.\n\nTo reduce the Maven spam you can pass the `-q` option, to debug pass the `-e` option.\n\n## Conclusion ##\nThat's it, now you've got your first Mahout recommender. You can now experiment with various data and various recommenders as described in other Mahout recommender tutorials. Some of the good ones:\n * A good starting guide, from which the example is partly overtaken: http://philippeadjiman.com/blog/2009/11/11/flexible-collaborative-filtering-in-java-with-mahout-taste/\n * The definitive IBM guide: http://www.ibm.com/developerworks/java/library/j-mahout/\n\nThanks for reading, any suggestions, corrections and comments are appreciated.\n\nSources and further reading\n * https://cwiki.apache.org/confluence/display/MAHOUT/Mahout+Wiki\n * https://cwiki.apache.org/confluence/display/MAHOUT/BuildingMahout\n * http://maven.apache.org/plugins/maven-surefire-plugin/howto.html\n * http://www.jarvana.com/jarvana/view/org/apache/mahout/mahout-core/0.4/mahout-core-0.4-javadoc.jar!/index-all.html\n * http://shuyo.wordpress.com/2011/02/14/mahout-development-environment-with-maven-and-eclipse-2/"},{"alpha_fraction":0.5778035521507263,"alphanum_fraction":0.5778035521507263,"avg_line_length":38.21590805053711,"blob_id":"31bc717d67fe630404e9b123a8fffacc887f435a","content_id":"b3eb7cf40522a9aca1ea39fc12659a7f44225cba","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3451,"license_type":"no_license","max_line_length":108,"num_lines":88,"path":"/UnresystCD/code/adapter/unresyst/algorithm/compiling_algorithm.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The CompilingAlgorithm class\"\"\"\n\nfrom base import BaseAlgorithm\n\nclass CompilingAlgorithm(BaseAlgorithm):\n \"\"\"The algorithm that compiles aggregated similarities and biases with\n the predictions.\n \"\"\"\n def __init__(self, inner_algorithm, compilator):\n \"\"\"The initializer\"\"\"\n \n super(CompilingAlgorithm, self).__init__(inner_algorithm=inner_algorithm)\n \n self.compilator=compilator\n \"\"\"The compilator that will be used during the build\"\"\"\n\n def build(self, recommender_model):\n \"\"\"See the base class for documentation.\n \n Compiles and calls the inner algorithm build\n \"\"\" \n print \" Compiling aggregates and predictions.\"\n \n self.compilator.compile_all(recommender_model) \n \n\n print \"Predictions compiled. Building the inner algorithm...\"\n \n super(CompilingAlgorithm, self).build(recommender_model=recommender_model) \n\n \n def get_relationship_prediction(self, recommender_model, dn_subject, dn_object, remove_predicted):\n \"\"\"See the base class for the documentation.\n \n Get all available info for the pair.\n \"\"\"\n \n # get what we have from the inner algo\n inner_prediction = super(CompilingAlgorithm, self).get_relationship_prediction(\n recommender_model=recommender_model,\n dn_subject=dn_subject,\n dn_object=dn_object,\n remove_predicted=remove_predicted)\n \n # if it's trivial, than return it\n if remove_predicted and inner_prediction.is_trivial:\n return inner_prediction\n \n # if it's not uncertain, it was already available for the inner algorithm,\n # return it\n if not inner_prediction.is_uncertain:\n return inner_prediction\n \n # otherwise compile the prediction from all available info \n prediction = self.compilator.compile_prediction(\n recommender_model=recommender_model,\n dn_subject=dn_subject,\n dn_object=dn_object)\n\n # if it found something, return it \n if prediction: \n return prediction\n \n # otherwise return the uncertain\n return self._get_uncertain_prediction(\n recommender_model=recommender_model, \n dn_subject=dn_subject, \n dn_object=dn_object\n ) \n \n def get_recommendations(self, recommender_model, dn_subject, count, expectancy_limit, remove_predicted):\n \"\"\"See the base class for the documentation. \n\n Here: fill the recommendations until the count\n with random predictions, fill all info we know, or create uncertain \n predictions.\n \"\"\"\n # get what we have from the inner algo\n inner_recs = super(CompilingAlgorithm, self).get_recommendations(\n recommender_model=recommender_model,\n dn_subject=dn_subject,\n count=count,\n expectancy_limit=expectancy_limit,\n remove_predicted=remove_predicted)\n\n # add there some random and find info for them \n # to samy jako nahore\n return inner_recs\n"},{"alpha_fraction":0.5001169443130493,"alphanum_fraction":0.5038596391677856,"avg_line_length":35.85344696044922,"blob_id":"7c921da5cce0e18772c9d5d422e47874ee9e14b0","content_id":"db3e6d7cc76017b8f98f4769ac32aca2d7aac2a7","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":4275,"license_type":"no_license","max_line_length":150,"num_lines":116,"path":"/UnresystCD/code/adapter/unresyst/recommender/rank_evaluation.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Evaluator defining a method for the average prediction rank\"\"\"\n\nfrom unresyst.exceptions import EmptyTestSetError\nfrom unresyst.constants import *\nfrom unresyst.recommender.evaluation import BaseEvaluator\n\ndef index(seq, f):\n \"\"\"Return the index of the first item in seq where f(item) == True.\"\"\"\n return next((i for i in xrange(len(seq)) if f(seq[i])), None)\n\n\nclass RankEvaluator(BaseEvaluator):\n \"\"\"The evaluator for counting the average prediction rank\"\"\"\n \n SUBJ_IDS = []\n \"\"\"A list of subjects (domain specific ids) that are to be tested\"\"\"\n \n \n @classmethod\n def evaluate_predictions(cls, recommender, save_predictions=False):\n \"\"\"See the base class for documentation.\n \n Here it counts the rank, without calling any metric.\n \"\"\"\n \n # get the test pairs\n qs_pairs = cls._get_cleared_pairs()\n \n # evaluate the rank for each subject\n # \n\n obj_count = recommender.objects.count()\n rank_sum = 0\n rank_count = 0\n\n print \"Evaluating %d subjects\" % len(cls.SUBJ_IDS) \n \n \n # go through the wished subjects\n for subj, i in zip(recommender.subjects.filter(pk__in=cls.SUBJ_IDS), range(1, len(cls.SUBJ_IDS))):\n \n # if the test set for the subject is emty go ahead\n test_set_subj = qs_pairs.filter(subj=subj) \n if not test_set_subj:\n continue\n \n # create an array of object predictions for the subject\n #\n neg_array = []\n pos_array = []\n uncertain_count = 0 \n \n for obj in recommender.objects.all():\n \n # get the prediction\n exp_prediction = recommender.predict_relationship(subj, obj)\n \n if exp_prediction is None or exp_prediction.is_uncertain: \n uncertain_count += 1\n \n # otherwise add it to the right array\n else:\n arr = neg_array if exp_prediction.expectancy > UNCERTAIN_PREDICTION_VALUE \\\n else pos_array\n \n arr.append((obj, exp_prediction.expectancy))\n\n print \"all predicted\"\n\n # sort the array by expectancy\n pos_array.sort(key=lambda el: el[1], reverse=True) \n neg_array.sort(key=lambda el: el[1], reverse=True) \n\n print \"all sorted\" \n pos_count = len(pos_array)\n neg_count = len(neg_array)\n \n # count the rank that will be given to uncertains.\n if uncertain_count:\n \n # the rank of the first uncertain\n rank_first = float(pos_count) / (obj_count -1)\n \n # the rank of the last uncertain\n rank_last = float(pos_count + uncertain_count) / (obj_count - 1)\n \n uncertain_rank = (rank_first + rank_last) / 2\n \n rank_count += test_set_subj.count()\n \n # count the rank for all objects for the subjects in the test set\n for obj in test_set_subj:\n \n # try finding it in positives\n ipos = index(pos_array, lambda el: el[0] == obj)\n if not (ipos is None):\n rank_sum += float(ipos) / (obj_count - 1)\n continue\n \n # try finding it in negatives\n ineg = index(neg_array, lambda el: el[0] == obj)\n if not (ineg is None):\n rank_sum += float(pos_count + uncertain_count + ineg) / (obj_count - 1)\n continue\n \n # otherwise it's uncertain\n rank_sum += uncertain_rank\n \n print \"%d subjects evaluated. current avg rank: %f, current rank count: %d\" % (i, rank_sum/rank_count if rank_count else -1.0, rank_count)\n \n\n res = rank_sum / rank_count\n \n print \"Average rank: %f\" % res\n \n return res\n"},{"alpha_fraction":0.8055555820465088,"alphanum_fraction":0.8055555820465088,"avg_line_length":70,"blob_id":"ff968e097fa8518be7d988877b40629eb655decc","content_id":"435580c197e421c8796039673b75a2a29be25919","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":72,"license_type":"no_license","max_line_length":70,"num_lines":1,"path":"/UnresystCD/code/mahout/mahoutrec/unresystboolrecommend.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"mvn -q exec:java -Dexec.mainClass=\"com.unresyst.UnresystBoolRecommend\" \n"},{"alpha_fraction":0.653235673904419,"alphanum_fraction":0.6556776762008667,"avg_line_length":34.0428581237793,"blob_id":"c55413980319a261b4eb939e2134eb85490aabc3","content_id":"9496068ad98a18b21d0150eb245db04ed942f8c8","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2457,"license_type":"no_license","max_line_length":82,"num_lines":70,"path":"/UnresystCD/code/adapter/unresyst/models/aggregator.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Models which instances are created by the aggregator package.\n\nThe aggregated relationship.\n\"\"\"\n\nfrom django.db import models\n\nfrom base import BaseRelationshipInstance\nfrom unresyst.constants import *\n\nclass AggregatedRelationshipInstance(BaseRelationshipInstance):\n \"\"\"A representation of an aggregated relationship between two subject/objects\n \n There can be only one aggregated relationship for each subject/object pair\n for a recommender.\n \"\"\"\n\n expectancy = models.FloatField()\n \"\"\"The probability of the relationship between subject/object.\n A number from [0, 1].\n \"\"\"\n \n recommender = models.ForeignKey('unresyst.Recommender')\n \"\"\"The recommender it belongs to\"\"\"\n \n relationship_type = models.CharField(max_length=MAX_LENGTH_RELATIONSHIP_TYPE, \n choices=RELATIONSHIP_TYPE_CHOICES)\n \"\"\"A string indicating whether it's a S-O, O-O, S-S, or SO-SO relationship.\"\"\"\n \n additional_unique = ('recommender', )\n \"\"\"There can be multiple pairs for one recommender\"\"\"\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\" \n ret = super(AggregatedRelationshipInstance, self).__unicode__()\n\n return ret + \", %f\" % (self.expectancy) \n \n class Meta:\n app_label = 'unresyst' \n \n unique_together = ('subject_object1', 'subject_object2', 'recommender')\n \"\"\"For each recommender there can be only one subject-object pair.\"\"\"\n \nclass AggregatedBiasInstance(models.Model):\n \"\"\"An aggregated bias of a subjectobject\"\"\" \n \n expectancy = models.FloatField()\n \"\"\"The probability of the subject/object being in the predicted_relationship\n A number from [0, 1].\n \"\"\"\n \n subject_object = models.ForeignKey('unresyst.SubjectObject')\n \"\"\"The biased subject/object.\"\"\" \n \n recommender = models.ForeignKey('unresyst.Recommender')\n \"\"\"The recommender it belongs to\"\"\"\n \n description = models.TextField(default='', blank=True)\n \"\"\"The filled description of the aggregated bias.\"\"\" \n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\" \n return u\"%s: %f\" % (self.subject_object, self.expectancy)\n \n class Meta:\n app_label = 'unresyst' \n \n unique_together = ('subject_object', 'recommender')\n \"\"\"For each recommender there can be only one subject-object pair.\"\"\" \n"},{"alpha_fraction":0.6512749791145325,"alphanum_fraction":0.6529979109764099,"avg_line_length":41.67647171020508,"blob_id":"bfc9e424f3f5bdfbb30e9355241bfcb7ef2d306e","content_id":"ed6202f65567939db68bc509050d78c74a0f14f9","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Java","length_bytes":2902,"license_type":"no_license","max_line_length":129,"num_lines":68,"path":"/code/mahout/mahoutrec/src/main/java/com/unresyst/UnresystRecommend.java","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"package com.unresyst;\n\nimport java.io.File;\nimport java.io.FileNotFoundException;\nimport java.util.List;\nimport java.io.IOException;\nimport java.io.PrintWriter;\nimport java.io.OutputStreamWriter;\nimport java.io.FileOutputStream;\nimport java.nio.charset.Charset;\n\nimport org.apache.commons.cli2.OptionException; \nimport org.apache.mahout.cf.taste.common.TasteException;\nimport org.apache.mahout.cf.taste.impl.model.file.FileDataModel;\nimport org.apache.mahout.cf.taste.impl.recommender.CachingRecommender;\nimport org.apache.mahout.cf.taste.impl.recommender.slopeone.SlopeOneRecommender;\nimport org.apache.mahout.cf.taste.model.DataModel;\nimport org.apache.mahout.cf.taste.recommender.RecommendedItem;\nimport org.apache.mahout.cf.taste.impl.common.LongPrimitiveIterator;\nimport org.apache.mahout.common.FileLineIterable;\nimport org.apache.mahout.common.IOUtils;\n\npublic class UnresystRecommend {\n \n public static void main(String[] args) throws FileNotFoundException, TasteException, IOException, OptionException {\n\n // parameters: train data filename, number of recommendations per user , output filename \n String trainFilename = args[0]; \n int recCount = Integer.parseInt(args[1]);\n String outputFilename = args[2];\n \n // create data source (model) - from the train csv file \n File ratingsFile = new File(trainFilename); \n DataModel model = new FileDataModel(ratingsFile); \n \n // create a simple recommender on our data\n CachingRecommender cachingRecommender = new CachingRecommender(new SlopeOneRecommender(model));\n \n // create recommendations for all users and write them to the output file\n //\n \n // open the file to write to\n File resultFile = new File(outputFilename);\n PrintWriter writer = new PrintWriter(new OutputStreamWriter(new FileOutputStream(resultFile), Charset.forName(\"UTF-8\")));\n \n // for all users\n for (LongPrimitiveIterator it = model.getUserIDs(); it.hasNext();){\n long userId = it.nextLong();\n \n // get the recommendations for the user\n List recommendations = cachingRecommender.recommend(userId, recCount); \n \n \n // print the list of recommendations for each recommendation\n for (RecommendedItem recommendedItem : recommendations) {\n\n long itemId = recommendedItem.getItemID();\n float prediction = recommendedItem.getValue();\n \n writer.println(String.format(\"%d,%d,%f\", userId, itemId, prediction));\n }\n }\n \n // writer spam\n writer.flush();\n IOUtils.quietClose(writer); \n }\n}\n"},{"alpha_fraction":0.6050138473510742,"alphanum_fraction":0.6061981916427612,"avg_line_length":28.97633171081543,"blob_id":"621e8ed2a71bc636df4a679453b9be65b4736211","content_id":"68f32ee383b09220c15fbba11b7af8ff59160436","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":5066,"license_type":"no_license","max_line_length":95,"num_lines":169,"path":"/code/adapter/travel/models.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Models for the travel agency dataset\"\"\"\n\nfrom django.db import models\n\n\nfrom constants import *\nfrom unresyst.models import BaseEvaluationPair\n\nclass User(models.Model):\n \"\"\"The user\"\"\"\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return 'user_%d' % self.id \n\nclass Session(models.Model):\n \"\"\"User's session with the system\"\"\"\n \n user = models.ForeignKey('travel.User')\n \"\"\"The user\"\"\"\n \n session_no = models.PositiveIntegerField()\n \"\"\"The number of session for the user\"\"\"\n \n class Meta:\n unique_together = ('user', 'session_no')\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return u'user %s: session: %d' % (self.user, self.session_no)\n \nclass Country(models.Model):\n \"\"\"The country for a tour.\"\"\"\n \n name = models.CharField(max_length=MAX_LENGTH_NAME, unique=True)\n \"\"\"The name of the country\"\"\"\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n\n\nclass TourType(models.Model):\n \"\"\"A model for a type of the tour.\"\"\" \n \n name = models.CharField(max_length=MAX_LENGTH_NAME, unique=True)\n \"\"\"The name of the type.\"\"\" \n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n\n\nclass Tour(models.Model):\n \"\"\"The tour\"\"\" \n \n name = models.CharField(max_length=MAX_LENGTH_NAME, unique=True)\n \"\"\"The name\"\"\"\n \n url = models.URLField(max_length=MAX_LENGTH_URL, verify_exists=False)\n \"\"\"The url without the final id and without GET parameters\"\"\"\n \n country = models.ForeignKey('travel.Country')\n \"\"\"The country the tour is to\"\"\"\n \n tour_type = models.ForeignKey('travel.TourType')\n \"\"\"The type of the tour\"\"\"\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n\n\nclass Action(models.Model):\n \"\"\"An abstract class for a user action\"\"\"\n\n session = models.ForeignKey('travel.Session')\n \"\"\"The session in which the action was taken (identifies the user)\"\"\"\n \n tour = models.ForeignKey('travel.Tour')\n \"\"\"The tour on which the action was taken\"\"\"\n \n timestamp = models.DateTimeField()\n \"\"\"The date and time the track was played\"\"\"\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return u\"%s: %s\" % (self.session.user, self.tour)\n \n class Meta:\n abstract = True \n \nclass Order(Action):\n \"\"\"User ordered a tour\"\"\"\n pass\n\nclass Question(Action):\n \"\"\"User asked a question about a tour\"\"\"\n pass\n \nclass Click(Action):\n \"\"\"User clicked on something on the tour profile\"\"\"\n pass\n \nclass MouseMove(Action):\n \"\"\"User moved the mouse over something on the tour profile\"\"\"\n pass\n \nclass ViewProfile(Action):\n \"\"\"User viewed the profile of the tour\"\"\"\n \n duration = models.PositiveIntegerField(null=True)\n \"\"\"How long the user has been viewing the profile, \n In seconds.\"\"\" \n \n\nclass TourOrderEvalPair(BaseEvaluationPair):\n \"\"\"An abstract class for the artist evaluators\"\"\"\n\n subj = models.ForeignKey('travel.User')\n \"\"\"The subject\"\"\"\n \n obj = models.ForeignKey('travel.Tour')\n \"\"\"The object\"\"\"\n \n test_ratio = 0.2\n \"\"\"The ratio of pairs to select to test pairs\"\"\"\n\n class Meta:\n app_label = 'travel' \n \n @classmethod \n def select(cls, i=0):\n \"\"\"See the base class for the documentation.\"\"\" \n \n all_count = Order.objects.count() \n test_count = int(cls.test_ratio * all_count ) \n train_count = all_count - test_count\n\n min_stamp_test = Order.objects.order_by('-timestamp')[test_count-1].timestamp\n\n test_pairs=Order.objects.filter(timestamp__gte=min_stamp_test)\n train_pairs = Order.objects.filter(timestamp__lt=min_stamp_test)\n \n # remove all other feedback newer than the test timestamp\n ViewProfile.objects.filter(timestamp__gt=min_stamp_test).delete()\n MouseMove.objects.filter(timestamp__gt=min_stamp_test).delete()\n Click.objects.filter(timestamp__gt=min_stamp_test).delete()\n Question.objects.filter(timestamp__gt=min_stamp_test).delete()\n \n # take 1/5 of the orders, remove them and put them to test data \n # save to test, remove from build\n for order in test_pairs.iterator():\n\n # create the test pair for the order\n TourOrderEvalPair.objects.create(\n subj=order.session.user,\n obj=order.tour,\n expected_expectancy=EXPECTED_EXPECTANCY_ORDERED,\n )\n \n \n # delete the order\n order.delete()\n \n print \"%d test pairs selected from total %d pairs.\" % (test_count, all_count) \n\n def get_success(self):\n return True\n"},{"alpha_fraction":0.6163253784179688,"alphanum_fraction":0.6197383403778076,"avg_line_length":32.46666717529297,"blob_id":"5191888627dd960a304bff6a9351cab6d5565e4d","content_id":"d455b849ec7714785a1e0ffbf66fbc292018548f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3516,"license_type":"no_license","max_line_length":87,"num_lines":105,"path":"/UnresystCD/code/adapter/unresyst/models/base.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Base classes for the unresyst application.\"\"\"\n\nfrom django.db import models\nfrom django.contrib.contenttypes.models import ContentType\n\nfrom symmetric import SymmetricalRelationship\nfrom unresyst.constants import *\n\nclass ContentTypeModel(models.Model):\n \"\"\"An abstract base class for all models having some subclasses whose\n instances shouldn't be always returned\n \n Be careful when using multiple inheritance: this class shouldbe used\n as the last and save() shouldn't be overriden in none of the other\n base classes\n \"\"\"\n\n content_type = models.ForeignKey(ContentType,editable=False, null=True)\n \"\"\"The actual type of the object.\"\"\"\n \n @classmethod\n def remove_subclass_objects(cls):\n \"\"\"Get instances only of this class\"\"\"\n # get the content type\n cont_type = ContentType.objects.get_for_model(cls)\n \n # return only this cont type\n return cls.objects.filter(content_type=cont_type)\n \n def as_leaf_class(self):\n \"\"\"Get the object as the whole inherited class\"\"\"\n\n # get the leaf class\n content_type = self.content_type\n Model = content_type.model_class()\n \n # get the appropriate object\n if (Model == self.__class__):\n return self\n return Model.objects.get(pk=self.pk) \n\n def save(self,*args, **kwargs):\n \"\"\"Save with the right content type\"\"\"\n \n if not self.content_type:\n self.content_type = ContentType.objects.get_for_model(self.__class__)\n \n super(ContentTypeModel, self).save(*args, **kwargs)\n\n\n class Meta:\n abstract = True \n app_label = 'unresyst' \n\n\nclass BaseRelationshipInstance(SymmetricalRelationship):\n \"\"\"An abstract class, the base class of all s-o, s-s, o-o relationships.\"\"\"\n\n subject_object1 = models.ForeignKey('unresyst.SubjectObject', \\\n related_name='%(class)s_relationships1')\n \"\"\"The first subject/object that is in the relationship.\"\"\" \n \n subject_object2 = models.ForeignKey('unresyst.SubjectObject', \\\n related_name='%(class)s_relationships2')\n \"\"\"The second subject/object that is in the relationship\"\"\" \n \n description = models.TextField(default='', blank=True)\n \"\"\"The description of the relationship/rule instance.\"\"\" \n \n \n attr_name1 = 'subject_object1'\n \"\"\"Overriden attribute name 1\"\"\"\n \n attr_name2 = 'subject_object2'\n \"\"\"Overriden attribute name 2\"\"\"\n\n class Meta:\n abstract = True \n app_label = 'unresyst' \n\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return u\"(%s, %s)\" % (self.subject_object1, self.subject_object2) \n\n\nclass BaseRelationshipDefinition(ContentTypeModel):\n \"\"\"A definition of the relationship that should be predicted. There's only\n one for a recommender.\n \"\"\" \n \n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"The name of the relationship\"\"\" \n \n recommender = models.ForeignKey('unresyst.Recommender')\n \"\"\"The recommender to which the definition belongs. Each recommender has\n exactly one predicted relationship.\n \"\"\" \n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n \n class Meta:\n app_label = 'unresyst'\n\n\n"},{"alpha_fraction":0.6709677577018738,"alphanum_fraction":0.7419354915618896,"avg_line_length":29.799999237060547,"blob_id":"1ea1e45a1f2dcdcfa4e4e78454aaa69a127b4577","content_id":"25141e17388802b89ba36cbcaf6eeb24da4e6e83","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":155,"license_type":"no_license","max_line_length":43,"num_lines":5,"path":"/UnresystCD/code/dataset_scripts/process_flixster.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"sort -n links.txt -o slinks.txt\nsort -n ratings.txt -o sratings.txt\n\nhead -15000 sratings.txt > mini/ratings.tsv\nhead -105000 slinks.txt > mini/links.tsv\n\n"},{"alpha_fraction":0.6352763175964355,"alphanum_fraction":0.6361846923828125,"avg_line_length":31.69801902770996,"blob_id":"9d281e87a9264e93f7f2b601d8131a225df6259b","content_id":"cf3bc001473d3162fc7420079e877bb8cd03be11","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":6605,"license_type":"no_license","max_line_length":87,"num_lines":202,"path":"/UnresystCD/code/adapter/unresyst/recommender/base.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Module containing the definition of BaseRecommender: the base class \nfor user-defined recommenders.\n\"\"\"\n\nfrom unresyst.constants import *\n\nclass BaseRecommender(object):\n \"\"\"The base class for all recommenders.\n \n Defines and the interface of the recommender. All methods and data fields\n are empty.\n \"\"\"\n\n # Methods:\n #\n #\n\n # Build phase:\n # \n @classmethod\n def build(cls):\n \"\"\"Build the recommender. Process all the rules and relationships \n in order to be able to provide recommendations.\n \"\"\"\n pass\n \n # Recommend phase:\n #\n \n @classmethod \n def predict_relationship(cls, subject, object_, save_to_db=False): \n \"\"\"Get the prediction of the given relationship type between the subject\n and the object. \n \n @type subject: domain specific subject\n @param subject: the subject\n \n @type object_: domain specific object\n @param object_: the object\n \n @type save_to_db: bool\n @param save_to_db: should the obtained predictions be saved, so that it\n doesn't have to be computed again when asked for the pair next time\n \n @rtype: RelationshipPrediction\n @return: An instance representing the prediction of the relationship \n between the subject and the object, containing the estimated \n probability and explanation. If the prediction isn't known, returns\n an prediction with expectancy 0.5 and empty explanation\n \n @raise InvalidParameterError: if the given subject or object doesn't \n have a domain neutral representation in the unresyst database. \n Either the recommender hasn't been built or the subject/object was\n added later without updating the recommender.\n \"\"\"\n pass\n \n @classmethod\n def get_recommendations(cls, subject, count=None):\n \"\"\"Get recommendations for the given subject.\n\n @type subject: domain specific subject\n @param subject: the subject\n \n @type count: int\n @param count: a maximum number of objects to be recommended, if not given,\n the default_recommendation_count from the recommender class is used.\n \n @rtype: list of RelationshipPrediction\n @return: An instance representing the prediction of the relationship \n between the subject and the object, containing the estimated \n probability and explanation. If the prediction isn't known, returns\n an prediction with expectancy 0.5 and empty explanation\n\n @raise InvalidParameterError: if the given subject doesn't \n have a domain neutral representation in the unresyst database. \n Either the recommender hasn't been built or the subject was\n added later without updating the recommender. \n \"\"\"\n pass\n \n \n # Update phase:\n # \n @classmethod \n def add_subject(cls, subject):\n \"\"\"Add subject to the recommender.\"\"\"\n pass\n\n @classmethod\n def add_object(cls, object_):\n \"\"\"Add object to the recommender.\"\"\"\n pass\n\n @classmethod\n def update_subject(cls, subject):\n \"\"\"Update the subject in the recommender, including its relationships\n and applied rules\"\"\"\n pass\n\n\n @classmethod \n def update_object(cls, object_):\n \"\"\"Update the object in the recommender, including its relationships\n and applied rules\"\"\"\n pass\n\n\n @classmethod\n def remove_subject(cls, subject):\n \"\"\"Remove the subject from the recommender, including its relationships\n and applied rules\"\"\"\n pass\n\n\n @classmethod \n def remove_object(cls, object_):\n \"\"\"Remove the object from the recommender, including its relationships\n and applied rules\"\"\"\n pass \n \n \n \n # Domain specific data. empty, will be overriden in the domain specific recommender\n #\n\n name = \"\"\n \"\"\"The name of the recommender\"\"\"\n \n subjects = None\n \"\"\"The objects to who the recommender will recommend.\n Requires the following interface:\n on subjects manager:\n - iterator(): get an iterator on the collection\n - all(): get all sujbects\n - get(id=..) get the object with the given id\n queryset:\n - exists(): is there something in the queryset?\n on each subject instance:\n - id: an integer id of the subject \n - __unicode__(): printable string\n \"\"\"\n \n objects = None\n \"\"\"The objects that will be recommended.\"\"\" \n\n predicted_relationship = None\n \"\"\"The relationship that will be predicted\"\"\"\n \n relationships = ()\n \"\"\"Relationships among the subjects and objects in the domain\"\"\"\n \n rules = ()\n \"\"\"Rules that can be applied to the domain\"\"\"\n \n cluster_sets = ()\n \"\"\"Clusters to which subjects and objects can be divided\"\"\"\n \n biases = ()\n \"\"\"Bias - predisposition of subjects/objects to be in the predicted_relationship\"\"\"\n\n random_recommendation_description = None\n \"\"\"The description that will be used as a description \n to random recommendations.\"\"\"\n\n # Auxiliary class attributes\n _recommender_model = None\n \"\"\"The database model instance belonging to the class\"\"\"\n \n # Class configuration - the behaviour of the layers below the recommender\n # Empty, will be overriden by the Recommender class\n #\n \n abstractor = None\n \"\"\"The class that will be used for the abstractor level. Can be \n overriden in subclasses\"\"\" \n \n algorithm = None\n \"\"\"The class that will be used for the algorithm level. Can be \n overriden in subclasses\"\"\"\n \n remove_predicted_from_recommendations = None\n \"\"\"Should the instances of the predicted_relationship be removed from\n recommendation list?\n \"\"\"\n \n save_all_to_predictions = None\n \"\"\"Should also instances of predicted_relationship be saved to predictions?\n \"\"\"\n \n recommendation_expectancy_limit = None\n \"\"\"The limit for expectancy above which the objects can be recommended.\n If not none, only objects with expectancy above the limit are recommended.\n A reasonable limit is 0.5 - the recommendations then don't include random \n objects.\n \"\"\"\n \n verbose_build = None\n \"\"\"Should messages be printed during the build?\"\"\"\n \n explicit_rating_rule = None\n \"\"\"If given, this rule is exported, not the predicted_relationship\"\"\"\n"},{"alpha_fraction":0.8416422009468079,"alphanum_fraction":0.8416422009468079,"avg_line_length":36.88888931274414,"blob_id":"5f3d97050e753b19fe0d90682b372202a501009f","content_id":"d6ffb603555e521ec623ae4497ae716e0fe36042","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":341,"license_type":"no_license","max_line_length":69,"num_lines":9,"path":"/UnresystCD/code/adapter/unresyst/algorithm/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The package contains all classes for the algorithm layer.\n\nThe algorithm layer can be replaced by a common algorithm exploiting \nthe known preferences, like collaborative filtering.\n\"\"\"\n\nfrom simple_algorithm import SimpleAlgorithm\nfrom aggregating_algorithm import AggregatingAlgorithm\nfrom compiling_algorithm import CompilingAlgorithm\n"},{"alpha_fraction":0.6346863508224487,"alphanum_fraction":0.6346863508224487,"avg_line_length":19.820512771606445,"blob_id":"ff32dd7225c9e17f47ae2c0d66a81b75a053a4cd","content_id":"ae5df24917045a80cc0081f9e32c1d5fa3829312","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":813,"license_type":"no_license","max_line_length":132,"num_lines":39,"path":"/UnresystCD/code/adapter/lastfmeval.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"#!/bin/sh\n\n# run the evaluation- recommendations. parameters \n# novel -> NovelArtistRecommender, \n# nenovel -> ArtistRecommender\n# dontbuild-> don't build\n\nBUILD=true\nEVAL=''\nREC=''\n\nfor param in $*;\ndo \n case $param in\n 'novel')\n EVAL='NovelArtistRankEvaluator'\n REC='NovelArtistRecommender'\n ;;\n 'nenovel')\n EVAL='ArtistRankEvaluator'\n REC='ArtistRecommender'\n ;;\n 'dontbuild')\n BUILD=false\n ;;\n esac\ndone\n\nif [ $BUILD = true ]\nthen\n # build\n echo \"Building...\"\n echo \"from lastfm.recommender import *; $REC.build()\" | python ./manage.py shell\nfi\n\necho \"Evaluating recommendations...\"\necho \"from lastfm.evaluation import *; from lastfm.recommender import *; $EVAL.evaluate_predictions($REC)\"| python ./manage.py shell\n\necho \"\"\n\n"},{"alpha_fraction":0.6459985375404358,"alphanum_fraction":0.6459985375404358,"avg_line_length":20.671875,"blob_id":"240d9c3e17fedad9e7ad2dc48ce541cc1373512c","content_id":"acecebc085fa5fe72c71f0bf3e43802cb269eee3","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":1387,"license_type":"no_license","max_line_length":132,"num_lines":64,"path":"/UnresystCD/code/adapter/build.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"# build the shoe recommender\n\nLASTFM=false\nNLASTFM=false\nDEMO=false\nADEMO=false\nFLIXSTER=false\nTRAVEL=false\n\nfor param in $*;\ndo \n case $param in\n 'lastfm')\n LASTFM=true\n ;;\n 'nlastfm')\n NLASTFM=true\n ;;\n 'demo')\n DEMO=true\n ;;\n 'flixster')\n FLIXSTER=true\n ;;\n 'travel')\n TRAVEL=true\n ;;\n 'ademo')\n ADEMO=true\n ;; \n esac\ndone\n\nif [ $DEMO = true ]\nthen\n echo \"from demo.recommender import ShoeRecommender; ShoeRecommender.build(); quit();\" | python ./manage.py shell\nfi\n\nif [ $ADEMO = true ]\nthen\n echo \"from demo.recommender import AverageRecommender; AverageRecommender.build(); quit();\" | python ./manage.py shell\nfi\n\nif [ $LASTFM = true ]\nthen\n echo \"from lastfm.recommender import ArtistRecommender; ArtistRecommender.build(); quit();\" | python ./manage.py shell\nfi\n\nif [ $NLASTFM = true ]\nthen\n echo \"from lastfm.recommender import NovelArtistRecommender; NovelArtistRecommender.build(); quit();\" | python ./manage.py shell\nfi\n\nif [ $FLIXSTER = true ]\nthen\n echo \"from flixster.recommender import MovieRecommender; MovieRecommender.build(); quit();\" | python ./manage.py shell\nfi\n\nif [ $TRAVEL = true ]\nthen\n echo \"from travel.recommender import OrderTourRecommender; OrderTourRecommender.build(); quit();\" | python ./manage.py shell\nfi\n\necho \"\"\n"},{"alpha_fraction":0.8062015771865845,"alphanum_fraction":0.8062015771865845,"avg_line_length":31.25,"blob_id":"03d93fb7e5af5f9cfcc13af9d980d345de9381c7","content_id":"728db3828f4936747110a00115337c277963c69e","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":129,"license_type":"no_license","max_line_length":81,"num_lines":4,"path":"/UnresystCD/code/adapter/unresyst/recommender/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Package holding the BaseRecommender class, holding the interface for the whole\nunresyst.\"\"\"\n\nfrom base import BaseRecommender\n"},{"alpha_fraction":0.6535211205482483,"alphanum_fraction":0.6788732409477234,"avg_line_length":21.125,"blob_id":"51a21eb6ed1100fac55f6654d42dd24e300e396c","content_id":"46b996f5dd70d8892ad94fe110f7a4e6d7be3cfd","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":355,"license_type":"no_license","max_line_length":61,"num_lines":16,"path":"/code/adapter/flixster/constants.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Constants for flixster\"\"\"\n\nMIN_POSITIVE_RATING = 3.0\n\"\"\"Minimal rating taken as positive\"\"\"\n\nMIN_HIGH_RATING = 4.0\n\"\"\"Minimal rating taken as \"high\".\"\"\"\n\nMAX_LOW_RATING = 1.0\n\"\"\"Maximal rating taken as \"low\".\"\"\"\n\nMAX_STARS = 5\n\"\"\"Maximal number of starts in rating\"\"\"\n\nMAX_TOLERANCE = 1.0\n\"\"\"Size of the interval to take a prediction as successful\"\"\"\n\n"},{"alpha_fraction":0.5889149904251099,"alphanum_fraction":0.598078191280365,"avg_line_length":36.93692398071289,"blob_id":"dc6153f14a58a55b7cb4816548e8b114e8d4c65d","content_id":"5ad83dc9f7b4f6ed08692ad2a5336acebffd92d5","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":24664,"license_type":"no_license","max_line_length":112,"num_lines":650,"path":"/UnresystCD/code/adapter/unresyst/recommender/rules.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The classes for representing business rules and relationships\"\"\"\n\nfrom unresyst.constants import *\n\nfrom unresyst.models.abstractor import *\nfrom unresyst.models.common import SubjectObject\nfrom unresyst.exceptions import DescriptionKeyError, ConfigurationError\n\nclass BaseRelationship(object):\n \"\"\"A base class for representing all relationships and rules.\n \n A subclass for all classes representing a relationship between entities (not necessarily \n of the same type). Contains the condition that is true between and only \n between the entities that are in the given relationship.\n \"\"\"\n \n def __init__(self, name, condition=None, description=None, generator=None):\n \"\"\"The constructor.\"\"\"\n \n self.name = name\n \"\"\"The name of the rule/relationship.\"\"\"\n \n self.condition = condition\n \"\"\"A boolean function that represents the condition. If True \n for the given pair of entities, there's the Relationship between the \n entities. \n Should be simple.\n \"\"\"\n \n self.description = description\n \"\"\"A string describing the rule. It can contain placeholders for entities: \n \n - %(subject)s, %(object)s for subject-object relationships and rules\n - %(subject1)s, %(subject2)s for subject-subject relationships \n and rules\n - %(object1)s, %(object2)s for object-object relationships and rules\n - %(subjectobject1)s, %(subjectobject2)s for recommenders where \n subject domain is the same as object domain \n \"\"\"\n \n self.generator = generator\n \"\"\"A generator returning pairs of objects that are in the relationship/\n the rule applies to them.\n For performance reasons - if given, the pairs will be taken from it \n without the need for evaluating the condition for each possible pair.\n \"\"\"\n \n DESCRIPTION_FORMAT_DICT = {\n RELATIONSHIP_TYPE_SUBJECT_OBJECT: \n (FORMAT_STR_SUBJECT, FORMAT_STR_OBJECT),\n RELATIONSHIP_TYPE_SUBJECT_SUBJECT: \n (FORMAT_STR_SUBJECT1, FORMAT_STR_SUBJECT2),\n RELATIONSHIP_TYPE_OBJECT_OBJECT: \n (FORMAT_STR_OBJECT1, FORMAT_STR_OBJECT2),\n RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT: \n (FORMAT_STR_SUBJECTOBJECT1, FORMAT_STR_SUBJECTOBJECT2)\n }\n \"\"\"A dictionary relationship type (e.g. 'S-O') a pair of formatting strings \n for description, e.g. ('subject', 'object').\n \"\"\"\n\n relationship_type = RELATIONSHIP_TYPE_SUBJECT_OBJECT\n \"\"\"The type of the relationship S-O\"\"\"\n \n InstanceClass = RelationshipInstance\n \"\"\"The model class used for representing instances of the rule/relationship\"\"\"\n \n DefinitionClass = RuleRelationshipDefinition\n \"\"\"The model class used for representing the definition of the \n rule/relationship\n \"\"\"\n \n is_symmetric = False\n \"\"\"Is the rule/relationship between the entities of the same domain?\n True for S-S, O-O, SO-SO\n \"\"\"\n \n def get_filled_description(self, arg1, arg2):\n \"\"\"Get description for a rule/relationship instance, between \n arg1 and arg2. \n \n @type arg1: models.SubjectObject\n @param arg1: the first subjectobject in the relationship/rule\n\n @type arg2: models.SubjectObject\n @param arg2: the second subjectobject in the relationship/rule\n \n @rtype: str\n @return: a string with filled gaps for entities.\n \"\"\"\n # get the format strings, e.g. ('subject', 'object')\n format_strings = self.DESCRIPTION_FORMAT_DICT[self.relationship_type]\n \n # create the formating strings to be passed to description\n format_dict = {\n format_strings[0]: arg1,\n format_strings[1]: arg2\n }\n if self.description is None:\n desc = ''\n else:\n try: \n desc = self.description % format_dict\n except KeyError, e:\n raise DescriptionKeyError(\n message=\"There's an invalid key in description\",\n recommender=self.recommender,\n name=self.name, \n key=e.__str__(), \n permitted_keys=format_dict.keys()\n )\n \n return desc\n \n def get_create_definition_kwargs(self):\n \"\"\"Get dictionary of parameters for the definition model constructor. \n \n @rtype: dictionary string: object\n @return: the kwargs of the definition model constructor \n \"\"\"\n return {\n \"name\": self.name,\n \"recommender\": self.recommender._get_recommender_model(),\n }\n \n def get_additional_instance_kwargs(self, ds_arg1, ds_arg2):\n \"\"\"Get dictionary of additional kwargs for creating the rule/relationship\n instance. \n \n @type ds_arg1: django.db.models.manager.Manager\n @param ds_arg1: the first argument of the rule instance - domain specific\n \n @type ds_arg2: django.db.models.manager.Manager\n @param ds_arg2: the second argument of the rule instance - domain specific\n \n @rtype: dictionary string: object\n @return: additional keyword args for creating rule/relationship instance \n \"\"\"\n return {}\n \n \n def evaluate_on_dn_args(self, dn_arg1, dn_arg2, definition): \n \"\"\"Evaluates the rule on the given arguments. If evaluated positively,\n a new rule/relationship instance is saved.\n \n @type dn_arg1: models.SubjectObject\n @param dn_arg1: a domain neutral representation of the first entity.\n\n @type dn_arg2: models.SubjectObject\n @param dn_arg2: a domain neutral representation of the second entity.\n \n @type definition: models.abstractor.RuleRelationshipDefinition\n @param definition: the model representing the rule/relationship \n definition\n \n @rtype: int\n @return: 1 if something has benn created, 0 if not\n \"\"\"\n \n # get the domain specific objects for our universal representations\n #\n arg1_manager = self.recommender._get_entity_manager(dn_arg1.entity_type)\n ds_arg1 = dn_arg1.get_domain_specific_entity(entity_manager=arg1_manager)\n \n arg2_manager = self.recommender._get_entity_manager(dn_arg2.entity_type)\n ds_arg2 = dn_arg2.get_domain_specific_entity(entity_manager=arg2_manager)\n\n # if the condition is satisfied\n if self.condition(ds_arg1, ds_arg2):\n \n self._perform_save_instance(definition, ds_arg1, ds_arg2, dn_arg1, dn_arg2)\n return 1\n \n return 0\n\n @classmethod\n def order_arguments(cls, dn_arg1, dn_arg2):\n \"\"\"Order the arguments as they appear in the relationship.\"\"\"\n a1, a2, x, x = cls._order_in_pair(dn_arg1, dn_arg2, None, None)\n \n return (a1, a2)\n\n @classmethod\n def _order_in_pair(cls, dn_arg1, dn_arg2, ds_arg1, ds_arg2):\n \"\"\"Swap the arguments in the rule/relationships so that the first\n has a lower id than the second\n \"\"\"\n\n # if the second argument has lower id than the first, swap them\n if dn_arg2.pk < dn_arg1.pk: \n return (dn_arg2, dn_arg1, ds_arg2, ds_arg1)\n\n # otherwise not\n return (dn_arg1, dn_arg2, ds_arg1, ds_arg2)\n \n\n def _perform_save_instance(self, definition, ds_arg1, ds_arg2, dn_arg1, dn_arg2):\n \"\"\"Perform the action of creating and saving the instance\"\"\"\n \n # order the instances in pairs as the class requires\n dn_arg1, dn_arg2, ds_arg1, ds_arg2 = self._order_in_pair(dn_arg1, dn_arg2, ds_arg1, ds_arg2)\n \n add_kwargs = self.get_additional_instance_kwargs(ds_arg1, ds_arg2)\n \n # create a rule/relationship instance\n instance = self.InstanceClass(\n definition=definition,\n subject_object1=dn_arg1,\n subject_object2=dn_arg2,\n description=self.get_filled_description(dn_arg1, dn_arg2),\n **add_kwargs)\n \n instance.save()\n\n\n def save_instance(self, ds_arg1, ds_arg2, definition): \n \"\"\"Save an instance of the rule/relationship for the given args.\n \n @type ds_arg1: domain specific subject/object\n @param ds_arg1: a domain specific representation of the first entity.\n\n @type ds_arg2: domain specific subject/object\n @param ds_arg2: a domain specific representation of the second entity.\n \n @type definition: models.abstractor.RuleRelationshipDefinition\n @param definition: the model representing the rule/relationship \n definition\n \n @raise ConfigurationError: thrown if the condition doesn't evaluate \n to true on the given pair\n \"\"\"\n if not (self.condition is None) and not self.condition(ds_arg1, ds_arg2):\n raise ConfigurationError(\n message=(\"The condition wasn't evaluated as true for the pair \" + \n \"%s, %s, even though it was returned by the generator.\") % (ds_arg1, ds_arg2), \n recommender=self.recommender, \n parameter_name='rules/relationships', \n parameter_value=self.name)\n \n # convert the domain specific to domain neutral \n #\n arg1_ent_type, arg2_ent_type = self.relationship_type.split(RELATIONSHIP_TYPE_SEPARATOR) \n\n dn_arg1 = SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=ds_arg1, \n entity_type=arg1_ent_type, \n recommender=definition.recommender)\n\n dn_arg2 = SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=ds_arg2, \n entity_type=arg2_ent_type, \n recommender=definition.recommender)\n \n # create and save the instance\n self._perform_save_instance(definition, ds_arg1, ds_arg2, dn_arg1, dn_arg2) \n \n \n \n def evaluate(self):\n \"\"\"Evaluate the rule on all subjects/objects - pairs.\n \n Creates and saves the rule/relationship definition, creates and saves\n rule instances. \n \"\"\"\n \n # obtain the kwargs for creating the definition\n def_kwargs = self.get_create_definition_kwargs()\n\n # create and save the definition\n definition = self.DefinitionClass(**def_kwargs)\n definition.save()\n\n i = 0 \n # if we have a generator, use it for looping through pairs\n if not (self.generator is None):\n \n\n # loop through pairs, save the rule/relationship instances\n for ds_arg1, ds_arg2 in self.generator():\n self.save_instance(ds_arg1, ds_arg2, definition)\n i += 1\n \n print \" %d instances of rule/rel %s created\" % (i, self.name)\n \n # that's it\n return\n \n \n # otherwise take the entities one by one\n \n # parse what should be used as condition args\n arg1_s, arg2_s = self.relationship_type.split(RELATIONSHIP_TYPE_SEPARATOR) \n \n if arg1_s == arg2_s:\n \n # loop only through the matrix members below the diagonal \n # \n i = 0\n for arg1, arg2 in SubjectObject.unique_pairs(\n recommender=self.recommender._get_recommender_model(),\n entity_type=arg1_s): \n # evaluate it\n i += self.evaluate_on_dn_args(arg1, arg2, definition)\n \n\n \n else:\n # filter subjectobjects for my recommender\n qs_recommender = SubjectObject.objects.filter(\n recommender=self.recommender._get_recommender_model())\n \n # go through all things that have to be as first and as second param\n for arg1 in qs_recommender.filter(entity_type=arg1_s).iterator():\n for arg2 in qs_recommender.filter(entity_type=arg2_s).iterator():\n \n # evaluate the rule/relationship on the given args\n i += self.evaluate_on_dn_args(arg1, arg2, definition)\n\n print \" %d instances of rule/rel %s created\" % (i, self.name)\n\n \n def export(self, f):\n \"\"\"Export the relationship as lines to the given file object.\n \n Exports the relationship in form:\n subject_id, object_id[,confidence]\\n\n \n confidence is exported only if it's relevant,\n no matter if it's positive or not\n \n @type f: file\n @param f: the file to write to\n \n @raise: ConfigurationError: if the rule/relationship doesn't have\n a generator.\n \"\"\"\n \n # if we don't have a generator raise an error\n if self.generator is None:\n raise ConfigurationError(\n message=\"The relationship to export is missing a generator.\", \n recommender=cls, \n parameter_name='?', \n parameter_value=cls.name)\n \n i = 0\n \n # loop through pairs, export the rule/relationship instances\n for ds_arg1, ds_arg2 in self.generator():\n \n # create the common part\n linestr = \"%s,%s\" % (ds_arg1.pk, ds_arg2.pk)\n \n # get the confidence if provided\n adkwargs = self.get_additional_instance_kwargs(ds_arg1, ds_arg2)\n \n # if confidence provided, append it\n if adkwargs.has_key(EXPECTANCY_KWARG_NAME):\n linestr += \",%f\" % adkwargs[EXPECTANCY_KWARG_NAME]\n \n linestr += '\\n'\n \n # write it to the file\n f.write(linestr)\n \n i += 1\n \n print \" %d instances of rule/rel %s exported\" % (i, self.name)\n\n \nclass PredictedRelationship(BaseRelationship):\n \"\"\"A class for representing the predicted relationship.\"\"\"\n\n DefinitionClass = PredictedRelationshipDefinition\n \"\"\"The model class used for representing the definition of the \n rule/relationship\n \"\"\" \n \n @classmethod\n def _order_in_pair(cls, dn_arg1, dn_arg2, ds_arg1, ds_arg2):\n \"\"\"Swap the arguments in the relationship so that the first\n is always a subject and second the object.\n \"\"\" \n \n # for the SO entities we apply the normal policy\n if dn_arg1.entity_type == ENTITY_TYPE_SUBJECTOBJECT:\n return super(PredictedRelationship, cls)._order_in_pair(dn_arg1, dn_arg2, ds_arg1, ds_arg2)\n\n # otherwise we put subjects as first\n \n # if the first is object, swap\n if dn_arg1.entity_type == ENTITY_TYPE_OBJECT:\n return (dn_arg2, dn_arg1, ds_arg2, ds_arg1)\n \n # if not, keep\n return (dn_arg1, dn_arg2, ds_arg1, ds_arg2)\n\nclass ExplicitSubjectObjectRule(BaseRelationship):\n \"\"\"A class for representing explicit preference like rating, for attributes where\n the user can express both positive and negative preference.\"\"\"\n\n relationship_type = RELATIONSHIP_TYPE_SUBJECT_OBJECT\n \"\"\"The type of the relationship S-O\"\"\" \n\n DefinitionClass = ExplicitRuleDefinition\n \"\"\"The model class used for representing the definition of the \n rule/relationship\n \"\"\" \n \n InstanceClass = ExplicitRuleInstance\n \"\"\"The model class used for representing instances of \n the rule/relationship\"\"\"\n \n \n def __init__(self, name, expectancy, condition=None, description=None, generator=None):\n \"\"\"The constructor.\"\"\"\n \n super(ExplicitSubjectObjectRule, self).__init__(\n name=name, \n condition=condition, \n description=description, \n generator=generator)\n \n self.expectancy = expectancy\n \"\"\"A function taking a subject and an object, giving the explicit preference\n normalized to [0, 1].\n \"\"\"\n \n def get_additional_instance_kwargs(self, ds_arg1, ds_arg2):\n \"\"\"See the base class for documentation\n \n @raise ConfigurationError: if the confidence function returns a value\n outside [0, 1]\n \"\"\"\n ret_dict = super(ExplicitSubjectObjectRule, self)\\\n .get_additional_instance_kwargs(ds_arg1, ds_arg2)\n \n # call the user-defined confidence method\n expectancy = self.expectancy(ds_arg1, ds_arg2)\n \n if not (MIN_EXPECTANCY <= expectancy <= MAX_EXPECTANCY):\n raise ConfigurationError(\n message=(\"The rule '%s' has expectancy %f, for the\" + \\\n \" pair (%s, %s). Should be between 0 and 1.\") % \\\n (self.name, expectancy, ds_arg1, ds_arg2),\n recommender=self.recommender,\n parameter_name=\"Recommender.rules\",\n parameter_value=self.recommender.rules\n )\n \n ret_dict[EXPECTANCY_KWARG_NAME] = expectancy\n return ret_dict \n\n @classmethod\n def _order_in_pair(cls, dn_arg1, dn_arg2, ds_arg1, ds_arg2):\n return PredictedRelationship._order_in_pair(dn_arg1, dn_arg2, ds_arg1, ds_arg2)\n \nclass _WeightedRelationship(BaseRelationship):\n \"\"\"A class representing a relationship with a weight.\"\"\"\n\n DefinitionClass = RuleRelationshipDefinition\n \"\"\"The model class used for representing the definition of the \n rule/relationship\n \"\"\" \n \n def __init__(self, name, is_positive, weight, condition=None, description=None, generator=None):\n \"\"\"The constructor.\"\"\"\n \n super(_WeightedRelationship, self).__init__(\n name=name, \n condition=condition, \n description=description, \n generator=generator)\n \n self.is_positive = is_positive\n \"\"\"Is the relationship positive to the predicted relationship?\"\"\"\n \n self.weight = weight\n \"\"\"A float number from [0, 1] representing the *static* weight of the rule. \n It doesn't depend on the entity pair.\n \"\"\"\n\n\n def get_create_definition_kwargs(self):\n \"\"\"Get dictionary of parameters for the definition model constructor. \n \n Add the weight to the parameters. \n \n @rtype: dictionary string: object\n @return: the kwargs of the definition model constructor \n \n @raise ConfigurationError: if the weight isn't from [0, 1]\n \"\"\"\n ret_dict = super(_WeightedRelationship, self).get_create_definition_kwargs()\n \n ret_dict['is_positive'] = self.is_positive\n\n if not (MIN_WEIGHT <= self.weight <= MAX_WEIGHT):\n raise ConfigurationError(\n message=(\"The rule/relationship '%s' has weight %f,\" + \\\n \" should be between 0 and 1.\") % (self.name, self.weight),\n recommender=self.recommender,\n parameter_name=\"Recommender.rules or Recommender.relationships\",\n parameter_value=(self.recommender.rules, self.recommender.relationships)\n )\n \n ret_dict['weight'] = self.weight\n ret_dict[\"relationship_type\"] = self.relationship_type \n \n return ret_dict\n \n \n \n\nclass SubjectObjectRelationship(_WeightedRelationship):\n \"\"\"A class for representing subject-object preference for recommendation\"\"\"\n \n relationship_type = RELATIONSHIP_TYPE_SUBJECT_OBJECT\n \"\"\"The type of the relationship S-O\"\"\" \n\n @classmethod\n def _order_in_pair(cls, dn_arg1, dn_arg2, ds_arg1, ds_arg2):\n return PredictedRelationship._order_in_pair(dn_arg1, dn_arg2, ds_arg1, ds_arg2)\n\nclass _SimilarityRelationship(_WeightedRelationship):\n \"\"\"A base class (abstract) for all relationships operating between the same type \n and meaning similarity.\n \"\"\"\n \n is_symmetric = True\n \"\"\"For documentation see the base class\"\"\"\n \n\nclass ObjectSimilarityRelationship(_SimilarityRelationship):\n \"\"\"A class for representing inter-object similarity.\"\"\" \n\n relationship_type = RELATIONSHIP_TYPE_OBJECT_OBJECT\n \"\"\"The type of the relationship O-O\"\"\" \n\n\nclass SubjectSimilarityRelationship(_SimilarityRelationship):\n \"\"\"A class for representing inter-subject similarity.\"\"\"\n\n relationship_type = RELATIONSHIP_TYPE_SUBJECT_SUBJECT\n \"\"\"The type of the relationship S-S\"\"\"\n\n\nclass SubjectObjectSimilarityRelationship(_SimilarityRelationship):\n \"\"\"A class used only when subject domain equals object domain. \n For representing inter-entity similarity.\n \"\"\"\n \n relationship_type = RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT\n \"\"\"The type of the relationship SO-SO\"\"\"\n\n\n# rules:\n# \n\nclass _BaseRule(_WeightedRelationship):\n \"\"\"A base class for all rules (abstract).\"\"\"\n \n InstanceClass = RuleInstance\n \"\"\"The model class used for representing instances of \n the rule/relationship\"\"\"\n \n def __init__(self, name, is_positive, weight, confidence, condition=None, description=None, generator=None):\n \"\"\"The constructor.\"\"\" \n\n super(_BaseRule, self).__init__(\n name=name, \n condition=condition,\n is_positive=is_positive, \n weight=weight, \n description=description, \n generator=generator)\n \n self.confidence = confidence\n \"\"\"A float function giving values from [0, 1] representing the \n the confidence of the rule on the given pair. \n It's dynamic, depends on the entity pair.\n \"\"\" \n \n def get_additional_instance_kwargs(self, ds_arg1, ds_arg2):\n \"\"\"See the base class for documentation\n \n @raise ConfigurationError: if the confidence function returns a value\n outside [0, 1]\n \"\"\"\n ret_dict = super(_BaseRule, self).get_additional_instance_kwargs(\n ds_arg1, ds_arg2)\n \n # call the user-defined confidence method\n confidence = self.confidence(ds_arg1, ds_arg2)\n \n if not (MIN_CONFIDENCE <= confidence <= MAX_CONFIDENCE):\n raise ConfigurationError(\n message=(\"The rule '%s' has a confidence %f, for the\" + \\\n \" pair (%s, %s). Should be between 0 and 1.\") % \\\n (self.name, confidence, ds_arg1, ds_arg2),\n recommender=self.recommender,\n parameter_name=\"Recommender.rules\",\n parameter_value=self.recommender.rules\n )\n \n ret_dict[CONFIDENCE_KWARG_NAME] = confidence\n return ret_dict\n \n \n# confidence by taky mohla vracet string s doplnujicim vysvetlenim, \n\nclass _SimilarityRule(_BaseRule):\n \"\"\"A base class (abstract) for all rules operating between the same type \n and meaning similarity.\"\"\"\n\n is_symmetric = True\n \"\"\"For documentation see the base class\"\"\" \n \n\nclass ObjectSimilarityRule(_SimilarityRule):\n \"\"\"A class for representing inter-object similarity.\"\"\" \n\n relationship_type = RELATIONSHIP_TYPE_OBJECT_OBJECT\n \"\"\"The type of the relationship O-O\"\"\" \n\n\nclass SubjectSimilarityRule(_SimilarityRule):\n \"\"\"A class for representing inter-subject similarity.\"\"\"\n\n relationship_type = RELATIONSHIP_TYPE_SUBJECT_SUBJECT\n \"\"\"The type of the relationship S-S\"\"\" \n\n\nclass SubjectObjectRule(_BaseRule):\n \"\"\"A class for representing subject-object preference for recommendation\"\"\"\n\n relationship_type = RELATIONSHIP_TYPE_SUBJECT_OBJECT\n \"\"\"The type of the relationship S-O\"\"\" \n\n @classmethod\n def _order_in_pair(cls, dn_arg1, dn_arg2, ds_arg1, ds_arg2):\n return PredictedRelationship._order_in_pair(dn_arg1, dn_arg2, ds_arg1, ds_arg2)\n \nclass SubjectObjectSimilarityRule(_SimilarityRelationship):\n \"\"\"A class used only when subject domain equals object domain. \n For representing inter-entity rule.\n \"\"\"\n \n relationship_type = RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT\n \"\"\"The type of the relationship SO-SO\"\"\"\n\n \n"},{"alpha_fraction":0.6410256624221802,"alphanum_fraction":0.6538461446762085,"avg_line_length":25,"blob_id":"8bd5bbee2b8e0dfe1070d9756b7487bd06c7114b","content_id":"c5c7bf0826c526e2a2d602c31e306b308e2b8563","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":156,"license_type":"no_license","max_line_length":62,"num_lines":6,"path":"/code/adapter/dump.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"#!/bin/sh\n\n# dump the database to given file (relative from ../db/dumps/)\nmysqldump --opt -u root adapter > ../db/dumps/$1\n\necho \"dumped to ../db/dumps/$1\"\n"},{"alpha_fraction":0.6247366070747375,"alphanum_fraction":0.626033365726471,"avg_line_length":37.316768646240234,"blob_id":"cca61489cbc79c8f905c9f2ec184854d1dec0025","content_id":"c8c35fff934115f4b210c138e687c3751a25175c","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":6169,"license_type":"no_license","max_line_length":89,"num_lines":161,"path":"/UnresystCD/code/adapter/unresyst/models/common.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The models that are common to the whole unresyst application.\"\"\"\n\nfrom django.db import models\n\nfrom unresyst.constants import *\n\nclass Recommender(models.Model):\n \"\"\"The representation of a recommender. \n \n There can be multiple recommenders for one parent system.\n \"\"\"\n \n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"The name of the recommender\"\"\"\n \n class_name = models.CharField(max_length=MAX_LENGTH_CLASS_NAME, unique=True)\n \"\"\"The name of the recommender class. Has to be unique.\"\"\"\n\n are_subjects_objects = models.BooleanField()\n \"\"\"Are subjects == objects for the recommender?\"\"\"\n \n is_built = models.BooleanField()\n \"\"\"Is the recommender built?\"\"\"\n \n random_recommendation_description = models.TextField(default='', blank=True)\n \"\"\"The description of the relationship/rule instance.\"\"\" \n \n remove_predicted_from_recommendations = models.BooleanField()\n \"\"\"Should the objects that are already \"liked\" be removed from \n recommendations?\"\"\"\n \n class Meta:\n app_label = 'unresyst'\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n\n\nclass SubjectObject(models.Model):\n \"\"\"The common representation for a subject and an object.\"\"\"\n \n id_in_specific = models.CharField(max_length=MAX_LENGTH_ID)\n \"\"\"The id of the subject/object in the domain-specific system.\"\"\"\n \n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"A textual characterization of the subject/object\"\"\"\n \n entity_type = models.CharField(max_length=MAX_LENGTH_ENTITY_TYPE, \\\n choices=ENTITY_TYPE_CHOICES)\n \"\"\"A string indicating whether it's a subject, object or both.s/o/so\"\"\"\n \n recommender = models.ForeignKey('unresyst.Recommender')\n \"\"\"The recommender to which the subject/object belongs.\"\"\"\n \n\n class Meta:\n app_label = 'unresyst' \n \n unique_together = ('id_in_specific', 'entity_type', 'recommender')\n \"\"\"There can be only one subject/object with the given id and \n recommender.\n \"\"\"\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n \n @classmethod\n def get_domain_neutral_entity(cls, domain_specific_entity, entity_type, recommender):\n \"\"\"Get domain neutral representation of the given domain specific entity\n (subject/object/subjectobject)\n \n @type domain_specific_entity: django.db.models.Model\n @param domain_specific_entity: the domain specific subject/object/\n subjectobject for which the domain neutral repre should be got\n \n @type entity_type: str\n @param entity_type: 'S'/'O'/'SO' .. see constants, determines whether\n the entity is a subject, object or both\n \n @type recommender: models.Recommender\n @param recommended: the recommender the for which the entity should \n be got\n \n @rtype: models.SubjectObject\n @returns: the domain neutral representation of the entity \n \n @raise MultipleObjectsReturned: when there's a broken constraint in \n the unresyst database (hopefully never)\n @raise DoesNotExist: when the domain neutral representation for \n the given entity does not exist\n \"\"\"\n return cls.objects.get(\n id_in_specific=domain_specific_entity.pk,\n entity_type=entity_type,\n recommender=recommender)\n \n def get_domain_specific_entity(self, entity_manager):\n \"\"\"Get domain specific subject/object/both for this universal \n representation.\n \n @type entity_manager: django.db.models.manager.Manager\n @param entity_manager: the manager over the model containing \n the domain specific subjects/objects/bot\n \n @rtype: models.Model\n @returns: the domain specific entity for this universal representation\n \n @raise MultipleObjectsReturned: when there's a broken constraint in \n the client database (hopefully never)\n @raise DoesNotExist: when the domain specific entity for \n this universal entity does not exist\n \"\"\"\n return entity_manager.get(pk=self.id_in_specific)\n\n @classmethod\n def unique_pairs(cls, recommender, entity_type):\n \"\"\"A generator looping through the pairs of subjectobjects so that each two \n object set is returned only once. \n \n E.g. after (a, b), the (b, a) pair isn't returned. \n It doesn't return pairs like (a, a).\n \n Useful for symmetric rule and relationship evaluation.\n \n @type recommender: models.Recommender\n @param recommender: the recommender model for which the pairs should be\n obtained\n\n @type entity_type: str ('S', 'O' or 'SO') or None\n @param entity_type: the entity type from whic the pair should be taken,\n if None, all entity types are taken\n \n @rtype: generator for two-tuples\n @returns: pairs of subjectobjects entity_type entities belonging to \n the recommender.\n \"\"\"\n ent_type_kwargs = {} if entity_type is None \\\n else {'entity_type': entity_type}\n \n # get the subjectobjects to iterate over\n qs_entities = cls.objects.filter(\n recommender=recommender, \n **ent_type_kwargs).order_by('id')\n\n # the number of entities\n entity_count = qs_entities.count() \n \n # get the first argument and the number of entities that \n # will be taken as the second arg. Starting from 1, \n # finishing at .\n # The first entity will never be used as second argument \n for arg1, count in zip( \\\n qs_entities[1:].iterator(), \\\n range(1, entity_count)):\n\n # obtain only first count entities\n for arg2 in qs_entities[:count].iterator():\n\n yield (arg1, arg2)\n"},{"alpha_fraction":0.7341772317886353,"alphanum_fraction":0.7341772317886353,"avg_line_length":18.75,"blob_id":"be3fa2be861725aaa0b158caf2c22896177abf5b","content_id":"a301faf6bd839a0f7365d3ea650a3ddbf69c3595","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":79,"license_type":"no_license","max_line_length":44,"num_lines":4,"path":"/code/adapter/unresyst/tests/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The module containing tests for unresyst.\n\nIt uses the demo appliation.\n\"\"\"\n"},{"alpha_fraction":0.7977900505065918,"alphanum_fraction":0.8016999363899231,"avg_line_length":28.263681411743164,"blob_id":"087d0f7433293c7351326b79e62c3a817fdaf1c6","content_id":"5753a3ca5f8ff52fc6a7e772132a383b686a18d9","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":11765,"license_type":"no_license","max_line_length":149,"num_lines":402,"path":"/code/adapter/evaluate.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"#!/bin/sh\n\n# parametry: \n# 1. co chces evaluovat: lastfm, demo\n# 2. cislo iterace\n\n\nLASTFM=false\nDEMO=false\n\nfor param in $*;\ndo \n case $param in\n 'lastfm')\n LASTFM=true\n ;;\n 'demo')\n DEMO=true\n ;;\n esac\ndone\n\nif [ $LASTFM = true ]\nthen\n echo \"Creating test set.\"\n echo \"from lastfm.recommender import *; ArtistRecommender.ValidationPairClass.select_validation_pairs($2); quit()\" | python ./manage.py shell \n echo \"Building and evaluating recommender.\"\n echo \"from lastfm.recommender import *; ArtistRecommender.build(); ArtistRecommender.evaluate(); quit();\" | python ./manage.py shell\nfi\n\n# LASTFM\n#\n#\n\n# specificky pro lastfm - dva test sety\nfrom lastfm.models import *\nBaseArtistEvaluationPair.select()\n\n\n\n# evaluate novel mahout recommender:\n# \n\n# export training data\nfrom lastfm.mahout_recommender import *\nNovelMahoutArtistRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv')\n\n# export test data\nfrom lastfm.evaluation import *\nNovelArtistRankEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_test.csv')\n\n# run mahout train, test -> lastfm_predictions.csv\ncd ../mahout/mahoutrec\n./unresystpredict.sh lastfm\n\n# import predictions\nfrom lastfm.mahout_recommender import *\nNovelMahoutArtistRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_predictions.csv')\n\n# evaluate the predictions \nfrom lastfm.mahout_recommender import *\nfrom lastfm.evaluation import *\nNovelArtistRankEvaluator.evaluate_predictions(NovelMahoutArtistRecommender)\n\n# evaluate nenovel mahout recommender:\n# \n\n# export training data\nfrom lastfm.mahout_recommender import *\nMahoutArtistRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv')\n\n# export test data\nfrom lastfm.evaluation import *\nArtistRankEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_test.csv')\n\n# run mahout train, test -> lastfm_predictions.csv\ncd ../mahout/mahoutrec\n./unresystpredict.sh lastfm\n\n# import predictions\nfrom lastfm.mahout_recommender import *\nMahoutArtistRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_predictions.csv')\n\n# evaluate the predictions \nfrom lastfm.mahout_recommender import *\nfrom lastfm.evaluation import *\nArtistRankEvaluator.evaluate_predictions(MahoutArtistRecommender)\n\n\n\n# below\n#\n\n# build, \n#save_all_to_predictions must be true\n./build.sh lastfm\n\n# evaluate and save the predictions\nfrom lastfm.evaluation import *\nfrom lastfm.recommender import *\nNovelArtistRecommenderEvaluator.evaluate_predictions(NovelArtistRecommender, save_predictions=True)\n\n# export predictions\nfrom lastfm.recommender import *\nNovelArtistRecommender.export_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv')\n\n# export test data\nfrom lastfm.evaluation import *\nNovelArtistRankEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_test.csv')\n\n# run mahout train, test -> flixster_predictions.csv\ncd ../mahout/mahoutrec\n./unresystpredict.sh lastfm\n\n# update unresyst predictions with the obtained\nfrom lastfm.recommender import *\nNovelArtistRecommender.update_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_predictions.csv')\n\n# run evaluation\n./lastfmeval.sh novel dontbuild\n\n\n\n# recommendations - netreba\n# \n\n# evaluate unresyst\nfrom lastfm.evaluation import *\nfrom lastfm.recommender import *\nArtistRecommenderEvaluator.evaluate_recommendations(ArtistRecommender, 10)\n\n# evaluate mahout recommender\n# nenovel:\n\n# export training data\nfrom lastfm.mahout_recommender import *\nMahoutArtistRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv')\n\n# run mahout train, test -> lastfm_recommendations.csv\ncd ../mahout/mahoutrec\n./unresystrecommend.sh\n\n# import predictions\nfrom lastfm.mahout_recommender import *\nMahoutArtistRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_recommendations.csv')\n\n# evaluate the recommendations \nfrom lastfm.mahout_recommender import *\nfrom lastfm.evaluation import *\nArtistRecommenderEvaluator.evaluate_recommendations(MahoutArtistRecommender, 10)\n\n# evaluate mahout recommender\n# novel:\n\n# export training data\nfrom lastfm.mahout_recommender import *\nNovelMahoutArtistRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv')\n\n# run mahout train, test -> lastfm_recommendations.csv\ncd ../mahout/mahoutrec\n./unresystrecommend.sh\n\n# import predictions\nfrom lastfm.mahout_recommender import *\nNovelMahoutArtistRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_recommendations.csv')\n\n# evaluate the recommendations \nfrom lastfm.mahout_recommender import *\nfrom lastfm.evaluation import *\nNovelArtistRecommenderEvaluator.evaluate_recommendations(NovelMahoutArtistRecommender, 10)\n\n\n# Flixster\n#\n#\n\n# select test data - obecne\nfrom flixster.evaluation import *\nMovieRecommenderEvaluator.select_evaluation_pairs()\n\n# evaluate mahout recommender - predictions:\n#\n\n# export training data\nfrom flixster.mahout_recommender import *\nMahoutMovieRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv')\n\n# export test data\nfrom flixster.evaluation import *\nMovieRecommenderEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_test.csv')\n\n# run mahout train, test -> lastfm_predictions.csv\ncd ../mahout/mahoutrec\n./unresystpredict.sh flixster\n\n# import predictions\nfrom flixster.mahout_recommender import *\nMahoutMovieRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_predictions.csv')\n\n# evaluate the predictions \nfrom flixster.mahout_recommender import *\nfrom flixster.evaluation import *\nMovieRecommenderEvaluator.evaluate_predictions(MahoutMovieRecommender)\n\n\n# evaluate mahout recommender - recommendations\n# \n\n# export training data\nfrom flixster.mahout_recommender import *\nMahoutMovieRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv')\n\n# run mahout train, test -> flixster_recommendations.csv\ncd ../mahout/mahoutrec\n./unresystrecommend.sh flixster\n\n# import predictions\nfrom flixster.mahout_recommender import *\nMahoutMovieRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_recommendations.csv')\n\n# evaluate the recommendations \nfrom flixster.mahout_recommender import *\nfrom flixster.evaluation import *\nMovieRecommenderEvaluator.evaluate_recommendations(MahoutMovieRecommender, 10)\n\n\n# hybrid\n#\n\n# build save_all_to_predictions must be true\n./build.sh flixster\n\n# export as usual\n\n# export training data\nfrom flixster.mahout_recommender import *\nMahoutMovieRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv')\n\n# export test data\nfrom flixster.evaluation import *\nMovieRecommenderEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_test.csv')\n\n# run mahout train, test -> flixster_predictions.csv\ncd ../mahout/mahoutrec\n./unresystpredict.sh flixster\n\n# evaluate what we have and save the predictions\nfrom flixster.evaluation import *\nfrom flixster.recommender import *\nMovieRecommenderEvaluator.evaluate_predictions(MovieRecommender, save_predictions=True)\n\n# update unresyst predictions with the obtained\nfrom flixster.recommender import *\nMovieRecommender.update_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_predictions.csv')\n\n# run evaluation\n./flixstereval.sh preds dontbuild\n\n\n# below\n#\n\n# build, \n#save_all_to_predictions must be true\n./build.sh flixster\n\n# evaluate and save the predictions\nfrom flixster.evaluation import *\nfrom flixster.recommender import *\nMovieRecommenderEvaluator.evaluate_predictions(MovieRecommender, save_predictions=True)\n\n# export predictions\nfrom flixster.recommender import *\nMovieRecommender.export_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv')\n\n# export test data\nfrom flixster.evaluation import *\nMovieRecommenderEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_test.csv')\n\n# run mahout train, test -> flixster_predictions.csv\ncd ../mahout/mahoutrec\n./unresystpredict.sh flixster\n\n# import predictions\nfrom flixster.mahout_recommender import *\nMahoutMovieRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_predictions.csv')\n\n# evaluate the predictions \nfrom flixster.mahout_recommender import *\nfrom flixster.evaluation import *\nMovieRecommenderEvaluator.evaluate_predictions(MahoutMovieRecommender)\n\n\n# Travel\n#\n#\n\n# select test data - obecne\nfrom travel.models import *\nTourOrderEvalPair.select()\n\n# evaluate mahout recommender - recommendations\n# \n\n# export training data\nfrom travel.mahout_recommender import *\nMahoutOrderTourRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_train.csv')\n\n# run mahout train, test -> travel_recommendations.csv\ncd ../mahout/mahoutrec\n./unresystrecommend.sh travel \n\n# import predictions\nfrom travel.mahout_recommender import *\nMahoutOrderTourRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_recommendations.csv')\n\n# evaluate the recommendations \nfrom travel.mahout_recommender import *\nfrom travel.evaluation import *\nOrderTourRecommenderEvaluator.evaluate_recommendations(MahoutOrderTourRecommender, 10)\n\n# evaluate mahout recommender - predictions\n# \n\n# export training data\nfrom travel.mahout_recommender import *\nMahoutOrderTourRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_train.csv')\n\n# export test data\nfrom travel.evaluation import *\nOrderTourRecommenderEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_test.csv')\n\n# run mahout train, test -> travel_recommendations.csv\ncd ../mahout/mahoutrec\n./unresystpredict.sh travel \n\n# import predictions\nfrom travel.mahout_recommender import *\nMahoutOrderTourRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_predictions.csv')\n\n# evaluate the predictions\nfrom travel.evaluation import *\nfrom travel.mahout_recommender import *\nOrderTourRankEvaluator.evaluate_predictions(MahoutOrderTourRecommender)\n\n\n# below\n#\n\n# build, \n#save_all_to_predictions must be true\n./build.sh travel\n\n# evaluate and save the predictions\nfrom travel.evaluation import *\nfrom travel.recommender import *\nOrderTourRecommenderEvaluator.evaluate_predictions(OrderTourRecommender, save_predictions=True)\n\n# export predictions\nfrom travel.recommender import *\nOrderTourRecommender.export_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_train.csv')\n\n# export test data\nfrom travel.evaluation import *\nOrderTourRecommenderEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_test.csv')\n\n# run mahout train, test -> flixster_predictions.csv\ncd ../mahout/mahoutrec\n./unresystpredict.sh travel\n\n# update unresyst predictions with the obtained\nfrom travel.recommender import *\nOrderTourRecommender.update_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_predictions.csv')\n\n# run evaluation\n./traveleval.sh preds dontbuild\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n#############################\n\n# import predictions\nfrom travel.mahout_recommender import *\nMahoutOrderTourRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_predictions.csv')\n\n# evaluate the predictions \nfrom travel.mahout_recommender import *\nfrom travel.evaluation import *\nOrderTourRankEvaluator.evaluate_predictions(MahoutOrderTourRecommender)\n\n"},{"alpha_fraction":0.5666928291320801,"alphanum_fraction":0.5733941793441772,"avg_line_length":42.68224334716797,"blob_id":"289c86e2785511d7f624b95de018ebc8dd829c1d","content_id":"450b9f3c477029466ae38cc5422f510325ccfca6","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":14027,"license_type":"no_license","max_line_length":118,"num_lines":321,"path":"/code/adapter/unresyst/algorithm/algorithm.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The main classes of the algorithm package\"\"\"\n\nfrom django.db.models import Q\nfrom django.db.models import Avg\n\nfrom base import BaseAlgorithm\nfrom unresyst.constants import *\nfrom unresyst.models.abstractor import PredictedRelationshipDefinition, \\\n RelationshipInstance\nfrom unresyst.models.aggregator import AggregatedRelationshipInstance\nfrom unresyst.models.algorithm import RelationshipPredictionInstance\nfrom unresyst.models.common import SubjectObject\nfrom unresyst.recommender.rules import BaseRelationship\n\nclass Algorithm(BaseAlgorithm):\n \"\"\"The original deprecated implementation mixing combinator, compilator and algorithm. \n \n A simple implementation of a recommender algorithm.\n \n The remove_predicted_from_recommendations is implemented by adding a zero\n prediction for all pairs in predicted relationship. \n \n The get_recommendation function returns only recommendations with expectancy\n above zero or the expectancy_limit.\n \"\"\"\n \n N_NEIGHBOURHOOD = 10\n \"\"\"The maximum size of the neighbourhood, from which the similar items are taken\"\"\"\n\n # Build phase:\n #\n \n def build(self, recommender_model):\n \"\"\"Build the recommender - create the instances of the \n RelationshipPredictionInstance model where there is some simple \n prediction available. Where there isn't, leave it.\n \"\"\" \n\n\n\n # Recommend phase:\n #\n\n\n def get_relationship_prediction(self, recommender_model, dn_subject, dn_object, remove_predicted):\n \"\"\"See the base class for the documentation.\n \"\"\"\n # if predicted should be removed and the pair is in the predicted_rel, \n # return the special expectancy value\n if remove_predicted: \n # all predicted relationships\n qs_predicted = RelationshipInstance.filter_predicted(recommender_model)\n\n # the relationship between dn_subject and dn_object\n qs_predicted_rel = RelationshipInstance.filter_relationships(dn_subject, dn_object, queryset=qs_predicted)\n\n # if the prediction for the pair exists\n if qs_predicted_rel:\n \n # return the special expectancy value \n assert len(qs_predicted_rel) == 1\n predicted = qs_predicted_rel[0]\n\n return self._get_already_in_relatinship_prediction(\n recommender_model=recommender_model,\n predicted_relationship=predicted)\n \n \n # filter the predictions for recommender\n qs_rec_pred = RelationshipPredictionInstance.objects.filter(\n recommender=recommender_model)\n\n # filter the predictions for the given pair \n qs_pred = RelationshipPredictionInstance.filter_relationships(\n object1=dn_subject,\n object2=dn_object,\n queryset=qs_rec_pred) \n \n # if available return it\n if qs_pred: \n assert len(qs_pred) == 1\n return qs_pred[0]\n \n # if it's not available, maybe it wasn't in the N_NEIGHBOURHOOD, \n # so try finding it in aggregates \n so1, so2 = BaseRelationship.order_arguments(dn_subject, dn_object)\n\n # the Subject-object aggregate (just in case it hasn't been built) \n qs_rels = AggregatedRelationshipInstance.objects.filter(\n subject_object1=so1,\n subject_object2=so2,\n recommender=recommender_model)\n \n # if found return it\n if qs_rels:\n assert len(qs_rels) == 1\n\n pred = RelationshipPredictionInstance(\n subject_object1=dn_subject,\n subject_object2=dn_object,\n description=qs_rels[0].description,\n recommender=recommender_model,\n expectancy=qs_rels[0].expectancy\n )\n pred.save()\n return pred\n\n # the definition of the predicted relationship\n d = PredictedRelationshipDefinition.objects.get(recommender=recommender_model)\n \n # try finding the similar entity to the one liked by so1\n # content-based \n \n # get similarities starting the traverse with the similarity.\n qs_sim1 = AggregatedRelationshipInstance.objects\\\n .filter(recommender=recommender_model)\\\n .filter(\n Q(\n # take so2 as stable - in the subject_object2 position of the similarity\n subject_object2=so2,\n \n # traverse from the other object in similarity (subject_object1) through\n # the relationship instance, its subject (subject_object1) must be so1\n subject_object1__relationshipinstance_relationships2__subject_object1=so1,\n \n # the relationship instance definition must be the predicted relationship def\n subject_object1__relationshipinstance_relationships2__definition=d) | \\\n Q( \n # take so2 as stable again, now in the subject_object1 position of the similarity\n subject_object1=so2, \n \n # traverse from the other through relationship to so1\n subject_object2__relationshipinstance_relationships2__subject_object1=so1,\n \n # the definition again must be the predicted\n subject_object2__relationshipinstance_relationships2__definition=d))\n\n # if found return\n if qs_sim1:\n\n # if found return the average TODO nejak jinak\n avg = qs_sim1.aggregate(Avg('expectancy'))\n\n pred = RelationshipPredictionInstance(\n subject_object1=dn_subject,\n subject_object2=dn_object,\n description=qs_sim1[0].description,\n recommender=recommender_model,\n expectancy=avg['expectancy__avg']\n )\n pred.save()\n return pred \n\n \n # try finding the similar entity (user) to entity that liked so2\n # cf\n qs_sim2 = AggregatedRelationshipInstance.objects\\\n .filter(recommender=recommender_model)\\\n .filter(\n Q(\n # take so1 as stable - in the subject_object2 position of the similarity\n subject_object2=so1,\n \n # traverse from the other object in similarity (subject_object1) through\n # the relationship instance, its object (subject_object2) must be so2\n subject_object1__relationshipinstance_relationships1__subject_object2=so2,\n \n # the relationship instance definition must be the predicted relationship def\n subject_object1__relationshipinstance_relationships1__definition=d) | \\\n Q( \n # take so1 as stable again, now in the subject_object1 position of the similarity\n subject_object1=so1, \n \n # traverse from the other through relationship to so2\n subject_object2__relationshipinstance_relationships1__subject_object2=so2,\n \n # the definition again must be the predicted\n subject_object2__relationshipinstance_relationships1__definition=d))\n\n # if found return the average TODO nejak jinak\n if qs_sim2:\n avg = qs_sim2.aggregate(Avg('expectancy'))\n \n pred = RelationshipPredictionInstance(\n subject_object1=dn_subject,\n subject_object2=dn_object,\n description=qs_sim2[0].description,\n recommender=recommender_model,\n expectancy=avg['expectancy__avg']\n )\n pred.save()\n\n return pred \n \n \n # return the uncertain\n return self._get_uncertain_prediction(\n recommender_model=recommender_model, \n dn_subject=dn_subject, \n dn_object=dn_object\n )\n \n @classmethod\n def get_recommendations(self, recommender_model, dn_subject, count, expectancy_limit, remove_predicted):\n \"\"\"See the base class for documentation.\n \"\"\" \n \n # get the recommendations ordered by the expectancy from the largest\n recommendations = RelationshipPredictionInstance\\\n .get_relationships(obj=dn_subject)\\\n .filter(recommender=recommender_model, \n expectancy__gt=expectancy_limit)\\\n .order_by('-expectancy') \n \n # remove the predicted from the recommendations if it should be done \n if remove_predicted:\n \n # get objects that are already liked\n #\n \n qs_predicted = RelationshipInstance.filter_predicted(recommender_model) \n\n # get ids of subjectobjects where dn_subject appears\n qs_predicted = qs_predicted.filter( \n Q(subject_object1=dn_subject) | Q(subject_object2=dn_subject) \n ).values_list('subject_object1__pk', 'subject_object2__pk')\n \n # flatten it and take only the objects\n import itertools \n predicted_obj_ids = [i for i in itertools.chain(*qs_predicted) if i <> dn_subject.pk] \n\n # remove the already liked objects\n #\n \n # if it is a subjectobject we have to take both places\n if dn_subject.entity_type == ENTITY_TYPE_SUBJECTOBJECT:\n recommendations = recommendations.exclude(\n Q(subject_object1__id__in=predicted_obj_ids) \n | Q(subject_object2__id__in=predicted_obj_ids)\n )\n # otherwise one field is enough\n else:\n recommendations = recommendations.exclude(\n subject_object2__id__in=predicted_obj_ids\n )\n \n # if there should be more recommendations than we have and \n # the expectancy limit is below the uncertain, add uncertain \n # predictions\n if recommendations.count() < count \\\n and expectancy_limit < UNCERTAIN_PREDICTION_VALUE: \n\n object_ent_type = ENTITY_TYPE_SUBJECTOBJECT \\\n if recommender_model.are_subjects_objects \\\n else ENTITY_TYPE_OBJECT \n\n # all objects excluding the objects recommended to dn_subject\n # the last exclude is just for SO entity_type - we won't recommend \n # subject to himself\n uncertain_objects = SubjectObject.objects\\\n .filter(entity_type=object_ent_type, recommender=recommender_model)\\\n .exclude(relationshippredictioninstance_relationships1__subject_object2=dn_subject)\\\n .exclude(relationshippredictioninstance_relationships2__subject_object1=dn_subject)\\\n .exclude(pk=dn_subject.pk)\n \n # if predicted should be removed remove them\n if remove_predicted:\n uncertain_objects = uncertain_objects.exclude(id__in=predicted_obj_ids)\n \n # divide the recommendations into groups exp => uncertain, exp < uncertain\n positive_preds = recommendations.filter(\n expectancy__gte=UNCERTAIN_PREDICTION_VALUE)\n negative_preds = recommendations.filter(\n expectancy__lt=UNCERTAIN_PREDICTION_VALUE)\n \n # how many non-positive recommendations we want\n required_count = count - positive_preds.count()\n uncertain_count = uncertain_objects.count()\n \n # if there're enough uncertains to fill the recommendations, \n # clear the negative\n if uncertain_count >= required_count:\n uncertain_required_count = required_count\n negative_preds = []\n else:\n # otherwise we want all uncertains we have\n uncertain_required_count = uncertain_count\n\n negative_required_count = required_count - uncertain_count\n \n # and for the rest use negative - only those above limit\n negative_preds = negative_preds.filter(\n expectancy__gt=expectancy_limit)[:negative_required_count]\n\n uncertain_preds = []\n \n # construct the uncertain recommendations\n for obj in uncertain_objects[:uncertain_required_count]:\n pred = self._get_uncertain_prediction(recommender_model, dn_subject, obj)\n uncertain_preds.append(pred)\n\n return list(positive_preds) + uncertain_preds + list(negative_preds)\n\n # apply the limit for count\n return list(recommendations[:count]) \n \n\n\n\nclass PredictOnlyAlgorithm(Algorithm):\n \"\"\"An algorithm that doesn't need to be built, performs only predictions.\"\"\"\n \n @classmethod\n def build(self, recommender_model):\n \"\"\"Do nothing\"\"\"\n return\n\n @classmethod\n def get_recommendations(self, recommender_model, dn_subject, count, expectancy_limit, remove_predicted): \n \"\"\"Raise an error\"\"\"\n raise NotImplementedError() \n\n"},{"alpha_fraction":0.8039215803146362,"alphanum_fraction":0.8039215803146362,"avg_line_length":28.14285659790039,"blob_id":"a484a682b0fc5c982c0fc6ff7cf9830676019511","content_id":"64faa2b3c3f268bf2b3bbde6854976dfc739be87","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":204,"license_type":"no_license","max_line_length":65,"num_lines":7,"path":"/UnresystCD/code/adapter/unresyst/models/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The models for universal object and subject representation.\"\"\"\n\nfrom common import *\nfrom abstractor import *\nfrom aggregator import *\nfrom algorithm import *\nfrom evaluation import BaseEvaluationPair\n"},{"alpha_fraction":0.767624020576477,"alphanum_fraction":0.7780678868293762,"avg_line_length":41.66666793823242,"blob_id":"2e74aacfa2f641a63cab4215ad7f65d7f104fbb7","content_id":"6c8174bf9ac2980362b139acc357b55dc8bfb19b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":383,"license_type":"no_license","max_line_length":140,"num_lines":9,"path":"/README.md","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"**a recommender you can't resist ...**\n\nA prototype of a Universal Recommender System created as a part of my Master thesis.\n\nThe thesis was selected to the second round of the 2011 ACM Spy competition. See http://www.acm-spy.cz/o-soutezi-acm-spy/galerie-nejlepsich/\n\n[Mahout tutorial created in the process](docs/CreateMahoutRecommender.md)\n\n"},{"alpha_fraction":0.6669912338256836,"alphanum_fraction":0.6689386367797852,"avg_line_length":21.30434799194336,"blob_id":"0d4fcd4835f521b92feba9b34fdf57d2934fa3db","content_id":"dc47bba31c34243b602de93b48388b652365c4d7","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":1027,"license_type":"no_license","max_line_length":180,"num_lines":46,"path":"/code/adapter/flixstereval.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"#!/bin/sh\n\n# run the evaluation- recommendations. parameters \n# recs -> recommendations\n# preds -> predictions\n# dontbuild-> don't build\n\nBUILD=true\nRECS=false\nPREDS=false\n\nfor param in $*;\ndo \n case $param in\n 'recs')\n RECS=true\n ;;\n 'preds')\n PREDS=true\n ;;\n 'dontbuild')\n BUILD=false\n ;;\n esac\ndone\n\nif [ $BUILD = true ]\nthen\n # build\n echo \"Building...\"\n echo \"from flixster.recommender import *; MovieRecommender.build()\" | python ./manage.py shell\nfi\n\nif [ $RECS = true ]\nthen\n echo \"Evaluating recommendations...\"\n echo \"from flixster.evaluation import *; from flixster.recommender import *; MovieRecommenderEvaluator.evaluate_recommendations(MovieRecommender, 10)\"| python ./manage.py shell\nfi\n\nif [ $PREDS = true ]\nthen\n echo \"Evaluating predictions...\"\n echo \"from flixster.evaluation import *; from flixster.recommender import *; MovieRecommenderEvaluator.evaluate_predictions(MovieRecommender)\"| python ./manage.py shell\nfi\n\necho \"\"\n\n"},{"alpha_fraction":0.6990595459938049,"alphanum_fraction":0.7241379022598267,"avg_line_length":25.41666603088379,"blob_id":"d3ada56d0457dd1687f89e0a71e4c80664ce37f2","content_id":"41e744138d8492ced894d29c82cb75314744fc87","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":319,"license_type":"no_license","max_line_length":62,"num_lines":12,"path":"/UnresystCD/code/adapter/lastfm/constants.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The constants for last.fm\"\"\"\n\nMAX_LENGTH_NAME = 200\n\nCROSS_VALIDATION_COUNT = 5\n\"\"\"The number of test sets for the cross-validation\"\"\"\n\nSUCCESS_LIMIT = 0.5\n\"\"\"The limit above which the pair is considered successful.\"\"\"\n\nEXPECTED_EXPECTANCY_LISTENED = 1.0\n\"\"\"The expected expectancy for the listened to artists.\"\"\"\n\n\n"},{"alpha_fraction":0.5753098130226135,"alphanum_fraction":0.586749255657196,"avg_line_length":35.48695755004883,"blob_id":"91f7f98bac6c9825d846bc4908b450a2160faab3","content_id":"88fedd9c7d16e9e763b85e8db64f69241d3ae8f7","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":12588,"license_type":"no_license","max_line_length":102,"num_lines":345,"path":"/UnresystCD/code/adapter/unresyst/models/symmetric.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Support for symmetric relationships\"\"\"\n\nfrom django.db import models\nfrom django.db.models import Q\nfrom django.core.exceptions import MultipleObjectsReturned\n\nfrom unresyst.exceptions import SymmetryError, UnresystError as BaseError\n\nclass SymmetricalRelationship(models.Model):\n \"\"\"A base class for all symmetrical relationship models. The relationship can't be \n between the same object.\n \n Main custom methods:\n - rel.get_related(obj): object related to obj on rel\n - rel.contains_object(obj): does the rel contain obj?\n\n Only for additional_unique == ():\n - cls.get_related_objects(obj): objects related to obj\n - cls.get_relationships(obj): relationship objects related to obj\n - cls.get_relationship(obj1, obj2): get relationship between obj1 and obj2, or None\n - cls.are_related(obj1, obj2): are obj1 and obj2 related?\n\n Also for additional_unique <> ():\n - cls.filter_relationships(obj1, obj2, queryset=None): filter relationships between\n obj1 and obj2, from queryset or from all relationships\n\n \"\"\"\n \n attr_name1 = ''\n \"\"\"The name of the first attribute pointing on some model. Should be overriden\"\"\"\n \n attr_name2 = ''\n \"\"\"The name of the second attribute pointing on some model. Should be overriden\"\"\"\n \n additional_unique = ()\n \"\"\"A tuple containing names of attributes that are unique together with the symmetrical\n relationship pair. Like unique_together = (('attr1', 'attr2'), 'additional_unique')\n \"\"\"\n\n class Meta:\n abstract = True\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n object1 = getattr(self, self.attr_name1)\n object2 = getattr(self, self.attr_name2)\n \n return u\"%s - %s\" % (object1, object2)\n \n def save(self, *args, **kwargs):\n \"\"\"Overrides the save method in order to control the symmetricity.\n\n @raise SymmetryError: if there's something wrong with the saved entities in the\n means of symmetry \n \"\"\"\n \n # get the objects that should be saved as related\n object1 = getattr(self, self.attr_name1)\n object2 = getattr(self, self.attr_name2)\n \n # if they're identical, raise an error\n # the equality of entity models is already solved by overriding the __eq__ method\n if object1 == object2: \n raise SymmetryError(\n message=\"The related objects can't be identical\",\n object1=object1, \n object2=object2)\n \n # if the relationship exists in some direction\n # raise another error.\n # \n # try the saving direction\n self.__check_save(object1, object2)\n \n # try the other direction\n self.__check_save(object2, object1)\n \n # save the relationship \n return super(SymmetricalRelationship, self).save(*args, **kwargs) \n \n \n def __check_save(self, object1, object2):\n \"\"\"Check whether the the relationship between the objects in the given direction\n can be saved. If not, throw an exception. Includes checks for update.\"\"\"\n \n # create a dictionary of additional kwargs\n additional_kwargs = {}\n for ad_un in self.additional_unique:\n additional_kwargs[ad_un] = getattr(self, ad_un)\n \n # try finding the relation in the database\n rel = self.__class__.__get_relation(object1, object2, **additional_kwargs)\n\n # if we have found a relation between object1 and object2\n if rel:\n \n # if it's new value to insert (no id set)\n # or it's a value to update, and the found relationship isn't the same \n # as the one we're saving\n if not self.id or \\\n (self.id and self.id != rel.id):\n \n # than it's an error. \n raise SymmetryError(\n message=\"The relationship is already in the database.\",\n object1=object1, \n object2=object2) \n\n @classmethod\n def __get_related_objects_in_attribute(cls, obj, attr_name1, attr_name2):\n \"\"\"Get objects related to obj that are in the attr_name2 attribute...\n obj is in the attr_name1 attribute.\"\"\"\n \n # construct the kwargs\n kwargs = {attr_name1: obj}\n \n # get the relationships where obj is in the first attribute\n relationships = cls.objects.filter(**kwargs)\n \n # get the objects that are in the second attribute\n related_objects = [getattr(relationship, attr_name2) for relationship in relationships]\n \n return related_objects\n \n\n @classmethod \n def get_related_objects(cls, obj):\n \"\"\"Get objects that are in relationship with obj.\n \n Returns just objects (not the whole binding items).\n \n @type obj: core.models.BaseEntity\n @param obj: the object whose related objects should be got\n \n @rtype: list of core.models.BaseEntity\n @return: the list of objects related to obj\n \"\"\"\n # relationships where obj is as the first\n related_objects1 = cls.__get_related_objects_in_attribute(obj, cls.attr_name1, cls.attr_name2)\n\n # relationships where obj is as the second \n related_objects2 = cls.__get_related_objects_in_attribute(obj, cls.attr_name2, cls.attr_name1)\n \n return related_objects1 + related_objects2\n \n @classmethod \n def get_relationships(cls, obj):\n \"\"\"Get bindings for objects that are in relationship with the object obj.\n \n Returns whole binding items (not just objects).\n \n The related objects can be got using get_related. Example:\n\n connections = get_relationships(pepa)\n \n for connection in connections:\n print connection.connection_comment\n \n object_connected_to_pepa = connection.get_related(pepa)\n # ...\n \n @type obj: core.models.BaseEntity\n @param obj: the object whose relationships should be got\n \n @rtype: QuerySet of relationship class\n @return: relationships related to obj \n \"\"\"\n # get relationships where there is cls at first or second place\n # \n kwargs1 = {cls.attr_name1: obj}\n kwargs2 = {cls.attr_name2: obj}\n\n from django.db.models import Q\n q = cls.objects.filter(\n Q(**kwargs1) | Q(**kwargs2)\n )\n\n return q\n \n \n def get_related(self, obj):\n \"\"\"On a relationship, get the object that is related to obj, no matter in which\n attribute it is.\n \n To be used in the combination with the get_relationships method.\n\n @type obj: core.models.BaseEntity\n @param obj: the object whose related object should be got\n \n @rtype: core.models.BaseEntity\n @return: the object in the relationship no matter in which attribute it is.\n \n @raise SymmetryError: if the obj isn't a part of the relationship on which \n the method is called\n \"\"\"\n # get the objects in the relationship\n object1 = getattr(self, self.attr_name1)\n object2 = getattr(self, self.attr_name2)\n \n # if obj ist the first one, return the second\n # the equality of entity models is already solved by overriding the __eq__ method\n if obj == object1: \n return object2\n \n # if it's the second, return the first\n if obj == object2:\n return object1\n \n # otherwise return an error\n raise SymmetryError(\n message=\"The relationship object doesn't contain the '%s' object\" % obj,\n object1=object1, \n object2=object2) \n \n def contains_object(self, obj):\n \"\"\"Does the relationship contain the obj object? On one of the sides.\n \n @type obj: object\n @param obj: the object to check whether it's in the relationship\n \n @rtype: bool\n @return: Whether the object is contained on one side of the relationship\n \"\"\"\n\n # get the objects in the relationship\n object1 = getattr(self, self.attr_name1)\n object2 = getattr(self, self.attr_name2) \n \n return obj==object1 or obj==object2\n\n @classmethod\n def __contains_relation(cls, object1, object2):\n \"\"\"Does model in database contain the relationship of object1 and object2?\"\"\"\n \n return bool(cls.__get_relation(object1, object2))\n \n @classmethod\n def __get_relation(cls, object1, object2, **kwargs): \n \"\"\"Get the relation of the two objects in the given direction. If it doesn't exist\n return None.\n kwargs are additional arguments for filter, when additional_unique <> ()\n \"\"\"\n \n # prepare kwargs for the filter function\n filter_kwargs = {\n cls.attr_name1: object1,\n cls.attr_name2: object2\n }\n filter_kwargs.update(kwargs)\n \n # try to get the relationship\n query = cls.objects.filter(**filter_kwargs)\n\n # if some exists\n if query:\n if query.count() > 1:\n raise MultipleObjectsReturned(\n 'Too many obtained relationships for the pair %s, %s.' % (object1, object2))\n \n # return it \n return query[0]\n \n # otherwise there's nothing\n return None \n \n \n @classmethod\n def get_relationship(cls, object1, object2):\n \"\"\"Get the relationship between the two objects if they're related.\n If not return None.\n \n Takes into account both directions of the symmetry.\n \n @type object1: Model\n @param object1: the first object \n \n @type object2: Model\n @param object2: the second object \n \n @rtype: relationship object\n @return: the relationship between the two objects or None if they aren't related\n \"\"\" \n # try the first direction\n relationship = cls.__get_relation(object1, object2)\n if relationship:\n return relationship \n \n # try the second\n return cls.__get_relation(object2, object1) \n\n @classmethod\n def filter_relationships(cls, object1, object2, queryset=None):\n \"\"\"Filter relationships so that it contains only the ones between \n object1 and object2. \n For models where sets (object1, object2) are unique does the same \n as get_relationships (except returning a queryset)\n\n @type object1: Model\n @param object1: the first object \n \n @type object2: Model\n @param object2: the second object \n \n @type queryset: django queryset\n @param queryset: the queryset to filter (optional). If not passed, all\n objects of the model are filtered, otherwise the passed queryset is\n filtered.\n \n @rtype: django queryset\n @return: queryset containing relationships between object1 and object2\n \"\"\"\n if queryset is None:\n queryset = cls.objects.all()\n # prepare the filter kwargs:\n kwargs1 = {\n cls.attr_name1: object1,\n cls.attr_name2: object2\n } \n kwargs2 = {\n cls.attr_name1: object2,\n cls.attr_name2: object1\n }\n\n # filter relationships that have the pair in the given or the opposite\n # direction \n return queryset.filter(Q(**kwargs1) | Q(**kwargs2)) \n\n @classmethod\n def are_related(cls, object1, object2):\n \"\"\"Is the object related to other object?\n \n Takes into account both directions of the symmetry.\n \n @type object1: Model\n @param object1: the first object \n \n @type object2: Model\n @param object2: the second object \n \n @rtype: bool\n @return: True if they're in relationship, False otherwise. \n \"\"\" \n \n return cls.__contains_relation(object1, object2) \\\n or cls.__contains_relation(object2, object1)\n"},{"alpha_fraction":0.5918495059013367,"alphanum_fraction":0.5918495059013367,"avg_line_length":29.673076629638672,"blob_id":"c3de9e6129bcafd863cd330ac281345fb563cc8c","content_id":"7749ad6678f29b47d38275ab1ff6d16d157c5669","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1595,"license_type":"no_license","max_line_length":83,"num_lines":52,"path":"/UnresystCD/code/adapter/unresyst/recommender/predictions.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The prediction classes. Instances of the class are returned by \nthe recommender.\n\"\"\"\n\nclass RelationshipPrediction(object):\n \"\"\"The prediction of the predicted_relationship appearing between\n the given subject-object pair.\n \n @type subject: the domain-specific subject\n @ivar subject: the subject \n \n @type object_: the domain-specific object\n @ivar object_: the object\n \n @type expectancy: float\n @ivar expectancy: the estimated probability of the predict_relationship\n occuring between the subject and the object\n \n @type explanation: str\n @ivar explanation: the explanation for the prediction \n \"\"\"\n \n def __init__(self, subject, object_, expectancy, is_uncertain, explanation=''):\n \"\"\"The initializer\"\"\"\n \n self.subject = subject\n \"\"\"The subject\"\"\"\n \n self.object_ = object_\n \"\"\"The object\"\"\"\n \n self.expectancy = expectancy\n \"\"\"The estimated probability of the predicted_relationship\n occuring between the subject and the object.\n \"\"\"\n \n self.is_uncertain = is_uncertain\n \"\"\"Is the prediction made without having any information available?\"\"\"\n \n self.explanation = explanation\n \"\"\"The explanation for the prediction\"\"\"\n\n def __unicode__(self):\n return u\"%s <- %s: %f, %s\" % (\n self.subject, \n self.object_, \n self.expectancy, \n self.explanation\n )\n \n def __repr__(self):\n return \"< %s >\" % str(self.__unicode__())\n"},{"alpha_fraction":0.5630881786346436,"alphanum_fraction":0.5655280351638794,"avg_line_length":37.911563873291016,"blob_id":"0116196a027d45f1e52daa1218f4d26d0ea5222f","content_id":"3e2969c510c10aabccea92e3412b06b5eac85d9a","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":5738,"license_type":"no_license","max_line_length":102,"num_lines":147,"path":"/UnresystCD/code/adapter/unresyst/recommender/clusters.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The classes for defining clusters by users\"\"\"\n\nfrom unresyst.models.abstractor import ClusterSet, Cluster, ClusterMember\nfrom unresyst.models.common import SubjectObject\nfrom unresyst.exceptions import ConfigurationError\nfrom unresyst.constants import *\n\nclass BaseClusterSet(object):\n \"\"\"The base class for all clusters sets. Cluster set is a set of clusters,\n e.g. cluster set 'Gender' contains clusters 'Male', 'Female'.\"\"\"\n \n entity_type = None\n \"\"\"The type of the entities in the clusters\"\"\" \n \n entity_format_str = None\n \"\"\"Formating string used in the description\"\"\" \n \n def __init__(self, name, weight, filter_entities, get_cluster_confidence_pairs, description=None):\n \"\"\"The initializer\"\"\"\n\n self.name = name\n \"\"\"The name of the rule/relationship.\"\"\"\n \n self.weight = weight\n \"\"\"The significance of the similarity inferred from two subject/objects\n belonging to the same cluster. A number from [0, 1]\"\"\"\n \n self.filter_entities = filter_entities\n \"\"\"A queryset of all entities of the given type (s/o/so) belonging\n to some cluster.\n \"\"\"\n \n self.get_cluster_confidence_pairs = get_cluster_confidence_pairs\n \"\"\"A generator returning pairs of cluster, confidence for the given entity.\n E.g. for 'Liars' it returns (('Punk', 0.41), ('Indie', 0.72))\n Parameters: domain specific entity\n Return: the pairs: (name of the cluster, confidence - number from [0, 1])\n \"\"\"\n \n self.description = description\n \"\"\"The description of the membership. Can contain: placeholders for \n subject/object/subjectobject and cluster.\"\"\"\n\n \n def evaluate(self):\n \"\"\"Crate the cluster set in the database, its clusters, bindings \n of subjectobjects to the clusters.\n \"\"\"\n \n if not (MIN_WEIGHT <= self.weight <= MAX_WEIGHT):\n raise ConfigurationError(\n message=(\"The set '%s' provides weight %f,\" + \n \" should be between 0 and 1. .\"\n ) % (self.name, self.weight),\n recommender=self.recommender,\n parameter_name=\"Recommender.cluster_sets\",\n parameter_value=(self.recommender.cluster_sets)\n )\n \n recommender_model = self.recommender._get_recommender_model()\n\n # create the cluster set in the database\n cluster_set = ClusterSet(\n name=self.name,\n recommender=recommender_model,\n entity_type=self.entity_type,\n weight=self.weight)\n \n cluster_set.save()\n \n # go through the entities create clusters on demand\n #\n \n for ds_entity in self.filter_entities:\n \n # get entity cluster-confidence pairs\n cluster_conf_pairs = self.get_cluster_confidence_pairs(ds_entity)\n \n # convert the entity to universal\n dn_entity = SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=ds_entity, \n entity_type=self.entity_type, \n recommender=recommender_model)\n \n # go through the entity clusters\n for cluster_name, confidence in cluster_conf_pairs:\n \n # if confidence invalid through an error\n if not (MIN_CONFIDENCE <= confidence <= MAX_CONFIDENCE):\n raise ConfigurationError(\n message=(\"The cluster set '%s' provides confidence %f,\" + \n \" should be between 0 and 1. For cluster %s.\"\n ) % (self.name, confidence, cluster_name),\n recommender=self.recommender,\n parameter_name=\"Recommender.cluster_sets\",\n parameter_value=(self.recommender.cluster_sets)\n )\n \n # get or create the cluster \n cluster, x = Cluster.objects.get_or_create(\n name=cluster_name[:MAX_LENGTH_NAME],\n cluster_set=cluster_set)\n\n # evaluate the description\n if self.description:\n description = self.description % {\n self.entity_format_str: dn_entity.name,\n FORMAT_STR_CLUSTER: cluster_name} \n else:\n description = '' \n \n # save the binding of the cluster to the dn_entity\n member = ClusterMember.objects.create(\n cluster=cluster,\n member=dn_entity,\n confidence=confidence,\n description=description)\n \n print \" %d clusters and %d cluster members for '%s' cluster set created.\" \\\n % (Cluster.objects.filter(cluster_set=cluster_set).count(), \n ClusterMember.objects.filter(cluster__cluster_set=cluster_set).count(),\n self.name)\n \n \n \n\nclass SubjectClusterSet(BaseClusterSet):\n \"\"\"Cluster set for subjects\"\"\"\n \n entity_type = ENTITY_TYPE_SUBJECT\n \n entity_format_str = FORMAT_STR_SUBJECT\n\n \nclass ObjectClusterSet(BaseClusterSet):\n \"\"\"Cluster set for objects\"\"\"\n \n entity_type = ENTITY_TYPE_OBJECT\n\n entity_format_str = FORMAT_STR_OBJECT\n \nclass SubjectObjectClusterSet(BaseClusterSet):\n \"\"\"Clusters for recommenders where subject domain == object domain\"\"\"\n \n entity_type = ENTITY_TYPE_SUBJECTOBJECT\n \n entity_format_str = FORMAT_STR_SUBJECTOBJECT\n \n \n"},{"alpha_fraction":0.6540930867195129,"alphanum_fraction":0.6540930867195129,"avg_line_length":35.64706039428711,"blob_id":"4d7905dc524c2474531e811f3a1aae03f13e6d2f","content_id":"f8b6a8b980ba159a6a68de40716df35f3b869fa6","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1246,"license_type":"no_license","max_line_length":91,"num_lines":34,"path":"/code/adapter/unresyst/algorithm/aggregating_algorithm.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The aggregating algorithm class\"\"\"\n\nfrom base import BaseAlgorithm\n\nclass AggregatingAlgorithm(BaseAlgorithm):\n \"\"\"The algorithm that aggregates the similarity relationships and biases,\n uses the compiling algorithm as its inner_algorithm.\n \"\"\"\n \n def __init__(self, inner_algorithm, aggregator):\n \"\"\"The initializer\"\"\"\n \n super(AggregatingAlgorithm, self).__init__(inner_algorithm=inner_algorithm)\n \n self.aggregator=aggregator\n \"\"\"The aggregator that will be used for aggregating the relationships\n and biases\"\"\"\n \n def build(self, recommender_model):\n \"\"\"See the base class for documentation.\n \n Aggregates and calls the inner algorithm build\n \"\"\"\n \n # aggregate the relationships and rules\n self.aggregator.aggregate_rules_relationships(\n recommender_model=recommender_model) \n \n # aggregate the biases\n self.aggregator.aggregate_biases(recommender_model=recommender_model)\n \n print \"Rules, relationships and biases aggregated. Building the inner algorithm...\"\n \n super(AggregatingAlgorithm, self).build(recommender_model=recommender_model)\n"},{"alpha_fraction":0.800000011920929,"alphanum_fraction":0.800000011920929,"avg_line_length":34,"blob_id":"679060bf3bd6466a49a1124ea8f36296e390d08a","content_id":"833db83b59cc5685a7a8c8a23cf84f40fb5faad7","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":105,"license_type":"no_license","max_line_length":64,"num_lines":3,"path":"/code/adapter/unresyst/abstractor/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The package contains all classes for the abstractor layer.\"\"\"\n\nfrom abstractor import BasicAbstractor\n"},{"alpha_fraction":0.649068295955658,"alphanum_fraction":0.6537266969680786,"avg_line_length":25.58333396911621,"blob_id":"60fb7f1a8af2838251c1467c5ac671819f0b936d","content_id":"24ec29c615b0a824413690c8dc8206e22887bb0d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1288,"license_type":"no_license","max_line_length":66,"num_lines":48,"path":"/UnresystCD/code/adapter/demo/views.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The views for the demo application\"\"\"\n\nfrom django.http import HttpResponse\nfrom django.template import RequestContext, Context, loader\nfrom django.shortcuts import render_to_response, get_object_or_404\n\n\nfrom models import User\nfrom recommender import ShoeRecommender\n\ndef view_recommendations(request, user_name):\n \"\"\"The view for the demo recommendations\"\"\"\n \n # get the user and user list from db\n user = get_object_or_404(User, name__iexact=user_name)\n user_list = User.objects.all()\n \n # create user recommendations\n recommendations = ShoeRecommender.get_recommendations(user)\n \n context = {\n 'user': user,\n 'user_list': user_list,\n 'recommendations': recommendations,\n 'show_expectancy': True,\n }\n\n return render_to_response(\n 'demo/shoe_recommendations.html',\n dictionary=context, \n context_instance=RequestContext(request)\n )\n \ndef view_home_page(request):\n \"\"\"The view for the demo home page\"\"\"\n\n # get the user list from db\n user_list = User.objects.all()\n \n context = {\n 'user_list': user_list,\n }\n\n return render_to_response(\n 'demo/home_page.html',\n dictionary=context, \n context_instance=RequestContext(request)\n ) \n \n \n"},{"alpha_fraction":0.7196261882781982,"alphanum_fraction":0.7196261882781982,"avg_line_length":34.66666793823242,"blob_id":"289625693b56eb880e7853b71d3a9da2dfe44d3a","content_id":"88fdadf5a9aed29723056a56032d930de9d36b85","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":107,"license_type":"no_license","max_line_length":80,"num_lines":3,"path":"/code/adapter/unresyst/tests/test_update.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Testing the update\"\"\"\n\n# idea - porovnat vysledky recommenderu po updatu a po rebuildu - mely by byt stejne\n"},{"alpha_fraction":0.521481454372406,"alphanum_fraction":0.5237036943435669,"avg_line_length":41.44094467163086,"blob_id":"f4c944268e3c3cf02f08c1ccd2f6057c2662cc16","content_id":"034d4cc30e78aff70b78a057e7aa4392dfe5b7f3","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":5400,"license_type":"no_license","max_line_length":91,"num_lines":127,"path":"/code/adapter/unresyst/tests/test_base.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The base classes for the tests used in unresyst\"\"\"\n\nfrom django.test import TestCase\n\nfrom unresyst.models.common import SubjectObject \n\nfrom demo.recommender import ShoeRecommender, AverageRecommender\nfrom demo.models import User, ShoePair\n\nclass DBTestCase(TestCase):\n \"\"\"A base class for all tests which need database testing data\"\"\"\n\n def setUp(self):\n \"\"\"Insert data into the database\"\"\"\n\n # insert test data\n from demo.save_data import save_data\n save_data()\n \n def save_entities(self):\n \"\"\"Save instances of the entities to the testcase isntance\n to be called in subclasses.\n \n Not distinguishing the recommender.\"\"\"\n \n self.specific_entities = {\n 'Alice': User.objects.get(name=\"Alice\"),\n 'Bob': User.objects.get(name=\"Bob\"),\n 'Cindy': User.objects.get(name=\"Cindy\"),\n 'Daisy': User.objects.get(name=\"Daisy\"),\n 'Edgar': User.objects.get(name=\"Edgar\"),\n 'Fionna': User.objects.get(name=\"Fionna\"),\n 'Sneakers': ShoePair.objects.get(name=\"Sneakers\"),\n \"Rubber Shoes\": ShoePair.objects.get(name=\"Rubber Shoes\"),\n 'RS 130': ShoePair.objects.get(name='RS 130'),\n 'Design Shoes': ShoePair.objects.get(name='Design Shoes'),\n 'Octane SL': ShoePair.objects.get(name='Octane SL'), \n }\n \n rm = self.recommender._get_recommender_model()\n self.universal_entities = {\n 'Alice': SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=self.specific_entities['Alice'], \n entity_type='S', \n recommender=rm),\n 'Bob': SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=self.specific_entities['Bob'], \n entity_type='S', \n recommender=rm),\n 'Cindy': SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=self.specific_entities['Cindy'], \n entity_type='S', \n recommender=rm),\n 'Daisy': SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=self.specific_entities['Daisy'], \n entity_type='S', \n recommender=rm),\n 'Edgar': SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=self.specific_entities['Edgar'], \n entity_type='S', \n recommender=rm),\n 'Fionna': SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=self.specific_entities['Fionna'], \n entity_type='S', \n recommender=rm), \n 'Sneakers': SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=self.specific_entities['Sneakers'], \n entity_type='O', \n recommender=rm),\n 'Rubber Shoes': SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=self.specific_entities['Rubber Shoes'], \n entity_type='O', \n recommender=rm),\n 'RS 130': SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=self.specific_entities['RS 130'], \n entity_type='O', \n recommender=rm),\n 'Design Shoes': SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=self.specific_entities['Design Shoes'], \n entity_type='O', \n recommender=rm), \n 'Octane SL': SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=self.specific_entities['Octane SL'], \n entity_type='O', \n recommender=rm), \n } \n\nclass TestBuild(DBTestCase):\n \"\"\"The base class performing build in the setup.\"\"\"\n\n def setUp(self):\n \"\"\"The setup for all tests - build the recommender\"\"\"\n super(TestBuild, self).setUp()\n\n # call the tested function \n ShoeRecommender.build()\n\n self.recommender = ShoeRecommender\n\nclass TestBuildAverage(DBTestCase):\n \"\"\"The base class performing AverageRecommender build in the setup\"\"\"\n \n def setUp(self):\n \"\"\"The setup for all tests - build the recommender\"\"\"\n super(TestBuildAverage, self).setUp()\n\n # call the tested function \n AverageRecommender.build()\n \n self.recommender = AverageRecommender\n \n self.save_entities()\n \n\nclass TestEntities(TestBuild):\n \"\"\"The base class adding universal and specific entitits \n to the test instance\n \"\"\"\n \n def setUp(self):\n \"\"\"Obtain specific and universal subject objects \n and store them in the test instance\n \"\"\" \n \n super(TestEntities, self).setUp()\n \n self.save_entities()\n \n\n"},{"alpha_fraction":0.7019230723381042,"alphanum_fraction":0.7211538553237915,"avg_line_length":28.714284896850586,"blob_id":"74f9f8da729072e90e61e65cb1f7c03ef577fb22","content_id":"e9d1b16262fe2758fb6a519a8879452fd278a510","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":208,"license_type":"no_license","max_line_length":69,"num_lines":7,"path":"/code/adapter/demo/constants.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Constants for the demo application.\"\"\"\n\nMAX_LENGTH_NAME = 40\n\"\"\"The maximum length of the name in the universal representation.\"\"\"\n\nMAX_LENGTH_IMAGE_PATH = 50\n\"\"\"The maximum length of the path to image\"\"\"\n"},{"alpha_fraction":0.5149863958358765,"alphanum_fraction":0.5599455237388611,"avg_line_length":32.318180084228516,"blob_id":"9abe4d84c17f395a26390477ed959a3bad67f354","content_id":"1492be4f5e1bc974e9d0d001fa3fba7c297b9b3e","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":734,"license_type":"no_license","max_line_length":79,"num_lines":22,"path":"/UnresystCD/code/dataset_scripts/process_travel.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"grep diplomka_time_order travel.sql -m 1 -n\n\n# shorten, remove unused\ntail -n +38 diplomka\\(2\\).sql > travel1.sql \nhead -n 302073 travel1.sql > travel2.sql\n\n# away with the insert part \nsed -e 's/.*(//' travel2.sql > travel3.sql \n\n# away with the end \nsed -e 's/);.*//' travel3.sql > travel4.sql \n\n# remove ', ' from places between appostrophes\nsed -e \"s/', '\\(.[^']*\\),\\([^']*\\)', '/', '\\1\\2', '/\" travel4.sql > travel5.sql\nsed -e \"s/', '\\(.[^']*\\),\\([^']*\\)', '/', '\\1\\2', '/\" travel5.sql > travel6.sql\nsed -e \"s/', '\\(.[^']*\\),\\([^']*\\)', '/', '\\1\\2', '/\" travel6.sql > travel7.sql\n\n# away with spaces behind commas \nsed -e 's/, /,/g' travel7.sql > travel8.sql\n\n# away with appostrophes \nsed -e \"s/'//g\" travel8.sql > travel9.sql\n\n"},{"alpha_fraction":0.6000116467475891,"alphanum_fraction":0.6046186089515686,"avg_line_length":42.78717803955078,"blob_id":"4b60e53462fc33d0c54b4c6ff40808d6e82690b0","content_id":"06cf1f57e05adba80edb00c93f1600db52142aaa","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":17148,"license_type":"no_license","max_line_length":126,"num_lines":390,"path":"/UnresystCD/code/adapter/unresyst/combinator/base.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Base classes for the combinator layer:\n - BaseCombinator\n\"\"\"\nfrom unresyst.exceptions import CombinatorError\nfrom unresyst.models.abstractor import RelationshipInstance, RuleInstance, \\\n ExplicitRuleInstance, PredictedRelationshipDefinition, ClusterMember\nfrom unresyst.models.aggregator import AggregatedRelationshipInstance, AggregatedBiasInstance\nfrom unresyst.models.algorithm import RelationshipPredictionInstance\nfrom unresyst.constants import *\n\nclass BaseCombinator(object):\n \"\"\"The base class defining the interface of all combinators.\n interface methods:\n - combine_pair_similarities\n - combine_entity_biases\n - combine_pair_prediction_elements\n - choose_promising_objects\n \n methods to be overriden:\n - _combine \n \n helper methods for subclasses:\n - _concat_descriptions \n \"\"\"\n \n DIVISOR = 3\n \"\"\"A constant for dividing the relationship members that would otherwise be \n too numerous\"\"\"\n \n def __init__(self):\n \"\"\"The initializer\"\"\"\n \n self.top_bias_objects = None\n\n def _checked_combine(self, combination_elements, ResultClass):\n \"\"\"Check if something was given and call the overriden _combine method\n \"\"\"\n if not combination_elements:\n raise CombinatorError(\"No combination_elements given\")\n return self._combine(combination_elements, ResultClass)\n \n def _combine(self, combination_elements, ResultClass):\n \"\"\"Combine the combination elements to produce an instance \n of the ResultClass, filling in its expectancy and description.\n \n Overriden in subclasses\n\n @type combination_elements: a list of BaseCombinationElement\n @param combination_elements: the elements that should be combined\n\n @type ResultClass: class\n @param ResultClass: the class which instance will be returned\n \n @rtype: ResultClass\n @return: the combination class with filled expectancy and description\n \"\"\"\n pass\n\n @staticmethod\n def _concat_descriptions(element_list):\n \"\"\"Concat descriptions of the elements in the element_list.\n \n @type element_list: a list of BaseCombinationElement\n @param element_list: a list of elements in order that they should\n appear\n \n @rtype: str\n @return: the string concatenation of the element descriptions\n \"\"\"\n list_len = len(element_list)\n \n assert list_len > 0\n \n if list_len == 1:\n return element_list[0].get_description()\n \n # shorten the list in case it's too long\n short_list = element_list[:MAX_REASONS_DESCRIPTION]\n \n # join the descriptions\n joined_desc = ' '.join([\"%s: %s\" % ((REASON_STR % i), e.get_description()) \\\n for e, i in zip(short_list, range(1, list_len + 1))])\n \n # if the list was shortened add a message\n if list_len > MAX_REASONS_DESCRIPTION:\n joined_desc += ' ' + MORE_REASONS_STR % (list_len - MAX_REASONS_DESCRIPTION)\n \n return joined_desc\n\n \n\n def combine_pair_similarities(self, combination_elements):\n \"\"\"Aggregate similarities of the given pair S-S, O-O, or SO-SO. \n \n @type combination_elements: a list of BaseCombinationElement\n @param combination_elements: the elements that should be combined \n \n @rtype: AggregatedRelationshipInstance\n @return: the aggregated relationship with filled expectancy and\n description, other fields are empty.\n \"\"\" \n return self._checked_combine(\n combination_elements=combination_elements, \n ResultClass=AggregatedRelationshipInstance)\n \n\n def combine_entity_biases(self, entity_biases):\n \"\"\"Aggregate biases of the given entity S, O, or SO.\n \n @type entitiy_biases: QuerySet, each member having a get_expectancy()\n method\n @param entitiy_biases: the biases to aggregate\n \n @rtype: float\n @return: the expectancy that the entity will be in \n a predicted_relationship, aggregated from its biases\n \"\"\"\n return self._checked_combine(\n combination_elements=entity_biases,\n ResultClass=AggregatedBiasInstance)\n\n\n def combine_pair_prediction_elements(self, combination_elements):\n \"\"\"Combine all preference sources producing predictions\"\"\"\n return self._checked_combine(\n combination_elements=combination_elements,\n ResultClass=RelationshipPredictionInstance)\n\n \n def choose_promising_objects(self, dn_subject, min_count):\n \"\"\"Choose at least min_count objects that are likely to be interesting\n for the dn_subject, using various preference sources, choosing by \n a heuristic.\n \n Get maximum of min_count objects from each category\n - aggregated object bias \n - s-o relationships \n - predicted_relationship + object_similarities\n - predicted_relationship + subject similarities\n - predicted_relationship + subject cluster memberships \n - predicted_relationship + object cluster memberships \n \n @type dn_subject: SubjectObject\n @param dn_subject: the subject to choose the objects for\n \n @type min_count: int\n @param min_count: the minimum count that is chosen (if it's available)\n \n @rtype: iterable of SubjectObject\n @return: the promising objects\n \"\"\" \n \n if not min_count:\n return [] \n \n recommender_model = dn_subject.recommender\n \n object_ent_type = ENTITY_TYPE_SUBJECTOBJECT \\\n if recommender_model.are_subjects_objects else \\\n ENTITY_TYPE_OBJECT\n \n # aggregated object bias\n #\n \n # use some caching\n if self.top_bias_objects is None:\n \n qs_biases = AggregatedBiasInstance.objects\\\n .filter(recommender=recommender_model, \n subject_object__entity_type=object_ent_type,\n expectancy__gt=UNCERTAIN_PREDICTION_VALUE)\\\n .order_by('-expectancy')\n \n self.top_bias_objects = [(b.subject_object, b.expectancy) for b in qs_biases[:min_count]]\n \n top_bias_objs = self.top_bias_objects\n \n # s-o rules, relationships, explicit rules\n #\n \n top_rel_objs = self._get_promising_objects_rules_relationships(\n dn_subject=dn_subject,\n min_count=min_count,\n recommender_model=recommender_model)\n \n \n # predicted_relationship + object_similarities\n # predicted_relationship + subject similarities\n \n top_sim_objs = self._get_promising_objects_similarities(\n dn_subject=dn_subject,\n min_count=min_count,\n recommender_model=recommender_model)\n \n \n\n # predicted_relationship + subject cluster memberships \n # predicted_relationship + object cluster memberships\n \n cluster_objs = self._get_promising_objects_clusters(\n dn_subject=dn_subject,\n min_count=min_count,\n recommender_model=recommender_model)\n\n # concat all the lists and sort by expectancy\n ret_list = top_bias_objs + top_rel_objs + top_sim_objs + cluster_objs\n \n # if the liked should be removed, remove the liked objects from promising\n if recommender_model.remove_predicted_from_recommendations:\n\n # get ids of the liked objects for dn subject\n pred_obj_ids = RelationshipInstance\\\n .filter_predicted(recommender_model=recommender_model)\\\n .filter(subject_object1=dn_subject)\\\n .values_list('subject_object2__pk', flat=True)\n \n pred_obj_ids = set(pred_obj_ids)\n \n # remove the liked objects from ret_list\n ret_list = filter(lambda pair: not(pair[0].pk in pred_obj_ids), ret_list)\n \n ret_list.sort(key=lambda pair: pair[1], reverse=True) \n \n # remove the duplicates and take only some of the first\n return list(set([obj for obj, x in ret_list]))[:int(PROMISING_RATE*min_count)]\n\n def _get_promising_objects_clusters(self, dn_subject, min_count, recommender_model):\n \"\"\"Get promising objects from predicted_relationship + cluster membership\n\n @return: list of pairs (object, expectancy)\n \"\"\"\n\n # the definition of the predicted relationship\n d = PredictedRelationshipDefinition.objects.get(recommender=recommender_model)\n \n # taking similar subjects from clusters\n #\n sim_subj_objs = []\n\n # go through the subject's memberships\n for cm1 in dn_subject.clustermember_set.order_by('-confidence')[:min_count/(2*self.DIVISOR)]:\n\n # take the most similar subjects\n for cm2 in cm1.cluster.clustermember_set.exclude(id=cm1.id).order_by('-confidence')[:min_count/(2*self.DIVISOR)]:\n \n # take objects connected with them\n for rel in cm2.member.relationshipinstance_relationships1.filter(definition=d)[:min_count/self.DIVISOR]:\n \n # use a approximation for overall expectancy\n sim_subj_objs.append((rel.subject_object2, \n (cm1.confidence * cm2.confidence)/2 + UNCERTAIN_PREDICTION_VALUE))\n \n # taking similar objects from clusters\n #\n \n qs_memberships = ClusterMember.objects.filter(\n cluster__cluster_set__recommender=recommender_model,\n cluster__clustermember__member__relationshipinstance_relationships2__definition=d,\n cluster__clustermember__member__relationshipinstance_relationships2__subject_object1=dn_subject)\\\n .order_by('-confidence', '-cluster__clustermember__confidence')\n\n sim_obj_objs = [(cm.member, cm.confidence/2 + UNCERTAIN_PREDICTION_VALUE) \\\n for cm in qs_memberships[:min_count]]\n \n return sim_subj_objs + sim_obj_objs\n \n def _get_promising_objects_rules_relationships(self, dn_subject, min_count, recommender_model):\n \"\"\"Get promising objects from subject-object rules, relationships,\n explicit rules\n \n @return: list of pairs (object, expectancy) \n \"\"\"\n \n # get what is available in instances\n\n # the relationship type we're interested in\n rel_type = RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT \\\n if recommender_model.are_subjects_objects else \\\n RELATIONSHIP_TYPE_SUBJECT_OBJECT \n \n # relationship instances by weight, some rules also may be chosen\n qs_rel = RelationshipInstance.objects.filter(\n definition__recommender=recommender_model,\n definition__rulerelationshipdefinition__relationship_type=rel_type,\n subject_object1=dn_subject,\n definition__rulerelationshipdefinition__is_positive=True)\\\n .distinct()\\\n .order_by('-definition__rulerelationshipdefinition__weight')\n\n # rule instances by confidence\n qs_rule = RuleInstance.objects.filter(\n definition__recommender=recommender_model,\n definition__rulerelationshipdefinition__relationship_type=rel_type,\n subject_object1=dn_subject,\n definition__rulerelationshipdefinition__is_positive=True)\\\n .distinct()\\\n .order_by('-confidence') \n\n # explicit feedback \n qs_exp_rel = ExplicitRuleInstance.objects.filter( \n definition__recommender=recommender_model,\n subject_object1=dn_subject,\n expectancy__gt=UNCERTAIN_PREDICTION_VALUE)\\\n .distinct()\\\n .order_by('-expectancy')\n\n return [(rel.subject_object2, rel.get_expectancy()) for rel in qs_rel[:min_count]] + \\\n [(rel.subject_object2, rel.get_expectancy()) for rel in qs_rule[:min_count]] + \\\n [(rel.subject_object2, rel.expectancy) for rel in qs_exp_rel[:min_count]] \n\n \n def _get_promising_objects_similarities(self, dn_subject, min_count, recommender_model):\n \"\"\"Get promising objects from \n predicted_relationship + aggregated similarities\n \n @return: list of pairs (object, expectancy)\n \"\"\"\n \n # the definition of the predicted relationship\n d = PredictedRelationshipDefinition.objects.get(recommender=recommender_model)\n \n # try finding the similar entities to the ones liked by dn_subject\n # content-based \n \n # get similarities starting the traverse with the similarity.\n cont_qs_sim1 = AggregatedRelationshipInstance.objects\\\n .filter(recommender=recommender_model, expectancy__gt=UNCERTAIN_PREDICTION_VALUE)\\\n .filter(\n # traverse from the other object in similarity (subject_object1) through\n # the relationship instance, its subject (subject_object1) must be so1\n subject_object1__relationshipinstance_relationships2__subject_object1=dn_subject,\n \n # the relationship instance definition must be the predicted relationship def\n subject_object1__relationshipinstance_relationships2__definition=d)\\\n .distinct()\\\n .order_by('-expectancy')\n\n cont_objs_sim1 = [(sim.subject_object2, sim.expectancy) for sim in cont_qs_sim1[:min_count]]\n\n cont_qs_sim2 = AggregatedRelationshipInstance.objects\\\n .filter(recommender=recommender_model, expectancy__gt=UNCERTAIN_PREDICTION_VALUE)\\\n .filter(\n # traverse from the other through relationship to so1\n subject_object2__relationshipinstance_relationships2__subject_object1=dn_subject,\n \n # the definition again must be the predicted\n subject_object2__relationshipinstance_relationships2__definition=d)\\\n .distinct()\\\n .order_by('-expectancy')\n \n cont_objs_sim2 = [(sim.subject_object1, sim.expectancy) for sim in cont_qs_sim2[:min_count]]\n \n # try finding the similar entity (user) to entity that liked so2\n # cf\n \n # when subject is in the second position in similarity\n cf_qs_sim1 = AggregatedRelationshipInstance.objects\\\n .filter(recommender=recommender_model, expectancy__gt=UNCERTAIN_PREDICTION_VALUE)\\\n .filter(\n # take dn_subject as stable - in the subject_object2 position of the similarity\n subject_object2=dn_subject,\n \n # the relationship instance definition must be the predicted relationship def\n subject_object1__relationshipinstance_relationships1__definition=d)\\\n .distinct()\\\n .order_by('-expectancy')\n\n # get the object behind the predicted_relationship \n cf_objs_sim1 = [(pref.subject_object2, sim.expectancy) \\\n for sim in cf_qs_sim1[:min_count/self.DIVISOR] \\\n for pref in sim.subject_object1.relationshipinstance_relationships1.filter(definition=d)[:min_count/self.DIVISOR]]\n\n # when subject is in the first position in similarity \n cf_qs_sim2 = AggregatedRelationshipInstance.objects\\\n .filter(recommender=recommender_model, expectancy__gt=UNCERTAIN_PREDICTION_VALUE)\\\n .filter(\n # take dn_subject as stable again, now in the subject_object1 position of the similarity\n subject_object1=dn_subject, \n \n # the definition again must be the predicted\n subject_object2__relationshipinstance_relationships1__definition=d)\\\n .distinct()\\\n .order_by('-expectancy')\n \n # get the object behind the predicted_relationship \n cf_objs_sim2 = [(pref.subject_object2, sim.expectancy) \\\n for sim in cf_qs_sim2[:min_count/self.DIVISOR] \\\n for pref in sim.subject_object2.relationshipinstance_relationships1.filter(definition=d)[:min_count/self.DIVISOR]]\n \n return cont_objs_sim1 + cont_objs_sim2 + cf_objs_sim1 + cf_objs_sim2\n \n \n\n \n \n \n \n"},{"alpha_fraction":0.407888799905777,"alphanum_fraction":0.44402703642845154,"avg_line_length":35.519229888916016,"blob_id":"77be0dd2ab0e6d6b28632408f382e8cc098d9a7b","content_id":"9799347ab014baef3e09e67ea2786a89cb0ee061","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":13310,"license_type":"no_license","max_line_length":243,"num_lines":364,"path":"/code/adapter/unresyst/tests/test_combine_compile.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Tests for combinator, compilator.\"\"\"\nfrom nose.tools import eq_, assert_almost_equal\n\nfrom unresyst.models.common import SubjectObject\nfrom unresyst.combinator.base import BaseCombinator\nfrom unresyst.compilator.base import BaseCompilator\n\nfrom test_base import TestBuildAverage\n\nMIN_COUNT = 5\nPLACES = 4\n\nclass TestCombinator(TestBuildAverage):\n \"\"\"Tests for the base combinator\"\"\"\n\n EXPECTED_PROMISING_OBJECTS = {\n 'Alice': (('Sneakers', 'Design Shoes'), ('Rubber Shoes', 'Sneakers', 'Rubber Shoes', 'Design Shoes'), ('Rubber Shoes', 'RS 130', 'Sneakers'), ('Sneakers', 'Rubber Shoes')),\n 'Bob': (('Sneakers', 'Design Shoes'), ('Sneakers', 'Sneakers', 'Rubber Shoes'), ('Rubber Shoes', 'RS 130', 'Sneakers'), ('Sneakers', 'Rubber Shoes')),\n 'Cindy': ((), ('Rubber Shoes', 'RS 130'), (), ('Sneakers', 'Rubber Shoes')),\n 'Daisy': ((), ('RS 130',), (), ('Sneakers', 'Rubber Shoes')),\n 'Edgar': ((), (), ('Sneakers',), ('Sneakers', 'Rubber Shoes'), ('RS 130', 'Octane SL')),\n 'Fionna': ((), (), ('Rubber Shoes',), ('Sneakers', 'Rubber Shoes'), ('RS 130',)),\n }\n \"\"\"Dictionary subject: tuple( tuple from clusters, tuple from s-o rules, tuple from similarities, tuple from biases)\n \"\"\"\n\n def dtest_choose_promising_objects(self):\n \"\"\"Test choosing promising objects for subjects\"\"\"\n\n r = self.recommender._get_recommender_model()\n bc = BaseCombinator()\n\n for subj in SubjectObject.objects.filter(recommender=r, entity_type='S'):\n\n # choose it\n promobjs = bc.choose_promising_objects(dn_subject=subj, min_count=MIN_COUNT)\n \n # compare it\n eq_((subj, set(promobjs)), (subj, set([self.universal_entities[oname] for tup in self.EXPECTED_PROMISING_OBJECTS[subj.name] for oname in tup])))\n \nclass TestCompilator(TestBuildAverage):\n \"\"\"Tests for the base compilator\"\"\"\n EXPECTED_COMBINATION_ELEMENTS = {\n \n ('Alice', 'Sneakers'): (\n [(0.566667, 'User Alice likes many shoe pairs.'), (0.766667, 'Shoe pair Sneakers is popular')], \n [(0.550000, 'User Alice is from the same city as the manufacturer of Sneakers.')], \n [],\n [(0.612500, 'Similarity: Reason 1: Users Alice and Bob live in the same city. Reason 2: Users Alice and Bob are about the same age. And: User Bob likes shoes Sneakers.')],\n [],\n [],\n ),\n \n ('Alice', 'Rubber Shoes'): (\n [(0.566667, 'User Alice likes many shoe pairs.'), (0.633333, 'Shoe pair Rubber Shoes is popular')], \n [(0.700000, 'User Alice has viewed Rubber Shoes.'), (0.550000, 'User Alice is from the same city as the manufacturer of Rubber Shoes.')], \n [(0.625000, 'User Alice likes shoes Sneakers. And similarity: Reason 1: The shoe pairs Sneakers and Rubber Shoes share some keywords. Reason 2: Shoes Sneakers and Rubber Shoes were made by the same manufacturer.')],\n [],\n [],\n [],\n ), \n \n ('Alice', 'RS 130'): (\n [(0.566667, 'User Alice likes many shoe pairs.'), (0.633333, 'Shoe pair RS 130 is popular')], \n [(0.075000, \"Alice is from south, so RS 130 can't be recommended to him/her.\")],\n [],\n [],\n [],\n [],\n ),\n \n ('Alice', 'Design Shoes'): (\n [(0.566667, 'User Alice likes many shoe pairs.')], \n [],\n [(0.625000, 'User Alice likes shoes Sneakers. And similarity: Reason 1: Sneakers belong to the Casual category. Design Shoes belong to the Casual category. Reason 2: The shoe pairs Sneakers and Design Shoes share some keywords.')],\n [],\n [],\n [],\n ),\n \n ('Alice', 'Octane SL'): (\n [(0.566667, 'User Alice likes many shoe pairs.')], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Bob', 'Sneakers'): (\n [(0.566667, 'User Bob likes many shoe pairs.'), (0.766667, 'Shoe pair Sneakers is popular')], \n [(0.700000, 'User Bob has viewed Sneakers.'), (0.550000, 'User Bob is from the same city as the manufacturer of Sneakers.')], \n [],\n [(0.612500, 'Similarity: Reason 1: Users Alice and Bob live in the same city. Reason 2: Users Alice and Bob are about the same age. And: User Alice likes shoes Sneakers.')],\n [],\n [],\n ), \n \n ('Bob', 'Rubber Shoes'): (\n [(0.566667, 'User Bob likes many shoe pairs.'), (0.633333, 'Shoe pair Rubber Shoes is popular')], \n [(0.550000, 'User Bob is from the same city as the manufacturer of Rubber Shoes.')], \n [(0.625000, 'User Bob likes shoes Sneakers. And similarity: Reason 1: The shoe pairs Sneakers and Rubber Shoes share some keywords. Reason 2: Shoes Sneakers and Rubber Shoes were made by the same manufacturer.')], \n [],\n [], \n [],\n ),\n \n ('Bob', 'RS 130'): (\n [(0.566667, 'User Bob likes many shoe pairs.'), (0.633333, 'Shoe pair RS 130 is popular')], \n [(0.075000, \"Bob is from south, so RS 130 can't be recommended to him/her.\")], \n [],\n [],\n [],\n [],\n ),\n \n ('Bob', 'Design Shoes'): (\n [(0.566667, 'User Bob likes many shoe pairs.')], \n [],\n [(0.625000, 'User Bob likes shoes Sneakers. And similarity: Reason 1: Sneakers belong to the Casual category. Design Shoes belong to the Casual category. Reason 2: The shoe pairs Sneakers and Design Shoes share some keywords.')],\n [],\n [],\n [],\n ),\n \n ('Bob', 'Octane SL'): (\n [(0.566667, 'User Bob likes many shoe pairs.')], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Cindy', 'Sneakers'): (\n [(0.766667, 'Shoe pair Sneakers is popular')], \n [],\n [],\n [],\n [],\n [(0.550000, 'Similarity: Cindy has searched for the word Comfortable. Bob has searched for the word Comfortable. And: User Bob likes shoes Sneakers.')],\n ),\n\n \n ('Cindy', 'Rubber Shoes'): (\n [(0.633333, 'Shoe pair Rubber Shoes is popular')], \n [(0.700000, 'User Cindy has viewed Rubber Shoes.')],\n [],\n [],\n [],\n [],\n ),\n \n ('Cindy', 'RS 130'): (\n [(0.633333, 'Shoe pair RS 130 is popular')], \n [(0.550000, 'User Cindy is from the same city as the manufacturer of RS 130.')],\n [],\n [],\n [],\n [],\n ),\n \n ('Cindy', 'Design Shoes'): (\n [], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Cindy', 'Octane SL'): (\n [], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Daisy', 'Sneakers'): (\n [(0.766667, 'Shoe pair Sneakers is popular')], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Daisy', 'Rubber Shoes'): (\n [(0.633333, 'Shoe pair Rubber Shoes is popular')], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Daisy', 'RS 130'): (\n [(0.633333, 'Shoe pair RS 130 is popular')], \n [(0.550000, 'User Daisy is from the same city as the manufacturer of RS 130.')],\n [],\n [],\n [],\n [],\n ), \n \n ('Daisy', 'Design Shoes'): (\n [], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Daisy', 'Octane SL'): (\n [], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Edgar', 'Sneakers'): (\n [(0.633333, 'User Edgar likes many shoe pairs.'), (0.766667, 'Shoe pair Sneakers is popular')],\n [],\n [(0.625000, 'User Edgar likes shoes Rubber Shoes. And similarity: Reason 1: The shoe pairs Sneakers and Rubber Shoes share some keywords. Reason 2: Shoes Sneakers and Rubber Shoes were made by the same manufacturer.')],\n [],\n [],\n [],\n ),\n \n ('Edgar', 'Rubber Shoes'): (\n [(0.633333, 'User Edgar likes many shoe pairs.'), (0.633333, 'Shoe pair Rubber Shoes is popular')], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Edgar', 'RS 130'): (\n [(0.633333, 'User Edgar likes many shoe pairs.'), (0.633333, 'Shoe pair RS 130 is popular')], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Edgar', 'Design Shoes'): (\n [(0.633333, 'User Edgar likes many shoe pairs.')], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Edgar', 'Octane SL'): (\n [(0.633333, 'User Edgar likes many shoe pairs.')], \n [],\n [],\n [],\n [(0.650000, 'User Edgar likes shoes RS 130. And similarity: RS 130 belong to the For Sports category. Octane SL belong to the For Sports category.')],\n [],\n ),\n \n ('Fionna', 'Sneakers'): (\n [(0.766667, 'Shoe pair Sneakers is popular')], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Fionna', 'Rubber Shoes'): (\n [(0.633333, 'Shoe pair Rubber Shoes is popular')], \n [],\n [],\n [(0.600000, 'Similarity: Users Edgar and Fionna are about the same age. And: User Edgar likes shoes Rubber Shoes.')],\n [],\n [],\n ), \n \n ('Fionna', 'RS 130'): (\n [(0.633333, 'Shoe pair RS 130 is popular')], \n [],\n [],\n [(0.600000, 'Similarity: Users Edgar and Fionna are about the same age. And: User Edgar likes shoes RS 130.')],\n [],\n [],\n ),\n \n ('Fionna', 'Design Shoes'): (\n [], \n [],\n [],\n [],\n [],\n [],\n ),\n \n ('Fionna', 'Octane SL'): (\n [], \n [],\n [],\n [],\n [],\n [],\n ), \n }\n \"\"\"Dictionary s-o pair: tuple of elements \n (the pair: expectancy, description): \n - bias, \n - s-o relationships\n - predicted + object similarity\n - predicted + subject similarity\n - predicted + object cluster\n - predicted + subject cluster\n \"\"\"\n \n\n \n def dtest_get_pair_combination_elements(self):\n \"\"\"Test getting all we know for the given s-o pairs\"\"\"\n\n r = self.recommender._get_recommender_model()\n \n bc = BaseCompilator()\n\n # for all subject - object pairs\n for subj in SubjectObject.objects.filter(recommender=r, entity_type='S'):\n\n for obj in SubjectObject.objects.filter(recommender=r, entity_type='O'):\n \n dn_subject = self.universal_entities[subj.name]\n dn_object = self.universal_entities[obj.name]\n \n # call the tested function \n els = bc.get_pair_combination_elements(dn_subject, dn_object)\n \n expected_data = self.EXPECTED_COMBINATION_ELEMENTS[(subj.name, obj.name)]\n \n # flatten the expected data\n flat_data = [pair for listt in expected_data for pair in listt]\n\n # assert the length is right \n eq_(len(flat_data), len(els), \"Expected: %d, Obtained %d. For pair %s, %s Obtained data: %s\" % \\\n (len(flat_data), len(els), subj, obj, els))\n \n for el in els:\n \n # try finding the description in the expected data\n found = filter(lambda pair: pair[1]==el.get_description(), flat_data)\n \n eq_(len(found), 1, \"The description %s wasn't found for the pair %s, %s\" % (el.get_description(), subj, obj))\n found = found[0]\n \n assert_almost_equal(found[0], el.get_expectancy(), PLACES,\n \"The expectancy is wrong for pair %s, %s. Expected %f, Got %f\" % (subj, obj, found[0], el.get_expectancy()))\n \n"},{"alpha_fraction":0.5384615659713745,"alphanum_fraction":0.5641025900840759,"avg_line_length":18.5,"blob_id":"5e4066ae41edc0aa7c1a7f8ec796d070d77f0c22","content_id":"39d9a160c41fc8946eb2a6a18ea979a1b2062a6d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":39,"license_type":"no_license","max_line_length":27,"num_lines":2,"path":"/code/adapter/gr.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"#/bin/bash\ngrep -r --include=*.py $1 *\n"},{"alpha_fraction":0.6993424892425537,"alphanum_fraction":0.7065152525901794,"avg_line_length":29.418182373046875,"blob_id":"a060d34a1df70ccb30aba92abd8b1b6bdac1f4d0","content_id":"c2c0fa1dbb3cd32ee0d63d48f1879eeaaccdf503","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":1673,"license_type":"no_license","max_line_length":287,"num_lines":55,"path":"/code/mahout/mahoutrec/unresystpredict.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"#!/bin/sh\n\n# params:\n# lastfm\n# flixster\n# travel\n\n\nLASTFM=false\nFLIXSTER=false\nTRAVEL=false\n\nFLIXSTERBELOW=false\n\nfor param in $*;\ndo \n case $param in\n 'lastfm')\n LASTFM=true\n ;;\n 'flixster')\n FLIXSTER=true\n ;;\n 'travel')\n TRAVEL=true\n ;; \n 'flixsterbelow')\n FLIXSTERBELOW=true\n ;; \n esac\ndone\n\nif [ $LASTFM = true ]\nthen\n mvn -e exec:java -Dexec.mainClass=\"com.unresyst.UnresystPredict\" -Dexec.args=\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_test.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_predictions.csv\" \nfi\n\nif [ $FLIXSTER = true ]\nthen\n mvn -e exec:java -Dexec.mainClass=\"com.unresyst.UnresystPredict\" -Dexec.args=\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_test.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_predictions.csv\" \nfi\n\nif [ $FLIXSTERBELOW = true ]\nthen\n mvn -e exec:java -Dexec.mainClass=\"com.unresyst.UnresystPredict\" -Dexec.args=\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_test.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_predictions.csv\" \nfi\n\n\nif [ $TRAVEL = true ]\nthen\n mvn -e exec:java -Dexec.mainClass=\"com.unresyst.UnresystPredict\" -Dexec.args=\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_train.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_test.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_predictions.csv\" \nfi\n\n\necho \"\"\n"},{"alpha_fraction":0.7361376881599426,"alphanum_fraction":0.7361376881599426,"avg_line_length":29.705883026123047,"blob_id":"b1c76c010eee71b762a42787202fa779a27e40b0","content_id":"69152065ee0d7f36460fb4c1a9775c6e64d39541","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":523,"license_type":"no_license","max_line_length":63,"num_lines":17,"path":"/code/adapter/flixster/evaluation.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The evaluators for the last.fm recommenders\"\"\"\n\nfrom unresyst.recommender.evaluation import BaseEvaluator\nfrom models import MovieEvaluationPair\nfrom unresyst.recommender.metrics import rmse, precision_recall\n\nclass MovieRecommenderEvaluator(BaseEvaluator):\n \"\"\"The evaluator of the artist recommender\"\"\"\n \n EvaluationPairModel = MovieEvaluationPair\n \"\"\"The model - pairs\"\"\"\n \n prediction_metric = rmse\n \"\"\"The metric\"\"\"\n \n recommendation_metric = precision_recall\n \"\"\"The other metric\"\"\"\n\n"},{"alpha_fraction":0.7560126781463623,"alphanum_fraction":0.7569620013237,"avg_line_length":76.0243911743164,"blob_id":"e4e4234d3bd3cfeffc22e3d5b6a28a141d698c91","content_id":"032105fd7530c2e4ce9bb1f334b90dd1659d3b43","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Text","length_bytes":3160,"license_type":"no_license","max_line_length":345,"num_lines":41,"path":"/UnresystCD/readme.txt","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"Welcome to the Unresyst CD!\n---------------------------\n\nCurrent versions of all contained files can be found on http://code.google.com/p/unresyst/\n\nThe directory structure of the CD is the following:\n\n - docs: documentation of the unresyst prototype implementation.\n - diagrams: various diagrams used in the thesis, created in dia, a program for drawing structured diagrams http://live.gnome.org/Dia\n - evaluation: spreadsheets containing the results of unresyst evaluation, including charts\n - thesis: the thesis in the linked pdf format and latex source codes\n - mindmaps: mind maps used for presenting ideas about Unresyst. Created in XMind version: 3.2.0 http://www.xmind.net/\n - pics: pictures used in the thesis\n - presentations: presentations on the thesis from the User-Web seminar.\n - datamodels: data model visualization for each dataset application and for the whole unresyst.\n \n \n \n - code: contains all source code files belonging to the thesis.\n\n - adapter: adapter is a django project home directory (see http://www.djangoproject.com/), there are the following applications in the project. Each application has its own directory. The documentation for the project can be found in the docs/epydoc folder.\n\n - unresyst: the universal recommender application containing the implementation of the universal recommedner\n - lastfm: an application containing data models and a Unresyst recommender for the last.fm dataset.\n - flixster: application for the flixster dataset, the same structure as lastfm\n - travel: application for the travel agency dataset, the same structure as lastfm\n - demo: a demo application used in the tests\n - csv: a directory for temporary csv files used for the external mahout recommender\n - logs: a directory of text files where the evaluation of the recommendations is logged\n - evaluate.txt: a list of commands for running an external recommender for each application\n - example.py: an example of an Unresyst recommender, presented on the User-Web seminar\n - flixstereval.sh, lastfmeval.sh, traveleval.sh: scripts for running an evaluation on example datasets, see the script code for help\n - dump.sh, load.sh: scripts for opperating with the MySQL database\n - setup.sh: a script for loading data from dataset to the corresponing application models, the script just uses python scripts save_data.py, defined in each of the applications. \n - unresyst/tests/: tests for the implementation, the tests can be run using the command 'python ./manage.py test unresyst\"\n\n - dataset_scripts: shell scripts for reducing the datasets as described in the thesis.\n\n - mahout: contains an implementation of the Slope One recommender that was used for a comparison with unresyst. The recommender is implemented in Java using the Mahout recommender framework. See http://mahout.apache.org/ and http://code.google.com/p/unresyst/wiki/CreateMahoutRecommender for download and instalation of the Mahout framework.\n \n - unresystrecommend.sh, unresystpredict.sh: the scripts used for running the recommender on our data sets.\n\n\n"},{"alpha_fraction":0.5913978219032288,"alphanum_fraction":0.5913978219032288,"avg_line_length":45,"blob_id":"a7006a8fa9180fdf69189b409144c4558add2b25","content_id":"fbf1ff2571e73a191b467e04bafc59305f331b18","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":93,"license_type":"no_license","max_line_length":45,"num_lines":2,"path":"/code/adapter/rmpyc.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"find . -type f -name \\*.pyc -print | xargs rm\nfind . -type f -name \\*.py~ -print | xargs rm\n\n"},{"alpha_fraction":0.5462448000907898,"alphanum_fraction":0.5462448000907898,"avg_line_length":31.022472381591797,"blob_id":"b606e14182912067aa748b05ac5b1b12f75be5d5","content_id":"644ee4d31f9a0251faad680e4db4b9700bfa0bf1","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2876,"license_type":"no_license","max_line_length":81,"num_lines":89,"path":"/code/adapter/unresyst/abstractor/abstractor.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The main class of the abstractor package - Abstractor\"\"\"\n\nfrom base import BaseAbstractor\nfrom unresyst.models.common import SubjectObject, Recommender as RecommenderModel\nfrom unresyst.constants import *\n\nclass BasicAbstractor(BaseAbstractor):\n \"\"\"The basic implementation of the Abstractor class\"\"\"\n\n\n def create_subjectobjects(self, recommender_model, subjects, objects):\n \"\"\"See the base class for documentation.\"\"\"\n \n # if subjects are the same as objects, use the \"so\" entity type \n so = ENTITY_TYPE_SUBJECTOBJECT if subjects == objects else \"\"\n\n # create them again. subjects:\n for subj in subjects.iterator():\n \n # create the subject\n subob = SubjectObject(\n id_in_specific=subj.pk,\n name=subj.__unicode__()[:MAX_LENGTH_NAME],\n entity_type=so or ENTITY_TYPE_SUBJECT,\n recommender=recommender_model\n )\n \n # save it\n subob.save()\n \n print \" %d subjects created\" % subjects.count()\n \n # for recommenders where subjects==objects, that's it\n if so:\n return\n \n # create objects: \n for obj in objects.iterator():\n \n # create the object\n subob = SubjectObject(\n id_in_specific=obj.pk,\n name=obj.__unicode__()[:MAX_LENGTH_NAME],\n entity_type=ENTITY_TYPE_OBJECT,\n recommender=recommender_model\n )\n \n # save it\n subob.save()\n \n print \" %d objects created\" % objects.count()\n\n\n \n def create_predicted_relationship_instances(self, predicted_relationship):\n \"\"\"See the base class for documentation.\"\"\" \n \n # evaluate the relationship for all possible subjectobjects\n predicted_relationship.evaluate()\n \n \n def create_relationship_instances(self, relationships):\n \"\"\"See the base class for documentation.\"\"\"\n\n # evaluate all relationships\n for rel in relationships:\n rel.evaluate()\n\n \n def create_rule_instances(self, rules):\n \"\"\"See the base class for documentation.\"\"\"\n \n # eveluate all rules\n for rule in rules:\n rule.evaluate()\n\n def create_clusters(self, cluster_sets):\n \"\"\"See the base class for documentation.\"\"\"\n \n # evaluate all cluster sets\n for cluster_set in cluster_sets:\n cluster_set.evaluate() \n \n def create_biases(self, biases):\n \"\"\"See the base class for documentation.\"\"\" \n \n # evaluate the biases\n for bias in biases:\n bias.evaluate()\n \n \n"},{"alpha_fraction":0.5340961217880249,"alphanum_fraction":0.5447750091552734,"avg_line_length":34.819671630859375,"blob_id":"8491b9f2e8b52afe3779e5f92e6d4edc5a28c3d5","content_id":"cc7eb7e3479bbc007ca79cd3972ef4f5c0018a4e","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":6555,"license_type":"no_license","max_line_length":82,"num_lines":183,"path":"/code/adapter/recommender_example.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"An example of an Unresyst recommender for the last.fm domain.\nThe addaptation rules include:\n - subject-object preference \n - subject-subject similarity\n - object-object similarity\n - object bias\n\nRecommenders used during the evaluation tests can be found in:\n - lastfm/recommender.py\n - travel/recommender.py\n - flixster/recommender.py\n\"\"\"\n\nfrom unresyst import *\nfrom models import *\n\nAGE_DIFFERENCE = 38 - 17\n\"\"\"The age difference between the oldest and the yongest user\"\"\"\n\nMAX_PLAY_COUNT = 542\n\"\"\"The maximum play count for an artist in the period\"\"\"\n\nN_MIN_PLAY_COUNT = 100\n\"\"\"The minimum play count for an artist to apply the bias\"\"\"\n\nPERIOD_START_DATE = datetime.date(2010, 9, 1)\nPERIOD_END_DATE = datetime.date(2010, 12, 31)\n\n\nclass ArtistRecommender(Recommender):\n \"\"\"A recommender recommending artists (musicians) that \n the user can like.\n \"\"\" \n \n name = \" Artist Recommender\"\n \"\"\"The name\"\"\" \n \n subjects = User.objects\n \"\"\"The subjects to who the recommender will recommend.\"\"\"\n \n objects = Artist.objects\n \"\"\"The objects that will be recommended.\"\"\"\n \n predicted_relationship = PredictedRelationship( \n name=\"User listens to artist's tracks.\",\n \n # gives true for user, artist pairs where the user have listened\n # to the artist\n condition=lambda user, artist: \\\n user.scrobble_set.filter(track__artist=artist).exists(), \n \n description=\"\"\"User %(subject)s listens to the %(object)s's tracks.\"\"\"\n )\n \"\"\"The relationship that will be predicted\"\"\"\n \n # the class contains definitions for business rules\n rules = (\n \n # don't recommend artists with male-specific tags to females\n SubjectObjectRule(\n name=\"Don't recommend male music to female users.\",\n\n # the user is a female and the artist was tagged by\n # a male-specific tag\n condition=lambda user, artist: user.gender == 'f' and \\\n artist.artisttag_set.filter(tag__gender_specific='m').exists(),\n \n # it's a negative rule\n is_positive=False,\n \n weight=0.5,\n \n # the more male-specific tags the artist has, the higher is \n # the rule confidence. Normalized by the artist tag count\n confidence=lambda user, artist: float(\n artist.artisttag_set.filter(tag__gender_specific='m').count())/ \\\n artist.artisttag_set.count(),\n \n description=\"Artist %(object)s isn't recommended to %(subject)s, \" +\n \"because the artist is considered male-specific.\"\n ),\n \n \n # users of similar age are similar\n SubjectSimilarityRule(\n name=\"Users with similar age.\",\n \n # both users have given their age and the difference \n # is lower than five\n condition=lambda user1, user2: \n user1.age and user2.age and abs(user1.age - user2.age) <= 5,\n \n is_positive=True, \n \n weight=0.5,\n \n # a magic linear confidence function\n confidence=lambda user1, user2: \n 1 - float(abs(user1.age - user2.age))/AGE_DIFFERENCE,\n \n description=\"Users %(subject1)s and %(subject2)s are about \" + \n \"the same age.\"\n ), \n \n # artists sharing some tags are similar\n ObjectSimilarityRule(\n name=\"Artists sharing some tags.\",\n\n # both artists have some tags and they share at least one tag\n condition=lambda artist1, artist2: \\\n artist1.artisttag_set.exists() and \\\n artist2.artisttag_set.exists() and \\\n artist1.artisttag_set.filter(\n tag__id__in=artist2.artisttag_set.values_list('tag__id')\n ).exists(),\n \n # it's a positive rule\n is_positive=True,\n \n weight=0.5,\n \n # The more tags the artists have in common, the higher is \n # the similarity confidence\n confidence=lambda artist1, artist2: \\\n float(artist1.artisttag_set.filter(\n tag__id__in=artist2.artisttag_set.values_list('tag__id')\n ).count()) / \\\n min(artist1.artisttag_set.count(), artist2.artisttag_set.count()),\n \n description=\"Artists %(object1)s and %(object2)s are similar \" +\n \"because they share some tags.\"\n ),\n )\n \n relationships = (\n # if two users are from the same country, they are similar\n SubjectSimilarityRelationship(\n name=\"Users living in the same country\",\n \n # both users have given their country and it's the same\n condition=lambda user1, user2:\n user1.country and \\\n user2.country and \\\n user1.country == user2.country,\n \n # it's relationship positive to similarity\n is_positive=True, \n \n weight=0.5, \n \n description=\"Users %(subject1)s and %(subject2)s are from \" + \\\n \" the same country.\",\n ),\n )\n \n biases = (\n ObjectBias(\n name=\"Artists whose tracks have been listened a lot recently.\",\n \n description=\"%(object)s has been listened much recently.\",\n \n # take only artists with more than the minimal play count\n # in the given period\n condition=lambda artist: \\\n artist.track_set\\\n .filter(scrobble__timestamp__range=\n (PERIOD_START_DATE, PERIOD_END_DATE))\\\n .aggregate(Count('scrobble')) > N_MIN_PLAY_COUNT,\n \n weight=0.5,\n \n # it's a positive bias\n is_positive=True,\n \n # the number of scrobbles for the artist divided by the max.\n confidence=lambda artist: \\\n float(artist.track_set\\\n .filter(scrobble__timestamp__range=\n (PERIOD_START_DATE, PERIOD_END_DATE))\\\n .annotate(scrobble_count=Count('scrobble'))\\\n .aggregate(Sum('scrobble_count')))/MAX_PLAY_COUNT\n ),\n )\n"},{"alpha_fraction":0.5987638235092163,"alphanum_fraction":0.6003090143203735,"avg_line_length":32.84716033935547,"blob_id":"b8d5a394c3f4eb03cbd9ec92f382c65b8c472e4a","content_id":"a196c296da5fc0ad056f21531bcc8a73ef3c3a3d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":7766,"license_type":"no_license","max_line_length":107,"num_lines":229,"path":"/UnresystCD/code/adapter/unresyst/exceptions.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Errors thrown by the Unresyst application.\"\"\"\n\nclass UnresystError(Exception):\n \"\"\"The base exception for all exceptions.\n \n Defines the message property.\n \"\"\"\n def _get_message(self): \n return self._message\n def _set_message(self, message): \n self._message = message\n message = property(_get_message, _set_message)\n\nclass CombinatorError(UnresystError):\n \"\"\"An error in combinator\"\"\"\n pass\n \nclass RecommenderError(UnresystError):\n \"\"\"The base class for exceptions related to a recommender.\n\n @type message: string\n @ivar message: the reason for the error \n \n @type recommender: recommender.Recommender\n @ivar recommender: the recommender on which the error occured \n \"\"\" \n \n def __init__(self, message, recommender):\n \"\"\"The constructor.\"\"\" \n\n self.message = message\n \"\"\"The message saying the reason for the error\"\"\" \n \n self.recommender = recommender\n \"\"\"The recommender for which the error is raised\"\"\" \n \nclass RecommenderBuildError(RecommenderError):\n \"\"\"An error during the build of the recommender.\"\"\"\n pass\n\nclass RecommenderNotBuiltError(RecommenderError):\n \"\"\"Exception meaning that the recommender hasn't been built even though\n it should have been.\n \n @type message: string\n @ivar message: the reason for the error \n \n @type recommender: recommender.Recommender\n @ivar recommender: the recommender on which the error occured\n \"\"\" \n\n\n def __str__(self):\n return (\"The recommender needs to be built to perform the action.\\n\" + \\\n \" message: %s\\n\" + \\\n \" recommender: %s\\n\") \\\n % (self.message, self.recommender.name)\n \n\nclass ConfigurationError(RecommenderError):\n \"\"\"Exception meaning that something is wrong in the recommender configuration. \n\n @type message: string\n @ivar message: the reason why the parameter is invalid\n\n @type recommender: recommender.Recommender\n @ivar recommender: the recommender on which the error occured\n \n @type parameter_name: string\n @ivar parameter_name: the name of the invalid parameter\n \n @type parameter_value: object\n @ivar parameter_value: the current (invalid) parameter value \n \"\"\"\n \n def __init__(self, message, recommender, parameter_name, parameter_value):\n \"\"\"The constructor.\"\"\" \n\n super(ConfigurationError, self).__init__(message, recommender)\n \n self.parameter_name = parameter_name\n \"\"\"The name of the invalid parameter\"\"\"\n \n self.parameter_value = parameter_value\n \"\"\"The current (invalid) value of the parameter\"\"\" \n\n def __str__(self):\n return (\"A configuration parameter is invalid.\\n\" + \\\n \" message: %s\\n\" + \\\n \" recommender: %s\\n\" + \\\n \" parameter name: %s\\n\" + \\\n \" parameter value: %s\\n\") \\\n % (self.message, self.recommender.name, self.parameter_name, self.parameter_value)\n\n\nclass RuleRelationshipError(RecommenderError):\n \"\"\"An error in the configuration of rules/relationships.\n\n @type message: str\n @ivar message: additional message\n\n @type recommender: recommender.Recommender\n @ivar recommender: the recommender on which the error occured\n \n @type name: str\n @ivar name: the name of the rule/relationship where the error occured \n \"\"\"\n \n def __init__(self, message, recommender, name):\n \"\"\"The constructor.\"\"\" \n\n super(RuleRelationshipError, self).__init__(message, recommender)\n \n self.name = name\n \"\"\"The rule/relationship name\"\"\"\n \n\nclass DescriptionKeyError(RuleRelationshipError):\n \"\"\"Exception meaning that there are wrong strings in the rule/relationship\n description. \n\n @type message: str\n @ivar message: additional message\n \n @type recommender: recommender.Recommender\n @ivar recommender: the recommender on which the error occured\n \n @type name: str\n @ivar name: the name of the rule/relationship where the error occured\n \n @type key: str\n @ivar key: the key that shouldn't have been there\n \n @type permitted_keys: str list\n @ivar permitted_keys: the keys that can be in the description\n \"\"\"\n \n def __init__(self, message, recommender, name, key, permitted_keys):\n \"\"\"The constructor.\"\"\" \n\n super(DescriptionKeyError, self).__init__(message, recommender, name)\n \n self.key = key\n \"\"\"The string key that is wrong\"\"\"\n \n self.permitted_keys = permitted_keys\n \"\"\"The list of permitted keys\"\"\"\n\n\n def __str__(self):\n return (\"There's an invalid format key '%s' in the rule/relationship description.\\n\" + \\\n \" message: %s\\n\" + \\\n \" recommender: %s\\n\" + \\\n \" rule/relationship name: %s\\n\" + \\\n \" the invalid key: %s\\n\" + \\\n \" The permitted keys are: %s\") \\\n % (self.key, self.message, self.recommender.name, self.name, self.key, self.permitted_keys) \n\nclass InvalidParameterError(RecommenderError):\n \"\"\"An exception raised when a parameter passed to a function is invalid.\n\n @type message: string\n @ivar message: the reason why the parameter is invalid\n\n @type recommender: recommender.Recommender\n @ivar recommender: the recommender on which the error occured\n \n @type parameter_name: string\n @ivar parameter_name: the name of the invalid parameter\n \n @type parameter_value: object\n @ivar parameter_value: the current (invalid) parameter value\n \n \"\"\"\n def __init__(self, message, recommender, parameter_name, parameter_value):\n \"\"\"The constructor.\"\"\" \n\n super(InvalidParameterError, self).__init__(message, recommender)\n \n self.parameter_name = parameter_name\n \"\"\"The name of the invalid parameter\"\"\"\n \n self.parameter_value = parameter_value\n \"\"\"The current (invalid) value of the parameter\"\"\" \n \n def __str__(self):\n return (\"The parameter passed to the function is invalid.\\n\" + \\\n \" message: %s\\n\" + \\\n \" recommender: %s\\n\" + \\\n \" parameter name: %s\\n\" + \\\n \" parameter value: %s\") \\\n % (self.message, self.recommender.name, self.parameter_name, self.parameter_value)\n \n \nclass SymmetryError(UnresystError):\n \"\"\"An error for handling symmetric relationship errors.\n \n @type message: string\n @ivar message: the reason for the error\n \n @type object1: object\n @ivar object1: the first object in the relationship\n \n @type object2: object\n @ivar object2: the second object in the relationship\n \"\"\"\n \n def __init__(self, message, object1, object2):\n \"\"\"The constructor\"\"\"\n\n self.message = message\n \"\"\"the reason for the error\"\"\"\n \n self.object1 = object1\n \"\"\"the first object in the relationship\"\"\"\n \n self.object2 = object2 \n \"\"\"the second object in the relationship\"\"\"\n \n def __str__(self):\n return (\"There's an error in relationship symmetry.\\n\" + \\\n \" message: %s\\n\" + \\\n \" first object in relationship: %s\\n\" + \\\n \" second object in relationship: %s\") \\\n % (self.message, self.object1, self.object2) \n \nclass EmptyTestSetError(UnresystError):\n \"\"\"An error indicating that the test set is empty.\"\"\"\n pass \n"},{"alpha_fraction":0.8471337556838989,"alphanum_fraction":0.8471337556838989,"avg_line_length":38.25,"blob_id":"b12b94b6c2d011fd8bd11d69d21be903195767e9","content_id":"043a96d55420568d1741424d9735fb4331d54093","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":157,"license_type":"no_license","max_line_length":52,"num_lines":4,"path":"/code/adapter/unresyst/compilator/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The layer making predictions from aggregates\"\"\"\n\nfrom get_first_compilator import GetFirstCompilator\nfrom combining_compilator import CombiningCompilator\n"},{"alpha_fraction":0.5678645968437195,"alphanum_fraction":0.5681615471839905,"avg_line_length":32.83333206176758,"blob_id":"e1f2f7507d3a580f771c0a1f8cb7ad15ac713776","content_id":"e8abee01b6e2d8adcc5c76312f84296b7ed1ff34","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":6734,"license_type":"no_license","max_line_length":88,"num_lines":198,"path":"/code/adapter/unresyst/recommender/external_recommender.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"A module for recommender making predictions in an outside engine\"\"\"\nimport csv\n\nfrom base import BaseRecommender\nfrom predictions import RelationshipPrediction\nfrom unresyst.exceptions import RecommenderError, RecommenderNotBuiltError\nfrom unresyst.models.algorithm import ExternalPrediction\nfrom unresyst.models.common import Recommender as RecommenderModel\nfrom unresyst.constants import *\n\nclass ExternalRecommender(BaseRecommender):\n \"\"\"A class representing an outside-world recommender with an Unresyst \n interface.\n \"\"\"\n \n PredictionModel = ExternalPrediction\n \"\"\"The model where predictions are stored\"\"\" \n \n # build phase:\n #\n \n @classmethod\n def build(cls):\n \"\"\"Raise an error, as thist can't be done here, but outside.\n \n @raise RecommenderError: always\n \"\"\"\n raise RecommenderError(\"The method is invalid for the external\" + \\\n \" recommender. Call the export_data, build it outside and \" + \\\n \"import_predictions.\", cls)\n\n\n @classmethod\n def export_data(cls, filename):\n \"\"\"Export the data from subject and object managers to a csv \n file of the given name.\n \n @type filename: str\n @param filename: the full path to the file\n \n @raise FileNotExists and other file open errors.\n \"\"\"\n \n with open(filename, 'w') as f:\n \n rel = cls.explicit_rating_rule \\\n if cls.explicit_rating_rule else cls.predicted_relationship\n \n # export the relationships to the given file\n rel.export(f)\n \n @classmethod\n def import_predictions(cls, filename):\n \"\"\"Load predictions from the given csv file.\n \n Creates the recommender model for the recommender and imports the \n predictions from the given file. The file has to be in format:\n ,,\\n\n \n @type filename: str\n @param filename: the full path to the file\n\n @raise FileNotExists and other file open errors.\n \"\"\"\n \n cls._print('Deleting old predictions...')\n \n \n # if the recommender with the given name exists, delete it,\n RecommenderModel.objects.filter(class_name=cls.__name__).delete()\n \n # create a new recommender and save it\n recommender_model = RecommenderModel(\n class_name=cls.__name__,\n name=cls.name,\n is_built=False,\n are_subjects_objects=False\n ) \n recommender_model.save() \n \n cls._print('Importing new predictions...')\n \n # open the csv reader\n reader = csv.reader(open(filename, \"rb\"), delimiter=',', quoting=csv.QUOTE_NONE)\n \n # parse the csv line by line\n for subj_id, obj_id, expectancy in reader:\n \n # parse the values to the right types\n subj_id = int(subj_id)\n obj_id = int(obj_id)\n expectancy = float(expectancy)\n \n # if it's already imported, go to the next pair\n if cls.PredictionModel.objects.filter(\n subj_id=subj_id, obj_id=obj_id, recommender=recommender_model):\n continue\n \n # create a prediction\n cls.PredictionModel.objects.create(\n subj_id=subj_id,\n obj_id=obj_id,\n recommender=recommender_model,\n expectancy=expectancy\n )\n \n recommender_model.is_built=True\n recommender_model.save() \n \n cls._print('Done.')\n\n \n # recommend phase\n #\n @classmethod \n def predict_relationship(cls, subject, object_, save_to_db=False):\n \"\"\"See the base class for the explanation\n\n If the prediction doesn't exist retruns None.\n \"\"\"\n\n recommender_model = cls._get_recommender_model()\n \n # if the recommender isn't built raise an error\n if not recommender_model or not recommender_model.is_built:\n raise RecommenderNotBuiltError(\n message=\"Build the recommender prior to performing the \" + \\\n \"predict_relationship action.\",\n recommender=cls\n )\n \n # get the prediction if it exists\n qs_prediction_model = cls.PredictionModel.objects.filter(\n recommender=recommender_model,\n subj_id=subject.pk,\n obj_id=object_.pk\n )\n \n # if not return none\n if not qs_prediction_model:\n return None\n \n # if so extract the expectancy\n assert qs_prediction_model.count() == 1 \n expectancy = qs_prediction_model[0].expectancy\n \n # create and return the outer-world object\n prediction = RelationshipPrediction(\n subject=subject,\n object_=object_,\n expectancy=expectancy,\n is_uncertain=abs(expectancy - UNCERTAIN_PREDICTION_VALUE) < EXP_PRECISION\n ) \n return prediction\n\n\n @classmethod\n def get_recommendations(cls, subject, count=None): \n \"\"\"For documentation, see the base class\"\"\"\n \n recommender_model = cls._get_recommender_model()\n \n # if the recommender isn't built raise an error\n if not recommender_model or not recommender_model.is_built:\n raise RecommenderNotBuiltError(\n message=\"Build the recommender prior to performing the \" + \\\n \"get_recommendations action.\",\n recommender=cls\n )\n \n # if count wasn't given take the default one\n if not count:\n count = cls.default_recommendation_count\n \n # get the prediction objects\n qs_predictions = cls.PredictionModel.objects.filter(\n subj_id=subject.pk,\n recommender=recommender_model)\\\n .order_by('-expectancy')[:count]\n \n recommendations = []\n \n # go through the obtained predictions\n for pred_model in qs_predictions:\n\n # get the object by the id\n object_=cls.objects.get(pk=pred_model.obj_id)\n \n # create the outer-world object\n prediction = RelationshipPrediction(\n subject=subject,\n object_=object_,\n expectancy=pred_model.expectancy\n ) \n \n recommendations.append(prediction)\n\n return recommendations\n \n \n \n \n\n"},{"alpha_fraction":0.8885714411735535,"alphanum_fraction":0.8885714411735535,"avg_line_length":49,"blob_id":"70352b51a1d48ab49add548b003010777b50fb5b","content_id":"273cda81444a14fdebc1b6b8b4d4649311fd0271","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":350,"license_type":"no_license","max_line_length":100,"num_lines":7,"path":"/UnresystCD/code/adapter/unresyst/combinator/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The combining layer of Unresyst\"\"\"\n\nfrom function_combinator import FunctionCombinator\nfrom combination_element import RelSimilarityCombinationElement, ClusterSimilarityCombinationElement\nfrom average_combinator import AverageCombinator\nfrom twisted_average import TwistedAverageCombinator\nfrom confidence_factor import ConfidenceFactorCombinator\n"},{"alpha_fraction":0.6129541993141174,"alphanum_fraction":0.6200631856918335,"avg_line_length":36.235294342041016,"blob_id":"5bb1986eef9f3ddc7f24b5f21b64cb4710226911","content_id":"b0eed7bd7e60bb70c8a550796fc9139839a1ae65","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1266,"license_type":"no_license","max_line_length":102,"num_lines":34,"path":"/UnresystCD/code/adapter/unresyst/combinator/twisted_average.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The twisted average combinator\"\"\"\n\n\nfrom base import BaseCombinator\n\nclass TwistedAverageCombinator(BaseCombinator):\n \"\"\"A combinator using weighted average\n \"\"\"\n \n def _combine(self, combination_elements, ResultClass):\n \"\"\"See the base class for documentation\n See the thesis text for the explanation of twisted average\n \"\"\"\n \n # number of positive elements\n num_positive = len(filter(lambda ce: ce.get_positiveness(), combination_elements))\n \n # the difference between the number of positive and negative\n pos_dif = abs(len(combination_elements) - 2 * num_positive)\n \n # the average expectancy\n avgexp = sum([ce.get_expectancy() for ce in combination_elements]) / len(combination_elements)\n \n # select the formula according to the expectancy\n if avgexp <= 0.5:\n res_exp = pow(2, pos_dif) * pow(avgexp, pos_dif + 1)\n else:\n res_exp = 1 - abs(pow(2, pos_dif) * pow((avgexp - 1), pos_dif + 1))\n\n # concat the description\n desc = self._concat_descriptions(combination_elements)\n \n # return the resulting class\n return ResultClass(expectancy=res_exp, description=desc)\n"},{"alpha_fraction":0.5863401889801025,"alphanum_fraction":0.5906894207000732,"avg_line_length":43.88768005371094,"blob_id":"210ebfcf21c921cbe01a6946738b98bfdbeade9f","content_id":"6705c2e5e9ccff81de0fd76b46a172caf86cb155","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":12416,"license_type":"no_license","max_line_length":111,"num_lines":276,"path":"/UnresystCD/code/adapter/unresyst/compilator/base.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The base class for all compilators BaseCompilator\"\"\"\nfrom django.db.models import Q\n\nfrom unresyst.constants import *\nfrom unresyst.combinator.combination_element import BiasAggregateCombinationElement, \\\n SubjectObjectRelCombinationElement, PredictedPlusObjectSimilarityCombinationElement, \\\n PredictedPlusSubjectSimilarityCombinationElement, PredictedPlusObjectClusterMemberCombinationElement, \\\n PredictedPlusSubjectClusterMemberCombinationElement, ClusterSimilarityCombinationElement\nfrom unresyst.models.aggregator import AggregatedBiasInstance, AggregatedRelationshipInstance\nfrom unresyst.models.abstractor import PredictedRelationshipDefinition, RelationshipInstance, \\\n ClusterMember, ExplicitRuleInstance\n\n\nclass BaseCompilator(object):\n \"\"\"The base class for all compilators\"\"\"\n \n def __init__(\n self, \n combinator=None, \n depth=DEFAULT_COMPILATOR_DEPTH, \n breadth=DEFAULT_COMPILATOR_BREADTH,\n pair_depth=DEFAULT_COMPILATOR_PAIR_DEPTH):\n \"\"\"The initializer\"\"\"\n\n self.combinator = combinator\n \"\"\"The combinator that will be used in compiling (e.g. the combination of cluster members\n to get similarity. If None, the get first strategy is used.\n \"\"\" \n \n self.depth = depth\n \"\"\"The depth until where the comiplates should be done\"\"\"\n \n self.breadth = breadth\n \"\"\"The number of neighbours that will be taken during the compilation\"\"\" \n \n self.pair_depth = pair_depth\n \"\"\"The number of combination elements that can be taken for each pair from\n each group. Used only for cluster membership.\n \"\"\"\n\n \n\n def compile_all(self, recommender_model):\n \"\"\"Compile preferences, known relationships + similarities. \n \"\"\"\n pass\n \n \n def get_pair_combination_elements(self, dn_subject, dn_object):\n \"\"\"Find all we know about the relationship of dn_subject and\n dn_subject using:\n - aggregated bias for both\n - s-o relationships (all)\n - predicted_relationship + object_similarities\n - predicted_relationship + subject similarities\n - predicted_relationship + subject cluster memberships (pairs not covered by similarities)\n - predicted_relationship + object cluster memberships (pairs not covered by similarities)\n --------- depth COMPILATOR_DEPTH_ONE_UNSURE\n \n @type dn_subject, dn_object: SubjectObject \n @param dn_subject, dn_object: the pair to get combination elements for\n \n @rtype: iterable of BaseCombinationElement\n @return: the list of all we know about the pair\n \"\"\"\n recommender_model = dn_subject.recommender\n els = []\n other_objs = []\n other_subjs = [] \n \n predicted_def = PredictedRelationshipDefinition.objects.get(\n recommender=recommender_model)\n # aggregated bias for both\n qs_bias = AggregatedBiasInstance.objects.filter(\n subject_object__id__in=[dn_subject.id, dn_object.id])\n \n for bias in qs_bias:\n els.append(BiasAggregateCombinationElement(bias_aggregate=bias))\n\n # s-o relationships (all)\n #\n \n qs_rels = RelationshipInstance.objects\\\n .exclude(definition=predicted_def)\\\n .filter(definition__recommender=recommender_model)\\\n .filter(subject_object1=dn_subject, subject_object2=dn_object)\\\n \n for rel in qs_rels:\n els.append(SubjectObjectRelCombinationElement(rel_instance=rel))\n \n # explicit also \n qs_expl_rels = ExplicitRuleInstance.objects\\\n .filter(definition__recommender=recommender_model)\\\n .filter(subject_object1=dn_subject, subject_object2=dn_object)\\\n \n for rel in qs_expl_rels:\n els.append(SubjectObjectRelCombinationElement(rel_instance=rel))\n \n # predicted_relationship + object_similarities\n #\n\n # get similarities starting the traverse with the similarity.\n qs_sim1 = AggregatedRelationshipInstance.objects\\\n .filter(recommender=recommender_model)\\\n .filter(\n Q(\n # take dn_object as stable - in the subject_object2 position of the similarity\n subject_object2=dn_object,\n \n # traverse from the other object in similarity (subject_object1) through\n # the relationship instance, its subject (subject_object1) must be dn_subject\n subject_object1__relationshipinstance_relationships2__subject_object1=dn_subject,\n \n # the relationship instance definition must be the predicted relationship def\n subject_object1__relationshipinstance_relationships2__definition=predicted_def) | \\\n Q( \n # take dn_object as stable again, now in the subject_object1 position of the similarity\n subject_object1=dn_object, \n \n # traverse from the other through relationship to dn_subject\n subject_object2__relationshipinstance_relationships2__subject_object1=dn_subject,\n \n # the definition again must be the predicted\n subject_object2__relationshipinstance_relationships2__definition=predicted_def)).distinct()\n \n\n \n # create combination elements\n for sim_rel in qs_sim1:\n \n # get the object that is similar to dn_object\n other_obj = sim_rel.get_related(dn_object)\n \n other_objs.append(other_obj)\n \n # get the predicted relationship to the other object\n predicted_rel = RelationshipInstance.objects.get(\n definition=predicted_def, \n subject_object1=dn_subject,\n subject_object2=other_obj)\n \n els.append(PredictedPlusObjectSimilarityCombinationElement(\n predicted_rel=predicted_rel,\n similarity_aggregate=sim_rel))\n \n\n \n # predicted_relationship + subject similarities\n #\n qs_sim2 = AggregatedRelationshipInstance.objects\\\n .filter(recommender=recommender_model)\\\n .filter(\n Q(\n # take dn_subject as stable - in the subject_object2 position of the similarity\n subject_object2=dn_subject,\n \n # traverse from the other object in similarity (subject_object1) through\n # the relationship instance, its object (subject_object2) must be dn_object\n subject_object1__relationshipinstance_relationships1__subject_object2=dn_object,\n \n # the relationship instance definition must be the predicted relationship def\n subject_object1__relationshipinstance_relationships1__definition=predicted_def) | \\\n Q( \n # take dn_subject as stable again, now in the subject_object1 position of the similarity\n subject_object1=dn_subject, \n \n # traverse from the other through relationship to dn_object\n subject_object2__relationshipinstance_relationships1__subject_object2=dn_object,\n \n # the definition again must be the predicted\n subject_object2__relationshipinstance_relationships1__definition=predicted_def)).distinct()\n \n # create combination elements\n for sim_rel in qs_sim2:\n \n # get the object that is similar to dn_object\n other_subj = sim_rel.get_related(dn_subject)\n\n other_subjs.append(other_subj)\n \n \n # get the predicted relationship to the other subject\n predicted_rel = RelationshipInstance.objects.get(\n definition=predicted_def, \n subject_object1=other_subj,\n subject_object2=dn_object)\n \n els.append(PredictedPlusSubjectSimilarityCombinationElement(\n predicted_rel=predicted_rel,\n similarity_aggregate=sim_rel))\n\n \n \n # predicted_relationship + object cluster memberships (pairs not covered by similarities) \n # \n \n # take the second membership from the object (first on the path from subject)\n # exclude the trivial cases\n qs_memberships_o = ClusterMember.objects.filter(\n cluster__cluster_set__recommender=recommender_model,\n cluster__clustermember__member=dn_object,\n member__relationshipinstance_relationships2__definition=predicted_def,\n member__relationshipinstance_relationships2__subject_object1=dn_subject)\\\n .exclude(member=dn_object)\\\n .distinct()\\\n .order_by('-confidence')\n \n # create the combination elements \n for cm in qs_memberships_o[:self.pair_depth]:\n \n # dont include the ones that are already there because of the similarity\n if cm.member in other_objs:\n continue\n \n # get the predicted relationship to the subject\n predicted_rel = RelationshipInstance.objects.get(\n definition=predicted_def, \n subject_object1=dn_subject,\n subject_object2=cm.member)\n \n # get the other membership\n second_membership = ClusterMember.objects.get(\n cluster=cm.cluster,\n member=dn_object)\n \n # create the combination element out of the memberships \n ce = ClusterSimilarityCombinationElement(\n cluster_members=(cm, second_membership)) \n \n els.append(PredictedPlusObjectClusterMemberCombinationElement(\n predicted_rel=predicted_rel,\n cluster_combination_element=ce)) \n \n \n # predicted_relationship + subject cluster memberships (pairs not covered by similarities) \n # \n \n # take the second membership from the subject\n # exclude the trivial cases\n qs_memberships_s = ClusterMember.objects.filter(\n cluster__cluster_set__recommender=recommender_model,\n cluster__clustermember__member=dn_subject,\n member__relationshipinstance_relationships1__definition=predicted_def,\n member__relationshipinstance_relationships1__subject_object2=dn_object)\\\n .exclude(member=dn_subject)\\\n .distinct()\\\n .order_by('-confidence')\n \n # create the combination elements \n for cm in qs_memberships_s[:self.pair_depth]:\n \n # dont include the ones that are already there because of the similarity\n if cm.member in other_subjs:\n continue\n \n # get the predicted relationship to the object\n predicted_rel = RelationshipInstance.objects.get(\n definition=predicted_def, \n subject_object1=cm.member,\n subject_object2=dn_object)\n \n # get the other membership\n first_membership = ClusterMember.objects.get(\n cluster=cm.cluster,\n member=dn_subject)\n \n # create the combination element out of the memberships \n ce = ClusterSimilarityCombinationElement(\n cluster_members=(first_membership, cm)) \n \n # append the element to the result \n els.append(PredictedPlusSubjectClusterMemberCombinationElement(\n predicted_rel=predicted_rel,\n cluster_combination_element=ce)) \n \n return els\n \n \n \n"},{"alpha_fraction":0.6250635385513306,"alphanum_fraction":0.6255719661712646,"avg_line_length":29.703125,"blob_id":"25bec476c43ad9c0f3a1bb9e4445a55935df12d9","content_id":"b1a85e4a4706479fa3844af78cb7d5d6652427f7","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3934,"license_type":"no_license","max_line_length":80,"num_lines":128,"path":"/UnresystCD/code/adapter/demo/models.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Demo models for the demo application\"\"\"\n\nfrom django.db import models\n\nfrom constants import *\n\nclass ShoePair(models.Model):\n \"\"\"A model for a pair of shoes.\"\"\"\n \n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"The shoe model name\"\"\"\n \n manufacturer = models.ForeignKey('Manufacturer', null=True)\n \"\"\"The manufacturer who made the shoes.\"\"\"\n \n for_winter = models.BooleanField(default=False)\n \"\"\"Are the shoes suitable for winter?\"\"\"\n \n keywords = models.ManyToManyField('Keyword')\n \"\"\"The keywords associated with the shoes.\"\"\"\n\n image_path = models.CharField(max_length=MAX_LENGTH_IMAGE_PATH)\n \"\"\"The path to the shoe image, relative to the MEDIA_ROOT\"\"\"\n \n category = models.ForeignKey('ShoeCategory', null=True)\n \"\"\"The category where the shoe pair belongs\"\"\"\n \n def get_avatar_path(self):\n return self.image_path\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n\n\nclass ShoeCategory(models.Model): \n \"\"\"A category of shoes, like casual, sport, ...\"\"\"\n \n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"The shoe category name\"\"\"\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name\n\n\nclass Keyword(models.Model):\n \"\"\"A model for a keyword associated with shoe pair.\"\"\"\n \n word = models.CharField(max_length=MAX_LENGTH_NAME, unique=True)\n \"\"\"The key word.\"\"\" \n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.word\n\nclass User(models.Model):\n \"\"\"A model for the users of the demo system.\"\"\"\n\n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"The user's name\"\"\"\n\n age = models.PositiveIntegerField(null=True, default=None)\n \"\"\"The age of the user.\"\"\"\n \n likes_shoes = models.ManyToManyField('ShoePair', related_name='likers')\n \"\"\"User's favorite shoes.\"\"\"\n \n viewed_shoes = models.ManyToManyField('ShoePair', related_name='viewers')\n \"\"\"Shoes the user has viewed.\"\"\"\n \n home_city = models.ForeignKey('City', null=True, default=None) \n \"\"\"The city where the user lives.\"\"\"\n \n words_searched = models.ManyToManyField('Keyword', related_name='searchers')\n \"\"\"The words searched by the user.\"\"\"\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name\n\nclass ShoeRating(models.Model):\n \"\"\"An explicit rating by user to a shoe pair\"\"\"\n \n user = models.ForeignKey('User')\n \"\"\"The rater\"\"\"\n \n shoe_pair = models.ForeignKey('ShoePair')\n \"\"\"The rated shoe pair\"\"\"\n \n stars = models.PositiveIntegerField()\n \"\"\"Number of stars from 0 to 5 given to the shoepair\"\"\"\n\n class Meta:\n unique_together = ('user', 'shoe_pair')\n \"\"\"There can be only one rating from a user to a shoe pair\"\"\"\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return u\"%s: %s (%d)\" % (self.user, self.shoe_pair, self.stars) \n \n\nclass Manufacturer(models.Model):\n \"\"\"A model for a shoe manufacturer\"\"\"\n \n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"The name of the manufacturer\"\"\"\n \n home_city = models.ForeignKey('City', null=True)\n \"\"\"The city where the manufacturer seats.\"\"\"\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n\n\nclass City(models.Model):\n \"\"\"A model for a city.\"\"\" \n \n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"The name of the city.\"\"\"\n \n in_south = models.BooleanField(default=False)\n \"\"\"Is the city in south?\"\"\"\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n"},{"alpha_fraction":0.7131147384643555,"alphanum_fraction":0.7418032884597778,"avg_line_length":26,"blob_id":"5459b44dd127d7846518d124979b6801463a65b7","content_id":"0cce9ecc7bab47dcb095cc4beb97f35d24fb5761","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":244,"license_type":"no_license","max_line_length":85,"num_lines":9,"path":"/code/adapter/modelviz.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"#/bin/bash\n\n# 1 argument: the application name for which models visualization should be performed\n# there must exist the 'modelvizs' directory\n\npython modelviz.py $1 > $1.dot\ndot $1.dot -Tpng -o modelvizs/$1.png\nrm $1.dot\neog modelvizs/$1.png \n"},{"alpha_fraction":0.5691642761230469,"alphanum_fraction":0.5788905024528503,"avg_line_length":29.80555534362793,"blob_id":"e013e8a26d2b1c4a91c20d6a2c421f099f3254c3","content_id":"bd40d365043ff62f86670e6a64dd0a739646a891","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":5552,"license_type":"no_license","max_line_length":104,"num_lines":180,"path":"/UnresystCD/code/adapter/lastfm/recommender.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The configuration for the last.fm recommender\"\"\"\nfrom django.db.models import Sum, Count\n\nfrom unresyst import *\n\nfrom models import *\nfrom constants import *\n\n\nAGE_DIFFERENCE = 38 - 17\n\nREGISTERED_DIFFERENCE = 733696 - 731857\n\nMAX_SCROBBLE_COUNT = 85\n\nSCROBBLE_DIFFERENCE = MAX_SCROBBLE_COUNT - 1\n\ndef _listens_artist_generator():\n \"\"\"The generator to the predicted relationship\"\"\"\n for u in User.objects.iterator():\n for a in Artist.objects.filter(track__scrobble__user=u).distinct().iterator():\n yield (u, a)\n\ndef _get_artist_tag_pairs(artist):\n \"\"\"Get list tags, confidence for all tags he/she has\"\"\" \n \n sumcount = artist.artisttag_set.aggregate(Sum('count'))['count__sum']\n \n # get only tags that are shared between at least two artists \n artist_tags = artist.artisttag_set.annotate(tagcount=Count('tag__artisttag')).filter(tagcount__gt=1)\n\n # confidence is counted as \n # the count of the tag / how many times the artist was tagged.\n return [(at.tag.name, float(at.count)/sumcount) \\\n for at in artist_tags]\n\ndef _tag_similarity_generator():\n \"\"\"Generate the pairs that share some tags\"\"\"\n \n # the artists that were tagged at least once\n qs_tagged_artists = Artist.objects.filter(artisttag__isnull=False).distinct()\n count = qs_tagged_artists.count()\n \n for a1, count in zip( qs_tagged_artists[1:].iterator(), \\\n range(1, count)):\n\n # obtain only first count entities\n for a2 in qs_tagged_artists[:count].iterator():\n if a1.artisttag_set.filter(\n tag__id__in=a2.artisttag_set.values_list('tag__id')\n ).distinct().count() > 45:\n yield (a1, a2) \n\ndef _gender_specific_generator():\n \"\"\"Generate pairs for the gender-specific tag rule\"\"\"\n \n qs_tagged_artists = Artist.objects.filter(artisttag__tag__gender_specific='m').distinct()\n \n for a in qs_tagged_artists:\n for u in User.objects.filter(gender='f'):\n yield (u, a)\n\nclass NovelArtistRecommender(Recommender):\n \"\"\"A recommender for discovering previously unheard artists\"\"\" \n\n name = \"Novel Artist Recommender\"\n \"\"\"The name\"\"\" \n \n subjects = User.objects\n \"\"\"The subjects to who the recommender will recommend.\"\"\"\n \n objects = Artist.objects\n \"\"\"The objects that will be recommended.\"\"\" \n\n predicted_relationship = PredictedRelationship( \n name=\"User listens to artist's tracks.\",\n condition=None, \n description=\"\"\"User %(subject)s listens to the %(object)s's tracks.\"\"\",\n generator=_listens_artist_generator\n )\n \"\"\"The relationship that will be predicted\"\"\"\n\n \"\"\"\n clusters:\n user:\n - country\n - gender\n \"\"\"\n \n relationships = ()\n \n rules = (\n # users of similar age are similar\n SubjectSimilarityRule(\n name=\"Users with similar age.\",\n \n # both users have given their age and the difference \n # is lower than five\n condition=lambda user1, user2: \n user1.age and user2.age and abs(user1.age - user2.age) <= 5,\n \n is_positive=True, \n \n weight=0.4,\n \n # a magic linear confidence function\n confidence=lambda user1, user2: \n 1 - float(abs(user1.age - user2.age))/AGE_DIFFERENCE,\n \n description=\"Users %(subject1)s and %(subject2)s are about \" + \n \"the same age.\"\n ), \n ) \n\n cluster_sets = (\n\n \n # user - gender\n SubjectClusterSet(\n \n name='User gender.',\n \n weight=0.5,\n \n # users that have a gender (filled)\n filter_entities=User.objects.exclude(gender=''),\n \n get_cluster_confidence_pairs=lambda user: ((user.gender, 1),),\n \n description=\"%(subject)s's gender is %(cluster)s.\"\n \n ),\n \n # user - country\n SubjectClusterSet(\n \n name='User country.',\n \n weight=0.1,\n \n # users that have country filled\n filter_entities=User.objects.filter(country__isnull=False),\n \n get_cluster_confidence_pairs=lambda user: ((user.country.name, 1),),\n \n description=\"%(subject)s is from %(cluster)s.\"\n ),\n )\n \n save_all_to_predictions = True\n\n random_recommendation_description = \"Recommending a random artist to the user.\"\n \n \n\nclass ArtistRecommender(NovelArtistRecommender):\n \"\"\"A recommender for recommending items no matter if they were heard or not.\"\"\"\n\n name = \"Artist Recommender\"\n \"\"\"The name\"\"\" \n \n remove_predicted_from_recommendations = False \n \"\"\"The already heard artists can appear in recommendations\"\"\"\n \n rules = NovelArtistRecommender.rules + ((SubjectObjectRule(\n name=\"User has listened to the artist.\",\n generator=_listens_artist_generator,\n weight=0.5,\n is_positive=True,\n description=\"%(subject)s listened to %(object)s.\",\n\n # the number of user's scrobbles on artist divided by the number of\n # user's scrobbles overall\n confidence=lambda s, o: \n float(Scrobble.objects.filter(user=s, track__artist=o).count())\\\n /Scrobble.objects.filter(user=s).count(),\n \n \n )),\n )\n \n"},{"alpha_fraction":0.5535823702812195,"alphanum_fraction":0.5584813356399536,"avg_line_length":28.472726821899414,"blob_id":"3bdb67cdeb002662d10c41d4a19ab745c5a01cd5","content_id":"ba2543ea7825f6c68f2f9989d2aaad071fe2cebd","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3266,"license_type":"no_license","max_line_length":99,"num_lines":110,"path":"/code/adapter/flixster/models.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Models for the flixster app\"\"\"\nfrom django.db import models\nfrom django.db.models import Min\n\nfrom unresyst.models import BaseEvaluationPair\nfrom constants import *\n\nclass User(models.Model):\n \"\"\"The user\"\"\"\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return 'user_%d' % self.id \n \n \nclass Movie(models.Model): \n \"\"\"The movie\"\"\"\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return 'movie_%d' % self.id \n\n \nclass Friend(models.Model):\n \"\"\"A user is a friend of another user. Symmetric. Smaller id first.\"\"\"\n\n friend1 = models.ForeignKey('flixster.User', related_name='friends1')\n \"\"\"The first user in the friend relationship\"\"\" \n \n friend2 = models.ForeignKey('flixster.User', related_name='friends2')\n \"\"\"The second user in the friend relationship\"\"\" \n \n class Meta:\n unique_together = ('friend1', 'friend2') \n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return u'(%s, %s)' % (self.friend1, self.friend2)\n\n \nclass Rating(models.Model):\n \"\"\"A user rating a movie\"\"\"\n \n user = models.ForeignKey('flixster.User')\n \"\"\"The user\"\"\"\n \n movie = models.ForeignKey('flixster.Movie')\n \"\"\"The movie\"\"\"\n \n rating = models.DecimalField(max_digits=2, decimal_places=1)\n \"\"\"The rating\"\"\"\n\n class Meta:\n unique_together = ('user', 'movie')\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return u'%s - %s: %f' % (self.user, self.movie, self.rating)\n\n\nclass MovieEvaluationPair(BaseEvaluationPair):\n \"\"\"Movie test pairs\"\"\"\n\n subj = models.ForeignKey('flixster.User')\n \"\"\"The subject\"\"\"\n \n obj = models.ForeignKey('flixster.Movie')\n \"\"\"The object\"\"\"\n \n test_ratio = 0.2\n \"\"\"The ratio of pairs to select to test pairs\"\"\"\n\n class Meta:\n app_label = 'flixster' \n \n @classmethod \n def select(cls, i=0):\n \"\"\"See the base class for the documentation.\"\"\"\n\n n = int(1 / cls.test_ratio) \n all_count = Rating.objects.count() \n \n it = 0 \n # take every n-th rating, remove it and put it to test data \n # save to test, remove from build\n for rating in Rating.objects.order_by('id').iterator():\n\n it += 1\n\n # if we aren't on the nth object go ahead\n if it % n != i:\n continue\n \n # save the test object and delete it from train\n cls.objects.create(\n subj=rating.user,\n obj=rating.movie,\n expected_expectancy=rating.rating/MAX_STARS)\n \n rating.delete()\n \n \n test_count = cls.objects.count()\n \n print \"%d test pairs selected from total %d pairs\" % (test_count, all_count) \n\n\n def get_success(self):\n \"\"\"See the base class for the documentation.\"\"\"\n return abs(self.obtained_expectancy - self.expected_expectancy) < MAX_TOLERANCE / MAX_STARS\n \n \n \n"},{"alpha_fraction":0.6610608100891113,"alphanum_fraction":0.6610608100891113,"avg_line_length":31.20833396911621,"blob_id":"3334f4d95e591e313845bdaff67bb82b060bdc3d","content_id":"efc32801bbcc7d2e860c6653c2dacbea11ff5a43","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":773,"license_type":"no_license","max_line_length":94,"num_lines":24,"path":"/code/adapter/urls.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"from django.conf.urls.defaults import *\n\nfrom settings import SERVE_STATIC_FILES, MEDIA_ROOT\n\n# Uncomment the next two lines to enable the admin:\n# from django.contrib import admin\n# admin.autodiscover()\n\nurlpatterns = patterns('',\n \n # the demo app \n (r'^', include('demo.urls', namespace='demo'))\n # Uncomment the admin/doc line below to enable admin documentation:\n # (r'^admin/doc/', include('django.contrib.admindocs.urls')),\n\n # Uncomment the next line to enable the admin:\n # (r'^admin/', include(admin.site.urls)),\n \n)\nif SERVE_STATIC_FILES:\n urlpatterns += patterns('',\n # for working with static files (for development only)\n (r'^media/(?P.*)$', 'django.views.static.serve', {'document_root': MEDIA_ROOT}) \n )\n"},{"alpha_fraction":0.5111602544784546,"alphanum_fraction":0.5206112861633301,"avg_line_length":35.74074172973633,"blob_id":"648a211ed35ae2ae4920e4491edf2db3d0765fc8","content_id":"4ce47597afb1a0b95f3938f187880230f8333f0c","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":4973,"license_type":"no_license","max_line_length":114,"num_lines":135,"path":"/code/adapter/travel/recbckp.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":" rules = ( \n # click = sign of preference\n SubjectObjectRule(\n name='User has clicked on something on the tour profile.',\n weight=0.5,\n condition=None,\n is_positive=True,\n description='User %(subject)s has clicked on something on the %(object)s profile.',\n # pairs that user has clicked on the tour\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\n for uid, tid in Click.objects.values_list('session__user', 'tour').distinct()),\n # the average is around 3, so take 1/6. so that 3 points to the middle.\n confidence=lambda u, t: min(float(Click.objects.filter(tour=t, session__user=u).count())/6, 1.0),\n ),\n \n # question .. also a sign of preference\n SubjectObjectRule(\n name='User has asked about the tour.',\n weight=0.5,\n condition=None,\n is_positive=True,\n description='User %(subject)s has asked about %(object)s.',\n # pairs that user has asked on the tour\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\n for uid, tid in Question.objects.values_list('session__user', 'tour').distinct()),\n confidence=lambda u, t: min(float(Question.objects.filter(tour=t, session__user=u).count())/2, 1.0),\n ),\n \n # mouse move .. also a sign of preference\n SubjectObjectRule(\n name='User has moved the mouse on the tour profile.',\n weight=0.5,\n condition=None,\n is_positive=True,\n description='User %(subject)s has moved the mouse on %(object)s.',\n # pairs that user has moved on the tour\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\n for uid, tid in MouseMove.objects.values_list('session__user', 'tour').distinct()),\n confidence=lambda u, t: min(float(MouseMove.objects.filter(tour=t, session__user=u).count())/18, 1.0),\n ),\n \n # view profile \n SubjectObjectRule(\n name='User has viewed the tour profile page.',\n weight=0.5,\n condition=None,\n is_positive=True,\n description='User %(subject)s has viewed %(object)s.',\n # pairs that user has viewed the tour\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\n for uid, tid in ViewProfile.objects.values_list('session__user', 'tour').distinct()),\n # how many times * how long\n confidence=_viewed_profile_confidence\n ),\n\n \n )\n \n cluster_sets = (\n # cluster - tour types\n ObjectClusterSet(\n\n name=\"Tour type cluster set.\",\n\n weight=0.5,\n \n filter_entities=Tour.objects.all(),\n \n get_cluster_confidence_pairs=lambda tour: ((tour.tour_type.name, 1),),\n \n description=\"The tour %(object)s has type %(cluster)s.\",\n ),\n \n # cluster - countries\n ObjectClusterSet(\n\n name=\"Country cluster set.\",\n\n weight=0.5,\n \n filter_entities=Tour.objects.all(),\n \n get_cluster_confidence_pairs=lambda tour: ((tour.country.name, 1),),\n \n description=\"The tour %(object)s is to %(cluster)s country.\",\n ),\n\n )\n \n biases = (\n # multiply viewed tours\n ObjectBias(\n name=\"Most viewed tours.\",\n \n description=\"Tour %(object)s is much viewed\",\n \n weight=0.5,\n \n is_positive=True,\n \n generator=lambda: Tour.objects.annotate(hh=Count('viewprofile')).filter(hh__gt=2).distinct(),\n \n confidence=lambda t: min(float(t.viewprofile_set.count())/4, 1.0)\n ),\n \n # multiply clicked tours\n ObjectBias(\n name=\"Most clicked tours.\",\n \n description=\"Tour %(object)s is often clicked on.\",\n \n weight=0.5,\n \n is_positive=True,\n \n generator=lambda: Tour.objects.annotate(hh=Count('click')).filter(hh__gt=1).distinct(),\n \n confidence=lambda t: min(float(t.click_set.count())/2, 1.0)\n ),\n \n # multiply mouse moved tours\n ObjectBias(\n name=\"Most mouse moved tours.\",\n \n description=\"Tour %(object)s is often mouse moved.\",\n \n weight=0.5,\n \n is_positive=True,\n \n generator=lambda: Tour.objects.annotate(hh=Count('mousemove')).filter(hh__gt=6).distinct(),\n \n confidence=lambda t: min(float(t.mousemove_set.count())/12, 1.0)\n ),\n )\n\n\n\n \n\n"},{"alpha_fraction":0.8060836791992188,"alphanum_fraction":0.8060836791992188,"avg_line_length":28.22222137451172,"blob_id":"03439cbcfe150685e23ccce6e7ba2bcdefedf629","content_id":"83dd87a6ac3bcf90c1e71fdb345da51af7d52ce5","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":263,"license_type":"no_license","max_line_length":57,"num_lines":9,"path":"/code/adapter/unresyst/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"A package holding the whole unresyst application.\n\nThe only class to be used outside is the BaseRecommender.\n\"\"\"\n\nfrom recommender.recommender import Recommender\nfrom recommender.rules import *\nfrom recommender.clusters import *\nfrom recommender.bias import *\n"},{"alpha_fraction":0.5788713097572327,"alphanum_fraction":0.5795742869377136,"avg_line_length":35.88328552246094,"blob_id":"85ee8cdf4107b400d05f293407aa78ee45437b76","content_id":"0755784d1a25ffa1dfb225da4565b41a8ff03c88","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":25605,"license_type":"no_license","max_line_length":120,"num_lines":694,"path":"/code/adapter/unresyst/recommender/recommender.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Module containing the definition of BaseRecommender: the base class \nfor user-defined recommenders.\n\nContents:\n - Recommender: the recommender for client subclassing\n - MetaRecommender: the metaclass for creating recommender classes\n\"\"\"\nimport math\nimport copy\nimport csv\n\nfrom base import BaseRecommender\nfrom predictions import RelationshipPrediction\nfrom unresyst.constants import *\nfrom unresyst.exceptions import ConfigurationError, InvalidParameterError, \\\n RecommenderNotBuiltError\nfrom unresyst.abstractor import BasicAbstractor \nfrom unresyst.aggregator import LinearAggregator, CombiningAggregator\nfrom unresyst.algorithm import SimpleAlgorithm, AggregatingAlgorithm, CompilingAlgorithm\nfrom unresyst.models.common import SubjectObject, Recommender as RecommenderModel\nfrom unresyst.compilator import GetFirstCompilator, CombiningCompilator\nfrom unresyst.combinator import AverageCombinator, TwistedAverageCombinator, ConfidenceFactorCombinator\nfrom unresyst.models.abstractor import RelationshipInstance, ExplicitRuleInstance\nfrom unresyst.models.algorithm import RelationshipPredictionInstance\n\ndef _assign_recommender(list_rels, recommender):\n \"\"\"Go throuth the list, if the items have the \"recommender\" attribute,\n create a copy to the returning list, if not put it there directly\n \n @param list_rels: the list of rules/relationships\n \n @param recommender: the recommender they belong to \n \n @rtype: tuple of rules/relationships\n @return: the tuple of new or copied rules/relationships\n \"\"\"\n new_rels = []\n \n # go through the rels \n for rel in list_rels:\n \n # if they already have a recommender assigned, copy them\n if hasattr(rel, \"recommender\"):\n rel = copy.copy(rel)\n \n # assign the recommender and append it \n rel.recommender = recommender\n new_rels.append(rel)\n \n return tuple(new_rels)\n\nclass MetaRecommender(type):\n \"\"\"The meta-class adding a reference to the class for the contained\n rules and relationships.\n \"\"\"\n \n def __init__(cls, name, bases, dct): \n \"\"\"The class initializer.\n \n Adds the reference to the recommender class to all of the rules\n and relationships.\n \n Tries searching for the recommender model in the database.\n \"\"\"\n \n super(MetaRecommender, cls).__init__(name, bases, dct)\n \n # add the recommender class to the predicted relationship\n if cls.predicted_relationship: \n\n # if it already has a recommender assigned, create a copy.\n if hasattr(cls.predicted_relationship, 'recommender'):\n cls.predicted_relationship = copy.copy(cls.predicted_relationship)\n \n cls.predicted_relationship.recommender = cls\n \n # then to the rules\n if cls.rules:\n cls.rules = _assign_recommender(list_rels=cls.rules, recommender=cls) \n \n # finally to the relationships\n if cls.relationships:\n cls.relationships = _assign_recommender(\n list_rels=cls.relationships, \n recommender=cls)\n \n # moreover to the cluster sets\n if cls.cluster_sets:\n cls.cluster_sets = _assign_recommender(\n list_rels=cls.cluster_sets,\n recommender=cls) \n \n # and to the biases\n if cls.biases:\n cls.biases = _assign_recommender(\n list_rels=cls.biases,\n recommender=cls) \n\n\nclass Recommender(BaseRecommender):\n \"\"\"The base class for all user-defined recommenders.\n \n Implements the interface of the recommender. It doesn't hold any \n domain specific data.\n \n Defines default behaviour assigning the classes for the layers. \n \n The predicted relationship is ignored when building and updating the \n recommender. If it's needed, one should create a rule/relationship \n with the appropriate condition.\n \n The contained rules and relationships have to be included in only one\n recommender class. There can't be a rule/relationship instance \n which is contained in two recommender classes.\n \"\"\"\n __metaclass__ = MetaRecommender\n \"\"\"A metaclass putting the reference to the recommender to all member\n rules and relationships.\n \"\"\" \n \n # Build phase:\n #\n \n @classmethod\n def build(cls):\n \"\"\"For documentation, see the base class\"\"\"\n \n \n # validate the subclass data\n # \n \n # Subjects, objects non-empty\n if not cls.subjects.all().exists():\n raise ConfigurationError(\n message=\"No subjects given\",\n recommender=cls,\n parameter_name=\"Recommender.subjects\",\n parameter_value=cls.subjects\n )\n \n if not cls.objects.all().exists():\n raise ConfigurationError(\n message=\"No objects given\",\n recommender=cls,\n parameter_name=\"Recommender.objects\",\n parameter_value=cls.objects\n )\n \n \n # predicted relationship given\n if not cls.predicted_relationship:\n raise ConfigurationError(\n message=\"No predicted relationship given\",\n recommender=cls,\n parameter_name=\"Recommender.predicted_relationship\",\n parameter_value=cls.predicted_relationship\n )\n \n # rules and relationships don't have to be given\n \n cls._print('Recommender validated, deleting old objects...')\n \n # if the recommender with the given name exists, delete it,\n RecommenderModel.objects.filter(class_name=cls.__name__).delete()\n \n # create a new recommender and save it, keep it in the class\n recommender_model = RecommenderModel(\n class_name=cls.__name__,\n name=cls.name,\n is_built=False,\n are_subjects_objects=(cls.subjects == cls.objects),\n random_recommendation_description=cls.random_recommendation_description,\n remove_predicted_from_recommendations=cls.remove_predicted_from_recommendations\n ) \n recommender_model.save() \n \n # build the recommender model\n #\n #\n cls._print(\"Old objects deleted. Creating universal subjectobjects...\")\n \n # Abstractor\n #\n \n # create the domain neutral representation for objects and subjects\n cls.abstractor.create_subjectobjects(\n recommender_model=recommender_model,\n subjects=cls.subjects, \n objects=cls.objects\n )\n \n cls._print(\"Universal subject and object representations created. Creating predicted_relationship instances...\")\n \n # create the relationship instances for the predicted relationship\n cls.abstractor.create_predicted_relationship_instances( \n predicted_relationship=cls.predicted_relationship \n )\n \n cls._print(\"Predicted relationship instances created. Creating relationship instances...\")\n \n # create relationship instances between subjects/objects \n cls.abstractor.create_relationship_instances(\n relationships=cls.relationships\n ) \n \n cls._print(\"Relationship instances created. Creating rule instances...\")\n \n # evaluate rules and make rule instances between the affected \n # subjects/objects\n cls.abstractor.create_rule_instances(rules=cls.rules)\n \n cls._print(\"Rule instances created. Creating clusters...\")\n \n # evaluate the clusters and their members\n cls.abstractor.create_clusters(cluster_sets=cls.cluster_sets)\n \n cls._print(\"Clusters created. Creating biases...\")\n \n # evaluate the biases\n cls.abstractor.create_biases(biases=cls.biases)\n \n cls._print(\"Biases created. Aggregating...\")\n\n \n \n # Algorithm\n # \n # build the algorithm model from the aggregated relationships\n cls.algorithm.build(recommender_model=recommender_model)\n \n cls._print(\"Algorithm built.\")\n \n # if it should be done and predicted should be removed, \n # save predicted_rel to predictions\n if cls.remove_predicted_from_recommendations and cls.save_all_to_predictions:\n \n cls._print(\"Saving explicit/predicted to predictions...\")\n \n # if explicit relationship is available, get its instances\n # if not, get the predicted_rel\n qs_predicted_rels = ExplicitRuleInstance.objects.filter(definition__recommender=recommender_model) \\\n if cls.explicit_rating_rule else \\\n RelationshipInstance.filter_predicted(recommender_model=recommender_model)\n \n for ri in qs_predicted_rels:\n \n # get the expectancy of the rating or the trivial\n expectancy = ri.expectancy if cls.explicit_rating_rule else TRIVIAL_EXPECTANCY \n \n rpi, created = RelationshipPredictionInstance.objects.get_or_create(\n subject_object1=ri.subject_object1,\n subject_object2=ri.subject_object2,\n recommender=recommender_model,\n defaults={\n 'expectancy': expectancy,\n 'is_trivial': True,\n 'description': ri.description,\n }\n )\n \n # if it was found update it to the predicted\n if not created:\n rpi.expectancy = expectancy\n rpi.is_trivial = True\n rpi.description = ri.description\n \n # save it if created or not \n rpi.save()\n \n # mark the recommender as built, save it and keep it in the class\n recommender_model.is_built = True\n recommender_model.save()\n \n cls._print('Done')\n\n\n # Recommend phase:\n # \n \n @classmethod \n def predict_relationship(cls, subject, object_, save_to_db=False):\n \"\"\"For documentation, see the base class\"\"\" \n \n recommender_model = cls._get_recommender_model()\n # if the recommender isn't built raise an error\n if not recommender_model or not recommender_model.is_built:\n raise RecommenderNotBuiltError(\n message=\"Build the recommender prior to performing the \" + \\\n \"predict_relationship action.\",\n recommender=cls\n )\n \n \n subject_ent_type = ENTITY_TYPE_SUBJECT \\\n if not recommender_model.are_subjects_objects \\\n else ENTITY_TYPE_SUBJECTOBJECT\n\n object_ent_type = ENTITY_TYPE_OBJECT \\\n if not recommender_model.are_subjects_objects \\\n else ENTITY_TYPE_SUBJECTOBJECT\n\n # get the domain neutral representations for the subject and object\n try:\n dn_subject = SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=subject,\n entity_type=subject_ent_type,\n recommender=recommender_model\n )\n except SubjectObject.DoesNotExist, e:\n raise InvalidParameterError(\n message=\"The subject wasn't found in the recommender database.\" + \\\n \"Try rebuilding the recommender. Exception: %s\" % e,\n recommender=cls,\n parameter_name='subject', \n parameter_value=subject) \n\n try: \n dn_object = SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=object_,\n entity_type=object_ent_type,\n recommender=recommender_model\n )\n except SubjectObject.DoesNotExist, e:\n raise InvalidParameterError(\n message=\"The object wasn't found in the recommender database.\" + \\\n \"Try rebuilding the recommender. Exception: %s\" % e,\n recommender=cls,\n parameter_name='object_', \n parameter_value=object_) \n \n # get the prediction from the algorithm\n prediction_model = cls.algorithm.get_relationship_prediction(\n recommender_model=recommender_model,\n dn_subject=dn_subject,\n dn_object=dn_object,\n remove_predicted=cls.remove_predicted_from_recommendations\n )\n \n # if it should be done and we know something about the pair\n if save_to_db and not prediction_model.is_uncertain:\n prediction_model.save()\n \n # create and return the outer-world object\n prediction = RelationshipPrediction(\n subject=subject,\n object_=object_,\n expectancy=prediction_model.expectancy,\n explanation=prediction_model.description,\n is_uncertain=prediction_model.is_uncertain\n ) \n return prediction\n\n\n @classmethod\n def get_recommendations(cls, subject, count=None): \n \"\"\"For documentation, see the base class\"\"\"\n \n recommender_model = cls._get_recommender_model()\n \n # if the recommender isn't built raise an error\n if not recommender_model or not recommender_model.is_built:\n raise RecommenderNotBuiltError(\n message=\"Build the recommender prior to performing the \" + \\\n \"get_recommendations action.\",\n recommender=cls\n )\n \n # if count wasn't given take the default one\n if not count:\n count = cls.default_recommendation_count\n \n subject_ent_type = ENTITY_TYPE_SUBJECT \\\n if not recommender_model.are_subjects_objects \\\n else ENTITY_TYPE_SUBJECTOBJECT\n \n # convert the the subject to domain neutral\n try:\n dn_subject = SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=subject,\n entity_type=subject_ent_type,\n recommender=recommender_model\n )\n except SubjectObject.DoesNotExist, e:\n raise InvalidParameterError(\n message=\"The subject wasn't found in the recommender database.\" + \\\n \"Try rebuilding the recommender. Exception: %s\" % e,\n recommender=cls,\n parameter_name='subject', \n parameter_value=subject) \n\n limit = cls.recommendation_expectancy_limit \\\n if not cls.recommendation_expectancy_limit is None else 0\n\n # get the recommendations from the algorithm\n prediction_models = cls.algorithm.get_recommendations(\n recommender_model=recommender_model,\n dn_subject=dn_subject,\n count=count,\n expectancy_limit=limit,\n remove_predicted=cls.remove_predicted_from_recommendations\n )\n \n recommendations = []\n \n # go through the obtained predictions\n for pred_model in prediction_models:\n\n # obtain the object from the prediction\n dn_object = pred_model.get_related(dn_subject)\n \n # get its domain specific representation\n object_ = dn_object.get_domain_specific_entity(entity_manager=cls.objects)\n \n # create the outer-world object\n prediction = RelationshipPrediction(\n subject=subject,\n object_=object_,\n expectancy=pred_model.expectancy,\n explanation=pred_model.description,\n is_uncertain=pred_model.is_uncertain\n ) \n \n recommendations.append(prediction)\n\n return recommendations\n\n @classmethod\n def export_predictions(cls, filename):\n \"\"\"Export all predictions to a csv \n file of the given name.\n \n @type filename: str\n @param filename: the full path to the file\n \n @raise FileNotExists and other file open errors.\n \"\"\"\n recommender_model = cls._get_recommender_model()\n \n with open(filename, 'w') as f:\n \n i = 0\n \n # loop through pairs, export the prediction instances\n for rpi in RelationshipPredictionInstance.objects.filter(recommender=recommender_model):\n \n # create the common part\n linestr = \"%s,%s,%s\\n\" % (\n rpi.subject_object1.id_in_specific, \n rpi.subject_object2.id_in_specific,\n rpi.expectancy) \n \n # write it to the file\n f.write(linestr)\n \n i += 1\n \n print \" %d predictions exported\" % i\n\n @classmethod\n def update_predictions(cls, filename):\n \"\"\"Load predictions for pairs that are unknown in our recommender from \n the given csv file.\n \n The file has to be in format:\n ,,\\n\n \n @type filename: str\n @param filename: the full path to the file\n\n @raise FileNotExists and other file open errors.\n \"\"\"\n recommender_model = cls._get_recommender_model() \n \n # open the csv reader\n reader = csv.reader(open(filename, \"rb\"), delimiter=',', quoting=csv.QUOTE_NONE)\n \n i = 0\n # parse the csv line by line\n for subj_id, obj_id, expectancy in reader:\n \n # parse the values to the right types\n subj_id = int(subj_id)\n obj_id = int(obj_id)\n expectancy = float(expectancy)\n \n dn_subject = SubjectObject.objects.get(\n entity_type=ENTITY_TYPE_SUBJECT, \n id_in_specific=subj_id,\n recommender=recommender_model)\n dn_object = SubjectObject.objects.get(\n entity_type=ENTITY_TYPE_OBJECT, \n id_in_specific=obj_id,\n recommender=recommender_model)\n \n # if it's already imported, go to the next pair\n if RelationshipPredictionInstance.objects.filter(\n subject_object1=dn_subject, \n subject_object2=dn_object, \n recommender=recommender_model).exists():\n continue\n \n # create a prediction\n RelationshipPredictionInstance.objects.create(\n subject_object1=dn_subject,\n subject_object2=dn_object,\n recommender=recommender_model,\n expectancy=expectancy\n )\n\n i += 1\n\n print \"%d new predictions imported\" % i\n \n \n \n \n \n # Update phase:\n # \n @classmethod\n def add_subject(cls, subject):\n \"\"\"For documentation, see the base class\"\"\"\n \n # add the object to abstractor rule/relationship instances, \n # see what has changed\n instance_changes = cls.Abstractor.add_subject(\n recommender=cls.__name__,\n subject=subject\n )\n \n # let the Aggregator update the aggregates\n aggregated_changes = cls.Aggregator.update(\n recommender=cls.__name__,\n instance_changes=instance_changes\n )\n \n # let the Algorithm update its structures\n cls.Algorithm.update(\n recommender=cls.__name__,\n aggregated_changes=aggregated_changes\n )\n\n\n @classmethod\n def add_object(cls, object_):\n \"\"\"For documentation, see the base class\"\"\"\n \n # add the object to abstractor rule/relationship instances\n # see what has changed\n instance_changes = cls.Abstractor.add_object(\n recommender=cls.__name__,\n object_=object_\n )\n # pak nejak podobne - asi na to udelat fci. \n\n @classmethod\n def update_subject(cls, subject):\n \"\"\"For documentation, see the base class\"\"\"\n\n # update the subject in abstractor rule/relationship instances\n # see what has changed\n instance_changes = cls.Abstractor.update_subject(\n recommender=cls.__name__,\n subject=subject\n )\n # pak nejak podobne - asi na to udelat fci.\n\n\n @classmethod \n def update_object(cls, object_):\n \"\"\"For documentation, see the base class\"\"\"\n\n # update the object in abstractor rule/relationship instances\n # see what has changed\n instance_changes = cls.Abstractor.update_object(\n recommender=cls.__name__,\n object_=object_\n )\n # pak nejak podobne - asi na to udelat fci. \n \n @classmethod\n def remove_subject(cls, subject):\n \"\"\"For documentation, see the base class\"\"\"\n\n # remove the subject from abstractor rule/relationship instances\n # see what has changed\n instance_changes = cls.Abstractor.remove_subject(\n recommender=cls.__name__,\n subject=subject\n )\n # pak nejak podobne - asi na to udelat fci.\n\n\n @classmethod \n def remove_object(cls, object_):\n \"\"\"For documentation, see the base class\"\"\"\n\n # update the object in abstractor rule/relationship instances\n # see what has changed\n instance_changes = cls.Abstractor.remove_object(\n recommender=cls.__name__,\n object_=object_\n )\n # pak nejak podobne - asi na to udelat fci. \n \n \n # Class configuration - the behaviour of the layers below the recommender\n # Can be overriden in user defined subclasses\n \n abstractor = BasicAbstractor()\n \"\"\"The class that will be used for the abstractor level. Can be \n overriden in suclasses\"\"\" \n \n algorithm = AggregatingAlgorithm(\n inner_algorithm=CompilingAlgorithm(\n inner_algorithm=SimpleAlgorithm(\n inner_algorithm=None\n ),\n compilator=CombiningCompilator(combinator=TwistedAverageCombinator(), breadth=0) #\n ),\n aggregator=CombiningAggregator(combinator=TwistedAverageCombinator())\n )\n \"\"\"The default algorithm setup.\n The class that will be used for the algorithm level. Can be \n overriden in subclasses\"\"\" \n \n # TwistedAverageCombinator\n # AverageCombinator\n # ConfidenceFactorCombinator\n \n default_recommendation_count = DEFAULT_RECOMMENDATION_COUNT\n \"\"\"The defaul count of the obtained recommended objects\"\"\"\n \n remove_predicted_from_recommendations = True\n \"\"\"The entity pairs that already have the predicted_relationship\n between them, will be removed from recommendation list.\n \"\"\"\n \n random_recommendation_description = \"A random object\"\n \"\"\"The description for random recommendations Can be overriden in\n subclass\"\"\"\n \n verbose_build = True\n \"\"\"Should messages be printed during the build?\"\"\"\n \n save_all_to_predictions = True\n \n # Auxiliary methods - not to be used from outside the application\n # \n @classmethod\n def _get_recommender_model(cls):\n \"\"\"Get the recommender model belonging to the class\"\"\"\n \n # can't be caching 'cause database can die out without notifying the \n # if it's already saved, return it\n #if cls._recommender_model:\n # return cls._recommender_model\n \n # otherwise try finding it in database \n models = RecommenderModel.objects.filter(class_name=cls.__name__)\n \n # if the recommender was found, assign it to the class and return it\n if models: \n assert len(models) == 1 \n return models[0]\n \n # if not return None\n return None \n \n @classmethod\n def _get_entity_manager(cls, entity_type):\n \"\"\"Get the manager from the recommender for the given entity type.\n \n @type entity_type: str\n @param entity_type: the type of the entity 'S'/'O'/'SO'\n \n @rtype: django.db.models.manager.Manager\n @return: the manager over the domain specific entities.\n \n @raise KeyError: when the entity_type is invalid \n \"\"\"\n \n manager_dict = {\n ENTITY_TYPE_SUBJECT: cls.subjects,\n ENTITY_TYPE_OBJECT: cls.objects, \n ENTITY_TYPE_SUBJECTOBJECT: cls.objects # or sujbects if you care\n }\n \n return manager_dict[entity_type]\n \n @classmethod\n def _print(cls, msg):\n if cls.verbose_build:\n print msg \n"},{"alpha_fraction":0.715423583984375,"alphanum_fraction":0.7286990284919739,"avg_line_length":27.57241439819336,"blob_id":"a6e4e96e9efadd85c404c0083ff7cea0be5b242a","content_id":"eb3b0ec2bf38722d752add83f7bf080bf7a09b14","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":4143,"license_type":"no_license","max_line_length":87,"num_lines":145,"path":"/UnresystCD/code/adapter/unresyst/constants.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The constants used in Unresyst\"\"\"\n\n_ = lambda x: x\n\nMAX_LENGTH_NAME = 80\n\"\"\"The maximum length of the name in the universal representation.\"\"\"\n\nMAX_LENGTH_CLASS_NAME = 40\n\"\"\"The maximum length of the class name.\"\"\"\n\nMAX_LENGTH_ENTITY_TYPE = 2\n\"\"\"The maximum length of the entity type string\"\"\"\n\nMAX_LENGTH_RELATIONSHIP_TYPE = 5\n\"\"\"The maximum length of the relationship type string\"\"\"\n\nMAX_LENGTH_ID = 50\n\"\"\"The maximum length of the id in parent system\"\"\"\n\nENTITY_TYPE_SUBJECT = 'S'\n\"\"\"The subject entity type\"\"\"\n\nENTITY_TYPE_OBJECT = 'O'\n\"\"\"The object entity type\"\"\"\n\nENTITY_TYPE_SUBJECTOBJECT = 'SO'\n\"\"\"The subject object entity type\"\"\"\n\nENTITY_TYPE_CHOICES = (\n # a subject:\n (ENTITY_TYPE_SUBJECT, _('Subject')),\n \n # an object:\n (ENTITY_TYPE_OBJECT, _('Object')),\n \n # when subject domain is the same as object domain\n (ENTITY_TYPE_SUBJECTOBJECT, _('Subject == Object')),\n)\n\"\"\"Choices for the entity_type field\"\"\"\n\nRELATIONSHIP_TYPE_SEPARATOR = '-'\n\nRELATIONSHIP_TYPE_SUBJECT_OBJECT = \\\n ENTITY_TYPE_SUBJECT + RELATIONSHIP_TYPE_SEPARATOR + ENTITY_TYPE_OBJECT\n\"\"\"Subject-object relationship type\"\"\"\n\nRELATIONSHIP_TYPE_SUBJECT_SUBJECT = \\\n ENTITY_TYPE_SUBJECT + RELATIONSHIP_TYPE_SEPARATOR + ENTITY_TYPE_SUBJECT \n\"\"\"Subject-subject relationship type\"\"\"\n\nRELATIONSHIP_TYPE_OBJECT_OBJECT = \\\n ENTITY_TYPE_OBJECT + RELATIONSHIP_TYPE_SEPARATOR + ENTITY_TYPE_OBJECT\n\"\"\"Object-object relatioship type\"\"\"\n\nRELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT = \\\n ENTITY_TYPE_SUBJECTOBJECT + RELATIONSHIP_TYPE_SEPARATOR + ENTITY_TYPE_SUBJECTOBJECT\n\"\"\"Subjectobject-subjectobject relatioship type\"\"\"\n\nRELATIONSHIP_TYPE_CHOICES = (\n # a subject-object relationship\n (RELATIONSHIP_TYPE_SUBJECT_OBJECT, _('Subejct-Object')),\n \n # a subject-subject relationship\n (RELATIONSHIP_TYPE_SUBJECT_SUBJECT, _('Subject-Subject')),\n \n # an object-object relationship\n (RELATIONSHIP_TYPE_OBJECT_OBJECT, _('Object-Object')),\n \n # a relationship for recommender where subject domain equals object domain\n (RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT, _('SubjectObject-SubjectObject')),\n)\n\"\"\"Choices for the relationship_type field\"\"\"\n\nFORMAT_STR_SUBJECT = \"subject\"\nFORMAT_STR_OBJECT = \"object\"\nFORMAT_STR_SUBJECT1 = \"subject1\"\nFORMAT_STR_SUBJECT2 = \"subject2\"\nFORMAT_STR_OBJECT1 = \"object1\"\nFORMAT_STR_OBJECT2 = \"object2\"\nFORMAT_STR_SUBJECTOBJECT = 'subjectobject'\nFORMAT_STR_SUBJECTOBJECT1 = \"subjectobject1\"\nFORMAT_STR_SUBJECTOBJECT2 = \"subjectobject2\"\nFORMAT_STR_CLUSTER = \"cluster\"\n\"\"\"Format strings used in the explanation field.\"\"\"\n\nCOMPILATOR_DEPTH_ONE_UNSURE = 1\n\"\"\"One unsure relationship is used [+ predicted_relationship]\"\"\"\n\nCOMPILATOR_DEPTH_COMING_SOON = 2\n\"\"\"Uvidime\"\"\"\n\nDEFAULT_RECOMMENDATION_COUNT = 10\n\"\"\"The defaul count of the obtained recommended objects\"\"\"\n\nDEFAULT_COMPILATOR_BREADTH = 10\n\"\"\"The default neighbourhood size for the compilator\"\"\"\n\nDEFAULT_COMPILATOR_PAIR_DEPTH = 10\n\"\"\"The default compilation element count taken for a pair for a group\"\"\"\n\nDEFAULT_COMPILATOR_DEPTH = COMPILATOR_DEPTH_ONE_UNSURE\n\"\"\"Take only one unsure relationship\"\"\"\n\nUNCERTAIN_PREDICTION_VALUE = 0.5\n\"\"\"The value that is returned when the prediction for the pair isn't \navailable\"\"\"\n\nALREADY_IN_REL_PREDICTION_VALUE = 1.0\n\"\"\"The value that is returned when the pair already is in the \npredicted_relationship\"\"\"\n\nMIN_WEIGHT = 0.0\nMAX_WEIGHT = 1.0\n\"\"\"Weight limits\"\"\"\n\nMIN_CONFIDENCE = 0.0\nMAX_CONFIDENCE = 1.0\n\"\"\"Confidence limits\"\"\"\n\nMIN_EXPECTANCY = 0.0\nMAX_EXPECTANCY = 1.0\n\"\"\"Expectancy limits\"\"\"\n\nTRIVIAL_EXPECTANCY = 1.0\n\"\"\"The expectancy that is given for the trivial cases - already \nlistened to the artist.\"\"\"\n\nCONFIDENCE_KWARG_NAME = 'confidence'\nEXPECTANCY_KWARG_NAME = 'expectancy'\n\nREASON_STR = \"Reason %d\"\n\nLOG_RECOMMENDATIONS_FILENAME = 'recommendations.txt'\nLOG_PREDICTIONS_FILENAME = 'predictions.txt'\nLOG_HITS_FILENAME = 'hits.txt'\n\nMAX_REASONS_DESCRIPTION = 5\n\nMORE_REASONS_STR = \"... and %d other reasons.\"\n\nPROMISING_RATE = 1.2\n\"\"\"The rate by which the compiler breadth is multiplied to get the number\nof promising objects to be inspected\"\"\"\n\nEXP_PRECISION = 0.00001\n"},{"alpha_fraction":0.6530148983001709,"alphanum_fraction":0.6531229615211487,"avg_line_length":32.28776931762695,"blob_id":"18f51e08a7555f5c74b5e0faa02ff606f12983dc","content_id":"fd528e0cb2670f8ca1e5c3a9e6767d53e99cefe4","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":9254,"license_type":"no_license","max_line_length":126,"num_lines":278,"path":"/code/adapter/unresyst/combinator/combination_element.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The elements that are given to combinators to combine them. \nInstancies of the classes represent knowledge that we have about a pair\nto count similarity/preference.\"\"\"\n\nfrom unresyst.constants import *\n\ndef _get_expectancy_positiveness(expectancy):\n return expectancy > UNCERTAIN_PREDICTION_VALUE\n\nclass BaseCombinationElement(object):\n \"\"\"The base for all elements that are being combined in combinator\"\"\"\n \n def __init__(self):\n \"\"\"Initialize the members\"\"\"\n self._positiveness = None\n self._expectancy = None\n self._description = None\n \n def get_positiveness(self):\n \"\"\"Is the element positive/negative to the combination?\"\"\"\n if self._positiveness is None:\n self._positiveness = self._get_positiveness()\n \n return self._positiveness\n \n def get_expectancy(self):\n \"\"\"The expectancy of the element\"\"\"\n \n if self._expectancy is None:\n self._expectancy = self._get_expectancy()\n \n return self._expectancy\n \n def get_description(self):\n \"\"\"The description of the element\"\"\" \n \n if self._description is None:\n self._description = self._get_description()\n \n return self._description\n\n def _get_positiveness(self):\n pass\n\n def _get_expectancy(self):\n pass\n \n def _get_description(self):\n pass\n \n def __repr__(self):\n return \"<%f, %s, %s>\" % (self.get_expectancy(), self.get_positiveness(), self.get_description()) \n\nclass SubjectObjectRelCombinationElement(BaseCombinationElement):\n \"\"\"The relationship between a subject and an object meaning the preference\n in means of the predicted_relationship. \n \n For compilator. \n \"\"\"\n \n def __init__(self, rel_instance):\n \"\"\"The initializer\"\"\"\n \n super(SubjectObjectRelCombinationElement, self).__init__()\n \n self.rel_instance = rel_instance\n\n def _get_positiveness(self):\n return self.rel_instance.definition.as_leaf_class().is_positive\n\n def _get_expectancy(self):\n return self.rel_instance.get_expectancy() \n \n def _get_description(self):\n return self.rel_instance.description \n \n \nclass _SimilarityCombinationElement(BaseCombinationElement):\n \"\"\"The element of a similarity combination\"\"\"\n pass\n \nclass RelSimilarityCombinationElement(_SimilarityCombinationElement):\n \"\"\"The similarity coming out of a rule/relationship\n \n For Aggregator.\n \"\"\"\n \n def __init__(self, rel_instance):\n \"\"\"The initializer\"\"\"\n \n super(RelSimilarityCombinationElement, self).__init__()\n \n self.rel_instance = rel_instance\n \"\"\"The rule/relationship instance the similarity is obtained from\n @type rel_instance: RelationshipInstance\n \"\"\"\n \n def _get_positiveness(self):\n return self.rel_instance.definition.as_leaf_class().is_positive\n\n def _get_expectancy(self):\n return self.rel_instance.get_expectancy() \n \n def _get_description(self):\n return self.rel_instance.description \n\n\nclass ClusterSimilarityCombinationElement(_SimilarityCombinationElement):\n \"\"\"The similarity coming out of a common cluster membership\n\n For combinator and aggregator.\n \"\"\"\n\n def __init__(self, cluster_members):\n \"\"\"The initializer\"\"\"\n \n super(ClusterSimilarityCombinationElement, self).__init__()\n \n self.cluster_members = cluster_members\n \"\"\"The pair of cluster members that caused the similarity\n @type cluster_members: pair ClusterMember, ClusterMember\n \"\"\"\n \n def _get_positiveness(self):\n \"\"\"Cluster membership is always positive\"\"\"\n return True\n \n def _get_expectancy(self):\n \"\"\"Return the product of the reasoning members\"\"\"\n \n return self.cluster_members[0].get_pair_expectancy(cluster_member_pair=self.cluster_members)\n \n def _get_description(self):\n \"\"\"Return the concatenaged member descriptions\"\"\"\n return \" \".join(cm.description for cm in self.cluster_members)\n \nclass BiasCombinationElement(BaseCombinationElement):\n \"\"\"The biase of a subject/object coming to the combination\n For aggregator.\n \"\"\"\n \n def __init__(self, bias_instance):\n \"\"\"The initializer\"\"\"\n \n super(BiasCombinationElement, self).__init__()\n \n self.bias_instance = bias_instance\n \"\"\"Bias instance. \n @type bias_instance: BiasInstance\n \"\"\" \n\n\n def _get_positiveness(self):\n return self.bias_instance.definition.is_positive\n \n def _get_expectancy(self):\n return self.bias_instance.get_expectancy()\n\n def _get_description(self):\n return self.bias_instance.description\n\nclass BiasAggregateCombinationElement(BaseCombinationElement):\n \"\"\"The aggregated bias of a subject/object. \n For compilator.\n \"\"\"\n \n def __init__(self, bias_aggregate):\n \"\"\"The initializer\"\"\"\n \n super(BiasAggregateCombinationElement, self).__init__()\n \n self.bias_aggregate = bias_aggregate\n \"\"\"The aggregated bias.\n @type bias_aggregate: AggregatedBiasInstance\n \"\"\"\n \n def _get_positiveness(self): \n exp = self.get_expectancy()\n return _get_expectancy_positiveness(exp)\n \n def _get_expectancy(self):\n return self.bias_aggregate.expectancy\n \n def _get_description(self):\n return self.bias_aggregate.description\n\n\nclass _PredictedPlusSimilarityCombinationElement(BaseCombinationElement):\n \"\"\"A sign of preference made of known predicted relationship plus similarity.\n For compilator.\n \"\"\"\n \n def __init__(self, predicted_rel, similarity_aggregate):\n \"\"\"The initializer\"\"\"\n \n super(_PredictedPlusSimilarityCombinationElement, self).__init__()\n \n self.predicted_rel = predicted_rel\n \"\"\"The predicted relationship to/from some other object/subject\n @type predicted_rel: RelationshipInstance binded to PredictedRelationshipDefinition\n \"\"\"\n \n self.similarity_aggregate = similarity_aggregate\n \"\"\"The aggregated similarity between the subjects/objects\n @type similarity_aggregate: AggregatedRelationshipInstance\n \"\"\"\n\n def _get_positiveness(self): \n exp = self.get_expectancy()\n return _get_expectancy_positiveness(exp)\n \n def _get_expectancy(self):\n return self.similarity_aggregate.expectancy\n \n\n\nclass PredictedPlusObjectSimilarityCombinationElement(_PredictedPlusSimilarityCombinationElement):\n \"\"\"Predicted relationship plus similarity of objects.\n For compilator.\n \"\"\"\n def _get_description(self):\n return \"%s And similarity: %s\" % (self.predicted_rel.description, self.similarity_aggregate.description)\n\nclass PredictedPlusSubjectSimilarityCombinationElement(_PredictedPlusSimilarityCombinationElement):\n \"\"\"Predicted relationship plus similarity of subjects.\n For compilator.\n \"\"\"\n def _get_description(self):\n return \"Similarity: %s And: %s\" % (self.similarity_aggregate.description, self.predicted_rel.description)\n\n\nclass _PredictedPlusClusterMemberCombinationElement(BaseCombinationElement):\n \"\"\"A sign of preference made of known predicted relationship plus similarity\n coming out of two cluster memberships.\n\n For compilator.\n \"\"\"\n\n def __init__(self, predicted_rel, cluster_combination_element):\n \"\"\"The initializer\"\"\"\n \n super(_PredictedPlusClusterMemberCombinationElement, self).__init__()\n \n self.predicted_rel = predicted_rel\n \"\"\"The predicted relationship to/from some other object/subject\n @type predicted_rel: RelationshipInstance binded to PredictedRelationshipDefinition\n \"\"\"\n \n self.cluster_combination_element = cluster_combination_element\n \"\"\"The pair of cluster memeberships meaning the similarity. In the order as if coming\n from the subject to the object.\n @type cluster_membership_pair: ClusterSimilarityCombinationElement\n \"\"\" \n \n \n def _get_positiveness(self): \n return True\n \n def _get_expectancy(self):\n return self.cluster_combination_element.get_expectancy()\n \n \n\nclass PredictedPlusObjectClusterMemberCombinationElement(_PredictedPlusClusterMemberCombinationElement):\n \"\"\"Predicted relationship plus cluster membership of objects.\n\n For compilator.\n \"\"\"\n def _get_description(self):\n return \"%s And similarity: %s\" % (self.predicted_rel.description, self.cluster_combination_element.get_description())\n\nclass PredictedPlusSubjectClusterMemberCombinationElement(_PredictedPlusClusterMemberCombinationElement):\n \"\"\"Predicted relationship plus cluster membership of subjects.\n\n For compilator.\n \"\"\"\n def _get_description(self):\n return \"Similarity: %s And: %s\" % (self.cluster_combination_element.get_description(), self.predicted_rel.description)\n"},{"alpha_fraction":0.6421499252319336,"alphanum_fraction":0.6584158539772034,"avg_line_length":23.379310607910156,"blob_id":"41a1d8c02a6c91077ddaa1edff6bc895d5cd9457","content_id":"56b5bc0bc1ec5156785f4f3dbcdcdd87f27ddaca","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1414,"license_type":"no_license","max_line_length":63,"num_lines":58,"path":"/UnresystCD/code/adapter/lastfm/evaluation.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The evaluators for the last.fm recommenders\"\"\"\n\nfrom unresyst.recommender.evaluation import BaseEvaluator\nfrom unresyst.recommender.rank_evaluation import RankEvaluator\nfrom models import ArtistEvalPair, NovelArtistEvalPair\nfrom unresyst.recommender.metrics import rmse, precision_recall\n\nclass ArtistRecommenderEvaluator(BaseEvaluator):\n \"\"\"The evaluator of the artist recommender\"\"\"\n \n EvaluationPairModel = ArtistEvalPair\n \"\"\"The model - pairs\"\"\"\n \n prediction_metric = rmse\n \"\"\"The metric\"\"\"\n \n recommendation_metric = precision_recall\n \"\"\"The other metric\"\"\"\n \nclass ArtistRankEvaluator(RankEvaluator):\n \"\"\"Evaluating the rank metric\"\"\"\n \n EvaluationPairModel = ArtistEvalPair\n \"\"\"The model - pairs\"\"\"\n \n SUBJ_IDS = [\n 6L,\n 11L,\n 14L,\n 36L,\n 38L,\n 53L,\n 55L,\n 59L,\n 61L,\n 79L,\n 81L,\n 90L,\n ]\n\n \nclass NovelArtistRecommenderEvaluator(BaseEvaluator):\n \"\"\"The evaluator of the novel artist recommender\"\"\"\n \n EvaluationPairModel = NovelArtistEvalPair\n \n prediction_metric = rmse\n \n recommendation_metric = precision_recall \n\n \nclass NovelArtistRankEvaluator(RankEvaluator):\n \"\"\"Evaluation of the rank metric\"\"\"\n \n EvaluationPairModel = NovelArtistEvalPair\n \"\"\"The model - pairs\"\"\"\n \n SUBJ_IDS = ArtistRankEvaluator.SUBJ_IDS\n"},{"alpha_fraction":0.6785714030265808,"alphanum_fraction":0.680672287940979,"avg_line_length":33,"blob_id":"00c30c4627da8a8eb34e37fc4c92ac1e3c5d7d6e","content_id":"9f3810ae018f50fc898187da60be831f6161be69","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":476,"license_type":"no_license","max_line_length":92,"num_lines":14,"path":"/code/adapter/unresyst/combinator/function_combinator.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Combinator using the magic function.\"\"\"\n\nfrom base import BaseCombinator\n\nclass FunctionCombinator(BaseCombinator):\n \"\"\"A combinator using a special function to combine.\n \"\"\"\n \n def _combine(self, combination_elements, ResultClass):\n \"\"\"See the base class for documentation\"\"\"\n \n # narvat to do magicke funkce a hotovka (spocitat prumer a pocet positive, negative)\n res = ResultClass(expectancy=1, description='')\n return res\n"},{"alpha_fraction":0.48442137241363525,"alphanum_fraction":0.48961424827575684,"avg_line_length":33.89610290527344,"blob_id":"ba5938caabc8a4048c33761bc89052173ba4d996","content_id":"f3491924fcd8f3deade3395e3f275f7523756234","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2696,"license_type":"no_license","max_line_length":142,"num_lines":77,"path":"/UnresystCD/code/adapter/flixster/recbckp.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":" relationships = (\n # users in friendship\n SubjectSimilarityRelationship(\n name=\"Users are friends.\",\n \n generator=lambda: [(f.friend1, f.friend2) for f in Friend.objects.all()], \n \n is_positive=True, \n \n weight=0.5, \n \n description=\"Users %(subject1)s and %(subject2)s are friends.\",\n ),\n )\n \"\"\"The relationships\"\"\"\n \n biases = (\n # people giving high ratings\n SubjectBias(\n name=\"Users giving high ratings.\",\n \n description=\"User %(subject)s gives high ratings.\",\n \n weight=0.5, \n \n is_positive=True,\n \n generator=lambda: User.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__gt=str(MIN_HIGH_RATING)), \n \n confidence=lambda user: user.rating_set.aggregate(Avg('rating'))['rating__avg'] - MIN_HIGH_RATING\n ),\n \n # highly rated movies\n ObjectBias(\n name=\"High-rated movies.\",\n \n description=\"Movie %(object)s is high-rated\",\n \n weight=0.5,\n \n is_positive=True,\n \n generator=lambda: Movie.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__gt=str(MIN_HIGH_RATING)),\n \n confidence=lambda movie: movie.rating_set.aggregate(Avg('rating'))['rating__avg'] - MIN_HIGH_RATING\n ),\n \n # people giving low ratings\n SubjectBias(\n name=\"Users giving low ratings.\",\n \n description=\"User %(subject)s gives low ratings.\",\n \n weight=0.5, \n \n is_positive=False,\n \n generator=lambda: User.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__lt=str(MAX_LOW_RATING)), \n \n confidence=lambda user: MAX_LOW_RATING - user.rating_set.aggregate(Avg('rating'))['rating__avg'] \n ),\n \n # low-rated movies\n ObjectBias(\n name=\"Low-rated movies.\",\n \n description=\"Movie %(object)s is low-rated\",\n \n weight=0.5,\n \n is_positive=False,\n \n generator=lambda: Movie.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__lt=str(MAX_LOW_RATING)),\n \n confidence=lambda movie: MAX_LOW_RATING - movie.rating_set.aggregate(Avg('rating'))['rating__avg']\n ), \n )\n \n"},{"alpha_fraction":0.737500011920929,"alphanum_fraction":0.737500011920929,"avg_line_length":25.66666603088379,"blob_id":"c3752c1c077960e0f5b803ab01e63cf04f6a1dae","content_id":"9faa7d4d0922076951d8647a2b6aad672ceadc57","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":80,"license_type":"no_license","max_line_length":54,"num_lines":3,"path":"/UnresystCD/code/adapter/lastfm/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Application holding the data and a recommender for \nthe Last.fm dataset.\n\"\"\"\n"},{"alpha_fraction":0.5477624535560608,"alphanum_fraction":0.5494836568832397,"avg_line_length":34.72093200683594,"blob_id":"3edd44a8d58f005f8feca242a252d6b64be52355","content_id":"5e27c0fec155d93316bdf2840109384444a73d1f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":4648,"license_type":"no_license","max_line_length":91,"num_lines":129,"path":"/code/adapter/unresyst/recommender/bias.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Classes to represent bias of subjects and objects\"\"\"\n\nfrom unresyst.constants import *\n\nfrom unresyst.models.abstractor import BiasDefinition, BiasInstance\nfrom unresyst.models.common import SubjectObject\n\nclass _BaseBias(object):\n \"\"\"The base class for all bias clases\"\"\"\n \n entity_type = None\n \"\"\"The entity type the bias is for s/o/so\"\"\"\n \n format_string = None\n \"\"\"The format string appearing in the description\"\"\"\n \n def __init__(self, name, generator, is_positive, confidence, weight, description=None):\n \"\"\"The constructor.\"\"\"\n \n self.name = name\n \"\"\"The name of the bias.\"\"\" \n \n self.description = description\n \"\"\"A string describing the rule. It can contain placeholders for entities: \n \n - %(subject)s for subject bias\n - %(object)s for object bias\n - %(subjectobject)s for subjectobject bias\n \"\"\"\n \n self.is_positive = is_positive\n \"\"\"Is the bias positive to the predicted relationship?\"\"\"\n \n self.generator = generator\n \"\"\"A generator returning subjects/objects that are affected by the bias.\n \"\"\" \n \n self.weight = weight\n \"\"\"A float number from [0, 1] representing the *static* weight of the bias\n It doesn't depend on the entity.\n \"\"\"\n \n self.confidence = confidence\n \"\"\"A float function giving values from [0, 1] representing the \n the confidence of the bias of the entity. \n It's dynamic, depends on the entity.\n \"\"\" \n \n def evaluate(self):\n \"\"\"Crate bias definitions and the instances in the database.\n \"\"\"\n \n if not (MIN_WEIGHT <= self.weight <= MAX_WEIGHT):\n raise ConfigurationError(\n message=(\"The bias '%s' provides weight %f,\" + \n \" should be between 0 and 1. .\"\n ) % (self.name, self.weight),\n recommender=self.recommender,\n parameter_name=\"Recommender.biases\",\n parameter_value=(self.recommender.biases)\n )\n \n recommender_model = self.recommender._get_recommender_model()\n\n # create the definition in the database\n definition = BiasDefinition.objects.create(\n name=self.name,\n recommender=recommender_model,\n entity_type=self.entity_type,\n weight=self.weight,\n is_positive=self.is_positive\n ) \n \n # go through the affected entities create bias instances\n # \n for ds_entity in self.generator():\n \n # convert the entity to universal\n dn_entity = SubjectObject.get_domain_neutral_entity(\n domain_specific_entity=ds_entity, \n entity_type=self.entity_type, \n recommender=recommender_model)\n \n # count the confidence by the provided function \n confidence = self.confidence(ds_entity) \n \n # if confidence invalid through an error\n if not (MIN_CONFIDENCE <= confidence <= MAX_CONFIDENCE):\n raise ConfigurationError(\n message=(\"The bias '%s' provides confidence %f,\" + \n \" should be between 0 and 1. \"\n ) % (self.name, confidence),\n recommender=self.recommender,\n parameter_name=\"Recommender.biases\",\n parameter_value=(self.recommender.biases)\n )\n \n # fill the description\n description = self.description % {self.format_string: dn_entity.name}\n \n # create the instance\n BiasInstance.objects.create(\n subject_object=dn_entity,\n confidence=confidence,\n definition=definition,\n description=description\n )\n \n print \" %d bias instances for bias %s created.\" % \\\n (BiasInstance.objects.filter(definition=definition).count(), self.name)\n \nclass SubjectBias(_BaseBias):\n\n entity_type = ENTITY_TYPE_SUBJECT\n \n format_string = FORMAT_STR_SUBJECT\n\n \nclass ObjectBias(_BaseBias):\n\n entity_type = ENTITY_TYPE_OBJECT\n \n format_string = FORMAT_STR_OBJECT\n\nclass SubjectObjectBias(_BaseBias):\n\n entity_type = ENTITY_TYPE_SUBJECTOBJECT\n \n format_string = FORMAT_STR_SUBJECTOBJECT\n \n \n \n \n \n \n \n \n"},{"alpha_fraction":0.7523809671401978,"alphanum_fraction":0.7523809671401978,"avg_line_length":34,"blob_id":"a6e704907d6452c58a8ebf2456f216f50160faf4","content_id":"52b390f506337b8b7fef12287b254a20f114033b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":105,"license_type":"no_license","max_line_length":65,"num_lines":3,"path":"/code/adapter/demo/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The demo application showing the possibilities of Unresyst on \na simple example of a shoe e-shop.\n\"\"\"\n"},{"alpha_fraction":0.5544632077217102,"alphanum_fraction":0.5560361742973328,"avg_line_length":30.33333396911621,"blob_id":"5ebfcc473dcb3e61526b2323283908dc81c385ad","content_id":"e9e5c5a3660dc082cb5cf53b0b46bc019c456c48","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2543,"license_type":"no_license","max_line_length":91,"num_lines":81,"path":"/code/adapter/unresyst/models/evaluation.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Models related to recommender evaluation\"\"\"\n\nfrom django.db import models\n\nclass BaseEvaluationPair(models.Model):\n \"\"\"An abstract base class for a test set pair. To be used for\n testing the correctness of the predicted expectancy.\n \n The subclass has to have attributes:\n - subj - a foreign key to subject\n - obj - a foreign key to object \n \"\"\" \n \n obtained_expectancy = models.FloatField(null=True)\n \"\"\"The predicted expectancy obtained by the predict_relationship recommender\n method.\n \"\"\"\n \n expected_expectancy = models.FloatField()\n \"\"\"The expected expectancy value for the pair\"\"\"\n \n is_successful = models.NullBooleanField(null=True)\n \"\"\"A sign whether the pair prediction was successful\"\"\"\n \n class Meta:\n abstract = True\n app_label = 'unresyst'\n\n def __unicode__(self):\n \"\"\"Return a textual representation.\"\"\" \n \n return u\"%s - %s: exp: %f, got: %f\" % \\\n (self.subj, self.obj, self.expected_expectancy, self.obtained_expectancy or -1)\n \n @classmethod\n def export(cls, f):\n \"\"\"Export evaluation pairs to a csv file of the given name.\n \n @type f: file\n @param f: open file to write to \n \"\"\" \n i = 0\n f.write(\"# userId, itemId\\n\")\n \n # loop through the pairs, \n for subj_id, obj_id in cls.objects.values_list('subj__pk', 'obj__pk'):\n \n # create the line\n linestr = \"%s,%s\\n\" % (subj_id, obj_id) \n \n # write it to the file\n f.write(linestr)\n \n i += 1\n \n print \" %d evaluation pairs exported\" % i\n \n # to be implemented by subclasses\n # \n \n @classmethod\n def select(cls, i=0):\n \"\"\"Select the pairs for a validation, save them to the database and\n remove them from the system data. The method is called in every \n iteration.\n \n To be implemented by the subclass.\n\n @type i: int\n @param i: the number of iteration of the validation (useful when\n cross-validation is done)\n \"\"\"\n raise NotImplementedError() \n\n def get_prediction_success(self):\n \"\"\"Count whether the pair was successful. The obtained expectancy has\n to be filled first.\n \n To be implemented by the subclass.\n \"\"\" \n raise NotImplementedError()\n \n"},{"alpha_fraction":0.5600860714912415,"alphanum_fraction":0.5629558563232422,"avg_line_length":39.66829299926758,"blob_id":"3444088d69f1d0c78a17e8b5561cba93a6144850","content_id":"a1a2fb041241abef63a71ce3ab40ccce1c4753de","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":8363,"license_type":"no_license","max_line_length":96,"num_lines":205,"path":"/code/adapter/unresyst/aggregator/linear_aggregator.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The classes of the aggregator package:\n - LinearAggregator: everything is combined linearly\n - CombiningAggregator: uses the Combinator class for aggregating\n\"\"\"\n\nfrom django.db.models import Count\nfrom django.db.models import Q\n\n\nfrom base import BaseAggregator\nfrom unresyst.models.abstractor import RelationshipInstance, \\\n PredictedRelationshipDefinition\nfrom unresyst.models.aggregator import AggregatedRelationshipInstance, \\\n AggregatedBiasInstance\nfrom unresyst.models.common import SubjectObject\nfrom unresyst.exceptions import InvalidParameterError\n\nclass LinearAggregator(BaseAggregator):\n \"\"\"The class aggregating rule/relationship instances to one for each pair\n of entities. \n \n It ignores the predicted_relationship instances.\n \n A better aggregator could add the positive and subtract the negative \n expectances somehow.\n \"\"\"\n\n def aggregate_rules_relationships(self, recommender_model):\n \"\"\"For documentation see the base class.\n \n Linearly combines the rule and relationship instances.\n \n The descriptions of the aggregates are made by joining the descriptions\n of the aggregated instances. The descriptions are ordered by the expectancy\n of their owners, the highest expectancy comes first.\n \"\"\"\n \n # if there's something in the database for the recommender\n # throw an error\n if AggregatedRelationshipInstance.objects\\\n .filter(recommender=recommender_model).exists():\n \n raise InvalidParameterError(\n message=\"There're unexpected aggregated instances for the recommender.\", \n recommender=recomender_model,\n parameter_name=\"recommender_model\", \n parameter_value=recommender_model)\n\n # aggregate it\n # \n \n # take all rule/relationship instances, that don't belong \n # to the predicted_relationship\n # order them by the first and the second\n predicted_def = PredictedRelationshipDefinition.objects.get(\n recommender=recommender_model)\n instance_qs = RelationshipInstance.objects\\\n .exclude(definition=predicted_def)\\\n .filter(definition__recommender=recommender_model)\\\n .order_by('subject_object1__id', 'subject_object2__id')\n \n # if there's nothing to aggregate, schluss\n if not instance_qs:\n return\n \n first_inst = instance_qs[0]\n \n # continuously built aggregated instance\n # initialize it wit the first instance\n cont_inst = AggregatedRelationshipInstance(\n subject_object1=first_inst.subject_object1,\n subject_object2=first_inst.subject_object2, \n relationship_type=first_inst.definition.as_leaf_class().relationship_type,\n recommender=recommender_model)\n\n exp = first_inst.get_expectancy()\n exp_sum = exp\n count = 1 \n # a list of pairs (expectancy, description) \n desc_list = [(exp, first_inst.description), ]\n \n # go through the rel instances \n for instance in instance_qs.exclude(pk=first_inst.pk).iterator(): \n \n # if the pair has chnged from the last pair, save what we've got and \n # start anew\n if cont_inst.subject_object1 <> instance.subject_object1 \\\n or cont_inst.subject_object2 <> instance.subject_object2:\n \n # count the average expectancy\n cont_inst.expectancy = float(exp_sum) / count\n \n # sort the description list by expectancy and join it\n desc_list.sort(key=lambda pair: pair[0], reverse=True) \n cont_inst.description = ' '.join([desc for x, desc in desc_list])\n\n # save the current instance\n cont_inst.save()\n \n # start a new continuously aggregated instance\n cont_inst = AggregatedRelationshipInstance(\n subject_object1=instance.subject_object1,\n subject_object2=instance.subject_object2, \n relationship_type=instance.definition.as_leaf_class().relationship_type,\n recommender=recommender_model)\n \n exp = instance.get_expectancy()\n exp_sum = exp\n count = 1 \n \n # a list of pairs (expectancy, description) \n desc_list = [(exp, instance.description), ]\n \n # otherwise aggregate \n else: \n exp = instance.get_expectancy() \n exp_sum += exp\n count += 1\n \n desc_list.append((exp, instance.description))\n \n \n # count and save the last we have\n # \n \n # count the average expectancy\n cont_inst.expectancy = float(exp_sum) / count\n \n # sort the description list by expectancy and join it\n desc_list.sort(key=lambda pair: pair[0], reverse=True) \n cont_inst.description = ' '.join([desc for x, desc in desc_list])\n\n # save the last instance\n cont_inst.save()\n \n print \" %d rule/relationship aggregates created\" % \\\n AggregatedRelationshipInstance.objects.filter(recommender=recommender_model).count()\n \n\n def aggregate_biases(self, recommender_model):\n \"\"\"For documentation see the base class.\n \n Linearly combines the bias instances.\n \n The descriptions of the aggregates are made by joining the descriptions\n of the aggregated instances. The descriptions are ordered by the expectancy\n of their owners, the highest expectancy comes first.\n \"\"\"\n \n # if there's something in the database for the recommender\n # throw an error\n if AggregatedBiasInstance.objects\\\n .filter(recommender=recommender_model).exists():\n \n raise InvalidParameterError(\n message=\"There're unexpected aggregated instances for the recommender.\", \n recommender=recommender_model,\n parameter_name=\"recommender_model\", \n parameter_value=recommender_model)\n\n # aggregate it\n # \n \n # take all subjectobjects that have some biases\n qs_biased_so = SubjectObject.objects\\\n .filter(recommender=recommender_model)\\\n .annotate(num_bias=Count('biasinstance'))\\\n .filter(num_bias__gt=0)\n \n count = 0\n \n # go through the biased subjectobjects\n for so in qs_biased_so.iterator(): \n \n # a list of pairs (expectancy, description) \n desc_list = [] \n \n exp_sum = 0\n # go through their biases\n for bias in so.biasinstance_set.all():\n\n # count the expectancy for the bias\n exp = bias.get_expectancy()\n \n exp_sum += exp\n \n desc_list.append((exp, bias.description))\n\n # count the average expectancy\n avg_exp = float(exp_sum) / so.biasinstance_set.count()\n \n # sort the description list by expectancy and join it\n desc_list.sort(key=lambda pair: pair[0], reverse=True) \n desc = ' '.join([d for x, d in desc_list])\n \n # create and save the model\n AggregatedBiasInstance.objects.create(\n expectancy=avg_exp,\n subject_object=so,\n recommender=recommender_model,\n description=desc\n )\n \n print \" %d bias aggregates created\" % \\\n AggregatedBiasInstance.objects.filter(recommender=recommender_model).count()\n \n\n"},{"alpha_fraction":0.5359686017036438,"alphanum_fraction":0.5438196063041687,"avg_line_length":32.28571319580078,"blob_id":"f28cfdf2ed5c979ea0bbe9bd739c6a0fb1f2cf27","content_id":"04fb8d402910bbd2313c1f781bc9bdc4d8f0beb4","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":10954,"license_type":"no_license","max_line_length":133,"num_lines":329,"path":"/UnresystCD/code/adapter/demo/recommender.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The recommender used in the parent system \n(the unresyst.BaseRecommender subclass)\n\"\"\"\nfrom django.db.models import Count\n\nfrom unresyst import *\nfrom models import * \nfrom unresyst.algorithm import *\nfrom unresyst.compilator import *\nfrom unresyst.aggregator import *\nfrom unresyst.combinator import *\n\n# helper functions: \n#\n \ndef _get_keyword_set(o):\n \"\"\"Get set of keywords for the given shoepair\n \"\"\"\n return set([kw.word for kw in o.keywords.all()]) \n\ndef _keyword_set_similarity(o1, o2):\n \"\"\"A function measuring similarity of keywords for shoepairs o1 and o2.\n \n Counted as the size of the intersection divided by the size of the \n smaller set of keywords.\n \"\"\"\n # make sets of the keywords\n keyword_set1 = _get_keyword_set(o1)\n keyword_set2 = _get_keyword_set(o2)\n \n # the intersection of the keywords\n keyword_intersection = keyword_set1.intersection(keyword_set2)\n \n # size of the smaller set\n min_len = min(len(keyword_set1), len(keyword_set2))\n \n # they have some keywords in common, so min_len is never 0\n \n #the final measure\n return float(len(keyword_intersection))/min_len\n\ndef _likes_shoes_generator():\n \"\"\"A generator for the user likes shoes relationship\"\"\" \n for u in User.objects.iterator():\n for s in u.likes_shoes.iterator():\n yield (u, s)\n\ndef _shoelikers():\n \"\"\"A generator returning users liking at least one shoepair\"\"\"\n for u in User.objects.annotate(num_shoes=Count('likes_shoes')).filter(num_shoes__gte=1):\n yield u\n\n\ndef _popular_shoes():\n \"\"\"A generator returning shoes that are liked by at least one user\"\"\"\n for s in ShoePair.objects.annotate(num_likers=Count('likers')).filter(num_likers__gte=1):\n yield s\n\ndef _viewed_shoes_generator():\n \"\"\"A generator for the user has viewed shoes relationship\"\"\" \n for u in User.objects.iterator():\n for s in u.viewed_shoes.iterator():\n yield (u, s)\n\ndef _south_generator():\n \"\"\"A generator for the rule: if user is from south don't recommend \n him/her winter shoes.\n \"\"\"\n for u in User.objects.filter(home_city__in_south=True).iterator():\n for s in ShoePair.objects.filter(for_winter=True).iterator():\n yield (u, s)\n\n# the recommender:\n# \n\nclass ShoeRecommender(Recommender):\n \"\"\"A BaseRecommender subclass holding all domain-specific data\"\"\"\n\n name = \"Shoe Recommender\"\n \"\"\"The name\"\"\" \n \n subjects = User.objects\n \"\"\"The objects to who the recommender will recommend.\"\"\"\n \n objects = ShoePair.objects\n \"\"\"The objects that will be recommended.\"\"\" \n\n predicted_relationship = PredictedRelationship( \n name=\"User likes shoes.\",\n condition=lambda s, o: \n o in s.likes_shoes.all(), \n description=\"\"\"User %(subject)s likes shoes %(object)s.\"\"\",\n generator=_likes_shoes_generator\n )\n \"\"\"The relationship that will be predicted\"\"\"\n \n relationships = (\n \n # if the user has viewed the shoes it's a sign of preference\n SubjectObjectRelationship(\n name=\"User has viewed shoes.\",\n \n condition=lambda s, o:\n o in s.viewed_shoes.all(), \n\n is_positive=True, \n \n weight=0.4, \n \n description=\"User %(subject)s has viewed %(object)s.\",\n \n generator=_viewed_shoes_generator\n ),\n \n # if the user is from the same city as the shoe manufacturer, he might like it\n SubjectObjectRelationship(\n name='User lives in the same city as the shoe manufacturer.',\n \n condition=lambda s, o:\n o.manufacturer and s.home_city and o.manufacturer.home_city == s.home_city,\n\n is_positive=True, \n \n weight=0.1, \n \n description=\"User %(subject)s is from the same city as the manufacturer of %(object)s.\"\n ),\n \n # if users live in the same city, they are considered similar\n SubjectSimilarityRelationship(\n name=\"Users live in the same city.\",\n \n condition=lambda s1, s2:\n s1.home_city and s1.home_city == s2.home_city, \n \n is_positive=True, \n \n weight=0.3, \n \n description=\"Users %(subject1)s and %(subject2)s live in the same city.\"\n ),\n \n # if shoes were made by the same manufacturer, they are considered \n # similar\n ObjectSimilarityRelationship(\n name=\"Shoes were made by the same manufacturer.\",\n \n condition=lambda o1, o2:\n o1.manufacturer and o1.manufacturer == o2.manufacturer,\n \n is_positive=True, \n \n weight=0.1,\n \n description=\"Shoes %(object1)s and %(object2)s were made by\" + \\\n \" the same manufacturer.\"\n ) \n )\n \"\"\"Relationships among the subjects and objects in the domain\"\"\"\n \n \n rules = (\n # if user lives on south, don't recommend him winter shoes \n SubjectObjectRule( \n name=\"Don't recommend winter shoes for southern users.\",\n # is the user from a southern city and shoes for winter?\n condition=lambda s, o: \n s.home_city.in_south and o.for_winter, \n \n is_positive=False,\n \n weight=0.85, \n \n confidence=lambda s, o: 1, \n \n description=\"%(subject)s is from south, so %(object)s can't \" + \n \"be recommended to him/her.\",\n\n generator=_south_generator\n ),\n \n # if users are the same age +- year they are similar\n SubjectSimilarityRule(\n name=\"Users with similar age.\",\n \n # both users have given their age\n condition=lambda s1, s2: \n s1.age and s2.age and s1.age -1 <= s2.age <= s2.age + 1,\n \n is_positive=True, \n \n weight=0.2,\n \n # a magic linear confidence function\n confidence=lambda s1, s2: \n 1 - 0.25 * abs(s1.age - s2.age),\n \n description=\"Users %(subject1)s and %(subject2)s are about \" + \n \"the same age.\"\n ),\n \n # if shoes have common keywords, they are similar.\n ObjectSimilarityRule(\n name=\"Shoes with common keywords.\",\n \n # shoes have some common keywords, if both empty, it's false\n condition=lambda o1, o2: \n bool(_get_keyword_set(o1).intersection(_get_keyword_set(o2))),\n \n is_positive=True,\n \n weight=0.4,\n \n # the size of the intersection / the size of the smaller set\n confidence=_keyword_set_similarity,\n \n description=\"The shoe pairs %(object1)s and %(object2)s \" + \n \"share some keywords.\"\n ),\n \n # explicit rating\n ExplicitSubjectObjectRule(\n name=\"Shoe rating.\",\n \n condition=None,\n \n description=\"User %(subject)s has rated %(object)s.\",\n \n # all pairs user, rated shoes\n generator=lambda: [(r.user, r.shoe_pair) for r in ShoeRating.objects.all()],\n \n # the number of stars divided by five\n expectancy=lambda s, o:float(ShoeRating.objects.get(user=s, shoe_pair=o).stars) / 5,\n ),\n )\n \"\"\"Rules that can be applied to the domain\"\"\"\n \n cluster_sets = (\n # shoe category cluster\n ObjectClusterSet(\n\n name=\"Shoe category cluster set.\",\n\n weight=0.3,\n \n filter_entities=ShoePair.objects.filter(category__isnull=False),\n \n get_cluster_confidence_pairs=lambda shoe: ((shoe.category.name, 1),),\n \n description=\"%(object)s belong to the %(cluster)s category.\",\n ),\n \n # searched keywords cluster\n SubjectClusterSet(\n name=\"Keyword search cluster set.\",\n \n weight=0.4,\n \n filter_entities=User.objects.filter(words_searched__isnull=False).distinct(),\n \n # clusters are words, confidence is 1/number of searched words for user\n get_cluster_confidence_pairs=lambda user: [(w.word, 1.0/user.words_searched.count()) for w in user.words_searched.all()],\n \n description=\"%(subject)s has searched for the word %(cluster)s.\",\n ),\n )\n \n biases = (\n # people liking many shoepairs are more likely to like some more\n SubjectBias(\n name=\"Users liking many shoes.\",\n \n description=\"User %(subject)s likes many shoe pairs.\",\n \n weight=0.4, \n \n is_positive=True,\n \n generator=_shoelikers, \n \n confidence=lambda user: float(user.likes_shoes.count())/3\n ),\n \n # multiply liked shoes are more likely to be liked\n ObjectBias(\n name=\"Popular shoes\",\n \n description=\"Shoe pair %(object)s is popular\",\n \n weight=0.8,\n \n is_positive=True,\n \n generator=_popular_shoes,\n \n confidence=lambda shoe: float(shoe.likers.count())/3\n ),\n )\n\n random_recommendation_description = \"Recommending a random shoe pair to the user.\"\n \n algorithm = AggregatingAlgorithm(\n inner_algorithm=CompilingAlgorithm(\n inner_algorithm=SimpleAlgorithm(\n inner_algorithm=None\n ),\n compilator=GetFirstCompilator()\n ),\n aggregator=LinearAggregator()\n )\n \"\"\"The most basic algorithm is used\"\"\"\n\nShoeRecommender.explicit_rating_rule = ShoeRecommender.rules[3]\n\nclass AverageRecommender(ShoeRecommender):\n \n name = \"Advanced shoe recommender\"\n #remove_predicted_from_recommendations = False\n \n algorithm = AggregatingAlgorithm(\n inner_algorithm=CompilingAlgorithm(\n inner_algorithm=SimpleAlgorithm(\n inner_algorithm=None\n ),\n compilator=CombiningCompilator(combinator=AverageCombinator())\n ),\n aggregator=CombiningAggregator(combinator=AverageCombinator())\n )\n \"\"\"The normal algorithm is used\"\"\"\n\n\n\n"},{"alpha_fraction":0.735551655292511,"alphanum_fraction":0.735551655292511,"avg_line_length":30.66666603088379,"blob_id":"c3566ae29c10998605ef4d74844566ae225b99e5","content_id":"c9c8dd13df97c1175d2cf4da2fcb1d6afa953e0e","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":571,"license_type":"no_license","max_line_length":76,"num_lines":18,"path":"/UnresystCD/code/adapter/travel/mahout_recommender.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The representation of the Mahout recommender in Unresyst.\"\"\"\n\nfrom unresyst.recommender.external_recommender import ExternalRecommender\nfrom recommender import OrderTourRecommender\n\nfrom models import *\n\nclass MahoutOrderTourRecommender(ExternalRecommender, OrderTourRecommender):\n \"\"\"An external order tour recommender\"\"\"\n \n name = \"Mahout order tour recommender\"\n \"\"\"The name\"\"\"\n \n subjects = User.objects\n \"\"\"The objects to who the recommender will recommend.\"\"\"\n \n objects = Tour.objects\n \"\"\"The objects that will be recommended.\"\"\" \n"},{"alpha_fraction":0.6030217409133911,"alphanum_fraction":0.6056657433509827,"avg_line_length":31.300613403320312,"blob_id":"66a0c43d7c3b33d128d124f8c9c4d591f58f0069","content_id":"503f619f12d00e51283694f39285e255e14f54a6","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":5295,"license_type":"no_license","max_line_length":93,"num_lines":163,"path":"/code/adapter/unresyst/abstractor/base.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The module defines base class for the abstractor package.\"\"\"\n\nclass BaseAbstractor(object):\n \"\"\"The base (abstract) class for all abstractors. Defines the interface.\"\"\"\n \n # Build phase:\n #\n \n \n def create_subjectobjects(self, recommender_model, subjects, objects):\n \"\"\"Create abstract representation of subjects and objects.\n \n @type recommender_model: models.Recommender\n @param recommender_model: the recommender model instance\n \n @type subjects: django.db.models.manager.Manager\n @param subjects: the manager above the subject model\n \n @type objects: django.db.models.manager.Manager\n @param objects: the manager above the object model\n \n \"\"\"\n pass\n \n \n \n def create_predicted_relationship_instances(self, predicted_relationship):\n \"\"\"Create the instances of the predicted relationship.\n \n Create bindings on subject-object pairs for pairs between which there's\n the predicted relationship.\n \n @type predicted_relationship: rules.Relationship\n @param predicted_relationship: the definition of relationship to be predicted \n \"\"\"\n pass\n\n \n \n def create_relationship_instances(self, relationships):\n \"\"\"Create the instances of relationships relevant for recommendation.\n \n Create bindings on pairs of subject/objects, for pairs between where \n the relationship is. \n \n @type relationships: list of rules._WeightedRelationship subclass instances\n @param relatioships: recommender relationships to be instantiated \n\n @raise ConfigurationError: if the weight of some relatioship is\n outside [0, 1] \n \"\"\"\n pass\n\n \n \n def create_rule_instances(self, rules):\n \"\"\"Create the instances of rules.\n \n Create bindings on pairs of subject/objects, for pairs between where \n the rule applies.\n\n @type rules: list of rules._BaseRule subclass instances\n @param rules: recommender rules to be instantiated\n \n @raise ConfigurationError: if the weight of some rule is outside [0, 1]\n or some confidence function returns a value outside [0, 1].\n \"\"\"\n pass\n\n \n \n \n def create_clusters(self, cluster_sets):\n \"\"\"Create clusters defined in the given sets\n \n Crates the cluster sets in the database, their clusters, bindings \n of subjectobjects to the clusters.\n \n @type cluster_sets: a list of clusters.BaseClusterSet subclass instances\n @param cluster_sets: recommender cluster sets to be evaluated\n \n @raise ConfigurationError: if the weight of some cluster set is outside [0, 1]\n or some cluster member confidence is outside [0, 1].\n \"\"\"\n pass\n \n \n def create_biases(self, biases):\n \"\"\"Create biases defined by the user.\n \n Creates bias definitions and instances in the database.\n \n @type biases: a list of bias._BaseBias instances\n @param biases: biases to be evaluated\n \n @raise ConfigurationError: if the weight of some cluster set is outside [0, 1]\n or some cluster member confidence is outside [0, 1]. \n \"\"\"\n pass\n\n \n # Update phase:\n # \n \n \n def add_subject(self, recommender, subject):\n \"\"\"Add the subject to the abstract subjects, to the relationship and\n rule instances.\n \n @rtype: uvidime\n @return: the changes performed on the rule and relationship instances.\n \"\"\"\n pass\n \n \n def add_object(self, recommender, object_):\n \"\"\"Add the object to the abstract objects, to the relationship and\n rule instances.\n \n @rtype: uvidime\n @return: the changes performed on the rule and relationship instances.\n \"\"\"\n pass\n \n \n def update_subject(self, recommender, subject):\n \"\"\"Update the subject in the abstract subjects, in the relationship and\n rule instances.\n \n @rtype: uvidime\n @return: the changes performed on the rule and relationship instances.\n \"\"\"\n pass\n \n \n def update_object(self, recommender, object_):\n \"\"\"Update the object in the abstract objects, in the relationship and\n rule instances.\n \n @rtype: uvidime\n @return: the changes performed on the rule and relationship instances.\n \"\"\"\n pass \n \n \n def remove_subject(self, recommender, subject):\n \"\"\"Remove the subject from the abstract subjects, its relationships and\n rule instances.\n \n @rtype: uvidime\n @return: the changes performed on the rule and relationship instances.\n \"\"\"\n pass\n \n \n def remove_object(self, recommender, object_):\n \"\"\"Update the object from the abstract objects, its relationship and\n rule instances.\n \n @rtype: uvidime\n @return: the changes performed on the rule and relationship instances.\n \"\"\"\n pass \n \n \n"},{"alpha_fraction":0.5553778409957886,"alphanum_fraction":0.5639743804931641,"avg_line_length":35.445255279541016,"blob_id":"16ac98ed4ffaeb4faf46f8db3777add92cba6db0","content_id":"e4b8c7014c26aa3d937d4bff79aba92a06a557bf","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":5002,"license_type":"no_license","max_line_length":167,"num_lines":137,"path":"/code/adapter/travel/recommender.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The configuration for the travel agency recommender\"\"\"\nfrom django.db.models import Sum, Count, Avg\n\nfrom unresyst import *\n\nfrom models import *\nfrom constants import *\n\ndef _viewed_profile_confidence(u, t):\n viewed_profs = ViewProfile.objects.filter(tour=t, session__user=u)\n view_count = viewed_profs.count()\n avg_duration = viewed_profs.aggregate(Avg('duration'))['duration__avg']\n \n return min((float(view_count)/6) * avg_duration/160, 1.0)\n\n# predicted bude order\n# remove_predicted_from_recommendations = True\n# podpurny budou ty vtipy\n\nclass OrderTourRecommender(Recommender):\n \"\"\"A recommender for suggesting what tour the user should order.\"\"\" \n\n name = \"Order Tour Recommender\"\n \"\"\"The name\"\"\" \n \n subjects = User.objects\n \"\"\"The objects to who the recommender will recommend.\"\"\"\n \n objects = Tour.objects\n \"\"\"The objects that will be recommended.\"\"\" \n\n random_recommendation_description = \"Recommending a random tour to the user.\"\n\n predicted_relationship = PredictedRelationship( \n name=\"User has ordered the tour.\",\n condition=None, \n description=\"\"\"User %(subject)s has ordered %(object)s.\"\"\",\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) for uid, tid in Order.objects.values_list('session__user', 'tour').distinct()), \n )\n \"\"\"The relationship that will be predicted\"\"\"\n\n \n relationships = ()\n \n rules = ( \n # click = sign of preference\n SubjectObjectRule(\n name='User has clicked on something on the tour profile.',\n weight=0.05,\n condition=None,\n is_positive=True,\n description='User %(subject)s has clicked on something on the %(object)s profile.',\n # pairs that user has clicked on the tour\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\n for uid, tid in Click.objects.values_list('session__user', 'tour').distinct()),\n # the average is around 3, so take 1/6. so that 3 points to the middle.\n confidence=lambda u, t: min(float(Click.objects.filter(tour=t, session__user=u).count())/6, 1.0),\n ), \n \n # mouse move .. also a sign of preference\n SubjectObjectRule(\n name='User has moved the mouse on the tour profile.',\n weight=0.1,\n condition=None,\n is_positive=True,\n description='User %(subject)s has moved the mouse on %(object)s.',\n # pairs that user has moved on the tour\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\n for uid, tid in MouseMove.objects.values_list('session__user', 'tour').distinct()),\n confidence=lambda u, t: min(float(MouseMove.objects.filter(tour=t, session__user=u).count())/18, 1.0),\n ),\n \n # view profile \n SubjectObjectRule(\n name='User has viewed the tour profile page.',\n weight=0.1,\n condition=None,\n is_positive=True,\n description='User %(subject)s has viewed %(object)s.',\n # pairs that user has viewed the tour\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\n for uid, tid in ViewProfile.objects.values_list('session__user', 'tour').distinct()),\n # how many times * how long\n confidence=_viewed_profile_confidence\n ),\n\n \n )\n \n\n biases = (\n # multiply viewed tours\n ObjectBias(\n name=\"Most viewed tours.\",\n \n description=\"Tour %(object)s is much viewed\",\n \n weight=0.3,\n \n is_positive=True,\n \n generator=lambda: Tour.objects.annotate(hh=Count('viewprofile')).filter(hh__gt=2).distinct(),\n \n confidence=lambda t: min(float(t.viewprofile_set.count())/4, 1.0)\n ),\n \n \n # multiply mouse moved tours\n ObjectBias(\n name=\"Most mouse moved tours.\",\n \n description=\"Tour %(object)s is often mouse moved.\",\n \n weight=0.2,\n \n is_positive=True,\n \n generator=lambda: Tour.objects.annotate(hh=Count('mousemove')).filter(hh__gt=6).distinct(),\n \n confidence=lambda t: min(float(t.mousemove_set.count())/12, 1.0)\n ),\n # multiply clicked tours\n ObjectBias(\n name=\"Most clicked tours.\",\n \n description=\"Tour %(object)s is often clicked on.\",\n \n weight=0.5,\n \n is_positive=True,\n \n generator=lambda: Tour.objects.annotate(hh=Count('click')).filter(hh__gt=1).distinct(),\n \n confidence=lambda t: min(float(t.click_set.count())/2, 1.0)\n ),\n \n )\n\n\n\n \n\n"},{"alpha_fraction":0.5668229460716248,"alphanum_fraction":0.580405056476593,"avg_line_length":45.32362747192383,"blob_id":"6c9ac1aa037d4d4fa059f3f62df8d209180e2318","content_id":"3a4ab5042f1421a2c4c0f84ee73ebf9d9139d5a8","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":41378,"license_type":"no_license","max_line_length":197,"num_lines":893,"path":"/UnresystCD/code/adapter/unresyst/tests/test_build.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Tests for the building phase.\n\nEach layer has a separate test:\n - recommender\n - abstractor\n - aggregator\n - algorithm \n\nThe tests are always started by running the recommender.build method.\n\"\"\"\n\nfrom nose.tools import eq_, assert_raises, assert_almost_equal\nfrom django.db.models import Q\n\nfrom unresyst import Recommender\nfrom unresyst.models.common import SubjectObject, Recommender as RecommenderModel\nfrom unresyst.models.abstractor import PredictedRelationshipDefinition, \\\n RelationshipInstance, RuleInstance, RuleRelationshipDefinition, ClusterSet, \\\n BiasDefinition, ExplicitRuleDefinition, ExplicitRuleInstance\nfrom unresyst.models.aggregator import AggregatedRelationshipInstance, AggregatedBiasInstance\nfrom unresyst.models.algorithm import RelationshipPredictionInstance \nfrom test_base import TestBuild, TestEntities, DBTestCase, TestBuildAverage\nfrom unresyst.exceptions import ConfigurationError, DescriptionKeyError\nfrom unresyst.recommender.rules import ExplicitSubjectObjectRule\n\nfrom demo.recommender import ShoeRecommender\nfrom demo.models import User, ShoePair\n\nPLACES = 4\n\"\"\"How many places are counted for expectancy accuracy\"\"\"\n\nclass TestRecommender(TestBuild):\n \"\"\"Test case for the recommender class\"\"\"\n \n def test_recommender_created(self):\n \"\"\"Test that the recommender was created during the build\"\"\" \n \n # get it from db\n rec = RecommenderModel.objects.filter(\n class_name=ShoeRecommender.__name__)\n \n # assert there's one \n eq_(rec.count(), 1, \"There should be right one recommender %s.\" \\\n % ShoeRecommender.name)\n \n rec = rec[0]\n \n # assert it has the right name\n eq_(rec.name, ShoeRecommender.name)\n \n # assert the subjects != objects\n eq_(rec.are_subjects_objects, \\\n ShoeRecommender.subjects == ShoeRecommender.objects)\n \n # assert the model is saved in the recommender\n eq_(ShoeRecommender._get_recommender_model(), rec) \n\n def test_cascade_delete(self):\n \"\"\"Test that the rebuild deletes all that should be deleted\"\"\"\n \n # get and delete alice\n a = User.objects.get(name='Alice') \n a.delete()\n \n # build the recommender again\n ShoeRecommender.build()\n \n # assert there's nothing:\n \n # in SubjectObject\n eq_(SubjectObject.objects.filter(name='Alice').count(), 0)\n \n # in RelationshipInstance\n eq_(RelationshipInstance.objects.filter(subject_object1__name='Alice').count(), 0)\n eq_(RelationshipInstance.objects.filter(subject_object2__name='Alice').count(), 0) \n \n # in RuleInstance\n eq_(RuleInstance.objects.filter(subject_object1__name='Alice').count(), 0)\n eq_(RuleInstance.objects.filter(subject_object2__name='Alice').count(), 0)\n \n # in AggregatedRelationshipInstance\n eq_(AggregatedRelationshipInstance.objects.filter(subject_object1__name='Alice').count(), 0)\n eq_(AggregatedRelationshipInstance.objects.filter(subject_object2__name='Alice').count(), 0)\n \n # in RelationshipPredictionInstance\n eq_(RelationshipPredictionInstance.objects.filter(subject_object1__name='Alice').count(), 0)\n eq_(RelationshipPredictionInstance.objects.filter(subject_object2__name='Alice').count(), 0) \n\n\nclass TestAbstractor(TestEntities):\n \"\"\"Testing the abstractor in the build phase\"\"\" \n \n # creating subjectobjects:\n #\n \n def test_create_subjectobjects(self):\n \"\"\"Test creating universal subjects, objects\"\"\"\n \n # get the specific entities\n subjects = User.objects.all()\n objects = ShoePair.objects.all()\n \n universal = SubjectObject.objects.all()\n \n # assert the counts fit\n eq_(subjects.count() + objects.count(), universal.count())\n \n # assert all subjects and objects are there\n for entities, entity_type in ((subjects, 'S'), (objects, 'O')):\n for en in entities:\n\n # the universal representation of entity\n en_un = universal.filter(id_in_specific=en.id, entity_type=entity_type)\n \n # assert there's one \n eq_(en_un.count(), 1, \"There should be right one represenation\" + \\\n \"of %s, there are %d\" % (en, en_un.count()))\n \n en_un = en_un[0] \n \n # assert it has the right properties \n\n # name\n eq_(en_un.name, en.__unicode__())\n \n # recommender\n eq_(en_un.recommender.class_name, ShoeRecommender.__name__)\n\n \n # predicted relationship\n # \n \n def test_create_predicted_relationship_definition(self):\n \"\"\"Test creating the definition of the predicted relationship\"\"\"\n \n # get definitions related to the shoe recommender\n qs = PredictedRelationshipDefinition.remove_subclass_objects().filter(\n recommender=ShoeRecommender._get_recommender_model())\n \n # get how many definitions there are\n # should be one\n eq_(qs.count(), 1)\n \n definition = qs[0]\n \n # test it has the right name\n eq_(definition.name, \"User likes shoes.\")\n\n \n EXPECTED_PREDICTED_RELATIONSHIP_INSTANCES = (\n ('Alice', 'Sneakers', \"User Alice likes shoes Sneakers.\"),\n ('Bob', 'Sneakers', \"User Bob likes shoes Sneakers.\"),\n ('Edgar', 'Rubber Shoes', \"User Edgar likes shoes Rubber Shoes.\"), \n ('Edgar', 'RS 130', \"User Edgar likes shoes RS 130.\"), \n ) \n\n \n def test_create_predicted_relationship_instances(self):\n \"\"\"Test creating instances of the predicted relationship.\"\"\"\n \n # get the relationship definition\n definition = PredictedRelationshipDefinition.remove_subclass_objects().get( \n recommender=ShoeRecommender._get_recommender_model())\n \n # get instances of the predicted relationship\n instances = RelationshipInstance.remove_subclass_objects().filter(definition=definition)\n \n eq_(len(self.EXPECTED_PREDICTED_RELATIONSHIP_INSTANCES), instances.count(),\n \"Expected: %d, obtained: %d. Obtained: %s\" % \\\n (len(self.EXPECTED_PREDICTED_RELATIONSHIP_INSTANCES), instances.count(), instances))\n \n for expected_data in self.EXPECTED_PREDICTED_RELATIONSHIP_INSTANCES:\n\n # instances that have either one or another order of the subjobjects\n rel_instance = instances.filter(\n subject_object1=self.universal_entities[expected_data[0]], \n subject_object2=self.universal_entities[expected_data[1]]\n )\n \n # there should be one\n eq_(rel_instance.count(), 1)\n \n instance = rel_instance[0]\n \n # test it has the right description\n eq_(instance.description, expected_data[2]) \n\n\n # relationships:\n #\n\n def test_create_relationship_definitions(self):\n \"\"\"Test creating the definitions of the relationships\"\"\"\n \n self._test_definitions(ShoeRecommender.relationships)\n\n \n EXPECTED_RELATIONSHIP_DICT = {\n # relationship name, list of instances - (entity1, entity2, description)\n # for symmetric relationships the description is a pair - both orderings\n \"User has viewed shoes.\": (\n ('Alice', 'Rubber Shoes', \"User Alice has viewed Rubber Shoes.\"),\n ('Bob', 'Sneakers', \"User Bob has viewed Sneakers.\"),\n ('Cindy', 'Rubber Shoes', \"User Cindy has viewed Rubber Shoes.\"),\n ),\n \"Users live in the same city.\": (\n ('Alice', 'Bob', (\"Users Alice and Bob live in the same city.\",\n \"Users Bob and Alice live in the same city.\")),\n ('Cindy', 'Daisy', (\"Users Cindy and Daisy live in the same city.\",\n \"Users Daisy and Cindy live in the same city.\")),\n ),\n \"Shoes were made by the same manufacturer.\": (\n ('Sneakers', 'Rubber Shoes', \n (\"Shoes Sneakers and Rubber Shoes were made by the same manufacturer.\",\n \"Shoes Rubber Shoes and Sneakers were made by the same manufacturer.\")),\n ),\n 'User lives in the same city as the shoe manufacturer.': (\n ('Alice', 'Rubber Shoes', \"User Alice is from the same city as the manufacturer of Rubber Shoes.\"),\n ('Bob', 'Rubber Shoes', \"User Bob is from the same city as the manufacturer of Rubber Shoes.\"),\n ('Alice', 'Sneakers', \"User Alice is from the same city as the manufacturer of Sneakers.\"),\n ('Bob', 'Sneakers', \"User Bob is from the same city as the manufacturer of Sneakers.\"),\n ('Cindy', 'RS 130', \"User Cindy is from the same city as the manufacturer of RS 130.\"),\n ('Daisy', 'RS 130', \"User Daisy is from the same city as the manufacturer of RS 130.\")\n ),\n }\n \n \n def test_create_relationship_instances(self):\n \"\"\"Test creating instances of all the defined relationships.\"\"\"\n \n self._test_instances(\n def_list=ShoeRecommender.relationships,\n expected_dict=self.EXPECTED_RELATIONSHIP_DICT,\n instance_manager=RelationshipInstance.objects) \n \n # rules:\n #\n \n def test_create_rule_definitions(self):\n \"\"\"Test creating the definitions of the rules\"\"\"\n \n # explicit rules are tested separately\n rules = filter(lambda r: r.__class__ != ExplicitSubjectObjectRule, ShoeRecommender.rules)\n self._test_definitions(rules)\n\n\n EXPECTED_RULE_DICT = {\n # rule name, list of instances - (entity1, entity2, description, expectancy)\n # for symmetric rules the description is a pair - both orderings\n \"Don't recommend winter shoes for southern users.\": (\n ('Alice', 'RS 130', \"Alice is from south, so RS 130 can't be recommended to him/her.\", 1),\n ('Bob', 'RS 130', \"Bob is from south, so RS 130 can't be recommended to him/her.\", 1),\n ),\n \"Users with similar age.\": (\n ('Alice', 'Bob', (\"Users Alice and Bob are about the same age.\", \n \"Users Bob and Alice are about the same age.\"), 0.75),\n ('Edgar', 'Fionna', (\"Users Edgar and Fionna are about the same age.\", \n \"Fionna and Edgar are about the same age.\"), 1),\n ),\n \"Shoes with common keywords.\": (\n ('Rubber Shoes', 'Sneakers', (\"The shoe pairs Rubber Shoes and Sneakers share some keywords.\",\n \"The shoe pairs Sneakers and Rubber Shoes share some keywords.\"), 1),\n ('Sneakers', 'Design Shoes', (\"The shoe pairs Sneakers and Design Shoes share some keywords.\",\n \"The shoe pairs Design Shoes and Sneakers share some keywords.\"), 0.5), \n ),\n } \n\n def test_create_rule_instances(self):\n \"\"\"Test creating instances of all the defined rules.\"\"\"\n\n # explicit rules are tested separately\n rules = filter(lambda r: r.__class__ != ExplicitSubjectObjectRule, ShoeRecommender.rules)\n \n self._test_instances(\n def_list=rules,\n expected_dict=self.EXPECTED_RULE_DICT,\n instance_manager=RuleInstance.objects)\n \n # auxiliary functions:\n # \n \n def _test_instances(self, def_list, expected_dict, instance_manager):\n \"\"\"Test creating instances.\n \n @type def_list: list of Rule/Relationship definitions \n @param def_list: a list of definitions to check\n \n @type expected_dict: dict str: tuple\n @param expected_dict: dictionary of expected instance attributes\n \n @type instance_manager: django.db.models.manager.Manager\n @param instance_manager: the manager above rule/relationship instances\n \"\"\"\n # get definitions related to the shoe recommender\n qs = RuleRelationshipDefinition.objects.filter(\n recommender=ShoeRecommender._get_recommender_model()) \n \n # for each defined type there should be a model definition in db\n for r in def_list:\n\n # get the expected data\n expected_relationships = expected_dict[r.name] \n\n # get the definition model by name (shouldn't throw an error)\n definition = qs.get(name=r.name)\n \n instances = instance_manager.filter(definition=definition)\n \n # expect there are as many relationship instances as expected\n eq_(instances.count(), len(expected_relationships))\n \n for expected_data in expected_relationships:\n \n # instances that have either one or another order of the subjobjects\n rel_instance = instances.filter(\n Q(subject_object1=self.universal_entities[expected_data[0]], \n subject_object2=self.universal_entities[expected_data[1]]) | \\\n Q(subject_object1=self.universal_entities[expected_data[1]], \n subject_object2=self.universal_entities[expected_data[0]])\n ) \n \n # there should be one\n eq_(rel_instance.count(), 1)\n \n instance = rel_instance[0]\n \n # assert the right order in the relationship\n if instance.definition.as_leaf_class().relationship_type == 'S-O': \n eq_(instance.subject_object1.entity_type, 'S')\n eq_(instance.subject_object2.entity_type, 'O')\n else:\n assert instance.subject_object1.id < instance.subject_object2.id\n \n # test it has the right description\n # the diferentiation for symmetric relationships - the order \n # of the entities in the description can be arbitrary\n if r.is_symmetric:\n desc_tuple = expected_data[2]\n assert instance.description == desc_tuple[0] or \\\n instance.description == desc_tuple[1], \\\n \"The description '%s' is wrong. Should be '%s' or '%s'\" % \\\n (instance.description, desc_tuple[0], desc_tuple[1])\n else:\n eq_(instance.description, expected_data[2]) \n \n # for rules \n if len(expected_data) == 4:\n eq_(instance.confidence, expected_data[3], \n \"Got confidence: %f, expected: %f for instance: %s\" % (instance.confidence, expected_data[3], instance))\n \n\n def _test_definitions(self, def_list):\n \"\"\"Test the definitions.\n \n @type def_list: an iterable of RuleRelationshipDefinition model\n @param def_list: list of definitions to check.\n \"\"\"\n # get definitions related to the shoe recommender\n qs = RuleRelationshipDefinition.objects.filter(\n recommender=ShoeRecommender._get_recommender_model()) \n \n # explicit rules are tested separately\n rules = filter(lambda r: r.__class__ != ExplicitSubjectObjectRule, ShoeRecommender.rules)\n \n # assert the number of definitions is right\n eq_(qs.count(), len(ShoeRecommender.relationships) + len(rules))\n \n # for each defined type there should be a model definition in db\n for r in def_list:\n\n # get the definition model by name (shouldn't throw an error)\n rmodel = qs.get(name=r.name)\n \n # assert the positiveness, weight and relationship type are right\n eq_(rmodel.is_positive, r.is_positive)\n eq_(rmodel.weight, r.weight) \n eq_(rmodel.relationship_type, r.relationship_type) \n\n\n EXP_CLUSTER_SETS = {\n \"Shoe category cluster set.\": \n (\n {'For Sports': (('RS 130', 1, \"RS 130 belong to the For Sports category.\"), \n (\"Octane SL\", 1, \"Octane SL belong to the For Sports category.\")),\n 'Casual': ((\"Design Shoes\", 1, \"Design Shoes belong to the Casual category.\"), \n (\"Sneakers\", 1, \"Sneakers belong to the Casual category.\")),},\n 'O', 0.3),\n \"Keyword search cluster set.\":\n (\n {'Sporty': (\n ('Bob', 0.5, \"Bob has searched for the word Sporty.\"),\n ),\n 'Comfortable': (\n ('Bob', 0.5, \"Bob has searched for the word Comfortable.\"),\n ('Cindy', 0.5, \"Cindy has searched for the word Comfortable.\"),\n ),\n 'Cool':(\n ('Cindy', 0.5, \"Cindy has searched for the word Cool.\"),\n ),\n },\n 'S', 0.4),\n }\n \"\"\"A dictionary: cluster set name: contained clusters, entity type, weight,\n the members: object name, confidence, description.\"\"\"\n \n def test_clusters(self):\n \"\"\"Test that the clusters are created as expected\"\"\"\n \n # get cluster sets from db\n cluster_sets = ClusterSet.objects.filter(\n recommender=ShoeRecommender._get_recommender_model())\n \n # assert there are as many as expected\n eq_(len(self.EXP_CLUSTER_SETS), cluster_sets.count())\n \n for cs in cluster_sets:\n\n # get the expected contents\n clusters, ent_type, weight = self.EXP_CLUSTER_SETS[cs.name]\n \n eq_(ent_type, cs.entity_type)\n eq_(weight, cs.weight)\n eq_(len(clusters), cs.cluster_set.count())\n \n # go through the clusters\n for c in cs.cluster_set.all():\n\n cluster_members = clusters[c.name]\n \n eq_(len(cluster_members), c.clustermember_set.count())\n \n for cm in c.clustermember_set.all():\n # expect the combination of entity and confidence is \n # in the expected\n assert (cm.member.name, cm.confidence, cm.description) in cluster_members, \\\n \"The member '%s' with confidence '%f' and description '%s' isn't one of the expected %s\" % (\n cm.member, cm.confidence, cm.description, cluster_members)\n\n EXP_BIASES = {\n \"Users liking many shoes.\": \n {\n 'Alice': (0.333333, \"User Alice likes many shoe pairs.\"),\n 'Bob': (0.333333, \"User Bob likes many shoe pairs.\"),\n 'Edgar': (0.66666, \"User Edgar likes many shoe pairs.\"),\n },\n \"Popular shoes\":\n {\n 'Sneakers': (0.6666666, \"Shoe pair Sneakers is popular\"),\n 'Rubber Shoes': (0.333333, \"Shoe pair Rubber Shoes is popular\"),\n 'RS 130': (0.333333, \"Shoe pair RS 130 is popular\"),\n },\n }\n \"\"\"A dictionary: cluster set name: contained clusters, entity type, weight\"\"\" \n \n def test_bias(self):\n \"\"\"Test that biases were created as expected\"\"\"\n \n # get bias definitions from db\n bias_defs = BiasDefinition.objects.filter(\n recommender=ShoeRecommender._get_recommender_model())\n \n # assert there are as many as expected\n eq_(len(self.EXP_BIASES), bias_defs.count())\n \n for bias_def in bias_defs:\n \n # assert it's in the expected\n assert self.EXP_BIASES.has_key(bias_def.name)\n\n # get the expected contents \n exp_biases = self.EXP_BIASES[bias_def.name]\n \n eq_(len(exp_biases), bias_def.biasinstance_set.count(), \n \"Wrong count for definition '%s' should be %s, is %s\" % \\\n (bias_def.name, exp_biases, bias_def.biasinstance_set.all()))\n \n # go through the instances connected to definition\n for bias in bias_def.biasinstance_set.all():\n \n # assert it's in the expected biases\n assert exp_biases.has_key(bias.subject_object.name)\n \n # assert the properties are as expected\n exp_conf, exp_desc = exp_biases[bias.subject_object.name]\n \n assert_almost_equal(exp_conf, bias.confidence, PLACES)\n eq_(exp_desc, bias.description) \n \n EXP_EXPLICIT_RULES = {\n \"Shoe rating.\": \n {\n ('Alice', 'Design Shoes'): (0.8, \"User Alice has rated Design Shoes.\"),\n ('Bob', 'Design Shoes'): (0.2, \"User Bob has rated Design Shoes.\"),\n },\n }\n \n def test_explicit_rules(self):\n \"\"\"Test the explicit rules and their definitions \n were created as expected\"\"\"\n \n rule_defs = ExplicitRuleDefinition.objects.all()\n \n # expect the definition count fits\n eq_(len(self.EXP_EXPLICIT_RULES), rule_defs.count())\n \n for rule_def in rule_defs:\n \n # assert the def is there\n assert self.EXP_EXPLICIT_RULES.has_key(rule_def.name)\n \n # get its expected rules\n exp_rules = self.EXP_EXPLICIT_RULES[rule_def.name]\n \n rules = ExplicitRuleInstance.objects.filter(definition=rule_def)\n \n for rule in rules:\n \n # assert the pair is expected\n assert exp_rules.has_key((rule.subject_object1.name, rule.subject_object2.name)), \\\n \"The pair %s, %s isn't expected\" % (rule.subject_object1.name, rule.subject_object2.name)\n\n exp_rule = exp_rules[(rule.subject_object1.name, rule.subject_object2.name)]\n \n assert_almost_equal(exp_rule[0], rule.expectancy, PLACES)\n eq_(exp_rule[1], rule.description)\n \n \nclass TestAbstractorRecommenderErrors(DBTestCase):\n \"\"\"Test various errors thrown by Abstractor and/or Recommender and/or Algorithm\"\"\"\n\n def test_invalid_relationship_weight(self):\n \"\"\"Test if the exception is raised for a relatioship with invalid weight\"\"\"\n \n # set some invalid weight, assert it throws the error \n w = ShoeRecommender.relationships[1].weight \n ShoeRecommender.relationships[1].weight = 1.5\n\n assert_raises(ConfigurationError, ShoeRecommender.build)\n\n # restore the original value\n ShoeRecommender.relationships[1].weight = w\n\n \n def test_invalid_rule_weight(self):\n \"\"\"Test if the exception is raised for a rule with invalid weight\"\"\"\n \n # set some invalid weight, assert it throws the error \n w = ShoeRecommender.rules[2].weight \n ShoeRecommender.rules[2].weight = 1.5\n\n assert_raises(ConfigurationError, ShoeRecommender.build)\n\n # restore the original value\n ShoeRecommender.rules[2].weight = w \n\n\n def test_invalid_confidence(self):\n \"\"\"Test if the exception is raised for a rule with invalid confidence\"\"\"\n \n # set some invalid weight, assert it throws the error \n c = ShoeRecommender.rules[0].confidence\n ShoeRecommender.rules[0].confidence = lambda a, b: 1.3\n\n assert_raises(ConfigurationError, ShoeRecommender.build)\n\n # restore the original value\n ShoeRecommender.rules[0].confidence = c\n\n def test_empty_predicted_relationship(self):\n \"\"\"Test building a recommender with emtpy predicted relationship\"\"\"\n \n # save and delete the predicted_relationship\n pr = ShoeRecommender.predicted_relationship\n ShoeRecommender.predicted_relationship = None\n \n # assert it raises an error\n assert_raises(ConfigurationError, ShoeRecommender.build)\n \n # restore the original predicted_relationship\n ShoeRecommender.predicted_relationship = pr\n \n def test_empty_subjects(self):\n \"\"\"Test building a recommender with empty subjects\"\"\"\n \n # delete all users\n User.objects.all().delete()\n \n # assert it raises an error\n assert_raises(ConfigurationError, ShoeRecommender.build) \n\n def test_empty_objects(self):\n \"\"\"Test building a recommender with empty objects\"\"\"\n \n # delete all shoes\n ShoePair.objects.all().delete()\n \n # assert it raises an error\n assert_raises(ConfigurationError, ShoeRecommender.build) \n \n def test_rule_description_error(self):\n \"\"\"Test building a recommender with strange rule description\"\"\"\n \n # save and mess up the description\n d = ShoeRecommender.rules[1].description \n ShoeRecommender.rules[1].description = \"\"\"Some messed %(crap)s description\"\"\"\n \n # assert it raises an error\n assert_raises(DescriptionKeyError, ShoeRecommender.build) \n \n # restore the original \n ShoeRecommender.rules[1].description = d\n\n def test_relationship_description_error(self):\n \"\"\"Test building a recommender with strange relationship description\"\"\"\n \n # save and mess up the description\n d = ShoeRecommender.relationships[0].description \n ShoeRecommender.relationships[0].description = \"\"\"Some other messed %(other_crap)s description\"\"\"\n \n # assert it raises an error\n assert_raises(DescriptionKeyError, ShoeRecommender.build) \n \n # restore the original \n ShoeRecommender.relationships[0].description = d\n\n def test_no_key_description_error(self):\n \"\"\"Test building a recommender with a description without keys\"\"\"\n \n # save and mess up the description\n d = ShoeRecommender.rules[2].description \n ShoeRecommender.rules[2].description = \"\"\"No-key description.\"\"\"\n \n # assert it doesn't raise any error\n ShoeRecommender.build()\n \n # restore the original \n ShoeRecommender.rules[2].description = d\n \n def test_empty_rules_rels(self):\n \"\"\"Test building the recommender with empty rules and relationships\"\"\"\n \n # save and delete rules and relationships\n rules = ShoeRecommender.rules\n relationships = ShoeRecommender.relationships\n \n ShoeRecommender.rules = ()\n ShoeRecommender.relationships = ()\n \n # build it, test it doesn't freeze\n ShoeRecommender.build() \n \n # restore\n ShoeRecommender.rules = rules\n ShoeRecommender.relationships = relationships\n \ndef _count_exp(conf):\n return 0.5 + conf/2\n \ndef _count_neg_exp(conf): \n return 0.5 - conf/2\n \nclass TestAggregator(TestEntities):\n \"\"\"Testing the building phase of the Linear Aggregator\"\"\"\n \n EXPECTED_AGGREGATES = {\n # S-O\n ('Alice', 'RS 130'): (_count_neg_exp(0.85), 'S-O', \n (\"Alice is from south, so RS 130 can't be recommended to him/her.\",)), # 0.075\n ('Alice', 'Rubber Shoes'): ((_count_exp(0.4) + _count_exp(0.1))/2, 'S-O',\n ('User Alice has viewed Rubber Shoes. User Alice is from the same city as the manufacturer of Rubber Shoes.',)), # 0.625\n ('Alice', 'Sneakers'): (_count_exp(0.1), 'S-O', \n (\"User Alice is from the same city as the manufacturer of Sneakers.\",)), # 0.55\n \n ('Bob', 'RS 130'): (_count_neg_exp(0.85), 'S-O', \n (\"Bob is from south, so RS 130 can't be recommended to him/her.\",)), # 0.075\n ('Bob', 'Rubber Shoes'): (_count_exp(0.1), 'S-O',\n (\"User Bob is from the same city as the manufacturer of Rubber Shoes.\",)), # 0.55\n ('Bob', 'Sneakers'): ((_count_exp(0.4) + _count_exp(0.1))/2, 'S-O',\n (\"User Bob has viewed Sneakers. User Bob is from the same city as the manufacturer of Sneakers.\",)), # 0.625\n\n ('Cindy', 'RS 130'): (_count_exp(0.1), 'S-O',\n (\"User Cindy is from the same city as the manufacturer of RS 130.\",)), # 0.55\n ('Cindy', 'Rubber Shoes'): (_count_exp(0.4), 'S-O',\n (\"User Cindy has viewed Rubber Shoes.\",)), # 0.7\n \n ('Daisy', 'RS 130'): (_count_exp(0.1), 'S-O',\n (\"User Daisy is from the same city as the manufacturer of RS 130.\",)), # 0.55\n \n # S-S\n ('Alice', 'Bob'): ((_count_exp(0.75 * 0.2) + _count_exp(0.3))/2, 'S-S', (\n \"Users Alice and Bob live in the same city. Users Alice and Bob are about the same age.\",\n \"Users Alice and Bob live in the same city. Users Bob and Alice are about the same age.\",\n \"Users Alice and Bob live in the same city. Users Bob and Alice are about the same age.\",\n \"Users Bob and Alice live in the same city. Users Bob and Alice are about the same age.\")), # 0.6125\n ('Cindy', 'Daisy'): (_count_exp(0.3), 'S-S', (\n \"Users Cindy and Daisy live in the same city.\",\n \"Users Daisy and Cindy live in the same city.\")), #0.65\n ('Edgar', 'Fionna'): (_count_exp(1 * 0.2), 'S-S', (\n \"Users Edgar and Fionna are about the same age.\",\n \"Users Fionna and Edgar are about the same age.\")),\n \n # O-O\n ('Rubber Shoes', 'Sneakers'): ((_count_exp(0.4) + _count_exp(0.1))/2, 'O-O', (\n \"The shoe pairs Rubber Shoes and Sneakers share some keywords. Shoes Sneakers and Rubber Shoes were made by the same manufacturer.\",\n \"The shoe pairs Rubber Shoes and Sneakers share some keywords. Shoes Rubber Shoes and Sneakers were made by the same manufacturer.\",\n \"The shoe pairs Sneakers and Rubber Shoes share some keywords. Shoes Sneakers and Rubber Shoes were made by the same manufacturer.\",\n \"The shoe pairs Sneakers and Rubber Shoes share some keywords. Shoes Rubber Shoes and Sneakers were made by the same manufacturer.\")), # 0.625\n ('Sneakers', 'Design Shoes'): (_count_exp(0.2), 'O-O',(\n \"The shoe pairs Sneakers and Design Shoes share some keywords.\",\n \"The shoe pairs Design Shoes and Sneakers share some keywords.\",)), # 0.6 \n }\n \"\"\"A dictionary: pair of entities : expectancy, entity_type, description\"\"\" \n\n def test_aggregates_created(self):\n \"\"\"Test that the aggregates were created as expected\"\"\"\n \n # filter aggregated instances for my recommender\n aggr_instances = AggregatedRelationshipInstance.objects.filter(\n recommender=self.recommender._get_recommender_model())\n \n # assert it has the expected length \n eq_(aggr_instances.count(), len(self.EXPECTED_AGGREGATES))\n \n for aggr_inst in aggr_instances.iterator():\n pair1 = (aggr_inst.subject_object1.name, aggr_inst.subject_object2.name)\n pair2 = (aggr_inst.subject_object2.name, aggr_inst.subject_object1.name) \n \n # try getting the instance from expected in both directions\n if self.EXPECTED_AGGREGATES.has_key(pair1):\n expected_expectancy, expected_rel_type, expected_descs = self.EXPECTED_AGGREGATES[pair1]\n else:\n if self.EXPECTED_AGGREGATES.has_key(pair2):\n expected_expectancy,expected_rel_type, expected_descs = self.EXPECTED_AGGREGATES[pair2]\n else:\n # if not found it's unexpected.\n assert False, \\\n \"Unexpected aggregate between '%s' and '%s' expectancy: %f\" % \\\n (pair1 + (aggr_inst.expectancy, ))\n \n # assert the order in the aggregate is right\n if aggr_inst.relationship_type == 'S-O':\n assert aggr_inst.subject_object1.entity_type == 'S' and aggr_inst.subject_object2.entity_type == 'O'\n else:\n assert aggr_inst.subject_object1.id < aggr_inst.subject_object2.id\n \n # assert the expectancy is as expected \n assert_almost_equal(aggr_inst.expectancy, expected_expectancy, PLACES,\n \"Expectancy is '%f' should be '%f' for the pair %s, %s, Obtained description: %s\" % \\\n ((aggr_inst.expectancy, expected_expectancy) + pair1 + (aggr_inst.description,))) \n \n # assert the relationship type is as expected \n eq_(aggr_inst.relationship_type, expected_rel_type,\n \"Relationship type is '%s' should be '%s' for the pair %s, %s\" % \\\n ((aggr_inst.relationship_type, expected_rel_type) + pair1)) \n \n # assert the description is as expected \n assert aggr_inst.description in expected_descs, \\\n \"Description is '%s' should be one of '%s' for the pair %s, %s\" % \\\n ((aggr_inst.description, expected_descs) + pair1) \n\n EXP_AGGR_BIASES = {\n 'Alice': (_count_exp(0.4 * 0.333333), \"User Alice likes many shoe pairs.\"),\n 'Bob': (_count_exp(0.4 * 0.333333), \"User Bob likes many shoe pairs.\"),\n 'Edgar': (_count_exp(0.4 * 0.66666), \"User Edgar likes many shoe pairs.\"),\n 'Sneakers': (_count_exp(0.8 * 0.6666666), \"Shoe pair Sneakers is popular\"),\n 'Rubber Shoes': (_count_exp(0.8 * 0.333333), \"Shoe pair Rubber Shoes is popular\"),\n 'RS 130': (_count_exp(0.8 * 0.333333), \"Shoe pair RS 130 is popular\"),\n }\n \n def test_aggregated_biases_created(self):\n \"\"\"Test that the bias aggregates were created as expected\"\"\"\n \n # get the aggregates, check count\n # \n aggregates = AggregatedBiasInstance.objects.filter(recommender=self.recommender._get_recommender_model()) \n eq_(aggregates.count(), len(self.EXP_AGGR_BIASES))\n \n # go through created aggregates, assert its in expected and \n # with expected attrs\n for aggr in aggregates:\n \n assert self.EXP_AGGR_BIASES.has_key(aggr.subject_object.name)\n exp_expectancy, exp_desc = self.EXP_AGGR_BIASES[aggr.subject_object.name]\n \n assert_almost_equal(aggr.expectancy, exp_expectancy, PLACES,\n \"Expected expectancy %f, got %f for %s\" % \\\n (exp_expectancy, aggr.expectancy, aggr.subject_object.name) )\n \n eq_(aggr.description, exp_desc) \n \nclass TestAggregatorAverage(TestBuildAverage): \n \"\"\"Testing the aggregator of the average build\"\"\"\n\n def test_aggregated_biases_created(self):\n \"\"\"AverageRecommender: Test that the bias aggregates were created as expected\"\"\" \n # call the same method on the normal aggregator\n ta = TestAggregator(\"test_aggregated_biases_created\")\n ta.recommender = self.recommender\n ta.test_aggregated_biases_created()\n EXPECTED_AGGREGATES = {\n # S-S\n ('Alice', 'Bob'): ((_count_exp(0.75 * 0.2) + _count_exp(0.3))/2, 'S-S', (\n \"Reason 1: Users Alice and Bob live in the same city. Reason 2: Users Alice and Bob are about the same age.\",\n \"Reason 1: Users Alice and Bob live in the same city. Reason 2: Users Bob and Alice are about the same age.\",\n \"Reason 1: Users Alice and Bob live in the same city. Reason 2: Users Bob and Alice are about the same age.\",\n \"Reason 1: Users Bob and Alice live in the same city. Reason 2: Users Bob and Alice are about the same age.\")), # 0.6125\n ('Cindy', 'Daisy'): (_count_exp(0.3), 'S-S', (\n \"Users Cindy and Daisy live in the same city.\",\n \"Users Daisy and Cindy live in the same city.\")), #0.65\n ('Edgar', 'Fionna'): (_count_exp(1 * 0.2), 'S-S', (\n \"Users Edgar and Fionna are about the same age.\",\n \"Users Fionna and Edgar are about the same age.\")),\n \n # O-O\n ('Rubber Shoes', 'Sneakers'): ((_count_exp(0.4) + _count_exp(0.1))/2, 'O-O', (\n \"Reason 1: The shoe pairs Rubber Shoes and Sneakers share some keywords. Reason 2: Shoes Sneakers and Rubber Shoes were made by the same manufacturer.\",\n \"Reason 1: The shoe pairs Rubber Shoes and Sneakers share some keywords. Reason 2: Shoes Rubber Shoes and Sneakers were made by the same manufacturer.\",\n \"Reason 1: The shoe pairs Sneakers and Rubber Shoes share some keywords. Reason 2: Shoes Sneakers and Rubber Shoes were made by the same manufacturer.\",\n \"Reason 1: The shoe pairs Sneakers and Rubber Shoes share some keywords. Reason 2: Shoes Rubber Shoes and Sneakers were made by the same manufacturer.\")), # 0.625\n ('Sneakers', 'Design Shoes'): (0.625, 'O-O',(\n \"Reason 1: Sneakers belong to the Casual category. Design Shoes belong to the Casual category. Reason 2: The shoe pairs Sneakers and Design Shoes share some keywords.\",\n \"Reason 1: Sneakers belong to the Casual category. Design Shoes belong to the Casual category. Reason 2: The shoe pairs Design Shoes and Sneakers share some keywords.\",\n \"Reason 1: Design Shoes belong to the Casual category. Sneakers belong to the Casual category. Reason 2: The shoe pairs Sneakers and Design Shoes share some keywords.\",\n \"Reason 1: Design Shoes belong to the Casual category. Sneakers belong to the Casual category. Reason 2: The shoe pairs Design Shoes and Sneakers share some keywords.\",)), # 0.6 \n }\n\n def test_aggregates_created(self):\n \"\"\"AverageRecommender: Test that the aggregates were created as expected\"\"\"\n \n # instantiate the common aggregator\n ta = TestAggregator(\"test_aggregates_created\")\n ta.recommender = self.recommender\n\n # leave in the aggregates only similarity relationship \n ta.EXPECTED_AGGREGATES = self.EXPECTED_AGGREGATES\n\n \n # call the same method on the normal aggregator\n ta.test_aggregates_created()\n \n \nclass DTestAlgorithm(TestEntities):\n \"\"\"Testing the building phase of the SimpleAlgorithm\"\"\" \n \n \n EXPECTED_PREDICTIONS = {\n # S-O\n ('Alice', 'RS 130'): _count_neg_exp(0.85), # 0.075\n ('Alice', 'Rubber Shoes'): (_count_exp(0.4) + _count_exp(0.1))/2, # 0.625\n ('Alice', 'Sneakers'): _count_exp(0.1), # 0.55\n ('Alice', 'Design Shoes'): _count_exp(0.2), # 0.6\n \n ('Bob', 'RS 130'): _count_neg_exp(0.85), # 0.075\n ('Bob', 'Rubber Shoes'): _count_exp(0.1), # 0.55\n ('Bob', 'Sneakers'): (_count_exp(0.4) + _count_exp(0.1))/2, # 0.625\n ('Bob', 'Design Shoes'): _count_exp(0.2), # 0.6\n\n ('Cindy', 'RS 130'): _count_exp(0.1), # 0.55\n ('Cindy', 'Rubber Shoes'): _count_exp(0.4), # 0.7 \n \n ('Daisy', 'RS 130'): _count_exp(0.1), # 0.55\n \n ('Edgar', 'Sneakers'): (_count_exp(0.1) + _count_exp(0.4))/2, # 0.625 \n \n ('Fionna', 'Rubber Shoes'): _count_exp(0.2),\n ('Fionna', 'RS 130'): _count_exp(0.2), # 0.6 \n } \n \n def dtest_predictions_created(self):\n \"\"\"Test that the predictions were created as expected\"\"\"\n \n # filter aggregated instances for my recommender\n pred_instances = RelationshipPredictionInstance.objects.filter(\n recommender=ShoeRecommender._get_recommender_model())\n \n # assert it has the expected length \n eq_(pred_instances.count(), len(self.EXPECTED_PREDICTIONS), \n \"expected: %d, obtained: %d, obtained instances: %s\" % \\\n (len(self.EXPECTED_PREDICTIONS), pred_instances.count(), pred_instances))\n \n for pred_inst in pred_instances.iterator():\n pair1 = (pred_inst.subject_object1.name, pred_inst.subject_object2.name) \n \n # try getting the instance from expected in both directions\n assert self.EXPECTED_PREDICTIONS.has_key(pair1), \\\n \"Unexpected prediction between '%s' and '%s' expectancy: %f\" % \\\n (pair1 + (pred_inst.expectancy, ))\n expected_prediction = self.EXPECTED_PREDICTIONS[pair1]\n\n # assert the expectancy is as expected \n assert_almost_equal(pred_inst.expectancy, expected_prediction, PLACES, \n \"Prediction is '%f' should be '%f' for the pair %s, %s\" % \\\n ((pred_inst.expectancy, expected_prediction) + pair1)) \n \n\n"},{"alpha_fraction":0.5023733973503113,"alphanum_fraction":0.5142405033111572,"avg_line_length":30.600000381469727,"blob_id":"100bdb93f6cbb8886489104c5cf6d157b1ba74eb","content_id":"22af5dba242c3bd59bcfac3c2710b10a4c089a63","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1264,"license_type":"no_license","max_line_length":79,"num_lines":40,"path":"/UnresystCD/code/adapter/unresyst/combinator/confidence_factor.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Combinator using confidence factor calculus.\"\"\"\n\nfrom base import BaseCombinator\n\n\n \n\nclass ConfidenceFactorCombinator(BaseCombinator):\n \"\"\"A combinator using weighted average\n \"\"\"\n \n def _combine(self, combination_elements, ResultClass):\n \"\"\"See the base class for documentation\"\"\"\n \n res_exp = combination_elements[0].get_expectancy()\n \n # go through the combination elements\n for ce in combination_elements[1:]:\n \n # convert to confidence factors\n res_cf = 2 * res_exp - 1\n ce_cf = 2 * ce.get_expectancy() -1\n \n # count the confidence factor combination\n if res_cf > 0 and ce_cf > 0: \n comb_cf = res_cf + ce_cf * (1 - res_cf)\n \n elif res_cf < 0 and ce_cf < 0:\n comb_cf = res_cf + ce_cf * (1 + res_cf)\n \n else:\n comb_cf = (res_cf * ce_cf) / (1 - min(abs(res_cf), abs(ce_cf)))\n \n # and back to expectancy\n res_exp = (comb_cf + 1) / 2\n \n\n desc = self._concat_descriptions(combination_elements)\n \n return ResultClass(expectancy=res_exp, description=desc)\n"},{"alpha_fraction":0.5262203812599182,"alphanum_fraction":0.5283953547477722,"avg_line_length":32.290321350097656,"blob_id":"ef402b712baa9bec74f7b91d52c575dc73ac448b","content_id":"fef96f14982817083f7791d2f051bc5be587bcdb","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":8276,"license_type":"no_license","max_line_length":98,"num_lines":248,"path":"/code/adapter/unresyst/recommender/evaluation.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The classes related to recommender evaluation\"\"\"\nimport os.path\nimport codecs\n\nfrom unresyst.exceptions import EmptyTestSetError\nfrom unresyst.constants import *\n\nfrom settings import LOG_DIRECTORY\n\nclass BaseEvaluator(object):\n \"\"\"The base class for all evaluators\"\"\"\n \n EvaluationPairModel = None\n \"\"\"The model of the class containing testing pairs. To be overriden.\n A EvaluationPair subclass.\n \"\"\"\n \n prediction_metric = None\n \"\"\"A function taking the test pair model that is called on the evaluated pairs\n to obtain a number \n \"\"\"\n \n recommendation_metric = None\n \"\"\"A function taking the test pair model and number of desired recommendations,\n that is called on the evaluated pair model to obtain a number \n \"\"\"\n \n # methods to be called before the build:\n # \n \n @classmethod\n def select_evaluation_pairs(cls):\n \"\"\"Select the pairs for recommender evaluation.\n \n The selected pairs will be added to the EvaluationPairModel and\n removed from the models where they are now\n \"\"\"\n cls.EvaluationPairModel.select()\n \n \n @classmethod\n def export_evaluation_pairs(cls, filename):\n \"\"\"Export the evaluation pair ids to a csv \n file of the given name.\n \n @type filename: str\n @param filename: the full path to the file\n \n @raise FileNotExists and other file open errors.\n \"\"\"\n \n with open(filename, 'w') as f:\n \n # export the relationships to the given file\n cls.EvaluationPairModel.export(f)\n\n \n # methods to be called after the build:\n # \n \n @classmethod\n def _get_cleared_pairs(cls):\n \"\"\"Get the test pairs, check whether they aren't empty and\n clear the results.\n \n @rtype: QuerySet\n @return: the test pairs\n \n @raise EmptyTestSetError: if the set is empty\n \"\"\"\n # get the dataset and test if not empty\n qs_pairs = cls.EvaluationPairModel.objects.all() \n \n # remove all previous success and obtained expectancies\n qs_pairs.update(obtained_expectancy=None, is_successful=None) \n \n if not qs_pairs:\n raise EmptyTestSetError(\"Call the select_validation_pairs()\"+ \\\n \" method first\") \n \n return qs_pairs\n\n @staticmethod\n def _open_logfile(directory, filename):\n \"\"\"Open a log file\n \n @rtype: file\n \"\"\"\n if directory and filename:\n full_name = os.path.join(directory, filename)\n return codecs.open(full_name, mode='w', encoding='utf-8')\n\n return None \n\n @classmethod\n def evaluate_predictions(cls, recommender, save_predictions=False):\n \"\"\"Evaluate each evaluation pair by calling the predict_relationship\n method.\n \n The results are written to the pair model\n \n @type recommender: Recommender\n @param recommender: the built recommender on which the predictions\n should be evaluated \n \n @type save_predictions: bool\n @param save_predictions: should the generated predictions be saved to db?\n \n @rtype: float\n @return: the result of the metric \n\n @raise EmptyTestSetError: if the test set is empty \n \"\"\"\n # if the settings are given, open a log file and log\n fpreds = cls._open_logfile(LOG_DIRECTORY, LOG_PREDICTIONS_FILENAME)\n fhits = cls._open_logfile(LOG_DIRECTORY, LOG_HITS_FILENAME)\n \n # get the pairs\n qs_pairs = cls._get_cleared_pairs()\n all_count = qs_pairs.count() \n\n # initialize \n i = 0\n succ_count = 0\n non_triv_count = 0\n \n print \"Processing %d pairs...\" % all_count\n \n # go through the pairs\n for pair in qs_pairs.iterator():\n \n # evaluate\n prediction = recommender.predict_relationship(pair.subj, pair.obj, save_predictions)\n \n # log\n if fpreds:\n fpreds.write(u\"%s\\n\" % prediction.__unicode__())\n \n pair.obtained_expectancy = prediction.expectancy\n pair.is_successful = pair.get_success()\n pair.save() \n \n i += 1\n \n if i % 100 == 0:\n print \"%d pairs processed\" % i\n \n if pair.is_successful:\n succ_count += 1\n \n if fhits:\n fhits.write(u'%s\\n' % prediction.__unicode__())\n \n if pair.obtained_expectancy < TRIVIAL_EXPECTANCY:\n non_triv_count += 1\n \n # this isn't very best practice-following but we don't care for file corruption\n if fhits:\n fhits.close()\n \n if fpreds:\n fpreds.close()\n \n # count and print the success rate \n cls.success_rate = float(succ_count)/ all_count\n\n print \"Success rate: %f (%d/%d)\" % (cls.success_rate, succ_count, all_count)\n print \"%d of %d successful predictions were non-trivial.\" % (non_triv_count, succ_count)\n \n # count the metric \n return cls.prediction_metric()\n \n @classmethod\n def evaluate_recommendations(cls, recommender, count):\n \"\"\"Evaluate recommendations obtained for the subjects in the test\n pairs.\n \n @type recommender: Recommender\n @param recommender: the built recommender on which the predictions\n should be evaluated\n \n @type count: int\n @param count: the number of recommendations to get \n \n @rtype: float\n @return: the result of the metric \n \"\"\"\n # if the settings are given, open a log file and log \n fhits = cls._open_logfile(LOG_DIRECTORY, LOG_HITS_FILENAME)\n frecs = cls._open_logfile(LOG_DIRECTORY, LOG_RECOMMENDATIONS_FILENAME)\n \n # get the cleared pair dataset \n qs_pairs = cls._get_cleared_pairs()\n \n # take ids of subjects from the test pairs \n subj_id_list = qs_pairs.values_list('subj__pk',flat=True).distinct()\n \n # take the whole subjects\n qs_subjects = recommender.subjects.filter(pk__in=subj_id_list) \n\n i = 0 \n hit_count = 0 \n for subj in qs_subjects: \n \n i += 1\n \n if i % 10 == 0:\n print \"%d subjects processed\" % i\n \n # for each subject in the test pair get its recommendations\n recommendations = recommender.get_recommendations(subj, count) \n \n \n # mark each recommended object that is in the test set as successful\n #\n for rec in recommendations:\n\n # if it should be logged, log it\n if frecs:\n frecs.write(u'%s\\n' % rec.__unicode__())\n \n obj = rec.object_\n \n # filter the test pair\n qs_pair = cls.EvaluationPairModel.objects.filter(subj=subj, obj=obj)\n \n # if it's in the test pair mark it as successful\n for pair in qs_pair:\n \n hit_count += 1\n \n pair.is_successful = True\n pair.save()\n \n if fhits:\n fhits.write(u'%s\\n' % rec.__unicode__())\n\n # this isn't very best practice-following but we don't care for file corruption\n if fhits:\n fhits.close()\n \n if frecs:\n frecs.close()\n \n print \"%d hits recorded, counting metric...\" % hit_count\n \n # count the metric \n return cls.recommendation_metric(count)\n \n \n \n\n"},{"alpha_fraction":0.7510204315185547,"alphanum_fraction":0.7530612349510193,"avg_line_length":36.69230651855469,"blob_id":"dd47d8c9649e3b251edb94912656715064a4e277","content_id":"6a020e161580ee21d4d2238f749bfa815261ef8d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":490,"license_type":"no_license","max_line_length":75,"num_lines":13,"path":"/UnresystCD/code/adapter/flixster/mahout_recommender.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The representation of the Mahout recommender in Unresyst.\"\"\"\n\nfrom unresyst.recommender.external_recommender import ExternalRecommender\nfrom recommender import MovieRecommender\n\nclass MahoutMovieRecommender(ExternalRecommender, MovieRecommender):\n \"\"\"An external artist recommender - both novel and non-novel artists\"\"\"\n \n name = \"Mahout movie recommender\"\n \"\"\"The name\"\"\"\n \n explicit_rating_rule = MovieRecommender.rules[0]\n \"\"\"The first one is the explicit one\"\"\"\n"},{"alpha_fraction":0.6203219890594482,"alphanum_fraction":0.6234704852104187,"avg_line_length":32.91990280151367,"blob_id":"042d921424fee5827dca328ed2ba70d90411e1bd","content_id":"a1b0dbb96e6b153bf61ff07f114a55cc9171088c","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":13975,"license_type":"no_license","max_line_length":100,"num_lines":412,"path":"/UnresystCD/code/adapter/unresyst/models/abstractor.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Models which instances are created by the abstractor package.\n\nA representation of each defined role and relationship.\n\"\"\"\n\nfrom django.db import models\n\nfrom unresyst.constants import *\nfrom base import BaseRelationshipInstance, ContentTypeModel, \\\n BaseRelationshipDefinition\nfrom unresyst.exceptions import InvalidParameterError\n\ndef _count_expectancy(is_positive, weight, confidence=1):\n \"\"\"Count the expectancy for the instance as 1/2 +- weighted_confidence/2 \n .. depending on is_positive.\n \n @type is_positive: bool\n @param is_positive: is the rule/relationship/bias positive for the \n predicted_relationship?\n \n @type weight: float\n @param weight: the static weight of the rule/relationship/bias\n\n @type confidence: float\n @param confidence: the confidence of the rule/relationship/bias depending\n on the entity\n \n @rtype: float\n @return: the probability of the predicted_relationship appearing between\n the entities in the pair.\n \"\"\"\n \n # get factor \n factor = 1 if is_positive else -1 \n \n return 0.5 + factor * ((weight * confidence) / 2)\n\n\n# Definitions:\n# \n\nclass PredictedRelationshipDefinition(BaseRelationshipDefinition):\n \"\"\"A definition of a the predicted relationship\"\"\"\n\n class Meta:\n app_label = 'unresyst' \n \nclass ExplicitRuleDefinition(BaseRelationshipDefinition):\n \"\"\"A definition of an explicit feedback rule\"\"\"\n \n class Meta:\n app_label = 'unresyst' \n\n\nclass RuleRelationshipDefinition(BaseRelationshipDefinition):\n \"\"\"A definition of a rule/relationship. Represents the rule/relationship\n given by the domain expert in the database.\"\"\" \n \n weight = models.FloatField()\n \"\"\"The weight of the rule/relationship. A number from [0, 1].\"\"\"\n \n relationship_type = models.CharField(max_length=MAX_LENGTH_RELATIONSHIP_TYPE, \n choices=RELATIONSHIP_TYPE_CHOICES)\n \"\"\"A string indicating whether it's a subject, object or both.s/o/so\"\"\"\n \n is_positive = models.BooleanField()\n \"\"\"Is the rule/relationship positive for the predicted_relationship?\"\"\"\n\n class Meta:\n app_label = 'unresyst' \n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\" \n ret = super(RuleRelationshipDefinition, self).__unicode__()\n\n return ret + \"%f, positive: %s\" % (self.weight, self.is_positive)\n \n# instances:\n#\n\nclass RelationshipInstance(BaseRelationshipInstance, ContentTypeModel):\n \"\"\"The relationship between two subject/objects.\n \n All relationships are \"symetrical\" in sense that if there is \n a relationship in the given direction there can't be one in the opposite.\n \n Weight of the relationship is in the definition.\n \n The model also contains the predicted relationship instances.\n \"\"\"\n \n definition = models.ForeignKey('unresyst.BaseRelationshipDefinition')\n \"\"\"The definition of the relationship. Pointing to the base, because\n it has to be also for the predicted relationship instances.\n \"\"\"\n \n additional_unique = ('definition', )\n \"\"\"There can be multiple pairs for one recommender\"\"\"\n \n class Meta:\n app_label = 'unresyst' \n \n unique_together = ('subject_object1', 'subject_object2', 'definition')\n \"\"\"For each definition there can be only one subject-object pair.\"\"\"\n\n def get_expectancy(self, _redirect_to_leaf=True):\n \"\"\"Get the instance expectancy counted \n \n @type redirect_to_leaf: bool\n @param redirect_to_leaf: if True, the call is redirected to the leaf class.\n When called from outside _redirect_to_leaf should always be True, unless\n it's sure it's called on the leaf class instance.\n \n @rtype: float from [0, 1]\n @return: the aggregated expectancy.\n \"\"\"\n \n # redirect to leaf or not\n if _redirect_to_leaf:\n return self.as_leaf_class().get_expectancy(_redirect_to_leaf=False)\n \n # get the whole object for definition\n leaf_definition = self.definition.as_leaf_class() \n \n return _count_expectancy(\n is_positive=leaf_definition.is_positive,\n weight=leaf_definition.weight)\n\n @classmethod\n def filter_predicted(cls, recommender_model):\n \"\"\"Get queryset of the instances of the predicted relationship.\n \n @type recommender_model: models.common.Recommender\n @param recommender_model: the recommender whose instances should\n be aggregated. \n\n @rtype: QuerySet\n @returns: instances of the predicted relationship for the given\n recommender\n \"\"\"\n # get the predicted relationship definition\n pred_rel_def = PredictedRelationshipDefinition.objects.get(\n recommender=recommender_model)\n \n return cls.objects.filter(definition=pred_rel_def) \n\n\nclass ExplicitRuleInstance(BaseRelationshipInstance):\n \"\"\"The explicit preference of a subject to an object.\n \n The relationship isn't weighted. \n \"\"\"\n \n definition = models.ForeignKey('unresyst.ExplicitRuleDefinition')\n \"\"\"The definition of the relationship. \n \"\"\"\n\n expectancy = models.FloatField()\n \"\"\"The normalized explicit preference of the given subject to the given \n object.\n A number from [0, 1].\n \"\"\"\n \n additional_unique = ('definition', )\n \"\"\"There can be multiple pairs for one recommender\"\"\"\n \n class Meta:\n app_label = 'unresyst' \n \n unique_together = ('subject_object1', 'subject_object2', 'definition')\n \"\"\"For each definition there can be only one subject-object pair.\"\"\"\n\n def get_expectancy(self, _redirect_to_leaf=True):\n \"\"\"Get the instance expectancy counted \n \n @type redirect_to_leaf: bool\n @param redirect_to_leaf: to stay compatible with other classes\n \n @rtype: float from [0, 1]\n @return: the expectancy.\n \"\"\" \n return self.expectancy \n\n\nclass RuleInstance(RelationshipInstance):\n \"\"\"The rule applied to a pair of subjects/objects.\"\"\" \n \n confidence = models.FloatField()\n \"\"\"The confidence of the given rule being positive/negative (depending on\n definition.is_positive) in means of the predicted_relationship\n A number from [0, 1].\n \"\"\"\n \n def get_expectancy(self, _redirect_to_leaf=True):\n \"\"\"Get the instance expectancy counted as 1/2 +- weight*confidence/2 .. depending on whether\n the relationship is positive.\n\n @type redirect_to_leaf: bool\n @param redirect_to_leaf: if True, the call is redirected to the leaf class.\n When called from outside _redirect_to_leaf should always be True, unless\n it's sure it's called on the leaf class instance. \n \n @rtype: float from [0, 1]\n @return: the aggregated expectancy.\n \"\"\"\n # get the whole object for definition\n leaf_definition = self.definition.as_leaf_class()\n \n return _count_expectancy(\n is_positive=leaf_definition.is_positive, \n weight=leaf_definition.weight,\n confidence=self.confidence) \n \n class Meta:\n app_label = 'unresyst'\n \n# clusters:\n# \n\nclass ClusterSet(models.Model):\n \"\"\"A set of clusters defined by the user. E.g. 'Gender'.\"\"\"\n\n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"A textual characterization of the cluster set.\"\"\" \n \n recommender = models.ForeignKey('unresyst.Recommender')\n \"\"\"The recommender to which the cluster set belongs.\"\"\" \n\n entity_type = models.CharField(max_length=MAX_LENGTH_ENTITY_TYPE, \\\n choices=ENTITY_TYPE_CHOICES)\n \"\"\"A string indicating whether the contained clusters are for subjects,\n objects or both.s/o/so\"\"\"\n\n weight = models.FloatField()\n \"\"\"The weight of the similarity inferred from two subject/objects \n belonging to one cluster. A number from [0, 1].\n \"\"\"\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n \n class Meta:\n app_label = 'unresyst' \n \n unique_together = ('name', 'recommender')\n \"\"\"There can be only one cluster set with the given name for \n a recommender.\n \"\"\"\n\n\nclass Cluster(models.Model):\n \"\"\"A cluster (class of subjects/objects) defined by the user.\n E.g. 'Female'\"\"\"\n \n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"A textual characterization of the cluster\"\"\"\n \n cluster_set = models.ForeignKey('unresyst.ClusterSet')\n \"\"\"The set the cluster belongs to\"\"\"\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return u\"%s - %s\" % (self.cluster_set, self.name)\n \n class Meta:\n app_label = 'unresyst' \n \n unique_together = ('name', 'cluster_set')\n \"\"\"There can be only one cluster cluster set of the given name in the \n cluster set.\n \"\"\"\n\nclass ClusterMember(models.Model):\n \"\"\"A membership of a subject/object in the cluster.\"\"\"\n \n cluster = models.ForeignKey('unresyst.Cluster')\n \"\"\"The cluster\"\"\"\n \n member = models.ForeignKey('unresyst.SubjectObject')\n \"\"\"The cluster member\"\"\"\n \n description = models.TextField(default='', blank=True)\n \"\"\"The description of the membership.\"\"\" \n \n confidence = models.FloatField()\n \"\"\"The confidence of the member belonging to the cluster.\n A number from [0, 1].\n \"\"\"\n \n @classmethod\n def get_pair_expectancy(cls, cluster_member_pair):\n \"\"\"Return the expectancy that will be used for similarity counting \n for cluster members\n \n @type cluster_member_pair: a tuple (pair) of ClusterMember\n @param cluster_member_pair: cluster members for counting the similarity\n of the members.\n \n @rtype: float\n @return: the expectancy that the members are similar\n \"\"\"\n cm1, cm2 = cluster_member_pair\n \n # if they aren't from the same cluster raise an error\n if cm1.cluster != cm2.cluster:\n raise InvalidParameterError(\n message=\"The members to combine don't belong to the same cluster.\",\n recommender=cm1.cluster.cluster_set.recommender, \n parameter_name='cluster_member_pair', \n parameter_value=cluster_member_pair)\n \n conf = cm1.confidence * cm2.confidence\n \n return _count_expectancy(\n is_positive=True,\n weight=cm1.cluster.cluster_set.weight,\n confidence=conf)\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return u\"(%s, %s): %f\" % (self.cluster, self.member, self.confidence)\n \n \n class Meta:\n app_label = 'unresyst' \n \n unique_together = ('cluster', 'member')\n \"\"\"There can be only one membership for a subjectobject in the given \n cluster.\n \"\"\" \n\n# Bias:\n#\n \nclass BiasDefinition(models.Model):\n \"\"\"The definition of a bias\"\"\"\n \n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"The name of the bias\"\"\" \n \n recommender = models.ForeignKey('unresyst.Recommender')\n \"\"\"The recommender to which the definition belongs. \n \"\"\"\n \n weight = models.FloatField()\n \"\"\"The weight of the bias\n belonging to one cluster. A number from [0, 1].\n \"\"\"\n \n entity_type = models.CharField(max_length=MAX_LENGTH_ENTITY_TYPE, \\\n choices=ENTITY_TYPE_CHOICES)\n \"\"\"A string indicating whether the contained bias is for subjects,\n objects or both.s/o/so\"\"\"\n \n is_positive = models.BooleanField()\n \"\"\"Is the bias positive (adding a probability) for the predicted_relationship?\"\"\"\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance.\"\"\"\n return self.name\n \n class Meta:\n app_label = 'unresyst' \n unique_together = ('name', 'recommender')\n \"\"\"There can be only one bias definition with the given name for \n a recommender.\n \"\"\"\n \n \nclass BiasInstance(models.Model):\n \"\"\"A bias for a particular subject/object.\"\"\" \n \n confidence = models.FloatField()\n \"\"\"The confidence of the given bias being positive/negative (depending on\n definition.is_positive) in means of the predicted_relationship\n A number from [0, 1].\n \"\"\" \n \n subject_object = models.ForeignKey('unresyst.SubjectObject')\n \"\"\"The biased subject/object.\"\"\"\n \n definition = models.ForeignKey('unresyst.BiasDefinition')\n \"\"\"The definition of the bias.\n \"\"\"\n \n description = models.TextField(default='', blank=True)\n \"\"\"The filled description of the bias.\"\"\" \n\n \n class Meta:\n app_label = 'unresyst' \n unique_together = ('definition', 'subject_object')\n \"\"\"There can be only one bias for each definition on one subjectobject\n a recommender.\n \"\"\" \n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance.\"\"\"\n return u\"%s: %s\" % (self.definition, self.subject_object)\n\n def get_expectancy(self):\n \"\"\"Get the expectancy (probability that the subjectobject will be \n in the predicted_relationship, according to the bias).\n \n @rtype: float\n @return: expectancy\n \"\"\"\n return _count_expectancy(\n is_positive=self.definition.is_positive,\n weight=self.definition.weight,\n confidence=self.confidence)\n"},{"alpha_fraction":0.6214788556098938,"alphanum_fraction":0.6214788556098938,"avg_line_length":31.75,"blob_id":"85e0020087a2ab13f553e98f84f39e95e03b0b20","content_id":"8f12c0e536194f11e08ce5be4657ed527d83d9e0","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1704,"license_type":"no_license","max_line_length":83,"num_lines":52,"path":"/code/adapter/unresyst/aggregator/base.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The module defines base class for the aggregator package.\"\"\"\n\nclass BaseAggregator(object):\n \"\"\"The base (abstract) class for all aggregators. Defines the interface.\"\"\" \n \n # Build phase:\n #\n\n def __init__(self, combinator=None):\n \"\"\"The initializer\"\"\"\n \n self.combinator = combinator\n \"\"\"The combinator that should be used during aggregating\"\"\"\n\n \n def aggregate_rules_relationships(cls, recommender_model):\n \"\"\"Aggregate the rule and relationship instances.\n \n Aggregates instances of the relationships and rules from \n the recommender_model to create AggregatedRelationshipInstance\n instances. Instances are saved to db.\n \n @type recommender_model: models.common.Recommender\n @param recommender_model: the recommender whose instances should\n be aggregated. \n \"\"\"\n pass\n \n def aggregate_biases(cls, recommender_model):\n \"\"\"Aggregate bias instances.\n \n Aggregates instances of the biases from the recommender_model \n to create AggregatedBiasInstance instances. Instances are saved to db.\n \n @type recommender_model: models.common.Recommender\n @param recommender_model: the recommender whose instances should\n be aggregated. \n \"\"\"\n pass\n\n\n # Update phase:\n # \n \n @classmethod\n def update(cls, recommender, instance_changes):\n \"\"\"Apdate the aggregated relationships according to the instance \n changes.\n \n @rtype: uvidime\n @return: the changes performed on the aggregated relationships. \n \"\"\"\n\n"},{"alpha_fraction":0.5711066722869873,"alphanum_fraction":0.6090385317802429,"avg_line_length":31.73770523071289,"blob_id":"0ce8beca8e8c9058007d21a35e17f232ffdfa132","content_id":"cc21b4cc085d14fb0118532c1aa738c285744b73","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":7988,"license_type":"no_license","max_line_length":145,"num_lines":244,"path":"/code/adapter/lastfm/models.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The models for the last.fm datasets.\"\"\"\n\n\n\"\"\"\nuser_000001\t2009-05-04T23:08:57Z\tf1b1cf71-bd35-4e99-8624-24a6e15f133a\tDeep Dish\t\tFuck Me Im Famous (Pacha Ibiza)-09-28-2007\nuser_000001\t2009-04-01T14:24:52Z\t6c639b1f-5390-4ec9-b3f8-7afe1562e107\tLady Alma\t\tRunning For Nothing [(Instrumental)]/(Instrumental)\nuser_000001\t2009-02-05T22:58:19Z\t440cb8ef-cee3-4711-8408-b1bd6e93f390\tTowa Tei\tc605c224-245b-4107-b9cd-ab1b2c6adb0a\tMind Wall (Feat. Miho Hatori)\n\n#id\tgender\tage\tcountry\tregistered\nuser_000001\tm\t\tJapan\tAug 13, 2006\nuser_000002\tf\t\tPeru\tFeb 24, 2006\nuser_000003\tm\t22\tUnited States\tOct 30, 2005\nuser_000004\tf\t\t\tApr 26, 2006\nuser_000005\tm\t\tBulgaria\tJun 29, 2006\nuser_000006\t\t24\tRussian Federation\tMay 18, 2006\nuser_000007\tf\t\tUnited States\tJan 22, 2006\nuser_000008\tm\t23\tSlovakia\tSep 28, 2006\nuser_000009\tf\t19\tUnited States\tJan 13, 2007\nuser_000010\tm\t19\tPoland\tMay 4, 2006\nuser_000011\tm\t21\tFinland\tSep 8, 2005\n\"\"\"\n\nfrom django.db import models\nfrom django.db.models import Min\n\nfrom unresyst.models import BaseEvaluationPair\n\n\n\nfrom constants import *\n\nclass User(models.Model):\n \"\"\"The user\"\"\"\n \n gender = models.CharField(max_length=1, default='')\n \"\"\"The english name of the enumerator. Should be translated by ugettext before \n displaying.\"\"\"\n \n age = models.PositiveIntegerField(null=True, default=None)\n \"\"\"The age of the user.\"\"\"\n \n registered = models.DateTimeField(null=True, default=None)\n \"\"\"Date when the user was registered\"\"\"\n\n country = models.ForeignKey('Country', null=True)\n \"\"\"The country the user is from\"\"\" \n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return 'user_%d' % self.id \n \n \nclass Country(models.Model):\n \"\"\"A model for a country.\"\"\" \n \n name = models.CharField(max_length=MAX_LENGTH_NAME, unique=True)\n \"\"\"The name of the city.\"\"\" \n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n \n \nclass Track(models.Model):\n \"\"\"A track (song)\"\"\"\n \n guid = models.CharField(unique=True, max_length=40)\n \"\"\"The id of the track in the musicbrainz database\"\"\"\n \n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"The name of the track.\"\"\"\n\n artist = models.ForeignKey('Artist')\n \"\"\"The author of the track\"\"\"\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return u\"%s: %s\" % (self.artist, self.name)\n\n\nclass Artist(models.Model):\n \"\"\"An artist (band or singer)\"\"\"\n \n guid = models.CharField(unique=True, max_length=40)\n \"\"\"The id of the singer in the musicbrainz database\"\"\"\n \n name = models.CharField(max_length=MAX_LENGTH_NAME)\n \"\"\"The name of the track.\"\"\"\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n\n\nclass Scrobble(models.Model):\n \"\"\"A representation of the fact that a user has played a track\"\"\"\n \n user = models.ForeignKey('User')\n \"\"\"The user\"\"\"\n \n track = models.ForeignKey('Track')\n \"\"\"The played track\"\"\"\n \n timestamp = models.DateTimeField()\n \"\"\"The date and time the track was played\"\"\"\n \n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return u\"(%s, %s)\" % (self.user, self.track)\n\n\nclass ArtistTag(models.Model):\n \"\"\"A representation a artist being tagged by the tag\"\"\"\n \n artist = models.ForeignKey('Artist')\n \"\"\"The artist\"\"\"\n \n tag = models.ForeignKey('Tag')\n \"\"\"The tag\"\"\" \n \n count = models.PositiveIntegerField()\n \"\"\"How many times the artist was tagged by the tag\"\"\"\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return u\"(%s, %s: %d)\" % (self.artist, self.tag, self.count) \n \n\n \nclass Tag(models.Model):\n \"\"\"A representation of a tag (label) in the system\"\"\"\n \n name = models.CharField(max_length=MAX_LENGTH_NAME, unique=True)\n \"\"\"The label.\"\"\"\n \n gender_specific = models.CharField(max_length=1, default='')\n \"\"\"If the tag is preferred by a gender, the gender is here\"\"\"\n\n def __unicode__(self):\n \"\"\"Return a printable representation of the instance\"\"\"\n return self.name \n\n\nclass BaseArtistEvaluationPair(BaseEvaluationPair):\n \"\"\"An abstract class for the artist evaluators\"\"\"\n\n subj = models.ForeignKey('lastfm.User')\n \"\"\"The subject\"\"\"\n \n obj = models.ForeignKey('lastfm.Artist')\n \"\"\"The object\"\"\"\n \n test_ratio = 0.2\n \"\"\"The ratio of pairs to select to test pairs\"\"\"\n\n class Meta:\n abstract = True \n \n @classmethod \n def select(cls, i=0):\n \"\"\"See the base class for the documentation.\"\"\"\n all_count = Scrobble.objects.count() \n test_count = int(cls.test_ratio * all_count ) \n train_count = all_count - test_count\n \n test_pairs = Scrobble.objects.order_by('-timestamp')[:test_count]\n\n min_stamp_test = test_pairs.aggregate(Min('timestamp'))['timestamp__min']\n train_pairs = Scrobble.objects.filter(timestamp__lt=min_stamp_test)\n \n # take 1/5 of the scrobbles, remove them and put them to test data \n # save to test, remove from build\n for scrobble in test_pairs.iterator():\n\n # create the test pair for the radio recommender\n val_pair = ArtistEvalPair(\n subj=scrobble.user,\n obj=scrobble.track.artist,\n expected_expectancy=EXPECTED_EXPECTANCY_LISTENED,\n )\n \n # if not in train pairs \n if not train_pairs.filter(user=val_pair.subj, track__artist=val_pair.obj).exists():\n\n # and not already in novel test_pairs\n if not NovelArtistEvalPair.objects.filter(subj=val_pair.subj, obj=val_pair.obj).exists():\n \n # add it also to novel test pairs\n n_val_pair = NovelArtistEvalPair(\n subj=val_pair.subj,\n obj=val_pair.obj,\n expected_expectancy=val_pair.expected_expectancy\n )\n n_val_pair.save()\n\n val_pair.save()\n \n # so wie so delete\n scrobble.delete()\n \n nontrivial_count = NovelArtistEvalPair.objects.count()\n \n print \"Test pairs selected from total %d pairs, %d of %d are non-trivial\" % (all_count, nontrivial_count, test_count) \n\n def get_success(self):\n \"\"\"See the base class for the documentation.\"\"\"\n return self.obtained_expectancy > SUCCESS_LIMIT\n \n\n \n \nclass ArtistEvalPair(BaseArtistEvaluationPair):\n \"\"\"An artist - user pair for validation purposes\n \n All test pairs, no matter if contained in training data or not. \n \"\"\" \n\n @classmethod\n def select_random(cls, i=0):\n \"\"\"UNUSED\"\"\"\n\n scrobble_count = Scrobble.objects.all().count()\n \n # get a queryset containing every n-th scrobble\n id_list = range(1 + i, scrobble_count, CROSS_VALIDATION_COUNT)\n qs_validation_scrobbles = Scrobble.objects.filter(id__in=id_list)\n \n # remove the scrobbles and add them to the class\n for scrobble in qs_validation_scrobbles.iterator():\n \n # create and save the validation pair\n val_pair = cls(\n subj=scrobble.user,\n obj=scrobble.track.artist,\n expected_expectancy=EXPECTED_EXPECTANCY_LISTENED) \n val_pair.save()\n \n # remove the scrobble\n scrobble.delete()\n\nclass NovelArtistEvalPair(BaseArtistEvaluationPair):\n \"\"\"Only test pairs that aren't included in the test set\"\"\"\n class Meta:\n unique_together = ('subj', 'obj')\n"},{"alpha_fraction":0.6770642399787903,"alphanum_fraction":0.6880733966827393,"avg_line_length":24.34883689880371,"blob_id":"133fb2b44e39a49aa2bcbc39b4e0b2716922b53c","content_id":"c38b85c4a195f444066545a6bcbd5e58d58cf32c","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":1090,"license_type":"no_license","max_line_length":231,"num_lines":43,"path":"/code/mahout/mahoutrec/unresystrecommend.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"#!/bin/sh\n\n# params:\n# lastfm\n# flixster\n# travel\n\n\nLASTFM=false\nFLIXSTER=false\nTRAVEL=false\n\nfor param in $*;\ndo \n case $param in\n 'lastfm')\n LASTFM=true\n ;;\n 'flixster')\n FLIXSTER=true\n ;;\n 'travel')\n TRAVEL=true\n ;; \n esac\ndone\n\nif [ $LASTFM = true ]\nthen\n mvn -e exec:java -Dexec.mainClass=\"com.unresyst.UnresystRecommend\" -Dexec.args=\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv 10 /home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_recommendations.csv\" \nfi\n\nif [ $FLIXSTER = true ]\nthen\n mvn -e exec:java -Dexec.mainClass=\"com.unresyst.UnresystRecommend\" -Dexec.args=\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv 10 /home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_recommendations.csv\" \nfi\n\nif [ $TRAVEL = true ]\nthen\n mvn -e exec:java -Dexec.mainClass=\"com.unresyst.UnresystRecommend\" -Dexec.args=\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_train.csv 10 /home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_recommendations.csv\" \nfi\n\necho \"\"\n"},{"alpha_fraction":0.7937915921211243,"alphanum_fraction":0.7937915921211243,"avg_line_length":55.375,"blob_id":"09a23a1f8ea432afe93770da81d70f433cb4da6d","content_id":"627982f133521056ba61ba62704fe4769be69279","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":451,"license_type":"no_license","max_line_length":110,"num_lines":8,"path":"/UnresystCD/code/adapter/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The home of the adapter django project. \n\nContained applications:\n - unresyst: the universal recommender application containing the implementation of the universal recommedner\n - lastfm: an application containing data models and a Unresyst recommender for the last.fm dataset.\n - flixster: application for the flixster dataset, the same structure as lastfm\n - travel: application for the travel agency dataset, the same structure as lastfm\n\"\"\"\n"},{"alpha_fraction":0.6514772176742554,"alphanum_fraction":0.6554832458496094,"avg_line_length":31.639345169067383,"blob_id":"5cff1464dd01874dd53e68affc31a072f820b58a","content_id":"fb84b5471f3093d38a41c109c7e10a8e55cf2638","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1997,"license_type":"no_license","max_line_length":93,"num_lines":61,"path":"/UnresystCD/code/adapter/unresyst/models/algorithm.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Models which instances are created by the algorithm package.\n\nThe preference between the subject and object in the means of \nthe predicted relationship.\n\"\"\"\n\nfrom django.db import models\n\nfrom base import BaseRelationshipInstance\n\nclass RelationshipPredictionInstance(BaseRelationshipInstance):\n \"\"\"The preference between the subject and object in the means of \n the predicted relationship.\n \"\"\"\n \n expectancy = models.FloatField()\n \"\"\"The probability of the predicted relationship appearance between \n the subject and the object. A number from [0, 1]. \n \"\"\"\n \n recommender = models.ForeignKey('unresyst.Recommender')\n \"\"\"The recommender it belongs to\"\"\"\n \n is_uncertain = models.BooleanField(default=False)\n \"\"\"Is the prediction uncertain? (We don't know anything about the pair)\"\"\"\n \n is_trivial = models.BooleanField(default=False)\n \"\"\"Is the prediction trivial? (Already contained in the train data)\"\"\"\n \n class Meta:\n app_label = 'unresyst' \n unique_together = ('subject_object1', 'subject_object2', 'recommender')\n\n def __unicode__(self):\n return \"(%s, %s), %f\" % (self.subject_object1, self.subject_object2, self.expectancy)\n\n \nclass ExternalPrediction(models.Model):\n \"\"\"A prediction taken from an external recommender.\n \"\"\" \n \n subj_id = models.IntegerField()\n \"\"\"The id of the subject\"\"\"\n \n obj_id = models.IntegerField()\n \"\"\"The if of the object\"\"\"\n \n expectancy = models.FloatField()\n \"\"\"The probability of the predicted relationship appearance between \n the subject and the object. A number from [0, 1]. \n \"\"\" \n \n recommender = models.ForeignKey('unresyst.Recommender')\n \"\"\"The recommender it belongs to\"\"\" \n \n class Meta:\n app_label = 'unresyst' \n unique_together = ('subj_id', 'obj_id', 'recommender')\n\n def __unicode__(self):\n return \"(%s, %s), %f\" % (self.subj_id, self.obj_id, self.expectancy) \n"},{"alpha_fraction":0.5508274435997009,"alphanum_fraction":0.5508274435997009,"avg_line_length":22.27777862548828,"blob_id":"b3e488fef73f2bf74b2bd4f721d639a4d867b199","content_id":"4160764ea3b97a155d2f0c99fd77eab10980d7ce","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":423,"license_type":"no_license","max_line_length":68,"num_lines":18,"path":"/UnresystCD/code/adapter/demo/urls.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Urls for the demo application\"\"\"\n\nfrom django.conf.urls.defaults import *\n\nurlpatterns = patterns('demo.views',\n \n # shoe recommendations\n url(regex=r'^user/(?P[\\w-]+)/shoe-recommendations/$',\n view='view_recommendations', \n name=\"recommendations\"\n ),\n\n # home page\n url(regex=r'^$',\n view='view_home_page',\n name='home_page'\n ) \n) \n"},{"alpha_fraction":0.5290427803993225,"alphanum_fraction":0.5329925417900085,"avg_line_length":31.786258697509766,"blob_id":"ad1cde0ed7345882ec924ec7c3e2862b9cfa53d7","content_id":"7b0edd5c2b209e4c3de06222c5a0623a05db94da","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":4304,"license_type":"no_license","max_line_length":142,"num_lines":131,"path":"/UnresystCD/code/adapter/flixster/recommender.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The flixster recommender\"\"\"\n\nfrom django.db.models import Avg\n\nfrom unresyst import *\nfrom models import *\nfrom constants import *\n\ndef _rated_positively_generator():\n \"\"\"The generator to the predicted relationship\"\"\"\n for r in Rating.objects.filter(rating__gt=str(MIN_POSITIVE_RATING)):\n yield (r.user, r.movie)\n\nclass MovieRecommender(Recommender):\n \"\"\"The flixster movie recommender\"\"\"\n\n name = \"Flixster Movie Recommender\"\n \"\"\"The name\"\"\" \n \n subjects = User.objects\n \"\"\"The objects to who the recommender will recommend.\"\"\"\n \n objects = Movie.objects\n \"\"\"The objects that will be recommended.\"\"\" \n\n predicted_relationship = PredictedRelationship( \n name=\"User has rated the movie positively.\",\n condition=None, \n description=\"\"\"User %(subject)s has rated the %(object)s positively.\"\"\",\n generator=_rated_positively_generator,\n )\n \"\"\"The relationship that will be predicted\"\"\"\n \n rules = (\n # explicit rating\n ExplicitSubjectObjectRule(\n name=\"Movie rating.\",\n \n description=\"User %(subject)s has rated %(object)s.\",\n \n # all pairs user, rated movie\n generator=lambda: [(r.user, r.movie) for r in Rating.objects.all()],\n \n # the number of stars divided by five\n expectancy=lambda s, o:float(Rating.objects.get(user=s, movie=o).rating) / MAX_STARS,\n ),\n )\n \"\"\"The rules\"\"\"\n\n relationships = (\n # users in friendship\n SubjectSimilarityRelationship(\n name=\"Users are friends.\",\n \n generator=lambda: [(f.friend1, f.friend2) for f in Friend.objects.all()], \n \n is_positive=True, \n \n weight=0.1, \n \n description=\"Users %(subject1)s and %(subject2)s are friends.\",\n ),\n )\n \"\"\"The relationships\"\"\"\n \n biases = (\n \n # people giving high ratings\n SubjectBias(\n name=\"Users giving high ratings.\",\n \n description=\"User %(subject)s gives high ratings.\",\n \n weight=1.0, \n \n is_positive=True,\n \n generator=lambda: User.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__gt=str(MIN_HIGH_RATING)), \n \n confidence=lambda user: user.rating_set.aggregate(Avg('rating'))['rating__avg'] - MIN_HIGH_RATING\n ),\n \n # highly rated movies\n ObjectBias(\n name=\"High-rated movies.\",\n \n description=\"Movie %(object)s is high-rated\",\n \n weight=1.0,\n \n is_positive=True,\n \n generator=lambda: Movie.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__gt=str(MIN_HIGH_RATING)),\n \n confidence=lambda movie: movie.rating_set.aggregate(Avg('rating'))['rating__avg'] - MIN_HIGH_RATING\n ),\n \n # people giving low ratings\n SubjectBias(\n name=\"Users giving low ratings.\",\n \n description=\"User %(subject)s gives low ratings.\",\n \n weight=0.25, \n \n is_positive=False,\n \n generator=lambda: User.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__lt=str(MAX_LOW_RATING)), \n \n confidence=lambda user: MAX_LOW_RATING - user.rating_set.aggregate(Avg('rating'))['rating__avg'] \n ),\n \n # low-rated movies\n ObjectBias(\n name=\"Low-rated movies.\",\n \n description=\"Movie %(object)s is low-rated\",\n \n weight=0.25,\n \n is_positive=False,\n \n generator=lambda: Movie.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__lt=str(MAX_LOW_RATING)),\n \n confidence=lambda movie: MAX_LOW_RATING - movie.rating_set.aggregate(Avg('rating'))['rating__avg']\n ), \n \n \n )\n\nMovieRecommender.explicit_rating_rule = MovieRecommender.rules[0] \n \n"},{"alpha_fraction":0.6794871687889099,"alphanum_fraction":0.692307710647583,"avg_line_length":18.5,"blob_id":"e1ecc50b56bc9bcd857699933ad1343cd1eb6e3f","content_id":"2c326e581ce012116baf0dbe73f6af9e0588ee6b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":78,"license_type":"no_license","max_line_length":39,"num_lines":4,"path":"/code/adapter/load.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"#!/bin/sh\n\n# load the database from the given file\nmysql -u root adapter < $1\n"},{"alpha_fraction":0.5593509078025818,"alphanum_fraction":0.5637439489364624,"avg_line_length":39.85714340209961,"blob_id":"de67b101372b2d6ca574db8bb84fed49f465a847","content_id":"fddd4bd233c7e3d4c7308f87998218274cd809ca","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":11154,"license_type":"no_license","max_line_length":96,"num_lines":273,"path":"/code/adapter/unresyst/compilator/get_first_compilator.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The basic compilator used in Unresyst\"\"\"\n\nfrom base import BaseCompilator\nfrom unresyst.constants import *\nfrom unresyst.models.aggregator import AggregatedRelationshipInstance\nfrom unresyst.models.abstractor import RelationshipInstance\nfrom unresyst.models.algorithm import RelationshipPredictionInstance\n\nclass GetFirstCompilator(BaseCompilator):\n \"\"\"Compilator using the first relationship it finds to create a prediction\"\"\"\n \n def __init__(\n self, \n depth=DEFAULT_COMPILATOR_DEPTH, \n breadth=DEFAULT_COMPILATOR_BREADTH, \n pair_depth=DEFAULT_COMPILATOR_PAIR_DEPTH):\n \"\"\"The initializer\"\"\" \n \n super(GetFirstCompilator, self).__init__(\n combinator=None, \n depth=depth, \n breadth=breadth, \n pair_depth=pair_depth) \n\n\n def compile_all(self, recommender_model):\n \"\"\"Compile preferences, known relationships + similarities.\n \"\"\"\n self.compile_aggregates(recommender_model)\n\n print \" Compiling similar objects.\"\n \n # if subjects == objects\n if recommender_model.are_subjects_objects:\n\n # take similar on both sides\n self._compile_similar_subjectobjects(recommender_model)\n\n else:\n \n # take similar to the ones we already have (content-based recommender)\n self._compile_similar_objects(recommender_model)\n print \" Done. Compiling similar subjects.\"\n\n # take liked objects of similar users (almost collaborative filtering)\n self._compile_similar_subjects(recommender_model) \n \n\n #TODO pryc, asi nebude potreba\n def compile_aggregates(self, recommender_model):\n \"\"\"Create predictions from aggregates\"\"\" \n \n # filter only S-O or SO-SO aggregates\n #\n rel_type = RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT \\\n if recommender_model.are_subjects_objects else \\\n RELATIONSHIP_TYPE_SUBJECT_OBJECT \n \n qs_aggr = AggregatedRelationshipInstance.objects.filter(\n recommender=recommender_model,\n relationship_type=rel_type)\n\n # go through the aggregates, create predictions and save them\n for aggr in qs_aggr.iterator():\n\n # order the arguments as they should be \n so1, so2 = self._order_in_pair(aggr.subject_object1, aggr.subject_object2)\n\n prediction = RelationshipPredictionInstance(\n subject_object1=so1,\n subject_object2=so2,\n description=aggr.description,\n expectancy=aggr.expectancy,\n recommender=recommender_model)\n \n prediction.save() \n \n print \" %d aggregated predictions created\" % qs_aggr.count()\n \n \n def _compile_similar_objects(self, recommender_model):\n \"\"\"Create predictions by adding objects similar to ones the objects\n in predicted_relationship - that's a content-based recommender\n \"\"\" \n \n self._compile_similar_entities(\n recommender_model=recommender_model,\n start_entity_type=ENTITY_TYPE_SUBJECT)\n \n def _compile_similar_subjects(self, recommender_model):\n \"\"\"Create predictions by adding objects that similar subjects liked\n - that's collaborative filtering\n \"\"\" \n \n self._compile_similar_entities(\n recommender_model=recommender_model,\n start_entity_type=ENTITY_TYPE_OBJECT)\n\n\n def _compile_similar_subjectobjects(self, recommender_model):\n \"\"\"Create predictions based on similarity for recommenders where \n subjects==objects\n \"\"\"\n \n # firstly the normal direction\n self._compile_similar_entities(\n recommender_model=recommender_model,\n start_entity_type=ENTITY_TYPE_SUBJECTOBJECT,\n reverse=False)\n \n # secondly the opposite\n self._compile_similar_entities(\n recommender_model=recommender_model,\n start_entity_type=ENTITY_TYPE_SUBJECTOBJECT,\n reverse=True)\n\n def _order_in_pair(self, arg1, arg2):\n \"\"\"Swap the arguments in the rule/relationships so that the first\n has a lower id than the second (for subjectobjects), or the subject\n is the first (for others)\n \"\"\"\n # for subjectobject return ordered by pk\n if arg1.entity_type == ENTITY_TYPE_SUBJECTOBJECT:\n if arg2.pk < arg1.pk:\n return (arg2, arg1)\n return (arg1, arg2)\n # for others return subject first \n if arg2.entity_type == ENTITY_TYPE_SUBJECT:\n return (arg2, arg1)\n return (arg1, arg2)\n\n SIMILARITY_RELATIONSHIP_TYPES = {\n ENTITY_TYPE_SUBJECT: RELATIONSHIP_TYPE_OBJECT_OBJECT,\n ENTITY_TYPE_OBJECT: RELATIONSHIP_TYPE_SUBJECT_SUBJECT,\n ENTITY_TYPE_SUBJECTOBJECT: RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT\n } \n \"\"\"Dictionary entity type: relationship entity type.\n Key is the entity type where the traversing for similarity starts, value\n is the relationship type that should be traversed for similarity.\n \"\"\"\n \n ORDERING = {\n ENTITY_TYPE_OBJECT: \"subject_object1__pk\",\n ENTITY_TYPE_SUBJECT: \"subject_object2__pk\",\n ENTITY_TYPE_SUBJECTOBJECT: \"subject_object1__pk\",\n }\n \"\"\"Dictinary entity type (starting) - name of the attribute in PredictedRelationship, \n that should be used for sorting\"\"\"\n \n \n def _compile_similar_entities(self, recommender_model, start_entity_type, reverse=False):\n \"\"\"Create predictions from start_entity_type objects, looking for \n similar entities in end_entity_type\n \n @type recommender_model: models.common.Recommender\n @param recommender_model: the model the for which the predictions\n should be built\n\n @type start_entity_type: str\n @param start_entity_type: the entity type, where to start searching for\n similar entities. E.g. if start_entity_type is 'S', similar 'O' \n will be added to predictions to each 'S'.\n\n @type reverse: bool\n @param reverse: relevant only if start_entity_type=='SO', indicates \n whether the relationships will be traversed in the reverse order. \n \"\"\" \n \n # get the predictions for this recommender\n qs_predictions = RelationshipPredictionInstance.objects.filter(\n recommender=recommender_model)\n\n ordering = self.ORDERING[start_entity_type]\n \n # go through the predicted relationship instances \n # (objects that subjects liked)\n # ordered as needed\n qs_pred_rel_instances = RelationshipInstance\\\n .filter_predicted(recommender_model)\\\n .order_by(ordering)\\\n .select_related(depth=1)\n \n # the caching variable \n last_fin = None \n last_qs = None \n \n print \"Predicted relationship count: %d\" % qs_pred_rel_instances.count()\n \n i = 0\n \n # count neighbourhood\n count_n = 0\n \n # count all \n count_all = 0\n \n for pred_inst in qs_pred_rel_instances.iterator():\n \n i += 1\n # get the subject and object from the relationship instance\n start, fin = (pred_inst.subject_object1, pred_inst.subject_object2) \\\n if pred_inst.subject_object1.entity_type == start_entity_type \\\n else (pred_inst.subject_object2, pred_inst.subject_object1)\n \n # if they are subjectobjects and reversed swap them\n if start_entity_type == ENTITY_TYPE_SUBJECTOBJECT and reverse:\n start, fin = fin, start\n \n similarity_relationship_type = self.SIMILARITY_RELATIONSHIP_TYPES[start_entity_type]\n \n # if the ending entity is the same as the last time, use the cached qs\n if last_fin == fin:\n qs_similar_rels = last_qs\n else: \n # get objects similar to obj (whole relationships) - only breadth highest\n qs_similar_rels = AggregatedRelationshipInstance\\\n .get_relationships(fin)\\\n .filter(relationship_type=similarity_relationship_type)\\\n .order_by('-expectancy')\n \n # current count of theoretically available predictions\n cur_count_all = qs_similar_rels.count()\n \n qs_similar_rels = qs_similar_rels[:self.breadth]\\\n .select_related(depth=1)\n\n # keep it for the next time\n last_fin = fin\n last_qs = qs_similar_rels\n \n\n count = qs_similar_rels.count()\n\n count_all += cur_count_all\n count_n += count\n \n if count and i % 1000 == 0:\n print \"similar count: %d; relationships processed: %d\" % (count, i)\n import gc; gc.collect()\n \n # go through them \n for similar_rel in qs_similar_rels.iterator():\n \n # get the object similar to obj from the relationship\n similar_fin = similar_rel.get_related(fin)\n\n # find out whether there exists a prediction for the pair\n qs_pair_predictions = RelationshipPredictionInstance\\\n .filter_relationships(\n object1=start,\n object2=similar_fin,\n queryset=qs_predictions)\n\n # if exists, keep it there, ignore\n if qs_pair_predictions.exists():\n continue\n \n # order the arguments as they should be \n so1, so2 = self._order_in_pair(start, similar_fin) \n\n # if not, create it with the attributes of the similarity \n # relationship instance\n prediction = RelationshipPredictionInstance(\n subject_object1=so1,\n subject_object2=so2,\n description=similar_rel.description,\n expectancy=similar_rel.expectancy,\n recommender=recommender_model)\n \n prediction.save() \n \n print \"For starting entity type %s, %d out of %d possible relationships created\" \\\n % (start_entity_type, count_n, count_all)\n"},{"alpha_fraction":0.5061048865318298,"alphanum_fraction":0.5157351493835449,"avg_line_length":34.650306701660156,"blob_id":"b7c406a6e4ef6fc0fe902a5c1c7208eac1bf1810","content_id":"d2021e3563ad203310c773135625624c376a9466","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":5815,"license_type":"no_license","max_line_length":146,"num_lines":163,"path":"/UnresystCD/code/adapter/lastfm/recbckp.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":" rules = ( \n \n # don't recommend artists with male-specific tags to females\n SubjectObjectRule(\n name=\"Don't recommend male music to female users.\",\n\n # the user is a female and the artist was tagged by\n # a male-specific tag\n condition=lambda user, artist: user.gender == 'f' and \\\n artist.artisttag_set.filter(tag__gender_specific='m').exists()\n \n # it's a negative rule\n is_positive=False,\n \n weight=0.5,\n \n # the more male-specific tags the artist has, the higher is \n # the rule confidence. Normalized by the artist tag count\n confidence=lambda user, artist: float(\n artist.artisttag_set.filter(tag__gender_specific='m').count())/ \\\n artist.artisttag_set.count(),\n \n description=\"Artist %(object)s isn't recommended to %(subject)s, \" +\n \"because the artist is considered male-specific.\"\n )\n \n \n # users of similar age are similar\n SubjectSimilarityRule(\n name=\"Users with similar age.\",\n \n # both users have given their age and the difference \n # is lower than five\n condition=lambda user1, user2: \n user1.age and user2.age and abs(user1.age - user2.age) <= 5,\n \n is_positive=True, \n \n weight=0.5,\n \n # a magic linear confidence function\n confidence=lambda user1, user2: \n 1 - float(abs(user1.age - user2.age))/AGE_DIFFERENCE,\n \n description=\"Users %(subject1)s and %(subject2)s are about \" + \n \"the same age.\"\n ), \n \n # artists sharing some tags are similar\n ObjectSimilarityRule(\n name=\"Artists sharing some tags.\",\n\n # both artists have some tags and they share at least one tag\n # generator - take artists having some tags, compare them one to one\n generator=_tag_similarity_generator,\n \n \n # it's a positive rule\n is_positive=True,\n \n weight=0.5,\n \n # The more tags the artists have in common, the higher is \n # the similarity confidence\n confidence=lambda artist1, artist2: \\\n float(artist1.artisttag_set.filter(\n tag__id__in=artist2.artisttag_set.values_list('tag__id')\n ).count()) / \\\n min(artist1.artisttag_set.count(), artist2.artisttag_set.count()),\n \n description=\"Artists %(object1)s and %(object2)s are similar \" + \\\n \"because they share some tags.\"\n ), \n \n \n # if the users were registered in a similar period, the're similar\n SubjectSimilarityRule(\n name='Users registered in similar time.',\n \n condition=lambda s1, s2:\n s1.registered and s2.registered and \\\n abs(s1.registered.toordinal() - s2.registered.toordinal()) < REGISTERED_DIFFERENCE / 5,\n \n is_positive=True,\n weight=0.5,\n \n confidence=lambda s1, s2:\n 1 - float(abs(s1.registered.toordinal() - s2.registered.toordinal()))/REGISTERED_DIFFERENCE,\n \n description=\"Users %(subject1)s and %(subject2)s were registered in similar times\",\n ),\n )\n \n cluster_sets = (\n\n \n # user - gender\n SubjectClusterSet(\n \n name='User gender.',\n \n weight=0.5,\n \n # users that have a gender (filled)\n filter_entities=User.objects.exclude(gender=''),\n \n get_cluster_confidence_pairs=lambda user: ((user.gender, 1),),\n \n description=\"%(subject)s's gender is %(cluster)s.\"\n \n ),\n \n # user - country\n SubjectClusterSet(\n \n name='User country.',\n \n weight=0.5,\n \n # users that have country filled\n filter_entities=User.objects.filter(country__isnull=False),\n \n get_cluster_confidence_pairs=lambda user: ((user.country.name, 1),),\n \n description=\"%(subject)s is from %(cluster)s.\"\n ),\n )\n \n biases = (\n ObjectBias(\n name=\"Artists whose tracks were listened the most.\",\n description=\"%(object)s is much listened.\",\n weight=0.5,\n is_positive=True,\n # users whose tracks were listened more than the half of the most listened\n generator=lambda: Artist.objects.annotate(listen_count=Count('track__scrobble')).filter(listen_count__gt=MAX_SCROBBLE_COUNT/7),\n \n # the number of scrobbles for the artist divided by the max.\n confidence=lambda a: float(a.track_set.annotate(scrobble_count=Count('scrobble')).aggregate(Sum('scrobble_count')))/MAX_SCROBBLE_COUNT\n ),\n \n ) \n \n \n \n \n \n \n \n # tag clusters\n ObjectClusterSet(\n\n name=\"Artist tags.\",\n\n weight=0.5,\n \n # artists that are tagged by a tag that another artist also has\n filter_entities=Artist.objects.annotate(shared_count=Count('artisttag__tag__artisttag')).filter(shared_count__gt=1).distinct(),\n \n get_cluster_confidence_pairs=_get_artist_tag_pairs,\n \n description=\"%(object)s was tagged as %(cluster)s.\",\n ),\n"},{"alpha_fraction":0.5247273445129395,"alphanum_fraction":0.5608845353126526,"avg_line_length":26.657024383544922,"blob_id":"94c6ae64ba47d9b97d88b8812b92c95e32dfa0b7","content_id":"a1c2c5fae057744714787879bd4a97cd774cc388","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":6693,"license_type":"no_license","max_line_length":159,"num_lines":242,"path":"/UnresystCD/code/adapter/lastfm/save_data.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Saving data from the last.fm datasets\"\"\"\n\nimport csv\nimport os\nfrom datetime import datetime\nfrom dateutil.parser import parse\nfrom dateutil.tz import tzutc\n\nfrom models import *\n\nTZ = tzutc()\n\"\"\"The used timezone\"\"\"\n\nGENDER_SPECIFIC_TAGS = {\n 'punk': 'm',\n 'metal': 'm',\n 'punk rock': 'm',\n 'hardcore': 'm',\n 'hard rock': 'm',\n 'industrial': 'm',\n 'post-punk': 'm',\n 'post-rock': 'm',\n 'heavy metal': 'm',\n 'metalcore': 'm',\n 'death metal': 'm',\n 'noise': 'm',\n 'hardcore punk': 'm',\n\n 'female': 'f',\n 'beautiful': 'f',\n 'Love': 'f',\n 'lovely': 'f',\n 'sweet': 'f',\n 'singer-songwriter': 'f',\n 'singer songwriter': 'f',\n 'woman': 'f',\n 'mellow': 'f',\n 'dream pop': 'f',\n 'soft': 'f',\n 'emotional': 'f',\n 'cut': 'f',\n}\n\"\"\"Tags that are specific for the given gender.\n\"\"\"\n\n\"\"\"\n7141 artists\n992 users\n19903 tracks\n22666 scrobbles\n\n\n\n# relative paths to the dataset files\nfilename_scrobbles = '../../datasets/lastfm/1K/tracks30000.tsv'\nfilename_users = '../../datasets/lastfm/1K/userid-profile.tsv'\nfilename_tags = '../../datasets/lastfm/Lastfm-ArtistTags2007/artist_tags200K.tsv'\n\"\"\"\n\n\"\"\"\n2348 artists\n100 users\n5057 tracks\n5532 scrobbles\n9120 tags\n28588 tagartists\n322 tagged artists\n\nmax age 38\nmin age 17\n\"\"\"\n# relative paths to the dataset files\nfilename_scrobbles = '../../datasets/lastfm/mini/scrobbles.tsv'\nfilename_users = '../../datasets/lastfm/mini/users.tsv'\nfilename_tags = '../../datasets/lastfm/mini/tags.tsv'\n\nseparator = '\\t'\n\ndef save_data():\n \"\"\"Save the data\"\"\"\n _parse_users(filename_users)\n print \"Users saved.\"\n \n _parse_scrobbles(filename_scrobbles)\n print \"Scrobbles saved.\"\n \n _parse_tags(filename_tags)\n print \"Tags saved.\"\n\n\"\"\"\nuser_000001\tm\t\tJapan\tAug 13, 2006\nuser_000002\tf\t\tPeru\tFeb 24, 2006\nuser_000003\tm\t22\tUnited States\tOct 30, 2005\n\"\"\"\n\ndef _parse_users(filename):\n \"\"\"Parse the user csv file\"\"\"\n \n filename = _get_abs_path(filename)\n\n # open the csv reader\n reader = csv.reader(open(filename, \"rb\"), delimiter=separator, quoting=csv.QUOTE_NONE)\n \n for user_id, gender, age, country, reg_date in reader:\n \n # parse the user id\n user_id = _parse_user_id(user_id)\n \n # replace the empty string by None\n age = None if not age else age\n \n # find or create the country\n if country:\n country_model, created = Country.objects.get_or_create(name=country)\n else:\n country_model = None\n \n # parse the date\n reg_date = None if not reg_date else datetime.strptime(reg_date, '%b %d, %Y')\n \n # create and save the user\n user = User(\n id=user_id,\n gender=gender,\n age=age,\n country=country_model,\n registered=reg_date)\n \n user.save()\n\n\"\"\"\n user_000639 2009-04-08T01:57:47Z 15676fc4-ba0b-4871-ac8d-ef058895b075 The Dogs D'Amour 6cc252d0-3f42-4fd3-a70f-c8ff8b693aa4 How Do You Fall in Love Again\n\"\"\"\n\ndef _parse_scrobbles(filename):\n \"\"\"Parse the scrobbles by the users\"\"\"\n\n filename = _get_abs_path(filename)\n \n reader = csv.reader(open(filename, \"rb\"), delimiter=separator, quoting=csv.QUOTE_NONE)\n for user_id, timestamp, artist_guid, artist_name, track_guid, track_name in reader:\n\n if reader.line_num % 1000 == 0:\n print '%s lines processed' % reader.line_num \n \n # if the track isn't in the musicbrainz db, continue\n if not track_guid or not artist_guid:\n continue\n \n # parse the user id and find the user\n user_id = _parse_user_id(user_id)\n user = User.objects.get(id=user_id)\n \n # parse the date and put it to the UTC timezone\n timestamp = parse(timestamp).astimezone(TZ).replace(tzinfo=None)\n \n # get or create the artist\n artist, created = Artist.objects.get_or_create(\n guid=artist_guid,\n defaults={'name': artist_name})\n \n # get or crate the track\n track, created = Track.objects.get_or_create(\n guid=track_guid, \n defaults= {'artist': artist, 'name': track_name})\n \n \n # create and save the scrobble\n scrobble = Scrobble(\n user=user,\n timestamp=timestamp,\n track=track)\n \n scrobble.save() \n \n\n\"\"\"\n 11eabe0c-2638-4808-92f9-1dbd9c453429Deerhoofamerican14\n 11eabe0c-2638-4808-92f9-1dbd9c453429Deerhoofanimals5\n 11eabe0c-2638-4808-92f9-1dbd9c453429Deerhoofart punk2\n\"\"\" \n\ndef _parse_tags(filename):\n \"\"\"Parse the tags\"\"\"\n\n filename = _get_abs_path(filename)\n \n reader = csv.reader(open(filename, \"rb\"), delimiter=separator, quoting=csv.QUOTE_NONE)\n \n last_artist = Artist.objects.all()[0]\n \n for artist_guid, artist_name, tag_name, count in reader:\n\n if reader.line_num % 1000 == 0:\n print '%s lines processed' % reader.line_num \n \n # some caching for more speed\n if artist_guid == last_artist.guid:\n artist = last_artist\n else: \n # try finding the artist\n qs_artist = Artist.objects.filter(guid=artist_guid)\n \n # if not found go ahead\n if not qs_artist:\n continue\n \n artist = qs_artist[0]\n \n # if the name is strange, go ahead\n if artist.name != artist_name:\n continue\n \n # get the gender specificity\n if GENDER_SPECIFIC_TAGS.has_key(tag_name):\n gender_specific = GENDER_SPECIFIC_TAGS[tag_name]\n else:\n gender_specific = ''\n \n # get or create the tag\n tag, created = Tag.objects.get_or_create(\n name=tag_name, \n defaults={'gender_specific': gender_specific,}) \n \n # create and save the tag for the artist \n artist_tag = ArtistTag(\n artist=artist,\n tag=tag,\n count=count)\n \n artist_tag.save() \n \n last_artist = artist \n \n \ndef _get_abs_path(filename):\n \"\"\"Get the absolute path from the relative\"\"\"\n return os.path.join(os.path.dirname(__file__), filename)\n \ndef _parse_user_id(user_id):\n \"\"\"Get the integer id from the id string\"\"\"\n return int(user_id.split('_')[1])\n"},{"alpha_fraction":0.7558139562606812,"alphanum_fraction":0.7558139562606812,"avg_line_length":27.66666603088379,"blob_id":"0cba3557c9d5312194d5d95f8c3b15e006820481","content_id":"7301d27729a9c8643e0f847085cd196f315d649d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":86,"license_type":"no_license","max_line_length":54,"num_lines":3,"path":"/UnresystCD/code/adapter/travel/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Application holding the data and a recommender for \nthe travel agency dataset.\n\"\"\"\n"},{"alpha_fraction":0.7663999795913696,"alphanum_fraction":0.7663999795913696,"avg_line_length":38.0625,"blob_id":"2829d1381995c103e68d0afad6485f4b995717a9","content_id":"7e1456486a0e2d20d616f2940869865ca593246f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":625,"license_type":"no_license","max_line_length":80,"num_lines":16,"path":"/code/adapter/lastfm/mahout_recommender.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The representation of the Mahout recommender in Unresyst.\"\"\"\n\nfrom unresyst.recommender.external_recommender import ExternalRecommender\nfrom recommender import ArtistRecommender, NovelArtistRecommender\n\nclass MahoutArtistRecommender(ExternalRecommender, ArtistRecommender):\n \"\"\"An external artist recommender - both novel and non-novel artists\"\"\"\n \n name = \"Mahout artist recommender\"\n \"\"\"The name\"\"\"\n \nclass NovelMahoutArtistRecommender(ExternalRecommender, NovelArtistRecommender):\n \"\"\"External recommender for novel artists only\"\"\"\n \n name = \"Novel Mahout artist recommender\"\n \"\"\"The name\"\"\"\n"},{"alpha_fraction":0.7530864477157593,"alphanum_fraction":0.7530864477157593,"avg_line_length":26,"blob_id":"2cc663c22636ab3d282b4a3c9e490836aa81eb0f","content_id":"447bdfd1b9fd756e0f53ccfe0aae9a1f97c85ff8","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":81,"license_type":"no_license","max_line_length":54,"num_lines":3,"path":"/code/adapter/flixster/__init__.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Application holding the data and a recommender for \nthe Flixster dataset.\n\"\"\"\n"},{"alpha_fraction":0.5634920597076416,"alphanum_fraction":0.5634920597076416,"avg_line_length":30.25,"blob_id":"9536a8e1ff67ca685db02dbfb9009d15abafec92","content_id":"2cdd98fc94573b198df43eb36e8ea8ab59478acf","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":126,"license_type":"no_license","max_line_length":75,"num_lines":4,"path":"/code/adapter/diff.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"#!/bin/bash\n\n# diffovat v meldu zmenene soubory.\nsvn diff --diff-cmd meld `svn stat | grep \"M \" | sed -e 's/M //'`\n\n"},{"alpha_fraction":0.6452485918998718,"alphanum_fraction":0.6459249258041382,"avg_line_length":32.98850631713867,"blob_id":"ab43235feb01e51bb9cf20b564b11fb6ae1c6eb9","content_id":"9203604c83a4ceb6773366cba34dbea8b44ca395","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2957,"license_type":"no_license","max_line_length":100,"num_lines":87,"path":"/UnresystCD/code/adapter/unresyst/recommender/metrics.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The functions used as metrics for the evaluators\"\"\"\nimport math\n\n@classmethod\ndef rmse(cls):\n \"\"\"Count and print the result of the RMSE on the evaluation pair model\n \n @type EvaluationPairModel: a model, BaseEvaluationPair subclass\n @param EvaluationPairModel: a model to evaluate on \n \n @rtype: float\n @return: the RMSE of the predictions\n \"\"\"\n \n # get the expected, obtained expectancy pairs\n exp_pairs = cls.EvaluationPairModel.objects.values_list(\n 'expected_expectancy', 'obtained_expectancy') \n \n # count the sum of the squares of the differences\n square_sum = sum([pow(exp_e - obt_e, 2) for exp_e, obt_e in exp_pairs])\n \n # get the rmse\n ret_rmse = math.sqrt(square_sum / len(exp_pairs))\n print \"RMSE: %f\" % ret_rmse\n \n return ret_rmse\n\n \n@classmethod \ndef precision_recall(cls, count):\n \"\"\"Count and print the Precision/Recall measure. \n \n Both are counted as per-user average\n \n Hit: an object that is both in the recommendation and in the evaluation set.\n \n Precision: number of hits / recommendation list size (=count)\n If a hit appears for an subject-object that is multiple time present in the\n test set, it's counted only once.\n \n Recall: number of hits / number of objects for the subject in the evaluation test\n If a hit appears for an subject-object that is multiple time present in the\n test set, it's counted multiple times.\n \n @type count: int\n @param count: the number of objects in recommendation lists\n \n @rtype: pair float float\n @return: precision, recall \n \"\"\"\n \n # hits are successful pairs\n qs_hit_pairs = cls.EvaluationPairModel.objects.filter(is_successful=True)\n \n # take ids of subjects from the test pairs \n subj_id_list = cls.EvaluationPairModel.objects.values_list('subj__pk', flat=True).distinct() \n \n subj_count = len(subj_id_list)\n \n unique_hit_pairs = qs_hit_pairs.values_list('subj__pk', 'obj__pk').distinct()\n \n # count the precision as explained in docstring\n # as count is the same for all users this can be counted at once\n precision = float(len(unique_hit_pairs)) / (subj_count * count)\n print \"Precision: %f\" % precision\n \n recall_sum = 0\n \n # count the average recall / user\n for subj_id in subj_id_list:\n \n # count the parameters per user \n user_hit_count = qs_hit_pairs.filter(subj__pk=subj_id).count()\n possible_hit_count = cls.EvaluationPairModel.objects.filter(subj__pk=subj_id).count()\n \n subj_recall = float(user_hit_count) / possible_hit_count\n\n #print \"%d precision: %f\" % (subj_id, float(user_hit_count)/count)\n #print \"%d recall: %f\" % (subj_id, subj_recall)\n \n recall_sum += subj_recall\n \n recall = recall_sum / subj_count \n\n print \"Recall: %f\" % recall\n \n return (precision, recall)\n"},{"alpha_fraction":0.5109958052635193,"alphanum_fraction":0.5757350921630859,"avg_line_length":21.685392379760742,"blob_id":"65928876245a1c3845fe7a61dea8aea090510c12","content_id":"1bf2649748b12465206a5401d6fa857b5b29019b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":4047,"license_type":"no_license","max_line_length":90,"num_lines":178,"path":"/code/adapter/flixster/save_data.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"Saving data from csv files\"\"\"\nimport csv\nimport os\n\nfrom django.db.models import Count\n\nfrom models import *\n\n# relative paths to the dataset files\nfilename_links = '../../datasets/flixster/mini/links.tsv'\nfilename_ratings = '../../datasets/flixster/mini/ratings.tsv'\n\nseparator = '\\t'\n\n\"\"\"\nBefore the removal:\n\n>>> User.objects.count()\n1269\n>>> Friend.objects.count()\n4439\n>>> Movie.objects.count()\n4237\n>>> Rating.objects.count()\n14195\n\nAfter the removal:\n\n>>> User.objects.count()\n418\n>>> Friend.objects.count()\n1051\n\n>>> MovieRecommenderEvaluator.select_evaluation_pairs()\n2839 test pairs selected from total 14195 pairs\n\"\"\"\n\nmax_user_id = 1520\n\ndef _get_abs_path(filename):\n \"\"\"Get the absolute path from the relative\"\"\"\n return os.path.join(os.path.dirname(__file__), filename)\n\ndef save_data():\n \"\"\"Save the data\"\"\"\n _parse_links(filename_links)\n print \"Users and links saved.\"\n \n _parse_ratings(filename_ratings)\n print \"Ratings saved.\" \n \n # remove things we don't like\n \n # remove users without rating - they can't be tested anyhow \n User.objects.annotate(rating_count=Count('rating')).filter(rating_count=0).delete()\n print \"Disliked stuff deleted.\" \n\n\n\"\"\"\n6\t1349\n6\t1370\n6\t1375\n6\t1761\n6\t1770\n6\t1775\n6\t1785\n6\t1797\n6\t1803\n6\t1810\n6\t1819\n6\t1825\n6\t1831\n\"\"\" \n \ndef _parse_links(filename):\n \"\"\"Parse the user csv file\"\"\"\n \n filename = _get_abs_path(filename)\n \n last_user = User(id=0) \n\n # open the csv reader\n reader = csv.reader(open(filename, \"rb\"), delimiter=separator, quoting=csv.QUOTE_NONE)\n \n for user_id, friend_id in reader:\n \n if reader.line_num % 5000 == 0:\n print '%s lines processed' % reader.line_num \n \n # parse the user ids\n user_id = int(user_id)\n friend_id = int(friend_id) \n \n # skip high friends\n if friend_id > max_user_id:\n continue\n \n # if it's about the last user, use him, otherwise create him\n if user_id == last_user.id: \n user = last_user\n else:\n user, created = User.objects.get_or_create(id=user_id)\n \n # get or create the friend\n friend, created = User.objects.get_or_create(id=friend_id) \n \n f1, f2 = (user, friend) if user.id < friend.id else (friend, user)\n \n # create the friendship\n Friend.objects.get_or_create(\n friend1=f1,\n friend2=f2)\n \n # save the user for the next time\n last_user = user\n\"\"\"\n6\t57699\t3\n7\t18858\t4\n8\t20644\t4\n8\t43310\t4\n8\t5806\t4\n9\t33430\t4\n11\t12145\t4\n11\t12913\t3\n11\t12920\t5\n11\t14034\t3.5\n11\t14277\t4\n11\t14506\t4\n11\t14889\t5\n11\t15154\t3.5\n11\t15815\t5\n11\t16476\t4.5\n11\t18181\t5\n11\t18874\t5\n\"\"\"\n\ndef _parse_ratings(filename):\n \"\"\"Parse the user csv file\"\"\"\n\n filename = _get_abs_path(filename)\n \n last_user = User(id=0) \n\n # open the csv reader\n reader = csv.reader(open(filename, \"rb\"), delimiter=separator, quoting=csv.QUOTE_NONE)\n \n for user_id, movie_id, rating in reader:\n \n if reader.line_num % 1000 == 0:\n print '%s lines processed' % reader.line_num \n \n # parse the ids and rating\n user_id = int(user_id)\n movie_id = long(movie_id) \n rating = rating\n\n # if it's about the last user, use him, otherwise get him\n if user_id == last_user.id: \n user = last_user\n else:\n # take only users we already have\n user = User.objects.filter(id=user_id) \n if not user:\n continue\n \n user=user[0]\n \n # get or create the movie\n movie, created = Movie.objects.get_or_create(id=movie_id)\n \n # create the rating\n Rating.objects.create(\n user=user, \n movie=movie,\n rating=rating)\n \n # save the user for the next time\n last_user = user\n \n"},{"alpha_fraction":0.6356180310249329,"alphanum_fraction":0.6367896795272827,"avg_line_length":38.39230728149414,"blob_id":"ceb6010c1aebeea284fd0d160826418e3b6dfca1","content_id":"14522cb466bf528486bf4859ebe693234d0f7c19","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":5121,"license_type":"no_license","max_line_length":108,"num_lines":130,"path":"/code/adapter/unresyst/algorithm/base.py","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"\"\"\"The base abstract class for all recommender algorithms.\"\"\"\n\nfrom unresyst.constants import *\nfrom unresyst.models.algorithm import RelationshipPredictionInstance\n\nclass BaseAlgorithm(object):\n \"\"\"The interface provided to the other packages. The methods in the interface\n are implemented by the subclasses\"\"\"\n \n def __init__(self, inner_algorithm=None):\n \"\"\"The initializer\"\"\"\n \n self.inner_algorithm = inner_algorithm\n \"\"\"The inner algorithm\"\"\"\n \n # Build phase:\n #\n \n def build(self, recommender_model):\n \"\"\"Build the recommender, so that the given relationship can be\n predicted easily.\n \n @type recommender_model: models.common.Recommender\n @param recommender_model: the recommender whose instances should\n be aggregated. \n\n \"\"\"\n if self.inner_algorithm:\n self.inner_algorithm.build(recommender_model=recommender_model)\n \n \n # Recommend phase:\n #\n \n def get_relationship_prediction(self, recommender_model, dn_subject, dn_object, remove_predicted):\n \"\"\"Get the prediction of the appearance of the predicted_relationship.\n\n @type recommender_model: models.common.Recommender\n @param recommender_model: the recommender whose instances should\n be aggregated. \n \n @type dn_subject: models.common.SubjectObject\n @param dn_subject: the domain neutral subject\n \n @type dn_object: models.common.SubjectObject\n @param dn_object: the domain neutral object\n\n @type remove_predicted: bool\n @param remove_predicted: should pairs already having \n predicted_relationship between them get the special expectancy value?\n \n @rtype: models.algorithm.RelationshipPredictionInstance\n @return: the model instance for the prediction \n \"\"\"\n if self.inner_algorithm:\n return self.inner_algorithm.get_relationship_prediction(\n recommender_model=recommender_model, \n dn_subject=dn_subject, \n dn_object=dn_object, \n remove_predicted=remove_predicted)\n\n \n def get_recommendations(self, recommender_model, dn_subject, count, expectancy_limit, remove_predicted):\n \"\"\"Get the recommendations for the given subject\n\n @type recommender_model: models.common.Recommender\n @param recommender_model: the recommender whose instances should\n be aggregated. \n\n @type dn_subject: models.common.SubjectObject\n @param dn_subject: the domain neutral subject\n\n @type count: int\n @param count: how many recommendations should be obtained\n \n @type expectancy_limit: float\n @param expectancy_limit: the lower limit for object expectancy, only\n objects with expectancy higher than limit are recommended.\n \n @type remove_predicted: bool\n @param remove_predicted: should pairs already having \n predicted_relationship between them be removed from recommendations?\n \n @rtype: a list of models.algorithm.RelationshipPredictionInstance\n @return: the predictions of the objects recommended to the subject\n \"\"\"\n if self.inner_algorithm:\n return self.inner_algorithm.get_recommendations(\n recommender_model=recommender_model,\n dn_subject=dn_subject,\n count=count,\n expectancy_limit=expectancy_limit,\n remove_predicted=remove_predicted)\n\n @staticmethod\n def _get_uncertain_prediction(recommender_model, dn_subject, dn_object):\n \"\"\"Get the prediction for a pair for which nothing is known\"\"\"\n \n return RelationshipPredictionInstance(\n subject_object1=dn_subject,\n subject_object2=dn_object,\n description=recommender_model.random_recommendation_description,\n recommender=recommender_model,\n expectancy=UNCERTAIN_PREDICTION_VALUE,\n is_uncertain=True\n ) \n\n @staticmethod\n def _get_already_in_relatinship_prediction(recommender_model, predicted_relationship):\n \"\"\"Get the prediction for a pair which is already in the predicted_relationship.\n Valid only for recommenders with remove_predicted=True.\n \"\"\"\n \n return RelationshipPredictionInstance(\n subject_object1=predicted_relationship.subject_object1,\n subject_object2=predicted_relationship.subject_object2,\n description=predicted_relationship.description,\n recommender=recommender_model,\n expectancy=ALREADY_IN_REL_PREDICTION_VALUE,\n is_trivial=True\n ) \n\n # Update phase:\n # \n \n def update(self, recommender, aggregated_changes):\n \"\"\"Update the recommender algorithm structures according to the changes\n in the aggregated relationships.\n \n \"\"\"\n"},{"alpha_fraction":0.6363636255264282,"alphanum_fraction":0.63840651512146,"avg_line_length":18.959182739257812,"blob_id":"5686f75d1fbc60c81183db4c172749f280dd75d6","content_id":"b006b353ab6ace2bb63d19825a6415d705edf892","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":979,"license_type":"no_license","max_line_length":178,"num_lines":49,"path":"/UnresystCD/code/adapter/traveleval.sh","repo_name":"cvengros/unresyst","src_encoding":"UTF-8","text":"#!/bin/sh\n\n\n\nBUILD=true\nEVAL=''\nREC=''\nRECS=false\nPREDS=false\n\nfor param in $*;\ndo \n case $param in\n 'recs')\n EVAL='OrderTourRecommenderEvaluator'\n REC='OrderTourRecommender'\n RECS=true\n ;;\n 'preds')\n PREDS=true\n ;;\n 'dontbuild')\n BUILD=false\n ;;\n esac\ndone\n\nif [ $BUILD = true ]\nthen\n # build\n echo \"Building...\"\n echo \"from travel.recommender import *; OrderTourRecommender.build()\" | python ./manage.py shell\nfi\n\nif [ $RECS = true ]\nthen\n echo \"Evaluating recommendations...\"\n echo \"from travel.evaluation import *; from travel.recommender import *; $EVAL.evaluate_recommendations($REC, 10); quit()\"| python ./manage.py shell\nfi\n\nif [ $PREDS = true ]\nthen\n echo \"Evaluating predicitons...\"\n echo \"from travel.evaluation import *; from travel.recommender import *; OrderTourRankEvaluator.evaluate_predictions(OrderTourRecommender); quit()\" | python ./manage.py shell\nfi\n\n\n\necho \"\"\n\n"}],"string":"[\n {\n \"alpha_fraction\": 0.8373494148254395,\n \"alphanum_fraction\": 0.8373494148254395,\n \"avg_line_length\": 40.5,\n \"blob_id\": \"d65849e870bb0af3006bd0087f2c786bde12eab5\",\n \"content_id\": \"5624145863bed58df98504244c5cc053f0e7fe7f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 166,\n \"license_type\": \"no_license\",\n \"max_line_length\": 64,\n \"num_lines\": 4,\n \"path\": \"/UnresystCD/code/adapter/unresyst/aggregator/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The package contains all classes for the aggregator layer.\\\"\\\"\\\"\\n\\nfrom linear_aggregator import LinearAggregator\\nfrom combining_aggregator import CombiningAggregator\\n\"\n },\n {\n \"alpha_fraction\": 0.5338044762611389,\n \"alphanum_fraction\": 0.5361601710319519,\n \"avg_line_length\": 38.09345626831055,\n \"blob_id\": \"b360191817d609d4eb2a543c9173515247c4e54b\",\n \"content_id\": \"4154f74a3e83c119f5c22aa26eee045400afac5e\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 4245,\n \"license_type\": \"no_license\",\n \"max_line_length\": 118,\n \"num_lines\": 107,\n \"path\": \"/code/adapter/unresyst/algorithm/simple_algorithm.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"A trivial algorithm returning what it has\\\"\\\"\\\"\\n\\nfrom base import BaseAlgorithm\\nfrom unresyst.models.abstractor import RelationshipInstance, \\\\\\n PredictedRelationshipDefinition\\nfrom unresyst.models.algorithm import RelationshipPredictionInstance\\n\\nclass SimpleAlgorithm(BaseAlgorithm):\\n \\n # Build phase:\\n #\\n \\n def build(cls, recommender_model): \\n \\\"\\\"\\\"Do nothing\\\"\\\"\\\"\\n return\\n\\n # Recommend phase:\\n #\\n \\n def get_relationship_prediction(self, recommender_model, dn_subject, dn_object, remove_predicted):\\n \\\"\\\"\\\"See the base class for the documentation.\\n \\n Here - Handle remove_predicted, return the prediction if available, \\n if not, return uncertain. \\n \\\"\\\"\\\" \\n\\n # if predicted should be removed and the pair is in the predicted_rel, \\n # return the special expectancy value\\n if remove_predicted: \\n # all predicted relationships\\n qs_predicted = RelationshipInstance.filter_predicted(recommender_model)\\n\\n # the relationship between dn_subject and dn_object\\n qs_predicted_rel = RelationshipInstance.filter_relationships(dn_subject, dn_object, queryset=qs_predicted)\\n\\n # if the prediction for the pair exists\\n if qs_predicted_rel:\\n \\n # return the special expectancy value \\n assert len(qs_predicted_rel) == 1\\n predicted = qs_predicted_rel[0]\\n\\n return self._get_already_in_relatinship_prediction(\\n recommender_model=recommender_model,\\n predicted_relationship=predicted)\\n \\n # if prediction available, return it\\n # \\n \\n # filter the predictions for recommender\\n qs_pred = RelationshipPredictionInstance.objects.filter(\\n recommender=recommender_model,\\n subject_object1=dn_subject,\\n subject_object2=dn_object)\\n \\n # if available return it\\n if qs_pred: \\n assert len(qs_pred) == 1\\n return qs_pred[0] \\n \\n # otherwise return the uncertain\\n return self._get_uncertain_prediction(\\n recommender_model=recommender_model, \\n dn_subject=dn_subject, \\n dn_object=dn_object\\n ) \\n \\n def get_recommendations(self, recommender_model, dn_subject, count, expectancy_limit, remove_predicted):\\n \\\"\\\"\\\"See the base class for the documentation.\\n \\n Here - Handle remove_predicted and expectancy_limit, return the predictions \\n we have. \\n \\n \\\"\\\"\\\" \\n if remove_predicted:\\n \\n # create kwargs to exclude already liked objects\\n #\\n \\n # get objects that are already liked\\n #\\n qs_predicted = RelationshipInstance.filter_predicted(recommender_model) \\n\\n # get ids of subjectobjects where dn_subject appears\\n predicted_obj_ids = qs_predicted.filter( \\n subject_object1=dn_subject).values_list('subject_object2__pk', flat=True)\\n \\n # construct the arguments to the exclude method.\\n exclude_args = {\\n # the ones that have the predicted def\\n 'subject_object2__pk__in': predicted_obj_ids,\\n }\\n \\n else:\\n exclude_args = {} \\n \\n # get the recommendations ordered by the expectancy from the largest\\n recommendations = RelationshipPredictionInstance.objects\\\\\\n .filter(\\n subject_object1=dn_subject,\\n recommender=recommender_model, \\n expectancy__gt=expectancy_limit)\\\\\\n .exclude(**exclude_args)\\\\\\n .distinct()\\\\\\n .order_by('-expectancy')\\n \\n return list(recommendations[:count]) \\n \\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6227883696556091,\n \"alphanum_fraction\": 0.6234961152076721,\n \"avg_line_length\": 19.171428680419922,\n \"blob_id\": \"7b151d19420122a4838ff7654ddc2b093c930e8a\",\n \"content_id\": \"537dc6792f7ef7acf9687c2c64c514516378f46d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 1413,\n \"license_type\": \"no_license\",\n \"max_line_length\": 113,\n \"num_lines\": 70,\n \"path\": \"/code/adapter/setup.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/bin/sh\\n\\n# parametry: \\n# co chces vlozit do databaze: lastfm, demo\\n# dontdrop: pokud se preda, nedropuje se databaze\\n\\n\\nLASTFM=false\\nDEMO=false\\nFLIXSTER=false\\nDONTDROP=false\\nTRAVEL=false\\n\\nfor param in $*;\\ndo \\n case $param in\\n 'lastfm')\\n LASTFM=true\\n ;;\\n 'demo')\\n DEMO=true\\n ;;\\n 'flixster')\\n FLIXSTER=true\\n ;;\\n 'travel')\\n TRAVEL=true\\n ;;\\n 'dontdrop')\\n DONTDROP=true\\n ;;\\n esac\\ndone\\n\\nif [ $DONTDROP = false ]\\nthen\\n # drop and create database\\n echo \\\"Dropping and creating database.\\\"\\n echo \\\"DROP DATABASE IF EXISTS adapter; CREATE DATABASE adapter CHARACTER SET utf8;\\\" | mysql --user=root mysql\\nfi\\n\\n# syncdb\\npython ./manage.py syncdb --noinput > /dev/null\\n\\n\\nif [ $DEMO = true ]\\nthen\\n echo \\\"Adding demo data.\\\"\\n echo \\\"from demo.save_data import save_data; save_data(); quit();\\\" | python ./manage.py shell\\nfi\\n\\nif [ $LASTFM = true ]\\nthen\\n echo \\\"Adding last.fm data.\\\"\\n echo \\\"from lastfm.save_data import save_data; save_data(); quit();\\\" | python ./manage.py shell\\nfi\\n\\nif [ $FLIXSTER = true ]\\nthen\\n echo \\\"Adding flixster data.\\\"\\n echo \\\"from flixster.save_data import save_data; save_data(); quit();\\\" | python ./manage.py shell\\nfi\\n\\nif [ $TRAVEL = true ]\\nthen\\n echo \\\"Adding travel data.\\\"\\n echo \\\"from travel.save_data import save_data; save_data(); quit();\\\" | python ./manage.py shell\\nfi\\n\\necho \\\"\\\" \\n\"\n },\n {\n \"alpha_fraction\": 0.6679438352584839,\n \"alphanum_fraction\": 0.6679438352584839,\n \"avg_line_length\": 40.21052551269531,\n \"blob_id\": \"a3c7981df115919cbbc332ce1d569cc02eabfccd\",\n \"content_id\": \"15c9154a172b54dab069ff3133db48e3b6696594\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 783,\n \"license_type\": \"no_license\",\n \"max_line_length\": 102,\n \"num_lines\": 19,\n \"path\": \"/UnresystCD/code/adapter/unresyst/combinator/average_combinator.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Combinator using a weighted average.\\\"\\\"\\\"\\n\\nfrom base import BaseCombinator\\n\\nclass AverageCombinator(BaseCombinator):\\n \\\"\\\"\\\"A combinator using weighted average\\n \\\"\\\"\\\"\\n \\n def _combine(self, combination_elements, ResultClass):\\n \\\"\\\"\\\"See the base class for documentation\\\"\\\"\\\"\\n \\n # sort it in order to provide the right order of explanations\\n combination_elements.sort(key=lambda el: el.get_expectancy(), reverse=True) \\n\\n # count the average and concatenation of the descriptions\\n avgexp = sum([ce.get_expectancy() for ce in combination_elements]) / len(combination_elements)\\n desc = self._concat_descriptions(combination_elements)\\n \\n return ResultClass(expectancy=avgexp, description=desc)\\n\"\n },\n {\n \"alpha_fraction\": 0.543181300163269,\n \"alphanum_fraction\": 0.5789534449577332,\n \"avg_line_length\": 29.939067840576172,\n \"blob_id\": \"3657a4cf62ca68f99f3d0e4c08af7f73cb470560\",\n \"content_id\": \"c6f1c1c9ec8d9c2e872e62ca38d245a783ff6b0f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 8638,\n \"license_type\": \"no_license\",\n \"max_line_length\": 168,\n \"num_lines\": 279,\n \"path\": \"/code/adapter/travel/save_data.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Saving data from the last.fm datasets\\\"\\\"\\\"\\n\\nimport csv\\nimport os\\nimport re\\nfrom datetime import datetime\\nfrom dateutil.parser import parse\\n\\nfrom models import *\\n\\n\\n\\nregexp_url = r'/zajezdy/zobrazit/(?P[\\\\w-]+)/(?P[\\\\w-]+)/(?P[\\\\w-]+)/*'\\n\\n# relative paths to the dataset files\\nfilename_all = '../../datasets/travel/travel.csv'\\n\\nseparator = ','\\n\\ndecay_secs = 1200\\n\\nACTIONS = {\\n 'HREF_CLICK': Click,\\n 'ON_PAGE': MouseMove,\\n 'OBJECT_INTERACTION': {\\n 'ORDER': Order,\\n 'QUESTION': Question,\\n },\\n 'PAGE_OPEN': ViewProfile,\\n 'PAGE_CLOSE': None,\\n}\\n\\\"\\\"\\\"Dictionary action: class to save\\\"\\\"\\\"\\n\\ndef save_data():\\n \\\"\\\"\\\"Save the data\\\"\\\"\\\"\\n \\n _parse_all(filename_all)\\n print \\\"All saved.\\\"\\n \\n\\n\\\"\\\"\\\"\\n1,3,1,2010-04-23 11:37:34,PAGE_OPEN,REF: http://slantour.cz/zajezdy/katalog/za-sportem,/\\n2,3,1,2010-04-23 11:37:35,PAGE_CLOSE,,/\\n4,3,1,2010-04-23 11:37:36,PAGE_OPEN,REF: http://slantour.cz/,/zajezdy/katalog/pobytove-zajezdy\\n5,3,1,2010-04-23 11:37:37,PAGE_CLOSE,,/zajezdy/katalog/pobytove-zajezdy\\n7,3,1,2010-04-23 11:37:38,PAGE_OPEN,REF: http://slantour.cz/zajezdy/katalog/pobytove-zajezdy,/zajezdy/katalog/pobytove-zajezdy/ceska-republika\\n8,3,1,2010-04-23 11:37:39,PAGE_CLOSE,,/zajezdy/katalog/pobytove-zajezdy/ceska-republika\\n\\nBefore removal:\\n>>> from travel.models import *\\n>>> from django.db.models import *\\n>>> User.objects.count()\\n12991\\n>>> Click.objects.count()\\n17426\\n>>> Question.objects.count()\\n21\\n>>> Order.objects.count()\\n227\\n>>> MouseMove.objects.count()\\n68759\\n>>> Session.objects.count()\\n14409\\n>>> Tour.objects.count()\\n702\\n>>> TourType.objects.count()\\n6\\n>>> TourType.objects.all()\\n[, , , , , ]\\n>>> Country.objects.count()\\n50\\n>>> User.objects.annotate(scount=Count('session')).filter(scount__gt=1).count()\\n834\\n>>> ViewProfile.objects.count()\\n40766\\n\\n>>> User.objects.annotate(ocount=Count('session__order')).filter(ocount__gte=1).count()\\n193\\n>>> User.objects.annotate(ocount=Count('session__order')).filter(ocount__gte=1).annotate(vcount=Count('session__viewprofile')).aggregate(Avg('vcount'))\\n{'vcount__avg': 9.3161000000000005}\\n>>> User.objects.annotate(ocount=Count('session__order')).filter(ocount__gte=1).annotate(vcount=Count('session__mousemove')).aggregate(Avg('vcount'))\\n{'vcount__avg': 28.590699999999998}\\n>>> User.objects.annotate(ocount=Count('session__order')).filter(ocount__gte=1).annotate(vcount=Count('session__question')).aggregate(Sum('vcount'))\\n{'vcount__sum': 5}\\n>>> User.objects.annotate(ocount=Count('session__order')).filter(ocount__gte=1).annotate(vcount=Count('session__click')).aggregate(Avg('vcount'))\\n{'vcount__avg': 3.9430000000000001}\\n\\nAfter removal:\\n>>> User.objects.count()\\n193\\n>>> Click.objects.count()\\n672\\n>>> Click.objects.values_list('session__user', 'tour').distinct().count()\\n226\\n>>> Question.objects.count()\\n5\\n>>> Question.objects.values_list('session__user', 'tour').distinct().count()\\n4\\n>>> Order.objects.count()\\n227\\n>>> Order.objects.values_list('session__user', 'tour').distinct().count()\\n194\\n>>> MouseMove.objects.count()\\n4570\\n>>> MouseMove.objects.values_list('session__user', 'tour').distinct().count()\\n551\\n>>> Session.objects.count()\\n333\\n>>> Tour.objects.count()\\n753\\n>>> TourType.objects.count()\\n6\\n>>> Country.objects.count()\\n50\\n>>> ViewProfile.objects.count()\\n1551\\n>>> ViewProfile.objects.values_list('session__user', 'tour').distinct().count()\\n619\\n>>> ViewProfile.objects.aggregate(Max('duration'))\\n{'duration__max': 1195}\\n>>> 1195/60\\n19\\n>>> ViewProfile.objects.aggregate(Avg('duration'))\\n{'duration__avg': 88.494500000000002}\\n>>> User.objects.annotate(hh=Count('session__click')).filter(hh=0).count()\\n67\\n>>> User.objects.annotate(hh=Count('session__mousemove')).filter(hh=0).count()\\n0\\n>>> User.objects.annotate(hh=Count('session__viewprofile')).filter(hh=0).count() \\n8\\n>>> Tour.objects.filter(viewprofile__isnull=False).distinct().count()\\n288\\n\\nAfter 0.2 division\\n45 test pairs selected from total 227 pairs.\\n>>> ViewProfile.objects.count()\\n1226\\n>>> MouseMove.objects.count()\\n3645\\n>>> Click.objects.count()\\n522\\n>>> Question.objects.count()\\n4\\n\\n\\n\\\"\\\"\\\"\\n\\ndef _parse_all(filename):\\n \\\"\\\"\\\"Parse the mega csv file\\\"\\\"\\\"\\n \\n filename = _get_abs_path(filename) \\n\\n # open the csv reader\\n reader = csv.reader(open(filename, \\\"rb\\\"), delimiter=separator, quoting=csv.QUOTE_NONE)\\n\\n # compile the pattern\\n pattern = re.compile(regexp_url)\\n \\n # open pages \\n open_pages = {}\\n \\n try:\\n for x, user_id, user_session_id, timestamp, action, action_parameter, object_ in reader:\\n \\n if reader.line_num % 5000 == 0:\\n print '%s lines processed' % reader.line_num \\n \\n cur_timestamp = parse(timestamp)\\n\\n # do the cleanup\\n open_pages = dict((k, view_page) for (k, view_page) in open_pages.items() \\n if (cur_timestamp - view_page.timestamp).seconds > decay_secs)\\n \\n # parse the ids\\n user_id = int(user_id)\\n \\n # take only users that bought something\\n if not (user_id in USERS_IDS_ORDER):\\n continue\\n \\n # find out whether it's done on an interesting object - tour.\\n # try matching the object, if it's not, go on\\n m = pattern.match(object_)\\n \\n if not m:\\n continue\\n\\n # try getting the action class, if not available, continue\\n try:\\n act_class = ACTIONS[action]\\n except KeyError:\\n continue\\n \\n # for object interaction it's a bit more difficult\\n if action == 'OBJECT_INTERACTION': \\n try:\\n act_class = act_class[action_parameter]\\n except KeyError:\\n continue\\n \\n \\n # prepare the user\\n #\\n\\n # parse the ids\\n user_session_id = int(user_session_id)\\n \\n # get or create the user, session\\n user, x = User.objects.get_or_create(id=user_id)\\n session, x = Session.objects.get_or_create(user=user, session_no=user_session_id)\\n \\n \\n # prepare the tour\\n #\\n \\n # get the matched stuff\\n country_name = m.group('country_name')\\n tour_type_name = m.group('tour_type')\\n tour_name = m.group('tour_name')\\n \\n # get or create the country, tour type.\\n country, x = Country.objects.get_or_create(name=country_name)\\n tour_type, x = TourType.objects.get_or_create(name=tour_type_name)\\n tour, x = Tour.objects.get_or_create(\\n name=tour_name, \\n defaults={\\n 'country': country,\\n 'tour_type': tour_type,\\n 'url': object_,\\n }\\n )\\n \\n # parse the date\\n timestamp = parse(timestamp)\\n \\n # open/close page need special handling\\n #\\n \\n if action == 'PAGE_OPEN':\\n \\n view_page = act_class(\\n session=session,\\n tour=tour,\\n timestamp=timestamp)\\n \\n # save the action to the dictionary waiting for the close event\\n open_pages[(tour, user)] = view_page\\n \\n continue\\n \\n if action == 'PAGE_CLOSE':\\n\\n # try finding the open, if not present, go ahead\\n try:\\n view_page = open_pages[(tour, user)]\\n except KeyError:\\n continue\\n \\n # count the duration and save the event if reasonable\\n view_page.duration = (timestamp - view_page.timestamp).seconds\\n\\n if view_page.duration < decay_secs:\\n view_page.save()\\n\\n continue\\n \\n # create the action\\n act_class.objects.create(\\n session=session,\\n tour=tour,\\n timestamp=timestamp)\\n except:\\n print \\\"error on line %d\\\" % reader.line_num\\n raise\\n\\n \\ndef _get_abs_path(filename):\\n \\\"\\\"\\\"Get the absolute path from the relative\\\"\\\"\\\"\\n return os.path.join(os.path.dirname(__file__), filename)\\n \\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5779221057891846,\n \"alphanum_fraction\": 0.5811688303947449,\n \"avg_line_length\": 41.573768615722656,\n \"blob_id\": \"ae8bdf6dbdf6d68324e4adf9ea944a09bec444db\",\n \"content_id\": \"3827e6e651659b6d0d80cecd983cb3378fa667d4\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5236,\n \"license_type\": \"no_license\",\n \"max_line_length\": 119,\n \"num_lines\": 122,\n \"path\": \"/UnresystCD/code/adapter/unresyst/aggregator/combining_aggregator.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The aggregator that uses a combinator to aggregate\\\"\\\"\\\"\\n\\nfrom base import BaseAggregator\\nfrom unresyst.constants import *\\nfrom unresyst.models.abstractor import RelationshipInstance, \\\\\\n PredictedRelationshipDefinition, Cluster, BiasInstance\\nfrom unresyst.models.aggregator import AggregatedRelationshipInstance, \\\\\\n AggregatedBiasInstance\\nfrom unresyst.combinator.combination_element import RelSimilarityCombinationElement, \\\\\\n ClusterSimilarityCombinationElement, BiasCombinationElement \\n\\nclass CombiningAggregator(BaseAggregator):\\n \\\"\\\"\\\"A class using unresyst.combinator for creating aggregates\\\"\\\"\\\"\\n \\n def aggregate_rules_relationships(self, recommender_model):\\n \\\"\\\"\\\"See the base class for documentation.\\n \\n Does only similarity aggregation + clusters, the preference aggregation is left out\\n for the algorithm.\\n \\\"\\\"\\\"\\n # get pairs that have some similarities\\n # \\n \\n # take all rule/relationship instances, that don't belong \\n # to the predicted_relationship - their id unique pairs\\n #\\n \\n predicted_def = PredictedRelationshipDefinition.objects.get(\\n recommender=recommender_model)\\n\\n qs_similarities = RelationshipInstance.objects\\\\\\n .exclude(definition=predicted_def)\\\\\\n .filter(definition__recommender=recommender_model)\\\\\\n .filter(definition__rulerelationshipdefinition__relationship_type__in=[\\n RELATIONSHIP_TYPE_OBJECT_OBJECT,\\n RELATIONSHIP_TYPE_SUBJECT_SUBJECT,\\n RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT]) \\n \\n # the unique pairs of entities between which there's some similarity\\n qs_id_pairs = qs_similarities\\\\\\n .values_list('subject_object1__id', 'subject_object2__id')\\\\\\n .distinct()\\n \\n # for all pairs that have some similarity\\n for id1, id2 in qs_id_pairs.iterator():\\n \\n combination_elements = []\\n\\n # similarities:\\n #\\n \\n # filter similarities that are for the pair\\n # only one direction is needed, because we're taking it from the same source\\n qs_pair_similarity = qs_similarities.filter(\\n subject_object1__id=id1, \\n subject_object2__id=id2)\\n \\n # save the entity type - there must be always at least\\n # one similarity relationship for the pair, so this is\\n # always filled.\\n relationship_type = qs_pair_similarity[0].definition.as_leaf_class().relationship_type\\n \\n # create combining elements from them\\n for pair_sim in qs_pair_similarity.iterator():\\n\\n el = RelSimilarityCombinationElement(rel_instance=pair_sim)\\n combination_elements.append(el)\\n \\n # clusters:\\n #\\n \\n # get clusters the two have in common\\n qs_common_clusters = Cluster.objects.filter(clustermember__member__id=id1)\\\\\\n .filter(clustermember__member__id=id2)\\n \\n # create the compiling elements from them\\n for c in qs_common_clusters.iterator():\\n \\n el = ClusterSimilarityCombinationElement(\\n cluster_members=[\\n c.clustermember_set.get(member__id=id1),\\n c.clustermember_set.get(member__id=id2)]\\n )\\n combination_elements.append(el) \\n \\n # aggregation and save:\\n #\\n \\n # aggregate the similarity through the combinator\\n aggr = self.combinator.combine_pair_similarities(\\n combination_elements=combination_elements)\\n \\n # create and save the aggregated similarity instance\\n aggr.subject_object1_id = id1\\n aggr.subject_object2_id = id2\\n aggr.recommender = recommender_model\\n aggr.relationship_type = relationship_type\\n aggr.save()\\n\\n \\n def aggregate_biases(self, recommender_model):\\n \\\"\\\"\\\"See the base class for documentation.\\n \\\"\\\"\\\"\\n \\n # all available biases\\n qs_biases = BiasInstance.objects.filter(definition__recommender=recommender_model)\\n \\n # get entities that have some biases: subject/object anything\\n qs_ids = qs_biases.values_list('subject_object__id', flat=True).distinct()\\n \\n for ent_id in qs_ids:\\n \\n # get the biases for the entity\\n ent_biases = [BiasCombinationElement(bias_instance=b) for b in qs_biases.filter(subject_object__id=ent_id)]\\n \\n # throw them to the combinator\\n aggr = self.combinator.combine_entity_biases(entity_biases=ent_biases)\\n \\n # fill the missing fields and save\\n aggr.subject_object_id = ent_id\\n aggr.recommender = recommender_model\\n aggr.save()\\n \\n \\n\"\n },\n {\n \"alpha_fraction\": 0.5399905443191528,\n \"alphanum_fraction\": 0.5416469573974609,\n \"avg_line_length\": 39.08571243286133,\n \"blob_id\": \"f38ea2de0a428fe42518e9412d418ab5fea47683\",\n \"content_id\": \"de94d585369e1dd52723d67f7d3d2341ba290dcf\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 4226,\n \"license_type\": \"no_license\",\n \"max_line_length\": 125,\n \"num_lines\": 105,\n \"path\": \"/UnresystCD/code/adapter/unresyst/compilator/combining_compilator.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The combining compilator class\\\"\\\"\\\"\\n\\nfrom base import BaseCompilator\\nfrom unresyst.constants import *\\nfrom unresyst.models.common import SubjectObject\\nfrom unresyst.exceptions import RecommenderBuildError\\n\\nclass CombiningCompilator(BaseCompilator):\\n \\\"\\\"\\\"The compilator using the given combinator to combine the predictions\\n that it reveals\\n \\\"\\\"\\\" \\n \\n def __init__(\\n self, \\n combinator, \\n depth=DEFAULT_COMPILATOR_DEPTH, \\n breadth=DEFAULT_COMPILATOR_BREADTH,\\n pair_depth=DEFAULT_COMPILATOR_PAIR_DEPTH):\\n \\\"\\\"\\\"The initializer, combinator is not optional.\\\"\\\"\\\" \\n \\n super(CombiningCompilator, self).__init__(\\n combinator=combinator, \\n depth=depth, \\n breadth=breadth,\\n pair_depth=pair_depth)\\n \\n def compile_prediction(self, recommender_model, dn_subject, dn_object):\\n \\\"\\\"\\\"Create a prediction using all available information and the instance combinator.\\n \\n @type recommender_model: models.common.Recommender\\n @param recommender_model: the recommender model\\n \\n @type dn_subject: SubjectObject\\n @param dn_subject: the domain neutral subject\\n \\n @type dn_object: SubjectObject\\n @param dn_object: the domain neutral object\\n \\n @rtype: RelationshipPredictionInstance\\n @return: the prediction from what we know, if we don't know anything return None.\\n \\\"\\\"\\\"\\n # find all we know about the subject - object pair - biases, similarities, clusters,... \\n els = self.get_pair_combination_elements(dn_subject=dn_subject, dn_object=dn_object)\\n \\n if not els:\\n return None\\n \\n # pass it all to the combinator to get the prediction\\n pred = self.combinator.combine_pair_prediction_elements(combination_elements=els)\\n\\n # fill the missing fields in the prediction and save\\n pred.recommender = recommender_model\\n pred.subject_object1 = dn_subject\\n pred.subject_object2 = dn_object\\n return pred\\n\\n \\n def compile_all(self, recommender_model):\\n \\\"\\\"\\\"Compile preferences, known relationships + similarities.\\n \\\"\\\"\\\"\\n \\n if not self.breadth:\\n return\\n \\n # go through the sujbects (or subjectobjects)\\n #\\n \\n subject_ent_type = ENTITY_TYPE_SUBJECTOBJECT if recommender_model.are_subjects_objects \\\\\\n else ENTITY_TYPE_SUBJECT \\n \\n qs_subjects = SubjectObject.objects.filter(\\n recommender=recommender_model, \\n entity_type=subject_ent_type)\\n\\n print \\\" Compiling predictions for %d subjects.\\\" % qs_subjects.count()\\n i = 0\\n \\n for subj in qs_subjects.iterator():\\n \\n # get the most promising objects for the subject\\n promising_objects = self.combinator.choose_promising_objects(\\n dn_subject=subj, \\n min_count=self.breadth) \\n \\n if i % 20 == 0:\\n print \\\" %d subjects processed. Current promising object count: %d\\\" % (i, len(promising_objects))\\n i += 1\\n \\n # go through the promising objects\\n for obj in promising_objects:\\n \\n # compile the predition and save it\\n pred = self.compile_prediction(\\n recommender_model=recommender_model,\\n dn_subject=subj,\\n dn_object=obj)\\n\\n # if the compilate is empty, it's an error \\n if not pred:\\n raise RecommenderBuildError(\\n message=\\\"Nothing found for the pair %s, %s, although the object was in promising.\\\" \\\\\\n % (subj, obj),\\n recommender=recommender_model)\\n \\n pred.save()\\n \\n\"\n },\n {\n \"alpha_fraction\": 0.6433120965957642,\n \"alphanum_fraction\": 0.7579618096351624,\n \"avg_line_length\": 50,\n \"blob_id\": \"913794b8d0e2e76a7cd46eeed143ad41269ec4e2\",\n \"content_id\": \"5cb1ba6533861bde80ac13036b5c8d536c48e490\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 157,\n \"license_type\": \"no_license\",\n \"max_line_length\": 100,\n \"num_lines\": 3,\n \"path\": \"/UnresystCD/code/dataset_scripts/process_lastfm.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\nsed -n '1~750p' /home/pcv/Desktop/userid-timestamp-artid-artname-traid-traname.tsv > tracks30000.tsv\\n\\nhead -200000 artist_tags.tsv > artist_tags200K.tsv \\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.7053459286689758,\n \"alphanum_fraction\": 0.7182297110557556,\n \"avg_line_length\": 43.11070251464844,\n \"blob_id\": \"a9d168831822b1242bfbc717004f4fb39d178621\",\n \"content_id\": \"f6e44ad26ed136b9bcb9d8cb905bf8ad1b561740\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 11953,\n \"license_type\": \"no_license\",\n \"max_line_length\": 496,\n \"num_lines\": 271,\n \"path\": \"/docs/CreateMahoutRecommender.md\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Mahout for Newbies: How to Create a Recommender #\\nA short tutorial on actions you have to take to create your first Mahout recommender.\\n\\n## Introduction ##\\n\\n[Mahout](http://mahout.apache.org/) is a library containing scalable implementation of many machine learning algorithms, notably the collaborative filtering and other algorithms used in recommender systems.\\n\\nThere are many tutorials on creating simple or more complex recommenders but for Mahout newbies it's quite difficult to cope with all the technical details you have to do to make Mahout work in your way.\\n\\nI'm not an Eclipse fan, so all instructions use only the shell console. I was running the examples on Ubuntu 10.04, but it should be working on most of the linux distros and even on windows with cygwin.\\n\\nThis page is aiming to be a step-by-step description for Mahout/Maven beginners. I'm also a Mahout/Maven beginner, so any corrections and comments are more then welcome.\\n\\n## Part 1: Installation ##\\nThere's a good [installation guide in the Mahout wiki](https://cwiki.apache.org/confluence/display/MAHOUT/BuildingMahout).\\n\\nI'm running the 0.4 Mahout release, so all stuff in this article is primarily for that, but should be working also on trunk and other releases.\\n\\nAs described in the guide, after JDK and Maven you should do something like:\\n```\\ncd core\\nmvn compile\\nmvn install\\n```\\n\\nA few things to note here:\\n * If you installation fails, try the fix proposed in the tutorial:\\n```\\nrm -rf ~/.m2/\\nmvn clean install\\n```\\n * If this doesn't help, try again\\n```\\nmvn install\\n```\\nand look to the `core/target/surefire-reports/`. If you can find the cause of the problem there, you're a lucky person. If not, you can seek help on the `user@mahout.apache.org` mailing list as [I successfully did.](http://comments.gmane.org/gmane.comp.apache.mahout.user/5881)\\n * Alternatively, you can try running the install without tests\\n```\\nmvn -Dmaven.test.skip=true install\\n```\\n\\n## Part 2: Creating a simple recommender ##\\n\\nNow that you have mahout successfully installed, you can try some built-in examples, as [described in the Mahout wiki](https://cwiki.apache.org/confluence/display/MAHOUT/RecommendationExamples). Firstly you have to download some dataset, and then pass it as the `-i` parameter. It's a kind of a detective work but you can find some hints in the source code of the examples. I've successfully run the recommender on [the Book Crossing dataset](http://www.informatik.uni-freiburg.de/~cziegler/BX/)\\n\\nAnd now to creating your own recommender on your own data. Editing the examples isn't a good idea, because the code gets quite messy and all the paths are terribly long. Instead of that follow the oncoming steps.\\n\\n### Create a Maven project ###\\nFirstly, you have to create a maven project. Go to the Mahout directory (the parent of the core directory) and run something like:\\n```\\nmvn archetype:create -DarchetypeGroupId=org.apache.maven.archetypes -DgroupId=com.unresyst -DartifactId=mahoutrec\\n```\\nThis creates an empty project called `mahoutrec` with the package namespace `com.unresyst`. Now change to the `mahoutrec` directory. You can try out the new project by running:\\n```\\nmvn compile\\nmvn exec:java -Dexec.mainClass=\\\"com.unresyst.App\\\"\\n```\\nThis should print the Hello world message. If not try passing the `-e` option for the second command and try to realize what's going wrong.\\n\\nMore info on creating a Maven project can be found in [the official guide](http://maven.apache.org/guides/getting-started/)\\n\\n### Set the project dependencies ###\\nNow that you have your maven project, let's set it up for your Mahout instalation.\\n\\nOpen the `pom.xml` file from your project directory (`mahoutrec`) in your favourite text editor. And now:\\n * add the Mahout dependencies to your `pom.xml`. The following `` elements must be under the `` parent element.\\n```\\n \\n org.apache.mahout \\n mahout-core \\n 0.4 \\n \\n \\n org.apache.mahout \\n mahout-math \\n 0.4 \\n \\n \\n org.apache.mahout \\n mahout-math \\n 0.4 \\n test-jar \\n test \\n \\n \\n org.apache.mahout \\n mahout-utils \\n 0.4 \\n \\n```\\n\\n> If you are using other than the 0.4 version replace the 0.4 by the number of your version.\\n * Set the relative path to the parent project `pom.xml` by adding the `` element to the `` element:\\n```\\n ../pom.xml \\n```\\nYour whole `pom.xml` should look like that:\\n```\\n\\n\\n 4.0.0 \\n \\n mahout \\n org.apache.mahout \\n 0.4 \\n ../pom.xml \\n \\n com.unresyst \\n mahoutrec \\n 1.0-SNAPSHOT \\n mahoutrec \\n http://maven.apache.org \\n \\n \\n junit \\n junit \\n 3.8.1 \\n test \\n \\n \\n org.apache.mahout \\n mahout-core \\n 0.4 \\n \\n \\n org.apache.mahout \\n mahout-math \\n 0.4 \\n \\n \\n org.apache.mahout \\n mahout-math \\n 0.4 \\n test-jar \\n test \\n \\n \\n org.apache.mahout \\n mahout-utils \\n 0.4 \\n \\n \\n \\n UTF-8 \\n \\n \\n```\\n\\nNow try compiling your project by running\\n```\\nmvn compile\\n```\\nYou should get a Build successful message.\\n\\n### Create a simple recommender ###\\nAnd now to the promised creation of your own recommender running on your own data. Firstly we'll create some dummy data to feed the recommender. Create a `datasets` directory in your project directory and create the `dummy-bool.csv` file with the following contents:\\n```\\n#userId,itemId\\n1,3\\n1,4\\n2,44\\n2,46\\n3,3\\n3,5\\n3,6\\n4,3\\n4,5\\n4,11\\n4,44\\n5,1\\n5,2\\n5,4\\n```\\nThe first line is a comment, all the others are boolean preferences for items expressed by the users. Boolean preferences are useful when you don't have any explicit user-item ratings in your data. A boolean preference can mean e.g. that the user has viewed a profile page of the item. The example file states that the user id `1` has viewed the item id `3`, and item id `4`, user `2` item `44` and `46`, and so on.\\n\\nNow, let's create the recommender. Create a file UnresystBoolRecommend.java in the well hidden subddirectory `src/main/java/com/unresyst/` of your project directory. Place the following content to the file\\n```\\npackage com.unresyst;\\n\\nimport java.io.File;\\nimport java.io.FileNotFoundException;\\nimport java.util.List;\\nimport java.io.IOException;\\n\\nimport org.apache.commons.cli2.OptionException; \\nimport org.apache.mahout.cf.taste.common.TasteException;\\nimport org.apache.mahout.cf.taste.impl.model.file.FileDataModel;\\nimport org.apache.mahout.cf.taste.impl.recommender.CachingRecommender;\\nimport org.apache.mahout.cf.taste.impl.recommender.slopeone.SlopeOneRecommender;\\nimport org.apache.mahout.cf.taste.model.DataModel;\\nimport org.apache.mahout.cf.taste.recommender.RecommendedItem;\\nimport org.apache.mahout.cf.taste.impl.common.LongPrimitiveIterator;\\n\\npublic class UnresystBoolRecommend {\\n \\n public static void main(String... args) throws FileNotFoundException, TasteException, IOException, OptionException {\\n \\n // create data source (model) - from the csv file \\n File ratingsFile = new File(\\\"datasets/dummy-bool.csv\\\"); \\n DataModel model = new FileDataModel(ratingsFile);\\n \\n // create a simple recommender on our data\\n CachingRecommender cachingRecommender = new CachingRecommender(new SlopeOneRecommender(model));\\n\\n // for all users\\n for (LongPrimitiveIterator it = model.getUserIDs(); it.hasNext();){\\n long userId = it.nextLong();\\n \\n // get the recommendations for the user\\n List recommendations = cachingRecommender.recommend(userId, 10);\\n \\n // if empty write something\\n if (recommendations.size() == 0){\\n System.out.print(\\\"User \\\");\\n System.out.print(userId);\\n System.out.println(\\\": no recommendations\\\");\\n }\\n \\n // print the list of recommendations for each \\n for (RecommendedItem recommendedItem : recommendations) {\\n System.out.print(\\\"User \\\");\\n System.out.print(userId);\\n System.out.print(\\\": \\\");\\n System.out.println(recommendedItem);\\n }\\n } \\n }\\n}\\n```\\nSome notes to that:\\n * If Java isn't your tongue language, be aware of the package declaration on the first line. If you omit that the Maven runner will never find your class.\\n * In the main function we first create a data model from our csv file, then we instantiate a simple Slope One recommender filled with the data. Then we iterate over all users in the dataset and we print recommendations for them.\\n * Sorry for my poor java.\\n\\nTo run the recommender on our dummy data, do:\\n```\\nmvn compile\\n```\\nThis should end up by BUILD SUCCESSFUL. Then run the main function:\\n```\\nmvn exec:java -Dexec.mainClass=\\\"com.unresyst.UnresystBoolRecommend\\\" \\n```\\n\\nIn the middle of the Maven spam you should be able to see:\\n```\\nUser 1: RecommendedItem[item:5, value:1.0]\\nUser 2: RecommendedItem[item:5, value:1.0]\\nUser 2: RecommendedItem[item:3, value:1.0]\\nUser 3: no recommendations\\nUser 4: no recommendations\\nUser 5: RecommendedItem[item:5, value:1.0]\\nUser 5: RecommendedItem[item:3, value:1.0]\\n```\\nThese are recommendations for our users. Our recommender recommends the item `5` to the user `1`, with certainty 1.0 (this is a constant value for boolean recommenders), items `5` and `3` to user 2 and so on.\\n\\nTo reduce the Maven spam you can pass the `-q` option, to debug pass the `-e` option.\\n\\n## Conclusion ##\\nThat's it, now you've got your first Mahout recommender. You can now experiment with various data and various recommenders as described in other Mahout recommender tutorials. Some of the good ones:\\n * A good starting guide, from which the example is partly overtaken: http://philippeadjiman.com/blog/2009/11/11/flexible-collaborative-filtering-in-java-with-mahout-taste/\\n * The definitive IBM guide: http://www.ibm.com/developerworks/java/library/j-mahout/\\n\\nThanks for reading, any suggestions, corrections and comments are appreciated.\\n\\nSources and further reading\\n * https://cwiki.apache.org/confluence/display/MAHOUT/Mahout+Wiki\\n * https://cwiki.apache.org/confluence/display/MAHOUT/BuildingMahout\\n * http://maven.apache.org/plugins/maven-surefire-plugin/howto.html\\n * http://www.jarvana.com/jarvana/view/org/apache/mahout/mahout-core/0.4/mahout-core-0.4-javadoc.jar!/index-all.html\\n * http://shuyo.wordpress.com/2011/02/14/mahout-development-environment-with-maven-and-eclipse-2/\"\n },\n {\n \"alpha_fraction\": 0.5778035521507263,\n \"alphanum_fraction\": 0.5778035521507263,\n \"avg_line_length\": 38.21590805053711,\n \"blob_id\": \"31bc717d67fe630404e9b123a8fffacc887f435a\",\n \"content_id\": \"b3eb7cf40522a9aca1ea39fc12659a7f44225cba\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3451,\n \"license_type\": \"no_license\",\n \"max_line_length\": 108,\n \"num_lines\": 88,\n \"path\": \"/UnresystCD/code/adapter/unresyst/algorithm/compiling_algorithm.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The CompilingAlgorithm class\\\"\\\"\\\"\\n\\nfrom base import BaseAlgorithm\\n\\nclass CompilingAlgorithm(BaseAlgorithm):\\n \\\"\\\"\\\"The algorithm that compiles aggregated similarities and biases with\\n the predictions.\\n \\\"\\\"\\\"\\n def __init__(self, inner_algorithm, compilator):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n super(CompilingAlgorithm, self).__init__(inner_algorithm=inner_algorithm)\\n \\n self.compilator=compilator\\n \\\"\\\"\\\"The compilator that will be used during the build\\\"\\\"\\\"\\n\\n def build(self, recommender_model):\\n \\\"\\\"\\\"See the base class for documentation.\\n \\n Compiles and calls the inner algorithm build\\n \\\"\\\"\\\" \\n print \\\" Compiling aggregates and predictions.\\\"\\n \\n self.compilator.compile_all(recommender_model) \\n \\n\\n print \\\"Predictions compiled. Building the inner algorithm...\\\"\\n \\n super(CompilingAlgorithm, self).build(recommender_model=recommender_model) \\n\\n \\n def get_relationship_prediction(self, recommender_model, dn_subject, dn_object, remove_predicted):\\n \\\"\\\"\\\"See the base class for the documentation.\\n \\n Get all available info for the pair.\\n \\\"\\\"\\\"\\n \\n # get what we have from the inner algo\\n inner_prediction = super(CompilingAlgorithm, self).get_relationship_prediction(\\n recommender_model=recommender_model,\\n dn_subject=dn_subject,\\n dn_object=dn_object,\\n remove_predicted=remove_predicted)\\n \\n # if it's trivial, than return it\\n if remove_predicted and inner_prediction.is_trivial:\\n return inner_prediction\\n \\n # if it's not uncertain, it was already available for the inner algorithm,\\n # return it\\n if not inner_prediction.is_uncertain:\\n return inner_prediction\\n \\n # otherwise compile the prediction from all available info \\n prediction = self.compilator.compile_prediction(\\n recommender_model=recommender_model,\\n dn_subject=dn_subject,\\n dn_object=dn_object)\\n\\n # if it found something, return it \\n if prediction: \\n return prediction\\n \\n # otherwise return the uncertain\\n return self._get_uncertain_prediction(\\n recommender_model=recommender_model, \\n dn_subject=dn_subject, \\n dn_object=dn_object\\n ) \\n \\n def get_recommendations(self, recommender_model, dn_subject, count, expectancy_limit, remove_predicted):\\n \\\"\\\"\\\"See the base class for the documentation. \\n\\n Here: fill the recommendations until the count\\n with random predictions, fill all info we know, or create uncertain \\n predictions.\\n \\\"\\\"\\\"\\n # get what we have from the inner algo\\n inner_recs = super(CompilingAlgorithm, self).get_recommendations(\\n recommender_model=recommender_model,\\n dn_subject=dn_subject,\\n count=count,\\n expectancy_limit=expectancy_limit,\\n remove_predicted=remove_predicted)\\n\\n # add there some random and find info for them \\n # to samy jako nahore\\n return inner_recs\\n\"\n },\n {\n \"alpha_fraction\": 0.5001169443130493,\n \"alphanum_fraction\": 0.5038596391677856,\n \"avg_line_length\": 35.85344696044922,\n \"blob_id\": \"7c921da5cce0e18772c9d5d422e47874ee9e14b0\",\n \"content_id\": \"db3e6d7cc76017b8f98f4769ac32aca2d7aac2a7\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 4275,\n \"license_type\": \"no_license\",\n \"max_line_length\": 150,\n \"num_lines\": 116,\n \"path\": \"/UnresystCD/code/adapter/unresyst/recommender/rank_evaluation.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Evaluator defining a method for the average prediction rank\\\"\\\"\\\"\\n\\nfrom unresyst.exceptions import EmptyTestSetError\\nfrom unresyst.constants import *\\nfrom unresyst.recommender.evaluation import BaseEvaluator\\n\\ndef index(seq, f):\\n \\\"\\\"\\\"Return the index of the first item in seq where f(item) == True.\\\"\\\"\\\"\\n return next((i for i in xrange(len(seq)) if f(seq[i])), None)\\n\\n\\nclass RankEvaluator(BaseEvaluator):\\n \\\"\\\"\\\"The evaluator for counting the average prediction rank\\\"\\\"\\\"\\n \\n SUBJ_IDS = []\\n \\\"\\\"\\\"A list of subjects (domain specific ids) that are to be tested\\\"\\\"\\\"\\n \\n \\n @classmethod\\n def evaluate_predictions(cls, recommender, save_predictions=False):\\n \\\"\\\"\\\"See the base class for documentation.\\n \\n Here it counts the rank, without calling any metric.\\n \\\"\\\"\\\"\\n \\n # get the test pairs\\n qs_pairs = cls._get_cleared_pairs()\\n \\n # evaluate the rank for each subject\\n # \\n\\n obj_count = recommender.objects.count()\\n rank_sum = 0\\n rank_count = 0\\n\\n print \\\"Evaluating %d subjects\\\" % len(cls.SUBJ_IDS) \\n \\n \\n # go through the wished subjects\\n for subj, i in zip(recommender.subjects.filter(pk__in=cls.SUBJ_IDS), range(1, len(cls.SUBJ_IDS))):\\n \\n # if the test set for the subject is emty go ahead\\n test_set_subj = qs_pairs.filter(subj=subj) \\n if not test_set_subj:\\n continue\\n \\n # create an array of object predictions for the subject\\n #\\n neg_array = []\\n pos_array = []\\n uncertain_count = 0 \\n \\n for obj in recommender.objects.all():\\n \\n # get the prediction\\n exp_prediction = recommender.predict_relationship(subj, obj)\\n \\n if exp_prediction is None or exp_prediction.is_uncertain: \\n uncertain_count += 1\\n \\n # otherwise add it to the right array\\n else:\\n arr = neg_array if exp_prediction.expectancy > UNCERTAIN_PREDICTION_VALUE \\\\\\n else pos_array\\n \\n arr.append((obj, exp_prediction.expectancy))\\n\\n print \\\"all predicted\\\"\\n\\n # sort the array by expectancy\\n pos_array.sort(key=lambda el: el[1], reverse=True) \\n neg_array.sort(key=lambda el: el[1], reverse=True) \\n\\n print \\\"all sorted\\\" \\n pos_count = len(pos_array)\\n neg_count = len(neg_array)\\n \\n # count the rank that will be given to uncertains.\\n if uncertain_count:\\n \\n # the rank of the first uncertain\\n rank_first = float(pos_count) / (obj_count -1)\\n \\n # the rank of the last uncertain\\n rank_last = float(pos_count + uncertain_count) / (obj_count - 1)\\n \\n uncertain_rank = (rank_first + rank_last) / 2\\n \\n rank_count += test_set_subj.count()\\n \\n # count the rank for all objects for the subjects in the test set\\n for obj in test_set_subj:\\n \\n # try finding it in positives\\n ipos = index(pos_array, lambda el: el[0] == obj)\\n if not (ipos is None):\\n rank_sum += float(ipos) / (obj_count - 1)\\n continue\\n \\n # try finding it in negatives\\n ineg = index(neg_array, lambda el: el[0] == obj)\\n if not (ineg is None):\\n rank_sum += float(pos_count + uncertain_count + ineg) / (obj_count - 1)\\n continue\\n \\n # otherwise it's uncertain\\n rank_sum += uncertain_rank\\n \\n print \\\"%d subjects evaluated. current avg rank: %f, current rank count: %d\\\" % (i, rank_sum/rank_count if rank_count else -1.0, rank_count)\\n \\n\\n res = rank_sum / rank_count\\n \\n print \\\"Average rank: %f\\\" % res\\n \\n return res\\n\"\n },\n {\n \"alpha_fraction\": 0.8055555820465088,\n \"alphanum_fraction\": 0.8055555820465088,\n \"avg_line_length\": 70,\n \"blob_id\": \"ff968e097fa8518be7d988877b40629eb655decc\",\n \"content_id\": \"435580c197e421c8796039673b75a2a29be25919\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 72,\n \"license_type\": \"no_license\",\n \"max_line_length\": 70,\n \"num_lines\": 1,\n \"path\": \"/UnresystCD/code/mahout/mahoutrec/unresystboolrecommend.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"mvn -q exec:java -Dexec.mainClass=\\\"com.unresyst.UnresystBoolRecommend\\\" \\n\"\n },\n {\n \"alpha_fraction\": 0.653235673904419,\n \"alphanum_fraction\": 0.6556776762008667,\n \"avg_line_length\": 34.0428581237793,\n \"blob_id\": \"c55413980319a261b4eb939e2134eb85490aabc3\",\n \"content_id\": \"9496068ad98a18b21d0150eb245db04ed942f8c8\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2457,\n \"license_type\": \"no_license\",\n \"max_line_length\": 82,\n \"num_lines\": 70,\n \"path\": \"/UnresystCD/code/adapter/unresyst/models/aggregator.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Models which instances are created by the aggregator package.\\n\\nThe aggregated relationship.\\n\\\"\\\"\\\"\\n\\nfrom django.db import models\\n\\nfrom base import BaseRelationshipInstance\\nfrom unresyst.constants import *\\n\\nclass AggregatedRelationshipInstance(BaseRelationshipInstance):\\n \\\"\\\"\\\"A representation of an aggregated relationship between two subject/objects\\n \\n There can be only one aggregated relationship for each subject/object pair\\n for a recommender.\\n \\\"\\\"\\\"\\n\\n expectancy = models.FloatField()\\n \\\"\\\"\\\"The probability of the relationship between subject/object.\\n A number from [0, 1].\\n \\\"\\\"\\\"\\n \\n recommender = models.ForeignKey('unresyst.Recommender')\\n \\\"\\\"\\\"The recommender it belongs to\\\"\\\"\\\"\\n \\n relationship_type = models.CharField(max_length=MAX_LENGTH_RELATIONSHIP_TYPE, \\n choices=RELATIONSHIP_TYPE_CHOICES)\\n \\\"\\\"\\\"A string indicating whether it's a S-O, O-O, S-S, or SO-SO relationship.\\\"\\\"\\\"\\n \\n additional_unique = ('recommender', )\\n \\\"\\\"\\\"There can be multiple pairs for one recommender\\\"\\\"\\\"\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\" \\n ret = super(AggregatedRelationshipInstance, self).__unicode__()\\n\\n return ret + \\\", %f\\\" % (self.expectancy) \\n \\n class Meta:\\n app_label = 'unresyst' \\n \\n unique_together = ('subject_object1', 'subject_object2', 'recommender')\\n \\\"\\\"\\\"For each recommender there can be only one subject-object pair.\\\"\\\"\\\"\\n \\nclass AggregatedBiasInstance(models.Model):\\n \\\"\\\"\\\"An aggregated bias of a subjectobject\\\"\\\"\\\" \\n \\n expectancy = models.FloatField()\\n \\\"\\\"\\\"The probability of the subject/object being in the predicted_relationship\\n A number from [0, 1].\\n \\\"\\\"\\\"\\n \\n subject_object = models.ForeignKey('unresyst.SubjectObject')\\n \\\"\\\"\\\"The biased subject/object.\\\"\\\"\\\" \\n \\n recommender = models.ForeignKey('unresyst.Recommender')\\n \\\"\\\"\\\"The recommender it belongs to\\\"\\\"\\\"\\n \\n description = models.TextField(default='', blank=True)\\n \\\"\\\"\\\"The filled description of the aggregated bias.\\\"\\\"\\\" \\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\" \\n return u\\\"%s: %f\\\" % (self.subject_object, self.expectancy)\\n \\n class Meta:\\n app_label = 'unresyst' \\n \\n unique_together = ('subject_object', 'recommender')\\n \\\"\\\"\\\"For each recommender there can be only one subject-object pair.\\\"\\\"\\\" \\n\"\n },\n {\n \"alpha_fraction\": 0.6512749791145325,\n \"alphanum_fraction\": 0.6529979109764099,\n \"avg_line_length\": 41.67647171020508,\n \"blob_id\": \"bfc9e424f3f5bdfbb30e9355241bfcb7ef2d306e\",\n \"content_id\": \"ed6202f65567939db68bc509050d78c74a0f14f9\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Java\",\n \"length_bytes\": 2902,\n \"license_type\": \"no_license\",\n \"max_line_length\": 129,\n \"num_lines\": 68,\n \"path\": \"/code/mahout/mahoutrec/src/main/java/com/unresyst/UnresystRecommend.java\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"package com.unresyst;\\n\\nimport java.io.File;\\nimport java.io.FileNotFoundException;\\nimport java.util.List;\\nimport java.io.IOException;\\nimport java.io.PrintWriter;\\nimport java.io.OutputStreamWriter;\\nimport java.io.FileOutputStream;\\nimport java.nio.charset.Charset;\\n\\nimport org.apache.commons.cli2.OptionException; \\nimport org.apache.mahout.cf.taste.common.TasteException;\\nimport org.apache.mahout.cf.taste.impl.model.file.FileDataModel;\\nimport org.apache.mahout.cf.taste.impl.recommender.CachingRecommender;\\nimport org.apache.mahout.cf.taste.impl.recommender.slopeone.SlopeOneRecommender;\\nimport org.apache.mahout.cf.taste.model.DataModel;\\nimport org.apache.mahout.cf.taste.recommender.RecommendedItem;\\nimport org.apache.mahout.cf.taste.impl.common.LongPrimitiveIterator;\\nimport org.apache.mahout.common.FileLineIterable;\\nimport org.apache.mahout.common.IOUtils;\\n\\npublic class UnresystRecommend {\\n \\n public static void main(String[] args) throws FileNotFoundException, TasteException, IOException, OptionException {\\n\\n // parameters: train data filename, number of recommendations per user , output filename \\n String trainFilename = args[0]; \\n int recCount = Integer.parseInt(args[1]);\\n String outputFilename = args[2];\\n \\n // create data source (model) - from the train csv file \\n File ratingsFile = new File(trainFilename); \\n DataModel model = new FileDataModel(ratingsFile); \\n \\n // create a simple recommender on our data\\n CachingRecommender cachingRecommender = new CachingRecommender(new SlopeOneRecommender(model));\\n \\n // create recommendations for all users and write them to the output file\\n //\\n \\n // open the file to write to\\n File resultFile = new File(outputFilename);\\n PrintWriter writer = new PrintWriter(new OutputStreamWriter(new FileOutputStream(resultFile), Charset.forName(\\\"UTF-8\\\")));\\n \\n // for all users\\n for (LongPrimitiveIterator it = model.getUserIDs(); it.hasNext();){\\n long userId = it.nextLong();\\n \\n // get the recommendations for the user\\n List recommendations = cachingRecommender.recommend(userId, recCount); \\n \\n \\n // print the list of recommendations for each recommendation\\n for (RecommendedItem recommendedItem : recommendations) {\\n\\n long itemId = recommendedItem.getItemID();\\n float prediction = recommendedItem.getValue();\\n \\n writer.println(String.format(\\\"%d,%d,%f\\\", userId, itemId, prediction));\\n }\\n }\\n \\n // writer spam\\n writer.flush();\\n IOUtils.quietClose(writer); \\n }\\n}\\n\"\n },\n {\n \"alpha_fraction\": 0.6050138473510742,\n \"alphanum_fraction\": 0.6061981916427612,\n \"avg_line_length\": 28.97633171081543,\n \"blob_id\": \"621e8ed2a71bc636df4a679453b9be65b4736211\",\n \"content_id\": \"68f32ee383b09220c15fbba11b7af8ff59160436\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5066,\n \"license_type\": \"no_license\",\n \"max_line_length\": 95,\n \"num_lines\": 169,\n \"path\": \"/code/adapter/travel/models.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Models for the travel agency dataset\\\"\\\"\\\"\\n\\nfrom django.db import models\\n\\n\\nfrom constants import *\\nfrom unresyst.models import BaseEvaluationPair\\n\\nclass User(models.Model):\\n \\\"\\\"\\\"The user\\\"\\\"\\\"\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return 'user_%d' % self.id \\n\\nclass Session(models.Model):\\n \\\"\\\"\\\"User's session with the system\\\"\\\"\\\"\\n \\n user = models.ForeignKey('travel.User')\\n \\\"\\\"\\\"The user\\\"\\\"\\\"\\n \\n session_no = models.PositiveIntegerField()\\n \\\"\\\"\\\"The number of session for the user\\\"\\\"\\\"\\n \\n class Meta:\\n unique_together = ('user', 'session_no')\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return u'user %s: session: %d' % (self.user, self.session_no)\\n \\nclass Country(models.Model):\\n \\\"\\\"\\\"The country for a tour.\\\"\\\"\\\"\\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME, unique=True)\\n \\\"\\\"\\\"The name of the country\\\"\\\"\\\"\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n\\n\\nclass TourType(models.Model):\\n \\\"\\\"\\\"A model for a type of the tour.\\\"\\\"\\\" \\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME, unique=True)\\n \\\"\\\"\\\"The name of the type.\\\"\\\"\\\" \\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n\\n\\nclass Tour(models.Model):\\n \\\"\\\"\\\"The tour\\\"\\\"\\\" \\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME, unique=True)\\n \\\"\\\"\\\"The name\\\"\\\"\\\"\\n \\n url = models.URLField(max_length=MAX_LENGTH_URL, verify_exists=False)\\n \\\"\\\"\\\"The url without the final id and without GET parameters\\\"\\\"\\\"\\n \\n country = models.ForeignKey('travel.Country')\\n \\\"\\\"\\\"The country the tour is to\\\"\\\"\\\"\\n \\n tour_type = models.ForeignKey('travel.TourType')\\n \\\"\\\"\\\"The type of the tour\\\"\\\"\\\"\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n\\n\\nclass Action(models.Model):\\n \\\"\\\"\\\"An abstract class for a user action\\\"\\\"\\\"\\n\\n session = models.ForeignKey('travel.Session')\\n \\\"\\\"\\\"The session in which the action was taken (identifies the user)\\\"\\\"\\\"\\n \\n tour = models.ForeignKey('travel.Tour')\\n \\\"\\\"\\\"The tour on which the action was taken\\\"\\\"\\\"\\n \\n timestamp = models.DateTimeField()\\n \\\"\\\"\\\"The date and time the track was played\\\"\\\"\\\"\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return u\\\"%s: %s\\\" % (self.session.user, self.tour)\\n \\n class Meta:\\n abstract = True \\n \\nclass Order(Action):\\n \\\"\\\"\\\"User ordered a tour\\\"\\\"\\\"\\n pass\\n\\nclass Question(Action):\\n \\\"\\\"\\\"User asked a question about a tour\\\"\\\"\\\"\\n pass\\n \\nclass Click(Action):\\n \\\"\\\"\\\"User clicked on something on the tour profile\\\"\\\"\\\"\\n pass\\n \\nclass MouseMove(Action):\\n \\\"\\\"\\\"User moved the mouse over something on the tour profile\\\"\\\"\\\"\\n pass\\n \\nclass ViewProfile(Action):\\n \\\"\\\"\\\"User viewed the profile of the tour\\\"\\\"\\\"\\n \\n duration = models.PositiveIntegerField(null=True)\\n \\\"\\\"\\\"How long the user has been viewing the profile, \\n In seconds.\\\"\\\"\\\" \\n \\n\\nclass TourOrderEvalPair(BaseEvaluationPair):\\n \\\"\\\"\\\"An abstract class for the artist evaluators\\\"\\\"\\\"\\n\\n subj = models.ForeignKey('travel.User')\\n \\\"\\\"\\\"The subject\\\"\\\"\\\"\\n \\n obj = models.ForeignKey('travel.Tour')\\n \\\"\\\"\\\"The object\\\"\\\"\\\"\\n \\n test_ratio = 0.2\\n \\\"\\\"\\\"The ratio of pairs to select to test pairs\\\"\\\"\\\"\\n\\n class Meta:\\n app_label = 'travel' \\n \\n @classmethod \\n def select(cls, i=0):\\n \\\"\\\"\\\"See the base class for the documentation.\\\"\\\"\\\" \\n \\n all_count = Order.objects.count() \\n test_count = int(cls.test_ratio * all_count ) \\n train_count = all_count - test_count\\n\\n min_stamp_test = Order.objects.order_by('-timestamp')[test_count-1].timestamp\\n\\n test_pairs=Order.objects.filter(timestamp__gte=min_stamp_test)\\n train_pairs = Order.objects.filter(timestamp__lt=min_stamp_test)\\n \\n # remove all other feedback newer than the test timestamp\\n ViewProfile.objects.filter(timestamp__gt=min_stamp_test).delete()\\n MouseMove.objects.filter(timestamp__gt=min_stamp_test).delete()\\n Click.objects.filter(timestamp__gt=min_stamp_test).delete()\\n Question.objects.filter(timestamp__gt=min_stamp_test).delete()\\n \\n # take 1/5 of the orders, remove them and put them to test data \\n # save to test, remove from build\\n for order in test_pairs.iterator():\\n\\n # create the test pair for the order\\n TourOrderEvalPair.objects.create(\\n subj=order.session.user,\\n obj=order.tour,\\n expected_expectancy=EXPECTED_EXPECTANCY_ORDERED,\\n )\\n \\n \\n # delete the order\\n order.delete()\\n \\n print \\\"%d test pairs selected from total %d pairs.\\\" % (test_count, all_count) \\n\\n def get_success(self):\\n return True\\n\"\n },\n {\n \"alpha_fraction\": 0.6163253784179688,\n \"alphanum_fraction\": 0.6197383403778076,\n \"avg_line_length\": 32.46666717529297,\n \"blob_id\": \"5191888627dd960a304bff6a9351cab6d5565e4d\",\n \"content_id\": \"d455b849ec7714785a1e0ffbf66fbc292018548f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3516,\n \"license_type\": \"no_license\",\n \"max_line_length\": 87,\n \"num_lines\": 105,\n \"path\": \"/UnresystCD/code/adapter/unresyst/models/base.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Base classes for the unresyst application.\\\"\\\"\\\"\\n\\nfrom django.db import models\\nfrom django.contrib.contenttypes.models import ContentType\\n\\nfrom symmetric import SymmetricalRelationship\\nfrom unresyst.constants import *\\n\\nclass ContentTypeModel(models.Model):\\n \\\"\\\"\\\"An abstract base class for all models having some subclasses whose\\n instances shouldn't be always returned\\n \\n Be careful when using multiple inheritance: this class shouldbe used\\n as the last and save() shouldn't be overriden in none of the other\\n base classes\\n \\\"\\\"\\\"\\n\\n content_type = models.ForeignKey(ContentType,editable=False, null=True)\\n \\\"\\\"\\\"The actual type of the object.\\\"\\\"\\\"\\n \\n @classmethod\\n def remove_subclass_objects(cls):\\n \\\"\\\"\\\"Get instances only of this class\\\"\\\"\\\"\\n # get the content type\\n cont_type = ContentType.objects.get_for_model(cls)\\n \\n # return only this cont type\\n return cls.objects.filter(content_type=cont_type)\\n \\n def as_leaf_class(self):\\n \\\"\\\"\\\"Get the object as the whole inherited class\\\"\\\"\\\"\\n\\n # get the leaf class\\n content_type = self.content_type\\n Model = content_type.model_class()\\n \\n # get the appropriate object\\n if (Model == self.__class__):\\n return self\\n return Model.objects.get(pk=self.pk) \\n\\n def save(self,*args, **kwargs):\\n \\\"\\\"\\\"Save with the right content type\\\"\\\"\\\"\\n \\n if not self.content_type:\\n self.content_type = ContentType.objects.get_for_model(self.__class__)\\n \\n super(ContentTypeModel, self).save(*args, **kwargs)\\n\\n\\n class Meta:\\n abstract = True \\n app_label = 'unresyst' \\n\\n\\nclass BaseRelationshipInstance(SymmetricalRelationship):\\n \\\"\\\"\\\"An abstract class, the base class of all s-o, s-s, o-o relationships.\\\"\\\"\\\"\\n\\n subject_object1 = models.ForeignKey('unresyst.SubjectObject', \\\\\\n related_name='%(class)s_relationships1')\\n \\\"\\\"\\\"The first subject/object that is in the relationship.\\\"\\\"\\\" \\n \\n subject_object2 = models.ForeignKey('unresyst.SubjectObject', \\\\\\n related_name='%(class)s_relationships2')\\n \\\"\\\"\\\"The second subject/object that is in the relationship\\\"\\\"\\\" \\n \\n description = models.TextField(default='', blank=True)\\n \\\"\\\"\\\"The description of the relationship/rule instance.\\\"\\\"\\\" \\n \\n \\n attr_name1 = 'subject_object1'\\n \\\"\\\"\\\"Overriden attribute name 1\\\"\\\"\\\"\\n \\n attr_name2 = 'subject_object2'\\n \\\"\\\"\\\"Overriden attribute name 2\\\"\\\"\\\"\\n\\n class Meta:\\n abstract = True \\n app_label = 'unresyst' \\n\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return u\\\"(%s, %s)\\\" % (self.subject_object1, self.subject_object2) \\n\\n\\nclass BaseRelationshipDefinition(ContentTypeModel):\\n \\\"\\\"\\\"A definition of the relationship that should be predicted. There's only\\n one for a recommender.\\n \\\"\\\"\\\" \\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"The name of the relationship\\\"\\\"\\\" \\n \\n recommender = models.ForeignKey('unresyst.Recommender')\\n \\\"\\\"\\\"The recommender to which the definition belongs. Each recommender has\\n exactly one predicted relationship.\\n \\\"\\\"\\\" \\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n \\n class Meta:\\n app_label = 'unresyst'\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6709677577018738,\n \"alphanum_fraction\": 0.7419354915618896,\n \"avg_line_length\": 29.799999237060547,\n \"blob_id\": \"1ea1e45a1f2dcdcfa4e4e78454aaa69a127b4577\",\n \"content_id\": \"25141e17388802b89ba36cbcaf6eeb24da4e6e83\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 155,\n \"license_type\": \"no_license\",\n \"max_line_length\": 43,\n \"num_lines\": 5,\n \"path\": \"/UnresystCD/code/dataset_scripts/process_flixster.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"sort -n links.txt -o slinks.txt\\nsort -n ratings.txt -o sratings.txt\\n\\nhead -15000 sratings.txt > mini/ratings.tsv\\nhead -105000 slinks.txt > mini/links.tsv\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6352763175964355,\n \"alphanum_fraction\": 0.6361846923828125,\n \"avg_line_length\": 31.69801902770996,\n \"blob_id\": \"9d281e87a9264e93f7f2b601d8131a225df6259b\",\n \"content_id\": \"cf3bc001473d3162fc7420079e877bb8cd03be11\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 6605,\n \"license_type\": \"no_license\",\n \"max_line_length\": 87,\n \"num_lines\": 202,\n \"path\": \"/UnresystCD/code/adapter/unresyst/recommender/base.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Module containing the definition of BaseRecommender: the base class \\nfor user-defined recommenders.\\n\\\"\\\"\\\"\\n\\nfrom unresyst.constants import *\\n\\nclass BaseRecommender(object):\\n \\\"\\\"\\\"The base class for all recommenders.\\n \\n Defines and the interface of the recommender. All methods and data fields\\n are empty.\\n \\\"\\\"\\\"\\n\\n # Methods:\\n #\\n #\\n\\n # Build phase:\\n # \\n @classmethod\\n def build(cls):\\n \\\"\\\"\\\"Build the recommender. Process all the rules and relationships \\n in order to be able to provide recommendations.\\n \\\"\\\"\\\"\\n pass\\n \\n # Recommend phase:\\n #\\n \\n @classmethod \\n def predict_relationship(cls, subject, object_, save_to_db=False): \\n \\\"\\\"\\\"Get the prediction of the given relationship type between the subject\\n and the object. \\n \\n @type subject: domain specific subject\\n @param subject: the subject\\n \\n @type object_: domain specific object\\n @param object_: the object\\n \\n @type save_to_db: bool\\n @param save_to_db: should the obtained predictions be saved, so that it\\n doesn't have to be computed again when asked for the pair next time\\n \\n @rtype: RelationshipPrediction\\n @return: An instance representing the prediction of the relationship \\n between the subject and the object, containing the estimated \\n probability and explanation. If the prediction isn't known, returns\\n an prediction with expectancy 0.5 and empty explanation\\n \\n @raise InvalidParameterError: if the given subject or object doesn't \\n have a domain neutral representation in the unresyst database. \\n Either the recommender hasn't been built or the subject/object was\\n added later without updating the recommender.\\n \\\"\\\"\\\"\\n pass\\n \\n @classmethod\\n def get_recommendations(cls, subject, count=None):\\n \\\"\\\"\\\"Get recommendations for the given subject.\\n\\n @type subject: domain specific subject\\n @param subject: the subject\\n \\n @type count: int\\n @param count: a maximum number of objects to be recommended, if not given,\\n the default_recommendation_count from the recommender class is used.\\n \\n @rtype: list of RelationshipPrediction\\n @return: An instance representing the prediction of the relationship \\n between the subject and the object, containing the estimated \\n probability and explanation. If the prediction isn't known, returns\\n an prediction with expectancy 0.5 and empty explanation\\n\\n @raise InvalidParameterError: if the given subject doesn't \\n have a domain neutral representation in the unresyst database. \\n Either the recommender hasn't been built or the subject was\\n added later without updating the recommender. \\n \\\"\\\"\\\"\\n pass\\n \\n \\n # Update phase:\\n # \\n @classmethod \\n def add_subject(cls, subject):\\n \\\"\\\"\\\"Add subject to the recommender.\\\"\\\"\\\"\\n pass\\n\\n @classmethod\\n def add_object(cls, object_):\\n \\\"\\\"\\\"Add object to the recommender.\\\"\\\"\\\"\\n pass\\n\\n @classmethod\\n def update_subject(cls, subject):\\n \\\"\\\"\\\"Update the subject in the recommender, including its relationships\\n and applied rules\\\"\\\"\\\"\\n pass\\n\\n\\n @classmethod \\n def update_object(cls, object_):\\n \\\"\\\"\\\"Update the object in the recommender, including its relationships\\n and applied rules\\\"\\\"\\\"\\n pass\\n\\n\\n @classmethod\\n def remove_subject(cls, subject):\\n \\\"\\\"\\\"Remove the subject from the recommender, including its relationships\\n and applied rules\\\"\\\"\\\"\\n pass\\n\\n\\n @classmethod \\n def remove_object(cls, object_):\\n \\\"\\\"\\\"Remove the object from the recommender, including its relationships\\n and applied rules\\\"\\\"\\\"\\n pass \\n \\n \\n \\n # Domain specific data. empty, will be overriden in the domain specific recommender\\n #\\n\\n name = \\\"\\\"\\n \\\"\\\"\\\"The name of the recommender\\\"\\\"\\\"\\n \\n subjects = None\\n \\\"\\\"\\\"The objects to who the recommender will recommend.\\n Requires the following interface:\\n on subjects manager:\\n - iterator(): get an iterator on the collection\\n - all(): get all sujbects\\n - get(id=..) get the object with the given id\\n queryset:\\n - exists(): is there something in the queryset?\\n on each subject instance:\\n - id: an integer id of the subject \\n - __unicode__(): printable string\\n \\\"\\\"\\\"\\n \\n objects = None\\n \\\"\\\"\\\"The objects that will be recommended.\\\"\\\"\\\" \\n\\n predicted_relationship = None\\n \\\"\\\"\\\"The relationship that will be predicted\\\"\\\"\\\"\\n \\n relationships = ()\\n \\\"\\\"\\\"Relationships among the subjects and objects in the domain\\\"\\\"\\\"\\n \\n rules = ()\\n \\\"\\\"\\\"Rules that can be applied to the domain\\\"\\\"\\\"\\n \\n cluster_sets = ()\\n \\\"\\\"\\\"Clusters to which subjects and objects can be divided\\\"\\\"\\\"\\n \\n biases = ()\\n \\\"\\\"\\\"Bias - predisposition of subjects/objects to be in the predicted_relationship\\\"\\\"\\\"\\n\\n random_recommendation_description = None\\n \\\"\\\"\\\"The description that will be used as a description \\n to random recommendations.\\\"\\\"\\\"\\n\\n # Auxiliary class attributes\\n _recommender_model = None\\n \\\"\\\"\\\"The database model instance belonging to the class\\\"\\\"\\\"\\n \\n # Class configuration - the behaviour of the layers below the recommender\\n # Empty, will be overriden by the Recommender class\\n #\\n \\n abstractor = None\\n \\\"\\\"\\\"The class that will be used for the abstractor level. Can be \\n overriden in subclasses\\\"\\\"\\\" \\n \\n algorithm = None\\n \\\"\\\"\\\"The class that will be used for the algorithm level. Can be \\n overriden in subclasses\\\"\\\"\\\"\\n \\n remove_predicted_from_recommendations = None\\n \\\"\\\"\\\"Should the instances of the predicted_relationship be removed from\\n recommendation list?\\n \\\"\\\"\\\"\\n \\n save_all_to_predictions = None\\n \\\"\\\"\\\"Should also instances of predicted_relationship be saved to predictions?\\n \\\"\\\"\\\"\\n \\n recommendation_expectancy_limit = None\\n \\\"\\\"\\\"The limit for expectancy above which the objects can be recommended.\\n If not none, only objects with expectancy above the limit are recommended.\\n A reasonable limit is 0.5 - the recommendations then don't include random \\n objects.\\n \\\"\\\"\\\"\\n \\n verbose_build = None\\n \\\"\\\"\\\"Should messages be printed during the build?\\\"\\\"\\\"\\n \\n explicit_rating_rule = None\\n \\\"\\\"\\\"If given, this rule is exported, not the predicted_relationship\\\"\\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.8416422009468079,\n \"alphanum_fraction\": 0.8416422009468079,\n \"avg_line_length\": 36.88888931274414,\n \"blob_id\": \"5f3d97050e753b19fe0d90682b372202a501009f\",\n \"content_id\": \"d6ffb603555e521ec623ae4497ae716e0fe36042\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 341,\n \"license_type\": \"no_license\",\n \"max_line_length\": 69,\n \"num_lines\": 9,\n \"path\": \"/UnresystCD/code/adapter/unresyst/algorithm/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The package contains all classes for the algorithm layer.\\n\\nThe algorithm layer can be replaced by a common algorithm exploiting \\nthe known preferences, like collaborative filtering.\\n\\\"\\\"\\\"\\n\\nfrom simple_algorithm import SimpleAlgorithm\\nfrom aggregating_algorithm import AggregatingAlgorithm\\nfrom compiling_algorithm import CompilingAlgorithm\\n\"\n },\n {\n \"alpha_fraction\": 0.6346863508224487,\n \"alphanum_fraction\": 0.6346863508224487,\n \"avg_line_length\": 19.820512771606445,\n \"blob_id\": \"ff32dd7225c9e17f47ae2c0d66a81b75a053a4cd\",\n \"content_id\": \"ae5df24917045a80cc0081f9e32c1d5fa3829312\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 813,\n \"license_type\": \"no_license\",\n \"max_line_length\": 132,\n \"num_lines\": 39,\n \"path\": \"/UnresystCD/code/adapter/lastfmeval.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/bin/sh\\n\\n# run the evaluation- recommendations. parameters \\n# novel -> NovelArtistRecommender, \\n# nenovel -> ArtistRecommender\\n# dontbuild-> don't build\\n\\nBUILD=true\\nEVAL=''\\nREC=''\\n\\nfor param in $*;\\ndo \\n case $param in\\n 'novel')\\n EVAL='NovelArtistRankEvaluator'\\n REC='NovelArtistRecommender'\\n ;;\\n 'nenovel')\\n EVAL='ArtistRankEvaluator'\\n REC='ArtistRecommender'\\n ;;\\n 'dontbuild')\\n BUILD=false\\n ;;\\n esac\\ndone\\n\\nif [ $BUILD = true ]\\nthen\\n # build\\n echo \\\"Building...\\\"\\n echo \\\"from lastfm.recommender import *; $REC.build()\\\" | python ./manage.py shell\\nfi\\n\\necho \\\"Evaluating recommendations...\\\"\\necho \\\"from lastfm.evaluation import *; from lastfm.recommender import *; $EVAL.evaluate_predictions($REC)\\\"| python ./manage.py shell\\n\\necho \\\"\\\"\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6459985375404358,\n \"alphanum_fraction\": 0.6459985375404358,\n \"avg_line_length\": 20.671875,\n \"blob_id\": \"240d9c3e17fedad9e7ad2dc48ce541cc1373512c\",\n \"content_id\": \"acecebc085fa5fe72c71f0bf3e43802cb269eee3\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 1387,\n \"license_type\": \"no_license\",\n \"max_line_length\": 132,\n \"num_lines\": 64,\n \"path\": \"/UnresystCD/code/adapter/build.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# build the shoe recommender\\n\\nLASTFM=false\\nNLASTFM=false\\nDEMO=false\\nADEMO=false\\nFLIXSTER=false\\nTRAVEL=false\\n\\nfor param in $*;\\ndo \\n case $param in\\n 'lastfm')\\n LASTFM=true\\n ;;\\n 'nlastfm')\\n NLASTFM=true\\n ;;\\n 'demo')\\n DEMO=true\\n ;;\\n 'flixster')\\n FLIXSTER=true\\n ;;\\n 'travel')\\n TRAVEL=true\\n ;;\\n 'ademo')\\n ADEMO=true\\n ;; \\n esac\\ndone\\n\\nif [ $DEMO = true ]\\nthen\\n echo \\\"from demo.recommender import ShoeRecommender; ShoeRecommender.build(); quit();\\\" | python ./manage.py shell\\nfi\\n\\nif [ $ADEMO = true ]\\nthen\\n echo \\\"from demo.recommender import AverageRecommender; AverageRecommender.build(); quit();\\\" | python ./manage.py shell\\nfi\\n\\nif [ $LASTFM = true ]\\nthen\\n echo \\\"from lastfm.recommender import ArtistRecommender; ArtistRecommender.build(); quit();\\\" | python ./manage.py shell\\nfi\\n\\nif [ $NLASTFM = true ]\\nthen\\n echo \\\"from lastfm.recommender import NovelArtistRecommender; NovelArtistRecommender.build(); quit();\\\" | python ./manage.py shell\\nfi\\n\\nif [ $FLIXSTER = true ]\\nthen\\n echo \\\"from flixster.recommender import MovieRecommender; MovieRecommender.build(); quit();\\\" | python ./manage.py shell\\nfi\\n\\nif [ $TRAVEL = true ]\\nthen\\n echo \\\"from travel.recommender import OrderTourRecommender; OrderTourRecommender.build(); quit();\\\" | python ./manage.py shell\\nfi\\n\\necho \\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.8062015771865845,\n \"alphanum_fraction\": 0.8062015771865845,\n \"avg_line_length\": 31.25,\n \"blob_id\": \"03d93fb7e5af5f9cfcc13af9d980d345de9381c7\",\n \"content_id\": \"728db3828f4936747110a00115337c277963c69e\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 129,\n \"license_type\": \"no_license\",\n \"max_line_length\": 81,\n \"num_lines\": 4,\n \"path\": \"/UnresystCD/code/adapter/unresyst/recommender/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Package holding the BaseRecommender class, holding the interface for the whole\\nunresyst.\\\"\\\"\\\"\\n\\nfrom base import BaseRecommender\\n\"\n },\n {\n \"alpha_fraction\": 0.6535211205482483,\n \"alphanum_fraction\": 0.6788732409477234,\n \"avg_line_length\": 21.125,\n \"blob_id\": \"51a21eb6ed1100fac55f6654d42dd24e300e396c\",\n \"content_id\": \"46b996f5dd70d8892ad94fe110f7a4e6d7be3cfd\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 355,\n \"license_type\": \"no_license\",\n \"max_line_length\": 61,\n \"num_lines\": 16,\n \"path\": \"/code/adapter/flixster/constants.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Constants for flixster\\\"\\\"\\\"\\n\\nMIN_POSITIVE_RATING = 3.0\\n\\\"\\\"\\\"Minimal rating taken as positive\\\"\\\"\\\"\\n\\nMIN_HIGH_RATING = 4.0\\n\\\"\\\"\\\"Minimal rating taken as \\\"high\\\".\\\"\\\"\\\"\\n\\nMAX_LOW_RATING = 1.0\\n\\\"\\\"\\\"Maximal rating taken as \\\"low\\\".\\\"\\\"\\\"\\n\\nMAX_STARS = 5\\n\\\"\\\"\\\"Maximal number of starts in rating\\\"\\\"\\\"\\n\\nMAX_TOLERANCE = 1.0\\n\\\"\\\"\\\"Size of the interval to take a prediction as successful\\\"\\\"\\\"\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5889149904251099,\n \"alphanum_fraction\": 0.598078191280365,\n \"avg_line_length\": 36.93692398071289,\n \"blob_id\": \"dc6153f14a58a55b7cb4816548e8b114e8d4c65d\",\n \"content_id\": \"5ad83dc9f7b4f6ed08692ad2a5336acebffd92d5\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 24664,\n \"license_type\": \"no_license\",\n \"max_line_length\": 112,\n \"num_lines\": 650,\n \"path\": \"/UnresystCD/code/adapter/unresyst/recommender/rules.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The classes for representing business rules and relationships\\\"\\\"\\\"\\n\\nfrom unresyst.constants import *\\n\\nfrom unresyst.models.abstractor import *\\nfrom unresyst.models.common import SubjectObject\\nfrom unresyst.exceptions import DescriptionKeyError, ConfigurationError\\n\\nclass BaseRelationship(object):\\n \\\"\\\"\\\"A base class for representing all relationships and rules.\\n \\n A subclass for all classes representing a relationship between entities (not necessarily \\n of the same type). Contains the condition that is true between and only \\n between the entities that are in the given relationship.\\n \\\"\\\"\\\"\\n \\n def __init__(self, name, condition=None, description=None, generator=None):\\n \\\"\\\"\\\"The constructor.\\\"\\\"\\\"\\n \\n self.name = name\\n \\\"\\\"\\\"The name of the rule/relationship.\\\"\\\"\\\"\\n \\n self.condition = condition\\n \\\"\\\"\\\"A boolean function that represents the condition. If True \\n for the given pair of entities, there's the Relationship between the \\n entities. \\n Should be simple.\\n \\\"\\\"\\\"\\n \\n self.description = description\\n \\\"\\\"\\\"A string describing the rule. It can contain placeholders for entities: \\n \\n - %(subject)s, %(object)s for subject-object relationships and rules\\n - %(subject1)s, %(subject2)s for subject-subject relationships \\n and rules\\n - %(object1)s, %(object2)s for object-object relationships and rules\\n - %(subjectobject1)s, %(subjectobject2)s for recommenders where \\n subject domain is the same as object domain \\n \\\"\\\"\\\"\\n \\n self.generator = generator\\n \\\"\\\"\\\"A generator returning pairs of objects that are in the relationship/\\n the rule applies to them.\\n For performance reasons - if given, the pairs will be taken from it \\n without the need for evaluating the condition for each possible pair.\\n \\\"\\\"\\\"\\n \\n DESCRIPTION_FORMAT_DICT = {\\n RELATIONSHIP_TYPE_SUBJECT_OBJECT: \\n (FORMAT_STR_SUBJECT, FORMAT_STR_OBJECT),\\n RELATIONSHIP_TYPE_SUBJECT_SUBJECT: \\n (FORMAT_STR_SUBJECT1, FORMAT_STR_SUBJECT2),\\n RELATIONSHIP_TYPE_OBJECT_OBJECT: \\n (FORMAT_STR_OBJECT1, FORMAT_STR_OBJECT2),\\n RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT: \\n (FORMAT_STR_SUBJECTOBJECT1, FORMAT_STR_SUBJECTOBJECT2)\\n }\\n \\\"\\\"\\\"A dictionary relationship type (e.g. 'S-O') a pair of formatting strings \\n for description, e.g. ('subject', 'object').\\n \\\"\\\"\\\"\\n\\n relationship_type = RELATIONSHIP_TYPE_SUBJECT_OBJECT\\n \\\"\\\"\\\"The type of the relationship S-O\\\"\\\"\\\"\\n \\n InstanceClass = RelationshipInstance\\n \\\"\\\"\\\"The model class used for representing instances of the rule/relationship\\\"\\\"\\\"\\n \\n DefinitionClass = RuleRelationshipDefinition\\n \\\"\\\"\\\"The model class used for representing the definition of the \\n rule/relationship\\n \\\"\\\"\\\"\\n \\n is_symmetric = False\\n \\\"\\\"\\\"Is the rule/relationship between the entities of the same domain?\\n True for S-S, O-O, SO-SO\\n \\\"\\\"\\\"\\n \\n def get_filled_description(self, arg1, arg2):\\n \\\"\\\"\\\"Get description for a rule/relationship instance, between \\n arg1 and arg2. \\n \\n @type arg1: models.SubjectObject\\n @param arg1: the first subjectobject in the relationship/rule\\n\\n @type arg2: models.SubjectObject\\n @param arg2: the second subjectobject in the relationship/rule\\n \\n @rtype: str\\n @return: a string with filled gaps for entities.\\n \\\"\\\"\\\"\\n # get the format strings, e.g. ('subject', 'object')\\n format_strings = self.DESCRIPTION_FORMAT_DICT[self.relationship_type]\\n \\n # create the formating strings to be passed to description\\n format_dict = {\\n format_strings[0]: arg1,\\n format_strings[1]: arg2\\n }\\n if self.description is None:\\n desc = ''\\n else:\\n try: \\n desc = self.description % format_dict\\n except KeyError, e:\\n raise DescriptionKeyError(\\n message=\\\"There's an invalid key in description\\\",\\n recommender=self.recommender,\\n name=self.name, \\n key=e.__str__(), \\n permitted_keys=format_dict.keys()\\n )\\n \\n return desc\\n \\n def get_create_definition_kwargs(self):\\n \\\"\\\"\\\"Get dictionary of parameters for the definition model constructor. \\n \\n @rtype: dictionary string: object\\n @return: the kwargs of the definition model constructor \\n \\\"\\\"\\\"\\n return {\\n \\\"name\\\": self.name,\\n \\\"recommender\\\": self.recommender._get_recommender_model(),\\n }\\n \\n def get_additional_instance_kwargs(self, ds_arg1, ds_arg2):\\n \\\"\\\"\\\"Get dictionary of additional kwargs for creating the rule/relationship\\n instance. \\n \\n @type ds_arg1: django.db.models.manager.Manager\\n @param ds_arg1: the first argument of the rule instance - domain specific\\n \\n @type ds_arg2: django.db.models.manager.Manager\\n @param ds_arg2: the second argument of the rule instance - domain specific\\n \\n @rtype: dictionary string: object\\n @return: additional keyword args for creating rule/relationship instance \\n \\\"\\\"\\\"\\n return {}\\n \\n \\n def evaluate_on_dn_args(self, dn_arg1, dn_arg2, definition): \\n \\\"\\\"\\\"Evaluates the rule on the given arguments. If evaluated positively,\\n a new rule/relationship instance is saved.\\n \\n @type dn_arg1: models.SubjectObject\\n @param dn_arg1: a domain neutral representation of the first entity.\\n\\n @type dn_arg2: models.SubjectObject\\n @param dn_arg2: a domain neutral representation of the second entity.\\n \\n @type definition: models.abstractor.RuleRelationshipDefinition\\n @param definition: the model representing the rule/relationship \\n definition\\n \\n @rtype: int\\n @return: 1 if something has benn created, 0 if not\\n \\\"\\\"\\\"\\n \\n # get the domain specific objects for our universal representations\\n #\\n arg1_manager = self.recommender._get_entity_manager(dn_arg1.entity_type)\\n ds_arg1 = dn_arg1.get_domain_specific_entity(entity_manager=arg1_manager)\\n \\n arg2_manager = self.recommender._get_entity_manager(dn_arg2.entity_type)\\n ds_arg2 = dn_arg2.get_domain_specific_entity(entity_manager=arg2_manager)\\n\\n # if the condition is satisfied\\n if self.condition(ds_arg1, ds_arg2):\\n \\n self._perform_save_instance(definition, ds_arg1, ds_arg2, dn_arg1, dn_arg2)\\n return 1\\n \\n return 0\\n\\n @classmethod\\n def order_arguments(cls, dn_arg1, dn_arg2):\\n \\\"\\\"\\\"Order the arguments as they appear in the relationship.\\\"\\\"\\\"\\n a1, a2, x, x = cls._order_in_pair(dn_arg1, dn_arg2, None, None)\\n \\n return (a1, a2)\\n\\n @classmethod\\n def _order_in_pair(cls, dn_arg1, dn_arg2, ds_arg1, ds_arg2):\\n \\\"\\\"\\\"Swap the arguments in the rule/relationships so that the first\\n has a lower id than the second\\n \\\"\\\"\\\"\\n\\n # if the second argument has lower id than the first, swap them\\n if dn_arg2.pk < dn_arg1.pk: \\n return (dn_arg2, dn_arg1, ds_arg2, ds_arg1)\\n\\n # otherwise not\\n return (dn_arg1, dn_arg2, ds_arg1, ds_arg2)\\n \\n\\n def _perform_save_instance(self, definition, ds_arg1, ds_arg2, dn_arg1, dn_arg2):\\n \\\"\\\"\\\"Perform the action of creating and saving the instance\\\"\\\"\\\"\\n \\n # order the instances in pairs as the class requires\\n dn_arg1, dn_arg2, ds_arg1, ds_arg2 = self._order_in_pair(dn_arg1, dn_arg2, ds_arg1, ds_arg2)\\n \\n add_kwargs = self.get_additional_instance_kwargs(ds_arg1, ds_arg2)\\n \\n # create a rule/relationship instance\\n instance = self.InstanceClass(\\n definition=definition,\\n subject_object1=dn_arg1,\\n subject_object2=dn_arg2,\\n description=self.get_filled_description(dn_arg1, dn_arg2),\\n **add_kwargs)\\n \\n instance.save()\\n\\n\\n def save_instance(self, ds_arg1, ds_arg2, definition): \\n \\\"\\\"\\\"Save an instance of the rule/relationship for the given args.\\n \\n @type ds_arg1: domain specific subject/object\\n @param ds_arg1: a domain specific representation of the first entity.\\n\\n @type ds_arg2: domain specific subject/object\\n @param ds_arg2: a domain specific representation of the second entity.\\n \\n @type definition: models.abstractor.RuleRelationshipDefinition\\n @param definition: the model representing the rule/relationship \\n definition\\n \\n @raise ConfigurationError: thrown if the condition doesn't evaluate \\n to true on the given pair\\n \\\"\\\"\\\"\\n if not (self.condition is None) and not self.condition(ds_arg1, ds_arg2):\\n raise ConfigurationError(\\n message=(\\\"The condition wasn't evaluated as true for the pair \\\" + \\n \\\"%s, %s, even though it was returned by the generator.\\\") % (ds_arg1, ds_arg2), \\n recommender=self.recommender, \\n parameter_name='rules/relationships', \\n parameter_value=self.name)\\n \\n # convert the domain specific to domain neutral \\n #\\n arg1_ent_type, arg2_ent_type = self.relationship_type.split(RELATIONSHIP_TYPE_SEPARATOR) \\n\\n dn_arg1 = SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=ds_arg1, \\n entity_type=arg1_ent_type, \\n recommender=definition.recommender)\\n\\n dn_arg2 = SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=ds_arg2, \\n entity_type=arg2_ent_type, \\n recommender=definition.recommender)\\n \\n # create and save the instance\\n self._perform_save_instance(definition, ds_arg1, ds_arg2, dn_arg1, dn_arg2) \\n \\n \\n \\n def evaluate(self):\\n \\\"\\\"\\\"Evaluate the rule on all subjects/objects - pairs.\\n \\n Creates and saves the rule/relationship definition, creates and saves\\n rule instances. \\n \\\"\\\"\\\"\\n \\n # obtain the kwargs for creating the definition\\n def_kwargs = self.get_create_definition_kwargs()\\n\\n # create and save the definition\\n definition = self.DefinitionClass(**def_kwargs)\\n definition.save()\\n\\n i = 0 \\n # if we have a generator, use it for looping through pairs\\n if not (self.generator is None):\\n \\n\\n # loop through pairs, save the rule/relationship instances\\n for ds_arg1, ds_arg2 in self.generator():\\n self.save_instance(ds_arg1, ds_arg2, definition)\\n i += 1\\n \\n print \\\" %d instances of rule/rel %s created\\\" % (i, self.name)\\n \\n # that's it\\n return\\n \\n \\n # otherwise take the entities one by one\\n \\n # parse what should be used as condition args\\n arg1_s, arg2_s = self.relationship_type.split(RELATIONSHIP_TYPE_SEPARATOR) \\n \\n if arg1_s == arg2_s:\\n \\n # loop only through the matrix members below the diagonal \\n # \\n i = 0\\n for arg1, arg2 in SubjectObject.unique_pairs(\\n recommender=self.recommender._get_recommender_model(),\\n entity_type=arg1_s): \\n # evaluate it\\n i += self.evaluate_on_dn_args(arg1, arg2, definition)\\n \\n\\n \\n else:\\n # filter subjectobjects for my recommender\\n qs_recommender = SubjectObject.objects.filter(\\n recommender=self.recommender._get_recommender_model())\\n \\n # go through all things that have to be as first and as second param\\n for arg1 in qs_recommender.filter(entity_type=arg1_s).iterator():\\n for arg2 in qs_recommender.filter(entity_type=arg2_s).iterator():\\n \\n # evaluate the rule/relationship on the given args\\n i += self.evaluate_on_dn_args(arg1, arg2, definition)\\n\\n print \\\" %d instances of rule/rel %s created\\\" % (i, self.name)\\n\\n \\n def export(self, f):\\n \\\"\\\"\\\"Export the relationship as lines to the given file object.\\n \\n Exports the relationship in form:\\n subject_id, object_id[,confidence]\\\\n\\n \\n confidence is exported only if it's relevant,\\n no matter if it's positive or not\\n \\n @type f: file\\n @param f: the file to write to\\n \\n @raise: ConfigurationError: if the rule/relationship doesn't have\\n a generator.\\n \\\"\\\"\\\"\\n \\n # if we don't have a generator raise an error\\n if self.generator is None:\\n raise ConfigurationError(\\n message=\\\"The relationship to export is missing a generator.\\\", \\n recommender=cls, \\n parameter_name='?', \\n parameter_value=cls.name)\\n \\n i = 0\\n \\n # loop through pairs, export the rule/relationship instances\\n for ds_arg1, ds_arg2 in self.generator():\\n \\n # create the common part\\n linestr = \\\"%s,%s\\\" % (ds_arg1.pk, ds_arg2.pk)\\n \\n # get the confidence if provided\\n adkwargs = self.get_additional_instance_kwargs(ds_arg1, ds_arg2)\\n \\n # if confidence provided, append it\\n if adkwargs.has_key(EXPECTANCY_KWARG_NAME):\\n linestr += \\\",%f\\\" % adkwargs[EXPECTANCY_KWARG_NAME]\\n \\n linestr += '\\\\n'\\n \\n # write it to the file\\n f.write(linestr)\\n \\n i += 1\\n \\n print \\\" %d instances of rule/rel %s exported\\\" % (i, self.name)\\n\\n \\nclass PredictedRelationship(BaseRelationship):\\n \\\"\\\"\\\"A class for representing the predicted relationship.\\\"\\\"\\\"\\n\\n DefinitionClass = PredictedRelationshipDefinition\\n \\\"\\\"\\\"The model class used for representing the definition of the \\n rule/relationship\\n \\\"\\\"\\\" \\n \\n @classmethod\\n def _order_in_pair(cls, dn_arg1, dn_arg2, ds_arg1, ds_arg2):\\n \\\"\\\"\\\"Swap the arguments in the relationship so that the first\\n is always a subject and second the object.\\n \\\"\\\"\\\" \\n \\n # for the SO entities we apply the normal policy\\n if dn_arg1.entity_type == ENTITY_TYPE_SUBJECTOBJECT:\\n return super(PredictedRelationship, cls)._order_in_pair(dn_arg1, dn_arg2, ds_arg1, ds_arg2)\\n\\n # otherwise we put subjects as first\\n \\n # if the first is object, swap\\n if dn_arg1.entity_type == ENTITY_TYPE_OBJECT:\\n return (dn_arg2, dn_arg1, ds_arg2, ds_arg1)\\n \\n # if not, keep\\n return (dn_arg1, dn_arg2, ds_arg1, ds_arg2)\\n\\nclass ExplicitSubjectObjectRule(BaseRelationship):\\n \\\"\\\"\\\"A class for representing explicit preference like rating, for attributes where\\n the user can express both positive and negative preference.\\\"\\\"\\\"\\n\\n relationship_type = RELATIONSHIP_TYPE_SUBJECT_OBJECT\\n \\\"\\\"\\\"The type of the relationship S-O\\\"\\\"\\\" \\n\\n DefinitionClass = ExplicitRuleDefinition\\n \\\"\\\"\\\"The model class used for representing the definition of the \\n rule/relationship\\n \\\"\\\"\\\" \\n \\n InstanceClass = ExplicitRuleInstance\\n \\\"\\\"\\\"The model class used for representing instances of \\n the rule/relationship\\\"\\\"\\\"\\n \\n \\n def __init__(self, name, expectancy, condition=None, description=None, generator=None):\\n \\\"\\\"\\\"The constructor.\\\"\\\"\\\"\\n \\n super(ExplicitSubjectObjectRule, self).__init__(\\n name=name, \\n condition=condition, \\n description=description, \\n generator=generator)\\n \\n self.expectancy = expectancy\\n \\\"\\\"\\\"A function taking a subject and an object, giving the explicit preference\\n normalized to [0, 1].\\n \\\"\\\"\\\"\\n \\n def get_additional_instance_kwargs(self, ds_arg1, ds_arg2):\\n \\\"\\\"\\\"See the base class for documentation\\n \\n @raise ConfigurationError: if the confidence function returns a value\\n outside [0, 1]\\n \\\"\\\"\\\"\\n ret_dict = super(ExplicitSubjectObjectRule, self)\\\\\\n .get_additional_instance_kwargs(ds_arg1, ds_arg2)\\n \\n # call the user-defined confidence method\\n expectancy = self.expectancy(ds_arg1, ds_arg2)\\n \\n if not (MIN_EXPECTANCY <= expectancy <= MAX_EXPECTANCY):\\n raise ConfigurationError(\\n message=(\\\"The rule '%s' has expectancy %f, for the\\\" + \\\\\\n \\\" pair (%s, %s). Should be between 0 and 1.\\\") % \\\\\\n (self.name, expectancy, ds_arg1, ds_arg2),\\n recommender=self.recommender,\\n parameter_name=\\\"Recommender.rules\\\",\\n parameter_value=self.recommender.rules\\n )\\n \\n ret_dict[EXPECTANCY_KWARG_NAME] = expectancy\\n return ret_dict \\n\\n @classmethod\\n def _order_in_pair(cls, dn_arg1, dn_arg2, ds_arg1, ds_arg2):\\n return PredictedRelationship._order_in_pair(dn_arg1, dn_arg2, ds_arg1, ds_arg2)\\n \\nclass _WeightedRelationship(BaseRelationship):\\n \\\"\\\"\\\"A class representing a relationship with a weight.\\\"\\\"\\\"\\n\\n DefinitionClass = RuleRelationshipDefinition\\n \\\"\\\"\\\"The model class used for representing the definition of the \\n rule/relationship\\n \\\"\\\"\\\" \\n \\n def __init__(self, name, is_positive, weight, condition=None, description=None, generator=None):\\n \\\"\\\"\\\"The constructor.\\\"\\\"\\\"\\n \\n super(_WeightedRelationship, self).__init__(\\n name=name, \\n condition=condition, \\n description=description, \\n generator=generator)\\n \\n self.is_positive = is_positive\\n \\\"\\\"\\\"Is the relationship positive to the predicted relationship?\\\"\\\"\\\"\\n \\n self.weight = weight\\n \\\"\\\"\\\"A float number from [0, 1] representing the *static* weight of the rule. \\n It doesn't depend on the entity pair.\\n \\\"\\\"\\\"\\n\\n\\n def get_create_definition_kwargs(self):\\n \\\"\\\"\\\"Get dictionary of parameters for the definition model constructor. \\n \\n Add the weight to the parameters. \\n \\n @rtype: dictionary string: object\\n @return: the kwargs of the definition model constructor \\n \\n @raise ConfigurationError: if the weight isn't from [0, 1]\\n \\\"\\\"\\\"\\n ret_dict = super(_WeightedRelationship, self).get_create_definition_kwargs()\\n \\n ret_dict['is_positive'] = self.is_positive\\n\\n if not (MIN_WEIGHT <= self.weight <= MAX_WEIGHT):\\n raise ConfigurationError(\\n message=(\\\"The rule/relationship '%s' has weight %f,\\\" + \\\\\\n \\\" should be between 0 and 1.\\\") % (self.name, self.weight),\\n recommender=self.recommender,\\n parameter_name=\\\"Recommender.rules or Recommender.relationships\\\",\\n parameter_value=(self.recommender.rules, self.recommender.relationships)\\n )\\n \\n ret_dict['weight'] = self.weight\\n ret_dict[\\\"relationship_type\\\"] = self.relationship_type \\n \\n return ret_dict\\n \\n \\n \\n\\nclass SubjectObjectRelationship(_WeightedRelationship):\\n \\\"\\\"\\\"A class for representing subject-object preference for recommendation\\\"\\\"\\\"\\n \\n relationship_type = RELATIONSHIP_TYPE_SUBJECT_OBJECT\\n \\\"\\\"\\\"The type of the relationship S-O\\\"\\\"\\\" \\n\\n @classmethod\\n def _order_in_pair(cls, dn_arg1, dn_arg2, ds_arg1, ds_arg2):\\n return PredictedRelationship._order_in_pair(dn_arg1, dn_arg2, ds_arg1, ds_arg2)\\n\\nclass _SimilarityRelationship(_WeightedRelationship):\\n \\\"\\\"\\\"A base class (abstract) for all relationships operating between the same type \\n and meaning similarity.\\n \\\"\\\"\\\"\\n \\n is_symmetric = True\\n \\\"\\\"\\\"For documentation see the base class\\\"\\\"\\\"\\n \\n\\nclass ObjectSimilarityRelationship(_SimilarityRelationship):\\n \\\"\\\"\\\"A class for representing inter-object similarity.\\\"\\\"\\\" \\n\\n relationship_type = RELATIONSHIP_TYPE_OBJECT_OBJECT\\n \\\"\\\"\\\"The type of the relationship O-O\\\"\\\"\\\" \\n\\n\\nclass SubjectSimilarityRelationship(_SimilarityRelationship):\\n \\\"\\\"\\\"A class for representing inter-subject similarity.\\\"\\\"\\\"\\n\\n relationship_type = RELATIONSHIP_TYPE_SUBJECT_SUBJECT\\n \\\"\\\"\\\"The type of the relationship S-S\\\"\\\"\\\"\\n\\n\\nclass SubjectObjectSimilarityRelationship(_SimilarityRelationship):\\n \\\"\\\"\\\"A class used only when subject domain equals object domain. \\n For representing inter-entity similarity.\\n \\\"\\\"\\\"\\n \\n relationship_type = RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT\\n \\\"\\\"\\\"The type of the relationship SO-SO\\\"\\\"\\\"\\n\\n\\n# rules:\\n# \\n\\nclass _BaseRule(_WeightedRelationship):\\n \\\"\\\"\\\"A base class for all rules (abstract).\\\"\\\"\\\"\\n \\n InstanceClass = RuleInstance\\n \\\"\\\"\\\"The model class used for representing instances of \\n the rule/relationship\\\"\\\"\\\"\\n \\n def __init__(self, name, is_positive, weight, confidence, condition=None, description=None, generator=None):\\n \\\"\\\"\\\"The constructor.\\\"\\\"\\\" \\n\\n super(_BaseRule, self).__init__(\\n name=name, \\n condition=condition,\\n is_positive=is_positive, \\n weight=weight, \\n description=description, \\n generator=generator)\\n \\n self.confidence = confidence\\n \\\"\\\"\\\"A float function giving values from [0, 1] representing the \\n the confidence of the rule on the given pair. \\n It's dynamic, depends on the entity pair.\\n \\\"\\\"\\\" \\n \\n def get_additional_instance_kwargs(self, ds_arg1, ds_arg2):\\n \\\"\\\"\\\"See the base class for documentation\\n \\n @raise ConfigurationError: if the confidence function returns a value\\n outside [0, 1]\\n \\\"\\\"\\\"\\n ret_dict = super(_BaseRule, self).get_additional_instance_kwargs(\\n ds_arg1, ds_arg2)\\n \\n # call the user-defined confidence method\\n confidence = self.confidence(ds_arg1, ds_arg2)\\n \\n if not (MIN_CONFIDENCE <= confidence <= MAX_CONFIDENCE):\\n raise ConfigurationError(\\n message=(\\\"The rule '%s' has a confidence %f, for the\\\" + \\\\\\n \\\" pair (%s, %s). Should be between 0 and 1.\\\") % \\\\\\n (self.name, confidence, ds_arg1, ds_arg2),\\n recommender=self.recommender,\\n parameter_name=\\\"Recommender.rules\\\",\\n parameter_value=self.recommender.rules\\n )\\n \\n ret_dict[CONFIDENCE_KWARG_NAME] = confidence\\n return ret_dict\\n \\n \\n# confidence by taky mohla vracet string s doplnujicim vysvetlenim, \\n\\nclass _SimilarityRule(_BaseRule):\\n \\\"\\\"\\\"A base class (abstract) for all rules operating between the same type \\n and meaning similarity.\\\"\\\"\\\"\\n\\n is_symmetric = True\\n \\\"\\\"\\\"For documentation see the base class\\\"\\\"\\\" \\n \\n\\nclass ObjectSimilarityRule(_SimilarityRule):\\n \\\"\\\"\\\"A class for representing inter-object similarity.\\\"\\\"\\\" \\n\\n relationship_type = RELATIONSHIP_TYPE_OBJECT_OBJECT\\n \\\"\\\"\\\"The type of the relationship O-O\\\"\\\"\\\" \\n\\n\\nclass SubjectSimilarityRule(_SimilarityRule):\\n \\\"\\\"\\\"A class for representing inter-subject similarity.\\\"\\\"\\\"\\n\\n relationship_type = RELATIONSHIP_TYPE_SUBJECT_SUBJECT\\n \\\"\\\"\\\"The type of the relationship S-S\\\"\\\"\\\" \\n\\n\\nclass SubjectObjectRule(_BaseRule):\\n \\\"\\\"\\\"A class for representing subject-object preference for recommendation\\\"\\\"\\\"\\n\\n relationship_type = RELATIONSHIP_TYPE_SUBJECT_OBJECT\\n \\\"\\\"\\\"The type of the relationship S-O\\\"\\\"\\\" \\n\\n @classmethod\\n def _order_in_pair(cls, dn_arg1, dn_arg2, ds_arg1, ds_arg2):\\n return PredictedRelationship._order_in_pair(dn_arg1, dn_arg2, ds_arg1, ds_arg2)\\n \\nclass SubjectObjectSimilarityRule(_SimilarityRelationship):\\n \\\"\\\"\\\"A class used only when subject domain equals object domain. \\n For representing inter-entity rule.\\n \\\"\\\"\\\"\\n \\n relationship_type = RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT\\n \\\"\\\"\\\"The type of the relationship SO-SO\\\"\\\"\\\"\\n\\n \\n\"\n },\n {\n \"alpha_fraction\": 0.6410256624221802,\n \"alphanum_fraction\": 0.6538461446762085,\n \"avg_line_length\": 25,\n \"blob_id\": \"8bd5bbee2b8e0dfe1070d9756b7487bd06c7114b\",\n \"content_id\": \"c5c7bf0826c526e2a2d602c31e306b308e2b8563\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 156,\n \"license_type\": \"no_license\",\n \"max_line_length\": 62,\n \"num_lines\": 6,\n \"path\": \"/code/adapter/dump.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/bin/sh\\n\\n# dump the database to given file (relative from ../db/dumps/)\\nmysqldump --opt -u root adapter > ../db/dumps/$1\\n\\necho \\\"dumped to ../db/dumps/$1\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.6247366070747375,\n \"alphanum_fraction\": 0.626033365726471,\n \"avg_line_length\": 37.316768646240234,\n \"blob_id\": \"cca61489cbc79c8f905c9f2ec184854d1dec0025\",\n \"content_id\": \"c8c35fff934115f4b210c138e687c3751a25175c\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 6169,\n \"license_type\": \"no_license\",\n \"max_line_length\": 89,\n \"num_lines\": 161,\n \"path\": \"/UnresystCD/code/adapter/unresyst/models/common.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The models that are common to the whole unresyst application.\\\"\\\"\\\"\\n\\nfrom django.db import models\\n\\nfrom unresyst.constants import *\\n\\nclass Recommender(models.Model):\\n \\\"\\\"\\\"The representation of a recommender. \\n \\n There can be multiple recommenders for one parent system.\\n \\\"\\\"\\\"\\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"The name of the recommender\\\"\\\"\\\"\\n \\n class_name = models.CharField(max_length=MAX_LENGTH_CLASS_NAME, unique=True)\\n \\\"\\\"\\\"The name of the recommender class. Has to be unique.\\\"\\\"\\\"\\n\\n are_subjects_objects = models.BooleanField()\\n \\\"\\\"\\\"Are subjects == objects for the recommender?\\\"\\\"\\\"\\n \\n is_built = models.BooleanField()\\n \\\"\\\"\\\"Is the recommender built?\\\"\\\"\\\"\\n \\n random_recommendation_description = models.TextField(default='', blank=True)\\n \\\"\\\"\\\"The description of the relationship/rule instance.\\\"\\\"\\\" \\n \\n remove_predicted_from_recommendations = models.BooleanField()\\n \\\"\\\"\\\"Should the objects that are already \\\"liked\\\" be removed from \\n recommendations?\\\"\\\"\\\"\\n \\n class Meta:\\n app_label = 'unresyst'\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n\\n\\nclass SubjectObject(models.Model):\\n \\\"\\\"\\\"The common representation for a subject and an object.\\\"\\\"\\\"\\n \\n id_in_specific = models.CharField(max_length=MAX_LENGTH_ID)\\n \\\"\\\"\\\"The id of the subject/object in the domain-specific system.\\\"\\\"\\\"\\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"A textual characterization of the subject/object\\\"\\\"\\\"\\n \\n entity_type = models.CharField(max_length=MAX_LENGTH_ENTITY_TYPE, \\\\\\n choices=ENTITY_TYPE_CHOICES)\\n \\\"\\\"\\\"A string indicating whether it's a subject, object or both.s/o/so\\\"\\\"\\\"\\n \\n recommender = models.ForeignKey('unresyst.Recommender')\\n \\\"\\\"\\\"The recommender to which the subject/object belongs.\\\"\\\"\\\"\\n \\n\\n class Meta:\\n app_label = 'unresyst' \\n \\n unique_together = ('id_in_specific', 'entity_type', 'recommender')\\n \\\"\\\"\\\"There can be only one subject/object with the given id and \\n recommender.\\n \\\"\\\"\\\"\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n \\n @classmethod\\n def get_domain_neutral_entity(cls, domain_specific_entity, entity_type, recommender):\\n \\\"\\\"\\\"Get domain neutral representation of the given domain specific entity\\n (subject/object/subjectobject)\\n \\n @type domain_specific_entity: django.db.models.Model\\n @param domain_specific_entity: the domain specific subject/object/\\n subjectobject for which the domain neutral repre should be got\\n \\n @type entity_type: str\\n @param entity_type: 'S'/'O'/'SO' .. see constants, determines whether\\n the entity is a subject, object or both\\n \\n @type recommender: models.Recommender\\n @param recommended: the recommender the for which the entity should \\n be got\\n \\n @rtype: models.SubjectObject\\n @returns: the domain neutral representation of the entity \\n \\n @raise MultipleObjectsReturned: when there's a broken constraint in \\n the unresyst database (hopefully never)\\n @raise DoesNotExist: when the domain neutral representation for \\n the given entity does not exist\\n \\\"\\\"\\\"\\n return cls.objects.get(\\n id_in_specific=domain_specific_entity.pk,\\n entity_type=entity_type,\\n recommender=recommender)\\n \\n def get_domain_specific_entity(self, entity_manager):\\n \\\"\\\"\\\"Get domain specific subject/object/both for this universal \\n representation.\\n \\n @type entity_manager: django.db.models.manager.Manager\\n @param entity_manager: the manager over the model containing \\n the domain specific subjects/objects/bot\\n \\n @rtype: models.Model\\n @returns: the domain specific entity for this universal representation\\n \\n @raise MultipleObjectsReturned: when there's a broken constraint in \\n the client database (hopefully never)\\n @raise DoesNotExist: when the domain specific entity for \\n this universal entity does not exist\\n \\\"\\\"\\\"\\n return entity_manager.get(pk=self.id_in_specific)\\n\\n @classmethod\\n def unique_pairs(cls, recommender, entity_type):\\n \\\"\\\"\\\"A generator looping through the pairs of subjectobjects so that each two \\n object set is returned only once. \\n \\n E.g. after (a, b), the (b, a) pair isn't returned. \\n It doesn't return pairs like (a, a).\\n \\n Useful for symmetric rule and relationship evaluation.\\n \\n @type recommender: models.Recommender\\n @param recommender: the recommender model for which the pairs should be\\n obtained\\n\\n @type entity_type: str ('S', 'O' or 'SO') or None\\n @param entity_type: the entity type from whic the pair should be taken,\\n if None, all entity types are taken\\n \\n @rtype: generator for two-tuples\\n @returns: pairs of subjectobjects entity_type entities belonging to \\n the recommender.\\n \\\"\\\"\\\"\\n ent_type_kwargs = {} if entity_type is None \\\\\\n else {'entity_type': entity_type}\\n \\n # get the subjectobjects to iterate over\\n qs_entities = cls.objects.filter(\\n recommender=recommender, \\n **ent_type_kwargs).order_by('id')\\n\\n # the number of entities\\n entity_count = qs_entities.count() \\n \\n # get the first argument and the number of entities that \\n # will be taken as the second arg. Starting from 1, \\n # finishing at .\\n # The first entity will never be used as second argument \\n for arg1, count in zip( \\\\\\n qs_entities[1:].iterator(), \\\\\\n range(1, entity_count)):\\n\\n # obtain only first count entities\\n for arg2 in qs_entities[:count].iterator():\\n\\n yield (arg1, arg2)\\n\"\n },\n {\n \"alpha_fraction\": 0.7341772317886353,\n \"alphanum_fraction\": 0.7341772317886353,\n \"avg_line_length\": 18.75,\n \"blob_id\": \"be3fa2be861725aaa0b158caf2c22896177abf5b\",\n \"content_id\": \"a301faf6bd839a0f7365d3ea650a3ddbf69c3595\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 79,\n \"license_type\": \"no_license\",\n \"max_line_length\": 44,\n \"num_lines\": 4,\n \"path\": \"/code/adapter/unresyst/tests/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The module containing tests for unresyst.\\n\\nIt uses the demo appliation.\\n\\\"\\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.7977900505065918,\n \"alphanum_fraction\": 0.8016999363899231,\n \"avg_line_length\": 28.263681411743164,\n \"blob_id\": \"087d0f7433293c7351326b79e62c3a817fdaf1c6\",\n \"content_id\": \"5753a3ca5f8ff52fc6a7e772132a383b686a18d9\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 11765,\n \"license_type\": \"no_license\",\n \"max_line_length\": 149,\n \"num_lines\": 402,\n \"path\": \"/code/adapter/evaluate.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/bin/sh\\n\\n# parametry: \\n# 1. co chces evaluovat: lastfm, demo\\n# 2. cislo iterace\\n\\n\\nLASTFM=false\\nDEMO=false\\n\\nfor param in $*;\\ndo \\n case $param in\\n 'lastfm')\\n LASTFM=true\\n ;;\\n 'demo')\\n DEMO=true\\n ;;\\n esac\\ndone\\n\\nif [ $LASTFM = true ]\\nthen\\n echo \\\"Creating test set.\\\"\\n echo \\\"from lastfm.recommender import *; ArtistRecommender.ValidationPairClass.select_validation_pairs($2); quit()\\\" | python ./manage.py shell \\n echo \\\"Building and evaluating recommender.\\\"\\n echo \\\"from lastfm.recommender import *; ArtistRecommender.build(); ArtistRecommender.evaluate(); quit();\\\" | python ./manage.py shell\\nfi\\n\\n# LASTFM\\n#\\n#\\n\\n# specificky pro lastfm - dva test sety\\nfrom lastfm.models import *\\nBaseArtistEvaluationPair.select()\\n\\n\\n\\n# evaluate novel mahout recommender:\\n# \\n\\n# export training data\\nfrom lastfm.mahout_recommender import *\\nNovelMahoutArtistRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv')\\n\\n# export test data\\nfrom lastfm.evaluation import *\\nNovelArtistRankEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_test.csv')\\n\\n# run mahout train, test -> lastfm_predictions.csv\\ncd ../mahout/mahoutrec\\n./unresystpredict.sh lastfm\\n\\n# import predictions\\nfrom lastfm.mahout_recommender import *\\nNovelMahoutArtistRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_predictions.csv')\\n\\n# evaluate the predictions \\nfrom lastfm.mahout_recommender import *\\nfrom lastfm.evaluation import *\\nNovelArtistRankEvaluator.evaluate_predictions(NovelMahoutArtistRecommender)\\n\\n# evaluate nenovel mahout recommender:\\n# \\n\\n# export training data\\nfrom lastfm.mahout_recommender import *\\nMahoutArtistRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv')\\n\\n# export test data\\nfrom lastfm.evaluation import *\\nArtistRankEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_test.csv')\\n\\n# run mahout train, test -> lastfm_predictions.csv\\ncd ../mahout/mahoutrec\\n./unresystpredict.sh lastfm\\n\\n# import predictions\\nfrom lastfm.mahout_recommender import *\\nMahoutArtistRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_predictions.csv')\\n\\n# evaluate the predictions \\nfrom lastfm.mahout_recommender import *\\nfrom lastfm.evaluation import *\\nArtistRankEvaluator.evaluate_predictions(MahoutArtistRecommender)\\n\\n\\n\\n# below\\n#\\n\\n# build, \\n#save_all_to_predictions must be true\\n./build.sh lastfm\\n\\n# evaluate and save the predictions\\nfrom lastfm.evaluation import *\\nfrom lastfm.recommender import *\\nNovelArtistRecommenderEvaluator.evaluate_predictions(NovelArtistRecommender, save_predictions=True)\\n\\n# export predictions\\nfrom lastfm.recommender import *\\nNovelArtistRecommender.export_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv')\\n\\n# export test data\\nfrom lastfm.evaluation import *\\nNovelArtistRankEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_test.csv')\\n\\n# run mahout train, test -> flixster_predictions.csv\\ncd ../mahout/mahoutrec\\n./unresystpredict.sh lastfm\\n\\n# update unresyst predictions with the obtained\\nfrom lastfm.recommender import *\\nNovelArtistRecommender.update_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_predictions.csv')\\n\\n# run evaluation\\n./lastfmeval.sh novel dontbuild\\n\\n\\n\\n# recommendations - netreba\\n# \\n\\n# evaluate unresyst\\nfrom lastfm.evaluation import *\\nfrom lastfm.recommender import *\\nArtistRecommenderEvaluator.evaluate_recommendations(ArtistRecommender, 10)\\n\\n# evaluate mahout recommender\\n# nenovel:\\n\\n# export training data\\nfrom lastfm.mahout_recommender import *\\nMahoutArtistRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv')\\n\\n# run mahout train, test -> lastfm_recommendations.csv\\ncd ../mahout/mahoutrec\\n./unresystrecommend.sh\\n\\n# import predictions\\nfrom lastfm.mahout_recommender import *\\nMahoutArtistRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_recommendations.csv')\\n\\n# evaluate the recommendations \\nfrom lastfm.mahout_recommender import *\\nfrom lastfm.evaluation import *\\nArtistRecommenderEvaluator.evaluate_recommendations(MahoutArtistRecommender, 10)\\n\\n# evaluate mahout recommender\\n# novel:\\n\\n# export training data\\nfrom lastfm.mahout_recommender import *\\nNovelMahoutArtistRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv')\\n\\n# run mahout train, test -> lastfm_recommendations.csv\\ncd ../mahout/mahoutrec\\n./unresystrecommend.sh\\n\\n# import predictions\\nfrom lastfm.mahout_recommender import *\\nNovelMahoutArtistRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_recommendations.csv')\\n\\n# evaluate the recommendations \\nfrom lastfm.mahout_recommender import *\\nfrom lastfm.evaluation import *\\nNovelArtistRecommenderEvaluator.evaluate_recommendations(NovelMahoutArtistRecommender, 10)\\n\\n\\n# Flixster\\n#\\n#\\n\\n# select test data - obecne\\nfrom flixster.evaluation import *\\nMovieRecommenderEvaluator.select_evaluation_pairs()\\n\\n# evaluate mahout recommender - predictions:\\n#\\n\\n# export training data\\nfrom flixster.mahout_recommender import *\\nMahoutMovieRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv')\\n\\n# export test data\\nfrom flixster.evaluation import *\\nMovieRecommenderEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_test.csv')\\n\\n# run mahout train, test -> lastfm_predictions.csv\\ncd ../mahout/mahoutrec\\n./unresystpredict.sh flixster\\n\\n# import predictions\\nfrom flixster.mahout_recommender import *\\nMahoutMovieRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_predictions.csv')\\n\\n# evaluate the predictions \\nfrom flixster.mahout_recommender import *\\nfrom flixster.evaluation import *\\nMovieRecommenderEvaluator.evaluate_predictions(MahoutMovieRecommender)\\n\\n\\n# evaluate mahout recommender - recommendations\\n# \\n\\n# export training data\\nfrom flixster.mahout_recommender import *\\nMahoutMovieRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv')\\n\\n# run mahout train, test -> flixster_recommendations.csv\\ncd ../mahout/mahoutrec\\n./unresystrecommend.sh flixster\\n\\n# import predictions\\nfrom flixster.mahout_recommender import *\\nMahoutMovieRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_recommendations.csv')\\n\\n# evaluate the recommendations \\nfrom flixster.mahout_recommender import *\\nfrom flixster.evaluation import *\\nMovieRecommenderEvaluator.evaluate_recommendations(MahoutMovieRecommender, 10)\\n\\n\\n# hybrid\\n#\\n\\n# build save_all_to_predictions must be true\\n./build.sh flixster\\n\\n# export as usual\\n\\n# export training data\\nfrom flixster.mahout_recommender import *\\nMahoutMovieRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv')\\n\\n# export test data\\nfrom flixster.evaluation import *\\nMovieRecommenderEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_test.csv')\\n\\n# run mahout train, test -> flixster_predictions.csv\\ncd ../mahout/mahoutrec\\n./unresystpredict.sh flixster\\n\\n# evaluate what we have and save the predictions\\nfrom flixster.evaluation import *\\nfrom flixster.recommender import *\\nMovieRecommenderEvaluator.evaluate_predictions(MovieRecommender, save_predictions=True)\\n\\n# update unresyst predictions with the obtained\\nfrom flixster.recommender import *\\nMovieRecommender.update_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_predictions.csv')\\n\\n# run evaluation\\n./flixstereval.sh preds dontbuild\\n\\n\\n# below\\n#\\n\\n# build, \\n#save_all_to_predictions must be true\\n./build.sh flixster\\n\\n# evaluate and save the predictions\\nfrom flixster.evaluation import *\\nfrom flixster.recommender import *\\nMovieRecommenderEvaluator.evaluate_predictions(MovieRecommender, save_predictions=True)\\n\\n# export predictions\\nfrom flixster.recommender import *\\nMovieRecommender.export_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv')\\n\\n# export test data\\nfrom flixster.evaluation import *\\nMovieRecommenderEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_test.csv')\\n\\n# run mahout train, test -> flixster_predictions.csv\\ncd ../mahout/mahoutrec\\n./unresystpredict.sh flixster\\n\\n# import predictions\\nfrom flixster.mahout_recommender import *\\nMahoutMovieRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_predictions.csv')\\n\\n# evaluate the predictions \\nfrom flixster.mahout_recommender import *\\nfrom flixster.evaluation import *\\nMovieRecommenderEvaluator.evaluate_predictions(MahoutMovieRecommender)\\n\\n\\n# Travel\\n#\\n#\\n\\n# select test data - obecne\\nfrom travel.models import *\\nTourOrderEvalPair.select()\\n\\n# evaluate mahout recommender - recommendations\\n# \\n\\n# export training data\\nfrom travel.mahout_recommender import *\\nMahoutOrderTourRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_train.csv')\\n\\n# run mahout train, test -> travel_recommendations.csv\\ncd ../mahout/mahoutrec\\n./unresystrecommend.sh travel \\n\\n# import predictions\\nfrom travel.mahout_recommender import *\\nMahoutOrderTourRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_recommendations.csv')\\n\\n# evaluate the recommendations \\nfrom travel.mahout_recommender import *\\nfrom travel.evaluation import *\\nOrderTourRecommenderEvaluator.evaluate_recommendations(MahoutOrderTourRecommender, 10)\\n\\n# evaluate mahout recommender - predictions\\n# \\n\\n# export training data\\nfrom travel.mahout_recommender import *\\nMahoutOrderTourRecommender.export_data('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_train.csv')\\n\\n# export test data\\nfrom travel.evaluation import *\\nOrderTourRecommenderEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_test.csv')\\n\\n# run mahout train, test -> travel_recommendations.csv\\ncd ../mahout/mahoutrec\\n./unresystpredict.sh travel \\n\\n# import predictions\\nfrom travel.mahout_recommender import *\\nMahoutOrderTourRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_predictions.csv')\\n\\n# evaluate the predictions\\nfrom travel.evaluation import *\\nfrom travel.mahout_recommender import *\\nOrderTourRankEvaluator.evaluate_predictions(MahoutOrderTourRecommender)\\n\\n\\n# below\\n#\\n\\n# build, \\n#save_all_to_predictions must be true\\n./build.sh travel\\n\\n# evaluate and save the predictions\\nfrom travel.evaluation import *\\nfrom travel.recommender import *\\nOrderTourRecommenderEvaluator.evaluate_predictions(OrderTourRecommender, save_predictions=True)\\n\\n# export predictions\\nfrom travel.recommender import *\\nOrderTourRecommender.export_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_train.csv')\\n\\n# export test data\\nfrom travel.evaluation import *\\nOrderTourRecommenderEvaluator.export_evaluation_pairs('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_test.csv')\\n\\n# run mahout train, test -> flixster_predictions.csv\\ncd ../mahout/mahoutrec\\n./unresystpredict.sh travel\\n\\n# update unresyst predictions with the obtained\\nfrom travel.recommender import *\\nOrderTourRecommender.update_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_predictions.csv')\\n\\n# run evaluation\\n./traveleval.sh preds dontbuild\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n#############################\\n\\n# import predictions\\nfrom travel.mahout_recommender import *\\nMahoutOrderTourRecommender.import_predictions('/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_predictions.csv')\\n\\n# evaluate the predictions \\nfrom travel.mahout_recommender import *\\nfrom travel.evaluation import *\\nOrderTourRankEvaluator.evaluate_predictions(MahoutOrderTourRecommender)\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5666928291320801,\n \"alphanum_fraction\": 0.5733941793441772,\n \"avg_line_length\": 42.68224334716797,\n \"blob_id\": \"289c86e2785511d7f624b95de018ebc8dd829c1d\",\n \"content_id\": \"450b9f3c477029466ae38cc5422f510325ccfca6\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 14027,\n \"license_type\": \"no_license\",\n \"max_line_length\": 118,\n \"num_lines\": 321,\n \"path\": \"/code/adapter/unresyst/algorithm/algorithm.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The main classes of the algorithm package\\\"\\\"\\\"\\n\\nfrom django.db.models import Q\\nfrom django.db.models import Avg\\n\\nfrom base import BaseAlgorithm\\nfrom unresyst.constants import *\\nfrom unresyst.models.abstractor import PredictedRelationshipDefinition, \\\\\\n RelationshipInstance\\nfrom unresyst.models.aggregator import AggregatedRelationshipInstance\\nfrom unresyst.models.algorithm import RelationshipPredictionInstance\\nfrom unresyst.models.common import SubjectObject\\nfrom unresyst.recommender.rules import BaseRelationship\\n\\nclass Algorithm(BaseAlgorithm):\\n \\\"\\\"\\\"The original deprecated implementation mixing combinator, compilator and algorithm. \\n \\n A simple implementation of a recommender algorithm.\\n \\n The remove_predicted_from_recommendations is implemented by adding a zero\\n prediction for all pairs in predicted relationship. \\n \\n The get_recommendation function returns only recommendations with expectancy\\n above zero or the expectancy_limit.\\n \\\"\\\"\\\"\\n \\n N_NEIGHBOURHOOD = 10\\n \\\"\\\"\\\"The maximum size of the neighbourhood, from which the similar items are taken\\\"\\\"\\\"\\n\\n # Build phase:\\n #\\n \\n def build(self, recommender_model):\\n \\\"\\\"\\\"Build the recommender - create the instances of the \\n RelationshipPredictionInstance model where there is some simple \\n prediction available. Where there isn't, leave it.\\n \\\"\\\"\\\" \\n\\n\\n\\n # Recommend phase:\\n #\\n\\n\\n def get_relationship_prediction(self, recommender_model, dn_subject, dn_object, remove_predicted):\\n \\\"\\\"\\\"See the base class for the documentation.\\n \\\"\\\"\\\"\\n # if predicted should be removed and the pair is in the predicted_rel, \\n # return the special expectancy value\\n if remove_predicted: \\n # all predicted relationships\\n qs_predicted = RelationshipInstance.filter_predicted(recommender_model)\\n\\n # the relationship between dn_subject and dn_object\\n qs_predicted_rel = RelationshipInstance.filter_relationships(dn_subject, dn_object, queryset=qs_predicted)\\n\\n # if the prediction for the pair exists\\n if qs_predicted_rel:\\n \\n # return the special expectancy value \\n assert len(qs_predicted_rel) == 1\\n predicted = qs_predicted_rel[0]\\n\\n return self._get_already_in_relatinship_prediction(\\n recommender_model=recommender_model,\\n predicted_relationship=predicted)\\n \\n \\n # filter the predictions for recommender\\n qs_rec_pred = RelationshipPredictionInstance.objects.filter(\\n recommender=recommender_model)\\n\\n # filter the predictions for the given pair \\n qs_pred = RelationshipPredictionInstance.filter_relationships(\\n object1=dn_subject,\\n object2=dn_object,\\n queryset=qs_rec_pred) \\n \\n # if available return it\\n if qs_pred: \\n assert len(qs_pred) == 1\\n return qs_pred[0]\\n \\n # if it's not available, maybe it wasn't in the N_NEIGHBOURHOOD, \\n # so try finding it in aggregates \\n so1, so2 = BaseRelationship.order_arguments(dn_subject, dn_object)\\n\\n # the Subject-object aggregate (just in case it hasn't been built) \\n qs_rels = AggregatedRelationshipInstance.objects.filter(\\n subject_object1=so1,\\n subject_object2=so2,\\n recommender=recommender_model)\\n \\n # if found return it\\n if qs_rels:\\n assert len(qs_rels) == 1\\n\\n pred = RelationshipPredictionInstance(\\n subject_object1=dn_subject,\\n subject_object2=dn_object,\\n description=qs_rels[0].description,\\n recommender=recommender_model,\\n expectancy=qs_rels[0].expectancy\\n )\\n pred.save()\\n return pred\\n\\n # the definition of the predicted relationship\\n d = PredictedRelationshipDefinition.objects.get(recommender=recommender_model)\\n \\n # try finding the similar entity to the one liked by so1\\n # content-based \\n \\n # get similarities starting the traverse with the similarity.\\n qs_sim1 = AggregatedRelationshipInstance.objects\\\\\\n .filter(recommender=recommender_model)\\\\\\n .filter(\\n Q(\\n # take so2 as stable - in the subject_object2 position of the similarity\\n subject_object2=so2,\\n \\n # traverse from the other object in similarity (subject_object1) through\\n # the relationship instance, its subject (subject_object1) must be so1\\n subject_object1__relationshipinstance_relationships2__subject_object1=so1,\\n \\n # the relationship instance definition must be the predicted relationship def\\n subject_object1__relationshipinstance_relationships2__definition=d) | \\\\\\n Q( \\n # take so2 as stable again, now in the subject_object1 position of the similarity\\n subject_object1=so2, \\n \\n # traverse from the other through relationship to so1\\n subject_object2__relationshipinstance_relationships2__subject_object1=so1,\\n \\n # the definition again must be the predicted\\n subject_object2__relationshipinstance_relationships2__definition=d))\\n\\n # if found return\\n if qs_sim1:\\n\\n # if found return the average TODO nejak jinak\\n avg = qs_sim1.aggregate(Avg('expectancy'))\\n\\n pred = RelationshipPredictionInstance(\\n subject_object1=dn_subject,\\n subject_object2=dn_object,\\n description=qs_sim1[0].description,\\n recommender=recommender_model,\\n expectancy=avg['expectancy__avg']\\n )\\n pred.save()\\n return pred \\n\\n \\n # try finding the similar entity (user) to entity that liked so2\\n # cf\\n qs_sim2 = AggregatedRelationshipInstance.objects\\\\\\n .filter(recommender=recommender_model)\\\\\\n .filter(\\n Q(\\n # take so1 as stable - in the subject_object2 position of the similarity\\n subject_object2=so1,\\n \\n # traverse from the other object in similarity (subject_object1) through\\n # the relationship instance, its object (subject_object2) must be so2\\n subject_object1__relationshipinstance_relationships1__subject_object2=so2,\\n \\n # the relationship instance definition must be the predicted relationship def\\n subject_object1__relationshipinstance_relationships1__definition=d) | \\\\\\n Q( \\n # take so1 as stable again, now in the subject_object1 position of the similarity\\n subject_object1=so1, \\n \\n # traverse from the other through relationship to so2\\n subject_object2__relationshipinstance_relationships1__subject_object2=so2,\\n \\n # the definition again must be the predicted\\n subject_object2__relationshipinstance_relationships1__definition=d))\\n\\n # if found return the average TODO nejak jinak\\n if qs_sim2:\\n avg = qs_sim2.aggregate(Avg('expectancy'))\\n \\n pred = RelationshipPredictionInstance(\\n subject_object1=dn_subject,\\n subject_object2=dn_object,\\n description=qs_sim2[0].description,\\n recommender=recommender_model,\\n expectancy=avg['expectancy__avg']\\n )\\n pred.save()\\n\\n return pred \\n \\n \\n # return the uncertain\\n return self._get_uncertain_prediction(\\n recommender_model=recommender_model, \\n dn_subject=dn_subject, \\n dn_object=dn_object\\n )\\n \\n @classmethod\\n def get_recommendations(self, recommender_model, dn_subject, count, expectancy_limit, remove_predicted):\\n \\\"\\\"\\\"See the base class for documentation.\\n \\\"\\\"\\\" \\n \\n # get the recommendations ordered by the expectancy from the largest\\n recommendations = RelationshipPredictionInstance\\\\\\n .get_relationships(obj=dn_subject)\\\\\\n .filter(recommender=recommender_model, \\n expectancy__gt=expectancy_limit)\\\\\\n .order_by('-expectancy') \\n \\n # remove the predicted from the recommendations if it should be done \\n if remove_predicted:\\n \\n # get objects that are already liked\\n #\\n \\n qs_predicted = RelationshipInstance.filter_predicted(recommender_model) \\n\\n # get ids of subjectobjects where dn_subject appears\\n qs_predicted = qs_predicted.filter( \\n Q(subject_object1=dn_subject) | Q(subject_object2=dn_subject) \\n ).values_list('subject_object1__pk', 'subject_object2__pk')\\n \\n # flatten it and take only the objects\\n import itertools \\n predicted_obj_ids = [i for i in itertools.chain(*qs_predicted) if i <> dn_subject.pk] \\n\\n # remove the already liked objects\\n #\\n \\n # if it is a subjectobject we have to take both places\\n if dn_subject.entity_type == ENTITY_TYPE_SUBJECTOBJECT:\\n recommendations = recommendations.exclude(\\n Q(subject_object1__id__in=predicted_obj_ids) \\n | Q(subject_object2__id__in=predicted_obj_ids)\\n )\\n # otherwise one field is enough\\n else:\\n recommendations = recommendations.exclude(\\n subject_object2__id__in=predicted_obj_ids\\n )\\n \\n # if there should be more recommendations than we have and \\n # the expectancy limit is below the uncertain, add uncertain \\n # predictions\\n if recommendations.count() < count \\\\\\n and expectancy_limit < UNCERTAIN_PREDICTION_VALUE: \\n\\n object_ent_type = ENTITY_TYPE_SUBJECTOBJECT \\\\\\n if recommender_model.are_subjects_objects \\\\\\n else ENTITY_TYPE_OBJECT \\n\\n # all objects excluding the objects recommended to dn_subject\\n # the last exclude is just for SO entity_type - we won't recommend \\n # subject to himself\\n uncertain_objects = SubjectObject.objects\\\\\\n .filter(entity_type=object_ent_type, recommender=recommender_model)\\\\\\n .exclude(relationshippredictioninstance_relationships1__subject_object2=dn_subject)\\\\\\n .exclude(relationshippredictioninstance_relationships2__subject_object1=dn_subject)\\\\\\n .exclude(pk=dn_subject.pk)\\n \\n # if predicted should be removed remove them\\n if remove_predicted:\\n uncertain_objects = uncertain_objects.exclude(id__in=predicted_obj_ids)\\n \\n # divide the recommendations into groups exp => uncertain, exp < uncertain\\n positive_preds = recommendations.filter(\\n expectancy__gte=UNCERTAIN_PREDICTION_VALUE)\\n negative_preds = recommendations.filter(\\n expectancy__lt=UNCERTAIN_PREDICTION_VALUE)\\n \\n # how many non-positive recommendations we want\\n required_count = count - positive_preds.count()\\n uncertain_count = uncertain_objects.count()\\n \\n # if there're enough uncertains to fill the recommendations, \\n # clear the negative\\n if uncertain_count >= required_count:\\n uncertain_required_count = required_count\\n negative_preds = []\\n else:\\n # otherwise we want all uncertains we have\\n uncertain_required_count = uncertain_count\\n\\n negative_required_count = required_count - uncertain_count\\n \\n # and for the rest use negative - only those above limit\\n negative_preds = negative_preds.filter(\\n expectancy__gt=expectancy_limit)[:negative_required_count]\\n\\n uncertain_preds = []\\n \\n # construct the uncertain recommendations\\n for obj in uncertain_objects[:uncertain_required_count]:\\n pred = self._get_uncertain_prediction(recommender_model, dn_subject, obj)\\n uncertain_preds.append(pred)\\n\\n return list(positive_preds) + uncertain_preds + list(negative_preds)\\n\\n # apply the limit for count\\n return list(recommendations[:count]) \\n \\n\\n\\n\\nclass PredictOnlyAlgorithm(Algorithm):\\n \\\"\\\"\\\"An algorithm that doesn't need to be built, performs only predictions.\\\"\\\"\\\"\\n \\n @classmethod\\n def build(self, recommender_model):\\n \\\"\\\"\\\"Do nothing\\\"\\\"\\\"\\n return\\n\\n @classmethod\\n def get_recommendations(self, recommender_model, dn_subject, count, expectancy_limit, remove_predicted): \\n \\\"\\\"\\\"Raise an error\\\"\\\"\\\"\\n raise NotImplementedError() \\n\\n\"\n },\n {\n \"alpha_fraction\": 0.8039215803146362,\n \"alphanum_fraction\": 0.8039215803146362,\n \"avg_line_length\": 28.14285659790039,\n \"blob_id\": \"a484a682b0fc5c982c0fc6ff7cf9830676019511\",\n \"content_id\": \"64faa2b3c3f268bf2b3bbde6854976dfc739be87\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 204,\n \"license_type\": \"no_license\",\n \"max_line_length\": 65,\n \"num_lines\": 7,\n \"path\": \"/UnresystCD/code/adapter/unresyst/models/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The models for universal object and subject representation.\\\"\\\"\\\"\\n\\nfrom common import *\\nfrom abstractor import *\\nfrom aggregator import *\\nfrom algorithm import *\\nfrom evaluation import BaseEvaluationPair\\n\"\n },\n {\n \"alpha_fraction\": 0.767624020576477,\n \"alphanum_fraction\": 0.7780678868293762,\n \"avg_line_length\": 41.66666793823242,\n \"blob_id\": \"2e74aacfa2f641a63cab4215ad7f65d7f104fbb7\",\n \"content_id\": \"6c8174bf9ac2980362b139acc357b55dc8bfb19b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 383,\n \"license_type\": \"no_license\",\n \"max_line_length\": 140,\n \"num_lines\": 9,\n \"path\": \"/README.md\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"**a recommender you can't resist ...**\\n\\nA prototype of a Universal Recommender System created as a part of my Master thesis.\\n\\nThe thesis was selected to the second round of the 2011 ACM Spy competition. See http://www.acm-spy.cz/o-soutezi-acm-spy/galerie-nejlepsich/\\n\\n[Mahout tutorial created in the process](docs/CreateMahoutRecommender.md)\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6669912338256836,\n \"alphanum_fraction\": 0.6689386367797852,\n \"avg_line_length\": 21.30434799194336,\n \"blob_id\": \"0d4fcd4835f521b92feba9b34fdf57d2934fa3db\",\n \"content_id\": \"dc47bba31c34243b602de93b48388b652365c4d7\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 1027,\n \"license_type\": \"no_license\",\n \"max_line_length\": 180,\n \"num_lines\": 46,\n \"path\": \"/code/adapter/flixstereval.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/bin/sh\\n\\n# run the evaluation- recommendations. parameters \\n# recs -> recommendations\\n# preds -> predictions\\n# dontbuild-> don't build\\n\\nBUILD=true\\nRECS=false\\nPREDS=false\\n\\nfor param in $*;\\ndo \\n case $param in\\n 'recs')\\n RECS=true\\n ;;\\n 'preds')\\n PREDS=true\\n ;;\\n 'dontbuild')\\n BUILD=false\\n ;;\\n esac\\ndone\\n\\nif [ $BUILD = true ]\\nthen\\n # build\\n echo \\\"Building...\\\"\\n echo \\\"from flixster.recommender import *; MovieRecommender.build()\\\" | python ./manage.py shell\\nfi\\n\\nif [ $RECS = true ]\\nthen\\n echo \\\"Evaluating recommendations...\\\"\\n echo \\\"from flixster.evaluation import *; from flixster.recommender import *; MovieRecommenderEvaluator.evaluate_recommendations(MovieRecommender, 10)\\\"| python ./manage.py shell\\nfi\\n\\nif [ $PREDS = true ]\\nthen\\n echo \\\"Evaluating predictions...\\\"\\n echo \\\"from flixster.evaluation import *; from flixster.recommender import *; MovieRecommenderEvaluator.evaluate_predictions(MovieRecommender)\\\"| python ./manage.py shell\\nfi\\n\\necho \\\"\\\"\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6990595459938049,\n \"alphanum_fraction\": 0.7241379022598267,\n \"avg_line_length\": 25.41666603088379,\n \"blob_id\": \"d3ada56d0457dd1687f89e0a71e4c80664ce37f2\",\n \"content_id\": \"41e744138d8492ced894d29c82cb75314744fc87\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 319,\n \"license_type\": \"no_license\",\n \"max_line_length\": 62,\n \"num_lines\": 12,\n \"path\": \"/UnresystCD/code/adapter/lastfm/constants.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The constants for last.fm\\\"\\\"\\\"\\n\\nMAX_LENGTH_NAME = 200\\n\\nCROSS_VALIDATION_COUNT = 5\\n\\\"\\\"\\\"The number of test sets for the cross-validation\\\"\\\"\\\"\\n\\nSUCCESS_LIMIT = 0.5\\n\\\"\\\"\\\"The limit above which the pair is considered successful.\\\"\\\"\\\"\\n\\nEXPECTED_EXPECTANCY_LISTENED = 1.0\\n\\\"\\\"\\\"The expected expectancy for the listened to artists.\\\"\\\"\\\"\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5753098130226135,\n \"alphanum_fraction\": 0.586749255657196,\n \"avg_line_length\": 35.48695755004883,\n \"blob_id\": \"91f7f98bac6c9825d846bc4908b450a2160faab3\",\n \"content_id\": \"88fedd9c7d16e9e763b85e8db64f69241d3ae8f7\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 12588,\n \"license_type\": \"no_license\",\n \"max_line_length\": 102,\n \"num_lines\": 345,\n \"path\": \"/UnresystCD/code/adapter/unresyst/models/symmetric.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Support for symmetric relationships\\\"\\\"\\\"\\n\\nfrom django.db import models\\nfrom django.db.models import Q\\nfrom django.core.exceptions import MultipleObjectsReturned\\n\\nfrom unresyst.exceptions import SymmetryError, UnresystError as BaseError\\n\\nclass SymmetricalRelationship(models.Model):\\n \\\"\\\"\\\"A base class for all symmetrical relationship models. The relationship can't be \\n between the same object.\\n \\n Main custom methods:\\n - rel.get_related(obj): object related to obj on rel\\n - rel.contains_object(obj): does the rel contain obj?\\n\\n Only for additional_unique == ():\\n - cls.get_related_objects(obj): objects related to obj\\n - cls.get_relationships(obj): relationship objects related to obj\\n - cls.get_relationship(obj1, obj2): get relationship between obj1 and obj2, or None\\n - cls.are_related(obj1, obj2): are obj1 and obj2 related?\\n\\n Also for additional_unique <> ():\\n - cls.filter_relationships(obj1, obj2, queryset=None): filter relationships between\\n obj1 and obj2, from queryset or from all relationships\\n\\n \\\"\\\"\\\"\\n \\n attr_name1 = ''\\n \\\"\\\"\\\"The name of the first attribute pointing on some model. Should be overriden\\\"\\\"\\\"\\n \\n attr_name2 = ''\\n \\\"\\\"\\\"The name of the second attribute pointing on some model. Should be overriden\\\"\\\"\\\"\\n \\n additional_unique = ()\\n \\\"\\\"\\\"A tuple containing names of attributes that are unique together with the symmetrical\\n relationship pair. Like unique_together = (('attr1', 'attr2'), 'additional_unique')\\n \\\"\\\"\\\"\\n\\n class Meta:\\n abstract = True\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n object1 = getattr(self, self.attr_name1)\\n object2 = getattr(self, self.attr_name2)\\n \\n return u\\\"%s - %s\\\" % (object1, object2)\\n \\n def save(self, *args, **kwargs):\\n \\\"\\\"\\\"Overrides the save method in order to control the symmetricity.\\n\\n @raise SymmetryError: if there's something wrong with the saved entities in the\\n means of symmetry \\n \\\"\\\"\\\"\\n \\n # get the objects that should be saved as related\\n object1 = getattr(self, self.attr_name1)\\n object2 = getattr(self, self.attr_name2)\\n \\n # if they're identical, raise an error\\n # the equality of entity models is already solved by overriding the __eq__ method\\n if object1 == object2: \\n raise SymmetryError(\\n message=\\\"The related objects can't be identical\\\",\\n object1=object1, \\n object2=object2)\\n \\n # if the relationship exists in some direction\\n # raise another error.\\n # \\n # try the saving direction\\n self.__check_save(object1, object2)\\n \\n # try the other direction\\n self.__check_save(object2, object1)\\n \\n # save the relationship \\n return super(SymmetricalRelationship, self).save(*args, **kwargs) \\n \\n \\n def __check_save(self, object1, object2):\\n \\\"\\\"\\\"Check whether the the relationship between the objects in the given direction\\n can be saved. If not, throw an exception. Includes checks for update.\\\"\\\"\\\"\\n \\n # create a dictionary of additional kwargs\\n additional_kwargs = {}\\n for ad_un in self.additional_unique:\\n additional_kwargs[ad_un] = getattr(self, ad_un)\\n \\n # try finding the relation in the database\\n rel = self.__class__.__get_relation(object1, object2, **additional_kwargs)\\n\\n # if we have found a relation between object1 and object2\\n if rel:\\n \\n # if it's new value to insert (no id set)\\n # or it's a value to update, and the found relationship isn't the same \\n # as the one we're saving\\n if not self.id or \\\\\\n (self.id and self.id != rel.id):\\n \\n # than it's an error. \\n raise SymmetryError(\\n message=\\\"The relationship is already in the database.\\\",\\n object1=object1, \\n object2=object2) \\n\\n @classmethod\\n def __get_related_objects_in_attribute(cls, obj, attr_name1, attr_name2):\\n \\\"\\\"\\\"Get objects related to obj that are in the attr_name2 attribute...\\n obj is in the attr_name1 attribute.\\\"\\\"\\\"\\n \\n # construct the kwargs\\n kwargs = {attr_name1: obj}\\n \\n # get the relationships where obj is in the first attribute\\n relationships = cls.objects.filter(**kwargs)\\n \\n # get the objects that are in the second attribute\\n related_objects = [getattr(relationship, attr_name2) for relationship in relationships]\\n \\n return related_objects\\n \\n\\n @classmethod \\n def get_related_objects(cls, obj):\\n \\\"\\\"\\\"Get objects that are in relationship with obj.\\n \\n Returns just objects (not the whole binding items).\\n \\n @type obj: core.models.BaseEntity\\n @param obj: the object whose related objects should be got\\n \\n @rtype: list of core.models.BaseEntity\\n @return: the list of objects related to obj\\n \\\"\\\"\\\"\\n # relationships where obj is as the first\\n related_objects1 = cls.__get_related_objects_in_attribute(obj, cls.attr_name1, cls.attr_name2)\\n\\n # relationships where obj is as the second \\n related_objects2 = cls.__get_related_objects_in_attribute(obj, cls.attr_name2, cls.attr_name1)\\n \\n return related_objects1 + related_objects2\\n \\n @classmethod \\n def get_relationships(cls, obj):\\n \\\"\\\"\\\"Get bindings for objects that are in relationship with the object obj.\\n \\n Returns whole binding items (not just objects).\\n \\n The related objects can be got using get_related. Example:\\n\\n connections = get_relationships(pepa)\\n \\n for connection in connections:\\n print connection.connection_comment\\n \\n object_connected_to_pepa = connection.get_related(pepa)\\n # ...\\n \\n @type obj: core.models.BaseEntity\\n @param obj: the object whose relationships should be got\\n \\n @rtype: QuerySet of relationship class\\n @return: relationships related to obj \\n \\\"\\\"\\\"\\n # get relationships where there is cls at first or second place\\n # \\n kwargs1 = {cls.attr_name1: obj}\\n kwargs2 = {cls.attr_name2: obj}\\n\\n from django.db.models import Q\\n q = cls.objects.filter(\\n Q(**kwargs1) | Q(**kwargs2)\\n )\\n\\n return q\\n \\n \\n def get_related(self, obj):\\n \\\"\\\"\\\"On a relationship, get the object that is related to obj, no matter in which\\n attribute it is.\\n \\n To be used in the combination with the get_relationships method.\\n\\n @type obj: core.models.BaseEntity\\n @param obj: the object whose related object should be got\\n \\n @rtype: core.models.BaseEntity\\n @return: the object in the relationship no matter in which attribute it is.\\n \\n @raise SymmetryError: if the obj isn't a part of the relationship on which \\n the method is called\\n \\\"\\\"\\\"\\n # get the objects in the relationship\\n object1 = getattr(self, self.attr_name1)\\n object2 = getattr(self, self.attr_name2)\\n \\n # if obj ist the first one, return the second\\n # the equality of entity models is already solved by overriding the __eq__ method\\n if obj == object1: \\n return object2\\n \\n # if it's the second, return the first\\n if obj == object2:\\n return object1\\n \\n # otherwise return an error\\n raise SymmetryError(\\n message=\\\"The relationship object doesn't contain the '%s' object\\\" % obj,\\n object1=object1, \\n object2=object2) \\n \\n def contains_object(self, obj):\\n \\\"\\\"\\\"Does the relationship contain the obj object? On one of the sides.\\n \\n @type obj: object\\n @param obj: the object to check whether it's in the relationship\\n \\n @rtype: bool\\n @return: Whether the object is contained on one side of the relationship\\n \\\"\\\"\\\"\\n\\n # get the objects in the relationship\\n object1 = getattr(self, self.attr_name1)\\n object2 = getattr(self, self.attr_name2) \\n \\n return obj==object1 or obj==object2\\n\\n @classmethod\\n def __contains_relation(cls, object1, object2):\\n \\\"\\\"\\\"Does model in database contain the relationship of object1 and object2?\\\"\\\"\\\"\\n \\n return bool(cls.__get_relation(object1, object2))\\n \\n @classmethod\\n def __get_relation(cls, object1, object2, **kwargs): \\n \\\"\\\"\\\"Get the relation of the two objects in the given direction. If it doesn't exist\\n return None.\\n kwargs are additional arguments for filter, when additional_unique <> ()\\n \\\"\\\"\\\"\\n \\n # prepare kwargs for the filter function\\n filter_kwargs = {\\n cls.attr_name1: object1,\\n cls.attr_name2: object2\\n }\\n filter_kwargs.update(kwargs)\\n \\n # try to get the relationship\\n query = cls.objects.filter(**filter_kwargs)\\n\\n # if some exists\\n if query:\\n if query.count() > 1:\\n raise MultipleObjectsReturned(\\n 'Too many obtained relationships for the pair %s, %s.' % (object1, object2))\\n \\n # return it \\n return query[0]\\n \\n # otherwise there's nothing\\n return None \\n \\n \\n @classmethod\\n def get_relationship(cls, object1, object2):\\n \\\"\\\"\\\"Get the relationship between the two objects if they're related.\\n If not return None.\\n \\n Takes into account both directions of the symmetry.\\n \\n @type object1: Model\\n @param object1: the first object \\n \\n @type object2: Model\\n @param object2: the second object \\n \\n @rtype: relationship object\\n @return: the relationship between the two objects or None if they aren't related\\n \\\"\\\"\\\" \\n # try the first direction\\n relationship = cls.__get_relation(object1, object2)\\n if relationship:\\n return relationship \\n \\n # try the second\\n return cls.__get_relation(object2, object1) \\n\\n @classmethod\\n def filter_relationships(cls, object1, object2, queryset=None):\\n \\\"\\\"\\\"Filter relationships so that it contains only the ones between \\n object1 and object2. \\n For models where sets (object1, object2) are unique does the same \\n as get_relationships (except returning a queryset)\\n\\n @type object1: Model\\n @param object1: the first object \\n \\n @type object2: Model\\n @param object2: the second object \\n \\n @type queryset: django queryset\\n @param queryset: the queryset to filter (optional). If not passed, all\\n objects of the model are filtered, otherwise the passed queryset is\\n filtered.\\n \\n @rtype: django queryset\\n @return: queryset containing relationships between object1 and object2\\n \\\"\\\"\\\"\\n if queryset is None:\\n queryset = cls.objects.all()\\n # prepare the filter kwargs:\\n kwargs1 = {\\n cls.attr_name1: object1,\\n cls.attr_name2: object2\\n } \\n kwargs2 = {\\n cls.attr_name1: object2,\\n cls.attr_name2: object1\\n }\\n\\n # filter relationships that have the pair in the given or the opposite\\n # direction \\n return queryset.filter(Q(**kwargs1) | Q(**kwargs2)) \\n\\n @classmethod\\n def are_related(cls, object1, object2):\\n \\\"\\\"\\\"Is the object related to other object?\\n \\n Takes into account both directions of the symmetry.\\n \\n @type object1: Model\\n @param object1: the first object \\n \\n @type object2: Model\\n @param object2: the second object \\n \\n @rtype: bool\\n @return: True if they're in relationship, False otherwise. \\n \\\"\\\"\\\" \\n \\n return cls.__contains_relation(object1, object2) \\\\\\n or cls.__contains_relation(object2, object1)\\n\"\n },\n {\n \"alpha_fraction\": 0.5918495059013367,\n \"alphanum_fraction\": 0.5918495059013367,\n \"avg_line_length\": 29.673076629638672,\n \"blob_id\": \"c3de9e6129bcafd863cd330ac281345fb563cc8c\",\n \"content_id\": \"7749ad6678f29b47d38275ab1ff6d16d157c5669\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1595,\n \"license_type\": \"no_license\",\n \"max_line_length\": 83,\n \"num_lines\": 52,\n \"path\": \"/UnresystCD/code/adapter/unresyst/recommender/predictions.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The prediction classes. Instances of the class are returned by \\nthe recommender.\\n\\\"\\\"\\\"\\n\\nclass RelationshipPrediction(object):\\n \\\"\\\"\\\"The prediction of the predicted_relationship appearing between\\n the given subject-object pair.\\n \\n @type subject: the domain-specific subject\\n @ivar subject: the subject \\n \\n @type object_: the domain-specific object\\n @ivar object_: the object\\n \\n @type expectancy: float\\n @ivar expectancy: the estimated probability of the predict_relationship\\n occuring between the subject and the object\\n \\n @type explanation: str\\n @ivar explanation: the explanation for the prediction \\n \\\"\\\"\\\"\\n \\n def __init__(self, subject, object_, expectancy, is_uncertain, explanation=''):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n self.subject = subject\\n \\\"\\\"\\\"The subject\\\"\\\"\\\"\\n \\n self.object_ = object_\\n \\\"\\\"\\\"The object\\\"\\\"\\\"\\n \\n self.expectancy = expectancy\\n \\\"\\\"\\\"The estimated probability of the predicted_relationship\\n occuring between the subject and the object.\\n \\\"\\\"\\\"\\n \\n self.is_uncertain = is_uncertain\\n \\\"\\\"\\\"Is the prediction made without having any information available?\\\"\\\"\\\"\\n \\n self.explanation = explanation\\n \\\"\\\"\\\"The explanation for the prediction\\\"\\\"\\\"\\n\\n def __unicode__(self):\\n return u\\\"%s <- %s: %f, %s\\\" % (\\n self.subject, \\n self.object_, \\n self.expectancy, \\n self.explanation\\n )\\n \\n def __repr__(self):\\n return \\\"< %s >\\\" % str(self.__unicode__())\\n\"\n },\n {\n \"alpha_fraction\": 0.5630881786346436,\n \"alphanum_fraction\": 0.5655280351638794,\n \"avg_line_length\": 37.911563873291016,\n \"blob_id\": \"0116196a027d45f1e52daa1218f4d26d0ea5222f\",\n \"content_id\": \"3e2969c510c10aabccea92e3412b06b5eac85d9a\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5738,\n \"license_type\": \"no_license\",\n \"max_line_length\": 102,\n \"num_lines\": 147,\n \"path\": \"/UnresystCD/code/adapter/unresyst/recommender/clusters.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The classes for defining clusters by users\\\"\\\"\\\"\\n\\nfrom unresyst.models.abstractor import ClusterSet, Cluster, ClusterMember\\nfrom unresyst.models.common import SubjectObject\\nfrom unresyst.exceptions import ConfigurationError\\nfrom unresyst.constants import *\\n\\nclass BaseClusterSet(object):\\n \\\"\\\"\\\"The base class for all clusters sets. Cluster set is a set of clusters,\\n e.g. cluster set 'Gender' contains clusters 'Male', 'Female'.\\\"\\\"\\\"\\n \\n entity_type = None\\n \\\"\\\"\\\"The type of the entities in the clusters\\\"\\\"\\\" \\n \\n entity_format_str = None\\n \\\"\\\"\\\"Formating string used in the description\\\"\\\"\\\" \\n \\n def __init__(self, name, weight, filter_entities, get_cluster_confidence_pairs, description=None):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n\\n self.name = name\\n \\\"\\\"\\\"The name of the rule/relationship.\\\"\\\"\\\"\\n \\n self.weight = weight\\n \\\"\\\"\\\"The significance of the similarity inferred from two subject/objects\\n belonging to the same cluster. A number from [0, 1]\\\"\\\"\\\"\\n \\n self.filter_entities = filter_entities\\n \\\"\\\"\\\"A queryset of all entities of the given type (s/o/so) belonging\\n to some cluster.\\n \\\"\\\"\\\"\\n \\n self.get_cluster_confidence_pairs = get_cluster_confidence_pairs\\n \\\"\\\"\\\"A generator returning pairs of cluster, confidence for the given entity.\\n E.g. for 'Liars' it returns (('Punk', 0.41), ('Indie', 0.72))\\n Parameters: domain specific entity\\n Return: the pairs: (name of the cluster, confidence - number from [0, 1])\\n \\\"\\\"\\\"\\n \\n self.description = description\\n \\\"\\\"\\\"The description of the membership. Can contain: placeholders for \\n subject/object/subjectobject and cluster.\\\"\\\"\\\"\\n\\n \\n def evaluate(self):\\n \\\"\\\"\\\"Crate the cluster set in the database, its clusters, bindings \\n of subjectobjects to the clusters.\\n \\\"\\\"\\\"\\n \\n if not (MIN_WEIGHT <= self.weight <= MAX_WEIGHT):\\n raise ConfigurationError(\\n message=(\\\"The set '%s' provides weight %f,\\\" + \\n \\\" should be between 0 and 1. .\\\"\\n ) % (self.name, self.weight),\\n recommender=self.recommender,\\n parameter_name=\\\"Recommender.cluster_sets\\\",\\n parameter_value=(self.recommender.cluster_sets)\\n )\\n \\n recommender_model = self.recommender._get_recommender_model()\\n\\n # create the cluster set in the database\\n cluster_set = ClusterSet(\\n name=self.name,\\n recommender=recommender_model,\\n entity_type=self.entity_type,\\n weight=self.weight)\\n \\n cluster_set.save()\\n \\n # go through the entities create clusters on demand\\n #\\n \\n for ds_entity in self.filter_entities:\\n \\n # get entity cluster-confidence pairs\\n cluster_conf_pairs = self.get_cluster_confidence_pairs(ds_entity)\\n \\n # convert the entity to universal\\n dn_entity = SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=ds_entity, \\n entity_type=self.entity_type, \\n recommender=recommender_model)\\n \\n # go through the entity clusters\\n for cluster_name, confidence in cluster_conf_pairs:\\n \\n # if confidence invalid through an error\\n if not (MIN_CONFIDENCE <= confidence <= MAX_CONFIDENCE):\\n raise ConfigurationError(\\n message=(\\\"The cluster set '%s' provides confidence %f,\\\" + \\n \\\" should be between 0 and 1. For cluster %s.\\\"\\n ) % (self.name, confidence, cluster_name),\\n recommender=self.recommender,\\n parameter_name=\\\"Recommender.cluster_sets\\\",\\n parameter_value=(self.recommender.cluster_sets)\\n )\\n \\n # get or create the cluster \\n cluster, x = Cluster.objects.get_or_create(\\n name=cluster_name[:MAX_LENGTH_NAME],\\n cluster_set=cluster_set)\\n\\n # evaluate the description\\n if self.description:\\n description = self.description % {\\n self.entity_format_str: dn_entity.name,\\n FORMAT_STR_CLUSTER: cluster_name} \\n else:\\n description = '' \\n \\n # save the binding of the cluster to the dn_entity\\n member = ClusterMember.objects.create(\\n cluster=cluster,\\n member=dn_entity,\\n confidence=confidence,\\n description=description)\\n \\n print \\\" %d clusters and %d cluster members for '%s' cluster set created.\\\" \\\\\\n % (Cluster.objects.filter(cluster_set=cluster_set).count(), \\n ClusterMember.objects.filter(cluster__cluster_set=cluster_set).count(),\\n self.name)\\n \\n \\n \\n\\nclass SubjectClusterSet(BaseClusterSet):\\n \\\"\\\"\\\"Cluster set for subjects\\\"\\\"\\\"\\n \\n entity_type = ENTITY_TYPE_SUBJECT\\n \\n entity_format_str = FORMAT_STR_SUBJECT\\n\\n \\nclass ObjectClusterSet(BaseClusterSet):\\n \\\"\\\"\\\"Cluster set for objects\\\"\\\"\\\"\\n \\n entity_type = ENTITY_TYPE_OBJECT\\n\\n entity_format_str = FORMAT_STR_OBJECT\\n \\nclass SubjectObjectClusterSet(BaseClusterSet):\\n \\\"\\\"\\\"Clusters for recommenders where subject domain == object domain\\\"\\\"\\\"\\n \\n entity_type = ENTITY_TYPE_SUBJECTOBJECT\\n \\n entity_format_str = FORMAT_STR_SUBJECTOBJECT\\n \\n \\n\"\n },\n {\n \"alpha_fraction\": 0.6540930867195129,\n \"alphanum_fraction\": 0.6540930867195129,\n \"avg_line_length\": 35.64706039428711,\n \"blob_id\": \"4d7905dc524c2474531e811f3a1aae03f13e6d2f\",\n \"content_id\": \"f8b6a8b980ba159a6a68de40716df35f3b869fa6\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1246,\n \"license_type\": \"no_license\",\n \"max_line_length\": 91,\n \"num_lines\": 34,\n \"path\": \"/code/adapter/unresyst/algorithm/aggregating_algorithm.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The aggregating algorithm class\\\"\\\"\\\"\\n\\nfrom base import BaseAlgorithm\\n\\nclass AggregatingAlgorithm(BaseAlgorithm):\\n \\\"\\\"\\\"The algorithm that aggregates the similarity relationships and biases,\\n uses the compiling algorithm as its inner_algorithm.\\n \\\"\\\"\\\"\\n \\n def __init__(self, inner_algorithm, aggregator):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n super(AggregatingAlgorithm, self).__init__(inner_algorithm=inner_algorithm)\\n \\n self.aggregator=aggregator\\n \\\"\\\"\\\"The aggregator that will be used for aggregating the relationships\\n and biases\\\"\\\"\\\"\\n \\n def build(self, recommender_model):\\n \\\"\\\"\\\"See the base class for documentation.\\n \\n Aggregates and calls the inner algorithm build\\n \\\"\\\"\\\"\\n \\n # aggregate the relationships and rules\\n self.aggregator.aggregate_rules_relationships(\\n recommender_model=recommender_model) \\n \\n # aggregate the biases\\n self.aggregator.aggregate_biases(recommender_model=recommender_model)\\n \\n print \\\"Rules, relationships and biases aggregated. Building the inner algorithm...\\\"\\n \\n super(AggregatingAlgorithm, self).build(recommender_model=recommender_model)\\n\"\n },\n {\n \"alpha_fraction\": 0.800000011920929,\n \"alphanum_fraction\": 0.800000011920929,\n \"avg_line_length\": 34,\n \"blob_id\": \"679060bf3bd6466a49a1124ea8f36296e390d08a\",\n \"content_id\": \"833db83b59cc5685a7a8c8a23cf84f40fb5faad7\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 105,\n \"license_type\": \"no_license\",\n \"max_line_length\": 64,\n \"num_lines\": 3,\n \"path\": \"/code/adapter/unresyst/abstractor/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The package contains all classes for the abstractor layer.\\\"\\\"\\\"\\n\\nfrom abstractor import BasicAbstractor\\n\"\n },\n {\n \"alpha_fraction\": 0.649068295955658,\n \"alphanum_fraction\": 0.6537266969680786,\n \"avg_line_length\": 25.58333396911621,\n \"blob_id\": \"60fb7f1a8af2838251c1467c5ac671819f0b936d\",\n \"content_id\": \"24ec29c615b0a824413690c8dc8206e22887bb0d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1288,\n \"license_type\": \"no_license\",\n \"max_line_length\": 66,\n \"num_lines\": 48,\n \"path\": \"/UnresystCD/code/adapter/demo/views.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The views for the demo application\\\"\\\"\\\"\\n\\nfrom django.http import HttpResponse\\nfrom django.template import RequestContext, Context, loader\\nfrom django.shortcuts import render_to_response, get_object_or_404\\n\\n\\nfrom models import User\\nfrom recommender import ShoeRecommender\\n\\ndef view_recommendations(request, user_name):\\n \\\"\\\"\\\"The view for the demo recommendations\\\"\\\"\\\"\\n \\n # get the user and user list from db\\n user = get_object_or_404(User, name__iexact=user_name)\\n user_list = User.objects.all()\\n \\n # create user recommendations\\n recommendations = ShoeRecommender.get_recommendations(user)\\n \\n context = {\\n 'user': user,\\n 'user_list': user_list,\\n 'recommendations': recommendations,\\n 'show_expectancy': True,\\n }\\n\\n return render_to_response(\\n 'demo/shoe_recommendations.html',\\n dictionary=context, \\n context_instance=RequestContext(request)\\n )\\n \\ndef view_home_page(request):\\n \\\"\\\"\\\"The view for the demo home page\\\"\\\"\\\"\\n\\n # get the user list from db\\n user_list = User.objects.all()\\n \\n context = {\\n 'user_list': user_list,\\n }\\n\\n return render_to_response(\\n 'demo/home_page.html',\\n dictionary=context, \\n context_instance=RequestContext(request)\\n ) \\n \\n \\n\"\n },\n {\n \"alpha_fraction\": 0.7196261882781982,\n \"alphanum_fraction\": 0.7196261882781982,\n \"avg_line_length\": 34.66666793823242,\n \"blob_id\": \"289625693b56eb880e7853b71d3a9da2dfe44d3a\",\n \"content_id\": \"88fdadf5a9aed29723056a56032d930de9d36b85\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 107,\n \"license_type\": \"no_license\",\n \"max_line_length\": 80,\n \"num_lines\": 3,\n \"path\": \"/code/adapter/unresyst/tests/test_update.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Testing the update\\\"\\\"\\\"\\n\\n# idea - porovnat vysledky recommenderu po updatu a po rebuildu - mely by byt stejne\\n\"\n },\n {\n \"alpha_fraction\": 0.521481454372406,\n \"alphanum_fraction\": 0.5237036943435669,\n \"avg_line_length\": 41.44094467163086,\n \"blob_id\": \"f4c944268e3c3cf02f08c1ccd2f6057c2662cc16\",\n \"content_id\": \"034d4cc30e78aff70b78a057e7aa4392dfe5b7f3\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5400,\n \"license_type\": \"no_license\",\n \"max_line_length\": 91,\n \"num_lines\": 127,\n \"path\": \"/code/adapter/unresyst/tests/test_base.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The base classes for the tests used in unresyst\\\"\\\"\\\"\\n\\nfrom django.test import TestCase\\n\\nfrom unresyst.models.common import SubjectObject \\n\\nfrom demo.recommender import ShoeRecommender, AverageRecommender\\nfrom demo.models import User, ShoePair\\n\\nclass DBTestCase(TestCase):\\n \\\"\\\"\\\"A base class for all tests which need database testing data\\\"\\\"\\\"\\n\\n def setUp(self):\\n \\\"\\\"\\\"Insert data into the database\\\"\\\"\\\"\\n\\n # insert test data\\n from demo.save_data import save_data\\n save_data()\\n \\n def save_entities(self):\\n \\\"\\\"\\\"Save instances of the entities to the testcase isntance\\n to be called in subclasses.\\n \\n Not distinguishing the recommender.\\\"\\\"\\\"\\n \\n self.specific_entities = {\\n 'Alice': User.objects.get(name=\\\"Alice\\\"),\\n 'Bob': User.objects.get(name=\\\"Bob\\\"),\\n 'Cindy': User.objects.get(name=\\\"Cindy\\\"),\\n 'Daisy': User.objects.get(name=\\\"Daisy\\\"),\\n 'Edgar': User.objects.get(name=\\\"Edgar\\\"),\\n 'Fionna': User.objects.get(name=\\\"Fionna\\\"),\\n 'Sneakers': ShoePair.objects.get(name=\\\"Sneakers\\\"),\\n \\\"Rubber Shoes\\\": ShoePair.objects.get(name=\\\"Rubber Shoes\\\"),\\n 'RS 130': ShoePair.objects.get(name='RS 130'),\\n 'Design Shoes': ShoePair.objects.get(name='Design Shoes'),\\n 'Octane SL': ShoePair.objects.get(name='Octane SL'), \\n }\\n \\n rm = self.recommender._get_recommender_model()\\n self.universal_entities = {\\n 'Alice': SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=self.specific_entities['Alice'], \\n entity_type='S', \\n recommender=rm),\\n 'Bob': SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=self.specific_entities['Bob'], \\n entity_type='S', \\n recommender=rm),\\n 'Cindy': SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=self.specific_entities['Cindy'], \\n entity_type='S', \\n recommender=rm),\\n 'Daisy': SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=self.specific_entities['Daisy'], \\n entity_type='S', \\n recommender=rm),\\n 'Edgar': SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=self.specific_entities['Edgar'], \\n entity_type='S', \\n recommender=rm),\\n 'Fionna': SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=self.specific_entities['Fionna'], \\n entity_type='S', \\n recommender=rm), \\n 'Sneakers': SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=self.specific_entities['Sneakers'], \\n entity_type='O', \\n recommender=rm),\\n 'Rubber Shoes': SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=self.specific_entities['Rubber Shoes'], \\n entity_type='O', \\n recommender=rm),\\n 'RS 130': SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=self.specific_entities['RS 130'], \\n entity_type='O', \\n recommender=rm),\\n 'Design Shoes': SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=self.specific_entities['Design Shoes'], \\n entity_type='O', \\n recommender=rm), \\n 'Octane SL': SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=self.specific_entities['Octane SL'], \\n entity_type='O', \\n recommender=rm), \\n } \\n\\nclass TestBuild(DBTestCase):\\n \\\"\\\"\\\"The base class performing build in the setup.\\\"\\\"\\\"\\n\\n def setUp(self):\\n \\\"\\\"\\\"The setup for all tests - build the recommender\\\"\\\"\\\"\\n super(TestBuild, self).setUp()\\n\\n # call the tested function \\n ShoeRecommender.build()\\n\\n self.recommender = ShoeRecommender\\n\\nclass TestBuildAverage(DBTestCase):\\n \\\"\\\"\\\"The base class performing AverageRecommender build in the setup\\\"\\\"\\\"\\n \\n def setUp(self):\\n \\\"\\\"\\\"The setup for all tests - build the recommender\\\"\\\"\\\"\\n super(TestBuildAverage, self).setUp()\\n\\n # call the tested function \\n AverageRecommender.build()\\n \\n self.recommender = AverageRecommender\\n \\n self.save_entities()\\n \\n\\nclass TestEntities(TestBuild):\\n \\\"\\\"\\\"The base class adding universal and specific entitits \\n to the test instance\\n \\\"\\\"\\\"\\n \\n def setUp(self):\\n \\\"\\\"\\\"Obtain specific and universal subject objects \\n and store them in the test instance\\n \\\"\\\"\\\" \\n \\n super(TestEntities, self).setUp()\\n \\n self.save_entities()\\n \\n\\n\"\n },\n {\n \"alpha_fraction\": 0.7019230723381042,\n \"alphanum_fraction\": 0.7211538553237915,\n \"avg_line_length\": 28.714284896850586,\n \"blob_id\": \"74f9f8da729072e90e61e65cb1f7c03ef577fb22\",\n \"content_id\": \"e9d1b16262fe2758fb6a519a8879452fd278a510\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 208,\n \"license_type\": \"no_license\",\n \"max_line_length\": 69,\n \"num_lines\": 7,\n \"path\": \"/code/adapter/demo/constants.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Constants for the demo application.\\\"\\\"\\\"\\n\\nMAX_LENGTH_NAME = 40\\n\\\"\\\"\\\"The maximum length of the name in the universal representation.\\\"\\\"\\\"\\n\\nMAX_LENGTH_IMAGE_PATH = 50\\n\\\"\\\"\\\"The maximum length of the path to image\\\"\\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.5149863958358765,\n \"alphanum_fraction\": 0.5599455237388611,\n \"avg_line_length\": 32.318180084228516,\n \"blob_id\": \"9abe4d84c17f395a26390477ed959a3bad67f354\",\n \"content_id\": \"1492be4f5e1bc974e9d0d001fa3fba7c297b9b3e\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 734,\n \"license_type\": \"no_license\",\n \"max_line_length\": 79,\n \"num_lines\": 22,\n \"path\": \"/UnresystCD/code/dataset_scripts/process_travel.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"grep diplomka_time_order travel.sql -m 1 -n\\n\\n# shorten, remove unused\\ntail -n +38 diplomka\\\\(2\\\\).sql > travel1.sql \\nhead -n 302073 travel1.sql > travel2.sql\\n\\n# away with the insert part \\nsed -e 's/.*(//' travel2.sql > travel3.sql \\n\\n# away with the end \\nsed -e 's/);.*//' travel3.sql > travel4.sql \\n\\n# remove ', ' from places between appostrophes\\nsed -e \\\"s/', '\\\\(.[^']*\\\\),\\\\([^']*\\\\)', '/', '\\\\1\\\\2', '/\\\" travel4.sql > travel5.sql\\nsed -e \\\"s/', '\\\\(.[^']*\\\\),\\\\([^']*\\\\)', '/', '\\\\1\\\\2', '/\\\" travel5.sql > travel6.sql\\nsed -e \\\"s/', '\\\\(.[^']*\\\\),\\\\([^']*\\\\)', '/', '\\\\1\\\\2', '/\\\" travel6.sql > travel7.sql\\n\\n# away with spaces behind commas \\nsed -e 's/, /,/g' travel7.sql > travel8.sql\\n\\n# away with appostrophes \\nsed -e \\\"s/'//g\\\" travel8.sql > travel9.sql\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6000116467475891,\n \"alphanum_fraction\": 0.6046186089515686,\n \"avg_line_length\": 42.78717803955078,\n \"blob_id\": \"4b60e53462fc33d0c54b4c6ff40808d6e82690b0\",\n \"content_id\": \"06cf1f57e05adba80edb00c93f1600db52142aaa\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 17148,\n \"license_type\": \"no_license\",\n \"max_line_length\": 126,\n \"num_lines\": 390,\n \"path\": \"/UnresystCD/code/adapter/unresyst/combinator/base.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Base classes for the combinator layer:\\n - BaseCombinator\\n\\\"\\\"\\\"\\nfrom unresyst.exceptions import CombinatorError\\nfrom unresyst.models.abstractor import RelationshipInstance, RuleInstance, \\\\\\n ExplicitRuleInstance, PredictedRelationshipDefinition, ClusterMember\\nfrom unresyst.models.aggregator import AggregatedRelationshipInstance, AggregatedBiasInstance\\nfrom unresyst.models.algorithm import RelationshipPredictionInstance\\nfrom unresyst.constants import *\\n\\nclass BaseCombinator(object):\\n \\\"\\\"\\\"The base class defining the interface of all combinators.\\n interface methods:\\n - combine_pair_similarities\\n - combine_entity_biases\\n - combine_pair_prediction_elements\\n - choose_promising_objects\\n \\n methods to be overriden:\\n - _combine \\n \\n helper methods for subclasses:\\n - _concat_descriptions \\n \\\"\\\"\\\"\\n \\n DIVISOR = 3\\n \\\"\\\"\\\"A constant for dividing the relationship members that would otherwise be \\n too numerous\\\"\\\"\\\"\\n \\n def __init__(self):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n self.top_bias_objects = None\\n\\n def _checked_combine(self, combination_elements, ResultClass):\\n \\\"\\\"\\\"Check if something was given and call the overriden _combine method\\n \\\"\\\"\\\"\\n if not combination_elements:\\n raise CombinatorError(\\\"No combination_elements given\\\")\\n return self._combine(combination_elements, ResultClass)\\n \\n def _combine(self, combination_elements, ResultClass):\\n \\\"\\\"\\\"Combine the combination elements to produce an instance \\n of the ResultClass, filling in its expectancy and description.\\n \\n Overriden in subclasses\\n\\n @type combination_elements: a list of BaseCombinationElement\\n @param combination_elements: the elements that should be combined\\n\\n @type ResultClass: class\\n @param ResultClass: the class which instance will be returned\\n \\n @rtype: ResultClass\\n @return: the combination class with filled expectancy and description\\n \\\"\\\"\\\"\\n pass\\n\\n @staticmethod\\n def _concat_descriptions(element_list):\\n \\\"\\\"\\\"Concat descriptions of the elements in the element_list.\\n \\n @type element_list: a list of BaseCombinationElement\\n @param element_list: a list of elements in order that they should\\n appear\\n \\n @rtype: str\\n @return: the string concatenation of the element descriptions\\n \\\"\\\"\\\"\\n list_len = len(element_list)\\n \\n assert list_len > 0\\n \\n if list_len == 1:\\n return element_list[0].get_description()\\n \\n # shorten the list in case it's too long\\n short_list = element_list[:MAX_REASONS_DESCRIPTION]\\n \\n # join the descriptions\\n joined_desc = ' '.join([\\\"%s: %s\\\" % ((REASON_STR % i), e.get_description()) \\\\\\n for e, i in zip(short_list, range(1, list_len + 1))])\\n \\n # if the list was shortened add a message\\n if list_len > MAX_REASONS_DESCRIPTION:\\n joined_desc += ' ' + MORE_REASONS_STR % (list_len - MAX_REASONS_DESCRIPTION)\\n \\n return joined_desc\\n\\n \\n\\n def combine_pair_similarities(self, combination_elements):\\n \\\"\\\"\\\"Aggregate similarities of the given pair S-S, O-O, or SO-SO. \\n \\n @type combination_elements: a list of BaseCombinationElement\\n @param combination_elements: the elements that should be combined \\n \\n @rtype: AggregatedRelationshipInstance\\n @return: the aggregated relationship with filled expectancy and\\n description, other fields are empty.\\n \\\"\\\"\\\" \\n return self._checked_combine(\\n combination_elements=combination_elements, \\n ResultClass=AggregatedRelationshipInstance)\\n \\n\\n def combine_entity_biases(self, entity_biases):\\n \\\"\\\"\\\"Aggregate biases of the given entity S, O, or SO.\\n \\n @type entitiy_biases: QuerySet, each member having a get_expectancy()\\n method\\n @param entitiy_biases: the biases to aggregate\\n \\n @rtype: float\\n @return: the expectancy that the entity will be in \\n a predicted_relationship, aggregated from its biases\\n \\\"\\\"\\\"\\n return self._checked_combine(\\n combination_elements=entity_biases,\\n ResultClass=AggregatedBiasInstance)\\n\\n\\n def combine_pair_prediction_elements(self, combination_elements):\\n \\\"\\\"\\\"Combine all preference sources producing predictions\\\"\\\"\\\"\\n return self._checked_combine(\\n combination_elements=combination_elements,\\n ResultClass=RelationshipPredictionInstance)\\n\\n \\n def choose_promising_objects(self, dn_subject, min_count):\\n \\\"\\\"\\\"Choose at least min_count objects that are likely to be interesting\\n for the dn_subject, using various preference sources, choosing by \\n a heuristic.\\n \\n Get maximum of min_count objects from each category\\n - aggregated object bias \\n - s-o relationships \\n - predicted_relationship + object_similarities\\n - predicted_relationship + subject similarities\\n - predicted_relationship + subject cluster memberships \\n - predicted_relationship + object cluster memberships \\n \\n @type dn_subject: SubjectObject\\n @param dn_subject: the subject to choose the objects for\\n \\n @type min_count: int\\n @param min_count: the minimum count that is chosen (if it's available)\\n \\n @rtype: iterable of SubjectObject\\n @return: the promising objects\\n \\\"\\\"\\\" \\n \\n if not min_count:\\n return [] \\n \\n recommender_model = dn_subject.recommender\\n \\n object_ent_type = ENTITY_TYPE_SUBJECTOBJECT \\\\\\n if recommender_model.are_subjects_objects else \\\\\\n ENTITY_TYPE_OBJECT\\n \\n # aggregated object bias\\n #\\n \\n # use some caching\\n if self.top_bias_objects is None:\\n \\n qs_biases = AggregatedBiasInstance.objects\\\\\\n .filter(recommender=recommender_model, \\n subject_object__entity_type=object_ent_type,\\n expectancy__gt=UNCERTAIN_PREDICTION_VALUE)\\\\\\n .order_by('-expectancy')\\n \\n self.top_bias_objects = [(b.subject_object, b.expectancy) for b in qs_biases[:min_count]]\\n \\n top_bias_objs = self.top_bias_objects\\n \\n # s-o rules, relationships, explicit rules\\n #\\n \\n top_rel_objs = self._get_promising_objects_rules_relationships(\\n dn_subject=dn_subject,\\n min_count=min_count,\\n recommender_model=recommender_model)\\n \\n \\n # predicted_relationship + object_similarities\\n # predicted_relationship + subject similarities\\n \\n top_sim_objs = self._get_promising_objects_similarities(\\n dn_subject=dn_subject,\\n min_count=min_count,\\n recommender_model=recommender_model)\\n \\n \\n\\n # predicted_relationship + subject cluster memberships \\n # predicted_relationship + object cluster memberships\\n \\n cluster_objs = self._get_promising_objects_clusters(\\n dn_subject=dn_subject,\\n min_count=min_count,\\n recommender_model=recommender_model)\\n\\n # concat all the lists and sort by expectancy\\n ret_list = top_bias_objs + top_rel_objs + top_sim_objs + cluster_objs\\n \\n # if the liked should be removed, remove the liked objects from promising\\n if recommender_model.remove_predicted_from_recommendations:\\n\\n # get ids of the liked objects for dn subject\\n pred_obj_ids = RelationshipInstance\\\\\\n .filter_predicted(recommender_model=recommender_model)\\\\\\n .filter(subject_object1=dn_subject)\\\\\\n .values_list('subject_object2__pk', flat=True)\\n \\n pred_obj_ids = set(pred_obj_ids)\\n \\n # remove the liked objects from ret_list\\n ret_list = filter(lambda pair: not(pair[0].pk in pred_obj_ids), ret_list)\\n \\n ret_list.sort(key=lambda pair: pair[1], reverse=True) \\n \\n # remove the duplicates and take only some of the first\\n return list(set([obj for obj, x in ret_list]))[:int(PROMISING_RATE*min_count)]\\n\\n def _get_promising_objects_clusters(self, dn_subject, min_count, recommender_model):\\n \\\"\\\"\\\"Get promising objects from predicted_relationship + cluster membership\\n\\n @return: list of pairs (object, expectancy)\\n \\\"\\\"\\\"\\n\\n # the definition of the predicted relationship\\n d = PredictedRelationshipDefinition.objects.get(recommender=recommender_model)\\n \\n # taking similar subjects from clusters\\n #\\n sim_subj_objs = []\\n\\n # go through the subject's memberships\\n for cm1 in dn_subject.clustermember_set.order_by('-confidence')[:min_count/(2*self.DIVISOR)]:\\n\\n # take the most similar subjects\\n for cm2 in cm1.cluster.clustermember_set.exclude(id=cm1.id).order_by('-confidence')[:min_count/(2*self.DIVISOR)]:\\n \\n # take objects connected with them\\n for rel in cm2.member.relationshipinstance_relationships1.filter(definition=d)[:min_count/self.DIVISOR]:\\n \\n # use a approximation for overall expectancy\\n sim_subj_objs.append((rel.subject_object2, \\n (cm1.confidence * cm2.confidence)/2 + UNCERTAIN_PREDICTION_VALUE))\\n \\n # taking similar objects from clusters\\n #\\n \\n qs_memberships = ClusterMember.objects.filter(\\n cluster__cluster_set__recommender=recommender_model,\\n cluster__clustermember__member__relationshipinstance_relationships2__definition=d,\\n cluster__clustermember__member__relationshipinstance_relationships2__subject_object1=dn_subject)\\\\\\n .order_by('-confidence', '-cluster__clustermember__confidence')\\n\\n sim_obj_objs = [(cm.member, cm.confidence/2 + UNCERTAIN_PREDICTION_VALUE) \\\\\\n for cm in qs_memberships[:min_count]]\\n \\n return sim_subj_objs + sim_obj_objs\\n \\n def _get_promising_objects_rules_relationships(self, dn_subject, min_count, recommender_model):\\n \\\"\\\"\\\"Get promising objects from subject-object rules, relationships,\\n explicit rules\\n \\n @return: list of pairs (object, expectancy) \\n \\\"\\\"\\\"\\n \\n # get what is available in instances\\n\\n # the relationship type we're interested in\\n rel_type = RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT \\\\\\n if recommender_model.are_subjects_objects else \\\\\\n RELATIONSHIP_TYPE_SUBJECT_OBJECT \\n \\n # relationship instances by weight, some rules also may be chosen\\n qs_rel = RelationshipInstance.objects.filter(\\n definition__recommender=recommender_model,\\n definition__rulerelationshipdefinition__relationship_type=rel_type,\\n subject_object1=dn_subject,\\n definition__rulerelationshipdefinition__is_positive=True)\\\\\\n .distinct()\\\\\\n .order_by('-definition__rulerelationshipdefinition__weight')\\n\\n # rule instances by confidence\\n qs_rule = RuleInstance.objects.filter(\\n definition__recommender=recommender_model,\\n definition__rulerelationshipdefinition__relationship_type=rel_type,\\n subject_object1=dn_subject,\\n definition__rulerelationshipdefinition__is_positive=True)\\\\\\n .distinct()\\\\\\n .order_by('-confidence') \\n\\n # explicit feedback \\n qs_exp_rel = ExplicitRuleInstance.objects.filter( \\n definition__recommender=recommender_model,\\n subject_object1=dn_subject,\\n expectancy__gt=UNCERTAIN_PREDICTION_VALUE)\\\\\\n .distinct()\\\\\\n .order_by('-expectancy')\\n\\n return [(rel.subject_object2, rel.get_expectancy()) for rel in qs_rel[:min_count]] + \\\\\\n [(rel.subject_object2, rel.get_expectancy()) for rel in qs_rule[:min_count]] + \\\\\\n [(rel.subject_object2, rel.expectancy) for rel in qs_exp_rel[:min_count]] \\n\\n \\n def _get_promising_objects_similarities(self, dn_subject, min_count, recommender_model):\\n \\\"\\\"\\\"Get promising objects from \\n predicted_relationship + aggregated similarities\\n \\n @return: list of pairs (object, expectancy)\\n \\\"\\\"\\\"\\n \\n # the definition of the predicted relationship\\n d = PredictedRelationshipDefinition.objects.get(recommender=recommender_model)\\n \\n # try finding the similar entities to the ones liked by dn_subject\\n # content-based \\n \\n # get similarities starting the traverse with the similarity.\\n cont_qs_sim1 = AggregatedRelationshipInstance.objects\\\\\\n .filter(recommender=recommender_model, expectancy__gt=UNCERTAIN_PREDICTION_VALUE)\\\\\\n .filter(\\n # traverse from the other object in similarity (subject_object1) through\\n # the relationship instance, its subject (subject_object1) must be so1\\n subject_object1__relationshipinstance_relationships2__subject_object1=dn_subject,\\n \\n # the relationship instance definition must be the predicted relationship def\\n subject_object1__relationshipinstance_relationships2__definition=d)\\\\\\n .distinct()\\\\\\n .order_by('-expectancy')\\n\\n cont_objs_sim1 = [(sim.subject_object2, sim.expectancy) for sim in cont_qs_sim1[:min_count]]\\n\\n cont_qs_sim2 = AggregatedRelationshipInstance.objects\\\\\\n .filter(recommender=recommender_model, expectancy__gt=UNCERTAIN_PREDICTION_VALUE)\\\\\\n .filter(\\n # traverse from the other through relationship to so1\\n subject_object2__relationshipinstance_relationships2__subject_object1=dn_subject,\\n \\n # the definition again must be the predicted\\n subject_object2__relationshipinstance_relationships2__definition=d)\\\\\\n .distinct()\\\\\\n .order_by('-expectancy')\\n \\n cont_objs_sim2 = [(sim.subject_object1, sim.expectancy) for sim in cont_qs_sim2[:min_count]]\\n \\n # try finding the similar entity (user) to entity that liked so2\\n # cf\\n \\n # when subject is in the second position in similarity\\n cf_qs_sim1 = AggregatedRelationshipInstance.objects\\\\\\n .filter(recommender=recommender_model, expectancy__gt=UNCERTAIN_PREDICTION_VALUE)\\\\\\n .filter(\\n # take dn_subject as stable - in the subject_object2 position of the similarity\\n subject_object2=dn_subject,\\n \\n # the relationship instance definition must be the predicted relationship def\\n subject_object1__relationshipinstance_relationships1__definition=d)\\\\\\n .distinct()\\\\\\n .order_by('-expectancy')\\n\\n # get the object behind the predicted_relationship \\n cf_objs_sim1 = [(pref.subject_object2, sim.expectancy) \\\\\\n for sim in cf_qs_sim1[:min_count/self.DIVISOR] \\\\\\n for pref in sim.subject_object1.relationshipinstance_relationships1.filter(definition=d)[:min_count/self.DIVISOR]]\\n\\n # when subject is in the first position in similarity \\n cf_qs_sim2 = AggregatedRelationshipInstance.objects\\\\\\n .filter(recommender=recommender_model, expectancy__gt=UNCERTAIN_PREDICTION_VALUE)\\\\\\n .filter(\\n # take dn_subject as stable again, now in the subject_object1 position of the similarity\\n subject_object1=dn_subject, \\n \\n # the definition again must be the predicted\\n subject_object2__relationshipinstance_relationships1__definition=d)\\\\\\n .distinct()\\\\\\n .order_by('-expectancy')\\n \\n # get the object behind the predicted_relationship \\n cf_objs_sim2 = [(pref.subject_object2, sim.expectancy) \\\\\\n for sim in cf_qs_sim2[:min_count/self.DIVISOR] \\\\\\n for pref in sim.subject_object2.relationshipinstance_relationships1.filter(definition=d)[:min_count/self.DIVISOR]]\\n \\n return cont_objs_sim1 + cont_objs_sim2 + cf_objs_sim1 + cf_objs_sim2\\n \\n \\n\\n \\n \\n \\n \\n\"\n },\n {\n \"alpha_fraction\": 0.407888799905777,\n \"alphanum_fraction\": 0.44402703642845154,\n \"avg_line_length\": 35.519229888916016,\n \"blob_id\": \"77be0dd2ab0e6d6b28632408f382e8cc098d9a7b\",\n \"content_id\": \"9799347ab014baef3e09e67ea2786a89cb0ee061\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 13310,\n \"license_type\": \"no_license\",\n \"max_line_length\": 243,\n \"num_lines\": 364,\n \"path\": \"/code/adapter/unresyst/tests/test_combine_compile.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Tests for combinator, compilator.\\\"\\\"\\\"\\nfrom nose.tools import eq_, assert_almost_equal\\n\\nfrom unresyst.models.common import SubjectObject\\nfrom unresyst.combinator.base import BaseCombinator\\nfrom unresyst.compilator.base import BaseCompilator\\n\\nfrom test_base import TestBuildAverage\\n\\nMIN_COUNT = 5\\nPLACES = 4\\n\\nclass TestCombinator(TestBuildAverage):\\n \\\"\\\"\\\"Tests for the base combinator\\\"\\\"\\\"\\n\\n EXPECTED_PROMISING_OBJECTS = {\\n 'Alice': (('Sneakers', 'Design Shoes'), ('Rubber Shoes', 'Sneakers', 'Rubber Shoes', 'Design Shoes'), ('Rubber Shoes', 'RS 130', 'Sneakers'), ('Sneakers', 'Rubber Shoes')),\\n 'Bob': (('Sneakers', 'Design Shoes'), ('Sneakers', 'Sneakers', 'Rubber Shoes'), ('Rubber Shoes', 'RS 130', 'Sneakers'), ('Sneakers', 'Rubber Shoes')),\\n 'Cindy': ((), ('Rubber Shoes', 'RS 130'), (), ('Sneakers', 'Rubber Shoes')),\\n 'Daisy': ((), ('RS 130',), (), ('Sneakers', 'Rubber Shoes')),\\n 'Edgar': ((), (), ('Sneakers',), ('Sneakers', 'Rubber Shoes'), ('RS 130', 'Octane SL')),\\n 'Fionna': ((), (), ('Rubber Shoes',), ('Sneakers', 'Rubber Shoes'), ('RS 130',)),\\n }\\n \\\"\\\"\\\"Dictionary subject: tuple( tuple from clusters, tuple from s-o rules, tuple from similarities, tuple from biases)\\n \\\"\\\"\\\"\\n\\n def dtest_choose_promising_objects(self):\\n \\\"\\\"\\\"Test choosing promising objects for subjects\\\"\\\"\\\"\\n\\n r = self.recommender._get_recommender_model()\\n bc = BaseCombinator()\\n\\n for subj in SubjectObject.objects.filter(recommender=r, entity_type='S'):\\n\\n # choose it\\n promobjs = bc.choose_promising_objects(dn_subject=subj, min_count=MIN_COUNT)\\n \\n # compare it\\n eq_((subj, set(promobjs)), (subj, set([self.universal_entities[oname] for tup in self.EXPECTED_PROMISING_OBJECTS[subj.name] for oname in tup])))\\n \\nclass TestCompilator(TestBuildAverage):\\n \\\"\\\"\\\"Tests for the base compilator\\\"\\\"\\\"\\n EXPECTED_COMBINATION_ELEMENTS = {\\n \\n ('Alice', 'Sneakers'): (\\n [(0.566667, 'User Alice likes many shoe pairs.'), (0.766667, 'Shoe pair Sneakers is popular')], \\n [(0.550000, 'User Alice is from the same city as the manufacturer of Sneakers.')], \\n [],\\n [(0.612500, 'Similarity: Reason 1: Users Alice and Bob live in the same city. Reason 2: Users Alice and Bob are about the same age. And: User Bob likes shoes Sneakers.')],\\n [],\\n [],\\n ),\\n \\n ('Alice', 'Rubber Shoes'): (\\n [(0.566667, 'User Alice likes many shoe pairs.'), (0.633333, 'Shoe pair Rubber Shoes is popular')], \\n [(0.700000, 'User Alice has viewed Rubber Shoes.'), (0.550000, 'User Alice is from the same city as the manufacturer of Rubber Shoes.')], \\n [(0.625000, 'User Alice likes shoes Sneakers. And similarity: Reason 1: The shoe pairs Sneakers and Rubber Shoes share some keywords. Reason 2: Shoes Sneakers and Rubber Shoes were made by the same manufacturer.')],\\n [],\\n [],\\n [],\\n ), \\n \\n ('Alice', 'RS 130'): (\\n [(0.566667, 'User Alice likes many shoe pairs.'), (0.633333, 'Shoe pair RS 130 is popular')], \\n [(0.075000, \\\"Alice is from south, so RS 130 can't be recommended to him/her.\\\")],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Alice', 'Design Shoes'): (\\n [(0.566667, 'User Alice likes many shoe pairs.')], \\n [],\\n [(0.625000, 'User Alice likes shoes Sneakers. And similarity: Reason 1: Sneakers belong to the Casual category. Design Shoes belong to the Casual category. Reason 2: The shoe pairs Sneakers and Design Shoes share some keywords.')],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Alice', 'Octane SL'): (\\n [(0.566667, 'User Alice likes many shoe pairs.')], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Bob', 'Sneakers'): (\\n [(0.566667, 'User Bob likes many shoe pairs.'), (0.766667, 'Shoe pair Sneakers is popular')], \\n [(0.700000, 'User Bob has viewed Sneakers.'), (0.550000, 'User Bob is from the same city as the manufacturer of Sneakers.')], \\n [],\\n [(0.612500, 'Similarity: Reason 1: Users Alice and Bob live in the same city. Reason 2: Users Alice and Bob are about the same age. And: User Alice likes shoes Sneakers.')],\\n [],\\n [],\\n ), \\n \\n ('Bob', 'Rubber Shoes'): (\\n [(0.566667, 'User Bob likes many shoe pairs.'), (0.633333, 'Shoe pair Rubber Shoes is popular')], \\n [(0.550000, 'User Bob is from the same city as the manufacturer of Rubber Shoes.')], \\n [(0.625000, 'User Bob likes shoes Sneakers. And similarity: Reason 1: The shoe pairs Sneakers and Rubber Shoes share some keywords. Reason 2: Shoes Sneakers and Rubber Shoes were made by the same manufacturer.')], \\n [],\\n [], \\n [],\\n ),\\n \\n ('Bob', 'RS 130'): (\\n [(0.566667, 'User Bob likes many shoe pairs.'), (0.633333, 'Shoe pair RS 130 is popular')], \\n [(0.075000, \\\"Bob is from south, so RS 130 can't be recommended to him/her.\\\")], \\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Bob', 'Design Shoes'): (\\n [(0.566667, 'User Bob likes many shoe pairs.')], \\n [],\\n [(0.625000, 'User Bob likes shoes Sneakers. And similarity: Reason 1: Sneakers belong to the Casual category. Design Shoes belong to the Casual category. Reason 2: The shoe pairs Sneakers and Design Shoes share some keywords.')],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Bob', 'Octane SL'): (\\n [(0.566667, 'User Bob likes many shoe pairs.')], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Cindy', 'Sneakers'): (\\n [(0.766667, 'Shoe pair Sneakers is popular')], \\n [],\\n [],\\n [],\\n [],\\n [(0.550000, 'Similarity: Cindy has searched for the word Comfortable. Bob has searched for the word Comfortable. And: User Bob likes shoes Sneakers.')],\\n ),\\n\\n \\n ('Cindy', 'Rubber Shoes'): (\\n [(0.633333, 'Shoe pair Rubber Shoes is popular')], \\n [(0.700000, 'User Cindy has viewed Rubber Shoes.')],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Cindy', 'RS 130'): (\\n [(0.633333, 'Shoe pair RS 130 is popular')], \\n [(0.550000, 'User Cindy is from the same city as the manufacturer of RS 130.')],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Cindy', 'Design Shoes'): (\\n [], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Cindy', 'Octane SL'): (\\n [], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Daisy', 'Sneakers'): (\\n [(0.766667, 'Shoe pair Sneakers is popular')], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Daisy', 'Rubber Shoes'): (\\n [(0.633333, 'Shoe pair Rubber Shoes is popular')], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Daisy', 'RS 130'): (\\n [(0.633333, 'Shoe pair RS 130 is popular')], \\n [(0.550000, 'User Daisy is from the same city as the manufacturer of RS 130.')],\\n [],\\n [],\\n [],\\n [],\\n ), \\n \\n ('Daisy', 'Design Shoes'): (\\n [], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Daisy', 'Octane SL'): (\\n [], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Edgar', 'Sneakers'): (\\n [(0.633333, 'User Edgar likes many shoe pairs.'), (0.766667, 'Shoe pair Sneakers is popular')],\\n [],\\n [(0.625000, 'User Edgar likes shoes Rubber Shoes. And similarity: Reason 1: The shoe pairs Sneakers and Rubber Shoes share some keywords. Reason 2: Shoes Sneakers and Rubber Shoes were made by the same manufacturer.')],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Edgar', 'Rubber Shoes'): (\\n [(0.633333, 'User Edgar likes many shoe pairs.'), (0.633333, 'Shoe pair Rubber Shoes is popular')], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Edgar', 'RS 130'): (\\n [(0.633333, 'User Edgar likes many shoe pairs.'), (0.633333, 'Shoe pair RS 130 is popular')], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Edgar', 'Design Shoes'): (\\n [(0.633333, 'User Edgar likes many shoe pairs.')], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Edgar', 'Octane SL'): (\\n [(0.633333, 'User Edgar likes many shoe pairs.')], \\n [],\\n [],\\n [],\\n [(0.650000, 'User Edgar likes shoes RS 130. And similarity: RS 130 belong to the For Sports category. Octane SL belong to the For Sports category.')],\\n [],\\n ),\\n \\n ('Fionna', 'Sneakers'): (\\n [(0.766667, 'Shoe pair Sneakers is popular')], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Fionna', 'Rubber Shoes'): (\\n [(0.633333, 'Shoe pair Rubber Shoes is popular')], \\n [],\\n [],\\n [(0.600000, 'Similarity: Users Edgar and Fionna are about the same age. And: User Edgar likes shoes Rubber Shoes.')],\\n [],\\n [],\\n ), \\n \\n ('Fionna', 'RS 130'): (\\n [(0.633333, 'Shoe pair RS 130 is popular')], \\n [],\\n [],\\n [(0.600000, 'Similarity: Users Edgar and Fionna are about the same age. And: User Edgar likes shoes RS 130.')],\\n [],\\n [],\\n ),\\n \\n ('Fionna', 'Design Shoes'): (\\n [], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ),\\n \\n ('Fionna', 'Octane SL'): (\\n [], \\n [],\\n [],\\n [],\\n [],\\n [],\\n ), \\n }\\n \\\"\\\"\\\"Dictionary s-o pair: tuple of elements \\n (the pair: expectancy, description): \\n - bias, \\n - s-o relationships\\n - predicted + object similarity\\n - predicted + subject similarity\\n - predicted + object cluster\\n - predicted + subject cluster\\n \\\"\\\"\\\"\\n \\n\\n \\n def dtest_get_pair_combination_elements(self):\\n \\\"\\\"\\\"Test getting all we know for the given s-o pairs\\\"\\\"\\\"\\n\\n r = self.recommender._get_recommender_model()\\n \\n bc = BaseCompilator()\\n\\n # for all subject - object pairs\\n for subj in SubjectObject.objects.filter(recommender=r, entity_type='S'):\\n\\n for obj in SubjectObject.objects.filter(recommender=r, entity_type='O'):\\n \\n dn_subject = self.universal_entities[subj.name]\\n dn_object = self.universal_entities[obj.name]\\n \\n # call the tested function \\n els = bc.get_pair_combination_elements(dn_subject, dn_object)\\n \\n expected_data = self.EXPECTED_COMBINATION_ELEMENTS[(subj.name, obj.name)]\\n \\n # flatten the expected data\\n flat_data = [pair for listt in expected_data for pair in listt]\\n\\n # assert the length is right \\n eq_(len(flat_data), len(els), \\\"Expected: %d, Obtained %d. For pair %s, %s Obtained data: %s\\\" % \\\\\\n (len(flat_data), len(els), subj, obj, els))\\n \\n for el in els:\\n \\n # try finding the description in the expected data\\n found = filter(lambda pair: pair[1]==el.get_description(), flat_data)\\n \\n eq_(len(found), 1, \\\"The description %s wasn't found for the pair %s, %s\\\" % (el.get_description(), subj, obj))\\n found = found[0]\\n \\n assert_almost_equal(found[0], el.get_expectancy(), PLACES,\\n \\\"The expectancy is wrong for pair %s, %s. Expected %f, Got %f\\\" % (subj, obj, found[0], el.get_expectancy()))\\n \\n\"\n },\n {\n \"alpha_fraction\": 0.5384615659713745,\n \"alphanum_fraction\": 0.5641025900840759,\n \"avg_line_length\": 18.5,\n \"blob_id\": \"5e4066ae41edc0aa7c1a7f8ec796d070d77f0c22\",\n \"content_id\": \"39d9a160c41fc8946eb2a6a18ea979a1b2062a6d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 39,\n \"license_type\": \"no_license\",\n \"max_line_length\": 27,\n \"num_lines\": 2,\n \"path\": \"/code/adapter/gr.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#/bin/bash\\ngrep -r --include=*.py $1 *\\n\"\n },\n {\n \"alpha_fraction\": 0.6993424892425537,\n \"alphanum_fraction\": 0.7065152525901794,\n \"avg_line_length\": 29.418182373046875,\n \"blob_id\": \"a060d34a1df70ccb30aba92abd8b1b6bdac1f4d0\",\n \"content_id\": \"c2c0fa1dbb3cd32ee0d63d48f1879eeaaccdf503\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 1673,\n \"license_type\": \"no_license\",\n \"max_line_length\": 287,\n \"num_lines\": 55,\n \"path\": \"/code/mahout/mahoutrec/unresystpredict.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/bin/sh\\n\\n# params:\\n# lastfm\\n# flixster\\n# travel\\n\\n\\nLASTFM=false\\nFLIXSTER=false\\nTRAVEL=false\\n\\nFLIXSTERBELOW=false\\n\\nfor param in $*;\\ndo \\n case $param in\\n 'lastfm')\\n LASTFM=true\\n ;;\\n 'flixster')\\n FLIXSTER=true\\n ;;\\n 'travel')\\n TRAVEL=true\\n ;; \\n 'flixsterbelow')\\n FLIXSTERBELOW=true\\n ;; \\n esac\\ndone\\n\\nif [ $LASTFM = true ]\\nthen\\n mvn -e exec:java -Dexec.mainClass=\\\"com.unresyst.UnresystPredict\\\" -Dexec.args=\\\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_test.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_predictions.csv\\\" \\nfi\\n\\nif [ $FLIXSTER = true ]\\nthen\\n mvn -e exec:java -Dexec.mainClass=\\\"com.unresyst.UnresystPredict\\\" -Dexec.args=\\\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_test.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_predictions.csv\\\" \\nfi\\n\\nif [ $FLIXSTERBELOW = true ]\\nthen\\n mvn -e exec:java -Dexec.mainClass=\\\"com.unresyst.UnresystPredict\\\" -Dexec.args=\\\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_test.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_predictions.csv\\\" \\nfi\\n\\n\\nif [ $TRAVEL = true ]\\nthen\\n mvn -e exec:java -Dexec.mainClass=\\\"com.unresyst.UnresystPredict\\\" -Dexec.args=\\\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_train.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_test.csv /home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_predictions.csv\\\" \\nfi\\n\\n\\necho \\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.7361376881599426,\n \"alphanum_fraction\": 0.7361376881599426,\n \"avg_line_length\": 29.705883026123047,\n \"blob_id\": \"b1c76c010eee71b762a42787202fa779a27e40b0\",\n \"content_id\": \"69152065ee0d7f36460fb4c1a9775c6e64d39541\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 523,\n \"license_type\": \"no_license\",\n \"max_line_length\": 63,\n \"num_lines\": 17,\n \"path\": \"/code/adapter/flixster/evaluation.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The evaluators for the last.fm recommenders\\\"\\\"\\\"\\n\\nfrom unresyst.recommender.evaluation import BaseEvaluator\\nfrom models import MovieEvaluationPair\\nfrom unresyst.recommender.metrics import rmse, precision_recall\\n\\nclass MovieRecommenderEvaluator(BaseEvaluator):\\n \\\"\\\"\\\"The evaluator of the artist recommender\\\"\\\"\\\"\\n \\n EvaluationPairModel = MovieEvaluationPair\\n \\\"\\\"\\\"The model - pairs\\\"\\\"\\\"\\n \\n prediction_metric = rmse\\n \\\"\\\"\\\"The metric\\\"\\\"\\\"\\n \\n recommendation_metric = precision_recall\\n \\\"\\\"\\\"The other metric\\\"\\\"\\\"\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.7560126781463623,\n \"alphanum_fraction\": 0.7569620013237,\n \"avg_line_length\": 76.0243911743164,\n \"blob_id\": \"e4e4234d3bd3cfeffc22e3d5b6a28a141d698c91\",\n \"content_id\": \"032105fd7530c2e4ce9bb1f334b90dd1659d3b43\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Text\",\n \"length_bytes\": 3160,\n \"license_type\": \"no_license\",\n \"max_line_length\": 345,\n \"num_lines\": 41,\n \"path\": \"/UnresystCD/readme.txt\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"Welcome to the Unresyst CD!\\n---------------------------\\n\\nCurrent versions of all contained files can be found on http://code.google.com/p/unresyst/\\n\\nThe directory structure of the CD is the following:\\n\\n - docs: documentation of the unresyst prototype implementation.\\n - diagrams: various diagrams used in the thesis, created in dia, a program for drawing structured diagrams http://live.gnome.org/Dia\\n - evaluation: spreadsheets containing the results of unresyst evaluation, including charts\\n - thesis: the thesis in the linked pdf format and latex source codes\\n - mindmaps: mind maps used for presenting ideas about Unresyst. Created in XMind version: 3.2.0 http://www.xmind.net/\\n - pics: pictures used in the thesis\\n - presentations: presentations on the thesis from the User-Web seminar.\\n - datamodels: data model visualization for each dataset application and for the whole unresyst.\\n \\n \\n \\n - code: contains all source code files belonging to the thesis.\\n\\n - adapter: adapter is a django project home directory (see http://www.djangoproject.com/), there are the following applications in the project. Each application has its own directory. The documentation for the project can be found in the docs/epydoc folder.\\n\\n - unresyst: the universal recommender application containing the implementation of the universal recommedner\\n - lastfm: an application containing data models and a Unresyst recommender for the last.fm dataset.\\n - flixster: application for the flixster dataset, the same structure as lastfm\\n - travel: application for the travel agency dataset, the same structure as lastfm\\n - demo: a demo application used in the tests\\n - csv: a directory for temporary csv files used for the external mahout recommender\\n - logs: a directory of text files where the evaluation of the recommendations is logged\\n - evaluate.txt: a list of commands for running an external recommender for each application\\n - example.py: an example of an Unresyst recommender, presented on the User-Web seminar\\n - flixstereval.sh, lastfmeval.sh, traveleval.sh: scripts for running an evaluation on example datasets, see the script code for help\\n - dump.sh, load.sh: scripts for opperating with the MySQL database\\n - setup.sh: a script for loading data from dataset to the corresponing application models, the script just uses python scripts save_data.py, defined in each of the applications. \\n - unresyst/tests/: tests for the implementation, the tests can be run using the command 'python ./manage.py test unresyst\\\"\\n\\n - dataset_scripts: shell scripts for reducing the datasets as described in the thesis.\\n\\n - mahout: contains an implementation of the Slope One recommender that was used for a comparison with unresyst. The recommender is implemented in Java using the Mahout recommender framework. See http://mahout.apache.org/ and http://code.google.com/p/unresyst/wiki/CreateMahoutRecommender for download and instalation of the Mahout framework.\\n \\n - unresystrecommend.sh, unresystpredict.sh: the scripts used for running the recommender on our data sets.\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5913978219032288,\n \"alphanum_fraction\": 0.5913978219032288,\n \"avg_line_length\": 45,\n \"blob_id\": \"a7006a8fa9180fdf69189b409144c4558add2b25\",\n \"content_id\": \"fbf1ff2571e73a191b467e04bafc59305f331b18\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 93,\n \"license_type\": \"no_license\",\n \"max_line_length\": 45,\n \"num_lines\": 2,\n \"path\": \"/code/adapter/rmpyc.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"find . -type f -name \\\\*.pyc -print | xargs rm\\nfind . -type f -name \\\\*.py~ -print | xargs rm\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5462448000907898,\n \"alphanum_fraction\": 0.5462448000907898,\n \"avg_line_length\": 31.022472381591797,\n \"blob_id\": \"b606e14182912067aa748b05ac5b1b12f75be5d5\",\n \"content_id\": \"644ee4d31f9a0251faad680e4db4b9700bfa0bf1\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2876,\n \"license_type\": \"no_license\",\n \"max_line_length\": 81,\n \"num_lines\": 89,\n \"path\": \"/code/adapter/unresyst/abstractor/abstractor.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The main class of the abstractor package - Abstractor\\\"\\\"\\\"\\n\\nfrom base import BaseAbstractor\\nfrom unresyst.models.common import SubjectObject, Recommender as RecommenderModel\\nfrom unresyst.constants import *\\n\\nclass BasicAbstractor(BaseAbstractor):\\n \\\"\\\"\\\"The basic implementation of the Abstractor class\\\"\\\"\\\"\\n\\n\\n def create_subjectobjects(self, recommender_model, subjects, objects):\\n \\\"\\\"\\\"See the base class for documentation.\\\"\\\"\\\"\\n \\n # if subjects are the same as objects, use the \\\"so\\\" entity type \\n so = ENTITY_TYPE_SUBJECTOBJECT if subjects == objects else \\\"\\\"\\n\\n # create them again. subjects:\\n for subj in subjects.iterator():\\n \\n # create the subject\\n subob = SubjectObject(\\n id_in_specific=subj.pk,\\n name=subj.__unicode__()[:MAX_LENGTH_NAME],\\n entity_type=so or ENTITY_TYPE_SUBJECT,\\n recommender=recommender_model\\n )\\n \\n # save it\\n subob.save()\\n \\n print \\\" %d subjects created\\\" % subjects.count()\\n \\n # for recommenders where subjects==objects, that's it\\n if so:\\n return\\n \\n # create objects: \\n for obj in objects.iterator():\\n \\n # create the object\\n subob = SubjectObject(\\n id_in_specific=obj.pk,\\n name=obj.__unicode__()[:MAX_LENGTH_NAME],\\n entity_type=ENTITY_TYPE_OBJECT,\\n recommender=recommender_model\\n )\\n \\n # save it\\n subob.save()\\n \\n print \\\" %d objects created\\\" % objects.count()\\n\\n\\n \\n def create_predicted_relationship_instances(self, predicted_relationship):\\n \\\"\\\"\\\"See the base class for documentation.\\\"\\\"\\\" \\n \\n # evaluate the relationship for all possible subjectobjects\\n predicted_relationship.evaluate()\\n \\n \\n def create_relationship_instances(self, relationships):\\n \\\"\\\"\\\"See the base class for documentation.\\\"\\\"\\\"\\n\\n # evaluate all relationships\\n for rel in relationships:\\n rel.evaluate()\\n\\n \\n def create_rule_instances(self, rules):\\n \\\"\\\"\\\"See the base class for documentation.\\\"\\\"\\\"\\n \\n # eveluate all rules\\n for rule in rules:\\n rule.evaluate()\\n\\n def create_clusters(self, cluster_sets):\\n \\\"\\\"\\\"See the base class for documentation.\\\"\\\"\\\"\\n \\n # evaluate all cluster sets\\n for cluster_set in cluster_sets:\\n cluster_set.evaluate() \\n \\n def create_biases(self, biases):\\n \\\"\\\"\\\"See the base class for documentation.\\\"\\\"\\\" \\n \\n # evaluate the biases\\n for bias in biases:\\n bias.evaluate()\\n \\n \\n\"\n },\n {\n \"alpha_fraction\": 0.5340961217880249,\n \"alphanum_fraction\": 0.5447750091552734,\n \"avg_line_length\": 34.819671630859375,\n \"blob_id\": \"8491b9f2e8b52afe3779e5f92e6d4edc5a28c3d5\",\n \"content_id\": \"cc7eb7e3479bbc007ca79cd3972ef4f5c0018a4e\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 6555,\n \"license_type\": \"no_license\",\n \"max_line_length\": 82,\n \"num_lines\": 183,\n \"path\": \"/code/adapter/recommender_example.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"An example of an Unresyst recommender for the last.fm domain.\\nThe addaptation rules include:\\n - subject-object preference \\n - subject-subject similarity\\n - object-object similarity\\n - object bias\\n\\nRecommenders used during the evaluation tests can be found in:\\n - lastfm/recommender.py\\n - travel/recommender.py\\n - flixster/recommender.py\\n\\\"\\\"\\\"\\n\\nfrom unresyst import *\\nfrom models import *\\n\\nAGE_DIFFERENCE = 38 - 17\\n\\\"\\\"\\\"The age difference between the oldest and the yongest user\\\"\\\"\\\"\\n\\nMAX_PLAY_COUNT = 542\\n\\\"\\\"\\\"The maximum play count for an artist in the period\\\"\\\"\\\"\\n\\nN_MIN_PLAY_COUNT = 100\\n\\\"\\\"\\\"The minimum play count for an artist to apply the bias\\\"\\\"\\\"\\n\\nPERIOD_START_DATE = datetime.date(2010, 9, 1)\\nPERIOD_END_DATE = datetime.date(2010, 12, 31)\\n\\n\\nclass ArtistRecommender(Recommender):\\n \\\"\\\"\\\"A recommender recommending artists (musicians) that \\n the user can like.\\n \\\"\\\"\\\" \\n \\n name = \\\" Artist Recommender\\\"\\n \\\"\\\"\\\"The name\\\"\\\"\\\" \\n \\n subjects = User.objects\\n \\\"\\\"\\\"The subjects to who the recommender will recommend.\\\"\\\"\\\"\\n \\n objects = Artist.objects\\n \\\"\\\"\\\"The objects that will be recommended.\\\"\\\"\\\"\\n \\n predicted_relationship = PredictedRelationship( \\n name=\\\"User listens to artist's tracks.\\\",\\n \\n # gives true for user, artist pairs where the user have listened\\n # to the artist\\n condition=lambda user, artist: \\\\\\n user.scrobble_set.filter(track__artist=artist).exists(), \\n \\n description=\\\"\\\"\\\"User %(subject)s listens to the %(object)s's tracks.\\\"\\\"\\\"\\n )\\n \\\"\\\"\\\"The relationship that will be predicted\\\"\\\"\\\"\\n \\n # the class contains definitions for business rules\\n rules = (\\n \\n # don't recommend artists with male-specific tags to females\\n SubjectObjectRule(\\n name=\\\"Don't recommend male music to female users.\\\",\\n\\n # the user is a female and the artist was tagged by\\n # a male-specific tag\\n condition=lambda user, artist: user.gender == 'f' and \\\\\\n artist.artisttag_set.filter(tag__gender_specific='m').exists(),\\n \\n # it's a negative rule\\n is_positive=False,\\n \\n weight=0.5,\\n \\n # the more male-specific tags the artist has, the higher is \\n # the rule confidence. Normalized by the artist tag count\\n confidence=lambda user, artist: float(\\n artist.artisttag_set.filter(tag__gender_specific='m').count())/ \\\\\\n artist.artisttag_set.count(),\\n \\n description=\\\"Artist %(object)s isn't recommended to %(subject)s, \\\" +\\n \\\"because the artist is considered male-specific.\\\"\\n ),\\n \\n \\n # users of similar age are similar\\n SubjectSimilarityRule(\\n name=\\\"Users with similar age.\\\",\\n \\n # both users have given their age and the difference \\n # is lower than five\\n condition=lambda user1, user2: \\n user1.age and user2.age and abs(user1.age - user2.age) <= 5,\\n \\n is_positive=True, \\n \\n weight=0.5,\\n \\n # a magic linear confidence function\\n confidence=lambda user1, user2: \\n 1 - float(abs(user1.age - user2.age))/AGE_DIFFERENCE,\\n \\n description=\\\"Users %(subject1)s and %(subject2)s are about \\\" + \\n \\\"the same age.\\\"\\n ), \\n \\n # artists sharing some tags are similar\\n ObjectSimilarityRule(\\n name=\\\"Artists sharing some tags.\\\",\\n\\n # both artists have some tags and they share at least one tag\\n condition=lambda artist1, artist2: \\\\\\n artist1.artisttag_set.exists() and \\\\\\n artist2.artisttag_set.exists() and \\\\\\n artist1.artisttag_set.filter(\\n tag__id__in=artist2.artisttag_set.values_list('tag__id')\\n ).exists(),\\n \\n # it's a positive rule\\n is_positive=True,\\n \\n weight=0.5,\\n \\n # The more tags the artists have in common, the higher is \\n # the similarity confidence\\n confidence=lambda artist1, artist2: \\\\\\n float(artist1.artisttag_set.filter(\\n tag__id__in=artist2.artisttag_set.values_list('tag__id')\\n ).count()) / \\\\\\n min(artist1.artisttag_set.count(), artist2.artisttag_set.count()),\\n \\n description=\\\"Artists %(object1)s and %(object2)s are similar \\\" +\\n \\\"because they share some tags.\\\"\\n ),\\n )\\n \\n relationships = (\\n # if two users are from the same country, they are similar\\n SubjectSimilarityRelationship(\\n name=\\\"Users living in the same country\\\",\\n \\n # both users have given their country and it's the same\\n condition=lambda user1, user2:\\n user1.country and \\\\\\n user2.country and \\\\\\n user1.country == user2.country,\\n \\n # it's relationship positive to similarity\\n is_positive=True, \\n \\n weight=0.5, \\n \\n description=\\\"Users %(subject1)s and %(subject2)s are from \\\" + \\\\\\n \\\" the same country.\\\",\\n ),\\n )\\n \\n biases = (\\n ObjectBias(\\n name=\\\"Artists whose tracks have been listened a lot recently.\\\",\\n \\n description=\\\"%(object)s has been listened much recently.\\\",\\n \\n # take only artists with more than the minimal play count\\n # in the given period\\n condition=lambda artist: \\\\\\n artist.track_set\\\\\\n .filter(scrobble__timestamp__range=\\n (PERIOD_START_DATE, PERIOD_END_DATE))\\\\\\n .aggregate(Count('scrobble')) > N_MIN_PLAY_COUNT,\\n \\n weight=0.5,\\n \\n # it's a positive bias\\n is_positive=True,\\n \\n # the number of scrobbles for the artist divided by the max.\\n confidence=lambda artist: \\\\\\n float(artist.track_set\\\\\\n .filter(scrobble__timestamp__range=\\n (PERIOD_START_DATE, PERIOD_END_DATE))\\\\\\n .annotate(scrobble_count=Count('scrobble'))\\\\\\n .aggregate(Sum('scrobble_count')))/MAX_PLAY_COUNT\\n ),\\n )\\n\"\n },\n {\n \"alpha_fraction\": 0.5987638235092163,\n \"alphanum_fraction\": 0.6003090143203735,\n \"avg_line_length\": 32.84716033935547,\n \"blob_id\": \"b8d5a394c3f4eb03cbd9ec92f382c65b8c472e4a\",\n \"content_id\": \"a196c296da5fc0ad056f21531bcc8a73ef3c3a3d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 7766,\n \"license_type\": \"no_license\",\n \"max_line_length\": 107,\n \"num_lines\": 229,\n \"path\": \"/UnresystCD/code/adapter/unresyst/exceptions.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Errors thrown by the Unresyst application.\\\"\\\"\\\"\\n\\nclass UnresystError(Exception):\\n \\\"\\\"\\\"The base exception for all exceptions.\\n \\n Defines the message property.\\n \\\"\\\"\\\"\\n def _get_message(self): \\n return self._message\\n def _set_message(self, message): \\n self._message = message\\n message = property(_get_message, _set_message)\\n\\nclass CombinatorError(UnresystError):\\n \\\"\\\"\\\"An error in combinator\\\"\\\"\\\"\\n pass\\n \\nclass RecommenderError(UnresystError):\\n \\\"\\\"\\\"The base class for exceptions related to a recommender.\\n\\n @type message: string\\n @ivar message: the reason for the error \\n \\n @type recommender: recommender.Recommender\\n @ivar recommender: the recommender on which the error occured \\n \\\"\\\"\\\" \\n \\n def __init__(self, message, recommender):\\n \\\"\\\"\\\"The constructor.\\\"\\\"\\\" \\n\\n self.message = message\\n \\\"\\\"\\\"The message saying the reason for the error\\\"\\\"\\\" \\n \\n self.recommender = recommender\\n \\\"\\\"\\\"The recommender for which the error is raised\\\"\\\"\\\" \\n \\nclass RecommenderBuildError(RecommenderError):\\n \\\"\\\"\\\"An error during the build of the recommender.\\\"\\\"\\\"\\n pass\\n\\nclass RecommenderNotBuiltError(RecommenderError):\\n \\\"\\\"\\\"Exception meaning that the recommender hasn't been built even though\\n it should have been.\\n \\n @type message: string\\n @ivar message: the reason for the error \\n \\n @type recommender: recommender.Recommender\\n @ivar recommender: the recommender on which the error occured\\n \\\"\\\"\\\" \\n\\n\\n def __str__(self):\\n return (\\\"The recommender needs to be built to perform the action.\\\\n\\\" + \\\\\\n \\\" message: %s\\\\n\\\" + \\\\\\n \\\" recommender: %s\\\\n\\\") \\\\\\n % (self.message, self.recommender.name)\\n \\n\\nclass ConfigurationError(RecommenderError):\\n \\\"\\\"\\\"Exception meaning that something is wrong in the recommender configuration. \\n\\n @type message: string\\n @ivar message: the reason why the parameter is invalid\\n\\n @type recommender: recommender.Recommender\\n @ivar recommender: the recommender on which the error occured\\n \\n @type parameter_name: string\\n @ivar parameter_name: the name of the invalid parameter\\n \\n @type parameter_value: object\\n @ivar parameter_value: the current (invalid) parameter value \\n \\\"\\\"\\\"\\n \\n def __init__(self, message, recommender, parameter_name, parameter_value):\\n \\\"\\\"\\\"The constructor.\\\"\\\"\\\" \\n\\n super(ConfigurationError, self).__init__(message, recommender)\\n \\n self.parameter_name = parameter_name\\n \\\"\\\"\\\"The name of the invalid parameter\\\"\\\"\\\"\\n \\n self.parameter_value = parameter_value\\n \\\"\\\"\\\"The current (invalid) value of the parameter\\\"\\\"\\\" \\n\\n def __str__(self):\\n return (\\\"A configuration parameter is invalid.\\\\n\\\" + \\\\\\n \\\" message: %s\\\\n\\\" + \\\\\\n \\\" recommender: %s\\\\n\\\" + \\\\\\n \\\" parameter name: %s\\\\n\\\" + \\\\\\n \\\" parameter value: %s\\\\n\\\") \\\\\\n % (self.message, self.recommender.name, self.parameter_name, self.parameter_value)\\n\\n\\nclass RuleRelationshipError(RecommenderError):\\n \\\"\\\"\\\"An error in the configuration of rules/relationships.\\n\\n @type message: str\\n @ivar message: additional message\\n\\n @type recommender: recommender.Recommender\\n @ivar recommender: the recommender on which the error occured\\n \\n @type name: str\\n @ivar name: the name of the rule/relationship where the error occured \\n \\\"\\\"\\\"\\n \\n def __init__(self, message, recommender, name):\\n \\\"\\\"\\\"The constructor.\\\"\\\"\\\" \\n\\n super(RuleRelationshipError, self).__init__(message, recommender)\\n \\n self.name = name\\n \\\"\\\"\\\"The rule/relationship name\\\"\\\"\\\"\\n \\n\\nclass DescriptionKeyError(RuleRelationshipError):\\n \\\"\\\"\\\"Exception meaning that there are wrong strings in the rule/relationship\\n description. \\n\\n @type message: str\\n @ivar message: additional message\\n \\n @type recommender: recommender.Recommender\\n @ivar recommender: the recommender on which the error occured\\n \\n @type name: str\\n @ivar name: the name of the rule/relationship where the error occured\\n \\n @type key: str\\n @ivar key: the key that shouldn't have been there\\n \\n @type permitted_keys: str list\\n @ivar permitted_keys: the keys that can be in the description\\n \\\"\\\"\\\"\\n \\n def __init__(self, message, recommender, name, key, permitted_keys):\\n \\\"\\\"\\\"The constructor.\\\"\\\"\\\" \\n\\n super(DescriptionKeyError, self).__init__(message, recommender, name)\\n \\n self.key = key\\n \\\"\\\"\\\"The string key that is wrong\\\"\\\"\\\"\\n \\n self.permitted_keys = permitted_keys\\n \\\"\\\"\\\"The list of permitted keys\\\"\\\"\\\"\\n\\n\\n def __str__(self):\\n return (\\\"There's an invalid format key '%s' in the rule/relationship description.\\\\n\\\" + \\\\\\n \\\" message: %s\\\\n\\\" + \\\\\\n \\\" recommender: %s\\\\n\\\" + \\\\\\n \\\" rule/relationship name: %s\\\\n\\\" + \\\\\\n \\\" the invalid key: %s\\\\n\\\" + \\\\\\n \\\" The permitted keys are: %s\\\") \\\\\\n % (self.key, self.message, self.recommender.name, self.name, self.key, self.permitted_keys) \\n\\nclass InvalidParameterError(RecommenderError):\\n \\\"\\\"\\\"An exception raised when a parameter passed to a function is invalid.\\n\\n @type message: string\\n @ivar message: the reason why the parameter is invalid\\n\\n @type recommender: recommender.Recommender\\n @ivar recommender: the recommender on which the error occured\\n \\n @type parameter_name: string\\n @ivar parameter_name: the name of the invalid parameter\\n \\n @type parameter_value: object\\n @ivar parameter_value: the current (invalid) parameter value\\n \\n \\\"\\\"\\\"\\n def __init__(self, message, recommender, parameter_name, parameter_value):\\n \\\"\\\"\\\"The constructor.\\\"\\\"\\\" \\n\\n super(InvalidParameterError, self).__init__(message, recommender)\\n \\n self.parameter_name = parameter_name\\n \\\"\\\"\\\"The name of the invalid parameter\\\"\\\"\\\"\\n \\n self.parameter_value = parameter_value\\n \\\"\\\"\\\"The current (invalid) value of the parameter\\\"\\\"\\\" \\n \\n def __str__(self):\\n return (\\\"The parameter passed to the function is invalid.\\\\n\\\" + \\\\\\n \\\" message: %s\\\\n\\\" + \\\\\\n \\\" recommender: %s\\\\n\\\" + \\\\\\n \\\" parameter name: %s\\\\n\\\" + \\\\\\n \\\" parameter value: %s\\\") \\\\\\n % (self.message, self.recommender.name, self.parameter_name, self.parameter_value)\\n \\n \\nclass SymmetryError(UnresystError):\\n \\\"\\\"\\\"An error for handling symmetric relationship errors.\\n \\n @type message: string\\n @ivar message: the reason for the error\\n \\n @type object1: object\\n @ivar object1: the first object in the relationship\\n \\n @type object2: object\\n @ivar object2: the second object in the relationship\\n \\\"\\\"\\\"\\n \\n def __init__(self, message, object1, object2):\\n \\\"\\\"\\\"The constructor\\\"\\\"\\\"\\n\\n self.message = message\\n \\\"\\\"\\\"the reason for the error\\\"\\\"\\\"\\n \\n self.object1 = object1\\n \\\"\\\"\\\"the first object in the relationship\\\"\\\"\\\"\\n \\n self.object2 = object2 \\n \\\"\\\"\\\"the second object in the relationship\\\"\\\"\\\"\\n \\n def __str__(self):\\n return (\\\"There's an error in relationship symmetry.\\\\n\\\" + \\\\\\n \\\" message: %s\\\\n\\\" + \\\\\\n \\\" first object in relationship: %s\\\\n\\\" + \\\\\\n \\\" second object in relationship: %s\\\") \\\\\\n % (self.message, self.object1, self.object2) \\n \\nclass EmptyTestSetError(UnresystError):\\n \\\"\\\"\\\"An error indicating that the test set is empty.\\\"\\\"\\\"\\n pass \\n\"\n },\n {\n \"alpha_fraction\": 0.8471337556838989,\n \"alphanum_fraction\": 0.8471337556838989,\n \"avg_line_length\": 38.25,\n \"blob_id\": \"b12b94b6c2d011fd8bd11d69d21be903195767e9\",\n \"content_id\": \"043a96d55420568d1741424d9735fb4331d54093\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 157,\n \"license_type\": \"no_license\",\n \"max_line_length\": 52,\n \"num_lines\": 4,\n \"path\": \"/code/adapter/unresyst/compilator/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The layer making predictions from aggregates\\\"\\\"\\\"\\n\\nfrom get_first_compilator import GetFirstCompilator\\nfrom combining_compilator import CombiningCompilator\\n\"\n },\n {\n \"alpha_fraction\": 0.5678645968437195,\n \"alphanum_fraction\": 0.5681615471839905,\n \"avg_line_length\": 32.83333206176758,\n \"blob_id\": \"e1f2f7507d3a580f771c0a1f8cb7ad15ac713776\",\n \"content_id\": \"e8abee01b6e2d8adcc5c76312f84296b7ed1ff34\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 6734,\n \"license_type\": \"no_license\",\n \"max_line_length\": 88,\n \"num_lines\": 198,\n \"path\": \"/code/adapter/unresyst/recommender/external_recommender.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"A module for recommender making predictions in an outside engine\\\"\\\"\\\"\\nimport csv\\n\\nfrom base import BaseRecommender\\nfrom predictions import RelationshipPrediction\\nfrom unresyst.exceptions import RecommenderError, RecommenderNotBuiltError\\nfrom unresyst.models.algorithm import ExternalPrediction\\nfrom unresyst.models.common import Recommender as RecommenderModel\\nfrom unresyst.constants import *\\n\\nclass ExternalRecommender(BaseRecommender):\\n \\\"\\\"\\\"A class representing an outside-world recommender with an Unresyst \\n interface.\\n \\\"\\\"\\\"\\n \\n PredictionModel = ExternalPrediction\\n \\\"\\\"\\\"The model where predictions are stored\\\"\\\"\\\" \\n \\n # build phase:\\n #\\n \\n @classmethod\\n def build(cls):\\n \\\"\\\"\\\"Raise an error, as thist can't be done here, but outside.\\n \\n @raise RecommenderError: always\\n \\\"\\\"\\\"\\n raise RecommenderError(\\\"The method is invalid for the external\\\" + \\\\\\n \\\" recommender. Call the export_data, build it outside and \\\" + \\\\\\n \\\"import_predictions.\\\", cls)\\n\\n\\n @classmethod\\n def export_data(cls, filename):\\n \\\"\\\"\\\"Export the data from subject and object managers to a csv \\n file of the given name.\\n \\n @type filename: str\\n @param filename: the full path to the file\\n \\n @raise FileNotExists and other file open errors.\\n \\\"\\\"\\\"\\n \\n with open(filename, 'w') as f:\\n \\n rel = cls.explicit_rating_rule \\\\\\n if cls.explicit_rating_rule else cls.predicted_relationship\\n \\n # export the relationships to the given file\\n rel.export(f)\\n \\n @classmethod\\n def import_predictions(cls, filename):\\n \\\"\\\"\\\"Load predictions from the given csv file.\\n \\n Creates the recommender model for the recommender and imports the \\n predictions from the given file. The file has to be in format:\\n ,,\\\\n\\n \\n @type filename: str\\n @param filename: the full path to the file\\n\\n @raise FileNotExists and other file open errors.\\n \\\"\\\"\\\"\\n \\n cls._print('Deleting old predictions...')\\n \\n \\n # if the recommender with the given name exists, delete it,\\n RecommenderModel.objects.filter(class_name=cls.__name__).delete()\\n \\n # create a new recommender and save it\\n recommender_model = RecommenderModel(\\n class_name=cls.__name__,\\n name=cls.name,\\n is_built=False,\\n are_subjects_objects=False\\n ) \\n recommender_model.save() \\n \\n cls._print('Importing new predictions...')\\n \\n # open the csv reader\\n reader = csv.reader(open(filename, \\\"rb\\\"), delimiter=',', quoting=csv.QUOTE_NONE)\\n \\n # parse the csv line by line\\n for subj_id, obj_id, expectancy in reader:\\n \\n # parse the values to the right types\\n subj_id = int(subj_id)\\n obj_id = int(obj_id)\\n expectancy = float(expectancy)\\n \\n # if it's already imported, go to the next pair\\n if cls.PredictionModel.objects.filter(\\n subj_id=subj_id, obj_id=obj_id, recommender=recommender_model):\\n continue\\n \\n # create a prediction\\n cls.PredictionModel.objects.create(\\n subj_id=subj_id,\\n obj_id=obj_id,\\n recommender=recommender_model,\\n expectancy=expectancy\\n )\\n \\n recommender_model.is_built=True\\n recommender_model.save() \\n \\n cls._print('Done.')\\n\\n \\n # recommend phase\\n #\\n @classmethod \\n def predict_relationship(cls, subject, object_, save_to_db=False):\\n \\\"\\\"\\\"See the base class for the explanation\\n\\n If the prediction doesn't exist retruns None.\\n \\\"\\\"\\\"\\n\\n recommender_model = cls._get_recommender_model()\\n \\n # if the recommender isn't built raise an error\\n if not recommender_model or not recommender_model.is_built:\\n raise RecommenderNotBuiltError(\\n message=\\\"Build the recommender prior to performing the \\\" + \\\\\\n \\\"predict_relationship action.\\\",\\n recommender=cls\\n )\\n \\n # get the prediction if it exists\\n qs_prediction_model = cls.PredictionModel.objects.filter(\\n recommender=recommender_model,\\n subj_id=subject.pk,\\n obj_id=object_.pk\\n )\\n \\n # if not return none\\n if not qs_prediction_model:\\n return None\\n \\n # if so extract the expectancy\\n assert qs_prediction_model.count() == 1 \\n expectancy = qs_prediction_model[0].expectancy\\n \\n # create and return the outer-world object\\n prediction = RelationshipPrediction(\\n subject=subject,\\n object_=object_,\\n expectancy=expectancy,\\n is_uncertain=abs(expectancy - UNCERTAIN_PREDICTION_VALUE) < EXP_PRECISION\\n ) \\n return prediction\\n\\n\\n @classmethod\\n def get_recommendations(cls, subject, count=None): \\n \\\"\\\"\\\"For documentation, see the base class\\\"\\\"\\\"\\n \\n recommender_model = cls._get_recommender_model()\\n \\n # if the recommender isn't built raise an error\\n if not recommender_model or not recommender_model.is_built:\\n raise RecommenderNotBuiltError(\\n message=\\\"Build the recommender prior to performing the \\\" + \\\\\\n \\\"get_recommendations action.\\\",\\n recommender=cls\\n )\\n \\n # if count wasn't given take the default one\\n if not count:\\n count = cls.default_recommendation_count\\n \\n # get the prediction objects\\n qs_predictions = cls.PredictionModel.objects.filter(\\n subj_id=subject.pk,\\n recommender=recommender_model)\\\\\\n .order_by('-expectancy')[:count]\\n \\n recommendations = []\\n \\n # go through the obtained predictions\\n for pred_model in qs_predictions:\\n\\n # get the object by the id\\n object_=cls.objects.get(pk=pred_model.obj_id)\\n \\n # create the outer-world object\\n prediction = RelationshipPrediction(\\n subject=subject,\\n object_=object_,\\n expectancy=pred_model.expectancy\\n ) \\n \\n recommendations.append(prediction)\\n\\n return recommendations\\n \\n \\n \\n \\n\\n\"\n },\n {\n \"alpha_fraction\": 0.8885714411735535,\n \"alphanum_fraction\": 0.8885714411735535,\n \"avg_line_length\": 49,\n \"blob_id\": \"70352b51a1d48ab49add548b003010777b50fb5b\",\n \"content_id\": \"273cda81444a14fdebc1b6b8b4d4649311fd0271\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 350,\n \"license_type\": \"no_license\",\n \"max_line_length\": 100,\n \"num_lines\": 7,\n \"path\": \"/UnresystCD/code/adapter/unresyst/combinator/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The combining layer of Unresyst\\\"\\\"\\\"\\n\\nfrom function_combinator import FunctionCombinator\\nfrom combination_element import RelSimilarityCombinationElement, ClusterSimilarityCombinationElement\\nfrom average_combinator import AverageCombinator\\nfrom twisted_average import TwistedAverageCombinator\\nfrom confidence_factor import ConfidenceFactorCombinator\\n\"\n },\n {\n \"alpha_fraction\": 0.6129541993141174,\n \"alphanum_fraction\": 0.6200631856918335,\n \"avg_line_length\": 36.235294342041016,\n \"blob_id\": \"5bb1986eef9f3ddc7f24b5f21b64cb4710226911\",\n \"content_id\": \"b0eed7bd7e60bb70c8a550796fc9139839a1ae65\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1266,\n \"license_type\": \"no_license\",\n \"max_line_length\": 102,\n \"num_lines\": 34,\n \"path\": \"/UnresystCD/code/adapter/unresyst/combinator/twisted_average.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The twisted average combinator\\\"\\\"\\\"\\n\\n\\nfrom base import BaseCombinator\\n\\nclass TwistedAverageCombinator(BaseCombinator):\\n \\\"\\\"\\\"A combinator using weighted average\\n \\\"\\\"\\\"\\n \\n def _combine(self, combination_elements, ResultClass):\\n \\\"\\\"\\\"See the base class for documentation\\n See the thesis text for the explanation of twisted average\\n \\\"\\\"\\\"\\n \\n # number of positive elements\\n num_positive = len(filter(lambda ce: ce.get_positiveness(), combination_elements))\\n \\n # the difference between the number of positive and negative\\n pos_dif = abs(len(combination_elements) - 2 * num_positive)\\n \\n # the average expectancy\\n avgexp = sum([ce.get_expectancy() for ce in combination_elements]) / len(combination_elements)\\n \\n # select the formula according to the expectancy\\n if avgexp <= 0.5:\\n res_exp = pow(2, pos_dif) * pow(avgexp, pos_dif + 1)\\n else:\\n res_exp = 1 - abs(pow(2, pos_dif) * pow((avgexp - 1), pos_dif + 1))\\n\\n # concat the description\\n desc = self._concat_descriptions(combination_elements)\\n \\n # return the resulting class\\n return ResultClass(expectancy=res_exp, description=desc)\\n\"\n },\n {\n \"alpha_fraction\": 0.5863401889801025,\n \"alphanum_fraction\": 0.5906894207000732,\n \"avg_line_length\": 43.88768005371094,\n \"blob_id\": \"210ebfcf21c921cbe01a6946738b98bfdbeade9f\",\n \"content_id\": \"6705c2e5e9ccff81de0fd76b46a172caf86cb155\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 12416,\n \"license_type\": \"no_license\",\n \"max_line_length\": 111,\n \"num_lines\": 276,\n \"path\": \"/UnresystCD/code/adapter/unresyst/compilator/base.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The base class for all compilators BaseCompilator\\\"\\\"\\\"\\nfrom django.db.models import Q\\n\\nfrom unresyst.constants import *\\nfrom unresyst.combinator.combination_element import BiasAggregateCombinationElement, \\\\\\n SubjectObjectRelCombinationElement, PredictedPlusObjectSimilarityCombinationElement, \\\\\\n PredictedPlusSubjectSimilarityCombinationElement, PredictedPlusObjectClusterMemberCombinationElement, \\\\\\n PredictedPlusSubjectClusterMemberCombinationElement, ClusterSimilarityCombinationElement\\nfrom unresyst.models.aggregator import AggregatedBiasInstance, AggregatedRelationshipInstance\\nfrom unresyst.models.abstractor import PredictedRelationshipDefinition, RelationshipInstance, \\\\\\n ClusterMember, ExplicitRuleInstance\\n\\n\\nclass BaseCompilator(object):\\n \\\"\\\"\\\"The base class for all compilators\\\"\\\"\\\"\\n \\n def __init__(\\n self, \\n combinator=None, \\n depth=DEFAULT_COMPILATOR_DEPTH, \\n breadth=DEFAULT_COMPILATOR_BREADTH,\\n pair_depth=DEFAULT_COMPILATOR_PAIR_DEPTH):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n\\n self.combinator = combinator\\n \\\"\\\"\\\"The combinator that will be used in compiling (e.g. the combination of cluster members\\n to get similarity. If None, the get first strategy is used.\\n \\\"\\\"\\\" \\n \\n self.depth = depth\\n \\\"\\\"\\\"The depth until where the comiplates should be done\\\"\\\"\\\"\\n \\n self.breadth = breadth\\n \\\"\\\"\\\"The number of neighbours that will be taken during the compilation\\\"\\\"\\\" \\n \\n self.pair_depth = pair_depth\\n \\\"\\\"\\\"The number of combination elements that can be taken for each pair from\\n each group. Used only for cluster membership.\\n \\\"\\\"\\\"\\n\\n \\n\\n def compile_all(self, recommender_model):\\n \\\"\\\"\\\"Compile preferences, known relationships + similarities. \\n \\\"\\\"\\\"\\n pass\\n \\n \\n def get_pair_combination_elements(self, dn_subject, dn_object):\\n \\\"\\\"\\\"Find all we know about the relationship of dn_subject and\\n dn_subject using:\\n - aggregated bias for both\\n - s-o relationships (all)\\n - predicted_relationship + object_similarities\\n - predicted_relationship + subject similarities\\n - predicted_relationship + subject cluster memberships (pairs not covered by similarities)\\n - predicted_relationship + object cluster memberships (pairs not covered by similarities)\\n --------- depth COMPILATOR_DEPTH_ONE_UNSURE\\n \\n @type dn_subject, dn_object: SubjectObject \\n @param dn_subject, dn_object: the pair to get combination elements for\\n \\n @rtype: iterable of BaseCombinationElement\\n @return: the list of all we know about the pair\\n \\\"\\\"\\\"\\n recommender_model = dn_subject.recommender\\n els = []\\n other_objs = []\\n other_subjs = [] \\n \\n predicted_def = PredictedRelationshipDefinition.objects.get(\\n recommender=recommender_model)\\n # aggregated bias for both\\n qs_bias = AggregatedBiasInstance.objects.filter(\\n subject_object__id__in=[dn_subject.id, dn_object.id])\\n \\n for bias in qs_bias:\\n els.append(BiasAggregateCombinationElement(bias_aggregate=bias))\\n\\n # s-o relationships (all)\\n #\\n \\n qs_rels = RelationshipInstance.objects\\\\\\n .exclude(definition=predicted_def)\\\\\\n .filter(definition__recommender=recommender_model)\\\\\\n .filter(subject_object1=dn_subject, subject_object2=dn_object)\\\\\\n \\n for rel in qs_rels:\\n els.append(SubjectObjectRelCombinationElement(rel_instance=rel))\\n \\n # explicit also \\n qs_expl_rels = ExplicitRuleInstance.objects\\\\\\n .filter(definition__recommender=recommender_model)\\\\\\n .filter(subject_object1=dn_subject, subject_object2=dn_object)\\\\\\n \\n for rel in qs_expl_rels:\\n els.append(SubjectObjectRelCombinationElement(rel_instance=rel))\\n \\n # predicted_relationship + object_similarities\\n #\\n\\n # get similarities starting the traverse with the similarity.\\n qs_sim1 = AggregatedRelationshipInstance.objects\\\\\\n .filter(recommender=recommender_model)\\\\\\n .filter(\\n Q(\\n # take dn_object as stable - in the subject_object2 position of the similarity\\n subject_object2=dn_object,\\n \\n # traverse from the other object in similarity (subject_object1) through\\n # the relationship instance, its subject (subject_object1) must be dn_subject\\n subject_object1__relationshipinstance_relationships2__subject_object1=dn_subject,\\n \\n # the relationship instance definition must be the predicted relationship def\\n subject_object1__relationshipinstance_relationships2__definition=predicted_def) | \\\\\\n Q( \\n # take dn_object as stable again, now in the subject_object1 position of the similarity\\n subject_object1=dn_object, \\n \\n # traverse from the other through relationship to dn_subject\\n subject_object2__relationshipinstance_relationships2__subject_object1=dn_subject,\\n \\n # the definition again must be the predicted\\n subject_object2__relationshipinstance_relationships2__definition=predicted_def)).distinct()\\n \\n\\n \\n # create combination elements\\n for sim_rel in qs_sim1:\\n \\n # get the object that is similar to dn_object\\n other_obj = sim_rel.get_related(dn_object)\\n \\n other_objs.append(other_obj)\\n \\n # get the predicted relationship to the other object\\n predicted_rel = RelationshipInstance.objects.get(\\n definition=predicted_def, \\n subject_object1=dn_subject,\\n subject_object2=other_obj)\\n \\n els.append(PredictedPlusObjectSimilarityCombinationElement(\\n predicted_rel=predicted_rel,\\n similarity_aggregate=sim_rel))\\n \\n\\n \\n # predicted_relationship + subject similarities\\n #\\n qs_sim2 = AggregatedRelationshipInstance.objects\\\\\\n .filter(recommender=recommender_model)\\\\\\n .filter(\\n Q(\\n # take dn_subject as stable - in the subject_object2 position of the similarity\\n subject_object2=dn_subject,\\n \\n # traverse from the other object in similarity (subject_object1) through\\n # the relationship instance, its object (subject_object2) must be dn_object\\n subject_object1__relationshipinstance_relationships1__subject_object2=dn_object,\\n \\n # the relationship instance definition must be the predicted relationship def\\n subject_object1__relationshipinstance_relationships1__definition=predicted_def) | \\\\\\n Q( \\n # take dn_subject as stable again, now in the subject_object1 position of the similarity\\n subject_object1=dn_subject, \\n \\n # traverse from the other through relationship to dn_object\\n subject_object2__relationshipinstance_relationships1__subject_object2=dn_object,\\n \\n # the definition again must be the predicted\\n subject_object2__relationshipinstance_relationships1__definition=predicted_def)).distinct()\\n \\n # create combination elements\\n for sim_rel in qs_sim2:\\n \\n # get the object that is similar to dn_object\\n other_subj = sim_rel.get_related(dn_subject)\\n\\n other_subjs.append(other_subj)\\n \\n \\n # get the predicted relationship to the other subject\\n predicted_rel = RelationshipInstance.objects.get(\\n definition=predicted_def, \\n subject_object1=other_subj,\\n subject_object2=dn_object)\\n \\n els.append(PredictedPlusSubjectSimilarityCombinationElement(\\n predicted_rel=predicted_rel,\\n similarity_aggregate=sim_rel))\\n\\n \\n \\n # predicted_relationship + object cluster memberships (pairs not covered by similarities) \\n # \\n \\n # take the second membership from the object (first on the path from subject)\\n # exclude the trivial cases\\n qs_memberships_o = ClusterMember.objects.filter(\\n cluster__cluster_set__recommender=recommender_model,\\n cluster__clustermember__member=dn_object,\\n member__relationshipinstance_relationships2__definition=predicted_def,\\n member__relationshipinstance_relationships2__subject_object1=dn_subject)\\\\\\n .exclude(member=dn_object)\\\\\\n .distinct()\\\\\\n .order_by('-confidence')\\n \\n # create the combination elements \\n for cm in qs_memberships_o[:self.pair_depth]:\\n \\n # dont include the ones that are already there because of the similarity\\n if cm.member in other_objs:\\n continue\\n \\n # get the predicted relationship to the subject\\n predicted_rel = RelationshipInstance.objects.get(\\n definition=predicted_def, \\n subject_object1=dn_subject,\\n subject_object2=cm.member)\\n \\n # get the other membership\\n second_membership = ClusterMember.objects.get(\\n cluster=cm.cluster,\\n member=dn_object)\\n \\n # create the combination element out of the memberships \\n ce = ClusterSimilarityCombinationElement(\\n cluster_members=(cm, second_membership)) \\n \\n els.append(PredictedPlusObjectClusterMemberCombinationElement(\\n predicted_rel=predicted_rel,\\n cluster_combination_element=ce)) \\n \\n \\n # predicted_relationship + subject cluster memberships (pairs not covered by similarities) \\n # \\n \\n # take the second membership from the subject\\n # exclude the trivial cases\\n qs_memberships_s = ClusterMember.objects.filter(\\n cluster__cluster_set__recommender=recommender_model,\\n cluster__clustermember__member=dn_subject,\\n member__relationshipinstance_relationships1__definition=predicted_def,\\n member__relationshipinstance_relationships1__subject_object2=dn_object)\\\\\\n .exclude(member=dn_subject)\\\\\\n .distinct()\\\\\\n .order_by('-confidence')\\n \\n # create the combination elements \\n for cm in qs_memberships_s[:self.pair_depth]:\\n \\n # dont include the ones that are already there because of the similarity\\n if cm.member in other_subjs:\\n continue\\n \\n # get the predicted relationship to the object\\n predicted_rel = RelationshipInstance.objects.get(\\n definition=predicted_def, \\n subject_object1=cm.member,\\n subject_object2=dn_object)\\n \\n # get the other membership\\n first_membership = ClusterMember.objects.get(\\n cluster=cm.cluster,\\n member=dn_subject)\\n \\n # create the combination element out of the memberships \\n ce = ClusterSimilarityCombinationElement(\\n cluster_members=(first_membership, cm)) \\n \\n # append the element to the result \\n els.append(PredictedPlusSubjectClusterMemberCombinationElement(\\n predicted_rel=predicted_rel,\\n cluster_combination_element=ce)) \\n \\n return els\\n \\n \\n \\n\"\n },\n {\n \"alpha_fraction\": 0.6250635385513306,\n \"alphanum_fraction\": 0.6255719661712646,\n \"avg_line_length\": 29.703125,\n \"blob_id\": \"25bec476c43ad9c0f3a1bb9e4445a55935df12d9\",\n \"content_id\": \"b1a85e4a4706479fa3844af78cb7d5d6652427f7\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3934,\n \"license_type\": \"no_license\",\n \"max_line_length\": 80,\n \"num_lines\": 128,\n \"path\": \"/UnresystCD/code/adapter/demo/models.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Demo models for the demo application\\\"\\\"\\\"\\n\\nfrom django.db import models\\n\\nfrom constants import *\\n\\nclass ShoePair(models.Model):\\n \\\"\\\"\\\"A model for a pair of shoes.\\\"\\\"\\\"\\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"The shoe model name\\\"\\\"\\\"\\n \\n manufacturer = models.ForeignKey('Manufacturer', null=True)\\n \\\"\\\"\\\"The manufacturer who made the shoes.\\\"\\\"\\\"\\n \\n for_winter = models.BooleanField(default=False)\\n \\\"\\\"\\\"Are the shoes suitable for winter?\\\"\\\"\\\"\\n \\n keywords = models.ManyToManyField('Keyword')\\n \\\"\\\"\\\"The keywords associated with the shoes.\\\"\\\"\\\"\\n\\n image_path = models.CharField(max_length=MAX_LENGTH_IMAGE_PATH)\\n \\\"\\\"\\\"The path to the shoe image, relative to the MEDIA_ROOT\\\"\\\"\\\"\\n \\n category = models.ForeignKey('ShoeCategory', null=True)\\n \\\"\\\"\\\"The category where the shoe pair belongs\\\"\\\"\\\"\\n \\n def get_avatar_path(self):\\n return self.image_path\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n\\n\\nclass ShoeCategory(models.Model): \\n \\\"\\\"\\\"A category of shoes, like casual, sport, ...\\\"\\\"\\\"\\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"The shoe category name\\\"\\\"\\\"\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name\\n\\n\\nclass Keyword(models.Model):\\n \\\"\\\"\\\"A model for a keyword associated with shoe pair.\\\"\\\"\\\"\\n \\n word = models.CharField(max_length=MAX_LENGTH_NAME, unique=True)\\n \\\"\\\"\\\"The key word.\\\"\\\"\\\" \\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.word\\n\\nclass User(models.Model):\\n \\\"\\\"\\\"A model for the users of the demo system.\\\"\\\"\\\"\\n\\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"The user's name\\\"\\\"\\\"\\n\\n age = models.PositiveIntegerField(null=True, default=None)\\n \\\"\\\"\\\"The age of the user.\\\"\\\"\\\"\\n \\n likes_shoes = models.ManyToManyField('ShoePair', related_name='likers')\\n \\\"\\\"\\\"User's favorite shoes.\\\"\\\"\\\"\\n \\n viewed_shoes = models.ManyToManyField('ShoePair', related_name='viewers')\\n \\\"\\\"\\\"Shoes the user has viewed.\\\"\\\"\\\"\\n \\n home_city = models.ForeignKey('City', null=True, default=None) \\n \\\"\\\"\\\"The city where the user lives.\\\"\\\"\\\"\\n \\n words_searched = models.ManyToManyField('Keyword', related_name='searchers')\\n \\\"\\\"\\\"The words searched by the user.\\\"\\\"\\\"\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name\\n\\nclass ShoeRating(models.Model):\\n \\\"\\\"\\\"An explicit rating by user to a shoe pair\\\"\\\"\\\"\\n \\n user = models.ForeignKey('User')\\n \\\"\\\"\\\"The rater\\\"\\\"\\\"\\n \\n shoe_pair = models.ForeignKey('ShoePair')\\n \\\"\\\"\\\"The rated shoe pair\\\"\\\"\\\"\\n \\n stars = models.PositiveIntegerField()\\n \\\"\\\"\\\"Number of stars from 0 to 5 given to the shoepair\\\"\\\"\\\"\\n\\n class Meta:\\n unique_together = ('user', 'shoe_pair')\\n \\\"\\\"\\\"There can be only one rating from a user to a shoe pair\\\"\\\"\\\"\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return u\\\"%s: %s (%d)\\\" % (self.user, self.shoe_pair, self.stars) \\n \\n\\nclass Manufacturer(models.Model):\\n \\\"\\\"\\\"A model for a shoe manufacturer\\\"\\\"\\\"\\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"The name of the manufacturer\\\"\\\"\\\"\\n \\n home_city = models.ForeignKey('City', null=True)\\n \\\"\\\"\\\"The city where the manufacturer seats.\\\"\\\"\\\"\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n\\n\\nclass City(models.Model):\\n \\\"\\\"\\\"A model for a city.\\\"\\\"\\\" \\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"The name of the city.\\\"\\\"\\\"\\n \\n in_south = models.BooleanField(default=False)\\n \\\"\\\"\\\"Is the city in south?\\\"\\\"\\\"\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n\"\n },\n {\n \"alpha_fraction\": 0.7131147384643555,\n \"alphanum_fraction\": 0.7418032884597778,\n \"avg_line_length\": 26,\n \"blob_id\": \"5459b44dd127d7846518d124979b6801463a65b7\",\n \"content_id\": \"0cce9ecc7bab47dcb095cc4beb97f35d24fb5761\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 244,\n \"license_type\": \"no_license\",\n \"max_line_length\": 85,\n \"num_lines\": 9,\n \"path\": \"/code/adapter/modelviz.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#/bin/bash\\n\\n# 1 argument: the application name for which models visualization should be performed\\n# there must exist the 'modelvizs' directory\\n\\npython modelviz.py $1 > $1.dot\\ndot $1.dot -Tpng -o modelvizs/$1.png\\nrm $1.dot\\neog modelvizs/$1.png \\n\"\n },\n {\n \"alpha_fraction\": 0.5691642761230469,\n \"alphanum_fraction\": 0.5788905024528503,\n \"avg_line_length\": 29.80555534362793,\n \"blob_id\": \"e013e8a26d2b1c4a91c20d6a2c421f099f3254c3\",\n \"content_id\": \"bd40d365043ff62f86670e6a64dd0a739646a891\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5552,\n \"license_type\": \"no_license\",\n \"max_line_length\": 104,\n \"num_lines\": 180,\n \"path\": \"/UnresystCD/code/adapter/lastfm/recommender.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The configuration for the last.fm recommender\\\"\\\"\\\"\\nfrom django.db.models import Sum, Count\\n\\nfrom unresyst import *\\n\\nfrom models import *\\nfrom constants import *\\n\\n\\nAGE_DIFFERENCE = 38 - 17\\n\\nREGISTERED_DIFFERENCE = 733696 - 731857\\n\\nMAX_SCROBBLE_COUNT = 85\\n\\nSCROBBLE_DIFFERENCE = MAX_SCROBBLE_COUNT - 1\\n\\ndef _listens_artist_generator():\\n \\\"\\\"\\\"The generator to the predicted relationship\\\"\\\"\\\"\\n for u in User.objects.iterator():\\n for a in Artist.objects.filter(track__scrobble__user=u).distinct().iterator():\\n yield (u, a)\\n\\ndef _get_artist_tag_pairs(artist):\\n \\\"\\\"\\\"Get list tags, confidence for all tags he/she has\\\"\\\"\\\" \\n \\n sumcount = artist.artisttag_set.aggregate(Sum('count'))['count__sum']\\n \\n # get only tags that are shared between at least two artists \\n artist_tags = artist.artisttag_set.annotate(tagcount=Count('tag__artisttag')).filter(tagcount__gt=1)\\n\\n # confidence is counted as \\n # the count of the tag / how many times the artist was tagged.\\n return [(at.tag.name, float(at.count)/sumcount) \\\\\\n for at in artist_tags]\\n\\ndef _tag_similarity_generator():\\n \\\"\\\"\\\"Generate the pairs that share some tags\\\"\\\"\\\"\\n \\n # the artists that were tagged at least once\\n qs_tagged_artists = Artist.objects.filter(artisttag__isnull=False).distinct()\\n count = qs_tagged_artists.count()\\n \\n for a1, count in zip( qs_tagged_artists[1:].iterator(), \\\\\\n range(1, count)):\\n\\n # obtain only first count entities\\n for a2 in qs_tagged_artists[:count].iterator():\\n if a1.artisttag_set.filter(\\n tag__id__in=a2.artisttag_set.values_list('tag__id')\\n ).distinct().count() > 45:\\n yield (a1, a2) \\n\\ndef _gender_specific_generator():\\n \\\"\\\"\\\"Generate pairs for the gender-specific tag rule\\\"\\\"\\\"\\n \\n qs_tagged_artists = Artist.objects.filter(artisttag__tag__gender_specific='m').distinct()\\n \\n for a in qs_tagged_artists:\\n for u in User.objects.filter(gender='f'):\\n yield (u, a)\\n\\nclass NovelArtistRecommender(Recommender):\\n \\\"\\\"\\\"A recommender for discovering previously unheard artists\\\"\\\"\\\" \\n\\n name = \\\"Novel Artist Recommender\\\"\\n \\\"\\\"\\\"The name\\\"\\\"\\\" \\n \\n subjects = User.objects\\n \\\"\\\"\\\"The subjects to who the recommender will recommend.\\\"\\\"\\\"\\n \\n objects = Artist.objects\\n \\\"\\\"\\\"The objects that will be recommended.\\\"\\\"\\\" \\n\\n predicted_relationship = PredictedRelationship( \\n name=\\\"User listens to artist's tracks.\\\",\\n condition=None, \\n description=\\\"\\\"\\\"User %(subject)s listens to the %(object)s's tracks.\\\"\\\"\\\",\\n generator=_listens_artist_generator\\n )\\n \\\"\\\"\\\"The relationship that will be predicted\\\"\\\"\\\"\\n\\n \\\"\\\"\\\"\\n clusters:\\n user:\\n - country\\n - gender\\n \\\"\\\"\\\"\\n \\n relationships = ()\\n \\n rules = (\\n # users of similar age are similar\\n SubjectSimilarityRule(\\n name=\\\"Users with similar age.\\\",\\n \\n # both users have given their age and the difference \\n # is lower than five\\n condition=lambda user1, user2: \\n user1.age and user2.age and abs(user1.age - user2.age) <= 5,\\n \\n is_positive=True, \\n \\n weight=0.4,\\n \\n # a magic linear confidence function\\n confidence=lambda user1, user2: \\n 1 - float(abs(user1.age - user2.age))/AGE_DIFFERENCE,\\n \\n description=\\\"Users %(subject1)s and %(subject2)s are about \\\" + \\n \\\"the same age.\\\"\\n ), \\n ) \\n\\n cluster_sets = (\\n\\n \\n # user - gender\\n SubjectClusterSet(\\n \\n name='User gender.',\\n \\n weight=0.5,\\n \\n # users that have a gender (filled)\\n filter_entities=User.objects.exclude(gender=''),\\n \\n get_cluster_confidence_pairs=lambda user: ((user.gender, 1),),\\n \\n description=\\\"%(subject)s's gender is %(cluster)s.\\\"\\n \\n ),\\n \\n # user - country\\n SubjectClusterSet(\\n \\n name='User country.',\\n \\n weight=0.1,\\n \\n # users that have country filled\\n filter_entities=User.objects.filter(country__isnull=False),\\n \\n get_cluster_confidence_pairs=lambda user: ((user.country.name, 1),),\\n \\n description=\\\"%(subject)s is from %(cluster)s.\\\"\\n ),\\n )\\n \\n save_all_to_predictions = True\\n\\n random_recommendation_description = \\\"Recommending a random artist to the user.\\\"\\n \\n \\n\\nclass ArtistRecommender(NovelArtistRecommender):\\n \\\"\\\"\\\"A recommender for recommending items no matter if they were heard or not.\\\"\\\"\\\"\\n\\n name = \\\"Artist Recommender\\\"\\n \\\"\\\"\\\"The name\\\"\\\"\\\" \\n \\n remove_predicted_from_recommendations = False \\n \\\"\\\"\\\"The already heard artists can appear in recommendations\\\"\\\"\\\"\\n \\n rules = NovelArtistRecommender.rules + ((SubjectObjectRule(\\n name=\\\"User has listened to the artist.\\\",\\n generator=_listens_artist_generator,\\n weight=0.5,\\n is_positive=True,\\n description=\\\"%(subject)s listened to %(object)s.\\\",\\n\\n # the number of user's scrobbles on artist divided by the number of\\n # user's scrobbles overall\\n confidence=lambda s, o: \\n float(Scrobble.objects.filter(user=s, track__artist=o).count())\\\\\\n /Scrobble.objects.filter(user=s).count(),\\n \\n \\n )),\\n )\\n \\n\"\n },\n {\n \"alpha_fraction\": 0.5535823702812195,\n \"alphanum_fraction\": 0.5584813356399536,\n \"avg_line_length\": 28.472726821899414,\n \"blob_id\": \"3bdb67cdeb002662d10c41d4a19ab745c5a01cd5\",\n \"content_id\": \"ba2543ea7825f6c68f2f9989d2aaad071fe2cebd\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3266,\n \"license_type\": \"no_license\",\n \"max_line_length\": 99,\n \"num_lines\": 110,\n \"path\": \"/code/adapter/flixster/models.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Models for the flixster app\\\"\\\"\\\"\\nfrom django.db import models\\nfrom django.db.models import Min\\n\\nfrom unresyst.models import BaseEvaluationPair\\nfrom constants import *\\n\\nclass User(models.Model):\\n \\\"\\\"\\\"The user\\\"\\\"\\\"\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return 'user_%d' % self.id \\n \\n \\nclass Movie(models.Model): \\n \\\"\\\"\\\"The movie\\\"\\\"\\\"\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return 'movie_%d' % self.id \\n\\n \\nclass Friend(models.Model):\\n \\\"\\\"\\\"A user is a friend of another user. Symmetric. Smaller id first.\\\"\\\"\\\"\\n\\n friend1 = models.ForeignKey('flixster.User', related_name='friends1')\\n \\\"\\\"\\\"The first user in the friend relationship\\\"\\\"\\\" \\n \\n friend2 = models.ForeignKey('flixster.User', related_name='friends2')\\n \\\"\\\"\\\"The second user in the friend relationship\\\"\\\"\\\" \\n \\n class Meta:\\n unique_together = ('friend1', 'friend2') \\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return u'(%s, %s)' % (self.friend1, self.friend2)\\n\\n \\nclass Rating(models.Model):\\n \\\"\\\"\\\"A user rating a movie\\\"\\\"\\\"\\n \\n user = models.ForeignKey('flixster.User')\\n \\\"\\\"\\\"The user\\\"\\\"\\\"\\n \\n movie = models.ForeignKey('flixster.Movie')\\n \\\"\\\"\\\"The movie\\\"\\\"\\\"\\n \\n rating = models.DecimalField(max_digits=2, decimal_places=1)\\n \\\"\\\"\\\"The rating\\\"\\\"\\\"\\n\\n class Meta:\\n unique_together = ('user', 'movie')\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return u'%s - %s: %f' % (self.user, self.movie, self.rating)\\n\\n\\nclass MovieEvaluationPair(BaseEvaluationPair):\\n \\\"\\\"\\\"Movie test pairs\\\"\\\"\\\"\\n\\n subj = models.ForeignKey('flixster.User')\\n \\\"\\\"\\\"The subject\\\"\\\"\\\"\\n \\n obj = models.ForeignKey('flixster.Movie')\\n \\\"\\\"\\\"The object\\\"\\\"\\\"\\n \\n test_ratio = 0.2\\n \\\"\\\"\\\"The ratio of pairs to select to test pairs\\\"\\\"\\\"\\n\\n class Meta:\\n app_label = 'flixster' \\n \\n @classmethod \\n def select(cls, i=0):\\n \\\"\\\"\\\"See the base class for the documentation.\\\"\\\"\\\"\\n\\n n = int(1 / cls.test_ratio) \\n all_count = Rating.objects.count() \\n \\n it = 0 \\n # take every n-th rating, remove it and put it to test data \\n # save to test, remove from build\\n for rating in Rating.objects.order_by('id').iterator():\\n\\n it += 1\\n\\n # if we aren't on the nth object go ahead\\n if it % n != i:\\n continue\\n \\n # save the test object and delete it from train\\n cls.objects.create(\\n subj=rating.user,\\n obj=rating.movie,\\n expected_expectancy=rating.rating/MAX_STARS)\\n \\n rating.delete()\\n \\n \\n test_count = cls.objects.count()\\n \\n print \\\"%d test pairs selected from total %d pairs\\\" % (test_count, all_count) \\n\\n\\n def get_success(self):\\n \\\"\\\"\\\"See the base class for the documentation.\\\"\\\"\\\"\\n return abs(self.obtained_expectancy - self.expected_expectancy) < MAX_TOLERANCE / MAX_STARS\\n \\n \\n \\n\"\n },\n {\n \"alpha_fraction\": 0.6610608100891113,\n \"alphanum_fraction\": 0.6610608100891113,\n \"avg_line_length\": 31.20833396911621,\n \"blob_id\": \"3334f4d95e591e313845bdaff67bb82b060bdc3d\",\n \"content_id\": \"efc32801bbcc7d2e860c6653c2dacbea11ff5a43\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 773,\n \"license_type\": \"no_license\",\n \"max_line_length\": 94,\n \"num_lines\": 24,\n \"path\": \"/code/adapter/urls.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from django.conf.urls.defaults import *\\n\\nfrom settings import SERVE_STATIC_FILES, MEDIA_ROOT\\n\\n# Uncomment the next two lines to enable the admin:\\n# from django.contrib import admin\\n# admin.autodiscover()\\n\\nurlpatterns = patterns('',\\n \\n # the demo app \\n (r'^', include('demo.urls', namespace='demo'))\\n # Uncomment the admin/doc line below to enable admin documentation:\\n # (r'^admin/doc/', include('django.contrib.admindocs.urls')),\\n\\n # Uncomment the next line to enable the admin:\\n # (r'^admin/', include(admin.site.urls)),\\n \\n)\\nif SERVE_STATIC_FILES:\\n urlpatterns += patterns('',\\n # for working with static files (for development only)\\n (r'^media/(?P.*)$', 'django.views.static.serve', {'document_root': MEDIA_ROOT}) \\n )\\n\"\n },\n {\n \"alpha_fraction\": 0.5111602544784546,\n \"alphanum_fraction\": 0.5206112861633301,\n \"avg_line_length\": 35.74074172973633,\n \"blob_id\": \"648a211ed35ae2ae4920e4491edf2db3d0765fc8\",\n \"content_id\": \"4ce47597afb1a0b95f3938f187880230f8333f0c\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 4973,\n \"license_type\": \"no_license\",\n \"max_line_length\": 114,\n \"num_lines\": 135,\n \"path\": \"/code/adapter/travel/recbckp.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \" rules = ( \\n # click = sign of preference\\n SubjectObjectRule(\\n name='User has clicked on something on the tour profile.',\\n weight=0.5,\\n condition=None,\\n is_positive=True,\\n description='User %(subject)s has clicked on something on the %(object)s profile.',\\n # pairs that user has clicked on the tour\\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\\\\n for uid, tid in Click.objects.values_list('session__user', 'tour').distinct()),\\n # the average is around 3, so take 1/6. so that 3 points to the middle.\\n confidence=lambda u, t: min(float(Click.objects.filter(tour=t, session__user=u).count())/6, 1.0),\\n ),\\n \\n # question .. also a sign of preference\\n SubjectObjectRule(\\n name='User has asked about the tour.',\\n weight=0.5,\\n condition=None,\\n is_positive=True,\\n description='User %(subject)s has asked about %(object)s.',\\n # pairs that user has asked on the tour\\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\\\\n for uid, tid in Question.objects.values_list('session__user', 'tour').distinct()),\\n confidence=lambda u, t: min(float(Question.objects.filter(tour=t, session__user=u).count())/2, 1.0),\\n ),\\n \\n # mouse move .. also a sign of preference\\n SubjectObjectRule(\\n name='User has moved the mouse on the tour profile.',\\n weight=0.5,\\n condition=None,\\n is_positive=True,\\n description='User %(subject)s has moved the mouse on %(object)s.',\\n # pairs that user has moved on the tour\\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\\\\n for uid, tid in MouseMove.objects.values_list('session__user', 'tour').distinct()),\\n confidence=lambda u, t: min(float(MouseMove.objects.filter(tour=t, session__user=u).count())/18, 1.0),\\n ),\\n \\n # view profile \\n SubjectObjectRule(\\n name='User has viewed the tour profile page.',\\n weight=0.5,\\n condition=None,\\n is_positive=True,\\n description='User %(subject)s has viewed %(object)s.',\\n # pairs that user has viewed the tour\\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\\\\n for uid, tid in ViewProfile.objects.values_list('session__user', 'tour').distinct()),\\n # how many times * how long\\n confidence=_viewed_profile_confidence\\n ),\\n\\n \\n )\\n \\n cluster_sets = (\\n # cluster - tour types\\n ObjectClusterSet(\\n\\n name=\\\"Tour type cluster set.\\\",\\n\\n weight=0.5,\\n \\n filter_entities=Tour.objects.all(),\\n \\n get_cluster_confidence_pairs=lambda tour: ((tour.tour_type.name, 1),),\\n \\n description=\\\"The tour %(object)s has type %(cluster)s.\\\",\\n ),\\n \\n # cluster - countries\\n ObjectClusterSet(\\n\\n name=\\\"Country cluster set.\\\",\\n\\n weight=0.5,\\n \\n filter_entities=Tour.objects.all(),\\n \\n get_cluster_confidence_pairs=lambda tour: ((tour.country.name, 1),),\\n \\n description=\\\"The tour %(object)s is to %(cluster)s country.\\\",\\n ),\\n\\n )\\n \\n biases = (\\n # multiply viewed tours\\n ObjectBias(\\n name=\\\"Most viewed tours.\\\",\\n \\n description=\\\"Tour %(object)s is much viewed\\\",\\n \\n weight=0.5,\\n \\n is_positive=True,\\n \\n generator=lambda: Tour.objects.annotate(hh=Count('viewprofile')).filter(hh__gt=2).distinct(),\\n \\n confidence=lambda t: min(float(t.viewprofile_set.count())/4, 1.0)\\n ),\\n \\n # multiply clicked tours\\n ObjectBias(\\n name=\\\"Most clicked tours.\\\",\\n \\n description=\\\"Tour %(object)s is often clicked on.\\\",\\n \\n weight=0.5,\\n \\n is_positive=True,\\n \\n generator=lambda: Tour.objects.annotate(hh=Count('click')).filter(hh__gt=1).distinct(),\\n \\n confidence=lambda t: min(float(t.click_set.count())/2, 1.0)\\n ),\\n \\n # multiply mouse moved tours\\n ObjectBias(\\n name=\\\"Most mouse moved tours.\\\",\\n \\n description=\\\"Tour %(object)s is often mouse moved.\\\",\\n \\n weight=0.5,\\n \\n is_positive=True,\\n \\n generator=lambda: Tour.objects.annotate(hh=Count('mousemove')).filter(hh__gt=6).distinct(),\\n \\n confidence=lambda t: min(float(t.mousemove_set.count())/12, 1.0)\\n ),\\n )\\n\\n\\n\\n \\n\\n\"\n },\n {\n \"alpha_fraction\": 0.8060836791992188,\n \"alphanum_fraction\": 0.8060836791992188,\n \"avg_line_length\": 28.22222137451172,\n \"blob_id\": \"03439cbcfe150685e23ccce6e7ba2bcdefedf629\",\n \"content_id\": \"83dd87a6ac3bcf90c1e71fdb345da51af7d52ce5\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 263,\n \"license_type\": \"no_license\",\n \"max_line_length\": 57,\n \"num_lines\": 9,\n \"path\": \"/code/adapter/unresyst/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"A package holding the whole unresyst application.\\n\\nThe only class to be used outside is the BaseRecommender.\\n\\\"\\\"\\\"\\n\\nfrom recommender.recommender import Recommender\\nfrom recommender.rules import *\\nfrom recommender.clusters import *\\nfrom recommender.bias import *\\n\"\n },\n {\n \"alpha_fraction\": 0.5788713097572327,\n \"alphanum_fraction\": 0.5795742869377136,\n \"avg_line_length\": 35.88328552246094,\n \"blob_id\": \"85ee8cdf4107b400d05f293407aa78ee45437b76\",\n \"content_id\": \"0755784d1a25ffa1dfb225da4565b41a8ff03c88\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 25605,\n \"license_type\": \"no_license\",\n \"max_line_length\": 120,\n \"num_lines\": 694,\n \"path\": \"/code/adapter/unresyst/recommender/recommender.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Module containing the definition of BaseRecommender: the base class \\nfor user-defined recommenders.\\n\\nContents:\\n - Recommender: the recommender for client subclassing\\n - MetaRecommender: the metaclass for creating recommender classes\\n\\\"\\\"\\\"\\nimport math\\nimport copy\\nimport csv\\n\\nfrom base import BaseRecommender\\nfrom predictions import RelationshipPrediction\\nfrom unresyst.constants import *\\nfrom unresyst.exceptions import ConfigurationError, InvalidParameterError, \\\\\\n RecommenderNotBuiltError\\nfrom unresyst.abstractor import BasicAbstractor \\nfrom unresyst.aggregator import LinearAggregator, CombiningAggregator\\nfrom unresyst.algorithm import SimpleAlgorithm, AggregatingAlgorithm, CompilingAlgorithm\\nfrom unresyst.models.common import SubjectObject, Recommender as RecommenderModel\\nfrom unresyst.compilator import GetFirstCompilator, CombiningCompilator\\nfrom unresyst.combinator import AverageCombinator, TwistedAverageCombinator, ConfidenceFactorCombinator\\nfrom unresyst.models.abstractor import RelationshipInstance, ExplicitRuleInstance\\nfrom unresyst.models.algorithm import RelationshipPredictionInstance\\n\\ndef _assign_recommender(list_rels, recommender):\\n \\\"\\\"\\\"Go throuth the list, if the items have the \\\"recommender\\\" attribute,\\n create a copy to the returning list, if not put it there directly\\n \\n @param list_rels: the list of rules/relationships\\n \\n @param recommender: the recommender they belong to \\n \\n @rtype: tuple of rules/relationships\\n @return: the tuple of new or copied rules/relationships\\n \\\"\\\"\\\"\\n new_rels = []\\n \\n # go through the rels \\n for rel in list_rels:\\n \\n # if they already have a recommender assigned, copy them\\n if hasattr(rel, \\\"recommender\\\"):\\n rel = copy.copy(rel)\\n \\n # assign the recommender and append it \\n rel.recommender = recommender\\n new_rels.append(rel)\\n \\n return tuple(new_rels)\\n\\nclass MetaRecommender(type):\\n \\\"\\\"\\\"The meta-class adding a reference to the class for the contained\\n rules and relationships.\\n \\\"\\\"\\\"\\n \\n def __init__(cls, name, bases, dct): \\n \\\"\\\"\\\"The class initializer.\\n \\n Adds the reference to the recommender class to all of the rules\\n and relationships.\\n \\n Tries searching for the recommender model in the database.\\n \\\"\\\"\\\"\\n \\n super(MetaRecommender, cls).__init__(name, bases, dct)\\n \\n # add the recommender class to the predicted relationship\\n if cls.predicted_relationship: \\n\\n # if it already has a recommender assigned, create a copy.\\n if hasattr(cls.predicted_relationship, 'recommender'):\\n cls.predicted_relationship = copy.copy(cls.predicted_relationship)\\n \\n cls.predicted_relationship.recommender = cls\\n \\n # then to the rules\\n if cls.rules:\\n cls.rules = _assign_recommender(list_rels=cls.rules, recommender=cls) \\n \\n # finally to the relationships\\n if cls.relationships:\\n cls.relationships = _assign_recommender(\\n list_rels=cls.relationships, \\n recommender=cls)\\n \\n # moreover to the cluster sets\\n if cls.cluster_sets:\\n cls.cluster_sets = _assign_recommender(\\n list_rels=cls.cluster_sets,\\n recommender=cls) \\n \\n # and to the biases\\n if cls.biases:\\n cls.biases = _assign_recommender(\\n list_rels=cls.biases,\\n recommender=cls) \\n\\n\\nclass Recommender(BaseRecommender):\\n \\\"\\\"\\\"The base class for all user-defined recommenders.\\n \\n Implements the interface of the recommender. It doesn't hold any \\n domain specific data.\\n \\n Defines default behaviour assigning the classes for the layers. \\n \\n The predicted relationship is ignored when building and updating the \\n recommender. If it's needed, one should create a rule/relationship \\n with the appropriate condition.\\n \\n The contained rules and relationships have to be included in only one\\n recommender class. There can't be a rule/relationship instance \\n which is contained in two recommender classes.\\n \\\"\\\"\\\"\\n __metaclass__ = MetaRecommender\\n \\\"\\\"\\\"A metaclass putting the reference to the recommender to all member\\n rules and relationships.\\n \\\"\\\"\\\" \\n \\n # Build phase:\\n #\\n \\n @classmethod\\n def build(cls):\\n \\\"\\\"\\\"For documentation, see the base class\\\"\\\"\\\"\\n \\n \\n # validate the subclass data\\n # \\n \\n # Subjects, objects non-empty\\n if not cls.subjects.all().exists():\\n raise ConfigurationError(\\n message=\\\"No subjects given\\\",\\n recommender=cls,\\n parameter_name=\\\"Recommender.subjects\\\",\\n parameter_value=cls.subjects\\n )\\n \\n if not cls.objects.all().exists():\\n raise ConfigurationError(\\n message=\\\"No objects given\\\",\\n recommender=cls,\\n parameter_name=\\\"Recommender.objects\\\",\\n parameter_value=cls.objects\\n )\\n \\n \\n # predicted relationship given\\n if not cls.predicted_relationship:\\n raise ConfigurationError(\\n message=\\\"No predicted relationship given\\\",\\n recommender=cls,\\n parameter_name=\\\"Recommender.predicted_relationship\\\",\\n parameter_value=cls.predicted_relationship\\n )\\n \\n # rules and relationships don't have to be given\\n \\n cls._print('Recommender validated, deleting old objects...')\\n \\n # if the recommender with the given name exists, delete it,\\n RecommenderModel.objects.filter(class_name=cls.__name__).delete()\\n \\n # create a new recommender and save it, keep it in the class\\n recommender_model = RecommenderModel(\\n class_name=cls.__name__,\\n name=cls.name,\\n is_built=False,\\n are_subjects_objects=(cls.subjects == cls.objects),\\n random_recommendation_description=cls.random_recommendation_description,\\n remove_predicted_from_recommendations=cls.remove_predicted_from_recommendations\\n ) \\n recommender_model.save() \\n \\n # build the recommender model\\n #\\n #\\n cls._print(\\\"Old objects deleted. Creating universal subjectobjects...\\\")\\n \\n # Abstractor\\n #\\n \\n # create the domain neutral representation for objects and subjects\\n cls.abstractor.create_subjectobjects(\\n recommender_model=recommender_model,\\n subjects=cls.subjects, \\n objects=cls.objects\\n )\\n \\n cls._print(\\\"Universal subject and object representations created. Creating predicted_relationship instances...\\\")\\n \\n # create the relationship instances for the predicted relationship\\n cls.abstractor.create_predicted_relationship_instances( \\n predicted_relationship=cls.predicted_relationship \\n )\\n \\n cls._print(\\\"Predicted relationship instances created. Creating relationship instances...\\\")\\n \\n # create relationship instances between subjects/objects \\n cls.abstractor.create_relationship_instances(\\n relationships=cls.relationships\\n ) \\n \\n cls._print(\\\"Relationship instances created. Creating rule instances...\\\")\\n \\n # evaluate rules and make rule instances between the affected \\n # subjects/objects\\n cls.abstractor.create_rule_instances(rules=cls.rules)\\n \\n cls._print(\\\"Rule instances created. Creating clusters...\\\")\\n \\n # evaluate the clusters and their members\\n cls.abstractor.create_clusters(cluster_sets=cls.cluster_sets)\\n \\n cls._print(\\\"Clusters created. Creating biases...\\\")\\n \\n # evaluate the biases\\n cls.abstractor.create_biases(biases=cls.biases)\\n \\n cls._print(\\\"Biases created. Aggregating...\\\")\\n\\n \\n \\n # Algorithm\\n # \\n # build the algorithm model from the aggregated relationships\\n cls.algorithm.build(recommender_model=recommender_model)\\n \\n cls._print(\\\"Algorithm built.\\\")\\n \\n # if it should be done and predicted should be removed, \\n # save predicted_rel to predictions\\n if cls.remove_predicted_from_recommendations and cls.save_all_to_predictions:\\n \\n cls._print(\\\"Saving explicit/predicted to predictions...\\\")\\n \\n # if explicit relationship is available, get its instances\\n # if not, get the predicted_rel\\n qs_predicted_rels = ExplicitRuleInstance.objects.filter(definition__recommender=recommender_model) \\\\\\n if cls.explicit_rating_rule else \\\\\\n RelationshipInstance.filter_predicted(recommender_model=recommender_model)\\n \\n for ri in qs_predicted_rels:\\n \\n # get the expectancy of the rating or the trivial\\n expectancy = ri.expectancy if cls.explicit_rating_rule else TRIVIAL_EXPECTANCY \\n \\n rpi, created = RelationshipPredictionInstance.objects.get_or_create(\\n subject_object1=ri.subject_object1,\\n subject_object2=ri.subject_object2,\\n recommender=recommender_model,\\n defaults={\\n 'expectancy': expectancy,\\n 'is_trivial': True,\\n 'description': ri.description,\\n }\\n )\\n \\n # if it was found update it to the predicted\\n if not created:\\n rpi.expectancy = expectancy\\n rpi.is_trivial = True\\n rpi.description = ri.description\\n \\n # save it if created or not \\n rpi.save()\\n \\n # mark the recommender as built, save it and keep it in the class\\n recommender_model.is_built = True\\n recommender_model.save()\\n \\n cls._print('Done')\\n\\n\\n # Recommend phase:\\n # \\n \\n @classmethod \\n def predict_relationship(cls, subject, object_, save_to_db=False):\\n \\\"\\\"\\\"For documentation, see the base class\\\"\\\"\\\" \\n \\n recommender_model = cls._get_recommender_model()\\n # if the recommender isn't built raise an error\\n if not recommender_model or not recommender_model.is_built:\\n raise RecommenderNotBuiltError(\\n message=\\\"Build the recommender prior to performing the \\\" + \\\\\\n \\\"predict_relationship action.\\\",\\n recommender=cls\\n )\\n \\n \\n subject_ent_type = ENTITY_TYPE_SUBJECT \\\\\\n if not recommender_model.are_subjects_objects \\\\\\n else ENTITY_TYPE_SUBJECTOBJECT\\n\\n object_ent_type = ENTITY_TYPE_OBJECT \\\\\\n if not recommender_model.are_subjects_objects \\\\\\n else ENTITY_TYPE_SUBJECTOBJECT\\n\\n # get the domain neutral representations for the subject and object\\n try:\\n dn_subject = SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=subject,\\n entity_type=subject_ent_type,\\n recommender=recommender_model\\n )\\n except SubjectObject.DoesNotExist, e:\\n raise InvalidParameterError(\\n message=\\\"The subject wasn't found in the recommender database.\\\" + \\\\\\n \\\"Try rebuilding the recommender. Exception: %s\\\" % e,\\n recommender=cls,\\n parameter_name='subject', \\n parameter_value=subject) \\n\\n try: \\n dn_object = SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=object_,\\n entity_type=object_ent_type,\\n recommender=recommender_model\\n )\\n except SubjectObject.DoesNotExist, e:\\n raise InvalidParameterError(\\n message=\\\"The object wasn't found in the recommender database.\\\" + \\\\\\n \\\"Try rebuilding the recommender. Exception: %s\\\" % e,\\n recommender=cls,\\n parameter_name='object_', \\n parameter_value=object_) \\n \\n # get the prediction from the algorithm\\n prediction_model = cls.algorithm.get_relationship_prediction(\\n recommender_model=recommender_model,\\n dn_subject=dn_subject,\\n dn_object=dn_object,\\n remove_predicted=cls.remove_predicted_from_recommendations\\n )\\n \\n # if it should be done and we know something about the pair\\n if save_to_db and not prediction_model.is_uncertain:\\n prediction_model.save()\\n \\n # create and return the outer-world object\\n prediction = RelationshipPrediction(\\n subject=subject,\\n object_=object_,\\n expectancy=prediction_model.expectancy,\\n explanation=prediction_model.description,\\n is_uncertain=prediction_model.is_uncertain\\n ) \\n return prediction\\n\\n\\n @classmethod\\n def get_recommendations(cls, subject, count=None): \\n \\\"\\\"\\\"For documentation, see the base class\\\"\\\"\\\"\\n \\n recommender_model = cls._get_recommender_model()\\n \\n # if the recommender isn't built raise an error\\n if not recommender_model or not recommender_model.is_built:\\n raise RecommenderNotBuiltError(\\n message=\\\"Build the recommender prior to performing the \\\" + \\\\\\n \\\"get_recommendations action.\\\",\\n recommender=cls\\n )\\n \\n # if count wasn't given take the default one\\n if not count:\\n count = cls.default_recommendation_count\\n \\n subject_ent_type = ENTITY_TYPE_SUBJECT \\\\\\n if not recommender_model.are_subjects_objects \\\\\\n else ENTITY_TYPE_SUBJECTOBJECT\\n \\n # convert the the subject to domain neutral\\n try:\\n dn_subject = SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=subject,\\n entity_type=subject_ent_type,\\n recommender=recommender_model\\n )\\n except SubjectObject.DoesNotExist, e:\\n raise InvalidParameterError(\\n message=\\\"The subject wasn't found in the recommender database.\\\" + \\\\\\n \\\"Try rebuilding the recommender. Exception: %s\\\" % e,\\n recommender=cls,\\n parameter_name='subject', \\n parameter_value=subject) \\n\\n limit = cls.recommendation_expectancy_limit \\\\\\n if not cls.recommendation_expectancy_limit is None else 0\\n\\n # get the recommendations from the algorithm\\n prediction_models = cls.algorithm.get_recommendations(\\n recommender_model=recommender_model,\\n dn_subject=dn_subject,\\n count=count,\\n expectancy_limit=limit,\\n remove_predicted=cls.remove_predicted_from_recommendations\\n )\\n \\n recommendations = []\\n \\n # go through the obtained predictions\\n for pred_model in prediction_models:\\n\\n # obtain the object from the prediction\\n dn_object = pred_model.get_related(dn_subject)\\n \\n # get its domain specific representation\\n object_ = dn_object.get_domain_specific_entity(entity_manager=cls.objects)\\n \\n # create the outer-world object\\n prediction = RelationshipPrediction(\\n subject=subject,\\n object_=object_,\\n expectancy=pred_model.expectancy,\\n explanation=pred_model.description,\\n is_uncertain=pred_model.is_uncertain\\n ) \\n \\n recommendations.append(prediction)\\n\\n return recommendations\\n\\n @classmethod\\n def export_predictions(cls, filename):\\n \\\"\\\"\\\"Export all predictions to a csv \\n file of the given name.\\n \\n @type filename: str\\n @param filename: the full path to the file\\n \\n @raise FileNotExists and other file open errors.\\n \\\"\\\"\\\"\\n recommender_model = cls._get_recommender_model()\\n \\n with open(filename, 'w') as f:\\n \\n i = 0\\n \\n # loop through pairs, export the prediction instances\\n for rpi in RelationshipPredictionInstance.objects.filter(recommender=recommender_model):\\n \\n # create the common part\\n linestr = \\\"%s,%s,%s\\\\n\\\" % (\\n rpi.subject_object1.id_in_specific, \\n rpi.subject_object2.id_in_specific,\\n rpi.expectancy) \\n \\n # write it to the file\\n f.write(linestr)\\n \\n i += 1\\n \\n print \\\" %d predictions exported\\\" % i\\n\\n @classmethod\\n def update_predictions(cls, filename):\\n \\\"\\\"\\\"Load predictions for pairs that are unknown in our recommender from \\n the given csv file.\\n \\n The file has to be in format:\\n ,,\\\\n\\n \\n @type filename: str\\n @param filename: the full path to the file\\n\\n @raise FileNotExists and other file open errors.\\n \\\"\\\"\\\"\\n recommender_model = cls._get_recommender_model() \\n \\n # open the csv reader\\n reader = csv.reader(open(filename, \\\"rb\\\"), delimiter=',', quoting=csv.QUOTE_NONE)\\n \\n i = 0\\n # parse the csv line by line\\n for subj_id, obj_id, expectancy in reader:\\n \\n # parse the values to the right types\\n subj_id = int(subj_id)\\n obj_id = int(obj_id)\\n expectancy = float(expectancy)\\n \\n dn_subject = SubjectObject.objects.get(\\n entity_type=ENTITY_TYPE_SUBJECT, \\n id_in_specific=subj_id,\\n recommender=recommender_model)\\n dn_object = SubjectObject.objects.get(\\n entity_type=ENTITY_TYPE_OBJECT, \\n id_in_specific=obj_id,\\n recommender=recommender_model)\\n \\n # if it's already imported, go to the next pair\\n if RelationshipPredictionInstance.objects.filter(\\n subject_object1=dn_subject, \\n subject_object2=dn_object, \\n recommender=recommender_model).exists():\\n continue\\n \\n # create a prediction\\n RelationshipPredictionInstance.objects.create(\\n subject_object1=dn_subject,\\n subject_object2=dn_object,\\n recommender=recommender_model,\\n expectancy=expectancy\\n )\\n\\n i += 1\\n\\n print \\\"%d new predictions imported\\\" % i\\n \\n \\n \\n \\n \\n # Update phase:\\n # \\n @classmethod\\n def add_subject(cls, subject):\\n \\\"\\\"\\\"For documentation, see the base class\\\"\\\"\\\"\\n \\n # add the object to abstractor rule/relationship instances, \\n # see what has changed\\n instance_changes = cls.Abstractor.add_subject(\\n recommender=cls.__name__,\\n subject=subject\\n )\\n \\n # let the Aggregator update the aggregates\\n aggregated_changes = cls.Aggregator.update(\\n recommender=cls.__name__,\\n instance_changes=instance_changes\\n )\\n \\n # let the Algorithm update its structures\\n cls.Algorithm.update(\\n recommender=cls.__name__,\\n aggregated_changes=aggregated_changes\\n )\\n\\n\\n @classmethod\\n def add_object(cls, object_):\\n \\\"\\\"\\\"For documentation, see the base class\\\"\\\"\\\"\\n \\n # add the object to abstractor rule/relationship instances\\n # see what has changed\\n instance_changes = cls.Abstractor.add_object(\\n recommender=cls.__name__,\\n object_=object_\\n )\\n # pak nejak podobne - asi na to udelat fci. \\n\\n @classmethod\\n def update_subject(cls, subject):\\n \\\"\\\"\\\"For documentation, see the base class\\\"\\\"\\\"\\n\\n # update the subject in abstractor rule/relationship instances\\n # see what has changed\\n instance_changes = cls.Abstractor.update_subject(\\n recommender=cls.__name__,\\n subject=subject\\n )\\n # pak nejak podobne - asi na to udelat fci.\\n\\n\\n @classmethod \\n def update_object(cls, object_):\\n \\\"\\\"\\\"For documentation, see the base class\\\"\\\"\\\"\\n\\n # update the object in abstractor rule/relationship instances\\n # see what has changed\\n instance_changes = cls.Abstractor.update_object(\\n recommender=cls.__name__,\\n object_=object_\\n )\\n # pak nejak podobne - asi na to udelat fci. \\n \\n @classmethod\\n def remove_subject(cls, subject):\\n \\\"\\\"\\\"For documentation, see the base class\\\"\\\"\\\"\\n\\n # remove the subject from abstractor rule/relationship instances\\n # see what has changed\\n instance_changes = cls.Abstractor.remove_subject(\\n recommender=cls.__name__,\\n subject=subject\\n )\\n # pak nejak podobne - asi na to udelat fci.\\n\\n\\n @classmethod \\n def remove_object(cls, object_):\\n \\\"\\\"\\\"For documentation, see the base class\\\"\\\"\\\"\\n\\n # update the object in abstractor rule/relationship instances\\n # see what has changed\\n instance_changes = cls.Abstractor.remove_object(\\n recommender=cls.__name__,\\n object_=object_\\n )\\n # pak nejak podobne - asi na to udelat fci. \\n \\n \\n # Class configuration - the behaviour of the layers below the recommender\\n # Can be overriden in user defined subclasses\\n \\n abstractor = BasicAbstractor()\\n \\\"\\\"\\\"The class that will be used for the abstractor level. Can be \\n overriden in suclasses\\\"\\\"\\\" \\n \\n algorithm = AggregatingAlgorithm(\\n inner_algorithm=CompilingAlgorithm(\\n inner_algorithm=SimpleAlgorithm(\\n inner_algorithm=None\\n ),\\n compilator=CombiningCompilator(combinator=TwistedAverageCombinator(), breadth=0) #\\n ),\\n aggregator=CombiningAggregator(combinator=TwistedAverageCombinator())\\n )\\n \\\"\\\"\\\"The default algorithm setup.\\n The class that will be used for the algorithm level. Can be \\n overriden in subclasses\\\"\\\"\\\" \\n \\n # TwistedAverageCombinator\\n # AverageCombinator\\n # ConfidenceFactorCombinator\\n \\n default_recommendation_count = DEFAULT_RECOMMENDATION_COUNT\\n \\\"\\\"\\\"The defaul count of the obtained recommended objects\\\"\\\"\\\"\\n \\n remove_predicted_from_recommendations = True\\n \\\"\\\"\\\"The entity pairs that already have the predicted_relationship\\n between them, will be removed from recommendation list.\\n \\\"\\\"\\\"\\n \\n random_recommendation_description = \\\"A random object\\\"\\n \\\"\\\"\\\"The description for random recommendations Can be overriden in\\n subclass\\\"\\\"\\\"\\n \\n verbose_build = True\\n \\\"\\\"\\\"Should messages be printed during the build?\\\"\\\"\\\"\\n \\n save_all_to_predictions = True\\n \\n # Auxiliary methods - not to be used from outside the application\\n # \\n @classmethod\\n def _get_recommender_model(cls):\\n \\\"\\\"\\\"Get the recommender model belonging to the class\\\"\\\"\\\"\\n \\n # can't be caching 'cause database can die out without notifying the \\n # if it's already saved, return it\\n #if cls._recommender_model:\\n # return cls._recommender_model\\n \\n # otherwise try finding it in database \\n models = RecommenderModel.objects.filter(class_name=cls.__name__)\\n \\n # if the recommender was found, assign it to the class and return it\\n if models: \\n assert len(models) == 1 \\n return models[0]\\n \\n # if not return None\\n return None \\n \\n @classmethod\\n def _get_entity_manager(cls, entity_type):\\n \\\"\\\"\\\"Get the manager from the recommender for the given entity type.\\n \\n @type entity_type: str\\n @param entity_type: the type of the entity 'S'/'O'/'SO'\\n \\n @rtype: django.db.models.manager.Manager\\n @return: the manager over the domain specific entities.\\n \\n @raise KeyError: when the entity_type is invalid \\n \\\"\\\"\\\"\\n \\n manager_dict = {\\n ENTITY_TYPE_SUBJECT: cls.subjects,\\n ENTITY_TYPE_OBJECT: cls.objects, \\n ENTITY_TYPE_SUBJECTOBJECT: cls.objects # or sujbects if you care\\n }\\n \\n return manager_dict[entity_type]\\n \\n @classmethod\\n def _print(cls, msg):\\n if cls.verbose_build:\\n print msg \\n\"\n },\n {\n \"alpha_fraction\": 0.715423583984375,\n \"alphanum_fraction\": 0.7286990284919739,\n \"avg_line_length\": 27.57241439819336,\n \"blob_id\": \"a6e4e96e9efadd85c404c0083ff7cea0be5b242a\",\n \"content_id\": \"eb3b0ec2bf38722d752add83f7bf080bf7a09b14\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 4143,\n \"license_type\": \"no_license\",\n \"max_line_length\": 87,\n \"num_lines\": 145,\n \"path\": \"/UnresystCD/code/adapter/unresyst/constants.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The constants used in Unresyst\\\"\\\"\\\"\\n\\n_ = lambda x: x\\n\\nMAX_LENGTH_NAME = 80\\n\\\"\\\"\\\"The maximum length of the name in the universal representation.\\\"\\\"\\\"\\n\\nMAX_LENGTH_CLASS_NAME = 40\\n\\\"\\\"\\\"The maximum length of the class name.\\\"\\\"\\\"\\n\\nMAX_LENGTH_ENTITY_TYPE = 2\\n\\\"\\\"\\\"The maximum length of the entity type string\\\"\\\"\\\"\\n\\nMAX_LENGTH_RELATIONSHIP_TYPE = 5\\n\\\"\\\"\\\"The maximum length of the relationship type string\\\"\\\"\\\"\\n\\nMAX_LENGTH_ID = 50\\n\\\"\\\"\\\"The maximum length of the id in parent system\\\"\\\"\\\"\\n\\nENTITY_TYPE_SUBJECT = 'S'\\n\\\"\\\"\\\"The subject entity type\\\"\\\"\\\"\\n\\nENTITY_TYPE_OBJECT = 'O'\\n\\\"\\\"\\\"The object entity type\\\"\\\"\\\"\\n\\nENTITY_TYPE_SUBJECTOBJECT = 'SO'\\n\\\"\\\"\\\"The subject object entity type\\\"\\\"\\\"\\n\\nENTITY_TYPE_CHOICES = (\\n # a subject:\\n (ENTITY_TYPE_SUBJECT, _('Subject')),\\n \\n # an object:\\n (ENTITY_TYPE_OBJECT, _('Object')),\\n \\n # when subject domain is the same as object domain\\n (ENTITY_TYPE_SUBJECTOBJECT, _('Subject == Object')),\\n)\\n\\\"\\\"\\\"Choices for the entity_type field\\\"\\\"\\\"\\n\\nRELATIONSHIP_TYPE_SEPARATOR = '-'\\n\\nRELATIONSHIP_TYPE_SUBJECT_OBJECT = \\\\\\n ENTITY_TYPE_SUBJECT + RELATIONSHIP_TYPE_SEPARATOR + ENTITY_TYPE_OBJECT\\n\\\"\\\"\\\"Subject-object relationship type\\\"\\\"\\\"\\n\\nRELATIONSHIP_TYPE_SUBJECT_SUBJECT = \\\\\\n ENTITY_TYPE_SUBJECT + RELATIONSHIP_TYPE_SEPARATOR + ENTITY_TYPE_SUBJECT \\n\\\"\\\"\\\"Subject-subject relationship type\\\"\\\"\\\"\\n\\nRELATIONSHIP_TYPE_OBJECT_OBJECT = \\\\\\n ENTITY_TYPE_OBJECT + RELATIONSHIP_TYPE_SEPARATOR + ENTITY_TYPE_OBJECT\\n\\\"\\\"\\\"Object-object relatioship type\\\"\\\"\\\"\\n\\nRELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT = \\\\\\n ENTITY_TYPE_SUBJECTOBJECT + RELATIONSHIP_TYPE_SEPARATOR + ENTITY_TYPE_SUBJECTOBJECT\\n\\\"\\\"\\\"Subjectobject-subjectobject relatioship type\\\"\\\"\\\"\\n\\nRELATIONSHIP_TYPE_CHOICES = (\\n # a subject-object relationship\\n (RELATIONSHIP_TYPE_SUBJECT_OBJECT, _('Subejct-Object')),\\n \\n # a subject-subject relationship\\n (RELATIONSHIP_TYPE_SUBJECT_SUBJECT, _('Subject-Subject')),\\n \\n # an object-object relationship\\n (RELATIONSHIP_TYPE_OBJECT_OBJECT, _('Object-Object')),\\n \\n # a relationship for recommender where subject domain equals object domain\\n (RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT, _('SubjectObject-SubjectObject')),\\n)\\n\\\"\\\"\\\"Choices for the relationship_type field\\\"\\\"\\\"\\n\\nFORMAT_STR_SUBJECT = \\\"subject\\\"\\nFORMAT_STR_OBJECT = \\\"object\\\"\\nFORMAT_STR_SUBJECT1 = \\\"subject1\\\"\\nFORMAT_STR_SUBJECT2 = \\\"subject2\\\"\\nFORMAT_STR_OBJECT1 = \\\"object1\\\"\\nFORMAT_STR_OBJECT2 = \\\"object2\\\"\\nFORMAT_STR_SUBJECTOBJECT = 'subjectobject'\\nFORMAT_STR_SUBJECTOBJECT1 = \\\"subjectobject1\\\"\\nFORMAT_STR_SUBJECTOBJECT2 = \\\"subjectobject2\\\"\\nFORMAT_STR_CLUSTER = \\\"cluster\\\"\\n\\\"\\\"\\\"Format strings used in the explanation field.\\\"\\\"\\\"\\n\\nCOMPILATOR_DEPTH_ONE_UNSURE = 1\\n\\\"\\\"\\\"One unsure relationship is used [+ predicted_relationship]\\\"\\\"\\\"\\n\\nCOMPILATOR_DEPTH_COMING_SOON = 2\\n\\\"\\\"\\\"Uvidime\\\"\\\"\\\"\\n\\nDEFAULT_RECOMMENDATION_COUNT = 10\\n\\\"\\\"\\\"The defaul count of the obtained recommended objects\\\"\\\"\\\"\\n\\nDEFAULT_COMPILATOR_BREADTH = 10\\n\\\"\\\"\\\"The default neighbourhood size for the compilator\\\"\\\"\\\"\\n\\nDEFAULT_COMPILATOR_PAIR_DEPTH = 10\\n\\\"\\\"\\\"The default compilation element count taken for a pair for a group\\\"\\\"\\\"\\n\\nDEFAULT_COMPILATOR_DEPTH = COMPILATOR_DEPTH_ONE_UNSURE\\n\\\"\\\"\\\"Take only one unsure relationship\\\"\\\"\\\"\\n\\nUNCERTAIN_PREDICTION_VALUE = 0.5\\n\\\"\\\"\\\"The value that is returned when the prediction for the pair isn't \\navailable\\\"\\\"\\\"\\n\\nALREADY_IN_REL_PREDICTION_VALUE = 1.0\\n\\\"\\\"\\\"The value that is returned when the pair already is in the \\npredicted_relationship\\\"\\\"\\\"\\n\\nMIN_WEIGHT = 0.0\\nMAX_WEIGHT = 1.0\\n\\\"\\\"\\\"Weight limits\\\"\\\"\\\"\\n\\nMIN_CONFIDENCE = 0.0\\nMAX_CONFIDENCE = 1.0\\n\\\"\\\"\\\"Confidence limits\\\"\\\"\\\"\\n\\nMIN_EXPECTANCY = 0.0\\nMAX_EXPECTANCY = 1.0\\n\\\"\\\"\\\"Expectancy limits\\\"\\\"\\\"\\n\\nTRIVIAL_EXPECTANCY = 1.0\\n\\\"\\\"\\\"The expectancy that is given for the trivial cases - already \\nlistened to the artist.\\\"\\\"\\\"\\n\\nCONFIDENCE_KWARG_NAME = 'confidence'\\nEXPECTANCY_KWARG_NAME = 'expectancy'\\n\\nREASON_STR = \\\"Reason %d\\\"\\n\\nLOG_RECOMMENDATIONS_FILENAME = 'recommendations.txt'\\nLOG_PREDICTIONS_FILENAME = 'predictions.txt'\\nLOG_HITS_FILENAME = 'hits.txt'\\n\\nMAX_REASONS_DESCRIPTION = 5\\n\\nMORE_REASONS_STR = \\\"... and %d other reasons.\\\"\\n\\nPROMISING_RATE = 1.2\\n\\\"\\\"\\\"The rate by which the compiler breadth is multiplied to get the number\\nof promising objects to be inspected\\\"\\\"\\\"\\n\\nEXP_PRECISION = 0.00001\\n\"\n },\n {\n \"alpha_fraction\": 0.6530148983001709,\n \"alphanum_fraction\": 0.6531229615211487,\n \"avg_line_length\": 32.28776931762695,\n \"blob_id\": \"18f51e08a7555f5c74b5e0faa02ff606f12983dc\",\n \"content_id\": \"fd528e0cb2670f8ca1e5c3a9e6767d53e99cefe4\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 9254,\n \"license_type\": \"no_license\",\n \"max_line_length\": 126,\n \"num_lines\": 278,\n \"path\": \"/code/adapter/unresyst/combinator/combination_element.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The elements that are given to combinators to combine them. \\nInstancies of the classes represent knowledge that we have about a pair\\nto count similarity/preference.\\\"\\\"\\\"\\n\\nfrom unresyst.constants import *\\n\\ndef _get_expectancy_positiveness(expectancy):\\n return expectancy > UNCERTAIN_PREDICTION_VALUE\\n\\nclass BaseCombinationElement(object):\\n \\\"\\\"\\\"The base for all elements that are being combined in combinator\\\"\\\"\\\"\\n \\n def __init__(self):\\n \\\"\\\"\\\"Initialize the members\\\"\\\"\\\"\\n self._positiveness = None\\n self._expectancy = None\\n self._description = None\\n \\n def get_positiveness(self):\\n \\\"\\\"\\\"Is the element positive/negative to the combination?\\\"\\\"\\\"\\n if self._positiveness is None:\\n self._positiveness = self._get_positiveness()\\n \\n return self._positiveness\\n \\n def get_expectancy(self):\\n \\\"\\\"\\\"The expectancy of the element\\\"\\\"\\\"\\n \\n if self._expectancy is None:\\n self._expectancy = self._get_expectancy()\\n \\n return self._expectancy\\n \\n def get_description(self):\\n \\\"\\\"\\\"The description of the element\\\"\\\"\\\" \\n \\n if self._description is None:\\n self._description = self._get_description()\\n \\n return self._description\\n\\n def _get_positiveness(self):\\n pass\\n\\n def _get_expectancy(self):\\n pass\\n \\n def _get_description(self):\\n pass\\n \\n def __repr__(self):\\n return \\\"<%f, %s, %s>\\\" % (self.get_expectancy(), self.get_positiveness(), self.get_description()) \\n\\nclass SubjectObjectRelCombinationElement(BaseCombinationElement):\\n \\\"\\\"\\\"The relationship between a subject and an object meaning the preference\\n in means of the predicted_relationship. \\n \\n For compilator. \\n \\\"\\\"\\\"\\n \\n def __init__(self, rel_instance):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n super(SubjectObjectRelCombinationElement, self).__init__()\\n \\n self.rel_instance = rel_instance\\n\\n def _get_positiveness(self):\\n return self.rel_instance.definition.as_leaf_class().is_positive\\n\\n def _get_expectancy(self):\\n return self.rel_instance.get_expectancy() \\n \\n def _get_description(self):\\n return self.rel_instance.description \\n \\n \\nclass _SimilarityCombinationElement(BaseCombinationElement):\\n \\\"\\\"\\\"The element of a similarity combination\\\"\\\"\\\"\\n pass\\n \\nclass RelSimilarityCombinationElement(_SimilarityCombinationElement):\\n \\\"\\\"\\\"The similarity coming out of a rule/relationship\\n \\n For Aggregator.\\n \\\"\\\"\\\"\\n \\n def __init__(self, rel_instance):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n super(RelSimilarityCombinationElement, self).__init__()\\n \\n self.rel_instance = rel_instance\\n \\\"\\\"\\\"The rule/relationship instance the similarity is obtained from\\n @type rel_instance: RelationshipInstance\\n \\\"\\\"\\\"\\n \\n def _get_positiveness(self):\\n return self.rel_instance.definition.as_leaf_class().is_positive\\n\\n def _get_expectancy(self):\\n return self.rel_instance.get_expectancy() \\n \\n def _get_description(self):\\n return self.rel_instance.description \\n\\n\\nclass ClusterSimilarityCombinationElement(_SimilarityCombinationElement):\\n \\\"\\\"\\\"The similarity coming out of a common cluster membership\\n\\n For combinator and aggregator.\\n \\\"\\\"\\\"\\n\\n def __init__(self, cluster_members):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n super(ClusterSimilarityCombinationElement, self).__init__()\\n \\n self.cluster_members = cluster_members\\n \\\"\\\"\\\"The pair of cluster members that caused the similarity\\n @type cluster_members: pair ClusterMember, ClusterMember\\n \\\"\\\"\\\"\\n \\n def _get_positiveness(self):\\n \\\"\\\"\\\"Cluster membership is always positive\\\"\\\"\\\"\\n return True\\n \\n def _get_expectancy(self):\\n \\\"\\\"\\\"Return the product of the reasoning members\\\"\\\"\\\"\\n \\n return self.cluster_members[0].get_pair_expectancy(cluster_member_pair=self.cluster_members)\\n \\n def _get_description(self):\\n \\\"\\\"\\\"Return the concatenaged member descriptions\\\"\\\"\\\"\\n return \\\" \\\".join(cm.description for cm in self.cluster_members)\\n \\nclass BiasCombinationElement(BaseCombinationElement):\\n \\\"\\\"\\\"The biase of a subject/object coming to the combination\\n For aggregator.\\n \\\"\\\"\\\"\\n \\n def __init__(self, bias_instance):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n super(BiasCombinationElement, self).__init__()\\n \\n self.bias_instance = bias_instance\\n \\\"\\\"\\\"Bias instance. \\n @type bias_instance: BiasInstance\\n \\\"\\\"\\\" \\n\\n\\n def _get_positiveness(self):\\n return self.bias_instance.definition.is_positive\\n \\n def _get_expectancy(self):\\n return self.bias_instance.get_expectancy()\\n\\n def _get_description(self):\\n return self.bias_instance.description\\n\\nclass BiasAggregateCombinationElement(BaseCombinationElement):\\n \\\"\\\"\\\"The aggregated bias of a subject/object. \\n For compilator.\\n \\\"\\\"\\\"\\n \\n def __init__(self, bias_aggregate):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n super(BiasAggregateCombinationElement, self).__init__()\\n \\n self.bias_aggregate = bias_aggregate\\n \\\"\\\"\\\"The aggregated bias.\\n @type bias_aggregate: AggregatedBiasInstance\\n \\\"\\\"\\\"\\n \\n def _get_positiveness(self): \\n exp = self.get_expectancy()\\n return _get_expectancy_positiveness(exp)\\n \\n def _get_expectancy(self):\\n return self.bias_aggregate.expectancy\\n \\n def _get_description(self):\\n return self.bias_aggregate.description\\n\\n\\nclass _PredictedPlusSimilarityCombinationElement(BaseCombinationElement):\\n \\\"\\\"\\\"A sign of preference made of known predicted relationship plus similarity.\\n For compilator.\\n \\\"\\\"\\\"\\n \\n def __init__(self, predicted_rel, similarity_aggregate):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n super(_PredictedPlusSimilarityCombinationElement, self).__init__()\\n \\n self.predicted_rel = predicted_rel\\n \\\"\\\"\\\"The predicted relationship to/from some other object/subject\\n @type predicted_rel: RelationshipInstance binded to PredictedRelationshipDefinition\\n \\\"\\\"\\\"\\n \\n self.similarity_aggregate = similarity_aggregate\\n \\\"\\\"\\\"The aggregated similarity between the subjects/objects\\n @type similarity_aggregate: AggregatedRelationshipInstance\\n \\\"\\\"\\\"\\n\\n def _get_positiveness(self): \\n exp = self.get_expectancy()\\n return _get_expectancy_positiveness(exp)\\n \\n def _get_expectancy(self):\\n return self.similarity_aggregate.expectancy\\n \\n\\n\\nclass PredictedPlusObjectSimilarityCombinationElement(_PredictedPlusSimilarityCombinationElement):\\n \\\"\\\"\\\"Predicted relationship plus similarity of objects.\\n For compilator.\\n \\\"\\\"\\\"\\n def _get_description(self):\\n return \\\"%s And similarity: %s\\\" % (self.predicted_rel.description, self.similarity_aggregate.description)\\n\\nclass PredictedPlusSubjectSimilarityCombinationElement(_PredictedPlusSimilarityCombinationElement):\\n \\\"\\\"\\\"Predicted relationship plus similarity of subjects.\\n For compilator.\\n \\\"\\\"\\\"\\n def _get_description(self):\\n return \\\"Similarity: %s And: %s\\\" % (self.similarity_aggregate.description, self.predicted_rel.description)\\n\\n\\nclass _PredictedPlusClusterMemberCombinationElement(BaseCombinationElement):\\n \\\"\\\"\\\"A sign of preference made of known predicted relationship plus similarity\\n coming out of two cluster memberships.\\n\\n For compilator.\\n \\\"\\\"\\\"\\n\\n def __init__(self, predicted_rel, cluster_combination_element):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n super(_PredictedPlusClusterMemberCombinationElement, self).__init__()\\n \\n self.predicted_rel = predicted_rel\\n \\\"\\\"\\\"The predicted relationship to/from some other object/subject\\n @type predicted_rel: RelationshipInstance binded to PredictedRelationshipDefinition\\n \\\"\\\"\\\"\\n \\n self.cluster_combination_element = cluster_combination_element\\n \\\"\\\"\\\"The pair of cluster memeberships meaning the similarity. In the order as if coming\\n from the subject to the object.\\n @type cluster_membership_pair: ClusterSimilarityCombinationElement\\n \\\"\\\"\\\" \\n \\n \\n def _get_positiveness(self): \\n return True\\n \\n def _get_expectancy(self):\\n return self.cluster_combination_element.get_expectancy()\\n \\n \\n\\nclass PredictedPlusObjectClusterMemberCombinationElement(_PredictedPlusClusterMemberCombinationElement):\\n \\\"\\\"\\\"Predicted relationship plus cluster membership of objects.\\n\\n For compilator.\\n \\\"\\\"\\\"\\n def _get_description(self):\\n return \\\"%s And similarity: %s\\\" % (self.predicted_rel.description, self.cluster_combination_element.get_description())\\n\\nclass PredictedPlusSubjectClusterMemberCombinationElement(_PredictedPlusClusterMemberCombinationElement):\\n \\\"\\\"\\\"Predicted relationship plus cluster membership of subjects.\\n\\n For compilator.\\n \\\"\\\"\\\"\\n def _get_description(self):\\n return \\\"Similarity: %s And: %s\\\" % (self.cluster_combination_element.get_description(), self.predicted_rel.description)\\n\"\n },\n {\n \"alpha_fraction\": 0.6421499252319336,\n \"alphanum_fraction\": 0.6584158539772034,\n \"avg_line_length\": 23.379310607910156,\n \"blob_id\": \"41a1d8c02a6c91077ddaa1edff6bc895d5cd9457\",\n \"content_id\": \"56b5bc0bc1ec5156785f4f3dbcdcdd87f27ddaca\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1414,\n \"license_type\": \"no_license\",\n \"max_line_length\": 63,\n \"num_lines\": 58,\n \"path\": \"/UnresystCD/code/adapter/lastfm/evaluation.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The evaluators for the last.fm recommenders\\\"\\\"\\\"\\n\\nfrom unresyst.recommender.evaluation import BaseEvaluator\\nfrom unresyst.recommender.rank_evaluation import RankEvaluator\\nfrom models import ArtistEvalPair, NovelArtistEvalPair\\nfrom unresyst.recommender.metrics import rmse, precision_recall\\n\\nclass ArtistRecommenderEvaluator(BaseEvaluator):\\n \\\"\\\"\\\"The evaluator of the artist recommender\\\"\\\"\\\"\\n \\n EvaluationPairModel = ArtistEvalPair\\n \\\"\\\"\\\"The model - pairs\\\"\\\"\\\"\\n \\n prediction_metric = rmse\\n \\\"\\\"\\\"The metric\\\"\\\"\\\"\\n \\n recommendation_metric = precision_recall\\n \\\"\\\"\\\"The other metric\\\"\\\"\\\"\\n \\nclass ArtistRankEvaluator(RankEvaluator):\\n \\\"\\\"\\\"Evaluating the rank metric\\\"\\\"\\\"\\n \\n EvaluationPairModel = ArtistEvalPair\\n \\\"\\\"\\\"The model - pairs\\\"\\\"\\\"\\n \\n SUBJ_IDS = [\\n 6L,\\n 11L,\\n 14L,\\n 36L,\\n 38L,\\n 53L,\\n 55L,\\n 59L,\\n 61L,\\n 79L,\\n 81L,\\n 90L,\\n ]\\n\\n \\nclass NovelArtistRecommenderEvaluator(BaseEvaluator):\\n \\\"\\\"\\\"The evaluator of the novel artist recommender\\\"\\\"\\\"\\n \\n EvaluationPairModel = NovelArtistEvalPair\\n \\n prediction_metric = rmse\\n \\n recommendation_metric = precision_recall \\n\\n \\nclass NovelArtistRankEvaluator(RankEvaluator):\\n \\\"\\\"\\\"Evaluation of the rank metric\\\"\\\"\\\"\\n \\n EvaluationPairModel = NovelArtistEvalPair\\n \\\"\\\"\\\"The model - pairs\\\"\\\"\\\"\\n \\n SUBJ_IDS = ArtistRankEvaluator.SUBJ_IDS\\n\"\n },\n {\n \"alpha_fraction\": 0.6785714030265808,\n \"alphanum_fraction\": 0.680672287940979,\n \"avg_line_length\": 33,\n \"blob_id\": \"00c30c4627da8a8eb34e37fc4c92ac1e3c5d7d6e\",\n \"content_id\": \"9f3810ae018f50fc898187da60be831f6161be69\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 476,\n \"license_type\": \"no_license\",\n \"max_line_length\": 92,\n \"num_lines\": 14,\n \"path\": \"/code/adapter/unresyst/combinator/function_combinator.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Combinator using the magic function.\\\"\\\"\\\"\\n\\nfrom base import BaseCombinator\\n\\nclass FunctionCombinator(BaseCombinator):\\n \\\"\\\"\\\"A combinator using a special function to combine.\\n \\\"\\\"\\\"\\n \\n def _combine(self, combination_elements, ResultClass):\\n \\\"\\\"\\\"See the base class for documentation\\\"\\\"\\\"\\n \\n # narvat to do magicke funkce a hotovka (spocitat prumer a pocet positive, negative)\\n res = ResultClass(expectancy=1, description='')\\n return res\\n\"\n },\n {\n \"alpha_fraction\": 0.48442137241363525,\n \"alphanum_fraction\": 0.48961424827575684,\n \"avg_line_length\": 33.89610290527344,\n \"blob_id\": \"ba5938caabc8a4048c33761bc89052173ba4d996\",\n \"content_id\": \"f3491924fcd8f3deade3395e3f275f7523756234\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2696,\n \"license_type\": \"no_license\",\n \"max_line_length\": 142,\n \"num_lines\": 77,\n \"path\": \"/UnresystCD/code/adapter/flixster/recbckp.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \" relationships = (\\n # users in friendship\\n SubjectSimilarityRelationship(\\n name=\\\"Users are friends.\\\",\\n \\n generator=lambda: [(f.friend1, f.friend2) for f in Friend.objects.all()], \\n \\n is_positive=True, \\n \\n weight=0.5, \\n \\n description=\\\"Users %(subject1)s and %(subject2)s are friends.\\\",\\n ),\\n )\\n \\\"\\\"\\\"The relationships\\\"\\\"\\\"\\n \\n biases = (\\n # people giving high ratings\\n SubjectBias(\\n name=\\\"Users giving high ratings.\\\",\\n \\n description=\\\"User %(subject)s gives high ratings.\\\",\\n \\n weight=0.5, \\n \\n is_positive=True,\\n \\n generator=lambda: User.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__gt=str(MIN_HIGH_RATING)), \\n \\n confidence=lambda user: user.rating_set.aggregate(Avg('rating'))['rating__avg'] - MIN_HIGH_RATING\\n ),\\n \\n # highly rated movies\\n ObjectBias(\\n name=\\\"High-rated movies.\\\",\\n \\n description=\\\"Movie %(object)s is high-rated\\\",\\n \\n weight=0.5,\\n \\n is_positive=True,\\n \\n generator=lambda: Movie.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__gt=str(MIN_HIGH_RATING)),\\n \\n confidence=lambda movie: movie.rating_set.aggregate(Avg('rating'))['rating__avg'] - MIN_HIGH_RATING\\n ),\\n \\n # people giving low ratings\\n SubjectBias(\\n name=\\\"Users giving low ratings.\\\",\\n \\n description=\\\"User %(subject)s gives low ratings.\\\",\\n \\n weight=0.5, \\n \\n is_positive=False,\\n \\n generator=lambda: User.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__lt=str(MAX_LOW_RATING)), \\n \\n confidence=lambda user: MAX_LOW_RATING - user.rating_set.aggregate(Avg('rating'))['rating__avg'] \\n ),\\n \\n # low-rated movies\\n ObjectBias(\\n name=\\\"Low-rated movies.\\\",\\n \\n description=\\\"Movie %(object)s is low-rated\\\",\\n \\n weight=0.5,\\n \\n is_positive=False,\\n \\n generator=lambda: Movie.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__lt=str(MAX_LOW_RATING)),\\n \\n confidence=lambda movie: MAX_LOW_RATING - movie.rating_set.aggregate(Avg('rating'))['rating__avg']\\n ), \\n )\\n \\n\"\n },\n {\n \"alpha_fraction\": 0.737500011920929,\n \"alphanum_fraction\": 0.737500011920929,\n \"avg_line_length\": 25.66666603088379,\n \"blob_id\": \"c3752c1c077960e0f5b803ab01e63cf04f6a1dae\",\n \"content_id\": \"9faa7d4d0922076951d8647a2b6aad672ceadc57\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 80,\n \"license_type\": \"no_license\",\n \"max_line_length\": 54,\n \"num_lines\": 3,\n \"path\": \"/UnresystCD/code/adapter/lastfm/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Application holding the data and a recommender for \\nthe Last.fm dataset.\\n\\\"\\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.5477624535560608,\n \"alphanum_fraction\": 0.5494836568832397,\n \"avg_line_length\": 34.72093200683594,\n \"blob_id\": \"3edd44a8d58f005f8feca242a252d6b64be52355\",\n \"content_id\": \"5e27c0fec155d93316bdf2840109384444a73d1f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 4648,\n \"license_type\": \"no_license\",\n \"max_line_length\": 91,\n \"num_lines\": 129,\n \"path\": \"/code/adapter/unresyst/recommender/bias.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Classes to represent bias of subjects and objects\\\"\\\"\\\"\\n\\nfrom unresyst.constants import *\\n\\nfrom unresyst.models.abstractor import BiasDefinition, BiasInstance\\nfrom unresyst.models.common import SubjectObject\\n\\nclass _BaseBias(object):\\n \\\"\\\"\\\"The base class for all bias clases\\\"\\\"\\\"\\n \\n entity_type = None\\n \\\"\\\"\\\"The entity type the bias is for s/o/so\\\"\\\"\\\"\\n \\n format_string = None\\n \\\"\\\"\\\"The format string appearing in the description\\\"\\\"\\\"\\n \\n def __init__(self, name, generator, is_positive, confidence, weight, description=None):\\n \\\"\\\"\\\"The constructor.\\\"\\\"\\\"\\n \\n self.name = name\\n \\\"\\\"\\\"The name of the bias.\\\"\\\"\\\" \\n \\n self.description = description\\n \\\"\\\"\\\"A string describing the rule. It can contain placeholders for entities: \\n \\n - %(subject)s for subject bias\\n - %(object)s for object bias\\n - %(subjectobject)s for subjectobject bias\\n \\\"\\\"\\\"\\n \\n self.is_positive = is_positive\\n \\\"\\\"\\\"Is the bias positive to the predicted relationship?\\\"\\\"\\\"\\n \\n self.generator = generator\\n \\\"\\\"\\\"A generator returning subjects/objects that are affected by the bias.\\n \\\"\\\"\\\" \\n \\n self.weight = weight\\n \\\"\\\"\\\"A float number from [0, 1] representing the *static* weight of the bias\\n It doesn't depend on the entity.\\n \\\"\\\"\\\"\\n \\n self.confidence = confidence\\n \\\"\\\"\\\"A float function giving values from [0, 1] representing the \\n the confidence of the bias of the entity. \\n It's dynamic, depends on the entity.\\n \\\"\\\"\\\" \\n \\n def evaluate(self):\\n \\\"\\\"\\\"Crate bias definitions and the instances in the database.\\n \\\"\\\"\\\"\\n \\n if not (MIN_WEIGHT <= self.weight <= MAX_WEIGHT):\\n raise ConfigurationError(\\n message=(\\\"The bias '%s' provides weight %f,\\\" + \\n \\\" should be between 0 and 1. .\\\"\\n ) % (self.name, self.weight),\\n recommender=self.recommender,\\n parameter_name=\\\"Recommender.biases\\\",\\n parameter_value=(self.recommender.biases)\\n )\\n \\n recommender_model = self.recommender._get_recommender_model()\\n\\n # create the definition in the database\\n definition = BiasDefinition.objects.create(\\n name=self.name,\\n recommender=recommender_model,\\n entity_type=self.entity_type,\\n weight=self.weight,\\n is_positive=self.is_positive\\n ) \\n \\n # go through the affected entities create bias instances\\n # \\n for ds_entity in self.generator():\\n \\n # convert the entity to universal\\n dn_entity = SubjectObject.get_domain_neutral_entity(\\n domain_specific_entity=ds_entity, \\n entity_type=self.entity_type, \\n recommender=recommender_model)\\n \\n # count the confidence by the provided function \\n confidence = self.confidence(ds_entity) \\n \\n # if confidence invalid through an error\\n if not (MIN_CONFIDENCE <= confidence <= MAX_CONFIDENCE):\\n raise ConfigurationError(\\n message=(\\\"The bias '%s' provides confidence %f,\\\" + \\n \\\" should be between 0 and 1. \\\"\\n ) % (self.name, confidence),\\n recommender=self.recommender,\\n parameter_name=\\\"Recommender.biases\\\",\\n parameter_value=(self.recommender.biases)\\n )\\n \\n # fill the description\\n description = self.description % {self.format_string: dn_entity.name}\\n \\n # create the instance\\n BiasInstance.objects.create(\\n subject_object=dn_entity,\\n confidence=confidence,\\n definition=definition,\\n description=description\\n )\\n \\n print \\\" %d bias instances for bias %s created.\\\" % \\\\\\n (BiasInstance.objects.filter(definition=definition).count(), self.name)\\n \\nclass SubjectBias(_BaseBias):\\n\\n entity_type = ENTITY_TYPE_SUBJECT\\n \\n format_string = FORMAT_STR_SUBJECT\\n\\n \\nclass ObjectBias(_BaseBias):\\n\\n entity_type = ENTITY_TYPE_OBJECT\\n \\n format_string = FORMAT_STR_OBJECT\\n\\nclass SubjectObjectBias(_BaseBias):\\n\\n entity_type = ENTITY_TYPE_SUBJECTOBJECT\\n \\n format_string = FORMAT_STR_SUBJECTOBJECT\\n \\n \\n \\n \\n \\n \\n \\n \\n\"\n },\n {\n \"alpha_fraction\": 0.7523809671401978,\n \"alphanum_fraction\": 0.7523809671401978,\n \"avg_line_length\": 34,\n \"blob_id\": \"a6e704907d6452c58a8ebf2456f216f50160faf4\",\n \"content_id\": \"52b390f506337b8b7fef12287b254a20f114033b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 105,\n \"license_type\": \"no_license\",\n \"max_line_length\": 65,\n \"num_lines\": 3,\n \"path\": \"/code/adapter/demo/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The demo application showing the possibilities of Unresyst on \\na simple example of a shoe e-shop.\\n\\\"\\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.5544632077217102,\n \"alphanum_fraction\": 0.5560361742973328,\n \"avg_line_length\": 30.33333396911621,\n \"blob_id\": \"5ebfcc473dcb3e61526b2323283908dc81c385ad\",\n \"content_id\": \"e9e5c5a3660dc082cb5cf53b0b46bc019c456c48\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2543,\n \"license_type\": \"no_license\",\n \"max_line_length\": 91,\n \"num_lines\": 81,\n \"path\": \"/code/adapter/unresyst/models/evaluation.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Models related to recommender evaluation\\\"\\\"\\\"\\n\\nfrom django.db import models\\n\\nclass BaseEvaluationPair(models.Model):\\n \\\"\\\"\\\"An abstract base class for a test set pair. To be used for\\n testing the correctness of the predicted expectancy.\\n \\n The subclass has to have attributes:\\n - subj - a foreign key to subject\\n - obj - a foreign key to object \\n \\\"\\\"\\\" \\n \\n obtained_expectancy = models.FloatField(null=True)\\n \\\"\\\"\\\"The predicted expectancy obtained by the predict_relationship recommender\\n method.\\n \\\"\\\"\\\"\\n \\n expected_expectancy = models.FloatField()\\n \\\"\\\"\\\"The expected expectancy value for the pair\\\"\\\"\\\"\\n \\n is_successful = models.NullBooleanField(null=True)\\n \\\"\\\"\\\"A sign whether the pair prediction was successful\\\"\\\"\\\"\\n \\n class Meta:\\n abstract = True\\n app_label = 'unresyst'\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a textual representation.\\\"\\\"\\\" \\n \\n return u\\\"%s - %s: exp: %f, got: %f\\\" % \\\\\\n (self.subj, self.obj, self.expected_expectancy, self.obtained_expectancy or -1)\\n \\n @classmethod\\n def export(cls, f):\\n \\\"\\\"\\\"Export evaluation pairs to a csv file of the given name.\\n \\n @type f: file\\n @param f: open file to write to \\n \\\"\\\"\\\" \\n i = 0\\n f.write(\\\"# userId, itemId\\\\n\\\")\\n \\n # loop through the pairs, \\n for subj_id, obj_id in cls.objects.values_list('subj__pk', 'obj__pk'):\\n \\n # create the line\\n linestr = \\\"%s,%s\\\\n\\\" % (subj_id, obj_id) \\n \\n # write it to the file\\n f.write(linestr)\\n \\n i += 1\\n \\n print \\\" %d evaluation pairs exported\\\" % i\\n \\n # to be implemented by subclasses\\n # \\n \\n @classmethod\\n def select(cls, i=0):\\n \\\"\\\"\\\"Select the pairs for a validation, save them to the database and\\n remove them from the system data. The method is called in every \\n iteration.\\n \\n To be implemented by the subclass.\\n\\n @type i: int\\n @param i: the number of iteration of the validation (useful when\\n cross-validation is done)\\n \\\"\\\"\\\"\\n raise NotImplementedError() \\n\\n def get_prediction_success(self):\\n \\\"\\\"\\\"Count whether the pair was successful. The obtained expectancy has\\n to be filled first.\\n \\n To be implemented by the subclass.\\n \\\"\\\"\\\" \\n raise NotImplementedError()\\n \\n\"\n },\n {\n \"alpha_fraction\": 0.5600860714912415,\n \"alphanum_fraction\": 0.5629558563232422,\n \"avg_line_length\": 39.66829299926758,\n \"blob_id\": \"3444088d69f1d0c78a17e8b5561cba93a6144850\",\n \"content_id\": \"a1a2fb041241abef63a71ce3ab40ccce1c4753de\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 8363,\n \"license_type\": \"no_license\",\n \"max_line_length\": 96,\n \"num_lines\": 205,\n \"path\": \"/code/adapter/unresyst/aggregator/linear_aggregator.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The classes of the aggregator package:\\n - LinearAggregator: everything is combined linearly\\n - CombiningAggregator: uses the Combinator class for aggregating\\n\\\"\\\"\\\"\\n\\nfrom django.db.models import Count\\nfrom django.db.models import Q\\n\\n\\nfrom base import BaseAggregator\\nfrom unresyst.models.abstractor import RelationshipInstance, \\\\\\n PredictedRelationshipDefinition\\nfrom unresyst.models.aggregator import AggregatedRelationshipInstance, \\\\\\n AggregatedBiasInstance\\nfrom unresyst.models.common import SubjectObject\\nfrom unresyst.exceptions import InvalidParameterError\\n\\nclass LinearAggregator(BaseAggregator):\\n \\\"\\\"\\\"The class aggregating rule/relationship instances to one for each pair\\n of entities. \\n \\n It ignores the predicted_relationship instances.\\n \\n A better aggregator could add the positive and subtract the negative \\n expectances somehow.\\n \\\"\\\"\\\"\\n\\n def aggregate_rules_relationships(self, recommender_model):\\n \\\"\\\"\\\"For documentation see the base class.\\n \\n Linearly combines the rule and relationship instances.\\n \\n The descriptions of the aggregates are made by joining the descriptions\\n of the aggregated instances. The descriptions are ordered by the expectancy\\n of their owners, the highest expectancy comes first.\\n \\\"\\\"\\\"\\n \\n # if there's something in the database for the recommender\\n # throw an error\\n if AggregatedRelationshipInstance.objects\\\\\\n .filter(recommender=recommender_model).exists():\\n \\n raise InvalidParameterError(\\n message=\\\"There're unexpected aggregated instances for the recommender.\\\", \\n recommender=recomender_model,\\n parameter_name=\\\"recommender_model\\\", \\n parameter_value=recommender_model)\\n\\n # aggregate it\\n # \\n \\n # take all rule/relationship instances, that don't belong \\n # to the predicted_relationship\\n # order them by the first and the second\\n predicted_def = PredictedRelationshipDefinition.objects.get(\\n recommender=recommender_model)\\n instance_qs = RelationshipInstance.objects\\\\\\n .exclude(definition=predicted_def)\\\\\\n .filter(definition__recommender=recommender_model)\\\\\\n .order_by('subject_object1__id', 'subject_object2__id')\\n \\n # if there's nothing to aggregate, schluss\\n if not instance_qs:\\n return\\n \\n first_inst = instance_qs[0]\\n \\n # continuously built aggregated instance\\n # initialize it wit the first instance\\n cont_inst = AggregatedRelationshipInstance(\\n subject_object1=first_inst.subject_object1,\\n subject_object2=first_inst.subject_object2, \\n relationship_type=first_inst.definition.as_leaf_class().relationship_type,\\n recommender=recommender_model)\\n\\n exp = first_inst.get_expectancy()\\n exp_sum = exp\\n count = 1 \\n # a list of pairs (expectancy, description) \\n desc_list = [(exp, first_inst.description), ]\\n \\n # go through the rel instances \\n for instance in instance_qs.exclude(pk=first_inst.pk).iterator(): \\n \\n # if the pair has chnged from the last pair, save what we've got and \\n # start anew\\n if cont_inst.subject_object1 <> instance.subject_object1 \\\\\\n or cont_inst.subject_object2 <> instance.subject_object2:\\n \\n # count the average expectancy\\n cont_inst.expectancy = float(exp_sum) / count\\n \\n # sort the description list by expectancy and join it\\n desc_list.sort(key=lambda pair: pair[0], reverse=True) \\n cont_inst.description = ' '.join([desc for x, desc in desc_list])\\n\\n # save the current instance\\n cont_inst.save()\\n \\n # start a new continuously aggregated instance\\n cont_inst = AggregatedRelationshipInstance(\\n subject_object1=instance.subject_object1,\\n subject_object2=instance.subject_object2, \\n relationship_type=instance.definition.as_leaf_class().relationship_type,\\n recommender=recommender_model)\\n \\n exp = instance.get_expectancy()\\n exp_sum = exp\\n count = 1 \\n \\n # a list of pairs (expectancy, description) \\n desc_list = [(exp, instance.description), ]\\n \\n # otherwise aggregate \\n else: \\n exp = instance.get_expectancy() \\n exp_sum += exp\\n count += 1\\n \\n desc_list.append((exp, instance.description))\\n \\n \\n # count and save the last we have\\n # \\n \\n # count the average expectancy\\n cont_inst.expectancy = float(exp_sum) / count\\n \\n # sort the description list by expectancy and join it\\n desc_list.sort(key=lambda pair: pair[0], reverse=True) \\n cont_inst.description = ' '.join([desc for x, desc in desc_list])\\n\\n # save the last instance\\n cont_inst.save()\\n \\n print \\\" %d rule/relationship aggregates created\\\" % \\\\\\n AggregatedRelationshipInstance.objects.filter(recommender=recommender_model).count()\\n \\n\\n def aggregate_biases(self, recommender_model):\\n \\\"\\\"\\\"For documentation see the base class.\\n \\n Linearly combines the bias instances.\\n \\n The descriptions of the aggregates are made by joining the descriptions\\n of the aggregated instances. The descriptions are ordered by the expectancy\\n of their owners, the highest expectancy comes first.\\n \\\"\\\"\\\"\\n \\n # if there's something in the database for the recommender\\n # throw an error\\n if AggregatedBiasInstance.objects\\\\\\n .filter(recommender=recommender_model).exists():\\n \\n raise InvalidParameterError(\\n message=\\\"There're unexpected aggregated instances for the recommender.\\\", \\n recommender=recommender_model,\\n parameter_name=\\\"recommender_model\\\", \\n parameter_value=recommender_model)\\n\\n # aggregate it\\n # \\n \\n # take all subjectobjects that have some biases\\n qs_biased_so = SubjectObject.objects\\\\\\n .filter(recommender=recommender_model)\\\\\\n .annotate(num_bias=Count('biasinstance'))\\\\\\n .filter(num_bias__gt=0)\\n \\n count = 0\\n \\n # go through the biased subjectobjects\\n for so in qs_biased_so.iterator(): \\n \\n # a list of pairs (expectancy, description) \\n desc_list = [] \\n \\n exp_sum = 0\\n # go through their biases\\n for bias in so.biasinstance_set.all():\\n\\n # count the expectancy for the bias\\n exp = bias.get_expectancy()\\n \\n exp_sum += exp\\n \\n desc_list.append((exp, bias.description))\\n\\n # count the average expectancy\\n avg_exp = float(exp_sum) / so.biasinstance_set.count()\\n \\n # sort the description list by expectancy and join it\\n desc_list.sort(key=lambda pair: pair[0], reverse=True) \\n desc = ' '.join([d for x, d in desc_list])\\n \\n # create and save the model\\n AggregatedBiasInstance.objects.create(\\n expectancy=avg_exp,\\n subject_object=so,\\n recommender=recommender_model,\\n description=desc\\n )\\n \\n print \\\" %d bias aggregates created\\\" % \\\\\\n AggregatedBiasInstance.objects.filter(recommender=recommender_model).count()\\n \\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5359686017036438,\n \"alphanum_fraction\": 0.5438196063041687,\n \"avg_line_length\": 32.28571319580078,\n \"blob_id\": \"f28cfdf2ed5c979ea0bbe9bd739c6a0fb1f2cf27\",\n \"content_id\": \"04fb8d402910bbd2313c1f781bc9bdc4d8f0beb4\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 10954,\n \"license_type\": \"no_license\",\n \"max_line_length\": 133,\n \"num_lines\": 329,\n \"path\": \"/UnresystCD/code/adapter/demo/recommender.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The recommender used in the parent system \\n(the unresyst.BaseRecommender subclass)\\n\\\"\\\"\\\"\\nfrom django.db.models import Count\\n\\nfrom unresyst import *\\nfrom models import * \\nfrom unresyst.algorithm import *\\nfrom unresyst.compilator import *\\nfrom unresyst.aggregator import *\\nfrom unresyst.combinator import *\\n\\n# helper functions: \\n#\\n \\ndef _get_keyword_set(o):\\n \\\"\\\"\\\"Get set of keywords for the given shoepair\\n \\\"\\\"\\\"\\n return set([kw.word for kw in o.keywords.all()]) \\n\\ndef _keyword_set_similarity(o1, o2):\\n \\\"\\\"\\\"A function measuring similarity of keywords for shoepairs o1 and o2.\\n \\n Counted as the size of the intersection divided by the size of the \\n smaller set of keywords.\\n \\\"\\\"\\\"\\n # make sets of the keywords\\n keyword_set1 = _get_keyword_set(o1)\\n keyword_set2 = _get_keyword_set(o2)\\n \\n # the intersection of the keywords\\n keyword_intersection = keyword_set1.intersection(keyword_set2)\\n \\n # size of the smaller set\\n min_len = min(len(keyword_set1), len(keyword_set2))\\n \\n # they have some keywords in common, so min_len is never 0\\n \\n #the final measure\\n return float(len(keyword_intersection))/min_len\\n\\ndef _likes_shoes_generator():\\n \\\"\\\"\\\"A generator for the user likes shoes relationship\\\"\\\"\\\" \\n for u in User.objects.iterator():\\n for s in u.likes_shoes.iterator():\\n yield (u, s)\\n\\ndef _shoelikers():\\n \\\"\\\"\\\"A generator returning users liking at least one shoepair\\\"\\\"\\\"\\n for u in User.objects.annotate(num_shoes=Count('likes_shoes')).filter(num_shoes__gte=1):\\n yield u\\n\\n\\ndef _popular_shoes():\\n \\\"\\\"\\\"A generator returning shoes that are liked by at least one user\\\"\\\"\\\"\\n for s in ShoePair.objects.annotate(num_likers=Count('likers')).filter(num_likers__gte=1):\\n yield s\\n\\ndef _viewed_shoes_generator():\\n \\\"\\\"\\\"A generator for the user has viewed shoes relationship\\\"\\\"\\\" \\n for u in User.objects.iterator():\\n for s in u.viewed_shoes.iterator():\\n yield (u, s)\\n\\ndef _south_generator():\\n \\\"\\\"\\\"A generator for the rule: if user is from south don't recommend \\n him/her winter shoes.\\n \\\"\\\"\\\"\\n for u in User.objects.filter(home_city__in_south=True).iterator():\\n for s in ShoePair.objects.filter(for_winter=True).iterator():\\n yield (u, s)\\n\\n# the recommender:\\n# \\n\\nclass ShoeRecommender(Recommender):\\n \\\"\\\"\\\"A BaseRecommender subclass holding all domain-specific data\\\"\\\"\\\"\\n\\n name = \\\"Shoe Recommender\\\"\\n \\\"\\\"\\\"The name\\\"\\\"\\\" \\n \\n subjects = User.objects\\n \\\"\\\"\\\"The objects to who the recommender will recommend.\\\"\\\"\\\"\\n \\n objects = ShoePair.objects\\n \\\"\\\"\\\"The objects that will be recommended.\\\"\\\"\\\" \\n\\n predicted_relationship = PredictedRelationship( \\n name=\\\"User likes shoes.\\\",\\n condition=lambda s, o: \\n o in s.likes_shoes.all(), \\n description=\\\"\\\"\\\"User %(subject)s likes shoes %(object)s.\\\"\\\"\\\",\\n generator=_likes_shoes_generator\\n )\\n \\\"\\\"\\\"The relationship that will be predicted\\\"\\\"\\\"\\n \\n relationships = (\\n \\n # if the user has viewed the shoes it's a sign of preference\\n SubjectObjectRelationship(\\n name=\\\"User has viewed shoes.\\\",\\n \\n condition=lambda s, o:\\n o in s.viewed_shoes.all(), \\n\\n is_positive=True, \\n \\n weight=0.4, \\n \\n description=\\\"User %(subject)s has viewed %(object)s.\\\",\\n \\n generator=_viewed_shoes_generator\\n ),\\n \\n # if the user is from the same city as the shoe manufacturer, he might like it\\n SubjectObjectRelationship(\\n name='User lives in the same city as the shoe manufacturer.',\\n \\n condition=lambda s, o:\\n o.manufacturer and s.home_city and o.manufacturer.home_city == s.home_city,\\n\\n is_positive=True, \\n \\n weight=0.1, \\n \\n description=\\\"User %(subject)s is from the same city as the manufacturer of %(object)s.\\\"\\n ),\\n \\n # if users live in the same city, they are considered similar\\n SubjectSimilarityRelationship(\\n name=\\\"Users live in the same city.\\\",\\n \\n condition=lambda s1, s2:\\n s1.home_city and s1.home_city == s2.home_city, \\n \\n is_positive=True, \\n \\n weight=0.3, \\n \\n description=\\\"Users %(subject1)s and %(subject2)s live in the same city.\\\"\\n ),\\n \\n # if shoes were made by the same manufacturer, they are considered \\n # similar\\n ObjectSimilarityRelationship(\\n name=\\\"Shoes were made by the same manufacturer.\\\",\\n \\n condition=lambda o1, o2:\\n o1.manufacturer and o1.manufacturer == o2.manufacturer,\\n \\n is_positive=True, \\n \\n weight=0.1,\\n \\n description=\\\"Shoes %(object1)s and %(object2)s were made by\\\" + \\\\\\n \\\" the same manufacturer.\\\"\\n ) \\n )\\n \\\"\\\"\\\"Relationships among the subjects and objects in the domain\\\"\\\"\\\"\\n \\n \\n rules = (\\n # if user lives on south, don't recommend him winter shoes \\n SubjectObjectRule( \\n name=\\\"Don't recommend winter shoes for southern users.\\\",\\n # is the user from a southern city and shoes for winter?\\n condition=lambda s, o: \\n s.home_city.in_south and o.for_winter, \\n \\n is_positive=False,\\n \\n weight=0.85, \\n \\n confidence=lambda s, o: 1, \\n \\n description=\\\"%(subject)s is from south, so %(object)s can't \\\" + \\n \\\"be recommended to him/her.\\\",\\n\\n generator=_south_generator\\n ),\\n \\n # if users are the same age +- year they are similar\\n SubjectSimilarityRule(\\n name=\\\"Users with similar age.\\\",\\n \\n # both users have given their age\\n condition=lambda s1, s2: \\n s1.age and s2.age and s1.age -1 <= s2.age <= s2.age + 1,\\n \\n is_positive=True, \\n \\n weight=0.2,\\n \\n # a magic linear confidence function\\n confidence=lambda s1, s2: \\n 1 - 0.25 * abs(s1.age - s2.age),\\n \\n description=\\\"Users %(subject1)s and %(subject2)s are about \\\" + \\n \\\"the same age.\\\"\\n ),\\n \\n # if shoes have common keywords, they are similar.\\n ObjectSimilarityRule(\\n name=\\\"Shoes with common keywords.\\\",\\n \\n # shoes have some common keywords, if both empty, it's false\\n condition=lambda o1, o2: \\n bool(_get_keyword_set(o1).intersection(_get_keyword_set(o2))),\\n \\n is_positive=True,\\n \\n weight=0.4,\\n \\n # the size of the intersection / the size of the smaller set\\n confidence=_keyword_set_similarity,\\n \\n description=\\\"The shoe pairs %(object1)s and %(object2)s \\\" + \\n \\\"share some keywords.\\\"\\n ),\\n \\n # explicit rating\\n ExplicitSubjectObjectRule(\\n name=\\\"Shoe rating.\\\",\\n \\n condition=None,\\n \\n description=\\\"User %(subject)s has rated %(object)s.\\\",\\n \\n # all pairs user, rated shoes\\n generator=lambda: [(r.user, r.shoe_pair) for r in ShoeRating.objects.all()],\\n \\n # the number of stars divided by five\\n expectancy=lambda s, o:float(ShoeRating.objects.get(user=s, shoe_pair=o).stars) / 5,\\n ),\\n )\\n \\\"\\\"\\\"Rules that can be applied to the domain\\\"\\\"\\\"\\n \\n cluster_sets = (\\n # shoe category cluster\\n ObjectClusterSet(\\n\\n name=\\\"Shoe category cluster set.\\\",\\n\\n weight=0.3,\\n \\n filter_entities=ShoePair.objects.filter(category__isnull=False),\\n \\n get_cluster_confidence_pairs=lambda shoe: ((shoe.category.name, 1),),\\n \\n description=\\\"%(object)s belong to the %(cluster)s category.\\\",\\n ),\\n \\n # searched keywords cluster\\n SubjectClusterSet(\\n name=\\\"Keyword search cluster set.\\\",\\n \\n weight=0.4,\\n \\n filter_entities=User.objects.filter(words_searched__isnull=False).distinct(),\\n \\n # clusters are words, confidence is 1/number of searched words for user\\n get_cluster_confidence_pairs=lambda user: [(w.word, 1.0/user.words_searched.count()) for w in user.words_searched.all()],\\n \\n description=\\\"%(subject)s has searched for the word %(cluster)s.\\\",\\n ),\\n )\\n \\n biases = (\\n # people liking many shoepairs are more likely to like some more\\n SubjectBias(\\n name=\\\"Users liking many shoes.\\\",\\n \\n description=\\\"User %(subject)s likes many shoe pairs.\\\",\\n \\n weight=0.4, \\n \\n is_positive=True,\\n \\n generator=_shoelikers, \\n \\n confidence=lambda user: float(user.likes_shoes.count())/3\\n ),\\n \\n # multiply liked shoes are more likely to be liked\\n ObjectBias(\\n name=\\\"Popular shoes\\\",\\n \\n description=\\\"Shoe pair %(object)s is popular\\\",\\n \\n weight=0.8,\\n \\n is_positive=True,\\n \\n generator=_popular_shoes,\\n \\n confidence=lambda shoe: float(shoe.likers.count())/3\\n ),\\n )\\n\\n random_recommendation_description = \\\"Recommending a random shoe pair to the user.\\\"\\n \\n algorithm = AggregatingAlgorithm(\\n inner_algorithm=CompilingAlgorithm(\\n inner_algorithm=SimpleAlgorithm(\\n inner_algorithm=None\\n ),\\n compilator=GetFirstCompilator()\\n ),\\n aggregator=LinearAggregator()\\n )\\n \\\"\\\"\\\"The most basic algorithm is used\\\"\\\"\\\"\\n\\nShoeRecommender.explicit_rating_rule = ShoeRecommender.rules[3]\\n\\nclass AverageRecommender(ShoeRecommender):\\n \\n name = \\\"Advanced shoe recommender\\\"\\n #remove_predicted_from_recommendations = False\\n \\n algorithm = AggregatingAlgorithm(\\n inner_algorithm=CompilingAlgorithm(\\n inner_algorithm=SimpleAlgorithm(\\n inner_algorithm=None\\n ),\\n compilator=CombiningCompilator(combinator=AverageCombinator())\\n ),\\n aggregator=CombiningAggregator(combinator=AverageCombinator())\\n )\\n \\\"\\\"\\\"The normal algorithm is used\\\"\\\"\\\"\\n\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.735551655292511,\n \"alphanum_fraction\": 0.735551655292511,\n \"avg_line_length\": 30.66666603088379,\n \"blob_id\": \"c3566ae29c10998605ef4d74844566ae225b99e5\",\n \"content_id\": \"c9c8dd13df97c1175d2cf4da2fcb1d6afa953e0e\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 571,\n \"license_type\": \"no_license\",\n \"max_line_length\": 76,\n \"num_lines\": 18,\n \"path\": \"/UnresystCD/code/adapter/travel/mahout_recommender.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The representation of the Mahout recommender in Unresyst.\\\"\\\"\\\"\\n\\nfrom unresyst.recommender.external_recommender import ExternalRecommender\\nfrom recommender import OrderTourRecommender\\n\\nfrom models import *\\n\\nclass MahoutOrderTourRecommender(ExternalRecommender, OrderTourRecommender):\\n \\\"\\\"\\\"An external order tour recommender\\\"\\\"\\\"\\n \\n name = \\\"Mahout order tour recommender\\\"\\n \\\"\\\"\\\"The name\\\"\\\"\\\"\\n \\n subjects = User.objects\\n \\\"\\\"\\\"The objects to who the recommender will recommend.\\\"\\\"\\\"\\n \\n objects = Tour.objects\\n \\\"\\\"\\\"The objects that will be recommended.\\\"\\\"\\\" \\n\"\n },\n {\n \"alpha_fraction\": 0.6030217409133911,\n \"alphanum_fraction\": 0.6056657433509827,\n \"avg_line_length\": 31.300613403320312,\n \"blob_id\": \"66a0c43d7c3b33d128d124f8c9c4d591f58f0069\",\n \"content_id\": \"503f619f12d00e51283694f39285e255e14f54a6\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5295,\n \"license_type\": \"no_license\",\n \"max_line_length\": 93,\n \"num_lines\": 163,\n \"path\": \"/code/adapter/unresyst/abstractor/base.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The module defines base class for the abstractor package.\\\"\\\"\\\"\\n\\nclass BaseAbstractor(object):\\n \\\"\\\"\\\"The base (abstract) class for all abstractors. Defines the interface.\\\"\\\"\\\"\\n \\n # Build phase:\\n #\\n \\n \\n def create_subjectobjects(self, recommender_model, subjects, objects):\\n \\\"\\\"\\\"Create abstract representation of subjects and objects.\\n \\n @type recommender_model: models.Recommender\\n @param recommender_model: the recommender model instance\\n \\n @type subjects: django.db.models.manager.Manager\\n @param subjects: the manager above the subject model\\n \\n @type objects: django.db.models.manager.Manager\\n @param objects: the manager above the object model\\n \\n \\\"\\\"\\\"\\n pass\\n \\n \\n \\n def create_predicted_relationship_instances(self, predicted_relationship):\\n \\\"\\\"\\\"Create the instances of the predicted relationship.\\n \\n Create bindings on subject-object pairs for pairs between which there's\\n the predicted relationship.\\n \\n @type predicted_relationship: rules.Relationship\\n @param predicted_relationship: the definition of relationship to be predicted \\n \\\"\\\"\\\"\\n pass\\n\\n \\n \\n def create_relationship_instances(self, relationships):\\n \\\"\\\"\\\"Create the instances of relationships relevant for recommendation.\\n \\n Create bindings on pairs of subject/objects, for pairs between where \\n the relationship is. \\n \\n @type relationships: list of rules._WeightedRelationship subclass instances\\n @param relatioships: recommender relationships to be instantiated \\n\\n @raise ConfigurationError: if the weight of some relatioship is\\n outside [0, 1] \\n \\\"\\\"\\\"\\n pass\\n\\n \\n \\n def create_rule_instances(self, rules):\\n \\\"\\\"\\\"Create the instances of rules.\\n \\n Create bindings on pairs of subject/objects, for pairs between where \\n the rule applies.\\n\\n @type rules: list of rules._BaseRule subclass instances\\n @param rules: recommender rules to be instantiated\\n \\n @raise ConfigurationError: if the weight of some rule is outside [0, 1]\\n or some confidence function returns a value outside [0, 1].\\n \\\"\\\"\\\"\\n pass\\n\\n \\n \\n \\n def create_clusters(self, cluster_sets):\\n \\\"\\\"\\\"Create clusters defined in the given sets\\n \\n Crates the cluster sets in the database, their clusters, bindings \\n of subjectobjects to the clusters.\\n \\n @type cluster_sets: a list of clusters.BaseClusterSet subclass instances\\n @param cluster_sets: recommender cluster sets to be evaluated\\n \\n @raise ConfigurationError: if the weight of some cluster set is outside [0, 1]\\n or some cluster member confidence is outside [0, 1].\\n \\\"\\\"\\\"\\n pass\\n \\n \\n def create_biases(self, biases):\\n \\\"\\\"\\\"Create biases defined by the user.\\n \\n Creates bias definitions and instances in the database.\\n \\n @type biases: a list of bias._BaseBias instances\\n @param biases: biases to be evaluated\\n \\n @raise ConfigurationError: if the weight of some cluster set is outside [0, 1]\\n or some cluster member confidence is outside [0, 1]. \\n \\\"\\\"\\\"\\n pass\\n\\n \\n # Update phase:\\n # \\n \\n \\n def add_subject(self, recommender, subject):\\n \\\"\\\"\\\"Add the subject to the abstract subjects, to the relationship and\\n rule instances.\\n \\n @rtype: uvidime\\n @return: the changes performed on the rule and relationship instances.\\n \\\"\\\"\\\"\\n pass\\n \\n \\n def add_object(self, recommender, object_):\\n \\\"\\\"\\\"Add the object to the abstract objects, to the relationship and\\n rule instances.\\n \\n @rtype: uvidime\\n @return: the changes performed on the rule and relationship instances.\\n \\\"\\\"\\\"\\n pass\\n \\n \\n def update_subject(self, recommender, subject):\\n \\\"\\\"\\\"Update the subject in the abstract subjects, in the relationship and\\n rule instances.\\n \\n @rtype: uvidime\\n @return: the changes performed on the rule and relationship instances.\\n \\\"\\\"\\\"\\n pass\\n \\n \\n def update_object(self, recommender, object_):\\n \\\"\\\"\\\"Update the object in the abstract objects, in the relationship and\\n rule instances.\\n \\n @rtype: uvidime\\n @return: the changes performed on the rule and relationship instances.\\n \\\"\\\"\\\"\\n pass \\n \\n \\n def remove_subject(self, recommender, subject):\\n \\\"\\\"\\\"Remove the subject from the abstract subjects, its relationships and\\n rule instances.\\n \\n @rtype: uvidime\\n @return: the changes performed on the rule and relationship instances.\\n \\\"\\\"\\\"\\n pass\\n \\n \\n def remove_object(self, recommender, object_):\\n \\\"\\\"\\\"Update the object from the abstract objects, its relationship and\\n rule instances.\\n \\n @rtype: uvidime\\n @return: the changes performed on the rule and relationship instances.\\n \\\"\\\"\\\"\\n pass \\n \\n \\n\"\n },\n {\n \"alpha_fraction\": 0.5553778409957886,\n \"alphanum_fraction\": 0.5639743804931641,\n \"avg_line_length\": 35.445255279541016,\n \"blob_id\": \"16ac98ed4ffaeb4faf46f8db3777add92cba6db0\",\n \"content_id\": \"e4b8c7014c26aa3d937d4bff79aba92a06a557bf\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5002,\n \"license_type\": \"no_license\",\n \"max_line_length\": 167,\n \"num_lines\": 137,\n \"path\": \"/code/adapter/travel/recommender.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The configuration for the travel agency recommender\\\"\\\"\\\"\\nfrom django.db.models import Sum, Count, Avg\\n\\nfrom unresyst import *\\n\\nfrom models import *\\nfrom constants import *\\n\\ndef _viewed_profile_confidence(u, t):\\n viewed_profs = ViewProfile.objects.filter(tour=t, session__user=u)\\n view_count = viewed_profs.count()\\n avg_duration = viewed_profs.aggregate(Avg('duration'))['duration__avg']\\n \\n return min((float(view_count)/6) * avg_duration/160, 1.0)\\n\\n# predicted bude order\\n# remove_predicted_from_recommendations = True\\n# podpurny budou ty vtipy\\n\\nclass OrderTourRecommender(Recommender):\\n \\\"\\\"\\\"A recommender for suggesting what tour the user should order.\\\"\\\"\\\" \\n\\n name = \\\"Order Tour Recommender\\\"\\n \\\"\\\"\\\"The name\\\"\\\"\\\" \\n \\n subjects = User.objects\\n \\\"\\\"\\\"The objects to who the recommender will recommend.\\\"\\\"\\\"\\n \\n objects = Tour.objects\\n \\\"\\\"\\\"The objects that will be recommended.\\\"\\\"\\\" \\n\\n random_recommendation_description = \\\"Recommending a random tour to the user.\\\"\\n\\n predicted_relationship = PredictedRelationship( \\n name=\\\"User has ordered the tour.\\\",\\n condition=None, \\n description=\\\"\\\"\\\"User %(subject)s has ordered %(object)s.\\\"\\\"\\\",\\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) for uid, tid in Order.objects.values_list('session__user', 'tour').distinct()), \\n )\\n \\\"\\\"\\\"The relationship that will be predicted\\\"\\\"\\\"\\n\\n \\n relationships = ()\\n \\n rules = ( \\n # click = sign of preference\\n SubjectObjectRule(\\n name='User has clicked on something on the tour profile.',\\n weight=0.05,\\n condition=None,\\n is_positive=True,\\n description='User %(subject)s has clicked on something on the %(object)s profile.',\\n # pairs that user has clicked on the tour\\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\\\\n for uid, tid in Click.objects.values_list('session__user', 'tour').distinct()),\\n # the average is around 3, so take 1/6. so that 3 points to the middle.\\n confidence=lambda u, t: min(float(Click.objects.filter(tour=t, session__user=u).count())/6, 1.0),\\n ), \\n \\n # mouse move .. also a sign of preference\\n SubjectObjectRule(\\n name='User has moved the mouse on the tour profile.',\\n weight=0.1,\\n condition=None,\\n is_positive=True,\\n description='User %(subject)s has moved the mouse on %(object)s.',\\n # pairs that user has moved on the tour\\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\\\\n for uid, tid in MouseMove.objects.values_list('session__user', 'tour').distinct()),\\n confidence=lambda u, t: min(float(MouseMove.objects.filter(tour=t, session__user=u).count())/18, 1.0),\\n ),\\n \\n # view profile \\n SubjectObjectRule(\\n name='User has viewed the tour profile page.',\\n weight=0.1,\\n condition=None,\\n is_positive=True,\\n description='User %(subject)s has viewed %(object)s.',\\n # pairs that user has viewed the tour\\n generator=lambda: ((User.objects.get(pk=uid), Tour.objects.get(pk=tid)) \\\\\\n for uid, tid in ViewProfile.objects.values_list('session__user', 'tour').distinct()),\\n # how many times * how long\\n confidence=_viewed_profile_confidence\\n ),\\n\\n \\n )\\n \\n\\n biases = (\\n # multiply viewed tours\\n ObjectBias(\\n name=\\\"Most viewed tours.\\\",\\n \\n description=\\\"Tour %(object)s is much viewed\\\",\\n \\n weight=0.3,\\n \\n is_positive=True,\\n \\n generator=lambda: Tour.objects.annotate(hh=Count('viewprofile')).filter(hh__gt=2).distinct(),\\n \\n confidence=lambda t: min(float(t.viewprofile_set.count())/4, 1.0)\\n ),\\n \\n \\n # multiply mouse moved tours\\n ObjectBias(\\n name=\\\"Most mouse moved tours.\\\",\\n \\n description=\\\"Tour %(object)s is often mouse moved.\\\",\\n \\n weight=0.2,\\n \\n is_positive=True,\\n \\n generator=lambda: Tour.objects.annotate(hh=Count('mousemove')).filter(hh__gt=6).distinct(),\\n \\n confidence=lambda t: min(float(t.mousemove_set.count())/12, 1.0)\\n ),\\n # multiply clicked tours\\n ObjectBias(\\n name=\\\"Most clicked tours.\\\",\\n \\n description=\\\"Tour %(object)s is often clicked on.\\\",\\n \\n weight=0.5,\\n \\n is_positive=True,\\n \\n generator=lambda: Tour.objects.annotate(hh=Count('click')).filter(hh__gt=1).distinct(),\\n \\n confidence=lambda t: min(float(t.click_set.count())/2, 1.0)\\n ),\\n \\n )\\n\\n\\n\\n \\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5668229460716248,\n \"alphanum_fraction\": 0.580405056476593,\n \"avg_line_length\": 45.32362747192383,\n \"blob_id\": \"6c9ac1aa037d4d4fa059f3f62df8d209180e2318\",\n \"content_id\": \"3a4ab5042f1421a2c4c0f84ee73ebf9d9139d5a8\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 41378,\n \"license_type\": \"no_license\",\n \"max_line_length\": 197,\n \"num_lines\": 893,\n \"path\": \"/UnresystCD/code/adapter/unresyst/tests/test_build.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Tests for the building phase.\\n\\nEach layer has a separate test:\\n - recommender\\n - abstractor\\n - aggregator\\n - algorithm \\n\\nThe tests are always started by running the recommender.build method.\\n\\\"\\\"\\\"\\n\\nfrom nose.tools import eq_, assert_raises, assert_almost_equal\\nfrom django.db.models import Q\\n\\nfrom unresyst import Recommender\\nfrom unresyst.models.common import SubjectObject, Recommender as RecommenderModel\\nfrom unresyst.models.abstractor import PredictedRelationshipDefinition, \\\\\\n RelationshipInstance, RuleInstance, RuleRelationshipDefinition, ClusterSet, \\\\\\n BiasDefinition, ExplicitRuleDefinition, ExplicitRuleInstance\\nfrom unresyst.models.aggregator import AggregatedRelationshipInstance, AggregatedBiasInstance\\nfrom unresyst.models.algorithm import RelationshipPredictionInstance \\nfrom test_base import TestBuild, TestEntities, DBTestCase, TestBuildAverage\\nfrom unresyst.exceptions import ConfigurationError, DescriptionKeyError\\nfrom unresyst.recommender.rules import ExplicitSubjectObjectRule\\n\\nfrom demo.recommender import ShoeRecommender\\nfrom demo.models import User, ShoePair\\n\\nPLACES = 4\\n\\\"\\\"\\\"How many places are counted for expectancy accuracy\\\"\\\"\\\"\\n\\nclass TestRecommender(TestBuild):\\n \\\"\\\"\\\"Test case for the recommender class\\\"\\\"\\\"\\n \\n def test_recommender_created(self):\\n \\\"\\\"\\\"Test that the recommender was created during the build\\\"\\\"\\\" \\n \\n # get it from db\\n rec = RecommenderModel.objects.filter(\\n class_name=ShoeRecommender.__name__)\\n \\n # assert there's one \\n eq_(rec.count(), 1, \\\"There should be right one recommender %s.\\\" \\\\\\n % ShoeRecommender.name)\\n \\n rec = rec[0]\\n \\n # assert it has the right name\\n eq_(rec.name, ShoeRecommender.name)\\n \\n # assert the subjects != objects\\n eq_(rec.are_subjects_objects, \\\\\\n ShoeRecommender.subjects == ShoeRecommender.objects)\\n \\n # assert the model is saved in the recommender\\n eq_(ShoeRecommender._get_recommender_model(), rec) \\n\\n def test_cascade_delete(self):\\n \\\"\\\"\\\"Test that the rebuild deletes all that should be deleted\\\"\\\"\\\"\\n \\n # get and delete alice\\n a = User.objects.get(name='Alice') \\n a.delete()\\n \\n # build the recommender again\\n ShoeRecommender.build()\\n \\n # assert there's nothing:\\n \\n # in SubjectObject\\n eq_(SubjectObject.objects.filter(name='Alice').count(), 0)\\n \\n # in RelationshipInstance\\n eq_(RelationshipInstance.objects.filter(subject_object1__name='Alice').count(), 0)\\n eq_(RelationshipInstance.objects.filter(subject_object2__name='Alice').count(), 0) \\n \\n # in RuleInstance\\n eq_(RuleInstance.objects.filter(subject_object1__name='Alice').count(), 0)\\n eq_(RuleInstance.objects.filter(subject_object2__name='Alice').count(), 0)\\n \\n # in AggregatedRelationshipInstance\\n eq_(AggregatedRelationshipInstance.objects.filter(subject_object1__name='Alice').count(), 0)\\n eq_(AggregatedRelationshipInstance.objects.filter(subject_object2__name='Alice').count(), 0)\\n \\n # in RelationshipPredictionInstance\\n eq_(RelationshipPredictionInstance.objects.filter(subject_object1__name='Alice').count(), 0)\\n eq_(RelationshipPredictionInstance.objects.filter(subject_object2__name='Alice').count(), 0) \\n\\n\\nclass TestAbstractor(TestEntities):\\n \\\"\\\"\\\"Testing the abstractor in the build phase\\\"\\\"\\\" \\n \\n # creating subjectobjects:\\n #\\n \\n def test_create_subjectobjects(self):\\n \\\"\\\"\\\"Test creating universal subjects, objects\\\"\\\"\\\"\\n \\n # get the specific entities\\n subjects = User.objects.all()\\n objects = ShoePair.objects.all()\\n \\n universal = SubjectObject.objects.all()\\n \\n # assert the counts fit\\n eq_(subjects.count() + objects.count(), universal.count())\\n \\n # assert all subjects and objects are there\\n for entities, entity_type in ((subjects, 'S'), (objects, 'O')):\\n for en in entities:\\n\\n # the universal representation of entity\\n en_un = universal.filter(id_in_specific=en.id, entity_type=entity_type)\\n \\n # assert there's one \\n eq_(en_un.count(), 1, \\\"There should be right one represenation\\\" + \\\\\\n \\\"of %s, there are %d\\\" % (en, en_un.count()))\\n \\n en_un = en_un[0] \\n \\n # assert it has the right properties \\n\\n # name\\n eq_(en_un.name, en.__unicode__())\\n \\n # recommender\\n eq_(en_un.recommender.class_name, ShoeRecommender.__name__)\\n\\n \\n # predicted relationship\\n # \\n \\n def test_create_predicted_relationship_definition(self):\\n \\\"\\\"\\\"Test creating the definition of the predicted relationship\\\"\\\"\\\"\\n \\n # get definitions related to the shoe recommender\\n qs = PredictedRelationshipDefinition.remove_subclass_objects().filter(\\n recommender=ShoeRecommender._get_recommender_model())\\n \\n # get how many definitions there are\\n # should be one\\n eq_(qs.count(), 1)\\n \\n definition = qs[0]\\n \\n # test it has the right name\\n eq_(definition.name, \\\"User likes shoes.\\\")\\n\\n \\n EXPECTED_PREDICTED_RELATIONSHIP_INSTANCES = (\\n ('Alice', 'Sneakers', \\\"User Alice likes shoes Sneakers.\\\"),\\n ('Bob', 'Sneakers', \\\"User Bob likes shoes Sneakers.\\\"),\\n ('Edgar', 'Rubber Shoes', \\\"User Edgar likes shoes Rubber Shoes.\\\"), \\n ('Edgar', 'RS 130', \\\"User Edgar likes shoes RS 130.\\\"), \\n ) \\n\\n \\n def test_create_predicted_relationship_instances(self):\\n \\\"\\\"\\\"Test creating instances of the predicted relationship.\\\"\\\"\\\"\\n \\n # get the relationship definition\\n definition = PredictedRelationshipDefinition.remove_subclass_objects().get( \\n recommender=ShoeRecommender._get_recommender_model())\\n \\n # get instances of the predicted relationship\\n instances = RelationshipInstance.remove_subclass_objects().filter(definition=definition)\\n \\n eq_(len(self.EXPECTED_PREDICTED_RELATIONSHIP_INSTANCES), instances.count(),\\n \\\"Expected: %d, obtained: %d. Obtained: %s\\\" % \\\\\\n (len(self.EXPECTED_PREDICTED_RELATIONSHIP_INSTANCES), instances.count(), instances))\\n \\n for expected_data in self.EXPECTED_PREDICTED_RELATIONSHIP_INSTANCES:\\n\\n # instances that have either one or another order of the subjobjects\\n rel_instance = instances.filter(\\n subject_object1=self.universal_entities[expected_data[0]], \\n subject_object2=self.universal_entities[expected_data[1]]\\n )\\n \\n # there should be one\\n eq_(rel_instance.count(), 1)\\n \\n instance = rel_instance[0]\\n \\n # test it has the right description\\n eq_(instance.description, expected_data[2]) \\n\\n\\n # relationships:\\n #\\n\\n def test_create_relationship_definitions(self):\\n \\\"\\\"\\\"Test creating the definitions of the relationships\\\"\\\"\\\"\\n \\n self._test_definitions(ShoeRecommender.relationships)\\n\\n \\n EXPECTED_RELATIONSHIP_DICT = {\\n # relationship name, list of instances - (entity1, entity2, description)\\n # for symmetric relationships the description is a pair - both orderings\\n \\\"User has viewed shoes.\\\": (\\n ('Alice', 'Rubber Shoes', \\\"User Alice has viewed Rubber Shoes.\\\"),\\n ('Bob', 'Sneakers', \\\"User Bob has viewed Sneakers.\\\"),\\n ('Cindy', 'Rubber Shoes', \\\"User Cindy has viewed Rubber Shoes.\\\"),\\n ),\\n \\\"Users live in the same city.\\\": (\\n ('Alice', 'Bob', (\\\"Users Alice and Bob live in the same city.\\\",\\n \\\"Users Bob and Alice live in the same city.\\\")),\\n ('Cindy', 'Daisy', (\\\"Users Cindy and Daisy live in the same city.\\\",\\n \\\"Users Daisy and Cindy live in the same city.\\\")),\\n ),\\n \\\"Shoes were made by the same manufacturer.\\\": (\\n ('Sneakers', 'Rubber Shoes', \\n (\\\"Shoes Sneakers and Rubber Shoes were made by the same manufacturer.\\\",\\n \\\"Shoes Rubber Shoes and Sneakers were made by the same manufacturer.\\\")),\\n ),\\n 'User lives in the same city as the shoe manufacturer.': (\\n ('Alice', 'Rubber Shoes', \\\"User Alice is from the same city as the manufacturer of Rubber Shoes.\\\"),\\n ('Bob', 'Rubber Shoes', \\\"User Bob is from the same city as the manufacturer of Rubber Shoes.\\\"),\\n ('Alice', 'Sneakers', \\\"User Alice is from the same city as the manufacturer of Sneakers.\\\"),\\n ('Bob', 'Sneakers', \\\"User Bob is from the same city as the manufacturer of Sneakers.\\\"),\\n ('Cindy', 'RS 130', \\\"User Cindy is from the same city as the manufacturer of RS 130.\\\"),\\n ('Daisy', 'RS 130', \\\"User Daisy is from the same city as the manufacturer of RS 130.\\\")\\n ),\\n }\\n \\n \\n def test_create_relationship_instances(self):\\n \\\"\\\"\\\"Test creating instances of all the defined relationships.\\\"\\\"\\\"\\n \\n self._test_instances(\\n def_list=ShoeRecommender.relationships,\\n expected_dict=self.EXPECTED_RELATIONSHIP_DICT,\\n instance_manager=RelationshipInstance.objects) \\n \\n # rules:\\n #\\n \\n def test_create_rule_definitions(self):\\n \\\"\\\"\\\"Test creating the definitions of the rules\\\"\\\"\\\"\\n \\n # explicit rules are tested separately\\n rules = filter(lambda r: r.__class__ != ExplicitSubjectObjectRule, ShoeRecommender.rules)\\n self._test_definitions(rules)\\n\\n\\n EXPECTED_RULE_DICT = {\\n # rule name, list of instances - (entity1, entity2, description, expectancy)\\n # for symmetric rules the description is a pair - both orderings\\n \\\"Don't recommend winter shoes for southern users.\\\": (\\n ('Alice', 'RS 130', \\\"Alice is from south, so RS 130 can't be recommended to him/her.\\\", 1),\\n ('Bob', 'RS 130', \\\"Bob is from south, so RS 130 can't be recommended to him/her.\\\", 1),\\n ),\\n \\\"Users with similar age.\\\": (\\n ('Alice', 'Bob', (\\\"Users Alice and Bob are about the same age.\\\", \\n \\\"Users Bob and Alice are about the same age.\\\"), 0.75),\\n ('Edgar', 'Fionna', (\\\"Users Edgar and Fionna are about the same age.\\\", \\n \\\"Fionna and Edgar are about the same age.\\\"), 1),\\n ),\\n \\\"Shoes with common keywords.\\\": (\\n ('Rubber Shoes', 'Sneakers', (\\\"The shoe pairs Rubber Shoes and Sneakers share some keywords.\\\",\\n \\\"The shoe pairs Sneakers and Rubber Shoes share some keywords.\\\"), 1),\\n ('Sneakers', 'Design Shoes', (\\\"The shoe pairs Sneakers and Design Shoes share some keywords.\\\",\\n \\\"The shoe pairs Design Shoes and Sneakers share some keywords.\\\"), 0.5), \\n ),\\n } \\n\\n def test_create_rule_instances(self):\\n \\\"\\\"\\\"Test creating instances of all the defined rules.\\\"\\\"\\\"\\n\\n # explicit rules are tested separately\\n rules = filter(lambda r: r.__class__ != ExplicitSubjectObjectRule, ShoeRecommender.rules)\\n \\n self._test_instances(\\n def_list=rules,\\n expected_dict=self.EXPECTED_RULE_DICT,\\n instance_manager=RuleInstance.objects)\\n \\n # auxiliary functions:\\n # \\n \\n def _test_instances(self, def_list, expected_dict, instance_manager):\\n \\\"\\\"\\\"Test creating instances.\\n \\n @type def_list: list of Rule/Relationship definitions \\n @param def_list: a list of definitions to check\\n \\n @type expected_dict: dict str: tuple\\n @param expected_dict: dictionary of expected instance attributes\\n \\n @type instance_manager: django.db.models.manager.Manager\\n @param instance_manager: the manager above rule/relationship instances\\n \\\"\\\"\\\"\\n # get definitions related to the shoe recommender\\n qs = RuleRelationshipDefinition.objects.filter(\\n recommender=ShoeRecommender._get_recommender_model()) \\n \\n # for each defined type there should be a model definition in db\\n for r in def_list:\\n\\n # get the expected data\\n expected_relationships = expected_dict[r.name] \\n\\n # get the definition model by name (shouldn't throw an error)\\n definition = qs.get(name=r.name)\\n \\n instances = instance_manager.filter(definition=definition)\\n \\n # expect there are as many relationship instances as expected\\n eq_(instances.count(), len(expected_relationships))\\n \\n for expected_data in expected_relationships:\\n \\n # instances that have either one or another order of the subjobjects\\n rel_instance = instances.filter(\\n Q(subject_object1=self.universal_entities[expected_data[0]], \\n subject_object2=self.universal_entities[expected_data[1]]) | \\\\\\n Q(subject_object1=self.universal_entities[expected_data[1]], \\n subject_object2=self.universal_entities[expected_data[0]])\\n ) \\n \\n # there should be one\\n eq_(rel_instance.count(), 1)\\n \\n instance = rel_instance[0]\\n \\n # assert the right order in the relationship\\n if instance.definition.as_leaf_class().relationship_type == 'S-O': \\n eq_(instance.subject_object1.entity_type, 'S')\\n eq_(instance.subject_object2.entity_type, 'O')\\n else:\\n assert instance.subject_object1.id < instance.subject_object2.id\\n \\n # test it has the right description\\n # the diferentiation for symmetric relationships - the order \\n # of the entities in the description can be arbitrary\\n if r.is_symmetric:\\n desc_tuple = expected_data[2]\\n assert instance.description == desc_tuple[0] or \\\\\\n instance.description == desc_tuple[1], \\\\\\n \\\"The description '%s' is wrong. Should be '%s' or '%s'\\\" % \\\\\\n (instance.description, desc_tuple[0], desc_tuple[1])\\n else:\\n eq_(instance.description, expected_data[2]) \\n \\n # for rules \\n if len(expected_data) == 4:\\n eq_(instance.confidence, expected_data[3], \\n \\\"Got confidence: %f, expected: %f for instance: %s\\\" % (instance.confidence, expected_data[3], instance))\\n \\n\\n def _test_definitions(self, def_list):\\n \\\"\\\"\\\"Test the definitions.\\n \\n @type def_list: an iterable of RuleRelationshipDefinition model\\n @param def_list: list of definitions to check.\\n \\\"\\\"\\\"\\n # get definitions related to the shoe recommender\\n qs = RuleRelationshipDefinition.objects.filter(\\n recommender=ShoeRecommender._get_recommender_model()) \\n \\n # explicit rules are tested separately\\n rules = filter(lambda r: r.__class__ != ExplicitSubjectObjectRule, ShoeRecommender.rules)\\n \\n # assert the number of definitions is right\\n eq_(qs.count(), len(ShoeRecommender.relationships) + len(rules))\\n \\n # for each defined type there should be a model definition in db\\n for r in def_list:\\n\\n # get the definition model by name (shouldn't throw an error)\\n rmodel = qs.get(name=r.name)\\n \\n # assert the positiveness, weight and relationship type are right\\n eq_(rmodel.is_positive, r.is_positive)\\n eq_(rmodel.weight, r.weight) \\n eq_(rmodel.relationship_type, r.relationship_type) \\n\\n\\n EXP_CLUSTER_SETS = {\\n \\\"Shoe category cluster set.\\\": \\n (\\n {'For Sports': (('RS 130', 1, \\\"RS 130 belong to the For Sports category.\\\"), \\n (\\\"Octane SL\\\", 1, \\\"Octane SL belong to the For Sports category.\\\")),\\n 'Casual': ((\\\"Design Shoes\\\", 1, \\\"Design Shoes belong to the Casual category.\\\"), \\n (\\\"Sneakers\\\", 1, \\\"Sneakers belong to the Casual category.\\\")),},\\n 'O', 0.3),\\n \\\"Keyword search cluster set.\\\":\\n (\\n {'Sporty': (\\n ('Bob', 0.5, \\\"Bob has searched for the word Sporty.\\\"),\\n ),\\n 'Comfortable': (\\n ('Bob', 0.5, \\\"Bob has searched for the word Comfortable.\\\"),\\n ('Cindy', 0.5, \\\"Cindy has searched for the word Comfortable.\\\"),\\n ),\\n 'Cool':(\\n ('Cindy', 0.5, \\\"Cindy has searched for the word Cool.\\\"),\\n ),\\n },\\n 'S', 0.4),\\n }\\n \\\"\\\"\\\"A dictionary: cluster set name: contained clusters, entity type, weight,\\n the members: object name, confidence, description.\\\"\\\"\\\"\\n \\n def test_clusters(self):\\n \\\"\\\"\\\"Test that the clusters are created as expected\\\"\\\"\\\"\\n \\n # get cluster sets from db\\n cluster_sets = ClusterSet.objects.filter(\\n recommender=ShoeRecommender._get_recommender_model())\\n \\n # assert there are as many as expected\\n eq_(len(self.EXP_CLUSTER_SETS), cluster_sets.count())\\n \\n for cs in cluster_sets:\\n\\n # get the expected contents\\n clusters, ent_type, weight = self.EXP_CLUSTER_SETS[cs.name]\\n \\n eq_(ent_type, cs.entity_type)\\n eq_(weight, cs.weight)\\n eq_(len(clusters), cs.cluster_set.count())\\n \\n # go through the clusters\\n for c in cs.cluster_set.all():\\n\\n cluster_members = clusters[c.name]\\n \\n eq_(len(cluster_members), c.clustermember_set.count())\\n \\n for cm in c.clustermember_set.all():\\n # expect the combination of entity and confidence is \\n # in the expected\\n assert (cm.member.name, cm.confidence, cm.description) in cluster_members, \\\\\\n \\\"The member '%s' with confidence '%f' and description '%s' isn't one of the expected %s\\\" % (\\n cm.member, cm.confidence, cm.description, cluster_members)\\n\\n EXP_BIASES = {\\n \\\"Users liking many shoes.\\\": \\n {\\n 'Alice': (0.333333, \\\"User Alice likes many shoe pairs.\\\"),\\n 'Bob': (0.333333, \\\"User Bob likes many shoe pairs.\\\"),\\n 'Edgar': (0.66666, \\\"User Edgar likes many shoe pairs.\\\"),\\n },\\n \\\"Popular shoes\\\":\\n {\\n 'Sneakers': (0.6666666, \\\"Shoe pair Sneakers is popular\\\"),\\n 'Rubber Shoes': (0.333333, \\\"Shoe pair Rubber Shoes is popular\\\"),\\n 'RS 130': (0.333333, \\\"Shoe pair RS 130 is popular\\\"),\\n },\\n }\\n \\\"\\\"\\\"A dictionary: cluster set name: contained clusters, entity type, weight\\\"\\\"\\\" \\n \\n def test_bias(self):\\n \\\"\\\"\\\"Test that biases were created as expected\\\"\\\"\\\"\\n \\n # get bias definitions from db\\n bias_defs = BiasDefinition.objects.filter(\\n recommender=ShoeRecommender._get_recommender_model())\\n \\n # assert there are as many as expected\\n eq_(len(self.EXP_BIASES), bias_defs.count())\\n \\n for bias_def in bias_defs:\\n \\n # assert it's in the expected\\n assert self.EXP_BIASES.has_key(bias_def.name)\\n\\n # get the expected contents \\n exp_biases = self.EXP_BIASES[bias_def.name]\\n \\n eq_(len(exp_biases), bias_def.biasinstance_set.count(), \\n \\\"Wrong count for definition '%s' should be %s, is %s\\\" % \\\\\\n (bias_def.name, exp_biases, bias_def.biasinstance_set.all()))\\n \\n # go through the instances connected to definition\\n for bias in bias_def.biasinstance_set.all():\\n \\n # assert it's in the expected biases\\n assert exp_biases.has_key(bias.subject_object.name)\\n \\n # assert the properties are as expected\\n exp_conf, exp_desc = exp_biases[bias.subject_object.name]\\n \\n assert_almost_equal(exp_conf, bias.confidence, PLACES)\\n eq_(exp_desc, bias.description) \\n \\n EXP_EXPLICIT_RULES = {\\n \\\"Shoe rating.\\\": \\n {\\n ('Alice', 'Design Shoes'): (0.8, \\\"User Alice has rated Design Shoes.\\\"),\\n ('Bob', 'Design Shoes'): (0.2, \\\"User Bob has rated Design Shoes.\\\"),\\n },\\n }\\n \\n def test_explicit_rules(self):\\n \\\"\\\"\\\"Test the explicit rules and their definitions \\n were created as expected\\\"\\\"\\\"\\n \\n rule_defs = ExplicitRuleDefinition.objects.all()\\n \\n # expect the definition count fits\\n eq_(len(self.EXP_EXPLICIT_RULES), rule_defs.count())\\n \\n for rule_def in rule_defs:\\n \\n # assert the def is there\\n assert self.EXP_EXPLICIT_RULES.has_key(rule_def.name)\\n \\n # get its expected rules\\n exp_rules = self.EXP_EXPLICIT_RULES[rule_def.name]\\n \\n rules = ExplicitRuleInstance.objects.filter(definition=rule_def)\\n \\n for rule in rules:\\n \\n # assert the pair is expected\\n assert exp_rules.has_key((rule.subject_object1.name, rule.subject_object2.name)), \\\\\\n \\\"The pair %s, %s isn't expected\\\" % (rule.subject_object1.name, rule.subject_object2.name)\\n\\n exp_rule = exp_rules[(rule.subject_object1.name, rule.subject_object2.name)]\\n \\n assert_almost_equal(exp_rule[0], rule.expectancy, PLACES)\\n eq_(exp_rule[1], rule.description)\\n \\n \\nclass TestAbstractorRecommenderErrors(DBTestCase):\\n \\\"\\\"\\\"Test various errors thrown by Abstractor and/or Recommender and/or Algorithm\\\"\\\"\\\"\\n\\n def test_invalid_relationship_weight(self):\\n \\\"\\\"\\\"Test if the exception is raised for a relatioship with invalid weight\\\"\\\"\\\"\\n \\n # set some invalid weight, assert it throws the error \\n w = ShoeRecommender.relationships[1].weight \\n ShoeRecommender.relationships[1].weight = 1.5\\n\\n assert_raises(ConfigurationError, ShoeRecommender.build)\\n\\n # restore the original value\\n ShoeRecommender.relationships[1].weight = w\\n\\n \\n def test_invalid_rule_weight(self):\\n \\\"\\\"\\\"Test if the exception is raised for a rule with invalid weight\\\"\\\"\\\"\\n \\n # set some invalid weight, assert it throws the error \\n w = ShoeRecommender.rules[2].weight \\n ShoeRecommender.rules[2].weight = 1.5\\n\\n assert_raises(ConfigurationError, ShoeRecommender.build)\\n\\n # restore the original value\\n ShoeRecommender.rules[2].weight = w \\n\\n\\n def test_invalid_confidence(self):\\n \\\"\\\"\\\"Test if the exception is raised for a rule with invalid confidence\\\"\\\"\\\"\\n \\n # set some invalid weight, assert it throws the error \\n c = ShoeRecommender.rules[0].confidence\\n ShoeRecommender.rules[0].confidence = lambda a, b: 1.3\\n\\n assert_raises(ConfigurationError, ShoeRecommender.build)\\n\\n # restore the original value\\n ShoeRecommender.rules[0].confidence = c\\n\\n def test_empty_predicted_relationship(self):\\n \\\"\\\"\\\"Test building a recommender with emtpy predicted relationship\\\"\\\"\\\"\\n \\n # save and delete the predicted_relationship\\n pr = ShoeRecommender.predicted_relationship\\n ShoeRecommender.predicted_relationship = None\\n \\n # assert it raises an error\\n assert_raises(ConfigurationError, ShoeRecommender.build)\\n \\n # restore the original predicted_relationship\\n ShoeRecommender.predicted_relationship = pr\\n \\n def test_empty_subjects(self):\\n \\\"\\\"\\\"Test building a recommender with empty subjects\\\"\\\"\\\"\\n \\n # delete all users\\n User.objects.all().delete()\\n \\n # assert it raises an error\\n assert_raises(ConfigurationError, ShoeRecommender.build) \\n\\n def test_empty_objects(self):\\n \\\"\\\"\\\"Test building a recommender with empty objects\\\"\\\"\\\"\\n \\n # delete all shoes\\n ShoePair.objects.all().delete()\\n \\n # assert it raises an error\\n assert_raises(ConfigurationError, ShoeRecommender.build) \\n \\n def test_rule_description_error(self):\\n \\\"\\\"\\\"Test building a recommender with strange rule description\\\"\\\"\\\"\\n \\n # save and mess up the description\\n d = ShoeRecommender.rules[1].description \\n ShoeRecommender.rules[1].description = \\\"\\\"\\\"Some messed %(crap)s description\\\"\\\"\\\"\\n \\n # assert it raises an error\\n assert_raises(DescriptionKeyError, ShoeRecommender.build) \\n \\n # restore the original \\n ShoeRecommender.rules[1].description = d\\n\\n def test_relationship_description_error(self):\\n \\\"\\\"\\\"Test building a recommender with strange relationship description\\\"\\\"\\\"\\n \\n # save and mess up the description\\n d = ShoeRecommender.relationships[0].description \\n ShoeRecommender.relationships[0].description = \\\"\\\"\\\"Some other messed %(other_crap)s description\\\"\\\"\\\"\\n \\n # assert it raises an error\\n assert_raises(DescriptionKeyError, ShoeRecommender.build) \\n \\n # restore the original \\n ShoeRecommender.relationships[0].description = d\\n\\n def test_no_key_description_error(self):\\n \\\"\\\"\\\"Test building a recommender with a description without keys\\\"\\\"\\\"\\n \\n # save and mess up the description\\n d = ShoeRecommender.rules[2].description \\n ShoeRecommender.rules[2].description = \\\"\\\"\\\"No-key description.\\\"\\\"\\\"\\n \\n # assert it doesn't raise any error\\n ShoeRecommender.build()\\n \\n # restore the original \\n ShoeRecommender.rules[2].description = d\\n \\n def test_empty_rules_rels(self):\\n \\\"\\\"\\\"Test building the recommender with empty rules and relationships\\\"\\\"\\\"\\n \\n # save and delete rules and relationships\\n rules = ShoeRecommender.rules\\n relationships = ShoeRecommender.relationships\\n \\n ShoeRecommender.rules = ()\\n ShoeRecommender.relationships = ()\\n \\n # build it, test it doesn't freeze\\n ShoeRecommender.build() \\n \\n # restore\\n ShoeRecommender.rules = rules\\n ShoeRecommender.relationships = relationships\\n \\ndef _count_exp(conf):\\n return 0.5 + conf/2\\n \\ndef _count_neg_exp(conf): \\n return 0.5 - conf/2\\n \\nclass TestAggregator(TestEntities):\\n \\\"\\\"\\\"Testing the building phase of the Linear Aggregator\\\"\\\"\\\"\\n \\n EXPECTED_AGGREGATES = {\\n # S-O\\n ('Alice', 'RS 130'): (_count_neg_exp(0.85), 'S-O', \\n (\\\"Alice is from south, so RS 130 can't be recommended to him/her.\\\",)), # 0.075\\n ('Alice', 'Rubber Shoes'): ((_count_exp(0.4) + _count_exp(0.1))/2, 'S-O',\\n ('User Alice has viewed Rubber Shoes. User Alice is from the same city as the manufacturer of Rubber Shoes.',)), # 0.625\\n ('Alice', 'Sneakers'): (_count_exp(0.1), 'S-O', \\n (\\\"User Alice is from the same city as the manufacturer of Sneakers.\\\",)), # 0.55\\n \\n ('Bob', 'RS 130'): (_count_neg_exp(0.85), 'S-O', \\n (\\\"Bob is from south, so RS 130 can't be recommended to him/her.\\\",)), # 0.075\\n ('Bob', 'Rubber Shoes'): (_count_exp(0.1), 'S-O',\\n (\\\"User Bob is from the same city as the manufacturer of Rubber Shoes.\\\",)), # 0.55\\n ('Bob', 'Sneakers'): ((_count_exp(0.4) + _count_exp(0.1))/2, 'S-O',\\n (\\\"User Bob has viewed Sneakers. User Bob is from the same city as the manufacturer of Sneakers.\\\",)), # 0.625\\n\\n ('Cindy', 'RS 130'): (_count_exp(0.1), 'S-O',\\n (\\\"User Cindy is from the same city as the manufacturer of RS 130.\\\",)), # 0.55\\n ('Cindy', 'Rubber Shoes'): (_count_exp(0.4), 'S-O',\\n (\\\"User Cindy has viewed Rubber Shoes.\\\",)), # 0.7\\n \\n ('Daisy', 'RS 130'): (_count_exp(0.1), 'S-O',\\n (\\\"User Daisy is from the same city as the manufacturer of RS 130.\\\",)), # 0.55\\n \\n # S-S\\n ('Alice', 'Bob'): ((_count_exp(0.75 * 0.2) + _count_exp(0.3))/2, 'S-S', (\\n \\\"Users Alice and Bob live in the same city. Users Alice and Bob are about the same age.\\\",\\n \\\"Users Alice and Bob live in the same city. Users Bob and Alice are about the same age.\\\",\\n \\\"Users Alice and Bob live in the same city. Users Bob and Alice are about the same age.\\\",\\n \\\"Users Bob and Alice live in the same city. Users Bob and Alice are about the same age.\\\")), # 0.6125\\n ('Cindy', 'Daisy'): (_count_exp(0.3), 'S-S', (\\n \\\"Users Cindy and Daisy live in the same city.\\\",\\n \\\"Users Daisy and Cindy live in the same city.\\\")), #0.65\\n ('Edgar', 'Fionna'): (_count_exp(1 * 0.2), 'S-S', (\\n \\\"Users Edgar and Fionna are about the same age.\\\",\\n \\\"Users Fionna and Edgar are about the same age.\\\")),\\n \\n # O-O\\n ('Rubber Shoes', 'Sneakers'): ((_count_exp(0.4) + _count_exp(0.1))/2, 'O-O', (\\n \\\"The shoe pairs Rubber Shoes and Sneakers share some keywords. Shoes Sneakers and Rubber Shoes were made by the same manufacturer.\\\",\\n \\\"The shoe pairs Rubber Shoes and Sneakers share some keywords. Shoes Rubber Shoes and Sneakers were made by the same manufacturer.\\\",\\n \\\"The shoe pairs Sneakers and Rubber Shoes share some keywords. Shoes Sneakers and Rubber Shoes were made by the same manufacturer.\\\",\\n \\\"The shoe pairs Sneakers and Rubber Shoes share some keywords. Shoes Rubber Shoes and Sneakers were made by the same manufacturer.\\\")), # 0.625\\n ('Sneakers', 'Design Shoes'): (_count_exp(0.2), 'O-O',(\\n \\\"The shoe pairs Sneakers and Design Shoes share some keywords.\\\",\\n \\\"The shoe pairs Design Shoes and Sneakers share some keywords.\\\",)), # 0.6 \\n }\\n \\\"\\\"\\\"A dictionary: pair of entities : expectancy, entity_type, description\\\"\\\"\\\" \\n\\n def test_aggregates_created(self):\\n \\\"\\\"\\\"Test that the aggregates were created as expected\\\"\\\"\\\"\\n \\n # filter aggregated instances for my recommender\\n aggr_instances = AggregatedRelationshipInstance.objects.filter(\\n recommender=self.recommender._get_recommender_model())\\n \\n # assert it has the expected length \\n eq_(aggr_instances.count(), len(self.EXPECTED_AGGREGATES))\\n \\n for aggr_inst in aggr_instances.iterator():\\n pair1 = (aggr_inst.subject_object1.name, aggr_inst.subject_object2.name)\\n pair2 = (aggr_inst.subject_object2.name, aggr_inst.subject_object1.name) \\n \\n # try getting the instance from expected in both directions\\n if self.EXPECTED_AGGREGATES.has_key(pair1):\\n expected_expectancy, expected_rel_type, expected_descs = self.EXPECTED_AGGREGATES[pair1]\\n else:\\n if self.EXPECTED_AGGREGATES.has_key(pair2):\\n expected_expectancy,expected_rel_type, expected_descs = self.EXPECTED_AGGREGATES[pair2]\\n else:\\n # if not found it's unexpected.\\n assert False, \\\\\\n \\\"Unexpected aggregate between '%s' and '%s' expectancy: %f\\\" % \\\\\\n (pair1 + (aggr_inst.expectancy, ))\\n \\n # assert the order in the aggregate is right\\n if aggr_inst.relationship_type == 'S-O':\\n assert aggr_inst.subject_object1.entity_type == 'S' and aggr_inst.subject_object2.entity_type == 'O'\\n else:\\n assert aggr_inst.subject_object1.id < aggr_inst.subject_object2.id\\n \\n # assert the expectancy is as expected \\n assert_almost_equal(aggr_inst.expectancy, expected_expectancy, PLACES,\\n \\\"Expectancy is '%f' should be '%f' for the pair %s, %s, Obtained description: %s\\\" % \\\\\\n ((aggr_inst.expectancy, expected_expectancy) + pair1 + (aggr_inst.description,))) \\n \\n # assert the relationship type is as expected \\n eq_(aggr_inst.relationship_type, expected_rel_type,\\n \\\"Relationship type is '%s' should be '%s' for the pair %s, %s\\\" % \\\\\\n ((aggr_inst.relationship_type, expected_rel_type) + pair1)) \\n \\n # assert the description is as expected \\n assert aggr_inst.description in expected_descs, \\\\\\n \\\"Description is '%s' should be one of '%s' for the pair %s, %s\\\" % \\\\\\n ((aggr_inst.description, expected_descs) + pair1) \\n\\n EXP_AGGR_BIASES = {\\n 'Alice': (_count_exp(0.4 * 0.333333), \\\"User Alice likes many shoe pairs.\\\"),\\n 'Bob': (_count_exp(0.4 * 0.333333), \\\"User Bob likes many shoe pairs.\\\"),\\n 'Edgar': (_count_exp(0.4 * 0.66666), \\\"User Edgar likes many shoe pairs.\\\"),\\n 'Sneakers': (_count_exp(0.8 * 0.6666666), \\\"Shoe pair Sneakers is popular\\\"),\\n 'Rubber Shoes': (_count_exp(0.8 * 0.333333), \\\"Shoe pair Rubber Shoes is popular\\\"),\\n 'RS 130': (_count_exp(0.8 * 0.333333), \\\"Shoe pair RS 130 is popular\\\"),\\n }\\n \\n def test_aggregated_biases_created(self):\\n \\\"\\\"\\\"Test that the bias aggregates were created as expected\\\"\\\"\\\"\\n \\n # get the aggregates, check count\\n # \\n aggregates = AggregatedBiasInstance.objects.filter(recommender=self.recommender._get_recommender_model()) \\n eq_(aggregates.count(), len(self.EXP_AGGR_BIASES))\\n \\n # go through created aggregates, assert its in expected and \\n # with expected attrs\\n for aggr in aggregates:\\n \\n assert self.EXP_AGGR_BIASES.has_key(aggr.subject_object.name)\\n exp_expectancy, exp_desc = self.EXP_AGGR_BIASES[aggr.subject_object.name]\\n \\n assert_almost_equal(aggr.expectancy, exp_expectancy, PLACES,\\n \\\"Expected expectancy %f, got %f for %s\\\" % \\\\\\n (exp_expectancy, aggr.expectancy, aggr.subject_object.name) )\\n \\n eq_(aggr.description, exp_desc) \\n \\nclass TestAggregatorAverage(TestBuildAverage): \\n \\\"\\\"\\\"Testing the aggregator of the average build\\\"\\\"\\\"\\n\\n def test_aggregated_biases_created(self):\\n \\\"\\\"\\\"AverageRecommender: Test that the bias aggregates were created as expected\\\"\\\"\\\" \\n # call the same method on the normal aggregator\\n ta = TestAggregator(\\\"test_aggregated_biases_created\\\")\\n ta.recommender = self.recommender\\n ta.test_aggregated_biases_created()\\n EXPECTED_AGGREGATES = {\\n # S-S\\n ('Alice', 'Bob'): ((_count_exp(0.75 * 0.2) + _count_exp(0.3))/2, 'S-S', (\\n \\\"Reason 1: Users Alice and Bob live in the same city. Reason 2: Users Alice and Bob are about the same age.\\\",\\n \\\"Reason 1: Users Alice and Bob live in the same city. Reason 2: Users Bob and Alice are about the same age.\\\",\\n \\\"Reason 1: Users Alice and Bob live in the same city. Reason 2: Users Bob and Alice are about the same age.\\\",\\n \\\"Reason 1: Users Bob and Alice live in the same city. Reason 2: Users Bob and Alice are about the same age.\\\")), # 0.6125\\n ('Cindy', 'Daisy'): (_count_exp(0.3), 'S-S', (\\n \\\"Users Cindy and Daisy live in the same city.\\\",\\n \\\"Users Daisy and Cindy live in the same city.\\\")), #0.65\\n ('Edgar', 'Fionna'): (_count_exp(1 * 0.2), 'S-S', (\\n \\\"Users Edgar and Fionna are about the same age.\\\",\\n \\\"Users Fionna and Edgar are about the same age.\\\")),\\n \\n # O-O\\n ('Rubber Shoes', 'Sneakers'): ((_count_exp(0.4) + _count_exp(0.1))/2, 'O-O', (\\n \\\"Reason 1: The shoe pairs Rubber Shoes and Sneakers share some keywords. Reason 2: Shoes Sneakers and Rubber Shoes were made by the same manufacturer.\\\",\\n \\\"Reason 1: The shoe pairs Rubber Shoes and Sneakers share some keywords. Reason 2: Shoes Rubber Shoes and Sneakers were made by the same manufacturer.\\\",\\n \\\"Reason 1: The shoe pairs Sneakers and Rubber Shoes share some keywords. Reason 2: Shoes Sneakers and Rubber Shoes were made by the same manufacturer.\\\",\\n \\\"Reason 1: The shoe pairs Sneakers and Rubber Shoes share some keywords. Reason 2: Shoes Rubber Shoes and Sneakers were made by the same manufacturer.\\\")), # 0.625\\n ('Sneakers', 'Design Shoes'): (0.625, 'O-O',(\\n \\\"Reason 1: Sneakers belong to the Casual category. Design Shoes belong to the Casual category. Reason 2: The shoe pairs Sneakers and Design Shoes share some keywords.\\\",\\n \\\"Reason 1: Sneakers belong to the Casual category. Design Shoes belong to the Casual category. Reason 2: The shoe pairs Design Shoes and Sneakers share some keywords.\\\",\\n \\\"Reason 1: Design Shoes belong to the Casual category. Sneakers belong to the Casual category. Reason 2: The shoe pairs Sneakers and Design Shoes share some keywords.\\\",\\n \\\"Reason 1: Design Shoes belong to the Casual category. Sneakers belong to the Casual category. Reason 2: The shoe pairs Design Shoes and Sneakers share some keywords.\\\",)), # 0.6 \\n }\\n\\n def test_aggregates_created(self):\\n \\\"\\\"\\\"AverageRecommender: Test that the aggregates were created as expected\\\"\\\"\\\"\\n \\n # instantiate the common aggregator\\n ta = TestAggregator(\\\"test_aggregates_created\\\")\\n ta.recommender = self.recommender\\n\\n # leave in the aggregates only similarity relationship \\n ta.EXPECTED_AGGREGATES = self.EXPECTED_AGGREGATES\\n\\n \\n # call the same method on the normal aggregator\\n ta.test_aggregates_created()\\n \\n \\nclass DTestAlgorithm(TestEntities):\\n \\\"\\\"\\\"Testing the building phase of the SimpleAlgorithm\\\"\\\"\\\" \\n \\n \\n EXPECTED_PREDICTIONS = {\\n # S-O\\n ('Alice', 'RS 130'): _count_neg_exp(0.85), # 0.075\\n ('Alice', 'Rubber Shoes'): (_count_exp(0.4) + _count_exp(0.1))/2, # 0.625\\n ('Alice', 'Sneakers'): _count_exp(0.1), # 0.55\\n ('Alice', 'Design Shoes'): _count_exp(0.2), # 0.6\\n \\n ('Bob', 'RS 130'): _count_neg_exp(0.85), # 0.075\\n ('Bob', 'Rubber Shoes'): _count_exp(0.1), # 0.55\\n ('Bob', 'Sneakers'): (_count_exp(0.4) + _count_exp(0.1))/2, # 0.625\\n ('Bob', 'Design Shoes'): _count_exp(0.2), # 0.6\\n\\n ('Cindy', 'RS 130'): _count_exp(0.1), # 0.55\\n ('Cindy', 'Rubber Shoes'): _count_exp(0.4), # 0.7 \\n \\n ('Daisy', 'RS 130'): _count_exp(0.1), # 0.55\\n \\n ('Edgar', 'Sneakers'): (_count_exp(0.1) + _count_exp(0.4))/2, # 0.625 \\n \\n ('Fionna', 'Rubber Shoes'): _count_exp(0.2),\\n ('Fionna', 'RS 130'): _count_exp(0.2), # 0.6 \\n } \\n \\n def dtest_predictions_created(self):\\n \\\"\\\"\\\"Test that the predictions were created as expected\\\"\\\"\\\"\\n \\n # filter aggregated instances for my recommender\\n pred_instances = RelationshipPredictionInstance.objects.filter(\\n recommender=ShoeRecommender._get_recommender_model())\\n \\n # assert it has the expected length \\n eq_(pred_instances.count(), len(self.EXPECTED_PREDICTIONS), \\n \\\"expected: %d, obtained: %d, obtained instances: %s\\\" % \\\\\\n (len(self.EXPECTED_PREDICTIONS), pred_instances.count(), pred_instances))\\n \\n for pred_inst in pred_instances.iterator():\\n pair1 = (pred_inst.subject_object1.name, pred_inst.subject_object2.name) \\n \\n # try getting the instance from expected in both directions\\n assert self.EXPECTED_PREDICTIONS.has_key(pair1), \\\\\\n \\\"Unexpected prediction between '%s' and '%s' expectancy: %f\\\" % \\\\\\n (pair1 + (pred_inst.expectancy, ))\\n expected_prediction = self.EXPECTED_PREDICTIONS[pair1]\\n\\n # assert the expectancy is as expected \\n assert_almost_equal(pred_inst.expectancy, expected_prediction, PLACES, \\n \\\"Prediction is '%f' should be '%f' for the pair %s, %s\\\" % \\\\\\n ((pred_inst.expectancy, expected_prediction) + pair1)) \\n \\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5023733973503113,\n \"alphanum_fraction\": 0.5142405033111572,\n \"avg_line_length\": 30.600000381469727,\n \"blob_id\": \"100bdb93f6cbb8886489104c5cf6d157b1ba74eb\",\n \"content_id\": \"22af5dba242c3bd59bcfac3c2710b10a4c089a63\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1264,\n \"license_type\": \"no_license\",\n \"max_line_length\": 79,\n \"num_lines\": 40,\n \"path\": \"/UnresystCD/code/adapter/unresyst/combinator/confidence_factor.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Combinator using confidence factor calculus.\\\"\\\"\\\"\\n\\nfrom base import BaseCombinator\\n\\n\\n \\n\\nclass ConfidenceFactorCombinator(BaseCombinator):\\n \\\"\\\"\\\"A combinator using weighted average\\n \\\"\\\"\\\"\\n \\n def _combine(self, combination_elements, ResultClass):\\n \\\"\\\"\\\"See the base class for documentation\\\"\\\"\\\"\\n \\n res_exp = combination_elements[0].get_expectancy()\\n \\n # go through the combination elements\\n for ce in combination_elements[1:]:\\n \\n # convert to confidence factors\\n res_cf = 2 * res_exp - 1\\n ce_cf = 2 * ce.get_expectancy() -1\\n \\n # count the confidence factor combination\\n if res_cf > 0 and ce_cf > 0: \\n comb_cf = res_cf + ce_cf * (1 - res_cf)\\n \\n elif res_cf < 0 and ce_cf < 0:\\n comb_cf = res_cf + ce_cf * (1 + res_cf)\\n \\n else:\\n comb_cf = (res_cf * ce_cf) / (1 - min(abs(res_cf), abs(ce_cf)))\\n \\n # and back to expectancy\\n res_exp = (comb_cf + 1) / 2\\n \\n\\n desc = self._concat_descriptions(combination_elements)\\n \\n return ResultClass(expectancy=res_exp, description=desc)\\n\"\n },\n {\n \"alpha_fraction\": 0.5262203812599182,\n \"alphanum_fraction\": 0.5283953547477722,\n \"avg_line_length\": 32.290321350097656,\n \"blob_id\": \"ef402b712baa9bec74f7b91d52c575dc73ac448b\",\n \"content_id\": \"fef96f14982817083f7791d2f051bc5be587bcdb\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 8276,\n \"license_type\": \"no_license\",\n \"max_line_length\": 98,\n \"num_lines\": 248,\n \"path\": \"/code/adapter/unresyst/recommender/evaluation.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The classes related to recommender evaluation\\\"\\\"\\\"\\nimport os.path\\nimport codecs\\n\\nfrom unresyst.exceptions import EmptyTestSetError\\nfrom unresyst.constants import *\\n\\nfrom settings import LOG_DIRECTORY\\n\\nclass BaseEvaluator(object):\\n \\\"\\\"\\\"The base class for all evaluators\\\"\\\"\\\"\\n \\n EvaluationPairModel = None\\n \\\"\\\"\\\"The model of the class containing testing pairs. To be overriden.\\n A EvaluationPair subclass.\\n \\\"\\\"\\\"\\n \\n prediction_metric = None\\n \\\"\\\"\\\"A function taking the test pair model that is called on the evaluated pairs\\n to obtain a number \\n \\\"\\\"\\\"\\n \\n recommendation_metric = None\\n \\\"\\\"\\\"A function taking the test pair model and number of desired recommendations,\\n that is called on the evaluated pair model to obtain a number \\n \\\"\\\"\\\"\\n \\n # methods to be called before the build:\\n # \\n \\n @classmethod\\n def select_evaluation_pairs(cls):\\n \\\"\\\"\\\"Select the pairs for recommender evaluation.\\n \\n The selected pairs will be added to the EvaluationPairModel and\\n removed from the models where they are now\\n \\\"\\\"\\\"\\n cls.EvaluationPairModel.select()\\n \\n \\n @classmethod\\n def export_evaluation_pairs(cls, filename):\\n \\\"\\\"\\\"Export the evaluation pair ids to a csv \\n file of the given name.\\n \\n @type filename: str\\n @param filename: the full path to the file\\n \\n @raise FileNotExists and other file open errors.\\n \\\"\\\"\\\"\\n \\n with open(filename, 'w') as f:\\n \\n # export the relationships to the given file\\n cls.EvaluationPairModel.export(f)\\n\\n \\n # methods to be called after the build:\\n # \\n \\n @classmethod\\n def _get_cleared_pairs(cls):\\n \\\"\\\"\\\"Get the test pairs, check whether they aren't empty and\\n clear the results.\\n \\n @rtype: QuerySet\\n @return: the test pairs\\n \\n @raise EmptyTestSetError: if the set is empty\\n \\\"\\\"\\\"\\n # get the dataset and test if not empty\\n qs_pairs = cls.EvaluationPairModel.objects.all() \\n \\n # remove all previous success and obtained expectancies\\n qs_pairs.update(obtained_expectancy=None, is_successful=None) \\n \\n if not qs_pairs:\\n raise EmptyTestSetError(\\\"Call the select_validation_pairs()\\\"+ \\\\\\n \\\" method first\\\") \\n \\n return qs_pairs\\n\\n @staticmethod\\n def _open_logfile(directory, filename):\\n \\\"\\\"\\\"Open a log file\\n \\n @rtype: file\\n \\\"\\\"\\\"\\n if directory and filename:\\n full_name = os.path.join(directory, filename)\\n return codecs.open(full_name, mode='w', encoding='utf-8')\\n\\n return None \\n\\n @classmethod\\n def evaluate_predictions(cls, recommender, save_predictions=False):\\n \\\"\\\"\\\"Evaluate each evaluation pair by calling the predict_relationship\\n method.\\n \\n The results are written to the pair model\\n \\n @type recommender: Recommender\\n @param recommender: the built recommender on which the predictions\\n should be evaluated \\n \\n @type save_predictions: bool\\n @param save_predictions: should the generated predictions be saved to db?\\n \\n @rtype: float\\n @return: the result of the metric \\n\\n @raise EmptyTestSetError: if the test set is empty \\n \\\"\\\"\\\"\\n # if the settings are given, open a log file and log\\n fpreds = cls._open_logfile(LOG_DIRECTORY, LOG_PREDICTIONS_FILENAME)\\n fhits = cls._open_logfile(LOG_DIRECTORY, LOG_HITS_FILENAME)\\n \\n # get the pairs\\n qs_pairs = cls._get_cleared_pairs()\\n all_count = qs_pairs.count() \\n\\n # initialize \\n i = 0\\n succ_count = 0\\n non_triv_count = 0\\n \\n print \\\"Processing %d pairs...\\\" % all_count\\n \\n # go through the pairs\\n for pair in qs_pairs.iterator():\\n \\n # evaluate\\n prediction = recommender.predict_relationship(pair.subj, pair.obj, save_predictions)\\n \\n # log\\n if fpreds:\\n fpreds.write(u\\\"%s\\\\n\\\" % prediction.__unicode__())\\n \\n pair.obtained_expectancy = prediction.expectancy\\n pair.is_successful = pair.get_success()\\n pair.save() \\n \\n i += 1\\n \\n if i % 100 == 0:\\n print \\\"%d pairs processed\\\" % i\\n \\n if pair.is_successful:\\n succ_count += 1\\n \\n if fhits:\\n fhits.write(u'%s\\\\n' % prediction.__unicode__())\\n \\n if pair.obtained_expectancy < TRIVIAL_EXPECTANCY:\\n non_triv_count += 1\\n \\n # this isn't very best practice-following but we don't care for file corruption\\n if fhits:\\n fhits.close()\\n \\n if fpreds:\\n fpreds.close()\\n \\n # count and print the success rate \\n cls.success_rate = float(succ_count)/ all_count\\n\\n print \\\"Success rate: %f (%d/%d)\\\" % (cls.success_rate, succ_count, all_count)\\n print \\\"%d of %d successful predictions were non-trivial.\\\" % (non_triv_count, succ_count)\\n \\n # count the metric \\n return cls.prediction_metric()\\n \\n @classmethod\\n def evaluate_recommendations(cls, recommender, count):\\n \\\"\\\"\\\"Evaluate recommendations obtained for the subjects in the test\\n pairs.\\n \\n @type recommender: Recommender\\n @param recommender: the built recommender on which the predictions\\n should be evaluated\\n \\n @type count: int\\n @param count: the number of recommendations to get \\n \\n @rtype: float\\n @return: the result of the metric \\n \\\"\\\"\\\"\\n # if the settings are given, open a log file and log \\n fhits = cls._open_logfile(LOG_DIRECTORY, LOG_HITS_FILENAME)\\n frecs = cls._open_logfile(LOG_DIRECTORY, LOG_RECOMMENDATIONS_FILENAME)\\n \\n # get the cleared pair dataset \\n qs_pairs = cls._get_cleared_pairs()\\n \\n # take ids of subjects from the test pairs \\n subj_id_list = qs_pairs.values_list('subj__pk',flat=True).distinct()\\n \\n # take the whole subjects\\n qs_subjects = recommender.subjects.filter(pk__in=subj_id_list) \\n\\n i = 0 \\n hit_count = 0 \\n for subj in qs_subjects: \\n \\n i += 1\\n \\n if i % 10 == 0:\\n print \\\"%d subjects processed\\\" % i\\n \\n # for each subject in the test pair get its recommendations\\n recommendations = recommender.get_recommendations(subj, count) \\n \\n \\n # mark each recommended object that is in the test set as successful\\n #\\n for rec in recommendations:\\n\\n # if it should be logged, log it\\n if frecs:\\n frecs.write(u'%s\\\\n' % rec.__unicode__())\\n \\n obj = rec.object_\\n \\n # filter the test pair\\n qs_pair = cls.EvaluationPairModel.objects.filter(subj=subj, obj=obj)\\n \\n # if it's in the test pair mark it as successful\\n for pair in qs_pair:\\n \\n hit_count += 1\\n \\n pair.is_successful = True\\n pair.save()\\n \\n if fhits:\\n fhits.write(u'%s\\\\n' % rec.__unicode__())\\n\\n # this isn't very best practice-following but we don't care for file corruption\\n if fhits:\\n fhits.close()\\n \\n if frecs:\\n frecs.close()\\n \\n print \\\"%d hits recorded, counting metric...\\\" % hit_count\\n \\n # count the metric \\n return cls.recommendation_metric(count)\\n \\n \\n \\n\\n\"\n },\n {\n \"alpha_fraction\": 0.7510204315185547,\n \"alphanum_fraction\": 0.7530612349510193,\n \"avg_line_length\": 36.69230651855469,\n \"blob_id\": \"dd47d8c9649e3b251edb94912656715064a4e277\",\n \"content_id\": \"6a020e161580ee21d4d2238f749bfa815261ef8d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 490,\n \"license_type\": \"no_license\",\n \"max_line_length\": 75,\n \"num_lines\": 13,\n \"path\": \"/UnresystCD/code/adapter/flixster/mahout_recommender.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The representation of the Mahout recommender in Unresyst.\\\"\\\"\\\"\\n\\nfrom unresyst.recommender.external_recommender import ExternalRecommender\\nfrom recommender import MovieRecommender\\n\\nclass MahoutMovieRecommender(ExternalRecommender, MovieRecommender):\\n \\\"\\\"\\\"An external artist recommender - both novel and non-novel artists\\\"\\\"\\\"\\n \\n name = \\\"Mahout movie recommender\\\"\\n \\\"\\\"\\\"The name\\\"\\\"\\\"\\n \\n explicit_rating_rule = MovieRecommender.rules[0]\\n \\\"\\\"\\\"The first one is the explicit one\\\"\\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.6203219890594482,\n \"alphanum_fraction\": 0.6234704852104187,\n \"avg_line_length\": 32.91990280151367,\n \"blob_id\": \"042d921424fee5827dca328ed2ba70d90411e1bd\",\n \"content_id\": \"a1b0dbb96e6b153bf61ff07f114a55cc9171088c\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 13975,\n \"license_type\": \"no_license\",\n \"max_line_length\": 100,\n \"num_lines\": 412,\n \"path\": \"/UnresystCD/code/adapter/unresyst/models/abstractor.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Models which instances are created by the abstractor package.\\n\\nA representation of each defined role and relationship.\\n\\\"\\\"\\\"\\n\\nfrom django.db import models\\n\\nfrom unresyst.constants import *\\nfrom base import BaseRelationshipInstance, ContentTypeModel, \\\\\\n BaseRelationshipDefinition\\nfrom unresyst.exceptions import InvalidParameterError\\n\\ndef _count_expectancy(is_positive, weight, confidence=1):\\n \\\"\\\"\\\"Count the expectancy for the instance as 1/2 +- weighted_confidence/2 \\n .. depending on is_positive.\\n \\n @type is_positive: bool\\n @param is_positive: is the rule/relationship/bias positive for the \\n predicted_relationship?\\n \\n @type weight: float\\n @param weight: the static weight of the rule/relationship/bias\\n\\n @type confidence: float\\n @param confidence: the confidence of the rule/relationship/bias depending\\n on the entity\\n \\n @rtype: float\\n @return: the probability of the predicted_relationship appearing between\\n the entities in the pair.\\n \\\"\\\"\\\"\\n \\n # get factor \\n factor = 1 if is_positive else -1 \\n \\n return 0.5 + factor * ((weight * confidence) / 2)\\n\\n\\n# Definitions:\\n# \\n\\nclass PredictedRelationshipDefinition(BaseRelationshipDefinition):\\n \\\"\\\"\\\"A definition of a the predicted relationship\\\"\\\"\\\"\\n\\n class Meta:\\n app_label = 'unresyst' \\n \\nclass ExplicitRuleDefinition(BaseRelationshipDefinition):\\n \\\"\\\"\\\"A definition of an explicit feedback rule\\\"\\\"\\\"\\n \\n class Meta:\\n app_label = 'unresyst' \\n\\n\\nclass RuleRelationshipDefinition(BaseRelationshipDefinition):\\n \\\"\\\"\\\"A definition of a rule/relationship. Represents the rule/relationship\\n given by the domain expert in the database.\\\"\\\"\\\" \\n \\n weight = models.FloatField()\\n \\\"\\\"\\\"The weight of the rule/relationship. A number from [0, 1].\\\"\\\"\\\"\\n \\n relationship_type = models.CharField(max_length=MAX_LENGTH_RELATIONSHIP_TYPE, \\n choices=RELATIONSHIP_TYPE_CHOICES)\\n \\\"\\\"\\\"A string indicating whether it's a subject, object or both.s/o/so\\\"\\\"\\\"\\n \\n is_positive = models.BooleanField()\\n \\\"\\\"\\\"Is the rule/relationship positive for the predicted_relationship?\\\"\\\"\\\"\\n\\n class Meta:\\n app_label = 'unresyst' \\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\" \\n ret = super(RuleRelationshipDefinition, self).__unicode__()\\n\\n return ret + \\\"%f, positive: %s\\\" % (self.weight, self.is_positive)\\n \\n# instances:\\n#\\n\\nclass RelationshipInstance(BaseRelationshipInstance, ContentTypeModel):\\n \\\"\\\"\\\"The relationship between two subject/objects.\\n \\n All relationships are \\\"symetrical\\\" in sense that if there is \\n a relationship in the given direction there can't be one in the opposite.\\n \\n Weight of the relationship is in the definition.\\n \\n The model also contains the predicted relationship instances.\\n \\\"\\\"\\\"\\n \\n definition = models.ForeignKey('unresyst.BaseRelationshipDefinition')\\n \\\"\\\"\\\"The definition of the relationship. Pointing to the base, because\\n it has to be also for the predicted relationship instances.\\n \\\"\\\"\\\"\\n \\n additional_unique = ('definition', )\\n \\\"\\\"\\\"There can be multiple pairs for one recommender\\\"\\\"\\\"\\n \\n class Meta:\\n app_label = 'unresyst' \\n \\n unique_together = ('subject_object1', 'subject_object2', 'definition')\\n \\\"\\\"\\\"For each definition there can be only one subject-object pair.\\\"\\\"\\\"\\n\\n def get_expectancy(self, _redirect_to_leaf=True):\\n \\\"\\\"\\\"Get the instance expectancy counted \\n \\n @type redirect_to_leaf: bool\\n @param redirect_to_leaf: if True, the call is redirected to the leaf class.\\n When called from outside _redirect_to_leaf should always be True, unless\\n it's sure it's called on the leaf class instance.\\n \\n @rtype: float from [0, 1]\\n @return: the aggregated expectancy.\\n \\\"\\\"\\\"\\n \\n # redirect to leaf or not\\n if _redirect_to_leaf:\\n return self.as_leaf_class().get_expectancy(_redirect_to_leaf=False)\\n \\n # get the whole object for definition\\n leaf_definition = self.definition.as_leaf_class() \\n \\n return _count_expectancy(\\n is_positive=leaf_definition.is_positive,\\n weight=leaf_definition.weight)\\n\\n @classmethod\\n def filter_predicted(cls, recommender_model):\\n \\\"\\\"\\\"Get queryset of the instances of the predicted relationship.\\n \\n @type recommender_model: models.common.Recommender\\n @param recommender_model: the recommender whose instances should\\n be aggregated. \\n\\n @rtype: QuerySet\\n @returns: instances of the predicted relationship for the given\\n recommender\\n \\\"\\\"\\\"\\n # get the predicted relationship definition\\n pred_rel_def = PredictedRelationshipDefinition.objects.get(\\n recommender=recommender_model)\\n \\n return cls.objects.filter(definition=pred_rel_def) \\n\\n\\nclass ExplicitRuleInstance(BaseRelationshipInstance):\\n \\\"\\\"\\\"The explicit preference of a subject to an object.\\n \\n The relationship isn't weighted. \\n \\\"\\\"\\\"\\n \\n definition = models.ForeignKey('unresyst.ExplicitRuleDefinition')\\n \\\"\\\"\\\"The definition of the relationship. \\n \\\"\\\"\\\"\\n\\n expectancy = models.FloatField()\\n \\\"\\\"\\\"The normalized explicit preference of the given subject to the given \\n object.\\n A number from [0, 1].\\n \\\"\\\"\\\"\\n \\n additional_unique = ('definition', )\\n \\\"\\\"\\\"There can be multiple pairs for one recommender\\\"\\\"\\\"\\n \\n class Meta:\\n app_label = 'unresyst' \\n \\n unique_together = ('subject_object1', 'subject_object2', 'definition')\\n \\\"\\\"\\\"For each definition there can be only one subject-object pair.\\\"\\\"\\\"\\n\\n def get_expectancy(self, _redirect_to_leaf=True):\\n \\\"\\\"\\\"Get the instance expectancy counted \\n \\n @type redirect_to_leaf: bool\\n @param redirect_to_leaf: to stay compatible with other classes\\n \\n @rtype: float from [0, 1]\\n @return: the expectancy.\\n \\\"\\\"\\\" \\n return self.expectancy \\n\\n\\nclass RuleInstance(RelationshipInstance):\\n \\\"\\\"\\\"The rule applied to a pair of subjects/objects.\\\"\\\"\\\" \\n \\n confidence = models.FloatField()\\n \\\"\\\"\\\"The confidence of the given rule being positive/negative (depending on\\n definition.is_positive) in means of the predicted_relationship\\n A number from [0, 1].\\n \\\"\\\"\\\"\\n \\n def get_expectancy(self, _redirect_to_leaf=True):\\n \\\"\\\"\\\"Get the instance expectancy counted as 1/2 +- weight*confidence/2 .. depending on whether\\n the relationship is positive.\\n\\n @type redirect_to_leaf: bool\\n @param redirect_to_leaf: if True, the call is redirected to the leaf class.\\n When called from outside _redirect_to_leaf should always be True, unless\\n it's sure it's called on the leaf class instance. \\n \\n @rtype: float from [0, 1]\\n @return: the aggregated expectancy.\\n \\\"\\\"\\\"\\n # get the whole object for definition\\n leaf_definition = self.definition.as_leaf_class()\\n \\n return _count_expectancy(\\n is_positive=leaf_definition.is_positive, \\n weight=leaf_definition.weight,\\n confidence=self.confidence) \\n \\n class Meta:\\n app_label = 'unresyst'\\n \\n# clusters:\\n# \\n\\nclass ClusterSet(models.Model):\\n \\\"\\\"\\\"A set of clusters defined by the user. E.g. 'Gender'.\\\"\\\"\\\"\\n\\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"A textual characterization of the cluster set.\\\"\\\"\\\" \\n \\n recommender = models.ForeignKey('unresyst.Recommender')\\n \\\"\\\"\\\"The recommender to which the cluster set belongs.\\\"\\\"\\\" \\n\\n entity_type = models.CharField(max_length=MAX_LENGTH_ENTITY_TYPE, \\\\\\n choices=ENTITY_TYPE_CHOICES)\\n \\\"\\\"\\\"A string indicating whether the contained clusters are for subjects,\\n objects or both.s/o/so\\\"\\\"\\\"\\n\\n weight = models.FloatField()\\n \\\"\\\"\\\"The weight of the similarity inferred from two subject/objects \\n belonging to one cluster. A number from [0, 1].\\n \\\"\\\"\\\"\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n \\n class Meta:\\n app_label = 'unresyst' \\n \\n unique_together = ('name', 'recommender')\\n \\\"\\\"\\\"There can be only one cluster set with the given name for \\n a recommender.\\n \\\"\\\"\\\"\\n\\n\\nclass Cluster(models.Model):\\n \\\"\\\"\\\"A cluster (class of subjects/objects) defined by the user.\\n E.g. 'Female'\\\"\\\"\\\"\\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"A textual characterization of the cluster\\\"\\\"\\\"\\n \\n cluster_set = models.ForeignKey('unresyst.ClusterSet')\\n \\\"\\\"\\\"The set the cluster belongs to\\\"\\\"\\\"\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return u\\\"%s - %s\\\" % (self.cluster_set, self.name)\\n \\n class Meta:\\n app_label = 'unresyst' \\n \\n unique_together = ('name', 'cluster_set')\\n \\\"\\\"\\\"There can be only one cluster cluster set of the given name in the \\n cluster set.\\n \\\"\\\"\\\"\\n\\nclass ClusterMember(models.Model):\\n \\\"\\\"\\\"A membership of a subject/object in the cluster.\\\"\\\"\\\"\\n \\n cluster = models.ForeignKey('unresyst.Cluster')\\n \\\"\\\"\\\"The cluster\\\"\\\"\\\"\\n \\n member = models.ForeignKey('unresyst.SubjectObject')\\n \\\"\\\"\\\"The cluster member\\\"\\\"\\\"\\n \\n description = models.TextField(default='', blank=True)\\n \\\"\\\"\\\"The description of the membership.\\\"\\\"\\\" \\n \\n confidence = models.FloatField()\\n \\\"\\\"\\\"The confidence of the member belonging to the cluster.\\n A number from [0, 1].\\n \\\"\\\"\\\"\\n \\n @classmethod\\n def get_pair_expectancy(cls, cluster_member_pair):\\n \\\"\\\"\\\"Return the expectancy that will be used for similarity counting \\n for cluster members\\n \\n @type cluster_member_pair: a tuple (pair) of ClusterMember\\n @param cluster_member_pair: cluster members for counting the similarity\\n of the members.\\n \\n @rtype: float\\n @return: the expectancy that the members are similar\\n \\\"\\\"\\\"\\n cm1, cm2 = cluster_member_pair\\n \\n # if they aren't from the same cluster raise an error\\n if cm1.cluster != cm2.cluster:\\n raise InvalidParameterError(\\n message=\\\"The members to combine don't belong to the same cluster.\\\",\\n recommender=cm1.cluster.cluster_set.recommender, \\n parameter_name='cluster_member_pair', \\n parameter_value=cluster_member_pair)\\n \\n conf = cm1.confidence * cm2.confidence\\n \\n return _count_expectancy(\\n is_positive=True,\\n weight=cm1.cluster.cluster_set.weight,\\n confidence=conf)\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return u\\\"(%s, %s): %f\\\" % (self.cluster, self.member, self.confidence)\\n \\n \\n class Meta:\\n app_label = 'unresyst' \\n \\n unique_together = ('cluster', 'member')\\n \\\"\\\"\\\"There can be only one membership for a subjectobject in the given \\n cluster.\\n \\\"\\\"\\\" \\n\\n# Bias:\\n#\\n \\nclass BiasDefinition(models.Model):\\n \\\"\\\"\\\"The definition of a bias\\\"\\\"\\\"\\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"The name of the bias\\\"\\\"\\\" \\n \\n recommender = models.ForeignKey('unresyst.Recommender')\\n \\\"\\\"\\\"The recommender to which the definition belongs. \\n \\\"\\\"\\\"\\n \\n weight = models.FloatField()\\n \\\"\\\"\\\"The weight of the bias\\n belonging to one cluster. A number from [0, 1].\\n \\\"\\\"\\\"\\n \\n entity_type = models.CharField(max_length=MAX_LENGTH_ENTITY_TYPE, \\\\\\n choices=ENTITY_TYPE_CHOICES)\\n \\\"\\\"\\\"A string indicating whether the contained bias is for subjects,\\n objects or both.s/o/so\\\"\\\"\\\"\\n \\n is_positive = models.BooleanField()\\n \\\"\\\"\\\"Is the bias positive (adding a probability) for the predicted_relationship?\\\"\\\"\\\"\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance.\\\"\\\"\\\"\\n return self.name\\n \\n class Meta:\\n app_label = 'unresyst' \\n unique_together = ('name', 'recommender')\\n \\\"\\\"\\\"There can be only one bias definition with the given name for \\n a recommender.\\n \\\"\\\"\\\"\\n \\n \\nclass BiasInstance(models.Model):\\n \\\"\\\"\\\"A bias for a particular subject/object.\\\"\\\"\\\" \\n \\n confidence = models.FloatField()\\n \\\"\\\"\\\"The confidence of the given bias being positive/negative (depending on\\n definition.is_positive) in means of the predicted_relationship\\n A number from [0, 1].\\n \\\"\\\"\\\" \\n \\n subject_object = models.ForeignKey('unresyst.SubjectObject')\\n \\\"\\\"\\\"The biased subject/object.\\\"\\\"\\\"\\n \\n definition = models.ForeignKey('unresyst.BiasDefinition')\\n \\\"\\\"\\\"The definition of the bias.\\n \\\"\\\"\\\"\\n \\n description = models.TextField(default='', blank=True)\\n \\\"\\\"\\\"The filled description of the bias.\\\"\\\"\\\" \\n\\n \\n class Meta:\\n app_label = 'unresyst' \\n unique_together = ('definition', 'subject_object')\\n \\\"\\\"\\\"There can be only one bias for each definition on one subjectobject\\n a recommender.\\n \\\"\\\"\\\" \\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance.\\\"\\\"\\\"\\n return u\\\"%s: %s\\\" % (self.definition, self.subject_object)\\n\\n def get_expectancy(self):\\n \\\"\\\"\\\"Get the expectancy (probability that the subjectobject will be \\n in the predicted_relationship, according to the bias).\\n \\n @rtype: float\\n @return: expectancy\\n \\\"\\\"\\\"\\n return _count_expectancy(\\n is_positive=self.definition.is_positive,\\n weight=self.definition.weight,\\n confidence=self.confidence)\\n\"\n },\n {\n \"alpha_fraction\": 0.6214788556098938,\n \"alphanum_fraction\": 0.6214788556098938,\n \"avg_line_length\": 31.75,\n \"blob_id\": \"85e0020087a2ab13f553e98f84f39e95e03b0b20\",\n \"content_id\": \"8f12c0e536194f11e08ce5be4657ed527d83d9e0\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1704,\n \"license_type\": \"no_license\",\n \"max_line_length\": 83,\n \"num_lines\": 52,\n \"path\": \"/code/adapter/unresyst/aggregator/base.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The module defines base class for the aggregator package.\\\"\\\"\\\"\\n\\nclass BaseAggregator(object):\\n \\\"\\\"\\\"The base (abstract) class for all aggregators. Defines the interface.\\\"\\\"\\\" \\n \\n # Build phase:\\n #\\n\\n def __init__(self, combinator=None):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n self.combinator = combinator\\n \\\"\\\"\\\"The combinator that should be used during aggregating\\\"\\\"\\\"\\n\\n \\n def aggregate_rules_relationships(cls, recommender_model):\\n \\\"\\\"\\\"Aggregate the rule and relationship instances.\\n \\n Aggregates instances of the relationships and rules from \\n the recommender_model to create AggregatedRelationshipInstance\\n instances. Instances are saved to db.\\n \\n @type recommender_model: models.common.Recommender\\n @param recommender_model: the recommender whose instances should\\n be aggregated. \\n \\\"\\\"\\\"\\n pass\\n \\n def aggregate_biases(cls, recommender_model):\\n \\\"\\\"\\\"Aggregate bias instances.\\n \\n Aggregates instances of the biases from the recommender_model \\n to create AggregatedBiasInstance instances. Instances are saved to db.\\n \\n @type recommender_model: models.common.Recommender\\n @param recommender_model: the recommender whose instances should\\n be aggregated. \\n \\\"\\\"\\\"\\n pass\\n\\n\\n # Update phase:\\n # \\n \\n @classmethod\\n def update(cls, recommender, instance_changes):\\n \\\"\\\"\\\"Apdate the aggregated relationships according to the instance \\n changes.\\n \\n @rtype: uvidime\\n @return: the changes performed on the aggregated relationships. \\n \\\"\\\"\\\"\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5711066722869873,\n \"alphanum_fraction\": 0.6090385317802429,\n \"avg_line_length\": 31.73770523071289,\n \"blob_id\": \"0ce8beca8e8c9058007d21a35e17f232ffdfa132\",\n \"content_id\": \"cc21b4cc085d14fb0118532c1aa738c285744b73\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 7988,\n \"license_type\": \"no_license\",\n \"max_line_length\": 145,\n \"num_lines\": 244,\n \"path\": \"/code/adapter/lastfm/models.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The models for the last.fm datasets.\\\"\\\"\\\"\\n\\n\\n\\\"\\\"\\\"\\nuser_000001\\t2009-05-04T23:08:57Z\\tf1b1cf71-bd35-4e99-8624-24a6e15f133a\\tDeep Dish\\t\\tFuck Me Im Famous (Pacha Ibiza)-09-28-2007\\nuser_000001\\t2009-04-01T14:24:52Z\\t6c639b1f-5390-4ec9-b3f8-7afe1562e107\\tLady Alma\\t\\tRunning For Nothing [(Instrumental)]/(Instrumental)\\nuser_000001\\t2009-02-05T22:58:19Z\\t440cb8ef-cee3-4711-8408-b1bd6e93f390\\tTowa Tei\\tc605c224-245b-4107-b9cd-ab1b2c6adb0a\\tMind Wall (Feat. Miho Hatori)\\n\\n#id\\tgender\\tage\\tcountry\\tregistered\\nuser_000001\\tm\\t\\tJapan\\tAug 13, 2006\\nuser_000002\\tf\\t\\tPeru\\tFeb 24, 2006\\nuser_000003\\tm\\t22\\tUnited States\\tOct 30, 2005\\nuser_000004\\tf\\t\\t\\tApr 26, 2006\\nuser_000005\\tm\\t\\tBulgaria\\tJun 29, 2006\\nuser_000006\\t\\t24\\tRussian Federation\\tMay 18, 2006\\nuser_000007\\tf\\t\\tUnited States\\tJan 22, 2006\\nuser_000008\\tm\\t23\\tSlovakia\\tSep 28, 2006\\nuser_000009\\tf\\t19\\tUnited States\\tJan 13, 2007\\nuser_000010\\tm\\t19\\tPoland\\tMay 4, 2006\\nuser_000011\\tm\\t21\\tFinland\\tSep 8, 2005\\n\\\"\\\"\\\"\\n\\nfrom django.db import models\\nfrom django.db.models import Min\\n\\nfrom unresyst.models import BaseEvaluationPair\\n\\n\\n\\nfrom constants import *\\n\\nclass User(models.Model):\\n \\\"\\\"\\\"The user\\\"\\\"\\\"\\n \\n gender = models.CharField(max_length=1, default='')\\n \\\"\\\"\\\"The english name of the enumerator. Should be translated by ugettext before \\n displaying.\\\"\\\"\\\"\\n \\n age = models.PositiveIntegerField(null=True, default=None)\\n \\\"\\\"\\\"The age of the user.\\\"\\\"\\\"\\n \\n registered = models.DateTimeField(null=True, default=None)\\n \\\"\\\"\\\"Date when the user was registered\\\"\\\"\\\"\\n\\n country = models.ForeignKey('Country', null=True)\\n \\\"\\\"\\\"The country the user is from\\\"\\\"\\\" \\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return 'user_%d' % self.id \\n \\n \\nclass Country(models.Model):\\n \\\"\\\"\\\"A model for a country.\\\"\\\"\\\" \\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME, unique=True)\\n \\\"\\\"\\\"The name of the city.\\\"\\\"\\\" \\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n \\n \\nclass Track(models.Model):\\n \\\"\\\"\\\"A track (song)\\\"\\\"\\\"\\n \\n guid = models.CharField(unique=True, max_length=40)\\n \\\"\\\"\\\"The id of the track in the musicbrainz database\\\"\\\"\\\"\\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"The name of the track.\\\"\\\"\\\"\\n\\n artist = models.ForeignKey('Artist')\\n \\\"\\\"\\\"The author of the track\\\"\\\"\\\"\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return u\\\"%s: %s\\\" % (self.artist, self.name)\\n\\n\\nclass Artist(models.Model):\\n \\\"\\\"\\\"An artist (band or singer)\\\"\\\"\\\"\\n \\n guid = models.CharField(unique=True, max_length=40)\\n \\\"\\\"\\\"The id of the singer in the musicbrainz database\\\"\\\"\\\"\\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME)\\n \\\"\\\"\\\"The name of the track.\\\"\\\"\\\"\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n\\n\\nclass Scrobble(models.Model):\\n \\\"\\\"\\\"A representation of the fact that a user has played a track\\\"\\\"\\\"\\n \\n user = models.ForeignKey('User')\\n \\\"\\\"\\\"The user\\\"\\\"\\\"\\n \\n track = models.ForeignKey('Track')\\n \\\"\\\"\\\"The played track\\\"\\\"\\\"\\n \\n timestamp = models.DateTimeField()\\n \\\"\\\"\\\"The date and time the track was played\\\"\\\"\\\"\\n \\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return u\\\"(%s, %s)\\\" % (self.user, self.track)\\n\\n\\nclass ArtistTag(models.Model):\\n \\\"\\\"\\\"A representation a artist being tagged by the tag\\\"\\\"\\\"\\n \\n artist = models.ForeignKey('Artist')\\n \\\"\\\"\\\"The artist\\\"\\\"\\\"\\n \\n tag = models.ForeignKey('Tag')\\n \\\"\\\"\\\"The tag\\\"\\\"\\\" \\n \\n count = models.PositiveIntegerField()\\n \\\"\\\"\\\"How many times the artist was tagged by the tag\\\"\\\"\\\"\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return u\\\"(%s, %s: %d)\\\" % (self.artist, self.tag, self.count) \\n \\n\\n \\nclass Tag(models.Model):\\n \\\"\\\"\\\"A representation of a tag (label) in the system\\\"\\\"\\\"\\n \\n name = models.CharField(max_length=MAX_LENGTH_NAME, unique=True)\\n \\\"\\\"\\\"The label.\\\"\\\"\\\"\\n \\n gender_specific = models.CharField(max_length=1, default='')\\n \\\"\\\"\\\"If the tag is preferred by a gender, the gender is here\\\"\\\"\\\"\\n\\n def __unicode__(self):\\n \\\"\\\"\\\"Return a printable representation of the instance\\\"\\\"\\\"\\n return self.name \\n\\n\\nclass BaseArtistEvaluationPair(BaseEvaluationPair):\\n \\\"\\\"\\\"An abstract class for the artist evaluators\\\"\\\"\\\"\\n\\n subj = models.ForeignKey('lastfm.User')\\n \\\"\\\"\\\"The subject\\\"\\\"\\\"\\n \\n obj = models.ForeignKey('lastfm.Artist')\\n \\\"\\\"\\\"The object\\\"\\\"\\\"\\n \\n test_ratio = 0.2\\n \\\"\\\"\\\"The ratio of pairs to select to test pairs\\\"\\\"\\\"\\n\\n class Meta:\\n abstract = True \\n \\n @classmethod \\n def select(cls, i=0):\\n \\\"\\\"\\\"See the base class for the documentation.\\\"\\\"\\\"\\n all_count = Scrobble.objects.count() \\n test_count = int(cls.test_ratio * all_count ) \\n train_count = all_count - test_count\\n \\n test_pairs = Scrobble.objects.order_by('-timestamp')[:test_count]\\n\\n min_stamp_test = test_pairs.aggregate(Min('timestamp'))['timestamp__min']\\n train_pairs = Scrobble.objects.filter(timestamp__lt=min_stamp_test)\\n \\n # take 1/5 of the scrobbles, remove them and put them to test data \\n # save to test, remove from build\\n for scrobble in test_pairs.iterator():\\n\\n # create the test pair for the radio recommender\\n val_pair = ArtistEvalPair(\\n subj=scrobble.user,\\n obj=scrobble.track.artist,\\n expected_expectancy=EXPECTED_EXPECTANCY_LISTENED,\\n )\\n \\n # if not in train pairs \\n if not train_pairs.filter(user=val_pair.subj, track__artist=val_pair.obj).exists():\\n\\n # and not already in novel test_pairs\\n if not NovelArtistEvalPair.objects.filter(subj=val_pair.subj, obj=val_pair.obj).exists():\\n \\n # add it also to novel test pairs\\n n_val_pair = NovelArtistEvalPair(\\n subj=val_pair.subj,\\n obj=val_pair.obj,\\n expected_expectancy=val_pair.expected_expectancy\\n )\\n n_val_pair.save()\\n\\n val_pair.save()\\n \\n # so wie so delete\\n scrobble.delete()\\n \\n nontrivial_count = NovelArtistEvalPair.objects.count()\\n \\n print \\\"Test pairs selected from total %d pairs, %d of %d are non-trivial\\\" % (all_count, nontrivial_count, test_count) \\n\\n def get_success(self):\\n \\\"\\\"\\\"See the base class for the documentation.\\\"\\\"\\\"\\n return self.obtained_expectancy > SUCCESS_LIMIT\\n \\n\\n \\n \\nclass ArtistEvalPair(BaseArtistEvaluationPair):\\n \\\"\\\"\\\"An artist - user pair for validation purposes\\n \\n All test pairs, no matter if contained in training data or not. \\n \\\"\\\"\\\" \\n\\n @classmethod\\n def select_random(cls, i=0):\\n \\\"\\\"\\\"UNUSED\\\"\\\"\\\"\\n\\n scrobble_count = Scrobble.objects.all().count()\\n \\n # get a queryset containing every n-th scrobble\\n id_list = range(1 + i, scrobble_count, CROSS_VALIDATION_COUNT)\\n qs_validation_scrobbles = Scrobble.objects.filter(id__in=id_list)\\n \\n # remove the scrobbles and add them to the class\\n for scrobble in qs_validation_scrobbles.iterator():\\n \\n # create and save the validation pair\\n val_pair = cls(\\n subj=scrobble.user,\\n obj=scrobble.track.artist,\\n expected_expectancy=EXPECTED_EXPECTANCY_LISTENED) \\n val_pair.save()\\n \\n # remove the scrobble\\n scrobble.delete()\\n\\nclass NovelArtistEvalPair(BaseArtistEvaluationPair):\\n \\\"\\\"\\\"Only test pairs that aren't included in the test set\\\"\\\"\\\"\\n class Meta:\\n unique_together = ('subj', 'obj')\\n\"\n },\n {\n \"alpha_fraction\": 0.6770642399787903,\n \"alphanum_fraction\": 0.6880733966827393,\n \"avg_line_length\": 24.34883689880371,\n \"blob_id\": \"133fb2b44e39a49aa2bcbc39b4e0b2716922b53c\",\n \"content_id\": \"c38b85c4a195f444066545a6bcbd5e58d58cf32c\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 1090,\n \"license_type\": \"no_license\",\n \"max_line_length\": 231,\n \"num_lines\": 43,\n \"path\": \"/code/mahout/mahoutrec/unresystrecommend.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/bin/sh\\n\\n# params:\\n# lastfm\\n# flixster\\n# travel\\n\\n\\nLASTFM=false\\nFLIXSTER=false\\nTRAVEL=false\\n\\nfor param in $*;\\ndo \\n case $param in\\n 'lastfm')\\n LASTFM=true\\n ;;\\n 'flixster')\\n FLIXSTER=true\\n ;;\\n 'travel')\\n TRAVEL=true\\n ;; \\n esac\\ndone\\n\\nif [ $LASTFM = true ]\\nthen\\n mvn -e exec:java -Dexec.mainClass=\\\"com.unresyst.UnresystRecommend\\\" -Dexec.args=\\\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_train.csv 10 /home/pcv/diplomka2/svn/trunk/code/adapter/csv/lastfm_recommendations.csv\\\" \\nfi\\n\\nif [ $FLIXSTER = true ]\\nthen\\n mvn -e exec:java -Dexec.mainClass=\\\"com.unresyst.UnresystRecommend\\\" -Dexec.args=\\\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_train.csv 10 /home/pcv/diplomka2/svn/trunk/code/adapter/csv/flixster_recommendations.csv\\\" \\nfi\\n\\nif [ $TRAVEL = true ]\\nthen\\n mvn -e exec:java -Dexec.mainClass=\\\"com.unresyst.UnresystRecommend\\\" -Dexec.args=\\\"/home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_train.csv 10 /home/pcv/diplomka2/svn/trunk/code/adapter/csv/travel_recommendations.csv\\\" \\nfi\\n\\necho \\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.7937915921211243,\n \"alphanum_fraction\": 0.7937915921211243,\n \"avg_line_length\": 55.375,\n \"blob_id\": \"09a23a1f8ea432afe93770da81d70f433cb4da6d\",\n \"content_id\": \"627982f133521056ba61ba62704fe4769be69279\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 451,\n \"license_type\": \"no_license\",\n \"max_line_length\": 110,\n \"num_lines\": 8,\n \"path\": \"/UnresystCD/code/adapter/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The home of the adapter django project. \\n\\nContained applications:\\n - unresyst: the universal recommender application containing the implementation of the universal recommedner\\n - lastfm: an application containing data models and a Unresyst recommender for the last.fm dataset.\\n - flixster: application for the flixster dataset, the same structure as lastfm\\n - travel: application for the travel agency dataset, the same structure as lastfm\\n\\\"\\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.6514772176742554,\n \"alphanum_fraction\": 0.6554832458496094,\n \"avg_line_length\": 31.639345169067383,\n \"blob_id\": \"5cff1464dd01874dd53e68affc31a072f820b58a\",\n \"content_id\": \"fb84b5471f3093d38a41c109c7e10a8e55cf2638\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1997,\n \"license_type\": \"no_license\",\n \"max_line_length\": 93,\n \"num_lines\": 61,\n \"path\": \"/UnresystCD/code/adapter/unresyst/models/algorithm.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Models which instances are created by the algorithm package.\\n\\nThe preference between the subject and object in the means of \\nthe predicted relationship.\\n\\\"\\\"\\\"\\n\\nfrom django.db import models\\n\\nfrom base import BaseRelationshipInstance\\n\\nclass RelationshipPredictionInstance(BaseRelationshipInstance):\\n \\\"\\\"\\\"The preference between the subject and object in the means of \\n the predicted relationship.\\n \\\"\\\"\\\"\\n \\n expectancy = models.FloatField()\\n \\\"\\\"\\\"The probability of the predicted relationship appearance between \\n the subject and the object. A number from [0, 1]. \\n \\\"\\\"\\\"\\n \\n recommender = models.ForeignKey('unresyst.Recommender')\\n \\\"\\\"\\\"The recommender it belongs to\\\"\\\"\\\"\\n \\n is_uncertain = models.BooleanField(default=False)\\n \\\"\\\"\\\"Is the prediction uncertain? (We don't know anything about the pair)\\\"\\\"\\\"\\n \\n is_trivial = models.BooleanField(default=False)\\n \\\"\\\"\\\"Is the prediction trivial? (Already contained in the train data)\\\"\\\"\\\"\\n \\n class Meta:\\n app_label = 'unresyst' \\n unique_together = ('subject_object1', 'subject_object2', 'recommender')\\n\\n def __unicode__(self):\\n return \\\"(%s, %s), %f\\\" % (self.subject_object1, self.subject_object2, self.expectancy)\\n\\n \\nclass ExternalPrediction(models.Model):\\n \\\"\\\"\\\"A prediction taken from an external recommender.\\n \\\"\\\"\\\" \\n \\n subj_id = models.IntegerField()\\n \\\"\\\"\\\"The id of the subject\\\"\\\"\\\"\\n \\n obj_id = models.IntegerField()\\n \\\"\\\"\\\"The if of the object\\\"\\\"\\\"\\n \\n expectancy = models.FloatField()\\n \\\"\\\"\\\"The probability of the predicted relationship appearance between \\n the subject and the object. A number from [0, 1]. \\n \\\"\\\"\\\" \\n \\n recommender = models.ForeignKey('unresyst.Recommender')\\n \\\"\\\"\\\"The recommender it belongs to\\\"\\\"\\\" \\n \\n class Meta:\\n app_label = 'unresyst' \\n unique_together = ('subj_id', 'obj_id', 'recommender')\\n\\n def __unicode__(self):\\n return \\\"(%s, %s), %f\\\" % (self.subj_id, self.obj_id, self.expectancy) \\n\"\n },\n {\n \"alpha_fraction\": 0.5508274435997009,\n \"alphanum_fraction\": 0.5508274435997009,\n \"avg_line_length\": 22.27777862548828,\n \"blob_id\": \"b3e488fef73f2bf74b2bd4f721d639a4d867b199\",\n \"content_id\": \"4160764ea3b97a155d2f0c99fd77eab10980d7ce\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 423,\n \"license_type\": \"no_license\",\n \"max_line_length\": 68,\n \"num_lines\": 18,\n \"path\": \"/UnresystCD/code/adapter/demo/urls.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Urls for the demo application\\\"\\\"\\\"\\n\\nfrom django.conf.urls.defaults import *\\n\\nurlpatterns = patterns('demo.views',\\n \\n # shoe recommendations\\n url(regex=r'^user/(?P[\\\\w-]+)/shoe-recommendations/$',\\n view='view_recommendations', \\n name=\\\"recommendations\\\"\\n ),\\n\\n # home page\\n url(regex=r'^$',\\n view='view_home_page',\\n name='home_page'\\n ) \\n) \\n\"\n },\n {\n \"alpha_fraction\": 0.5290427803993225,\n \"alphanum_fraction\": 0.5329925417900085,\n \"avg_line_length\": 31.786258697509766,\n \"blob_id\": \"ad1cde0ed7345882ec924ec7c3e2862b9cfa53d7\",\n \"content_id\": \"7b0edd5c2b209e4c3de06222c5a0623a05db94da\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 4304,\n \"license_type\": \"no_license\",\n \"max_line_length\": 142,\n \"num_lines\": 131,\n \"path\": \"/UnresystCD/code/adapter/flixster/recommender.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The flixster recommender\\\"\\\"\\\"\\n\\nfrom django.db.models import Avg\\n\\nfrom unresyst import *\\nfrom models import *\\nfrom constants import *\\n\\ndef _rated_positively_generator():\\n \\\"\\\"\\\"The generator to the predicted relationship\\\"\\\"\\\"\\n for r in Rating.objects.filter(rating__gt=str(MIN_POSITIVE_RATING)):\\n yield (r.user, r.movie)\\n\\nclass MovieRecommender(Recommender):\\n \\\"\\\"\\\"The flixster movie recommender\\\"\\\"\\\"\\n\\n name = \\\"Flixster Movie Recommender\\\"\\n \\\"\\\"\\\"The name\\\"\\\"\\\" \\n \\n subjects = User.objects\\n \\\"\\\"\\\"The objects to who the recommender will recommend.\\\"\\\"\\\"\\n \\n objects = Movie.objects\\n \\\"\\\"\\\"The objects that will be recommended.\\\"\\\"\\\" \\n\\n predicted_relationship = PredictedRelationship( \\n name=\\\"User has rated the movie positively.\\\",\\n condition=None, \\n description=\\\"\\\"\\\"User %(subject)s has rated the %(object)s positively.\\\"\\\"\\\",\\n generator=_rated_positively_generator,\\n )\\n \\\"\\\"\\\"The relationship that will be predicted\\\"\\\"\\\"\\n \\n rules = (\\n # explicit rating\\n ExplicitSubjectObjectRule(\\n name=\\\"Movie rating.\\\",\\n \\n description=\\\"User %(subject)s has rated %(object)s.\\\",\\n \\n # all pairs user, rated movie\\n generator=lambda: [(r.user, r.movie) for r in Rating.objects.all()],\\n \\n # the number of stars divided by five\\n expectancy=lambda s, o:float(Rating.objects.get(user=s, movie=o).rating) / MAX_STARS,\\n ),\\n )\\n \\\"\\\"\\\"The rules\\\"\\\"\\\"\\n\\n relationships = (\\n # users in friendship\\n SubjectSimilarityRelationship(\\n name=\\\"Users are friends.\\\",\\n \\n generator=lambda: [(f.friend1, f.friend2) for f in Friend.objects.all()], \\n \\n is_positive=True, \\n \\n weight=0.1, \\n \\n description=\\\"Users %(subject1)s and %(subject2)s are friends.\\\",\\n ),\\n )\\n \\\"\\\"\\\"The relationships\\\"\\\"\\\"\\n \\n biases = (\\n \\n # people giving high ratings\\n SubjectBias(\\n name=\\\"Users giving high ratings.\\\",\\n \\n description=\\\"User %(subject)s gives high ratings.\\\",\\n \\n weight=1.0, \\n \\n is_positive=True,\\n \\n generator=lambda: User.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__gt=str(MIN_HIGH_RATING)), \\n \\n confidence=lambda user: user.rating_set.aggregate(Avg('rating'))['rating__avg'] - MIN_HIGH_RATING\\n ),\\n \\n # highly rated movies\\n ObjectBias(\\n name=\\\"High-rated movies.\\\",\\n \\n description=\\\"Movie %(object)s is high-rated\\\",\\n \\n weight=1.0,\\n \\n is_positive=True,\\n \\n generator=lambda: Movie.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__gt=str(MIN_HIGH_RATING)),\\n \\n confidence=lambda movie: movie.rating_set.aggregate(Avg('rating'))['rating__avg'] - MIN_HIGH_RATING\\n ),\\n \\n # people giving low ratings\\n SubjectBias(\\n name=\\\"Users giving low ratings.\\\",\\n \\n description=\\\"User %(subject)s gives low ratings.\\\",\\n \\n weight=0.25, \\n \\n is_positive=False,\\n \\n generator=lambda: User.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__lt=str(MAX_LOW_RATING)), \\n \\n confidence=lambda user: MAX_LOW_RATING - user.rating_set.aggregate(Avg('rating'))['rating__avg'] \\n ),\\n \\n # low-rated movies\\n ObjectBias(\\n name=\\\"Low-rated movies.\\\",\\n \\n description=\\\"Movie %(object)s is low-rated\\\",\\n \\n weight=0.25,\\n \\n is_positive=False,\\n \\n generator=lambda: Movie.objects.annotate(avg_rating=Avg('rating__rating')).filter(avg_rating__lt=str(MAX_LOW_RATING)),\\n \\n confidence=lambda movie: MAX_LOW_RATING - movie.rating_set.aggregate(Avg('rating'))['rating__avg']\\n ), \\n \\n \\n )\\n\\nMovieRecommender.explicit_rating_rule = MovieRecommender.rules[0] \\n \\n\"\n },\n {\n \"alpha_fraction\": 0.6794871687889099,\n \"alphanum_fraction\": 0.692307710647583,\n \"avg_line_length\": 18.5,\n \"blob_id\": \"e1ecc50b56bc9bcd857699933ad1343cd1eb6e3f\",\n \"content_id\": \"2c326e581ce012116baf0dbe73f6af9e0588ee6b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 78,\n \"license_type\": \"no_license\",\n \"max_line_length\": 39,\n \"num_lines\": 4,\n \"path\": \"/code/adapter/load.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/bin/sh\\n\\n# load the database from the given file\\nmysql -u root adapter < $1\\n\"\n },\n {\n \"alpha_fraction\": 0.5593509078025818,\n \"alphanum_fraction\": 0.5637439489364624,\n \"avg_line_length\": 39.85714340209961,\n \"blob_id\": \"de67b101372b2d6ca574db8bb84fed49f465a847\",\n \"content_id\": \"fddd4bd233c7e3d4c7308f87998218274cd809ca\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 11154,\n \"license_type\": \"no_license\",\n \"max_line_length\": 96,\n \"num_lines\": 273,\n \"path\": \"/code/adapter/unresyst/compilator/get_first_compilator.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The basic compilator used in Unresyst\\\"\\\"\\\"\\n\\nfrom base import BaseCompilator\\nfrom unresyst.constants import *\\nfrom unresyst.models.aggregator import AggregatedRelationshipInstance\\nfrom unresyst.models.abstractor import RelationshipInstance\\nfrom unresyst.models.algorithm import RelationshipPredictionInstance\\n\\nclass GetFirstCompilator(BaseCompilator):\\n \\\"\\\"\\\"Compilator using the first relationship it finds to create a prediction\\\"\\\"\\\"\\n \\n def __init__(\\n self, \\n depth=DEFAULT_COMPILATOR_DEPTH, \\n breadth=DEFAULT_COMPILATOR_BREADTH, \\n pair_depth=DEFAULT_COMPILATOR_PAIR_DEPTH):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\" \\n \\n super(GetFirstCompilator, self).__init__(\\n combinator=None, \\n depth=depth, \\n breadth=breadth, \\n pair_depth=pair_depth) \\n\\n\\n def compile_all(self, recommender_model):\\n \\\"\\\"\\\"Compile preferences, known relationships + similarities.\\n \\\"\\\"\\\"\\n self.compile_aggregates(recommender_model)\\n\\n print \\\" Compiling similar objects.\\\"\\n \\n # if subjects == objects\\n if recommender_model.are_subjects_objects:\\n\\n # take similar on both sides\\n self._compile_similar_subjectobjects(recommender_model)\\n\\n else:\\n \\n # take similar to the ones we already have (content-based recommender)\\n self._compile_similar_objects(recommender_model)\\n print \\\" Done. Compiling similar subjects.\\\"\\n\\n # take liked objects of similar users (almost collaborative filtering)\\n self._compile_similar_subjects(recommender_model) \\n \\n\\n #TODO pryc, asi nebude potreba\\n def compile_aggregates(self, recommender_model):\\n \\\"\\\"\\\"Create predictions from aggregates\\\"\\\"\\\" \\n \\n # filter only S-O or SO-SO aggregates\\n #\\n rel_type = RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT \\\\\\n if recommender_model.are_subjects_objects else \\\\\\n RELATIONSHIP_TYPE_SUBJECT_OBJECT \\n \\n qs_aggr = AggregatedRelationshipInstance.objects.filter(\\n recommender=recommender_model,\\n relationship_type=rel_type)\\n\\n # go through the aggregates, create predictions and save them\\n for aggr in qs_aggr.iterator():\\n\\n # order the arguments as they should be \\n so1, so2 = self._order_in_pair(aggr.subject_object1, aggr.subject_object2)\\n\\n prediction = RelationshipPredictionInstance(\\n subject_object1=so1,\\n subject_object2=so2,\\n description=aggr.description,\\n expectancy=aggr.expectancy,\\n recommender=recommender_model)\\n \\n prediction.save() \\n \\n print \\\" %d aggregated predictions created\\\" % qs_aggr.count()\\n \\n \\n def _compile_similar_objects(self, recommender_model):\\n \\\"\\\"\\\"Create predictions by adding objects similar to ones the objects\\n in predicted_relationship - that's a content-based recommender\\n \\\"\\\"\\\" \\n \\n self._compile_similar_entities(\\n recommender_model=recommender_model,\\n start_entity_type=ENTITY_TYPE_SUBJECT)\\n \\n def _compile_similar_subjects(self, recommender_model):\\n \\\"\\\"\\\"Create predictions by adding objects that similar subjects liked\\n - that's collaborative filtering\\n \\\"\\\"\\\" \\n \\n self._compile_similar_entities(\\n recommender_model=recommender_model,\\n start_entity_type=ENTITY_TYPE_OBJECT)\\n\\n\\n def _compile_similar_subjectobjects(self, recommender_model):\\n \\\"\\\"\\\"Create predictions based on similarity for recommenders where \\n subjects==objects\\n \\\"\\\"\\\"\\n \\n # firstly the normal direction\\n self._compile_similar_entities(\\n recommender_model=recommender_model,\\n start_entity_type=ENTITY_TYPE_SUBJECTOBJECT,\\n reverse=False)\\n \\n # secondly the opposite\\n self._compile_similar_entities(\\n recommender_model=recommender_model,\\n start_entity_type=ENTITY_TYPE_SUBJECTOBJECT,\\n reverse=True)\\n\\n def _order_in_pair(self, arg1, arg2):\\n \\\"\\\"\\\"Swap the arguments in the rule/relationships so that the first\\n has a lower id than the second (for subjectobjects), or the subject\\n is the first (for others)\\n \\\"\\\"\\\"\\n # for subjectobject return ordered by pk\\n if arg1.entity_type == ENTITY_TYPE_SUBJECTOBJECT:\\n if arg2.pk < arg1.pk:\\n return (arg2, arg1)\\n return (arg1, arg2)\\n # for others return subject first \\n if arg2.entity_type == ENTITY_TYPE_SUBJECT:\\n return (arg2, arg1)\\n return (arg1, arg2)\\n\\n SIMILARITY_RELATIONSHIP_TYPES = {\\n ENTITY_TYPE_SUBJECT: RELATIONSHIP_TYPE_OBJECT_OBJECT,\\n ENTITY_TYPE_OBJECT: RELATIONSHIP_TYPE_SUBJECT_SUBJECT,\\n ENTITY_TYPE_SUBJECTOBJECT: RELATIONSHIP_TYPE_SUBJECTOBJECT_SUBJECTOBJECT\\n } \\n \\\"\\\"\\\"Dictionary entity type: relationship entity type.\\n Key is the entity type where the traversing for similarity starts, value\\n is the relationship type that should be traversed for similarity.\\n \\\"\\\"\\\"\\n \\n ORDERING = {\\n ENTITY_TYPE_OBJECT: \\\"subject_object1__pk\\\",\\n ENTITY_TYPE_SUBJECT: \\\"subject_object2__pk\\\",\\n ENTITY_TYPE_SUBJECTOBJECT: \\\"subject_object1__pk\\\",\\n }\\n \\\"\\\"\\\"Dictinary entity type (starting) - name of the attribute in PredictedRelationship, \\n that should be used for sorting\\\"\\\"\\\"\\n \\n \\n def _compile_similar_entities(self, recommender_model, start_entity_type, reverse=False):\\n \\\"\\\"\\\"Create predictions from start_entity_type objects, looking for \\n similar entities in end_entity_type\\n \\n @type recommender_model: models.common.Recommender\\n @param recommender_model: the model the for which the predictions\\n should be built\\n\\n @type start_entity_type: str\\n @param start_entity_type: the entity type, where to start searching for\\n similar entities. E.g. if start_entity_type is 'S', similar 'O' \\n will be added to predictions to each 'S'.\\n\\n @type reverse: bool\\n @param reverse: relevant only if start_entity_type=='SO', indicates \\n whether the relationships will be traversed in the reverse order. \\n \\\"\\\"\\\" \\n \\n # get the predictions for this recommender\\n qs_predictions = RelationshipPredictionInstance.objects.filter(\\n recommender=recommender_model)\\n\\n ordering = self.ORDERING[start_entity_type]\\n \\n # go through the predicted relationship instances \\n # (objects that subjects liked)\\n # ordered as needed\\n qs_pred_rel_instances = RelationshipInstance\\\\\\n .filter_predicted(recommender_model)\\\\\\n .order_by(ordering)\\\\\\n .select_related(depth=1)\\n \\n # the caching variable \\n last_fin = None \\n last_qs = None \\n \\n print \\\"Predicted relationship count: %d\\\" % qs_pred_rel_instances.count()\\n \\n i = 0\\n \\n # count neighbourhood\\n count_n = 0\\n \\n # count all \\n count_all = 0\\n \\n for pred_inst in qs_pred_rel_instances.iterator():\\n \\n i += 1\\n # get the subject and object from the relationship instance\\n start, fin = (pred_inst.subject_object1, pred_inst.subject_object2) \\\\\\n if pred_inst.subject_object1.entity_type == start_entity_type \\\\\\n else (pred_inst.subject_object2, pred_inst.subject_object1)\\n \\n # if they are subjectobjects and reversed swap them\\n if start_entity_type == ENTITY_TYPE_SUBJECTOBJECT and reverse:\\n start, fin = fin, start\\n \\n similarity_relationship_type = self.SIMILARITY_RELATIONSHIP_TYPES[start_entity_type]\\n \\n # if the ending entity is the same as the last time, use the cached qs\\n if last_fin == fin:\\n qs_similar_rels = last_qs\\n else: \\n # get objects similar to obj (whole relationships) - only breadth highest\\n qs_similar_rels = AggregatedRelationshipInstance\\\\\\n .get_relationships(fin)\\\\\\n .filter(relationship_type=similarity_relationship_type)\\\\\\n .order_by('-expectancy')\\n \\n # current count of theoretically available predictions\\n cur_count_all = qs_similar_rels.count()\\n \\n qs_similar_rels = qs_similar_rels[:self.breadth]\\\\\\n .select_related(depth=1)\\n\\n # keep it for the next time\\n last_fin = fin\\n last_qs = qs_similar_rels\\n \\n\\n count = qs_similar_rels.count()\\n\\n count_all += cur_count_all\\n count_n += count\\n \\n if count and i % 1000 == 0:\\n print \\\"similar count: %d; relationships processed: %d\\\" % (count, i)\\n import gc; gc.collect()\\n \\n # go through them \\n for similar_rel in qs_similar_rels.iterator():\\n \\n # get the object similar to obj from the relationship\\n similar_fin = similar_rel.get_related(fin)\\n\\n # find out whether there exists a prediction for the pair\\n qs_pair_predictions = RelationshipPredictionInstance\\\\\\n .filter_relationships(\\n object1=start,\\n object2=similar_fin,\\n queryset=qs_predictions)\\n\\n # if exists, keep it there, ignore\\n if qs_pair_predictions.exists():\\n continue\\n \\n # order the arguments as they should be \\n so1, so2 = self._order_in_pair(start, similar_fin) \\n\\n # if not, create it with the attributes of the similarity \\n # relationship instance\\n prediction = RelationshipPredictionInstance(\\n subject_object1=so1,\\n subject_object2=so2,\\n description=similar_rel.description,\\n expectancy=similar_rel.expectancy,\\n recommender=recommender_model)\\n \\n prediction.save() \\n \\n print \\\"For starting entity type %s, %d out of %d possible relationships created\\\" \\\\\\n % (start_entity_type, count_n, count_all)\\n\"\n },\n {\n \"alpha_fraction\": 0.5061048865318298,\n \"alphanum_fraction\": 0.5157351493835449,\n \"avg_line_length\": 34.650306701660156,\n \"blob_id\": \"b7c406a6e4ef6fc0fe902a5c1c7208eac1bf1810\",\n \"content_id\": \"d2021e3563ad203310c773135625624c376a9466\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5815,\n \"license_type\": \"no_license\",\n \"max_line_length\": 146,\n \"num_lines\": 163,\n \"path\": \"/UnresystCD/code/adapter/lastfm/recbckp.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \" rules = ( \\n \\n # don't recommend artists with male-specific tags to females\\n SubjectObjectRule(\\n name=\\\"Don't recommend male music to female users.\\\",\\n\\n # the user is a female and the artist was tagged by\\n # a male-specific tag\\n condition=lambda user, artist: user.gender == 'f' and \\\\\\n artist.artisttag_set.filter(tag__gender_specific='m').exists()\\n \\n # it's a negative rule\\n is_positive=False,\\n \\n weight=0.5,\\n \\n # the more male-specific tags the artist has, the higher is \\n # the rule confidence. Normalized by the artist tag count\\n confidence=lambda user, artist: float(\\n artist.artisttag_set.filter(tag__gender_specific='m').count())/ \\\\\\n artist.artisttag_set.count(),\\n \\n description=\\\"Artist %(object)s isn't recommended to %(subject)s, \\\" +\\n \\\"because the artist is considered male-specific.\\\"\\n )\\n \\n \\n # users of similar age are similar\\n SubjectSimilarityRule(\\n name=\\\"Users with similar age.\\\",\\n \\n # both users have given their age and the difference \\n # is lower than five\\n condition=lambda user1, user2: \\n user1.age and user2.age and abs(user1.age - user2.age) <= 5,\\n \\n is_positive=True, \\n \\n weight=0.5,\\n \\n # a magic linear confidence function\\n confidence=lambda user1, user2: \\n 1 - float(abs(user1.age - user2.age))/AGE_DIFFERENCE,\\n \\n description=\\\"Users %(subject1)s and %(subject2)s are about \\\" + \\n \\\"the same age.\\\"\\n ), \\n \\n # artists sharing some tags are similar\\n ObjectSimilarityRule(\\n name=\\\"Artists sharing some tags.\\\",\\n\\n # both artists have some tags and they share at least one tag\\n # generator - take artists having some tags, compare them one to one\\n generator=_tag_similarity_generator,\\n \\n \\n # it's a positive rule\\n is_positive=True,\\n \\n weight=0.5,\\n \\n # The more tags the artists have in common, the higher is \\n # the similarity confidence\\n confidence=lambda artist1, artist2: \\\\\\n float(artist1.artisttag_set.filter(\\n tag__id__in=artist2.artisttag_set.values_list('tag__id')\\n ).count()) / \\\\\\n min(artist1.artisttag_set.count(), artist2.artisttag_set.count()),\\n \\n description=\\\"Artists %(object1)s and %(object2)s are similar \\\" + \\\\\\n \\\"because they share some tags.\\\"\\n ), \\n \\n \\n # if the users were registered in a similar period, the're similar\\n SubjectSimilarityRule(\\n name='Users registered in similar time.',\\n \\n condition=lambda s1, s2:\\n s1.registered and s2.registered and \\\\\\n abs(s1.registered.toordinal() - s2.registered.toordinal()) < REGISTERED_DIFFERENCE / 5,\\n \\n is_positive=True,\\n weight=0.5,\\n \\n confidence=lambda s1, s2:\\n 1 - float(abs(s1.registered.toordinal() - s2.registered.toordinal()))/REGISTERED_DIFFERENCE,\\n \\n description=\\\"Users %(subject1)s and %(subject2)s were registered in similar times\\\",\\n ),\\n )\\n \\n cluster_sets = (\\n\\n \\n # user - gender\\n SubjectClusterSet(\\n \\n name='User gender.',\\n \\n weight=0.5,\\n \\n # users that have a gender (filled)\\n filter_entities=User.objects.exclude(gender=''),\\n \\n get_cluster_confidence_pairs=lambda user: ((user.gender, 1),),\\n \\n description=\\\"%(subject)s's gender is %(cluster)s.\\\"\\n \\n ),\\n \\n # user - country\\n SubjectClusterSet(\\n \\n name='User country.',\\n \\n weight=0.5,\\n \\n # users that have country filled\\n filter_entities=User.objects.filter(country__isnull=False),\\n \\n get_cluster_confidence_pairs=lambda user: ((user.country.name, 1),),\\n \\n description=\\\"%(subject)s is from %(cluster)s.\\\"\\n ),\\n )\\n \\n biases = (\\n ObjectBias(\\n name=\\\"Artists whose tracks were listened the most.\\\",\\n description=\\\"%(object)s is much listened.\\\",\\n weight=0.5,\\n is_positive=True,\\n # users whose tracks were listened more than the half of the most listened\\n generator=lambda: Artist.objects.annotate(listen_count=Count('track__scrobble')).filter(listen_count__gt=MAX_SCROBBLE_COUNT/7),\\n \\n # the number of scrobbles for the artist divided by the max.\\n confidence=lambda a: float(a.track_set.annotate(scrobble_count=Count('scrobble')).aggregate(Sum('scrobble_count')))/MAX_SCROBBLE_COUNT\\n ),\\n \\n ) \\n \\n \\n \\n \\n \\n \\n \\n # tag clusters\\n ObjectClusterSet(\\n\\n name=\\\"Artist tags.\\\",\\n\\n weight=0.5,\\n \\n # artists that are tagged by a tag that another artist also has\\n filter_entities=Artist.objects.annotate(shared_count=Count('artisttag__tag__artisttag')).filter(shared_count__gt=1).distinct(),\\n \\n get_cluster_confidence_pairs=_get_artist_tag_pairs,\\n \\n description=\\\"%(object)s was tagged as %(cluster)s.\\\",\\n ),\\n\"\n },\n {\n \"alpha_fraction\": 0.5247273445129395,\n \"alphanum_fraction\": 0.5608845353126526,\n \"avg_line_length\": 26.657024383544922,\n \"blob_id\": \"94c6ae64ba47d9b97d88b8812b92c95e32dfa0b7\",\n \"content_id\": \"a1c2c5fae057744714787879bd4a97cd774cc388\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 6693,\n \"license_type\": \"no_license\",\n \"max_line_length\": 159,\n \"num_lines\": 242,\n \"path\": \"/UnresystCD/code/adapter/lastfm/save_data.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Saving data from the last.fm datasets\\\"\\\"\\\"\\n\\nimport csv\\nimport os\\nfrom datetime import datetime\\nfrom dateutil.parser import parse\\nfrom dateutil.tz import tzutc\\n\\nfrom models import *\\n\\nTZ = tzutc()\\n\\\"\\\"\\\"The used timezone\\\"\\\"\\\"\\n\\nGENDER_SPECIFIC_TAGS = {\\n 'punk': 'm',\\n 'metal': 'm',\\n 'punk rock': 'm',\\n 'hardcore': 'm',\\n 'hard rock': 'm',\\n 'industrial': 'm',\\n 'post-punk': 'm',\\n 'post-rock': 'm',\\n 'heavy metal': 'm',\\n 'metalcore': 'm',\\n 'death metal': 'm',\\n 'noise': 'm',\\n 'hardcore punk': 'm',\\n\\n 'female': 'f',\\n 'beautiful': 'f',\\n 'Love': 'f',\\n 'lovely': 'f',\\n 'sweet': 'f',\\n 'singer-songwriter': 'f',\\n 'singer songwriter': 'f',\\n 'woman': 'f',\\n 'mellow': 'f',\\n 'dream pop': 'f',\\n 'soft': 'f',\\n 'emotional': 'f',\\n 'cut': 'f',\\n}\\n\\\"\\\"\\\"Tags that are specific for the given gender.\\n\\\"\\\"\\\"\\n\\n\\\"\\\"\\\"\\n7141 artists\\n992 users\\n19903 tracks\\n22666 scrobbles\\n\\n\\n\\n# relative paths to the dataset files\\nfilename_scrobbles = '../../datasets/lastfm/1K/tracks30000.tsv'\\nfilename_users = '../../datasets/lastfm/1K/userid-profile.tsv'\\nfilename_tags = '../../datasets/lastfm/Lastfm-ArtistTags2007/artist_tags200K.tsv'\\n\\\"\\\"\\\"\\n\\n\\\"\\\"\\\"\\n2348 artists\\n100 users\\n5057 tracks\\n5532 scrobbles\\n9120 tags\\n28588 tagartists\\n322 tagged artists\\n\\nmax age 38\\nmin age 17\\n\\\"\\\"\\\"\\n# relative paths to the dataset files\\nfilename_scrobbles = '../../datasets/lastfm/mini/scrobbles.tsv'\\nfilename_users = '../../datasets/lastfm/mini/users.tsv'\\nfilename_tags = '../../datasets/lastfm/mini/tags.tsv'\\n\\nseparator = '\\\\t'\\n\\ndef save_data():\\n \\\"\\\"\\\"Save the data\\\"\\\"\\\"\\n _parse_users(filename_users)\\n print \\\"Users saved.\\\"\\n \\n _parse_scrobbles(filename_scrobbles)\\n print \\\"Scrobbles saved.\\\"\\n \\n _parse_tags(filename_tags)\\n print \\\"Tags saved.\\\"\\n\\n\\\"\\\"\\\"\\nuser_000001\\tm\\t\\tJapan\\tAug 13, 2006\\nuser_000002\\tf\\t\\tPeru\\tFeb 24, 2006\\nuser_000003\\tm\\t22\\tUnited States\\tOct 30, 2005\\n\\\"\\\"\\\"\\n\\ndef _parse_users(filename):\\n \\\"\\\"\\\"Parse the user csv file\\\"\\\"\\\"\\n \\n filename = _get_abs_path(filename)\\n\\n # open the csv reader\\n reader = csv.reader(open(filename, \\\"rb\\\"), delimiter=separator, quoting=csv.QUOTE_NONE)\\n \\n for user_id, gender, age, country, reg_date in reader:\\n \\n # parse the user id\\n user_id = _parse_user_id(user_id)\\n \\n # replace the empty string by None\\n age = None if not age else age\\n \\n # find or create the country\\n if country:\\n country_model, created = Country.objects.get_or_create(name=country)\\n else:\\n country_model = None\\n \\n # parse the date\\n reg_date = None if not reg_date else datetime.strptime(reg_date, '%b %d, %Y')\\n \\n # create and save the user\\n user = User(\\n id=user_id,\\n gender=gender,\\n age=age,\\n country=country_model,\\n registered=reg_date)\\n \\n user.save()\\n\\n\\\"\\\"\\\"\\n user_000639 2009-04-08T01:57:47Z 15676fc4-ba0b-4871-ac8d-ef058895b075 The Dogs D'Amour 6cc252d0-3f42-4fd3-a70f-c8ff8b693aa4 How Do You Fall in Love Again\\n\\\"\\\"\\\"\\n\\ndef _parse_scrobbles(filename):\\n \\\"\\\"\\\"Parse the scrobbles by the users\\\"\\\"\\\"\\n\\n filename = _get_abs_path(filename)\\n \\n reader = csv.reader(open(filename, \\\"rb\\\"), delimiter=separator, quoting=csv.QUOTE_NONE)\\n for user_id, timestamp, artist_guid, artist_name, track_guid, track_name in reader:\\n\\n if reader.line_num % 1000 == 0:\\n print '%s lines processed' % reader.line_num \\n \\n # if the track isn't in the musicbrainz db, continue\\n if not track_guid or not artist_guid:\\n continue\\n \\n # parse the user id and find the user\\n user_id = _parse_user_id(user_id)\\n user = User.objects.get(id=user_id)\\n \\n # parse the date and put it to the UTC timezone\\n timestamp = parse(timestamp).astimezone(TZ).replace(tzinfo=None)\\n \\n # get or create the artist\\n artist, created = Artist.objects.get_or_create(\\n guid=artist_guid,\\n defaults={'name': artist_name})\\n \\n # get or crate the track\\n track, created = Track.objects.get_or_create(\\n guid=track_guid, \\n defaults= {'artist': artist, 'name': track_name})\\n \\n \\n # create and save the scrobble\\n scrobble = Scrobble(\\n user=user,\\n timestamp=timestamp,\\n track=track)\\n \\n scrobble.save() \\n \\n\\n\\\"\\\"\\\"\\n 11eabe0c-2638-4808-92f9-1dbd9c453429Deerhoofamerican14\\n 11eabe0c-2638-4808-92f9-1dbd9c453429Deerhoofanimals5\\n 11eabe0c-2638-4808-92f9-1dbd9c453429Deerhoofart punk2\\n\\\"\\\"\\\" \\n\\ndef _parse_tags(filename):\\n \\\"\\\"\\\"Parse the tags\\\"\\\"\\\"\\n\\n filename = _get_abs_path(filename)\\n \\n reader = csv.reader(open(filename, \\\"rb\\\"), delimiter=separator, quoting=csv.QUOTE_NONE)\\n \\n last_artist = Artist.objects.all()[0]\\n \\n for artist_guid, artist_name, tag_name, count in reader:\\n\\n if reader.line_num % 1000 == 0:\\n print '%s lines processed' % reader.line_num \\n \\n # some caching for more speed\\n if artist_guid == last_artist.guid:\\n artist = last_artist\\n else: \\n # try finding the artist\\n qs_artist = Artist.objects.filter(guid=artist_guid)\\n \\n # if not found go ahead\\n if not qs_artist:\\n continue\\n \\n artist = qs_artist[0]\\n \\n # if the name is strange, go ahead\\n if artist.name != artist_name:\\n continue\\n \\n # get the gender specificity\\n if GENDER_SPECIFIC_TAGS.has_key(tag_name):\\n gender_specific = GENDER_SPECIFIC_TAGS[tag_name]\\n else:\\n gender_specific = ''\\n \\n # get or create the tag\\n tag, created = Tag.objects.get_or_create(\\n name=tag_name, \\n defaults={'gender_specific': gender_specific,}) \\n \\n # create and save the tag for the artist \\n artist_tag = ArtistTag(\\n artist=artist,\\n tag=tag,\\n count=count)\\n \\n artist_tag.save() \\n \\n last_artist = artist \\n \\n \\ndef _get_abs_path(filename):\\n \\\"\\\"\\\"Get the absolute path from the relative\\\"\\\"\\\"\\n return os.path.join(os.path.dirname(__file__), filename)\\n \\ndef _parse_user_id(user_id):\\n \\\"\\\"\\\"Get the integer id from the id string\\\"\\\"\\\"\\n return int(user_id.split('_')[1])\\n\"\n },\n {\n \"alpha_fraction\": 0.7558139562606812,\n \"alphanum_fraction\": 0.7558139562606812,\n \"avg_line_length\": 27.66666603088379,\n \"blob_id\": \"0cba3557c9d5312194d5d95f8c3b15e006820481\",\n \"content_id\": \"7301d27729a9c8643e0f847085cd196f315d649d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 86,\n \"license_type\": \"no_license\",\n \"max_line_length\": 54,\n \"num_lines\": 3,\n \"path\": \"/UnresystCD/code/adapter/travel/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Application holding the data and a recommender for \\nthe travel agency dataset.\\n\\\"\\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.7663999795913696,\n \"alphanum_fraction\": 0.7663999795913696,\n \"avg_line_length\": 38.0625,\n \"blob_id\": \"2829d1381995c103e68d0afad6485f4b995717a9\",\n \"content_id\": \"7e1456486a0e2d20d616f2940869865ca593246f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 625,\n \"license_type\": \"no_license\",\n \"max_line_length\": 80,\n \"num_lines\": 16,\n \"path\": \"/code/adapter/lastfm/mahout_recommender.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The representation of the Mahout recommender in Unresyst.\\\"\\\"\\\"\\n\\nfrom unresyst.recommender.external_recommender import ExternalRecommender\\nfrom recommender import ArtistRecommender, NovelArtistRecommender\\n\\nclass MahoutArtistRecommender(ExternalRecommender, ArtistRecommender):\\n \\\"\\\"\\\"An external artist recommender - both novel and non-novel artists\\\"\\\"\\\"\\n \\n name = \\\"Mahout artist recommender\\\"\\n \\\"\\\"\\\"The name\\\"\\\"\\\"\\n \\nclass NovelMahoutArtistRecommender(ExternalRecommender, NovelArtistRecommender):\\n \\\"\\\"\\\"External recommender for novel artists only\\\"\\\"\\\"\\n \\n name = \\\"Novel Mahout artist recommender\\\"\\n \\\"\\\"\\\"The name\\\"\\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.7530864477157593,\n \"alphanum_fraction\": 0.7530864477157593,\n \"avg_line_length\": 26,\n \"blob_id\": \"2cc663c22636ab3d282b4a3c9e490836aa81eb0f\",\n \"content_id\": \"447bdfd1b9fd756e0f53ccfe0aae9a1f97c85ff8\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 81,\n \"license_type\": \"no_license\",\n \"max_line_length\": 54,\n \"num_lines\": 3,\n \"path\": \"/code/adapter/flixster/__init__.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Application holding the data and a recommender for \\nthe Flixster dataset.\\n\\\"\\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.5634920597076416,\n \"alphanum_fraction\": 0.5634920597076416,\n \"avg_line_length\": 30.25,\n \"blob_id\": \"9536a8e1ff67ca685db02dbfb9009d15abafec92\",\n \"content_id\": \"2cdd98fc94573b198df43eb36e8ea8ab59478acf\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 126,\n \"license_type\": \"no_license\",\n \"max_line_length\": 75,\n \"num_lines\": 4,\n \"path\": \"/code/adapter/diff.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/bin/bash\\n\\n# diffovat v meldu zmenene soubory.\\nsvn diff --diff-cmd meld `svn stat | grep \\\"M \\\" | sed -e 's/M //'`\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6452485918998718,\n \"alphanum_fraction\": 0.6459249258041382,\n \"avg_line_length\": 32.98850631713867,\n \"blob_id\": \"ab43235feb01e51bb9cf20b564b11fb6ae1c6eb9\",\n \"content_id\": \"9203604c83a4ceb6773366cba34dbea8b44ca395\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2957,\n \"license_type\": \"no_license\",\n \"max_line_length\": 100,\n \"num_lines\": 87,\n \"path\": \"/UnresystCD/code/adapter/unresyst/recommender/metrics.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The functions used as metrics for the evaluators\\\"\\\"\\\"\\nimport math\\n\\n@classmethod\\ndef rmse(cls):\\n \\\"\\\"\\\"Count and print the result of the RMSE on the evaluation pair model\\n \\n @type EvaluationPairModel: a model, BaseEvaluationPair subclass\\n @param EvaluationPairModel: a model to evaluate on \\n \\n @rtype: float\\n @return: the RMSE of the predictions\\n \\\"\\\"\\\"\\n \\n # get the expected, obtained expectancy pairs\\n exp_pairs = cls.EvaluationPairModel.objects.values_list(\\n 'expected_expectancy', 'obtained_expectancy') \\n \\n # count the sum of the squares of the differences\\n square_sum = sum([pow(exp_e - obt_e, 2) for exp_e, obt_e in exp_pairs])\\n \\n # get the rmse\\n ret_rmse = math.sqrt(square_sum / len(exp_pairs))\\n print \\\"RMSE: %f\\\" % ret_rmse\\n \\n return ret_rmse\\n\\n \\n@classmethod \\ndef precision_recall(cls, count):\\n \\\"\\\"\\\"Count and print the Precision/Recall measure. \\n \\n Both are counted as per-user average\\n \\n Hit: an object that is both in the recommendation and in the evaluation set.\\n \\n Precision: number of hits / recommendation list size (=count)\\n If a hit appears for an subject-object that is multiple time present in the\\n test set, it's counted only once.\\n \\n Recall: number of hits / number of objects for the subject in the evaluation test\\n If a hit appears for an subject-object that is multiple time present in the\\n test set, it's counted multiple times.\\n \\n @type count: int\\n @param count: the number of objects in recommendation lists\\n \\n @rtype: pair float float\\n @return: precision, recall \\n \\\"\\\"\\\"\\n \\n # hits are successful pairs\\n qs_hit_pairs = cls.EvaluationPairModel.objects.filter(is_successful=True)\\n \\n # take ids of subjects from the test pairs \\n subj_id_list = cls.EvaluationPairModel.objects.values_list('subj__pk', flat=True).distinct() \\n \\n subj_count = len(subj_id_list)\\n \\n unique_hit_pairs = qs_hit_pairs.values_list('subj__pk', 'obj__pk').distinct()\\n \\n # count the precision as explained in docstring\\n # as count is the same for all users this can be counted at once\\n precision = float(len(unique_hit_pairs)) / (subj_count * count)\\n print \\\"Precision: %f\\\" % precision\\n \\n recall_sum = 0\\n \\n # count the average recall / user\\n for subj_id in subj_id_list:\\n \\n # count the parameters per user \\n user_hit_count = qs_hit_pairs.filter(subj__pk=subj_id).count()\\n possible_hit_count = cls.EvaluationPairModel.objects.filter(subj__pk=subj_id).count()\\n \\n subj_recall = float(user_hit_count) / possible_hit_count\\n\\n #print \\\"%d precision: %f\\\" % (subj_id, float(user_hit_count)/count)\\n #print \\\"%d recall: %f\\\" % (subj_id, subj_recall)\\n \\n recall_sum += subj_recall\\n \\n recall = recall_sum / subj_count \\n\\n print \\\"Recall: %f\\\" % recall\\n \\n return (precision, recall)\\n\"\n },\n {\n \"alpha_fraction\": 0.5109958052635193,\n \"alphanum_fraction\": 0.5757350921630859,\n \"avg_line_length\": 21.685392379760742,\n \"blob_id\": \"65928876245a1c3845fe7a61dea8aea090510c12\",\n \"content_id\": \"1bf2649748b12465206a5401d6fa857b5b29019b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 4047,\n \"license_type\": \"no_license\",\n \"max_line_length\": 90,\n \"num_lines\": 178,\n \"path\": \"/code/adapter/flixster/save_data.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"Saving data from csv files\\\"\\\"\\\"\\nimport csv\\nimport os\\n\\nfrom django.db.models import Count\\n\\nfrom models import *\\n\\n# relative paths to the dataset files\\nfilename_links = '../../datasets/flixster/mini/links.tsv'\\nfilename_ratings = '../../datasets/flixster/mini/ratings.tsv'\\n\\nseparator = '\\\\t'\\n\\n\\\"\\\"\\\"\\nBefore the removal:\\n\\n>>> User.objects.count()\\n1269\\n>>> Friend.objects.count()\\n4439\\n>>> Movie.objects.count()\\n4237\\n>>> Rating.objects.count()\\n14195\\n\\nAfter the removal:\\n\\n>>> User.objects.count()\\n418\\n>>> Friend.objects.count()\\n1051\\n\\n>>> MovieRecommenderEvaluator.select_evaluation_pairs()\\n2839 test pairs selected from total 14195 pairs\\n\\\"\\\"\\\"\\n\\nmax_user_id = 1520\\n\\ndef _get_abs_path(filename):\\n \\\"\\\"\\\"Get the absolute path from the relative\\\"\\\"\\\"\\n return os.path.join(os.path.dirname(__file__), filename)\\n\\ndef save_data():\\n \\\"\\\"\\\"Save the data\\\"\\\"\\\"\\n _parse_links(filename_links)\\n print \\\"Users and links saved.\\\"\\n \\n _parse_ratings(filename_ratings)\\n print \\\"Ratings saved.\\\" \\n \\n # remove things we don't like\\n \\n # remove users without rating - they can't be tested anyhow \\n User.objects.annotate(rating_count=Count('rating')).filter(rating_count=0).delete()\\n print \\\"Disliked stuff deleted.\\\" \\n\\n\\n\\\"\\\"\\\"\\n6\\t1349\\n6\\t1370\\n6\\t1375\\n6\\t1761\\n6\\t1770\\n6\\t1775\\n6\\t1785\\n6\\t1797\\n6\\t1803\\n6\\t1810\\n6\\t1819\\n6\\t1825\\n6\\t1831\\n\\\"\\\"\\\" \\n \\ndef _parse_links(filename):\\n \\\"\\\"\\\"Parse the user csv file\\\"\\\"\\\"\\n \\n filename = _get_abs_path(filename)\\n \\n last_user = User(id=0) \\n\\n # open the csv reader\\n reader = csv.reader(open(filename, \\\"rb\\\"), delimiter=separator, quoting=csv.QUOTE_NONE)\\n \\n for user_id, friend_id in reader:\\n \\n if reader.line_num % 5000 == 0:\\n print '%s lines processed' % reader.line_num \\n \\n # parse the user ids\\n user_id = int(user_id)\\n friend_id = int(friend_id) \\n \\n # skip high friends\\n if friend_id > max_user_id:\\n continue\\n \\n # if it's about the last user, use him, otherwise create him\\n if user_id == last_user.id: \\n user = last_user\\n else:\\n user, created = User.objects.get_or_create(id=user_id)\\n \\n # get or create the friend\\n friend, created = User.objects.get_or_create(id=friend_id) \\n \\n f1, f2 = (user, friend) if user.id < friend.id else (friend, user)\\n \\n # create the friendship\\n Friend.objects.get_or_create(\\n friend1=f1,\\n friend2=f2)\\n \\n # save the user for the next time\\n last_user = user\\n\\\"\\\"\\\"\\n6\\t57699\\t3\\n7\\t18858\\t4\\n8\\t20644\\t4\\n8\\t43310\\t4\\n8\\t5806\\t4\\n9\\t33430\\t4\\n11\\t12145\\t4\\n11\\t12913\\t3\\n11\\t12920\\t5\\n11\\t14034\\t3.5\\n11\\t14277\\t4\\n11\\t14506\\t4\\n11\\t14889\\t5\\n11\\t15154\\t3.5\\n11\\t15815\\t5\\n11\\t16476\\t4.5\\n11\\t18181\\t5\\n11\\t18874\\t5\\n\\\"\\\"\\\"\\n\\ndef _parse_ratings(filename):\\n \\\"\\\"\\\"Parse the user csv file\\\"\\\"\\\"\\n\\n filename = _get_abs_path(filename)\\n \\n last_user = User(id=0) \\n\\n # open the csv reader\\n reader = csv.reader(open(filename, \\\"rb\\\"), delimiter=separator, quoting=csv.QUOTE_NONE)\\n \\n for user_id, movie_id, rating in reader:\\n \\n if reader.line_num % 1000 == 0:\\n print '%s lines processed' % reader.line_num \\n \\n # parse the ids and rating\\n user_id = int(user_id)\\n movie_id = long(movie_id) \\n rating = rating\\n\\n # if it's about the last user, use him, otherwise get him\\n if user_id == last_user.id: \\n user = last_user\\n else:\\n # take only users we already have\\n user = User.objects.filter(id=user_id) \\n if not user:\\n continue\\n \\n user=user[0]\\n \\n # get or create the movie\\n movie, created = Movie.objects.get_or_create(id=movie_id)\\n \\n # create the rating\\n Rating.objects.create(\\n user=user, \\n movie=movie,\\n rating=rating)\\n \\n # save the user for the next time\\n last_user = user\\n \\n\"\n },\n {\n \"alpha_fraction\": 0.6356180310249329,\n \"alphanum_fraction\": 0.6367896795272827,\n \"avg_line_length\": 38.39230728149414,\n \"blob_id\": \"ceb6010c1aebeea284fd0d160826418e3b6dfca1\",\n \"content_id\": \"14522cb466bf528486bf4859ebe693234d0f7c19\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5121,\n \"license_type\": \"no_license\",\n \"max_line_length\": 108,\n \"num_lines\": 130,\n \"path\": \"/code/adapter/unresyst/algorithm/base.py\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"The base abstract class for all recommender algorithms.\\\"\\\"\\\"\\n\\nfrom unresyst.constants import *\\nfrom unresyst.models.algorithm import RelationshipPredictionInstance\\n\\nclass BaseAlgorithm(object):\\n \\\"\\\"\\\"The interface provided to the other packages. The methods in the interface\\n are implemented by the subclasses\\\"\\\"\\\"\\n \\n def __init__(self, inner_algorithm=None):\\n \\\"\\\"\\\"The initializer\\\"\\\"\\\"\\n \\n self.inner_algorithm = inner_algorithm\\n \\\"\\\"\\\"The inner algorithm\\\"\\\"\\\"\\n \\n # Build phase:\\n #\\n \\n def build(self, recommender_model):\\n \\\"\\\"\\\"Build the recommender, so that the given relationship can be\\n predicted easily.\\n \\n @type recommender_model: models.common.Recommender\\n @param recommender_model: the recommender whose instances should\\n be aggregated. \\n\\n \\\"\\\"\\\"\\n if self.inner_algorithm:\\n self.inner_algorithm.build(recommender_model=recommender_model)\\n \\n \\n # Recommend phase:\\n #\\n \\n def get_relationship_prediction(self, recommender_model, dn_subject, dn_object, remove_predicted):\\n \\\"\\\"\\\"Get the prediction of the appearance of the predicted_relationship.\\n\\n @type recommender_model: models.common.Recommender\\n @param recommender_model: the recommender whose instances should\\n be aggregated. \\n \\n @type dn_subject: models.common.SubjectObject\\n @param dn_subject: the domain neutral subject\\n \\n @type dn_object: models.common.SubjectObject\\n @param dn_object: the domain neutral object\\n\\n @type remove_predicted: bool\\n @param remove_predicted: should pairs already having \\n predicted_relationship between them get the special expectancy value?\\n \\n @rtype: models.algorithm.RelationshipPredictionInstance\\n @return: the model instance for the prediction \\n \\\"\\\"\\\"\\n if self.inner_algorithm:\\n return self.inner_algorithm.get_relationship_prediction(\\n recommender_model=recommender_model, \\n dn_subject=dn_subject, \\n dn_object=dn_object, \\n remove_predicted=remove_predicted)\\n\\n \\n def get_recommendations(self, recommender_model, dn_subject, count, expectancy_limit, remove_predicted):\\n \\\"\\\"\\\"Get the recommendations for the given subject\\n\\n @type recommender_model: models.common.Recommender\\n @param recommender_model: the recommender whose instances should\\n be aggregated. \\n\\n @type dn_subject: models.common.SubjectObject\\n @param dn_subject: the domain neutral subject\\n\\n @type count: int\\n @param count: how many recommendations should be obtained\\n \\n @type expectancy_limit: float\\n @param expectancy_limit: the lower limit for object expectancy, only\\n objects with expectancy higher than limit are recommended.\\n \\n @type remove_predicted: bool\\n @param remove_predicted: should pairs already having \\n predicted_relationship between them be removed from recommendations?\\n \\n @rtype: a list of models.algorithm.RelationshipPredictionInstance\\n @return: the predictions of the objects recommended to the subject\\n \\\"\\\"\\\"\\n if self.inner_algorithm:\\n return self.inner_algorithm.get_recommendations(\\n recommender_model=recommender_model,\\n dn_subject=dn_subject,\\n count=count,\\n expectancy_limit=expectancy_limit,\\n remove_predicted=remove_predicted)\\n\\n @staticmethod\\n def _get_uncertain_prediction(recommender_model, dn_subject, dn_object):\\n \\\"\\\"\\\"Get the prediction for a pair for which nothing is known\\\"\\\"\\\"\\n \\n return RelationshipPredictionInstance(\\n subject_object1=dn_subject,\\n subject_object2=dn_object,\\n description=recommender_model.random_recommendation_description,\\n recommender=recommender_model,\\n expectancy=UNCERTAIN_PREDICTION_VALUE,\\n is_uncertain=True\\n ) \\n\\n @staticmethod\\n def _get_already_in_relatinship_prediction(recommender_model, predicted_relationship):\\n \\\"\\\"\\\"Get the prediction for a pair which is already in the predicted_relationship.\\n Valid only for recommenders with remove_predicted=True.\\n \\\"\\\"\\\"\\n \\n return RelationshipPredictionInstance(\\n subject_object1=predicted_relationship.subject_object1,\\n subject_object2=predicted_relationship.subject_object2,\\n description=predicted_relationship.description,\\n recommender=recommender_model,\\n expectancy=ALREADY_IN_REL_PREDICTION_VALUE,\\n is_trivial=True\\n ) \\n\\n # Update phase:\\n # \\n \\n def update(self, recommender, aggregated_changes):\\n \\\"\\\"\\\"Update the recommender algorithm structures according to the changes\\n in the aggregated relationships.\\n \\n \\\"\\\"\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.6363636255264282,\n \"alphanum_fraction\": 0.63840651512146,\n \"avg_line_length\": 18.959182739257812,\n \"blob_id\": \"5686f75d1fbc60c81183db4c172749f280dd75d6\",\n \"content_id\": \"b006b353ab6ace2bb63d19825a6415d705edf892\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 979,\n \"license_type\": \"no_license\",\n \"max_line_length\": 178,\n \"num_lines\": 49,\n \"path\": \"/UnresystCD/code/adapter/traveleval.sh\",\n \"repo_name\": \"cvengros/unresyst\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/bin/sh\\n\\n\\n\\nBUILD=true\\nEVAL=''\\nREC=''\\nRECS=false\\nPREDS=false\\n\\nfor param in $*;\\ndo \\n case $param in\\n 'recs')\\n EVAL='OrderTourRecommenderEvaluator'\\n REC='OrderTourRecommender'\\n RECS=true\\n ;;\\n 'preds')\\n PREDS=true\\n ;;\\n 'dontbuild')\\n BUILD=false\\n ;;\\n esac\\ndone\\n\\nif [ $BUILD = true ]\\nthen\\n # build\\n echo \\\"Building...\\\"\\n echo \\\"from travel.recommender import *; OrderTourRecommender.build()\\\" | python ./manage.py shell\\nfi\\n\\nif [ $RECS = true ]\\nthen\\n echo \\\"Evaluating recommendations...\\\"\\n echo \\\"from travel.evaluation import *; from travel.recommender import *; $EVAL.evaluate_recommendations($REC, 10); quit()\\\"| python ./manage.py shell\\nfi\\n\\nif [ $PREDS = true ]\\nthen\\n echo \\\"Evaluating predicitons...\\\"\\n echo \\\"from travel.evaluation import *; from travel.recommender import *; OrderTourRankEvaluator.evaluate_predictions(OrderTourRecommender); quit()\\\" | python ./manage.py shell\\nfi\\n\\n\\n\\necho \\\"\\\"\\n\\n\"\n }\n]"},"num_files":{"kind":"number","value":104,"string":"104"}}},{"rowIdx":2980383,"cells":{"repo_name":{"kind":"string","value":"wilbarnes/jet-sweep"},"repo_url":{"kind":"string","value":"https://github.com/wilbarnes/jet-sweep"},"snapshot_id":{"kind":"string","value":"7ed19319e2d598447e23c245fb9386174bc6e2c0"},"revision_id":{"kind":"string","value":"1077ce600346420283dfcf12cda3111c62c47d1d"},"directory_id":{"kind":"string","value":"72ffe0a3498f59b47918abd56aeed4fb9f94e0ef"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2020-05-20T05:49:24.383657","string":"2020-05-20T05:49:24.383657"},"revision_date":{"kind":"timestamp","value":"2019-05-07T14:26:20","string":"2019-05-07T14:26:20"},"committer_date":{"kind":"timestamp","value":"2019-05-07T14:26:20","string":"2019-05-07T14:26:20"},"github_id":{"kind":"number","value":185415386,"string":"185,415,386"},"star_events_count":{"kind":"number","value":1,"string":"1"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"null"},"gha_updated_at":{"kind":"null"},"gha_pushed_at":{"kind":"null"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.7846994400024414,"alphanum_fraction":0.7890710234642029,"avg_line_length":82.18181610107422,"blob_id":"d8acf18340e62f39a5a04ad6ef5a511f6c67907c","content_id":"2164d431bac6cc227bc79b7d1b362601c68a677f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":915,"license_type":"no_license","max_line_length":187,"num_lines":11,"path":"/README.md","repo_name":"wilbarnes/jet-sweep","src_encoding":"UTF-8","text":"# jet-sweep\nMinimize idle crypto drag by temporarily sweeping your digital assets to higher interest accounts (in this case, Compound v2 Presidio). Open to exploring additional defi infrastructures. \n\nSparked from an idea on how to offset the rising MakerDAO stability fee costs. Goal is to use hot/cold idle ETH to capture interest overnight.\n\n'jet-sweep' named after the [spectacular football play.](https://youtu.be/atrBEX-GP7E?t=78) \n\n[Investopedia definition](https://www.investopedia.com/terms/s/sweepaccount.asp): \n* A sweep account automatically transfers cash funds into a safe but higher interest-earning investment option at the close of each business day, e.g. into a money market fund.\n* Sweep accounts try to minimize idle cash drag.\n* A sweep account service, however, may not always be free - and you might have to pay fees to your broker that might make the sweep not as attractive on a net basis.\n"},{"alpha_fraction":0.6104430556297302,"alphanum_fraction":0.6145569682121277,"avg_line_length":29.095237731933594,"blob_id":"f00a4d0e3349b10375e9aea04e4624708b42e53e","content_id":"c9cb992a4dbf45a72a3655a13f3f56987417076c","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":6320,"license_type":"no_license","max_line_length":106,"num_lines":210,"path":"/main.py","repo_name":"wilbarnes/jet-sweep","src_encoding":"UTF-8","text":"import os\nimport sys\nimport datetime\nimport time\nimport json\nimport argparse\nimport configparser\n\nfrom eth_utils import to_checksum_address\nfrom web3 import Web3\n\nfrom compoundv2 import Address\nfrom compoundv2.cETH import cETH\nfrom compoundv2.Comptroller import Comptroller\n\ndef main():\n \"\"\"\n Generic scripts for interacting with Compound v2 Presidio\n \"\"\"\n\n if args.enter in ('eth', 'Eth', 'ETH'):\n enter_markets_tx = contract_Comptroller.enter_markets(cETH_addr, address_one)\n print(enter_markets_tx)\n\n elif args.get_assets_in:\n get_assets_call = contract_Comptroller.get_assets_in(address_one)\n print(get_assets_call)\n\n elif args.get_balance:\n get_balance_call = contract_cETH.balanceOf(address_one.address)\n print(get_balance_call)\n\n elif args.get_liquidity:\n get_liquidity_call = contract_Comptroller.get_liquidity(address_one)\n print(get_liquidity_call)\n \n elif args.supply and args.eth:\n supply_eth = contract_cETH.mint(args.eth)\n tx = supply_eth.send_raw_transaction()\n print(web3.toHex(tx['transactionHash']))\n\n elif args.supply_rate:\n supply_rate_call = contract_cETH.supplyRatePerBlock()\n print(supply_rate_call / 1e18)\n\n elif args.exchange_rate:\n exchange_rate_call = contract_cETH.exchangeRateCurrent()\n print(exchange_rate_call / 1e18)\n\n elif args.redeem and args.eth:\n redeem_eth = contract_cETH.redeem(args.eth)\n print(redeem_eth)\n tx = redeem_eth.send_raw_transaction()\n print('ETH redeemed.')\n print(web3.toHex(tx['transactionHash']))\n\n elif args.sweep_overnight and args.eth:\n supply_eth = contract_cETH.mint(args.eth)\n tx = supply_eth.send_raw_transaction()\n get_balance_call = contract_cETH.balanceOf(address_one.address)\n print(get_balance_call)\n time.sleep(60 * 15)\n redeem_eth = contract_cETH.redeem(get_balance_call)\n tx = redeem_eth.send_raw_transaction()\n print('Sweep complete')\n print(web3.toHex(tx['transactionHash']))\n\n elif args.supply and args.dai:\n print('Supplying dai is not yet implemented.')\n\n elif args.borrow and args.eth:\n # requires market entry\n print('Borrowing eth is not yet implemented.')\n\n elif args.borrow and args.dai:\n # requires market entry\n print('Borrowing dai is not yet implemented.')\n\n else:\n print('Not a valid option.')\n\nif __name__ == \"__main__\":\n # set config parser and read config variables\n parser = argparse.ArgumentParser('Jet sweep your assets to higher earning contracts')\n config = configparser.ConfigParser()\n config.read('config.ini')\n\n private_key = config['ACCOUNT']['PrivateKey']\n address_one = Address(config['ACCOUNT']['Address'])\n infura_mainnet = config['PROVIDER']['InfuraMainnet']\n infura_rinkeby = config['PROVIDER']['InfuraRinkeby']\n infura_secret = config['PROVIDER']['InfuraProjSecret']\n\n # compound.finance contracts\n cDAI = to_checksum_address(config['COMPOUND-CONTRACTS']['cDAI'])\n cETH_addr = Address(config['COMPOUND-CONTRACTS']['cETH'])\n comptroller_addr = Address(config['COMPOUND-CONTRACTS']['Comptroller'])\n price_oracle = to_checksum_address(config['COMPOUND-CONTRACTS']['PriceOracle'])\n stable_interest = to_checksum_address(config['COMPOUND-CONTRACTS']['StableCoinInterestRateModel'])\n standard_interest = to_checksum_address(config['COMPOUND-CONTRACTS']['StandardInterestRateModel'])\n\n # arguments\n parser.add_argument(\n '--supply', \n dest='supply', \n action='store_true', \n help='supply asset'\n )\n\n parser.add_argument(\n '--borrow', \n dest='borrow', \n action='store_true', \n help='borrow asset'\n )\n\n parser.add_argument(\n '--redeem', \n dest='redeem', \n action='store_true', \n help='redeem asset'\n )\n\n parser.add_argument(\n '--supply-rate', \n dest='supply_rate', \n action='store_true', \n help='retrieve supply rate per block'\n )\n\n parser.add_argument(\n '--exchange-rate', \n dest='exchange_rate', \n action='store_true', \n help='retrieve current exchange rate'\n )\n\n parser.add_argument(\n '--eth', \n dest='eth', \n help='ether'\n )\n\n parser.add_argument(\n '--dai', \n dest='dai', \n help='dai'\n )\n\n # markets argument group\n markets_grp = parser.add_argument_group(\n title=\"ENTER/EXIT MARKETS\",\n description=\"commands for entering & exiting markets\"\n )\n\n markets_grp.add_argument(\n '--enter', \n dest='enter', \n help='enter specified market'\n )\n\n # reporting argument group\n reporting_grp = parser.add_argument_group(\n title=\"REPORTING\", \n description=\"short form reports of your open positions\"\n )\n\n reporting_grp.add_argument(\n '--get-assets-in', \n dest = 'get_assets_in', \n action = 'store_true', \n help = 'get list of entered markets'\n )\n\n reporting_grp.add_argument(\n '--get-liquidity',\n dest = 'get_liquidity',\n action = 'store_true',\n help = 'return estimated ether value of account\\'s collateral'\n )\n\n reporting_grp.add_argument(\n '--get-balance-of',\n dest = 'get_balance',\n action = 'store_true',\n help = 'get cToken balance of provided account'\n )\n\n sweep_grp = parser.add_argument_group(\n title=\"sweep options for your digital assets\",\n description='quick POC for sweeping your ETH into high interest accounts'\n )\n\n sweep_grp.add_argument(\n '--sweep-overnight',\n dest = 'sweep_overnight',\n action = 'store_true',\n help = 'sweep specified amount of eth to a high interest account once overnight'\n )\n\n\n args = parser.parse_args()\n\n web3 = Web3(Web3.HTTPProvider(infura_rinkeby))\n print('Web3 connected:', web3.isConnected())\n \n contract_cETH = cETH(web3, cETH_addr, address_one)\n contract_Comptroller = Comptroller(web3, comptroller_addr)\n\n main()\n"},{"alpha_fraction":0.7490196228027344,"alphanum_fraction":0.7647058963775635,"avg_line_length":27.33333396911621,"blob_id":"05a1ac376acd307d0393e3f7be0a2ad7cae892aa","content_id":"f99a9b7299d09f8d4cbfc64bfe0ee3a35ef6e966","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":255,"license_type":"no_license","max_line_length":54,"num_lines":9,"path":"/compoundv2/account.py","repo_name":"wilbarnes/jet-sweep","src_encoding":"UTF-8","text":"import os\nfrom eth_account import Account\nfrom web3 import Web3\nfrom typing import Optional\n\ndef private_key_to_account(web3: Web3):\n private_key = os.environ['ETH_PRIVATE_KEY']\n account = Account.privateKeyToAccount(private_key)\n return account\n"}],"string":"[\n {\n \"alpha_fraction\": 0.7846994400024414,\n \"alphanum_fraction\": 0.7890710234642029,\n \"avg_line_length\": 82.18181610107422,\n \"blob_id\": \"d8acf18340e62f39a5a04ad6ef5a511f6c67907c\",\n \"content_id\": \"2164d431bac6cc227bc79b7d1b362601c68a677f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 915,\n \"license_type\": \"no_license\",\n \"max_line_length\": 187,\n \"num_lines\": 11,\n \"path\": \"/README.md\",\n \"repo_name\": \"wilbarnes/jet-sweep\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# jet-sweep\\nMinimize idle crypto drag by temporarily sweeping your digital assets to higher interest accounts (in this case, Compound v2 Presidio). Open to exploring additional defi infrastructures. \\n\\nSparked from an idea on how to offset the rising MakerDAO stability fee costs. Goal is to use hot/cold idle ETH to capture interest overnight.\\n\\n'jet-sweep' named after the [spectacular football play.](https://youtu.be/atrBEX-GP7E?t=78) \\n\\n[Investopedia definition](https://www.investopedia.com/terms/s/sweepaccount.asp): \\n* A sweep account automatically transfers cash funds into a safe but higher interest-earning investment option at the close of each business day, e.g. into a money market fund.\\n* Sweep accounts try to minimize idle cash drag.\\n* A sweep account service, however, may not always be free - and you might have to pay fees to your broker that might make the sweep not as attractive on a net basis.\\n\"\n },\n {\n \"alpha_fraction\": 0.6104430556297302,\n \"alphanum_fraction\": 0.6145569682121277,\n \"avg_line_length\": 29.095237731933594,\n \"blob_id\": \"f00a4d0e3349b10375e9aea04e4624708b42e53e\",\n \"content_id\": \"c9cb992a4dbf45a72a3655a13f3f56987417076c\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 6320,\n \"license_type\": \"no_license\",\n \"max_line_length\": 106,\n \"num_lines\": 210,\n \"path\": \"/main.py\",\n \"repo_name\": \"wilbarnes/jet-sweep\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import os\\nimport sys\\nimport datetime\\nimport time\\nimport json\\nimport argparse\\nimport configparser\\n\\nfrom eth_utils import to_checksum_address\\nfrom web3 import Web3\\n\\nfrom compoundv2 import Address\\nfrom compoundv2.cETH import cETH\\nfrom compoundv2.Comptroller import Comptroller\\n\\ndef main():\\n \\\"\\\"\\\"\\n Generic scripts for interacting with Compound v2 Presidio\\n \\\"\\\"\\\"\\n\\n if args.enter in ('eth', 'Eth', 'ETH'):\\n enter_markets_tx = contract_Comptroller.enter_markets(cETH_addr, address_one)\\n print(enter_markets_tx)\\n\\n elif args.get_assets_in:\\n get_assets_call = contract_Comptroller.get_assets_in(address_one)\\n print(get_assets_call)\\n\\n elif args.get_balance:\\n get_balance_call = contract_cETH.balanceOf(address_one.address)\\n print(get_balance_call)\\n\\n elif args.get_liquidity:\\n get_liquidity_call = contract_Comptroller.get_liquidity(address_one)\\n print(get_liquidity_call)\\n \\n elif args.supply and args.eth:\\n supply_eth = contract_cETH.mint(args.eth)\\n tx = supply_eth.send_raw_transaction()\\n print(web3.toHex(tx['transactionHash']))\\n\\n elif args.supply_rate:\\n supply_rate_call = contract_cETH.supplyRatePerBlock()\\n print(supply_rate_call / 1e18)\\n\\n elif args.exchange_rate:\\n exchange_rate_call = contract_cETH.exchangeRateCurrent()\\n print(exchange_rate_call / 1e18)\\n\\n elif args.redeem and args.eth:\\n redeem_eth = contract_cETH.redeem(args.eth)\\n print(redeem_eth)\\n tx = redeem_eth.send_raw_transaction()\\n print('ETH redeemed.')\\n print(web3.toHex(tx['transactionHash']))\\n\\n elif args.sweep_overnight and args.eth:\\n supply_eth = contract_cETH.mint(args.eth)\\n tx = supply_eth.send_raw_transaction()\\n get_balance_call = contract_cETH.balanceOf(address_one.address)\\n print(get_balance_call)\\n time.sleep(60 * 15)\\n redeem_eth = contract_cETH.redeem(get_balance_call)\\n tx = redeem_eth.send_raw_transaction()\\n print('Sweep complete')\\n print(web3.toHex(tx['transactionHash']))\\n\\n elif args.supply and args.dai:\\n print('Supplying dai is not yet implemented.')\\n\\n elif args.borrow and args.eth:\\n # requires market entry\\n print('Borrowing eth is not yet implemented.')\\n\\n elif args.borrow and args.dai:\\n # requires market entry\\n print('Borrowing dai is not yet implemented.')\\n\\n else:\\n print('Not a valid option.')\\n\\nif __name__ == \\\"__main__\\\":\\n # set config parser and read config variables\\n parser = argparse.ArgumentParser('Jet sweep your assets to higher earning contracts')\\n config = configparser.ConfigParser()\\n config.read('config.ini')\\n\\n private_key = config['ACCOUNT']['PrivateKey']\\n address_one = Address(config['ACCOUNT']['Address'])\\n infura_mainnet = config['PROVIDER']['InfuraMainnet']\\n infura_rinkeby = config['PROVIDER']['InfuraRinkeby']\\n infura_secret = config['PROVIDER']['InfuraProjSecret']\\n\\n # compound.finance contracts\\n cDAI = to_checksum_address(config['COMPOUND-CONTRACTS']['cDAI'])\\n cETH_addr = Address(config['COMPOUND-CONTRACTS']['cETH'])\\n comptroller_addr = Address(config['COMPOUND-CONTRACTS']['Comptroller'])\\n price_oracle = to_checksum_address(config['COMPOUND-CONTRACTS']['PriceOracle'])\\n stable_interest = to_checksum_address(config['COMPOUND-CONTRACTS']['StableCoinInterestRateModel'])\\n standard_interest = to_checksum_address(config['COMPOUND-CONTRACTS']['StandardInterestRateModel'])\\n\\n # arguments\\n parser.add_argument(\\n '--supply', \\n dest='supply', \\n action='store_true', \\n help='supply asset'\\n )\\n\\n parser.add_argument(\\n '--borrow', \\n dest='borrow', \\n action='store_true', \\n help='borrow asset'\\n )\\n\\n parser.add_argument(\\n '--redeem', \\n dest='redeem', \\n action='store_true', \\n help='redeem asset'\\n )\\n\\n parser.add_argument(\\n '--supply-rate', \\n dest='supply_rate', \\n action='store_true', \\n help='retrieve supply rate per block'\\n )\\n\\n parser.add_argument(\\n '--exchange-rate', \\n dest='exchange_rate', \\n action='store_true', \\n help='retrieve current exchange rate'\\n )\\n\\n parser.add_argument(\\n '--eth', \\n dest='eth', \\n help='ether'\\n )\\n\\n parser.add_argument(\\n '--dai', \\n dest='dai', \\n help='dai'\\n )\\n\\n # markets argument group\\n markets_grp = parser.add_argument_group(\\n title=\\\"ENTER/EXIT MARKETS\\\",\\n description=\\\"commands for entering & exiting markets\\\"\\n )\\n\\n markets_grp.add_argument(\\n '--enter', \\n dest='enter', \\n help='enter specified market'\\n )\\n\\n # reporting argument group\\n reporting_grp = parser.add_argument_group(\\n title=\\\"REPORTING\\\", \\n description=\\\"short form reports of your open positions\\\"\\n )\\n\\n reporting_grp.add_argument(\\n '--get-assets-in', \\n dest = 'get_assets_in', \\n action = 'store_true', \\n help = 'get list of entered markets'\\n )\\n\\n reporting_grp.add_argument(\\n '--get-liquidity',\\n dest = 'get_liquidity',\\n action = 'store_true',\\n help = 'return estimated ether value of account\\\\'s collateral'\\n )\\n\\n reporting_grp.add_argument(\\n '--get-balance-of',\\n dest = 'get_balance',\\n action = 'store_true',\\n help = 'get cToken balance of provided account'\\n )\\n\\n sweep_grp = parser.add_argument_group(\\n title=\\\"sweep options for your digital assets\\\",\\n description='quick POC for sweeping your ETH into high interest accounts'\\n )\\n\\n sweep_grp.add_argument(\\n '--sweep-overnight',\\n dest = 'sweep_overnight',\\n action = 'store_true',\\n help = 'sweep specified amount of eth to a high interest account once overnight'\\n )\\n\\n\\n args = parser.parse_args()\\n\\n web3 = Web3(Web3.HTTPProvider(infura_rinkeby))\\n print('Web3 connected:', web3.isConnected())\\n \\n contract_cETH = cETH(web3, cETH_addr, address_one)\\n contract_Comptroller = Comptroller(web3, comptroller_addr)\\n\\n main()\\n\"\n },\n {\n \"alpha_fraction\": 0.7490196228027344,\n \"alphanum_fraction\": 0.7647058963775635,\n \"avg_line_length\": 27.33333396911621,\n \"blob_id\": \"05a1ac376acd307d0393e3f7be0a2ad7cae892aa\",\n \"content_id\": \"f99a9b7299d09f8d4cbfc64bfe0ee3a35ef6e966\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 255,\n \"license_type\": \"no_license\",\n \"max_line_length\": 54,\n \"num_lines\": 9,\n \"path\": \"/compoundv2/account.py\",\n \"repo_name\": \"wilbarnes/jet-sweep\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import os\\nfrom eth_account import Account\\nfrom web3 import Web3\\nfrom typing import Optional\\n\\ndef private_key_to_account(web3: Web3):\\n private_key = os.environ['ETH_PRIVATE_KEY']\\n account = Account.privateKeyToAccount(private_key)\\n return account\\n\"\n }\n]"},"num_files":{"kind":"number","value":3,"string":"3"}}},{"rowIdx":2980384,"cells":{"repo_name":{"kind":"string","value":"bjp12/recipe-book"},"repo_url":{"kind":"string","value":"https://github.com/bjp12/recipe-book"},"snapshot_id":{"kind":"string","value":"8b864adaae6c44d8792bd86eed913076bc4558a8"},"revision_id":{"kind":"string","value":"5ec410c8c830d5b4edb671bac64a68a10ee81140"},"directory_id":{"kind":"string","value":"ba811a13ff2f7ec0efcf049cbc7dfe5c9be5afea"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2020-06-02T20:12:14.310775","string":"2020-06-02T20:12:14.310775"},"revision_date":{"kind":"timestamp","value":"2019-12-04T01:10:25","string":"2019-12-04T01:10:25"},"committer_date":{"kind":"timestamp","value":"2019-12-04T01:10:25","string":"2019-12-04T01:10:25"},"github_id":{"kind":"number","value":191294186,"string":"191,294,186"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"null"},"gha_updated_at":{"kind":"null"},"gha_pushed_at":{"kind":"null"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.5807398557662964,"alphanum_fraction":0.594245433807373,"avg_line_length":27.88135528564453,"blob_id":"499ea62c45a96d493e62d092bcf549445eedd7f7","content_id":"496eb6d8e80f237dcb8a9400a7c0d8dc8cd51cb3","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"HTML","length_bytes":1703,"license_type":"no_license","max_line_length":133,"num_lines":59,"path":"/recipe book/templates/ingredients.html","repo_name":"bjp12/recipe-book","src_encoding":"UTF-8","text":"\n\n\n\tIngredient Picker \n\t{% include 'header.html' %}\n\t\n\n\n\n\n\tIngredient Picker \\n\\t{% include 'header.html' %}\\n\\t\\n\\n\\n\\n\\n\\t]*?href=(?:\\'|\\\")(?:https://|http://|ftp://|)(.*?)[\\\"|\\'|/|:]')\nresp = requests.get(html_file).text\nurl = link_pattern.findall(resp)\nurl2 = []\nfor i in url:\n\tif not i.startswith(\"..\"):\n\t\turl2.append(i)\nurl2 = sorted(list(set(url2)))\n# url = sorted(url)\n# url = set(url)\n# match_object = re.search(pattern, line)\nprint(*url2, sep='\\n')\n\n# import re\n# import requests\n#\n# resp = requests.get(input()).text\n# sites = set()\n# for site in re.findall(r'= neg)\n return pos >= neg\n\n def judge_any(pos, neg):\n # допускает элемент, если его допускает хотя бы одна функция (pos >= 1)\n return pos >= 1\n\n def judge_all(pos, neg):\n # допускает элемент, если его допускают все функции (neg == 0)\n return neg == 0\n\n def __init__(self, iterable, *funcs, judge=judge_any):\n # iterable - исходная последовательность\n # funcs - допускающие функции\n # judge - решающая функция\n self.i = 0\n self.judge = judge\n self.funcs = funcs\n self.iterable = iterable\n self.k = len(self.iterable)\n\n def __iter__(self):\n # возвращает итератор по результирующей последовательности\n return self\n\n def __next__(self):\n while self.i < self.k:\n pos = 0\n neg = 0\n for _ in self.funcs:\n if _(self.iterable[self.i]):\n pos += 1\n else:\n neg += 1\n self.i += 1\n if self.judge(pos, neg):\n return self.iterable[self.i - 1]\n raise StopIteration\n\n\n\n\n\n\n\ndef mul2(x):\n return x % 2 == 0\n\ndef mul3(x):\n return x % 3 == 0\n\ndef mul5(x):\n return x % 5 == 0\n\na = [i for i in range(31)] # [0, 1, 2, ... , 30]\nb = multifilter(a, mul2, mul3, mul5)\nprint(list(b))\nprint(list(b))\n# [0, 2, 3, 4, 5, 6, 8, 9, 10, 12, 14, 15, 16, 18, 20, 21, 22, 24, 25, 26, 27, 28, 30]\n\n\n# print(list(multifilter(a, mul2, mul3, mul5, judge=multifilter.judge_half)))\n# # [0, 6, 10, 12, 15, 18, 20, 24, 30]\n#\n# print(list(multifilter(a, mul2, mul3, mul5, judge=multifilter.judge_all)))\n# # [0, 30]\n\n\n# # с применением генератора в одной строке\n# class multifilter:\n# judge_half = lambda fx: sum(fx) >= len(fx) / 2\n# judge_any = lambda fx: any(fx)\n# judge_all = lambda fx: all(fx)\n#\n# def __init__(self, iterable, *funcs, judge=judge_any):\n# self.iterable = iterable\n# self.funcs = funcs\n# self.judge = judge\n#\n# def __iter__(self):\n# return (x for x in self.iterable if self.judge([f(x) for f in self.funcs]))\n\n\n# собственное решение через yield (вместо __next__)\n#\n# логика работы прописана в итераторе объекта __iter__\n# где пробегаемся по элементам исходной последовательности\n# просеиваем элемент через функции-фильтры\n# и если выбранная функция judge дала добро\n# то yield`им этот элемент\n#\n#\n#\n\n# # решение с yield в iter\n# class multifilter:\n# def judge_half(pos, neg):\n# return pos >= neg # допускает элемент, если его допускает хотя бы половина фукнций (pos >= neg)\n#\n# def judge_any(pos, neg):\n# return pos >= 1 # допускает элемент, если его допускает хотя бы одна функция (pos >= 1)\n#\n# def judge_all(pos, neg):\n# return neg == 0 # допускает элемент, если его допускают все функции (neg == 0)\n#\n# def __init__(self, iterable, *funcs, judge=judge_any):\n# self.lst = iterable # iterable - исходная последовательность\n# self.funcs = funcs # funcs - допускающие функции\n# self.judge = judge # judge - решающая функция\n#\n# def __iter__(self):\n# for el in self.lst:\n# pos, neg = 0, 0\n# for f in self.funcs:\n# pass\n# if f(el):\n# pos += 1\n# else:\n# neg += 1\n# if self.judge(pos, neg):\n# yield el"},{"alpha_fraction":0.5600000023841858,"alphanum_fraction":0.6200000047683716,"avg_line_length":19.200000762939453,"blob_id":"5c5a923429492556e6caf3b1e10d72a1675be624","content_id":"d108970f7d33e8326c36f572dfdc1e067014f74b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":100,"license_type":"no_license","max_line_length":30,"num_lines":5,"path":"/Python_stepik_2/2.5.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"def mod_checker(x, mod=0):\n\treturn lambda y: y % x == mod\n\nmod_3 = mod_checker(3, 1)\nprint(mod_3(4))"},{"alpha_fraction":0.7330895662307739,"alphanum_fraction":0.7586837410926819,"avg_line_length":41.07692337036133,"blob_id":"9511b7f652c20f4ee353c39976c5733ee7489a89","content_id":"0031654e8d97abca3d5f52e261eb223329012ae0","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":602,"license_type":"no_license","max_line_length":119,"num_lines":13,"path":"/Django_projects/vk_oauth/vk_oauth/apps/articles/models.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"from django.db import models\n\n\nclass Article(models.Model):\n\tarticle_title = models.CharField('название статьи', max_length=200) # VarChar in MySQL - 200-300 simbols?\n\tarticle_text = models.TextField('текст статьи')\n\tpub_date = models.DateTimeField('дата публикации')\n\n\nclass Comment(models.Model):\n\tarticle = models.ForeignKey(Article, on_delete=models.CASCADE) #on_delete - delete all its comments if delete article\n\tauthor_name = models.CharField('имя автора', max_length=50)\n\tcomment_text = models.CharField('текст комментария', max_length=200)\n"},{"alpha_fraction":0.3900621235370636,"alphanum_fraction":0.47577640414237976,"avg_line_length":30,"blob_id":"ed831e5b9684586bedf787e2de269b5b5c9b58da","content_id":"b0cbc3a72d140dd54b3fe72840316a26eb5813c9","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":805,"license_type":"no_license","max_line_length":84,"num_lines":26,"path":"/MyPr1/ex12.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"# import os\n# s = ''\n# t = ''\n# num = '1'\n# with open(r'C:\\Users\\arkka\\PycharmProjects\\MyPr1\\dataset_3363_2.txt', 'r') as inf:\n# line = inf.readline().strip()\n# for i in range(len(line)):\n# if line[i] in ['1', '2', '3', '4', '5', '6', '7', '8', '9', '0']:\n# num += line[i]\n# elif t != '':\n# s += t * int(num)\n# t = line[i]\n# num = ''\n# else:\n# t = line[i]\n# num = ''\n# s += t * int(num)\n# with open(r'C:\\Users\\arkka\\PycharmProjects\\MyPr1\\dataset_3363_2.txt', 'w') as inf:\n# inf.write(s + '\\n')\n\n\ns, d = 'D8s11T20s13r16V8O9o11f3V3S5B14T11m12n18i17T18x15d17V1b5l2x13I5t12q7c14', []\nfor i in s:\n if not i.isdigit(): d.append(i)\n else: d[-1] += i\nprint(*[i[0]*int(i[1:]) for i in d], sep='')"},{"alpha_fraction":0.6551724076271057,"alphanum_fraction":0.6551724076271057,"avg_line_length":18.44444465637207,"blob_id":"77448c835d5001a0646d2d046c6839203fb0585a","content_id":"e8b24b6ada1d429440b15ee85d309f6dc9748360","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":174,"license_type":"no_license","max_line_length":63,"num_lines":9,"path":"/Python_stepik_2/3.3.6.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import re\nimport sys\n\nfor line in sys.stdin:\n\tline = line.rstrip()\n\tprint(re.sub(\"human\", \"computer\", line))\n\n\n# print(re.sub(r'human', 'computer', sys.stdin.read()), end='')"},{"alpha_fraction":0.6403940916061401,"alphanum_fraction":0.6403940916061401,"avg_line_length":34.034481048583984,"blob_id":"50c84dbf33db6231fde4963ea5d754abfc39e455","content_id":"562320c846388f19bd4e002df680a8f95305cd71","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1339,"license_type":"no_license","max_line_length":79,"num_lines":29,"path":"/MyPr1/ex3.7.3.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"a = int(input())\nb = set()\n[b.add(*input().lower().split('\\n')) for i in range(a)]\nc = int(input())\nd = set()\nfor i in range(c):\n g = input().lower().split(' ')\n [d.add(i) for i in g]\n[print(i) for i in d if i not in b]\n\n# # # формируем множество известных слов на основании построчного ввода\n# # dic = {input().lower() for _ in range(int(input()))}\n# #\n# # # заводим пустое множество для приема текста\n# wrd = set()\n#\n# # т.к. текст построчно подается, а также в каждой строке несколько слов,\n# # то каждую строку превращаем во множество и добавляем в единое множество wrd\n# for _ in range(int(input())):\n# wrd |= {i.lower() for i in input().split()}\n#\n# # на вывод отправляем результат вычитания словарного множества dic\n# # из текстового множества wrd; впереди ставим *, чтобы раскрыть поэлементно\n# print(*(wrd-dic), sep=\"\\n\")\n\n\n# words = set(input().lower() for i in range(int(input())))\n# text = set(('\\n'.join(input().lower() for i in range(int(input())))).split())\n# print('\\n'.join(text - words))"},{"alpha_fraction":0.625261664390564,"alphanum_fraction":0.6329379081726074,"avg_line_length":27.117647171020508,"blob_id":"0d844e6856effdcc44170acb6c9ea5a4d85e3d2a","content_id":"63821cb3d99617e244763533ab9e20616cf83c1d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1433,"license_type":"no_license","max_line_length":109,"num_lines":51,"path":"/Python_stepik_2/3.7.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"# from xml.etree import ElementTree\nfrom xml.etree.ElementTree import XMLParser\n\n# s = input()\ns = '=n-1: i+=1\n elif i>=j and i+j>n-1: j-=1\n elif i>j+1 and i+j<=n-1: i-=1\nfor i in range(n):\n print(*t[i])"},{"alpha_fraction":0.5873016119003296,"alphanum_fraction":0.6142857074737549,"avg_line_length":21.535715103149414,"blob_id":"005820b914d674dddd4a9dcfadfea96bcf9bb6e6","content_id":"01be5c981785e04082250e457187692799543092","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":630,"license_type":"no_license","max_line_length":82,"num_lines":28,"path":"/Python_stepik_2/3.5.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import csv\n\nwith open(\"Crimes.csv\", 'r') as f:\n\treader = csv.reader(f)\n\tr = (i[5] for i in reader if '2015' in i[2])\n\td = dict()\n\tfor word in r:\n\t\tif word in d:\n\t\t\td[word] += 1\n\t\telse:\n\t\t\td[word] = 0\n\td = {k: v for k, v in sorted(d.items(), key=lambda item: item[1], reverse=True)}\n\tprint(d)\n\n# import csv\n# from collections import Counter as c\n#\n# with open('Crimes.csv') as f:\n# \tdata = csv.reader(f)\n# \tprint(c( row[5] for row in data if '2015' in row[2] ))\n\n\n# import csv\n# from collections import Counter as c\n#\n# with open('Crimes.csv') as f:\n# result = c(row[5] for row in csv.reader(f))\n# print(result.most_common(1))"},{"alpha_fraction":0.5309917330741882,"alphanum_fraction":0.6012396812438965,"avg_line_length":47.5,"blob_id":"894a1e63f1d5c279152c5761ab3fdea8c9e893c3","content_id":"f2e7d0583a9a34e69125528191df17a74c76998b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":484,"license_type":"no_license","max_line_length":89,"num_lines":10,"path":"/Python_stepik_2/2.4.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"with open('dataset_24465_4.txt', 'r') as f, open('test_copy.txt', 'w') as w:\n\tr = [line.strip() for line in f]\n\tw.write('\\n'.join([r[i] for i in range(len(r)-1, -1, -1)]))\n\n\n# with open('dataset_24465_4.txt', 'r') as fr, open('dataset_24465_4_w.txt', 'w') as fw:\n# fw.writelines(fr.readlines()[::-1])\n\n# with open('dataset_24465_4.txt') as inp, open('dataset_24465_4_write.txt', 'w') as out:\n# [out.write('{}\\n'.format(oln)) for oln in reversed([iln.rstrip() for iln in inp])]"},{"alpha_fraction":0.30249109864234924,"alphanum_fraction":0.3327402174472809,"avg_line_length":16.30769157409668,"blob_id":"041f5e7719f6532d4ba15a616b803453b610a024","content_id":"1884bb844d77b0de55d9955dc383d2cbaf5f84bd","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1124,"license_type":"no_license","max_line_length":51,"num_lines":65,"path":"/MyPr1/ex4.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"# n = int(input())\n# i = 0\n# k = 0\n# j = 0\n# s = ''\n# if n == 0:\n# print(s)\n# else:\n# while k <= n:\n# i += 1\n# for j in range(i):\n# k += 1\n# if k < n:\n# # print(i, end=' ')\n# s += str(i) + ' '\n# elif k == n:\n# s += str(i)\n# print(s)\n\n# n, b = int(input()), []\n# c = []\n# for i in range(n):\n# if len(b) == n:\n# break\n# c = [i]*i\n# j = 0\n# b.append([i for j in range(i) if len(b) < n])\n# # while len(b) < n and len(c) - 1 >= j:\n# # b.append(c[j])\n# # j += 1\n# print(*b)\n\n# n = int(input())\n# a = []\n# i = 0\n# while len(a) < n:\n# # a += [i] * i\n# a.append([i] * i)\n# i += 1\n# print(*a[:n])\n\na = [1, 2, 3]\nb = [56, 67,78]\na +=b\na += [1, 2, 4]\na.append([1, 47, 34])\ns = str(a)\nprint(s)\n\n# n = int(input())\n# a = [for j in range(1) for i in range(n)]\n# a =\n\n# n = int(input())\n# i = 0\n# k = 0\n# j = 0\n# while True:\n# i += 1\n# for j in range(i):\n# k += 1\n# if k < n:\n# print(i, end=' ')\n# elif k == n:\n# print(i)"},{"alpha_fraction":0.592783510684967,"alphanum_fraction":0.592783510684967,"avg_line_length":15.891304016113281,"blob_id":"9a4584636b1e3f9faf237a84456262cdf7c82976","content_id":"68ab8d010646fd04205ba053d072d8285b2e4cf0","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":776,"license_type":"no_license","max_line_length":59,"num_lines":46,"path":"/Python_stepik_2/3.4.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import requests\nimport re\n\n\ndef url_in(in_wich, look_for):\n\tres = requests.get(in_wich)\n\t_lst = re.findall(r\"]*?href=\"(.*?)\"[^>]*?>')\n#\n# resp = requests.get(start_url).text\n# for url in link_pattern.findall(resp):\n# \tcur_resp = requests.get(url).text\n# \tif end_url in link_pattern.findall(cur_resp):\n# \t\tfound = True\n# \t\tbreak\n#\n# print(\"Yes\" if found else \"No\")"},{"alpha_fraction":0.6280944347381592,"alphanum_fraction":0.6511226296424866,"avg_line_length":24.130434036254883,"blob_id":"66a40d1235f2a4d661ae3a36eebe8d751ef77cd5","content_id":"75d464579f16a4c4a78f1a0e8f6d8a807227957f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2049,"license_type":"no_license","max_line_length":83,"num_lines":69,"path":"/Python_stepik_2/1.4.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"namespaces = {'global': None} # словарь пространств - пространство : родитель\nvars = {'global': set()} # словарь переменных - пространство, множество переменных\n\n\ndef create(par1, par2):\n\t# create – создать новое пространство\n\t# имен с именем внутри пространства \n\tnamespaces[par1] = par2\n\n\ndef add(par1, par2):\n\t# add – добавить в пространство переменную \n\tif par1 in vars:\n\t\tvars[par1].add(par2)\n\telse:\n\t\tvars[par1] = set()\n\t\tvars[par1].add(par2)\n\n\ndef get(par1, par2):\n\t# get – получить имя пространства,\n\t# из которого будет взята переменная при запросе и\n\t# з пространства , или None, если такого пространства не существует\n\tglobal namespaces\n\tglobal vars\n\tif par1 is None:\n\t\treturn None\n\telif par1 in vars.keys():\n\t\tif par2 in vars[par1]:\n\t\t\treturn par1\n\t\telse:\n\t\t\tpar1 = namespaces[par1]\n\t\t\treturn get(par1, par2)\n\telse:\n\t\tpar1 = namespaces[par1]\n\t\treturn get(par1, par2)\n\n\ns = ''\nn = int(input())\nfor i in range(n):\n\tcmd, _par1, _par2 = input().split()\n\tif cmd == 'create':\n\t\tcreate(_par1, _par2)\n\telif cmd == 'add':\n\t\tadd(_par1, _par2)\n\telse:\n\t\ts += str(get(_par1, _par2)) + '\\n'\nprint(s)\n\n\n# # Интересный вызов функции\n# info = dict({'global':[None]})\n#\n# def create(namespace, parent):\n# info.update({namespace:[parent]})\n#\n# def add(namespace, var):\n# info[namespace].append(var)\n#\n# def get(namespace, var):\n# while namespace != None and var not in info[namespace][1:]:\n# namespace = info[namespace][0]\n# print(namespace)\n#\n# operations = {'create': create, 'add': add, 'get': get}\n# for i in range(int(input())):\n# inp = input().split()\n# operations[inp[0]](inp[1], inp[2])\n\n\n\n"},{"alpha_fraction":0.6104294657707214,"alphanum_fraction":0.6134969592094421,"avg_line_length":20.799999237060547,"blob_id":"4d80c2eba527a1a603c817d147ed90daca136014","content_id":"fed059ab3b985423e22b183f04e6980dc5abf0ce","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":326,"license_type":"no_license","max_line_length":90,"num_lines":15,"path":"/Python_stepik_2/2.4.2.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import os\n\nr = []\nfor current_dir, dirs, files in os.walk('main'):\n\tif '.py' in [files[i][-3:] for i in range(len(files))]:\n\t\tr.append(current_dir)\nr.sort()\n[print(i) for i in r]\n\n\n# import os\n#\n# result = [cur_dir for cur_dir, dirs, files in os.walk(\"main\") if any([fl.endswith(\".py\")\n# for fl in files])]\n# print(result)"},{"alpha_fraction":0.6066176295280457,"alphanum_fraction":0.6176470518112183,"avg_line_length":18.5,"blob_id":"fbd2be53c828636b226023be726994de4963b171","content_id":"e10fbc5f3e4e967634f83cb3a754f71e4a341627","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":291,"license_type":"no_license","max_line_length":49,"num_lines":14,"path":"/Python_stepik_2/1.2.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"ans = 0\nn = len(objects)\nfor i in range(n): # доступная переменная objects\n\tf = True\n\tfor j in range(i + 1, n):\n\t\tif objects[i] is objects[j]:\n\t\t\tf = False\n\t\t\tbreak\n\tif f:\n\t\tans += 1\nprint(ans)\n\n# print(len(set(map(id, objects))))\n# print(len(set(id(i) for i in objects)))"},{"alpha_fraction":0.48148149251937866,"alphanum_fraction":0.5308641791343689,"avg_line_length":13.235294342041016,"blob_id":"b744ad8b51d816c7f194fc892ceb80a38c532fc1","content_id":"1a5c39abd529011c795f3fbd8295d0f70ecfc170","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":243,"license_type":"no_license","max_line_length":22,"num_lines":17,"path":"/Python_stepik_2/3.2.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"s = input()\na = input()\nb = input()\ns1 = ''\nif a == b and a in s:\n\tprint('Impossible')\nelse:\n\tfor i in range(1001):\n\t\ts1 = s.replace(a, b)\n\t\tif s == s1:\n\t\t\tprint(i)\n\t\t\tbreak\n\t\telif i == 1000:\n\t\t\tprint('Impossible')\n\t\t\tbreak\n\t\telse:\n\t\t\ts = s1\n\n"},{"alpha_fraction":0.47428572177886963,"alphanum_fraction":0.508571445941925,"avg_line_length":16.600000381469727,"blob_id":"30c02e851ef2d658967c9a1706fae0aa10ee758f","content_id":"facff2a774f1656fb875e1562134027243139119","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":175,"license_type":"no_license","max_line_length":46,"num_lines":10,"path":"/Python_stepik_2/1.3.2.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"n, k = map(int, input().split())\ndef c_rec(n, k):\n\tif k == 0:\n\t\treturn 1\n\telif k > n:\n\t\treturn 0\n\telse:\n\t\treturn c_rec(n - 1, k) + c_rec(n - 1, k - 1)\ny = c_rec(n, k)\nprint(y)"},{"alpha_fraction":0.6692307591438293,"alphanum_fraction":0.6769230961799622,"avg_line_length":20.83333396911621,"blob_id":"3bd03f4549980d40ba1b6501dfc517f664f6b515","content_id":"33e79c16eb95b553d6f58a52ed6ff41a8826afce","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":130,"license_type":"no_license","max_line_length":63,"num_lines":6,"path":"/Python_stepik_2/3.3.7.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import re\nimport sys\n\nfor line in sys.stdin:\n\tline = line.rstrip()\n\tprint(re.sub(r\"\\ba+\\b\", \"argh\", line, 1, flags=re.IGNORECASE))"}],"string":"[\n {\n \"alpha_fraction\": 0.7599999904632568,\n \"alphanum_fraction\": 0.7599999904632568,\n \"avg_line_length\": 15.777777671813965,\n \"blob_id\": \"4b3878dbf7e6f7ca78759c26d05edfe5ea5fb12d\",\n \"content_id\": \"674b63226989a3e5758f6090c591cd1405228dfe\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 159,\n \"license_type\": \"no_license\",\n \"max_line_length\": 36,\n \"num_lines\": 9,\n \"path\": \"/Django_projects/vk_oauth/vk_oauth/apps/articles/views.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from django.http import HttpResponse\\n\\n\\ndef index(request):\\n\\treturn HttpResponse(\\\"Привет, мир!\\\")\\n\\n\\ndef test(request):\\n\\treturn HttpResponse(\\\"Test page\\\")\"\n },\n {\n \"alpha_fraction\": 0.6798780560493469,\n \"alphanum_fraction\": 0.6829268336296082,\n \"avg_line_length\": 22.5,\n \"blob_id\": \"855fff72efeea51d23aca1e346a5a3815e793dba\",\n \"content_id\": \"eec7a8f21692a18311c71ffc0b3102879c53fa3d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 418,\n \"license_type\": \"no_license\",\n \"max_line_length\": 50,\n \"num_lines\": 14,\n \"path\": \"/Python_stepik_2/1.5.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"class MoneyBox:\\n\\n\\tdef __init__(self, capacity):\\n\\t\\t# конструктор с аргументом – вместимость копилки\\n\\t\\tself.capacity = capacity\\n\\t\\tself.count = 0\\n\\n\\tdef can_add(self, v):\\n\\t\\t# True, если можно добавить v монет, False иначе\\n\\t\\t\\treturn self.count + v <= self.capacity\\n\\n\\tdef add(self, v):\\n\\t\\t# положить v монет в копилку\\n\\t\\tself.count += v\"\n },\n {\n \"alpha_fraction\": 0.4110429584980011,\n \"alphanum_fraction\": 0.4519427418708801,\n \"avg_line_length\": 24.789474487304688,\n \"blob_id\": \"b64f815a42415b83cbb2d9cb49d13c89ba9e52f9\",\n \"content_id\": \"6216bec9c6c429d744dd7dee7fc93470e3e7d527\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 543,\n \"license_type\": \"no_license\",\n \"max_line_length\": 73,\n \"num_lines\": 19,\n \"path\": \"/MyPr1/ex3.7.4.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"a = ('\\\\n'.join(input().lower() for i in range(int(input())))).split('\\\\n')\\nx = 0\\ny = 0\\nfor s in a:\\n\\tif s[0:2] == 'юг':\\n\\t\\ty -= int(s[3:])\\n\\telif s[0:5] == 'запад':\\n\\t\\tx -= int(s[6:])\\n\\telif s[0:6] == 'восток':\\n\\t\\tx += int(s[7:])\\n\\telif s[0:5] == 'север':\\n\\t\\ty += int(s[6:])\\nprint(str(x) + ' ' + str(y))\\n\\n# s = {'север': 0, 'запад': 0, 'юг': 0, 'восток': 0}\\n# for i in range(int(input())):\\n# k = input().split()\\n# s[k[0]] += int(k[1])\\n# print(s['восток'] - s['запад'], s['север'] - s['юг'])\"\n },\n {\n \"alpha_fraction\": 0.6132075190544128,\n \"alphanum_fraction\": 0.6320754885673523,\n \"avg_line_length\": 16.83333396911621,\n \"blob_id\": \"aafb89885ca199c9b477a8768556461600af3b65\",\n \"content_id\": \"69ef3b2aa248b2fbc2f70f0b1236e878d72792df\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 106,\n \"license_type\": \"no_license\",\n \"max_line_length\": 39,\n \"num_lines\": 6,\n \"path\": \"/Python_stepik_2/3.3.9.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import re\\nimport sys\\n\\nfor line in sys.stdin:\\n\\tline = line.rstrip()\\n\\tprint(re.sub(r\\\"(\\\\w)\\\\1+\\\", r'\\\\1', line))\"\n },\n {\n \"alpha_fraction\": 0.540145993232727,\n \"alphanum_fraction\": 0.5839415788650513,\n \"avg_line_length\": 21.83333396911621,\n \"blob_id\": \"8a18f08225cde2f00b7e8b3282e28a98358f6075\",\n \"content_id\": \"ab9914576b0697636d9645953a329b02c8f00d76\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 137,\n \"license_type\": \"no_license\",\n \"max_line_length\": 62,\n \"num_lines\": 6,\n \"path\": \"/Python_stepik_2/test.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"a = set([1, 2, 3])\\nb = set([1, 3, 2])\\n\\nprint(a == b, type(a), sorted(b))\\n\\ns = \\\"abba com mother bill mother com abba dog abba mother com\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.5813953280448914,\n \"alphanum_fraction\": 0.6511628031730652,\n \"avg_line_length\": 31.125,\n \"blob_id\": \"4d30309fd74e9fd1f215536dba5bf6f235cae44f\",\n \"content_id\": \"e2fdc7c55b697444b689fad3df860ade54260594\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 258,\n \"license_type\": \"no_license\",\n \"max_line_length\": 82,\n \"num_lines\": 8,\n \"path\": \"/MyPr1/ex17.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import requests\\n\\nr = requests.get('https://stepic.org/media/attachments/course67/3.6.3/699991.txt')\\nb = r.text.strip()\\nwhile b[0:2] !='We':\\n r = requests.get('https://stepic.org/media/attachments/course67/3.6.3/' + b)\\n b = r.text.strip()\\n print(b)\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5072045922279358,\n \"alphanum_fraction\": 0.5072045922279358,\n \"avg_line_length\": 19.47058868408203,\n \"blob_id\": \"8cce65453e5ec2ce4f038b0a6cf68e7864f72aa4\",\n \"content_id\": \"c8562ace5d78ea7961d8806aed73796c3142b889\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 347,\n \"license_type\": \"no_license\",\n \"max_line_length\": 47,\n \"num_lines\": 17,\n \"path\": \"/MyPr1/ex19.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"a, b, c, d = input(), input(), input(), input()\\ng = {}\\nfor i in range(len(a)):\\n g[a[i]] = b[i]\\n\\ndef cypher(a, d):\\n b = ''\\n for i in range(len(a)):\\n b += d[a[i]]\\n return b\\ndef uncode(d):\\n uncode = {}\\n for key, value in d.items():\\n uncode[value] = key\\n return uncode\\nprint(cypher(c, g))\\nprint(cypher(d, uncode(g)))\"\n },\n {\n \"alpha_fraction\": 0.41081079840660095,\n \"alphanum_fraction\": 0.4216216206550598,\n \"avg_line_length\": 22.125,\n \"blob_id\": \"e35a367eab4f8d809beb1114231571f9ad791909\",\n \"content_id\": \"64157a156220be3636d99a1d2aeb86b1688f872d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 185,\n \"license_type\": \"no_license\",\n \"max_line_length\": 50,\n \"num_lines\": 8,\n \"path\": \"/MyPr1/ex10.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\nlst = [i.lower() for i in input().split()]\\ns = {}\\nfor i in lst:\\n if s.get(i) == None:\\n s[i] = 1\\n else:\\n s[i] += 1\\nprint(*[i + ' ' + str(s[i]) for i in s], sep='\\\\n')\"\n },\n {\n \"alpha_fraction\": 0.6003372669219971,\n \"alphanum_fraction\": 0.6003372669219971,\n \"avg_line_length\": 22.760000228881836,\n \"blob_id\": \"8f4afefea8a30bf4fe64ec43b19ba0d04d8d5c29\",\n \"content_id\": \"610afc42a15a1c74842f2e3f794d2dbaacb8aa13\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 624,\n \"license_type\": \"no_license\",\n \"max_line_length\": 102,\n \"num_lines\": 25,\n \"path\": \"/Python_stepik_2/3.5.2.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import json\\n\\n\\ndef count_descendant(Vertex): # для класса Vertex возвращает всех потомков []\\n\\tch = [Vertex]\\n\\tfor i in js:\\n\\t\\tif Vertex in i[\\\"parents\\\"]:\\n\\t\\t\\tch.append(i[\\\"name\\\"])\\n\\t\\t\\tch.extend(count_descendant(i[\\\"name\\\"]))\\n\\treturn ch\\n\\n\\njs = json.loads(input())\\nprint(*sorted([i[\\\"name\\\"] + ' : ' + str(len(set(count_descendant(i[\\\"name\\\"])))) for i in js]), sep='\\\\n')\\n\\n\\n\\n# import json\\n#\\n# cls = {c['name']: c['parents'] for c in json.loads(input())}\\n#\\n# isbase = lambda b, d: b == d or any(isbase(b, c) for c in cls[d])\\n#\\n# for p in sorted(cls):\\n# print(p, ':', len({c for c in cls if isbase(p, c)}))\"\n },\n {\n \"alpha_fraction\": 0.5525835752487183,\n \"alphanum_fraction\": 0.5610942244529724,\n \"avg_line_length\": 21.243244171142578,\n \"blob_id\": \"c01f5e9efb826bd08cadeae8fb6293a4299f3928\",\n \"content_id\": \"9f939bcde84169a0a33803f3395e4e3c06dd4c29\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1737,\n \"license_type\": \"no_license\",\n \"max_line_length\": 83,\n \"num_lines\": 74,\n \"path\": \"/Python_stepik_2/2.1.2.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"n = int(input())\\nVertex = {} # вершина : множество прямых предков\\n\\n\\ndef get_direct_ancestor(a, vertex): # возвращает множество всех предков\\n\\n\\tstack = []\\n\\tfor i in vertex[a]:\\n\\t\\tstack.append(i)\\n\\t\\tif i:\\n\\t\\t\\tstack.extend(get_direct_ancestor(i, vertex))\\n\\treturn set(stack)\\n\\n\\nfor i in range(n):\\n\\tVertexBuff = input().split()\\n\\tVertex[VertexBuff[0]] = set()\\n\\tif len(VertexBuff) != 1:\\n\\t\\t[Vertex[VertexBuff[0]].add(i) for i in VertexBuff[2:]]\\nq = int(input())\\nex = [input() for i in range(q)]\\nre = []\\nfor i in range(len(ex) - 1, 0, -1):\\n\\tif ex[i] not in Vertex:\\n\\t\\tcontinue\\n\\tanc = get_direct_ancestor(ex[i], Vertex)\\n\\tfor j in range(i):\\n\\t\\tif ex[i] in re:\\n\\t\\t\\tcontinue\\n\\t\\tif ex[j] == ex[i] or ex[j] in anc:\\n\\t\\t\\tre.append(ex[i])\\nfor i in range(len(re) - 1, -1, -1):\\n\\tprint(re[i])\\n\\n# \\tчто это за генератор?\\n# def checkdup(d):\\n# return cls[d] is None or any(map(checkdup, cls[d]))\\n#\\n# cls = {d: set(b[1:]) for _ in range(int(input())) for d, *b in [input().split()]}\\n#\\n# for _ in range(int(input())):\\n# c = input()\\n# if checkdup(c):\\n# print(c)\\n# cls[c] = None\\n\\n\\n# # код преподавателя\\n# n = int(input())\\n# classes = {}\\n# for i in range(n):\\n# line = input()\\n# parts = line.split(\\\" : \\\")\\n# cls = parts[0]\\n# if len(parts) == 1:\\n# classes[cls] = []\\n# else:\\n# classes[cls] = parts[1].split(\\\" \\\")\\n#\\n#\\n# def check(src, dest):\\n# if src == dest:\\n# return True\\n# return any([check(child, dest) for child in classes[src]])\\n#\\n#\\n# m = int(input())\\n# used = []\\n#\\n# for i in range(m):\\n# cls = input()\\n# if any([check(cls, used_one) for used_one in used]):\\n# print(cls)\\n# used.append(cls)\"\n },\n {\n \"alpha_fraction\": 0.3731343150138855,\n \"alphanum_fraction\": 0.45895522832870483,\n \"avg_line_length\": 32.54166793823242,\n \"blob_id\": \"8d45181b2b54373826f4c7fad51fd1cc2c095e86\",\n \"content_id\": \"c60dad311c2946929cf166fe94f10669a2d78117\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 814,\n \"license_type\": \"no_license\",\n \"max_line_length\": 161,\n \"num_lines\": 24,\n \"path\": \"/MyPr1/ex3.7.5.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"with open(r'C:\\\\Users\\\\arkka\\\\PycharmProjects\\\\MyPr1\\\\dataset_3380_5.txt', 'r') as file1:\\n\\td = {'1' : [0, 0], '2' : [0, 0], '3' : [0, 0], '4' : [0, 0], '5' : [0, 0], '6' : [0, 0], '7' : [0, 0], '8' : [0, 0], '9' : [0, 0], '10' : [0, 0], '11' : [0, 0]}\\n\\ta = file1.read().strip().split('\\\\n')\\n\\ta = [i.split() for i in a]\\nfor ai in a:\\n\\td[ai[0]][0] += int(ai[2])\\n\\td[ai[0]][1] += 1\\nfor key in d:\\n\\tif d[key] == [0, 0]:\\n\\t\\td[key] = '-'\\n\\telse:\\n\\t\\td[key] = str(d[key][0] / d[key][1])\\nprint(*[i + ' ' + d[i] for i in d], sep='\\\\n')\\n\\n# d = {i: [] for i in range(1,12)}\\n# with open(r'D:\\\\Новая папка\\\\dataset_3380_5.txt','r', encoding='utf-8') as f1:\\n# for i in f1:\\n# d[int(i.split()[0])].append(float(i.split()[2]))\\n#\\n# for i in range(1,12):\\n# if d[i]:\\n# print(i, sum(d[i])/len(d[i]))\\n# else:\\n# print(i, '-')\"\n },\n {\n \"alpha_fraction\": 0.6646825671195984,\n \"alphanum_fraction\": 0.6726190447807312,\n \"avg_line_length\": 23.047618865966797,\n \"blob_id\": \"580ec1d93ba093f7874df3dd57cfba997dee2469\",\n \"content_id\": \"e0c2899ee67f19c0085a2c0f2db0c77da087b1f0\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 506,\n \"license_type\": \"no_license\",\n \"max_line_length\": 49,\n \"num_lines\": 21,\n \"path\": \"/Python_stepik_2/2.1.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"try:\\n\\tfoo()\\nexcept (AssertionError, ZeroDivisionError) as e:\\n\\tprint(type(e).__name__)\\nexcept\\tArithmeticError as e:\\n\\tprint('ArithmeticError')\\n\\n\\n\\n\\n# except Exception as e:\\n\\t# print(str(type(e))[8:-2])\\n\\t# print(type(e).__name__)\\n# \\tif 'ZeroDivisionError' in str(type(e)):\\n# \\t\\tprint(\\\"ZeroDivisionError\\\")\\n# \\telif 'ArithmeticError' in str(type(e)):\\n# \\t\\tprint(\\\"ArithmeticError\\\")\\n# except AssertionError as e:\\n# \\tprint(type(e).__name__)\\n# print(type(AssertionError()))\\n# print(str(type(AssertionError()))[8:-2])\"\n },\n {\n \"alpha_fraction\": 0.6084848642349243,\n \"alphanum_fraction\": 0.6278787851333618,\n \"avg_line_length\": 29.592592239379883,\n \"blob_id\": \"996d0e463cd70fd5ec339de74ff800e8897c13b3\",\n \"content_id\": \"6206dc923a6a316974b4d2dd061dc33130338ddc\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1044,\n \"license_type\": \"no_license\",\n \"max_line_length\": 106,\n \"num_lines\": 27,\n \"path\": \"/Interview_Tasks/habr_1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Есть файл, в котором содержаться слова разделённые пробелом.\\n# Например: \\\"abba com mother bill mother com abba dog abba mother com\\\".\\n# Нужно найти и вывести тройку слов, которые чаще всего встречаются вместе (порядок не имеет значения).\\n# То есть в моём примере тройки слов это \\\"abba com mother\\\", \\\"com mother bill\\\", \\\"mother bill mother\\\" и т.д.\\n# Тут правильным ответом должно быть \\\"abba com mother\\\" (частота — 3 раза).\\n# https://habr.com/en/post/439576/\\n\\ns = \\\"abba com mother bill mother com abba dog abba mother com\\\".split()\\na = []\\nd = dict()\\nmax_i = 0\\nmax_s = ''\\nfor i in range(1,len(s)-1):\\n\\tb = sorted([s[i - 1], s[i], s[i + 1]])\\n\\ta.append(b)\\n\\tc = str(b)\\n\\tif b in a:\\n\\t\\tif c in d:\\n\\t\\t\\td[c] += 1\\n\\t\\telse:\\n\\t\\t\\td[c] = 1\\n\\t\\tif d[c] > max_i:\\n\\t\\t\\tmax_i = d[c]\\n\\t\\t\\tmax_s = s[i - 1] + ' ' + s[i] + ' ' + s[i + 1]\\n\\nprint(d)\\nprint(max_s)\"\n },\n {\n \"alpha_fraction\": 0.7368420958518982,\n \"alphanum_fraction\": 0.7368420958518982,\n \"avg_line_length\": 24.66666603088379,\n \"blob_id\": \"38aa4010566c6fbd377706c19ad8f48309279acd\",\n \"content_id\": \"9aef216ae3c16cf94ddfb72d5ffd4ef1de97cd77\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 76,\n \"license_type\": \"no_license\",\n \"max_line_length\": 32,\n \"num_lines\": 3,\n \"path\": \"/MyPr1/codewars1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"def number_to_string(num):\\n\\treturn str(num)\\nprint(number_to_string(input()))\"\n },\n {\n \"alpha_fraction\": 0.45652174949645996,\n \"alphanum_fraction\": 0.489130437374115,\n \"avg_line_length\": 22.08333396911621,\n \"blob_id\": \"60d9e47cd68d7b50be0b8e461cad6dcb72a286e3\",\n \"content_id\": \"7029cc47bb7d02a0a41ce363c0427fb5c75c8bc8\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 276,\n \"license_type\": \"no_license\",\n \"max_line_length\": 58,\n \"num_lines\": 12,\n \"path\": \"/MyPr1/example1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"numbers = [int(i) for i in input().split()]\\nsums = ''\\nl = len(numbers)\\nif l == 1:\\n sums = str(numbers[0])\\nelse:\\n j = 0\\n while j < l - 1:\\n sums += str(numbers[j - 1] + numbers[j + 1]) + ' '\\n j += 1\\n sums += str(numbers[l - 2] + numbers[0])\\nprint(sums)\"\n },\n {\n \"alpha_fraction\": 0.6374622583389282,\n \"alphanum_fraction\": 0.6374622583389282,\n \"avg_line_length\": 22.5,\n \"blob_id\": \"c82d11a570760586cd5505d197a92c2dd9d60bea\",\n \"content_id\": \"1795de39c5aecd22d5a29f8390c815f72f8bbd1f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 331,\n \"license_type\": \"no_license\",\n \"max_line_length\": 75,\n \"num_lines\": 14,\n \"path\": \"/Python_stepik_2/2.2.2.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import simplecrypt\\n\\nwith open(\\\"encrypted.bin\\\", \\\"rb\\\") as inp, open(\\\"passwords.txt\\\", \\\"r\\\") as pas:\\n\\tencrypted = inp.read()\\n\\t# a = pas.read().strip().split('\\\\n')\\n\\tfor i in pas:\\n\\t\\ta = i.strip()\\n\\t\\ttry:\\n\\t\\t\\ts = simplecrypt.decrypt(a, encrypted)\\n\\t\\texcept simplecrypt.DecryptionException:\\n\\t\\t\\tprint('Error')\\n\\t\\telse:\\n\\t\\t\\tprint(a)\\n\\t\\t\\tprint(s)\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.48759007453918457,\n \"alphanum_fraction\": 0.5572457909584045,\n \"avg_line_length\": 20.55172348022461,\n \"blob_id\": \"83e9480b9a6f6ceb6efa7fb275f7576a5cca2b56\",\n \"content_id\": \"9bdb3361b56da7a89e8de076809112b3b57c1a21\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1411,\n \"license_type\": \"no_license\",\n \"max_line_length\": 64,\n \"num_lines\": 58,\n \"path\": \"/Python_stepik_2/1.3.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"def closest_mod_5(x):\\n\\treturn (x if x % 5 == 0 else x + (5 - x % 5))\\nprint(closest_mod_5(int(input())))\\n\\n# closest_mod_5 = lambda x: (x + 4) // 5 * 5\\n\\n# a=1\\n# b=2\\n# lst = [5,6,7,8]\\n# dct = {\\\"1\\\":11,\\\"2\\\":22}\\n# print(a, b, lst, dct)\\n#\\n# def pr(aa, bb, cc, *llst, **ddct):\\n# print(aa) #=a\\n# print(bb) #=b\\n# print(cc) #возьмётся из списка\\n# print(llst) #остатки списка в виде кортежа\\n# print(lst) #глобальная переменная\\n# for k in ddct:\\n# print(k, ddct[k])\\n# print(ddct) #словарь в к-й попали dct, r=8,t=6\\n# print(*llst) #элементы кортежа\\n# print(*ddct) #ключи словаря\\n# print(*dct) #глобальная переменная\\n# pr(a, b, *lst, **dct, r=8,t=6)\\n# 1 2 [5, 6, 7, 8] {'1': 11, '2': 22}\\n# 1\\n# 2\\n# 5\\n# (6, 7, 8)\\n# [5, 6, 7, 8]\\n# 1 11\\n# 2 22\\n# r 8\\n# t 6\\n# {'1': 11, '2': 22, 'r': 8, 't': 6}\\n# 6 7 8\\n# 1 2 r t\\n# 1 2\\n\\nКод проверки времени выполнения функций\\nimport time\\ndef one(a, b, c):\\n\\treturn a + b + c\\n\\n\\ndef two(*args):\\n\\treturn args[0] + args[1] + args[2]\\ncount = 9000000\\n_startTime = time.time()\\nfor num in range(count):\\n\\tone(1, 2, 3)\\nprint(\\\"Positional test:\\\", time.time() - _startTime)\\n\\n_startTime = time.time()\\nfor num in range(count):\\n\\ttwo(1, 2, 3)\\nprint(\\\"Positional args as list test:\\\", time.time() - _startTime)\"\n },\n {\n \"alpha_fraction\": 0.4267834722995758,\n \"alphanum_fraction\": 0.46057572960853577,\n \"avg_line_length\": 24,\n \"blob_id\": \"bd284837e79d5efee2efbfea4ab10d0b29403308\",\n \"content_id\": \"1d245b720e7145422cada3b137b9ab314de46526\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 799,\n \"license_type\": \"no_license\",\n \"max_line_length\": 84,\n \"num_lines\": 32,\n \"path\": \"/MyPr1/ex13.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# s = ''\\n# a = []\\n# line = ''\\n# with open(r'C:\\\\Users\\\\arkka\\\\PycharmProjects\\\\MyPr1\\\\dataset_3363_3.txt', 'r') as inf:\\n# for line in inf:\\n# line += line.strip().lower() + ' '\\n# for ltr in line:\\n# if ltr != ' ':\\n# s += ltr\\n# else:\\n# a.append(s)\\n# s = ''\\n# s = {}\\n# max = 0\\n# max_w = ''\\n# for i in a:\\n# if s.get(i) == None:\\n# s[i] = 1\\n# else:\\n# s[i] += 1\\n# if max < s[i]:\\n# max = s[i]\\n# max_w = i\\n#\\n# print(max_w, s[max_w], sep=' ')\\n\\nwith open(r'C:\\\\Users\\\\arkka\\\\PycharmProjects\\\\MyPr1\\\\dataset_3363_3.txt', 'w') as inf:\\n inf.write('12\\\\n')\\n inf.write('1234\\\\n')\\nwith open(r'C:\\\\Users\\\\arkka\\\\PycharmProjects\\\\MyPr1\\\\dataset_3363_3.txt', 'r') as inf:\\n line = inf.read()\\nprint(line)\"\n },\n {\n \"alpha_fraction\": 0.6388888955116272,\n \"alphanum_fraction\": 0.6419752836227417,\n \"avg_line_length\": 17.05555534362793,\n \"blob_id\": \"645c605853ca2260fb5b6d03c272c9edd2db91d1\",\n \"content_id\": \"387366f1260bc2528fc049675effd4821597f6c4\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 324,\n \"license_type\": \"no_license\",\n \"max_line_length\": 41,\n \"num_lines\": 18,\n \"path\": \"/Python_stepik_2/3.3.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import re\\nimport sys\\n\\npattern = r\\\"cat\\\"\\nfor line in sys.stdin:\\n\\tline = line.rstrip()\\n\\tmatch_object = re.findall(pattern, line)\\n\\tif len(match_object) >= 2:\\n\\t\\tprint(line)\\n\\n#catcatsghdtscat\\n# import re\\n# import sys\\n#\\n# for line in sys.stdin:\\n# line = line.strip()\\n# if re.search(r\\\"cat.*cat\\\", line):\\n# print(line)\"\n },\n {\n \"alpha_fraction\": 0.40112993121147156,\n \"alphanum_fraction\": 0.4067796468734741,\n \"avg_line_length\": 13.75,\n \"blob_id\": \"fde79baf9072f84fb3ea4642d06117bd565e0577\",\n \"content_id\": \"b7630c4a4aa2a060ee60210b01eba1b6e9ff9acb\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 177,\n \"license_type\": \"no_license\",\n \"max_line_length\": 20,\n \"num_lines\": 12,\n \"path\": \"/MyPr1/ex11.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"def f(x):\\n return x + 1\\n\\nn = int(input())\\nb = {}\\nfor i in range(n):\\n a = int(input())\\n if a in b:\\n print(b[a])\\n else:\\n b[a] = f(a)\\n print(b[a])\\n\"\n },\n {\n \"alpha_fraction\": 0.7093023061752319,\n \"alphanum_fraction\": 0.7093023061752319,\n \"avg_line_length\": 18.22222137451172,\n \"blob_id\": \"70031e6e790ca23c3bd652b6d459d0e33adc016e\",\n \"content_id\": \"a92e5966fbcc3aa1fe1075550a55c2a9227150ec\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 172,\n \"license_type\": \"no_license\",\n \"max_line_length\": 40,\n \"num_lines\": 9,\n \"path\": \"/Python_stepik_2/3.3.2.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import re\\nimport sys\\n\\npattern = r\\\"\\\\bcat\\\\b\\\"\\nfor line in sys.stdin:\\n\\tline = line.rstrip()\\n\\tmatch_object = re.search(pattern, line)\\n\\tif match_object is not None:\\n\\t\\tprint(line)\"\n },\n {\n \"alpha_fraction\": 0.6648044586181641,\n \"alphanum_fraction\": 0.6703910827636719,\n \"avg_line_length\": 18.55555534362793,\n \"blob_id\": \"4b94f35ae6290fa404558cf9c042a767b450986b\",\n \"content_id\": \"0999e7eb1f35623347f73c00f6a7f13fc82813df\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 179,\n \"license_type\": \"no_license\",\n \"max_line_length\": 40,\n \"num_lines\": 9,\n \"path\": \"/Python_stepik_2/3.3.5.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import re\\nimport sys\\n\\npattern = r\\\"\\\\b(.+)\\\\1\\\\b\\\"\\nfor line in sys.stdin:\\n\\tline = line.rstrip()\\n\\tmatch_object = re.search(pattern, line)\\n\\tif match_object is not None:\\n\\t\\tprint(line)\\n\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6766917109489441,\n \"alphanum_fraction\": 0.7368420958518982,\n \"avg_line_length\": 32.5,\n \"blob_id\": \"eacd28b3222d72abd6ee9b5a5cadf68259965f66\",\n \"content_id\": \"44768accd9dc247f4d282df7bf7d3c275058889d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 133,\n \"license_type\": \"no_license\",\n \"max_line_length\": 79,\n \"num_lines\": 4,\n \"path\": \"/MyPr1/ex16.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import requests\\nr = requests.get('https://stepic.org/media/attachments/course67/3.6.2/983.txt')\\nb = r.text.splitlines()\\nprint(len(b))\"\n },\n {\n \"alpha_fraction\": 0.5966101884841919,\n \"alphanum_fraction\": 0.6203389763832092,\n \"avg_line_length\": 13.047618865966797,\n \"blob_id\": \"3c264c182393a726f6a223676a35d8b91de4efc9\",\n \"content_id\": \"9f3d578f29667b01761d0bde965d46e8a652cc44\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 295,\n \"license_type\": \"no_license\",\n \"max_line_length\": 61,\n \"num_lines\": 21,\n \"path\": \"/Python_stepik_2/2.3.2.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import math\\nimport itertools\\n\\ndef primes():\\n\\tcount = 2\\n\\n\\twhile True:\\n\\t\\tisprime = True\\n\\n\\t\\tfor x in range(2, int(math.sqrt(count) + 1)):\\n\\t\\t\\tif count % x == 0:\\n\\t\\t\\t\\tisprime = False\\n\\t\\t\\t\\tbreak\\n\\n\\t\\tif isprime:\\n\\t\\t\\tyield count\\n\\n\\t\\tcount += 1\\n\\n\\nprint(list(itertools.takewhile(lambda x: x <= 31, primes())))\\n\"\n },\n {\n \"alpha_fraction\": 0.3937007784843445,\n \"alphanum_fraction\": 0.4724409580230713,\n \"avg_line_length\": 13.222222328186035,\n \"blob_id\": \"82932a0f7775c5193fdb3f8ad6545894af82bb8c\",\n \"content_id\": \"498df92f3b6d3bfe1f7eb7dcc61a5815f93bbe01\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 127,\n \"license_type\": \"no_license\",\n \"max_line_length\": 20,\n \"num_lines\": 9,\n \"path\": \"/MyPr1/ex3.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"sum1 = 0\\nsum2 = 0\\nwhile True:\\n a = int(input())\\n sum1 += a\\n sum2 += a ** 2\\n if sum1 == 0:\\n break\\nprint(sum2)\"\n },\n {\n \"alpha_fraction\": 0.7222222089767456,\n \"alphanum_fraction\": 0.7222222089767456,\n \"avg_line_length\": 25.75,\n \"blob_id\": \"073cdce3d5ffc02e889f98a63ce3d4b5a6430941\",\n \"content_id\": \"f56217f6bd0f4d3391d252bb8323384fb3f6bf24\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 108,\n \"license_type\": \"no_license\",\n \"max_line_length\": 35,\n \"num_lines\": 4,\n \"path\": \"/Python_stepik_2/1.6.3.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"class LoggableList(list, Loggable):\\n\\tdef append(self, object):\\n\\t\\tsuper().append(object)\\n\\t\\tself.log(object)\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6722221970558167,\n \"alphanum_fraction\": 0.6722221970558167,\n \"avg_line_length\": 29.16666603088379,\n \"blob_id\": \"a38ba225262f080ba9dbfd617405fe24c81268ef\",\n \"content_id\": \"4c7fc8b5b23840f34b7397aead5d42590b045e09\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 180,\n \"license_type\": \"no_license\",\n \"max_line_length\": 81,\n \"num_lines\": 6,\n \"path\": \"/Python_stepik_2/2.2.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import datetime\\n\\nyyyy, mm, dd = input().split()\\ndelta = int(input())\\nd = datetime.date(int(yyyy), int(mm), int(dd)) + datetime.timedelta(days = delta)\\nprint(d.year, d.month, d.day)\"\n },\n {\n \"alpha_fraction\": 0.5845070481300354,\n \"alphanum_fraction\": 0.6338028311729431,\n \"avg_line_length\": 16.875,\n \"blob_id\": \"8a085c15551c7f991a014512138e557f5d65d3e3\",\n \"content_id\": \"1671f6a7d877cc5e4af713b7bf04bd9ce837d23c\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 142,\n \"license_type\": \"no_license\",\n \"max_line_length\": 42,\n \"num_lines\": 8,\n \"path\": \"/Python_stepik_2/3.3.10.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import re\\nimport sys\\n\\nfor line in sys.stdin:\\n\\tline = line.rstrip()\\n\\ts = re.fullmatch(r\\\"(1(01*0)*1|0)*\\\", line)\\n\\tif s is not None:\\n\\t\\tprint(s[0])\"\n },\n {\n \"alpha_fraction\": 0.549199104309082,\n \"alphanum_fraction\": 0.5606407523155212,\n \"avg_line_length\": 24.735294342041016,\n \"blob_id\": \"d8fa33e64296e1986a8a021b0cdc7866e6410383\",\n \"content_id\": \"d3d55582c8c2c6f4daba139e217bae296276038b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 874,\n \"license_type\": \"no_license\",\n \"max_line_length\": 112,\n \"num_lines\": 34,\n \"path\": \"/Python_stepik_2/3.4.2.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import re\\nimport requests\\n\\nhtml_file = input()\\n# html_file = 'http://pastebin.com/raw/7543p0ns'\\n\\nlink_pattern = re.compile(r']*?href=(?:\\\\'|\\\\\\\")(?:https://|http://|ftp://|)(.*?)[\\\\\\\"|\\\\'|/|:]')\\nresp = requests.get(html_file).text\\nurl = link_pattern.findall(resp)\\nurl2 = []\\nfor i in url:\\n\\tif not i.startswith(\\\"..\\\"):\\n\\t\\turl2.append(i)\\nurl2 = sorted(list(set(url2)))\\n# url = sorted(url)\\n# url = set(url)\\n# match_object = re.search(pattern, line)\\nprint(*url2, sep='\\\\n')\\n\\n# import re\\n# import requests\\n#\\n# resp = requests.get(input()).text\\n# sites = set()\\n# for site in re.findall(r'= neg)\\n return pos >= neg\\n\\n def judge_any(pos, neg):\\n # допускает элемент, если его допускает хотя бы одна функция (pos >= 1)\\n return pos >= 1\\n\\n def judge_all(pos, neg):\\n # допускает элемент, если его допускают все функции (neg == 0)\\n return neg == 0\\n\\n def __init__(self, iterable, *funcs, judge=judge_any):\\n # iterable - исходная последовательность\\n # funcs - допускающие функции\\n # judge - решающая функция\\n self.i = 0\\n self.judge = judge\\n self.funcs = funcs\\n self.iterable = iterable\\n self.k = len(self.iterable)\\n\\n def __iter__(self):\\n # возвращает итератор по результирующей последовательности\\n return self\\n\\n def __next__(self):\\n while self.i < self.k:\\n pos = 0\\n neg = 0\\n for _ in self.funcs:\\n if _(self.iterable[self.i]):\\n pos += 1\\n else:\\n neg += 1\\n self.i += 1\\n if self.judge(pos, neg):\\n return self.iterable[self.i - 1]\\n raise StopIteration\\n\\n\\n\\n\\n\\n\\n\\ndef mul2(x):\\n return x % 2 == 0\\n\\ndef mul3(x):\\n return x % 3 == 0\\n\\ndef mul5(x):\\n return x % 5 == 0\\n\\na = [i for i in range(31)] # [0, 1, 2, ... , 30]\\nb = multifilter(a, mul2, mul3, mul5)\\nprint(list(b))\\nprint(list(b))\\n# [0, 2, 3, 4, 5, 6, 8, 9, 10, 12, 14, 15, 16, 18, 20, 21, 22, 24, 25, 26, 27, 28, 30]\\n\\n\\n# print(list(multifilter(a, mul2, mul3, mul5, judge=multifilter.judge_half)))\\n# # [0, 6, 10, 12, 15, 18, 20, 24, 30]\\n#\\n# print(list(multifilter(a, mul2, mul3, mul5, judge=multifilter.judge_all)))\\n# # [0, 30]\\n\\n\\n# # с применением генератора в одной строке\\n# class multifilter:\\n# judge_half = lambda fx: sum(fx) >= len(fx) / 2\\n# judge_any = lambda fx: any(fx)\\n# judge_all = lambda fx: all(fx)\\n#\\n# def __init__(self, iterable, *funcs, judge=judge_any):\\n# self.iterable = iterable\\n# self.funcs = funcs\\n# self.judge = judge\\n#\\n# def __iter__(self):\\n# return (x for x in self.iterable if self.judge([f(x) for f in self.funcs]))\\n\\n\\n# собственное решение через yield (вместо __next__)\\n#\\n# логика работы прописана в итераторе объекта __iter__\\n# где пробегаемся по элементам исходной последовательности\\n# просеиваем элемент через функции-фильтры\\n# и если выбранная функция judge дала добро\\n# то yield`им этот элемент\\n#\\n#\\n#\\n\\n# # решение с yield в iter\\n# class multifilter:\\n# def judge_half(pos, neg):\\n# return pos >= neg # допускает элемент, если его допускает хотя бы половина фукнций (pos >= neg)\\n#\\n# def judge_any(pos, neg):\\n# return pos >= 1 # допускает элемент, если его допускает хотя бы одна функция (pos >= 1)\\n#\\n# def judge_all(pos, neg):\\n# return neg == 0 # допускает элемент, если его допускают все функции (neg == 0)\\n#\\n# def __init__(self, iterable, *funcs, judge=judge_any):\\n# self.lst = iterable # iterable - исходная последовательность\\n# self.funcs = funcs # funcs - допускающие функции\\n# self.judge = judge # judge - решающая функция\\n#\\n# def __iter__(self):\\n# for el in self.lst:\\n# pos, neg = 0, 0\\n# for f in self.funcs:\\n# pass\\n# if f(el):\\n# pos += 1\\n# else:\\n# neg += 1\\n# if self.judge(pos, neg):\\n# yield el\"\n },\n {\n \"alpha_fraction\": 0.5600000023841858,\n \"alphanum_fraction\": 0.6200000047683716,\n \"avg_line_length\": 19.200000762939453,\n \"blob_id\": \"5c5a923429492556e6caf3b1e10d72a1675be624\",\n \"content_id\": \"d108970f7d33e8326c36f572dfdc1e067014f74b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 100,\n \"license_type\": \"no_license\",\n \"max_line_length\": 30,\n \"num_lines\": 5,\n \"path\": \"/Python_stepik_2/2.5.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"def mod_checker(x, mod=0):\\n\\treturn lambda y: y % x == mod\\n\\nmod_3 = mod_checker(3, 1)\\nprint(mod_3(4))\"\n },\n {\n \"alpha_fraction\": 0.7330895662307739,\n \"alphanum_fraction\": 0.7586837410926819,\n \"avg_line_length\": 41.07692337036133,\n \"blob_id\": \"9511b7f652c20f4ee353c39976c5733ee7489a89\",\n \"content_id\": \"0031654e8d97abca3d5f52e261eb223329012ae0\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 602,\n \"license_type\": \"no_license\",\n \"max_line_length\": 119,\n \"num_lines\": 13,\n \"path\": \"/Django_projects/vk_oauth/vk_oauth/apps/articles/models.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from django.db import models\\n\\n\\nclass Article(models.Model):\\n\\tarticle_title = models.CharField('название статьи', max_length=200) # VarChar in MySQL - 200-300 simbols?\\n\\tarticle_text = models.TextField('текст статьи')\\n\\tpub_date = models.DateTimeField('дата публикации')\\n\\n\\nclass Comment(models.Model):\\n\\tarticle = models.ForeignKey(Article, on_delete=models.CASCADE) #on_delete - delete all its comments if delete article\\n\\tauthor_name = models.CharField('имя автора', max_length=50)\\n\\tcomment_text = models.CharField('текст комментария', max_length=200)\\n\"\n },\n {\n \"alpha_fraction\": 0.3900621235370636,\n \"alphanum_fraction\": 0.47577640414237976,\n \"avg_line_length\": 30,\n \"blob_id\": \"ed831e5b9684586bedf787e2de269b5b5c9b58da\",\n \"content_id\": \"b0cbc3a72d140dd54b3fe72840316a26eb5813c9\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 805,\n \"license_type\": \"no_license\",\n \"max_line_length\": 84,\n \"num_lines\": 26,\n \"path\": \"/MyPr1/ex12.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# import os\\n# s = ''\\n# t = ''\\n# num = '1'\\n# with open(r'C:\\\\Users\\\\arkka\\\\PycharmProjects\\\\MyPr1\\\\dataset_3363_2.txt', 'r') as inf:\\n# line = inf.readline().strip()\\n# for i in range(len(line)):\\n# if line[i] in ['1', '2', '3', '4', '5', '6', '7', '8', '9', '0']:\\n# num += line[i]\\n# elif t != '':\\n# s += t * int(num)\\n# t = line[i]\\n# num = ''\\n# else:\\n# t = line[i]\\n# num = ''\\n# s += t * int(num)\\n# with open(r'C:\\\\Users\\\\arkka\\\\PycharmProjects\\\\MyPr1\\\\dataset_3363_2.txt', 'w') as inf:\\n# inf.write(s + '\\\\n')\\n\\n\\ns, d = 'D8s11T20s13r16V8O9o11f3V3S5B14T11m12n18i17T18x15d17V1b5l2x13I5t12q7c14', []\\nfor i in s:\\n if not i.isdigit(): d.append(i)\\n else: d[-1] += i\\nprint(*[i[0]*int(i[1:]) for i in d], sep='')\"\n },\n {\n \"alpha_fraction\": 0.6551724076271057,\n \"alphanum_fraction\": 0.6551724076271057,\n \"avg_line_length\": 18.44444465637207,\n \"blob_id\": \"77448c835d5001a0646d2d046c6839203fb0585a\",\n \"content_id\": \"e8b24b6ada1d429440b15ee85d309f6dc9748360\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 174,\n \"license_type\": \"no_license\",\n \"max_line_length\": 63,\n \"num_lines\": 9,\n \"path\": \"/Python_stepik_2/3.3.6.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import re\\nimport sys\\n\\nfor line in sys.stdin:\\n\\tline = line.rstrip()\\n\\tprint(re.sub(\\\"human\\\", \\\"computer\\\", line))\\n\\n\\n# print(re.sub(r'human', 'computer', sys.stdin.read()), end='')\"\n },\n {\n \"alpha_fraction\": 0.6403940916061401,\n \"alphanum_fraction\": 0.6403940916061401,\n \"avg_line_length\": 34.034481048583984,\n \"blob_id\": \"50c84dbf33db6231fde4963ea5d754abfc39e455\",\n \"content_id\": \"562320c846388f19bd4e002df680a8f95305cd71\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1339,\n \"license_type\": \"no_license\",\n \"max_line_length\": 79,\n \"num_lines\": 29,\n \"path\": \"/MyPr1/ex3.7.3.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"a = int(input())\\nb = set()\\n[b.add(*input().lower().split('\\\\n')) for i in range(a)]\\nc = int(input())\\nd = set()\\nfor i in range(c):\\n g = input().lower().split(' ')\\n [d.add(i) for i in g]\\n[print(i) for i in d if i not in b]\\n\\n# # # формируем множество известных слов на основании построчного ввода\\n# # dic = {input().lower() for _ in range(int(input()))}\\n# #\\n# # # заводим пустое множество для приема текста\\n# wrd = set()\\n#\\n# # т.к. текст построчно подается, а также в каждой строке несколько слов,\\n# # то каждую строку превращаем во множество и добавляем в единое множество wrd\\n# for _ in range(int(input())):\\n# wrd |= {i.lower() for i in input().split()}\\n#\\n# # на вывод отправляем результат вычитания словарного множества dic\\n# # из текстового множества wrd; впереди ставим *, чтобы раскрыть поэлементно\\n# print(*(wrd-dic), sep=\\\"\\\\n\\\")\\n\\n\\n# words = set(input().lower() for i in range(int(input())))\\n# text = set(('\\\\n'.join(input().lower() for i in range(int(input())))).split())\\n# print('\\\\n'.join(text - words))\"\n },\n {\n \"alpha_fraction\": 0.625261664390564,\n \"alphanum_fraction\": 0.6329379081726074,\n \"avg_line_length\": 27.117647171020508,\n \"blob_id\": \"0d844e6856effdcc44170acb6c9ea5a4d85e3d2a\",\n \"content_id\": \"63821cb3d99617e244763533ab9e20616cf83c1d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1433,\n \"license_type\": \"no_license\",\n \"max_line_length\": 109,\n \"num_lines\": 51,\n \"path\": \"/Python_stepik_2/3.7.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# from xml.etree import ElementTree\\nfrom xml.etree.ElementTree import XMLParser\\n\\n# s = input()\\ns = '=n-1: i+=1\\n elif i>=j and i+j>n-1: j-=1\\n elif i>j+1 and i+j<=n-1: i-=1\\nfor i in range(n):\\n print(*t[i])\"\n },\n {\n \"alpha_fraction\": 0.5873016119003296,\n \"alphanum_fraction\": 0.6142857074737549,\n \"avg_line_length\": 21.535715103149414,\n \"blob_id\": \"005820b914d674dddd4a9dcfadfea96bcf9bb6e6\",\n \"content_id\": \"01be5c981785e04082250e457187692799543092\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 630,\n \"license_type\": \"no_license\",\n \"max_line_length\": 82,\n \"num_lines\": 28,\n \"path\": \"/Python_stepik_2/3.5.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import csv\\n\\nwith open(\\\"Crimes.csv\\\", 'r') as f:\\n\\treader = csv.reader(f)\\n\\tr = (i[5] for i in reader if '2015' in i[2])\\n\\td = dict()\\n\\tfor word in r:\\n\\t\\tif word in d:\\n\\t\\t\\td[word] += 1\\n\\t\\telse:\\n\\t\\t\\td[word] = 0\\n\\td = {k: v for k, v in sorted(d.items(), key=lambda item: item[1], reverse=True)}\\n\\tprint(d)\\n\\n# import csv\\n# from collections import Counter as c\\n#\\n# with open('Crimes.csv') as f:\\n# \\tdata = csv.reader(f)\\n# \\tprint(c( row[5] for row in data if '2015' in row[2] ))\\n\\n\\n# import csv\\n# from collections import Counter as c\\n#\\n# with open('Crimes.csv') as f:\\n# result = c(row[5] for row in csv.reader(f))\\n# print(result.most_common(1))\"\n },\n {\n \"alpha_fraction\": 0.5309917330741882,\n \"alphanum_fraction\": 0.6012396812438965,\n \"avg_line_length\": 47.5,\n \"blob_id\": \"894a1e63f1d5c279152c5761ab3fdea8c9e893c3\",\n \"content_id\": \"f2e7d0583a9a34e69125528191df17a74c76998b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 484,\n \"license_type\": \"no_license\",\n \"max_line_length\": 89,\n \"num_lines\": 10,\n \"path\": \"/Python_stepik_2/2.4.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"with open('dataset_24465_4.txt', 'r') as f, open('test_copy.txt', 'w') as w:\\n\\tr = [line.strip() for line in f]\\n\\tw.write('\\\\n'.join([r[i] for i in range(len(r)-1, -1, -1)]))\\n\\n\\n# with open('dataset_24465_4.txt', 'r') as fr, open('dataset_24465_4_w.txt', 'w') as fw:\\n# fw.writelines(fr.readlines()[::-1])\\n\\n# with open('dataset_24465_4.txt') as inp, open('dataset_24465_4_write.txt', 'w') as out:\\n# [out.write('{}\\\\n'.format(oln)) for oln in reversed([iln.rstrip() for iln in inp])]\"\n },\n {\n \"alpha_fraction\": 0.30249109864234924,\n \"alphanum_fraction\": 0.3327402174472809,\n \"avg_line_length\": 16.30769157409668,\n \"blob_id\": \"041f5e7719f6532d4ba15a616b803453b610a024\",\n \"content_id\": \"1884bb844d77b0de55d9955dc383d2cbaf5f84bd\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1124,\n \"license_type\": \"no_license\",\n \"max_line_length\": 51,\n \"num_lines\": 65,\n \"path\": \"/MyPr1/ex4.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# n = int(input())\\n# i = 0\\n# k = 0\\n# j = 0\\n# s = ''\\n# if n == 0:\\n# print(s)\\n# else:\\n# while k <= n:\\n# i += 1\\n# for j in range(i):\\n# k += 1\\n# if k < n:\\n# # print(i, end=' ')\\n# s += str(i) + ' '\\n# elif k == n:\\n# s += str(i)\\n# print(s)\\n\\n# n, b = int(input()), []\\n# c = []\\n# for i in range(n):\\n# if len(b) == n:\\n# break\\n# c = [i]*i\\n# j = 0\\n# b.append([i for j in range(i) if len(b) < n])\\n# # while len(b) < n and len(c) - 1 >= j:\\n# # b.append(c[j])\\n# # j += 1\\n# print(*b)\\n\\n# n = int(input())\\n# a = []\\n# i = 0\\n# while len(a) < n:\\n# # a += [i] * i\\n# a.append([i] * i)\\n# i += 1\\n# print(*a[:n])\\n\\na = [1, 2, 3]\\nb = [56, 67,78]\\na +=b\\na += [1, 2, 4]\\na.append([1, 47, 34])\\ns = str(a)\\nprint(s)\\n\\n# n = int(input())\\n# a = [for j in range(1) for i in range(n)]\\n# a =\\n\\n# n = int(input())\\n# i = 0\\n# k = 0\\n# j = 0\\n# while True:\\n# i += 1\\n# for j in range(i):\\n# k += 1\\n# if k < n:\\n# print(i, end=' ')\\n# elif k == n:\\n# print(i)\"\n },\n {\n \"alpha_fraction\": 0.592783510684967,\n \"alphanum_fraction\": 0.592783510684967,\n \"avg_line_length\": 15.891304016113281,\n \"blob_id\": \"9a4584636b1e3f9faf237a84456262cdf7c82976\",\n \"content_id\": \"68ab8d010646fd04205ba053d072d8285b2e4cf0\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 776,\n \"license_type\": \"no_license\",\n \"max_line_length\": 59,\n \"num_lines\": 46,\n \"path\": \"/Python_stepik_2/3.4.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import requests\\nimport re\\n\\n\\ndef url_in(in_wich, look_for):\\n\\tres = requests.get(in_wich)\\n\\t_lst = re.findall(r\\\"]*?href=\\\"(.*?)\\\"[^>]*?>')\\n#\\n# resp = requests.get(start_url).text\\n# for url in link_pattern.findall(resp):\\n# \\tcur_resp = requests.get(url).text\\n# \\tif end_url in link_pattern.findall(cur_resp):\\n# \\t\\tfound = True\\n# \\t\\tbreak\\n#\\n# print(\\\"Yes\\\" if found else \\\"No\\\")\"\n },\n {\n \"alpha_fraction\": 0.6280944347381592,\n \"alphanum_fraction\": 0.6511226296424866,\n \"avg_line_length\": 24.130434036254883,\n \"blob_id\": \"66a40d1235f2a4d661ae3a36eebe8d751ef77cd5\",\n \"content_id\": \"75d464579f16a4c4a78f1a0e8f6d8a807227957f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2049,\n \"license_type\": \"no_license\",\n \"max_line_length\": 83,\n \"num_lines\": 69,\n \"path\": \"/Python_stepik_2/1.4.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"namespaces = {'global': None} # словарь пространств - пространство : родитель\\nvars = {'global': set()} # словарь переменных - пространство, множество переменных\\n\\n\\ndef create(par1, par2):\\n\\t# create – создать новое пространство\\n\\t# имен с именем внутри пространства \\n\\tnamespaces[par1] = par2\\n\\n\\ndef add(par1, par2):\\n\\t# add – добавить в пространство переменную \\n\\tif par1 in vars:\\n\\t\\tvars[par1].add(par2)\\n\\telse:\\n\\t\\tvars[par1] = set()\\n\\t\\tvars[par1].add(par2)\\n\\n\\ndef get(par1, par2):\\n\\t# get – получить имя пространства,\\n\\t# из которого будет взята переменная при запросе и\\n\\t# з пространства , или None, если такого пространства не существует\\n\\tglobal namespaces\\n\\tglobal vars\\n\\tif par1 is None:\\n\\t\\treturn None\\n\\telif par1 in vars.keys():\\n\\t\\tif par2 in vars[par1]:\\n\\t\\t\\treturn par1\\n\\t\\telse:\\n\\t\\t\\tpar1 = namespaces[par1]\\n\\t\\t\\treturn get(par1, par2)\\n\\telse:\\n\\t\\tpar1 = namespaces[par1]\\n\\t\\treturn get(par1, par2)\\n\\n\\ns = ''\\nn = int(input())\\nfor i in range(n):\\n\\tcmd, _par1, _par2 = input().split()\\n\\tif cmd == 'create':\\n\\t\\tcreate(_par1, _par2)\\n\\telif cmd == 'add':\\n\\t\\tadd(_par1, _par2)\\n\\telse:\\n\\t\\ts += str(get(_par1, _par2)) + '\\\\n'\\nprint(s)\\n\\n\\n# # Интересный вызов функции\\n# info = dict({'global':[None]})\\n#\\n# def create(namespace, parent):\\n# info.update({namespace:[parent]})\\n#\\n# def add(namespace, var):\\n# info[namespace].append(var)\\n#\\n# def get(namespace, var):\\n# while namespace != None and var not in info[namespace][1:]:\\n# namespace = info[namespace][0]\\n# print(namespace)\\n#\\n# operations = {'create': create, 'add': add, 'get': get}\\n# for i in range(int(input())):\\n# inp = input().split()\\n# operations[inp[0]](inp[1], inp[2])\\n\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6104294657707214,\n \"alphanum_fraction\": 0.6134969592094421,\n \"avg_line_length\": 20.799999237060547,\n \"blob_id\": \"4d80c2eba527a1a603c817d147ed90daca136014\",\n \"content_id\": \"fed059ab3b985423e22b183f04e6980dc5abf0ce\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 326,\n \"license_type\": \"no_license\",\n \"max_line_length\": 90,\n \"num_lines\": 15,\n \"path\": \"/Python_stepik_2/2.4.2.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import os\\n\\nr = []\\nfor current_dir, dirs, files in os.walk('main'):\\n\\tif '.py' in [files[i][-3:] for i in range(len(files))]:\\n\\t\\tr.append(current_dir)\\nr.sort()\\n[print(i) for i in r]\\n\\n\\n# import os\\n#\\n# result = [cur_dir for cur_dir, dirs, files in os.walk(\\\"main\\\") if any([fl.endswith(\\\".py\\\")\\n# for fl in files])]\\n# print(result)\"\n },\n {\n \"alpha_fraction\": 0.6066176295280457,\n \"alphanum_fraction\": 0.6176470518112183,\n \"avg_line_length\": 18.5,\n \"blob_id\": \"fbd2be53c828636b226023be726994de4963b171\",\n \"content_id\": \"e10fbc5f3e4e967634f83cb3a754f71e4a341627\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 291,\n \"license_type\": \"no_license\",\n \"max_line_length\": 49,\n \"num_lines\": 14,\n \"path\": \"/Python_stepik_2/1.2.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"ans = 0\\nn = len(objects)\\nfor i in range(n): # доступная переменная objects\\n\\tf = True\\n\\tfor j in range(i + 1, n):\\n\\t\\tif objects[i] is objects[j]:\\n\\t\\t\\tf = False\\n\\t\\t\\tbreak\\n\\tif f:\\n\\t\\tans += 1\\nprint(ans)\\n\\n# print(len(set(map(id, objects))))\\n# print(len(set(id(i) for i in objects)))\"\n },\n {\n \"alpha_fraction\": 0.48148149251937866,\n \"alphanum_fraction\": 0.5308641791343689,\n \"avg_line_length\": 13.235294342041016,\n \"blob_id\": \"b744ad8b51d816c7f194fc892ceb80a38c532fc1\",\n \"content_id\": \"1a5c39abd529011c795f3fbd8295d0f70ecfc170\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 243,\n \"license_type\": \"no_license\",\n \"max_line_length\": 22,\n \"num_lines\": 17,\n \"path\": \"/Python_stepik_2/3.2.1.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"s = input()\\na = input()\\nb = input()\\ns1 = ''\\nif a == b and a in s:\\n\\tprint('Impossible')\\nelse:\\n\\tfor i in range(1001):\\n\\t\\ts1 = s.replace(a, b)\\n\\t\\tif s == s1:\\n\\t\\t\\tprint(i)\\n\\t\\t\\tbreak\\n\\t\\telif i == 1000:\\n\\t\\t\\tprint('Impossible')\\n\\t\\t\\tbreak\\n\\t\\telse:\\n\\t\\t\\ts = s1\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.47428572177886963,\n \"alphanum_fraction\": 0.508571445941925,\n \"avg_line_length\": 16.600000381469727,\n \"blob_id\": \"30c02e851ef2d658967c9a1706fae0aa10ee758f\",\n \"content_id\": \"facff2a774f1656fb875e1562134027243139119\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 175,\n \"license_type\": \"no_license\",\n \"max_line_length\": 46,\n \"num_lines\": 10,\n \"path\": \"/Python_stepik_2/1.3.2.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"n, k = map(int, input().split())\\ndef c_rec(n, k):\\n\\tif k == 0:\\n\\t\\treturn 1\\n\\telif k > n:\\n\\t\\treturn 0\\n\\telse:\\n\\t\\treturn c_rec(n - 1, k) + c_rec(n - 1, k - 1)\\ny = c_rec(n, k)\\nprint(y)\"\n },\n {\n \"alpha_fraction\": 0.6692307591438293,\n \"alphanum_fraction\": 0.6769230961799622,\n \"avg_line_length\": 20.83333396911621,\n \"blob_id\": \"3bd03f4549980d40ba1b6501dfc517f664f6b515\",\n \"content_id\": \"33e79c16eb95b553d6f58a52ed6ff41a8826afce\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 130,\n \"license_type\": \"no_license\",\n \"max_line_length\": 63,\n \"num_lines\": 6,\n \"path\": \"/Python_stepik_2/3.3.7.py\",\n \"repo_name\": \"Arkkav/git-learing\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import re\\nimport sys\\n\\nfor line in sys.stdin:\\n\\tline = line.rstrip()\\n\\tprint(re.sub(r\\\"\\\\ba+\\\\b\\\", \\\"argh\\\", line, 1, flags=re.IGNORECASE))\"\n }\n]"},"num_files":{"kind":"number","value":58,"string":"58"}}},{"rowIdx":2980391,"cells":{"repo_name":{"kind":"string","value":"SK-CERT/dell-storage-client"},"repo_url":{"kind":"string","value":"https://github.com/SK-CERT/dell-storage-client"},"snapshot_id":{"kind":"string","value":"1e37db96a49ff1eaaeca3a64e9ea4770bba07cb3"},"revision_id":{"kind":"string","value":"c72586f194211f4e6c3d00fe878802f42ac362cb"},"directory_id":{"kind":"string","value":"daa265d6802252ad5fd0ecdf866bb21e124f62e7"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2020-12-29T19:46:12.931765","string":"2020-12-29T19:46:12.931765"},"revision_date":{"kind":"timestamp","value":"2020-03-05T14:04:09","string":"2020-03-05T14:04:09"},"committer_date":{"kind":"timestamp","value":"2020-03-05T14:04:09","string":"2020-03-05T14:04:09"},"github_id":{"kind":"number","value":238711105,"string":"238,711,105"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"null"},"gha_updated_at":{"kind":"null"},"gha_pushed_at":{"kind":"null"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.6138086915016174,"alphanum_fraction":0.6164708733558655,"avg_line_length":39.752357482910156,"blob_id":"5b0c8f91ee180534215567d6e1dfad476456aaa9","content_id":"47f1ebd7f8b25adee0e1e722525b36758b61215b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":17279,"license_type":"no_license","max_line_length":120,"num_lines":424,"path":"/dell_storage_api/volume.py","repo_name":"SK-CERT/dell-storage-client","src_encoding":"UTF-8","text":"\"\"\" This module contains classes for management of volumes in Storage Center managed by Dell Storage Manager\"\"\"\nfrom typing import Dict, Any\n\nfrom requests import Session\n\nfrom dell_storage_api.storage_object import StorageObject, StorageObjectCollection, StorageObjectFolder\n\n\nclass Volume(StorageObject):\n \"\"\"\n Class that represents volume in Storage Center\n \"\"\"\n ENDPOINT = '/StorageCenter/ScVolume'\n VOLUME_ENDPOINT = '/StorageCenter/ScVolume/%s'\n MAPPING_ENDPOINT = '/StorageCenter/ScVolume/%s/MapToServer'\n MAPPING_PROFILE_ENDPOINT = '/StorageCenter/ScVolume/%s/MappingProfileList'\n UNMAPPING_ENDPOINT = '/StorageCenter/ScVolume/%s/Unmap'\n RECYCLE_ENDPOINT = '/StorageCenter/ScVolume/%s/Recycle'\n EXPAND_TO_SIZE_ENDPOINT = '/StorageCenter/ScVolume/%s/ExpandToSize'\n EXPAND_ENDPOINT = '/StorageCenter/ScVolume/%s/Expand'\n\n def __init__(self, req_session: Session, base_url: str, name: str, instance_id: str,\n parent_folder_id: str, wwid: str, status: str) -> None:\n super().__init__(req_session, base_url, name, instance_id)\n self.parent_folder_id = parent_folder_id\n self.wwid = wwid\n self.status = status\n\n @classmethod\n def from_json(cls, req_session: Session, base_url: str, source_dict: Dict[Any, Any]) -> 'Volume':\n \"\"\"\n Class method that creates instance of Volume class from supplied dictionary. Source dictionary is expected\n to contain at least 'instanceId', 'name', 'deviceId' and 'volumeFolder' keys. Value of 'volumeFolder' is\n expected to be of type dict, containing at least key 'instanceId'\n :param req_session: requests.Session object with stored login cookies. (passed down\n from dell_storage_api.session.DsmSession)\n :param base_url: base URL of DSM\n :param source_dict: Dictionary containing data about Volume object\n :return: instance of a Volume class\n \"\"\"\n return Volume(req_session=req_session,\n base_url=base_url,\n name=source_dict['name'],\n instance_id=source_dict['instanceId'],\n parent_folder_id=source_dict['volumeFolder']['instanceId'],\n wwid=source_dict['deviceId'],\n status=source_dict['status']\n )\n\n @property\n def mapping_url(self) -> str:\n \"\"\"\n Return complete URL for mapping this volume to specific server\n :return: URL for volume mapping\n \"\"\"\n return self.build_url(self.MAPPING_ENDPOINT)\n\n @property\n def unmapping_url(self) -> str:\n \"\"\"\n Return complete URL for unmapping this volume from servers to which it is currently mapped\n :return: URL for volume unmapping\n \"\"\"\n return self.build_url(self.UNMAPPING_ENDPOINT)\n\n @property\n def mapping_profile_url(self) -> str:\n \"\"\"\n Return complete URL to read mapping profile of this volume.\n :return: URL for volume mapping list\n \"\"\"\n return self.build_url(self.MAPPING_PROFILE_ENDPOINT)\n\n @property\n def recycle_url(self) -> str:\n \"\"\"\n Return complete URL for moving this volume to recycle bin\n :return: URL for moving this volume to recycle bin\n \"\"\"\n return self.build_url(self.RECYCLE_ENDPOINT)\n\n @property\n def delete_url(self) -> str:\n \"\"\"\n Return complete URL to permanently remove this volume.\n WARNING: this action can not be undone\n :return: URL for removing this volume\n \"\"\"\n return self.build_url(self.VOLUME_ENDPOINT)\n\n @property\n def expand_url(self) -> str:\n \"\"\"\n Return complete URL for expanding this volume's capacity by specified amount.\n :return: URL for expanding this volume by specified amount\n \"\"\"\n return self.build_url(self.EXPAND_ENDPOINT)\n\n @property\n def expand_to_size_url(self) -> str:\n \"\"\"\n Return complete URL for expanding this volume's capacity to specified size\n :return: URL for expanding this volume to specified size\n \"\"\"\n return self.build_url(self.EXPAND_TO_SIZE_ENDPOINT)\n\n @property\n def modify_url(self) -> str:\n \"\"\"\n Return complete URL for modification of this volume. Modifiable attributes are 'Name' and 'VolumeFolder'\n :return: URL for modification of this volume\n \"\"\"\n return self.build_url(self.VOLUME_ENDPOINT)\n\n @property\n def details_url(self) -> str:\n \"\"\"\n Return complete URL for getting details about this volume\n :return: URL for details about this volume\n \"\"\"\n return self.build_url(self.VOLUME_ENDPOINT)\n\n def map_to_server(self, server_id: str) -> bool:\n \"\"\"\n Perform API call to DSM that maps this volume to server with instance ID specified by parameter 'server_id'. If\n supplied server_id is instance ID of a cluster, this volume will be mapped to every server that is part of that\n cluster.\n This operation fails if volume is already mapped to some server.\n :param server_id: Instance ID of server to which this volume will be mapped\n :return: True if operation is successful, otherwise False\n \"\"\"\n success = False\n payload = {'Server': server_id}\n resp = self.session.post(self.mapping_url, json=payload)\n if resp.status_code == 200:\n success = True\n print(\"OK - Volume '%s' (%s) sucessfully mapped to server.\" % (self.name, self.instance_id))\n else:\n print(\"Error: Failed to map volume - %s\" % resp.json().get('result'))\n return success\n\n def unmap(self) -> bool:\n \"\"\"\n Perform API call to DSM that unmaps this volume from any servers it is currently mapped to.\n WARNING: unmapping volume from active servers will cause those servers to loose connectivity with this volume.\n :return: True if operation is successful, otherwise False\n \"\"\"\n success = False\n resp = self.session.post(self.unmapping_url)\n if resp.status_code == 204:\n success = True\n print('OK - Volume successfully unmapped')\n else:\n print('Error: Failed to unamp volume - %s' % resp.text)\n return success\n\n def mapping(self) -> object:\n \"\"\"\n \"\"\"\n success = False\n resp = self.session.get(self.mapping_profile_url)\n if resp.status_code == 200:\n success = True\n mapping_profiles = resp.json()\n if mapping_profiles:\n return mapping_profiles[0]['server']\n else:\n print(\"Error: Failed to get volume mapping list\")\n return None\n\n def expand(self, size: str) -> bool:\n \"\"\"\n Perform API call to DSM that expands this volume by specified amount.\n :param size: Size by which this volume will be expanded (e.g.: 10GB or 1.2TB)\n :return: True if operation is successful, otherwise False\n \"\"\"\n success = False\n payload = {\"ExpandAmount\": size}\n resp = self.session.post(self.expand_url, json=payload)\n if resp.status_code == 200:\n success = True\n print(\"OK - Volume expanded by %s\" % size)\n else:\n print(\"Error: Failed to expand volume - %s\" % resp.json().get('result'))\n return success\n\n def expand_to_size(self, size: str) -> bool:\n \"\"\"\n Perform API call to DSM that expands this volume to the specified size. This method can be used only to\n increase volume size, DSM is unable to shrink volumes\n :param size: Size to which this volume is expanded (e.g.: 500GB or 2.5TB)\n :return: True if operation is successful, otherwise False\n \"\"\"\n success = False\n payload = {\"NewSize\": size}\n resp = self.session.post(self.expand_to_size_url, json=payload)\n if resp.status_code == 200:\n success = True\n print(\"OK - Volume expanded to size %s\" % size)\n else:\n print(\"Error: Failed to expand volume - %s\" % resp.json().get('result'))\n return success\n\n def recycle(self) -> bool:\n \"\"\"\n Perform API call to DSM that moves this volume to recycle bin. Volumes in recycle bin can be restored.\n :return: True if operation is successful, otherwise False\n \"\"\"\n success = False\n resp = self.session.post(self.recycle_url)\n if resp.status_code == 204:\n success = True\n print('OK - Volume successfully moved to recycle bin')\n else:\n print('Error: Failed to recycle volume - %s' % resp.text)\n return success\n\n def delete(self) -> bool:\n \"\"\"\n Perform API call to DSM that permanently deletes this volume.\n WARNING: This action can not be undone.\n :return: True if operation is successful, otherwise False\n \"\"\"\n success = False\n resp = self.session.delete(self.delete_url)\n if resp.status_code == 200:\n success = True\n print(\"Ok - Volume successfully deleted\")\n else:\n print(\"Error: Failed to delete volume - %s\" % resp.json().get('result'))\n return success\n\n def _modify_volume(self, payload: Dict[str, str]) -> bool:\n \"\"\"\n Internal method that performs API call to DSM to modify volume properties. Only properties that are modifiable\n are 'Name' and 'VolumeFolder'.\n :param payload: Dictionary with modified properties and their new values (e.g.: {'Name': 'new_volume_name'})\n :return: True if operation is successful, otherwise False\n \"\"\"\n # TODO: Move common functionality (like modify/rename/move) to base class\n success = False\n resp = self.session.put(self.modify_url, json=payload)\n if resp.status_code == 200:\n success = True\n print(\"Ok - Volume modified\")\n else:\n print(\"Error: Failed to modify volume\")\n return success\n\n def rename(self, new_name: str) -> bool:\n \"\"\"\n Perform API call to DSM to change current volume name to the new value.\n Note: Volume names do not have to be unique even within same folder\n :param new_name: New name for this volume\n :return: True if operation is successful, otherwise False\n \"\"\"\n if self._modify_volume({\"Name\": new_name}):\n self.name = new_name\n return True\n else:\n return False\n\n def move_to_folder(self, volume_folder_id: str) -> bool:\n \"\"\"\n Perform API call to DSM to move this volume to the folder with instance ID specified by parameters\n volume_folder_id.\n :param volume_folder_id: Instance ID of a Volume folder to which this volume will be moved\n :return: True if operation is successful, otherwise False\n \"\"\"\n if self._modify_volume({\"VolumeFolder\": volume_folder_id}):\n self.parent_folder_id = volume_folder_id\n return True\n else:\n return False\n\n def details(self) -> Dict[str, Any]:\n \"\"\"\n Perform API call to DSM to get all information available about this volume and returns it in\n form of a dictionary\n :return: Dictionary containing details about this volume.\n \"\"\"\n result: Dict[str, Any] = {}\n resp = self.session.put(self.details_url)\n if resp.status_code == 200:\n result = resp.json()\n else:\n print(\"Error: Failed to fetch volume details\")\n return result\n\n\nclass VolumeCollection(StorageObjectCollection):\n \"\"\" Collection of volume folders\"\"\"\n\n def find_by_parent_folder(self, folder_id: str) -> 'VolumeCollection':\n \"\"\"\n Return subset VolumeCollection that contains only volumes with specified parent folder.\n :param folder_id: Instance ID of a folder whose children should be in the result\n :return: Volumes with common parent specified by folder_id\n \"\"\"\n result = VolumeCollection()\n for volume in self:\n if volume.parent_folder_id == folder_id:\n result.add(volume)\n return result\n\n\nclass VolumeFolder(StorageObjectFolder):\n \"\"\" Class representing Volume Folder\"\"\"\n ENDPOINT = '/StorageCenter/ScVolumeFolder'\n VOLUME_FOLDER_ENDPOINT = '/StorageCenter/ScVolumeFolder/%s'\n\n @classmethod\n def from_json(cls, req_session: Session, base_url: str, source_dict: Dict[Any, Any]) -> 'VolumeFolder':\n \"\"\"\n Class method that creates instance of VolumeFolder class from supplied dictionary. Source dictionary is expected\n to contain at least 'instanceId' and 'name' keys. Optional key 'parent' can be present and its value is\n expected to be dict containing key 'instanceId'.\n :param req_session: requests.Session object with stored login cookies. (passed down\n from dell_storage_api.session.DsmSession)\n :param base_url: base URL of DSM\n :param source_dict: Dictionary containing data about VolumeFolder object\n :return: instance of a VolumeFolder class\n \"\"\"\n return VolumeFolder(req_session=req_session,\n base_url=base_url,\n name=source_dict[\"name\"],\n instance_id=source_dict[\"instanceId\"],\n parent_id=source_dict.get('parent', {}).get('instanceId', None))\n\n @property\n def modify_url(self) -> str:\n \"\"\"\n Return complete URL for modification of this volume folder\n :return: URL for folder modification\n \"\"\"\n return self.build_url(self.VOLUME_FOLDER_ENDPOINT)\n\n @property\n def details_url(self) -> str:\n \"\"\"\n Return complete URL for getting details about this volume folder\n :return: URL for details about this volume folder\n \"\"\"\n return self.build_url(self.VOLUME_FOLDER_ENDPOINT)\n\n @property\n def delete_url(self) -> str:\n \"\"\"\n Return complete URL to permanently remove this volume Folder.\n WARNING: this action can not be undone\n :return: URL for removing this volume folder\n \"\"\"\n return self.build_url(self.VOLUME_FOLDER_ENDPOINT)\n\n def _modify_volume_folder(self, payload: Dict[str, str]) -> bool:\n \"\"\"\n Internal method that performs call to DMS to modify this volume folder. Only modifiable properties are\n 'Name' and 'Parent'.\n :param payload: Dictionary with modified properties and their new values (e.g.: {'Name': 'new_folder_name'})\n :return: True if operation is successful, otherwise False\n \"\"\"\n success = False\n resp = self.session.put(self.modify_url, json=payload)\n if resp.status_code == 200:\n success = True\n print(\"Ok - Volume folder modified\")\n else:\n print(\"Error: Failed to modify volume folder\")\n return success\n\n def rename(self, name: str) -> bool:\n \"\"\"\n Perform API call to DSM to change this volume folder name to the new value.\n Note: Volume Folder names do not have to be unique even within same parent folder\n :param name: New name for this volume folder\n :return: True if operation is successful, otherwise False\n \"\"\"\n if self._modify_volume_folder({\"Name\": name}):\n self.name = name\n return True\n else:\n return False\n\n def move_to_folder(self, parent_folder_id: str) -> bool:\n \"\"\"\n Perform API call to DSM to move this folder to different parent folder specified by Instance ID in\n 'parent_folder_id' parameter.\n :param parent_folder_id: Instance ID of the new parent folder\n :return: True if operation is successful, otherwise False\n \"\"\"\n if self._modify_volume_folder({\"VolumeFolder\": parent_folder_id}):\n self.parent_id = parent_folder_id\n return True\n else:\n return False\n\n def details(self) -> Dict[str, Any]:\n \"\"\"\n Perform API call to DSM to fetch details about this volume folder. Result is returned as dictionary\n :return: Dictionary containing details about this volume folder\n \"\"\"\n result: Dict[str, Any] = {}\n resp = self.session.put(self.details_url)\n if resp.status_code == 200:\n result = resp.json()\n else:\n print(\"Error: Failed to fetch volume folder details\")\n return result\n\n def delete(self) -> bool:\n \"\"\"\n Perform API call to DSM to permanently delete this volume folder.\n Note: Volume folder can not be removed if it contains volumes.\n WARNING: This action can not be undone\n :return: True if operation is successful, otherwise False\n \"\"\"\n success = False\n resp = self.session.delete(self.delete_url)\n if resp.status_code == 200:\n success = True\n print(\"Ok - Volume folder successfully deleted\")\n else:\n print(\"Error: Failed to delete volume folder - %s\" % resp.json().get('result'))\n return success\n"},{"alpha_fraction":0.6333118677139282,"alphanum_fraction":0.6353750228881836,"avg_line_length":43.31428527832031,"blob_id":"c94848c943ad0353550f780b1532f4296a431234","content_id":"5157cf487c4446bce91a145f3dc465adf6a75b79","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":23265,"license_type":"no_license","max_line_length":127,"num_lines":525,"path":"/bin/dell-storage-client","repo_name":"SK-CERT/dell-storage-client","src_encoding":"UTF-8","text":"#!/usr/bin/env python3\n\n# pylint: disable=C0103,C0111\nimport argparse\nimport getpass\nimport json\nfrom typing import Optional, List\n\nfrom texttable import Texttable\n\nfrom dell_storage_api import DsmSession, StorageCenter\n\nCMD_CONST_VOLUME = 'volume'\nCMD_CONST_VOLUME_CREATE = 'create'\nCMD_CONST_VOLUME_LIST = 'list'\nCMD_CONST_VOLUME_MAP = 'map'\nCMD_CONST_VOLUME_UNMAP = 'unmap'\n\nCMD_CONST_VOLUME_FOLDER = 'volume_folder'\nCMD_CONST_VOLUME_FOLDER_CREATE = 'create'\nCMD_CONST_VOLUME_FOLDER_LIST = 'list'\n\nCMD_CONST_STORAGE_CENTER = 'storage_center'\nCMD_CONST_STORAGE_CENTER_LIST = 'list'\n\nCMD_CONST_SERVER = 'server'\nCMD_CONST_SERVER_LIST = 'list'\n\nclass ReturnCode: # pylint: disable=R0903\n \"\"\"Convenience class that holds semantic return codes \"\"\"\n SUCCESS = 0\n FAILURE = 1\n\n\nclass ServerType:\n \"\"\"Convenience class that represents various types of Server object\"\"\"\n\n def __init__(self) -> None:\n self._server = 'Server'\n self._cluster = 'Cluster'\n self._all = 'All'\n self._all_types = [self._all, self._server, self._cluster]\n\n @property\n def server(self) -> str:\n \"\"\"\n Return string representing Physical server type\n :return: Physical server type\n \"\"\"\n return self._server\n\n @property\n def cluster(self) -> str:\n \"\"\"\n Return string representing Custer server type\n :return: Cluster server type\n \"\"\"\n return self._cluster\n\n @property\n def all_keyword(self) -> str:\n \"\"\"\n Return list of all possible server type keywords (including magic keyword 'All')\n :return: list of all server type keywords\n \"\"\"\n return self._all\n\n @property\n def all_types(self) -> List[str]:\n \"\"\"\n Magic keyword representing all/any server types\n :return: keyword representing all types\n \"\"\"\n return self._all_types\n\n\nSERVER_TYPES = ServerType()\n\ndef print_table(table):\n cli_args = parse_arguments()\n\n if cli_args.json:\n res = []\n for row in table._rows:\n line = {}\n for i in range(len(table._header)):\n line[table._header[i]] = row[i]\n res.append(line)\n text = json.dumps(res)\n else:\n text = table.draw()\n\n print (text)\n\n\ndef volume_create(storage: StorageCenter, name: str, size: str,\n unique_name: bool = True, folder_id: str = '', map_to_id: str = '') -> int:\n \"\"\"\n Create new volume in Storage Center.\n :param storage: Storage Center in which new volume will be created\n :param name: Name of the volume\n :param size: Size of the volume (e.g.: '500GB' or '1.5TB')\n :param unique_name: Should the volume creation fail if the valoume name already exists in this folder ?\n :param folder_id: Folder in which the volume will be created (Defaults to root folder)\n :param map_to_id: Instance ID of a server (or cluster) to which this volume should be mapped\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\n \"\"\"\n if map_to_id:\n mapping_server = storage.server_list().find_by_instance_id(map_to_id)\n if mapping_server is None:\n print(\"Volume can't be mapped to server with instance ID '%s'. No such server\" % map_to_id)\n return ReturnCode.FAILURE\n if unique_name and storage.volume_list().find_by_name(name):\n print(\"Volume with name '%s' already exists\" % name)\n return ReturnCode.FAILURE\n\n new_volume = storage.new_volume(name, size, folder_id)\n if new_volume is not None:\n print(\"OK - Volume '%s' created with instance ID '%s'\" % (new_volume.name, new_volume.instance_id))\n if map_to_id:\n if new_volume.map_to_server(map_to_id):\n return ReturnCode.SUCCESS\n else:\n return ReturnCode.FAILURE\n else:\n return ReturnCode.SUCCESS\n else:\n return ReturnCode.FAILURE\n\n\ndef volume_map(storage: StorageCenter, volume_id: str, server_id: str) -> int:\n \"\"\"\n Map existing volume to the server (or cluster).\n NOTE: Volume that is already mapped to server can not be mapped again without unmapping it first.\n :param storage: Storage Center where the volume is located\n :param volume_id: Instance ID of a Volume to be mapped\n :param server_id: Instance ID of a Server (or a cluster) to which volume will be mapped\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\n \"\"\"\n volume = storage.volume_list().find_by_instance_id(volume_id)\n if volume is None:\n print(\"Failed to map volume with ID '%s'. No such volume\" % volume_id)\n return ReturnCode.FAILURE\n server = storage.server_list().find_by_instance_id(server_id)\n if server is None:\n print(\"failed to map volume '%s' (%s) to server with ID '%s'. No such server\" % (volume.name,\n volume.instance_id,\n server_id))\n return ReturnCode.FAILURE\n if volume.map_to_server(server_id=server_id):\n return ReturnCode.SUCCESS\n else:\n return ReturnCode.FAILURE\n\n\ndef volume_unmap(storage: StorageCenter, volume_id: str) -> int:\n \"\"\"\n Unmap volume from all servers it is currently mapped to\n :param storage: Storage Center where volume is located\n :param volume_id: Instance ID of a volume to be unampped\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\n \"\"\"\n volume = storage.volume_list().find_by_instance_id(volume_id)\n if volume is None:\n print(\"Failed to unmap volume with ID '%s'. No such volume\" % volume_id)\n return ReturnCode.FAILURE\n if volume.unmap():\n return ReturnCode.SUCCESS\n else:\n return ReturnCode.FAILURE\n\n\ndef volume_list(storage: StorageCenter, folder_id: str = '', show_mapping: bool = False) -> int:\n \"\"\"\n Print table of Volumes present in Storage Center in specified volume folder.\n :param storage: Storage Center, from which to list volumes\n :param folder_id: Volume Folder, from which to list volumes (Defaults to root)\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\n \"\"\"\n table = Texttable(max_width=120)\n all_volumes = storage.volume_list()\n if folder_id:\n all_volumes = all_volumes.find_by_parent_folder(folder_id)\n\n if show_mapping:\n table.header(['volume', 'instance_id', 'parent_folder', 'wwid', 'status', 'mapping'])\n table.set_cols_dtype(['t', 't', 't', 't','t','t'])\n else:\n table.header(['volume', 'instance_id', 'parent_folder', 'wwid', 'status'])\n table.set_cols_dtype(['t', 't', 't', 't','t'])\n\n for volume in all_volumes:\n if show_mapping:\n mapping = volume.mapping()\n mapping_name = mapping['instanceName'] if mapping else None\n table.add_row([volume.name, volume.instance_id, volume.parent_folder_id, volume.wwid, volume.status, mapping_name])\n else:\n table.add_row([volume.name, volume.instance_id, volume.parent_folder_id, volume.wwid, volume.status])\n\n print_table(table)\n\n return ReturnCode.SUCCESS\n\n\ndef volume_folder_list(storage: StorageCenter, parent_id: str = '') -> int:\n \"\"\"\n Print table of Volume Folders present in Storage Center in specified parent Volume Folder\n :param storage: Storage Center, from which to list volume folders\n :param parent_id: Parent volume folder, from which the child volume folders will be listed (Defaults to root)\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\n \"\"\"\n # TODO: Adaptable table width\n table = Texttable(max_width=120)\n table.header(['folder', 'instance_id', 'parent_instance_id'])\n table.set_cols_dtype(['t', 't', 't'])\n folder_list = storage.volume_folder_list()\n if parent_id:\n folder_list = folder_list.find_by_parent_id(parent_id)\n for folder in folder_list:\n table.add_row([folder.name, folder.instance_id, folder.parent_id])\n print_table(table)\n\n return ReturnCode.SUCCESS\n\n\ndef volume_folder_create(storage: StorageCenter, folder_name: str, folder_parent_id: str = '',\n unique_name: bool = True) -> int:\n \"\"\"\n Create new volume folder in Storage Center\n :param storage: Storage Center where the new folder will be created\n :param folder_name: Name of the new folder\n :param folder_parent_id: Instance ID of a parent folder for the new volume folder\n :param unique_name: Should the volume creation fail if volume with same name already exists within parent folder\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\n \"\"\"\n if unique_name and storage.volume_folder_list().find_by_name(folder_name):\n print(\"Volume folder with name '%s' already exists\" % folder_name)\n return ReturnCode.FAILURE\n if storage.new_volume_folder(folder_name, folder_parent_id):\n return ReturnCode.SUCCESS\n else:\n return ReturnCode.FAILURE\n\n\ndef server_list(storage: StorageCenter, object_type: str) -> int:\n \"\"\"\n Print table of Servers defined in Storage Center.\n :param storage: Storage Center from which servers will be listed\n :param object_type: Limit output only to Servers of specific\n type (e.g.: SERVER_TYPES.server or SERVER_TYPES.cluster)\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\n \"\"\"\n table = Texttable(max_width=120)\n table.header(['server', 'type', 'instance_id'])\n table.set_cols_dtype(['t', 't', 't'])\n\n servers = storage.server_list()\n if object_type == SERVER_TYPES.server:\n servers = servers.filter_physical_servers()\n elif object_type == SERVER_TYPES.cluster:\n servers = servers.filter_clusters()\n\n for server in servers:\n table.add_row([server.name,\n server.pretty_type(),\n server.instance_id])\n print_table(table)\n return ReturnCode.SUCCESS\n\n\ndef storage_center_list(session: DsmSession) -> int:\n \"\"\"\n Print table of Storage Centers connected to Dell Storage Manager\n :param session: Authenticated session with Dell Storage Manager\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\n \"\"\"\n storage_centers = session.storage_centers()\n table = Texttable()\n table.header([\"name\", \"ip\", \"instance_id\", \"serial\"])\n table.set_cols_dtype(['t', 't', 't', 't'])\n\n for storage_center in storage_centers:\n table.add_row([storage_center.name,\n storage_center.ip_addr,\n storage_center.instance_id,\n storage_center.serial_num])\n print_table(table)\n\n return ReturnCode.SUCCESS\n\n\ndef _find_storage_center(session: DsmSession, instance_id: str) -> Optional[StorageCenter]:\n \"\"\"\n Find and return Storage Center with specified Instance ID connected to Dell Storage manager. If no such\n Storage Center is found, return None.\n :param session: Authenticated session with Dell Storage Manager\n :param instance_id: Instance ID of a Storage Center that should be returned\n :return: Storage Center with Given Instance ID or None if no such ID is found\n \"\"\"\n storage_center = session.storage_centers().find_by_instance_id(instance_id)\n if storage_center is None:\n print(\"Failed to find storage center with instance ID '%s'. Try listing all storage \"\n \"centers with command 'storage_center list'\" % instance_id)\n return None\n else:\n return storage_center # type: ignore\n\n\ndef exit_cli(session: DsmSession, return_code: int) -> None:\n \"\"\"\n Perform Session logout and exit program\n :param session: Session with Dell Storage Manager\n :param return_code: Return Code to exit with\n :return: None\n \"\"\"\n session.logout(silent=True)\n exit(return_code)\n\n\ndef parse_arguments() -> argparse.Namespace: # pylint: disable=R0914\n \"\"\"\n Define and parse CLI commands, subcommands and arguments using argparse module\n :return: argparse.Namespace with parsed CLI argument values\n \"\"\"\n parser = argparse.ArgumentParser()\n\n # General options\n parser.add_argument('-H', '--host', required=True, help=\"Hostname or IP address of Dell Storage Center\")\n parser.add_argument('-P', '--port', default=3033, help=\"Management port of Dell storage Center\")\n parser.add_argument('-u', '--user', help='Login username')\n parser.add_argument('-p', '--password', help='Login password')\n parser.add_argument('-j', '--json', action='store_true', help='Output in JSON format')\n\n # Top level subcommands\n command_parser = parser.add_subparsers(dest='command')\n\n # Storage Center subcommands\n storage_center_parser = command_parser.add_parser(CMD_CONST_STORAGE_CENTER)\n storage_center_parser_cmd = storage_center_parser.add_subparsers(dest='storage_center_commands')\n # List Storage centers\n storage_center_parser_cmd.add_parser(CMD_CONST_STORAGE_CENTER_LIST)\n\n # Volume subcommands\n volume_parser = command_parser.add_parser(CMD_CONST_VOLUME)\n volume_parser_cmd = volume_parser.add_subparsers(dest='volume_commands')\n # Create volume\n volume_create_args = volume_parser_cmd.add_parser(CMD_CONST_VOLUME_CREATE)\n volume_create_args.add_argument('-s', '--size', required=True, help='Size of the new volume. Example: \"500GB\"')\n volume_create_args.add_argument('-n', '--name', required=True, help='Name of the new volume')\n volume_create_args.add_argument('-f', '--folder-id', dest='folder_id', help='Instance ID of parent folder')\n volume_create_args.add_argument('-m' '--map-to-server', dest='map_to_server',\n help='Instance ID of server to which new volume will be mapped')\n volume_create_args.add_argument('-S', '--storage-id', required=True, dest='storage_id',\n help='Instance ID of storage center where the volume will be created')\n volume_create_args.add_argument('-Q', '--non-unique-name', default=False, action='store_true',\n help='If this flag is present, volume creation wont fail if there is another '\n 'volume with the same name')\n # List Volumes\n volume_list_args = volume_parser_cmd.add_parser(CMD_CONST_VOLUME_LIST)\n volume_list_args.add_argument('-S', '--storage-id', required=True, dest='storage_id',\n help='Instance ID of storage center from which the volumes will be listed')\n volume_list_args.add_argument('-f', '--folder-id', dest='folder_id', help='Instance ID of folder from '\n 'which the volumes will be listed')\n volume_list_args.add_argument('-m', '--show-mapping', dest='show_mapping', action=\"store_true\",\n help='Show mapping profile (this will slow down things!)')\n\n # Map Volume\n volume_map_args = volume_parser_cmd.add_parser(CMD_CONST_VOLUME_MAP)\n volume_map_args.add_argument('-S', '--storage-id', required=True, dest='storage_id',\n help='Instance ID of storage center where volume is located')\n volume_map_args.add_argument('-v', '--volume-id', dest='volume_id', help='Instance ID of volume to be mapped')\n volume_map_args.add_argument('-m' '--map-to-server', dest='map_to_server',\n help='Instance ID of server to which this volume will be mapped')\n # Unmap Volume\n volume_unmap_args = volume_parser_cmd.add_parser(CMD_CONST_VOLUME_UNMAP)\n volume_unmap_args.add_argument('-S', '--storage-id', required=True, dest='storage_id',\n help='Instance ID of storage center where volume is located')\n volume_unmap_args.add_argument('-v', '--volume-id', dest='volume_id', help='Instance ID of volume to be unmapped')\n\n # Volume Folder subcommands\n volume_folder_parser = command_parser.add_parser('volume_folder')\n volume_folder_parser_cmd = volume_folder_parser.add_subparsers(dest='volume_folder_commands')\n # List volume folders\n volume_folder_list_args = volume_folder_parser_cmd.add_parser(CMD_CONST_VOLUME_FOLDER_LIST)\n volume_folder_list_args.add_argument('-S' '--storage-id', required=True, dest='storage_id',\n help='Instance ID of storage center from which, volume folders will be listed')\n volume_folder_list_args.add_argument('-f', '--folder-id', dest='folder_id',\n help='Instance ID of folder from which the child volumes will be listed')\n\n # Create volume folder\n volume_folder_create_args = volume_folder_parser_cmd.add_parser(CMD_CONST_VOLUME_FOLDER_CREATE)\n volume_folder_create_args.add_argument('-n', '--name', help='Name of the new folder')\n volume_folder_create_args.add_argument('-f', '--folder-id', dest='folder_id', help='Instance ID of parent folder')\n volume_folder_create_args.add_argument('-S' '--storage-id', required=True, dest='storage_id',\n help='Instance ID of storage center from which,'\n 'volume folders will be listed')\n volume_folder_create_args.add_argument('-Q', '--non-unique-name', default=False, action='store_true',\n help='If this flag is present, folder creation wont fail if there is another'\n ' folder with the same name in specified parent folder')\n\n # Server subcommands\n server_parser = command_parser.add_parser('server')\n server_parser_cmd = server_parser.add_subparsers(dest='server_commands')\n # List Servers\n server_list_args = server_parser_cmd.add_parser(CMD_CONST_SERVER_LIST)\n server_list_args.add_argument('-t', '--type', default=SERVER_TYPES.all_keyword, choices=SERVER_TYPES.all_types,\n help='Print only server object of selected type. '\n '(Default=%s)' % SERVER_TYPES.all_keyword)\n server_list_args.add_argument('-S' '--storage-id', required=True, dest='storage_id',\n help='Instance ID of storage center from which, servers will be listed')\n return parser.parse_args()\n\n\ndef execute_command(args: argparse.Namespace, session: DsmSession) -> int: # pylint: disable=R0912,R0915\n \"\"\"\n Execute CLI command based on parsed arguments and parameters\n :param args: Parsed argparse CLI arguments\n :param session: Authenticated session with Dell Storage manager\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a performed command\n \"\"\"\n # Volume commands\n if args.command == CMD_CONST_VOLUME:\n # Create Volume\n if args.volume_commands == CMD_CONST_VOLUME_CREATE:\n unique_name = not args.non_unique_name\n storage_center = _find_storage_center(session, args.storage_id)\n if storage_center is None:\n ret_code = ReturnCode.FAILURE\n else:\n ret_code = volume_create(storage=storage_center, name=args.name, size=args.size,\n unique_name=unique_name, map_to_id=args.map_to_server)\n # List Volumes\n elif args.volume_commands == CMD_CONST_VOLUME_LIST:\n storage_center = _find_storage_center(session, args.storage_id)\n if storage_center is None:\n ret_code = ReturnCode.FAILURE\n else:\n parent_id = args.folder_id or ''\n show_mapping = args.show_mapping or False\n ret_code = volume_list(storage_center, parent_id, show_mapping)\n elif args.volume_commands == CMD_CONST_VOLUME_MAP:\n storage_center = _find_storage_center(session, args.storage_id)\n if storage_center is None:\n ret_code = ReturnCode.FAILURE\n else:\n ret_code = volume_map(storage_center, args.volume_id, args.map_to_server)\n elif args.volume_commands == CMD_CONST_VOLUME_UNMAP:\n storage_center = _find_storage_center(session, args.storage_id)\n if storage_center is None:\n ret_code = ReturnCode.FAILURE\n else:\n ret_code = volume_unmap(storage_center, args.volume_id)\n # Default branch\n else:\n ret_code = ReturnCode.FAILURE\n\n # Storage Center commands\n elif args.command == CMD_CONST_STORAGE_CENTER:\n # List Storage Centers\n if args.storage_center_commands == CMD_CONST_STORAGE_CENTER_LIST:\n ret_code = storage_center_list(session)\n # Default branch\n else:\n ret_code = ReturnCode.FAILURE\n\n # Volume Folder commands\n elif args.command == CMD_CONST_VOLUME_FOLDER:\n # List Volume folders\n if args.volume_folder_commands == CMD_CONST_VOLUME_FOLDER_LIST:\n storage_center = _find_storage_center(session, args.storage_id)\n if storage_center is None:\n ret_code = ReturnCode.FAILURE\n else:\n parent_folder_id = args.folder_id or ''\n ret_code = volume_folder_list(storage_center, parent_folder_id)\n # Create volume folder\n elif args.volume_folder_commands == CMD_CONST_VOLUME_FOLDER_CREATE:\n unique_name = not args.non_unique_name\n storage_center = _find_storage_center(session, args.storage_id)\n if storage_center is None:\n ret_code = ReturnCode.FAILURE\n else:\n ret_code = volume_folder_create(storage_center, args.name, args.folder_id, unique_name)\n # Default branch\n else:\n ret_code = ReturnCode.FAILURE\n # Server commands\n elif args.command == CMD_CONST_SERVER:\n # List servers\n if args.server_commands == CMD_CONST_SERVER_LIST:\n storage_center = _find_storage_center(session, args.storage_id)\n if storage_center is not None:\n ret_code = server_list(storage_center, args.type)\n else:\n ret_code = ReturnCode.FAILURE\n else:\n ret_code = ReturnCode.FAILURE\n else:\n ret_code = ReturnCode.FAILURE\n\n return ret_code\n\n\ndef main() -> None:\n # parse CLI arguments\n cli_args = parse_arguments()\n\n # Request missing arguments via CLI dialog\n if not cli_args.user:\n cli_args.user = input(\"Username: \")\n\n if not cli_args.password:\n cli_args.password = getpass.getpass()\n\n # Initialize Session with Storage controller\n scm_session = DsmSession(cli_args.user, cli_args.password, cli_args.host, cli_args.port, verify_cert=False)\n if not scm_session.login():\n exit(ReturnCode.SUCCESS)\n\n success = execute_command(cli_args, scm_session)\n exit_cli(scm_session, success)\n\n\nif __name__ == '__main__':\n main()\n"},{"alpha_fraction":0.6169469952583313,"alphanum_fraction":0.6184279918670654,"avg_line_length":42.96744155883789,"blob_id":"5132fe82785279d3a829aaf79dd09d7c91394e9c","content_id":"aa59ee756b6896087da07d54f2425b0e9878c0e6","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":9453,"license_type":"no_license","max_line_length":117,"num_lines":215,"path":"/dell_storage_api/storage_center.py","repo_name":"SK-CERT/dell-storage-client","src_encoding":"UTF-8","text":"\"\"\" This module contains classes that represent Storage Centers managed by Dell Storage manager (DSM) \"\"\"\nfrom typing import Any, Dict, Optional\n\nimport requests\n\nfrom dell_storage_api.storage_object import StorageObject, StorageObjectFolder, StorageObjectCollection, \\\n StorageObjectFolderCollection\nfrom dell_storage_api.volume import Volume, VolumeCollection, VolumeFolder\nfrom dell_storage_api.server import Server, ServerCollection\n\n\nclass StorageCenter(StorageObject):\n \"\"\"\n Class representing physical Storage Center managed by DSM\n \"\"\"\n SERVER_FOLDER_LIST_ENDPOINT = '/StorageCenter/StorageCenter/%s/ServerFolderList'\n SERVER_LIST_ENDPOINT = '/StorageCenter/StorageCenter/%s/ServerList'\n SERVER_FOLDER_ENDPOINT = '/StorageCenter/ScServerFolder/%s'\n\n VOLUME_FOLDER_LIST_ENDPOINT = '/StorageCenter/StorageCenter/%s/VolumeFolderList'\n VOLUME_LIST_ENDPOINT = '/StorageCenter/StorageCenter/%s/VolumeList'\n\n def __init__(self, req_session: requests.Session, base_url: str, name: str,\n instance_id: str, serial_num: str, ip_addr: str) -> None:\n super(StorageCenter, self).__init__(req_session, base_url, name, instance_id)\n self.serial_num = serial_num\n self.ip_addr = ip_addr\n\n @property\n def server_folder_list_url(self) -> str:\n \"\"\"\n Return complete URL to list all server folders in this Storage Center\n :return: URL for Server Folder listing\n \"\"\"\n return self.build_url(self.SERVER_FOLDER_LIST_ENDPOINT)\n\n @property\n def server_list_url(self) -> str:\n \"\"\"\n Return complete URL to list all servers defined in this Storage Center\n :return: URL for Server listing\n \"\"\"\n return self.build_url(self.SERVER_LIST_ENDPOINT)\n\n @property\n def volume_folder_list_url(self) -> str:\n \"\"\"\n Return complete URL to list all volume folders in this Storage Center\n :return: URL for Volume Folder listing\n \"\"\"\n return self.build_url(self.VOLUME_FOLDER_LIST_ENDPOINT)\n\n @property\n def volume_list_url(self) -> str:\n \"\"\"\n Return complete URL fo list all volumes present in this Storage Center\n :return: URL for Volume listning\n \"\"\"\n return self.build_url(self.VOLUME_LIST_ENDPOINT)\n\n def server_folder_list(self) -> StorageObjectFolderCollection:\n \"\"\"\n Return collection of all Server Folders in this Storage Center\n :return: Collection of all Server Folders\n \"\"\"\n return self._fetch_object_list(self.server_folder_list_url)\n\n def server_list(self) -> ServerCollection:\n \"\"\"\n Return collection of all servers defined in this Storage Center\n :return: Collection of all Servers\n \"\"\"\n result = ServerCollection()\n for server_data in self._fetch_object_list(self.server_list_url):\n result.add(Server.from_json(req_session=self.session,\n base_url=self.base_url,\n source_dict=server_data))\n return result\n\n def volume_folder_list(self) -> StorageObjectFolderCollection:\n \"\"\"\n Return collection of all Volume Folders in this Storage Center\n :return: Collection of all Volume Folders\n \"\"\"\n result = StorageObjectFolderCollection()\n for volume_folder_data in self._fetch_object_list(url=self.volume_folder_list_url):\n result.add(VolumeFolder.from_json(req_session=self.session,\n base_url=self.base_url,\n source_dict=volume_folder_data))\n return result\n\n def volume_list(self) -> VolumeCollection:\n \"\"\"\n Return collection of all volumes present in this Storage Center\n :return: Collection of all volumes\n \"\"\"\n result = VolumeCollection()\n for volume_data in self._fetch_object_list(url=self.volume_list_url):\n result.add(Volume.from_json(req_session=self.session,\n base_url=self.base_url,\n source_dict=volume_data))\n return result\n\n def _find_volume_folder_root(self) -> Optional[StorageObjectFolder]:\n \"\"\"\n Internal method to find root Volume Folder that contains all other Volumes and Volume Folders. Method returns\n None in case that there is problem with data fetching or lookup.\n :return: Root Volume Folder or None in case of failure\n \"\"\"\n all_folders = self.volume_folder_list()\n if not all_folders:\n print(\"Error: Failed to fetch volume folder list\")\n return None\n else:\n root_folder = self.volume_folder_list().root_folder()\n if root_folder is None:\n print(\"Error: Failed to lookup root volume folder in list of all folders. \"\n \"This really should not happen\")\n return None\n else:\n return root_folder\n\n def new_volume_folder(self, name: str, parent_folder_id: str = '') -> Optional[VolumeFolder]:\n \"\"\"\n Create new Volume Folder in Storage Center and return object representing this new folder. This method\n returns None in case there is a problem with folder creation.\n :param name: Name for the new Volume Folder\n :param parent_folder_id: Instance ID of parent folder. Defaults to root folder\n :return: new VolumeFolder object or None in case of failure\n \"\"\"\n if not parent_folder_id:\n parent_folder = self._find_volume_folder_root()\n if parent_folder is not None:\n parent_folder_id = parent_folder.instance_id\n else:\n print(\"Error: Failed to create new volume folder\")\n return None\n if parent_folder_id is None:\n print(\"Error: Failed to create new volume folder\")\n return None\n url = self.base_url + VolumeFolder.ENDPOINT\n payload = {\"Name\": name,\n \"Parent\": parent_folder_id,\n \"StorageCenter\": self.instance_id}\n resp = self.session.post(url, json=payload)\n if resp.status_code == 201:\n return VolumeFolder.from_json(self.session, self.base_url, resp.json())\n else:\n print(\"Error: Failed to create new volume folder.\")\n return None\n\n def new_volume(self, name: str, size: str, volume_folder_id: str = '') -> Optional[Volume]:\n \"\"\"\n Create new Volume in Storage Center and return object representing this new volume. This method returns None\n in case there is a problem with volume creation.\n :param name: Name of the new volume\n :param size: Size of the new volume (e.g: '100GB' or '1.5TB')\n :param volume_folder_id: Instance ID of folder in which this volume will be created. Defaults to root folder\n :return: new Volume object or None in case of failure\n \"\"\"\n if not volume_folder_id:\n volume_folder = self._find_volume_folder_root()\n if volume_folder is not None:\n volume_folder_id = volume_folder.instance_id\n else:\n print(\"Error: Failed to create new volume\")\n return None\n if volume_folder_id is None:\n print(\"Error: Failed to create new volume\")\n return None\n url = self.base_url + Volume.ENDPOINT\n payload = {\"Name\": name,\n \"Size\": size,\n \"StorageCenter\": self.instance_id,\n \"VolumeFolder\": volume_folder_id}\n resp = self.session.post(url, json=payload)\n if resp.status_code == 201:\n return Volume.from_json(self.session, self.base_url, resp.json())\n else:\n print(\"Error: Failed to create new volume. (%d) - %s\" % (resp.status_code, resp.text))\n return None\n\n def _fetch_object_list(self, url: str) -> Dict[Any, Any]:\n \"\"\"\n Internal generic method to fetch list of object from supplied URL. This method returns raw dictionary created\n from json in response body. This method returns empty dictionary if there is problem with data fetching\n :param url: URL of API endpoint that returns (json) list of objects\n :return: raw dictionary of objects returned by API endpoint\n \"\"\"\n result: Dict[Any, Any] = {}\n resp = self.session.get(url)\n if resp.status_code == 200:\n result = resp.json()\n else:\n print(\"Error: Failed to fetch object list (%d) - %s\" % (resp.status_code, resp.text))\n return result\n\n\nclass StorageCenterCollection(StorageObjectCollection):\n \"\"\"\n Collection representing multiple physical Storage Centers managed by single DSM\n \"\"\"\n\n def find_by_serial_num(self, serial_num: str) -> Optional[StorageCenter]:\n \"\"\"\n Return Storage Center object with given serial number. In case there is no such Storage Center, return None\n :param serial_num: Serial number of physical Storage Center\n :return: StorageCenter object with given serial number or None\n \"\"\"\n result = None\n for storage_center in self:\n if storage_center.serial_num == serial_num:\n result = storage_center\n break\n return result\n"},{"alpha_fraction":0.8702290058135986,"alphanum_fraction":0.8702290058135986,"avg_line_length":64.5,"blob_id":"1034ecb922e4f3409e8af44a44d510fd584cc141","content_id":"ab49b16462340df324c28b4516a9d66af5a492ec","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":131,"license_type":"no_license","max_line_length":82,"num_lines":2,"path":"/dell_storage_api/__init__.py","repo_name":"SK-CERT/dell-storage-client","src_encoding":"UTF-8","text":"from dell_storage_api.session import DsmSession\nfrom dell_storage_api.storage_center import StorageCenterCollection, StorageCenter\n"},{"alpha_fraction":0.6359223127365112,"alphanum_fraction":0.6361920237541199,"avg_line_length":37.625,"blob_id":"f9e3cf0324a8039bbcab451c9665fcf16f5fafdf","content_id":"0dbff666ff32d69a76022c0358f2fb131594468a","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3708,"license_type":"no_license","max_line_length":116,"num_lines":96,"path":"/dell_storage_api/server.py","repo_name":"SK-CERT/dell-storage-client","src_encoding":"UTF-8","text":"\"\"\" This module contains classes representing servers connected to the Storage Center \"\"\"\nfrom typing import Dict, Any\n\nfrom requests import Session\n\nfrom dell_storage_api.storage_object import StorageObject, StorageObjectCollection\n\n\nclass Server(StorageObject):\n \"\"\"\n Class representing server connected to Storage controller. Server can be of type 'Physical Server' or 'Cluster'\n where cluster is logical group of 0-N physical servers.\n Instance ID of a server can be used to map Volume to this particular server. This action makes volume accessible\n as a block device in this server. Mapping volume to 'Cluster' makes volume accessible to all physical servers in\n this cluster.\n \"\"\"\n\n TYPE_PHYSICAL_SERVER = 'ScPhysicalServer'\n TYPE_SERVER_CLUSTER = 'ScServerCluster'\n\n def __init__(self, req_session: Session, base_url: str, name: str,\n instance_id: str, object_type: str) -> None:\n super(Server, self).__init__(req_session=req_session, base_url=base_url, name=name, instance_id=instance_id)\n self.type = object_type\n\n @classmethod\n def from_json(cls, req_session: Session, base_url: str, source_dict: Dict[Any, Any]) -> 'Server':\n \"\"\"\n Class method that creates instance of Server class from supplied dictionary. Source dictionary is expected\n to contain at least 'instanceId', 'name' and 'objectType' keys.\n :param req_session: requests.Session object with stored login cookies. (passed down\n from dell_storage_api.session.DsmSession)\n :param base_url: base URL of DSM\n :param source_dict: Dictionary containing data about Server object\n :return: instance of a Server class\n \"\"\"\n return Server(req_session=req_session,\n base_url=base_url,\n name=source_dict['name'],\n instance_id=source_dict['instanceId'],\n object_type=source_dict['objectType'])\n\n def is_cluster(self) -> bool:\n \"\"\"\n Is this Server object of type 'Cluster'\n :return: True if this Server is Cluster, otherwise False\n \"\"\"\n return bool(self.type == self.TYPE_SERVER_CLUSTER)\n\n def is_physical_server(self) -> bool:\n \"\"\"\n Is this Server object of type 'Physical Server'\n :return: True if this Server is Physical server, otherwise False\n \"\"\"\n return bool(self.type == self.TYPE_PHYSICAL_SERVER)\n\n def pretty_type(self) -> str:\n \"\"\"\n Pretty print type of this server\n :return:\n \"\"\"\n if self.type == self.TYPE_SERVER_CLUSTER:\n result = 'Cluster'\n elif self.type == self.TYPE_PHYSICAL_SERVER:\n result = 'Physical Server'\n else:\n result = \"Unknown Type\"\n return result\n\n\nclass ServerCollection(StorageObjectCollection):\n \"\"\"\n Collection of Server objects\n \"\"\"\n\n def filter_physical_servers(self) -> 'ServerCollection':\n \"\"\"\n Return subset collection containing only servers of type 'Physical Server'\n :return: Subset containing only physical servers\n \"\"\"\n physical_servers = ServerCollection()\n for server in self:\n if server.is_physical_server():\n physical_servers.add(server)\n return physical_servers\n\n def filter_clusters(self) -> 'ServerCollection':\n \"\"\"\n Return subset collection containing only servers of type 'Cluster'\n :return: Subset containing only clusters\n \"\"\"\n clusters = ServerCollection()\n for server in self:\n if server.is_cluster():\n clusters.add(server)\n return clusters\n"},{"alpha_fraction":0.6443299055099487,"alphanum_fraction":0.6520618796348572,"avg_line_length":14.520000457763672,"blob_id":"1727478982528e5f490daac86c478ae868e2ce0f","content_id":"d8a07dfeaa79461dc6f619dd01149e53e9cb7bcb","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":388,"license_type":"no_license","max_line_length":38,"num_lines":25,"path":"/setup.py","repo_name":"SK-CERT/dell-storage-client","src_encoding":"UTF-8","text":"#!/usr/bin/env python3\n\nimport setuptools\n\nname = 'dell_storage_api'\nversion = '0.1'\nauthor = 'SK-CERT'\n\nscripts = ['bin/dell-storage-client',]\nrequires = [\n 'requests',\n 'texttable',\n 'pylint',\n 'mypy',\n]\n\npackages = setuptools.find_packages()\n\nsetuptools.setup(\n name=name,\n version=version,\n install_requires=requires,\n packages=packages,\n scripts=scripts\n)\n"},{"alpha_fraction":0.647972822189331,"alphanum_fraction":0.647972822189331,"avg_line_length":42.35789489746094,"blob_id":"6c23b70775a86a8a99cbb132841e862a5a3847cc","content_id":"23c451afc116cf8ce2f6a5391adefbbcfa71d4c6","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":8238,"license_type":"no_license","max_line_length":118,"num_lines":190,"path":"/dell_storage_api/storage_object.py","repo_name":"SK-CERT/dell-storage-client","src_encoding":"UTF-8","text":"\"\"\"\nThis module contains generic classes that serve as a base for creation of more specific classes that represent objects\nin Dell Storage Manager (DSM) API.\nBase classes:\n - StorageObject: Base for standalone objects (e.g.: Volumes, Servers)\n - StorageObjectFolder: Standalone objects can be grouped into folders in DSM. This object\n represents such folders\n - StorageObjectCollection: Collection of StorageObject instances\n - StorageObjectFolderCollection: Collection of StorageObjectFolder instances\n\"\"\"\nfrom collections import Iterable\nfrom typing import Any\nfrom typing import Optional, Iterator, List, Dict\n\nfrom requests import Session\n\n\nclass StorageObject:\n \"\"\"\n Base class for more specific classes that represent standalone objects in DSM api, such as\n 'Volume' or 'Server'.\n \"\"\"\n\n def __init__(self, req_session: Session, base_url: str, name: str, instance_id: str) -> None:\n self.session = req_session\n self.base_url = base_url\n self.name = name\n self.instance_id = instance_id\n\n def __str__(self) -> str:\n return \"%s: %s (%s)\" % (self.__class__, self.name, self.instance_id)\n\n def build_url(self, endpoint_url: str) -> str:\n \"\"\"\n Build complete URL to the API endpoint by replacing single formatting character (%s) in endpoint_url with\n instance_id of the object and combining it with base url.\n :param endpoint_url: URL of the api endpoint (e.g.: '/StorageCenter/ScVolume/%s/MapToServer')\n :return: Full URL to the api endpoint containing base_url, endpoint_url and instance ID of the storage object\n \"\"\"\n return self.base_url + endpoint_url % self.instance_id\n\n @classmethod\n def from_json(cls, req_session: Session, base_url: str, source_dict: Dict[Any, Any]) -> 'StorageObject':\n \"\"\"\n Class method that creates instance of StorageObject from supplied dictionary. Source dictionary is expected\n to contain at least 'instanceId' and 'name' keys.\n :param req_session: requests.Session object with stored login cookies. (passed down\n from dell_storage_api.session.DsmSession)\n :param base_url: base URL of DSM\n :param source_dict: Dictionary containing data about storage object\n :return: instance of a StorageObject class\n \"\"\"\n return StorageObject(req_session=req_session,\n base_url=base_url,\n instance_id=source_dict['instanceId'],\n name=source_dict['name'])\n\n\nclass StorageObjectFolder(StorageObject):\n \"\"\"\n Base class for more specific classes that represent folders for objects in DSM api, such as\n 'VolumeFolder' or 'ServerFolder'.\n \"\"\"\n\n def __init__(self, req_session: Session, base_url: str, name: str,\n instance_id: str, parent_id: Optional[str]) -> None:\n super().__init__(req_session, base_url, name, instance_id)\n self.parent_id = parent_id\n\n @property\n def is_root(self) -> bool:\n \"\"\"\n Is this folder a root in folder hierarchy?\n :return: True if this folder is root, otherwise False\n \"\"\"\n return self.parent_id is None\n\n @classmethod\n def from_json(cls, req_session: Session, base_url: str, source_dict: Dict[Any, Any]) -> 'StorageObjectFolder':\n \"\"\"\n Class method that creates instance of StorageObjectFolder from supplied dictionary. Source dictionary is\n expected to contain at least 'instanceId', 'name' and 'parent' keys. 'parent' key is optional (it can be\n missing in case the folder is root) but if it's present, it should be dictionary containing at least\n 'instanceId' key.\n :param req_session: requests.Session object with stored login cookies. (passed down\n from dell_storage_api.session.DsmSession)\n :param base_url: base URL of DSM\n :param source_dict: Dictionary containing data about storage object folder\n :return: instance of a StorageObjectFolder class\n \"\"\"\n return StorageObjectFolder(req_session=req_session,\n base_url=base_url,\n instance_id=source_dict['instanceId'],\n name=source_dict['name'],\n parent_id=source_dict.get('parent', {}).get('instanceId', None)\n )\n\n\nclass StorageObjectCollection(Iterable):\n \"\"\"\n Base class representing collection of StorageObject-s. Internally, StorageObjects are stored as dictionary\n indexed by object's instance ID.\n This class provides method for object searching by attributes that are common for every object type in\n DSM and those are 'name' and 'instanceId'.\n \"\"\"\n\n def __iter__(self) -> Iterator:\n return self._store.values().__iter__()\n\n def __bool__(self) -> bool:\n return bool(self._store)\n\n def __len__(self) -> int:\n return len(self._store)\n\n def __init__(self) -> None:\n self._store: Dict[str, StorageObject] = {}\n\n def add(self, storage_object: StorageObject) -> None:\n \"\"\"\n Add storage object to this collection\n :param storage_object: StorageObject to be added\n :return: None\n \"\"\"\n self._store[storage_object.instance_id] = storage_object\n\n def find_by_instance_id(self, instance_id: str) -> Optional[StorageObject]:\n \"\"\"\n Find StorageObject with specified instance_id in this collection. If no object is found, None is returned\n :param instance_id: Instance ID of a storage object to search for\n :return: StorageObject or None\n \"\"\"\n return self._store.get(instance_id, None)\n\n def find_by_name(self, name: str) -> 'StorageObjectCollection':\n \"\"\"\n Find StorageObject with specified name in this collection. Since names do not have to be unique, this method\n will always return StorageObjectCollection containing all the StorageObjects with given name. In case no\n StorageObject is found, returned collection is empty\n :param name: Name of the StorageObject to search for\n :return: StorageObjectCollection containing all the objects with given name\n \"\"\"\n result = StorageObjectCollection()\n for storage_object in self:\n if storage_object.name == name:\n result.add(storage_object)\n return result\n\n def all_objects(self) -> List[StorageObject]:\n \"\"\"\n Return all StorageObjects in this collection as a list\n :return: List of all storage objects in this collection\n \"\"\"\n return [item for item in self._store.values()]\n\n\nclass StorageObjectFolderCollection(StorageObjectCollection):\n \"\"\"\n Base class representing collection of StorageObjectFolder-s. Internally, StorageObjectFolders are stored as\n dictionary indexed by folders's instance ID.\n This class provides methods to search for root folders and to find folders taht belong to the specific\n parent folder.\n \"\"\"\n\n def root_folder(self) -> Optional[StorageObjectFolder]:\n \"\"\"\n Find and return StorageObjectFolder that is a root in folder hierarchy. If this collection does not contain\n root folder, return None\n :return: Root folder from this collection or None\n \"\"\"\n for folder in self:\n if folder.is_root:\n root = folder\n break\n else:\n root = None\n return root\n\n def find_by_parent_id(self, parent_id: str) -> 'StorageObjectFolderCollection':\n \"\"\"\n Create new StorageObjectFolderCollection that is subset of current collection and contains only\n StorageObjectFolders whose direct parent is folder with given parent_id (instance ID of a folder)\n :param parent_id: Instance ID of a parent folder\n :return: Folder collection containing folders with specific parent folder\n \"\"\"\n result = StorageObjectFolderCollection()\n for server_folder in self:\n if server_folder.parent_id == parent_id:\n result.add(server_folder)\n return result\n"},{"alpha_fraction":0.5912607312202454,"alphanum_fraction":0.5938395261764526,"avg_line_length":39.58139419555664,"blob_id":"e36b03ac1d0f2ab35e091d3ad6203db480dbeb02","content_id":"f8497d221e6dbed29ac409c8f74d7657d5375cbe","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":6980,"license_type":"no_license","max_line_length":119,"num_lines":172,"path":"/dell_storage_api/session.py","repo_name":"SK-CERT/dell-storage-client","src_encoding":"UTF-8","text":"\"\"\" This module contains Session for communication with Dell Storage Manager (DSM) API. \"\"\"\nfrom typing import Optional\n\nimport urllib3\nimport requests\nfrom requests.auth import HTTPBasicAuth\nfrom requests.structures import CaseInsensitiveDict\n\nfrom dell_storage_api.storage_center import StorageCenter, StorageCenterCollection\n\n\nclass DsmSession:\n \"\"\"\n This class represents HTTP Session with Dell Storage Manager (DSM). After successful login, underlying\n requests.Session object holds login cookie used to authorize all further requests to DSM API until its expiration.\n DsmSession object holds two important properties, 'base_url' and 'session' which are passed down to child objects\n like Storage Centers, Servers or Volumes. These child elements can then perform their own specific API calls to DSM\n by combining base_url, specific API endpoint and their unique Instance ID to create complete API endpoint URL and\n send requests to this complete endpoint using authenticated session.\n \"\"\"\n API_VERSION_HEADER = 'x-dell-api-verions'\n LOGIN_ENDPOINT = '/ApiConnection/Login'\n LOGOUT_ENDPOINT = '/ApiConnection/Logout'\n STORAGE_CENTER_LIST_ENDPOINT = '/ApiConnection/ApiConnection/%s/StorageCenterList'\n\n def __init__(self, username: str, password: str, host: str, port: int = 3033,\n api_version: str = '3.0', verify_cert: bool = True) -> None:\n self._host = host\n self._port = port\n self._username = username\n self._auth = HTTPBasicAuth(username, password)\n self._api_version = api_version\n self.base_url = 'https://%s:%s/api/rest' % (host, port)\n self.session = requests.Session()\n self.session.headers = CaseInsensitiveDict({'Content-Type': 'application/json',\n 'Accept': 'application/json',\n self.API_VERSION_HEADER: self._api_version})\n self.session.verify = verify_cert\n if not verify_cert:\n # Silence Warning about untrusted certificates if 'verify_cert' is None\n urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)\n self.conn_instance_id = None\n\n @property\n def api_version(self) -> str:\n \"\"\"\n Return API version used by DSM\n :return: DSM API version\n \"\"\"\n return self._api_version\n\n @api_version.setter\n def api_version(self, value: str) -> None:\n \"\"\"\n Set API version used by DSM\n :param value: New DSM API version\n :return: None\n \"\"\"\n self._api_version = value\n self.session.headers[self.API_VERSION_HEADER] = value\n\n @property\n def username(self) -> str:\n \"\"\"\n Return username used to authenticate this session with DSM\n :return: Username used for authentication with DSM\n \"\"\"\n return self._username\n\n @username.setter\n def username(self, value: str) -> None:\n \"\"\"\n Set username used to authenticate this session with DSM\n :param value: New Username\n :return: None\n \"\"\"\n self._username = value\n self._auth.username = value\n\n @property\n def login_url(self) -> str:\n \"\"\"\n Return complete URL to API login endpoint\n :return: URL for login to DSM\n \"\"\"\n return self.base_url + self.LOGIN_ENDPOINT\n\n @property\n def logout_url(self) -> str:\n \"\"\"\n Return complete URL for API logout endpoint\n :return: URL for logout from DSM\n \"\"\"\n return self.base_url + self.LOGOUT_ENDPOINT\n\n @property\n def sc_list_url(self) -> Optional[str]:\n \"\"\"\n Return complete URL for listing Storage Centers managed by this DSM\n :return: URL for Storage Center listing\n \"\"\"\n endpoint = self.STORAGE_CENTER_LIST_ENDPOINT % self.conn_instance_id if self.conn_instance_id else ''\n return (self.base_url + endpoint) if endpoint else None\n\n def set_password(self, password: str) -> None:\n \"\"\"\n Set new password for authentication with DSM\n :param password: New password value\n :return: None\n \"\"\"\n self._auth.password = password\n\n def login(self) -> bool:\n \"\"\"\n Perform call to API login endpoint. If authentication is successful, this method returns boolean value based\n on result of login call.\n :return: True if authentication completed successfully, otherwise False\n \"\"\"\n success = False\n resp = self.session.post(url=self.login_url, auth=self._auth)\n if resp.status_code == 200:\n reported_api_version = resp.json().get('apiVersion', None)\n if reported_api_version:\n self.api_version = reported_api_version\n try:\n self.conn_instance_id = resp.json()['instanceId']\n except KeyError:\n print(\"ERROR: SCM API did not report connection instance ID\")\n else:\n success = True\n else:\n print(\"ERROR: Login failed (%d) - %s\" % (resp.status_code, resp.text))\n return success\n\n def logout(self, silent: bool = False) -> None:\n \"\"\"\n Performs call to api logout endpoint. Status messages about result of logout operation can be silenced by\n specifying silent = True\n :param silent: Whether this method should print result of the logout operation\n :return: None\n \"\"\"\n resp = self.session.post(url=self.logout_url)\n if resp.status_code == 204:\n if not silent:\n print(\"Logout - OK\")\n else:\n if not silent:\n print(\"WARNING: Logout failed (%d) - %s\" % (resp.status_code, resp.text))\n\n def storage_centers(self) -> StorageCenterCollection:\n \"\"\"\n Return collection of storage centers managed by this DSM\n :return:\n \"\"\"\n url = self.sc_list_url\n storage_centers = StorageCenterCollection()\n if url is None:\n print(\"ERROR: Missing Connection ID, try logging in first\")\n else:\n resp = self.session.get(url=url)\n if resp.status_code == 200:\n for storage_center in resp.json():\n storage_centers.add(StorageCenter(req_session=self.session,\n base_url=self.base_url,\n name=storage_center['name'],\n instance_id=storage_center['instanceId'],\n serial_num=storage_center['scSerialNumber'],\n ip_addr=storage_center['hostOrIpAddress']))\n\n else:\n print(\"ERROR: Failed to load Storage Center list (%d) - %s\" % (resp.status_code, resp.text))\n return storage_centers\n"},{"alpha_fraction":0.7161290049552917,"alphanum_fraction":0.7354838848114014,"avg_line_length":16.22222137451172,"blob_id":"157de3ffabc6916f762c6effe424659be4fb4ca2","content_id":"15d693eea00399238d32b24242d4398002e28b6b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"INI","length_bytes":155,"license_type":"no_license","max_line_length":29,"num_lines":9,"path":"/mypy.ini","repo_name":"SK-CERT/dell-storage-client","src_encoding":"UTF-8","text":"[mypy]\npython_version = 3.6\ndisallow_untyped_defs = True\n\n[mypy-urllib3.*]\nignore_missing_imports = True\n\n[mypy-texttable.*]\nignore_missing_imports = True\n"},{"alpha_fraction":0.6545454263687134,"alphanum_fraction":0.7909091114997864,"avg_line_length":12.75,"blob_id":"20da65ffc3e5420e88a5e815c139c30b70fef0eb","content_id":"c8f3d0c535bc141e6bd1d27d3f78e50635f8f31d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"INI","length_bytes":110,"license_type":"no_license","max_line_length":25,"num_lines":8,"path":"/.pylintrc","repo_name":"SK-CERT/dell-storage-client","src_encoding":"UTF-8","text":"[MESSAGE CONTROLL]\ndisable=R1705,R0913,R0902\n\n[FORMAT]\nmax-line-length=120\n\n[REPORTS]\noutput-format=colorized\n"}],"string":"[\n {\n \"alpha_fraction\": 0.6138086915016174,\n \"alphanum_fraction\": 0.6164708733558655,\n \"avg_line_length\": 39.752357482910156,\n \"blob_id\": \"5b0c8f91ee180534215567d6e1dfad476456aaa9\",\n \"content_id\": \"47f1ebd7f8b25adee0e1e722525b36758b61215b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 17279,\n \"license_type\": \"no_license\",\n \"max_line_length\": 120,\n \"num_lines\": 424,\n \"path\": \"/dell_storage_api/volume.py\",\n \"repo_name\": \"SK-CERT/dell-storage-client\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\" This module contains classes for management of volumes in Storage Center managed by Dell Storage Manager\\\"\\\"\\\"\\nfrom typing import Dict, Any\\n\\nfrom requests import Session\\n\\nfrom dell_storage_api.storage_object import StorageObject, StorageObjectCollection, StorageObjectFolder\\n\\n\\nclass Volume(StorageObject):\\n \\\"\\\"\\\"\\n Class that represents volume in Storage Center\\n \\\"\\\"\\\"\\n ENDPOINT = '/StorageCenter/ScVolume'\\n VOLUME_ENDPOINT = '/StorageCenter/ScVolume/%s'\\n MAPPING_ENDPOINT = '/StorageCenter/ScVolume/%s/MapToServer'\\n MAPPING_PROFILE_ENDPOINT = '/StorageCenter/ScVolume/%s/MappingProfileList'\\n UNMAPPING_ENDPOINT = '/StorageCenter/ScVolume/%s/Unmap'\\n RECYCLE_ENDPOINT = '/StorageCenter/ScVolume/%s/Recycle'\\n EXPAND_TO_SIZE_ENDPOINT = '/StorageCenter/ScVolume/%s/ExpandToSize'\\n EXPAND_ENDPOINT = '/StorageCenter/ScVolume/%s/Expand'\\n\\n def __init__(self, req_session: Session, base_url: str, name: str, instance_id: str,\\n parent_folder_id: str, wwid: str, status: str) -> None:\\n super().__init__(req_session, base_url, name, instance_id)\\n self.parent_folder_id = parent_folder_id\\n self.wwid = wwid\\n self.status = status\\n\\n @classmethod\\n def from_json(cls, req_session: Session, base_url: str, source_dict: Dict[Any, Any]) -> 'Volume':\\n \\\"\\\"\\\"\\n Class method that creates instance of Volume class from supplied dictionary. Source dictionary is expected\\n to contain at least 'instanceId', 'name', 'deviceId' and 'volumeFolder' keys. Value of 'volumeFolder' is\\n expected to be of type dict, containing at least key 'instanceId'\\n :param req_session: requests.Session object with stored login cookies. (passed down\\n from dell_storage_api.session.DsmSession)\\n :param base_url: base URL of DSM\\n :param source_dict: Dictionary containing data about Volume object\\n :return: instance of a Volume class\\n \\\"\\\"\\\"\\n return Volume(req_session=req_session,\\n base_url=base_url,\\n name=source_dict['name'],\\n instance_id=source_dict['instanceId'],\\n parent_folder_id=source_dict['volumeFolder']['instanceId'],\\n wwid=source_dict['deviceId'],\\n status=source_dict['status']\\n )\\n\\n @property\\n def mapping_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL for mapping this volume to specific server\\n :return: URL for volume mapping\\n \\\"\\\"\\\"\\n return self.build_url(self.MAPPING_ENDPOINT)\\n\\n @property\\n def unmapping_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL for unmapping this volume from servers to which it is currently mapped\\n :return: URL for volume unmapping\\n \\\"\\\"\\\"\\n return self.build_url(self.UNMAPPING_ENDPOINT)\\n\\n @property\\n def mapping_profile_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL to read mapping profile of this volume.\\n :return: URL for volume mapping list\\n \\\"\\\"\\\"\\n return self.build_url(self.MAPPING_PROFILE_ENDPOINT)\\n\\n @property\\n def recycle_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL for moving this volume to recycle bin\\n :return: URL for moving this volume to recycle bin\\n \\\"\\\"\\\"\\n return self.build_url(self.RECYCLE_ENDPOINT)\\n\\n @property\\n def delete_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL to permanently remove this volume.\\n WARNING: this action can not be undone\\n :return: URL for removing this volume\\n \\\"\\\"\\\"\\n return self.build_url(self.VOLUME_ENDPOINT)\\n\\n @property\\n def expand_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL for expanding this volume's capacity by specified amount.\\n :return: URL for expanding this volume by specified amount\\n \\\"\\\"\\\"\\n return self.build_url(self.EXPAND_ENDPOINT)\\n\\n @property\\n def expand_to_size_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL for expanding this volume's capacity to specified size\\n :return: URL for expanding this volume to specified size\\n \\\"\\\"\\\"\\n return self.build_url(self.EXPAND_TO_SIZE_ENDPOINT)\\n\\n @property\\n def modify_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL for modification of this volume. Modifiable attributes are 'Name' and 'VolumeFolder'\\n :return: URL for modification of this volume\\n \\\"\\\"\\\"\\n return self.build_url(self.VOLUME_ENDPOINT)\\n\\n @property\\n def details_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL for getting details about this volume\\n :return: URL for details about this volume\\n \\\"\\\"\\\"\\n return self.build_url(self.VOLUME_ENDPOINT)\\n\\n def map_to_server(self, server_id: str) -> bool:\\n \\\"\\\"\\\"\\n Perform API call to DSM that maps this volume to server with instance ID specified by parameter 'server_id'. If\\n supplied server_id is instance ID of a cluster, this volume will be mapped to every server that is part of that\\n cluster.\\n This operation fails if volume is already mapped to some server.\\n :param server_id: Instance ID of server to which this volume will be mapped\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n success = False\\n payload = {'Server': server_id}\\n resp = self.session.post(self.mapping_url, json=payload)\\n if resp.status_code == 200:\\n success = True\\n print(\\\"OK - Volume '%s' (%s) sucessfully mapped to server.\\\" % (self.name, self.instance_id))\\n else:\\n print(\\\"Error: Failed to map volume - %s\\\" % resp.json().get('result'))\\n return success\\n\\n def unmap(self) -> bool:\\n \\\"\\\"\\\"\\n Perform API call to DSM that unmaps this volume from any servers it is currently mapped to.\\n WARNING: unmapping volume from active servers will cause those servers to loose connectivity with this volume.\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n success = False\\n resp = self.session.post(self.unmapping_url)\\n if resp.status_code == 204:\\n success = True\\n print('OK - Volume successfully unmapped')\\n else:\\n print('Error: Failed to unamp volume - %s' % resp.text)\\n return success\\n\\n def mapping(self) -> object:\\n \\\"\\\"\\\"\\n \\\"\\\"\\\"\\n success = False\\n resp = self.session.get(self.mapping_profile_url)\\n if resp.status_code == 200:\\n success = True\\n mapping_profiles = resp.json()\\n if mapping_profiles:\\n return mapping_profiles[0]['server']\\n else:\\n print(\\\"Error: Failed to get volume mapping list\\\")\\n return None\\n\\n def expand(self, size: str) -> bool:\\n \\\"\\\"\\\"\\n Perform API call to DSM that expands this volume by specified amount.\\n :param size: Size by which this volume will be expanded (e.g.: 10GB or 1.2TB)\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n success = False\\n payload = {\\\"ExpandAmount\\\": size}\\n resp = self.session.post(self.expand_url, json=payload)\\n if resp.status_code == 200:\\n success = True\\n print(\\\"OK - Volume expanded by %s\\\" % size)\\n else:\\n print(\\\"Error: Failed to expand volume - %s\\\" % resp.json().get('result'))\\n return success\\n\\n def expand_to_size(self, size: str) -> bool:\\n \\\"\\\"\\\"\\n Perform API call to DSM that expands this volume to the specified size. This method can be used only to\\n increase volume size, DSM is unable to shrink volumes\\n :param size: Size to which this volume is expanded (e.g.: 500GB or 2.5TB)\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n success = False\\n payload = {\\\"NewSize\\\": size}\\n resp = self.session.post(self.expand_to_size_url, json=payload)\\n if resp.status_code == 200:\\n success = True\\n print(\\\"OK - Volume expanded to size %s\\\" % size)\\n else:\\n print(\\\"Error: Failed to expand volume - %s\\\" % resp.json().get('result'))\\n return success\\n\\n def recycle(self) -> bool:\\n \\\"\\\"\\\"\\n Perform API call to DSM that moves this volume to recycle bin. Volumes in recycle bin can be restored.\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n success = False\\n resp = self.session.post(self.recycle_url)\\n if resp.status_code == 204:\\n success = True\\n print('OK - Volume successfully moved to recycle bin')\\n else:\\n print('Error: Failed to recycle volume - %s' % resp.text)\\n return success\\n\\n def delete(self) -> bool:\\n \\\"\\\"\\\"\\n Perform API call to DSM that permanently deletes this volume.\\n WARNING: This action can not be undone.\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n success = False\\n resp = self.session.delete(self.delete_url)\\n if resp.status_code == 200:\\n success = True\\n print(\\\"Ok - Volume successfully deleted\\\")\\n else:\\n print(\\\"Error: Failed to delete volume - %s\\\" % resp.json().get('result'))\\n return success\\n\\n def _modify_volume(self, payload: Dict[str, str]) -> bool:\\n \\\"\\\"\\\"\\n Internal method that performs API call to DSM to modify volume properties. Only properties that are modifiable\\n are 'Name' and 'VolumeFolder'.\\n :param payload: Dictionary with modified properties and their new values (e.g.: {'Name': 'new_volume_name'})\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n # TODO: Move common functionality (like modify/rename/move) to base class\\n success = False\\n resp = self.session.put(self.modify_url, json=payload)\\n if resp.status_code == 200:\\n success = True\\n print(\\\"Ok - Volume modified\\\")\\n else:\\n print(\\\"Error: Failed to modify volume\\\")\\n return success\\n\\n def rename(self, new_name: str) -> bool:\\n \\\"\\\"\\\"\\n Perform API call to DSM to change current volume name to the new value.\\n Note: Volume names do not have to be unique even within same folder\\n :param new_name: New name for this volume\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n if self._modify_volume({\\\"Name\\\": new_name}):\\n self.name = new_name\\n return True\\n else:\\n return False\\n\\n def move_to_folder(self, volume_folder_id: str) -> bool:\\n \\\"\\\"\\\"\\n Perform API call to DSM to move this volume to the folder with instance ID specified by parameters\\n volume_folder_id.\\n :param volume_folder_id: Instance ID of a Volume folder to which this volume will be moved\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n if self._modify_volume({\\\"VolumeFolder\\\": volume_folder_id}):\\n self.parent_folder_id = volume_folder_id\\n return True\\n else:\\n return False\\n\\n def details(self) -> Dict[str, Any]:\\n \\\"\\\"\\\"\\n Perform API call to DSM to get all information available about this volume and returns it in\\n form of a dictionary\\n :return: Dictionary containing details about this volume.\\n \\\"\\\"\\\"\\n result: Dict[str, Any] = {}\\n resp = self.session.put(self.details_url)\\n if resp.status_code == 200:\\n result = resp.json()\\n else:\\n print(\\\"Error: Failed to fetch volume details\\\")\\n return result\\n\\n\\nclass VolumeCollection(StorageObjectCollection):\\n \\\"\\\"\\\" Collection of volume folders\\\"\\\"\\\"\\n\\n def find_by_parent_folder(self, folder_id: str) -> 'VolumeCollection':\\n \\\"\\\"\\\"\\n Return subset VolumeCollection that contains only volumes with specified parent folder.\\n :param folder_id: Instance ID of a folder whose children should be in the result\\n :return: Volumes with common parent specified by folder_id\\n \\\"\\\"\\\"\\n result = VolumeCollection()\\n for volume in self:\\n if volume.parent_folder_id == folder_id:\\n result.add(volume)\\n return result\\n\\n\\nclass VolumeFolder(StorageObjectFolder):\\n \\\"\\\"\\\" Class representing Volume Folder\\\"\\\"\\\"\\n ENDPOINT = '/StorageCenter/ScVolumeFolder'\\n VOLUME_FOLDER_ENDPOINT = '/StorageCenter/ScVolumeFolder/%s'\\n\\n @classmethod\\n def from_json(cls, req_session: Session, base_url: str, source_dict: Dict[Any, Any]) -> 'VolumeFolder':\\n \\\"\\\"\\\"\\n Class method that creates instance of VolumeFolder class from supplied dictionary. Source dictionary is expected\\n to contain at least 'instanceId' and 'name' keys. Optional key 'parent' can be present and its value is\\n expected to be dict containing key 'instanceId'.\\n :param req_session: requests.Session object with stored login cookies. (passed down\\n from dell_storage_api.session.DsmSession)\\n :param base_url: base URL of DSM\\n :param source_dict: Dictionary containing data about VolumeFolder object\\n :return: instance of a VolumeFolder class\\n \\\"\\\"\\\"\\n return VolumeFolder(req_session=req_session,\\n base_url=base_url,\\n name=source_dict[\\\"name\\\"],\\n instance_id=source_dict[\\\"instanceId\\\"],\\n parent_id=source_dict.get('parent', {}).get('instanceId', None))\\n\\n @property\\n def modify_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL for modification of this volume folder\\n :return: URL for folder modification\\n \\\"\\\"\\\"\\n return self.build_url(self.VOLUME_FOLDER_ENDPOINT)\\n\\n @property\\n def details_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL for getting details about this volume folder\\n :return: URL for details about this volume folder\\n \\\"\\\"\\\"\\n return self.build_url(self.VOLUME_FOLDER_ENDPOINT)\\n\\n @property\\n def delete_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL to permanently remove this volume Folder.\\n WARNING: this action can not be undone\\n :return: URL for removing this volume folder\\n \\\"\\\"\\\"\\n return self.build_url(self.VOLUME_FOLDER_ENDPOINT)\\n\\n def _modify_volume_folder(self, payload: Dict[str, str]) -> bool:\\n \\\"\\\"\\\"\\n Internal method that performs call to DMS to modify this volume folder. Only modifiable properties are\\n 'Name' and 'Parent'.\\n :param payload: Dictionary with modified properties and their new values (e.g.: {'Name': 'new_folder_name'})\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n success = False\\n resp = self.session.put(self.modify_url, json=payload)\\n if resp.status_code == 200:\\n success = True\\n print(\\\"Ok - Volume folder modified\\\")\\n else:\\n print(\\\"Error: Failed to modify volume folder\\\")\\n return success\\n\\n def rename(self, name: str) -> bool:\\n \\\"\\\"\\\"\\n Perform API call to DSM to change this volume folder name to the new value.\\n Note: Volume Folder names do not have to be unique even within same parent folder\\n :param name: New name for this volume folder\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n if self._modify_volume_folder({\\\"Name\\\": name}):\\n self.name = name\\n return True\\n else:\\n return False\\n\\n def move_to_folder(self, parent_folder_id: str) -> bool:\\n \\\"\\\"\\\"\\n Perform API call to DSM to move this folder to different parent folder specified by Instance ID in\\n 'parent_folder_id' parameter.\\n :param parent_folder_id: Instance ID of the new parent folder\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n if self._modify_volume_folder({\\\"VolumeFolder\\\": parent_folder_id}):\\n self.parent_id = parent_folder_id\\n return True\\n else:\\n return False\\n\\n def details(self) -> Dict[str, Any]:\\n \\\"\\\"\\\"\\n Perform API call to DSM to fetch details about this volume folder. Result is returned as dictionary\\n :return: Dictionary containing details about this volume folder\\n \\\"\\\"\\\"\\n result: Dict[str, Any] = {}\\n resp = self.session.put(self.details_url)\\n if resp.status_code == 200:\\n result = resp.json()\\n else:\\n print(\\\"Error: Failed to fetch volume folder details\\\")\\n return result\\n\\n def delete(self) -> bool:\\n \\\"\\\"\\\"\\n Perform API call to DSM to permanently delete this volume folder.\\n Note: Volume folder can not be removed if it contains volumes.\\n WARNING: This action can not be undone\\n :return: True if operation is successful, otherwise False\\n \\\"\\\"\\\"\\n success = False\\n resp = self.session.delete(self.delete_url)\\n if resp.status_code == 200:\\n success = True\\n print(\\\"Ok - Volume folder successfully deleted\\\")\\n else:\\n print(\\\"Error: Failed to delete volume folder - %s\\\" % resp.json().get('result'))\\n return success\\n\"\n },\n {\n \"alpha_fraction\": 0.6333118677139282,\n \"alphanum_fraction\": 0.6353750228881836,\n \"avg_line_length\": 43.31428527832031,\n \"blob_id\": \"c94848c943ad0353550f780b1532f4296a431234\",\n \"content_id\": \"5157cf487c4446bce91a145f3dc465adf6a75b79\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 23265,\n \"license_type\": \"no_license\",\n \"max_line_length\": 127,\n \"num_lines\": 525,\n \"path\": \"/bin/dell-storage-client\",\n \"repo_name\": \"SK-CERT/dell-storage-client\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python3\\n\\n# pylint: disable=C0103,C0111\\nimport argparse\\nimport getpass\\nimport json\\nfrom typing import Optional, List\\n\\nfrom texttable import Texttable\\n\\nfrom dell_storage_api import DsmSession, StorageCenter\\n\\nCMD_CONST_VOLUME = 'volume'\\nCMD_CONST_VOLUME_CREATE = 'create'\\nCMD_CONST_VOLUME_LIST = 'list'\\nCMD_CONST_VOLUME_MAP = 'map'\\nCMD_CONST_VOLUME_UNMAP = 'unmap'\\n\\nCMD_CONST_VOLUME_FOLDER = 'volume_folder'\\nCMD_CONST_VOLUME_FOLDER_CREATE = 'create'\\nCMD_CONST_VOLUME_FOLDER_LIST = 'list'\\n\\nCMD_CONST_STORAGE_CENTER = 'storage_center'\\nCMD_CONST_STORAGE_CENTER_LIST = 'list'\\n\\nCMD_CONST_SERVER = 'server'\\nCMD_CONST_SERVER_LIST = 'list'\\n\\nclass ReturnCode: # pylint: disable=R0903\\n \\\"\\\"\\\"Convenience class that holds semantic return codes \\\"\\\"\\\"\\n SUCCESS = 0\\n FAILURE = 1\\n\\n\\nclass ServerType:\\n \\\"\\\"\\\"Convenience class that represents various types of Server object\\\"\\\"\\\"\\n\\n def __init__(self) -> None:\\n self._server = 'Server'\\n self._cluster = 'Cluster'\\n self._all = 'All'\\n self._all_types = [self._all, self._server, self._cluster]\\n\\n @property\\n def server(self) -> str:\\n \\\"\\\"\\\"\\n Return string representing Physical server type\\n :return: Physical server type\\n \\\"\\\"\\\"\\n return self._server\\n\\n @property\\n def cluster(self) -> str:\\n \\\"\\\"\\\"\\n Return string representing Custer server type\\n :return: Cluster server type\\n \\\"\\\"\\\"\\n return self._cluster\\n\\n @property\\n def all_keyword(self) -> str:\\n \\\"\\\"\\\"\\n Return list of all possible server type keywords (including magic keyword 'All')\\n :return: list of all server type keywords\\n \\\"\\\"\\\"\\n return self._all\\n\\n @property\\n def all_types(self) -> List[str]:\\n \\\"\\\"\\\"\\n Magic keyword representing all/any server types\\n :return: keyword representing all types\\n \\\"\\\"\\\"\\n return self._all_types\\n\\n\\nSERVER_TYPES = ServerType()\\n\\ndef print_table(table):\\n cli_args = parse_arguments()\\n\\n if cli_args.json:\\n res = []\\n for row in table._rows:\\n line = {}\\n for i in range(len(table._header)):\\n line[table._header[i]] = row[i]\\n res.append(line)\\n text = json.dumps(res)\\n else:\\n text = table.draw()\\n\\n print (text)\\n\\n\\ndef volume_create(storage: StorageCenter, name: str, size: str,\\n unique_name: bool = True, folder_id: str = '', map_to_id: str = '') -> int:\\n \\\"\\\"\\\"\\n Create new volume in Storage Center.\\n :param storage: Storage Center in which new volume will be created\\n :param name: Name of the volume\\n :param size: Size of the volume (e.g.: '500GB' or '1.5TB')\\n :param unique_name: Should the volume creation fail if the valoume name already exists in this folder ?\\n :param folder_id: Folder in which the volume will be created (Defaults to root folder)\\n :param map_to_id: Instance ID of a server (or cluster) to which this volume should be mapped\\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\\n \\\"\\\"\\\"\\n if map_to_id:\\n mapping_server = storage.server_list().find_by_instance_id(map_to_id)\\n if mapping_server is None:\\n print(\\\"Volume can't be mapped to server with instance ID '%s'. No such server\\\" % map_to_id)\\n return ReturnCode.FAILURE\\n if unique_name and storage.volume_list().find_by_name(name):\\n print(\\\"Volume with name '%s' already exists\\\" % name)\\n return ReturnCode.FAILURE\\n\\n new_volume = storage.new_volume(name, size, folder_id)\\n if new_volume is not None:\\n print(\\\"OK - Volume '%s' created with instance ID '%s'\\\" % (new_volume.name, new_volume.instance_id))\\n if map_to_id:\\n if new_volume.map_to_server(map_to_id):\\n return ReturnCode.SUCCESS\\n else:\\n return ReturnCode.FAILURE\\n else:\\n return ReturnCode.SUCCESS\\n else:\\n return ReturnCode.FAILURE\\n\\n\\ndef volume_map(storage: StorageCenter, volume_id: str, server_id: str) -> int:\\n \\\"\\\"\\\"\\n Map existing volume to the server (or cluster).\\n NOTE: Volume that is already mapped to server can not be mapped again without unmapping it first.\\n :param storage: Storage Center where the volume is located\\n :param volume_id: Instance ID of a Volume to be mapped\\n :param server_id: Instance ID of a Server (or a cluster) to which volume will be mapped\\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\\n \\\"\\\"\\\"\\n volume = storage.volume_list().find_by_instance_id(volume_id)\\n if volume is None:\\n print(\\\"Failed to map volume with ID '%s'. No such volume\\\" % volume_id)\\n return ReturnCode.FAILURE\\n server = storage.server_list().find_by_instance_id(server_id)\\n if server is None:\\n print(\\\"failed to map volume '%s' (%s) to server with ID '%s'. No such server\\\" % (volume.name,\\n volume.instance_id,\\n server_id))\\n return ReturnCode.FAILURE\\n if volume.map_to_server(server_id=server_id):\\n return ReturnCode.SUCCESS\\n else:\\n return ReturnCode.FAILURE\\n\\n\\ndef volume_unmap(storage: StorageCenter, volume_id: str) -> int:\\n \\\"\\\"\\\"\\n Unmap volume from all servers it is currently mapped to\\n :param storage: Storage Center where volume is located\\n :param volume_id: Instance ID of a volume to be unampped\\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\\n \\\"\\\"\\\"\\n volume = storage.volume_list().find_by_instance_id(volume_id)\\n if volume is None:\\n print(\\\"Failed to unmap volume with ID '%s'. No such volume\\\" % volume_id)\\n return ReturnCode.FAILURE\\n if volume.unmap():\\n return ReturnCode.SUCCESS\\n else:\\n return ReturnCode.FAILURE\\n\\n\\ndef volume_list(storage: StorageCenter, folder_id: str = '', show_mapping: bool = False) -> int:\\n \\\"\\\"\\\"\\n Print table of Volumes present in Storage Center in specified volume folder.\\n :param storage: Storage Center, from which to list volumes\\n :param folder_id: Volume Folder, from which to list volumes (Defaults to root)\\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\\n \\\"\\\"\\\"\\n table = Texttable(max_width=120)\\n all_volumes = storage.volume_list()\\n if folder_id:\\n all_volumes = all_volumes.find_by_parent_folder(folder_id)\\n\\n if show_mapping:\\n table.header(['volume', 'instance_id', 'parent_folder', 'wwid', 'status', 'mapping'])\\n table.set_cols_dtype(['t', 't', 't', 't','t','t'])\\n else:\\n table.header(['volume', 'instance_id', 'parent_folder', 'wwid', 'status'])\\n table.set_cols_dtype(['t', 't', 't', 't','t'])\\n\\n for volume in all_volumes:\\n if show_mapping:\\n mapping = volume.mapping()\\n mapping_name = mapping['instanceName'] if mapping else None\\n table.add_row([volume.name, volume.instance_id, volume.parent_folder_id, volume.wwid, volume.status, mapping_name])\\n else:\\n table.add_row([volume.name, volume.instance_id, volume.parent_folder_id, volume.wwid, volume.status])\\n\\n print_table(table)\\n\\n return ReturnCode.SUCCESS\\n\\n\\ndef volume_folder_list(storage: StorageCenter, parent_id: str = '') -> int:\\n \\\"\\\"\\\"\\n Print table of Volume Folders present in Storage Center in specified parent Volume Folder\\n :param storage: Storage Center, from which to list volume folders\\n :param parent_id: Parent volume folder, from which the child volume folders will be listed (Defaults to root)\\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\\n \\\"\\\"\\\"\\n # TODO: Adaptable table width\\n table = Texttable(max_width=120)\\n table.header(['folder', 'instance_id', 'parent_instance_id'])\\n table.set_cols_dtype(['t', 't', 't'])\\n folder_list = storage.volume_folder_list()\\n if parent_id:\\n folder_list = folder_list.find_by_parent_id(parent_id)\\n for folder in folder_list:\\n table.add_row([folder.name, folder.instance_id, folder.parent_id])\\n print_table(table)\\n\\n return ReturnCode.SUCCESS\\n\\n\\ndef volume_folder_create(storage: StorageCenter, folder_name: str, folder_parent_id: str = '',\\n unique_name: bool = True) -> int:\\n \\\"\\\"\\\"\\n Create new volume folder in Storage Center\\n :param storage: Storage Center where the new folder will be created\\n :param folder_name: Name of the new folder\\n :param folder_parent_id: Instance ID of a parent folder for the new volume folder\\n :param unique_name: Should the volume creation fail if volume with same name already exists within parent folder\\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\\n \\\"\\\"\\\"\\n if unique_name and storage.volume_folder_list().find_by_name(folder_name):\\n print(\\\"Volume folder with name '%s' already exists\\\" % folder_name)\\n return ReturnCode.FAILURE\\n if storage.new_volume_folder(folder_name, folder_parent_id):\\n return ReturnCode.SUCCESS\\n else:\\n return ReturnCode.FAILURE\\n\\n\\ndef server_list(storage: StorageCenter, object_type: str) -> int:\\n \\\"\\\"\\\"\\n Print table of Servers defined in Storage Center.\\n :param storage: Storage Center from which servers will be listed\\n :param object_type: Limit output only to Servers of specific\\n type (e.g.: SERVER_TYPES.server or SERVER_TYPES.cluster)\\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\\n \\\"\\\"\\\"\\n table = Texttable(max_width=120)\\n table.header(['server', 'type', 'instance_id'])\\n table.set_cols_dtype(['t', 't', 't'])\\n\\n servers = storage.server_list()\\n if object_type == SERVER_TYPES.server:\\n servers = servers.filter_physical_servers()\\n elif object_type == SERVER_TYPES.cluster:\\n servers = servers.filter_clusters()\\n\\n for server in servers:\\n table.add_row([server.name,\\n server.pretty_type(),\\n server.instance_id])\\n print_table(table)\\n return ReturnCode.SUCCESS\\n\\n\\ndef storage_center_list(session: DsmSession) -> int:\\n \\\"\\\"\\\"\\n Print table of Storage Centers connected to Dell Storage Manager\\n :param session: Authenticated session with Dell Storage Manager\\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a operation\\n \\\"\\\"\\\"\\n storage_centers = session.storage_centers()\\n table = Texttable()\\n table.header([\\\"name\\\", \\\"ip\\\", \\\"instance_id\\\", \\\"serial\\\"])\\n table.set_cols_dtype(['t', 't', 't', 't'])\\n\\n for storage_center in storage_centers:\\n table.add_row([storage_center.name,\\n storage_center.ip_addr,\\n storage_center.instance_id,\\n storage_center.serial_num])\\n print_table(table)\\n\\n return ReturnCode.SUCCESS\\n\\n\\ndef _find_storage_center(session: DsmSession, instance_id: str) -> Optional[StorageCenter]:\\n \\\"\\\"\\\"\\n Find and return Storage Center with specified Instance ID connected to Dell Storage manager. If no such\\n Storage Center is found, return None.\\n :param session: Authenticated session with Dell Storage Manager\\n :param instance_id: Instance ID of a Storage Center that should be returned\\n :return: Storage Center with Given Instance ID or None if no such ID is found\\n \\\"\\\"\\\"\\n storage_center = session.storage_centers().find_by_instance_id(instance_id)\\n if storage_center is None:\\n print(\\\"Failed to find storage center with instance ID '%s'. Try listing all storage \\\"\\n \\\"centers with command 'storage_center list'\\\" % instance_id)\\n return None\\n else:\\n return storage_center # type: ignore\\n\\n\\ndef exit_cli(session: DsmSession, return_code: int) -> None:\\n \\\"\\\"\\\"\\n Perform Session logout and exit program\\n :param session: Session with Dell Storage Manager\\n :param return_code: Return Code to exit with\\n :return: None\\n \\\"\\\"\\\"\\n session.logout(silent=True)\\n exit(return_code)\\n\\n\\ndef parse_arguments() -> argparse.Namespace: # pylint: disable=R0914\\n \\\"\\\"\\\"\\n Define and parse CLI commands, subcommands and arguments using argparse module\\n :return: argparse.Namespace with parsed CLI argument values\\n \\\"\\\"\\\"\\n parser = argparse.ArgumentParser()\\n\\n # General options\\n parser.add_argument('-H', '--host', required=True, help=\\\"Hostname or IP address of Dell Storage Center\\\")\\n parser.add_argument('-P', '--port', default=3033, help=\\\"Management port of Dell storage Center\\\")\\n parser.add_argument('-u', '--user', help='Login username')\\n parser.add_argument('-p', '--password', help='Login password')\\n parser.add_argument('-j', '--json', action='store_true', help='Output in JSON format')\\n\\n # Top level subcommands\\n command_parser = parser.add_subparsers(dest='command')\\n\\n # Storage Center subcommands\\n storage_center_parser = command_parser.add_parser(CMD_CONST_STORAGE_CENTER)\\n storage_center_parser_cmd = storage_center_parser.add_subparsers(dest='storage_center_commands')\\n # List Storage centers\\n storage_center_parser_cmd.add_parser(CMD_CONST_STORAGE_CENTER_LIST)\\n\\n # Volume subcommands\\n volume_parser = command_parser.add_parser(CMD_CONST_VOLUME)\\n volume_parser_cmd = volume_parser.add_subparsers(dest='volume_commands')\\n # Create volume\\n volume_create_args = volume_parser_cmd.add_parser(CMD_CONST_VOLUME_CREATE)\\n volume_create_args.add_argument('-s', '--size', required=True, help='Size of the new volume. Example: \\\"500GB\\\"')\\n volume_create_args.add_argument('-n', '--name', required=True, help='Name of the new volume')\\n volume_create_args.add_argument('-f', '--folder-id', dest='folder_id', help='Instance ID of parent folder')\\n volume_create_args.add_argument('-m' '--map-to-server', dest='map_to_server',\\n help='Instance ID of server to which new volume will be mapped')\\n volume_create_args.add_argument('-S', '--storage-id', required=True, dest='storage_id',\\n help='Instance ID of storage center where the volume will be created')\\n volume_create_args.add_argument('-Q', '--non-unique-name', default=False, action='store_true',\\n help='If this flag is present, volume creation wont fail if there is another '\\n 'volume with the same name')\\n # List Volumes\\n volume_list_args = volume_parser_cmd.add_parser(CMD_CONST_VOLUME_LIST)\\n volume_list_args.add_argument('-S', '--storage-id', required=True, dest='storage_id',\\n help='Instance ID of storage center from which the volumes will be listed')\\n volume_list_args.add_argument('-f', '--folder-id', dest='folder_id', help='Instance ID of folder from '\\n 'which the volumes will be listed')\\n volume_list_args.add_argument('-m', '--show-mapping', dest='show_mapping', action=\\\"store_true\\\",\\n help='Show mapping profile (this will slow down things!)')\\n\\n # Map Volume\\n volume_map_args = volume_parser_cmd.add_parser(CMD_CONST_VOLUME_MAP)\\n volume_map_args.add_argument('-S', '--storage-id', required=True, dest='storage_id',\\n help='Instance ID of storage center where volume is located')\\n volume_map_args.add_argument('-v', '--volume-id', dest='volume_id', help='Instance ID of volume to be mapped')\\n volume_map_args.add_argument('-m' '--map-to-server', dest='map_to_server',\\n help='Instance ID of server to which this volume will be mapped')\\n # Unmap Volume\\n volume_unmap_args = volume_parser_cmd.add_parser(CMD_CONST_VOLUME_UNMAP)\\n volume_unmap_args.add_argument('-S', '--storage-id', required=True, dest='storage_id',\\n help='Instance ID of storage center where volume is located')\\n volume_unmap_args.add_argument('-v', '--volume-id', dest='volume_id', help='Instance ID of volume to be unmapped')\\n\\n # Volume Folder subcommands\\n volume_folder_parser = command_parser.add_parser('volume_folder')\\n volume_folder_parser_cmd = volume_folder_parser.add_subparsers(dest='volume_folder_commands')\\n # List volume folders\\n volume_folder_list_args = volume_folder_parser_cmd.add_parser(CMD_CONST_VOLUME_FOLDER_LIST)\\n volume_folder_list_args.add_argument('-S' '--storage-id', required=True, dest='storage_id',\\n help='Instance ID of storage center from which, volume folders will be listed')\\n volume_folder_list_args.add_argument('-f', '--folder-id', dest='folder_id',\\n help='Instance ID of folder from which the child volumes will be listed')\\n\\n # Create volume folder\\n volume_folder_create_args = volume_folder_parser_cmd.add_parser(CMD_CONST_VOLUME_FOLDER_CREATE)\\n volume_folder_create_args.add_argument('-n', '--name', help='Name of the new folder')\\n volume_folder_create_args.add_argument('-f', '--folder-id', dest='folder_id', help='Instance ID of parent folder')\\n volume_folder_create_args.add_argument('-S' '--storage-id', required=True, dest='storage_id',\\n help='Instance ID of storage center from which,'\\n 'volume folders will be listed')\\n volume_folder_create_args.add_argument('-Q', '--non-unique-name', default=False, action='store_true',\\n help='If this flag is present, folder creation wont fail if there is another'\\n ' folder with the same name in specified parent folder')\\n\\n # Server subcommands\\n server_parser = command_parser.add_parser('server')\\n server_parser_cmd = server_parser.add_subparsers(dest='server_commands')\\n # List Servers\\n server_list_args = server_parser_cmd.add_parser(CMD_CONST_SERVER_LIST)\\n server_list_args.add_argument('-t', '--type', default=SERVER_TYPES.all_keyword, choices=SERVER_TYPES.all_types,\\n help='Print only server object of selected type. '\\n '(Default=%s)' % SERVER_TYPES.all_keyword)\\n server_list_args.add_argument('-S' '--storage-id', required=True, dest='storage_id',\\n help='Instance ID of storage center from which, servers will be listed')\\n return parser.parse_args()\\n\\n\\ndef execute_command(args: argparse.Namespace, session: DsmSession) -> int: # pylint: disable=R0912,R0915\\n \\\"\\\"\\\"\\n Execute CLI command based on parsed arguments and parameters\\n :param args: Parsed argparse CLI arguments\\n :param session: Authenticated session with Dell Storage manager\\n :return: ReturnCode.SUCCESS or ReturnCode.FAILURE based on the outcome of a performed command\\n \\\"\\\"\\\"\\n # Volume commands\\n if args.command == CMD_CONST_VOLUME:\\n # Create Volume\\n if args.volume_commands == CMD_CONST_VOLUME_CREATE:\\n unique_name = not args.non_unique_name\\n storage_center = _find_storage_center(session, args.storage_id)\\n if storage_center is None:\\n ret_code = ReturnCode.FAILURE\\n else:\\n ret_code = volume_create(storage=storage_center, name=args.name, size=args.size,\\n unique_name=unique_name, map_to_id=args.map_to_server)\\n # List Volumes\\n elif args.volume_commands == CMD_CONST_VOLUME_LIST:\\n storage_center = _find_storage_center(session, args.storage_id)\\n if storage_center is None:\\n ret_code = ReturnCode.FAILURE\\n else:\\n parent_id = args.folder_id or ''\\n show_mapping = args.show_mapping or False\\n ret_code = volume_list(storage_center, parent_id, show_mapping)\\n elif args.volume_commands == CMD_CONST_VOLUME_MAP:\\n storage_center = _find_storage_center(session, args.storage_id)\\n if storage_center is None:\\n ret_code = ReturnCode.FAILURE\\n else:\\n ret_code = volume_map(storage_center, args.volume_id, args.map_to_server)\\n elif args.volume_commands == CMD_CONST_VOLUME_UNMAP:\\n storage_center = _find_storage_center(session, args.storage_id)\\n if storage_center is None:\\n ret_code = ReturnCode.FAILURE\\n else:\\n ret_code = volume_unmap(storage_center, args.volume_id)\\n # Default branch\\n else:\\n ret_code = ReturnCode.FAILURE\\n\\n # Storage Center commands\\n elif args.command == CMD_CONST_STORAGE_CENTER:\\n # List Storage Centers\\n if args.storage_center_commands == CMD_CONST_STORAGE_CENTER_LIST:\\n ret_code = storage_center_list(session)\\n # Default branch\\n else:\\n ret_code = ReturnCode.FAILURE\\n\\n # Volume Folder commands\\n elif args.command == CMD_CONST_VOLUME_FOLDER:\\n # List Volume folders\\n if args.volume_folder_commands == CMD_CONST_VOLUME_FOLDER_LIST:\\n storage_center = _find_storage_center(session, args.storage_id)\\n if storage_center is None:\\n ret_code = ReturnCode.FAILURE\\n else:\\n parent_folder_id = args.folder_id or ''\\n ret_code = volume_folder_list(storage_center, parent_folder_id)\\n # Create volume folder\\n elif args.volume_folder_commands == CMD_CONST_VOLUME_FOLDER_CREATE:\\n unique_name = not args.non_unique_name\\n storage_center = _find_storage_center(session, args.storage_id)\\n if storage_center is None:\\n ret_code = ReturnCode.FAILURE\\n else:\\n ret_code = volume_folder_create(storage_center, args.name, args.folder_id, unique_name)\\n # Default branch\\n else:\\n ret_code = ReturnCode.FAILURE\\n # Server commands\\n elif args.command == CMD_CONST_SERVER:\\n # List servers\\n if args.server_commands == CMD_CONST_SERVER_LIST:\\n storage_center = _find_storage_center(session, args.storage_id)\\n if storage_center is not None:\\n ret_code = server_list(storage_center, args.type)\\n else:\\n ret_code = ReturnCode.FAILURE\\n else:\\n ret_code = ReturnCode.FAILURE\\n else:\\n ret_code = ReturnCode.FAILURE\\n\\n return ret_code\\n\\n\\ndef main() -> None:\\n # parse CLI arguments\\n cli_args = parse_arguments()\\n\\n # Request missing arguments via CLI dialog\\n if not cli_args.user:\\n cli_args.user = input(\\\"Username: \\\")\\n\\n if not cli_args.password:\\n cli_args.password = getpass.getpass()\\n\\n # Initialize Session with Storage controller\\n scm_session = DsmSession(cli_args.user, cli_args.password, cli_args.host, cli_args.port, verify_cert=False)\\n if not scm_session.login():\\n exit(ReturnCode.SUCCESS)\\n\\n success = execute_command(cli_args, scm_session)\\n exit_cli(scm_session, success)\\n\\n\\nif __name__ == '__main__':\\n main()\\n\"\n },\n {\n \"alpha_fraction\": 0.6169469952583313,\n \"alphanum_fraction\": 0.6184279918670654,\n \"avg_line_length\": 42.96744155883789,\n \"blob_id\": \"5132fe82785279d3a829aaf79dd09d7c91394e9c\",\n \"content_id\": \"aa59ee756b6896087da07d54f2425b0e9878c0e6\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 9453,\n \"license_type\": \"no_license\",\n \"max_line_length\": 117,\n \"num_lines\": 215,\n \"path\": \"/dell_storage_api/storage_center.py\",\n \"repo_name\": \"SK-CERT/dell-storage-client\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\" This module contains classes that represent Storage Centers managed by Dell Storage manager (DSM) \\\"\\\"\\\"\\nfrom typing import Any, Dict, Optional\\n\\nimport requests\\n\\nfrom dell_storage_api.storage_object import StorageObject, StorageObjectFolder, StorageObjectCollection, \\\\\\n StorageObjectFolderCollection\\nfrom dell_storage_api.volume import Volume, VolumeCollection, VolumeFolder\\nfrom dell_storage_api.server import Server, ServerCollection\\n\\n\\nclass StorageCenter(StorageObject):\\n \\\"\\\"\\\"\\n Class representing physical Storage Center managed by DSM\\n \\\"\\\"\\\"\\n SERVER_FOLDER_LIST_ENDPOINT = '/StorageCenter/StorageCenter/%s/ServerFolderList'\\n SERVER_LIST_ENDPOINT = '/StorageCenter/StorageCenter/%s/ServerList'\\n SERVER_FOLDER_ENDPOINT = '/StorageCenter/ScServerFolder/%s'\\n\\n VOLUME_FOLDER_LIST_ENDPOINT = '/StorageCenter/StorageCenter/%s/VolumeFolderList'\\n VOLUME_LIST_ENDPOINT = '/StorageCenter/StorageCenter/%s/VolumeList'\\n\\n def __init__(self, req_session: requests.Session, base_url: str, name: str,\\n instance_id: str, serial_num: str, ip_addr: str) -> None:\\n super(StorageCenter, self).__init__(req_session, base_url, name, instance_id)\\n self.serial_num = serial_num\\n self.ip_addr = ip_addr\\n\\n @property\\n def server_folder_list_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL to list all server folders in this Storage Center\\n :return: URL for Server Folder listing\\n \\\"\\\"\\\"\\n return self.build_url(self.SERVER_FOLDER_LIST_ENDPOINT)\\n\\n @property\\n def server_list_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL to list all servers defined in this Storage Center\\n :return: URL for Server listing\\n \\\"\\\"\\\"\\n return self.build_url(self.SERVER_LIST_ENDPOINT)\\n\\n @property\\n def volume_folder_list_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL to list all volume folders in this Storage Center\\n :return: URL for Volume Folder listing\\n \\\"\\\"\\\"\\n return self.build_url(self.VOLUME_FOLDER_LIST_ENDPOINT)\\n\\n @property\\n def volume_list_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL fo list all volumes present in this Storage Center\\n :return: URL for Volume listning\\n \\\"\\\"\\\"\\n return self.build_url(self.VOLUME_LIST_ENDPOINT)\\n\\n def server_folder_list(self) -> StorageObjectFolderCollection:\\n \\\"\\\"\\\"\\n Return collection of all Server Folders in this Storage Center\\n :return: Collection of all Server Folders\\n \\\"\\\"\\\"\\n return self._fetch_object_list(self.server_folder_list_url)\\n\\n def server_list(self) -> ServerCollection:\\n \\\"\\\"\\\"\\n Return collection of all servers defined in this Storage Center\\n :return: Collection of all Servers\\n \\\"\\\"\\\"\\n result = ServerCollection()\\n for server_data in self._fetch_object_list(self.server_list_url):\\n result.add(Server.from_json(req_session=self.session,\\n base_url=self.base_url,\\n source_dict=server_data))\\n return result\\n\\n def volume_folder_list(self) -> StorageObjectFolderCollection:\\n \\\"\\\"\\\"\\n Return collection of all Volume Folders in this Storage Center\\n :return: Collection of all Volume Folders\\n \\\"\\\"\\\"\\n result = StorageObjectFolderCollection()\\n for volume_folder_data in self._fetch_object_list(url=self.volume_folder_list_url):\\n result.add(VolumeFolder.from_json(req_session=self.session,\\n base_url=self.base_url,\\n source_dict=volume_folder_data))\\n return result\\n\\n def volume_list(self) -> VolumeCollection:\\n \\\"\\\"\\\"\\n Return collection of all volumes present in this Storage Center\\n :return: Collection of all volumes\\n \\\"\\\"\\\"\\n result = VolumeCollection()\\n for volume_data in self._fetch_object_list(url=self.volume_list_url):\\n result.add(Volume.from_json(req_session=self.session,\\n base_url=self.base_url,\\n source_dict=volume_data))\\n return result\\n\\n def _find_volume_folder_root(self) -> Optional[StorageObjectFolder]:\\n \\\"\\\"\\\"\\n Internal method to find root Volume Folder that contains all other Volumes and Volume Folders. Method returns\\n None in case that there is problem with data fetching or lookup.\\n :return: Root Volume Folder or None in case of failure\\n \\\"\\\"\\\"\\n all_folders = self.volume_folder_list()\\n if not all_folders:\\n print(\\\"Error: Failed to fetch volume folder list\\\")\\n return None\\n else:\\n root_folder = self.volume_folder_list().root_folder()\\n if root_folder is None:\\n print(\\\"Error: Failed to lookup root volume folder in list of all folders. \\\"\\n \\\"This really should not happen\\\")\\n return None\\n else:\\n return root_folder\\n\\n def new_volume_folder(self, name: str, parent_folder_id: str = '') -> Optional[VolumeFolder]:\\n \\\"\\\"\\\"\\n Create new Volume Folder in Storage Center and return object representing this new folder. This method\\n returns None in case there is a problem with folder creation.\\n :param name: Name for the new Volume Folder\\n :param parent_folder_id: Instance ID of parent folder. Defaults to root folder\\n :return: new VolumeFolder object or None in case of failure\\n \\\"\\\"\\\"\\n if not parent_folder_id:\\n parent_folder = self._find_volume_folder_root()\\n if parent_folder is not None:\\n parent_folder_id = parent_folder.instance_id\\n else:\\n print(\\\"Error: Failed to create new volume folder\\\")\\n return None\\n if parent_folder_id is None:\\n print(\\\"Error: Failed to create new volume folder\\\")\\n return None\\n url = self.base_url + VolumeFolder.ENDPOINT\\n payload = {\\\"Name\\\": name,\\n \\\"Parent\\\": parent_folder_id,\\n \\\"StorageCenter\\\": self.instance_id}\\n resp = self.session.post(url, json=payload)\\n if resp.status_code == 201:\\n return VolumeFolder.from_json(self.session, self.base_url, resp.json())\\n else:\\n print(\\\"Error: Failed to create new volume folder.\\\")\\n return None\\n\\n def new_volume(self, name: str, size: str, volume_folder_id: str = '') -> Optional[Volume]:\\n \\\"\\\"\\\"\\n Create new Volume in Storage Center and return object representing this new volume. This method returns None\\n in case there is a problem with volume creation.\\n :param name: Name of the new volume\\n :param size: Size of the new volume (e.g: '100GB' or '1.5TB')\\n :param volume_folder_id: Instance ID of folder in which this volume will be created. Defaults to root folder\\n :return: new Volume object or None in case of failure\\n \\\"\\\"\\\"\\n if not volume_folder_id:\\n volume_folder = self._find_volume_folder_root()\\n if volume_folder is not None:\\n volume_folder_id = volume_folder.instance_id\\n else:\\n print(\\\"Error: Failed to create new volume\\\")\\n return None\\n if volume_folder_id is None:\\n print(\\\"Error: Failed to create new volume\\\")\\n return None\\n url = self.base_url + Volume.ENDPOINT\\n payload = {\\\"Name\\\": name,\\n \\\"Size\\\": size,\\n \\\"StorageCenter\\\": self.instance_id,\\n \\\"VolumeFolder\\\": volume_folder_id}\\n resp = self.session.post(url, json=payload)\\n if resp.status_code == 201:\\n return Volume.from_json(self.session, self.base_url, resp.json())\\n else:\\n print(\\\"Error: Failed to create new volume. (%d) - %s\\\" % (resp.status_code, resp.text))\\n return None\\n\\n def _fetch_object_list(self, url: str) -> Dict[Any, Any]:\\n \\\"\\\"\\\"\\n Internal generic method to fetch list of object from supplied URL. This method returns raw dictionary created\\n from json in response body. This method returns empty dictionary if there is problem with data fetching\\n :param url: URL of API endpoint that returns (json) list of objects\\n :return: raw dictionary of objects returned by API endpoint\\n \\\"\\\"\\\"\\n result: Dict[Any, Any] = {}\\n resp = self.session.get(url)\\n if resp.status_code == 200:\\n result = resp.json()\\n else:\\n print(\\\"Error: Failed to fetch object list (%d) - %s\\\" % (resp.status_code, resp.text))\\n return result\\n\\n\\nclass StorageCenterCollection(StorageObjectCollection):\\n \\\"\\\"\\\"\\n Collection representing multiple physical Storage Centers managed by single DSM\\n \\\"\\\"\\\"\\n\\n def find_by_serial_num(self, serial_num: str) -> Optional[StorageCenter]:\\n \\\"\\\"\\\"\\n Return Storage Center object with given serial number. In case there is no such Storage Center, return None\\n :param serial_num: Serial number of physical Storage Center\\n :return: StorageCenter object with given serial number or None\\n \\\"\\\"\\\"\\n result = None\\n for storage_center in self:\\n if storage_center.serial_num == serial_num:\\n result = storage_center\\n break\\n return result\\n\"\n },\n {\n \"alpha_fraction\": 0.8702290058135986,\n \"alphanum_fraction\": 0.8702290058135986,\n \"avg_line_length\": 64.5,\n \"blob_id\": \"1034ecb922e4f3409e8af44a44d510fd584cc141\",\n \"content_id\": \"ab49b16462340df324c28b4516a9d66af5a492ec\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 131,\n \"license_type\": \"no_license\",\n \"max_line_length\": 82,\n \"num_lines\": 2,\n \"path\": \"/dell_storage_api/__init__.py\",\n \"repo_name\": \"SK-CERT/dell-storage-client\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from dell_storage_api.session import DsmSession\\nfrom dell_storage_api.storage_center import StorageCenterCollection, StorageCenter\\n\"\n },\n {\n \"alpha_fraction\": 0.6359223127365112,\n \"alphanum_fraction\": 0.6361920237541199,\n \"avg_line_length\": 37.625,\n \"blob_id\": \"f9e3cf0324a8039bbcab451c9665fcf16f5fafdf\",\n \"content_id\": \"0dbff666ff32d69a76022c0358f2fb131594468a\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3708,\n \"license_type\": \"no_license\",\n \"max_line_length\": 116,\n \"num_lines\": 96,\n \"path\": \"/dell_storage_api/server.py\",\n \"repo_name\": \"SK-CERT/dell-storage-client\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\" This module contains classes representing servers connected to the Storage Center \\\"\\\"\\\"\\nfrom typing import Dict, Any\\n\\nfrom requests import Session\\n\\nfrom dell_storage_api.storage_object import StorageObject, StorageObjectCollection\\n\\n\\nclass Server(StorageObject):\\n \\\"\\\"\\\"\\n Class representing server connected to Storage controller. Server can be of type 'Physical Server' or 'Cluster'\\n where cluster is logical group of 0-N physical servers.\\n Instance ID of a server can be used to map Volume to this particular server. This action makes volume accessible\\n as a block device in this server. Mapping volume to 'Cluster' makes volume accessible to all physical servers in\\n this cluster.\\n \\\"\\\"\\\"\\n\\n TYPE_PHYSICAL_SERVER = 'ScPhysicalServer'\\n TYPE_SERVER_CLUSTER = 'ScServerCluster'\\n\\n def __init__(self, req_session: Session, base_url: str, name: str,\\n instance_id: str, object_type: str) -> None:\\n super(Server, self).__init__(req_session=req_session, base_url=base_url, name=name, instance_id=instance_id)\\n self.type = object_type\\n\\n @classmethod\\n def from_json(cls, req_session: Session, base_url: str, source_dict: Dict[Any, Any]) -> 'Server':\\n \\\"\\\"\\\"\\n Class method that creates instance of Server class from supplied dictionary. Source dictionary is expected\\n to contain at least 'instanceId', 'name' and 'objectType' keys.\\n :param req_session: requests.Session object with stored login cookies. (passed down\\n from dell_storage_api.session.DsmSession)\\n :param base_url: base URL of DSM\\n :param source_dict: Dictionary containing data about Server object\\n :return: instance of a Server class\\n \\\"\\\"\\\"\\n return Server(req_session=req_session,\\n base_url=base_url,\\n name=source_dict['name'],\\n instance_id=source_dict['instanceId'],\\n object_type=source_dict['objectType'])\\n\\n def is_cluster(self) -> bool:\\n \\\"\\\"\\\"\\n Is this Server object of type 'Cluster'\\n :return: True if this Server is Cluster, otherwise False\\n \\\"\\\"\\\"\\n return bool(self.type == self.TYPE_SERVER_CLUSTER)\\n\\n def is_physical_server(self) -> bool:\\n \\\"\\\"\\\"\\n Is this Server object of type 'Physical Server'\\n :return: True if this Server is Physical server, otherwise False\\n \\\"\\\"\\\"\\n return bool(self.type == self.TYPE_PHYSICAL_SERVER)\\n\\n def pretty_type(self) -> str:\\n \\\"\\\"\\\"\\n Pretty print type of this server\\n :return:\\n \\\"\\\"\\\"\\n if self.type == self.TYPE_SERVER_CLUSTER:\\n result = 'Cluster'\\n elif self.type == self.TYPE_PHYSICAL_SERVER:\\n result = 'Physical Server'\\n else:\\n result = \\\"Unknown Type\\\"\\n return result\\n\\n\\nclass ServerCollection(StorageObjectCollection):\\n \\\"\\\"\\\"\\n Collection of Server objects\\n \\\"\\\"\\\"\\n\\n def filter_physical_servers(self) -> 'ServerCollection':\\n \\\"\\\"\\\"\\n Return subset collection containing only servers of type 'Physical Server'\\n :return: Subset containing only physical servers\\n \\\"\\\"\\\"\\n physical_servers = ServerCollection()\\n for server in self:\\n if server.is_physical_server():\\n physical_servers.add(server)\\n return physical_servers\\n\\n def filter_clusters(self) -> 'ServerCollection':\\n \\\"\\\"\\\"\\n Return subset collection containing only servers of type 'Cluster'\\n :return: Subset containing only clusters\\n \\\"\\\"\\\"\\n clusters = ServerCollection()\\n for server in self:\\n if server.is_cluster():\\n clusters.add(server)\\n return clusters\\n\"\n },\n {\n \"alpha_fraction\": 0.6443299055099487,\n \"alphanum_fraction\": 0.6520618796348572,\n \"avg_line_length\": 14.520000457763672,\n \"blob_id\": \"1727478982528e5f490daac86c478ae868e2ce0f\",\n \"content_id\": \"d8a07dfeaa79461dc6f619dd01149e53e9cb7bcb\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 388,\n \"license_type\": \"no_license\",\n \"max_line_length\": 38,\n \"num_lines\": 25,\n \"path\": \"/setup.py\",\n \"repo_name\": \"SK-CERT/dell-storage-client\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python3\\n\\nimport setuptools\\n\\nname = 'dell_storage_api'\\nversion = '0.1'\\nauthor = 'SK-CERT'\\n\\nscripts = ['bin/dell-storage-client',]\\nrequires = [\\n 'requests',\\n 'texttable',\\n 'pylint',\\n 'mypy',\\n]\\n\\npackages = setuptools.find_packages()\\n\\nsetuptools.setup(\\n name=name,\\n version=version,\\n install_requires=requires,\\n packages=packages,\\n scripts=scripts\\n)\\n\"\n },\n {\n \"alpha_fraction\": 0.647972822189331,\n \"alphanum_fraction\": 0.647972822189331,\n \"avg_line_length\": 42.35789489746094,\n \"blob_id\": \"6c23b70775a86a8a99cbb132841e862a5a3847cc\",\n \"content_id\": \"23c451afc116cf8ce2f6a5391adefbbcfa71d4c6\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 8238,\n \"license_type\": \"no_license\",\n \"max_line_length\": 118,\n \"num_lines\": 190,\n \"path\": \"/dell_storage_api/storage_object.py\",\n \"repo_name\": \"SK-CERT/dell-storage-client\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\"\\nThis module contains generic classes that serve as a base for creation of more specific classes that represent objects\\nin Dell Storage Manager (DSM) API.\\nBase classes:\\n - StorageObject: Base for standalone objects (e.g.: Volumes, Servers)\\n - StorageObjectFolder: Standalone objects can be grouped into folders in DSM. This object\\n represents such folders\\n - StorageObjectCollection: Collection of StorageObject instances\\n - StorageObjectFolderCollection: Collection of StorageObjectFolder instances\\n\\\"\\\"\\\"\\nfrom collections import Iterable\\nfrom typing import Any\\nfrom typing import Optional, Iterator, List, Dict\\n\\nfrom requests import Session\\n\\n\\nclass StorageObject:\\n \\\"\\\"\\\"\\n Base class for more specific classes that represent standalone objects in DSM api, such as\\n 'Volume' or 'Server'.\\n \\\"\\\"\\\"\\n\\n def __init__(self, req_session: Session, base_url: str, name: str, instance_id: str) -> None:\\n self.session = req_session\\n self.base_url = base_url\\n self.name = name\\n self.instance_id = instance_id\\n\\n def __str__(self) -> str:\\n return \\\"%s: %s (%s)\\\" % (self.__class__, self.name, self.instance_id)\\n\\n def build_url(self, endpoint_url: str) -> str:\\n \\\"\\\"\\\"\\n Build complete URL to the API endpoint by replacing single formatting character (%s) in endpoint_url with\\n instance_id of the object and combining it with base url.\\n :param endpoint_url: URL of the api endpoint (e.g.: '/StorageCenter/ScVolume/%s/MapToServer')\\n :return: Full URL to the api endpoint containing base_url, endpoint_url and instance ID of the storage object\\n \\\"\\\"\\\"\\n return self.base_url + endpoint_url % self.instance_id\\n\\n @classmethod\\n def from_json(cls, req_session: Session, base_url: str, source_dict: Dict[Any, Any]) -> 'StorageObject':\\n \\\"\\\"\\\"\\n Class method that creates instance of StorageObject from supplied dictionary. Source dictionary is expected\\n to contain at least 'instanceId' and 'name' keys.\\n :param req_session: requests.Session object with stored login cookies. (passed down\\n from dell_storage_api.session.DsmSession)\\n :param base_url: base URL of DSM\\n :param source_dict: Dictionary containing data about storage object\\n :return: instance of a StorageObject class\\n \\\"\\\"\\\"\\n return StorageObject(req_session=req_session,\\n base_url=base_url,\\n instance_id=source_dict['instanceId'],\\n name=source_dict['name'])\\n\\n\\nclass StorageObjectFolder(StorageObject):\\n \\\"\\\"\\\"\\n Base class for more specific classes that represent folders for objects in DSM api, such as\\n 'VolumeFolder' or 'ServerFolder'.\\n \\\"\\\"\\\"\\n\\n def __init__(self, req_session: Session, base_url: str, name: str,\\n instance_id: str, parent_id: Optional[str]) -> None:\\n super().__init__(req_session, base_url, name, instance_id)\\n self.parent_id = parent_id\\n\\n @property\\n def is_root(self) -> bool:\\n \\\"\\\"\\\"\\n Is this folder a root in folder hierarchy?\\n :return: True if this folder is root, otherwise False\\n \\\"\\\"\\\"\\n return self.parent_id is None\\n\\n @classmethod\\n def from_json(cls, req_session: Session, base_url: str, source_dict: Dict[Any, Any]) -> 'StorageObjectFolder':\\n \\\"\\\"\\\"\\n Class method that creates instance of StorageObjectFolder from supplied dictionary. Source dictionary is\\n expected to contain at least 'instanceId', 'name' and 'parent' keys. 'parent' key is optional (it can be\\n missing in case the folder is root) but if it's present, it should be dictionary containing at least\\n 'instanceId' key.\\n :param req_session: requests.Session object with stored login cookies. (passed down\\n from dell_storage_api.session.DsmSession)\\n :param base_url: base URL of DSM\\n :param source_dict: Dictionary containing data about storage object folder\\n :return: instance of a StorageObjectFolder class\\n \\\"\\\"\\\"\\n return StorageObjectFolder(req_session=req_session,\\n base_url=base_url,\\n instance_id=source_dict['instanceId'],\\n name=source_dict['name'],\\n parent_id=source_dict.get('parent', {}).get('instanceId', None)\\n )\\n\\n\\nclass StorageObjectCollection(Iterable):\\n \\\"\\\"\\\"\\n Base class representing collection of StorageObject-s. Internally, StorageObjects are stored as dictionary\\n indexed by object's instance ID.\\n This class provides method for object searching by attributes that are common for every object type in\\n DSM and those are 'name' and 'instanceId'.\\n \\\"\\\"\\\"\\n\\n def __iter__(self) -> Iterator:\\n return self._store.values().__iter__()\\n\\n def __bool__(self) -> bool:\\n return bool(self._store)\\n\\n def __len__(self) -> int:\\n return len(self._store)\\n\\n def __init__(self) -> None:\\n self._store: Dict[str, StorageObject] = {}\\n\\n def add(self, storage_object: StorageObject) -> None:\\n \\\"\\\"\\\"\\n Add storage object to this collection\\n :param storage_object: StorageObject to be added\\n :return: None\\n \\\"\\\"\\\"\\n self._store[storage_object.instance_id] = storage_object\\n\\n def find_by_instance_id(self, instance_id: str) -> Optional[StorageObject]:\\n \\\"\\\"\\\"\\n Find StorageObject with specified instance_id in this collection. If no object is found, None is returned\\n :param instance_id: Instance ID of a storage object to search for\\n :return: StorageObject or None\\n \\\"\\\"\\\"\\n return self._store.get(instance_id, None)\\n\\n def find_by_name(self, name: str) -> 'StorageObjectCollection':\\n \\\"\\\"\\\"\\n Find StorageObject with specified name in this collection. Since names do not have to be unique, this method\\n will always return StorageObjectCollection containing all the StorageObjects with given name. In case no\\n StorageObject is found, returned collection is empty\\n :param name: Name of the StorageObject to search for\\n :return: StorageObjectCollection containing all the objects with given name\\n \\\"\\\"\\\"\\n result = StorageObjectCollection()\\n for storage_object in self:\\n if storage_object.name == name:\\n result.add(storage_object)\\n return result\\n\\n def all_objects(self) -> List[StorageObject]:\\n \\\"\\\"\\\"\\n Return all StorageObjects in this collection as a list\\n :return: List of all storage objects in this collection\\n \\\"\\\"\\\"\\n return [item for item in self._store.values()]\\n\\n\\nclass StorageObjectFolderCollection(StorageObjectCollection):\\n \\\"\\\"\\\"\\n Base class representing collection of StorageObjectFolder-s. Internally, StorageObjectFolders are stored as\\n dictionary indexed by folders's instance ID.\\n This class provides methods to search for root folders and to find folders taht belong to the specific\\n parent folder.\\n \\\"\\\"\\\"\\n\\n def root_folder(self) -> Optional[StorageObjectFolder]:\\n \\\"\\\"\\\"\\n Find and return StorageObjectFolder that is a root in folder hierarchy. If this collection does not contain\\n root folder, return None\\n :return: Root folder from this collection or None\\n \\\"\\\"\\\"\\n for folder in self:\\n if folder.is_root:\\n root = folder\\n break\\n else:\\n root = None\\n return root\\n\\n def find_by_parent_id(self, parent_id: str) -> 'StorageObjectFolderCollection':\\n \\\"\\\"\\\"\\n Create new StorageObjectFolderCollection that is subset of current collection and contains only\\n StorageObjectFolders whose direct parent is folder with given parent_id (instance ID of a folder)\\n :param parent_id: Instance ID of a parent folder\\n :return: Folder collection containing folders with specific parent folder\\n \\\"\\\"\\\"\\n result = StorageObjectFolderCollection()\\n for server_folder in self:\\n if server_folder.parent_id == parent_id:\\n result.add(server_folder)\\n return result\\n\"\n },\n {\n \"alpha_fraction\": 0.5912607312202454,\n \"alphanum_fraction\": 0.5938395261764526,\n \"avg_line_length\": 39.58139419555664,\n \"blob_id\": \"e36b03ac1d0f2ab35e091d3ad6203db480dbeb02\",\n \"content_id\": \"f8497d221e6dbed29ac409c8f74d7657d5375cbe\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 6980,\n \"license_type\": \"no_license\",\n \"max_line_length\": 119,\n \"num_lines\": 172,\n \"path\": \"/dell_storage_api/session.py\",\n \"repo_name\": \"SK-CERT/dell-storage-client\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\" This module contains Session for communication with Dell Storage Manager (DSM) API. \\\"\\\"\\\"\\nfrom typing import Optional\\n\\nimport urllib3\\nimport requests\\nfrom requests.auth import HTTPBasicAuth\\nfrom requests.structures import CaseInsensitiveDict\\n\\nfrom dell_storage_api.storage_center import StorageCenter, StorageCenterCollection\\n\\n\\nclass DsmSession:\\n \\\"\\\"\\\"\\n This class represents HTTP Session with Dell Storage Manager (DSM). After successful login, underlying\\n requests.Session object holds login cookie used to authorize all further requests to DSM API until its expiration.\\n DsmSession object holds two important properties, 'base_url' and 'session' which are passed down to child objects\\n like Storage Centers, Servers or Volumes. These child elements can then perform their own specific API calls to DSM\\n by combining base_url, specific API endpoint and their unique Instance ID to create complete API endpoint URL and\\n send requests to this complete endpoint using authenticated session.\\n \\\"\\\"\\\"\\n API_VERSION_HEADER = 'x-dell-api-verions'\\n LOGIN_ENDPOINT = '/ApiConnection/Login'\\n LOGOUT_ENDPOINT = '/ApiConnection/Logout'\\n STORAGE_CENTER_LIST_ENDPOINT = '/ApiConnection/ApiConnection/%s/StorageCenterList'\\n\\n def __init__(self, username: str, password: str, host: str, port: int = 3033,\\n api_version: str = '3.0', verify_cert: bool = True) -> None:\\n self._host = host\\n self._port = port\\n self._username = username\\n self._auth = HTTPBasicAuth(username, password)\\n self._api_version = api_version\\n self.base_url = 'https://%s:%s/api/rest' % (host, port)\\n self.session = requests.Session()\\n self.session.headers = CaseInsensitiveDict({'Content-Type': 'application/json',\\n 'Accept': 'application/json',\\n self.API_VERSION_HEADER: self._api_version})\\n self.session.verify = verify_cert\\n if not verify_cert:\\n # Silence Warning about untrusted certificates if 'verify_cert' is None\\n urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)\\n self.conn_instance_id = None\\n\\n @property\\n def api_version(self) -> str:\\n \\\"\\\"\\\"\\n Return API version used by DSM\\n :return: DSM API version\\n \\\"\\\"\\\"\\n return self._api_version\\n\\n @api_version.setter\\n def api_version(self, value: str) -> None:\\n \\\"\\\"\\\"\\n Set API version used by DSM\\n :param value: New DSM API version\\n :return: None\\n \\\"\\\"\\\"\\n self._api_version = value\\n self.session.headers[self.API_VERSION_HEADER] = value\\n\\n @property\\n def username(self) -> str:\\n \\\"\\\"\\\"\\n Return username used to authenticate this session with DSM\\n :return: Username used for authentication with DSM\\n \\\"\\\"\\\"\\n return self._username\\n\\n @username.setter\\n def username(self, value: str) -> None:\\n \\\"\\\"\\\"\\n Set username used to authenticate this session with DSM\\n :param value: New Username\\n :return: None\\n \\\"\\\"\\\"\\n self._username = value\\n self._auth.username = value\\n\\n @property\\n def login_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL to API login endpoint\\n :return: URL for login to DSM\\n \\\"\\\"\\\"\\n return self.base_url + self.LOGIN_ENDPOINT\\n\\n @property\\n def logout_url(self) -> str:\\n \\\"\\\"\\\"\\n Return complete URL for API logout endpoint\\n :return: URL for logout from DSM\\n \\\"\\\"\\\"\\n return self.base_url + self.LOGOUT_ENDPOINT\\n\\n @property\\n def sc_list_url(self) -> Optional[str]:\\n \\\"\\\"\\\"\\n Return complete URL for listing Storage Centers managed by this DSM\\n :return: URL for Storage Center listing\\n \\\"\\\"\\\"\\n endpoint = self.STORAGE_CENTER_LIST_ENDPOINT % self.conn_instance_id if self.conn_instance_id else ''\\n return (self.base_url + endpoint) if endpoint else None\\n\\n def set_password(self, password: str) -> None:\\n \\\"\\\"\\\"\\n Set new password for authentication with DSM\\n :param password: New password value\\n :return: None\\n \\\"\\\"\\\"\\n self._auth.password = password\\n\\n def login(self) -> bool:\\n \\\"\\\"\\\"\\n Perform call to API login endpoint. If authentication is successful, this method returns boolean value based\\n on result of login call.\\n :return: True if authentication completed successfully, otherwise False\\n \\\"\\\"\\\"\\n success = False\\n resp = self.session.post(url=self.login_url, auth=self._auth)\\n if resp.status_code == 200:\\n reported_api_version = resp.json().get('apiVersion', None)\\n if reported_api_version:\\n self.api_version = reported_api_version\\n try:\\n self.conn_instance_id = resp.json()['instanceId']\\n except KeyError:\\n print(\\\"ERROR: SCM API did not report connection instance ID\\\")\\n else:\\n success = True\\n else:\\n print(\\\"ERROR: Login failed (%d) - %s\\\" % (resp.status_code, resp.text))\\n return success\\n\\n def logout(self, silent: bool = False) -> None:\\n \\\"\\\"\\\"\\n Performs call to api logout endpoint. Status messages about result of logout operation can be silenced by\\n specifying silent = True\\n :param silent: Whether this method should print result of the logout operation\\n :return: None\\n \\\"\\\"\\\"\\n resp = self.session.post(url=self.logout_url)\\n if resp.status_code == 204:\\n if not silent:\\n print(\\\"Logout - OK\\\")\\n else:\\n if not silent:\\n print(\\\"WARNING: Logout failed (%d) - %s\\\" % (resp.status_code, resp.text))\\n\\n def storage_centers(self) -> StorageCenterCollection:\\n \\\"\\\"\\\"\\n Return collection of storage centers managed by this DSM\\n :return:\\n \\\"\\\"\\\"\\n url = self.sc_list_url\\n storage_centers = StorageCenterCollection()\\n if url is None:\\n print(\\\"ERROR: Missing Connection ID, try logging in first\\\")\\n else:\\n resp = self.session.get(url=url)\\n if resp.status_code == 200:\\n for storage_center in resp.json():\\n storage_centers.add(StorageCenter(req_session=self.session,\\n base_url=self.base_url,\\n name=storage_center['name'],\\n instance_id=storage_center['instanceId'],\\n serial_num=storage_center['scSerialNumber'],\\n ip_addr=storage_center['hostOrIpAddress']))\\n\\n else:\\n print(\\\"ERROR: Failed to load Storage Center list (%d) - %s\\\" % (resp.status_code, resp.text))\\n return storage_centers\\n\"\n },\n {\n \"alpha_fraction\": 0.7161290049552917,\n \"alphanum_fraction\": 0.7354838848114014,\n \"avg_line_length\": 16.22222137451172,\n \"blob_id\": \"157de3ffabc6916f762c6effe424659be4fb4ca2\",\n \"content_id\": \"15d693eea00399238d32b24242d4398002e28b6b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"INI\",\n \"length_bytes\": 155,\n \"license_type\": \"no_license\",\n \"max_line_length\": 29,\n \"num_lines\": 9,\n \"path\": \"/mypy.ini\",\n \"repo_name\": \"SK-CERT/dell-storage-client\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"[mypy]\\npython_version = 3.6\\ndisallow_untyped_defs = True\\n\\n[mypy-urllib3.*]\\nignore_missing_imports = True\\n\\n[mypy-texttable.*]\\nignore_missing_imports = True\\n\"\n },\n {\n \"alpha_fraction\": 0.6545454263687134,\n \"alphanum_fraction\": 0.7909091114997864,\n \"avg_line_length\": 12.75,\n \"blob_id\": \"20da65ffc3e5420e88a5e815c139c30b70fef0eb\",\n \"content_id\": \"c8f3d0c535bc141e6bd1d27d3f78e50635f8f31d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"INI\",\n \"length_bytes\": 110,\n \"license_type\": \"no_license\",\n \"max_line_length\": 25,\n \"num_lines\": 8,\n \"path\": \"/.pylintrc\",\n \"repo_name\": \"SK-CERT/dell-storage-client\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"[MESSAGE CONTROLL]\\ndisable=R1705,R0913,R0902\\n\\n[FORMAT]\\nmax-line-length=120\\n\\n[REPORTS]\\noutput-format=colorized\\n\"\n }\n]"},"num_files":{"kind":"number","value":10,"string":"10"}}},{"rowIdx":2980392,"cells":{"repo_name":{"kind":"string","value":"eduardotrein/leonesto"},"repo_url":{"kind":"string","value":"https://github.com/eduardotrein/leonesto"},"snapshot_id":{"kind":"string","value":"669d24c3e71c6b827efec3d61cac06149c5ab390"},"revision_id":{"kind":"string","value":"ffb26d61a4d2218598246e6b81e881e41591b6f6"},"directory_id":{"kind":"string","value":"b8e8c5aff0219b800aab6b7538b57fb627ba489f"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2020-09-15T16:39:42.837649","string":"2020-09-15T16:39:42.837649"},"revision_date":{"kind":"timestamp","value":"2020-02-15T14:09:01","string":"2020-02-15T14:09:01"},"committer_date":{"kind":"timestamp","value":"2020-02-15T14:09:01","string":"2020-02-15T14:09:01"},"github_id":{"kind":"number","value":223505509,"string":"223,505,509"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"timestamp","value":"2019-11-22T23:57:01","string":"2019-11-22T23:57:01"},"gha_updated_at":{"kind":"timestamp","value":"2019-11-23T00:14:22","string":"2019-11-23T00:14:22"},"gha_pushed_at":{"kind":"timestamp","value":"2019-11-23T00:25:35","string":"2019-11-23T00:25:35"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.7779527306556702,"alphanum_fraction":0.7779527306556702,"avg_line_length":27.863636016845703,"blob_id":"8da6e6caeb7b19f8923ef08a99bb0f18470dd435","content_id":"201c13aac0c3039643052db8d9bf1543734b924c","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":637,"license_type":"no_license","max_line_length":66,"num_lines":22,"path":"/README.md","repo_name":"eduardotrein/leonesto","src_encoding":"UTF-8","text":"# leonesto\nInstruções\n\ncd /home/eduardo/Documentos/Estudo/git\n\ngit add README.md \ngit commit -m \"first commit\"\ngit config --global user.email \"leo.trein@gmail.com\"\ngit config --global user.name \"eduardotrein\"\ngit commit -m \"first commit\"\ngit remote add origin https://github.com/eduardotrein/leonesto.git\ngit push -u origin master\ngit add README.md \ngit commit -m \"Segundo Commit, com uma linha a mais\"\ngit remote add origin https://github.com/eduardotrein/leonesto.git\ngit push -u origin teste\ngit push -u origin master\n\n\n# Salvar senha para um repositorio\ngit config credential.helper store\ngit push https://github.com/eduardotrein/leonesto.git\n"},{"alpha_fraction":0.5714285969734192,"alphanum_fraction":0.6000000238418579,"avg_line_length":34,"blob_id":"9d627d7c7e9b6025ccaa38bd1576b8b5b074a2d8","content_id":"019279105d9bb565409b5b14d413fb25c87100ce","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":35,"license_type":"no_license","max_line_length":34,"num_lines":1,"path":"/python/hello_world.py","repo_name":"eduardotrein/leonesto","src_encoding":"UTF-8","text":"print( \"hello world!!!\",\"teste 5\")\n"}],"string":"[\n {\n \"alpha_fraction\": 0.7779527306556702,\n \"alphanum_fraction\": 0.7779527306556702,\n \"avg_line_length\": 27.863636016845703,\n \"blob_id\": \"8da6e6caeb7b19f8923ef08a99bb0f18470dd435\",\n \"content_id\": \"201c13aac0c3039643052db8d9bf1543734b924c\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 637,\n \"license_type\": \"no_license\",\n \"max_line_length\": 66,\n \"num_lines\": 22,\n \"path\": \"/README.md\",\n \"repo_name\": \"eduardotrein/leonesto\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# leonesto\\nInstruções\\n\\ncd /home/eduardo/Documentos/Estudo/git\\n\\ngit add README.md \\ngit commit -m \\\"first commit\\\"\\ngit config --global user.email \\\"leo.trein@gmail.com\\\"\\ngit config --global user.name \\\"eduardotrein\\\"\\ngit commit -m \\\"first commit\\\"\\ngit remote add origin https://github.com/eduardotrein/leonesto.git\\ngit push -u origin master\\ngit add README.md \\ngit commit -m \\\"Segundo Commit, com uma linha a mais\\\"\\ngit remote add origin https://github.com/eduardotrein/leonesto.git\\ngit push -u origin teste\\ngit push -u origin master\\n\\n\\n# Salvar senha para um repositorio\\ngit config credential.helper store\\ngit push https://github.com/eduardotrein/leonesto.git\\n\"\n },\n {\n \"alpha_fraction\": 0.5714285969734192,\n \"alphanum_fraction\": 0.6000000238418579,\n \"avg_line_length\": 34,\n \"blob_id\": \"9d627d7c7e9b6025ccaa38bd1576b8b5b074a2d8\",\n \"content_id\": \"019279105d9bb565409b5b14d413fb25c87100ce\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 35,\n \"license_type\": \"no_license\",\n \"max_line_length\": 34,\n \"num_lines\": 1,\n \"path\": \"/python/hello_world.py\",\n \"repo_name\": \"eduardotrein/leonesto\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"print( \\\"hello world!!!\\\",\\\"teste 5\\\")\\n\"\n }\n]"},"num_files":{"kind":"number","value":2,"string":"2"}}},{"rowIdx":2980393,"cells":{"repo_name":{"kind":"string","value":"pranavbhatt/py-practice"},"repo_url":{"kind":"string","value":"https://github.com/pranavbhatt/py-practice"},"snapshot_id":{"kind":"string","value":"ae9c2112d79677731834a56447dbb899a3916ce3"},"revision_id":{"kind":"string","value":"17865ef8c68b633d2fc5d12adcb0d7e93750ecb3"},"directory_id":{"kind":"string","value":"b1be4c4be336c98c20f77ce8e3ebc64b83ebf66c"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2022-07-27T18:12:43.153545","string":"2022-07-27T18:12:43.153545"},"revision_date":{"kind":"timestamp","value":"2018-04-05T22:54:55","string":"2018-04-05T22:54:55"},"committer_date":{"kind":"timestamp","value":"2018-04-05T22:54:55","string":"2018-04-05T22:54:55"},"github_id":{"kind":"number","value":124005405,"string":"124,005,405"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"timestamp","value":"2018-03-06T01:59:34","string":"2018-03-06T01:59:34"},"gha_updated_at":{"kind":"timestamp","value":"2018-04-05T22:55:06","string":"2018-04-05T22:55:06"},"gha_pushed_at":{"kind":"timestamp","value":"2022-07-06T19:45:21","string":"2022-07-06T19:45:21"},"gha_language":{"kind":"string","value":"Python"},"files":{"kind":"list like","value":[{"alpha_fraction":0.5813953280448914,"alphanum_fraction":0.7383720874786377,"avg_line_length":13.333333015441895,"blob_id":"60aada3f2799a6b55e4edef8f8b64ebecde273e0","content_id":"f49fa9a21c643d897cc0aa1ff1e3e49d62faf653","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Text","length_bytes":172,"license_type":"no_license","max_line_length":26,"num_lines":12,"path":"/requirements.txt","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"beautifulsoup4==4.6.0\ncoreapi==2.3.3\nDjango==2.0.3\ndjangorestframework==3.7.7\nlxml==4.1.1\nrequests==2.18.4\nurllib3==1.22\npylint==1.7.6\npytest\npytest-django\nipdb\npytest-cov\n"},{"alpha_fraction":0.8289473652839661,"alphanum_fraction":0.8289473652839661,"avg_line_length":24.33333396911621,"blob_id":"8b00524456745f36640ef884093bdd1ec33bc2bc","content_id":"86bf64d84afdf31355fd1f4bb027aaae6b3f4c4e","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":76,"license_type":"no_license","max_line_length":63,"num_lines":3,"path":"/scrabble-play/README.md","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"# Scrabble\n\nusing sowpod scrabble dictionary, to solve interesting problems\n"},{"alpha_fraction":0.4442289173603058,"alphanum_fraction":0.44616878032684326,"avg_line_length":25.461538314819336,"blob_id":"5352bc622caa2305a88354f19bcbb25a8c99bc30","content_id":"69652b36a87409f1507db0c7feb86c1b2af87c5f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1031,"license_type":"no_license","max_line_length":55,"num_lines":39,"path":"/problem_solving/binary_tree_level_order_traversal.py","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"# Definition for a binary tree node.\n# class TreeNode:\n# def __init__(self, x):\n# self.val = x\n# self.left = None\n# self.right = None\n\nclass Solution:\n def levelOrder(self, root):\n \"\"\"\n :type root: TreeNode\n :rtype: List[List[int]]\n \"\"\"\n result = []\n if not root:\n return result\n \n from collections import deque\n queue = deque()\n queue.append(root)\n \n while(len(queue) != 0):\n level_result = []\n next_level_queue = deque()\n \n while(len(queue) != 0):\n node = queue.popleft()\n level_result.append(node.val)\n \n if node.left:\n next_level_queue.append(node.left)\n \n if node.right:\n next_level_queue.append(node.right)\n \n result.append(level_result)\n queue = next_level_queue\n \n return result"},{"alpha_fraction":0.7410714030265808,"alphanum_fraction":0.7410714030265808,"avg_line_length":36.33333206176758,"blob_id":"c033e76aa054de6fd6c9fbdde2bf8472c119b44d","content_id":"afdfdbbca7b4743f4823a62ea5cac9f9f34b6530","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"INI","length_bytes":112,"license_type":"no_license","max_line_length":51,"num_lines":3,"path":"/tdd_example/pytest.ini","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"[pytest]\nDJANGO_SETTINGS_MODULE = tdd_example.test_settings \naddopts = --nomigrations --cov=. --cov-report=html\n"},{"alpha_fraction":0.4469964802265167,"alphanum_fraction":0.4469964802265167,"avg_line_length":34.4375,"blob_id":"6eb4b91428c94e4e461391800eabdcbdcd3beb38","content_id":"8e45c53fde56b0e8a65ae1f92375f4c933e994ec","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":566,"license_type":"no_license","max_line_length":68,"num_lines":16,"path":"/problem_solving/subdomain_visits.py","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"class Solution:\n def subdomainVisits(self, cpdomains):\n \"\"\"\n :type cpdomains: List[str]\n :rtype: List[str]\n \"\"\"\n map = {}\n for domain in cpdomains:\n count, *domains_arr = domain.replace(\" \",\".\").split(\".\")\n for i in range(len(domains_arr)):\n item = \".\".join(domains_arr[i:])\n if not item in map: \n map[item] = int(count)\n else: \n map[item] += int(count)\n return [\" \".join((str(v), k)) for k, v in map.items()]"},{"alpha_fraction":0.6356356143951416,"alphanum_fraction":0.6376376152038574,"avg_line_length":22.761905670166016,"blob_id":"a21b60db7324aa03f9f3775aeed4fe3471bba09c","content_id":"adc06493946193ead879c15249f1a689405a9fdf","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":999,"license_type":"no_license","max_line_length":75,"num_lines":42,"path":"/scrabble-play/solver.py","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"import scrabble\nimport string\n\n# print all words containing \"uu\"\nfor word in scrabble.word_list:\n if \"uu\" in word:\n print(word)\n\n# print all letters that never appear doubled\nfor letter in string.ascii_lowercase:\n exists = False\n for word in scrabble.word_list:\n if letter * 2 in word:\n exists = True\n break\n if not exists:\n print(\"There are no English words with a double {}\".format(letter))\n\n# find all words which contain all vowels\nVOWELS = \"aeiou\"\n\n\ndef has_all_vowels(word):\n for vowel in VOWELS:\n if vowel not in word:\n return False\n return True\n\n\nfor word in scrabble.word_list:\n if has_all_vowels(word):\n print(\"Word with all vowels is : {}\".format(word))\n\n# find the longest palindromes\nlongest_word = \"\"\n\nfor word in scrabble.word_list:\n if word != word[::-1]:\n if len(word) > len(longest_word):\n longest_word = word\n\nprint(\"{} is the longest palindrome\".format(longest_word))\n\n"},{"alpha_fraction":0.7780678868293762,"alphanum_fraction":0.7911227345466614,"avg_line_length":94.75,"blob_id":"71331e7225e578e73537f1965abf574a2a6d4b09","content_id":"fbdd9eee17bbbbbec203237f6b025cbcac72990b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":1149,"license_type":"no_license","max_line_length":130,"num_lines":12,"path":"/problem_solving/README.md","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"# Data Structure and Algorithms\n\n1. LRU Cache Implementation [(Description)](https://leetcode.com/problems/lru-cache/description/)\n2. Trie Implementation [(Description)](https://leetcode.com/problems/implement-trie-prefix-tree/description/)\n3. Maximum Subarray [(Description)](https://leetcode.com/problems/maximum-subarray/description/)\n4. Preorder Traversal (Iterative) [(Description)](https://leetcode.com/problems/binary-tree-preorder-traversal/description/)\n5. Inorder Traversal (Iterative) [(Description)](https://leetcode.com/problems/binary-tree-inorder-traversal/description/)\n6. Postorder Traversal (Iterative) [(Description)](https://leetcode.com/problems/binary-tree-inorder-traversal/description/)\n7. Search 2D Matrix [(Description)](https://leetcode.com/problems/search-a-2d-matrix/description/)\n8. Search 2D Matrix II [(Description)](https://leetcode.com/problems/search-a-2d-matrix-ii/description/)\n9. Binary Tree Level Order Traversal [(Description)](https://leetcode.com/problems/binary-tree-level-order-traversal/description/)\n10. Subdomain Visits [(Description)](https://leetcode.com/problems/subdomain-visit-count/description/)\n"},{"alpha_fraction":0.7570621371269226,"alphanum_fraction":0.7570621371269226,"avg_line_length":43,"blob_id":"26aac61eb6153b8de13f537f5e5692db134851d4","content_id":"968b1273872eaac18df58af47e436a0517800838","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":177,"license_type":"no_license","max_line_length":101,"num_lines":4,"path":"/apod_scraper/README.md","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"# Simple Scraper to download images \n\n* Scrape images from [Astronomy Picture of the Day - NASA](http://apod.nasa.gov/apod/archivepix.html)\n* Store images in a local directory\n\n"},{"alpha_fraction":0.6952117681503296,"alphanum_fraction":0.6970534324645996,"avg_line_length":30.941177368164062,"blob_id":"a64193ecd9174dcb835596a02dfbd2e6e6a312be","content_id":"5b421f0360465235b323faf1c997025650bc2149","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1086,"license_type":"no_license","max_line_length":83,"num_lines":34,"path":"/apod_scraper/scraper.py","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"import os\nimport pathlib\nimport shutil\n\nimport requests\nfrom bs4 import BeautifulSoup\nfrom requests.compat import urljoin\n\n# apod images directory name\nAPOD_IMAGES_DIRECTORY = \"apod_images\"\nBASE_URL = \"http://apod.nasa.gov/apod/archivepix.html\"\n\n# create a directory to store images, if it does not exist already\npathlib.Path(APOD_IMAGES_DIRECTORY).mkdir(exist_ok=True)\n\n# Downloading the index page\napod_base_page = requests.get(BASE_URL).text\n\nfor link in BeautifulSoup(apod_base_page, \"lxml\").find_all('a'):\n print(\"Following link:\", link)\n href = urljoin(BASE_URL, link[\"href\"])\n\n # Follow the link and pull down the image on that link page\n content = requests.get(href).text\n\n for img in BeautifulSoup(content, \"lxml\").find_all('img'):\n img_href = urljoin(href, img[\"src\"])\n\n print(\"Downloading image:\", img_href)\n img_name = img_href.split(\"/\")[-1]\n\n response = requests.get(img_href, stream=True)\n with open(os.path.join(APOD_IMAGES_DIRECTORY, img_name), 'wb') as out_file:\n shutil.copyfileobj(response.raw, out_file)\n"},{"alpha_fraction":0.7831325531005859,"alphanum_fraction":0.7831325531005859,"avg_line_length":40.5,"blob_id":"f067a65e5780c6e8801c88677c7339a1440f2687","content_id":"58dcc18ccbe22330699db5cc4bac69b6f6ae50a9","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":83,"license_type":"no_license","max_line_length":47,"num_lines":2,"path":"/drf_example/README.md","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"# Django Rest Framework Sample APP\n##### CRUD operations for Programming Languages\n"},{"alpha_fraction":0.4376237690448761,"alphanum_fraction":0.4455445408821106,"avg_line_length":24.25,"blob_id":"fb70b6df1d51590341a19cc972bb5d49a4f8fe0e","content_id":"816eeb874953b07e80374c4526a839897d92b16d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":505,"license_type":"no_license","max_line_length":49,"num_lines":20,"path":"/problem_solving/max_subarray.py","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"import sys\n\nclass Solution:\n def maxSubArray(self, nums):\n \"\"\"\n :type nums: List[int]\n :rtype: int\n \"\"\"\n max_so_far = -sys.maxsize - 1\n max_ending_here = 0\n \n for n in nums:\n max_ending_here = max_ending_here + n\n if max_so_far < max_ending_here:\n max_so_far = max_ending_here\n \n if max_ending_here < 0:\n max_ending_here = 0\n \n return max_so_far\n"},{"alpha_fraction":0.40407469868659973,"alphanum_fraction":0.4159592390060425,"avg_line_length":22.117647171020508,"blob_id":"798d7fe27c5c976d0106d7eb503528393f3c3063","content_id":"10b4b938933b5f23bd0a2335ebf8c10f089ba3bf","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1178,"license_type":"no_license","max_line_length":46,"num_lines":51,"path":"/problem_solving/search_2d_matrix.py","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"class Solution:\n def searchMatrix(self, matrix, target):\n if len(matrix) == 0:\n return False\n \"\"\"\n :type matrix: List[List[int]]\n :type target: int\n :rtype: bool\n \"\"\"\n m = len(matrix)\n n = len(matrix[0]);\n \n start = 0;\n end = m*n-1;\n \n while start<=end:\n mid = int(start + (end - start)/2)\n midX = int(mid/n)\n midY = int(mid%n)\n \n if matrix[midX][midY]==target:\n return True\n \n if matrix[midX][midY]= 0 :\n if target == matrix[i][j]:\n return True\n elif target > matrix[i][j]:\n i += 1\n else:\n j -= 1\n return False"},{"alpha_fraction":0.7463768124580383,"alphanum_fraction":0.7536231875419617,"avg_line_length":22,"blob_id":"ab477be9f7eb9f5ff080d57b404089c8864a334a","content_id":"8b593d9109aa33066b3ea84a04814c4e85ef8f82","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":138,"license_type":"no_license","max_line_length":35,"num_lines":6,"path":"/README.md","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"# py-practice\ncode to ramp up my py level\n\n## setup python using virtualenv\n* virtualenv --python=python3.x env\n* source env/bin/activate\n"},{"alpha_fraction":0.3109452724456787,"alphanum_fraction":0.3805970251560211,"avg_line_length":39.20000076293945,"blob_id":"bf917fd2370817bbaf641a7d2b0c32afc6a0b112","content_id":"79b88e286bd2e7d1a0ba155bbe6178f61cdcf12d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":402,"license_type":"no_license","max_line_length":56,"num_lines":10,"path":"/scrabble-play/scrabble.py","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"WORD_LIST = \"sowpods_letters.txt\"\nword_list = open(WORD_LIST).readlines()\nword_list = [word.lower().strip() for word in word_list]\n\nscores = {\"a\": 1, \"c\": 3, \"b\": 3, \"e\": 1, \"d\": 2,\n \"f\": 4, \"g\": 2, \"h\": 4, \"i\": 1, \"j\": 8,\n \"k\": 5, \"l\": 1, \"m\": 3, \"n\": 3, \"o\": 1,\n \"p\": 3, \"q\": 10, \"r\": 1, \"s\": 1, \"t\": 1,\n \"u\": 1, \"v\": 4, \"w\": 3, \"x\": 8, \"y\": 4,\n \"z\": 10}\n"},{"alpha_fraction":0.4772946834564209,"alphanum_fraction":0.47777777910232544,"avg_line_length":24.887500762939453,"blob_id":"e9b9625195d56ce42b2f824b0342cba2ab4d2e9e","content_id":"8f2e3d467444d4e048a94b97ad0d7045f10cf4ff","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2070,"license_type":"no_license","max_line_length":64,"num_lines":80,"path":"/problem_solving/lru_cache.py","repo_name":"pranavbhatt/py-practice","src_encoding":"UTF-8","text":"class LRUNode:\n def __init__(self, key=None, val=None):\n self.value = val\n self.key = key\n self.next = None\n self.prev = None\n\nclass LRUList:\n def __init__(self): \n self.head = LRUNode()\n self.tail = LRUNode()\n self.head.next = self.tail\n self.tail.prev = self.head\n \n def insert(self, node):\n #node is always added to front\n node.next = self.head.next\n node.prev = self.head\n self.head.next.prev = node\n self.head.next = node\n \n def delete(self, node):\n node.prev.next = node.next\n node.next.prev = node.prev\n node.next = None\n node.prev = None\n \n def pop(self):\n node = self.tail.prev\n #cache is empty\n if node == self.head:\n return\n self.delete(node)\n return node\n \n def make_recent(self, node):\n self.delete(node)\n self.insert(node)\n\n \nclass LRUCache:\n def __init__(self, capacity):\n \"\"\"\n :type capacity: int\n \"\"\"\n self.lru_list = LRUList()\n self.lru_map = {}\n self.capacity = capacity \n \n def get(self, key):\n \"\"\"\n :type key: int\n :rtype: int\n \"\"\"\n if key in self.lru_map:\n node = self.lru_map[key]\n self.lru_list.make_recent(node)\n return node.value\n return -1\n \n\n def put(self, key, value):\n \"\"\"\n :type key: int\n :type value: int\n :rtype: void\n \"\"\"\n # cache hit\n if key in self.lru_map:\n node = self.lru_map[key]\n node.value = value\n self.lru_list.make_recent(node)\n else:\n # cache full? pop lru item\n if len(self.lru_map) == self.capacity:\n node = self.lru_list.pop()\n del self.lru_map[node.key]\n node = LRUNode(key=key, val=value)\n self.lru_list.insert(node)\n self.lru_map[key] = node"}],"string":"[\n {\n \"alpha_fraction\": 0.5813953280448914,\n \"alphanum_fraction\": 0.7383720874786377,\n \"avg_line_length\": 13.333333015441895,\n \"blob_id\": \"60aada3f2799a6b55e4edef8f8b64ebecde273e0\",\n \"content_id\": \"f49fa9a21c643d897cc0aa1ff1e3e49d62faf653\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Text\",\n \"length_bytes\": 172,\n \"license_type\": \"no_license\",\n \"max_line_length\": 26,\n \"num_lines\": 12,\n \"path\": \"/requirements.txt\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"beautifulsoup4==4.6.0\\ncoreapi==2.3.3\\nDjango==2.0.3\\ndjangorestframework==3.7.7\\nlxml==4.1.1\\nrequests==2.18.4\\nurllib3==1.22\\npylint==1.7.6\\npytest\\npytest-django\\nipdb\\npytest-cov\\n\"\n },\n {\n \"alpha_fraction\": 0.8289473652839661,\n \"alphanum_fraction\": 0.8289473652839661,\n \"avg_line_length\": 24.33333396911621,\n \"blob_id\": \"8b00524456745f36640ef884093bdd1ec33bc2bc\",\n \"content_id\": \"86bf64d84afdf31355fd1f4bb027aaae6b3f4c4e\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 76,\n \"license_type\": \"no_license\",\n \"max_line_length\": 63,\n \"num_lines\": 3,\n \"path\": \"/scrabble-play/README.md\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Scrabble\\n\\nusing sowpod scrabble dictionary, to solve interesting problems\\n\"\n },\n {\n \"alpha_fraction\": 0.4442289173603058,\n \"alphanum_fraction\": 0.44616878032684326,\n \"avg_line_length\": 25.461538314819336,\n \"blob_id\": \"5352bc622caa2305a88354f19bcbb25a8c99bc30\",\n \"content_id\": \"69652b36a87409f1507db0c7feb86c1b2af87c5f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1031,\n \"license_type\": \"no_license\",\n \"max_line_length\": 55,\n \"num_lines\": 39,\n \"path\": \"/problem_solving/binary_tree_level_order_traversal.py\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Definition for a binary tree node.\\n# class TreeNode:\\n# def __init__(self, x):\\n# self.val = x\\n# self.left = None\\n# self.right = None\\n\\nclass Solution:\\n def levelOrder(self, root):\\n \\\"\\\"\\\"\\n :type root: TreeNode\\n :rtype: List[List[int]]\\n \\\"\\\"\\\"\\n result = []\\n if not root:\\n return result\\n \\n from collections import deque\\n queue = deque()\\n queue.append(root)\\n \\n while(len(queue) != 0):\\n level_result = []\\n next_level_queue = deque()\\n \\n while(len(queue) != 0):\\n node = queue.popleft()\\n level_result.append(node.val)\\n \\n if node.left:\\n next_level_queue.append(node.left)\\n \\n if node.right:\\n next_level_queue.append(node.right)\\n \\n result.append(level_result)\\n queue = next_level_queue\\n \\n return result\"\n },\n {\n \"alpha_fraction\": 0.7410714030265808,\n \"alphanum_fraction\": 0.7410714030265808,\n \"avg_line_length\": 36.33333206176758,\n \"blob_id\": \"c033e76aa054de6fd6c9fbdde2bf8472c119b44d\",\n \"content_id\": \"afdfdbbca7b4743f4823a62ea5cac9f9f34b6530\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"INI\",\n \"length_bytes\": 112,\n \"license_type\": \"no_license\",\n \"max_line_length\": 51,\n \"num_lines\": 3,\n \"path\": \"/tdd_example/pytest.ini\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"[pytest]\\nDJANGO_SETTINGS_MODULE = tdd_example.test_settings \\naddopts = --nomigrations --cov=. --cov-report=html\\n\"\n },\n {\n \"alpha_fraction\": 0.4469964802265167,\n \"alphanum_fraction\": 0.4469964802265167,\n \"avg_line_length\": 34.4375,\n \"blob_id\": \"6eb4b91428c94e4e461391800eabdcbdcd3beb38\",\n \"content_id\": \"8e45c53fde56b0e8a65ae1f92375f4c933e994ec\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 566,\n \"license_type\": \"no_license\",\n \"max_line_length\": 68,\n \"num_lines\": 16,\n \"path\": \"/problem_solving/subdomain_visits.py\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"class Solution:\\n def subdomainVisits(self, cpdomains):\\n \\\"\\\"\\\"\\n :type cpdomains: List[str]\\n :rtype: List[str]\\n \\\"\\\"\\\"\\n map = {}\\n for domain in cpdomains:\\n count, *domains_arr = domain.replace(\\\" \\\",\\\".\\\").split(\\\".\\\")\\n for i in range(len(domains_arr)):\\n item = \\\".\\\".join(domains_arr[i:])\\n if not item in map: \\n map[item] = int(count)\\n else: \\n map[item] += int(count)\\n return [\\\" \\\".join((str(v), k)) for k, v in map.items()]\"\n },\n {\n \"alpha_fraction\": 0.6356356143951416,\n \"alphanum_fraction\": 0.6376376152038574,\n \"avg_line_length\": 22.761905670166016,\n \"blob_id\": \"a21b60db7324aa03f9f3775aeed4fe3471bba09c\",\n \"content_id\": \"adc06493946193ead879c15249f1a689405a9fdf\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 999,\n \"license_type\": \"no_license\",\n \"max_line_length\": 75,\n \"num_lines\": 42,\n \"path\": \"/scrabble-play/solver.py\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import scrabble\\nimport string\\n\\n# print all words containing \\\"uu\\\"\\nfor word in scrabble.word_list:\\n if \\\"uu\\\" in word:\\n print(word)\\n\\n# print all letters that never appear doubled\\nfor letter in string.ascii_lowercase:\\n exists = False\\n for word in scrabble.word_list:\\n if letter * 2 in word:\\n exists = True\\n break\\n if not exists:\\n print(\\\"There are no English words with a double {}\\\".format(letter))\\n\\n# find all words which contain all vowels\\nVOWELS = \\\"aeiou\\\"\\n\\n\\ndef has_all_vowels(word):\\n for vowel in VOWELS:\\n if vowel not in word:\\n return False\\n return True\\n\\n\\nfor word in scrabble.word_list:\\n if has_all_vowels(word):\\n print(\\\"Word with all vowels is : {}\\\".format(word))\\n\\n# find the longest palindromes\\nlongest_word = \\\"\\\"\\n\\nfor word in scrabble.word_list:\\n if word != word[::-1]:\\n if len(word) > len(longest_word):\\n longest_word = word\\n\\nprint(\\\"{} is the longest palindrome\\\".format(longest_word))\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.7780678868293762,\n \"alphanum_fraction\": 0.7911227345466614,\n \"avg_line_length\": 94.75,\n \"blob_id\": \"71331e7225e578e73537f1965abf574a2a6d4b09\",\n \"content_id\": \"fbdd9eee17bbbbbec203237f6b025cbcac72990b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 1149,\n \"license_type\": \"no_license\",\n \"max_line_length\": 130,\n \"num_lines\": 12,\n \"path\": \"/problem_solving/README.md\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Data Structure and Algorithms\\n\\n1. LRU Cache Implementation [(Description)](https://leetcode.com/problems/lru-cache/description/)\\n2. Trie Implementation [(Description)](https://leetcode.com/problems/implement-trie-prefix-tree/description/)\\n3. Maximum Subarray [(Description)](https://leetcode.com/problems/maximum-subarray/description/)\\n4. Preorder Traversal (Iterative) [(Description)](https://leetcode.com/problems/binary-tree-preorder-traversal/description/)\\n5. Inorder Traversal (Iterative) [(Description)](https://leetcode.com/problems/binary-tree-inorder-traversal/description/)\\n6. Postorder Traversal (Iterative) [(Description)](https://leetcode.com/problems/binary-tree-inorder-traversal/description/)\\n7. Search 2D Matrix [(Description)](https://leetcode.com/problems/search-a-2d-matrix/description/)\\n8. Search 2D Matrix II [(Description)](https://leetcode.com/problems/search-a-2d-matrix-ii/description/)\\n9. Binary Tree Level Order Traversal [(Description)](https://leetcode.com/problems/binary-tree-level-order-traversal/description/)\\n10. Subdomain Visits [(Description)](https://leetcode.com/problems/subdomain-visit-count/description/)\\n\"\n },\n {\n \"alpha_fraction\": 0.7570621371269226,\n \"alphanum_fraction\": 0.7570621371269226,\n \"avg_line_length\": 43,\n \"blob_id\": \"26aac61eb6153b8de13f537f5e5692db134851d4\",\n \"content_id\": \"968b1273872eaac18df58af47e436a0517800838\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 177,\n \"license_type\": \"no_license\",\n \"max_line_length\": 101,\n \"num_lines\": 4,\n \"path\": \"/apod_scraper/README.md\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Simple Scraper to download images \\n\\n* Scrape images from [Astronomy Picture of the Day - NASA](http://apod.nasa.gov/apod/archivepix.html)\\n* Store images in a local directory\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6952117681503296,\n \"alphanum_fraction\": 0.6970534324645996,\n \"avg_line_length\": 30.941177368164062,\n \"blob_id\": \"a64193ecd9174dcb835596a02dfbd2e6e6a312be\",\n \"content_id\": \"5b421f0360465235b323faf1c997025650bc2149\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1086,\n \"license_type\": \"no_license\",\n \"max_line_length\": 83,\n \"num_lines\": 34,\n \"path\": \"/apod_scraper/scraper.py\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import os\\nimport pathlib\\nimport shutil\\n\\nimport requests\\nfrom bs4 import BeautifulSoup\\nfrom requests.compat import urljoin\\n\\n# apod images directory name\\nAPOD_IMAGES_DIRECTORY = \\\"apod_images\\\"\\nBASE_URL = \\\"http://apod.nasa.gov/apod/archivepix.html\\\"\\n\\n# create a directory to store images, if it does not exist already\\npathlib.Path(APOD_IMAGES_DIRECTORY).mkdir(exist_ok=True)\\n\\n# Downloading the index page\\napod_base_page = requests.get(BASE_URL).text\\n\\nfor link in BeautifulSoup(apod_base_page, \\\"lxml\\\").find_all('a'):\\n print(\\\"Following link:\\\", link)\\n href = urljoin(BASE_URL, link[\\\"href\\\"])\\n\\n # Follow the link and pull down the image on that link page\\n content = requests.get(href).text\\n\\n for img in BeautifulSoup(content, \\\"lxml\\\").find_all('img'):\\n img_href = urljoin(href, img[\\\"src\\\"])\\n\\n print(\\\"Downloading image:\\\", img_href)\\n img_name = img_href.split(\\\"/\\\")[-1]\\n\\n response = requests.get(img_href, stream=True)\\n with open(os.path.join(APOD_IMAGES_DIRECTORY, img_name), 'wb') as out_file:\\n shutil.copyfileobj(response.raw, out_file)\\n\"\n },\n {\n \"alpha_fraction\": 0.7831325531005859,\n \"alphanum_fraction\": 0.7831325531005859,\n \"avg_line_length\": 40.5,\n \"blob_id\": \"f067a65e5780c6e8801c88677c7339a1440f2687\",\n \"content_id\": \"58dcc18ccbe22330699db5cc4bac69b6f6ae50a9\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 83,\n \"license_type\": \"no_license\",\n \"max_line_length\": 47,\n \"num_lines\": 2,\n \"path\": \"/drf_example/README.md\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Django Rest Framework Sample APP\\n##### CRUD operations for Programming Languages\\n\"\n },\n {\n \"alpha_fraction\": 0.4376237690448761,\n \"alphanum_fraction\": 0.4455445408821106,\n \"avg_line_length\": 24.25,\n \"blob_id\": \"fb70b6df1d51590341a19cc972bb5d49a4f8fe0e\",\n \"content_id\": \"816eeb874953b07e80374c4526a839897d92b16d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 505,\n \"license_type\": \"no_license\",\n \"max_line_length\": 49,\n \"num_lines\": 20,\n \"path\": \"/problem_solving/max_subarray.py\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import sys\\n\\nclass Solution:\\n def maxSubArray(self, nums):\\n \\\"\\\"\\\"\\n :type nums: List[int]\\n :rtype: int\\n \\\"\\\"\\\"\\n max_so_far = -sys.maxsize - 1\\n max_ending_here = 0\\n \\n for n in nums:\\n max_ending_here = max_ending_here + n\\n if max_so_far < max_ending_here:\\n max_so_far = max_ending_here\\n \\n if max_ending_here < 0:\\n max_ending_here = 0\\n \\n return max_so_far\\n\"\n },\n {\n \"alpha_fraction\": 0.40407469868659973,\n \"alphanum_fraction\": 0.4159592390060425,\n \"avg_line_length\": 22.117647171020508,\n \"blob_id\": \"798d7fe27c5c976d0106d7eb503528393f3c3063\",\n \"content_id\": \"10b4b938933b5f23bd0a2335ebf8c10f089ba3bf\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1178,\n \"license_type\": \"no_license\",\n \"max_line_length\": 46,\n \"num_lines\": 51,\n \"path\": \"/problem_solving/search_2d_matrix.py\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"class Solution:\\n def searchMatrix(self, matrix, target):\\n if len(matrix) == 0:\\n return False\\n \\\"\\\"\\\"\\n :type matrix: List[List[int]]\\n :type target: int\\n :rtype: bool\\n \\\"\\\"\\\"\\n m = len(matrix)\\n n = len(matrix[0]);\\n \\n start = 0;\\n end = m*n-1;\\n \\n while start<=end:\\n mid = int(start + (end - start)/2)\\n midX = int(mid/n)\\n midY = int(mid%n)\\n \\n if matrix[midX][midY]==target:\\n return True\\n \\n if matrix[midX][midY]= 0 :\\n if target == matrix[i][j]:\\n return True\\n elif target > matrix[i][j]:\\n i += 1\\n else:\\n j -= 1\\n return False\"\n },\n {\n \"alpha_fraction\": 0.7463768124580383,\n \"alphanum_fraction\": 0.7536231875419617,\n \"avg_line_length\": 22,\n \"blob_id\": \"ab477be9f7eb9f5ff080d57b404089c8864a334a\",\n \"content_id\": \"8b593d9109aa33066b3ea84a04814c4e85ef8f82\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 138,\n \"license_type\": \"no_license\",\n \"max_line_length\": 35,\n \"num_lines\": 6,\n \"path\": \"/README.md\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# py-practice\\ncode to ramp up my py level\\n\\n## setup python using virtualenv\\n* virtualenv --python=python3.x env\\n* source env/bin/activate\\n\"\n },\n {\n \"alpha_fraction\": 0.3109452724456787,\n \"alphanum_fraction\": 0.3805970251560211,\n \"avg_line_length\": 39.20000076293945,\n \"blob_id\": \"bf917fd2370817bbaf641a7d2b0c32afc6a0b112\",\n \"content_id\": \"79b88e286bd2e7d1a0ba155bbe6178f61cdcf12d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 402,\n \"license_type\": \"no_license\",\n \"max_line_length\": 56,\n \"num_lines\": 10,\n \"path\": \"/scrabble-play/scrabble.py\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"WORD_LIST = \\\"sowpods_letters.txt\\\"\\nword_list = open(WORD_LIST).readlines()\\nword_list = [word.lower().strip() for word in word_list]\\n\\nscores = {\\\"a\\\": 1, \\\"c\\\": 3, \\\"b\\\": 3, \\\"e\\\": 1, \\\"d\\\": 2,\\n \\\"f\\\": 4, \\\"g\\\": 2, \\\"h\\\": 4, \\\"i\\\": 1, \\\"j\\\": 8,\\n \\\"k\\\": 5, \\\"l\\\": 1, \\\"m\\\": 3, \\\"n\\\": 3, \\\"o\\\": 1,\\n \\\"p\\\": 3, \\\"q\\\": 10, \\\"r\\\": 1, \\\"s\\\": 1, \\\"t\\\": 1,\\n \\\"u\\\": 1, \\\"v\\\": 4, \\\"w\\\": 3, \\\"x\\\": 8, \\\"y\\\": 4,\\n \\\"z\\\": 10}\\n\"\n },\n {\n \"alpha_fraction\": 0.4772946834564209,\n \"alphanum_fraction\": 0.47777777910232544,\n \"avg_line_length\": 24.887500762939453,\n \"blob_id\": \"e9b9625195d56ce42b2f824b0342cba2ab4d2e9e\",\n \"content_id\": \"8f2e3d467444d4e048a94b97ad0d7045f10cf4ff\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2070,\n \"license_type\": \"no_license\",\n \"max_line_length\": 64,\n \"num_lines\": 80,\n \"path\": \"/problem_solving/lru_cache.py\",\n \"repo_name\": \"pranavbhatt/py-practice\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"class LRUNode:\\n def __init__(self, key=None, val=None):\\n self.value = val\\n self.key = key\\n self.next = None\\n self.prev = None\\n\\nclass LRUList:\\n def __init__(self): \\n self.head = LRUNode()\\n self.tail = LRUNode()\\n self.head.next = self.tail\\n self.tail.prev = self.head\\n \\n def insert(self, node):\\n #node is always added to front\\n node.next = self.head.next\\n node.prev = self.head\\n self.head.next.prev = node\\n self.head.next = node\\n \\n def delete(self, node):\\n node.prev.next = node.next\\n node.next.prev = node.prev\\n node.next = None\\n node.prev = None\\n \\n def pop(self):\\n node = self.tail.prev\\n #cache is empty\\n if node == self.head:\\n return\\n self.delete(node)\\n return node\\n \\n def make_recent(self, node):\\n self.delete(node)\\n self.insert(node)\\n\\n \\nclass LRUCache:\\n def __init__(self, capacity):\\n \\\"\\\"\\\"\\n :type capacity: int\\n \\\"\\\"\\\"\\n self.lru_list = LRUList()\\n self.lru_map = {}\\n self.capacity = capacity \\n \\n def get(self, key):\\n \\\"\\\"\\\"\\n :type key: int\\n :rtype: int\\n \\\"\\\"\\\"\\n if key in self.lru_map:\\n node = self.lru_map[key]\\n self.lru_list.make_recent(node)\\n return node.value\\n return -1\\n \\n\\n def put(self, key, value):\\n \\\"\\\"\\\"\\n :type key: int\\n :type value: int\\n :rtype: void\\n \\\"\\\"\\\"\\n # cache hit\\n if key in self.lru_map:\\n node = self.lru_map[key]\\n node.value = value\\n self.lru_list.make_recent(node)\\n else:\\n # cache full? pop lru item\\n if len(self.lru_map) == self.capacity:\\n node = self.lru_list.pop()\\n del self.lru_map[node.key]\\n node = LRUNode(key=key, val=value)\\n self.lru_list.insert(node)\\n self.lru_map[key] = node\"\n }\n]"},"num_files":{"kind":"number","value":15,"string":"15"}}},{"rowIdx":2980394,"cells":{"repo_name":{"kind":"string","value":"diwadd/PathImageSegmentation"},"repo_url":{"kind":"string","value":"https://github.com/diwadd/PathImageSegmentation"},"snapshot_id":{"kind":"string","value":"f487a114f600449c6a1538ee3f1eabe2ee6f8bb3"},"revision_id":{"kind":"string","value":"2c9efb7f69e5cfe8de21d04a16e676cc6686fc7c"},"directory_id":{"kind":"string","value":"77f53a619bed85e9d91d508d6d39fc4dbf328087"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2021-01-02T22:51:54.661577","string":"2021-01-02T22:51:54.661577"},"revision_date":{"kind":"timestamp","value":"2017-08-07T19:51:24","string":"2017-08-07T19:51:24"},"committer_date":{"kind":"timestamp","value":"2017-08-07T19:52:01","string":"2017-08-07T19:52:01"},"github_id":{"kind":"number","value":99406410,"string":"99,406,410"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"timestamp","value":"2017-08-05T07:18:26","string":"2017-08-05T07:18:26"},"gha_updated_at":{"kind":"timestamp","value":"2017-08-05T07:19:03","string":"2017-08-05T07:19:03"},"gha_pushed_at":{"kind":"timestamp","value":"2017-08-05T07:21:45","string":"2017-08-05T07:21:45"},"gha_language":{"kind":"string","value":"Python"},"files":{"kind":"list like","value":[{"alpha_fraction":0.5162613987922668,"alphanum_fraction":0.5376899838447571,"avg_line_length":27.236051559448242,"blob_id":"f4f5e6d5ae43ac3cdc357185b87dd2acf5131652","content_id":"54ba787e17929fc6556bfb7c5fb5507a2faeb89d","detected_licenses":["MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":6580,"license_type":"permissive","max_line_length":80,"num_lines":233,"path":"/main_training.py","repo_name":"diwadd/PathImageSegmentation","src_encoding":"UTF-8","text":"import random\nimport sys\nimport glob\n\nimport numpy as np\n\nfrom keras import backend as K\nfrom keras.models import load_model\nfrom keras.applications.vgg16 import VGG16\nfrom keras import optimizers\n\nimport tensorflow as tf\n\nimport data_handling as dh\nimport deep_models as dm\n\nrandom.seed(111)\n\n\n\ndef data_generator(file_name_list, noli=20):\n \n \"\"\"\n noli - number of loaded images per yield\n \n \"\"\"\n\n while True:\n n = len(file_name_list)\n number_of_image_loads = round(n / noli)\n ptr = 0\n # print(\"n: \" + str(n))\n # print(\"number_of_image_loads: \" + str(number_of_image_loads))\n\n for i in range(number_of_image_loads):\n # print(\"We are a i: \" + str(i))\n # create numpy arrays of input data\n # and labels, from each line in the file\n mini_batch_fnl = file_name_list[ptr:(ptr + noli)]\n ptr = ptr + noli\n\n x_data, y_data = dh.load_data_from_npz(mini_batch_fnl)\n yield (x_data, y_data)\n\n\ndef evaluate(model, file_name_list, noli=20):\n \n \"\"\"\n noli - number of loaded images per yield\n \n \"\"\"\n\n n = len(file_name_list)\n number_of_image_loads = round(n / noli)\n ptr = 0\n # print(\"n: \" + str(n))\n # print(\"number_of_image_loads: \" + str(number_of_image_loads))\n\n mean_loss = 0.0\n mean_acc = 0.0\n for i in range(number_of_image_loads):\n # print(\"We are a i: \" + str(i))\n # create numpy arrays of input data\n # and labels, from each line in the file\n mini_batch_fnl = file_name_list[ptr:(ptr + noli)]\n ptr = ptr + noli\n\n x_data, y_data = dh.load_data_from_npz(mini_batch_fnl)\n score = model.evaluate(x_data, y_data, verbose=0)\n\n #print(\"score: \" + str(score))\n\n local_loss = score[0]\n local_acc = score[1]\n\n mean_loss += local_loss\n mean_acc += local_acc\n mean_loss /= number_of_image_loads\n mean_acc /= number_of_image_loads\n print(\"Mean loss: \" + str(mean_loss) + \" - mean acc: \" + str(mean_acc))\n \n\n\nif __name__ == \"__main__\":\n print(\"Starting main.\")\n\n # ------------------------------------------------------------------------\n # Load data\n # ------------------------------------------------------------------------\n\n data_file_names = glob.glob(\"augmented_images/*npz\")\n\n n_files = len(data_file_names)\n\n loaded_data = np.load(data_file_names[0])\n image = loaded_data[\"image\"]\n label = loaded_data[\"label\"]\n\n iw, ih, ic = image.shape\n nw, nh, nc= label.shape\n\n random.shuffle(data_file_names)\n train_fraction = 0.6\n valid_fraction = 1.0 - train_fraction\n test_fraction = 0.5\n\n x_train_fnl = data_file_names[0:int(train_fraction*n_files)]\n temp = data_file_names[int(train_fraction*n_files):]\n x_valid_fnl = temp[0:int(len(temp)/2)]\n x_test_fnl = temp[int(len(temp)/2):]\n\n print(\"First train file: \" + str(x_train_fnl[0]))\n\n print(\"Number of train data files: \" + str(len(x_train_fnl)))\n print(\"Number of valid data files: \" + str(len(x_valid_fnl)))\n print(\"Number of test data files: \" + str(len(x_test_fnl)))\n\n K.get_session()\n\n\n new_model = False\n if (new_model == True):\n print(\"Creating a new model!\")\n\n\n #model = FCN(input_shape=(iw, ih, ic), classes=2, \n # weights='imagenet', trainable_encoder=True)\n\n model = dm.vgg16_16s_fcn(iw, \n ih, \n ic,\n dropout=0.5,\n alpha=0.0001,\n classes=2)\n\n opt = optimizers.SGD(lr=0.0001, momentum=0.9, clipvalue=0.5)\n opt = optimizers.Adam(1e-4, clipvalue=0.5)\n model.compile(loss=\"categorical_crossentropy\",\n optimizer=opt,\n metrics=[\"accuracy\"])\n\n\n vgg16_model = VGG16(weights='imagenet', include_top=False)\n vgg16_model.compile(loss=\"categorical_crossentropy\",\n optimizer=\"adadelta\")\n\n index = 12\n print(\"Our model config: \")\n print(model.layers[index].get_config())\n\n print(\"VGG16 model config: \")\n print(vgg16_model.layers[index].get_config())\n\n print(\"Our model:\")\n layer1 = model.layers[index].get_weights()\n print(layer1[0].shape)\n print(layer1[0][0,0,0,1:5])\n\n print(\"VGG16 model:\")\n layer1 = vgg16_model.layers[index].get_weights()\n print(layer1[0].shape)\n print(layer1[0][0,0,0,1:5])\n\n\n for i in range(len(vgg16_model.layers)):\n\n print(\"\\n\\n\\n Layer number: \" + str(i))\n\n layer_config = vgg16_model.layers[i].get_config()\n print(\"Setting name: \" + str(layer_config['name']))\n if layer_config['name'] == 'block5_pool':\n break\n\n layer_weights = vgg16_model.layers[i].get_weights()\n\n if len(layer_weights) != 0:\n print(layer1[0].shape)\n model.layers[i].set_weights(vgg16_model.layers[i].get_weights())\n else:\n print(\"No wieghts to set!\")\n\n print(\"After setting the weights...\")\n\n print(\"Our model:\")\n layer1 = model.layers[index].get_weights()\n print(layer1[0].shape)\n print(layer1[0][0,0,0,1:5])\n\n print(\"VGG16 model:\")\n layer1 = vgg16_model.layers[index].get_weights()\n print(layer1[0].shape)\n print(layer1[0][0,0,0,1:5])\n\n\n\n #model.compile(optimizer='adadelta',\n # loss='categorical_crossentropy',\n # metrics=['accuracy'])\n\n model.summary()\n\n\n else:\n model = load_model(\"model.h5\")\n print(\"Model loaded!\")\n model.summary()\n\n noli = 4\n n = len(x_train_fnl)\n number_of_image_loads = round(n / noli)\n print(\"Number of image loads: \" + str(number_of_image_loads))\n n_epochs = 100 \n n_sub_epochs = 5\n\n print(\"Pre training evaluation:\")\n evaluate(model, x_valid_fnl, noli=4)\n\n for i in range(n_epochs):\n print(\"Global epoch: \" + str(i) + \"/\" + str(n_epochs))\n model.fit_generator(data_generator(data_file_names, noli=noli),\n steps_per_epoch=number_of_image_loads,\n epochs=n_sub_epochs)\n print(\"Validation data mean loss: \")\n evaluate(model, x_valid_fnl, noli=4)\n\n\n model.save(\"model.h5\")\n\n print(\"Test data mean loss: \")\n evaluate(model, x_test_fnl, noli=64)\n\n\n K.clear_session()\n\n"},{"alpha_fraction":0.3254545331001282,"alphanum_fraction":0.4509090781211853,"avg_line_length":17.33333396911621,"blob_id":"d43756e65b061b6515c1c210768cfda3c2a3006e","content_id":"e96cd63a8c5fb40927d1ca53e57575a5d8d16330","detected_licenses":["MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":550,"license_type":"permissive","max_line_length":74,"num_lines":30,"path":"/feeling_binary_crossentropy.py","repo_name":"diwadd/PathImageSegmentation","src_encoding":"UTF-8","text":"import numpy as np\n\n\ndef bc(t, o):\n return -np.mean(t*np.log(o) + (1-t)*np.log(1-o))\n\n\nt = np.array([ 0, 0, 0, 1, 1, 1, 0, 0, 0, 1])\no = np.array([0.10, 0.10, 0.10, 0.95, 0.90, 0.90, 0.10, 0.10, 0.10, 0.90])\n\nprint(bc(t, o))\n\n\nn = 10000\nt = np.random.randint(2, size=n)\no = np.array([0.01 for i in range(n)])\n\nfor i in range(n):\n if t[i] == 1:\n o[i] = 0.99\n\nfor i in range(int(0.013*n)):\n if t[i] == 1:\n o[i] = 0.001\n elif t[i] == 0:\n o[i] = 0.999\n else:\n pass\n\nprint(bc(t, o))\n"},{"alpha_fraction":0.5835299491882324,"alphanum_fraction":0.6379121541976929,"avg_line_length":47.25833511352539,"blob_id":"4d97324a8ad7a9c93d3b2aac198fdd0e3cabdfb1","content_id":"4dcefe5660c67311c4f3a21fb3bb58b36f0c330d","detected_licenses":["MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":17377,"license_type":"permissive","max_line_length":194,"num_lines":360,"path":"/deep_models.py","repo_name":"diwadd/PathImageSegmentation","src_encoding":"UTF-8","text":"import keras\n\nfrom keras.models import Model\nfrom keras.models import Sequential\nfrom keras.layers import Flatten\nfrom keras.layers import Dense\nfrom keras.layers import Input\nfrom keras.layers import Conv2D\nfrom keras.layers import Conv2DTranspose\nfrom keras.layers import Cropping2D\nfrom keras.layers import MaxPooling2D\nfrom keras.layers import GlobalMaxPooling2D\nfrom keras.layers import Activation\nfrom keras.layers import GlobalAveragePooling2D\n\n\nfrom keras.layers import Dropout \nfrom keras.layers import BatchNormalization\nfrom keras.layers.advanced_activations import LeakyReLU\n\nfrom keras import regularizers\nfrom keras import backend as K\n\ndef basic_model(iw=500, # Input width\n ih=500, # Input height \n ic=3,\n ow=100, # Output width\n oh=100, # Output heigth\n dropout=0.9,\n alpha=0.0):\n\n input_image = Input((iw, ih, ic))\n\n x = Conv2D(16, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_1_layer_1\")(input_image)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(16, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_1_layer_2\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\"block_1_pooling\")(x)\n\n x = Conv2D(32, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_2_layer_1\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(32, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_2_layer_2\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\"block_2_pooling\")(x)\n\n x = Conv2D(64, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_3_layer_1\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(64, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_3_layer_2\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(64, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_3_layer_3\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\"block_3_pooling\")(x)\n\n x = Conv2D(128, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_4_layer_1\")(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(128, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_4_layer_2\")(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(128, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_4_layer_3\")(x)\n x = LeakyReLU(alpha)(x)\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\"block_4_pooling\")(x)\n\n x = Conv2D(128, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_5_layer_1\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(128, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_5_layer_2\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(128, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_5_layer_3\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\"block_5_pooling\")(x)\n\n x = Conv2D(256, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_7_layer_1\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(256, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_7_layer_2\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(256, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_7_layer_3\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\"block_7_pooling\")(x)\n\n x = Flatten(name=\"flatten\")(x)\n x = Dense(4096, activation=\"linear\", name=\"full_connected_1\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Dropout(dropout)(x)\n\n x = Dense(4096, activation=\"linear\", name=\"full_connected_2\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Dropout(dropout)(x)\n\n x = Dense(ow*oh, activation=\"sigmoid\", name=\"predictions\")(x)\n\n model = Model(input_image, x, name=\"vgg16_based\")\n\n model.compile(loss=\"binary_crossentropy\",\n optimizer=\"adadelta\")\n\n print(\"\\n ---> Model summary <--- \\n\")\n model.summary()\n\n return model\n\ndef basic_model_pooling(iw=500, # Input width\n ih=500, # Input height \n ic=3,\n ow=100, # Output width\n oh=100, # Output heigth\n dropout=0.9,\n alpha=0.001):\n\n input_image = Input((iw, ih, ic))\n\n x = Conv2D(16, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_1_layer_1\")(input_image)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(16, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_1_layer_2\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\"block_1_pooling\")(x)\n\n x = Conv2D(32, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_2_layer_1\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(32, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_2_layer_2\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\"block_2_pooling\")(x)\n\n x = Conv2D(64, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_3_layer_1\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(64, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_3_layer_2\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(64, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_3_layer_3\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\"block_3_pooling\")(x)\n\n x = Conv2D(128, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_4_layer_1\")(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(128, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_4_layer_2\")(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(128, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_4_layer_3\")(x)\n x = LeakyReLU(alpha)(x)\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\"block_4_pooling\")(x)\n\n x = Conv2D(128, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_5_layer_1\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(128, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_5_layer_2\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(128, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_5_layer_3\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\"block_5_pooling\")(x)\n\n x = Conv2D(256, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_6_layer_1\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(512, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_6_layer_2\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n x = Conv2D(ow*oh, (3, 3), activation=\"linear\", padding=\"same\", name=\"block_6_layer_3\")(x)\n x = BatchNormalization()(x)\n x = LeakyReLU(alpha)(x)\n\n x = GlobalAveragePooling2D()(x)\n\n model = Model(input_image, x, name=\"vgg16_based\")\n\n model.compile(loss=\"binary_crossentropy\",\n optimizer=\"adadelta\")\n\n print(\"\\n ---> Model summary <--- \\n\")\n model.summary()\n\n return model\n\n\ndef vgg16_32s_fcn(iw=500, # Input width\n ih=500, # Input height \n ic=3,\n dropout=0.5,\n alpha=0.001,\n classes=2):\n # Based on:\n # Fully Convolutional Models for Semantic Segmentation\n # Evan Shelhamer*, Jonathan Long*, Trevor Darrell\n # PAMI 2016\n # arXiv:1605.06211\n\n reg_fun = regularizers.l2(alpha)\n\n input_image = Input((iw, ih, ic))\n\n # Conv 1\n x = Conv2D(64, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block1_conv1')(input_image)\n x = Conv2D(64, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block1_conv2')(x)\n pool1 = MaxPooling2D((2, 2), strides=(2, 2), name='block1_pool')(x)\n\n # Conv 2\n x = Conv2D(128, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block2_conv1')(pool1)\n x = Conv2D(128, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block2_conv2')(x)\n pool2 = MaxPooling2D((2, 2), strides=(2, 2), name='block2_pool')(x)\n\n # Conv 3\n x = Conv2D(256, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block3_conv1')(pool2)\n x = Conv2D(256, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block3_conv2')(x)\n x = Conv2D(256, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block3_conv3')(x)\n pool3 = MaxPooling2D((2, 2), strides=(2, 2), name='block3_pool')(x)\n\n # Conv 4\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block4_conv1')(pool3)\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block4_conv2')(x)\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block4_conv3')(x)\n pool4 = MaxPooling2D((2, 2), strides=(2, 2), name='block4_pool')(x)\n\n # Conv 5\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block5_conv1')(pool4)\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block5_conv2')(x)\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block5_conv3')(x)\n pool5 = MaxPooling2D((2, 2), strides=(2, 2), name='block5_pool')(x)\n\n # Fully Conv fc6\n fc6 = Conv2D(4096, (7, 7), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='fc6')(pool5)\n drop6 = Dropout(rate=dropout)(fc6)\n\n # Fully Conv fc7\n fc7 = Conv2D(4096, (1, 1), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='fc7')(drop6)\n drop7 = Dropout(rate=dropout)(fc7)\n\n score_fr = Conv2D(classes, (1, 1), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='score_fr')(drop7)\n\n upscore = Conv2DTranspose(classes, kernel_size=(64, 64), strides=(32, 32), kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='upscore')(score_fr)\n\n\n _, uw, uh, uc = upscore._keras_shape\n cw = (uw - iw)//2\n ch = (uh - ih)//2\n print(\"cw: \" + str(cw))\n print(\"ch: \" + str(ch))\n\n score = Cropping2D(cropping=(cw, ch))(upscore)\n output = Activation('softmax')(score)\n\n model = Model(input_image, output, name=\"vgg16_based\")\n\n return model\n\n\ndef vgg16_16s_fcn(iw=500, # Input width\n ih=500, # Input height \n ic=3,\n dropout=0.5,\n alpha=0.001,\n classes=2):\n # Based on:\n # Fully Convolutional Models for Semantic Segmentation\n # Evan Shelhamer*, Jonathan Long*, Trevor Darrell\n # PAMI 2016\n # arXiv:1605.06211\n\n reg_fun = regularizers.l2(alpha)\n\n input_image = Input((iw, ih, ic))\n\n # Conv 1\n x = Conv2D(64, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block1_conv1')(input_image)\n x = Conv2D(64, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block1_conv2')(x)\n pool1 = MaxPooling2D((2, 2), strides=(2, 2), name='block1_pool')(x)\n\n # Conv 2\n x = Conv2D(128, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block2_conv1')(pool1)\n x = Conv2D(128, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block2_conv2')(x)\n pool2 = MaxPooling2D((2, 2), strides=(2, 2), name='block2_pool')(x)\n\n # Conv 3\n x = Conv2D(256, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block3_conv1')(pool2)\n x = Conv2D(256, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block3_conv2')(x)\n x = Conv2D(256, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block3_conv3')(x)\n pool3 = MaxPooling2D((2, 2), strides=(2, 2), name='block3_pool')(x)\n\n # Conv 4\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block4_conv1')(pool3)\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block4_conv2')(x)\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block4_conv3')(x)\n pool4 = MaxPooling2D((2, 2), strides=(2, 2), name='block4_pool')(x)\n\n # Conv 5\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block5_conv1')(pool4)\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block5_conv2')(x)\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block5_conv3')(x)\n pool5 = MaxPooling2D((2, 2), strides=(2, 2), name='block5_pool')(x)\n\n # Fully Conv fc6\n fc6 = Conv2D(4096, (7, 7), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='fc6')(pool5)\n drop6 = Dropout(rate=dropout)(fc6)\n\n # Fully Conv fc7\n fc7 = Conv2D(4096, (1, 1), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='fc7')(drop6)\n drop7 = Dropout(rate=dropout)(fc7)\n\n score_fr = Conv2D(classes, (1, 1), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='score_fr')(drop7)\n upscore2 = Conv2DTranspose(classes, kernel_size=(4, 4), strides=(2, 2), kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='upscore2')(score_fr)\n\n\n score_pool4 = Conv2D(classes, (1, 1), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='score_pool4')(pool4)\n\n _, uw, uh, uc = upscore2._keras_shape\n _, sw, sh, sc = score_pool4._keras_shape\n\n if ((uw - sw) == 1) or ((uh - sh) == 1):\n cw1 = 1\n ch1 = 1\n cropping = ((cw1, 0),(ch1, 0))\n else:\n cw1 = (uw - sw)//2\n ch1 = (uh - sh)//2\n cropping = (cw1, ch1)\n\n print(\"cw1: \" + str(cw1))\n print(\"ch1: \" + str(ch1))\n\n print(\"upscore2._keras_shape \" + str(upscore2._keras_shape))\n print(\"score_pool4._keras_shape \" + str(score_pool4._keras_shape))\n\n # Technically score_pool4 should have a larger size then upscore2.\n # At least that is what follows from crop(n.score_pool4, n.upscore2).\n # This is, however, not the case and we nned to crop upscore2.\n\n score_pool4c = Cropping2D(cropping=cropping)(upscore2) \n fuse_pool4 = keras.layers.Add()([score_pool4c, score_pool4])\n\n upscore16 = Conv2DTranspose(classes, kernel_size=(32, 32), strides=(16, 16), kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='upscore16')(fuse_pool4)\n\n _, uw, uh, uc = upscore16._keras_shape\n cw2 = (uw - iw)//2\n ch2 = (uh - ih)//2\n #print(\"cw2: \" + str(cw2))\n #print(\"ch2: \" + str(ch2))\n\n score = Cropping2D(cropping=(cw2, ch2))(upscore16)\n output = Activation('softmax')(score)\n\n model = Model(input_image, output, name=\"vgg16_based\")\n\n return model\n\n\n\n\n"},{"alpha_fraction":0.5284159779548645,"alphanum_fraction":0.6118500828742981,"avg_line_length":29.518518447875977,"blob_id":"b84f2279d5954ca6a997eb93f107c605266f5e34","content_id":"4524546c634058dfdda3bc02f97a8e1542f85f12","detected_licenses":["MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":827,"license_type":"permissive","max_line_length":78,"num_lines":27,"path":"/check_numpy_saved_array.py","repo_name":"diwadd/PathImageSegmentation","src_encoding":"UTF-8","text":"import cv2\nimport numpy as np\n\nimport matplotlib.pyplot as plt\nimport data_handling as dh\n\nif __name__ == \"__main__\":\n\n #file_name = \"augmented_images/numpy_image_array_id_0_0p0_30_30_0p85.npz\"\n file_name = \"augmented_images/numpy_image_array_id_0_3p0_10_10_1p0.npz\"\n\n loaded_data = np.load(file_name)\n\n image = (255.0*loaded_data[\"image\"]).astype(np.uint8)\n label = (255.0*loaded_data[\"label\"]).astype(np.uint8)\n #label = np.reshape(label, (100, 100))\n\n print(\"label shape: \" + str(label.shape))\n\n #label = cv2.resize(label,\n # (500, 500),\n # interpolation = cv2.INTER_LINEAR)\n #mask = cv2.imread(\"augmented_images/truth_image_id_0_0p0_30_30_0p85.png\")\n\n dh.plot_three_images(image, label[:, :, 0], label[:, :, 1])\n\n print(label[100:120, 100:120])\n\n\n\n"},{"alpha_fraction":0.5060120224952698,"alphanum_fraction":0.5751503109931946,"avg_line_length":22.447059631347656,"blob_id":"fff37a204466de8854984b1f7e84c138a67c3e4d","content_id":"5294bab4360471e30e63e97980a084e7aa3c5f95","detected_licenses":["MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1996,"license_type":"permissive","max_line_length":116,"num_lines":85,"path":"/temp.py","repo_name":"diwadd/PathImageSegmentation","src_encoding":"UTF-8","text":"import random\nimport sys\nimport glob\n\nimport cv2\nimport numpy as np\n\nimport keras\n\n#from keras import backend as K\n#from keras.models import load_model\n#from tensorflow.python.ops import math_ops\n\n#import deep_models as dm\n\n\n# model = dm.vg16_fcn()\n# model.summary()\n# model.compile(loss=\"categorical_crossentropy\",\n# optimizer=\"adadelta\")\n\n\"\"\"\n\ndef cross_entropy(t, p):\n return -np.sum(t*np.log(p))\n\n\n\n\ntrue_arr = np.array([[[1.0, 0.0], [1.0, 0.0], [1.0, 0.0]], \n [[1.0, 0.0], [1.0, 0.0], [1.0, 0.0]],\n [[1.0, 0.0], [1.0, 0.0], [1.0, 0.0]],])\n\n\npred_arr = np.array([[[0.9, 0.1], [0.9, 0.1], [0.9, 0.1]], \n [[0.9, 0.1], [0.99, 0.01], [0.99, 0.01]],\n [[0.9, 0.1], [0.99, 0.01], [0.9, 0.01]]])\n\npred_npr = np.array([[[9.9, 1.1], [8.9, 0.1], [1.9, 6.1]], \n [[9.9, 2.1], [9.4, 9.6], [2.5, 5.5]],\n [[9.9, 1.1], [9.2, 0.8], [3.9, 4.1]]])\n\nprint(true_arr.shape)\n\n\nce = cross_entropy(true_arr, pred_arr)\nprint(\"ce: \" + str(ce))\n\n\"\"\"\n\n\n\"\"\"\nK.get_session()\n\ntrue_var = K.variable(true_arr)\npred_var = K.variable(pred_arr)\npred_npr = K.variable(pred_npr)\n\n# output = pred_var / math_ops.reduce_sum(pred_var, reduction_indices=len(pred_var.get_shape()) - 1, keep_dims=True)\n# print(K.eval(output))\n\nsx = K.softmax(pred_npr)\nen = K.categorical_crossentropy(pred_var, true_var, from_logits=False)\n\nsx_eval = K.eval(sx)\nprint(\"sx_eval\")\nprint(sx_eval)\n\nen_eval = K.eval(en)\nprint(\"en_eval\")\nprint(en_eval)\n\nK.clear_session()\n\"\"\"\n\ninput1 = keras.layers.Input(shape=(16,))\nx1 = keras.layers.Dense(8, activation='relu')(input1)\ninput2 = keras.layers.Input(shape=(32,))\nx2 = keras.layers.Dense(8, activation='relu')(input2)\nadded = keras.layers.Add()([x1, x2]) # equivalent to added = keras.layers.add([x1, x2])\n\nprint(\"added._keras_shape: \" + str(added._keras_shape)) \n\nout = keras.layers.Dense(4)(added)\nmodel = keras.models.Model(inputs=[input1, input2], outputs=out)\n\n\n\n"},{"alpha_fraction":0.5030834674835205,"alphanum_fraction":0.5170778036117554,"avg_line_length":28.787986755371094,"blob_id":"f3101c7144555b606110f8abc8fb73279bfa08e9","content_id":"e4a92105596c7d56ed2c7056c03044752ad9ec20","detected_licenses":["MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":8432,"license_type":"permissive","max_line_length":118,"num_lines":283,"path":"/data_handling.py","repo_name":"diwadd/PathImageSegmentation","src_encoding":"UTF-8","text":"import glob\nimport cv2\nimport math\nimport sys\nimport os\nimport shutil\n\nimport numpy as np\nimport matplotlib.pyplot as plt\n\nfrom keras.utils.np_utils import to_categorical\n\nDEFAULT_WIDTH = 500\nDEFAULT_HEIGHT = 500\n\ndef load_images(file_names_list):\n\n n_images = len(file_names_list)\n\n images = [None for i in range(n_images)]\n for i in range(n_images):\n images[i] = (cv2.imread(file_names_list[i]))\n\n return images\n\n\ndef read_data(data_dir=\"training/images/\"): \n\n train_file_names_list = glob.glob(data_dir + \"*.tif\")\n n_train_images = len(train_file_names_list)\n\n # The truth images have the same numbers.\n # We change the path and the extensions.\n # This ensures that train image will\n # correspond to truth image.\n truth_file_names_list = [None for i in range(n_train_images)]\n for i in range(n_train_images):\n\n if data_dir == \"training/images/\":\n truth_file_names_list[i] = train_file_names_list[i].replace(\".tif\",\"_mask.png\").replace(\"images\", \"truth\")\n elif data_dir == \"augmented_images/\":\n truth_file_names_list[i] = train_file_names_list[i].replace(\".tif\",\".png\").replace(\"train\", \"truth\")\n else:\n pass\n\n n_truth_images = len(truth_file_names_list)\n\n for i in range(n_train_images):\n tri = train_file_names_list[i]\n tui = truth_file_names_list[i]\n print(tri + \" \" + tui)\n\n train_images = load_images(train_file_names_list)\n truth_images = load_images(truth_file_names_list)\n\n return train_images, truth_images\n\n\ndef transform_label(label):\n\n lw, lh, lc = label.shape\n\n flat_label = np.zeros((lw,lh))\n\n for i in range(lw):\n for j in range(lh):\n if (label[i, j, 0] >= label[i, j, 1]):\n flat_label[i, j] = 0\n else:\n flat_label[i,j] = 1\n\n return flat_label\n\n\ndef resize_image_list(image_list, n_width, n_height):\n\n n_images = len(image_list)\n\n resized_image_list = [None for i in range(n_images)]\n for i in range(n_images):\n resized_image_list[i] = cv2.resize(image_list[i], \n (n_width, n_height), \n interpolation = cv2.INTER_LINEAR)\n\n return resized_image_list\n\n\ndef convert_image_list_to_grayscale(image_list):\n \n n_images = len(image_list)\n\n gray_scale_image_list = [None for i in range(n_images)]\n for i in range(n_images):\n gray_scale_image_list[i] = cv2.cvtColor(image_list[i], cv2.COLOR_BGR2GRAY)\n\n return gray_scale_image_list\n\n\ndef plot_two_images(bgr, gs):\n\n rows = 1\n cols = 2\n f, axs = plt.subplots(rows, cols)\n\n plt.subplot(rows, cols, 1)\n if len(bgr.shape) == 3:\n plt.imshow(cv2.cvtColor(bgr, cv2.COLOR_BGR2RGB))\n else:\n plt.imshow(bgr)\n plt.colorbar()\n\n\n plt.subplot(rows, cols, 2)\n if len(gs.shape) == 3:\n plt.imshow(cv2.cvtColor(gs, cv2.COLOR_BGR2RGB))\n else:\n plt.imshow(gs)\n plt.colorbar()\n\n plt.show()\n\n\ndef plot_three_images(bgr, gs, img):\n\n rows = 1\n cols = 3\n f, axs = plt.subplots(rows, cols)\n\n plt.subplot(rows, cols, 1)\n if len(bgr.shape) == 3:\n plt.imshow(cv2.cvtColor(bgr, cv2.COLOR_BGR2RGB))\n else:\n plt.imshow(bgr)\n plt.colorbar()\n\n\n plt.subplot(rows, cols, 2)\n if len(gs.shape) == 3:\n plt.imshow(cv2.cvtColor(gs, cv2.COLOR_BGR2RGB))\n else:\n plt.imshow(gs)\n plt.colorbar()\n\n plt.subplot(rows, cols, 3)\n if len(gs.shape) == 3:\n plt.imshow(cv2.cvtColor(img, cv2.COLOR_BGR2RGB))\n else:\n plt.imshow(img)\n plt.colorbar()\n\n plt.show()\n\n\ndef augment_data(train_images, \n truth_images,\n nw_image=250,\n nh_image=250,\n nw_label=50,\n nh_label=50,\n augmented_images_dir=\"augmented_images/\",\n ap=[[0.0, 1.0], [180.0, 1.0]],\n save_images=True):\n\n # ap - augment parameters\n # ap[i][0] - rotation angle\n # ap[i][1] - scale\n\n is_dir = os.path.isdir(augmented_images_dir)\n if (is_dir == True):\n shutil.rmtree(augmented_images_dir)\n os.makedirs(augmented_images_dir)\n else:\n os.makedirs(augmented_images_dir)\n\n\n n_images = len(train_images)\n iw, ih, ic = train_images[0].shape\n mw, mh, _ = truth_images[0].shape\n\n\n if ((iw != mw) or (mh != mh)):\n sys.exit(\"ERROR: Dimension mismatch!\")\n\n for i in range(n_images):\n print(\"Image: \" + str(i) + \"/\" + str(n_images), end=\"\\r\")\n for j in range(len(ap)):\n\n rotation_angle = ap[j][0]\n scale = ap[j][1]\n\n R = cv2.getRotationMatrix2D((iw/2, ih/2), rotation_angle, scale)\n\n rotated_train_image = cv2.warpAffine(train_images[i], R, (iw, ih))\n rotated_truth_image = cv2.warpAffine(truth_images[i], R, (mw, mh))\n\n x_shift = ap[j][2]\n y_shift = ap[j][3]\n\n M = np.float32([[1, 0, x_shift],[0, 1, y_shift]])\n\n shifted_train_image = cv2.warpAffine(rotated_train_image, M, (iw, ih))\n shifted_truth_image = cv2.warpAffine(rotated_truth_image, M, (mw, mh))\n\n grayscale_truth_image = cv2.cvtColor(shifted_truth_image, cv2.COLOR_BGR2GRAY)\n\n resized_train_image = cv2.resize(shifted_train_image, \n (nw_image, nh_image), \n interpolation = cv2.INTER_LINEAR)\n\n resized_truth_image = cv2.resize(grayscale_truth_image, \n (nw_label, nh_label), \n interpolation = cv2.INTER_LINEAR)\n\n if save_images == True:\n cv2.imwrite(augmented_images_dir + \\\n \"train_image_id_\" + \\\n str(i) + \"_\" + \\\n str(rotation_angle).replace(\".\",\"p\") + \"_\" + \\\n str(x_shift) + \"_\" + \\\n str(y_shift) + \"_\" + \\\n str(scale).replace(\".\",\"p\") + \".tif\", \\\n shifted_train_image)\n\n cv2.imwrite(augmented_images_dir + \\\n \"truth_image_id_\" + \\\n str(i) + \"_\" + \\\n str(rotation_angle).replace(\".\",\"p\") + \"_\" + \\\n str(x_shift) + \"_\" + \\\n str(y_shift) + \"_\" + \\\n str(scale).replace(\".\",\"p\") + \".png\", \\\n shifted_truth_image)\n\n fn = augmented_images_dir + \\\n \"numpy_image_array_id_\" + \\\n str(i) + \"_\" + \\\n str(rotation_angle).replace(\".\",\"p\") + \"_\" + \\\n str(x_shift) + \"_\" + \\\n str(y_shift) + \"_\" + \\\n str(scale).replace(\".\",\"p\") + \".npz\"\n\n resized_truth_image[resized_truth_image > 0] = 1\n\n # This part is for FCN.\n # We destinguish between good and bad tissue so\n # we have just two classes.\n n_classes = 2\n resized_truth_image = to_categorical(resized_truth_image, n_classes)\n resized_truth_image = np.reshape(resized_truth_image, (nw_label, nh_label, n_classes))\n\n np.savez_compressed(fn, \n image=resized_train_image/255.0, \n #label=np.reshape(resized_truth_image, (nw_label*nh_label, 1)))\n label=resized_truth_image)\n\n\ndef load_data_from_npz(file_name_list):\n\n n_files = len(file_name_list)\n if n_files == 0:\n sys.exit(\"ERROR: File name list empty.\")\n\n loaded_data = np.load(file_name_list[0])\n image = loaded_data[\"image\"].astype(np.float32)\n label = loaded_data[\"label\"].astype(np.float32)\n\n iw, ih, ic = image.shape\n mw, mh, mc = label.shape\n\n x_data = np.zeros((n_files, ih, iw, ic))\n y_data = np.zeros((n_files, mw, mh, mc))\n\n x_data[0, :, :, :] = image\n y_data[0, :, :, :] = label\n\n for i in range(1, n_files):\n loaded_data = np.load(file_name_list[i])\n image = loaded_data[\"image\"].astype(np.float32)\n label = loaded_data[\"label\"].astype(np.float32)\n\n x_data[i, :, :, :] = image\n y_data[i, :, :, :] = label\n\n return x_data, y_data\n\n\n"},{"alpha_fraction":0.4591751992702484,"alphanum_fraction":0.48214781284332275,"avg_line_length":30.657894134521484,"blob_id":"374121e5cff7ffc912991d74cd71ff3bfeeea051","content_id":"5cd75ac02c858f66f2b6780c4a89ddb87814c4a5","detected_licenses":["MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3613,"license_type":"permissive","max_line_length":102,"num_lines":114,"path":"/prepare_test_data.py","repo_name":"diwadd/PathImageSegmentation","src_encoding":"UTF-8","text":"import random\nimport sys\nimport glob\nimport os\nimport shutil\n\nimport cv2\nimport numpy as np\nfrom keras import backend as K\nfrom keras.models import load_model\n\nfrom keras_fcn.layers import BilinearUpSampling2D\n\nimport data_handling as dh\nimport deep_models as dm\n\ndef load_and_resize_image(image_file_name,\n nw_image=250,\n nh_image=250):\n\n image = cv2.imread(image_file_name)\n\n resized_image = cv2.resize(image, \n (nw_image, nh_image), \n interpolation = cv2.INTER_LINEAR)\n return resized_image\n\n\ndef prepare_data_for_dispatch(model_file_name,\n test_images_list,\n nw_image=250,\n nh_image=250,\n nw_label=50,\n nh_label=50,\n images_dir=\"images_for_dispatch/\"):\n\n is_dir = os.path.isdir(images_dir)\n if (is_dir == True):\n shutil.rmtree(images_dir)\n os.makedirs(images_dir)\n else:\n os.makedirs(images_dir)\n\n K.get_session()\n model = load_model(model_file_name, custom_objects={\"BilinearUpSampling2D\": BilinearUpSampling2D})\n\n n_train_images = len(test_images_list)\n for i in range(n_train_images):\n print(\"i: \" + str(i + 1))\n resized_image = load_and_resize_image(test_images_list[i],\n nw_image=nw_image,\n nh_image=nh_image)\n\n # The the i000000 image indicator\n # from training/images/i000000.tif\n if test_images_list[i][0:3] == \"tra\":\n trunk = test_images_list[i][16:-4]\n elif test_images_list[i][0:3] == \"tes\":\n trunk = test_images_list[i][15:-4]\n else:\n pass\n\n print(trunk)\n fn = images_dir + str(trunk) + \".npz\"\n\n _, _, ic = resized_image.shape\n\n resized_image = np.reshape(resized_image, (1, nw_image, nh_image, ic))/255.0\n label = model.predict(resized_image)\n label = np.reshape(label, (nw_label, nh_label, 2))\n\n label = dh.transform_label(label)\n\n label = cv2.resize(label,\n (dh.DEFAULT_WIDTH, dh.DEFAULT_HEIGHT),\n interpolation = cv2.INTER_LINEAR)\n\n f = open(images_dir + trunk + \"_mask.txt\", \"w\")\n\n for v in range(dh.DEFAULT_WIDTH):\n for w in range(dh.DEFAULT_HEIGHT):\n if (label[v][w] > 0.5):\n f.write(\"1\")\n else:\n f.write(\"0\")\n f.write(\"\\n\")\n\n f.close()\n\n K.clear_session()\n\n\nif __name__ == \"__main__\":\n print(\"Starting main.\")\n\n # ------------------------------------------------------------------------\n # Load data \n # ------------------------------------------------------------------------\n\n train_images_list = glob.glob(\"training/images/*.tif\")\n test_images_list = glob.glob(\"testing/images/*.tif\")\n\n image_list = train_images_list + test_images_list\n\n # model_file_name = \"vgg16_16s_fcn_model_after_global_epoch_62.h5\"\n # model_file_name = \"vgg16_32s_fcn_model.h5\" \n model_file_name = \"model.h5\" \n prepare_data_for_dispatch(model_file_name,\n image_list,\n nw_image=500,\n nh_image=500,\n nw_label=500,\n nh_label=500,\n images_dir=\"images_for_dispatch/\")\n\n\n\n\n"},{"alpha_fraction":0.7676056623458862,"alphanum_fraction":0.8098591566085815,"avg_line_length":34,"blob_id":"886cca198d27f93f29955002481ecd8110115ef3","content_id":"d1ba44351efd1b61f94bb3d2e535c7f3f7224af7","detected_licenses":["MIT"],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":142,"license_type":"permissive","max_line_length":48,"num_lines":4,"path":"/README.md","repo_name":"diwadd/PathImageSegmentation","src_encoding":"UTF-8","text":"We implement the FCN VGG16 32s and 16s networks.\n\nWe use these networks for segmentation in the \nPathological Image Segmentation Challenge.\n\n\n"},{"alpha_fraction":0.4825242757797241,"alphanum_fraction":0.5543689131736755,"avg_line_length":20.35416603088379,"blob_id":"aba5e2cd9eff6a25eac205395748a145a57d15c3","content_id":"de509ba4265b79b89ff6f9df7989583f5a8716e0","detected_licenses":["MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1030,"license_type":"permissive","max_line_length":116,"num_lines":48,"path":"/categorical_crossentropy_testing.py","repo_name":"diwadd/PathImageSegmentation","src_encoding":"UTF-8","text":"import random\nimport sys\nimport glob\n\nimport cv2\nimport numpy as np\n\nfrom keras import backend as K\nfrom tensorflow.python.ops import math_ops\n\nimport deep_models as dm\n\n\ndef cross_entropy(t, p):\n return -t*np.log(p)\n\n\n\n\ntrue_arr = np.array([[[1.0, 0.0], [1.0, 0.0], [1.0, 0.0]], \n [[1.0, 0.0], [1.0, 0.0], [1.0, 0.0]],\n [[1.0, 0.0], [1.0, 0.0], [1.0, 0.0]],])\n\n\npred_arr = np.array([[[0.9, 0.1], [0.9, 0.1], [0.9, 0.1]], \n [[0.9, 0.1], [0.4, 0.6], [0.5, 0.5]],\n [[0.9, 0.1], [0.2, 0.8], [0.9, 0.1]]])\n\nprint(true_arr.shape)\n\n\nce = cross_entropy(true_arr, pred_arr)\nprint(\"ce: \" + str(ce))\n\nK.get_session()\n\ntrue_var = K.variable(true_arr)\npred_var = K.variable(pred_arr)\n\n# output = pred_var / math_ops.reduce_sum(pred_var, reduction_indices=len(pred_var.get_shape()) - 1, keep_dims=True)\n# print(K.eval(output))\n\nen = K.categorical_crossentropy(pred_var, true_var, from_logits=False)\n\nen_eval = K.eval(en)\nprint(en_eval)\n\nK.clear_session()\n\n\n\n\n\n"}],"string":"[\n {\n \"alpha_fraction\": 0.5162613987922668,\n \"alphanum_fraction\": 0.5376899838447571,\n \"avg_line_length\": 27.236051559448242,\n \"blob_id\": \"f4f5e6d5ae43ac3cdc357185b87dd2acf5131652\",\n \"content_id\": \"54ba787e17929fc6556bfb7c5fb5507a2faeb89d\",\n \"detected_licenses\": [\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 6580,\n \"license_type\": \"permissive\",\n \"max_line_length\": 80,\n \"num_lines\": 233,\n \"path\": \"/main_training.py\",\n \"repo_name\": \"diwadd/PathImageSegmentation\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import random\\nimport sys\\nimport glob\\n\\nimport numpy as np\\n\\nfrom keras import backend as K\\nfrom keras.models import load_model\\nfrom keras.applications.vgg16 import VGG16\\nfrom keras import optimizers\\n\\nimport tensorflow as tf\\n\\nimport data_handling as dh\\nimport deep_models as dm\\n\\nrandom.seed(111)\\n\\n\\n\\ndef data_generator(file_name_list, noli=20):\\n \\n \\\"\\\"\\\"\\n noli - number of loaded images per yield\\n \\n \\\"\\\"\\\"\\n\\n while True:\\n n = len(file_name_list)\\n number_of_image_loads = round(n / noli)\\n ptr = 0\\n # print(\\\"n: \\\" + str(n))\\n # print(\\\"number_of_image_loads: \\\" + str(number_of_image_loads))\\n\\n for i in range(number_of_image_loads):\\n # print(\\\"We are a i: \\\" + str(i))\\n # create numpy arrays of input data\\n # and labels, from each line in the file\\n mini_batch_fnl = file_name_list[ptr:(ptr + noli)]\\n ptr = ptr + noli\\n\\n x_data, y_data = dh.load_data_from_npz(mini_batch_fnl)\\n yield (x_data, y_data)\\n\\n\\ndef evaluate(model, file_name_list, noli=20):\\n \\n \\\"\\\"\\\"\\n noli - number of loaded images per yield\\n \\n \\\"\\\"\\\"\\n\\n n = len(file_name_list)\\n number_of_image_loads = round(n / noli)\\n ptr = 0\\n # print(\\\"n: \\\" + str(n))\\n # print(\\\"number_of_image_loads: \\\" + str(number_of_image_loads))\\n\\n mean_loss = 0.0\\n mean_acc = 0.0\\n for i in range(number_of_image_loads):\\n # print(\\\"We are a i: \\\" + str(i))\\n # create numpy arrays of input data\\n # and labels, from each line in the file\\n mini_batch_fnl = file_name_list[ptr:(ptr + noli)]\\n ptr = ptr + noli\\n\\n x_data, y_data = dh.load_data_from_npz(mini_batch_fnl)\\n score = model.evaluate(x_data, y_data, verbose=0)\\n\\n #print(\\\"score: \\\" + str(score))\\n\\n local_loss = score[0]\\n local_acc = score[1]\\n\\n mean_loss += local_loss\\n mean_acc += local_acc\\n mean_loss /= number_of_image_loads\\n mean_acc /= number_of_image_loads\\n print(\\\"Mean loss: \\\" + str(mean_loss) + \\\" - mean acc: \\\" + str(mean_acc))\\n \\n\\n\\nif __name__ == \\\"__main__\\\":\\n print(\\\"Starting main.\\\")\\n\\n # ------------------------------------------------------------------------\\n # Load data\\n # ------------------------------------------------------------------------\\n\\n data_file_names = glob.glob(\\\"augmented_images/*npz\\\")\\n\\n n_files = len(data_file_names)\\n\\n loaded_data = np.load(data_file_names[0])\\n image = loaded_data[\\\"image\\\"]\\n label = loaded_data[\\\"label\\\"]\\n\\n iw, ih, ic = image.shape\\n nw, nh, nc= label.shape\\n\\n random.shuffle(data_file_names)\\n train_fraction = 0.6\\n valid_fraction = 1.0 - train_fraction\\n test_fraction = 0.5\\n\\n x_train_fnl = data_file_names[0:int(train_fraction*n_files)]\\n temp = data_file_names[int(train_fraction*n_files):]\\n x_valid_fnl = temp[0:int(len(temp)/2)]\\n x_test_fnl = temp[int(len(temp)/2):]\\n\\n print(\\\"First train file: \\\" + str(x_train_fnl[0]))\\n\\n print(\\\"Number of train data files: \\\" + str(len(x_train_fnl)))\\n print(\\\"Number of valid data files: \\\" + str(len(x_valid_fnl)))\\n print(\\\"Number of test data files: \\\" + str(len(x_test_fnl)))\\n\\n K.get_session()\\n\\n\\n new_model = False\\n if (new_model == True):\\n print(\\\"Creating a new model!\\\")\\n\\n\\n #model = FCN(input_shape=(iw, ih, ic), classes=2, \\n # weights='imagenet', trainable_encoder=True)\\n\\n model = dm.vgg16_16s_fcn(iw, \\n ih, \\n ic,\\n dropout=0.5,\\n alpha=0.0001,\\n classes=2)\\n\\n opt = optimizers.SGD(lr=0.0001, momentum=0.9, clipvalue=0.5)\\n opt = optimizers.Adam(1e-4, clipvalue=0.5)\\n model.compile(loss=\\\"categorical_crossentropy\\\",\\n optimizer=opt,\\n metrics=[\\\"accuracy\\\"])\\n\\n\\n vgg16_model = VGG16(weights='imagenet', include_top=False)\\n vgg16_model.compile(loss=\\\"categorical_crossentropy\\\",\\n optimizer=\\\"adadelta\\\")\\n\\n index = 12\\n print(\\\"Our model config: \\\")\\n print(model.layers[index].get_config())\\n\\n print(\\\"VGG16 model config: \\\")\\n print(vgg16_model.layers[index].get_config())\\n\\n print(\\\"Our model:\\\")\\n layer1 = model.layers[index].get_weights()\\n print(layer1[0].shape)\\n print(layer1[0][0,0,0,1:5])\\n\\n print(\\\"VGG16 model:\\\")\\n layer1 = vgg16_model.layers[index].get_weights()\\n print(layer1[0].shape)\\n print(layer1[0][0,0,0,1:5])\\n\\n\\n for i in range(len(vgg16_model.layers)):\\n\\n print(\\\"\\\\n\\\\n\\\\n Layer number: \\\" + str(i))\\n\\n layer_config = vgg16_model.layers[i].get_config()\\n print(\\\"Setting name: \\\" + str(layer_config['name']))\\n if layer_config['name'] == 'block5_pool':\\n break\\n\\n layer_weights = vgg16_model.layers[i].get_weights()\\n\\n if len(layer_weights) != 0:\\n print(layer1[0].shape)\\n model.layers[i].set_weights(vgg16_model.layers[i].get_weights())\\n else:\\n print(\\\"No wieghts to set!\\\")\\n\\n print(\\\"After setting the weights...\\\")\\n\\n print(\\\"Our model:\\\")\\n layer1 = model.layers[index].get_weights()\\n print(layer1[0].shape)\\n print(layer1[0][0,0,0,1:5])\\n\\n print(\\\"VGG16 model:\\\")\\n layer1 = vgg16_model.layers[index].get_weights()\\n print(layer1[0].shape)\\n print(layer1[0][0,0,0,1:5])\\n\\n\\n\\n #model.compile(optimizer='adadelta',\\n # loss='categorical_crossentropy',\\n # metrics=['accuracy'])\\n\\n model.summary()\\n\\n\\n else:\\n model = load_model(\\\"model.h5\\\")\\n print(\\\"Model loaded!\\\")\\n model.summary()\\n\\n noli = 4\\n n = len(x_train_fnl)\\n number_of_image_loads = round(n / noli)\\n print(\\\"Number of image loads: \\\" + str(number_of_image_loads))\\n n_epochs = 100 \\n n_sub_epochs = 5\\n\\n print(\\\"Pre training evaluation:\\\")\\n evaluate(model, x_valid_fnl, noli=4)\\n\\n for i in range(n_epochs):\\n print(\\\"Global epoch: \\\" + str(i) + \\\"/\\\" + str(n_epochs))\\n model.fit_generator(data_generator(data_file_names, noli=noli),\\n steps_per_epoch=number_of_image_loads,\\n epochs=n_sub_epochs)\\n print(\\\"Validation data mean loss: \\\")\\n evaluate(model, x_valid_fnl, noli=4)\\n\\n\\n model.save(\\\"model.h5\\\")\\n\\n print(\\\"Test data mean loss: \\\")\\n evaluate(model, x_test_fnl, noli=64)\\n\\n\\n K.clear_session()\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.3254545331001282,\n \"alphanum_fraction\": 0.4509090781211853,\n \"avg_line_length\": 17.33333396911621,\n \"blob_id\": \"d43756e65b061b6515c1c210768cfda3c2a3006e\",\n \"content_id\": \"e96cd63a8c5fb40927d1ca53e57575a5d8d16330\",\n \"detected_licenses\": [\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 550,\n \"license_type\": \"permissive\",\n \"max_line_length\": 74,\n \"num_lines\": 30,\n \"path\": \"/feeling_binary_crossentropy.py\",\n \"repo_name\": \"diwadd/PathImageSegmentation\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import numpy as np\\n\\n\\ndef bc(t, o):\\n return -np.mean(t*np.log(o) + (1-t)*np.log(1-o))\\n\\n\\nt = np.array([ 0, 0, 0, 1, 1, 1, 0, 0, 0, 1])\\no = np.array([0.10, 0.10, 0.10, 0.95, 0.90, 0.90, 0.10, 0.10, 0.10, 0.90])\\n\\nprint(bc(t, o))\\n\\n\\nn = 10000\\nt = np.random.randint(2, size=n)\\no = np.array([0.01 for i in range(n)])\\n\\nfor i in range(n):\\n if t[i] == 1:\\n o[i] = 0.99\\n\\nfor i in range(int(0.013*n)):\\n if t[i] == 1:\\n o[i] = 0.001\\n elif t[i] == 0:\\n o[i] = 0.999\\n else:\\n pass\\n\\nprint(bc(t, o))\\n\"\n },\n {\n \"alpha_fraction\": 0.5835299491882324,\n \"alphanum_fraction\": 0.6379121541976929,\n \"avg_line_length\": 47.25833511352539,\n \"blob_id\": \"4d97324a8ad7a9c93d3b2aac198fdd0e3cabdfb1\",\n \"content_id\": \"4dcefe5660c67311c4f3a21fb3bb58b36f0c330d\",\n \"detected_licenses\": [\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 17377,\n \"license_type\": \"permissive\",\n \"max_line_length\": 194,\n \"num_lines\": 360,\n \"path\": \"/deep_models.py\",\n \"repo_name\": \"diwadd/PathImageSegmentation\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import keras\\n\\nfrom keras.models import Model\\nfrom keras.models import Sequential\\nfrom keras.layers import Flatten\\nfrom keras.layers import Dense\\nfrom keras.layers import Input\\nfrom keras.layers import Conv2D\\nfrom keras.layers import Conv2DTranspose\\nfrom keras.layers import Cropping2D\\nfrom keras.layers import MaxPooling2D\\nfrom keras.layers import GlobalMaxPooling2D\\nfrom keras.layers import Activation\\nfrom keras.layers import GlobalAveragePooling2D\\n\\n\\nfrom keras.layers import Dropout \\nfrom keras.layers import BatchNormalization\\nfrom keras.layers.advanced_activations import LeakyReLU\\n\\nfrom keras import regularizers\\nfrom keras import backend as K\\n\\ndef basic_model(iw=500, # Input width\\n ih=500, # Input height \\n ic=3,\\n ow=100, # Output width\\n oh=100, # Output heigth\\n dropout=0.9,\\n alpha=0.0):\\n\\n input_image = Input((iw, ih, ic))\\n\\n x = Conv2D(16, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_1_layer_1\\\")(input_image)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(16, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_1_layer_2\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\\\"block_1_pooling\\\")(x)\\n\\n x = Conv2D(32, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_2_layer_1\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(32, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_2_layer_2\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\\\"block_2_pooling\\\")(x)\\n\\n x = Conv2D(64, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_3_layer_1\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(64, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_3_layer_2\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(64, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_3_layer_3\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\\\"block_3_pooling\\\")(x)\\n\\n x = Conv2D(128, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_4_layer_1\\\")(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(128, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_4_layer_2\\\")(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(128, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_4_layer_3\\\")(x)\\n x = LeakyReLU(alpha)(x)\\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\\\"block_4_pooling\\\")(x)\\n\\n x = Conv2D(128, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_5_layer_1\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(128, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_5_layer_2\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(128, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_5_layer_3\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\\\"block_5_pooling\\\")(x)\\n\\n x = Conv2D(256, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_7_layer_1\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(256, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_7_layer_2\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(256, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_7_layer_3\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\\\"block_7_pooling\\\")(x)\\n\\n x = Flatten(name=\\\"flatten\\\")(x)\\n x = Dense(4096, activation=\\\"linear\\\", name=\\\"full_connected_1\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Dropout(dropout)(x)\\n\\n x = Dense(4096, activation=\\\"linear\\\", name=\\\"full_connected_2\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Dropout(dropout)(x)\\n\\n x = Dense(ow*oh, activation=\\\"sigmoid\\\", name=\\\"predictions\\\")(x)\\n\\n model = Model(input_image, x, name=\\\"vgg16_based\\\")\\n\\n model.compile(loss=\\\"binary_crossentropy\\\",\\n optimizer=\\\"adadelta\\\")\\n\\n print(\\\"\\\\n ---> Model summary <--- \\\\n\\\")\\n model.summary()\\n\\n return model\\n\\ndef basic_model_pooling(iw=500, # Input width\\n ih=500, # Input height \\n ic=3,\\n ow=100, # Output width\\n oh=100, # Output heigth\\n dropout=0.9,\\n alpha=0.001):\\n\\n input_image = Input((iw, ih, ic))\\n\\n x = Conv2D(16, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_1_layer_1\\\")(input_image)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(16, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_1_layer_2\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\\\"block_1_pooling\\\")(x)\\n\\n x = Conv2D(32, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_2_layer_1\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(32, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_2_layer_2\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\\\"block_2_pooling\\\")(x)\\n\\n x = Conv2D(64, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_3_layer_1\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(64, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_3_layer_2\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(64, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_3_layer_3\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\\\"block_3_pooling\\\")(x)\\n\\n x = Conv2D(128, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_4_layer_1\\\")(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(128, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_4_layer_2\\\")(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(128, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_4_layer_3\\\")(x)\\n x = LeakyReLU(alpha)(x)\\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\\\"block_4_pooling\\\")(x)\\n\\n x = Conv2D(128, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_5_layer_1\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(128, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_5_layer_2\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(128, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_5_layer_3\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = MaxPooling2D((2, 2), strides=(2, 2), name=\\\"block_5_pooling\\\")(x)\\n\\n x = Conv2D(256, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_6_layer_1\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(512, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_6_layer_2\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n x = Conv2D(ow*oh, (3, 3), activation=\\\"linear\\\", padding=\\\"same\\\", name=\\\"block_6_layer_3\\\")(x)\\n x = BatchNormalization()(x)\\n x = LeakyReLU(alpha)(x)\\n\\n x = GlobalAveragePooling2D()(x)\\n\\n model = Model(input_image, x, name=\\\"vgg16_based\\\")\\n\\n model.compile(loss=\\\"binary_crossentropy\\\",\\n optimizer=\\\"adadelta\\\")\\n\\n print(\\\"\\\\n ---> Model summary <--- \\\\n\\\")\\n model.summary()\\n\\n return model\\n\\n\\ndef vgg16_32s_fcn(iw=500, # Input width\\n ih=500, # Input height \\n ic=3,\\n dropout=0.5,\\n alpha=0.001,\\n classes=2):\\n # Based on:\\n # Fully Convolutional Models for Semantic Segmentation\\n # Evan Shelhamer*, Jonathan Long*, Trevor Darrell\\n # PAMI 2016\\n # arXiv:1605.06211\\n\\n reg_fun = regularizers.l2(alpha)\\n\\n input_image = Input((iw, ih, ic))\\n\\n # Conv 1\\n x = Conv2D(64, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block1_conv1')(input_image)\\n x = Conv2D(64, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block1_conv2')(x)\\n pool1 = MaxPooling2D((2, 2), strides=(2, 2), name='block1_pool')(x)\\n\\n # Conv 2\\n x = Conv2D(128, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block2_conv1')(pool1)\\n x = Conv2D(128, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block2_conv2')(x)\\n pool2 = MaxPooling2D((2, 2), strides=(2, 2), name='block2_pool')(x)\\n\\n # Conv 3\\n x = Conv2D(256, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block3_conv1')(pool2)\\n x = Conv2D(256, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block3_conv2')(x)\\n x = Conv2D(256, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block3_conv3')(x)\\n pool3 = MaxPooling2D((2, 2), strides=(2, 2), name='block3_pool')(x)\\n\\n # Conv 4\\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block4_conv1')(pool3)\\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block4_conv2')(x)\\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block4_conv3')(x)\\n pool4 = MaxPooling2D((2, 2), strides=(2, 2), name='block4_pool')(x)\\n\\n # Conv 5\\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block5_conv1')(pool4)\\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block5_conv2')(x)\\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block5_conv3')(x)\\n pool5 = MaxPooling2D((2, 2), strides=(2, 2), name='block5_pool')(x)\\n\\n # Fully Conv fc6\\n fc6 = Conv2D(4096, (7, 7), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='fc6')(pool5)\\n drop6 = Dropout(rate=dropout)(fc6)\\n\\n # Fully Conv fc7\\n fc7 = Conv2D(4096, (1, 1), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='fc7')(drop6)\\n drop7 = Dropout(rate=dropout)(fc7)\\n\\n score_fr = Conv2D(classes, (1, 1), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='score_fr')(drop7)\\n\\n upscore = Conv2DTranspose(classes, kernel_size=(64, 64), strides=(32, 32), kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='upscore')(score_fr)\\n\\n\\n _, uw, uh, uc = upscore._keras_shape\\n cw = (uw - iw)//2\\n ch = (uh - ih)//2\\n print(\\\"cw: \\\" + str(cw))\\n print(\\\"ch: \\\" + str(ch))\\n\\n score = Cropping2D(cropping=(cw, ch))(upscore)\\n output = Activation('softmax')(score)\\n\\n model = Model(input_image, output, name=\\\"vgg16_based\\\")\\n\\n return model\\n\\n\\ndef vgg16_16s_fcn(iw=500, # Input width\\n ih=500, # Input height \\n ic=3,\\n dropout=0.5,\\n alpha=0.001,\\n classes=2):\\n # Based on:\\n # Fully Convolutional Models for Semantic Segmentation\\n # Evan Shelhamer*, Jonathan Long*, Trevor Darrell\\n # PAMI 2016\\n # arXiv:1605.06211\\n\\n reg_fun = regularizers.l2(alpha)\\n\\n input_image = Input((iw, ih, ic))\\n\\n # Conv 1\\n x = Conv2D(64, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block1_conv1')(input_image)\\n x = Conv2D(64, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block1_conv2')(x)\\n pool1 = MaxPooling2D((2, 2), strides=(2, 2), name='block1_pool')(x)\\n\\n # Conv 2\\n x = Conv2D(128, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block2_conv1')(pool1)\\n x = Conv2D(128, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block2_conv2')(x)\\n pool2 = MaxPooling2D((2, 2), strides=(2, 2), name='block2_pool')(x)\\n\\n # Conv 3\\n x = Conv2D(256, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block3_conv1')(pool2)\\n x = Conv2D(256, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block3_conv2')(x)\\n x = Conv2D(256, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block3_conv3')(x)\\n pool3 = MaxPooling2D((2, 2), strides=(2, 2), name='block3_pool')(x)\\n\\n # Conv 4\\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block4_conv1')(pool3)\\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block4_conv2')(x)\\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block4_conv3')(x)\\n pool4 = MaxPooling2D((2, 2), strides=(2, 2), name='block4_pool')(x)\\n\\n # Conv 5\\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block5_conv1')(pool4)\\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block5_conv2')(x)\\n x = Conv2D(512, (3, 3), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='block5_conv3')(x)\\n pool5 = MaxPooling2D((2, 2), strides=(2, 2), name='block5_pool')(x)\\n\\n # Fully Conv fc6\\n fc6 = Conv2D(4096, (7, 7), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='fc6')(pool5)\\n drop6 = Dropout(rate=dropout)(fc6)\\n\\n # Fully Conv fc7\\n fc7 = Conv2D(4096, (1, 1), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='fc7')(drop6)\\n drop7 = Dropout(rate=dropout)(fc7)\\n\\n score_fr = Conv2D(classes, (1, 1), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='score_fr')(drop7)\\n upscore2 = Conv2DTranspose(classes, kernel_size=(4, 4), strides=(2, 2), kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='upscore2')(score_fr)\\n\\n\\n score_pool4 = Conv2D(classes, (1, 1), activation='relu', padding='same', kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='score_pool4')(pool4)\\n\\n _, uw, uh, uc = upscore2._keras_shape\\n _, sw, sh, sc = score_pool4._keras_shape\\n\\n if ((uw - sw) == 1) or ((uh - sh) == 1):\\n cw1 = 1\\n ch1 = 1\\n cropping = ((cw1, 0),(ch1, 0))\\n else:\\n cw1 = (uw - sw)//2\\n ch1 = (uh - sh)//2\\n cropping = (cw1, ch1)\\n\\n print(\\\"cw1: \\\" + str(cw1))\\n print(\\\"ch1: \\\" + str(ch1))\\n\\n print(\\\"upscore2._keras_shape \\\" + str(upscore2._keras_shape))\\n print(\\\"score_pool4._keras_shape \\\" + str(score_pool4._keras_shape))\\n\\n # Technically score_pool4 should have a larger size then upscore2.\\n # At least that is what follows from crop(n.score_pool4, n.upscore2).\\n # This is, however, not the case and we nned to crop upscore2.\\n\\n score_pool4c = Cropping2D(cropping=cropping)(upscore2) \\n fuse_pool4 = keras.layers.Add()([score_pool4c, score_pool4])\\n\\n upscore16 = Conv2DTranspose(classes, kernel_size=(32, 32), strides=(16, 16), kernel_regularizer=regularizers.l2(alpha), bias_regularizer=regularizers.l2(alpha), name='upscore16')(fuse_pool4)\\n\\n _, uw, uh, uc = upscore16._keras_shape\\n cw2 = (uw - iw)//2\\n ch2 = (uh - ih)//2\\n #print(\\\"cw2: \\\" + str(cw2))\\n #print(\\\"ch2: \\\" + str(ch2))\\n\\n score = Cropping2D(cropping=(cw2, ch2))(upscore16)\\n output = Activation('softmax')(score)\\n\\n model = Model(input_image, output, name=\\\"vgg16_based\\\")\\n\\n return model\\n\\n\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5284159779548645,\n \"alphanum_fraction\": 0.6118500828742981,\n \"avg_line_length\": 29.518518447875977,\n \"blob_id\": \"b84f2279d5954ca6a997eb93f107c605266f5e34\",\n \"content_id\": \"4524546c634058dfdda3bc02f97a8e1542f85f12\",\n \"detected_licenses\": [\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 827,\n \"license_type\": \"permissive\",\n \"max_line_length\": 78,\n \"num_lines\": 27,\n \"path\": \"/check_numpy_saved_array.py\",\n \"repo_name\": \"diwadd/PathImageSegmentation\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import cv2\\nimport numpy as np\\n\\nimport matplotlib.pyplot as plt\\nimport data_handling as dh\\n\\nif __name__ == \\\"__main__\\\":\\n\\n #file_name = \\\"augmented_images/numpy_image_array_id_0_0p0_30_30_0p85.npz\\\"\\n file_name = \\\"augmented_images/numpy_image_array_id_0_3p0_10_10_1p0.npz\\\"\\n\\n loaded_data = np.load(file_name)\\n\\n image = (255.0*loaded_data[\\\"image\\\"]).astype(np.uint8)\\n label = (255.0*loaded_data[\\\"label\\\"]).astype(np.uint8)\\n #label = np.reshape(label, (100, 100))\\n\\n print(\\\"label shape: \\\" + str(label.shape))\\n\\n #label = cv2.resize(label,\\n # (500, 500),\\n # interpolation = cv2.INTER_LINEAR)\\n #mask = cv2.imread(\\\"augmented_images/truth_image_id_0_0p0_30_30_0p85.png\\\")\\n\\n dh.plot_three_images(image, label[:, :, 0], label[:, :, 1])\\n\\n print(label[100:120, 100:120])\\n\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5060120224952698,\n \"alphanum_fraction\": 0.5751503109931946,\n \"avg_line_length\": 22.447059631347656,\n \"blob_id\": \"fff37a204466de8854984b1f7e84c138a67c3e4d\",\n \"content_id\": \"5294bab4360471e30e63e97980a084e7aa3c5f95\",\n \"detected_licenses\": [\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1996,\n \"license_type\": \"permissive\",\n \"max_line_length\": 116,\n \"num_lines\": 85,\n \"path\": \"/temp.py\",\n \"repo_name\": \"diwadd/PathImageSegmentation\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import random\\nimport sys\\nimport glob\\n\\nimport cv2\\nimport numpy as np\\n\\nimport keras\\n\\n#from keras import backend as K\\n#from keras.models import load_model\\n#from tensorflow.python.ops import math_ops\\n\\n#import deep_models as dm\\n\\n\\n# model = dm.vg16_fcn()\\n# model.summary()\\n# model.compile(loss=\\\"categorical_crossentropy\\\",\\n# optimizer=\\\"adadelta\\\")\\n\\n\\\"\\\"\\\"\\n\\ndef cross_entropy(t, p):\\n return -np.sum(t*np.log(p))\\n\\n\\n\\n\\ntrue_arr = np.array([[[1.0, 0.0], [1.0, 0.0], [1.0, 0.0]], \\n [[1.0, 0.0], [1.0, 0.0], [1.0, 0.0]],\\n [[1.0, 0.0], [1.0, 0.0], [1.0, 0.0]],])\\n\\n\\npred_arr = np.array([[[0.9, 0.1], [0.9, 0.1], [0.9, 0.1]], \\n [[0.9, 0.1], [0.99, 0.01], [0.99, 0.01]],\\n [[0.9, 0.1], [0.99, 0.01], [0.9, 0.01]]])\\n\\npred_npr = np.array([[[9.9, 1.1], [8.9, 0.1], [1.9, 6.1]], \\n [[9.9, 2.1], [9.4, 9.6], [2.5, 5.5]],\\n [[9.9, 1.1], [9.2, 0.8], [3.9, 4.1]]])\\n\\nprint(true_arr.shape)\\n\\n\\nce = cross_entropy(true_arr, pred_arr)\\nprint(\\\"ce: \\\" + str(ce))\\n\\n\\\"\\\"\\\"\\n\\n\\n\\\"\\\"\\\"\\nK.get_session()\\n\\ntrue_var = K.variable(true_arr)\\npred_var = K.variable(pred_arr)\\npred_npr = K.variable(pred_npr)\\n\\n# output = pred_var / math_ops.reduce_sum(pred_var, reduction_indices=len(pred_var.get_shape()) - 1, keep_dims=True)\\n# print(K.eval(output))\\n\\nsx = K.softmax(pred_npr)\\nen = K.categorical_crossentropy(pred_var, true_var, from_logits=False)\\n\\nsx_eval = K.eval(sx)\\nprint(\\\"sx_eval\\\")\\nprint(sx_eval)\\n\\nen_eval = K.eval(en)\\nprint(\\\"en_eval\\\")\\nprint(en_eval)\\n\\nK.clear_session()\\n\\\"\\\"\\\"\\n\\ninput1 = keras.layers.Input(shape=(16,))\\nx1 = keras.layers.Dense(8, activation='relu')(input1)\\ninput2 = keras.layers.Input(shape=(32,))\\nx2 = keras.layers.Dense(8, activation='relu')(input2)\\nadded = keras.layers.Add()([x1, x2]) # equivalent to added = keras.layers.add([x1, x2])\\n\\nprint(\\\"added._keras_shape: \\\" + str(added._keras_shape)) \\n\\nout = keras.layers.Dense(4)(added)\\nmodel = keras.models.Model(inputs=[input1, input2], outputs=out)\\n\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5030834674835205,\n \"alphanum_fraction\": 0.5170778036117554,\n \"avg_line_length\": 28.787986755371094,\n \"blob_id\": \"f3101c7144555b606110f8abc8fb73279bfa08e9\",\n \"content_id\": \"e4a92105596c7d56ed2c7056c03044752ad9ec20\",\n \"detected_licenses\": [\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 8432,\n \"license_type\": \"permissive\",\n \"max_line_length\": 118,\n \"num_lines\": 283,\n \"path\": \"/data_handling.py\",\n \"repo_name\": \"diwadd/PathImageSegmentation\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import glob\\nimport cv2\\nimport math\\nimport sys\\nimport os\\nimport shutil\\n\\nimport numpy as np\\nimport matplotlib.pyplot as plt\\n\\nfrom keras.utils.np_utils import to_categorical\\n\\nDEFAULT_WIDTH = 500\\nDEFAULT_HEIGHT = 500\\n\\ndef load_images(file_names_list):\\n\\n n_images = len(file_names_list)\\n\\n images = [None for i in range(n_images)]\\n for i in range(n_images):\\n images[i] = (cv2.imread(file_names_list[i]))\\n\\n return images\\n\\n\\ndef read_data(data_dir=\\\"training/images/\\\"): \\n\\n train_file_names_list = glob.glob(data_dir + \\\"*.tif\\\")\\n n_train_images = len(train_file_names_list)\\n\\n # The truth images have the same numbers.\\n # We change the path and the extensions.\\n # This ensures that train image will\\n # correspond to truth image.\\n truth_file_names_list = [None for i in range(n_train_images)]\\n for i in range(n_train_images):\\n\\n if data_dir == \\\"training/images/\\\":\\n truth_file_names_list[i] = train_file_names_list[i].replace(\\\".tif\\\",\\\"_mask.png\\\").replace(\\\"images\\\", \\\"truth\\\")\\n elif data_dir == \\\"augmented_images/\\\":\\n truth_file_names_list[i] = train_file_names_list[i].replace(\\\".tif\\\",\\\".png\\\").replace(\\\"train\\\", \\\"truth\\\")\\n else:\\n pass\\n\\n n_truth_images = len(truth_file_names_list)\\n\\n for i in range(n_train_images):\\n tri = train_file_names_list[i]\\n tui = truth_file_names_list[i]\\n print(tri + \\\" \\\" + tui)\\n\\n train_images = load_images(train_file_names_list)\\n truth_images = load_images(truth_file_names_list)\\n\\n return train_images, truth_images\\n\\n\\ndef transform_label(label):\\n\\n lw, lh, lc = label.shape\\n\\n flat_label = np.zeros((lw,lh))\\n\\n for i in range(lw):\\n for j in range(lh):\\n if (label[i, j, 0] >= label[i, j, 1]):\\n flat_label[i, j] = 0\\n else:\\n flat_label[i,j] = 1\\n\\n return flat_label\\n\\n\\ndef resize_image_list(image_list, n_width, n_height):\\n\\n n_images = len(image_list)\\n\\n resized_image_list = [None for i in range(n_images)]\\n for i in range(n_images):\\n resized_image_list[i] = cv2.resize(image_list[i], \\n (n_width, n_height), \\n interpolation = cv2.INTER_LINEAR)\\n\\n return resized_image_list\\n\\n\\ndef convert_image_list_to_grayscale(image_list):\\n \\n n_images = len(image_list)\\n\\n gray_scale_image_list = [None for i in range(n_images)]\\n for i in range(n_images):\\n gray_scale_image_list[i] = cv2.cvtColor(image_list[i], cv2.COLOR_BGR2GRAY)\\n\\n return gray_scale_image_list\\n\\n\\ndef plot_two_images(bgr, gs):\\n\\n rows = 1\\n cols = 2\\n f, axs = plt.subplots(rows, cols)\\n\\n plt.subplot(rows, cols, 1)\\n if len(bgr.shape) == 3:\\n plt.imshow(cv2.cvtColor(bgr, cv2.COLOR_BGR2RGB))\\n else:\\n plt.imshow(bgr)\\n plt.colorbar()\\n\\n\\n plt.subplot(rows, cols, 2)\\n if len(gs.shape) == 3:\\n plt.imshow(cv2.cvtColor(gs, cv2.COLOR_BGR2RGB))\\n else:\\n plt.imshow(gs)\\n plt.colorbar()\\n\\n plt.show()\\n\\n\\ndef plot_three_images(bgr, gs, img):\\n\\n rows = 1\\n cols = 3\\n f, axs = plt.subplots(rows, cols)\\n\\n plt.subplot(rows, cols, 1)\\n if len(bgr.shape) == 3:\\n plt.imshow(cv2.cvtColor(bgr, cv2.COLOR_BGR2RGB))\\n else:\\n plt.imshow(bgr)\\n plt.colorbar()\\n\\n\\n plt.subplot(rows, cols, 2)\\n if len(gs.shape) == 3:\\n plt.imshow(cv2.cvtColor(gs, cv2.COLOR_BGR2RGB))\\n else:\\n plt.imshow(gs)\\n plt.colorbar()\\n\\n plt.subplot(rows, cols, 3)\\n if len(gs.shape) == 3:\\n plt.imshow(cv2.cvtColor(img, cv2.COLOR_BGR2RGB))\\n else:\\n plt.imshow(img)\\n plt.colorbar()\\n\\n plt.show()\\n\\n\\ndef augment_data(train_images, \\n truth_images,\\n nw_image=250,\\n nh_image=250,\\n nw_label=50,\\n nh_label=50,\\n augmented_images_dir=\\\"augmented_images/\\\",\\n ap=[[0.0, 1.0], [180.0, 1.0]],\\n save_images=True):\\n\\n # ap - augment parameters\\n # ap[i][0] - rotation angle\\n # ap[i][1] - scale\\n\\n is_dir = os.path.isdir(augmented_images_dir)\\n if (is_dir == True):\\n shutil.rmtree(augmented_images_dir)\\n os.makedirs(augmented_images_dir)\\n else:\\n os.makedirs(augmented_images_dir)\\n\\n\\n n_images = len(train_images)\\n iw, ih, ic = train_images[0].shape\\n mw, mh, _ = truth_images[0].shape\\n\\n\\n if ((iw != mw) or (mh != mh)):\\n sys.exit(\\\"ERROR: Dimension mismatch!\\\")\\n\\n for i in range(n_images):\\n print(\\\"Image: \\\" + str(i) + \\\"/\\\" + str(n_images), end=\\\"\\\\r\\\")\\n for j in range(len(ap)):\\n\\n rotation_angle = ap[j][0]\\n scale = ap[j][1]\\n\\n R = cv2.getRotationMatrix2D((iw/2, ih/2), rotation_angle, scale)\\n\\n rotated_train_image = cv2.warpAffine(train_images[i], R, (iw, ih))\\n rotated_truth_image = cv2.warpAffine(truth_images[i], R, (mw, mh))\\n\\n x_shift = ap[j][2]\\n y_shift = ap[j][3]\\n\\n M = np.float32([[1, 0, x_shift],[0, 1, y_shift]])\\n\\n shifted_train_image = cv2.warpAffine(rotated_train_image, M, (iw, ih))\\n shifted_truth_image = cv2.warpAffine(rotated_truth_image, M, (mw, mh))\\n\\n grayscale_truth_image = cv2.cvtColor(shifted_truth_image, cv2.COLOR_BGR2GRAY)\\n\\n resized_train_image = cv2.resize(shifted_train_image, \\n (nw_image, nh_image), \\n interpolation = cv2.INTER_LINEAR)\\n\\n resized_truth_image = cv2.resize(grayscale_truth_image, \\n (nw_label, nh_label), \\n interpolation = cv2.INTER_LINEAR)\\n\\n if save_images == True:\\n cv2.imwrite(augmented_images_dir + \\\\\\n \\\"train_image_id_\\\" + \\\\\\n str(i) + \\\"_\\\" + \\\\\\n str(rotation_angle).replace(\\\".\\\",\\\"p\\\") + \\\"_\\\" + \\\\\\n str(x_shift) + \\\"_\\\" + \\\\\\n str(y_shift) + \\\"_\\\" + \\\\\\n str(scale).replace(\\\".\\\",\\\"p\\\") + \\\".tif\\\", \\\\\\n shifted_train_image)\\n\\n cv2.imwrite(augmented_images_dir + \\\\\\n \\\"truth_image_id_\\\" + \\\\\\n str(i) + \\\"_\\\" + \\\\\\n str(rotation_angle).replace(\\\".\\\",\\\"p\\\") + \\\"_\\\" + \\\\\\n str(x_shift) + \\\"_\\\" + \\\\\\n str(y_shift) + \\\"_\\\" + \\\\\\n str(scale).replace(\\\".\\\",\\\"p\\\") + \\\".png\\\", \\\\\\n shifted_truth_image)\\n\\n fn = augmented_images_dir + \\\\\\n \\\"numpy_image_array_id_\\\" + \\\\\\n str(i) + \\\"_\\\" + \\\\\\n str(rotation_angle).replace(\\\".\\\",\\\"p\\\") + \\\"_\\\" + \\\\\\n str(x_shift) + \\\"_\\\" + \\\\\\n str(y_shift) + \\\"_\\\" + \\\\\\n str(scale).replace(\\\".\\\",\\\"p\\\") + \\\".npz\\\"\\n\\n resized_truth_image[resized_truth_image > 0] = 1\\n\\n # This part is for FCN.\\n # We destinguish between good and bad tissue so\\n # we have just two classes.\\n n_classes = 2\\n resized_truth_image = to_categorical(resized_truth_image, n_classes)\\n resized_truth_image = np.reshape(resized_truth_image, (nw_label, nh_label, n_classes))\\n\\n np.savez_compressed(fn, \\n image=resized_train_image/255.0, \\n #label=np.reshape(resized_truth_image, (nw_label*nh_label, 1)))\\n label=resized_truth_image)\\n\\n\\ndef load_data_from_npz(file_name_list):\\n\\n n_files = len(file_name_list)\\n if n_files == 0:\\n sys.exit(\\\"ERROR: File name list empty.\\\")\\n\\n loaded_data = np.load(file_name_list[0])\\n image = loaded_data[\\\"image\\\"].astype(np.float32)\\n label = loaded_data[\\\"label\\\"].astype(np.float32)\\n\\n iw, ih, ic = image.shape\\n mw, mh, mc = label.shape\\n\\n x_data = np.zeros((n_files, ih, iw, ic))\\n y_data = np.zeros((n_files, mw, mh, mc))\\n\\n x_data[0, :, :, :] = image\\n y_data[0, :, :, :] = label\\n\\n for i in range(1, n_files):\\n loaded_data = np.load(file_name_list[i])\\n image = loaded_data[\\\"image\\\"].astype(np.float32)\\n label = loaded_data[\\\"label\\\"].astype(np.float32)\\n\\n x_data[i, :, :, :] = image\\n y_data[i, :, :, :] = label\\n\\n return x_data, y_data\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.4591751992702484,\n \"alphanum_fraction\": 0.48214781284332275,\n \"avg_line_length\": 30.657894134521484,\n \"blob_id\": \"374121e5cff7ffc912991d74cd71ff3bfeeea051\",\n \"content_id\": \"5cd75ac02c858f66f2b6780c4a89ddb87814c4a5\",\n \"detected_licenses\": [\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3613,\n \"license_type\": \"permissive\",\n \"max_line_length\": 102,\n \"num_lines\": 114,\n \"path\": \"/prepare_test_data.py\",\n \"repo_name\": \"diwadd/PathImageSegmentation\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import random\\nimport sys\\nimport glob\\nimport os\\nimport shutil\\n\\nimport cv2\\nimport numpy as np\\nfrom keras import backend as K\\nfrom keras.models import load_model\\n\\nfrom keras_fcn.layers import BilinearUpSampling2D\\n\\nimport data_handling as dh\\nimport deep_models as dm\\n\\ndef load_and_resize_image(image_file_name,\\n nw_image=250,\\n nh_image=250):\\n\\n image = cv2.imread(image_file_name)\\n\\n resized_image = cv2.resize(image, \\n (nw_image, nh_image), \\n interpolation = cv2.INTER_LINEAR)\\n return resized_image\\n\\n\\ndef prepare_data_for_dispatch(model_file_name,\\n test_images_list,\\n nw_image=250,\\n nh_image=250,\\n nw_label=50,\\n nh_label=50,\\n images_dir=\\\"images_for_dispatch/\\\"):\\n\\n is_dir = os.path.isdir(images_dir)\\n if (is_dir == True):\\n shutil.rmtree(images_dir)\\n os.makedirs(images_dir)\\n else:\\n os.makedirs(images_dir)\\n\\n K.get_session()\\n model = load_model(model_file_name, custom_objects={\\\"BilinearUpSampling2D\\\": BilinearUpSampling2D})\\n\\n n_train_images = len(test_images_list)\\n for i in range(n_train_images):\\n print(\\\"i: \\\" + str(i + 1))\\n resized_image = load_and_resize_image(test_images_list[i],\\n nw_image=nw_image,\\n nh_image=nh_image)\\n\\n # The the i000000 image indicator\\n # from training/images/i000000.tif\\n if test_images_list[i][0:3] == \\\"tra\\\":\\n trunk = test_images_list[i][16:-4]\\n elif test_images_list[i][0:3] == \\\"tes\\\":\\n trunk = test_images_list[i][15:-4]\\n else:\\n pass\\n\\n print(trunk)\\n fn = images_dir + str(trunk) + \\\".npz\\\"\\n\\n _, _, ic = resized_image.shape\\n\\n resized_image = np.reshape(resized_image, (1, nw_image, nh_image, ic))/255.0\\n label = model.predict(resized_image)\\n label = np.reshape(label, (nw_label, nh_label, 2))\\n\\n label = dh.transform_label(label)\\n\\n label = cv2.resize(label,\\n (dh.DEFAULT_WIDTH, dh.DEFAULT_HEIGHT),\\n interpolation = cv2.INTER_LINEAR)\\n\\n f = open(images_dir + trunk + \\\"_mask.txt\\\", \\\"w\\\")\\n\\n for v in range(dh.DEFAULT_WIDTH):\\n for w in range(dh.DEFAULT_HEIGHT):\\n if (label[v][w] > 0.5):\\n f.write(\\\"1\\\")\\n else:\\n f.write(\\\"0\\\")\\n f.write(\\\"\\\\n\\\")\\n\\n f.close()\\n\\n K.clear_session()\\n\\n\\nif __name__ == \\\"__main__\\\":\\n print(\\\"Starting main.\\\")\\n\\n # ------------------------------------------------------------------------\\n # Load data \\n # ------------------------------------------------------------------------\\n\\n train_images_list = glob.glob(\\\"training/images/*.tif\\\")\\n test_images_list = glob.glob(\\\"testing/images/*.tif\\\")\\n\\n image_list = train_images_list + test_images_list\\n\\n # model_file_name = \\\"vgg16_16s_fcn_model_after_global_epoch_62.h5\\\"\\n # model_file_name = \\\"vgg16_32s_fcn_model.h5\\\" \\n model_file_name = \\\"model.h5\\\" \\n prepare_data_for_dispatch(model_file_name,\\n image_list,\\n nw_image=500,\\n nh_image=500,\\n nw_label=500,\\n nh_label=500,\\n images_dir=\\\"images_for_dispatch/\\\")\\n\\n\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.7676056623458862,\n \"alphanum_fraction\": 0.8098591566085815,\n \"avg_line_length\": 34,\n \"blob_id\": \"886cca198d27f93f29955002481ecd8110115ef3\",\n \"content_id\": \"d1ba44351efd1b61f94bb3d2e535c7f3f7224af7\",\n \"detected_licenses\": [\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 142,\n \"license_type\": \"permissive\",\n \"max_line_length\": 48,\n \"num_lines\": 4,\n \"path\": \"/README.md\",\n \"repo_name\": \"diwadd/PathImageSegmentation\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"We implement the FCN VGG16 32s and 16s networks.\\n\\nWe use these networks for segmentation in the \\nPathological Image Segmentation Challenge.\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.4825242757797241,\n \"alphanum_fraction\": 0.5543689131736755,\n \"avg_line_length\": 20.35416603088379,\n \"blob_id\": \"aba5e2cd9eff6a25eac205395748a145a57d15c3\",\n \"content_id\": \"de509ba4265b79b89ff6f9df7989583f5a8716e0\",\n \"detected_licenses\": [\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1030,\n \"license_type\": \"permissive\",\n \"max_line_length\": 116,\n \"num_lines\": 48,\n \"path\": \"/categorical_crossentropy_testing.py\",\n \"repo_name\": \"diwadd/PathImageSegmentation\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import random\\nimport sys\\nimport glob\\n\\nimport cv2\\nimport numpy as np\\n\\nfrom keras import backend as K\\nfrom tensorflow.python.ops import math_ops\\n\\nimport deep_models as dm\\n\\n\\ndef cross_entropy(t, p):\\n return -t*np.log(p)\\n\\n\\n\\n\\ntrue_arr = np.array([[[1.0, 0.0], [1.0, 0.0], [1.0, 0.0]], \\n [[1.0, 0.0], [1.0, 0.0], [1.0, 0.0]],\\n [[1.0, 0.0], [1.0, 0.0], [1.0, 0.0]],])\\n\\n\\npred_arr = np.array([[[0.9, 0.1], [0.9, 0.1], [0.9, 0.1]], \\n [[0.9, 0.1], [0.4, 0.6], [0.5, 0.5]],\\n [[0.9, 0.1], [0.2, 0.8], [0.9, 0.1]]])\\n\\nprint(true_arr.shape)\\n\\n\\nce = cross_entropy(true_arr, pred_arr)\\nprint(\\\"ce: \\\" + str(ce))\\n\\nK.get_session()\\n\\ntrue_var = K.variable(true_arr)\\npred_var = K.variable(pred_arr)\\n\\n# output = pred_var / math_ops.reduce_sum(pred_var, reduction_indices=len(pred_var.get_shape()) - 1, keep_dims=True)\\n# print(K.eval(output))\\n\\nen = K.categorical_crossentropy(pred_var, true_var, from_logits=False)\\n\\nen_eval = K.eval(en)\\nprint(en_eval)\\n\\nK.clear_session()\\n\\n\\n\\n\\n\\n\"\n }\n]"},"num_files":{"kind":"number","value":9,"string":"9"}}},{"rowIdx":2980395,"cells":{"repo_name":{"kind":"string","value":"Gaterny/ActressSearch"},"repo_url":{"kind":"string","value":"https://github.com/Gaterny/ActressSearch"},"snapshot_id":{"kind":"string","value":"0b5069d859543a2236061775366314c5fa8669dc"},"revision_id":{"kind":"string","value":"6529adba73bb4dec1ea974df922cfc51ddae6564"},"directory_id":{"kind":"string","value":"328cb8dd4496b553b8c27901e8faaa9a6f505177"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2020-04-12T19:55:14.052735","string":"2020-04-12T19:55:14.052735"},"revision_date":{"kind":"timestamp","value":"2018-12-22T08:25:40","string":"2018-12-22T08:25:40"},"committer_date":{"kind":"timestamp","value":"2018-12-22T08:25:40","string":"2018-12-22T08:25:40"},"github_id":{"kind":"number","value":162721411,"string":"162,721,411"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"null"},"gha_updated_at":{"kind":"null"},"gha_pushed_at":{"kind":"null"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.7684210538864136,"alphanum_fraction":0.7684210538864136,"avg_line_length":18,"blob_id":"a42c88c43c3ac7650292b0baa1ff2788e1bfc62a","content_id":"c0cb931b8e9799626fc94e49dc1873c6e82882e3","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":95,"license_type":"no_license","max_line_length":34,"num_lines":5,"path":"/Search/actressapp/apps.py","repo_name":"Gaterny/ActressSearch","src_encoding":"UTF-8","text":"from django.apps import AppConfig\n\n\nclass ActressappConfig(AppConfig):\n name = 'actressapp'\n"},{"alpha_fraction":0.5325379371643066,"alphanum_fraction":0.536876380443573,"avg_line_length":37.90140914916992,"blob_id":"ea45dbea034cf341db8c6b3fd299bcf69fb17d41","content_id":"7502c5c2cfb3697f5c07d175d7c3670e0f245790","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2856,"license_type":"no_license","max_line_length":92,"num_lines":71,"path":"/Actress/Actress/spiders/actress_spider.py","repo_name":"Gaterny/ActressSearch","src_encoding":"UTF-8","text":"# -*- coding: utf-8 -*-\nimport scrapy\nfrom Actress.items import ActressItem\n\n\nclass ActressSpiderSpider(scrapy.Spider):\n name = 'actress_spider'\n allowed_domains = ['javbus.com']\n start_urls = [\n 'https://www.javbus.com/actresses',\n 'https://www.javbus.com/uncensored/actresses'\n ]\n\n def parse(self, response):\n hrefs = response.xpath('//a[contains(@class, \"avatar-box\")]/@href').extract()\n for href in hrefs:\n yield scrapy.Request(url=href, callback=self.parse_detail)\n\n next_page = response.xpath('//a[@id=\"next\"]/@href').extract_first()\n if next is not None:\n next_url = response.urljoin(next_page)\n yield scrapy.Request(url=next_url, callback=self.parse)\n\n def parse_detail(self, response):\n item = ActressItem()\n name = response.xpath('//span[contains(@class, \"pb10\")]/text()').extract_first()\n\n # 以下信息有误，以后再处理\n # birth = response.xpath('//div[@class=\"photo-info\"]//p[1]/text()').extract_first()\n # age = response.xpath('//div[@class=\"photo-info\"]//p[2]/text()').extract_first()\n # # 腰围\n # waist = response.xpath('//div[@class=\"photo-info\"]//p[6]/text()').extract_first()\n # # 罩杯\n # cup = response.xpath('//div[@class=\"photo-info\"]//p[4]/text()').extract_first()\n # # 胸围\n # bust = response.xpath('//div[@class=\"photo-info\"]//p[5]/text()').extract_first()\n # # 臀围\n # hips = response.xpath('//div[@class=\"photo-info\"]//p[7]/text()').extract_first()\n # # 身高\n # height = response.xpath('//div[@class=\"photo-info\"]//p[3]/text()').extract_first()\n # 演员作品名\n film_list = response.xpath(\"//a[@class='movie-box']//span/text()\").extract()\n # 作品番号\n tags = response.xpath(\"//a[@class='movie-box']//date[1]/text()\").extract()\n # 作品时间\n dates = response.xpath(\"//a[@class='movie-box']//date[2]/text()\").extract()\n\n item['name'] = name\n # item['age'] = age\n # item['birth'] = birth\n # item['waist'] = waist\n # item['hips'] = hips\n # item['cup'] = cup\n # item['bust'] = bust\n # item['height'] = height\n\n films = []\n for film in film_list:\n # 去除换行符的干扰字符\n afilm = film.strip('\\t \\n \\r '' /')\n if afilm != '':\n films.append(afilm)\n for j in range(len(films)):\n item['film'] = films[j]\n item['tag'] = tags[j]\n item['date'] = dates[j]\n yield item\n next_page = response.xpath(\"//a[@id='next']/@href\").extract_first()\n if next_page:\n next_url = response.urljoin(next_page)\n yield scrapy.Request(url=next_url, callback=self.parse_detail)\n\n\n\n\n"},{"alpha_fraction":0.6349206566810608,"alphanum_fraction":0.637188196182251,"avg_line_length":20.536584854125977,"blob_id":"84bd4951e68106148d5f3779e2931a206633064f","content_id":"acdb78b9dddafd542af415028aada8862d7dfc0c","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":910,"license_type":"no_license","max_line_length":72,"num_lines":41,"path":"/Actress/Actress/models/es_type.py","repo_name":"Gaterny/ActressSearch","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n\nfrom elasticsearch_dsl.connections import connections\nfrom elasticsearch_dsl import Document, Text, Date, Keyword, Completion\nfrom elasticsearch_dsl.analysis import CustomAnalyzer as _CustomAnalyzer\n\n# 创建连接\nconnections.create_connection(hosts=['localhost'])\n\n\nclass CustomAnalyzer(_CustomAnalyzer):\n \"\"\"\n 自动补全\n \"\"\"\n def get_analysis_definition(self):\n return {}\n\n\nik_analyzer = CustomAnalyzer('ik_max_word', filter=['lowercase'])\n\n\nclass ActressType(Document):\n \"\"\"\n 类型配置\n \"\"\"\n suggest = Completion(analyzer=ik_analyzer)\n name = Text(analyzer='ik_max_word') # 分词\n date = Date()\n film = Text(analyzer='ik_max_word')\n tag = Keyword()\n\n class Index:\n name = 'actress'\n settings = {\n 'number_of_shards': 2,\n }\n\n\nif __name__ == \"__main__\":\n ActressType.init()"}],"string":"[\n {\n \"alpha_fraction\": 0.7684210538864136,\n \"alphanum_fraction\": 0.7684210538864136,\n \"avg_line_length\": 18,\n \"blob_id\": \"a42c88c43c3ac7650292b0baa1ff2788e1bfc62a\",\n \"content_id\": \"c0cb931b8e9799626fc94e49dc1873c6e82882e3\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 95,\n \"license_type\": \"no_license\",\n \"max_line_length\": 34,\n \"num_lines\": 5,\n \"path\": \"/Search/actressapp/apps.py\",\n \"repo_name\": \"Gaterny/ActressSearch\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from django.apps import AppConfig\\n\\n\\nclass ActressappConfig(AppConfig):\\n name = 'actressapp'\\n\"\n },\n {\n \"alpha_fraction\": 0.5325379371643066,\n \"alphanum_fraction\": 0.536876380443573,\n \"avg_line_length\": 37.90140914916992,\n \"blob_id\": \"ea45dbea034cf341db8c6b3fd299bcf69fb17d41\",\n \"content_id\": \"7502c5c2cfb3697f5c07d175d7c3670e0f245790\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2856,\n \"license_type\": \"no_license\",\n \"max_line_length\": 92,\n \"num_lines\": 71,\n \"path\": \"/Actress/Actress/spiders/actress_spider.py\",\n \"repo_name\": \"Gaterny/ActressSearch\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# -*- coding: utf-8 -*-\\nimport scrapy\\nfrom Actress.items import ActressItem\\n\\n\\nclass ActressSpiderSpider(scrapy.Spider):\\n name = 'actress_spider'\\n allowed_domains = ['javbus.com']\\n start_urls = [\\n 'https://www.javbus.com/actresses',\\n 'https://www.javbus.com/uncensored/actresses'\\n ]\\n\\n def parse(self, response):\\n hrefs = response.xpath('//a[contains(@class, \\\"avatar-box\\\")]/@href').extract()\\n for href in hrefs:\\n yield scrapy.Request(url=href, callback=self.parse_detail)\\n\\n next_page = response.xpath('//a[@id=\\\"next\\\"]/@href').extract_first()\\n if next is not None:\\n next_url = response.urljoin(next_page)\\n yield scrapy.Request(url=next_url, callback=self.parse)\\n\\n def parse_detail(self, response):\\n item = ActressItem()\\n name = response.xpath('//span[contains(@class, \\\"pb10\\\")]/text()').extract_first()\\n\\n # 以下信息有误，以后再处理\\n # birth = response.xpath('//div[@class=\\\"photo-info\\\"]//p[1]/text()').extract_first()\\n # age = response.xpath('//div[@class=\\\"photo-info\\\"]//p[2]/text()').extract_first()\\n # # 腰围\\n # waist = response.xpath('//div[@class=\\\"photo-info\\\"]//p[6]/text()').extract_first()\\n # # 罩杯\\n # cup = response.xpath('//div[@class=\\\"photo-info\\\"]//p[4]/text()').extract_first()\\n # # 胸围\\n # bust = response.xpath('//div[@class=\\\"photo-info\\\"]//p[5]/text()').extract_first()\\n # # 臀围\\n # hips = response.xpath('//div[@class=\\\"photo-info\\\"]//p[7]/text()').extract_first()\\n # # 身高\\n # height = response.xpath('//div[@class=\\\"photo-info\\\"]//p[3]/text()').extract_first()\\n # 演员作品名\\n film_list = response.xpath(\\\"//a[@class='movie-box']//span/text()\\\").extract()\\n # 作品番号\\n tags = response.xpath(\\\"//a[@class='movie-box']//date[1]/text()\\\").extract()\\n # 作品时间\\n dates = response.xpath(\\\"//a[@class='movie-box']//date[2]/text()\\\").extract()\\n\\n item['name'] = name\\n # item['age'] = age\\n # item['birth'] = birth\\n # item['waist'] = waist\\n # item['hips'] = hips\\n # item['cup'] = cup\\n # item['bust'] = bust\\n # item['height'] = height\\n\\n films = []\\n for film in film_list:\\n # 去除换行符的干扰字符\\n afilm = film.strip('\\\\t \\\\n \\\\r '' /')\\n if afilm != '':\\n films.append(afilm)\\n for j in range(len(films)):\\n item['film'] = films[j]\\n item['tag'] = tags[j]\\n item['date'] = dates[j]\\n yield item\\n next_page = response.xpath(\\\"//a[@id='next']/@href\\\").extract_first()\\n if next_page:\\n next_url = response.urljoin(next_page)\\n yield scrapy.Request(url=next_url, callback=self.parse_detail)\\n\\n\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6349206566810608,\n \"alphanum_fraction\": 0.637188196182251,\n \"avg_line_length\": 20.536584854125977,\n \"blob_id\": \"84bd4951e68106148d5f3779e2931a206633064f\",\n \"content_id\": \"acdb78b9dddafd542af415028aada8862d7dfc0c\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 910,\n \"license_type\": \"no_license\",\n \"max_line_length\": 72,\n \"num_lines\": 41,\n \"path\": \"/Actress/Actress/models/es_type.py\",\n \"repo_name\": \"Gaterny/ActressSearch\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n\\nfrom elasticsearch_dsl.connections import connections\\nfrom elasticsearch_dsl import Document, Text, Date, Keyword, Completion\\nfrom elasticsearch_dsl.analysis import CustomAnalyzer as _CustomAnalyzer\\n\\n# 创建连接\\nconnections.create_connection(hosts=['localhost'])\\n\\n\\nclass CustomAnalyzer(_CustomAnalyzer):\\n \\\"\\\"\\\"\\n 自动补全\\n \\\"\\\"\\\"\\n def get_analysis_definition(self):\\n return {}\\n\\n\\nik_analyzer = CustomAnalyzer('ik_max_word', filter=['lowercase'])\\n\\n\\nclass ActressType(Document):\\n \\\"\\\"\\\"\\n 类型配置\\n \\\"\\\"\\\"\\n suggest = Completion(analyzer=ik_analyzer)\\n name = Text(analyzer='ik_max_word') # 分词\\n date = Date()\\n film = Text(analyzer='ik_max_word')\\n tag = Keyword()\\n\\n class Index:\\n name = 'actress'\\n settings = {\\n 'number_of_shards': 2,\\n }\\n\\n\\nif __name__ == \\\"__main__\\\":\\n ActressType.init()\"\n }\n]"},"num_files":{"kind":"number","value":3,"string":"3"}}},{"rowIdx":2980396,"cells":{"repo_name":{"kind":"string","value":"karanbudhraja/PABMD"},"repo_url":{"kind":"string","value":"https://github.com/karanbudhraja/PABMD"},"snapshot_id":{"kind":"string","value":"d8622289ebb1dd6fff60be22034d7633dd9bcad5"},"revision_id":{"kind":"string","value":"29982bef9fbd89e05404adfa1afe15c7a5ccd350"},"directory_id":{"kind":"string","value":"63d0e4df6913ae7c7cedd8d72b074273d934c204"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2020-06-14T17:50:29.006765","string":"2020-06-14T17:50:29.006765"},"revision_date":{"kind":"timestamp","value":"2019-07-15T21:12:47","string":"2019-07-15T21:12:47"},"committer_date":{"kind":"timestamp","value":"2019-07-15T21:12:47","string":"2019-07-15T21:12:47"},"github_id":{"kind":"number","value":195070897,"string":"195,070,897"},"star_events_count":{"kind":"number","value":1,"string":"1"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"null"},"gha_updated_at":{"kind":"null"},"gha_pushed_at":{"kind":"null"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.7330960631370544,"alphanum_fraction":0.7348754405975342,"avg_line_length":20.615385055541992,"blob_id":"5d3f87b1129a1658751e1a880d143d5ffd134a8e","content_id":"0ba15ef118cabd0e3e620163c13f215df0612ea4","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":562,"license_type":"no_license","max_line_length":50,"num_lines":26,"path":"/app/do_pipeline.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"#!/bin/bash\n\n# go to target script directory\ncd ../src\nsource configuration.sh\n\n# copy to demonstration folder for lfd use\ncd ../data/demonstrations/$ABM\nrm -rf *\ncp -r ../../../app/matches .\n\n# move to appropriate directory to run lfd\ncd ../../../src\n\n# run the framework to suggest several alps\n# that would produce the demonstration slps\n./lfd.sh\n\n# simulate those alps\n# to check the actual slps that they correspond to\n./simulate.sh\n\n# filter from the suggested alps\n# use distance from demonstration slps\nexport USE_FILTER_DESCRIPTOR_SETTING=1\n./filter.sh\n"},{"alpha_fraction":0.7180395126342773,"alphanum_fraction":0.7912715673446655,"avg_line_length":89.23711395263672,"blob_id":"4f17937b5d1023ddb3b730250401bf2783d10d39","content_id":"8b149f38b51d1b6a006c27a550c274e59df63929","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":8753,"license_type":"no_license","max_line_length":155,"num_lines":97,"path":"/data/gifs_all/generate_gifs.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# convert demonstrations\nconvert -delay 50 -loop 0 ../demonstrations_all/flocking/simulated/1/*.png demonstration_1.gif\nconvert -delay 50 -loop 0 ../demonstrations_all/flocking/simulated/2/*.png demonstration_2.gif\nconvert -delay 50 -loop 0 ../demonstrations_all/flocking/simulated/3/*.png demonstration_3.gif\n\nconvert -delay 50 -loop 0 ../demonstrations_/flocking/1/*.png demonstration__1.gif\nconvert -delay 50 -loop 0 ../demonstrations_/flocking/2/*.png demonstration__2.gif\nconvert -delay 50 -loop 0 ../demonstrations_/flocking/3/*.png demonstration__3.gif\n\n# convert rotated images\nconvert -rotate 90 demonstration_1.gif demonstration_1_rotation.gif\n\n# convert translated images\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_200.png -background black -flatten 1108_200_translation.png\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_201.png -background black -flatten 1108_201_translation.png\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_202.png -background black -flatten 1108_202_translation.png\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_203.png -background black -flatten 1108_203_translation.png\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_204.png -background black -flatten 1108_204_translation.png\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_205.png -background black -flatten 1108_205_translation.png\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_206.png -background black -flatten 1108_206_translation.png\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_207.png -background black -flatten 1108_207_translation.png\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_208.png -background black -flatten 1108_208_translation.png\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_209.png -background black -flatten 1108_209_translation.png\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_210.png -background black -flatten 1108_210_translation.png\nconvert +repage 1108_200_translation.png 1108_200_translation.png\nconvert +repage 1108_201_translation.png 1108_201_translation.png\nconvert +repage 1108_202_translation.png 1108_202_translation.png\nconvert +repage 1108_203_translation.png 1108_203_translation.png\nconvert +repage 1108_204_translation.png 1108_204_translation.png\nconvert +repage 1108_205_translation.png 1108_205_translation.png\nconvert +repage 1108_206_translation.png 1108_206_translation.png\nconvert +repage 1108_207_translation.png 1108_207_translation.png\nconvert +repage 1108_208_translation.png 1108_208_translation.png\nconvert +repage 1108_209_translation.png 1108_209_translation.png\nconvert +repage 1108_210_translation.png 1108_210_translation.png\nconvert -delay 50 -loop 0 *_translation.png demonstration_1_translation.gif\nrm *_translation.png\n\n# convert point selection methods\nconvert -delay 50 -loop 0 ../predictions_all/point\\ selection/predictions_error_weighted/flocking/1/*.png point\\ selection/error_weighted_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/point\\ selection/predictions_error_weighted/flocking/2/*.png point\\ selection/error_weighted_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/point\\ selection/predictions_error_weighted/flocking/3/*.png point\\ selection/error_weighted_3.gif\n\nconvert -delay 50 -loop 0 ../predictions_all/point\\ selection/predictions_kmeans/flocking/1/*.png point\\ selection/kmeans_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/point\\ selection/predictions_kmeans/flocking/2/*.png point\\ selection/kmeans_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/point\\ selection/predictions_kmeans/flocking/3/*.png point\\ selection/kmeans_3.gif\n\nconvert -delay 50 -loop 0 ../predictions_all/point\\ selection/predictions_mean/flocking/1/*.png point\\ selection/mean_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/point\\ selection/predictions_mean/flocking/2/*.png point\\ selection/mean_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/point\\ selection/predictions_mean/flocking/3/*.png point\\ selection/mean_3.gif\n\nconvert -delay 50 -loop 0 ../predictions_all/point\\ selection/predictions_min_error/flocking/1/*.png point\\ selection/min_error_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/point\\ selection/predictions_min_error/flocking/2/*.png point\\ selection/min_error_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/point\\ selection/predictions_min_error/flocking/3/*.png point\\ selection/min_error_3.gif\n\n# convert feature construction methods\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_contour/flocking/1/*.png feature\\ construction/contour_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_contour/flocking/2/*.png feature\\ construction/contour_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_contour/flocking/3/*.png feature\\ construction/contour_3.gif\n\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_mnist/flocking/1/*.png feature\\ construction/mnist_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_mnist/flocking/2/*.png feature\\ construction/mnist_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_mnist/flocking/3/*.png feature\\ construction/mnist_3.gif\n\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_vae/flocking/1/*.png feature\\ construction/vae_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_vae/flocking/2/*.png feature\\ construction/vae_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_vae/flocking/3/*.png feature\\ construction/vae_3.gif\n\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_vgg16/flocking/1/*.png feature\\ construction/vgg16_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_vgg16/flocking/2/*.png feature\\ construction/vgg16_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_vgg16/flocking/3/*.png feature\\ construction/vgg16_3.gif\n\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_resnet50/flocking/1/*.png feature\\ construction/resnet50_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_resnet50/flocking/2/*.png feature\\ construction/resnet50_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_resnet50/flocking/3/*.png feature\\ construction/resnet50_3.gif\n\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_lstm/flocking/1/*.png feature\\ construction/lstm_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_lstm/flocking/2/*.png feature\\ construction/lstm_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_lstm/flocking/3/*.png feature\\ construction/lstm_3.gif\n\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_lbp/flocking/1/*.png feature\\ construction/lbp_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_lbp/flocking/2/*.png feature\\ construction/lbp_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/feature\\ construction/predictions_lbp/flocking/3/*.png feature\\ construction/lbp_3.gif\n\n# convert dataset analysis methods\nconvert -delay 50 -loop 0 ../predictions_all/dataset\\ analysis/error_weighted/1/unguided/predictions_best/flocking/1/*.png dataset\\ analysis/unguided_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/dataset\\ analysis/error_weighted/2/unguided/predictions_best/flocking/2/*.png dataset\\ analysis/unguided_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/dataset\\ analysis/error_weighted/3/unguided/predictions_best/flocking/3/*.png dataset\\ analysis/unguided_3.gif\n\nconvert -delay 50 -loop 0 ../predictions_all/dataset\\ analysis/error_weighted/1/guided/predictions_best/flocking/1/*.png dataset\\ analysis/guided_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/dataset\\ analysis/error_weighted/2/guided/predictions_best/flocking/2/*.png dataset\\ analysis/guided_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/dataset\\ analysis/error_weighted/3/guided/predictions_best/flocking/3/*.png dataset\\ analysis/guided_3.gif\n\n# convert filtering method\nconvert -delay 50 -loop 0 ../predictions_all/pruning\\ od/flocking/1/*.png pruning\\ od/pruning_od_1.gif\nconvert -delay 50 -loop 0 ../predictions_all/pruning\\ od/flocking/2/*.png pruning\\ od/pruning_od_2.gif\nconvert -delay 50 -loop 0 ../predictions_all/pruning\\ od/flocking/3/*.png pruning\\ od/pruning_od_3.gif\n"},{"alpha_fraction":0.6934612989425659,"alphanum_fraction":0.6946610808372498,"avg_line_length":28.76785659790039,"blob_id":"f71261bde834b7e609f89053df97ec68ea16d57a","content_id":"d6b7077a1fe8f912a5db018ef2b0b7f084d62254","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":1667,"license_type":"no_license","max_line_length":101,"num_lines":56,"path":"/src/simulate.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# load configuration\nsource configuration.sh\n\nexport SIMULATION_SCRIPT=\"sampling/\"$ABM\"/predict_\"$ABM$SIMULATION_EXTENSION\n\n# remove existing data\nrm -f suggested_slps.txt\nrm -rf sampling/$ABM/images_*\n\n# for line in file\nwhile read -r line\ndo\n # read the number of alps per line\n export ALPS_PER_LINE=$(echo $line | tr -cd \")\" | wc -c)\n \n # remove all spaces in line\n # assume just one configuration per line, for now\n # fix this later by adding a loop\n export line=$(echo $line | sed \"s/ //g\")\n\n # simulate using suggested alps\n export SLP_CONFIGURATIONS=$($SIMULATION_TOOL $SIMULATION_SCRIPT $line)\n\n # in case of images, slp configurations are not returned\n # compute slp configuration from images\n if [ $IMAGES == 1 ]; then\n\t# all folders\n\texport SLP_CONFIGURATIONS=\"[\"\n\t\n\t# get the last image folder\n\t# this is the most recently created one\n\t# get as many lines as the number of alps in a row of suggested_alps.txt\n\texport SLP_IMAGES_FOLDERS=$(ls \"sampling/\"$ABM | sort | grep images_ | tail -n $ALPS_PER_LINE)\n\n\tfor folder in $SLP_IMAGES_FOLDERS\n\tdo\n\t # for each folder\n\t export SLP_CONFIGURATION=$(python sampling/utils/process_folder_images.py sampling/$ABM/$folder)\n\t export CHARACTER_COUNT=$(echo $SLP_CONFIGURATIONS | wc -c)\n\n\t if [ $CHARACTER_COUNT == 2 ]; then\n\t\t# skip first occurrence\n\t\texport SEPARATOR=\"\"\n\t else\n\t\texport SEPARATOR=\", \"\n\t fi\n\t \n\t export SLP_CONFIGURATIONS=$SLP_CONFIGURATIONS$SEPARATOR$SLP_CONFIGURATION\n\tdone\n\n\t# add closing bracket\n\texport SLP_CONFIGURATIONS=$SLP_CONFIGURATIONS\"]\"\n fi\n \n echo $SLP_CONFIGURATIONS >> suggested_slps.txt\ndone < suggested_alps.txt\n"},{"alpha_fraction":0.6227115988731384,"alphanum_fraction":0.6569381952285767,"avg_line_length":34.224300384521484,"blob_id":"141089750c1e1d743522c8da4b7ca8169c5e5edd","content_id":"30743688501ce491903eea44f95e83b89162b92f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3769,"license_type":"no_license","max_line_length":111,"num_lines":107,"path":"/src/sampling/utils/keras/conv_lstm.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"\"\"\" This script demonstrates the use of a convolutional LSTM network.\nThis network is used to predict the next frame of an artificially\ngenerated movie which contains moving squares.\n\"\"\"\nfrom keras.models import Sequential\nfrom keras.layers.convolutional import Convolution3D\nfrom keras.layers.convolutional_recurrent import ConvLSTM2D\nfrom keras.layers.normalization import BatchNormalization\nfrom keras.layers import Dense, Activation\nfrom keras.layers.core import Reshape, Flatten\nimport numpy as np\nimport pylab as plt\n\nfrom skimage.io import imread\nfrom skimage.transform import resize\n\n# We create a layer which take as input movies of shape\n# (n_frames, width, height, channels) and returns a movie\n# of identical shape.\n\nseq = Sequential()\n\nseq.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\n input_shape=(None, 40, 40, 1),\n batch_input_shape=(None, 11, 40, 40, 1),\n border_mode='same', return_sequences=True))\nseq.add(BatchNormalization())\nseq.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\n border_mode='same', return_sequences=True))\nseq.add(BatchNormalization())\n\nseq.add(Flatten())\nseq.add(Dense(128))\nseq.add(Dense(1600*11))\nseq.add(Reshape((11, 40, 40, 1)))\n\nseq.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\n border_mode='same', return_sequences=True))\nseq.add(BatchNormalization())\nseq.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\n border_mode='same', return_sequences=True))\nseq.add(BatchNormalization())\nseq.add(Convolution3D(nb_filter=1, kernel_dim1=1, kernel_dim2=3,\n kernel_dim3=3, activation='sigmoid',\n border_mode='same', dim_ordering='tf'))\nseq.compile(loss='binary_crossentropy', optimizer='adadelta')\n\ndef image_data(file):\n img = imread(file, as_grey=True)\n img = resize(img, (40, 40))\n return img\n\n# Artificial data generation:\n# Generate movies with 3 to 7 moving squares inside.\n# The squares are of shape 1x1 or 2x2 pixels,\n# which move linearly over time.\n# For convenience we first create movies with bigger width and height (80x80)\n# and at the end we select a 40x40 window.\n\ndef generate_movies(n_samples=1200, n_frames=11, dataFolder=None):\n row = 40\n col = 40\n \n noisy_movies = np.zeros((n_samples, n_frames, row, col, 1), dtype=np.float)\n shifted_movies = np.zeros((n_samples, 11, row, col, 1), dtype=np.float)\n\n for i in range(n_samples):\n # read custom dataset\n for t in range(n_frames):\n # read and resize to 40x40\n x = image_data(dataFolder + \"/\" + str(i) + \"_\" + str(200+t) + \".png\")\n shifted_movies[i, t, :, :, 0] = x\n \n # figure out how to add noise to this data\n noise_f = (-1)**np.random.randint(0, 2, size=(40,40))\n noisy_movies[i, t, :, :, 0] = np.dot(x, noise_f)\n \n return noisy_movies, shifted_movies\n\n# Train the network\nnoisy_movies, shifted_movies = generate_movies(n_samples=10, dataFolder=\"../../../../_swarm-lfd-data/Flocking\")\n\n# original nb_epoch = 500\n\nseq.fit(noisy_movies, shifted_movies, batch_size=10,\n nb_epoch=1, validation_split=0.05)\n\n# save model as JSON\nmodel = seq\nmodelAsJson = model.to_json()\nwith open(\"conv_lstm_model.json\", \"w\") as jsonFile:\n jsonFile.write(modelAsJson)\n# serialize weights to HDF5\nmodel.save_weights('conv_lstm_model.h5')\n\n'''\n# Testing the network on one movie\n# feed it with the first 7 positions and then\n# predict the new positions\nwhich = 1004\ntrack = noisy_movies[which][:7, ::, ::, ::]\n\nfor j in range(16):\n new_pos = seq.predict(track[np.newaxis, ::, ::, ::, ::])\n new = new_pos[::, -1, ::, ::, ::]\n track = np.concatenate((track, new), axis=0)\n'''\n"},{"alpha_fraction":0.6192708611488342,"alphanum_fraction":0.6348958611488342,"avg_line_length":36.28155517578125,"blob_id":"4d28b5aed7764cedb4a8ac825ad0c4d186e0922b","content_id":"06d75d7b7bf87ef7eb4643f1dcd0b97d34278821","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3840,"license_type":"no_license","max_line_length":110,"num_lines":103,"path":"/snippets/flocking/plot_suggested_slps.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import matplotlib.pyplot as plt\nimport numpy as np\n\n#demonstrations = [\"a\", \"b\", \"c\"]\n#demonstrations = [\"_a\", \"_b\", \"_c\"]\n#demonstrations = [\"a_filtered\", \"b_filtered\", \"c_filtered\"]\n#demonstrations = [\"_a\"]\n#demonstrations = [\"a\"]\ndemonstrations = [\"a_filtered\"]\n\n#folderName = \"error_weighted\"\nfolderName = \"min_error\"\n\nnumberOfBins = 10\n\nwith open(\"../../src/sampling/flocking/scale_data_flocking_contour.txt\", \"r\") as inFile:\n allLines = inFile.readlines()\n maxValueArea = eval(allLines[0])[-2]\n maxValuePerimeter = eval(allLines[0])[-1]\n minValueArea = eval(allLines[1])[-2]\n minValuePerimeter = eval(allLines[1])[-1]\n\nfor index, item in enumerate(demonstrations):\n allPointsX = []\n allPointsY = []\n errors = []\n\n # scatter\n \n plt.figure()\n with open(folderName + \"/suggested_slps_\" + item + \".txt\", \"r\") as inFile:\n for line in inFile:\n point = eval(line)[0]\n plt.scatter(point[0], point[1], c=\"b\", marker=\"o\", alpha=0.3)\n allPointsX.append(point[0])\n allPointsY.append(point[1])\n\n # compute error\n with open(\"demonstration_slps.txt\", \"r\") as inFile:\n demonstration = list(eval(inFile.readlines()[index]))\n demonstration[0] = (demonstration[0]-minValueArea)/(maxValueArea-minValueArea)\n demonstration[1] = (demonstration[1]-minValuePerimeter)/(maxValuePerimeter-minValuePerimeter)\n\n _point = [x for x in point]\n _point[0] = (_point[0]-minValueArea)/(maxValueArea-minValueArea)\n _point[1] = (_point[1]-minValuePerimeter)/(maxValuePerimeter-minValuePerimeter)\n \n error = (_point[0]-demonstration[0])**2 + (_point[1]-demonstration[1])**2\n errors.append(error)\n \n plt.scatter(point[0], point[1], c=\"b\", marker=\"o\", alpha=0.3, label=\"Suggested SLPs\")\n\n with open(\"demonstration_slps.txt\", \"r\") as inFile:\n point = eval(inFile.readlines()[index])\n plt.scatter(point[0], point[1], c=\"r\", marker=\"s\", alpha=1.0, label=\"Demonstration SLPs\")\n\n plt.scatter(np.mean(allPointsX), np.mean(allPointsY), c=\"g\", marker=\"s\", alpha=1.0, label=\"Mean SLPs\")\n\n print np.mean(errors)\n print np.std(errors)\n \n ax = plt.gca()\n\n ticks = ax.get_xticks()\n minValue = minValueArea\n maxValue = maxValueArea\n scaledTicks = [(value-minValue)/(maxValue-minValue) for value in ticks]\n scaledTicks = [int(value*100)/100.0 for value in scaledTicks]\n ax.set_xticklabels(scaledTicks)\n \n ticks = ax.get_yticks()\n minValue = minValuePerimeter\n maxValue = maxValuePerimeter\n scaledTicks = [(value-minValue)/(maxValue-minValue) for value in ticks]\n scaledTicks = [int(value*100)/100.0 for value in scaledTicks]\n ax.set_yticklabels(scaledTicks)\n \n plt.xlabel(\"Maximum Area\")\n plt.ylabel(\"Maximum Perimeter\")\n plt.legend()\n plt.savefig(folderName + \"/suggested_slps_\" + item)\n\n # histogram\n\n plt.figure()\n allPointsX = [(x-minValueArea)/(maxValueArea-minValueArea) for x in allPointsX]\n allPointsY = [(y-minValuePerimeter)/(maxValuePerimeter-minValuePerimeter) for y in allPointsY] \n histogram, xEdges, yEdges = np.histogram2d(allPointsX, allPointsY, bins=numberOfBins, range=[[0,1],[0,1]])\n plt.imshow(histogram, interpolation='nearest', origin='low')\n plt.ylabel(\"Maximum Area\")\n plt.xlabel(\"Maximum Perimeter\")\n plt.colorbar(label=\"Number of Points\")\n ax = plt.gca()\n\n # works for normalized values\n ticks = ax.get_xticks()\n ticks = [value/numberOfBins for value in ticks]\n ax.set_xticklabels(ticks)\n ticks = ax.get_yticks()\n ticks = [value/numberOfBins for value in ticks]\n ax.set_yticklabels(ticks)\n\n plt.savefig(folderName + \"/histogram_suggested_slps_\" + item)\n"},{"alpha_fraction":0.7653061151504517,"alphanum_fraction":0.7653061151504517,"avg_line_length":22.058822631835938,"blob_id":"8a08b6001be36accd0cab4f82e0c7fcf76ae8c2b","content_id":"35b1b2cf3d2faa900f67278973de6649044f2377","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":392,"license_type":"no_license","max_line_length":51,"num_lines":17,"path":"/src/all.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# this script does everything\n# it is for reference of order for running scripts \n\n# sample alp, slp pairs\n./sampling.sh\n\n# run the framework to suggest several alps\n# that would produce the demonstration slps\n./lfd.sh\n\n# simulate those alps\n# to check the actual slps that they correspond to\n./simulate.sh\n\n# filter from the suggested alps\n# use distance from demonstration slps\n./filter.sh\n"},{"alpha_fraction":0.6927609443664551,"alphanum_fraction":0.6944444179534912,"avg_line_length":27.285715103149414,"blob_id":"1f367beb221899b69dff768fa432444a241ee570","content_id":"26d57a014a12e3080863e4b1531d5912ad5cf827","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1188,"license_type":"no_license","max_line_length":81,"num_lines":42,"path":"/src/sampling/utils/process_folder_images.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import sys\nimport featurize_image as fi\nimport glob\nimport os\nimport numpy as np\n\n# process a single folder\ndef process_folder(folderName):\n # get environment variables\n descriptorSetting = os.environ[\"DESCRIPTOR_SETTING\"]\n numDependent = eval(os.environ[\"NUM_DEPENDENT\"])\n\n # load feature extractor if needed\n net = fi.get_network(descriptorSetting)\n \n # initialize collection of dependent values\n dependentValuesFolder = []\n\n # expect everything in this folder to be an image\n for imageName in glob.glob(folderName + \"/*.*\"):\n # read image\n print >> sys.stderr, (\"now processing \" + imageName)\n \n # define dependent values\n dependentValuesImage = fi.get_features(net, imageName, descriptorSetting)\n\n # append to list\n dependentValuesFolder.append(dependentValuesImage)\n\n dependentValuesFolder = np.mean(dependentValuesFolder, axis=0)\n\n return tuple(dependentValuesFolder)\n\ndef main():\n folderName = sys.argv[1]\n dependentValuesFolder = process_folder(folderName)\n\n print dependentValuesFolder\n \nif __name__ == \"__main__\":\n # stuff only to run when not called via 'import' here\n main()\n"},{"alpha_fraction":0.594718337059021,"alphanum_fraction":0.6133802533149719,"avg_line_length":33.20481872558594,"blob_id":"12c0aaeaf20be00b0d3d582f52f094d74f35a25c","content_id":"1f4673ce4f279190e5181de29c228ab9faa40146","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2840,"license_type":"no_license","max_line_length":109,"num_lines":83,"path":"/data/domaindata/compute_statistics.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import numpy as np\n\n# results from batch data\nRESULTS_FILE = \"results_flocking.csv\"\nINPUT_NAME =\"\\\"Input.gifPrefix\\\"\"\n\ndef is_integer(x):\n try:\n int(x)\n return True\n except:\n return False\n\n# main code\nwith open(RESULTS_FILE) as results:\n headers = results.readline()\n\n #print headers\n \n headers = headers.split(\",\")\n inputIndex = headers.index(INPUT_NAME)\n answer1Index = inputIndex + 1\n answer2Index = inputIndex + 2\n answer3Index = inputIndex + 3\n answer4Index = inputIndex + 4\n\n # increment all by 3 for some reason\n # to get correct values in data\n inputIndex += 3\n answer1Index += 3\n answer2Index += 3\n answer3Index += 3\n answer4Index += 3\n \n # read result data\n resultData = {}\n for line in results:\n inputData = line.split(\",\")[inputIndex]\n answer1 = line.split(\",\")[answer1Index]\n answer2 = line.split(\",\")[answer2Index]\n answer3 = line.split(\",\")[answer3Index]\n answer4 = line.split(\",\")[answer4Index]\n\n #print line\n #print inputData\n #print answer1, answer2, answer3, answer4\n\n if(inputData not in resultData.keys()):\n resultData[inputData] = {}\n if(\"answer1\" not in resultData[inputData].keys()):\n resultData[inputData][\"answer1\"] = []\n if(\"answer2\" not in resultData[inputData].keys()):\n resultData[inputData][\"answer2\"] = []\n if(\"answer3\" not in resultData[inputData].keys()):\n resultData[inputData][\"answer3\"] = []\n\n # check control question\n if(eval(eval(answer4)) < 8):\n # control question failed\n print \"control question failed: \" + answer4\n continue\n\n # check for invalid responses\n invalidResponse = [type(eval(eval(x))) is not int for x in [answer1, answer2, answer3, answer4]]\n if(np.any(invalidResponse) == True):\n # invalid response\n print \"invalid response: \" + str([answer1, answer2, answer3, answer4])\n \n resultData[inputData][\"answer1\"].append(eval(answer1))\n resultData[inputData][\"answer2\"].append(eval(answer2))\n resultData[inputData][\"answer3\"].append(eval(answer3))\n\n # now process result data\n for inputData in resultData.keys():\n print inputData\n answerList = []\n for answer in resultData[inputData].keys():\n resultData[inputData][answer] = [eval(x) for x in resultData[inputData][answer] if is_integer(x)]\n #resultData[inputData][answer] = np.mean(resultData[inputData][answer])\n resultData[inputData][answer] = np.std(resultData[inputData][answer])\n print answer, resultData[inputData][answer]\n answerList.append(resultData[inputData][answer])\n print \"mean: \", np.mean(answerList)\n\n"},{"alpha_fraction":0.6615158319473267,"alphanum_fraction":0.6857984066009521,"avg_line_length":27.11724090576172,"blob_id":"59eda82a02810568e0283b7b641928c11df11c13","content_id":"438233b2382da77003e30b8fdd85da3f83593bdf","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":4077,"license_type":"no_license","max_line_length":89,"num_lines":145,"path":"/src/sampling/utils/keras/variational_autoencoder_model.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"'''This script demonstrates how to build a variational autoencoder with Keras.\n\nReference: \"Auto-Encoding Variational Bayes\" https://arxiv.org/abs/1312.6114\n'''\nimport numpy as np\nimport matplotlib.pyplot as plt\nimport glob\nfrom skimage.io import imread\nfrom skimage.transform import resize\n\nfrom keras.layers import Input, Dense, Lambda\nfrom keras.models import Model\nfrom keras import backend as K\nfrom keras import objectives\nfrom keras.datasets import mnist\nfrom keras.models import model_from_json\n\nimport math\n\n#batch_size = 100\nbatch_size = 1\noriginal_dim = 784\nlatent_dim = 128\nintermediate_dim = 256\n#nb_epoch = 50\nnb_epoch = 1\nepsilon_std = 0.01\n\nx = Input(batch_shape=(batch_size, original_dim))\nh = Dense(intermediate_dim, activation='relu')(x)\nz_mean = Dense(latent_dim)(h)\nz_log_var = Dense(latent_dim)(h)\n\n\ndef sampling(args):\n z_mean, z_log_var = args\n\n # karan\n #batch_size = 100\n #latent_dim = 128\n #epsilon_std = 0.01\n \n epsilon = K.random_normal(shape=(batch_size, latent_dim), mean=0.,\n std=epsilon_std)\n return z_mean + K.exp(z_log_var / 2) * epsilon\n\n# note that \"output_shape\" isn't necessary with the TensorFlow backend\nz = Lambda(sampling, output_shape=(latent_dim,))([z_mean, z_log_var])\n\n# we instantiate these layers separately so as to reuse them later\ndecoder_h = Dense(intermediate_dim, activation='relu')\ndecoder_mean = Dense(original_dim, activation='sigmoid')\nh_decoded = decoder_h(z)\nx_decoded_mean = decoder_mean(h_decoded)\n\n\ndef vae_loss(x, x_decoded_mean):\n xent_loss = original_dim * objectives.binary_crossentropy(x, x_decoded_mean)\n kl_loss = - 0.5 * K.sum(1 + z_log_var - K.square(z_mean) - K.exp(z_log_var), axis=-1)\n return xent_loss + kl_loss\n\nvae = Model(x, x_decoded_mean)\nvae.compile(optimizer='rmsprop', loss=vae_loss)\n\n# train the VAE on MNIST digits\n(x_train, y_train), (x_test, y_test) = mnist.load_data()\n\n# load all images in folder\n\ndef image_data(file):\n\timg = imread(file, as_grey=True)\n img = resize(img, (28, 28))\n img = img.flatten()\n return img\n\ndef folder_data(folder):\n\tfiles = glob.glob(folder + \"/*.png\")\n\tprint(\"Calculating descriptors. Number of images is\", len(files))\n\treturn [image_data(file) for file in files]\n\nfolder = \"images/1108\"\n#folder = \"../../../../_swarm-lfd-data/Flocking\"\nx_folder = folder_data(folder)\nsplitPoint = int(math.ceil(len(x_folder)*2.0/3.0))\nx_train = x_folder[:splitPoint]\nx_test = x_folder[splitPoint:]\n\n# artificially increase size of lists\n#x_train *= 100\n#x_test *= 100\n\nx_train = np.array(x_train)\nx_test = np.array(x_test)\n\nprint(len(x_train))\nprint(len(x_test))\n\nx_train = x_train.astype('float32') / 255.\nx_test = x_test.astype('float32') / 255.\nx_train = x_train.reshape((len(x_train), np.prod(x_train.shape[1:])))\nx_test = x_test.reshape((len(x_test), np.prod(x_test.shape[1:])))\n\nprint(x_train.shape)\nprint(x_train.shape[1:])\n\nvae.fit(x_train, x_train,\n shuffle=True,\n nb_epoch=nb_epoch,\n batch_size=batch_size,\n validation_data=(x_test, x_test))\n\n'''\n# save model as JSON\nmodel = vae\nmodelAsJson = model.to_json()\nwith open(\"variational_autoencoder_model.json\", \"w\") as jsonFile:\n jsonFile.write(modelAsJson)\n# serialize weights to HDF5\nmodel.save_weights('variational_autoencoder_model.h5')\n'''\n\n'''\n# load json and create model\njson_file = open('variational_autoencoder_model.json', 'r')\nloaded_model_json = json_file.read()\njson_file.close()\nloaded_model = model_from_json(loaded_model_json)\n# load weights into new model\nloaded_model.load_weights(\"variational_autoencoder_model.h5\")\n\n# get output for specific layer\n# with a Sequential model\nmodel = loaded_model\nprint(len(model.layers))\n\nimage = image_data(\"images/1108/1108_200.png\")\nimage = image.astype('float32') / 255.\nimage = image.reshape((1, np.prod(x_train.shape[1:])))\n\nget_layer_output = K.function([model.layers[0].input, K.learning_phase()],\n [model.layers[3].output])\n\nlayer_output = get_layer_output([image, 0])[0]\nprint(len(layer_output[0]))\n'''\n"},{"alpha_fraction":0.7346625924110413,"alphanum_fraction":0.7346625924110413,"avg_line_length":30.047618865966797,"blob_id":"ccf0ccb753afb63d35c23cf504ea9943b2c7a1ad","content_id":"f4408024e0b759e7488b6f73af44748bd904bdd0","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":652,"license_type":"no_license","max_line_length":54,"num_lines":21,"path":"/src/plot_feedback.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import sys\nimport matplotlib.pyplot as plt\nimport numpy as np\n\nwith open(\"feedback_output_difference.txt\") as inFile:\n difference = inFile.readlines()\n difference = [eval(x) for x in difference]\n print >> sys.stderr, np.sum(difference)\nplt.figure()\nplt.plot(difference, label=\"difference\")\nplt.legend()\nplt.savefig(\"feedback_output_difference.png\")\n\nwith open(\"feedback_input_difference.txt\") as inFile:\n difference = inFile.readlines()\n difference = [eval(x) for x in difference]\n print >> sys.stderr, np.sum(difference)\nplt.figure()\nplt.plot(difference, label=\"difference\")\nplt.legend()\nplt.savefig(\"feedback_input_difference.png\")\n"},{"alpha_fraction":0.6097992658615112,"alphanum_fraction":0.6257379055023193,"avg_line_length":20.71794891357422,"blob_id":"2f509e1321a4c951de5d0934a940ef1d429544c7","content_id":"beea8a739a0d74e34470bdd642aa59286529886b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1694,"license_type":"no_license","max_line_length":66,"num_lines":78,"path":"/src/sampling/schelling/predict_schelling.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"from model import SchellingModel\nimport random\nimport sys\nimport numpy as np\n\n# only use three of the several ALPs\n# density (0,1)\n# minority_pc (0,1)\n\n# use one model-level SLP\n# happy\n\n# constant ALPs\nHEIGHT = 20\nWIDTH = 20\nHOMOPHILY = 4\n\n# constants\n# max_iters in the example is specified as 1000\n# so we run for 1000 steps\nINDEPENDENT_VARIABLES = 2\nDEPENDENT_VARIABLES = 1\nSAMPLES = 10\nTIME_LAPSE = 1000\n\ndef main_schelling(density, minorityPc):\n # instantiate and run model\n model = SchellingModel(\n height=HEIGHT,\n width=WIDTH,\n density=density,\n minority_pc=minorityPc,\n homophily=HOMOPHILY)\n\n # run time lapse\n for i in range(TIME_LAPSE):\n try:\n # step\n model.step()\n except:\n # no empty cells\n pass\n \n # collect data for next steps\n dependentValues = []\n\n for i in range(SAMPLES):\n try:\n # step\n model.step()\n except:\n # saturated\n # no empty cells\n pass\n\n # read data\n data = model.datacollector.get_model_vars_dataframe() \n dependentValues.append(np.mean(list(data.happy)[-SAMPLES-1:]))\n\n return dependentValues\n \ndef main():\n # simulate using specific ALPs\n alpConfigurationList = eval(sys.argv[1])\n\n # store corresponding slps\n slpConfigurationList = []\n \n for alpConfiguration in alpConfigurationList:\n [density, minorityPc] = alpConfiguration\n \n # run model using those ALPs\n dependentValues = main_schelling(density, minorityPc)\n slpConfigurationList.append(dependentValues)\n\n print(slpConfigurationList)\n\nmain()\n"},{"alpha_fraction":0.6643073558807373,"alphanum_fraction":0.6723963618278503,"avg_line_length":28.08823585510254,"blob_id":"c60013c803d1d6cc6161f9ad0b394b6365866421","content_id":"08ff94a342466b67f7faef2c5f2157a1a3a6fca0","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":989,"license_type":"no_license","max_line_length":97,"num_lines":34,"path":"/data/domaindata/reduce.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import numpy as np\nfrom sklearn.decomposition import KernelPCA\nimport cPickle\n\nnumDependent = 50\n\n# read input\nwith open(\"processed_turbulence_vgg.txt\", \"r\") as inFile:\n lines = inFile.readlines()[1:]\n lines = [[eval(x) for x in line.split(\" \")] for line in lines]\n\nlines = np.array(lines)\nalps = lines[:,:2]\nslps = lines[:,2:]\nkpca = KernelPCA(n_components=numDependent, kernel=\"rbf\")\nmodel = kpca.fit(slps)\nslpsKpca = model.transform(slps)\nxKpca = np.concatenate((alps, slpsKpca), axis=1)\n\nrows, columns = xKpca.shape\n\n# write output\nwith open(\"processed_turbulence_vgg_reduced.txt\", \"w\") as outFile:\n outFile.write(\"i\"*2 + \"d\"*numDependent)\n\n for rowIndex in range(rows):\n string = \"\\n\"\n for columnIndex in range(columns):\n string += str(xKpca[rowIndex][columnIndex]) + \" \"\n string = string[:-1]\n outFile.write(string)\n\n# save transform\ncPickle.dump(model, open(\"../../src/sampling/utils/transforms/turbulence_reduced_model.p\", \"wb\"))\n"},{"alpha_fraction":0.6873920559883118,"alphanum_fraction":0.6925734281539917,"avg_line_length":24.130434036254883,"blob_id":"957e66a619fa6e0f75520edde500d22267dd878c","content_id":"f62f510793c3d6900bfd2cad563db98bf7e558cc","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":579,"license_type":"no_license","max_line_length":69,"num_lines":23,"path":"/src/do_feedback.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# clean\nrm -f feedback_*.txt\nrm -f suggested_*.txt\nrm -f queried_*.txt\n\n# run multiple times to generate multiple sets of images\nexport ITERATIONS=100\n\nfor i in $(seq \"$ITERATIONS\")\ndo\n # store the current iteration value\n echo $i > feedback_iteration.txt\n \n # run experiment\n # simulate results\n ./lfd.sh && ./simulate.sh\n\n # observe data for slps\n # append to file\n cat suggested_slps.txt >> feedback_suggested_slps.txt\n cat queried_slps.txt >> feedback_queried_slps.txt\n cat regression_parameter.txt >> feedback_regression_parameter.txt\ndone\n\n"},{"alpha_fraction":0.5935162305831909,"alphanum_fraction":0.6084787845611572,"avg_line_length":32.072166442871094,"blob_id":"a433d1c05e443d18949c083047c0db92919fd502","content_id":"23ac65e121d2a200e5cf53c09a4ae04d5e00b15f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3208,"license_type":"no_license","max_line_length":106,"num_lines":97,"path":"/snippets/eum/plot_suggested_slps.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import matplotlib.pyplot as plt\nimport numpy as np\n\n#demonstrations = [\"a\", \"b\", \"c\"]\n#demonstrations = [\"_a\", \"_b\", \"_c\"]\ndemonstrations = [\"a\"]\n#demonstrations = [\"_a\"]\n#demonstrations = [\"a_filtered\"]\n\nfolderName = \"error_weighted\"\n#folderName = \"min_error\"\n\nnumberOfBins = 10\n\nwith open(\"../../src/sampling/eum/scale_data_eum.txt\", \"r\") as inFile:\n allLines = inFile.readlines()\n maxValueMean = eval(allLines[0])[-2]\n maxValueMedian = eval(allLines[0])[-1]\n minValueMean = eval(allLines[1])[-2]\n minValueMedian = eval(allLines[1])[-1]\n\nfor index, item in enumerate(demonstrations):\n allPointsX = []\n allPointsY = []\n errors = []\n \n # scatter\n \n plt.figure()\n with open(folderName + \"/suggested_slps_\" + item + \".txt\", \"r\") as inFile:\n for line in inFile:\n try:\n point = eval(line)[0]\n plt.scatter(point[0], point[1], c=\"b\", marker=\"o\", alpha=0.3)\n allPointsX.append(point[0])\n allPointsY.append(point[1])\n\n # compute error\n with open(\"demonstration_slps.txt\", \"r\") as inFile:\n demonstration = eval(inFile.readlines()[index])\n error = (point[0]-demonstration[0])**2 + (point[1]-demonstration[1])**2\n errors.append(error)\n \n except:\n continue\n \n plt.scatter(point[0], point[1], c=\"b\", marker=\"o\", alpha=0.3, label=\"Suggested SLPs\")\n\n with open(\"demonstration_slps.txt\", \"r\") as inFile:\n point = eval(inFile.readlines()[index])\n plt.scatter(point[0], point[1], c=\"r\", marker=\"s\", alpha=1.0, label=\"Demonstration SLPs\")\n \n plt.scatter(np.mean(allPointsX), np.mean(allPointsY), c=\"g\", marker=\"s\", alpha=1.0, label=\"Mean SLPs\")\n\n print np.mean(errors)\n print np.std(errors)\n \n ax = plt.gca()\n\n ticks = ax.get_xticks()\n minValue = minValueMean\n maxValue = maxValueMean\n scaledTicks = [(value-minValue)/(maxValue-minValue) for value in ticks]\n scaledTicks = [int(value*100)/100.0 for value in scaledTicks]\n ax.set_xticklabels(scaledTicks)\n \n ticks = ax.get_yticks()\n minValue = minValueMedian\n maxValue = maxValueMedian\n scaledTicks = [(value-minValue)/(maxValue-minValue) for value in ticks]\n scaledTicks = [int(value*100)/100.0 for value in scaledTicks]\n ax.set_yticklabels(scaledTicks)\n\n plt.ylabel(\"Median\")\n plt.xlabel(\"Mean\")\n plt.legend()\n plt.savefig(folderName + \"/suggested_slps_\" + item)\n\n # histogram\n \n plt.figure()\n histogram, xEdges, yEdges = np.histogram2d(allPointsX, allPointsY, bins=numberOfBins)\n plt.imshow(histogram, interpolation='nearest', origin='low')\n plt.xlabel(\"Mean Final Position\")\n plt.ylabel(\"Median Final Position\")\n plt.colorbar(label=\"Number of Points\")\n ax = plt.gca()\n\n # works for normalized values\n ticks = ax.get_xticks()\n ticks = [value/numberOfBins for value in ticks]\n ax.set_xticklabels(ticks)\n ticks = ax.get_yticks()\n ticks = [value/numberOfBins for value in ticks]\n ax.set_yticklabels(ticks)\n\n plt.savefig(folderName + \"/histogram_suggested_slps_\" + item)\n"},{"alpha_fraction":0.6872605085372925,"alphanum_fraction":0.7131226062774658,"avg_line_length":30.1641788482666,"blob_id":"b8c6d0c896548676b6ee1e7450744ec14a4bc566","content_id":"318d23cb9b11601480f20c3fd535fa0566bf5212","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2088,"license_type":"no_license","max_line_length":74,"num_lines":67,"path":"/src/sampling/utils/keras/mnist_cnn_model.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"'''Trains a simple convnet on the MNIST dataset.\n\nGets to 99.25% test accuracy after 12 epochs\n(there is still a lot of margin for parameter tuning).\n16 seconds per epoch on a GRID K520 GPU.\n'''\n\nimport numpy as np\nfrom keras.datasets import mnist\nfrom keras.utils import np_utils\nfrom keras.models import model_from_json\nfrom skimage.io import imread\nfrom skimage.transform import resize\n\n# load json and create model\njson_file = open('mnist_cnn_model.json', 'r')\nloaded_model_json = json_file.read()\njson_file.close()\nloaded_model = model_from_json(loaded_model_json)\n# load weights into new model\nloaded_model.load_weights(\"mnist_cnn_model.h5\")\nprint(\"Loaded model from disk\")\n\n# the data, shuffled and split between train and test sets\n# input image dimensions\nnb_classes = 10\nimg_rows, img_cols = 28, 28\n(X_train, y_train), (X_test, y_test) = mnist.load_data()\nX_train = X_train.reshape(X_train.shape[0], 1, img_rows, img_cols)\nX_test = X_test.reshape(X_test.shape[0], 1, img_rows, img_cols)\nX_train = X_train.astype('float32')\nX_test = X_test.astype('float32')\nX_train /= 255\nX_test /= 255\nprint('X_train shape:', X_train.shape)\nprint(X_train.shape[0], 'train samples')\nprint(X_test.shape[0], 'test samples')\n\n# convert class vectors to binary class matrices\nY_train = np_utils.to_categorical(y_train, nb_classes)\nY_test = np_utils.to_categorical(y_test, nb_classes)\n\n# test model\nmodel = loaded_model\nmodel.compile(loss='categorical_crossentropy',\n optimizer='adadelta',\n metrics=['accuracy'])\n\n# import image and resize\nimage = imread(\"409_200.png\", as_grey=True)\nimage = resize(image, (28, 28))\n\nfrom skimage.io import imshow\n\nimage = image.reshape(1, 1, 28, 28)\nprint(image.shape)\nclasses = model.predict_classes(image)\nprint(classes)\n\n# get output for specific layer\nfrom keras import backend as K\n# with a Sequential model\nprint(len(model.layers))\nget_layer_output = K.function([model.layers[0].input, K.learning_phase()],\n [model.layers[7].output])\nlayer_output = get_layer_output([image, 0])[0]\nprint(layer_output)\n"},{"alpha_fraction":0.7064123153686523,"alphanum_fraction":0.7095919251441956,"avg_line_length":34.60377502441406,"blob_id":"f5c21909f4bdf333f86c6c98da59f185db0f0011","content_id":"2a7d3206d3758b6fad54c07d67e94b4f212fcf1a","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":1887,"license_type":"no_license","max_line_length":116,"num_lines":53,"path":"/src/sampling/sampling_process_raw.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# copy raw data file if needed\nif [ $IMAGES == 1 ]; then\n # raw data file is present in non-repository directory\n cp ../../../_swarm-lfd-data/$ABM/$ABM.txt .\nfi\n\n# separate head from rest of data\nhead -n 1 $ABM.txt > header.txt\ntail -n +2 $ABM.txt > $ABM\"_.txt\"\nmv $ABM\"_.txt\" $ABM.txt\n\n# process images if any\n# check if this abm depends on images\nif [ $IMAGES == 1 ]; then\n # images and raw data file are present in non-repository directory\n mv ../../../_swarm-lfd-data/$ABM/images $ABM/\n python utils/process_training_images.py\n mv $ABM/images ../../../_swarm-lfd-data/$ABM/\nfi\n\n# extract independent component\n# generate dependent component\n# combine results\n# overwrite header file\nexport INDEPENDENT_HEADER=\"i\"\nexport DEPENDENT_HEADER=\"d\"\nexport INDEPENDENT_HEADER_STRING=$(sed \"s/[^\"$INDEPENDENT_HEADER\"']//g\" <<< cat header.txt)\nexport DEPENDENT_HEADER_STRING=$(printf $DEPENDENT_HEADER {1..$NUM_DEPENDENT})\necho $INDEPENDENT_HEADER_STRING$DEPENDENT_HEADER_STRING > header.txt\n\n# prepare data for amf input\necho \"==>preparing data for amf\"\npython ../map.py $ABM.txt $ABM > processed_$ABM.txt\ncat header.txt processed_$ABM.txt > temp.txt\nmv temp.txt processed_$ABM.txt\nrm header.txt\nmv processed_$ABM.txt ../../data/domaindata\n\n# if image processing was used, append descriptor setting to processed filename\nif [ $IMAGES == 1 ]; then\n # generate lower case descriptor name\n export DESCRIPTOR_SETTING_LOWER=$(echo \"$DESCRIPTOR_SETTING\" | tr '[:upper:]' '[:lower:]')\n mv ../../data/domaindata/processed_$ABM.txt ../../data/domaindata/processed_$ABM\"_\"$DESCRIPTOR_SETTING_LOWER.txt\nfi\n\n# this file is no longer needed for computation\n# the forward and reverse mappings are built from the processed file\necho \"==>cleaning\"\nrm $ABM.txt\n\n# move scale data to the abm folder\n# assume only one scale data file exists here right now\nmv scale_data_$ABM*.txt $ABM\n"},{"alpha_fraction":0.6077530384063721,"alphanum_fraction":0.6122422218322754,"avg_line_length":44.515625,"blob_id":"e38b994312427bd1fb97cca60eade85e1263f253","content_id":"13a7dbdd0afcb62d7bc875cb3eafbe47016f7074","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":37869,"license_type":"no_license","max_line_length":285,"num_lines":832,"path":"/src/lfd.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# import libraries\nimport matplotlib\nmatplotlib.use('Agg')\n\nimport sys\nimport amf\nimport amf.misc as misc\n\nfrom sklearn import svm\nfrom sklearn.cluster import KMeans\nimport scipy.io as sio\n\nfrom mpl_toolkits.mplot3d import Axes3D\nimport matplotlib.pyplot as plt\n\nimport os\nimport time\nimport random\n\nimport cPickle\nimport gzip\nimport numpy as np\nfrom sets import Set\n\nimport sampling.utils.featurize_image as fi\nimport sampling.utils.outlier_detection as od\n\n# constants\nPLOT_COLOR = {0: \"r\", 1: \"b\", 2: \"g\", 3: \"c\", 4: \"m\", 5: \"y\", 6: \"k\"}\nPLOT_SHAPE = {0: \"o\", 1: \"^\", 2: \"s\", 3: \"d\", 4: \"+\", 5: \"v\", 6: \"*\"}\nPLOT_DEMONSTRATIONS = {0: \"a\", 1: \"b\", 2: \"c\", 3: \"d\", 4: \"e\", 5: \"f\", 6: \"g\"}\n\nRUNS_PER_SLPS = 1\n\nIMAGE_FEATURE_PATH_PREFIX = \"sampling/\"\n\n# experiment class\nclass Experiment(object):\n\n # initialize experiment parameters based on file\n def __init__(self):\n # read configuration\n configurationFileName = os.environ[\"CONFIGURATION_FILE\"]\n configuration = open(configurationFileName).read()\n execd = {}\n exec(configuration, execd)\n\n self.trainingSize = execd['TRAINING_SIZE']\n self.validationSize = execd['VALIDATION_SIZE']\n self.rmGranularity = execd['RM_GRANULARITY']\n self.numIndependent = execd['NUM_INDEPENDENT']\n self.regression = execd['REGRESSION']\n\n # changes for feedback\n if(eval(os.environ[\"FEEDBACK\"]) == 1):\n # randomness causes too much noise for feedback\n self.validationSize = 0\n\n # modify regression parameter\n if(os.path.isfile(\"regression_parameter.txt\") == True):\n with open(\"regression_parameter.txt\", \"r\") as inFile:\n regressionParameter = inFile.readlines()[0]\n #self.regression = self.regression.replace(\"()\", \"(alpha=np.exp((-1*\" + regressionParameter + \")))\")\n self.regression = self.regression.replace(\"()\", \"(n_neighbors=int(np.floor(\" + regressionParameter + \")))\")\n \n # get environment variables\n self.abm = os.environ[\"ABM\"]\n self.demonstrationFolder = os.environ[\"DEMONSTRATION_FOLDER\"]\n self.descriptorSetting = os.environ[\"DESCRIPTOR_SETTING\"]\n self.numDependent = eval(os.environ[\"NUM_DEPENDENT\"])\n self.dataFileName = os.environ[\"DATA_FILE\"]\n self.scaleDataFileName = os.environ[\"SCALE_DATA_FILE\"]\n self.images = eval(os.environ[\"IMAGES\"])\n self.writeToDisk = eval(os.environ[\"WRITE_TO_DISK\"])\n \n # reverse mapping configuration: ranges of agent level parameters\n self.rmConfiguration = tuple()\n for i in range(self.numIndependent):\n self.rmConfiguration += ((0.0, 1.0), )\n self.rmConfiguration = list(self.rmConfiguration)\n\n # for plotting\n self.rmQueryCount = 0\n \n # mapping models\n self.fm = None\n self.rm = None\n \n # retrieve reverse mapping\n def get_mappings(self, slpsList, saveData=False):\n # parse slps list\n demonstrationDescriptors = [depValues for depValues, plotColor, plotShape in slpsList]\n \n # load data\n print >> sys.stderr, \"->Loading data set \" + self.dataFileName\n fullData = amf.data.load(self.dataFileName)\n \n # forward mapping using largest training size\n print >> sys.stderr, \"->Trainining FM\"\n trainingSize = self.trainingSize\n dataset, validation = amf.data.random_subsets(fullData, [trainingSize, self.validationSize], demonstrationDescriptors)\n fm = amf.ForwardMapping(self.regression, self.numDependent)\n fm.train(dataset)\n \n if(saveData == True):\n # gather ALPs and SLPs\n # plot them in space to observe distribution\n alpPoints = [item[0] for item in dataset]\n slpPoints = [item[1] for item in dataset]\n c = \"r\"\n m = \"o\"\n \n # save ALP dataset\n with open(\"alp_dataset_\" + str(int(time.time())) + \".txt\", \"w\") as alpDatasetFile:\n for alpPoint in alpPoints:\n alpDatasetFile.write(str(alpPoint) + \"\\n\")\n \n # save SLP dataset\n with open(\"slp_dataset_\" + str(int(time.time())) + \".txt\", \"w\") as slpDatasetFile:\n for slpPoint in slpPoints:\n slpDatasetFile.write(str(slpPoint) + \"\\n\")\n\n # plot ALPs\n fig = plt.figure()\n ax = fig.add_subplot(111, projection='3d')\n plotName = \"alp_points_\" + str(int(time.time())) + \".png\"\n alpPoints = zip(*alpPoints)\n [xs, ys, zs] = alpPoints\n ax.scatter(xs, ys, zs, c=c, marker=m)\n ax.set_xlabel(\"Max-align-turn\")\n ax.set_ylabel(\"Max-cohere-turn\")\n ax.set_zlabel(\"Max-separate-turn\")\n plt.savefig(plotName)\n plt.close(fig)\n \n fig = plt.figure()\n plotName = \"alp_points_dim1_\" + str(int(time.time())) + \".png\"\n plt.plot(xs, xs, c+m)\n plt.xlabel(\"ALP Dimension 1\")\n plt.savefig(plotName)\n plt.close(fig)\n \n fig = plt.figure()\n plotName = \"alp_points_dim2_\" + str(int(time.time())) + \".png\"\n plt.plot(ys, ys, c+m)\n plt.xlabel(\"ALP Dimension 2\")\n plt.savefig(plotName)\n plt.close(fig)\n \n fig = plt.figure()\n plotName = \"alp_points_dim3_\" + str(int(time.time())) + \".png\"\n plt.plot(zs, zs, c+m)\n plt.xlabel(\"ALP Dimension 3\")\n plt.savefig(plotName)\n plt.close(fig)\n \n # plot SLPs\n fig = plt.figure()\n plotName = \"slp_points_\" + str(int(time.time())) + \".png\"\n slpPoints = zip(*slpPoints)\n [xs, ys] = slpPoints\n plt.plot(xs, ys, c+m)\n plt.xlabel(\"SLP Dimension 1\")\n plt.ylabel(\"SLP Dimension 2\")\n plt.savefig(plotName)\n plt.close(fig)\n \n fig = plt.figure()\n plotName = \"slp_points_dim1_\" + str(int(time.time())) + \".png\"\n plt.plot(xs, xs, c+m)\n plt.xlabel(\"SLP Dimension 1\")\n plt.savefig(plotName)\n plt.close(fig)\n \n fig = plt.figure()\n plotName = \"slp_points_dim2_\" + str(int(time.time())) + \".png\"\n plt.plot(ys, ys, c+m)\n plt.xlabel(\"SLP Dimension 2\")\n plt.savefig(plotName)\n plt.close(fig)\n\n # reverse mapping using largest granularity\n print >> sys.stderr, \"->Trainining RM\"\n granularity = self.rmGranularity\n rm = amf.ReverseMapping(fm, self.rmConfiguration, granularity, slpsList)\n\n return [fm, rm]\n\n # query reverse mapping for configurations\n def query_rm(self, fm, rm, slps, ax, c, m, plotConfigurations=False):\n intersections = rm.all_intersections(slps)\n allConfigurations = []\n\n #\n # filter from all query results\n # \n \n simplexNumbers = []\n edgeNumbers = []\n _allConfigurations = []\n\n # check for special case\n # if wiggle happened, then only one configuration is returned\n if(len(intersections[0]) == 1):\n # configuration corresponds to simplex number 0 and edge number 0\n # an intersection is stored as [simplexNumber][edgeNumber][configuration]\n if(type(intersections[0][0][0]) is tuple):\n # no simplex number is returned in this case\n wiggledConfiguration = intersections[0][0][0]\n else:\n # different formats of return\n # in this case, a default simplex and edge number is attached\n # not sure why\n wiggledConfiguration = intersections[0][0][1][0][1]\n \n allConfigurations = [list(wiggledConfiguration)] \n return allConfigurations\n \n for intersection in intersections:\n currentEdgeNumbers = {}\n _configurations = {}\n simplexNumbers.append([simplexNumber for (simplexNumber, configurations) in intersection])\n \n for simplexNumber, configurations in intersection:\n currentEdgeNumbers[simplexNumber] = [edgeNumber for (edgeNumber, configuration) in configurations]\n _configurations[simplexNumber] = {}\n for edgeNumber, configuration in configurations:\n allConfigurations.append(configuration)\n _configurations[simplexNumber][edgeNumber] = configuration\n \n edgeNumbers.append(currentEdgeNumbers)\n _allConfigurations.append(_configurations)\n \n # get common simplexes\n simplexNumbers = [Set(numbers) for numbers in simplexNumbers]\n commonSimplexes = simplexNumbers[0]\n for index in range(1, len(simplexNumbers)):\n currentSimplexes = simplexNumbers[index]\n commonSimplexes = commonSimplexes.intersection(currentSimplexes)\n\n # get common edges\n commonEdges = {}\n for key in commonSimplexes:\n currentCommonEdges = Set(edgeNumbers[0][key])\n for index in range(1, len(edgeNumbers)):\n currentEdges = Set(edgeNumbers[index][key])\n currentCommonEdges = currentCommonEdges.intersection(currentEdges)\n commonEdges[key] = currentCommonEdges\n overlapCount = [len(values) for values in list(commonEdges.values())]\n \n # filter common edges\n #overlapAmount = [len(commonEdges[key]) for key in commonEdges]\n #print >> sys.stderr, overlapAmount\n #plt.figure()\n #plt.hist(overlapAmount)\n #plt.savefig(\"overlapAmount_before_filtering.png\")\n\n if(len(overlapCount) > 0):\n #plt.figure()\n #plt.hist(overlapCount)\n #plt.show()\n #sys.exit()\n\n overlapMean = np.mean(overlapCount)\n commonEdges = {key : commonEdges[key] for key in commonEdges if len(commonEdges[key]) >= overlapMean}\n commonSimplexes = list(commonEdges.keys())\n\n #overlapAmount = [len(commonEdges[key]) for key in commonEdges]\n #print >> sys.stderr, overlapAmount\n #plt.figure()\n #plt.hist(overlapAmount)\n #plt.savefig(\"overlapAmount_after_filtering.png\")\n #sys.exit()\n \n # now we need to cycle through the configurations again\n # only pick ones corresponding to filtered edges\n # TODO: we may want to combine configurations across SLP intersections for optimality. can we?\n allConfigurationsCommon = []\n for commonSimplex in commonSimplexes:\n currentCommonEdges = commonEdges[commonSimplex]\n for commonEdge in currentCommonEdges:\n for alpIndex in range(len(simplexNumbers)):\n currentConfiguration = _allConfigurations[alpIndex][commonSimplex][commonEdge]\n #print >> sys.stderr, currentConfiguration\n allConfigurationsCommon.append(currentConfiguration)\n\n # filter configurations based on common simplexes\n if(eval(os.environ[\"CONFIGURATIONS_PRUNING\"]) == 1):\n allConfigurations = allConfigurationsCommon\n else:\n # no overlapping edges found\n # pruning cannot be used in this case\n pass\n\n # outlier detection code\n # outlier detection based on hubness\n # approximate labeling based on minority clustering\n\n if(eval(os.environ[\"CONFIGURATIONS_OUTLIER_DETECTION\"]) == 1):\n # outlier detection only valid for a dataset of minimum size\n # meanshift clustering sometimes fails if it somehow finds nans\n try:\n # do outlier detection\n [odPoints, labels] = od.get_approximate_labels(allConfigurations)\n [odPoints, labels] = od.do_lpod(odPoints, labels)\n \n allConfigurationsOd = []\n for configurationIndex in range(len(allConfigurations)):\n currentConfiguration = odPoints[configurationIndex] \n if(labels[configurationIndex] == 0):\n # not an outlier\n allConfigurationsOd.append(list(currentConfiguration))\n \n allConfigurations = allConfigurationsOd\n except:\n # outlier detection had an error\n # it is not used in this case\n pass\n \n # add solution to scatter plot\n if(plotConfigurations == True):\n print >> sys.stderr, \"->Plotting configurations\" \n allXs = []\n allYs = []\n allZs = []\n allErrors = []\n configurationsToPlot = []\n \n for configuration in allConfigurations:\n error = misc.list_sub(slps, fm.predict(configuration))\n error = sum([x**2 for x in error])\n configurationsToPlot.append((error, configuration))\n configurationsToPlot.sort() \n\n # filter configurations here\n # checking threshold effect\n allConfigurationsToPlot = [x for x in configurationsToPlot]\n histogramData = [error for error, configuration in allConfigurationsToPlot]\n mu = np.mean(histogramData)\n sigma = np.std(histogramData)\n #configurationsToPlot = [x for x in configurationsToPlot if (x[0] < (mu + 0.1*sigma) and x[0] > (mu - 0.1*sigma))]\n \n for error, configuration in configurationsToPlot:\n xs, ys, zs = configuration\n allXs.append(xs)\n allYs.append(ys)\n allZs.append(zs)\n allErrors.append(error)\n\n odPointsAllXs = [x[0] for x in odPoints]\n odPointsAllYs = [x[1] for x in odPoints]\n odPointsAllZs = [x[2] for x in odPoints]\n \n # plot point errors in 3d space \n plot = ax.scatter(allXs, allYs, allZs, c=allErrors, marker=m, cmap=\"cool\")\n plt.colorbar(plot, label=\"FM Prediction Error\")\n ax.set_xlabel(\"Max-align-turn\")\n ax.set_ylabel(\"Max-cohere-turn\")\n ax.set_zlabel(\"Max-separate-turn\")\n plt.savefig(\"solution_space.png\")\n\n # plot point labels in 3d space\n fig = plt.figure()\n ax = fig.add_subplot(111, projection='3d')\n cmapClusters = matplotlib.cm.get_cmap(\"cool\", 2)\n plot = ax.scatter(odPointsAllXs, odPointsAllYs, odPointsAllZs, c=labels, marker=m, cmap=cmapClusters)\n plt.colorbar(plot, label=\"Cluster\").set_ticks([0, 1])\n ax.set_xlabel(\"Max-align-turn\")\n ax.set_ylabel(\"Max-cohere-turn\")\n ax.set_zlabel(\"Max-separate-turn\")\n plt.savefig(\"solution_space_clusters.png\")\n \n # error histogram\n plt.figure()\n plt.hist(histogramData)\n plt.xlabel(\"FM Prediction Error\")\n plt.title(\"mu = \" + str(mu) + \", sigma = \" + str(sigma))\n plt.savefig(\"error_histogram\")\n \n projectDimensions = True\n if(projectDimensions == True):\n # plotting for single demonstration SLP configuration\n\n # plot projections with error\n plt.figure()\n plt.scatter(allXs, allYs, c=allErrors, marker=m, cmap=\"cool\")\n plt.colorbar(label=\"FM Prediction Error\")\n plt.gca().set_xlabel(\"Max-align-turn\")\n plt.gca().set_ylabel(\"Max-cohere-turn\")\n plt.savefig(\"solution_space_xy.png\")\n plt.figure()\n plt.scatter(allXs, allZs, c=allErrors, marker=m, cmap=\"cool\")\n plt.colorbar(label=\"FM Prediction Error\")\n plt.gca().set_xlabel(\"Max-align-turn\")\n plt.gca().set_ylabel(\"Max-separate-turn\")\n plt.savefig(\"solution_space_xz.png\")\n plt.figure()\n plt.scatter(allYs, allZs, c=allErrors, marker=m, cmap=\"cool\")\n plt.colorbar(label=\"FM Prediction Error\")\n plt.gca().set_xlabel(\"Max-cohere-turn\")\n plt.gca().set_ylabel(\"Max-separate-turn\")\n plt.savefig(\"solution_space_yz.png\")\n\n # plot projections with labels\n plt.figure()\n plt.scatter(odPointsAllXs, odPointsAllYs, c=labels, marker=m, cmap=cmapClusters)\n plt.colorbar(label=\"Cluster\").set_ticks([0, 1])\n plt.gca().set_xlabel(\"Max-align-turn\")\n plt.gca().set_ylabel(\"Max-cohere-turn\")\n plt.savefig(\"solution_space_xy_clusters.png\")\n plt.figure()\n plt.scatter(odPointsAllXs, odPointsAllZs, c=labels, marker=m, cmap=cmapClusters)\n plt.colorbar(label=\"Cluster\").set_ticks([0, 1])\n plt.gca().set_xlabel(\"Max-align-turn\")\n plt.gca().set_ylabel(\"Max-separate-turn\")\n plt.savefig(\"solution_space_xz_clusters.png\")\n plt.figure()\n plt.scatter(odPointsAllYs, odPointsAllZs, c=labels, marker=m, cmap=cmapClusters)\n plt.colorbar(label=\"Cluster\").set_ticks([0, 1])\n plt.gca().set_xlabel(\"Max-cohere-turn\")\n plt.gca().set_ylabel(\"Max-separate-turn\")\n plt.savefig(\"solution_space_yz_clusters.png\")\n\n # clear memory\n del allXs, allYs, allZs, allErrors, configurationsToPlot, allConfigurationsToPlot\n\n return allConfigurations\n\n # format configurations for classifiaction\n def format_data_xy(self, allConfigurationsList):\n X = []\n Y = []\n for index in range(len(allConfigurationsList)):\n X += allConfigurationsList[index]\n Y += [index]*len(allConfigurationsList[index])\n\n return [X, Y]\n\n # compute fm errors and select configuration for minimum\n def get_min_error_configurations(self, fm, allConfigurationsList, slps):\n minErrorConfigurations = []\n for allConfigurations in allConfigurationsList:\n # error[0] contains error value\n # error[1] contains configuration\n errors = [(misc.list_sub(slps, fm.predict(configuration)), configuration) for configuration in allConfigurations]\n errors = [(sum([x**2 for x in error[0]]), error[1]) for error in errors]\n errors.sort()\n minErrorConfigurations.append(errors[0][1])\n\n return minErrorConfigurations\n\n # compute average configuration\n def get_mean_configurations(self, allConfigurationsList):\n meanConfigurations = []\n for allConfigurations in allConfigurationsList:\n meanConfiguration = tuple(misc.col_average(allConfigurations))\n meanConfigurations.append(meanConfiguration)\n\n return meanConfigurations\n\n def process_error_weighted_configuration(self, configuration, fm, slps, errorWeightedConfiguration, normalizationFactor, zeroErrorConfigurations):\n prediction = fm.predict(configuration)\n error = 0\n\n for index in range(len(slps)):\n error += (slps[index] - prediction[index])**2\n \n if(error > 0):\n if(zeroErrorConfigurations == 0):\n # no zero error configuration found yet\n normalizationFactor += 1/error\n for index in range(len(errorWeightedConfiguration)):\n errorWeightedConfiguration[index] += configuration[index]/error\n else:\n # error is zero\n if(zeroErrorConfigurations == 0):\n # first time we found a zero error configuration\n for index in range(len(errorWeightedConfiguration)):\n # wipe old data\n errorWeightedConfiguration[index] = 0\n zeroErrorConfigurations += 1\n \n # add new data\n for index in range(len(errorWeightedConfiguration)):\n errorWeightedConfiguration[index] += configuration[index]\n\n return [errorWeightedConfiguration, normalizationFactor, zeroErrorConfigurations]\n \n # compute error weighted average configuration\n def get_error_weighted_configurations(self, fm, allConfigurationsList, slps):\n errorWeightedConfigurations = []\n listIndex = -1\n \n for allConfigurations in allConfigurationsList:\n errorWeightedConfiguration = [0]*self.numIndependent\n normalizationFactor = 0\n zeroErrorConfigurations = 0\n\n if(self.writeToDisk == 1):\n # read from disk\n listIndex += 1\n with gzip.open(\"configuration_\" + str(listIndex+1) + \".pgz\", 'rb') as configurationFile:\n while 1:\n try:\n # read configuration\n configuration = cPickle.load(configurationFile)[listIndex]\n [errorWeightedConfiguration, normalizationFactor, zeroErrorConfigurations] = self.process_error_weighted_configuration(configuration, fm, slps, errorWeightedConfiguration, normalizationFactor, zeroErrorConfigurations)\n except EOFError:\n break\n else:\n # read from ram\n for configuration in allConfigurations:\n [errorWeightedConfiguration, normalizationFactor, zeroErrorConfigurations] = self.process_error_weighted_configuration(configuration, fm, slps, errorWeightedConfiguration, normalizationFactor, zeroErrorConfigurations)\n \n # nomalization\n for index in range(len(errorWeightedConfiguration)):\n if(zeroErrorConfigurations == 0):\n errorWeightedConfiguration[index] /= float(normalizationFactor)\n else:\n # we found some zero error configurations\n # we will only consider those configurations when computing a representative point\n errorWeightedConfiguration[index] /= float(zeroErrorConfigurations)\n \n errorWeightedConfigurations.append(tuple(errorWeightedConfiguration))\n\n return errorWeightedConfigurations\n\n # compute classifier based optimal configuration\n def get_svm_configurations(self, allConfigurationsList):\n # train svm classifier on configurations\n classifier = svm.SVC(probability=True)\n [X, Y] = self.format_data_xy(allConfigurationsList)\n classifier.fit(X, Y)\n\n # select configurations which maximize probability\n svmConfigurations = []\n for index in range(len(allConfigurationsList)):\n # compute score for each probability and select maximum\n configurationProbabilities = classifier.predict_proba(allConfigurationsList[index])\n configurationProbabilities = [(configurationProbabilities[i][index], allConfigurationsList[index][i]) for i in range(len(configurationProbabilities))]\n configurationProbabilities.sort(reverse=True)\n svmConfiguration = tuple(configurationProbabilities[0][1])\n svmConfigurations.append(svmConfiguration)\n\n return svmConfigurations\n\n # compute based on k means clustering\n def get_kmeans_configurations(self, allConfigurationsList):\n # train svm classifier on configurations\n nClusters = len(allConfigurationsList)\n kmeans = KMeans(init='k-means++', n_clusters=nClusters)\n [X, Y] = self.format_data_xy(allConfigurationsList)\n kmeans.fit(X, Y)\n\n # select centroids\n centroids = kmeans.cluster_centers_\n \n return centroids\n\n # run experiment\n def run(self, slpsList, plotConfigurations=False, saveError=False):\n [fm, rm] = self.get_mappings(slpsList)\n\n # append to model list\n self.fm = fm\n self.rm = rm\n\n # plot solution space\n fig = plt.figure()\n ax = fig.add_subplot(111, projection='3d')\n\n print >> sys.stderr, \"->Querying RM for configurations\" \n\n startTime = time.time()\n allConfigurationsList = []\n proxyPlots = []\n labels = []\n for slps,c,m in slpsList:\n allConfigurations = self.query_rm(fm, rm, slps, ax, c, m, plotConfigurations)\n allConfigurationsList.append(allConfigurations)\n\n proxyPlot = matplotlib.lines.Line2D([0],[0], linestyle=\"none\", c=c, marker=m)\n proxyPlots.append(proxyPlot)\n labels.append(\"ALPs \" + PLOT_DEMONSTRATIONS[self.rmQueryCount])\n ax.legend(proxyPlots, labels, numpoints = 1)\n self.rmQueryCount += 1\n stopTime = time.time()\n\n plt.close(fig)\n\n # how do we pick one?\n print >> sys.stderr, \"configurations retrieved: \" + str([len(allConfigurations) for allConfigurations in allConfigurationsList]) + \" [in \" + str(stopTime-startTime) + \" seconds]\"\n print >> sys.stderr, \"->Selecting Configurations\"\n\n # downsample by num dep\n # required to make higher dimension feature vectors tractable\n # can be removed later when computation is manageable\n downsampledAllConfigurationsList = []\n for allConfigurations in allConfigurationsList:\n if(len(allConfigurations) > 1):\n downsampledAllConfigurationsList.append(random.sample(allConfigurations, len(allConfigurations)/(self.numDependent**self.numIndependent)))\n else:\n downsampledAllConfigurationsList.append([allConfigurations[0]])\n allConfigurationsList = downsampledAllConfigurationsList\n\n # check forward mapping error for each configuratoin returned\n # check location in point cloud\n startTime = time.time()\n #configurationList = self.get_mean_configurations(allConfigurationsList)\n #configurationList = self.get_min_error_configurations(fm, allConfigurationsList, slps)\n configurationList = self.get_error_weighted_configurations(fm, allConfigurationsList, slps)\n #configurationList = self.get_svm_configurations(allConfigurationsList)\n #configurationList = self.get_kmeans_configurations(allConfigurationsList)\n stopTime = time.time()\n\n # experiment complete\n print >> sys.stderr, \"->Done\" + \" [in \" + str(stopTime-startTime) + \" seconds]\"\n\n # compute error for each configuration\n # print error\n demonstrationIndex = 0\n for slps,c,m in slpsList:\n configuration = configurationList[demonstrationIndex]\n prediction = fm.predict(configuration)\n error = 0\n for index in range(len(slps)):\n error += (slps[index] - prediction[index])**2\n if(saveError == True):\n with open(\"suggested_alps_fm_prediction_error.txt\", \"a\") as predictionErrorFile:\n predictionErrorFile.write(str(error) + \"\\n\")\n demonstrationIndex += 1\n \n return [allConfigurationsList, configurationList]\n\n# learning from visual demonstrations\ndef lfd(plotConfigurations=False, saveError=False):\n # expects configuration file as input\n experiment = Experiment()\n\n # load network if needed\n net = fi.get_network(experiment.descriptorSetting, IMAGE_FEATURE_PATH_PREFIX)\n\n # traverse all demonstration data\n plotIndex = 0\n slpsList = []\n\n # process directories in sorted order\n sortedPaths = []\n for dirName, subdirList, fileList in os.walk(experiment.demonstrationFolder):\n for sortedSubdirName in sorted(subdirList):\n sortedPaths.append(dirName + \"/\" + sortedSubdirName)\n\n # get slps for demonstrations\n for sortedPath in sortedPaths:\n for dirName, subdirList, fileList in os.walk(sortedPath):\n # only use non empty subdirectories\n if(len(fileList) > 0):\n # traverse all files in this directory\n if(bool(experiment.images) == True):\n # read dependent values as images\n dependentValues = [tuple(fi.get_features(net, dirName + \"/\" + fileName, experiment.descriptorSetting)) for fileName in fileList]\n else:\n # read dependent values as text\n dependentValues = [tuple([eval(x) for x in open(dirName + \"/\" + fileName).readlines()[0].replace(\"\\n\", \"\").split(\" \")]) for fileName in fileList]\n \n dependentValues = tuple(misc.col_average(dependentValues))\n #print >> sys.stderr, \"dependent values: \" + str(dependentValues)\n \n # edit scale data file name based on descriptor setting\n scaleDataFileNamePrefix = experiment.scaleDataFileName.split(\".\")[0]\n scaleDataFileNameExtension = experiment.scaleDataFileName.split(\".\")[1]\n scaleDataFileName = scaleDataFileNamePrefix + \"_\" + experiment.abm\n if(bool(experiment.images) == True):\n scaleDataFileName += \"_\" + experiment.descriptorSetting.lower()\n scaleDataFileName += \".\" + scaleDataFileNameExtension\n \n # scale demonstration data\n with open(\"sampling/\" + experiment.abm + \"/\" + scaleDataFileName, \"r\") as scaleDataFile:\n # read in order of writing in map.py\n maxs = eval(scaleDataFile.readline().strip(\"\\n\"))\n mins = eval(scaleDataFile.readline().strip(\"\\n\"))\n \n dependentMaxs = maxs[experiment.numIndependent:]\n dependentMins = mins[experiment.numIndependent:]\n dependentValuesMax = {i: dependentMaxs[i] for i in range(len(dependentMaxs))}\n dependentValuesMin = {i: dependentMins[i] for i in range(len(dependentMins))}\n dependentValues = tuple((dependentValues[index] - dependentValuesMin[index])/(dependentValuesMax[index] - dependentValuesMin[index]) if (dependentValuesMax[index] > dependentValuesMin[index]) else dependentValuesMax[index] for index in range(len(dependentValues)))\n\n # reduce dependent values if neccessary\n if(eval(os.environ[\"REDUCED\"]) == 1): \n model = cPickle.load(open(\"sampling/utils/transforms/\" + experiment.abm +\"_reduced_model.p\", \"rb\")) \n dependentValues = model.transform(np.array(dependentValues).reshape(1, -1))\n dependentValues = tuple(dependentValues[0])\n\n # if feedback is on\n # use feedback to modify queried slps\n if((eval(os.environ[\"FEEDBACK\"]) == 1) and (os.path.isfile(\"queried_slps.txt\") == True)):\n with open(\"queried_slps.txt\", \"r\") as inFile:\n queriedSlps = eval(inFile.readlines()[0])\n slpsList.append((queriedSlps, PLOT_COLOR[plotIndex], PLOT_SHAPE[plotIndex]))\n else:\n slpsList.append((dependentValues, PLOT_COLOR[plotIndex], PLOT_SHAPE[plotIndex]))\n\n plotIndex += 1\n\n # default value for queried slps\n # same as current demonstration slps\n queriedSlps = dependentValues\n\n #\n # feedback code\n #\n \n # random initialization if feedback is not yet applicable\n if(eval(os.environ[\"FEEDBACK\"]) == 1):\n with open(\"feedback_iteration.txt\", \"r\") as inFile:\n iteration = eval(inFile.readlines()[0])\n if(iteration <= 2):\n with open(\"regression_parameter.txt\", \"w\") as outFile:\n regressionParameterLowerLimit = eval(os.environ[\"REGRESSION_PARAMETER_LOWER_LIMIT\"])\n regressionParameterUpperLimit = eval(os.environ[\"REGRESSION_PARAMETER_UPPER_LIMIT\"])\n regressionParameter = regressionParameterLowerLimit + (regressionParameterUpperLimit-regressionParameterLowerLimit)*random.random()\n outFile.write(str(regressionParameter) + \"\\n\")\n \n # TODO: currently only works for a single slp suggestion\n # check if file exists\n if((eval(os.environ[\"FEEDBACK\"]) == 1) and (os.path.isfile(\"queried_slps.txt\") == True)):\n # use feedback to modify queried slps\n with open(\"queried_slps.txt\", \"r\") as inFile:\n queriedSlps = eval(inFile.readlines()[0])\n\n # check if file exists\n if(os.path.isfile(\"suggested_slps.txt\") == True):\n with open(\"suggested_slps.txt\", \"r\") as inFile:\n suggestedSlps = eval(inFile.readlines()[0])[0]\n\n # check consistency of suggested slps with current configuration\n if(len(suggestedSlps) == len(dependentValues)):\n #\n # fm dataset optimization\n #\n \n # scale suggested slps\n # TODO: make this a common function since its used everywhere\n suggestedSlps = tuple((suggestedSlps[index] - dependentValuesMin[index])/(dependentValuesMax[index] - dependentValuesMin[index]) if (dependentValuesMax[index] > dependentValuesMin[index]) else dependentValuesMax[index] for index in range(len(suggestedSlps)))\n\n # the differential accounts for the distance from the demonstration slps \n # demonstration slps are currently stored as dependentValues\n with open(\"feedback_iteration.txt\", \"r\") as inFile:\n iteration = inFile.readlines()[0].strip(\"\\n\")\n alpha = eval(os.environ[\"LEARNING_RATE_FM_DATASET\"])\n alpha /= eval(iteration) \n queriedSlps = tuple([queriedSlps[index] + alpha*(dependentValues[index] - suggestedSlps[index]) for index in range(len(dependentValues))])\n \n # write data\n with open(\"feedback_input_difference.txt\", \"a\") as outFile:\n outFile.write(str(np.abs(dependentValues[0] - queriedSlps[0])) + \"\\n\")\n with open(\"feedback_output_difference.txt\", \"a\") as outFile:\n outFile.write(str(dependentValues[0] - suggestedSlps[0]) + \"\\n\")\n\n #\n # regression parameter optimization initialization\n #\n\n # if just one iteration so far, move in a random direction\n with open(\"feedback_iteration.txt\", \"r\") as inFile:\n iteration = eval(inFile.readlines()[0])\n if(iteration > 2):\n # now we do feedback stuff\n with open(\"feedback_regression_parameter.txt\", \"r\") as feedbackRegressionParameterFile:\n with open(\"feedback_output_difference.txt\", \"r\") as feedbackOutputDifferenceFile:\n regressionParameterLines = feedbackRegressionParameterFile.readlines()\n outputDifferenceLines = feedbackOutputDifferenceFile.readlines()\n currentRegressionParameter = eval(regressionParameterLines[-1])\n previousRegressionParameter = eval(regressionParameterLines[-2])\n regressionParameterChange = currentRegressionParameter - previousRegressionParameter\n currentOutputDifference = eval(outputDifferenceLines[-1])\n previousOutputDifference = eval(outputDifferenceLines[-2])\n outputDifferenceChange = currentOutputDifference - previousOutputDifference\n \n alpha = eval(os.environ[\"LEARNING_RATE_REGRESSION_PARAMETER\"])\n alpha /= iteration\n\n # if output difference is negative, we move in direction to change\n nextRegressionParameter = currentRegressionParameter - regressionParameterChange*outputDifferenceChange*alpha\n with open(\"regression_parameter.txt\", \"w\") as outFile:\n outFile.write(str(nextRegressionParameter) + \"\\n\")\n \n # TODO: currently assumes only a single demonstration at a time. fix for multiple demonstrations\n with open(\"queried_slps.txt\", \"w\") as outFile:\n outFile.write(str(queriedSlps) + \"\\n\")\n \n configurationListPerRun = []\n\n for run in range(RUNS_PER_SLPS):\n # run and query slps\n [allConfigurationsList, configurationList] = experiment.run(slpsList, plotConfigurations=plotConfigurations, saveError=saveError)\n\n # scale configurations back from normalized space\n independentMaxs = maxs[:experiment.numIndependent]\n independentMins = mins[:experiment.numIndependent]\n independentValuesMax = {i: independentMaxs[i] for i in range(len(independentMaxs))}\n independentValuesMin = {i: independentMins[i] for i in range(len(independentMins))}\n normalizedConfigurationList = []\n for configuration in configurationList:\n normalizedConfigurationList.append(tuple([independentValuesMin[cIndex] + (independentValuesMax[cIndex]-independentValuesMin[cIndex])*configuration[cIndex] for cIndex in range(len(configuration))]))\n configurationList = tuple(normalizedConfigurationList)\n configurationListPerRun.append(configurationList)\n \n # return configurations\n print >> sys.stderr, \"configurationList: \" + str(configurationList)\n\n # compute average over runs\n averageConfigurationList = []\n for slpsIndex in range(len(slpsList)):\n slpsConfigurations = tuple([configurationListOneRun[slpsIndex] for configurationListOneRun in configurationListPerRun])\n averageConfigurationList.append(tuple(misc.col_average(slpsConfigurations)))\n\n print >> sys.stderr, \"averageConfigurationList: \" + str(averageConfigurationList)\n sys.stdout.write(str(averageConfigurationList))\n\n if(\"smt\" in experiment.regression):\n # circumvent excessive prints from smt\n with open(\"temp_average_configuration_list.txt\", \"w\") as outFile:\n outFile.write(str(averageConfigurationList) + \"\\n\")\n\n# main function\ndef main():\n lfd(plotConfigurations=False, saveError=True)\n\n# execute main\nmain()\n"},{"alpha_fraction":0.80923992395401,"alphanum_fraction":0.8159463405609131,"avg_line_length":48.703704833984375,"blob_id":"867655188aa23578bcdd69bc7ce77b29870b3d76","content_id":"4c735cf47aaed6e293b4cb5b9959c5e2a5fa13f8","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":1342,"license_type":"no_license","max_line_length":117,"num_lines":27,"path":"/data/dataset_analysis/dataset_analysis.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# demonstration specification\nexport DEMONSTRATION_ID=1\n\n# steps across dimensions\nexport DIM1_STEP=0.1\nexport DIM2_STEP=0.1\n\n# clean old datasets\nrm -f $DEMONSTRATION_ID/predictions_best/slp_dataset.txt\nrm -f $DEMONSTRATION_ID/predictions_second_best/slp_dataset.txt\nrm -f $DEMONSTRATION_ID/predictions_third_best/slp_dataset.txt\nrm -f $DEMONSTRATION_ID/predictions_worst/slp_dataset.txt\nrm -f $DEMONSTRATION_ID/predictions_second_worst/slp_dataset.txt\nrm -f $DEMONSTRATION_ID/predictions_third_worst/slp_dataset.txt\n\n# make a copy of the datasets\n# uniform name for ease of use\n\ncp $DEMONSTRATION_ID/predictions_best/slp_dataset* $DEMONSTRATION_ID/predictions_best/slp_dataset.txt\ncp $DEMONSTRATION_ID/predictions_second_best/slp_dataset* $DEMONSTRATION_ID/predictions_second_best/slp_dataset.txt\ncp $DEMONSTRATION_ID/predictions_third_best/slp_dataset* $DEMONSTRATION_ID/predictions_third_best/slp_dataset.txt\ncp $DEMONSTRATION_ID/predictions_worst/slp_dataset* $DEMONSTRATION_ID/predictions_worst/slp_dataset.txt\ncp $DEMONSTRATION_ID/predictions_second_worst/slp_dataset* $DEMONSTRATION_ID/predictions_second_worst/slp_dataset.txt\ncp $DEMONSTRATION_ID/predictions_third_worst/slp_dataset* $DEMONSTRATION_ID/predictions_third_worst/slp_dataset.txt\n\n# invoke python script\npython dataset_analysis.py $DEMONSTRATION_ID $DIM1_STEP $DIM2_STEP\n"},{"alpha_fraction":0.5875629782676697,"alphanum_fraction":0.5924505591392517,"avg_line_length":34.4640007019043,"blob_id":"978978042f347b82c80ef1a467ac77c89710c4c0","content_id":"0dd2c28a3cbaf433e292feaf53f2e43974e9e05e","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":13299,"license_type":"no_license","max_line_length":181,"num_lines":375,"path":"/src/amf/rm.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# rm.py contains the ReverseMapping\n\nimport misc\nimport numpy\nimport sys\nimport multiprocessing as mp\nimport time\nimport itertools\nimport random\nimport os\nimport gzip\nimport cPickle\n\n# parallel processing constants\nBATCH_SIZE = 8\n\n# hack to avoid rounding error\ndef _round(n):\n return numpy.round(n, 5)\n\nclass RMSimplex(object):\n # Simplexes are kind of immutable...\n\n # The LAST variable of each corner vector is the dependent variable\n # (this is important)\n\n def __init__(self, corners, slpsList):\n \"\"\"\n A simplex.\n corners should have the form ((ind0, ind1, ...), (dep1, dep2, dep3))\n \"\"\"\n # We already know which corners are adjanced to which\n # (each corner shares an edge with every other corner)\n\n self._corners = tuple(corners)\n self._inds = tuple( corner[0] for corner in corners )\n self._deps = tuple( corner[1] for corner in corners )\n self._min = misc.col_min(self._deps)\n self._max = misc.col_max(self._deps) \n self._edges = []\n\n # initialize edges only if needed\n # only if this simplex contains the query slp\n # otherwise the slp has no utility\n isContains = False\n \n for querySlps,c,m in slpsList:\n for index in range(len(querySlps)):\n isContains = isContains or self.contains(index, querySlps[index])\n \n if(isContains == True):\n for cidx in range(len(self._corners)):\n corner = self._corners[cidx]\n for oidx in range(cidx + 1, len(self._corners)):\n other_corner = self._corners[oidx]\n self._edges.append((corner, other_corner))\n self._edges = tuple(self._edges)\n\n @property\n def corners(self):\n return self._corners\n\n @property\n def edges(self):\n return self._edges\n\n def contains(self, slp_num, value):\n \"\"\"Determines if the value for slp_num idx lives inside this simplex.\"\"\"\n return ((value >= self._min[slp_num]) and (value <= self._max[slp_num]))\n\n def intersections(self, slp_num, value):\n \"\"\"Returns the approximate places that the value plane intersects with this simplex\"\"\"\n inters = []\n for edgeNumber, (scorner, bcorner) in enumerate(self.edges):\n # maintain order of scorner <= bcorner\n if(scorner[1][slp_num] > bcorner[1][slp_num]):\n tempCorner = bcorner\n bcorner = scorner\n scorner = tempCorner\n\n # check to see that this value is inside these edges\n if not ((value[slp_num] >= scorner[1][slp_num]) and (value[slp_num] <= bcorner[1][slp_num])):\n continue\n\n # this is the distance between the 2 corner values\n distance = bcorner[1][slp_num] - scorner[1][slp_num]\n\n if(distance == 0):\n # return any end point value\n inter = tuple( (1.0)*s for s, b in zip(scorner[0], bcorner[0]) )\n else:\n # this is the distance from 'value' from the smaller corner\n # if value is close to smaller, this ratio will be close to zero\n # if value is closer to bigger, this ratio will be close to one\n ratio = (value[slp_num] - scorner[1][slp_num]) / distance\n inter = tuple( (1.0-ratio)*s + ratio * b for s, b in zip(scorner[0], bcorner[0]) )\n \n inters.append([edgeNumber, inter])\n\n return tuple(inters)\n\n def __repr__(self):\n return 'simplex:' + repr(self._edges)\n\nclass ReverseMapping(object):\n def __init__(self, trained_fm, ranges, granularity, slpsList):\n\n self.fm = trained_fm\n self.granularity = granularity\n self.ranges = ranges\n self.steps = tuple( (ma - mi) / float(self.granularity) for mi, ma in self.ranges )\n self.slpsList = slpsList\n\n # disk write stuff\n self.writeToDisk = eval(os.environ[\"WRITE_TO_DISK\"])\n self.allIntersectionsCalls = 0\n \n # the dimensionality of the configuration space\n self.dim = len(ranges)\n\n # the dimensionality of the SLP space\n self.slp_dim = None\n\n self.__bin__ = misc.binary(self.dim)\n\n self.knots = {}\n self.simplexes = []\n print >> sys.stderr, 'building knots'\n startTime = time.time()\n self.build_knots()\n stopTime = time.time()\n print >> sys.stderr, 'done building knots, now building simplexes' + \" [\" + str(len(self.knots)) + \" in \" + str(stopTime-startTime) + \" seconds]\"\n startTime = time.time() \n self.build_simplexes()\n stopTime = time.time()\n print >> sys.stderr, 'done building simplexes' + \" [in \" + str(stopTime-startTime) + \" seconds]\"\n\n def build_knots(self):\n self.build_knots_inds()\n self.build_knots_deps()\n\n def build_knots_inds(self):\n values = [list(numpy.arange(self.ranges[curDim][0], self.ranges[curDim][1], (self.ranges[curDim][1] - self.ranges[curDim][0])/self.granularity)) for curDim in range(self.dim)]\n indexes = [range(len(values[i])) for i in range(len(values))]\n valuesPoints = [x for x in itertools.product(*values)]\n indexesPoints = [x for x in itertools.product(*indexes)]\n\n # sort items for correspondence\n valuesPoints.sort()\n indexesPoints.sort()\n\n # assign key-value pairs\n self.knots = dict(zip(indexesPoints, valuesPoints))\n\n def process_build_knots_deps(self, q, new_idx):\n new_vector = self.knots[new_idx]\n predicted = self.fm.predict(new_vector)\n q.put([new_idx, predicted])\n\n def build_knots_deps_parallel(self):\n # setup a list of processes that we want to run\n q = mp.Queue()\n processes = [mp.Process(target=self.process_build_knots_deps, args=(q, new_idx)) for new_idx in self.knots] \n batchSize = BATCH_SIZE\n batches = [processes[i:i+batchSize] for i in range(0, len(processes), batchSize)]\n\n for batch in batches:\n # run processes\n for p in batch:\n p.start() \n for p in batch:\n [new_idx, predicted] = q.get()\n self.knots[new_idx] = predicted\n for p in batch:\n p.join() \n\n # use last predicted point in dependent parameter space for length\n # we can use any key for this\n self.slp_dim = len(self.knots.values()[0])\n\n def build_knots_deps(self):\n # predict for each point in independent parameter space\n for new_idx in self.knots:\n new_vector = self.knots[new_idx]\n predicted = self.fm.predict(new_vector)\n self.knots[new_idx] = predicted\n \n # use last predicted point in dependent parameter space for length\n self.slp_dim = len(predicted)\n \n def build_simplexes(self, col=None):\n for key in sorted(self.knots.keys()):\n # check to see if this root is on the fringe and thus not actually a root\n for idx in key:\n if idx == (self.granularity - 1):\n break\n else:\n # extract simplexes\n # only keep those simplexes which are actually useful\n sims = self.simplexes_from_cube(key)\n sims = [x for x in sims if(len(x._edges) > 0)]\n\n if(self.writeToDisk == 1):\n # save to compressed file\n with gzip.open(\"sims.pgz\", 'ab') as simsFile:\n cPickle.dump(sims, simsFile, -1)\n else:\n # save to ram\n self.simplexes.extend(sims)\n\n '''\n # karan: testing downsampling\n sampleAmount = 100000\n\n if(len(self.simplexes) > sampleAmount):\n # downsample\n # for computational tractability\n sampledSimplexes = dict(random.sample(self.simplexes.items(), sampleAmount))\n self.simplexes = sampledSimplexes\n '''\n \n def wiggle(self, root):\n \"\"\" Returns the closest point to this root that is an actual root. \"\"\"\n distance, knot, knotKey = min((misc.distance(root, self.knots[knotKey]), self.knots[knotKey], knotKey) for knotKey in self.knots.keys())\n print >> sys.stderr, \"BADNESS: Wiggling point %s to knot %s\" % (repr(root), repr(knot))\n\n return [knot, knotKey]\n\n def cube_at_root(self, root):\n \"\"\" Returns the members of this cube with this point as its root. \"\"\"\n corners = []\n\n for perm in self.__bin__:\n corner = []\n for idx in xrange(self.dim):\n corner.append(perm[idx] + root[idx])\n corner = tuple(corner)\n val = self.knots[corner]\n corners.append((self.translate_idx(corner), val))\n\n return tuple(corners)\n\n def translate_idx(self, vec_idx):\n \"\"\" Translates a index to a configuration\"\"\"\n corner = []\n for idx in xrange(self.dim):\n corner.append(vec_idx[idx] * self.steps[idx] + self.ranges[idx][0])\n\n return tuple(corner)\n\n def simplexes_from_cube(self, root):\n cube_corners = self.cube_at_root(root)\n # cube_at_root always has the first right corner root as [0]\n # at [-1] we have the other right corner root\n\n # special case for 1 dimensional configuration space\n if self.dim == 1:\n return (RMSimplex(cube_corners, self.slpsList),)\n else:\n return RMSimplex(cube_corners[:-1], self.slpsList), RMSimplex(cube_corners[1:], self.slpsList)\n\n def process_intersections(self, q, simplex, param_num, value):\n out = []\n inter = simplex.intersections(param_num, value)\n if len(inter) > 0:\n out.append(inter)\n\n q.put(out)\n \n def intersections_parallel(self, param_num, value):\n out = []\n\n # setup a list of processes that we want to run\n q = mp.Queue()\n processes = [mp.Process(target=self.process_intersections, args=(q, simplex, param_num, value)) for simplex in self.simplexes] \n batchSize = BATCH_SIZE\n batches = [processes[i:i+batchSize] for i in range(0, len(processes), batchSize)]\n\n for batch in batches:\n # run processes\n for p in batch:\n p.start() \n for p in batch:\n out += q.get()\n for p in batch:\n p.join() \n\n return tuple(out)\n \n def intersections(self, param_num, value):\n out = []\n\n if(self.writeToDisk == 1):\n # read from disk\n with gzip.open(\"sims.pgz\", 'rb') as simsFile:\n with gzip.open(\"configuration_\" + str(self.allIntersectionsCalls) + \".pgz\", 'ab') as configurationFile:\n while 1:\n try:\n sims = cPickle.load(simsFile)\n for simplex in sims:\n inter = simplex.intersections(param_num, value)\n if len(inter) > 0:\n for configurations in inter:\n for configuration in configurations:\n cPickle.dump(inter, configurationFile, -1)\n\t except EOFError:\n break\n else:\n # read from ram\n for simplexNumber, simplex in enumerate(self.simplexes):\n inter = simplex.intersections(param_num, value)\n if len(inter) > 0:\n out.append([simplexNumber, inter])\n\n return tuple(out)\n\n def process_all_intersections(self, q, d, value):\n q.put(self.intersections(d, value))\n \n def all_intersections_parallel(self, value):\n allIntersections = []\n \n # setup a list of processes that we want to run\n q = mp.Queue()\n processes = [mp.Process(target=self.process_all_intersections, args=(q, d, value)) for d in range(self.slp_dim)] \n batchSize = BATCH_SIZE\n batches = [processes[i:i+batchSize] for i in range(0, len(processes), batchSize)]\n\n for batch in batches:\n # run processes\n for p in batch:\n p.start() \n for p in batch:\n allIntersections.append(q.get())\n for p in batch:\n p.join() \n\n if(len(allIntersections[0]) > 0):\n return tuple(allIntersections)\n else:\n # empty intersection set. need to wiggle\n print >> sys.stderr, \"BADNESS: PARALLELISM NOT USEFUL BECAUSE OF WIGGLE!\"\n [wiggledValue, wiggledKey] = self.wiggle(value)\n return [[[tuple([x/float(self.granularity) for x in wiggledKey])]]]\n \n def all_intersections(self, value):\n # increment call counter\n self.allIntersectionsCalls += 1\n \n \"\"\" Finds intersections with each provided SLP \"\"\"\n allIntersections = [self.intersections(d, value) for d in range(self.slp_dim)]\n\n if(len(allIntersections[0]) > 0):\n return tuple(allIntersections)\n else:\n # empty intersection set. need to wiggle\n [wiggledValue, wiggledKey] = self.wiggle(value)\n return [[[tuple([x/float(self.granularity) for x in wiggledKey])]]]\n \n def distance_to(self, param_num, config, value):\n # i'm going to cheat for now and just pick the closest intersecting corner\n ints = self.intersections(param_num, value)\n\n dists = []\n for simplex in ints:\n for i in simplex:\n dists.append( (misc.distance(i, config)) )\n \n if len(dists) == 0:\n print >> sys.stderr, \"BADNESS: no intersections found for %s\" % repr(value)\n dist = None\n else:\n dist = min(dists)\n\n return dist\n"},{"alpha_fraction":0.7288135886192322,"alphanum_fraction":0.7288135886192322,"avg_line_length":13.75,"blob_id":"70e2bd8c28ffa2d4d8281f4a0f7c39228fb33b5b","content_id":"0a7fa60de76adc20c8a52e56fcce98102fdeb963","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":118,"license_type":"no_license","max_line_length":23,"num_lines":8,"path":"/src/sampling.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# load configuration\nsource configuration.sh\n\n# sample data\necho \"=>sampling data\"\ncd sampling\n./sampling.sh\ncd ../..\n"},{"alpha_fraction":0.6308630108833313,"alphanum_fraction":0.6374296545982361,"avg_line_length":26.50967788696289,"blob_id":"d5e3cd2f68cc6adb325d89933359324154c70220","content_id":"26f32adeb7f1598db01cad8611ba640f9415ecfd","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":4264,"license_type":"no_license","max_line_length":140,"num_lines":155,"path":"/src/sampling/eum/sample_eum_parallel.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import pandas as pd\nimport numpy as np\nimport sys\nimport multiprocessing as mp\n\nfrom negotiation_model import *\nfrom bdm_agent import *\n\n# parallel processing constants\nBATCH_SIZE = 8\n\n# seed this for fixed environment\n# for better replication of results\nrandom.seed(0)\n\n# constants\nINDEPENDENT_VARIABLES = 2\nDEPENDENT_VARIABLES = 2\nRUNS = 10000\nSAMPLES = 10\nTIME_LAPSE = 200\n\n# process single run data\ndef process_run(q, run, qValue, tValue, bookData, salienceValues):\n try:\n line = main_eum(qValue, tValue, bookData, salienceValues)\n except:\n line = \"\"\n \n # add newline for clean output\n line = \"\\n\" + line\n \n # write to queue\n q.put(line)\n\ndef get_salience_values(timeLapse, numberOfAgents):\n salienceValues = np.zeros((timeLapse, numberOfAgents))\n\n for stepNumber in range(timeLapse):\n for agentNumber in range(numberOfAgents):\n salienceValues[stepNumber][agentNumber] = random.random()\n\n return salienceValues\n\n# Defining the model objects\nclass BDMActor(NegotiationActor):\n DecisionClass = BDM_Agent\n\nclass NegotiationModel_(NegotiationModel):\n # Variables for median caching\n median_computed_last = -1\n median = -1\n \n def find_median(self):\n if self.median_computed_last != self.schedule.steps:\n self.median = super().find_median()\n self.median_computed_last = self.schedule.steps\n return self.median\n\nclass ModelOutput:\n def __init__(self, model):\n '''\n Store data from model run.\n '''\n self.agent_vars = model.datacollector.get_agent_vars_dataframe()\n self.model_vars = model.datacollector.get_model_vars_dataframe()\n self.log = model.log\n\ndef load_data():\n # Load data \n bookData = pd.read_csv(\"BDM_ColdWar.csv\")\n bookData.Position = (bookData.Position + 100)/200\n\n return bookData\n\ndef main_eum(qValue, tValue, bookData, salienceValues):\n # define agents\n agents = []\n\n for i, row in bookData.iterrows():\n newAgent = BDMActor(row.Country, row.Capability, row.Position, 1)\n newAgent.decision_model.Q = qValue\n newAgent.decision_model.T = tValue\n newAgent.salience = salienceValues[i]\n agents.append(newAgent)\n\n # instantiate model\n model = NegotiationModel_(agents)\n\n # run model\n for stepNumber in range(TIME_LAPSE):\n agentNumber = 0\n for agent in model.agents:\n #agent.salience = random.random()\n agent.salience = salienceValues[stepNumber][agentNumber]\n agentNumber += 1 \n model.step()\n\n # collect data for next steps\n dependentValues = []\n modelOutput = ModelOutput(model)\n dependentValues.append(list(modelOutput.model_vars[\"Median\"][-SAMPLES-1:]))\n dependentValues.append(list(modelOutput.model_vars[\"Mean\"][-SAMPLES-1:]))\n \n # print line corresponding to this execution\n line = str(qValue) + \" \" + str(tValue)\n for dependentValue in dependentValues:\n line += \" \" + str(dependentValue).replace(\"[\", \"\").replace(\"]\", \"\").replace(\" \", \"\")\n\n return line\n\ndef main():\n # header string\n print(\"i\"*INDEPENDENT_VARIABLES + \"d\"*DEPENDENT_VARIABLES)\n\n # read data\n bookData = load_data()\n\n # get salience values\n [rows, columns] = bookData.shape\n salienceValues = get_salience_values(TIME_LAPSE, rows)\n\n qValues = []\n tValues = []\n \n for run in range(RUNS):\n # sample random ALPs\n qValue = random.random()\n tValue = random.random()\n\n qValues.append(qValue)\n tValues.append(tValue)\n \n # setup a list of processes that we want to run\n q = mp.Queue()\n processes = [mp.Process(target=process_run, args=(q, run, qValues[run], tValues[run], bookData, salienceValues)) for run in range(RUNS)]\n batchSize = BATCH_SIZE\n batches = [processes[i:i+batchSize] for i in range(0, len(processes), batchSize)]\n\n outFile = open(\"output.txt\", \"a\")\n \n for batch in batches:\n # run processes\n for p in batch:\n p.start()\n for p in batch:\n line = q.get()\n outFile.write(line)\n # exit the completed processes\n for p in batch:\n p.join()\n\n outFile.close()\n\nmain()\n"},{"alpha_fraction":0.5772079825401306,"alphanum_fraction":0.5948718190193176,"avg_line_length":21.5,"blob_id":"172f17fb56c1f893cf528f867f348c507707bc76","content_id":"5a0207a4e6641ab024fb9863a512420f9f05a4ee","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1755,"license_type":"no_license","max_line_length":92,"num_lines":78,"path":"/src/sampling/schelling/sample_schelling.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"from model import SchellingModel\nimport random\n\n# only use three of the several ALPs\n# density (0,1)\n# minority_pc (0,1)\n\n# use one model-level SLP\n# happy\n\n# constant ALPs\nHEIGHT = 20\nWIDTH = 20\nHOMOPHILY = 4\n\n# constants\n# max_iters in the example is specified as 1000\n# so we run for 1000 steps\nINDEPENDENT_VARIABLES = 2\nDEPENDENT_VARIABLES = 1\nRUNS = 10000\nSAMPLES = 10\nTIME_LAPSE = 1000\n\ndef main_schelling(density, minorityPc):\n # instantiate and run model\n model = SchellingModel(\n height=HEIGHT,\n width=WIDTH,\n density=density,\n minority_pc=minorityPc,\n homophily=HOMOPHILY)\n\n # run time lapse\n for i in range(TIME_LAPSE):\n try:\n # step\n model.step()\n except:\n # no empty cells\n pass\n \n # collect data for next steps\n dependentValues = []\n\n for i in range(SAMPLES):\n try:\n # step\n model.step()\n except:\n # saturated\n # no empty cells\n pass\n\n # read data\n data = model.datacollector.get_model_vars_dataframe() \n dependentValues.append(list(data.happy)[-SAMPLES-1:])\n\n # print line corresponding to this execution\n line = str(density) + \" \" + str(minorityPc)\n for dependentValue in dependentValues:\n line += \" \" + str(dependentValue).replace(\"[\", \"\").replace(\"]\", \"\").replace(\" \", \"\")\n print(line)\n\n \ndef main():\n # header string\n print(\"i\"*INDEPENDENT_VARIABLES + \"d\"*DEPENDENT_VARIABLES)\n\n for run in range(RUNS):\n # sample random ALPs\n density = random.random() \n minorityPc = random.random() \n \n # run model using those ALPs\n main_schelling(density, minorityPc)\n\nmain()\n"},{"alpha_fraction":0.6653721928596497,"alphanum_fraction":0.6737864017486572,"avg_line_length":29.294116973876953,"blob_id":"106a3afa59283bbc5813e8bc9048073d87b51f60","content_id":"f7c189ef2ce03983c7a3db9a1fa8665624f9f79d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1545,"license_type":"no_license","max_line_length":111,"num_lines":51,"path":"/app/es_server.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"from elasticsearch import Elasticsearch\nfrom image_match.elasticsearch_driver import SignatureES\nimport glob\nfrom flask import Flask, render_template, request\nimport os\nimport time\n\napp = Flask(__name__)\napp.config[\"INPUT_FILE\"] = \"sketches/demonstration.png\"\napp.config[\"ABM\"] = \"flocking\"\napp.config[\"SES\"] = SignatureES(Elasticsearch(), distance_cutoff=1.0)\n\n@app.route('/initialize', methods=[\"GET\"])\ndef initialize():\n # clear existing indexes\n os.system(\"curl -XDELETE 'http://localhost:9200/_all'\")\n\n # sleep to be sure of synchronization\n time.sleep(60)\n \n # index items\n folderName = \"/home/karan/storage/workspaces/bitbucket/_swarm-lfd-data/\" + app.config[\"ABM\"] + \"/images_10\"\n #folderName = \"/home/karan/storage/workspaces/bitbucket/_swarm-lfd-data/\" + app.config[\"ABM\"] + \"/images\"\n allFileNames = glob.glob(folderName + \"/*.*\")\n\n for imageName in sorted(allFileNames):\n print('elastic search initialize: ' + imageName)\n app.config[\"SES\"].add_image(imageName)\n\n # sleep to be sure of synchronization\n time.sleep(60)\n \n return \"initialization complete\"\n\n\n@app.route('/match', methods=[\"GET\"])\ndef match():\n results = app.config[\"SES\"].search_image(app.config[\"INPUT_FILE\"])\n\n distanceResults = []\n for result in results:\n distanceResults.append((result['dist'], result['path']))\n\n # get top result\n [distance, path] = sorted(distanceResults)[0]\n os.system('cp ' + path + ' matches/')\n \n return \"matched\"\n\n# code run at app start\ninitialize()\n"},{"alpha_fraction":0.6937590837478638,"alphanum_fraction":0.698113203048706,"avg_line_length":25.461538314819336,"blob_id":"e035ad7ba763933b646f7bbf7f9c261f52b7789c","content_id":"dc860b5c8baa82f6f21ccb359bcc3ff3eafbc5b8","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":689,"license_type":"no_license","max_line_length":73,"num_lines":26,"path":"/src/lfd.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"#!/usr/bin/env bash\n# load configuration\nsource configuration.sh\n\n# run multiple times to generate multiple sets of images\nexport RUNS_PER_INPUT=1\n\n# remove existing data\nrm -f suggested_alps.txt\nrm -f suggested_alps_fm_prediction_error.txt\n\nfor i in $(seq \"$RUNS_PER_INPUT\")\ndo\n # run experiment\n echo \"=>running experiment\"\n export ALP_CONFIGURATIONS=$(python lfd.py)\n echo $ALP_CONFIGURATIONS >> suggested_alps.txt\n \n # circumvent excessive prints from smt \n mv temp_average_configuration_list.txt suggested_alps.txt 2>/dev/null\n \n if [ $WRITE_TO_DISK == 1 ]; then\n # remove compressed pickle files generated for this run\n rm *.pgz\n fi\ndone\n\n"},{"alpha_fraction":0.6831682920455933,"alphanum_fraction":0.7623762488365173,"avg_line_length":24.25,"blob_id":"a54ecdd6f4685d3e1936514766d0a201aa8a708e","content_id":"a509b2a730edd34ff284c45c582bca0af6df31d2","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":101,"license_type":"no_license","max_line_length":36,"num_lines":4,"path":"/app/run_es_server.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"sudo service elasticsearch start\n\nexport FLASK_APP=es_server.py\nflask run --host=0.0.0.0 --port=5001\n"},{"alpha_fraction":0.5775167942047119,"alphanum_fraction":0.5892617702484131,"avg_line_length":28.799999237060547,"blob_id":"6dd53eb41161c6b2279e68b5d703786da0993126","content_id":"0a54a5e6f680756fcb7a659d7b5332ebb0e23535","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Java","length_bytes":2980,"license_type":"no_license","max_line_length":115,"num_lines":100,"path":"/src/sampling/wolf_sheep/predict_wolf_sheep.java","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"/* package required for calls from parent folder */\npackage sampling.wolf_sheep;\n\n/* libraries included */\nimport org.nlogo.headless.HeadlessWorkspace;\nimport java.io.*;\nimport java.util.*;\n\npublic class predict_wolf_sheep {\n public static void main(String[] argv) {\n HeadlessWorkspace workspace =\n HeadlessWorkspace.newInstance() ;\n try {\n\t// evaluate to list\n\tString alpConfigurationListString = argv[0];\n\tList alpConfigurationList = new ArrayList(Arrays.asList(alpConfigurationListString.split(\"\\\\(\")));\n\n\n\t// skip first entry, which is empty\n\tfor (int i = 1; i < alpConfigurationList.size(); i++) {\n\t\tString alpConfiguration = alpConfigurationList.get(i);\n\n\t\t// clean string\n\t\talpConfiguration = alpConfiguration.replace(\")\", \"\").replace(\"]\", \"\");\n\n\t\t// extract variables\n\t\tArrayList alpValues = new ArrayList(Arrays.asList(alpConfiguration.split(\",\")));\n\n workspace.open(\"sampling/wolf_sheep/Wolf Sheep Predation.nlogo\");\n\n \tworkspace.command(\"set grass? true\");\n workspace.command(\"set initial-number-sheep 100\");\n workspace.command(\"set initial-number-wolves 50\"); \n \n \t// wolf-reproduce\n \tString k1 = alpValues.get(0);\n \t\n \t// sheep-reproduce\n\t\t\t\tString k2 = alpValues.get(1);\n\t\t\t\t\n\t\t\t\t//wolf-gain-from-food\n\t\t\t\tString k3 = alpValues.get(2);\n\t\t\t\t\n\t\t\t\t//sheep-gain-from-food\n\t\t\t\tString k4 = alpValues.get(3);\n\t\t\t\t\n\t\t\t\t//grass-regrowth-time\n\t\t\t\tString k5 = alpValues.get(4);\n\n\n\t\t\t\tworkspace.command(\"set wolf-reproduce \" + k1);\n\t\t\t\tworkspace.command(\"set sheep-reproduce \" + k2);\n\t\t\t\tworkspace.command(\"set wolf-gain-from-food \" + k3);\n\t\t\t\tworkspace.command(\"set sheep-gain-from-food \" + k4);\n\t\t\t\tworkspace.command(\"set grass-regrowth-time \" + k5);\n\n \tworkspace.command(\"setup\");\n \tworkspace.command(\"repeat 500 [ go ]\");\n \t\n \tString wolves = \"\" + workspace.report(\"count wolves\");\n \tString sheep = \"\" + workspace.report(\"count sheep\");\n \t\n \tfor(int j = 0; j < 750; j++) {\n \t\tworkspace.command(\"go\");\n \t\twolves += \",\" + workspace.report(\"count wolves\");\n \t\tsheep += \",\" + workspace.report(\"count sheep\");\n \t}\n \t\n\t\tString[] wolvesValues = wolves.split(\",\");\n\t\tfloat sum = 0;\n\t\tfloat count = 0;\n\t\tfor(String value: wolvesValues){\n\t\t\tsum += Float.parseFloat(value);\n\t\t\tcount += 1;\n\t\t}\n\t\tfloat wolvesAverage = sum/count;\n\n\t\tString[] sheepValues = sheep.split(\",\");\n\t\tsum = 0;\n\t\tcount = 0;\n\t\tfor(String value: sheepValues){\n\t\t\tsum += Float.parseFloat(value);\n\t\t\tcount += 1;\n\t\t}\n\t\tfloat sheepAverage = sum/count;\n\n \tSystem.out.print(\"[(\" + String.valueOf(wolvesAverage) + \", \" + String.valueOf(sheepAverage) + \")]\\n\");\n\n }\n \n \tworkspace.dispose();\n\n }\n catch(Exception ex) {\n ex.printStackTrace();\n }\n }\n \n \n}\n"},{"alpha_fraction":0.6532751321792603,"alphanum_fraction":0.6555458307266235,"avg_line_length":32.44444274902344,"blob_id":"f761ee9cb3b527326523d8f632fdc8728f1d7713","content_id":"b80cd3d6bab49d99e5289b85984b8d0acbde82a9","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":5725,"license_type":"no_license","max_line_length":153,"num_lines":171,"path":"/src/amf/data.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# data.py contains everything that has to do with dealing with the actual\n# data and manipulating data sets.\n\nimport random\nimport copy\nimport sys\nimport misc\nimport os\n\nIND = 5\nDEP = 6\n\ndef make_vartype(vartype_str):\n \"\"\"\n Takes a string that labels columns in a data set.\n vartype_string should either have 'i' or 'd' in it, e.g., \"iidid\"\n\n i - independent variable\n d - dependent variable\n \"\"\"\n # replace characters i and d with IND and DEP symbols.\n return tuple( (IND if c == 'i' else DEP) for c in vartype_str.strip() if not c.isspace())\n\ndef load(file_name):\n \"\"\"\n Loads a data set from a file.\n\n The data set should be space delimited to distinguish columns, and newline\n delimited to distinguish rows.\n\n The first line should contain something that is to be passed into make_vartype.\n\n \"\"\"\n\n file_h = open(file_name)\n\n vartype = make_vartype(file_h.readline().strip())\n\n dataset = []\n # go through each line and parse it\n for line in file_h.xreadlines():\n line = line.strip()\n\n # ignore empty lines\n if len(line) == 0:\n continue\n\n sline = line.split()\n if len(sline) != len(vartype):\n sys.stderr.write(\"Ignored line with wrong number of tokens: \" + line + \"\\n\")\n continue\n\n # split the data into two groups, depending on if they are independent or dependent variables.\n ind_list = []\n dep_list = []\n for idx, item in enumerate(sline):\n (ind_list if vartype[idx] == IND else dep_list).append(float(item) if item != 'NOENTRY' else 'NOENTRY')\n \n dataset.append( ( tuple(ind_list), tuple(dep_list) ) )\n\n dataset.sort()\n\n return tuple(dataset)\n\ndef random_subsets(data_set, n_list, demonstrationDescriptors=[]):\n \"\"\"\n Returns a random subsets of the dataset with sizes specified in the n list.\n There will be no duplicates between subsets.\n \"\"\"\n\n if sum(n_list) > len(data_set):\n raise IndexError, \"Requested size of subsets, from set of size %d, are too large: %s (total size %d)\" % (len (data_set), repr(n_list), sum(n_list))\n\n # generate a list of the indexes, then shuffle them. I will use these to randomly select from the data set.\n idxs = range(len(data_set))\n\n # this is where the randomness lies\n #random.seed(0)\n\n # idxs contain indexes of all elements in dataset\n # we should get the top sum(n_list) values\n # then use those to form random subsets of length specified by n_list\n\n # check number of demonstration descriptors\n demonstrationCount = len(demonstrationDescriptors)\n\n # obtain and shuffle indexes for randomness\n # divides equally per demonstration slp\n # works better when a single slp is specified\n\n # dataset selection used if set to true\n if(eval(os.environ[\"DATASET_SELECTION\"]) == 1):\n useClosestIdx = True\n else:\n useClosestIdx = False\n\n #if(type(depData) == tuple):\n if(useClosestIdx == True):\n # we need to divide the dataset to cover all of these\n # edit last item to account for correctness of sum\n # shuffle for random allotment of extra points\n nListSubLengths = [sum(n_list)/demonstrationCount]\n\n nListSubLengths[-1] += sum(n_list)%demonstrationCount\n random.shuffle(nListSubLengths)\n\n # indexes corresponding to good points\n goodIdxs = []\n \n # for each sublength\n # corresponding to each demonstration\n for index in range(len(nListSubLengths)):\n # if feedback is on\n # compute distances based on queried slps and not demonstration descriptors\n if((eval(os.environ[\"FEEDBACK\"]) == 1) and (os.path.isfile(\"queried_slps.txt\") == True)):\n with open(\"queried_slps.txt\", \"r\") as inFile:\n queriedSlps = [eval(inFile.readlines()[0])]\n distances = [(misc.distance(queriedSlps[index], data_set[idx][1]), idx) for idx in idxs]\n else:\n # get distances of all descriptors in dataset\n # with respect to the input descriptor\n distances = [(misc.distance(demonstrationDescriptors[index], data_set[idx][1]), idx) for idx in idxs]\n\n distances.sort()\n\n # extract first few as needed\n nListSubLength = nListSubLengths[index]\n goodDataPoints = distances[:nListSubLength]\n\n # append to list of all points\n goodIdxs += [idx for distance, idx in goodDataPoints]\n\n random.shuffle(goodIdxs)\n else:\n # point sampling has already been done from the full dataset\n # now we are dealing with individual dep value correspondences for a given ind configuration\n random.shuffle(idxs)\n \n # now form subsets based on sequence of indexes\n subsets = []\n cur_offset = 0\n for subset_size in n_list:\n #if(type(depData) == tuple):\n if(useClosestIdx == True):\n # we are selecting from the full dataset\n # this dataset does not contain individual slp correspondence\n subsets.append( tuple(data_set[idx] for idx in goodIdxs[cur_offset:cur_offset + subset_size]) )\n else:\n # point sampling has already been done from the full dataset\n # now we are dealing with individual dep value correspondences for a given ind configuration\n subsets.append( tuple(data_set[idx] for idx in idxs[cur_offset:cur_offset + subset_size]) )\n\n cur_offset += subset_size\n\n return subsets\n\ndef split_dep(data_set):\n \"\"\"\n Split the data set into one-data-set-per-dependent variable.\n \"\"\"\n\n num_dep = len(data_set[0][1])\n\n sets = [ [] for idx in range(num_dep) ]\n\n for ind, dep in data_set:\n # for each dependent variable here...\n for d_idx in range(num_dep):\n sets[d_idx].append((ind, dep[d_idx]))\n \n return sets\n\n\n\n\n\n\n"},{"alpha_fraction":0.638373613357544,"alphanum_fraction":0.660858154296875,"avg_line_length":37.121429443359375,"blob_id":"6b842ce6a5a8435cd45f99cfb693f0cb89905ba2","content_id":"e48c3eaf9d0e8475ff26fe2768343c05d4fb1ab4","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":5337,"license_type":"no_license","max_line_length":109,"num_lines":140,"path":"/data/dataset_analysis/dataset_analysis.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import matplotlib.pyplot as plt\nimport numpy\nimport sys\n\ndef read_dataset(fileName):\n dataset = []\n \n with open(fileName + \"/slp_dataset.txt\", \"r\") as slpDatasetFile:\n for line in slpDatasetFile:\n point = eval(line)\n dataset.append(point)\n\n return dataset\n\ndef polarize_matrix(matrix, threshold=0):\n # polarize matrix values\n [rows, columns] = matrix.shape\n\n for row in range(rows):\n for column in range(columns):\n if(matrix[row][column] > threshold):\n matrix[row][column] = 1\n else:\n matrix[row][column] = 0\n\ndef analyze_datasets(goodDatasetFolders=[], badDatasetFolders=[], figureSuffix = \"\"):\n # determine comparison method for quality of datasets\n\n # demonstration specific data\n demonstrationID = sys.argv[1]\n \n # get all points in good dataset\n goodDataset = []\n for folder in goodDatasetFolders:\n goodDataset += read_dataset(demonstrationID + \"/\" + \"predictions_\" + folder)\n\n # get all points in bad dataset\n badDataset = []\n for folder in badDatasetFolders:\n badDataset += read_dataset(demonstrationID + \"/\" + \"predictions_\" + folder)\n\n # define plot ranges\n # each value represents a threshold for a bucket\n dim1Step = eval(sys.argv[2])\n dim2Step = eval(sys.argv[3])\n dim1ThresholdValues = numpy.arange(0, 1, dim1Step)\n dim2ThresholdValues = numpy.arange(0, 1, dim2Step)\n\n # now compute matrix based on discrete values\n goodDensityMap = numpy.zeros((len(dim1ThresholdValues), len(dim2ThresholdValues)))\n badDensityMap = numpy.zeros((len(dim1ThresholdValues), len(dim2ThresholdValues)))\n\n # for good datasets\n for point in goodDataset:\n # determine bucket for this point\n [dim1Value, dim2Value] = point\n\n # will be an integer value between 0 and len(dim1ThresholdValues)-1\n # will be an integer value between 0 and len(dim2ThresholdValues)-1\n dim1Bucket = int(dim1Value/dim1Step)\n dim2Bucket = int(dim2Value/dim2Step)\n\n # account for upper limits\n # use capping\n if(dim1Bucket == len(dim1ThresholdValues)):\n dim1Bucket = len(dim1ThresholdValues) - 1\n if(dim2Bucket == len(dim2ThresholdValues)):\n dim2Bucket = len(dim2ThresholdValues) - 1\n\n # add count to matrix\n goodDensityMap[dim1Bucket][dim2Bucket] += 1\n\n # for bad datasets\n for point in badDataset:\n # determine bucket for this point\n [dim1Value, dim2Value] = point\n\n # will be an integer value between 0 and len(dim1ThresholdValues) - 1\n # will be an integer value between 0 and len(dim2ThresholdValues) - 1\n dim1Bucket = int(dim1Value/dim1Step) - 1\n dim2Bucket = int(dim2Value/dim2Step) - 1\n\n # account for upper limits\n # use capping\n if(dim1Bucket == len(dim1ThresholdValues)):\n dim1Bucket = len(dim1ThresholdValues) - 1\n if(dim2Bucket == len(dim2ThresholdValues)):\n dim2Bucket = len(dim2ThresholdValues) - 1\n \n # add count to matrix\n badDensityMap[dim1Bucket][dim2Bucket] += 1\n\n # compute difference\n # and normalize\n goodDifferenceDensityMap = goodDensityMap - badDensityMap\n badDifferenceDensityMap = badDensityMap - goodDensityMap\n\n polarize_matrix(goodDifferenceDensityMap, 0)\n polarize_matrix(badDifferenceDensityMap, 0)\n\n # splice matrices if needed\n # to zoom in on finer sections\n #goodDifferenceDensityMap = goodDifferenceDensityMap[:10, :10]\n #badDifferenceDensityMap = badDifferenceDensityMap[:10, :10]\n #dim1ThresholdValues = dim1ThresholdValues[:10]\n #dim2ThresholdValues = dim2ThresholdValues[:10]\n \n plt.figure()\n plt.imshow(goodDifferenceDensityMap, interpolation='nearest', cmap=plt.cm.gray)\n plt.ylabel(\"Maximum Area\")\n plt.xlabel(\"Maximum Perimeter\")\n #plt.title(\"Good Difference Density Map\")\n plt.yticks(range(len(dim1ThresholdValues)), dim1ThresholdValues)\n plt.xticks(range(len(dim2ThresholdValues)), dim2ThresholdValues)\n plt.savefig(demonstrationID + \"/\" + \"good_difference_density_map\" + figureSuffix + \".png\")\n plt.close(\"all\")\n \n plt.figure()\n plt.imshow(badDifferenceDensityMap, interpolation='nearest', cmap=plt.cm.gray)\n plt.ylabel(\"Maximum Area\")\n plt.xlabel(\"Maximum Perimeter\")\n #plt.title(\"Bad Difference Density Map\")\n plt.yticks(range(len(dim1ThresholdValues)), dim1ThresholdValues)\n plt.xticks(range(len(dim2ThresholdValues)), dim2ThresholdValues)\n plt.savefig(demonstrationID + \"/\" + \"bad_difference_density_map\" + figureSuffix + \".png\")\n plt.close(\"all\")\n \ndef main():\n analyze_datasets([\"best\", \"second_best\", \"third_best\"], [\"worst\", \"second_worst\", \"third_worst\"], \"_all\")\n analyze_datasets([\"best\"], [\"worst\"], \"_1_1\")\n analyze_datasets([\"best\"], [\"second_worst\"], \"_1_2\")\n analyze_datasets([\"best\"], [\"third_worst\"], \"_1_3\")\n analyze_datasets([\"second_best\"], [\"worst\"], \"_2_1\")\n analyze_datasets([\"second_best\"], [\"second_worst\"], \"_2_2\")\n analyze_datasets([\"second_best\"], [\"third_worst\"], \"_2_3\")\n analyze_datasets([\"third_best\"], [\"worst\"], \"_3_1\")\n analyze_datasets([\"third_best\"], [\"second_worst\"], \"_3_2\")\n analyze_datasets([\"third_best\"], [\"third_worst\"], \"_3_3\")\n \nmain()\n"},{"alpha_fraction":0.6425849199295044,"alphanum_fraction":0.6641530990600586,"avg_line_length":32.789039611816406,"blob_id":"9bc73487b95b665c1d75851548f8c07f7c7dad62","content_id":"57cbd55a319ed9d96c686d00e8557615b2489af6","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":12333,"license_type":"no_license","max_line_length":127,"num_lines":365,"path":"/src/sampling/utils/featurize_image.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# libraries used\nimport cv2\nimport numpy as np\nfrom skimage.feature import local_binary_pattern\nfrom skimage.io import imread\nfrom skimage.transform import resize\nfrom keras import backend as K\nfrom keras.models import model_from_json\nfrom keras.applications.vgg16 import VGG16\nfrom keras.applications.vgg16 import preprocess_input\nfrom keras.applications.resnet50 import ResNet50\nfrom keras.preprocessing import image\nfrom keras.preprocessing.image import load_img, img_to_array\n\nfrom keras.models import Sequential\nfrom keras.layers.convolutional import Convolution3D\nfrom keras.layers.convolutional_recurrent import ConvLSTM2D\nfrom keras.layers.normalization import BatchNormalization\nfrom keras.layers import Dense, Activation\nfrom keras.layers.core import Reshape, Flatten\n\nimport os\n\n# using magic numbers for now\n\n# feature extraction initialization\ndef get_network(descriptorSetting, pathPrefix=None):\n # load feature extractor if needed\n net = None\n\n # prefix to folder location\n if(pathPrefix == None):\n pathPrefix = \"\"\n\n functionName = \"get_\" + descriptorSetting.lower() + \"_network\"\n if(functionName in globals().keys()):\n net = globals()[functionName](pathPrefix)\n\n return net\n\n# feature extraction\ndef get_features(net, imageName, descriptorSetting):\n # initialization\n features = None\n\n functionName = \"get_\" + descriptorSetting.lower() + \"_features\"\n features = globals()[functionName](net, imageName)\n \n # return features\n return features\n\n# saved keras network\ndef get_saved_keras_network(modelFile, weightFile):\n # load and compile saved network\n jsonFile = open(modelFile, 'r')\n modelJson = jsonFile.read()\n jsonFile.close()\n net = model_from_json(modelJson)\n net.load_weights(weightFile)\n net.compile(loss='categorical_crossentropy', optimizer='adadelta', metrics=['accuracy'])\n\n # return network\n return net\n\n# contour information\ndef get_contour_features(net, imageName):\n imageData = cv2.imread(imageName, cv2.IMREAD_GRAYSCALE)\n\n # dilate for better contour extraction\n kernel = np.ones((5,5),np.uint8)\n dilatedImageData = cv2.dilate(imageData, kernel, iterations=4)\n #cv2.imwrite(imageName[:-4] + \"_dilated.png\", dilatedImageData)\n\n # extract contours\n ret, threshold = cv2.threshold(dilatedImageData, 127, 255, 0)\n contours, hierarchy = cv2.findContours(threshold, cv2.RETR_TREE, cv2.CHAIN_APPROX_NONE)\n contouredImageData = np.zeros(dilatedImageData.shape)\n for i in xrange(len(contours)):\n cv2.drawContours(contouredImageData, contours, i, (255, 255, 255), 1, 8, hierarchy) \n #cv2.imwrite(imageName[:-4] + \"_contoured.png\", contouredImageData)\n\n # extract largest contour area and perimeter\n areas = []\n perimeters = []\n\n for contour in contours:\n area = cv2.contourArea(contour)\n perimeter = cv2.arcLength(contour, True)\n areas.append(area)\n perimeters.append(perimeter)\n\n # define dependent values\n maxArea = max(areas)\n maxPerimeter = max(perimeters)\n features = [maxArea, maxPerimeter]\n return features\n\nimport sys\n\n# density information\ndef get_density_features(net, imageName, blocks=eval(os.environ[\"NUM_DEPENDENT\"])):\n # read image\n image = imread(imageName, as_grey=True)\n \n # resize to a pixels power of 2\n # required for equal division when splitting\n resizeDimension = 2**int(min([np.log2(x) for x in image.shape]))\n image = resize(image, (resizeDimension, resizeDimension), mode=\"wrap\")\n \n # polarize pixels\n threshold = np.amax(image)\n image[image < threshold] = 0\n image[image >= threshold] = 1\n\n # segment image\n # segments must be a power of 4\n # because each block is split into 4 sub blocks\n currentBlocksCount = 1\n blocksList = np.array([image])\n \n while(currentBlocksCount < blocks):\n # split horizontally\n subBlocksList = []\n for block in blocksList:\n subBlocks = np.hsplit(block, 2)\n subBlocksList.extend(subBlocks)\n blocksList = subBlocksList\n\n # split vertically\n subBlocksList = []\n for block in blocksList:\n subBlocks = np.vsplit(block, 2)\n subBlocksList.extend(subBlocks)\n blocksList = subBlocksList\n\n # update number of blocks\n currentBlocksCount *= 4\n \n # compute density\n # since all images are the same size, we can use count instead\n features = [x.sum() for x in blocksList]\n \n return features\n \n# vgg network\ndef get_vgg_network(pathPrefix):\n # pretrained vgg16 network\n net = VGG16(include_top=True, weights='imagenet', input_tensor=None)\n return net\n \n# vgg based features\ndef get_vgg_features(net, imageName):\n # preprocess image\n imageData = image.load_img(imageName, target_size=(224, 224))\n imageData = image.img_to_array(imageData)\n imageData = np.expand_dims(imageData, axis=0)\n imageData = preprocess_input(imageData)\n\n # final layer output\n features = net.predict(imageData)[0]\n return features\n\n# lbp features\ndef get_lbp_features(net, imageName):\n # settings for LBP\n METHOD = 'uniform'\n radius = 3\n nPoints = 8 * radius\n nBins = 128\n \n # dilate for better contour extraction\n imageData = cv2.imread(imageName, cv2.IMREAD_GRAYSCALE)\n kernel = np.ones((5,5),np.uint8)\n dilatedImageData = cv2.dilate(imageData, kernel, iterations=4)\n #cv2.imwrite(imageName[:-4] + \"_dilated.png\", dilatedImageData)\n\n # compute normalized histogram distribution\n lbp = local_binary_pattern(dilatedImageData, nPoints, radius, METHOD)\n hist, binEdges = np.histogram(lbp, bins=nBins, range=(0, nBins), density=True)\n\n # define dependent values\n features = hist\n return features\n \n# mnist network\ndef get_mnist_network(pathPrefix):\n # path to saved network\n modelFile = pathPrefix + \"keras/mnist_cnn_model.json\"\n weightFile = pathPrefix + \"keras/mnist_cnn_model.h5\"\n \n # return network\n net = get_saved_keras_network(modelFile, weightFile)\n return net\n\n# mnist based convolutional features\ndef get_mnist_features(net, imageName):\n # output layer specification\n LAYER_NUMBER = 7\n LEARNING_PHASE = 0\n\n # dilate for better contour extraction\n imageData = cv2.imread(imageName, cv2.IMREAD_GRAYSCALE)\n kernel = np.ones((5,5),np.uint8)\n dilatedImageData = cv2.dilate(imageData, kernel, iterations=4)\n #cv2.imwrite(imageName[:-4] + \"_dilated.png\", dilatedImageData)\n\n imageData = resize(dilatedImageData, (28, 28))\n imageData = imageData.reshape(1, 1, 28, 28)\n\n # layer output\n get_layer_output = K.function([net.layers[0].input, K.learning_phase()],\n [net.layers[LAYER_NUMBER].output])\n layerOutput = get_layer_output([imageData, LEARNING_PHASE])[0]\n features = layerOutput[0]\n return features\n\n# resnet network\ndef get_resnet50_network(pathPrefix):\n # pretrained resnet50 network\n net = ResNet50(include_top=True, weights='imagenet', input_tensor=None)\n return net\n\n# resnet based features\ndef get_resnet50_features(net, imageName):\n # preprocess image\n dimensionOrdering = K.image_dim_ordering()\n mean = (103.939, 116.779, 123.68)\n imageData = load_img(imageName, target_size=(224, 224))\n imageData = img_to_array(imageData, dim_ordering=dimensionOrdering)\n\n if dimensionOrdering == 'th':\n imageData[0, :, :] -= mean[0]\n imageData[1, :, :] -= mean[1]\n imageData[2, :, :] -= mean[2]\n # 'RGB'->'BGR'\n imageData = imageData[::-1, :, :]\n else:\n imageData[:, :, 0] -= mean[0]\n imageData[:, :, 1] -= mean[1]\n imageData[:, :, 2] -= mean[2]\n imageData = imageData[:, :, ::-1]\n\n imageData = np.expand_dims(imageData, axis=0)\n\n # final layer output\n features = net.predict(imageData)[0]\n return features\n\n# autoencoder network\ndef get_vae_network(pathPrefix):\n # path to saved network\n modelFile = pathPrefix + \"keras/variational_autoencoder_model.json\"\n weightFile = pathPrefix + \"keras/variational_autoencoder_model.h5\"\n\n # return network\n net = get_saved_keras_network(modelFile, weightFile)\n return net\n\n# autoencoder based features\ndef get_vae_features(net, imageName):\n # output layer specification\n LAYER_NUMBER = 3\n LEARNING_PHASE = 0\n X_TRAIN_SHAPE = tuple([1,784])\n\n # preprocess image\n imageData = imread(imageName, as_grey=True)\n imageData = resize(imageData, (28, 28))\n imageData = imageData.flatten()\n imageData = imageData.astype('float32') / 255.\n imageData = imageData.reshape((1, np.prod(X_TRAIN_SHAPE[1:])))\n\n # layer output\n get_layer_output = K.function([net.layers[0].input, K.learning_phase()],\n [net.layers[LAYER_NUMBER].output])\n layer_output = get_layer_output([imageData, LEARNING_PHASE])[0]\n features = layer_output[0]\n return features\n\n# conv lstm network\ndef get_lstm_network(pathPrefix):\n # path to saved network\n modelFile = pathPrefix + \"keras/conv_lstm_model.json\"\n weightFile = pathPrefix + \"keras/conv_lstm_model.h5\"\n\n #net = get_saved_keras_network_lstm(modelFile, weightFile)\n\n # manually specify network\n # need to do this because convLTSM layer gives error\n net = Sequential()\n net.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\n input_shape=(None, 40, 40, 1),\n batch_input_shape=(None, 11, 40, 40, 1),\n border_mode='same', return_sequences=True))\n net.add(BatchNormalization())\n net.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\n border_mode='same', return_sequences=True))\n net.add(BatchNormalization())\n net.add(Flatten())\n net.add(Dense(128))\n net.add(Dense(1600*11))\n net.add(Reshape((11, 40, 40, 1)))\n net.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\n border_mode='same', return_sequences=True))\n net.add(BatchNormalization())\n net.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\n border_mode='same', return_sequences=True))\n net.add(BatchNormalization())\n net.add(Convolution3D(nb_filter=1, kernel_dim1=1, kernel_dim2=3,\n kernel_dim3=3, activation='sigmoid',\n border_mode='same', dim_ordering='tf'))\n net.load_weights(weightFile)\n net.compile(loss='categorical_crossentropy', optimizer='adadelta', metrics=['accuracy'])\n\n # return network\n return net\n\n# convolutional lstm based features\ndef get_lstm_features(net, imageName):\n # output layer specification\n LAYER_NUMBER = 5\n LEARNING_PHASE = 0\n\n # generate movie\n rows = 40\n columns = 40\n numberOfSamples = 1\n numberOfFrames = 11\n shiftedMovie = np.zeros((numberOfSamples, numberOfFrames, rows, columns, 1), dtype=np.float)\n imageNameSplit = imageName.split(\"_\")\n\n # stitch all but last\n imageSeriesName = \"\"\n for index in range(len(imageNameSplit)-1):\n splitItem = imageNameSplit[index]\n imageSeriesName += splitItem + \"_\"\n imageSeriesName = imageSeriesName[:-1]\n\n # image series name should be the entire thing up to underscore \n for i in range(numberOfSamples):\n # read custom dataset\n for t in range(numberOfFrames):\n currentFrameName = imageSeriesName + \"_\" + str(200 + t) + \".png\"\n \n # read and resize to 40x40\n imageData = imread(currentFrameName, as_grey=True)\n imageData = resize(imageData, (40, 40)) \n shiftedMovie[i, t, :, :, 0] = imageData\n\n # layer output\n get_layer_output = K.function([net.layers[0].input, K.learning_phase()],\n [net.layers[LAYER_NUMBER].output])\n layer_output = get_layer_output([shiftedMovie, LEARNING_PHASE])[0]\n features = layer_output[0]\n return features\n\n# image search features\ndef get_image_match_features(net, imageName):\n # use a system call to invoke the python3 library using this image name\n os.system(\"python3 sampling/utils/get_image_match_features.py \" + imageName)\n\n # use temporary file for communication\n features = np.load(\"image_match_features.npy\")\n\n return features\n"},{"alpha_fraction":0.7101449370384216,"alphanum_fraction":0.7753623127937317,"avg_line_length":22,"blob_id":"096d119115dfd8c1aeb7a1118d0b6beb8c5647c3","content_id":"bdc936d5c28554fd29fc750fdb90f22ede3ee6e3","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":138,"license_type":"no_license","max_line_length":54,"num_lines":6,"path":"/src/domainconfs/flocking.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"TRAINING_SIZE = 330\nVALIDATION_SIZE = 100\nRM_GRANULARITY = 40\nREGRESSION = \"sklearn.neighbors.KNeighborsRegressor()\"\n\nNUM_INDEPENDENT = 3\n"},{"alpha_fraction":0.6039416193962097,"alphanum_fraction":0.6153117418289185,"avg_line_length":30.02941131591797,"blob_id":"1f1e7cbdbe4e657662a152df12145edfc3c467b7","content_id":"e875ace109e35918df732c93332cbba1d4dcda2f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":5277,"license_type":"no_license","max_line_length":137,"num_lines":170,"path":"/src/configuration.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"#!/usr/bin/env bash\n#######################################################################\n# this is the only thing that should be edited when using this script #\n#######################################################################\n\n# abm being used\nexport ABM=\"flocking\"\n#export ABM=\"civil_violence\"\n#export ABM=\"forest_fire\"\n#export ABM=\"aids\"\n#export ABM=\"wolf_sheep\"\n#export ABM=\"eum\"\n#export ABM=\"schelling\"\n#export ABM=\"turbulence\"\n\n# use of disk to reduce ram consumption\nexport WRITE_TO_DISK=0\n\n# use of demonstration images is off by default \nexport IMAGES=0\n\n# check if this abm depends on images\nif [ $ABM == \"flocking\" ] || [ $ABM == \"turbulence\" ]; then\n export IMAGES=1\nfi\n\n# if feature reduction has been used\nexport REDUCED=0\n\n# if image features are uniform across all stages of the framework\n# used during filtering\nexport IMAGE_FEATURIZATION_HOMOGENEOUS=0\n\n# descriptor setting if any\nexport DESCRIPTOR_SETTING=\"\"\n\n# check if this abm depends on images\n# specify which descriptor to use\nif [ $IMAGES == 1 ]; then\n # determine number of dependent variables\n #export DESCRIPTOR_SETTING=\"DENSITY\"\n export DESCRIPTOR_SETTING=\"CONTOUR\"\n #export DESCRIPTOR_SETTING=\"LBP\"\n #export DESCRIPTOR_SETTING=\"MNIST\"\n #export DESCRIPTOR_SETTING=\"VAE\"\n #export DESCRIPTOR_SETTING=\"LSTM\"\n #export DESCRIPTOR_SETTING=\"VGG\"\n #export DESCRIPTOR_SETTING=\"RESNET50\"\n #export DESCRIPTOR_SETTING=\"IMAGE_MATCH\"\nfi\n\n# check for override in case of filtering\nif [ ! -z $USE_FILTER_DESCRIPTOR_SETTING ]; then\n if [ $USE_FILTER_DESCRIPTOR_SETTING == 1 ]; then\n #export DESCRIPTOR_SETTING=\"RESNET50\"\n\t#export DESCRIPTOR_SETTING=\"IMAGE_MATCH\"\n\texport DESCRIPTOR_SETTING=\"CONTOUR\"\n fi\nfi\n\n# distance metric used for filtering\n# euclidean assumed by default\nexport DISTANCE_METHOD=\"EUCLIDEAN\"\n#export DISTANCE_METHOD=\"NORMALIZED\"\n\n############\n# feedback #\n############\n\nexport FEEDBACK=0\nexport LEARNING_RATE_FM_DATASET=0.0\nexport LEARNING_RATE_REGRESSION_PARAMETER=1.0\nexport REGRESSION_PARAMETER_LOWER_LIMIT=1\nexport REGRESSION_PARAMETER_UPPER_LIMIT=100\n\n#####################\n# dataset selection #\n#####################\n\nexport DATASET_SELECTION=1\n\n#########################\n# refine configurations #\n#########################\n\nexport CONFIGURATIONS_PRUNING=1\nexport CONFIGURATIONS_OUTLIER_DETECTION=1\n\n#############################\n# constants for this script #\n#############################\n\n# demonstration file\nexport DEMONSTRATION_FOLDER=\"../data/demonstrations/\"$ABM\n\n# configuration file\nexport CONFIGURATION_FILE=\"domainconfs/\"$ABM\".py\"\n\n# data file\n# if image processing was used, append descriptor setting to processed filename\nif [ $IMAGES == 1 ]; then\n export DESCRIPTOR_SETTING_LOWER_CASE=$(echo $DESCRIPTOR_SETTING | tr \"[:upper:]\" \"[:lower:]\")\n export DATA_FILE=\"../data/domaindata/processed_\"$ABM\"_\"$DESCRIPTOR_SETTING_LOWER_CASE\n\n if [ $REDUCED == 1 ]; then\n\t export DATA_FILE=$DATA_FILE\"_reduced\"\n fi\n \n export DATA_FILE=$DATA_FILE\".txt\"\nelse\n export DATA_FILE=\"../data/domaindata/processed_\"$ABM\".txt\"\nfi\n\n# scaling information\nexport SCALE_DATA_FILE=$(cat \"map.py\" | grep SCALE_DATA_FILE | head -n 1 | sed s/\\\"/\\\\n/g | sed -n 2p)\n\n##########################################\n# abm simulation tool and file extension #\n# one entry required per abm #\n##########################################\n\nif [ $ABM == \"civil_violence\" ] || [ $ABM == \"forest_fire\" ] || [ $ABM == \"eum\" ] || [ $ABM == \"schelling\" ]; then\n export SIMULATION_TOOL=\"python3\"\n export SIMULATION_EXTENSION=\".py\"\nelif [ $ABM == \"flocking\" ] || [ $ABM == \"aids\" ] || [ $ABM == \"wolf_sheep\" ] || [ $ABM == \"turbulence\" ]; then\n export SIMULATION_TOOL=\"java\"\n export SIMULATION_EXTENSION=\"\"\nfi\n\n############################################################\n# image featurization details, if any #\n# entry required for abms using images #\n# one entry required per abm for nuber of dependent values #\n############################################################\n\n# default number of dependent values\nexport NUM_DEPENDENT=0\n\n# check if this abm depends on images\nif [ $IMAGES == 1 ]; then\n # determine number of dependent variables\n # based on image feature descriptors\n \n if [ \"$DESCRIPTOR_SETTING\" == \"DENSITY\" ]; then\n\t export NUM_DEPENDENT=1\n elif [ \"$DESCRIPTOR_SETTING\" == \"CONTOUR\" ]; then\n\t export NUM_DEPENDENT=2\n elif [ \"$DESCRIPTOR_SETTING\" == \"LBP\" ]; then\n\t export NUM_DEPENDENT=128\n elif [ \"$DESCRIPTOR_SETTING\" == \"MNIST\" ]; then\n\t export NUM_DEPENDENT=128\n elif [ \"$DESCRIPTOR_SETTING\" == \"VAE\" ]; then\n\t export NUM_DEPENDENT=128\n elif [ \"$DESCRIPTOR_SETTING\" == \"LSTM\" ]; then\n\t export NUM_DEPENDENT=128\n elif [ \"$DESCRIPTOR_SETTING\" == \"VGG\" ]; then\n\t export NUM_DEPENDENT=1000\n elif [ \"$DESCRIPTOR_SETTING\" == \"RESNET50\" ]; then\n\t export NUM_DEPENDENT=1000\n elif [ \"$DESCRIPTOR_SETTING\" == \"IMAGE_MATCH\" ]; then\n\t export NUM_DEPENDENT=648\n fi\n\n if [ $REDUCED == 1 ]; then\n\t export NUM_DEPENDENT=50\n fi\nelse\n # read from domainconfs\n export NUM_DEPENDENT=$(python -c \"execd = {} ; exec(open(\\\"\"$CONFIGURATION_FILE\"\\\").read(), execd) ; print execd[\\\"NUM_DEPENDENT\\\"]\")\nfi\n\n\n"},{"alpha_fraction":0.8611111044883728,"alphanum_fraction":0.8611111044883728,"avg_line_length":23,"blob_id":"91a90a5b5253a3c94bfff9c6a258b67f658d2be8","content_id":"2c486c490670b67cd6854c51490477641d70b6fe","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":72,"license_type":"no_license","max_line_length":29,"num_lines":3,"path":"/src/amf/__init__.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import data\nfrom fm import ForwardMapping\nfrom rm import ReverseMapping\n"},{"alpha_fraction":0.7781955003738403,"alphanum_fraction":0.7819548845291138,"avg_line_length":19.461538314819336,"blob_id":"e5745ea83081fe921aafb9494a65c6615c46565d","content_id":"e9f129c08b7512e45d606f8ccfaf1ecdf4784d4e","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":266,"license_type":"no_license","max_line_length":47,"num_lines":13,"path":"/src/sampling/utils/get_image_match_features.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"from image_match.goldberg import ImageSignature\nimport sys\nimport numpy as np\n\n# input file\nimageName = sys.argv[1]\n\n# generate signature\ngis = ImageSignature()\nfeatures = gis.generate_signature(imageName)\n\n# save signature\nnp.save(\"image_match_features\", features)\n"},{"alpha_fraction":0.612515926361084,"alphanum_fraction":0.6263416409492493,"avg_line_length":27.48186492919922,"blob_id":"8b3e6e1918ce334b2a4a0e87e0cd1a80681be8e5","content_id":"cce442fd6bc75d265c10ec508ed72c6399a6dfba","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":5497,"license_type":"no_license","max_line_length":83,"num_lines":193,"path":"/data/dataset_analysis/plot_alps_slps.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# snippet from lfd.py\nfrom mpl_toolkits.mplot3d import Axes3D\nimport matplotlib.pyplot as plt\n\nwith open(\"1/predictions_best/alp_dataset_1488308839.txt\", \"r\") as alpDatasetFile:\n alpDataset = alpDatasetFile.readlines()\n alpPoints = [eval(x) for x in alpDataset]\n\n m = \"o\"\n c = \"r\"\n \n # plot ALPs\n fig = plt.figure()\n ax = fig.add_subplot(111, projection='3d')\n plotName = \"alp_points_best\" + \".png\"\n alpPoints = zip(*alpPoints)\n [xs, ys, zs] = alpPoints\n ax.scatter(xs, ys, zs, c=c, marker=m)\n ax.set_xlabel(\"Max-align-turn\")\n ax.set_ylabel(\"Max-cohere-turn\")\n ax.set_zlabel(\"Max-separate-turn\")\n plt.savefig(plotName)\n plt.close(fig)\n\n fig = plt.figure()\n plotName = \"alp_points_dim1_best\" + \".png\"\n plt.hist(xs, color=c)\n plt.xlabel(\"Max-align-turn\")\n plt.ylabel(\"Number of points in each value range\")\n plt.savefig(plotName)\n plt.close(fig)\n\n fig = plt.figure()\n plotName = \"alp_points_dim2_best\" + \".png\"\n plt.hist(ys, color=c)\n plt.xlabel(\"Max-cohere-turn\")\n plt.ylabel(\"Number of points in each value range\")\n plt.savefig(plotName)\n plt.close(fig)\n\n fig = plt.figure()\n plotName = \"alp_points_dim3_best\" + \".png\"\n plt.hist(zs, color=c)\n plt.xlabel(\"Max-separate-turn\")\n plt.ylabel(\"Number of points in each value range\")\n plt.savefig(plotName)\n plt.close(fig)\n\nwith open(\"1/predictions_worst/alp_dataset_1488319453.txt\", \"r\") as alpDatasetFile:\n alpDataset = alpDatasetFile.readlines()\n alpPoints = [eval(x) for x in alpDataset]\n\n m = \"o\"\n c = \"r\"\n \n # plot ALPs\n fig = plt.figure()\n ax = fig.add_subplot(111, projection='3d')\n plotName = \"alp_points_worst\" + \".png\"\n alpPoints = zip(*alpPoints)\n [xs, ys, zs] = alpPoints\n ax.scatter(xs, ys, zs, c=c, marker=m)\n ax.set_xlabel(\"Max-align-turn\")\n ax.set_ylabel(\"Max-cohere-turn\")\n ax.set_zlabel(\"Max-separate-turn\")\n plt.savefig(plotName)\n plt.close(fig)\n\n fig = plt.figure()\n plotName = \"alp_points_dim1_worst\" + \".png\"\n plt.hist(xs, color=c)\n plt.xlabel(\"Max-align-turn\")\n plt.ylabel(\"Number of points in each value range\")\n plt.savefig(plotName)\n plt.close(fig)\n\n fig = plt.figure()\n plotName = \"alp_points_dim2_worst\" + \".png\"\n plt.hist(ys, color=c)\n plt.xlabel(\"Max-cohere-turn\")\n plt.ylabel(\"Number of points in each value range\")\n plt.savefig(plotName)\n plt.close(fig)\n\n fig = plt.figure()\n plotName = \"alp_points_dim3_worst\" + \".png\"\n plt.hist(zs, color=c)\n plt.xlabel(\"Max-separate-turn\")\n plt.ylabel(\"Number of points in each value range\")\n plt.savefig(plotName)\n plt.close(fig)\n\nwith open(\"1/predictions_best/slp_dataset_1488308839.txt\", \"r\") as slpDatasetFile:\n slpDataset = slpDatasetFile.readlines()\n slpPoints = [eval(x) for x in slpDataset]\n\n m = \"o\"\n c = \"r\"\n \n # plot SLPs\n fig = plt.figure()\n plotName = \"slp_points_best\" + \".png\"\n slpPoints = zip(*slpPoints)\n [xs, ys] = slpPoints\n plt.plot(xs, ys, c+m)\n plt.xlabel(\"Maximum Area\")\n plt.ylabel(\"Maximum Perimeter\")\n plt.savefig(plotName)\n plt.close(fig)\n\n fig = plt.figure()\n plotName = \"slp_points_dim1_best\" + \".png\"\n plt.hist(xs, color=c)\n plt.xlabel(\"Maximum Area\")\n plt.ylabel(\"Number of points in each value range\")\n plt.savefig(plotName)\n plt.close(fig)\n \n fig = plt.figure()\n plotName = \"slp_points_dim2_best\" + \".png\"\n plt.hist(ys, color=c)\n plt.xlabel(\"Maximum Perimeter\")\n plt.ylabel(\"Number of points in each value range\")\n plt.savefig(plotName)\n plt.close(fig)\n\nwith open(\"1/predictions_worst/slp_dataset_1488319453.txt\", \"r\") as slpDatasetFile:\n slpDataset = slpDatasetFile.readlines()\n slpPoints = [eval(x) for x in slpDataset]\n\n m = \"o\"\n c = \"r\"\n \n # plot SLPs\n fig = plt.figure()\n plotName = \"slp_points_worst\" + \".png\"\n slpPoints = zip(*slpPoints)\n [xs, ys] = slpPoints\n plt.plot(xs, ys, c+m)\n plt.xlabel(\"Maximum Area\")\n plt.ylabel(\"Maximum Perimeter\")\n plt.savefig(plotName)\n plt.close(fig)\n\n fig = plt.figure()\n plotName = \"slp_points_dim1_worst\" + \".png\"\n plt.hist(xs, color=c)\n plt.xlabel(\"Maximum Area\")\n plt.ylabel(\"Number of points in each value range\")\n plt.savefig(plotName)\n plt.close(fig)\n \n fig = plt.figure()\n plotName = \"slp_points_dim2_worst\" + \".png\"\n plt.hist(ys, color=c)\n plt.xlabel(\"Maximum Perimeter\")\n plt.ylabel(\"Number of points in each value range\")\n plt.savefig(plotName)\n plt.close(fig)\n\nwith open(\"slp_dataset_1488996573.txt\", \"r\") as slpDatasetFile:\n slpDataset = slpDatasetFile.readlines()\n slpPoints = [eval(x) for x in slpDataset]\n\n m = \"o\"\n c = \"r\"\n \n # plot SLPs\n fig = plt.figure()\n plotName = \"slp_points_guided\" + \".png\"\n slpPoints = zip(*slpPoints)\n [xs, ys] = slpPoints\n plt.plot(xs, ys, c+m)\n plt.xlabel(\"Maximum Area\")\n plt.ylabel(\"Maximum Perimeter\")\n plt.savefig(plotName)\n plt.close(fig)\n\n fig = plt.figure()\n plotName = \"slp_points_dim1_guided\" + \".png\"\n plt.hist(xs, color=c)\n plt.xlabel(\"Maximum Area\")\n plt.ylabel(\"Number of points in each value range\")\n plt.savefig(plotName)\n plt.close(fig)\n \n fig = plt.figure()\n plotName = \"slp_points_dim2_guided\" + \".png\"\n plt.hist(ys, color=c)\n plt.xlabel(\"Maximum Perimeter\")\n plt.ylabel(\"Number of points in each value range\")\n plt.savefig(plotName)\n plt.close(fig)\n"},{"alpha_fraction":0.7403846383094788,"alphanum_fraction":0.7403846383094788,"avg_line_length":33.66666793823242,"blob_id":"9e9e52a4dc407c0845d04708b7e1dbeb1c6a2e3c","content_id":"486c468e6c5a58f1286525eeb12b49523ad3a940","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":104,"license_type":"no_license","max_line_length":65,"num_lines":3,"path":"/src/sampling/sampling_raw.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# sample data\necho \"==>sampling data\"\n$SIMULATION_TOOL $ABM/sample_$ABM$SIMULATION_EXTENSION > $ABM.txt\n"},{"alpha_fraction":0.587643027305603,"alphanum_fraction":0.612356960773468,"avg_line_length":25.325302124023438,"blob_id":"d859a85c93f92e7398f42515ddbe06ccd838d28b","content_id":"75b4b605c3ff80e7cd04c5b542a8378370bd666b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2185,"license_type":"no_license","max_line_length":99,"num_lines":83,"path":"/src/amf/misc.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# misc.py has a bunch of random functions that are used now and then.\n# This is intended as a set of \"helper\" tools, and probably do not need to be used by the user\n# hence, why it is not imported in __init__.py\n\nimport time\nimport math\nimport numpy\nimport sys\n\ndef distance(v1, v2):\n if len(v1) != len(v2):\n raise ValueError, \"distance: vectors %s and %s are different lengths.\" % (repr(v1), repr(v2))\n\n runsum = 0.0\n for idx in xrange(len(v1)):\n runsum += (v2[idx] - v1[idx]) ** 2\n\n return runsum / len(v1)\n\n #return sum( (x2 - x1)**2 for x1, x2 in zip(v1, v2) ) / float(len(v1))\n\n# subtract two lists from one another\ndef list_sub(v1, v2):\n return tuple( x1 - x2 for x1, x2 in zip(v1, v2) )\n\ndef list_add(v1, v2):\n return tuple( x1 + x2 for x1, x2 in zip(v1, v2) )\n\ndef col_average(lists):\n return [ col / len(lists) for col in col_sum(lists) ]\n\ndef average(l):\n return sum(l) / float(len(l))\n\n#def stddev(lists):\n# return tuple(numpy.std(l) for l in lists)\n\ndef col_sum(lists):\n return col_reduce(lambda x, y: x + y, lists)\n\ndef col_max(lists):\n return col_reduce(max, lists)\n\ndef col_min(lists):\n return col_reduce(min, lists)\n\ndef get_col(lists, col):\n \"\"\" Retrieve a single column from a data set as a list. \"\"\"\n\n return tuple( row[col] for row in lists )\n\ndef col_reduce(function, matrix):\n \"\"\" performs a 'reduce' operation on each column of a matrix \"\"\"\n\n cur = list(matrix[0])\n\n for row in matrix[1:]:\n for idx, item in enumerate(row):\n cur[idx] = function(cur[idx], item)\n\n return tuple(cur)\n\ndef xfrange(start, stop, step):\n if start > stop and step > 0.0 or start < stop and step < 0.0 or step == 0.0:\n sys.stderr.write(\"Bad start/stop/step! %f %f %f \\n\" % (start, stop, step))\n\n else:\n cur = 0.0\n while start + cur * step < stop:\n yield start + cur * step\n cur += 1.0\n \ndef binary(n):\n \"\"\" Generate all binary numbers of length n \"\"\"\n out = []\n _binary_helper(n, (), out)\n return tuple(out)\n\ndef _binary_helper(n, l, output):\n if n == 0:\n output.append(l)\n else:\n _binary_helper(n-1, l + (0,), output), _binary_helper(n-1, l + (1,), output)\n"},{"alpha_fraction":0.8258426785469055,"alphanum_fraction":0.8314606547355652,"avg_line_length":24.428571701049805,"blob_id":"ade6db38966377eae78579263b22144fdbdb265e","content_id":"509a662bbe5947b5f0167425788a2dd614a6fd88","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":178,"license_type":"no_license","max_line_length":41,"num_lines":7,"path":"/src/filter.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# load configuration\nexport USE_FILTER_DESCRIPTOR_SETTING=1\nsource configuration.sh\n\n# filter configurations per demonstration\necho \"=>filtering configurations\"\npython filter.py\n"},{"alpha_fraction":0.7972350120544434,"alphanum_fraction":0.8018433451652527,"avg_line_length":26.125,"blob_id":"481628707a47dff288c11b9e6d909231cc7349a3","content_id":"45f19fd24842bcc08b169ef136eb4ca63171e586","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":217,"license_type":"no_license","max_line_length":67,"num_lines":8,"path":"/src/match_sketches.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# use neural network image features for matching\nexport USE_LSH_DESCRIPTOR_SETTING=1\n\n# load configuration\nsource configuration.sh\n\n# execute script\npython match_sketches.py ../data/demonstrations_all/$ABM/sketched .\n"},{"alpha_fraction":0.664454996585846,"alphanum_fraction":0.6714060306549072,"avg_line_length":27.258928298950195,"blob_id":"e4774f1caf353de94b7c81dd6ef02ecbe90bf191","content_id":"1361346e20e305c02277554221ed63446a8c9e8b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3165,"license_type":"no_license","max_line_length":88,"num_lines":112,"path":"/src/sampling/eum/predict_eum.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import pandas as pd\nimport numpy as np\nimport sys\n\nfrom negotiation_model import *\nfrom bdm_agent import *\n\n# seed this for fixed environment\n# for better replication of results\nrandom.seed(0)\n\n# constants\nINDEPENDENT_VARIABLES = 2\nDEPENDENT_VARIABLES = 2\nSAMPLES = 10\nTIME_LAPSE = 200\n\ndef get_salience_values(timeLapse, numberOfAgents):\n salienceValues = np.zeros((timeLapse, numberOfAgents))\n\n for stepNumber in range(timeLapse):\n for agentNumber in range(numberOfAgents):\n salienceValues[stepNumber][agentNumber] = random.random()\n\n return salienceValues\n\n# Defining the model objects\nclass BDMActor(NegotiationActor):\n DecisionClass = BDM_Agent\n\nclass NegotiationModel_(NegotiationModel):\n # Variables for median caching\n median_computed_last = -1\n median = -1\n \n def find_median(self):\n if self.median_computed_last != self.schedule.steps:\n self.median = super().find_median()\n self.median_computed_last = self.schedule.steps\n return self.median\n\nclass ModelOutput:\n def __init__(self, model):\n '''\n Store data from model run.\n '''\n self.agent_vars = model.datacollector.get_agent_vars_dataframe()\n self.model_vars = model.datacollector.get_model_vars_dataframe()\n self.log = model.log\n\ndef load_data():\n # Load data \n bookData = pd.read_csv(\"sampling/eum/BDM_ColdWar.csv\")\n bookData.Position = (bookData.Position + 100)/200\n\n return bookData\n\ndef main_eum(qValue, tValue, bookData, salienceValues):\n # define agents\n agents = []\n\n for i, row in bookData.iterrows():\n newAgent = BDMActor(row.Country, row.Capability, row.Position, 1)\n newAgent.decision_model.Q = qValue\n newAgent.decision_model.T = tValue\n newAgent.salience = salienceValues[i]\n agents.append(newAgent)\n\n # instantiate model\n model = NegotiationModel_(agents)\n\n # run model\n for stepNumber in range(TIME_LAPSE):\n agentNumber = 0\n for agent in model.agents:\n #agent.salience = random.random()\n agent.salience = salienceValues[stepNumber][agentNumber]\n agentNumber += 1 \n model.step()\n\n # collect data for next steps\n dependentValues = []\n modelOutput = ModelOutput(model)\n dependentValues.append(np.mean(list(modelOutput.model_vars[\"Median\"][-SAMPLES-1:])))\n dependentValues.append(np.mean(list(modelOutput.model_vars[\"Mean\"][-SAMPLES-1:])))\n \n return tuple(dependentValues)\n \ndef main():\n # read data\n bookData = load_data()\n\n # get salience values\n [rows, columns] = bookData.shape\n salienceValues = get_salience_values(TIME_LAPSE, rows)\n\n # simulate using specific ALPs\n alpConfigurationList = eval(sys.argv[1])\n\n # store corresponding slps\n slpConfigurationList = []\n\n for alpConfiguration in alpConfigurationList:\n [qValue, tValue] = alpConfiguration\n\n # run model using those ALPs\n dependentValues = main_eum(qValue, tValue, bookData, salienceValues)\n slpConfigurationList.append(dependentValues)\n\n print(slpConfigurationList)\n \nmain()\n"},{"alpha_fraction":0.648815393447876,"alphanum_fraction":0.6504195332527161,"avg_line_length":46.11627960205078,"blob_id":"35dd288d1da85d245866b180a4d70d40dbca58ec","content_id":"49685798ac9fa1fd65d6d86ee36b4a5a04ac1523","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":8104,"license_type":"no_license","max_line_length":282,"num_lines":172,"path":"/src/filter.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import os\nimport sampling.utils.featurize_image as fi\nfrom sampling.utils.process_folder_images import process_folder\nimport sys\nimport inspect\nimport glob\nimport numpy as np\n\ncurrentDir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))\nparentDir = os.path.dirname(currentDir)\nsys.path.insert(0,parentDir)\n\nimport amf.misc as misc\n\ndef normalized_distance(_a, _b):\n _a = np.array(_a)\n _b = np.array(_b)\n\n b = _b.astype(int)\n a = _a.astype(int)\n norm_diff = np.linalg.norm(b - a)\n norm1 = np.linalg.norm(b)\n norm2 = np.linalg.norm(a)\n return norm_diff / (norm1 + norm2)\n\nclass FilterConfiguration(object):\n def __init__(self):\n # get environment variables\n self.abm = os.environ[\"ABM\"]\n self.demonstrationFolder = os.environ[\"DEMONSTRATION_FOLDER\"]\n self.descriptorSetting = os.environ[\"DESCRIPTOR_SETTING\"]\n self.numDependent = eval(os.environ[\"NUM_DEPENDENT\"])\n self.dataFileName = os.environ[\"DATA_FILE\"]\n self.scaleDataFileName = os.environ[\"SCALE_DATA_FILE\"]\n self.images = eval(os.environ[\"IMAGES\"])\n\n # read configuration\n configurationFileName = os.environ[\"CONFIGURATION_FILE\"]\n configuration = open(configurationFileName).read()\n execd = {}\n exec(configuration, execd)\n self.numIndependent = execd['NUM_INDEPENDENT']\n \n# filter selected configurations based on closeness to demonstrations\ndef select_demonstration_configuratons(filterConfiguration, scaleData):\n # load network if needed\n net = fi.get_network(filterConfiguration.descriptorSetting)\n\n # for normalization\n if(scaleData is not None):\n [dependentValuesMax, dependentValuesMin] = scaleData\n \n # process directories in sorted order\n sortedPaths = []\n for dirName, subdirList, fileList in os.walk(filterConfiguration.demonstrationFolder):\n for sortedSubdirName in sorted(subdirList):\n sortedPaths.append(dirName + \"/\" + sortedSubdirName)\n\n # get slps for demonstrations\n demonstrationSlpsList = []\n demonstrationNameList = []\n for sortedPath in sortedPaths:\n for dirName, subdirList, fileList in os.walk(sortedPath):\n # only use non empty subdirectories\n if(len(fileList) > 0):\n # traverse all files in this directory\n if(bool(filterConfiguration.images) == True):\n # read dependent values as images\n dependentValues = [tuple(fi.get_features(net, dirName + \"/\" + fileName, filterConfiguration.descriptorSetting)) for fileName in fileList]\n else:\n # read dependent values as text\n dependentValues = [tuple([eval(x) for x in open(dirName + \"/\" + fileName).readlines()[0].replace(\"\\n\", \"\").split(\" \")]) for fileName in fileList]\n\n dependentValues = tuple(misc.col_average(dependentValues))\n demonstrationSlpsList.append(dependentValues)\n demonstrationName = dirName.split(\"/\")[-1].strip(\" \")\n demonstrationNameList.append(demonstrationName)\n\n minDistanceAlpsList = []\n minDistanceSlpsList = []\n for demonstrationIndex in range(len(demonstrationSlpsList)):\n demonstrationSlps = demonstrationSlpsList[demonstrationIndex]\n\n # normalize demonstration slps\n if(scaleData is not None):\n demonstrationSlps = tuple((demonstrationSlps[index] - dependentValuesMin[index])/(dependentValuesMax[index] - dependentValuesMin[index]) if (dependentValuesMax[index] > dependentValuesMin[index]) else dependentValuesMax[index] for index in range(len(demonstrationSlps)))\n \n distanceList = []\n \n with open(\"suggested_slps.txt\", \"r\") as suggestedSlpsFile:\n lines = suggestedSlpsFile.readlines()\n \n for suggestionIndex in range(len(lines)):\n # get slp corresponding to this demonstration\n line = lines[suggestionIndex]\n suggestedSlps = eval(line)[demonstrationIndex]\n\n if(bool(filterConfiguration.images == True)):\n if(eval(os.environ[\"IMAGE_FEATURIZATION_HOMOGENEOUS\"]) == 0):\n # this is needed if the two image featurization stages are different\n folderNames = glob.glob(\"sampling/\" + filterConfiguration.abm + \"/images_*\")\n suggestionFolder = folderNames[suggestionIndex]\n suggestedSlps = process_folder(suggestionFolder)\n \n # normalize suggested slps\n if(scaleData is not None):\n suggestedSlps = tuple((suggestedSlps[index] - dependentValuesMin[index])/(dependentValuesMax[index] - dependentValuesMin[index]) if (dependentValuesMax[index] > dependentValuesMin[index]) else dependentValuesMax[index] for index in range(len(suggestedSlps)))\n \n # compute distance\n if(os.environ[\"DISTANCE_METHOD\"] == \"NORMALIZED\"):\n distance = normalized_distance(suggestedSlps, demonstrationSlps)\n else:\n # default option\n distance = misc.distance(suggestedSlps, demonstrationSlps)\n distanceList.append((distance, suggestionIndex, suggestedSlps))\n\n # select entry corresponding to smallest distance\n minDistanceIndex = min(distanceList)[1]\n minDistanceSlps = min(distanceList)[2]\n\n # read the suggested alps corresponding to that index\n with open(\"suggested_alps.txt\", \"r\") as suggestedAlpsFile:\n minDistanceAlps = eval(suggestedAlpsFile.readlines()[minDistanceIndex])[demonstrationIndex]\n \n # write to file\n directory = \"../data/predictions/\" + filterConfiguration.abm + \"/\" + demonstrationNameList[demonstrationIndex] \n if not os.path.exists(directory):\n os.makedirs(directory)\n with open(directory + \"/predicted_alps.txt\", \"w\") as predictionFile:\n predictionFile.write(str(minDistanceAlps) + \" \" + str(minDistanceSlps))\n\n if(bool(filterConfiguration.images == True)):\n # if images used, copy images\n # get all image simulation folders\n folderNames = glob.glob(\"sampling/\" + filterConfiguration.abm + \"/images_*\")\n minFolder = folderNames[minDistanceIndex]\n os.system(\"cp -r \" + minFolder + \" \" + directory)\n \n# filter configurations with command line parameters\ndef main():\n filterConfiguration = FilterConfiguration()\n\n # edit scale data file name based on descriptor setting\n scaleDataFileNamePrefix = filterConfiguration.scaleDataFileName.split(\".\")[0]\n scaleDataFileNameExtension = filterConfiguration.scaleDataFileName.split(\".\")[1]\n scaleDataFileName = scaleDataFileNamePrefix + \"_\" + filterConfiguration.abm\n if(bool(filterConfiguration.images) == True):\n scaleDataFileName += \"_\" + filterConfiguration.descriptorSetting.lower()\n scaleDataFileName += \".\" + scaleDataFileNameExtension\n\n # get scale information\n fileName = \"sampling/\" + filterConfiguration.abm + \"/\" + scaleDataFileName\n\n # scale data assumed not present by default\n scaleData = None\n\n if(os.path.exists(fileName) == True):\n with open(fileName, \"r\") as scaleDataFile:\n # read in order of writing in map.py\n maxs = eval(scaleDataFile.readline().strip(\"\\n\"))\n mins = eval(scaleDataFile.readline().strip(\"\\n\"))\n\n dependentMaxs = maxs[filterConfiguration.numIndependent:]\n dependentMins = mins[filterConfiguration.numIndependent:]\n dependentValuesMax = {i: dependentMaxs[i] for i in range(len(dependentMaxs))}\n dependentValuesMin = {i: dependentMins[i] for i in range(len(dependentMins))}\n scaleData = [dependentValuesMax, dependentValuesMin]\n\n select_demonstration_configuratons(filterConfiguration, scaleData)\n \n# execute main\nmain()\n"},{"alpha_fraction":0.593353271484375,"alphanum_fraction":0.6056758761405945,"avg_line_length":22.286956787109375,"blob_id":"81b80e6126f8ea8bfb5aacaaa1f1220f4c320dd7","content_id":"f57dbc1885b6e61862bd3d9c7a5dad507ac44b03","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2678,"license_type":"no_license","max_line_length":129,"num_lines":115,"path":"/src/sampling/schelling/sample_schelling_parallel.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"from model import SchellingModel\nimport random\nimport multiprocessing as mp\n\n# only use three of the several ALPs\n# density (0,1)\n# minority_pc (0,1)\n\n# use one model-level SLP\n# happy\n\n# parallel processing constants\nBATCH_SIZE = 8\n\n# constant ALPs\nHEIGHT = 20\nWIDTH = 20\nHOMOPHILY = 4\n\n# constants\n# max_iters in the example is specified as 1000\n# so we run for 1000 steps\nINDEPENDENT_VARIABLES = 2\nDEPENDENT_VARIABLES = 1\nRUNS = 10000\nSAMPLES = 10\nTIME_LAPSE = 1000\n\n# process single run data\ndef process_run(q, run, density, minorityPc):\n try:\n line = main_eum(density, minorityPc)\n except:\n line = \"\"\n \n # add newline for clean output\n line = \"\\n\" + line\n \n # write to queue\n q.put(line)\n \ndef main_schelling(density, minorityPc):\n # instantiate and run model\n model = SchellingModel(\n height=HEIGHT,\n width=WIDTH,\n density=density,\n minority_pc=minorityPc,\n homophily=HOMOPHILY)\n\n for i in range(SAMPLES):\n try:\n # step\n model.step()\n except:\n # saturated\n # no empty cells\n pass\n\n # collect data for next steps\n dependentValues = []\n\n for i in range(SAMPLES):\n # step\n model.step()\n\n # read data\n data = model.datacollector.get_model_vars_dataframe() \n dependentValues.append(list(data.happy)[-SAMPLES-1:])\n\n # print line corresponding to this execution\n line = str(density) + \" \" + str(minorityPc)\n for dependentValue in dependentValues:\n line += \" \" + str(dependentValue).replace(\"[\", \"\").replace(\"]\", \"\").replace(\" \", \"\")\n\n return line\n\n \ndef main():\n # header string\n print(\"i\"*INDEPENDENT_VARIABLES + \"d\"*DEPENDENT_VARIABLES)\n\n densityValues = []\n minorityPcValues = []\n \n for run in range(RUNS):\n # sample random ALPs\n density = random.random() \n minorityPc = random.random() \n\n densityValues.append(density)\n minorityPcValues.append(minorityPc)\n\n # setup a list of processes that we want to run\n q = mp.Queue()\n processes = [mp.Process(target=process_run, args=(q, run, densityValues[run], minorityPcValues[run])) for run in range(RUNS)]\n batchSize = BATCH_SIZE\n batches = [processes[i:i+batchSize] for i in range(0, len(processes), batchSize)]\n\n outFile = open(\"output.txt\", \"a\")\n \n for batch in batches:\n # run processes\n for p in batch:\n p.start()\n for p in batch:\n line = q.get()\n outFile.write(line)\n # exit the completed processes\n for p in batch:\n p.join()\n\n outFile.close()\n\nmain()\n"},{"alpha_fraction":0.5906005501747131,"alphanum_fraction":0.6073107123374939,"avg_line_length":31.457626342773438,"blob_id":"9a7f39e9b6aef59e42587e05febaddc8a1629db8","content_id":"25299fd63872ecf9830f3e159794c1cad234e2cf","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1915,"license_type":"no_license","max_line_length":106,"num_lines":59,"path":"/snippets/turbulence/plot_suggested_slps.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import matplotlib.pyplot as plt\nimport numpy as np\n\n#demonstrations = [\"a\", \"b\", \"c\"]\n#demonstrations = [\"_a\", \"_b\", \"_c\"]\ndemonstrations = [\"a\"]\n#demonstrations = [\"_a\"]\n#demonstrations = [\"a_filtered\"]\n\nfolderName = \"error_weighted\"\n#folderName = \"min_error\"\n\nwith open(\"../../src/sampling/turbulence/scale_data_turbulence_density.txt\", \"r\") as inFile:\n allLines = inFile.readlines()\n maxValueDensity = eval(allLines[0])[-1]\n minValueDensity = eval(allLines[1])[-1]\n\nfor index, item in enumerate(demonstrations):\n allPointsX = []\n allPointsY = []\n\n # scatter\n \n plt.figure()\n with open(folderName + \"/suggested_slps_\" + item + \".txt\", \"r\") as inFile:\n for line in inFile:\n point = eval(line)[0]\n\n # scale\n #point = tuple([(point[0]-minValueDensity)/(maxValueDensity-minValueDensity)])\n\n plt.scatter(point[0], point[0], c=\"b\", marker=\"o\", alpha=0.3)\n allPointsX.append(point[0])\n allPointsY.append(point[0])\n \n plt.scatter(point[0], point[0], c=\"b\", marker=\"o\", alpha=0.3, label=\"Suggested SLPs\")\n\n with open(\"demonstration_slps.txt\", \"r\") as inFile:\n point = eval(inFile.readlines()[index])\n\n # scale\n #point = tuple([(point[0]-minValueDensity)/(maxValueDensity-minValueDensity)])\n \n plt.scatter(point[0], point[0], c=\"r\", marker=\"s\", alpha=1.0, label=\"Demonstration SLPs\")\n\n plt.scatter(np.mean(allPointsX), np.mean(allPointsY), c=\"g\", marker=\"s\", alpha=1.0, label=\"Mean SLPs\")\n plt.xlabel(\"Density\")\n plt.ylabel(\"Density\")\n plt.legend()\n plt.savefig(folderName + \"/suggested_slps_\" + item)\n\n # histogram\n \n plt.figure()\n plt.hist(allPointsX, bins=50, range=(0,50))\n plt.xlabel(\"Density\")\n plt.ylabel(\"Number of Points\")\n plt.gca().set_ylim([0,100])\n plt.savefig(folderName + \"/histogram_suggested_slps_\" + item)\n"},{"alpha_fraction":0.699999988079071,"alphanum_fraction":0.7785714268684387,"avg_line_length":22.33333396911621,"blob_id":"f7f52b9a07b9f0ffd9d5d242d30cadc8e5c9381a","content_id":"c2c9a561ebca1d1d5f50d1ef74188a8d3d61f2bc","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":140,"license_type":"no_license","max_line_length":54,"num_lines":6,"path":"/src/domainconfs/turbulence.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"TRAINING_SIZE = 200\nVALIDATION_SIZE = 2500\nRM_GRANULARITY = 300\nREGRESSION = \"sklearn.neighbors.KNeighborsRegressor()\"\n\nNUM_INDEPENDENT = 2\n"},{"alpha_fraction":0.5873016119003296,"alphanum_fraction":0.5873016119003296,"avg_line_length":41,"blob_id":"7b5b921b865a519db7dd55473231f2561b10b310","content_id":"bcbbfa8916d5a7b31b6eeee281019ab173e38515","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":252,"license_type":"no_license","max_line_length":61,"num_lines":6,"path":"/data/domaindata/fix_results.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# needed in case two results answers the same name identifier\nwith open(\"results.csv\", \"r\") as inFile:\n with open(\"temp.csv\", \"w\") as outFile:\n for line in inFile:\n line = line.replace(\"|\", \"\\\",\\\"\")\n outFile.write(line)\n"},{"alpha_fraction":0.6573289632797241,"alphanum_fraction":0.6801303029060364,"avg_line_length":22.25757598876953,"blob_id":"56265a9ff22d8615011ce681c7f3c10fa2c4981b","content_id":"bef667f68ffdf3fceb5aef11dcc07b40212654f0","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1535,"license_type":"no_license","max_line_length":67,"num_lines":66,"path":"/src/sampling/forest_fire/predict_forest_fire.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"from model import ForestFire\nimport random\nimport sys\nimport numpy as np\n\n# only one ALP\n# density (0,1)\n\n# constant ALPs\nHEIGHT = 100\nWIDTH = 100\n\n# constants\n# max_iters in the example is specified as 1000\n# so we run for 1000 steps\nINDEPENDENT_VARIABLES = 1\nDEPENDENT_VARIABLES = 3\nRUNS = 10000\nSAMPLES = 10\nTIME_LAPSE = 200\n\ndef main_forest_fire(density=0.65):\n # instantiate and run model\n model = ForestFire(height=HEIGHT, width=WIDTH, density=density)\n\n # run time lapse\n for i in range(TIME_LAPSE):\n model.step()\n\n # collect data for next steps\n dependentValues = []\n\n for i in range(SAMPLES):\n # step\n model.step()\n\n # read data\n data = model.datacollector.get_model_vars_dataframe()\n burnedTrees = np.array(list(data.BurnedOut)[-SAMPLES-1:])\n fineTrees = np.array(list(data.Fine)[-SAMPLES-1:])\n initialTrees = burnedTrees + fineTrees\n dependentValues.append(np.mean(burnedTrees/initialTrees))\n\n return tuple(dependentValues)\n \ndef main():\n # fixed randomness\n # parimarily done to test feedback\n random.seed(0)\n \n # simulate using specific ALPs\n alpConfigurationList = eval(sys.argv[1])\n\n # store corresponding slps\n slpConfigurationList = []\n \n for alpConfiguration in alpConfigurationList:\n [density] = alpConfiguration\n \n # run model using those ALPs\n dependentValues = main_forest_fire(density)\n slpConfigurationList.append(dependentValues)\n\n print(slpConfigurationList)\n \nmain()\n"},{"alpha_fraction":0.6421149373054504,"alphanum_fraction":0.6500611305236816,"avg_line_length":27.452173233032227,"blob_id":"9f0341d6f2f2df8aa43a8dffb7056ede562fc780","content_id":"890c9096b744269d09268aa6c9619c8fa586c910","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3272,"license_type":"no_license","max_line_length":92,"num_lines":115,"path":"/src/sampling/eum/sample_eum.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import pandas as pd\nimport numpy as np\nimport sys\n\nfrom negotiation_model import *\nfrom bdm_agent import *\n\n# seed this for fixed environment\n# for better replication of results\nrandom.seed(0)\n\n# constants\nINDEPENDENT_VARIABLES = 2\nDEPENDENT_VARIABLES = 2\nRUNS = 10000\nSAMPLES = 10\nTIME_LAPSE = 200\n\ndef get_salience_values(timeLapse, numberOfAgents):\n salienceValues = np.zeros((timeLapse, numberOfAgents))\n\n for stepNumber in range(timeLapse):\n for agentNumber in range(numberOfAgents):\n salienceValues[stepNumber][agentNumber] = random.random()\n\n return salienceValues\n\n# Defining the model objects\nclass BDMActor(NegotiationActor):\n DecisionClass = BDM_Agent\n\nclass NegotiationModel_(NegotiationModel):\n # Variables for median caching\n median_computed_last = -1\n median = -1\n \n def find_median(self):\n if self.median_computed_last != self.schedule.steps:\n self.median = super().find_median()\n self.median_computed_last = self.schedule.steps\n return self.median\n\nclass ModelOutput:\n def __init__(self, model):\n '''\n Store data from model run.\n '''\n self.agent_vars = model.datacollector.get_agent_vars_dataframe()\n self.model_vars = model.datacollector.get_model_vars_dataframe()\n self.log = model.log\n\ndef load_data():\n # Load data \n bookData = pd.read_csv(\"BDM_ColdWar.csv\")\n bookData.Position = (bookData.Position + 100)/200\n\n return bookData\n\ndef main_eum(qValue, tValue, bookData, salienceValues):\n # define agents\n agents = []\n\n for i, row in bookData.iterrows():\n newAgent = BDMActor(row.Country, row.Capability, row.Position, 1)\n newAgent.decision_model.Q = qValue\n newAgent.decision_model.T = tValue\n newAgent.salience = salienceValues[i]\n agents.append(newAgent)\n\n # instantiate model\n model = NegotiationModel_(agents)\n\n # run model\n for stepNumber in range(TIME_LAPSE):\n agentNumber = 0\n for agent in model.agents:\n #agent.salience = random.random()\n agent.salience = salienceValues[stepNumber][agentNumber]\n agentNumber += 1 \n model.step()\n\n # collect data for next steps\n dependentValues = []\n modelOutput = ModelOutput(model)\n dependentValues.append(list(modelOutput.model_vars[\"Median\"][-SAMPLES-1:]))\n dependentValues.append(list(modelOutput.model_vars[\"Mean\"][-SAMPLES-1:]))\n \n # print line corresponding to this execution\n line = str(qValue) + \" \" + str(tValue)\n for dependentValue in dependentValues:\n line += \" \" + str(dependentValue).replace(\"[\", \"\").replace(\"]\", \"\").replace(\" \", \"\")\n print(line)\n\ndef main():\n # header string\n print(\"i\"*INDEPENDENT_VARIABLES + \"d\"*DEPENDENT_VARIABLES)\n\n # read data\n bookData = load_data()\n\n # get salience values\n [rows, columns] = bookData.shape\n salienceValues = get_salience_values(TIME_LAPSE, rows)\n \n for run in range(RUNS):\n print(run, file=sys.stderr)\n \n # sample random ALPs\n qValue = random.random()\n tValue = random.random()\n \n # run model using those ALPs\n main_eum(qValue, tValue, bookData, salienceValues)\n \nmain()\n"},{"alpha_fraction":0.5925377607345581,"alphanum_fraction":0.5987563133239746,"avg_line_length":39.2023811340332,"blob_id":"dab20f1fa3ceb3346a5f77d3e106f650a24c62ae","content_id":"d69d05eab60363b2473cf96ff9331a809e81d73e","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3377,"license_type":"no_license","max_line_length":103,"num_lines":84,"path":"/src/amf/fm.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# fm.py contains the ForwardMapping frontend.\n\nimport data\nimport sys\nimport numpy as np\nimport sklearn\nfrom sklearn.gaussian_process import GaussianProcessRegressor\nfrom smt.surrogate_models import RBF, KRG, KPLS, KPLSK\nimport os\n\n# This thing is the Forward Mapping\nclass ForwardMapping(object):\n\n # pass in the regression algorithm that we will be using, along with any initialization parameters\n def __init__(self, regression_class, num_dependent):\n \"\"\"\n Initialized a forward mapping.\n Pass in the regression class to be used, the initial parameters for initlializing this reg class.\n Also, pass the the number of dependent variables the data set will have.\n \"\"\"\n self._mappers = [eval(regression_class) for x in range(num_dependent)]\n\n def train(self, training_set):\n # remove data entries which contain nan\n training_set = [x for x in training_set if not np.isnan(x[0]).any()]\n training_set = [x for x in training_set if not np.isnan(x[1]).any()]\n training_sets = data.split_dep(training_set)\n\n for mapper, tset in zip(self._mappers, training_sets):\n X = np.array([list(x[0]) for x in tset])\n Y = np.array([x[1] for x in tset])\n \n if(\"smt\" in str(mapper)):\n # using surrogate modeling toolbox\n X = np.array([x[0] for x in X])\n VThreshold = 0.001\n\n XSafe = [X[0]]\n YSafe = [Y[0]]\n for index in range(1, len(X)):\n currentX = X[index]\n currentY = Y[index]\n XSafetyCheck = XSafe + [currentX]\n YSafetyCheck = YSafe + [currentY]\n\n # we need to only collect those points that do not case matrix inversion problems\n # check using svd\n configurationFileName = os.environ[\"CONFIGURATION_FILE\"]\n configuration = open(configurationFileName).read()\n execd = {}\n exec(configuration, execd)\n numIndependent = execd['NUM_INDEPENDENT']\n numDependent = eval(os.environ[\"NUM_DEPENDENT\"])\n XSafetyCheck = np.array(XSafetyCheck).reshape(numIndependent, -1)\n YSafetyCheck = np.array(YSafetyCheck).reshape(numDependent, -1)\n XYSafetyCheck = np.concatenate((XSafetyCheck, YSafetyCheck), axis=0)\n U, s, V = np.linalg.svd(XYSafetyCheck, full_matrices=True)\n V = np.absolute(V)\n\n if(np.min(V) > VThreshold):\n XSafe.append(currentX)\n YSafe.append(currentY)\n \n # convert to numpy array\n # train model\n XSafe = np.array(XSafe)\n YSafe = np.array(YSafe)\n mapper.set_training_values(XSafe, YSafe)\n mapper.train()\n else:\n mapper.fit(X, Y)\n\n def predict(self, configuration):\n prediction = []\n for mapper in self._mappers:\n if(\"smt\" in str(mapper)):\n mapperPrediction = mapper.predict_values(np.array(list(configuration)).reshape(1, -1))\n mapperPrediction = [x[0] for x in mapperPrediction]\n else:\n mapperPrediction = mapper.predict(np.array(list(configuration)).reshape(1, -1)) \n prediction.append(mapperPrediction)\n prediction = tuple([x[0] for x in prediction])\n \n return prediction\n"},{"alpha_fraction":0.7027027010917664,"alphanum_fraction":0.8108108043670654,"avg_line_length":36,"blob_id":"80bd3913abf6b1ca1504165909b145c267f58805","content_id":"8cb3915a5ec3995ca8a235d648d9cf09a614929c","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":37,"license_type":"no_license","max_line_length":36,"num_lines":1,"path":"/app/run_public_url.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"ngrok http 5000 -subdomain=swarm-lfd\n"},{"alpha_fraction":0.6554731726646423,"alphanum_fraction":0.6720453500747681,"avg_line_length":25.662790298461914,"blob_id":"c00719c19bf8a7489d616f74b4ffb39138f77786","content_id":"dfe90210275d5a2213fa24eaae6b80a264af5f29","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2293,"license_type":"no_license","max_line_length":89,"num_lines":86,"path":"/src/sampling/civil_violence/predict_civil_violence.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"from model import CivilViolenceModel\nimport random\nimport sys\nimport numpy as np\nimport os\n\n# only use three of the several ALPs\n# citizen_density (0,1)\n# cop_density (0,1)\n# legitimacy (0,1)\n\n# use three model-level SLPs\n# Quiescent\n# Active\n# Jailed\n\n# constant ALPs\nHEIGHT = 40\nWIDTH = 40\nCITIZEN_VISION = 7\nCOP_VISION = 7\nMAX_JAIL_TERM = 1000\n\n# constants\n# max_iters in the example is specified as 1000\n# so we run for 1000 steps\nINDEPENDENT_VARIABLES = 3\nDEPENDENT_VARIABLES = 3\nSAMPLES = 10\nTIME_LAPSE = 1000\n\ndef main_civil_violence(citizenDensity, copDensity, legitimacy):\n # instantiate and run model\n model = CivilViolenceModel(\n height=HEIGHT,\n width=WIDTH,\n citizen_vision=CITIZEN_VISION,\n cop_vision=COP_VISION,\n max_jail_term=MAX_JAIL_TERM,\n citizen_density=citizenDensity,\n cop_density=copDensity,\n legitimacy=legitimacy)\n\n # run time lapse\n for i in range(TIME_LAPSE):\n model.step()\n\n # collect data for next steps\n dependentValues = []\n\n for i in range(SAMPLES):\n # step\n model.step()\n\n # read data\n data = model.dc.get_model_vars_dataframe() \n dependentValues.append(np.mean(list(data.Quiescent)[-SAMPLES-1:]))\n dependentValues.append(np.mean(list(data.Active)[-SAMPLES-1:]))\n dependentValues.append(np.mean(list(data.Jailed)[-SAMPLES-1:]))\n\n return tuple(dependentValues)\n \ndef main():\n # simulate using specific ALPs\n alpConfigurationList = eval(sys.argv[1])\n\n # store corresponding slps\n slpConfigurationList = []\n \n for alpConfiguration in alpConfigurationList:\n [citizenDensity, copDensity, legitimacy] = alpConfiguration\n\n # cop density + citizen density should be less than 1\n if(citizenDensity + copDensity > 1):\n # invalid parameters\n # set to infinity so distance is infinite\n # this value will be considered as a bad suggestion when filtering\n dependentValues = [float(\"Inf\")]*eval(os.environ[\"NUM_DEPENDENT\"])\n else:\n # run model using those ALPs\n dependentValues = main_civil_violence(citizenDensity, copDensity, legitimacy)\n slpConfigurationList.append(dependentValues)\n\n print(slpConfigurationList)\n \nmain()\n"},{"alpha_fraction":0.6349663734436035,"alphanum_fraction":0.6551392674446106,"avg_line_length":25.024999618530273,"blob_id":"180657a6bfc68f62ae32fc65df08057a4dceb821","content_id":"0dac153578d624ace0ca139bc6150746db256084","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2082,"license_type":"no_license","max_line_length":92,"num_lines":80,"path":"/src/sampling/civil_violence/sample_civil_violence.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"from model import CivilViolenceModel\nimport random\n\n# only use three of the several ALPs\n# citizen_density (0,1)\n# cop_density (0,1)\n# legitimacy (0,1)\n\n# use three model-level SLPs\n# Quiescent\n# Active\n# Jailed\n\n# constant ALPs\nHEIGHT = 40\nWIDTH = 40\nCITIZEN_VISION = 7\nCOP_VISION = 7\nMAX_JAIL_TERM = 1000\n\n# constants\n# max_iters in the example is specified as 1000\n# so we run for 1000 steps\nINDEPENDENT_VARIABLES = 3\nDEPENDENT_VARIABLES = 3\nRUNS = 10000\nSAMPLES = 10\nTIME_LAPSE = 1000\n\ndef main_civil_violence(citizenDensity, copDensity, legitimacy):\n # instantiate and run model\n model = CivilViolenceModel(\n height=HEIGHT,\n width=WIDTH,\n citizen_vision=CITIZEN_VISION,\n cop_vision=COP_VISION,\n max_jail_term=MAX_JAIL_TERM,\n citizen_density=citizenDensity,\n cop_density=copDensity,\n legitimacy=legitimacy)\n\n # run time lapse\n for i in range(TIME_LAPSE):\n model.step()\n\n # collect data for next steps\n dependentValues = []\n\n for i in range(SAMPLES):\n # step\n model.step()\n\n # read data\n data = model.dc.get_model_vars_dataframe() \n dependentValues.append(list(data.Quiescent)[-SAMPLES-1:])\n dependentValues.append(list(data.Active)[-SAMPLES-1:])\n dependentValues.append(list(data.Jailed)[-SAMPLES-1:])\n\n # print line corresponding to this execution\n line = str(citizenDensity) + \" \" + str(copDensity) + \" \" + str(legitimacy)\n for dependentValue in dependentValues:\n line += \" \" + str(dependentValue).replace(\"[\", \"\").replace(\"]\", \"\").replace(\" \", \"\")\n print(line)\n\n \ndef main():\n # header string\n print(\"i\"*INDEPENDENT_VARIABLES + \"d\"*DEPENDENT_VARIABLES)\n\n for run in range(RUNS):\n # sample random ALPs\n # cop density + citizen density should be less than 1\n citizenDensity = random.random() \n copDensity = (1-citizenDensity)*random.random() \n legitimacy = random.random()\n \n # run model using those ALPs\n main_civil_violence(citizenDensity, copDensity, legitimacy)\n\nmain()\n"},{"alpha_fraction":0.6672787070274353,"alphanum_fraction":0.6705839037895203,"avg_line_length":31.807228088378906,"blob_id":"dc44bad69890e9034018b1c7ed7e876a945268d2","content_id":"4069ce92e26a40606230288b7df6e0d449728504","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2723,"license_type":"no_license","max_line_length":198,"num_lines":83,"path":"/src/map.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# This script takes each row of a data set and processes it for use with AMF\nimport sys\nimport math\nimport os\n\nno_entry = \"NOENTRY\"\nSCALE_DATA_FILE = \"scale_data.txt\"\n\ndef map_abm(tokens):\n \"\"\"\n The raw data set is a combination of floats followed by a combination of lists and looks like this:\n FLOAT FLOAT ... LIST LIST ...\n Non-lists are independent variables\n Lists are dependent variables\n It takes the average of the list and returns it with the independent variables.\n \"\"\"\n\n tokenAvg = []\n\n for token in tokens:\n if(type(token) == list):\n avg = sum(token)/len(token)\n else:\n avg = token\n\n tokenAvg.append(avg)\n\n return tuple(tokenAvg)\n\ndef linear_scale(dataset):\n \"\"\"Linearly scales an entire data set to a range of 0-1\"\"\"\n\n maxs = list(dataset[0])\n mins = list(dataset[0])\n\n for datarow in dataset:\n for idx, item in enumerate(datarow):\n if item == no_entry: continue\n\n maxs[idx] = max(maxs[idx], item)\n mins[idx] = min(mins[idx], item)\n\n scaleDataFileName = SCALE_DATA_FILE\n abm = os.environ[\"ABM\"]\n descriptorSetting = os.environ[\"DESCRIPTOR_SETTING\"]\n\n # edit scale data file name based on descriptor setting\n scaleDataFileNamePrefix = scaleDataFileName.split(\".\")[0]\n scaleDataFileNameExtension = scaleDataFileName.split(\".\")[1]\n scaleDataFileName = scaleDataFileNamePrefix + \"_\" + abm\n if(len(descriptorSetting) > 0):\n # valid descriptor\n scaleDataFileName += \"_\" + descriptorSetting.lower()\n scaleDataFileName += \".\" + scaleDataFileNameExtension\n \n with open(scaleDataFileName, \"w\") as scaleDataFile:\n scaleDataFile.write(str(maxs))\n scaleDataFile.write(\"\\n\")\n scaleDataFile.write(str(mins))\n\n return [ [ (no_entry if item == no_entry else maxs[idx] if maxs[idx] == mins[idx] else (item - mins[idx]) / (maxs[idx] - mins[idx])) for idx, item in enumerate(datarow) ] for datarow in dataset ]\n\ndef parse(data_file_path):\n \"\"\"\n Applies 'map_function' to each row (as a string) of a data set.\n 'map_function' should takes a list of tokens as the only parameter and return a string as the only parameter.\n Prints out the new data set.\n \"\"\"\n\n return [ [ float(token) if token.find(',') == -1 else [ float(item) for item in token.split(',') ] for token in line.split() ] for line in open(data_file_path).xreadlines() ]\n\ndef dump(dataset):\n print \"\\n\".join( \" \".join(str(item) for item in row) for row in dataset)\n\ndef process_abm():\n parsed = parse(sys.argv[1])\n averaged = map(map_abm, parsed)\n scaled = linear_scale(averaged)\n dump(scaled)\n\n# average over system level paramters in raw data file\n# raw data file is provided as argument\nprocess_abm()\n"},{"alpha_fraction":0.6269554495811462,"alphanum_fraction":0.6281588673591614,"avg_line_length":26.09782600402832,"blob_id":"910587596e08507a8f46aa10de2d59ca35161a15","content_id":"a29d16ac260c7abde4aa19fa97ee8bb0378abaa3","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2493,"license_type":"no_license","max_line_length":130,"num_lines":92,"path":"/src/sampling/utils/process_training_images.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# libraries used\nimport sys\nimport os\nimport featurize_image as fi\nimport multiprocessing as mp\nimport glob\n\n# parallel processing constants\nBATCH_SIZE = 8\n\n# process single run data\ndef process_run(q, run, runConfiguration, line):\n\n # unpack run configuration\n [inFolder, dataFile, numDependent, net, descriptorSetting] = runConfiguration\n\n # initialize collection of dependent values\n depValuesStrings = [\"\"]*numDependent\n\n for imageName in glob.glob(inFolder + \"/\" + str(run) + \"_*.*\"):\n # read image\n print >> sys.stderr, (\"now processing \" + imageName)\n \n # define dependent values\n depValues = fi.get_features(net, imageName, descriptorSetting)\n\n for index in range(len(depValues)):\n depValuesStrings[index] += str(depValues[index]) + \",\"\n\n # add value data to line\n for depValuesString in depValuesStrings:\n line += \" \" + depValuesString[:-1]\n\n # remove extra \",\" at end\n line = line + \"\\n\"\n\n # write to queue\n q.put(line)\n\n# read images and extract features as dependent variables\ndef main():\n\n # get environment variables\n abm = os.environ[\"ABM\"]\n descriptorSetting = os.environ[\"DESCRIPTOR_SETTING\"]\n numDependent = eval(os.environ[\"NUM_DEPENDENT\"])\n dataFileName = os.environ[\"DATA_FILE\"]\n\n dataFile = open(abm + \".txt\", \"r\")\n tempFile = open(\"temp.txt\", \"w\")\n\n # set up image reading\n inFolder = abm + \"/images\"\n\n # load feature extractor if needed\n net = fi.get_network(descriptorSetting)\n\n # read data\n lines = dataFile.readlines()\n\n # get number of runs\n runs = len(lines)\n \n # process runs\n runConfiguration = [inFolder, dataFile, numDependent, net, descriptorSetting]\n\n # setup a list of processes that we want to run\n q = mp.Queue()\n processes = [mp.Process(target=process_run, args=(q, run, runConfiguration, lines[run].strip(\" \\n\"))) for run in range(runs)] \n batchSize = BATCH_SIZE\n batches = [processes[i:i+batchSize] for i in range(0, len(processes), batchSize)]\n\n for batch in batches:\n # run processes\n for p in batch:\n p.start()\n for p in batch:\n line = q.get()\n tempFile.write(line)\n # exit the completed processes\n for p in batch:\n p.join()\n\n # close files and rename\n dataFile.close()\n tempFile.close()\n\n os.remove(abm + \".txt\")\n os.rename(\"temp.txt\", abm + \".txt\")\n\nif __name__ == \"__main__\":\n main()\n"},{"alpha_fraction":0.5741397738456726,"alphanum_fraction":0.5817003846168518,"avg_line_length":35.206703186035156,"blob_id":"9310d151e1411d6b2469b6a34580daca836eca07","content_id":"6e231729abd51e8557ec5f9e45a0633308bc892a","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":6481,"license_type":"no_license","max_line_length":121,"num_lines":179,"path":"/src/sampling/utils/outlier_detection.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import numpy as np\nimport scipy.io as sio\nfrom sets import Set\nfrom ltfatpy.sigproc.thresh import thresh\nfrom sklearn.cluster import KMeans, MeanShift\nfrom collections import Counter\n\nimport sys\n\ndef lpod(trainData, testData, trainLabels, testLabels, lNum, tol, lambdaValue):\n # this is a python port of the original matlab code provided by the authors\n # of the 2017 paper titled \"Efficient Outlier Detection for High-Dimensional Data\"\n \n # supress invalid division warning\n # this happens in the matlab code too so we're ok\n np.seterr(divide='ignore', invalid='ignore')\n \n # size of x and y\n # check for size mismatch\n [rtrainData,ctrainData] = trainData.shape\n [rtestData,ctestData] = testData.shape\n assert rtrainData == rtestData\n\n # allocate memory to the maximum size\n # T is the score matrix of trainData\n # P is the loading matrix of trainData\n # U is the score matrix of testData\n # Q is the loading matrix of testData\n # B is the matrix of regression coefficient\n # W is the weight matrix of trainData \n n = max(ctrainData, ctestData)\n T = np.zeros((rtrainData, n)) \n P = np.zeros((ctrainData, n)) \n U = np.zeros((rtestData, n)) \n Q = np.zeros((ctestData, n)) \n B = np.zeros((n, n)) \n W = P \n k = 0\n\n # iteration loop if residual is larger than specfied\n iteration = 1\n\n while(iteration < lNum and np.linalg.norm(testData) > tol and k < n):\n # choose the column of x has the largest square of sum as t\n # choose the column of y has the largest square of sum as u\n tidx = np.argmax(sum(np.multiply(trainData, trainData)))\n uidx = np.argmax(sum(np.multiply(testData, testData)))\n t1 = trainData[:, tidx]\n u = testData[:, uidx]\n t = np.zeros((rtrainData, 1))\n\n # iteration for outer modeling until convergence\n while (np.linalg.norm(t1 - t) > tol):\n # computing the weight vector of trainData\n # computing the score vector of trainData\n # obtaining q by soft-threshold\n w = np.matmul(trainData.transpose(), u) \n w = w / np.linalg.norm(w)\n t = t1\n t1 = np.matmul(trainData, w) \n q, dummy = thresh(np.matmul(testData.transpose(), t1), lambdaValue, thresh_type=\"soft\")\n \n # computing the loading vector of testData according to t\n # computing the score vector of testData\n q = q / np.linalg.norm(q)\n u = np.matmul(testData, q) \n \n # update p based on t\n # computing the loading vecor of trainData\n t = t1\n p = np.matmul(trainData.transpose(), t) / np.matmul(t.transpose(), t)\n pnorm = np.linalg.norm(p)\n p = p / pnorm \n t = t * pnorm\n w = w * pnorm\n\n # regression and residuals\n t = t.reshape((-1, 1))\n p = p.reshape((-1, 1))\n u = u.reshape((-1, 1))\n q = q.reshape((-1, 1))\n\n b = np.matmul(u.transpose(), t) / np.matmul(t.transpose(), t)\n trainData = trainData - np.matmul(t, p.transpose())\n testData = testData - (b * np.matmul(t, q.transpose()))\n\n # save iteration results to outputs:\n k += 1\n T[:, k] = list(t)\n P[:, k] = list(p)\n U[:, k] = list(u)\n Q[:, k] = list(q)\n W[:, k] = list(w)\n B[k, k] = b\n \n iteration += 1\n\n # end of iterations\n T = T[:, :k]\n P = P[:, :k]\n U = U[:, :k]\n Q = Q[:, :k]\n W = W[:, :k]\n B = B[:k, :k]\n\n # prediction stage:\n # predicting real-valued matrix for test_data\n # transforming outputs into the corresponding label matrix\n # 0.5 is the default threshold value\n outputs = np.matmul(np.matmul(np.matmul(P, B), Q.transpose()).transpose(), trainLabels)\n [numClass, numTesting] = outputs.shape\n predictionLabels = np.zeros((numClass, numTesting))\n for i in range(numTesting):\n for j in range(numClass):\n if(outputs[j,i] >= 0.5): \n predictionLabels[j,i] = 1\n else:\n predictionLabels[j,i] = 0\n\n return predictionLabels\n\ndef do_lpod(X, Y):\n # outlier detection\n Y = Y.astype(float).transpose()\n allDataLength = len(Y)\n trainDataFraction = 0.7\n trainDataLength = np.floor(trainDataFraction * allDataLength)\n \n # create logical index vector\n # randomize order to select random elements\n trainIndexes = np.random.choice(int(allDataLength), int(trainDataLength), replace=False)\n testIndexes = np.setdiff1d(range(allDataLength), trainIndexes)\n \n trainData = X[trainIndexes, :]\n trainLabels = Y[trainIndexes]\n testData = X[testIndexes, :]\n testLabels = Y[testIndexes]\n \n # hyper parameters\n # get predictions\n lNum = 50\n tol = 1e-10\n lambdaValue = 0.1\n predictionLabels = lpod(trainData.transpose(), testData.transpose(), trainLabels, testLabels, lNum, tol, lambdaValue)\n lpodX = np.concatenate((trainData, testData), axis=0)\n lpodY = np.concatenate((trainLabels, testLabels), axis=0)\n\n return [lpodX, lpodY]\n\ndef get_approximate_labels(allConfigurations):\n # cluster points\n # outlier is decided as the minority cluster\n # count population of each cluster\n # outliers are labeled as 1\n # inliers are labeled as 0\n # force 2 clusters as backup in case meanshift finds a single cluster\n X = np.array(allConfigurations)\n meanShift = MeanShift(bin_seeding=True)\n meanShift.fit(X)\n nClusters = 2\n kmeans = KMeans(init='k-means++', n_clusters=nClusters)\n kmeans.fit(X)\n labelCounts = Counter(meanShift.labels_)\n\n if(len(labelCounts.keys()) > 1):\n # multiple clusters found by meanshift\n labels = meanShift.labels_\n else:\n # single cluster found by meanshift\n # use kmeans clusters instead\n labels = kmeans.labels_\n \n labelCounts = Counter(labels)\n minLabel = min(labelCounts, key=labelCounts.get)\n labels = ['1' if x == minLabel else '0' for x in labels]\n labels = [eval(x) for x in labels]\n labels = np.array(labels, ndmin=2)\n \n return [X, labels]\n"},{"alpha_fraction":0.6721804738044739,"alphanum_fraction":0.6827067732810974,"avg_line_length":16.972972869873047,"blob_id":"feca9bd17df0f29a1e5cb57306497076c663157d","content_id":"29d5276f4eddeedd6e0e60a5b60961e973ddda45","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Java","length_bytes":665,"license_type":"no_license","max_line_length":62,"num_lines":37,"path":"/src/sampling/heatbugs/run_heatbugs.java","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import heatbugs.HeatBugs;\nimport sim.engine.*;\n\npublic class run_heatbugs extends SimState\n{\n\tpublic run_heatbugs(long seed)\n\t{\n\t\tsuper(seed);\t\n\t}\n\n\tpublic static void main(String[] args)\n\t{\n\t\t// initialize\n\t\tSimState state = new HeatBugs(System.currentTimeMillis());\n\t\t\n\t\t// set parameters\n\t\tHeatBugs heatBugs = (HeatBugs)state;\n\t\theatBugs.setDiffusionConstant(0.0);\n\n\t\t// run\n\t\tstate.start();\n\n\t\tdo\n\t\t{\n\t\t\tif (!state.schedule.step(state)) \n\t\t\t\tbreak;\n\t\t}\n\t\twhile(state.schedule.getSteps() < 5000);\n\t\t\n\t\t// save state as image\n\t\t// get some state information\n\t\tSystem.out.println(heatBugs.getRandomMovementProbability());\n\n\t\tstate.finish();\n\t\tSystem.exit(0);\n\t}\n}\n"},{"alpha_fraction":0.6717171669006348,"alphanum_fraction":0.6717171669006348,"avg_line_length":27.285715103149414,"blob_id":"024ca2f638c42bb5954b9c78058c458f022b7583","content_id":"2d8eb79139ebed4a23ead99f5fa490eaf658bb34","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":198,"license_type":"no_license","max_line_length":74,"num_lines":7,"path":"/snippets/plot_error.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import numpy as np\n\nwith open(\"../src/suggested_alps_fm_prediction_error.txt\", \"r\") as inFile:\n errors = [eval(x) for x in inFile.readlines()]\n\n print np.mean(errors)\n print np.std(errors)\n"},{"alpha_fraction":0.62347012758255,"alphanum_fraction":0.6335493326187134,"avg_line_length":26.50494956970215,"blob_id":"4299bab84a5f3e6c72a9b75a73bf215dd333796e","content_id":"39d4907cc464658a0fd1a679c1ee7b0268a5acea","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2778,"license_type":"no_license","max_line_length":105,"num_lines":101,"path":"/app/app_server.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import os\nimport base64\nimport glob\n\nfrom flask import Flask, render_template, request\nfrom werkzeug import secure_filename\n\n\nUPLOAD_FOLDER = \"sketches/\"\nUPLOAD_FILENAME = \"demonstration.png\"\nMATCH_FOLDER = \"matches/\"\nABM = \"flocking\"\n\napp = Flask(__name__)\napp.config[\"UPLOAD_FOLDER\"] = UPLOAD_FOLDER\n\n\n@app.route('/')\ndef root():\n return \"You are here: /\"\n\n\n@app.route(\"/index\")\ndef index():\n return render_template(\"index.html\")\n\n@app.route(\"/upload\", methods=[\"GET\", \"POST\"])\ndef upload_file():\n\n if (request.method == \"POST\"):\n # get the base64 encoded image\n encodedImage = request.form[\"encodedImage\"]\n\n # save to disk\n with open(\"demonstration.jpg\", \"wb\") as outFile:\n outFile.write(base64.b64decode(encodedImage))\n\n # convert to png\n os.system(\"convert demonstration.jpg demonstration.png\")\n os.system(\"rm -f demonstration.jpg\")\n \n # clear old data\n os.system(\"rm -f \" + UPLOAD_FOLDER + \"*\")\n os.system(\"rm -f \" + MATCH_FOLDER + \"*\")\n\n # save file in selected location\n fileName = os.path.join(app.config[\"UPLOAD_FOLDER\"], UPLOAD_FILENAME)\n os.system(\"mv demonstration.png \" + fileName)\n\n # initiate matching and call pipeline background process\n os.system(\"rm -rf \" + \"../data/predictions/\" + ABM + \"/\" + MATCH_FOLDER + \"/*\")\n os.system('curl localhost:5001/match')\n os.system(\"./do_pipeline.sh &\")\n \n # refresh will be a get request\n return render_template(\"redirect.html\")\n\n\n@app.route(\"/redirect\")\ndef redirect():\n # get background process status\n fileExists = os.path.isfile(\"../data/predictions/\" + ABM + \"/\" + MATCH_FOLDER + \"predicted_alps.txt\")\n\n return str(fileExists)\n\n\n@app.route('/results')\ndef results():\n return render_template(\"results.html\")\n\n\n@app.route('/alps')\ndef get_suggested_alps():\n # get suggested ALPs from file\n # this will be the response sent\n with open(\"../data/predictions/\" + ABM + \"/\" + MATCH_FOLDER + \"predicted_alps.txt\") as inFile:\n prediction = inFile.readlines()[0]\n suggestedAlps = prediction.split(\" (\")[0]\n\n return suggestedAlps\n\n\n@app.route('/match')\ndef get_match():\n # return matched image as json\n imageName = glob.glob(\"matches/*.png\")[0]\n with open(imageName, \"rb\") as imageFile:\n imageString = base64.b64encode(imageFile.read())\n\n return imageString\n\n\n@app.route('/prediction')\ndef get_prediction():\n # return matched image as json\n imageFolder = glob.glob(\"../data/predictions/\" + ABM + \"/\" + MATCH_FOLDER + \"/images_*\")[0]\n imageName = imageFolder + \"/0_200.png\"\n with open(imageName, \"rb\") as imageFile:\n imageString = base64.b64encode(imageFile.read())\n\n return imageString\n"},{"alpha_fraction":0.729411780834198,"alphanum_fraction":0.729411780834198,"avg_line_length":16,"blob_id":"9b5e514746b13a44a2de2147a8dee5e93c9a49cf","content_id":"c241871ffcbca08fc11761bb6eb8677ac11fe5ed","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":85,"license_type":"no_license","max_line_length":25,"num_lines":5,"path":"/src/sampling/sampling.sh","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"# generate raw file\n#./sampling_raw.sh\n\n# process raw file\n./sampling_process_raw.sh\n"},{"alpha_fraction":0.6430445909500122,"alphanum_fraction":0.654199481010437,"avg_line_length":42.485713958740234,"blob_id":"5b7dd6773462dab1fca49591af102fe140960d1b","content_id":"dda6cd16ec97f0010a4b62e76231ba9d1702d50f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1524,"license_type":"no_license","max_line_length":120,"num_lines":35,"path":"/data/dataset_analysis/simulations/plot_fm_prediction_error.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import matplotlib.pyplot as plt\nimport numpy as np\n\npointSelectionMethods = [\"error_weighted\", \"min_error\"]\ndemonstrations = [\"1\", \"2\", \"3\"]\n\n# plot fm prediction error for various demonstrations\nfor pointSelectionMethod in pointSelectionMethods:\n # plot unguided and guided data\n plt.figure()\n\n print(pointSelectionMethod)\n \n for demonstration in demonstrations:\n with open(pointSelectionMethod + \"/\" + demonstration + \"/unguided/fm_prediction_error.txt\") as unguidedDataFile:\n unguidedData = [eval(x) for x in unguidedDataFile.readlines()]\n with open(pointSelectionMethod + \"/\" + demonstration + \"/guided/fm_prediction_error.txt\") as guidedDataFile:\n guidedData = [eval(x) for x in guidedDataFile.readlines()]\n\n mean = np.mean(unguidedData)\n sigma = np.std(unguidedData)\n plot1 = plt.errorbar(eval(demonstration), mean, sigma, fmt='s', color='b', markersize=20, elinewidth=2)\n print(mean, sigma)\n \n mean = np.mean(guidedData)\n sigma = np.std(guidedData)\n plot2 = plt.errorbar(eval(demonstration), mean, sigma, fmt='s', color='g', markersize=20, elinewidth=2)\n print(mean, sigma)\n \n plt.gca().set_ylim(bottom=-0.01)\n plt.legend([plot1, plot2], [\"Unguided Data\", \"Guided Data\"], loc=\"best\")\n plt.xticks(range(len(demonstrations)+2), [\" \", \"a\", \"b\", \"c\", \" \"])\n plt.xlabel(\"Demonstration (Demo)\")\n plt.ylabel(\"FM Prediction Error\")\n plt.savefig(pointSelectionMethod + \".png\")\n\n\n"},{"alpha_fraction":0.628926694393158,"alphanum_fraction":0.6511780023574829,"avg_line_length":23.253969192504883,"blob_id":"303c78a0e88e4f0ff01976c0e5b4273178738f70","content_id":"8be2c577064da046efa5eff6857b099eb04f8b71","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1528,"license_type":"no_license","max_line_length":92,"num_lines":63,"path":"/src/sampling/forest_fire/sample_forest_fire.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"from model import ForestFire\nimport random\nimport numpy as np\n\n# only one ALP\n# density (0,1)\n\n# constant ALPs\nHEIGHT = 100\nWIDTH = 100\n\n# constants\n# max_iters in the example is specified as 1000\n# so we run for 1000 steps\nINDEPENDENT_VARIABLES = 1\nDEPENDENT_VARIABLES = 1\nRUNS = 10000\nSAMPLES = 10\nTIME_LAPSE = 200\n\ndef main_forest_fire(density=0.65):\n # instantiate and run model\n model = ForestFire(height=HEIGHT, width=WIDTH, density=density)\n\n # run time lapse\n for i in range(TIME_LAPSE):\n model.step()\n\n # collect data for next steps\n dependentValues = []\n\n for i in range(SAMPLES):\n # step\n model.step()\n\n # read data\n data = model.datacollector.get_model_vars_dataframe()\n burnedTrees = np.array(list(data.BurnedOut)[-SAMPLES-1:])\n fineTrees = np.array(list(data.Fine)[-SAMPLES-1:])\n initialTrees = burnedTrees + fineTrees\n dependentValues.append(list(burnedTrees/initialTrees))\n\n # print line corresponding to this execution\n line = str(density)\n for dependentValue in dependentValues:\n line += \" \" + str(dependentValue).replace(\"[\", \"\").replace(\"]\", \"\").replace(\" \", \"\")\n print(line)\n \ndef main():\n\n # header string\n print(\"i\"*INDEPENDENT_VARIABLES + \"d\"*DEPENDENT_VARIABLES)\n\n for run in range(RUNS):\n\n # sample random ALPs\n # cop density + citizen density should be less than 1 \n density = random.random()\n \n # run model using those ALPs\n main_forest_fire(density)\n\nmain()\n"},{"alpha_fraction":0.7006369233131409,"alphanum_fraction":0.7707006335258484,"avg_line_length":21.428571701049805,"blob_id":"3ca71683ff14d9dfb27a0017fde89a5bc7de12d7","content_id":"2ed16a83192c7e76496fc97354113e21bbd05dd0","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":157,"license_type":"no_license","max_line_length":54,"num_lines":7,"path":"/src/domainconfs/aids.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"TRAINING_SIZE = 300\nVALIDATION_SIZE = 1000\nRM_GRANULARITY = 55\nREGRESSION = \"sklearn.neighbors.KNeighborsRegressor()\"\n\nNUM_INDEPENDENT = 2\nNUM_DEPENDENT = 2\n"},{"alpha_fraction":0.6805850863456726,"alphanum_fraction":0.6828457713127136,"avg_line_length":41.97142791748047,"blob_id":"fbffb23a39ddbe55972330d45169fce39ff864ba","content_id":"03dc45db4a8abb59562161fde903a2043907ebc2","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":7520,"license_type":"no_license","max_line_length":287,"num_lines":175,"path":"/src/match_sketches.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"import os\nimport sampling.utils.featurize_image as fi\nimport sys\nimport glob\nimport amf.misc\nimport matplotlib.pyplot as plt\nfrom sklearn.neighbors import LSHForest\nimport numpy as np\nimport cPickle\nimport random\n\ndef initialize_featurization():\n # edit scale data file name based on descriptor setting\n scaleDataFileNamePrefix = os.environ[\"SCALE_DATA_FILE\"].split(\".\")[0]\n scaleDataFileNameExtension = os.environ[\"SCALE_DATA_FILE\"].split(\".\")[1]\n scaleDataFileName = scaleDataFileNamePrefix + \"_\" + os.environ[\"ABM\"]\n if(bool(eval(os.environ[\"IMAGES\"])) == True):\n scaleDataFileName += \"_\" + os.environ[\"DESCRIPTOR_SETTING\"].lower()\n scaleDataFileName += \".\" + scaleDataFileNameExtension\n\n # get scale information\n with open(\"sampling/\" + os.environ[\"ABM\"] + \"/\" + scaleDataFileName, \"r\") as scaleDataFile:\n # read in order of writing in map.py\n maxs = eval(scaleDataFile.readline().strip(\"\\n\"))\n mins = eval(scaleDataFile.readline().strip(\"\\n\"))\n\n dependentMaxs = maxs[-eval(os.environ[\"NUM_DEPENDENT\"]):]\n dependentMins = mins[-eval(os.environ[\"NUM_DEPENDENT\"]):]\n dependentValuesMax = {i: dependentMaxs[i] for i in range(len(dependentMaxs))}\n dependentValuesMin = {i: dependentMins[i] for i in range(len(dependentMins))}\n scaleData = [dependentValuesMax, dependentValuesMin]\n\n \n # load feature extractor if needed\n net = fi.get_network(os.environ[\"DESCRIPTOR_SETTING\"])\n\n return [net, scaleData]\n\ndef get_dependent_values_image(net, scaleData, imageName):\n # define dependent values\n dependentValuesImage = fi.get_features(net, imageName, os.environ[\"DESCRIPTOR_SETTING\"])\n # normalize\n [dependentValuesMax, dependentValuesMin] = scaleData\n dependentValuesImage = tuple((dependentValuesImage[index] - dependentValuesMin[index])/(dependentValuesMax[index] - dependentValuesMin[index]) if (dependentValuesMax[index] > dependentValuesMin[index]) else dependentValuesMax[index] for index in range(len(dependentValuesImage)))\n\n return dependentValuesImage\n\ndef get_dependent_values_folder(net, scaleData, folderName, downsample=1):\n # initialize collection of dependent values\n imageNamesFolder = []\n dependentValuesFolder = []\n \n # expect everything in this folder to be an image\n # read sketches data\n allFileNames = glob.glob(folderName + \"/*.*\")\n allFileNames = random.sample(allFileNames, int(len(allFileNames)/downsample))\n for imageName in sorted(allFileNames):\n # featurize image\n print >> sys.stderr, (\"now processing \" + imageName) \n dependentValuesImage = get_dependent_values_image(net, scaleData, imageName)\n\n # append to list\n imageNamesFolder.append(imageName)\n dependentValuesFolder.append(dependentValuesImage)\n\n return [imageNamesFolder, dependentValuesFolder]\n\ndef do_matching_lsh(net, scaleData, imageNamesSketchList, dependentValuesSketchList):\n lshModelFileName = \"../../_swarm-lfd-data/\" + os.environ[\"ABM\"] + \"/lsh_forest_model\" + \"_\" + os.environ[\"DESCRIPTOR_SETTING\"].lower() + \".p\"\n\n if(os.path.exists(lshModelFileName) == False):\n # no model found in directory\n # train a locality sensitive hashing model on the simulation data\n [imageNamesSimulationList, dependentValuesSimulationList] = get_dependent_values_folder(net, scaleData, \"../../_swarm-lfd-data/\" + os.environ[\"ABM\"] + \"/images\")\n lshForest = LSHForest()\n lshForest.fit(dependentValuesSimulationList)\n\n # use disk to cache results\n with open(lshModelFileName, \"wb\") as outFile:\n cPickle.dump(lshForest, outFile)\n with open(\"image_names_simulation.p\", \"wb\") as outFile:\n cPickle.dump(imageNamesSimulationList, outFile)\n else:\n # use existing model\n with open(lshModelFileName, \"rb\") as inFile:\n lshForest = cPickle.load(inFile)\n with open(\"../../_swarm-lfd-data/\" + os.environ[\"ABM\"] + \"/image_names_simulation.p\", \"rb\") as inFile:\n imageNamesSimulationList = cPickle.load(inFile)\n\n # now match\n print >> sys.stderr, \"\\n\"\n print >> sys.stderr, \"now matching\"\n matches = []\n for imageNameSketch, dependentValuesSketch in zip(imageNamesSketchList, dependentValuesSketchList):\n minDistance = float('Inf')\n minDistanceImage = None\n distances = []\n\n # use lsh to get neighbors\n [distances, indices] = lshForest.kneighbors(np.array(dependentValuesSketch).reshape(1, -1))\n\n # use indices to get closest images\n # read image name and image feature files\n # then only compare with the few retrieved\n # this is only useful if LSH features are different from features used in the loop\n for index in indices[0]:\n imageNameSimulation = imageNamesSimulationList[index]\n dependentValuesSimulation = get_dependent_values_image(net, scaleData, imageNameSimulation)\n\n # compute distance\n distance = amf.misc.distance(dependentValuesSketch, dependentValuesSimulation)\n if(distance < minDistance):\n minDistance = distance\n minDistanceImage = imageNameSimulation\n\n # pick minimum distance\n matches.append((imageNameSketch, minDistanceImage))\n\n # clear memory\n del lshForest\n\n return matches\n\ndef do_matching(net, scaleData, imageNamesSketchList, dependentValuesSketchList):\n\n # get raw features\n # [imageNamesSimulationList, dependentValuesSimulationList] = get_dependent_values_folder(net, scaleData, \"../../_swarm-lfd-data/\" + os.environ[\"ABM\"] + \"/images\", downsample=11000)\n [imageNamesSimulationList, dependentValuesSimulationList] = get_dependent_values_folder(net, scaleData, \"../../_swarm-lfd-data/\" + os.environ[\"ABM\"] + \"/images_1000\")\n\n # now match\n print >> sys.stderr, \"\\n\"\n print >> sys.stderr, \"now matching\"\n matches = []\n for imageNameSketch, dependentValuesSketch in zip(imageNamesSketchList, dependentValuesSketchList):\n minDistance = float('Inf')\n minDistanceImage = None\n distances = []\n\n for imageNameSimulation, dependentValuesSimulation in zip(imageNamesSimulationList, dependentValuesSimulationList):\n # compute distance\n distance = amf.misc.distance(dependentValuesSketch, dependentValuesSimulation)\n if(distance < minDistance):\n minDistance = distance\n minDistanceImage = imageNameSimulation\n\n # pick minimum distance\n matches.append((imageNameSketch, minDistanceImage))\n\n return matches\n\ndef main():\n # get sketches folder\n sketchesFolder = sys.argv[1]\n \n # initialize and retrieve features\n [net, scaleData] = initialize_featurization()\n [imageNamesSketchList, dependentValuesSketchList] = get_dependent_values_folder(net, scaleData, sketchesFolder)\n\n # matches = do_matching_lsh(net, scaleData, imageNamesSketchList, dependentValuesSketchList)\n matches = do_matching(net, scaleData, imageNamesSketchList, dependentValuesSketchList)\n\n # print matches\n # save matches\n simulationSaveFolder = sys.argv[2]\n print >> sys.stderr, \"\\n\"\n print >> sys.stderr, \"now printing matches:\"\n print >> sys.stderr, \"\\n\"\n for x, y in matches:\n print >> sys.stderr, x\n print >> sys.stderr, y\n print >> sys.stderr, \"\\n\"\n\n os.system(\"cp \" + y + \" \" + simulationSaveFolder)\n \nmain()\n"},{"alpha_fraction":0.7051281929016113,"alphanum_fraction":0.7692307829856873,"avg_line_length":21.285715103149414,"blob_id":"1fd3ccaa98a7e798627221abe13a8c2be0c6fee9","content_id":"e335465a61bbee713df5f15a1d4610f47810a677","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":156,"license_type":"no_license","max_line_length":54,"num_lines":7,"path":"/src/domainconfs/civil_violence.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"TRAINING_SIZE = 330\nVALIDATION_SIZE = 100\nRM_GRANULARITY = 40\nREGRESSION = \"sklearn.neighbors.KNeighborsRegressor()\"\n\nNUM_INDEPENDENT = 3\nNUM_DEPENDENT = 3\n"},{"alpha_fraction":0.6403161883354187,"alphanum_fraction":0.7351778745651245,"avg_line_length":22,"blob_id":"02ab3a125819d197a0b3c5083757951007e66128","content_id":"effe63c2db702f02d9b6fa78d4ead36edce8b704","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":253,"license_type":"no_license","max_line_length":54,"num_lines":11,"path":"/src/domainconfs/wolf_sheep.py","repo_name":"karanbudhraja/PABMD","src_encoding":"UTF-8","text":"#TRAINING_SIZES = range(40, 440, 40)\n#VALIDATION_SIZE = 200\n#RM_GRANULARITIES = range(15, 65, 10)\n\nTRAINING_SIZE = 40\nVALIDATION_SIZE = 10\nRM_GRANULARITY = 20\nREGRESSION = \"sklearn.neighbors.KNeighborsRegressor()\"\n\nNUM_INDEPENDENT = 5\nNUM_DEPENDENT = 2\n"}],"string":"[\n {\n \"alpha_fraction\": 0.7330960631370544,\n \"alphanum_fraction\": 0.7348754405975342,\n \"avg_line_length\": 20.615385055541992,\n \"blob_id\": \"5d3f87b1129a1658751e1a880d143d5ffd134a8e\",\n \"content_id\": \"0ba15ef118cabd0e3e620163c13f215df0612ea4\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 562,\n \"license_type\": \"no_license\",\n \"max_line_length\": 50,\n \"num_lines\": 26,\n \"path\": \"/app/do_pipeline.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/bin/bash\\n\\n# go to target script directory\\ncd ../src\\nsource configuration.sh\\n\\n# copy to demonstration folder for lfd use\\ncd ../data/demonstrations/$ABM\\nrm -rf *\\ncp -r ../../../app/matches .\\n\\n# move to appropriate directory to run lfd\\ncd ../../../src\\n\\n# run the framework to suggest several alps\\n# that would produce the demonstration slps\\n./lfd.sh\\n\\n# simulate those alps\\n# to check the actual slps that they correspond to\\n./simulate.sh\\n\\n# filter from the suggested alps\\n# use distance from demonstration slps\\nexport USE_FILTER_DESCRIPTOR_SETTING=1\\n./filter.sh\\n\"\n },\n {\n \"alpha_fraction\": 0.7180395126342773,\n \"alphanum_fraction\": 0.7912715673446655,\n \"avg_line_length\": 89.23711395263672,\n \"blob_id\": \"4f17937b5d1023ddb3b730250401bf2783d10d39\",\n \"content_id\": \"8b149f38b51d1b6a006c27a550c274e59df63929\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 8753,\n \"license_type\": \"no_license\",\n \"max_line_length\": 155,\n \"num_lines\": 97,\n \"path\": \"/data/gifs_all/generate_gifs.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# convert demonstrations\\nconvert -delay 50 -loop 0 ../demonstrations_all/flocking/simulated/1/*.png demonstration_1.gif\\nconvert -delay 50 -loop 0 ../demonstrations_all/flocking/simulated/2/*.png demonstration_2.gif\\nconvert -delay 50 -loop 0 ../demonstrations_all/flocking/simulated/3/*.png demonstration_3.gif\\n\\nconvert -delay 50 -loop 0 ../demonstrations_/flocking/1/*.png demonstration__1.gif\\nconvert -delay 50 -loop 0 ../demonstrations_/flocking/2/*.png demonstration__2.gif\\nconvert -delay 50 -loop 0 ../demonstrations_/flocking/3/*.png demonstration__3.gif\\n\\n# convert rotated images\\nconvert -rotate 90 demonstration_1.gif demonstration_1_rotation.gif\\n\\n# convert translated images\\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_200.png -background black -flatten 1108_200_translation.png\\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_201.png -background black -flatten 1108_201_translation.png\\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_202.png -background black -flatten 1108_202_translation.png\\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_203.png -background black -flatten 1108_203_translation.png\\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_204.png -background black -flatten 1108_204_translation.png\\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_205.png -background black -flatten 1108_205_translation.png\\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_206.png -background black -flatten 1108_206_translation.png\\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_207.png -background black -flatten 1108_207_translation.png\\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_208.png -background black -flatten 1108_208_translation.png\\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_209.png -background black -flatten 1108_209_translation.png\\nconvert -page +100+0 ../demonstrations_all/flocking/simulated/1/1108_210.png -background black -flatten 1108_210_translation.png\\nconvert +repage 1108_200_translation.png 1108_200_translation.png\\nconvert +repage 1108_201_translation.png 1108_201_translation.png\\nconvert +repage 1108_202_translation.png 1108_202_translation.png\\nconvert +repage 1108_203_translation.png 1108_203_translation.png\\nconvert +repage 1108_204_translation.png 1108_204_translation.png\\nconvert +repage 1108_205_translation.png 1108_205_translation.png\\nconvert +repage 1108_206_translation.png 1108_206_translation.png\\nconvert +repage 1108_207_translation.png 1108_207_translation.png\\nconvert +repage 1108_208_translation.png 1108_208_translation.png\\nconvert +repage 1108_209_translation.png 1108_209_translation.png\\nconvert +repage 1108_210_translation.png 1108_210_translation.png\\nconvert -delay 50 -loop 0 *_translation.png demonstration_1_translation.gif\\nrm *_translation.png\\n\\n# convert point selection methods\\nconvert -delay 50 -loop 0 ../predictions_all/point\\\\ selection/predictions_error_weighted/flocking/1/*.png point\\\\ selection/error_weighted_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/point\\\\ selection/predictions_error_weighted/flocking/2/*.png point\\\\ selection/error_weighted_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/point\\\\ selection/predictions_error_weighted/flocking/3/*.png point\\\\ selection/error_weighted_3.gif\\n\\nconvert -delay 50 -loop 0 ../predictions_all/point\\\\ selection/predictions_kmeans/flocking/1/*.png point\\\\ selection/kmeans_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/point\\\\ selection/predictions_kmeans/flocking/2/*.png point\\\\ selection/kmeans_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/point\\\\ selection/predictions_kmeans/flocking/3/*.png point\\\\ selection/kmeans_3.gif\\n\\nconvert -delay 50 -loop 0 ../predictions_all/point\\\\ selection/predictions_mean/flocking/1/*.png point\\\\ selection/mean_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/point\\\\ selection/predictions_mean/flocking/2/*.png point\\\\ selection/mean_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/point\\\\ selection/predictions_mean/flocking/3/*.png point\\\\ selection/mean_3.gif\\n\\nconvert -delay 50 -loop 0 ../predictions_all/point\\\\ selection/predictions_min_error/flocking/1/*.png point\\\\ selection/min_error_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/point\\\\ selection/predictions_min_error/flocking/2/*.png point\\\\ selection/min_error_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/point\\\\ selection/predictions_min_error/flocking/3/*.png point\\\\ selection/min_error_3.gif\\n\\n# convert feature construction methods\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_contour/flocking/1/*.png feature\\\\ construction/contour_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_contour/flocking/2/*.png feature\\\\ construction/contour_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_contour/flocking/3/*.png feature\\\\ construction/contour_3.gif\\n\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_mnist/flocking/1/*.png feature\\\\ construction/mnist_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_mnist/flocking/2/*.png feature\\\\ construction/mnist_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_mnist/flocking/3/*.png feature\\\\ construction/mnist_3.gif\\n\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_vae/flocking/1/*.png feature\\\\ construction/vae_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_vae/flocking/2/*.png feature\\\\ construction/vae_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_vae/flocking/3/*.png feature\\\\ construction/vae_3.gif\\n\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_vgg16/flocking/1/*.png feature\\\\ construction/vgg16_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_vgg16/flocking/2/*.png feature\\\\ construction/vgg16_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_vgg16/flocking/3/*.png feature\\\\ construction/vgg16_3.gif\\n\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_resnet50/flocking/1/*.png feature\\\\ construction/resnet50_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_resnet50/flocking/2/*.png feature\\\\ construction/resnet50_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_resnet50/flocking/3/*.png feature\\\\ construction/resnet50_3.gif\\n\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_lstm/flocking/1/*.png feature\\\\ construction/lstm_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_lstm/flocking/2/*.png feature\\\\ construction/lstm_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_lstm/flocking/3/*.png feature\\\\ construction/lstm_3.gif\\n\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_lbp/flocking/1/*.png feature\\\\ construction/lbp_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_lbp/flocking/2/*.png feature\\\\ construction/lbp_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/feature\\\\ construction/predictions_lbp/flocking/3/*.png feature\\\\ construction/lbp_3.gif\\n\\n# convert dataset analysis methods\\nconvert -delay 50 -loop 0 ../predictions_all/dataset\\\\ analysis/error_weighted/1/unguided/predictions_best/flocking/1/*.png dataset\\\\ analysis/unguided_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/dataset\\\\ analysis/error_weighted/2/unguided/predictions_best/flocking/2/*.png dataset\\\\ analysis/unguided_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/dataset\\\\ analysis/error_weighted/3/unguided/predictions_best/flocking/3/*.png dataset\\\\ analysis/unguided_3.gif\\n\\nconvert -delay 50 -loop 0 ../predictions_all/dataset\\\\ analysis/error_weighted/1/guided/predictions_best/flocking/1/*.png dataset\\\\ analysis/guided_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/dataset\\\\ analysis/error_weighted/2/guided/predictions_best/flocking/2/*.png dataset\\\\ analysis/guided_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/dataset\\\\ analysis/error_weighted/3/guided/predictions_best/flocking/3/*.png dataset\\\\ analysis/guided_3.gif\\n\\n# convert filtering method\\nconvert -delay 50 -loop 0 ../predictions_all/pruning\\\\ od/flocking/1/*.png pruning\\\\ od/pruning_od_1.gif\\nconvert -delay 50 -loop 0 ../predictions_all/pruning\\\\ od/flocking/2/*.png pruning\\\\ od/pruning_od_2.gif\\nconvert -delay 50 -loop 0 ../predictions_all/pruning\\\\ od/flocking/3/*.png pruning\\\\ od/pruning_od_3.gif\\n\"\n },\n {\n \"alpha_fraction\": 0.6934612989425659,\n \"alphanum_fraction\": 0.6946610808372498,\n \"avg_line_length\": 28.76785659790039,\n \"blob_id\": \"f71261bde834b7e609f89053df97ec68ea16d57a\",\n \"content_id\": \"d6b7077a1fe8f912a5db018ef2b0b7f084d62254\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 1667,\n \"license_type\": \"no_license\",\n \"max_line_length\": 101,\n \"num_lines\": 56,\n \"path\": \"/src/simulate.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# load configuration\\nsource configuration.sh\\n\\nexport SIMULATION_SCRIPT=\\\"sampling/\\\"$ABM\\\"/predict_\\\"$ABM$SIMULATION_EXTENSION\\n\\n# remove existing data\\nrm -f suggested_slps.txt\\nrm -rf sampling/$ABM/images_*\\n\\n# for line in file\\nwhile read -r line\\ndo\\n # read the number of alps per line\\n export ALPS_PER_LINE=$(echo $line | tr -cd \\\")\\\" | wc -c)\\n \\n # remove all spaces in line\\n # assume just one configuration per line, for now\\n # fix this later by adding a loop\\n export line=$(echo $line | sed \\\"s/ //g\\\")\\n\\n # simulate using suggested alps\\n export SLP_CONFIGURATIONS=$($SIMULATION_TOOL $SIMULATION_SCRIPT $line)\\n\\n # in case of images, slp configurations are not returned\\n # compute slp configuration from images\\n if [ $IMAGES == 1 ]; then\\n\\t# all folders\\n\\texport SLP_CONFIGURATIONS=\\\"[\\\"\\n\\t\\n\\t# get the last image folder\\n\\t# this is the most recently created one\\n\\t# get as many lines as the number of alps in a row of suggested_alps.txt\\n\\texport SLP_IMAGES_FOLDERS=$(ls \\\"sampling/\\\"$ABM | sort | grep images_ | tail -n $ALPS_PER_LINE)\\n\\n\\tfor folder in $SLP_IMAGES_FOLDERS\\n\\tdo\\n\\t # for each folder\\n\\t export SLP_CONFIGURATION=$(python sampling/utils/process_folder_images.py sampling/$ABM/$folder)\\n\\t export CHARACTER_COUNT=$(echo $SLP_CONFIGURATIONS | wc -c)\\n\\n\\t if [ $CHARACTER_COUNT == 2 ]; then\\n\\t\\t# skip first occurrence\\n\\t\\texport SEPARATOR=\\\"\\\"\\n\\t else\\n\\t\\texport SEPARATOR=\\\", \\\"\\n\\t fi\\n\\t \\n\\t export SLP_CONFIGURATIONS=$SLP_CONFIGURATIONS$SEPARATOR$SLP_CONFIGURATION\\n\\tdone\\n\\n\\t# add closing bracket\\n\\texport SLP_CONFIGURATIONS=$SLP_CONFIGURATIONS\\\"]\\\"\\n fi\\n \\n echo $SLP_CONFIGURATIONS >> suggested_slps.txt\\ndone < suggested_alps.txt\\n\"\n },\n {\n \"alpha_fraction\": 0.6227115988731384,\n \"alphanum_fraction\": 0.6569381952285767,\n \"avg_line_length\": 34.224300384521484,\n \"blob_id\": \"141089750c1e1d743522c8da4b7ca8169c5e5edd\",\n \"content_id\": \"30743688501ce491903eea44f95e83b89162b92f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3769,\n \"license_type\": \"no_license\",\n \"max_line_length\": 111,\n \"num_lines\": 107,\n \"path\": \"/src/sampling/utils/keras/conv_lstm.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\\"\\\"\\\" This script demonstrates the use of a convolutional LSTM network.\\nThis network is used to predict the next frame of an artificially\\ngenerated movie which contains moving squares.\\n\\\"\\\"\\\"\\nfrom keras.models import Sequential\\nfrom keras.layers.convolutional import Convolution3D\\nfrom keras.layers.convolutional_recurrent import ConvLSTM2D\\nfrom keras.layers.normalization import BatchNormalization\\nfrom keras.layers import Dense, Activation\\nfrom keras.layers.core import Reshape, Flatten\\nimport numpy as np\\nimport pylab as plt\\n\\nfrom skimage.io import imread\\nfrom skimage.transform import resize\\n\\n# We create a layer which take as input movies of shape\\n# (n_frames, width, height, channels) and returns a movie\\n# of identical shape.\\n\\nseq = Sequential()\\n\\nseq.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\\n input_shape=(None, 40, 40, 1),\\n batch_input_shape=(None, 11, 40, 40, 1),\\n border_mode='same', return_sequences=True))\\nseq.add(BatchNormalization())\\nseq.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\\n border_mode='same', return_sequences=True))\\nseq.add(BatchNormalization())\\n\\nseq.add(Flatten())\\nseq.add(Dense(128))\\nseq.add(Dense(1600*11))\\nseq.add(Reshape((11, 40, 40, 1)))\\n\\nseq.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\\n border_mode='same', return_sequences=True))\\nseq.add(BatchNormalization())\\nseq.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\\n border_mode='same', return_sequences=True))\\nseq.add(BatchNormalization())\\nseq.add(Convolution3D(nb_filter=1, kernel_dim1=1, kernel_dim2=3,\\n kernel_dim3=3, activation='sigmoid',\\n border_mode='same', dim_ordering='tf'))\\nseq.compile(loss='binary_crossentropy', optimizer='adadelta')\\n\\ndef image_data(file):\\n img = imread(file, as_grey=True)\\n img = resize(img, (40, 40))\\n return img\\n\\n# Artificial data generation:\\n# Generate movies with 3 to 7 moving squares inside.\\n# The squares are of shape 1x1 or 2x2 pixels,\\n# which move linearly over time.\\n# For convenience we first create movies with bigger width and height (80x80)\\n# and at the end we select a 40x40 window.\\n\\ndef generate_movies(n_samples=1200, n_frames=11, dataFolder=None):\\n row = 40\\n col = 40\\n \\n noisy_movies = np.zeros((n_samples, n_frames, row, col, 1), dtype=np.float)\\n shifted_movies = np.zeros((n_samples, 11, row, col, 1), dtype=np.float)\\n\\n for i in range(n_samples):\\n # read custom dataset\\n for t in range(n_frames):\\n # read and resize to 40x40\\n x = image_data(dataFolder + \\\"/\\\" + str(i) + \\\"_\\\" + str(200+t) + \\\".png\\\")\\n shifted_movies[i, t, :, :, 0] = x\\n \\n # figure out how to add noise to this data\\n noise_f = (-1)**np.random.randint(0, 2, size=(40,40))\\n noisy_movies[i, t, :, :, 0] = np.dot(x, noise_f)\\n \\n return noisy_movies, shifted_movies\\n\\n# Train the network\\nnoisy_movies, shifted_movies = generate_movies(n_samples=10, dataFolder=\\\"../../../../_swarm-lfd-data/Flocking\\\")\\n\\n# original nb_epoch = 500\\n\\nseq.fit(noisy_movies, shifted_movies, batch_size=10,\\n nb_epoch=1, validation_split=0.05)\\n\\n# save model as JSON\\nmodel = seq\\nmodelAsJson = model.to_json()\\nwith open(\\\"conv_lstm_model.json\\\", \\\"w\\\") as jsonFile:\\n jsonFile.write(modelAsJson)\\n# serialize weights to HDF5\\nmodel.save_weights('conv_lstm_model.h5')\\n\\n'''\\n# Testing the network on one movie\\n# feed it with the first 7 positions and then\\n# predict the new positions\\nwhich = 1004\\ntrack = noisy_movies[which][:7, ::, ::, ::]\\n\\nfor j in range(16):\\n new_pos = seq.predict(track[np.newaxis, ::, ::, ::, ::])\\n new = new_pos[::, -1, ::, ::, ::]\\n track = np.concatenate((track, new), axis=0)\\n'''\\n\"\n },\n {\n \"alpha_fraction\": 0.6192708611488342,\n \"alphanum_fraction\": 0.6348958611488342,\n \"avg_line_length\": 36.28155517578125,\n \"blob_id\": \"4d28b5aed7764cedb4a8ac825ad0c4d186e0922b\",\n \"content_id\": \"06d75d7b7bf87ef7eb4643f1dcd0b97d34278821\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3840,\n \"license_type\": \"no_license\",\n \"max_line_length\": 110,\n \"num_lines\": 103,\n \"path\": \"/snippets/flocking/plot_suggested_slps.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import matplotlib.pyplot as plt\\nimport numpy as np\\n\\n#demonstrations = [\\\"a\\\", \\\"b\\\", \\\"c\\\"]\\n#demonstrations = [\\\"_a\\\", \\\"_b\\\", \\\"_c\\\"]\\n#demonstrations = [\\\"a_filtered\\\", \\\"b_filtered\\\", \\\"c_filtered\\\"]\\n#demonstrations = [\\\"_a\\\"]\\n#demonstrations = [\\\"a\\\"]\\ndemonstrations = [\\\"a_filtered\\\"]\\n\\n#folderName = \\\"error_weighted\\\"\\nfolderName = \\\"min_error\\\"\\n\\nnumberOfBins = 10\\n\\nwith open(\\\"../../src/sampling/flocking/scale_data_flocking_contour.txt\\\", \\\"r\\\") as inFile:\\n allLines = inFile.readlines()\\n maxValueArea = eval(allLines[0])[-2]\\n maxValuePerimeter = eval(allLines[0])[-1]\\n minValueArea = eval(allLines[1])[-2]\\n minValuePerimeter = eval(allLines[1])[-1]\\n\\nfor index, item in enumerate(demonstrations):\\n allPointsX = []\\n allPointsY = []\\n errors = []\\n\\n # scatter\\n \\n plt.figure()\\n with open(folderName + \\\"/suggested_slps_\\\" + item + \\\".txt\\\", \\\"r\\\") as inFile:\\n for line in inFile:\\n point = eval(line)[0]\\n plt.scatter(point[0], point[1], c=\\\"b\\\", marker=\\\"o\\\", alpha=0.3)\\n allPointsX.append(point[0])\\n allPointsY.append(point[1])\\n\\n # compute error\\n with open(\\\"demonstration_slps.txt\\\", \\\"r\\\") as inFile:\\n demonstration = list(eval(inFile.readlines()[index]))\\n demonstration[0] = (demonstration[0]-minValueArea)/(maxValueArea-minValueArea)\\n demonstration[1] = (demonstration[1]-minValuePerimeter)/(maxValuePerimeter-minValuePerimeter)\\n\\n _point = [x for x in point]\\n _point[0] = (_point[0]-minValueArea)/(maxValueArea-minValueArea)\\n _point[1] = (_point[1]-minValuePerimeter)/(maxValuePerimeter-minValuePerimeter)\\n \\n error = (_point[0]-demonstration[0])**2 + (_point[1]-demonstration[1])**2\\n errors.append(error)\\n \\n plt.scatter(point[0], point[1], c=\\\"b\\\", marker=\\\"o\\\", alpha=0.3, label=\\\"Suggested SLPs\\\")\\n\\n with open(\\\"demonstration_slps.txt\\\", \\\"r\\\") as inFile:\\n point = eval(inFile.readlines()[index])\\n plt.scatter(point[0], point[1], c=\\\"r\\\", marker=\\\"s\\\", alpha=1.0, label=\\\"Demonstration SLPs\\\")\\n\\n plt.scatter(np.mean(allPointsX), np.mean(allPointsY), c=\\\"g\\\", marker=\\\"s\\\", alpha=1.0, label=\\\"Mean SLPs\\\")\\n\\n print np.mean(errors)\\n print np.std(errors)\\n \\n ax = plt.gca()\\n\\n ticks = ax.get_xticks()\\n minValue = minValueArea\\n maxValue = maxValueArea\\n scaledTicks = [(value-minValue)/(maxValue-minValue) for value in ticks]\\n scaledTicks = [int(value*100)/100.0 for value in scaledTicks]\\n ax.set_xticklabels(scaledTicks)\\n \\n ticks = ax.get_yticks()\\n minValue = minValuePerimeter\\n maxValue = maxValuePerimeter\\n scaledTicks = [(value-minValue)/(maxValue-minValue) for value in ticks]\\n scaledTicks = [int(value*100)/100.0 for value in scaledTicks]\\n ax.set_yticklabels(scaledTicks)\\n \\n plt.xlabel(\\\"Maximum Area\\\")\\n plt.ylabel(\\\"Maximum Perimeter\\\")\\n plt.legend()\\n plt.savefig(folderName + \\\"/suggested_slps_\\\" + item)\\n\\n # histogram\\n\\n plt.figure()\\n allPointsX = [(x-minValueArea)/(maxValueArea-minValueArea) for x in allPointsX]\\n allPointsY = [(y-minValuePerimeter)/(maxValuePerimeter-minValuePerimeter) for y in allPointsY] \\n histogram, xEdges, yEdges = np.histogram2d(allPointsX, allPointsY, bins=numberOfBins, range=[[0,1],[0,1]])\\n plt.imshow(histogram, interpolation='nearest', origin='low')\\n plt.ylabel(\\\"Maximum Area\\\")\\n plt.xlabel(\\\"Maximum Perimeter\\\")\\n plt.colorbar(label=\\\"Number of Points\\\")\\n ax = plt.gca()\\n\\n # works for normalized values\\n ticks = ax.get_xticks()\\n ticks = [value/numberOfBins for value in ticks]\\n ax.set_xticklabels(ticks)\\n ticks = ax.get_yticks()\\n ticks = [value/numberOfBins for value in ticks]\\n ax.set_yticklabels(ticks)\\n\\n plt.savefig(folderName + \\\"/histogram_suggested_slps_\\\" + item)\\n\"\n },\n {\n \"alpha_fraction\": 0.7653061151504517,\n \"alphanum_fraction\": 0.7653061151504517,\n \"avg_line_length\": 22.058822631835938,\n \"blob_id\": \"8a08b6001be36accd0cab4f82e0c7fcf76ae8c2b\",\n \"content_id\": \"35b1b2cf3d2faa900f67278973de6649044f2377\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 392,\n \"license_type\": \"no_license\",\n \"max_line_length\": 51,\n \"num_lines\": 17,\n \"path\": \"/src/all.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# this script does everything\\n# it is for reference of order for running scripts \\n\\n# sample alp, slp pairs\\n./sampling.sh\\n\\n# run the framework to suggest several alps\\n# that would produce the demonstration slps\\n./lfd.sh\\n\\n# simulate those alps\\n# to check the actual slps that they correspond to\\n./simulate.sh\\n\\n# filter from the suggested alps\\n# use distance from demonstration slps\\n./filter.sh\\n\"\n },\n {\n \"alpha_fraction\": 0.6927609443664551,\n \"alphanum_fraction\": 0.6944444179534912,\n \"avg_line_length\": 27.285715103149414,\n \"blob_id\": \"1f367beb221899b69dff768fa432444a241ee570\",\n \"content_id\": \"26d57a014a12e3080863e4b1531d5912ad5cf827\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1188,\n \"license_type\": \"no_license\",\n \"max_line_length\": 81,\n \"num_lines\": 42,\n \"path\": \"/src/sampling/utils/process_folder_images.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import sys\\nimport featurize_image as fi\\nimport glob\\nimport os\\nimport numpy as np\\n\\n# process a single folder\\ndef process_folder(folderName):\\n # get environment variables\\n descriptorSetting = os.environ[\\\"DESCRIPTOR_SETTING\\\"]\\n numDependent = eval(os.environ[\\\"NUM_DEPENDENT\\\"])\\n\\n # load feature extractor if needed\\n net = fi.get_network(descriptorSetting)\\n \\n # initialize collection of dependent values\\n dependentValuesFolder = []\\n\\n # expect everything in this folder to be an image\\n for imageName in glob.glob(folderName + \\\"/*.*\\\"):\\n # read image\\n print >> sys.stderr, (\\\"now processing \\\" + imageName)\\n \\n # define dependent values\\n dependentValuesImage = fi.get_features(net, imageName, descriptorSetting)\\n\\n # append to list\\n dependentValuesFolder.append(dependentValuesImage)\\n\\n dependentValuesFolder = np.mean(dependentValuesFolder, axis=0)\\n\\n return tuple(dependentValuesFolder)\\n\\ndef main():\\n folderName = sys.argv[1]\\n dependentValuesFolder = process_folder(folderName)\\n\\n print dependentValuesFolder\\n \\nif __name__ == \\\"__main__\\\":\\n # stuff only to run when not called via 'import' here\\n main()\\n\"\n },\n {\n \"alpha_fraction\": 0.594718337059021,\n \"alphanum_fraction\": 0.6133802533149719,\n \"avg_line_length\": 33.20481872558594,\n \"blob_id\": \"12c0aaeaf20be00b0d3d582f52f094d74f35a25c\",\n \"content_id\": \"1f4673ce4f279190e5181de29c228ab9faa40146\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2840,\n \"license_type\": \"no_license\",\n \"max_line_length\": 109,\n \"num_lines\": 83,\n \"path\": \"/data/domaindata/compute_statistics.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import numpy as np\\n\\n# results from batch data\\nRESULTS_FILE = \\\"results_flocking.csv\\\"\\nINPUT_NAME =\\\"\\\\\\\"Input.gifPrefix\\\\\\\"\\\"\\n\\ndef is_integer(x):\\n try:\\n int(x)\\n return True\\n except:\\n return False\\n\\n# main code\\nwith open(RESULTS_FILE) as results:\\n headers = results.readline()\\n\\n #print headers\\n \\n headers = headers.split(\\\",\\\")\\n inputIndex = headers.index(INPUT_NAME)\\n answer1Index = inputIndex + 1\\n answer2Index = inputIndex + 2\\n answer3Index = inputIndex + 3\\n answer4Index = inputIndex + 4\\n\\n # increment all by 3 for some reason\\n # to get correct values in data\\n inputIndex += 3\\n answer1Index += 3\\n answer2Index += 3\\n answer3Index += 3\\n answer4Index += 3\\n \\n # read result data\\n resultData = {}\\n for line in results:\\n inputData = line.split(\\\",\\\")[inputIndex]\\n answer1 = line.split(\\\",\\\")[answer1Index]\\n answer2 = line.split(\\\",\\\")[answer2Index]\\n answer3 = line.split(\\\",\\\")[answer3Index]\\n answer4 = line.split(\\\",\\\")[answer4Index]\\n\\n #print line\\n #print inputData\\n #print answer1, answer2, answer3, answer4\\n\\n if(inputData not in resultData.keys()):\\n resultData[inputData] = {}\\n if(\\\"answer1\\\" not in resultData[inputData].keys()):\\n resultData[inputData][\\\"answer1\\\"] = []\\n if(\\\"answer2\\\" not in resultData[inputData].keys()):\\n resultData[inputData][\\\"answer2\\\"] = []\\n if(\\\"answer3\\\" not in resultData[inputData].keys()):\\n resultData[inputData][\\\"answer3\\\"] = []\\n\\n # check control question\\n if(eval(eval(answer4)) < 8):\\n # control question failed\\n print \\\"control question failed: \\\" + answer4\\n continue\\n\\n # check for invalid responses\\n invalidResponse = [type(eval(eval(x))) is not int for x in [answer1, answer2, answer3, answer4]]\\n if(np.any(invalidResponse) == True):\\n # invalid response\\n print \\\"invalid response: \\\" + str([answer1, answer2, answer3, answer4])\\n \\n resultData[inputData][\\\"answer1\\\"].append(eval(answer1))\\n resultData[inputData][\\\"answer2\\\"].append(eval(answer2))\\n resultData[inputData][\\\"answer3\\\"].append(eval(answer3))\\n\\n # now process result data\\n for inputData in resultData.keys():\\n print inputData\\n answerList = []\\n for answer in resultData[inputData].keys():\\n resultData[inputData][answer] = [eval(x) for x in resultData[inputData][answer] if is_integer(x)]\\n #resultData[inputData][answer] = np.mean(resultData[inputData][answer])\\n resultData[inputData][answer] = np.std(resultData[inputData][answer])\\n print answer, resultData[inputData][answer]\\n answerList.append(resultData[inputData][answer])\\n print \\\"mean: \\\", np.mean(answerList)\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6615158319473267,\n \"alphanum_fraction\": 0.6857984066009521,\n \"avg_line_length\": 27.11724090576172,\n \"blob_id\": \"59eda82a02810568e0283b7b641928c11df11c13\",\n \"content_id\": \"438233b2382da77003e30b8fdd85da3f83593bdf\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 4077,\n \"license_type\": \"no_license\",\n \"max_line_length\": 89,\n \"num_lines\": 145,\n \"path\": \"/src/sampling/utils/keras/variational_autoencoder_model.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"'''This script demonstrates how to build a variational autoencoder with Keras.\\n\\nReference: \\\"Auto-Encoding Variational Bayes\\\" https://arxiv.org/abs/1312.6114\\n'''\\nimport numpy as np\\nimport matplotlib.pyplot as plt\\nimport glob\\nfrom skimage.io import imread\\nfrom skimage.transform import resize\\n\\nfrom keras.layers import Input, Dense, Lambda\\nfrom keras.models import Model\\nfrom keras import backend as K\\nfrom keras import objectives\\nfrom keras.datasets import mnist\\nfrom keras.models import model_from_json\\n\\nimport math\\n\\n#batch_size = 100\\nbatch_size = 1\\noriginal_dim = 784\\nlatent_dim = 128\\nintermediate_dim = 256\\n#nb_epoch = 50\\nnb_epoch = 1\\nepsilon_std = 0.01\\n\\nx = Input(batch_shape=(batch_size, original_dim))\\nh = Dense(intermediate_dim, activation='relu')(x)\\nz_mean = Dense(latent_dim)(h)\\nz_log_var = Dense(latent_dim)(h)\\n\\n\\ndef sampling(args):\\n z_mean, z_log_var = args\\n\\n # karan\\n #batch_size = 100\\n #latent_dim = 128\\n #epsilon_std = 0.01\\n \\n epsilon = K.random_normal(shape=(batch_size, latent_dim), mean=0.,\\n std=epsilon_std)\\n return z_mean + K.exp(z_log_var / 2) * epsilon\\n\\n# note that \\\"output_shape\\\" isn't necessary with the TensorFlow backend\\nz = Lambda(sampling, output_shape=(latent_dim,))([z_mean, z_log_var])\\n\\n# we instantiate these layers separately so as to reuse them later\\ndecoder_h = Dense(intermediate_dim, activation='relu')\\ndecoder_mean = Dense(original_dim, activation='sigmoid')\\nh_decoded = decoder_h(z)\\nx_decoded_mean = decoder_mean(h_decoded)\\n\\n\\ndef vae_loss(x, x_decoded_mean):\\n xent_loss = original_dim * objectives.binary_crossentropy(x, x_decoded_mean)\\n kl_loss = - 0.5 * K.sum(1 + z_log_var - K.square(z_mean) - K.exp(z_log_var), axis=-1)\\n return xent_loss + kl_loss\\n\\nvae = Model(x, x_decoded_mean)\\nvae.compile(optimizer='rmsprop', loss=vae_loss)\\n\\n# train the VAE on MNIST digits\\n(x_train, y_train), (x_test, y_test) = mnist.load_data()\\n\\n# load all images in folder\\n\\ndef image_data(file):\\n\\timg = imread(file, as_grey=True)\\n img = resize(img, (28, 28))\\n img = img.flatten()\\n return img\\n\\ndef folder_data(folder):\\n\\tfiles = glob.glob(folder + \\\"/*.png\\\")\\n\\tprint(\\\"Calculating descriptors. Number of images is\\\", len(files))\\n\\treturn [image_data(file) for file in files]\\n\\nfolder = \\\"images/1108\\\"\\n#folder = \\\"../../../../_swarm-lfd-data/Flocking\\\"\\nx_folder = folder_data(folder)\\nsplitPoint = int(math.ceil(len(x_folder)*2.0/3.0))\\nx_train = x_folder[:splitPoint]\\nx_test = x_folder[splitPoint:]\\n\\n# artificially increase size of lists\\n#x_train *= 100\\n#x_test *= 100\\n\\nx_train = np.array(x_train)\\nx_test = np.array(x_test)\\n\\nprint(len(x_train))\\nprint(len(x_test))\\n\\nx_train = x_train.astype('float32') / 255.\\nx_test = x_test.astype('float32') / 255.\\nx_train = x_train.reshape((len(x_train), np.prod(x_train.shape[1:])))\\nx_test = x_test.reshape((len(x_test), np.prod(x_test.shape[1:])))\\n\\nprint(x_train.shape)\\nprint(x_train.shape[1:])\\n\\nvae.fit(x_train, x_train,\\n shuffle=True,\\n nb_epoch=nb_epoch,\\n batch_size=batch_size,\\n validation_data=(x_test, x_test))\\n\\n'''\\n# save model as JSON\\nmodel = vae\\nmodelAsJson = model.to_json()\\nwith open(\\\"variational_autoencoder_model.json\\\", \\\"w\\\") as jsonFile:\\n jsonFile.write(modelAsJson)\\n# serialize weights to HDF5\\nmodel.save_weights('variational_autoencoder_model.h5')\\n'''\\n\\n'''\\n# load json and create model\\njson_file = open('variational_autoencoder_model.json', 'r')\\nloaded_model_json = json_file.read()\\njson_file.close()\\nloaded_model = model_from_json(loaded_model_json)\\n# load weights into new model\\nloaded_model.load_weights(\\\"variational_autoencoder_model.h5\\\")\\n\\n# get output for specific layer\\n# with a Sequential model\\nmodel = loaded_model\\nprint(len(model.layers))\\n\\nimage = image_data(\\\"images/1108/1108_200.png\\\")\\nimage = image.astype('float32') / 255.\\nimage = image.reshape((1, np.prod(x_train.shape[1:])))\\n\\nget_layer_output = K.function([model.layers[0].input, K.learning_phase()],\\n [model.layers[3].output])\\n\\nlayer_output = get_layer_output([image, 0])[0]\\nprint(len(layer_output[0]))\\n'''\\n\"\n },\n {\n \"alpha_fraction\": 0.7346625924110413,\n \"alphanum_fraction\": 0.7346625924110413,\n \"avg_line_length\": 30.047618865966797,\n \"blob_id\": \"ccf0ccb753afb63d35c23cf504ea9943b2c7a1ad\",\n \"content_id\": \"f4408024e0b759e7488b6f73af44748bd904bdd0\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 652,\n \"license_type\": \"no_license\",\n \"max_line_length\": 54,\n \"num_lines\": 21,\n \"path\": \"/src/plot_feedback.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import sys\\nimport matplotlib.pyplot as plt\\nimport numpy as np\\n\\nwith open(\\\"feedback_output_difference.txt\\\") as inFile:\\n difference = inFile.readlines()\\n difference = [eval(x) for x in difference]\\n print >> sys.stderr, np.sum(difference)\\nplt.figure()\\nplt.plot(difference, label=\\\"difference\\\")\\nplt.legend()\\nplt.savefig(\\\"feedback_output_difference.png\\\")\\n\\nwith open(\\\"feedback_input_difference.txt\\\") as inFile:\\n difference = inFile.readlines()\\n difference = [eval(x) for x in difference]\\n print >> sys.stderr, np.sum(difference)\\nplt.figure()\\nplt.plot(difference, label=\\\"difference\\\")\\nplt.legend()\\nplt.savefig(\\\"feedback_input_difference.png\\\")\\n\"\n },\n {\n \"alpha_fraction\": 0.6097992658615112,\n \"alphanum_fraction\": 0.6257379055023193,\n \"avg_line_length\": 20.71794891357422,\n \"blob_id\": \"2f509e1321a4c951de5d0934a940ef1d429544c7\",\n \"content_id\": \"beea8a739a0d74e34470bdd642aa59286529886b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1694,\n \"license_type\": \"no_license\",\n \"max_line_length\": 66,\n \"num_lines\": 78,\n \"path\": \"/src/sampling/schelling/predict_schelling.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from model import SchellingModel\\nimport random\\nimport sys\\nimport numpy as np\\n\\n# only use three of the several ALPs\\n# density (0,1)\\n# minority_pc (0,1)\\n\\n# use one model-level SLP\\n# happy\\n\\n# constant ALPs\\nHEIGHT = 20\\nWIDTH = 20\\nHOMOPHILY = 4\\n\\n# constants\\n# max_iters in the example is specified as 1000\\n# so we run for 1000 steps\\nINDEPENDENT_VARIABLES = 2\\nDEPENDENT_VARIABLES = 1\\nSAMPLES = 10\\nTIME_LAPSE = 1000\\n\\ndef main_schelling(density, minorityPc):\\n # instantiate and run model\\n model = SchellingModel(\\n height=HEIGHT,\\n width=WIDTH,\\n density=density,\\n minority_pc=minorityPc,\\n homophily=HOMOPHILY)\\n\\n # run time lapse\\n for i in range(TIME_LAPSE):\\n try:\\n # step\\n model.step()\\n except:\\n # no empty cells\\n pass\\n \\n # collect data for next steps\\n dependentValues = []\\n\\n for i in range(SAMPLES):\\n try:\\n # step\\n model.step()\\n except:\\n # saturated\\n # no empty cells\\n pass\\n\\n # read data\\n data = model.datacollector.get_model_vars_dataframe() \\n dependentValues.append(np.mean(list(data.happy)[-SAMPLES-1:]))\\n\\n return dependentValues\\n \\ndef main():\\n # simulate using specific ALPs\\n alpConfigurationList = eval(sys.argv[1])\\n\\n # store corresponding slps\\n slpConfigurationList = []\\n \\n for alpConfiguration in alpConfigurationList:\\n [density, minorityPc] = alpConfiguration\\n \\n # run model using those ALPs\\n dependentValues = main_schelling(density, minorityPc)\\n slpConfigurationList.append(dependentValues)\\n\\n print(slpConfigurationList)\\n\\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.6643073558807373,\n \"alphanum_fraction\": 0.6723963618278503,\n \"avg_line_length\": 28.08823585510254,\n \"blob_id\": \"c60013c803d1d6cc6161f9ad0b394b6365866421\",\n \"content_id\": \"08ff94a342466b67f7faef2c5f2157a1a3a6fca0\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 989,\n \"license_type\": \"no_license\",\n \"max_line_length\": 97,\n \"num_lines\": 34,\n \"path\": \"/data/domaindata/reduce.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import numpy as np\\nfrom sklearn.decomposition import KernelPCA\\nimport cPickle\\n\\nnumDependent = 50\\n\\n# read input\\nwith open(\\\"processed_turbulence_vgg.txt\\\", \\\"r\\\") as inFile:\\n lines = inFile.readlines()[1:]\\n lines = [[eval(x) for x in line.split(\\\" \\\")] for line in lines]\\n\\nlines = np.array(lines)\\nalps = lines[:,:2]\\nslps = lines[:,2:]\\nkpca = KernelPCA(n_components=numDependent, kernel=\\\"rbf\\\")\\nmodel = kpca.fit(slps)\\nslpsKpca = model.transform(slps)\\nxKpca = np.concatenate((alps, slpsKpca), axis=1)\\n\\nrows, columns = xKpca.shape\\n\\n# write output\\nwith open(\\\"processed_turbulence_vgg_reduced.txt\\\", \\\"w\\\") as outFile:\\n outFile.write(\\\"i\\\"*2 + \\\"d\\\"*numDependent)\\n\\n for rowIndex in range(rows):\\n string = \\\"\\\\n\\\"\\n for columnIndex in range(columns):\\n string += str(xKpca[rowIndex][columnIndex]) + \\\" \\\"\\n string = string[:-1]\\n outFile.write(string)\\n\\n# save transform\\ncPickle.dump(model, open(\\\"../../src/sampling/utils/transforms/turbulence_reduced_model.p\\\", \\\"wb\\\"))\\n\"\n },\n {\n \"alpha_fraction\": 0.6873920559883118,\n \"alphanum_fraction\": 0.6925734281539917,\n \"avg_line_length\": 24.130434036254883,\n \"blob_id\": \"957e66a619fa6e0f75520edde500d22267dd878c\",\n \"content_id\": \"f62f510793c3d6900bfd2cad563db98bf7e558cc\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 579,\n \"license_type\": \"no_license\",\n \"max_line_length\": 69,\n \"num_lines\": 23,\n \"path\": \"/src/do_feedback.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# clean\\nrm -f feedback_*.txt\\nrm -f suggested_*.txt\\nrm -f queried_*.txt\\n\\n# run multiple times to generate multiple sets of images\\nexport ITERATIONS=100\\n\\nfor i in $(seq \\\"$ITERATIONS\\\")\\ndo\\n # store the current iteration value\\n echo $i > feedback_iteration.txt\\n \\n # run experiment\\n # simulate results\\n ./lfd.sh && ./simulate.sh\\n\\n # observe data for slps\\n # append to file\\n cat suggested_slps.txt >> feedback_suggested_slps.txt\\n cat queried_slps.txt >> feedback_queried_slps.txt\\n cat regression_parameter.txt >> feedback_regression_parameter.txt\\ndone\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5935162305831909,\n \"alphanum_fraction\": 0.6084787845611572,\n \"avg_line_length\": 32.072166442871094,\n \"blob_id\": \"a433d1c05e443d18949c083047c0db92919fd502\",\n \"content_id\": \"23ac65e121d2a200e5cf53c09a4ae04d5e00b15f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3208,\n \"license_type\": \"no_license\",\n \"max_line_length\": 106,\n \"num_lines\": 97,\n \"path\": \"/snippets/eum/plot_suggested_slps.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import matplotlib.pyplot as plt\\nimport numpy as np\\n\\n#demonstrations = [\\\"a\\\", \\\"b\\\", \\\"c\\\"]\\n#demonstrations = [\\\"_a\\\", \\\"_b\\\", \\\"_c\\\"]\\ndemonstrations = [\\\"a\\\"]\\n#demonstrations = [\\\"_a\\\"]\\n#demonstrations = [\\\"a_filtered\\\"]\\n\\nfolderName = \\\"error_weighted\\\"\\n#folderName = \\\"min_error\\\"\\n\\nnumberOfBins = 10\\n\\nwith open(\\\"../../src/sampling/eum/scale_data_eum.txt\\\", \\\"r\\\") as inFile:\\n allLines = inFile.readlines()\\n maxValueMean = eval(allLines[0])[-2]\\n maxValueMedian = eval(allLines[0])[-1]\\n minValueMean = eval(allLines[1])[-2]\\n minValueMedian = eval(allLines[1])[-1]\\n\\nfor index, item in enumerate(demonstrations):\\n allPointsX = []\\n allPointsY = []\\n errors = []\\n \\n # scatter\\n \\n plt.figure()\\n with open(folderName + \\\"/suggested_slps_\\\" + item + \\\".txt\\\", \\\"r\\\") as inFile:\\n for line in inFile:\\n try:\\n point = eval(line)[0]\\n plt.scatter(point[0], point[1], c=\\\"b\\\", marker=\\\"o\\\", alpha=0.3)\\n allPointsX.append(point[0])\\n allPointsY.append(point[1])\\n\\n # compute error\\n with open(\\\"demonstration_slps.txt\\\", \\\"r\\\") as inFile:\\n demonstration = eval(inFile.readlines()[index])\\n error = (point[0]-demonstration[0])**2 + (point[1]-demonstration[1])**2\\n errors.append(error)\\n \\n except:\\n continue\\n \\n plt.scatter(point[0], point[1], c=\\\"b\\\", marker=\\\"o\\\", alpha=0.3, label=\\\"Suggested SLPs\\\")\\n\\n with open(\\\"demonstration_slps.txt\\\", \\\"r\\\") as inFile:\\n point = eval(inFile.readlines()[index])\\n plt.scatter(point[0], point[1], c=\\\"r\\\", marker=\\\"s\\\", alpha=1.0, label=\\\"Demonstration SLPs\\\")\\n \\n plt.scatter(np.mean(allPointsX), np.mean(allPointsY), c=\\\"g\\\", marker=\\\"s\\\", alpha=1.0, label=\\\"Mean SLPs\\\")\\n\\n print np.mean(errors)\\n print np.std(errors)\\n \\n ax = plt.gca()\\n\\n ticks = ax.get_xticks()\\n minValue = minValueMean\\n maxValue = maxValueMean\\n scaledTicks = [(value-minValue)/(maxValue-minValue) for value in ticks]\\n scaledTicks = [int(value*100)/100.0 for value in scaledTicks]\\n ax.set_xticklabels(scaledTicks)\\n \\n ticks = ax.get_yticks()\\n minValue = minValueMedian\\n maxValue = maxValueMedian\\n scaledTicks = [(value-minValue)/(maxValue-minValue) for value in ticks]\\n scaledTicks = [int(value*100)/100.0 for value in scaledTicks]\\n ax.set_yticklabels(scaledTicks)\\n\\n plt.ylabel(\\\"Median\\\")\\n plt.xlabel(\\\"Mean\\\")\\n plt.legend()\\n plt.savefig(folderName + \\\"/suggested_slps_\\\" + item)\\n\\n # histogram\\n \\n plt.figure()\\n histogram, xEdges, yEdges = np.histogram2d(allPointsX, allPointsY, bins=numberOfBins)\\n plt.imshow(histogram, interpolation='nearest', origin='low')\\n plt.xlabel(\\\"Mean Final Position\\\")\\n plt.ylabel(\\\"Median Final Position\\\")\\n plt.colorbar(label=\\\"Number of Points\\\")\\n ax = plt.gca()\\n\\n # works for normalized values\\n ticks = ax.get_xticks()\\n ticks = [value/numberOfBins for value in ticks]\\n ax.set_xticklabels(ticks)\\n ticks = ax.get_yticks()\\n ticks = [value/numberOfBins for value in ticks]\\n ax.set_yticklabels(ticks)\\n\\n plt.savefig(folderName + \\\"/histogram_suggested_slps_\\\" + item)\\n\"\n },\n {\n \"alpha_fraction\": 0.6872605085372925,\n \"alphanum_fraction\": 0.7131226062774658,\n \"avg_line_length\": 30.1641788482666,\n \"blob_id\": \"b8c6d0c896548676b6ee1e7450744ec14a4bc566\",\n \"content_id\": \"318d23cb9b11601480f20c3fd535fa0566bf5212\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2088,\n \"license_type\": \"no_license\",\n \"max_line_length\": 74,\n \"num_lines\": 67,\n \"path\": \"/src/sampling/utils/keras/mnist_cnn_model.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"'''Trains a simple convnet on the MNIST dataset.\\n\\nGets to 99.25% test accuracy after 12 epochs\\n(there is still a lot of margin for parameter tuning).\\n16 seconds per epoch on a GRID K520 GPU.\\n'''\\n\\nimport numpy as np\\nfrom keras.datasets import mnist\\nfrom keras.utils import np_utils\\nfrom keras.models import model_from_json\\nfrom skimage.io import imread\\nfrom skimage.transform import resize\\n\\n# load json and create model\\njson_file = open('mnist_cnn_model.json', 'r')\\nloaded_model_json = json_file.read()\\njson_file.close()\\nloaded_model = model_from_json(loaded_model_json)\\n# load weights into new model\\nloaded_model.load_weights(\\\"mnist_cnn_model.h5\\\")\\nprint(\\\"Loaded model from disk\\\")\\n\\n# the data, shuffled and split between train and test sets\\n# input image dimensions\\nnb_classes = 10\\nimg_rows, img_cols = 28, 28\\n(X_train, y_train), (X_test, y_test) = mnist.load_data()\\nX_train = X_train.reshape(X_train.shape[0], 1, img_rows, img_cols)\\nX_test = X_test.reshape(X_test.shape[0], 1, img_rows, img_cols)\\nX_train = X_train.astype('float32')\\nX_test = X_test.astype('float32')\\nX_train /= 255\\nX_test /= 255\\nprint('X_train shape:', X_train.shape)\\nprint(X_train.shape[0], 'train samples')\\nprint(X_test.shape[0], 'test samples')\\n\\n# convert class vectors to binary class matrices\\nY_train = np_utils.to_categorical(y_train, nb_classes)\\nY_test = np_utils.to_categorical(y_test, nb_classes)\\n\\n# test model\\nmodel = loaded_model\\nmodel.compile(loss='categorical_crossentropy',\\n optimizer='adadelta',\\n metrics=['accuracy'])\\n\\n# import image and resize\\nimage = imread(\\\"409_200.png\\\", as_grey=True)\\nimage = resize(image, (28, 28))\\n\\nfrom skimage.io import imshow\\n\\nimage = image.reshape(1, 1, 28, 28)\\nprint(image.shape)\\nclasses = model.predict_classes(image)\\nprint(classes)\\n\\n# get output for specific layer\\nfrom keras import backend as K\\n# with a Sequential model\\nprint(len(model.layers))\\nget_layer_output = K.function([model.layers[0].input, K.learning_phase()],\\n [model.layers[7].output])\\nlayer_output = get_layer_output([image, 0])[0]\\nprint(layer_output)\\n\"\n },\n {\n \"alpha_fraction\": 0.7064123153686523,\n \"alphanum_fraction\": 0.7095919251441956,\n \"avg_line_length\": 34.60377502441406,\n \"blob_id\": \"f5c21909f4bdf333f86c6c98da59f185db0f0011\",\n \"content_id\": \"2a7d3206d3758b6fad54c07d67e94b4f212fcf1a\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 1887,\n \"license_type\": \"no_license\",\n \"max_line_length\": 116,\n \"num_lines\": 53,\n \"path\": \"/src/sampling/sampling_process_raw.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# copy raw data file if needed\\nif [ $IMAGES == 1 ]; then\\n # raw data file is present in non-repository directory\\n cp ../../../_swarm-lfd-data/$ABM/$ABM.txt .\\nfi\\n\\n# separate head from rest of data\\nhead -n 1 $ABM.txt > header.txt\\ntail -n +2 $ABM.txt > $ABM\\\"_.txt\\\"\\nmv $ABM\\\"_.txt\\\" $ABM.txt\\n\\n# process images if any\\n# check if this abm depends on images\\nif [ $IMAGES == 1 ]; then\\n # images and raw data file are present in non-repository directory\\n mv ../../../_swarm-lfd-data/$ABM/images $ABM/\\n python utils/process_training_images.py\\n mv $ABM/images ../../../_swarm-lfd-data/$ABM/\\nfi\\n\\n# extract independent component\\n# generate dependent component\\n# combine results\\n# overwrite header file\\nexport INDEPENDENT_HEADER=\\\"i\\\"\\nexport DEPENDENT_HEADER=\\\"d\\\"\\nexport INDEPENDENT_HEADER_STRING=$(sed \\\"s/[^\\\"$INDEPENDENT_HEADER\\\"']//g\\\" <<< cat header.txt)\\nexport DEPENDENT_HEADER_STRING=$(printf $DEPENDENT_HEADER {1..$NUM_DEPENDENT})\\necho $INDEPENDENT_HEADER_STRING$DEPENDENT_HEADER_STRING > header.txt\\n\\n# prepare data for amf input\\necho \\\"==>preparing data for amf\\\"\\npython ../map.py $ABM.txt $ABM > processed_$ABM.txt\\ncat header.txt processed_$ABM.txt > temp.txt\\nmv temp.txt processed_$ABM.txt\\nrm header.txt\\nmv processed_$ABM.txt ../../data/domaindata\\n\\n# if image processing was used, append descriptor setting to processed filename\\nif [ $IMAGES == 1 ]; then\\n # generate lower case descriptor name\\n export DESCRIPTOR_SETTING_LOWER=$(echo \\\"$DESCRIPTOR_SETTING\\\" | tr '[:upper:]' '[:lower:]')\\n mv ../../data/domaindata/processed_$ABM.txt ../../data/domaindata/processed_$ABM\\\"_\\\"$DESCRIPTOR_SETTING_LOWER.txt\\nfi\\n\\n# this file is no longer needed for computation\\n# the forward and reverse mappings are built from the processed file\\necho \\\"==>cleaning\\\"\\nrm $ABM.txt\\n\\n# move scale data to the abm folder\\n# assume only one scale data file exists here right now\\nmv scale_data_$ABM*.txt $ABM\\n\"\n },\n {\n \"alpha_fraction\": 0.6077530384063721,\n \"alphanum_fraction\": 0.6122422218322754,\n \"avg_line_length\": 44.515625,\n \"blob_id\": \"e38b994312427bd1fb97cca60eade85e1263f253\",\n \"content_id\": \"13a7dbdd0afcb62d7bc875cb3eafbe47016f7074\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 37869,\n \"license_type\": \"no_license\",\n \"max_line_length\": 285,\n \"num_lines\": 832,\n \"path\": \"/src/lfd.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# import libraries\\nimport matplotlib\\nmatplotlib.use('Agg')\\n\\nimport sys\\nimport amf\\nimport amf.misc as misc\\n\\nfrom sklearn import svm\\nfrom sklearn.cluster import KMeans\\nimport scipy.io as sio\\n\\nfrom mpl_toolkits.mplot3d import Axes3D\\nimport matplotlib.pyplot as plt\\n\\nimport os\\nimport time\\nimport random\\n\\nimport cPickle\\nimport gzip\\nimport numpy as np\\nfrom sets import Set\\n\\nimport sampling.utils.featurize_image as fi\\nimport sampling.utils.outlier_detection as od\\n\\n# constants\\nPLOT_COLOR = {0: \\\"r\\\", 1: \\\"b\\\", 2: \\\"g\\\", 3: \\\"c\\\", 4: \\\"m\\\", 5: \\\"y\\\", 6: \\\"k\\\"}\\nPLOT_SHAPE = {0: \\\"o\\\", 1: \\\"^\\\", 2: \\\"s\\\", 3: \\\"d\\\", 4: \\\"+\\\", 5: \\\"v\\\", 6: \\\"*\\\"}\\nPLOT_DEMONSTRATIONS = {0: \\\"a\\\", 1: \\\"b\\\", 2: \\\"c\\\", 3: \\\"d\\\", 4: \\\"e\\\", 5: \\\"f\\\", 6: \\\"g\\\"}\\n\\nRUNS_PER_SLPS = 1\\n\\nIMAGE_FEATURE_PATH_PREFIX = \\\"sampling/\\\"\\n\\n# experiment class\\nclass Experiment(object):\\n\\n # initialize experiment parameters based on file\\n def __init__(self):\\n # read configuration\\n configurationFileName = os.environ[\\\"CONFIGURATION_FILE\\\"]\\n configuration = open(configurationFileName).read()\\n execd = {}\\n exec(configuration, execd)\\n\\n self.trainingSize = execd['TRAINING_SIZE']\\n self.validationSize = execd['VALIDATION_SIZE']\\n self.rmGranularity = execd['RM_GRANULARITY']\\n self.numIndependent = execd['NUM_INDEPENDENT']\\n self.regression = execd['REGRESSION']\\n\\n # changes for feedback\\n if(eval(os.environ[\\\"FEEDBACK\\\"]) == 1):\\n # randomness causes too much noise for feedback\\n self.validationSize = 0\\n\\n # modify regression parameter\\n if(os.path.isfile(\\\"regression_parameter.txt\\\") == True):\\n with open(\\\"regression_parameter.txt\\\", \\\"r\\\") as inFile:\\n regressionParameter = inFile.readlines()[0]\\n #self.regression = self.regression.replace(\\\"()\\\", \\\"(alpha=np.exp((-1*\\\" + regressionParameter + \\\")))\\\")\\n self.regression = self.regression.replace(\\\"()\\\", \\\"(n_neighbors=int(np.floor(\\\" + regressionParameter + \\\")))\\\")\\n \\n # get environment variables\\n self.abm = os.environ[\\\"ABM\\\"]\\n self.demonstrationFolder = os.environ[\\\"DEMONSTRATION_FOLDER\\\"]\\n self.descriptorSetting = os.environ[\\\"DESCRIPTOR_SETTING\\\"]\\n self.numDependent = eval(os.environ[\\\"NUM_DEPENDENT\\\"])\\n self.dataFileName = os.environ[\\\"DATA_FILE\\\"]\\n self.scaleDataFileName = os.environ[\\\"SCALE_DATA_FILE\\\"]\\n self.images = eval(os.environ[\\\"IMAGES\\\"])\\n self.writeToDisk = eval(os.environ[\\\"WRITE_TO_DISK\\\"])\\n \\n # reverse mapping configuration: ranges of agent level parameters\\n self.rmConfiguration = tuple()\\n for i in range(self.numIndependent):\\n self.rmConfiguration += ((0.0, 1.0), )\\n self.rmConfiguration = list(self.rmConfiguration)\\n\\n # for plotting\\n self.rmQueryCount = 0\\n \\n # mapping models\\n self.fm = None\\n self.rm = None\\n \\n # retrieve reverse mapping\\n def get_mappings(self, slpsList, saveData=False):\\n # parse slps list\\n demonstrationDescriptors = [depValues for depValues, plotColor, plotShape in slpsList]\\n \\n # load data\\n print >> sys.stderr, \\\"->Loading data set \\\" + self.dataFileName\\n fullData = amf.data.load(self.dataFileName)\\n \\n # forward mapping using largest training size\\n print >> sys.stderr, \\\"->Trainining FM\\\"\\n trainingSize = self.trainingSize\\n dataset, validation = amf.data.random_subsets(fullData, [trainingSize, self.validationSize], demonstrationDescriptors)\\n fm = amf.ForwardMapping(self.regression, self.numDependent)\\n fm.train(dataset)\\n \\n if(saveData == True):\\n # gather ALPs and SLPs\\n # plot them in space to observe distribution\\n alpPoints = [item[0] for item in dataset]\\n slpPoints = [item[1] for item in dataset]\\n c = \\\"r\\\"\\n m = \\\"o\\\"\\n \\n # save ALP dataset\\n with open(\\\"alp_dataset_\\\" + str(int(time.time())) + \\\".txt\\\", \\\"w\\\") as alpDatasetFile:\\n for alpPoint in alpPoints:\\n alpDatasetFile.write(str(alpPoint) + \\\"\\\\n\\\")\\n \\n # save SLP dataset\\n with open(\\\"slp_dataset_\\\" + str(int(time.time())) + \\\".txt\\\", \\\"w\\\") as slpDatasetFile:\\n for slpPoint in slpPoints:\\n slpDatasetFile.write(str(slpPoint) + \\\"\\\\n\\\")\\n\\n # plot ALPs\\n fig = plt.figure()\\n ax = fig.add_subplot(111, projection='3d')\\n plotName = \\\"alp_points_\\\" + str(int(time.time())) + \\\".png\\\"\\n alpPoints = zip(*alpPoints)\\n [xs, ys, zs] = alpPoints\\n ax.scatter(xs, ys, zs, c=c, marker=m)\\n ax.set_xlabel(\\\"Max-align-turn\\\")\\n ax.set_ylabel(\\\"Max-cohere-turn\\\")\\n ax.set_zlabel(\\\"Max-separate-turn\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n \\n fig = plt.figure()\\n plotName = \\\"alp_points_dim1_\\\" + str(int(time.time())) + \\\".png\\\"\\n plt.plot(xs, xs, c+m)\\n plt.xlabel(\\\"ALP Dimension 1\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n \\n fig = plt.figure()\\n plotName = \\\"alp_points_dim2_\\\" + str(int(time.time())) + \\\".png\\\"\\n plt.plot(ys, ys, c+m)\\n plt.xlabel(\\\"ALP Dimension 2\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n \\n fig = plt.figure()\\n plotName = \\\"alp_points_dim3_\\\" + str(int(time.time())) + \\\".png\\\"\\n plt.plot(zs, zs, c+m)\\n plt.xlabel(\\\"ALP Dimension 3\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n \\n # plot SLPs\\n fig = plt.figure()\\n plotName = \\\"slp_points_\\\" + str(int(time.time())) + \\\".png\\\"\\n slpPoints = zip(*slpPoints)\\n [xs, ys] = slpPoints\\n plt.plot(xs, ys, c+m)\\n plt.xlabel(\\\"SLP Dimension 1\\\")\\n plt.ylabel(\\\"SLP Dimension 2\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n \\n fig = plt.figure()\\n plotName = \\\"slp_points_dim1_\\\" + str(int(time.time())) + \\\".png\\\"\\n plt.plot(xs, xs, c+m)\\n plt.xlabel(\\\"SLP Dimension 1\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n \\n fig = plt.figure()\\n plotName = \\\"slp_points_dim2_\\\" + str(int(time.time())) + \\\".png\\\"\\n plt.plot(ys, ys, c+m)\\n plt.xlabel(\\\"SLP Dimension 2\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\n # reverse mapping using largest granularity\\n print >> sys.stderr, \\\"->Trainining RM\\\"\\n granularity = self.rmGranularity\\n rm = amf.ReverseMapping(fm, self.rmConfiguration, granularity, slpsList)\\n\\n return [fm, rm]\\n\\n # query reverse mapping for configurations\\n def query_rm(self, fm, rm, slps, ax, c, m, plotConfigurations=False):\\n intersections = rm.all_intersections(slps)\\n allConfigurations = []\\n\\n #\\n # filter from all query results\\n # \\n \\n simplexNumbers = []\\n edgeNumbers = []\\n _allConfigurations = []\\n\\n # check for special case\\n # if wiggle happened, then only one configuration is returned\\n if(len(intersections[0]) == 1):\\n # configuration corresponds to simplex number 0 and edge number 0\\n # an intersection is stored as [simplexNumber][edgeNumber][configuration]\\n if(type(intersections[0][0][0]) is tuple):\\n # no simplex number is returned in this case\\n wiggledConfiguration = intersections[0][0][0]\\n else:\\n # different formats of return\\n # in this case, a default simplex and edge number is attached\\n # not sure why\\n wiggledConfiguration = intersections[0][0][1][0][1]\\n \\n allConfigurations = [list(wiggledConfiguration)] \\n return allConfigurations\\n \\n for intersection in intersections:\\n currentEdgeNumbers = {}\\n _configurations = {}\\n simplexNumbers.append([simplexNumber for (simplexNumber, configurations) in intersection])\\n \\n for simplexNumber, configurations in intersection:\\n currentEdgeNumbers[simplexNumber] = [edgeNumber for (edgeNumber, configuration) in configurations]\\n _configurations[simplexNumber] = {}\\n for edgeNumber, configuration in configurations:\\n allConfigurations.append(configuration)\\n _configurations[simplexNumber][edgeNumber] = configuration\\n \\n edgeNumbers.append(currentEdgeNumbers)\\n _allConfigurations.append(_configurations)\\n \\n # get common simplexes\\n simplexNumbers = [Set(numbers) for numbers in simplexNumbers]\\n commonSimplexes = simplexNumbers[0]\\n for index in range(1, len(simplexNumbers)):\\n currentSimplexes = simplexNumbers[index]\\n commonSimplexes = commonSimplexes.intersection(currentSimplexes)\\n\\n # get common edges\\n commonEdges = {}\\n for key in commonSimplexes:\\n currentCommonEdges = Set(edgeNumbers[0][key])\\n for index in range(1, len(edgeNumbers)):\\n currentEdges = Set(edgeNumbers[index][key])\\n currentCommonEdges = currentCommonEdges.intersection(currentEdges)\\n commonEdges[key] = currentCommonEdges\\n overlapCount = [len(values) for values in list(commonEdges.values())]\\n \\n # filter common edges\\n #overlapAmount = [len(commonEdges[key]) for key in commonEdges]\\n #print >> sys.stderr, overlapAmount\\n #plt.figure()\\n #plt.hist(overlapAmount)\\n #plt.savefig(\\\"overlapAmount_before_filtering.png\\\")\\n\\n if(len(overlapCount) > 0):\\n #plt.figure()\\n #plt.hist(overlapCount)\\n #plt.show()\\n #sys.exit()\\n\\n overlapMean = np.mean(overlapCount)\\n commonEdges = {key : commonEdges[key] for key in commonEdges if len(commonEdges[key]) >= overlapMean}\\n commonSimplexes = list(commonEdges.keys())\\n\\n #overlapAmount = [len(commonEdges[key]) for key in commonEdges]\\n #print >> sys.stderr, overlapAmount\\n #plt.figure()\\n #plt.hist(overlapAmount)\\n #plt.savefig(\\\"overlapAmount_after_filtering.png\\\")\\n #sys.exit()\\n \\n # now we need to cycle through the configurations again\\n # only pick ones corresponding to filtered edges\\n # TODO: we may want to combine configurations across SLP intersections for optimality. can we?\\n allConfigurationsCommon = []\\n for commonSimplex in commonSimplexes:\\n currentCommonEdges = commonEdges[commonSimplex]\\n for commonEdge in currentCommonEdges:\\n for alpIndex in range(len(simplexNumbers)):\\n currentConfiguration = _allConfigurations[alpIndex][commonSimplex][commonEdge]\\n #print >> sys.stderr, currentConfiguration\\n allConfigurationsCommon.append(currentConfiguration)\\n\\n # filter configurations based on common simplexes\\n if(eval(os.environ[\\\"CONFIGURATIONS_PRUNING\\\"]) == 1):\\n allConfigurations = allConfigurationsCommon\\n else:\\n # no overlapping edges found\\n # pruning cannot be used in this case\\n pass\\n\\n # outlier detection code\\n # outlier detection based on hubness\\n # approximate labeling based on minority clustering\\n\\n if(eval(os.environ[\\\"CONFIGURATIONS_OUTLIER_DETECTION\\\"]) == 1):\\n # outlier detection only valid for a dataset of minimum size\\n # meanshift clustering sometimes fails if it somehow finds nans\\n try:\\n # do outlier detection\\n [odPoints, labels] = od.get_approximate_labels(allConfigurations)\\n [odPoints, labels] = od.do_lpod(odPoints, labels)\\n \\n allConfigurationsOd = []\\n for configurationIndex in range(len(allConfigurations)):\\n currentConfiguration = odPoints[configurationIndex] \\n if(labels[configurationIndex] == 0):\\n # not an outlier\\n allConfigurationsOd.append(list(currentConfiguration))\\n \\n allConfigurations = allConfigurationsOd\\n except:\\n # outlier detection had an error\\n # it is not used in this case\\n pass\\n \\n # add solution to scatter plot\\n if(plotConfigurations == True):\\n print >> sys.stderr, \\\"->Plotting configurations\\\" \\n allXs = []\\n allYs = []\\n allZs = []\\n allErrors = []\\n configurationsToPlot = []\\n \\n for configuration in allConfigurations:\\n error = misc.list_sub(slps, fm.predict(configuration))\\n error = sum([x**2 for x in error])\\n configurationsToPlot.append((error, configuration))\\n configurationsToPlot.sort() \\n\\n # filter configurations here\\n # checking threshold effect\\n allConfigurationsToPlot = [x for x in configurationsToPlot]\\n histogramData = [error for error, configuration in allConfigurationsToPlot]\\n mu = np.mean(histogramData)\\n sigma = np.std(histogramData)\\n #configurationsToPlot = [x for x in configurationsToPlot if (x[0] < (mu + 0.1*sigma) and x[0] > (mu - 0.1*sigma))]\\n \\n for error, configuration in configurationsToPlot:\\n xs, ys, zs = configuration\\n allXs.append(xs)\\n allYs.append(ys)\\n allZs.append(zs)\\n allErrors.append(error)\\n\\n odPointsAllXs = [x[0] for x in odPoints]\\n odPointsAllYs = [x[1] for x in odPoints]\\n odPointsAllZs = [x[2] for x in odPoints]\\n \\n # plot point errors in 3d space \\n plot = ax.scatter(allXs, allYs, allZs, c=allErrors, marker=m, cmap=\\\"cool\\\")\\n plt.colorbar(plot, label=\\\"FM Prediction Error\\\")\\n ax.set_xlabel(\\\"Max-align-turn\\\")\\n ax.set_ylabel(\\\"Max-cohere-turn\\\")\\n ax.set_zlabel(\\\"Max-separate-turn\\\")\\n plt.savefig(\\\"solution_space.png\\\")\\n\\n # plot point labels in 3d space\\n fig = plt.figure()\\n ax = fig.add_subplot(111, projection='3d')\\n cmapClusters = matplotlib.cm.get_cmap(\\\"cool\\\", 2)\\n plot = ax.scatter(odPointsAllXs, odPointsAllYs, odPointsAllZs, c=labels, marker=m, cmap=cmapClusters)\\n plt.colorbar(plot, label=\\\"Cluster\\\").set_ticks([0, 1])\\n ax.set_xlabel(\\\"Max-align-turn\\\")\\n ax.set_ylabel(\\\"Max-cohere-turn\\\")\\n ax.set_zlabel(\\\"Max-separate-turn\\\")\\n plt.savefig(\\\"solution_space_clusters.png\\\")\\n \\n # error histogram\\n plt.figure()\\n plt.hist(histogramData)\\n plt.xlabel(\\\"FM Prediction Error\\\")\\n plt.title(\\\"mu = \\\" + str(mu) + \\\", sigma = \\\" + str(sigma))\\n plt.savefig(\\\"error_histogram\\\")\\n \\n projectDimensions = True\\n if(projectDimensions == True):\\n # plotting for single demonstration SLP configuration\\n\\n # plot projections with error\\n plt.figure()\\n plt.scatter(allXs, allYs, c=allErrors, marker=m, cmap=\\\"cool\\\")\\n plt.colorbar(label=\\\"FM Prediction Error\\\")\\n plt.gca().set_xlabel(\\\"Max-align-turn\\\")\\n plt.gca().set_ylabel(\\\"Max-cohere-turn\\\")\\n plt.savefig(\\\"solution_space_xy.png\\\")\\n plt.figure()\\n plt.scatter(allXs, allZs, c=allErrors, marker=m, cmap=\\\"cool\\\")\\n plt.colorbar(label=\\\"FM Prediction Error\\\")\\n plt.gca().set_xlabel(\\\"Max-align-turn\\\")\\n plt.gca().set_ylabel(\\\"Max-separate-turn\\\")\\n plt.savefig(\\\"solution_space_xz.png\\\")\\n plt.figure()\\n plt.scatter(allYs, allZs, c=allErrors, marker=m, cmap=\\\"cool\\\")\\n plt.colorbar(label=\\\"FM Prediction Error\\\")\\n plt.gca().set_xlabel(\\\"Max-cohere-turn\\\")\\n plt.gca().set_ylabel(\\\"Max-separate-turn\\\")\\n plt.savefig(\\\"solution_space_yz.png\\\")\\n\\n # plot projections with labels\\n plt.figure()\\n plt.scatter(odPointsAllXs, odPointsAllYs, c=labels, marker=m, cmap=cmapClusters)\\n plt.colorbar(label=\\\"Cluster\\\").set_ticks([0, 1])\\n plt.gca().set_xlabel(\\\"Max-align-turn\\\")\\n plt.gca().set_ylabel(\\\"Max-cohere-turn\\\")\\n plt.savefig(\\\"solution_space_xy_clusters.png\\\")\\n plt.figure()\\n plt.scatter(odPointsAllXs, odPointsAllZs, c=labels, marker=m, cmap=cmapClusters)\\n plt.colorbar(label=\\\"Cluster\\\").set_ticks([0, 1])\\n plt.gca().set_xlabel(\\\"Max-align-turn\\\")\\n plt.gca().set_ylabel(\\\"Max-separate-turn\\\")\\n plt.savefig(\\\"solution_space_xz_clusters.png\\\")\\n plt.figure()\\n plt.scatter(odPointsAllYs, odPointsAllZs, c=labels, marker=m, cmap=cmapClusters)\\n plt.colorbar(label=\\\"Cluster\\\").set_ticks([0, 1])\\n plt.gca().set_xlabel(\\\"Max-cohere-turn\\\")\\n plt.gca().set_ylabel(\\\"Max-separate-turn\\\")\\n plt.savefig(\\\"solution_space_yz_clusters.png\\\")\\n\\n # clear memory\\n del allXs, allYs, allZs, allErrors, configurationsToPlot, allConfigurationsToPlot\\n\\n return allConfigurations\\n\\n # format configurations for classifiaction\\n def format_data_xy(self, allConfigurationsList):\\n X = []\\n Y = []\\n for index in range(len(allConfigurationsList)):\\n X += allConfigurationsList[index]\\n Y += [index]*len(allConfigurationsList[index])\\n\\n return [X, Y]\\n\\n # compute fm errors and select configuration for minimum\\n def get_min_error_configurations(self, fm, allConfigurationsList, slps):\\n minErrorConfigurations = []\\n for allConfigurations in allConfigurationsList:\\n # error[0] contains error value\\n # error[1] contains configuration\\n errors = [(misc.list_sub(slps, fm.predict(configuration)), configuration) for configuration in allConfigurations]\\n errors = [(sum([x**2 for x in error[0]]), error[1]) for error in errors]\\n errors.sort()\\n minErrorConfigurations.append(errors[0][1])\\n\\n return minErrorConfigurations\\n\\n # compute average configuration\\n def get_mean_configurations(self, allConfigurationsList):\\n meanConfigurations = []\\n for allConfigurations in allConfigurationsList:\\n meanConfiguration = tuple(misc.col_average(allConfigurations))\\n meanConfigurations.append(meanConfiguration)\\n\\n return meanConfigurations\\n\\n def process_error_weighted_configuration(self, configuration, fm, slps, errorWeightedConfiguration, normalizationFactor, zeroErrorConfigurations):\\n prediction = fm.predict(configuration)\\n error = 0\\n\\n for index in range(len(slps)):\\n error += (slps[index] - prediction[index])**2\\n \\n if(error > 0):\\n if(zeroErrorConfigurations == 0):\\n # no zero error configuration found yet\\n normalizationFactor += 1/error\\n for index in range(len(errorWeightedConfiguration)):\\n errorWeightedConfiguration[index] += configuration[index]/error\\n else:\\n # error is zero\\n if(zeroErrorConfigurations == 0):\\n # first time we found a zero error configuration\\n for index in range(len(errorWeightedConfiguration)):\\n # wipe old data\\n errorWeightedConfiguration[index] = 0\\n zeroErrorConfigurations += 1\\n \\n # add new data\\n for index in range(len(errorWeightedConfiguration)):\\n errorWeightedConfiguration[index] += configuration[index]\\n\\n return [errorWeightedConfiguration, normalizationFactor, zeroErrorConfigurations]\\n \\n # compute error weighted average configuration\\n def get_error_weighted_configurations(self, fm, allConfigurationsList, slps):\\n errorWeightedConfigurations = []\\n listIndex = -1\\n \\n for allConfigurations in allConfigurationsList:\\n errorWeightedConfiguration = [0]*self.numIndependent\\n normalizationFactor = 0\\n zeroErrorConfigurations = 0\\n\\n if(self.writeToDisk == 1):\\n # read from disk\\n listIndex += 1\\n with gzip.open(\\\"configuration_\\\" + str(listIndex+1) + \\\".pgz\\\", 'rb') as configurationFile:\\n while 1:\\n try:\\n # read configuration\\n configuration = cPickle.load(configurationFile)[listIndex]\\n [errorWeightedConfiguration, normalizationFactor, zeroErrorConfigurations] = self.process_error_weighted_configuration(configuration, fm, slps, errorWeightedConfiguration, normalizationFactor, zeroErrorConfigurations)\\n except EOFError:\\n break\\n else:\\n # read from ram\\n for configuration in allConfigurations:\\n [errorWeightedConfiguration, normalizationFactor, zeroErrorConfigurations] = self.process_error_weighted_configuration(configuration, fm, slps, errorWeightedConfiguration, normalizationFactor, zeroErrorConfigurations)\\n \\n # nomalization\\n for index in range(len(errorWeightedConfiguration)):\\n if(zeroErrorConfigurations == 0):\\n errorWeightedConfiguration[index] /= float(normalizationFactor)\\n else:\\n # we found some zero error configurations\\n # we will only consider those configurations when computing a representative point\\n errorWeightedConfiguration[index] /= float(zeroErrorConfigurations)\\n \\n errorWeightedConfigurations.append(tuple(errorWeightedConfiguration))\\n\\n return errorWeightedConfigurations\\n\\n # compute classifier based optimal configuration\\n def get_svm_configurations(self, allConfigurationsList):\\n # train svm classifier on configurations\\n classifier = svm.SVC(probability=True)\\n [X, Y] = self.format_data_xy(allConfigurationsList)\\n classifier.fit(X, Y)\\n\\n # select configurations which maximize probability\\n svmConfigurations = []\\n for index in range(len(allConfigurationsList)):\\n # compute score for each probability and select maximum\\n configurationProbabilities = classifier.predict_proba(allConfigurationsList[index])\\n configurationProbabilities = [(configurationProbabilities[i][index], allConfigurationsList[index][i]) for i in range(len(configurationProbabilities))]\\n configurationProbabilities.sort(reverse=True)\\n svmConfiguration = tuple(configurationProbabilities[0][1])\\n svmConfigurations.append(svmConfiguration)\\n\\n return svmConfigurations\\n\\n # compute based on k means clustering\\n def get_kmeans_configurations(self, allConfigurationsList):\\n # train svm classifier on configurations\\n nClusters = len(allConfigurationsList)\\n kmeans = KMeans(init='k-means++', n_clusters=nClusters)\\n [X, Y] = self.format_data_xy(allConfigurationsList)\\n kmeans.fit(X, Y)\\n\\n # select centroids\\n centroids = kmeans.cluster_centers_\\n \\n return centroids\\n\\n # run experiment\\n def run(self, slpsList, plotConfigurations=False, saveError=False):\\n [fm, rm] = self.get_mappings(slpsList)\\n\\n # append to model list\\n self.fm = fm\\n self.rm = rm\\n\\n # plot solution space\\n fig = plt.figure()\\n ax = fig.add_subplot(111, projection='3d')\\n\\n print >> sys.stderr, \\\"->Querying RM for configurations\\\" \\n\\n startTime = time.time()\\n allConfigurationsList = []\\n proxyPlots = []\\n labels = []\\n for slps,c,m in slpsList:\\n allConfigurations = self.query_rm(fm, rm, slps, ax, c, m, plotConfigurations)\\n allConfigurationsList.append(allConfigurations)\\n\\n proxyPlot = matplotlib.lines.Line2D([0],[0], linestyle=\\\"none\\\", c=c, marker=m)\\n proxyPlots.append(proxyPlot)\\n labels.append(\\\"ALPs \\\" + PLOT_DEMONSTRATIONS[self.rmQueryCount])\\n ax.legend(proxyPlots, labels, numpoints = 1)\\n self.rmQueryCount += 1\\n stopTime = time.time()\\n\\n plt.close(fig)\\n\\n # how do we pick one?\\n print >> sys.stderr, \\\"configurations retrieved: \\\" + str([len(allConfigurations) for allConfigurations in allConfigurationsList]) + \\\" [in \\\" + str(stopTime-startTime) + \\\" seconds]\\\"\\n print >> sys.stderr, \\\"->Selecting Configurations\\\"\\n\\n # downsample by num dep\\n # required to make higher dimension feature vectors tractable\\n # can be removed later when computation is manageable\\n downsampledAllConfigurationsList = []\\n for allConfigurations in allConfigurationsList:\\n if(len(allConfigurations) > 1):\\n downsampledAllConfigurationsList.append(random.sample(allConfigurations, len(allConfigurations)/(self.numDependent**self.numIndependent)))\\n else:\\n downsampledAllConfigurationsList.append([allConfigurations[0]])\\n allConfigurationsList = downsampledAllConfigurationsList\\n\\n # check forward mapping error for each configuratoin returned\\n # check location in point cloud\\n startTime = time.time()\\n #configurationList = self.get_mean_configurations(allConfigurationsList)\\n #configurationList = self.get_min_error_configurations(fm, allConfigurationsList, slps)\\n configurationList = self.get_error_weighted_configurations(fm, allConfigurationsList, slps)\\n #configurationList = self.get_svm_configurations(allConfigurationsList)\\n #configurationList = self.get_kmeans_configurations(allConfigurationsList)\\n stopTime = time.time()\\n\\n # experiment complete\\n print >> sys.stderr, \\\"->Done\\\" + \\\" [in \\\" + str(stopTime-startTime) + \\\" seconds]\\\"\\n\\n # compute error for each configuration\\n # print error\\n demonstrationIndex = 0\\n for slps,c,m in slpsList:\\n configuration = configurationList[demonstrationIndex]\\n prediction = fm.predict(configuration)\\n error = 0\\n for index in range(len(slps)):\\n error += (slps[index] - prediction[index])**2\\n if(saveError == True):\\n with open(\\\"suggested_alps_fm_prediction_error.txt\\\", \\\"a\\\") as predictionErrorFile:\\n predictionErrorFile.write(str(error) + \\\"\\\\n\\\")\\n demonstrationIndex += 1\\n \\n return [allConfigurationsList, configurationList]\\n\\n# learning from visual demonstrations\\ndef lfd(plotConfigurations=False, saveError=False):\\n # expects configuration file as input\\n experiment = Experiment()\\n\\n # load network if needed\\n net = fi.get_network(experiment.descriptorSetting, IMAGE_FEATURE_PATH_PREFIX)\\n\\n # traverse all demonstration data\\n plotIndex = 0\\n slpsList = []\\n\\n # process directories in sorted order\\n sortedPaths = []\\n for dirName, subdirList, fileList in os.walk(experiment.demonstrationFolder):\\n for sortedSubdirName in sorted(subdirList):\\n sortedPaths.append(dirName + \\\"/\\\" + sortedSubdirName)\\n\\n # get slps for demonstrations\\n for sortedPath in sortedPaths:\\n for dirName, subdirList, fileList in os.walk(sortedPath):\\n # only use non empty subdirectories\\n if(len(fileList) > 0):\\n # traverse all files in this directory\\n if(bool(experiment.images) == True):\\n # read dependent values as images\\n dependentValues = [tuple(fi.get_features(net, dirName + \\\"/\\\" + fileName, experiment.descriptorSetting)) for fileName in fileList]\\n else:\\n # read dependent values as text\\n dependentValues = [tuple([eval(x) for x in open(dirName + \\\"/\\\" + fileName).readlines()[0].replace(\\\"\\\\n\\\", \\\"\\\").split(\\\" \\\")]) for fileName in fileList]\\n \\n dependentValues = tuple(misc.col_average(dependentValues))\\n #print >> sys.stderr, \\\"dependent values: \\\" + str(dependentValues)\\n \\n # edit scale data file name based on descriptor setting\\n scaleDataFileNamePrefix = experiment.scaleDataFileName.split(\\\".\\\")[0]\\n scaleDataFileNameExtension = experiment.scaleDataFileName.split(\\\".\\\")[1]\\n scaleDataFileName = scaleDataFileNamePrefix + \\\"_\\\" + experiment.abm\\n if(bool(experiment.images) == True):\\n scaleDataFileName += \\\"_\\\" + experiment.descriptorSetting.lower()\\n scaleDataFileName += \\\".\\\" + scaleDataFileNameExtension\\n \\n # scale demonstration data\\n with open(\\\"sampling/\\\" + experiment.abm + \\\"/\\\" + scaleDataFileName, \\\"r\\\") as scaleDataFile:\\n # read in order of writing in map.py\\n maxs = eval(scaleDataFile.readline().strip(\\\"\\\\n\\\"))\\n mins = eval(scaleDataFile.readline().strip(\\\"\\\\n\\\"))\\n \\n dependentMaxs = maxs[experiment.numIndependent:]\\n dependentMins = mins[experiment.numIndependent:]\\n dependentValuesMax = {i: dependentMaxs[i] for i in range(len(dependentMaxs))}\\n dependentValuesMin = {i: dependentMins[i] for i in range(len(dependentMins))}\\n dependentValues = tuple((dependentValues[index] - dependentValuesMin[index])/(dependentValuesMax[index] - dependentValuesMin[index]) if (dependentValuesMax[index] > dependentValuesMin[index]) else dependentValuesMax[index] for index in range(len(dependentValues)))\\n\\n # reduce dependent values if neccessary\\n if(eval(os.environ[\\\"REDUCED\\\"]) == 1): \\n model = cPickle.load(open(\\\"sampling/utils/transforms/\\\" + experiment.abm +\\\"_reduced_model.p\\\", \\\"rb\\\")) \\n dependentValues = model.transform(np.array(dependentValues).reshape(1, -1))\\n dependentValues = tuple(dependentValues[0])\\n\\n # if feedback is on\\n # use feedback to modify queried slps\\n if((eval(os.environ[\\\"FEEDBACK\\\"]) == 1) and (os.path.isfile(\\\"queried_slps.txt\\\") == True)):\\n with open(\\\"queried_slps.txt\\\", \\\"r\\\") as inFile:\\n queriedSlps = eval(inFile.readlines()[0])\\n slpsList.append((queriedSlps, PLOT_COLOR[plotIndex], PLOT_SHAPE[plotIndex]))\\n else:\\n slpsList.append((dependentValues, PLOT_COLOR[plotIndex], PLOT_SHAPE[plotIndex]))\\n\\n plotIndex += 1\\n\\n # default value for queried slps\\n # same as current demonstration slps\\n queriedSlps = dependentValues\\n\\n #\\n # feedback code\\n #\\n \\n # random initialization if feedback is not yet applicable\\n if(eval(os.environ[\\\"FEEDBACK\\\"]) == 1):\\n with open(\\\"feedback_iteration.txt\\\", \\\"r\\\") as inFile:\\n iteration = eval(inFile.readlines()[0])\\n if(iteration <= 2):\\n with open(\\\"regression_parameter.txt\\\", \\\"w\\\") as outFile:\\n regressionParameterLowerLimit = eval(os.environ[\\\"REGRESSION_PARAMETER_LOWER_LIMIT\\\"])\\n regressionParameterUpperLimit = eval(os.environ[\\\"REGRESSION_PARAMETER_UPPER_LIMIT\\\"])\\n regressionParameter = regressionParameterLowerLimit + (regressionParameterUpperLimit-regressionParameterLowerLimit)*random.random()\\n outFile.write(str(regressionParameter) + \\\"\\\\n\\\")\\n \\n # TODO: currently only works for a single slp suggestion\\n # check if file exists\\n if((eval(os.environ[\\\"FEEDBACK\\\"]) == 1) and (os.path.isfile(\\\"queried_slps.txt\\\") == True)):\\n # use feedback to modify queried slps\\n with open(\\\"queried_slps.txt\\\", \\\"r\\\") as inFile:\\n queriedSlps = eval(inFile.readlines()[0])\\n\\n # check if file exists\\n if(os.path.isfile(\\\"suggested_slps.txt\\\") == True):\\n with open(\\\"suggested_slps.txt\\\", \\\"r\\\") as inFile:\\n suggestedSlps = eval(inFile.readlines()[0])[0]\\n\\n # check consistency of suggested slps with current configuration\\n if(len(suggestedSlps) == len(dependentValues)):\\n #\\n # fm dataset optimization\\n #\\n \\n # scale suggested slps\\n # TODO: make this a common function since its used everywhere\\n suggestedSlps = tuple((suggestedSlps[index] - dependentValuesMin[index])/(dependentValuesMax[index] - dependentValuesMin[index]) if (dependentValuesMax[index] > dependentValuesMin[index]) else dependentValuesMax[index] for index in range(len(suggestedSlps)))\\n\\n # the differential accounts for the distance from the demonstration slps \\n # demonstration slps are currently stored as dependentValues\\n with open(\\\"feedback_iteration.txt\\\", \\\"r\\\") as inFile:\\n iteration = inFile.readlines()[0].strip(\\\"\\\\n\\\")\\n alpha = eval(os.environ[\\\"LEARNING_RATE_FM_DATASET\\\"])\\n alpha /= eval(iteration) \\n queriedSlps = tuple([queriedSlps[index] + alpha*(dependentValues[index] - suggestedSlps[index]) for index in range(len(dependentValues))])\\n \\n # write data\\n with open(\\\"feedback_input_difference.txt\\\", \\\"a\\\") as outFile:\\n outFile.write(str(np.abs(dependentValues[0] - queriedSlps[0])) + \\\"\\\\n\\\")\\n with open(\\\"feedback_output_difference.txt\\\", \\\"a\\\") as outFile:\\n outFile.write(str(dependentValues[0] - suggestedSlps[0]) + \\\"\\\\n\\\")\\n\\n #\\n # regression parameter optimization initialization\\n #\\n\\n # if just one iteration so far, move in a random direction\\n with open(\\\"feedback_iteration.txt\\\", \\\"r\\\") as inFile:\\n iteration = eval(inFile.readlines()[0])\\n if(iteration > 2):\\n # now we do feedback stuff\\n with open(\\\"feedback_regression_parameter.txt\\\", \\\"r\\\") as feedbackRegressionParameterFile:\\n with open(\\\"feedback_output_difference.txt\\\", \\\"r\\\") as feedbackOutputDifferenceFile:\\n regressionParameterLines = feedbackRegressionParameterFile.readlines()\\n outputDifferenceLines = feedbackOutputDifferenceFile.readlines()\\n currentRegressionParameter = eval(regressionParameterLines[-1])\\n previousRegressionParameter = eval(regressionParameterLines[-2])\\n regressionParameterChange = currentRegressionParameter - previousRegressionParameter\\n currentOutputDifference = eval(outputDifferenceLines[-1])\\n previousOutputDifference = eval(outputDifferenceLines[-2])\\n outputDifferenceChange = currentOutputDifference - previousOutputDifference\\n \\n alpha = eval(os.environ[\\\"LEARNING_RATE_REGRESSION_PARAMETER\\\"])\\n alpha /= iteration\\n\\n # if output difference is negative, we move in direction to change\\n nextRegressionParameter = currentRegressionParameter - regressionParameterChange*outputDifferenceChange*alpha\\n with open(\\\"regression_parameter.txt\\\", \\\"w\\\") as outFile:\\n outFile.write(str(nextRegressionParameter) + \\\"\\\\n\\\")\\n \\n # TODO: currently assumes only a single demonstration at a time. fix for multiple demonstrations\\n with open(\\\"queried_slps.txt\\\", \\\"w\\\") as outFile:\\n outFile.write(str(queriedSlps) + \\\"\\\\n\\\")\\n \\n configurationListPerRun = []\\n\\n for run in range(RUNS_PER_SLPS):\\n # run and query slps\\n [allConfigurationsList, configurationList] = experiment.run(slpsList, plotConfigurations=plotConfigurations, saveError=saveError)\\n\\n # scale configurations back from normalized space\\n independentMaxs = maxs[:experiment.numIndependent]\\n independentMins = mins[:experiment.numIndependent]\\n independentValuesMax = {i: independentMaxs[i] for i in range(len(independentMaxs))}\\n independentValuesMin = {i: independentMins[i] for i in range(len(independentMins))}\\n normalizedConfigurationList = []\\n for configuration in configurationList:\\n normalizedConfigurationList.append(tuple([independentValuesMin[cIndex] + (independentValuesMax[cIndex]-independentValuesMin[cIndex])*configuration[cIndex] for cIndex in range(len(configuration))]))\\n configurationList = tuple(normalizedConfigurationList)\\n configurationListPerRun.append(configurationList)\\n \\n # return configurations\\n print >> sys.stderr, \\\"configurationList: \\\" + str(configurationList)\\n\\n # compute average over runs\\n averageConfigurationList = []\\n for slpsIndex in range(len(slpsList)):\\n slpsConfigurations = tuple([configurationListOneRun[slpsIndex] for configurationListOneRun in configurationListPerRun])\\n averageConfigurationList.append(tuple(misc.col_average(slpsConfigurations)))\\n\\n print >> sys.stderr, \\\"averageConfigurationList: \\\" + str(averageConfigurationList)\\n sys.stdout.write(str(averageConfigurationList))\\n\\n if(\\\"smt\\\" in experiment.regression):\\n # circumvent excessive prints from smt\\n with open(\\\"temp_average_configuration_list.txt\\\", \\\"w\\\") as outFile:\\n outFile.write(str(averageConfigurationList) + \\\"\\\\n\\\")\\n\\n# main function\\ndef main():\\n lfd(plotConfigurations=False, saveError=True)\\n\\n# execute main\\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.80923992395401,\n \"alphanum_fraction\": 0.8159463405609131,\n \"avg_line_length\": 48.703704833984375,\n \"blob_id\": \"867655188aa23578bcdd69bc7ce77b29870b3d76\",\n \"content_id\": \"4c735cf47aaed6e293b4cb5b9959c5e2a5fa13f8\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 1342,\n \"license_type\": \"no_license\",\n \"max_line_length\": 117,\n \"num_lines\": 27,\n \"path\": \"/data/dataset_analysis/dataset_analysis.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# demonstration specification\\nexport DEMONSTRATION_ID=1\\n\\n# steps across dimensions\\nexport DIM1_STEP=0.1\\nexport DIM2_STEP=0.1\\n\\n# clean old datasets\\nrm -f $DEMONSTRATION_ID/predictions_best/slp_dataset.txt\\nrm -f $DEMONSTRATION_ID/predictions_second_best/slp_dataset.txt\\nrm -f $DEMONSTRATION_ID/predictions_third_best/slp_dataset.txt\\nrm -f $DEMONSTRATION_ID/predictions_worst/slp_dataset.txt\\nrm -f $DEMONSTRATION_ID/predictions_second_worst/slp_dataset.txt\\nrm -f $DEMONSTRATION_ID/predictions_third_worst/slp_dataset.txt\\n\\n# make a copy of the datasets\\n# uniform name for ease of use\\n\\ncp $DEMONSTRATION_ID/predictions_best/slp_dataset* $DEMONSTRATION_ID/predictions_best/slp_dataset.txt\\ncp $DEMONSTRATION_ID/predictions_second_best/slp_dataset* $DEMONSTRATION_ID/predictions_second_best/slp_dataset.txt\\ncp $DEMONSTRATION_ID/predictions_third_best/slp_dataset* $DEMONSTRATION_ID/predictions_third_best/slp_dataset.txt\\ncp $DEMONSTRATION_ID/predictions_worst/slp_dataset* $DEMONSTRATION_ID/predictions_worst/slp_dataset.txt\\ncp $DEMONSTRATION_ID/predictions_second_worst/slp_dataset* $DEMONSTRATION_ID/predictions_second_worst/slp_dataset.txt\\ncp $DEMONSTRATION_ID/predictions_third_worst/slp_dataset* $DEMONSTRATION_ID/predictions_third_worst/slp_dataset.txt\\n\\n# invoke python script\\npython dataset_analysis.py $DEMONSTRATION_ID $DIM1_STEP $DIM2_STEP\\n\"\n },\n {\n \"alpha_fraction\": 0.5875629782676697,\n \"alphanum_fraction\": 0.5924505591392517,\n \"avg_line_length\": 34.4640007019043,\n \"blob_id\": \"978978042f347b82c80ef1a467ac77c89710c4c0\",\n \"content_id\": \"0dd2c28a3cbaf433e292feaf53f2e43974e9e05e\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 13299,\n \"license_type\": \"no_license\",\n \"max_line_length\": 181,\n \"num_lines\": 375,\n \"path\": \"/src/amf/rm.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# rm.py contains the ReverseMapping\\n\\nimport misc\\nimport numpy\\nimport sys\\nimport multiprocessing as mp\\nimport time\\nimport itertools\\nimport random\\nimport os\\nimport gzip\\nimport cPickle\\n\\n# parallel processing constants\\nBATCH_SIZE = 8\\n\\n# hack to avoid rounding error\\ndef _round(n):\\n return numpy.round(n, 5)\\n\\nclass RMSimplex(object):\\n # Simplexes are kind of immutable...\\n\\n # The LAST variable of each corner vector is the dependent variable\\n # (this is important)\\n\\n def __init__(self, corners, slpsList):\\n \\\"\\\"\\\"\\n A simplex.\\n corners should have the form ((ind0, ind1, ...), (dep1, dep2, dep3))\\n \\\"\\\"\\\"\\n # We already know which corners are adjanced to which\\n # (each corner shares an edge with every other corner)\\n\\n self._corners = tuple(corners)\\n self._inds = tuple( corner[0] for corner in corners )\\n self._deps = tuple( corner[1] for corner in corners )\\n self._min = misc.col_min(self._deps)\\n self._max = misc.col_max(self._deps) \\n self._edges = []\\n\\n # initialize edges only if needed\\n # only if this simplex contains the query slp\\n # otherwise the slp has no utility\\n isContains = False\\n \\n for querySlps,c,m in slpsList:\\n for index in range(len(querySlps)):\\n isContains = isContains or self.contains(index, querySlps[index])\\n \\n if(isContains == True):\\n for cidx in range(len(self._corners)):\\n corner = self._corners[cidx]\\n for oidx in range(cidx + 1, len(self._corners)):\\n other_corner = self._corners[oidx]\\n self._edges.append((corner, other_corner))\\n self._edges = tuple(self._edges)\\n\\n @property\\n def corners(self):\\n return self._corners\\n\\n @property\\n def edges(self):\\n return self._edges\\n\\n def contains(self, slp_num, value):\\n \\\"\\\"\\\"Determines if the value for slp_num idx lives inside this simplex.\\\"\\\"\\\"\\n return ((value >= self._min[slp_num]) and (value <= self._max[slp_num]))\\n\\n def intersections(self, slp_num, value):\\n \\\"\\\"\\\"Returns the approximate places that the value plane intersects with this simplex\\\"\\\"\\\"\\n inters = []\\n for edgeNumber, (scorner, bcorner) in enumerate(self.edges):\\n # maintain order of scorner <= bcorner\\n if(scorner[1][slp_num] > bcorner[1][slp_num]):\\n tempCorner = bcorner\\n bcorner = scorner\\n scorner = tempCorner\\n\\n # check to see that this value is inside these edges\\n if not ((value[slp_num] >= scorner[1][slp_num]) and (value[slp_num] <= bcorner[1][slp_num])):\\n continue\\n\\n # this is the distance between the 2 corner values\\n distance = bcorner[1][slp_num] - scorner[1][slp_num]\\n\\n if(distance == 0):\\n # return any end point value\\n inter = tuple( (1.0)*s for s, b in zip(scorner[0], bcorner[0]) )\\n else:\\n # this is the distance from 'value' from the smaller corner\\n # if value is close to smaller, this ratio will be close to zero\\n # if value is closer to bigger, this ratio will be close to one\\n ratio = (value[slp_num] - scorner[1][slp_num]) / distance\\n inter = tuple( (1.0-ratio)*s + ratio * b for s, b in zip(scorner[0], bcorner[0]) )\\n \\n inters.append([edgeNumber, inter])\\n\\n return tuple(inters)\\n\\n def __repr__(self):\\n return 'simplex:' + repr(self._edges)\\n\\nclass ReverseMapping(object):\\n def __init__(self, trained_fm, ranges, granularity, slpsList):\\n\\n self.fm = trained_fm\\n self.granularity = granularity\\n self.ranges = ranges\\n self.steps = tuple( (ma - mi) / float(self.granularity) for mi, ma in self.ranges )\\n self.slpsList = slpsList\\n\\n # disk write stuff\\n self.writeToDisk = eval(os.environ[\\\"WRITE_TO_DISK\\\"])\\n self.allIntersectionsCalls = 0\\n \\n # the dimensionality of the configuration space\\n self.dim = len(ranges)\\n\\n # the dimensionality of the SLP space\\n self.slp_dim = None\\n\\n self.__bin__ = misc.binary(self.dim)\\n\\n self.knots = {}\\n self.simplexes = []\\n print >> sys.stderr, 'building knots'\\n startTime = time.time()\\n self.build_knots()\\n stopTime = time.time()\\n print >> sys.stderr, 'done building knots, now building simplexes' + \\\" [\\\" + str(len(self.knots)) + \\\" in \\\" + str(stopTime-startTime) + \\\" seconds]\\\"\\n startTime = time.time() \\n self.build_simplexes()\\n stopTime = time.time()\\n print >> sys.stderr, 'done building simplexes' + \\\" [in \\\" + str(stopTime-startTime) + \\\" seconds]\\\"\\n\\n def build_knots(self):\\n self.build_knots_inds()\\n self.build_knots_deps()\\n\\n def build_knots_inds(self):\\n values = [list(numpy.arange(self.ranges[curDim][0], self.ranges[curDim][1], (self.ranges[curDim][1] - self.ranges[curDim][0])/self.granularity)) for curDim in range(self.dim)]\\n indexes = [range(len(values[i])) for i in range(len(values))]\\n valuesPoints = [x for x in itertools.product(*values)]\\n indexesPoints = [x for x in itertools.product(*indexes)]\\n\\n # sort items for correspondence\\n valuesPoints.sort()\\n indexesPoints.sort()\\n\\n # assign key-value pairs\\n self.knots = dict(zip(indexesPoints, valuesPoints))\\n\\n def process_build_knots_deps(self, q, new_idx):\\n new_vector = self.knots[new_idx]\\n predicted = self.fm.predict(new_vector)\\n q.put([new_idx, predicted])\\n\\n def build_knots_deps_parallel(self):\\n # setup a list of processes that we want to run\\n q = mp.Queue()\\n processes = [mp.Process(target=self.process_build_knots_deps, args=(q, new_idx)) for new_idx in self.knots] \\n batchSize = BATCH_SIZE\\n batches = [processes[i:i+batchSize] for i in range(0, len(processes), batchSize)]\\n\\n for batch in batches:\\n # run processes\\n for p in batch:\\n p.start() \\n for p in batch:\\n [new_idx, predicted] = q.get()\\n self.knots[new_idx] = predicted\\n for p in batch:\\n p.join() \\n\\n # use last predicted point in dependent parameter space for length\\n # we can use any key for this\\n self.slp_dim = len(self.knots.values()[0])\\n\\n def build_knots_deps(self):\\n # predict for each point in independent parameter space\\n for new_idx in self.knots:\\n new_vector = self.knots[new_idx]\\n predicted = self.fm.predict(new_vector)\\n self.knots[new_idx] = predicted\\n \\n # use last predicted point in dependent parameter space for length\\n self.slp_dim = len(predicted)\\n \\n def build_simplexes(self, col=None):\\n for key in sorted(self.knots.keys()):\\n # check to see if this root is on the fringe and thus not actually a root\\n for idx in key:\\n if idx == (self.granularity - 1):\\n break\\n else:\\n # extract simplexes\\n # only keep those simplexes which are actually useful\\n sims = self.simplexes_from_cube(key)\\n sims = [x for x in sims if(len(x._edges) > 0)]\\n\\n if(self.writeToDisk == 1):\\n # save to compressed file\\n with gzip.open(\\\"sims.pgz\\\", 'ab') as simsFile:\\n cPickle.dump(sims, simsFile, -1)\\n else:\\n # save to ram\\n self.simplexes.extend(sims)\\n\\n '''\\n # karan: testing downsampling\\n sampleAmount = 100000\\n\\n if(len(self.simplexes) > sampleAmount):\\n # downsample\\n # for computational tractability\\n sampledSimplexes = dict(random.sample(self.simplexes.items(), sampleAmount))\\n self.simplexes = sampledSimplexes\\n '''\\n \\n def wiggle(self, root):\\n \\\"\\\"\\\" Returns the closest point to this root that is an actual root. \\\"\\\"\\\"\\n distance, knot, knotKey = min((misc.distance(root, self.knots[knotKey]), self.knots[knotKey], knotKey) for knotKey in self.knots.keys())\\n print >> sys.stderr, \\\"BADNESS: Wiggling point %s to knot %s\\\" % (repr(root), repr(knot))\\n\\n return [knot, knotKey]\\n\\n def cube_at_root(self, root):\\n \\\"\\\"\\\" Returns the members of this cube with this point as its root. \\\"\\\"\\\"\\n corners = []\\n\\n for perm in self.__bin__:\\n corner = []\\n for idx in xrange(self.dim):\\n corner.append(perm[idx] + root[idx])\\n corner = tuple(corner)\\n val = self.knots[corner]\\n corners.append((self.translate_idx(corner), val))\\n\\n return tuple(corners)\\n\\n def translate_idx(self, vec_idx):\\n \\\"\\\"\\\" Translates a index to a configuration\\\"\\\"\\\"\\n corner = []\\n for idx in xrange(self.dim):\\n corner.append(vec_idx[idx] * self.steps[idx] + self.ranges[idx][0])\\n\\n return tuple(corner)\\n\\n def simplexes_from_cube(self, root):\\n cube_corners = self.cube_at_root(root)\\n # cube_at_root always has the first right corner root as [0]\\n # at [-1] we have the other right corner root\\n\\n # special case for 1 dimensional configuration space\\n if self.dim == 1:\\n return (RMSimplex(cube_corners, self.slpsList),)\\n else:\\n return RMSimplex(cube_corners[:-1], self.slpsList), RMSimplex(cube_corners[1:], self.slpsList)\\n\\n def process_intersections(self, q, simplex, param_num, value):\\n out = []\\n inter = simplex.intersections(param_num, value)\\n if len(inter) > 0:\\n out.append(inter)\\n\\n q.put(out)\\n \\n def intersections_parallel(self, param_num, value):\\n out = []\\n\\n # setup a list of processes that we want to run\\n q = mp.Queue()\\n processes = [mp.Process(target=self.process_intersections, args=(q, simplex, param_num, value)) for simplex in self.simplexes] \\n batchSize = BATCH_SIZE\\n batches = [processes[i:i+batchSize] for i in range(0, len(processes), batchSize)]\\n\\n for batch in batches:\\n # run processes\\n for p in batch:\\n p.start() \\n for p in batch:\\n out += q.get()\\n for p in batch:\\n p.join() \\n\\n return tuple(out)\\n \\n def intersections(self, param_num, value):\\n out = []\\n\\n if(self.writeToDisk == 1):\\n # read from disk\\n with gzip.open(\\\"sims.pgz\\\", 'rb') as simsFile:\\n with gzip.open(\\\"configuration_\\\" + str(self.allIntersectionsCalls) + \\\".pgz\\\", 'ab') as configurationFile:\\n while 1:\\n try:\\n sims = cPickle.load(simsFile)\\n for simplex in sims:\\n inter = simplex.intersections(param_num, value)\\n if len(inter) > 0:\\n for configurations in inter:\\n for configuration in configurations:\\n cPickle.dump(inter, configurationFile, -1)\\n\\t except EOFError:\\n break\\n else:\\n # read from ram\\n for simplexNumber, simplex in enumerate(self.simplexes):\\n inter = simplex.intersections(param_num, value)\\n if len(inter) > 0:\\n out.append([simplexNumber, inter])\\n\\n return tuple(out)\\n\\n def process_all_intersections(self, q, d, value):\\n q.put(self.intersections(d, value))\\n \\n def all_intersections_parallel(self, value):\\n allIntersections = []\\n \\n # setup a list of processes that we want to run\\n q = mp.Queue()\\n processes = [mp.Process(target=self.process_all_intersections, args=(q, d, value)) for d in range(self.slp_dim)] \\n batchSize = BATCH_SIZE\\n batches = [processes[i:i+batchSize] for i in range(0, len(processes), batchSize)]\\n\\n for batch in batches:\\n # run processes\\n for p in batch:\\n p.start() \\n for p in batch:\\n allIntersections.append(q.get())\\n for p in batch:\\n p.join() \\n\\n if(len(allIntersections[0]) > 0):\\n return tuple(allIntersections)\\n else:\\n # empty intersection set. need to wiggle\\n print >> sys.stderr, \\\"BADNESS: PARALLELISM NOT USEFUL BECAUSE OF WIGGLE!\\\"\\n [wiggledValue, wiggledKey] = self.wiggle(value)\\n return [[[tuple([x/float(self.granularity) for x in wiggledKey])]]]\\n \\n def all_intersections(self, value):\\n # increment call counter\\n self.allIntersectionsCalls += 1\\n \\n \\\"\\\"\\\" Finds intersections with each provided SLP \\\"\\\"\\\"\\n allIntersections = [self.intersections(d, value) for d in range(self.slp_dim)]\\n\\n if(len(allIntersections[0]) > 0):\\n return tuple(allIntersections)\\n else:\\n # empty intersection set. need to wiggle\\n [wiggledValue, wiggledKey] = self.wiggle(value)\\n return [[[tuple([x/float(self.granularity) for x in wiggledKey])]]]\\n \\n def distance_to(self, param_num, config, value):\\n # i'm going to cheat for now and just pick the closest intersecting corner\\n ints = self.intersections(param_num, value)\\n\\n dists = []\\n for simplex in ints:\\n for i in simplex:\\n dists.append( (misc.distance(i, config)) )\\n \\n if len(dists) == 0:\\n print >> sys.stderr, \\\"BADNESS: no intersections found for %s\\\" % repr(value)\\n dist = None\\n else:\\n dist = min(dists)\\n\\n return dist\\n\"\n },\n {\n \"alpha_fraction\": 0.7288135886192322,\n \"alphanum_fraction\": 0.7288135886192322,\n \"avg_line_length\": 13.75,\n \"blob_id\": \"70e2bd8c28ffa2d4d8281f4a0f7c39228fb33b5b\",\n \"content_id\": \"0a7fa60de76adc20c8a52e56fcce98102fdeb963\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 118,\n \"license_type\": \"no_license\",\n \"max_line_length\": 23,\n \"num_lines\": 8,\n \"path\": \"/src/sampling.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# load configuration\\nsource configuration.sh\\n\\n# sample data\\necho \\\"=>sampling data\\\"\\ncd sampling\\n./sampling.sh\\ncd ../..\\n\"\n },\n {\n \"alpha_fraction\": 0.6308630108833313,\n \"alphanum_fraction\": 0.6374296545982361,\n \"avg_line_length\": 26.50967788696289,\n \"blob_id\": \"d5e3cd2f68cc6adb325d89933359324154c70220\",\n \"content_id\": \"26f32adeb7f1598db01cad8611ba640f9415ecfd\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 4264,\n \"license_type\": \"no_license\",\n \"max_line_length\": 140,\n \"num_lines\": 155,\n \"path\": \"/src/sampling/eum/sample_eum_parallel.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import pandas as pd\\nimport numpy as np\\nimport sys\\nimport multiprocessing as mp\\n\\nfrom negotiation_model import *\\nfrom bdm_agent import *\\n\\n# parallel processing constants\\nBATCH_SIZE = 8\\n\\n# seed this for fixed environment\\n# for better replication of results\\nrandom.seed(0)\\n\\n# constants\\nINDEPENDENT_VARIABLES = 2\\nDEPENDENT_VARIABLES = 2\\nRUNS = 10000\\nSAMPLES = 10\\nTIME_LAPSE = 200\\n\\n# process single run data\\ndef process_run(q, run, qValue, tValue, bookData, salienceValues):\\n try:\\n line = main_eum(qValue, tValue, bookData, salienceValues)\\n except:\\n line = \\\"\\\"\\n \\n # add newline for clean output\\n line = \\\"\\\\n\\\" + line\\n \\n # write to queue\\n q.put(line)\\n\\ndef get_salience_values(timeLapse, numberOfAgents):\\n salienceValues = np.zeros((timeLapse, numberOfAgents))\\n\\n for stepNumber in range(timeLapse):\\n for agentNumber in range(numberOfAgents):\\n salienceValues[stepNumber][agentNumber] = random.random()\\n\\n return salienceValues\\n\\n# Defining the model objects\\nclass BDMActor(NegotiationActor):\\n DecisionClass = BDM_Agent\\n\\nclass NegotiationModel_(NegotiationModel):\\n # Variables for median caching\\n median_computed_last = -1\\n median = -1\\n \\n def find_median(self):\\n if self.median_computed_last != self.schedule.steps:\\n self.median = super().find_median()\\n self.median_computed_last = self.schedule.steps\\n return self.median\\n\\nclass ModelOutput:\\n def __init__(self, model):\\n '''\\n Store data from model run.\\n '''\\n self.agent_vars = model.datacollector.get_agent_vars_dataframe()\\n self.model_vars = model.datacollector.get_model_vars_dataframe()\\n self.log = model.log\\n\\ndef load_data():\\n # Load data \\n bookData = pd.read_csv(\\\"BDM_ColdWar.csv\\\")\\n bookData.Position = (bookData.Position + 100)/200\\n\\n return bookData\\n\\ndef main_eum(qValue, tValue, bookData, salienceValues):\\n # define agents\\n agents = []\\n\\n for i, row in bookData.iterrows():\\n newAgent = BDMActor(row.Country, row.Capability, row.Position, 1)\\n newAgent.decision_model.Q = qValue\\n newAgent.decision_model.T = tValue\\n newAgent.salience = salienceValues[i]\\n agents.append(newAgent)\\n\\n # instantiate model\\n model = NegotiationModel_(agents)\\n\\n # run model\\n for stepNumber in range(TIME_LAPSE):\\n agentNumber = 0\\n for agent in model.agents:\\n #agent.salience = random.random()\\n agent.salience = salienceValues[stepNumber][agentNumber]\\n agentNumber += 1 \\n model.step()\\n\\n # collect data for next steps\\n dependentValues = []\\n modelOutput = ModelOutput(model)\\n dependentValues.append(list(modelOutput.model_vars[\\\"Median\\\"][-SAMPLES-1:]))\\n dependentValues.append(list(modelOutput.model_vars[\\\"Mean\\\"][-SAMPLES-1:]))\\n \\n # print line corresponding to this execution\\n line = str(qValue) + \\\" \\\" + str(tValue)\\n for dependentValue in dependentValues:\\n line += \\\" \\\" + str(dependentValue).replace(\\\"[\\\", \\\"\\\").replace(\\\"]\\\", \\\"\\\").replace(\\\" \\\", \\\"\\\")\\n\\n return line\\n\\ndef main():\\n # header string\\n print(\\\"i\\\"*INDEPENDENT_VARIABLES + \\\"d\\\"*DEPENDENT_VARIABLES)\\n\\n # read data\\n bookData = load_data()\\n\\n # get salience values\\n [rows, columns] = bookData.shape\\n salienceValues = get_salience_values(TIME_LAPSE, rows)\\n\\n qValues = []\\n tValues = []\\n \\n for run in range(RUNS):\\n # sample random ALPs\\n qValue = random.random()\\n tValue = random.random()\\n\\n qValues.append(qValue)\\n tValues.append(tValue)\\n \\n # setup a list of processes that we want to run\\n q = mp.Queue()\\n processes = [mp.Process(target=process_run, args=(q, run, qValues[run], tValues[run], bookData, salienceValues)) for run in range(RUNS)]\\n batchSize = BATCH_SIZE\\n batches = [processes[i:i+batchSize] for i in range(0, len(processes), batchSize)]\\n\\n outFile = open(\\\"output.txt\\\", \\\"a\\\")\\n \\n for batch in batches:\\n # run processes\\n for p in batch:\\n p.start()\\n for p in batch:\\n line = q.get()\\n outFile.write(line)\\n # exit the completed processes\\n for p in batch:\\n p.join()\\n\\n outFile.close()\\n\\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.5772079825401306,\n \"alphanum_fraction\": 0.5948718190193176,\n \"avg_line_length\": 21.5,\n \"blob_id\": \"172f17fb56c1f893cf528f867f348c507707bc76\",\n \"content_id\": \"5a0207a4e6641ab024fb9863a512420f9f05a4ee\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1755,\n \"license_type\": \"no_license\",\n \"max_line_length\": 92,\n \"num_lines\": 78,\n \"path\": \"/src/sampling/schelling/sample_schelling.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from model import SchellingModel\\nimport random\\n\\n# only use three of the several ALPs\\n# density (0,1)\\n# minority_pc (0,1)\\n\\n# use one model-level SLP\\n# happy\\n\\n# constant ALPs\\nHEIGHT = 20\\nWIDTH = 20\\nHOMOPHILY = 4\\n\\n# constants\\n# max_iters in the example is specified as 1000\\n# so we run for 1000 steps\\nINDEPENDENT_VARIABLES = 2\\nDEPENDENT_VARIABLES = 1\\nRUNS = 10000\\nSAMPLES = 10\\nTIME_LAPSE = 1000\\n\\ndef main_schelling(density, minorityPc):\\n # instantiate and run model\\n model = SchellingModel(\\n height=HEIGHT,\\n width=WIDTH,\\n density=density,\\n minority_pc=minorityPc,\\n homophily=HOMOPHILY)\\n\\n # run time lapse\\n for i in range(TIME_LAPSE):\\n try:\\n # step\\n model.step()\\n except:\\n # no empty cells\\n pass\\n \\n # collect data for next steps\\n dependentValues = []\\n\\n for i in range(SAMPLES):\\n try:\\n # step\\n model.step()\\n except:\\n # saturated\\n # no empty cells\\n pass\\n\\n # read data\\n data = model.datacollector.get_model_vars_dataframe() \\n dependentValues.append(list(data.happy)[-SAMPLES-1:])\\n\\n # print line corresponding to this execution\\n line = str(density) + \\\" \\\" + str(minorityPc)\\n for dependentValue in dependentValues:\\n line += \\\" \\\" + str(dependentValue).replace(\\\"[\\\", \\\"\\\").replace(\\\"]\\\", \\\"\\\").replace(\\\" \\\", \\\"\\\")\\n print(line)\\n\\n \\ndef main():\\n # header string\\n print(\\\"i\\\"*INDEPENDENT_VARIABLES + \\\"d\\\"*DEPENDENT_VARIABLES)\\n\\n for run in range(RUNS):\\n # sample random ALPs\\n density = random.random() \\n minorityPc = random.random() \\n \\n # run model using those ALPs\\n main_schelling(density, minorityPc)\\n\\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.6653721928596497,\n \"alphanum_fraction\": 0.6737864017486572,\n \"avg_line_length\": 29.294116973876953,\n \"blob_id\": \"106a3afa59283bbc5813e8bc9048073d87b51f60\",\n \"content_id\": \"f7c189ef2ce03983c7a3db9a1fa8665624f9f79d\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1545,\n \"license_type\": \"no_license\",\n \"max_line_length\": 111,\n \"num_lines\": 51,\n \"path\": \"/app/es_server.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from elasticsearch import Elasticsearch\\nfrom image_match.elasticsearch_driver import SignatureES\\nimport glob\\nfrom flask import Flask, render_template, request\\nimport os\\nimport time\\n\\napp = Flask(__name__)\\napp.config[\\\"INPUT_FILE\\\"] = \\\"sketches/demonstration.png\\\"\\napp.config[\\\"ABM\\\"] = \\\"flocking\\\"\\napp.config[\\\"SES\\\"] = SignatureES(Elasticsearch(), distance_cutoff=1.0)\\n\\n@app.route('/initialize', methods=[\\\"GET\\\"])\\ndef initialize():\\n # clear existing indexes\\n os.system(\\\"curl -XDELETE 'http://localhost:9200/_all'\\\")\\n\\n # sleep to be sure of synchronization\\n time.sleep(60)\\n \\n # index items\\n folderName = \\\"/home/karan/storage/workspaces/bitbucket/_swarm-lfd-data/\\\" + app.config[\\\"ABM\\\"] + \\\"/images_10\\\"\\n #folderName = \\\"/home/karan/storage/workspaces/bitbucket/_swarm-lfd-data/\\\" + app.config[\\\"ABM\\\"] + \\\"/images\\\"\\n allFileNames = glob.glob(folderName + \\\"/*.*\\\")\\n\\n for imageName in sorted(allFileNames):\\n print('elastic search initialize: ' + imageName)\\n app.config[\\\"SES\\\"].add_image(imageName)\\n\\n # sleep to be sure of synchronization\\n time.sleep(60)\\n \\n return \\\"initialization complete\\\"\\n\\n\\n@app.route('/match', methods=[\\\"GET\\\"])\\ndef match():\\n results = app.config[\\\"SES\\\"].search_image(app.config[\\\"INPUT_FILE\\\"])\\n\\n distanceResults = []\\n for result in results:\\n distanceResults.append((result['dist'], result['path']))\\n\\n # get top result\\n [distance, path] = sorted(distanceResults)[0]\\n os.system('cp ' + path + ' matches/')\\n \\n return \\\"matched\\\"\\n\\n# code run at app start\\ninitialize()\\n\"\n },\n {\n \"alpha_fraction\": 0.6937590837478638,\n \"alphanum_fraction\": 0.698113203048706,\n \"avg_line_length\": 25.461538314819336,\n \"blob_id\": \"e035ad7ba763933b646f7bbf7f9c261f52b7789c\",\n \"content_id\": \"dc860b5c8baa82f6f21ccb359bcc3ff3eafbc5b8\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 689,\n \"license_type\": \"no_license\",\n \"max_line_length\": 73,\n \"num_lines\": 26,\n \"path\": \"/src/lfd.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env bash\\n# load configuration\\nsource configuration.sh\\n\\n# run multiple times to generate multiple sets of images\\nexport RUNS_PER_INPUT=1\\n\\n# remove existing data\\nrm -f suggested_alps.txt\\nrm -f suggested_alps_fm_prediction_error.txt\\n\\nfor i in $(seq \\\"$RUNS_PER_INPUT\\\")\\ndo\\n # run experiment\\n echo \\\"=>running experiment\\\"\\n export ALP_CONFIGURATIONS=$(python lfd.py)\\n echo $ALP_CONFIGURATIONS >> suggested_alps.txt\\n \\n # circumvent excessive prints from smt \\n mv temp_average_configuration_list.txt suggested_alps.txt 2>/dev/null\\n \\n if [ $WRITE_TO_DISK == 1 ]; then\\n # remove compressed pickle files generated for this run\\n rm *.pgz\\n fi\\ndone\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6831682920455933,\n \"alphanum_fraction\": 0.7623762488365173,\n \"avg_line_length\": 24.25,\n \"blob_id\": \"a54ecdd6f4685d3e1936514766d0a201aa8a708e\",\n \"content_id\": \"a509b2a730edd34ff284c45c582bca0af6df31d2\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 101,\n \"license_type\": \"no_license\",\n \"max_line_length\": 36,\n \"num_lines\": 4,\n \"path\": \"/app/run_es_server.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"sudo service elasticsearch start\\n\\nexport FLASK_APP=es_server.py\\nflask run --host=0.0.0.0 --port=5001\\n\"\n },\n {\n \"alpha_fraction\": 0.5775167942047119,\n \"alphanum_fraction\": 0.5892617702484131,\n \"avg_line_length\": 28.799999237060547,\n \"blob_id\": \"6dd53eb41161c6b2279e68b5d703786da0993126\",\n \"content_id\": \"0a54a5e6f680756fcb7a659d7b5332ebb0e23535\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Java\",\n \"length_bytes\": 2980,\n \"license_type\": \"no_license\",\n \"max_line_length\": 115,\n \"num_lines\": 100,\n \"path\": \"/src/sampling/wolf_sheep/predict_wolf_sheep.java\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"/* package required for calls from parent folder */\\npackage sampling.wolf_sheep;\\n\\n/* libraries included */\\nimport org.nlogo.headless.HeadlessWorkspace;\\nimport java.io.*;\\nimport java.util.*;\\n\\npublic class predict_wolf_sheep {\\n public static void main(String[] argv) {\\n HeadlessWorkspace workspace =\\n HeadlessWorkspace.newInstance() ;\\n try {\\n\\t// evaluate to list\\n\\tString alpConfigurationListString = argv[0];\\n\\tList alpConfigurationList = new ArrayList(Arrays.asList(alpConfigurationListString.split(\\\"\\\\\\\\(\\\")));\\n\\n\\n\\t// skip first entry, which is empty\\n\\tfor (int i = 1; i < alpConfigurationList.size(); i++) {\\n\\t\\tString alpConfiguration = alpConfigurationList.get(i);\\n\\n\\t\\t// clean string\\n\\t\\talpConfiguration = alpConfiguration.replace(\\\")\\\", \\\"\\\").replace(\\\"]\\\", \\\"\\\");\\n\\n\\t\\t// extract variables\\n\\t\\tArrayList alpValues = new ArrayList(Arrays.asList(alpConfiguration.split(\\\",\\\")));\\n\\n workspace.open(\\\"sampling/wolf_sheep/Wolf Sheep Predation.nlogo\\\");\\n\\n \\tworkspace.command(\\\"set grass? true\\\");\\n workspace.command(\\\"set initial-number-sheep 100\\\");\\n workspace.command(\\\"set initial-number-wolves 50\\\"); \\n \\n \\t// wolf-reproduce\\n \\tString k1 = alpValues.get(0);\\n \\t\\n \\t// sheep-reproduce\\n\\t\\t\\t\\tString k2 = alpValues.get(1);\\n\\t\\t\\t\\t\\n\\t\\t\\t\\t//wolf-gain-from-food\\n\\t\\t\\t\\tString k3 = alpValues.get(2);\\n\\t\\t\\t\\t\\n\\t\\t\\t\\t//sheep-gain-from-food\\n\\t\\t\\t\\tString k4 = alpValues.get(3);\\n\\t\\t\\t\\t\\n\\t\\t\\t\\t//grass-regrowth-time\\n\\t\\t\\t\\tString k5 = alpValues.get(4);\\n\\n\\n\\t\\t\\t\\tworkspace.command(\\\"set wolf-reproduce \\\" + k1);\\n\\t\\t\\t\\tworkspace.command(\\\"set sheep-reproduce \\\" + k2);\\n\\t\\t\\t\\tworkspace.command(\\\"set wolf-gain-from-food \\\" + k3);\\n\\t\\t\\t\\tworkspace.command(\\\"set sheep-gain-from-food \\\" + k4);\\n\\t\\t\\t\\tworkspace.command(\\\"set grass-regrowth-time \\\" + k5);\\n\\n \\tworkspace.command(\\\"setup\\\");\\n \\tworkspace.command(\\\"repeat 500 [ go ]\\\");\\n \\t\\n \\tString wolves = \\\"\\\" + workspace.report(\\\"count wolves\\\");\\n \\tString sheep = \\\"\\\" + workspace.report(\\\"count sheep\\\");\\n \\t\\n \\tfor(int j = 0; j < 750; j++) {\\n \\t\\tworkspace.command(\\\"go\\\");\\n \\t\\twolves += \\\",\\\" + workspace.report(\\\"count wolves\\\");\\n \\t\\tsheep += \\\",\\\" + workspace.report(\\\"count sheep\\\");\\n \\t}\\n \\t\\n\\t\\tString[] wolvesValues = wolves.split(\\\",\\\");\\n\\t\\tfloat sum = 0;\\n\\t\\tfloat count = 0;\\n\\t\\tfor(String value: wolvesValues){\\n\\t\\t\\tsum += Float.parseFloat(value);\\n\\t\\t\\tcount += 1;\\n\\t\\t}\\n\\t\\tfloat wolvesAverage = sum/count;\\n\\n\\t\\tString[] sheepValues = sheep.split(\\\",\\\");\\n\\t\\tsum = 0;\\n\\t\\tcount = 0;\\n\\t\\tfor(String value: sheepValues){\\n\\t\\t\\tsum += Float.parseFloat(value);\\n\\t\\t\\tcount += 1;\\n\\t\\t}\\n\\t\\tfloat sheepAverage = sum/count;\\n\\n \\tSystem.out.print(\\\"[(\\\" + String.valueOf(wolvesAverage) + \\\", \\\" + String.valueOf(sheepAverage) + \\\")]\\\\n\\\");\\n\\n }\\n \\n \\tworkspace.dispose();\\n\\n }\\n catch(Exception ex) {\\n ex.printStackTrace();\\n }\\n }\\n \\n \\n}\\n\"\n },\n {\n \"alpha_fraction\": 0.6532751321792603,\n \"alphanum_fraction\": 0.6555458307266235,\n \"avg_line_length\": 32.44444274902344,\n \"blob_id\": \"f761ee9cb3b527326523d8f632fdc8728f1d7713\",\n \"content_id\": \"b80cd3d6bab49d99e5289b85984b8d0acbde82a9\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5725,\n \"license_type\": \"no_license\",\n \"max_line_length\": 153,\n \"num_lines\": 171,\n \"path\": \"/src/amf/data.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# data.py contains everything that has to do with dealing with the actual\\n# data and manipulating data sets.\\n\\nimport random\\nimport copy\\nimport sys\\nimport misc\\nimport os\\n\\nIND = 5\\nDEP = 6\\n\\ndef make_vartype(vartype_str):\\n \\\"\\\"\\\"\\n Takes a string that labels columns in a data set.\\n vartype_string should either have 'i' or 'd' in it, e.g., \\\"iidid\\\"\\n\\n i - independent variable\\n d - dependent variable\\n \\\"\\\"\\\"\\n # replace characters i and d with IND and DEP symbols.\\n return tuple( (IND if c == 'i' else DEP) for c in vartype_str.strip() if not c.isspace())\\n\\ndef load(file_name):\\n \\\"\\\"\\\"\\n Loads a data set from a file.\\n\\n The data set should be space delimited to distinguish columns, and newline\\n delimited to distinguish rows.\\n\\n The first line should contain something that is to be passed into make_vartype.\\n\\n \\\"\\\"\\\"\\n\\n file_h = open(file_name)\\n\\n vartype = make_vartype(file_h.readline().strip())\\n\\n dataset = []\\n # go through each line and parse it\\n for line in file_h.xreadlines():\\n line = line.strip()\\n\\n # ignore empty lines\\n if len(line) == 0:\\n continue\\n\\n sline = line.split()\\n if len(sline) != len(vartype):\\n sys.stderr.write(\\\"Ignored line with wrong number of tokens: \\\" + line + \\\"\\\\n\\\")\\n continue\\n\\n # split the data into two groups, depending on if they are independent or dependent variables.\\n ind_list = []\\n dep_list = []\\n for idx, item in enumerate(sline):\\n (ind_list if vartype[idx] == IND else dep_list).append(float(item) if item != 'NOENTRY' else 'NOENTRY')\\n \\n dataset.append( ( tuple(ind_list), tuple(dep_list) ) )\\n\\n dataset.sort()\\n\\n return tuple(dataset)\\n\\ndef random_subsets(data_set, n_list, demonstrationDescriptors=[]):\\n \\\"\\\"\\\"\\n Returns a random subsets of the dataset with sizes specified in the n list.\\n There will be no duplicates between subsets.\\n \\\"\\\"\\\"\\n\\n if sum(n_list) > len(data_set):\\n raise IndexError, \\\"Requested size of subsets, from set of size %d, are too large: %s (total size %d)\\\" % (len (data_set), repr(n_list), sum(n_list))\\n\\n # generate a list of the indexes, then shuffle them. I will use these to randomly select from the data set.\\n idxs = range(len(data_set))\\n\\n # this is where the randomness lies\\n #random.seed(0)\\n\\n # idxs contain indexes of all elements in dataset\\n # we should get the top sum(n_list) values\\n # then use those to form random subsets of length specified by n_list\\n\\n # check number of demonstration descriptors\\n demonstrationCount = len(demonstrationDescriptors)\\n\\n # obtain and shuffle indexes for randomness\\n # divides equally per demonstration slp\\n # works better when a single slp is specified\\n\\n # dataset selection used if set to true\\n if(eval(os.environ[\\\"DATASET_SELECTION\\\"]) == 1):\\n useClosestIdx = True\\n else:\\n useClosestIdx = False\\n\\n #if(type(depData) == tuple):\\n if(useClosestIdx == True):\\n # we need to divide the dataset to cover all of these\\n # edit last item to account for correctness of sum\\n # shuffle for random allotment of extra points\\n nListSubLengths = [sum(n_list)/demonstrationCount]\\n\\n nListSubLengths[-1] += sum(n_list)%demonstrationCount\\n random.shuffle(nListSubLengths)\\n\\n # indexes corresponding to good points\\n goodIdxs = []\\n \\n # for each sublength\\n # corresponding to each demonstration\\n for index in range(len(nListSubLengths)):\\n # if feedback is on\\n # compute distances based on queried slps and not demonstration descriptors\\n if((eval(os.environ[\\\"FEEDBACK\\\"]) == 1) and (os.path.isfile(\\\"queried_slps.txt\\\") == True)):\\n with open(\\\"queried_slps.txt\\\", \\\"r\\\") as inFile:\\n queriedSlps = [eval(inFile.readlines()[0])]\\n distances = [(misc.distance(queriedSlps[index], data_set[idx][1]), idx) for idx in idxs]\\n else:\\n # get distances of all descriptors in dataset\\n # with respect to the input descriptor\\n distances = [(misc.distance(demonstrationDescriptors[index], data_set[idx][1]), idx) for idx in idxs]\\n\\n distances.sort()\\n\\n # extract first few as needed\\n nListSubLength = nListSubLengths[index]\\n goodDataPoints = distances[:nListSubLength]\\n\\n # append to list of all points\\n goodIdxs += [idx for distance, idx in goodDataPoints]\\n\\n random.shuffle(goodIdxs)\\n else:\\n # point sampling has already been done from the full dataset\\n # now we are dealing with individual dep value correspondences for a given ind configuration\\n random.shuffle(idxs)\\n \\n # now form subsets based on sequence of indexes\\n subsets = []\\n cur_offset = 0\\n for subset_size in n_list:\\n #if(type(depData) == tuple):\\n if(useClosestIdx == True):\\n # we are selecting from the full dataset\\n # this dataset does not contain individual slp correspondence\\n subsets.append( tuple(data_set[idx] for idx in goodIdxs[cur_offset:cur_offset + subset_size]) )\\n else:\\n # point sampling has already been done from the full dataset\\n # now we are dealing with individual dep value correspondences for a given ind configuration\\n subsets.append( tuple(data_set[idx] for idx in idxs[cur_offset:cur_offset + subset_size]) )\\n\\n cur_offset += subset_size\\n\\n return subsets\\n\\ndef split_dep(data_set):\\n \\\"\\\"\\\"\\n Split the data set into one-data-set-per-dependent variable.\\n \\\"\\\"\\\"\\n\\n num_dep = len(data_set[0][1])\\n\\n sets = [ [] for idx in range(num_dep) ]\\n\\n for ind, dep in data_set:\\n # for each dependent variable here...\\n for d_idx in range(num_dep):\\n sets[d_idx].append((ind, dep[d_idx]))\\n \\n return sets\\n\\n\\n\\n\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.638373613357544,\n \"alphanum_fraction\": 0.660858154296875,\n \"avg_line_length\": 37.121429443359375,\n \"blob_id\": \"6b842ce6a5a8435cd45f99cfb693f0cb89905ba2\",\n \"content_id\": \"e48c3eaf9d0e8475ff26fe2768343c05d4fb1ab4\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5337,\n \"license_type\": \"no_license\",\n \"max_line_length\": 109,\n \"num_lines\": 140,\n \"path\": \"/data/dataset_analysis/dataset_analysis.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import matplotlib.pyplot as plt\\nimport numpy\\nimport sys\\n\\ndef read_dataset(fileName):\\n dataset = []\\n \\n with open(fileName + \\\"/slp_dataset.txt\\\", \\\"r\\\") as slpDatasetFile:\\n for line in slpDatasetFile:\\n point = eval(line)\\n dataset.append(point)\\n\\n return dataset\\n\\ndef polarize_matrix(matrix, threshold=0):\\n # polarize matrix values\\n [rows, columns] = matrix.shape\\n\\n for row in range(rows):\\n for column in range(columns):\\n if(matrix[row][column] > threshold):\\n matrix[row][column] = 1\\n else:\\n matrix[row][column] = 0\\n\\ndef analyze_datasets(goodDatasetFolders=[], badDatasetFolders=[], figureSuffix = \\\"\\\"):\\n # determine comparison method for quality of datasets\\n\\n # demonstration specific data\\n demonstrationID = sys.argv[1]\\n \\n # get all points in good dataset\\n goodDataset = []\\n for folder in goodDatasetFolders:\\n goodDataset += read_dataset(demonstrationID + \\\"/\\\" + \\\"predictions_\\\" + folder)\\n\\n # get all points in bad dataset\\n badDataset = []\\n for folder in badDatasetFolders:\\n badDataset += read_dataset(demonstrationID + \\\"/\\\" + \\\"predictions_\\\" + folder)\\n\\n # define plot ranges\\n # each value represents a threshold for a bucket\\n dim1Step = eval(sys.argv[2])\\n dim2Step = eval(sys.argv[3])\\n dim1ThresholdValues = numpy.arange(0, 1, dim1Step)\\n dim2ThresholdValues = numpy.arange(0, 1, dim2Step)\\n\\n # now compute matrix based on discrete values\\n goodDensityMap = numpy.zeros((len(dim1ThresholdValues), len(dim2ThresholdValues)))\\n badDensityMap = numpy.zeros((len(dim1ThresholdValues), len(dim2ThresholdValues)))\\n\\n # for good datasets\\n for point in goodDataset:\\n # determine bucket for this point\\n [dim1Value, dim2Value] = point\\n\\n # will be an integer value between 0 and len(dim1ThresholdValues)-1\\n # will be an integer value between 0 and len(dim2ThresholdValues)-1\\n dim1Bucket = int(dim1Value/dim1Step)\\n dim2Bucket = int(dim2Value/dim2Step)\\n\\n # account for upper limits\\n # use capping\\n if(dim1Bucket == len(dim1ThresholdValues)):\\n dim1Bucket = len(dim1ThresholdValues) - 1\\n if(dim2Bucket == len(dim2ThresholdValues)):\\n dim2Bucket = len(dim2ThresholdValues) - 1\\n\\n # add count to matrix\\n goodDensityMap[dim1Bucket][dim2Bucket] += 1\\n\\n # for bad datasets\\n for point in badDataset:\\n # determine bucket for this point\\n [dim1Value, dim2Value] = point\\n\\n # will be an integer value between 0 and len(dim1ThresholdValues) - 1\\n # will be an integer value between 0 and len(dim2ThresholdValues) - 1\\n dim1Bucket = int(dim1Value/dim1Step) - 1\\n dim2Bucket = int(dim2Value/dim2Step) - 1\\n\\n # account for upper limits\\n # use capping\\n if(dim1Bucket == len(dim1ThresholdValues)):\\n dim1Bucket = len(dim1ThresholdValues) - 1\\n if(dim2Bucket == len(dim2ThresholdValues)):\\n dim2Bucket = len(dim2ThresholdValues) - 1\\n \\n # add count to matrix\\n badDensityMap[dim1Bucket][dim2Bucket] += 1\\n\\n # compute difference\\n # and normalize\\n goodDifferenceDensityMap = goodDensityMap - badDensityMap\\n badDifferenceDensityMap = badDensityMap - goodDensityMap\\n\\n polarize_matrix(goodDifferenceDensityMap, 0)\\n polarize_matrix(badDifferenceDensityMap, 0)\\n\\n # splice matrices if needed\\n # to zoom in on finer sections\\n #goodDifferenceDensityMap = goodDifferenceDensityMap[:10, :10]\\n #badDifferenceDensityMap = badDifferenceDensityMap[:10, :10]\\n #dim1ThresholdValues = dim1ThresholdValues[:10]\\n #dim2ThresholdValues = dim2ThresholdValues[:10]\\n \\n plt.figure()\\n plt.imshow(goodDifferenceDensityMap, interpolation='nearest', cmap=plt.cm.gray)\\n plt.ylabel(\\\"Maximum Area\\\")\\n plt.xlabel(\\\"Maximum Perimeter\\\")\\n #plt.title(\\\"Good Difference Density Map\\\")\\n plt.yticks(range(len(dim1ThresholdValues)), dim1ThresholdValues)\\n plt.xticks(range(len(dim2ThresholdValues)), dim2ThresholdValues)\\n plt.savefig(demonstrationID + \\\"/\\\" + \\\"good_difference_density_map\\\" + figureSuffix + \\\".png\\\")\\n plt.close(\\\"all\\\")\\n \\n plt.figure()\\n plt.imshow(badDifferenceDensityMap, interpolation='nearest', cmap=plt.cm.gray)\\n plt.ylabel(\\\"Maximum Area\\\")\\n plt.xlabel(\\\"Maximum Perimeter\\\")\\n #plt.title(\\\"Bad Difference Density Map\\\")\\n plt.yticks(range(len(dim1ThresholdValues)), dim1ThresholdValues)\\n plt.xticks(range(len(dim2ThresholdValues)), dim2ThresholdValues)\\n plt.savefig(demonstrationID + \\\"/\\\" + \\\"bad_difference_density_map\\\" + figureSuffix + \\\".png\\\")\\n plt.close(\\\"all\\\")\\n \\ndef main():\\n analyze_datasets([\\\"best\\\", \\\"second_best\\\", \\\"third_best\\\"], [\\\"worst\\\", \\\"second_worst\\\", \\\"third_worst\\\"], \\\"_all\\\")\\n analyze_datasets([\\\"best\\\"], [\\\"worst\\\"], \\\"_1_1\\\")\\n analyze_datasets([\\\"best\\\"], [\\\"second_worst\\\"], \\\"_1_2\\\")\\n analyze_datasets([\\\"best\\\"], [\\\"third_worst\\\"], \\\"_1_3\\\")\\n analyze_datasets([\\\"second_best\\\"], [\\\"worst\\\"], \\\"_2_1\\\")\\n analyze_datasets([\\\"second_best\\\"], [\\\"second_worst\\\"], \\\"_2_2\\\")\\n analyze_datasets([\\\"second_best\\\"], [\\\"third_worst\\\"], \\\"_2_3\\\")\\n analyze_datasets([\\\"third_best\\\"], [\\\"worst\\\"], \\\"_3_1\\\")\\n analyze_datasets([\\\"third_best\\\"], [\\\"second_worst\\\"], \\\"_3_2\\\")\\n analyze_datasets([\\\"third_best\\\"], [\\\"third_worst\\\"], \\\"_3_3\\\")\\n \\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.6425849199295044,\n \"alphanum_fraction\": 0.6641530990600586,\n \"avg_line_length\": 32.789039611816406,\n \"blob_id\": \"9bc73487b95b665c1d75851548f8c07f7c7dad62\",\n \"content_id\": \"57cbd55a319ed9d96c686d00e8557615b2489af6\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 12333,\n \"license_type\": \"no_license\",\n \"max_line_length\": 127,\n \"num_lines\": 365,\n \"path\": \"/src/sampling/utils/featurize_image.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# libraries used\\nimport cv2\\nimport numpy as np\\nfrom skimage.feature import local_binary_pattern\\nfrom skimage.io import imread\\nfrom skimage.transform import resize\\nfrom keras import backend as K\\nfrom keras.models import model_from_json\\nfrom keras.applications.vgg16 import VGG16\\nfrom keras.applications.vgg16 import preprocess_input\\nfrom keras.applications.resnet50 import ResNet50\\nfrom keras.preprocessing import image\\nfrom keras.preprocessing.image import load_img, img_to_array\\n\\nfrom keras.models import Sequential\\nfrom keras.layers.convolutional import Convolution3D\\nfrom keras.layers.convolutional_recurrent import ConvLSTM2D\\nfrom keras.layers.normalization import BatchNormalization\\nfrom keras.layers import Dense, Activation\\nfrom keras.layers.core import Reshape, Flatten\\n\\nimport os\\n\\n# using magic numbers for now\\n\\n# feature extraction initialization\\ndef get_network(descriptorSetting, pathPrefix=None):\\n # load feature extractor if needed\\n net = None\\n\\n # prefix to folder location\\n if(pathPrefix == None):\\n pathPrefix = \\\"\\\"\\n\\n functionName = \\\"get_\\\" + descriptorSetting.lower() + \\\"_network\\\"\\n if(functionName in globals().keys()):\\n net = globals()[functionName](pathPrefix)\\n\\n return net\\n\\n# feature extraction\\ndef get_features(net, imageName, descriptorSetting):\\n # initialization\\n features = None\\n\\n functionName = \\\"get_\\\" + descriptorSetting.lower() + \\\"_features\\\"\\n features = globals()[functionName](net, imageName)\\n \\n # return features\\n return features\\n\\n# saved keras network\\ndef get_saved_keras_network(modelFile, weightFile):\\n # load and compile saved network\\n jsonFile = open(modelFile, 'r')\\n modelJson = jsonFile.read()\\n jsonFile.close()\\n net = model_from_json(modelJson)\\n net.load_weights(weightFile)\\n net.compile(loss='categorical_crossentropy', optimizer='adadelta', metrics=['accuracy'])\\n\\n # return network\\n return net\\n\\n# contour information\\ndef get_contour_features(net, imageName):\\n imageData = cv2.imread(imageName, cv2.IMREAD_GRAYSCALE)\\n\\n # dilate for better contour extraction\\n kernel = np.ones((5,5),np.uint8)\\n dilatedImageData = cv2.dilate(imageData, kernel, iterations=4)\\n #cv2.imwrite(imageName[:-4] + \\\"_dilated.png\\\", dilatedImageData)\\n\\n # extract contours\\n ret, threshold = cv2.threshold(dilatedImageData, 127, 255, 0)\\n contours, hierarchy = cv2.findContours(threshold, cv2.RETR_TREE, cv2.CHAIN_APPROX_NONE)\\n contouredImageData = np.zeros(dilatedImageData.shape)\\n for i in xrange(len(contours)):\\n cv2.drawContours(contouredImageData, contours, i, (255, 255, 255), 1, 8, hierarchy) \\n #cv2.imwrite(imageName[:-4] + \\\"_contoured.png\\\", contouredImageData)\\n\\n # extract largest contour area and perimeter\\n areas = []\\n perimeters = []\\n\\n for contour in contours:\\n area = cv2.contourArea(contour)\\n perimeter = cv2.arcLength(contour, True)\\n areas.append(area)\\n perimeters.append(perimeter)\\n\\n # define dependent values\\n maxArea = max(areas)\\n maxPerimeter = max(perimeters)\\n features = [maxArea, maxPerimeter]\\n return features\\n\\nimport sys\\n\\n# density information\\ndef get_density_features(net, imageName, blocks=eval(os.environ[\\\"NUM_DEPENDENT\\\"])):\\n # read image\\n image = imread(imageName, as_grey=True)\\n \\n # resize to a pixels power of 2\\n # required for equal division when splitting\\n resizeDimension = 2**int(min([np.log2(x) for x in image.shape]))\\n image = resize(image, (resizeDimension, resizeDimension), mode=\\\"wrap\\\")\\n \\n # polarize pixels\\n threshold = np.amax(image)\\n image[image < threshold] = 0\\n image[image >= threshold] = 1\\n\\n # segment image\\n # segments must be a power of 4\\n # because each block is split into 4 sub blocks\\n currentBlocksCount = 1\\n blocksList = np.array([image])\\n \\n while(currentBlocksCount < blocks):\\n # split horizontally\\n subBlocksList = []\\n for block in blocksList:\\n subBlocks = np.hsplit(block, 2)\\n subBlocksList.extend(subBlocks)\\n blocksList = subBlocksList\\n\\n # split vertically\\n subBlocksList = []\\n for block in blocksList:\\n subBlocks = np.vsplit(block, 2)\\n subBlocksList.extend(subBlocks)\\n blocksList = subBlocksList\\n\\n # update number of blocks\\n currentBlocksCount *= 4\\n \\n # compute density\\n # since all images are the same size, we can use count instead\\n features = [x.sum() for x in blocksList]\\n \\n return features\\n \\n# vgg network\\ndef get_vgg_network(pathPrefix):\\n # pretrained vgg16 network\\n net = VGG16(include_top=True, weights='imagenet', input_tensor=None)\\n return net\\n \\n# vgg based features\\ndef get_vgg_features(net, imageName):\\n # preprocess image\\n imageData = image.load_img(imageName, target_size=(224, 224))\\n imageData = image.img_to_array(imageData)\\n imageData = np.expand_dims(imageData, axis=0)\\n imageData = preprocess_input(imageData)\\n\\n # final layer output\\n features = net.predict(imageData)[0]\\n return features\\n\\n# lbp features\\ndef get_lbp_features(net, imageName):\\n # settings for LBP\\n METHOD = 'uniform'\\n radius = 3\\n nPoints = 8 * radius\\n nBins = 128\\n \\n # dilate for better contour extraction\\n imageData = cv2.imread(imageName, cv2.IMREAD_GRAYSCALE)\\n kernel = np.ones((5,5),np.uint8)\\n dilatedImageData = cv2.dilate(imageData, kernel, iterations=4)\\n #cv2.imwrite(imageName[:-4] + \\\"_dilated.png\\\", dilatedImageData)\\n\\n # compute normalized histogram distribution\\n lbp = local_binary_pattern(dilatedImageData, nPoints, radius, METHOD)\\n hist, binEdges = np.histogram(lbp, bins=nBins, range=(0, nBins), density=True)\\n\\n # define dependent values\\n features = hist\\n return features\\n \\n# mnist network\\ndef get_mnist_network(pathPrefix):\\n # path to saved network\\n modelFile = pathPrefix + \\\"keras/mnist_cnn_model.json\\\"\\n weightFile = pathPrefix + \\\"keras/mnist_cnn_model.h5\\\"\\n \\n # return network\\n net = get_saved_keras_network(modelFile, weightFile)\\n return net\\n\\n# mnist based convolutional features\\ndef get_mnist_features(net, imageName):\\n # output layer specification\\n LAYER_NUMBER = 7\\n LEARNING_PHASE = 0\\n\\n # dilate for better contour extraction\\n imageData = cv2.imread(imageName, cv2.IMREAD_GRAYSCALE)\\n kernel = np.ones((5,5),np.uint8)\\n dilatedImageData = cv2.dilate(imageData, kernel, iterations=4)\\n #cv2.imwrite(imageName[:-4] + \\\"_dilated.png\\\", dilatedImageData)\\n\\n imageData = resize(dilatedImageData, (28, 28))\\n imageData = imageData.reshape(1, 1, 28, 28)\\n\\n # layer output\\n get_layer_output = K.function([net.layers[0].input, K.learning_phase()],\\n [net.layers[LAYER_NUMBER].output])\\n layerOutput = get_layer_output([imageData, LEARNING_PHASE])[0]\\n features = layerOutput[0]\\n return features\\n\\n# resnet network\\ndef get_resnet50_network(pathPrefix):\\n # pretrained resnet50 network\\n net = ResNet50(include_top=True, weights='imagenet', input_tensor=None)\\n return net\\n\\n# resnet based features\\ndef get_resnet50_features(net, imageName):\\n # preprocess image\\n dimensionOrdering = K.image_dim_ordering()\\n mean = (103.939, 116.779, 123.68)\\n imageData = load_img(imageName, target_size=(224, 224))\\n imageData = img_to_array(imageData, dim_ordering=dimensionOrdering)\\n\\n if dimensionOrdering == 'th':\\n imageData[0, :, :] -= mean[0]\\n imageData[1, :, :] -= mean[1]\\n imageData[2, :, :] -= mean[2]\\n # 'RGB'->'BGR'\\n imageData = imageData[::-1, :, :]\\n else:\\n imageData[:, :, 0] -= mean[0]\\n imageData[:, :, 1] -= mean[1]\\n imageData[:, :, 2] -= mean[2]\\n imageData = imageData[:, :, ::-1]\\n\\n imageData = np.expand_dims(imageData, axis=0)\\n\\n # final layer output\\n features = net.predict(imageData)[0]\\n return features\\n\\n# autoencoder network\\ndef get_vae_network(pathPrefix):\\n # path to saved network\\n modelFile = pathPrefix + \\\"keras/variational_autoencoder_model.json\\\"\\n weightFile = pathPrefix + \\\"keras/variational_autoencoder_model.h5\\\"\\n\\n # return network\\n net = get_saved_keras_network(modelFile, weightFile)\\n return net\\n\\n# autoencoder based features\\ndef get_vae_features(net, imageName):\\n # output layer specification\\n LAYER_NUMBER = 3\\n LEARNING_PHASE = 0\\n X_TRAIN_SHAPE = tuple([1,784])\\n\\n # preprocess image\\n imageData = imread(imageName, as_grey=True)\\n imageData = resize(imageData, (28, 28))\\n imageData = imageData.flatten()\\n imageData = imageData.astype('float32') / 255.\\n imageData = imageData.reshape((1, np.prod(X_TRAIN_SHAPE[1:])))\\n\\n # layer output\\n get_layer_output = K.function([net.layers[0].input, K.learning_phase()],\\n [net.layers[LAYER_NUMBER].output])\\n layer_output = get_layer_output([imageData, LEARNING_PHASE])[0]\\n features = layer_output[0]\\n return features\\n\\n# conv lstm network\\ndef get_lstm_network(pathPrefix):\\n # path to saved network\\n modelFile = pathPrefix + \\\"keras/conv_lstm_model.json\\\"\\n weightFile = pathPrefix + \\\"keras/conv_lstm_model.h5\\\"\\n\\n #net = get_saved_keras_network_lstm(modelFile, weightFile)\\n\\n # manually specify network\\n # need to do this because convLTSM layer gives error\\n net = Sequential()\\n net.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\\n input_shape=(None, 40, 40, 1),\\n batch_input_shape=(None, 11, 40, 40, 1),\\n border_mode='same', return_sequences=True))\\n net.add(BatchNormalization())\\n net.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\\n border_mode='same', return_sequences=True))\\n net.add(BatchNormalization())\\n net.add(Flatten())\\n net.add(Dense(128))\\n net.add(Dense(1600*11))\\n net.add(Reshape((11, 40, 40, 1)))\\n net.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\\n border_mode='same', return_sequences=True))\\n net.add(BatchNormalization())\\n net.add(ConvLSTM2D(nb_filter=40, nb_row=3, nb_col=3,\\n border_mode='same', return_sequences=True))\\n net.add(BatchNormalization())\\n net.add(Convolution3D(nb_filter=1, kernel_dim1=1, kernel_dim2=3,\\n kernel_dim3=3, activation='sigmoid',\\n border_mode='same', dim_ordering='tf'))\\n net.load_weights(weightFile)\\n net.compile(loss='categorical_crossentropy', optimizer='adadelta', metrics=['accuracy'])\\n\\n # return network\\n return net\\n\\n# convolutional lstm based features\\ndef get_lstm_features(net, imageName):\\n # output layer specification\\n LAYER_NUMBER = 5\\n LEARNING_PHASE = 0\\n\\n # generate movie\\n rows = 40\\n columns = 40\\n numberOfSamples = 1\\n numberOfFrames = 11\\n shiftedMovie = np.zeros((numberOfSamples, numberOfFrames, rows, columns, 1), dtype=np.float)\\n imageNameSplit = imageName.split(\\\"_\\\")\\n\\n # stitch all but last\\n imageSeriesName = \\\"\\\"\\n for index in range(len(imageNameSplit)-1):\\n splitItem = imageNameSplit[index]\\n imageSeriesName += splitItem + \\\"_\\\"\\n imageSeriesName = imageSeriesName[:-1]\\n\\n # image series name should be the entire thing up to underscore \\n for i in range(numberOfSamples):\\n # read custom dataset\\n for t in range(numberOfFrames):\\n currentFrameName = imageSeriesName + \\\"_\\\" + str(200 + t) + \\\".png\\\"\\n \\n # read and resize to 40x40\\n imageData = imread(currentFrameName, as_grey=True)\\n imageData = resize(imageData, (40, 40)) \\n shiftedMovie[i, t, :, :, 0] = imageData\\n\\n # layer output\\n get_layer_output = K.function([net.layers[0].input, K.learning_phase()],\\n [net.layers[LAYER_NUMBER].output])\\n layer_output = get_layer_output([shiftedMovie, LEARNING_PHASE])[0]\\n features = layer_output[0]\\n return features\\n\\n# image search features\\ndef get_image_match_features(net, imageName):\\n # use a system call to invoke the python3 library using this image name\\n os.system(\\\"python3 sampling/utils/get_image_match_features.py \\\" + imageName)\\n\\n # use temporary file for communication\\n features = np.load(\\\"image_match_features.npy\\\")\\n\\n return features\\n\"\n },\n {\n \"alpha_fraction\": 0.7101449370384216,\n \"alphanum_fraction\": 0.7753623127937317,\n \"avg_line_length\": 22,\n \"blob_id\": \"096d119115dfd8c1aeb7a1118d0b6beb8c5647c3\",\n \"content_id\": \"bdc936d5c28554fd29fc750fdb90f22ede3ee6e3\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 138,\n \"license_type\": \"no_license\",\n \"max_line_length\": 54,\n \"num_lines\": 6,\n \"path\": \"/src/domainconfs/flocking.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"TRAINING_SIZE = 330\\nVALIDATION_SIZE = 100\\nRM_GRANULARITY = 40\\nREGRESSION = \\\"sklearn.neighbors.KNeighborsRegressor()\\\"\\n\\nNUM_INDEPENDENT = 3\\n\"\n },\n {\n \"alpha_fraction\": 0.6039416193962097,\n \"alphanum_fraction\": 0.6153117418289185,\n \"avg_line_length\": 30.02941131591797,\n \"blob_id\": \"1f1e7cbdbe4e657662a152df12145edfc3c467b7\",\n \"content_id\": \"e875ace109e35918df732c93332cbba1d4dcda2f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 5277,\n \"license_type\": \"no_license\",\n \"max_line_length\": 137,\n \"num_lines\": 170,\n \"path\": \"/src/configuration.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env bash\\n#######################################################################\\n# this is the only thing that should be edited when using this script #\\n#######################################################################\\n\\n# abm being used\\nexport ABM=\\\"flocking\\\"\\n#export ABM=\\\"civil_violence\\\"\\n#export ABM=\\\"forest_fire\\\"\\n#export ABM=\\\"aids\\\"\\n#export ABM=\\\"wolf_sheep\\\"\\n#export ABM=\\\"eum\\\"\\n#export ABM=\\\"schelling\\\"\\n#export ABM=\\\"turbulence\\\"\\n\\n# use of disk to reduce ram consumption\\nexport WRITE_TO_DISK=0\\n\\n# use of demonstration images is off by default \\nexport IMAGES=0\\n\\n# check if this abm depends on images\\nif [ $ABM == \\\"flocking\\\" ] || [ $ABM == \\\"turbulence\\\" ]; then\\n export IMAGES=1\\nfi\\n\\n# if feature reduction has been used\\nexport REDUCED=0\\n\\n# if image features are uniform across all stages of the framework\\n# used during filtering\\nexport IMAGE_FEATURIZATION_HOMOGENEOUS=0\\n\\n# descriptor setting if any\\nexport DESCRIPTOR_SETTING=\\\"\\\"\\n\\n# check if this abm depends on images\\n# specify which descriptor to use\\nif [ $IMAGES == 1 ]; then\\n # determine number of dependent variables\\n #export DESCRIPTOR_SETTING=\\\"DENSITY\\\"\\n export DESCRIPTOR_SETTING=\\\"CONTOUR\\\"\\n #export DESCRIPTOR_SETTING=\\\"LBP\\\"\\n #export DESCRIPTOR_SETTING=\\\"MNIST\\\"\\n #export DESCRIPTOR_SETTING=\\\"VAE\\\"\\n #export DESCRIPTOR_SETTING=\\\"LSTM\\\"\\n #export DESCRIPTOR_SETTING=\\\"VGG\\\"\\n #export DESCRIPTOR_SETTING=\\\"RESNET50\\\"\\n #export DESCRIPTOR_SETTING=\\\"IMAGE_MATCH\\\"\\nfi\\n\\n# check for override in case of filtering\\nif [ ! -z $USE_FILTER_DESCRIPTOR_SETTING ]; then\\n if [ $USE_FILTER_DESCRIPTOR_SETTING == 1 ]; then\\n #export DESCRIPTOR_SETTING=\\\"RESNET50\\\"\\n\\t#export DESCRIPTOR_SETTING=\\\"IMAGE_MATCH\\\"\\n\\texport DESCRIPTOR_SETTING=\\\"CONTOUR\\\"\\n fi\\nfi\\n\\n# distance metric used for filtering\\n# euclidean assumed by default\\nexport DISTANCE_METHOD=\\\"EUCLIDEAN\\\"\\n#export DISTANCE_METHOD=\\\"NORMALIZED\\\"\\n\\n############\\n# feedback #\\n############\\n\\nexport FEEDBACK=0\\nexport LEARNING_RATE_FM_DATASET=0.0\\nexport LEARNING_RATE_REGRESSION_PARAMETER=1.0\\nexport REGRESSION_PARAMETER_LOWER_LIMIT=1\\nexport REGRESSION_PARAMETER_UPPER_LIMIT=100\\n\\n#####################\\n# dataset selection #\\n#####################\\n\\nexport DATASET_SELECTION=1\\n\\n#########################\\n# refine configurations #\\n#########################\\n\\nexport CONFIGURATIONS_PRUNING=1\\nexport CONFIGURATIONS_OUTLIER_DETECTION=1\\n\\n#############################\\n# constants for this script #\\n#############################\\n\\n# demonstration file\\nexport DEMONSTRATION_FOLDER=\\\"../data/demonstrations/\\\"$ABM\\n\\n# configuration file\\nexport CONFIGURATION_FILE=\\\"domainconfs/\\\"$ABM\\\".py\\\"\\n\\n# data file\\n# if image processing was used, append descriptor setting to processed filename\\nif [ $IMAGES == 1 ]; then\\n export DESCRIPTOR_SETTING_LOWER_CASE=$(echo $DESCRIPTOR_SETTING | tr \\\"[:upper:]\\\" \\\"[:lower:]\\\")\\n export DATA_FILE=\\\"../data/domaindata/processed_\\\"$ABM\\\"_\\\"$DESCRIPTOR_SETTING_LOWER_CASE\\n\\n if [ $REDUCED == 1 ]; then\\n\\t export DATA_FILE=$DATA_FILE\\\"_reduced\\\"\\n fi\\n \\n export DATA_FILE=$DATA_FILE\\\".txt\\\"\\nelse\\n export DATA_FILE=\\\"../data/domaindata/processed_\\\"$ABM\\\".txt\\\"\\nfi\\n\\n# scaling information\\nexport SCALE_DATA_FILE=$(cat \\\"map.py\\\" | grep SCALE_DATA_FILE | head -n 1 | sed s/\\\\\\\"/\\\\\\\\n/g | sed -n 2p)\\n\\n##########################################\\n# abm simulation tool and file extension #\\n# one entry required per abm #\\n##########################################\\n\\nif [ $ABM == \\\"civil_violence\\\" ] || [ $ABM == \\\"forest_fire\\\" ] || [ $ABM == \\\"eum\\\" ] || [ $ABM == \\\"schelling\\\" ]; then\\n export SIMULATION_TOOL=\\\"python3\\\"\\n export SIMULATION_EXTENSION=\\\".py\\\"\\nelif [ $ABM == \\\"flocking\\\" ] || [ $ABM == \\\"aids\\\" ] || [ $ABM == \\\"wolf_sheep\\\" ] || [ $ABM == \\\"turbulence\\\" ]; then\\n export SIMULATION_TOOL=\\\"java\\\"\\n export SIMULATION_EXTENSION=\\\"\\\"\\nfi\\n\\n############################################################\\n# image featurization details, if any #\\n# entry required for abms using images #\\n# one entry required per abm for nuber of dependent values #\\n############################################################\\n\\n# default number of dependent values\\nexport NUM_DEPENDENT=0\\n\\n# check if this abm depends on images\\nif [ $IMAGES == 1 ]; then\\n # determine number of dependent variables\\n # based on image feature descriptors\\n \\n if [ \\\"$DESCRIPTOR_SETTING\\\" == \\\"DENSITY\\\" ]; then\\n\\t export NUM_DEPENDENT=1\\n elif [ \\\"$DESCRIPTOR_SETTING\\\" == \\\"CONTOUR\\\" ]; then\\n\\t export NUM_DEPENDENT=2\\n elif [ \\\"$DESCRIPTOR_SETTING\\\" == \\\"LBP\\\" ]; then\\n\\t export NUM_DEPENDENT=128\\n elif [ \\\"$DESCRIPTOR_SETTING\\\" == \\\"MNIST\\\" ]; then\\n\\t export NUM_DEPENDENT=128\\n elif [ \\\"$DESCRIPTOR_SETTING\\\" == \\\"VAE\\\" ]; then\\n\\t export NUM_DEPENDENT=128\\n elif [ \\\"$DESCRIPTOR_SETTING\\\" == \\\"LSTM\\\" ]; then\\n\\t export NUM_DEPENDENT=128\\n elif [ \\\"$DESCRIPTOR_SETTING\\\" == \\\"VGG\\\" ]; then\\n\\t export NUM_DEPENDENT=1000\\n elif [ \\\"$DESCRIPTOR_SETTING\\\" == \\\"RESNET50\\\" ]; then\\n\\t export NUM_DEPENDENT=1000\\n elif [ \\\"$DESCRIPTOR_SETTING\\\" == \\\"IMAGE_MATCH\\\" ]; then\\n\\t export NUM_DEPENDENT=648\\n fi\\n\\n if [ $REDUCED == 1 ]; then\\n\\t export NUM_DEPENDENT=50\\n fi\\nelse\\n # read from domainconfs\\n export NUM_DEPENDENT=$(python -c \\\"execd = {} ; exec(open(\\\\\\\"\\\"$CONFIGURATION_FILE\\\"\\\\\\\").read(), execd) ; print execd[\\\\\\\"NUM_DEPENDENT\\\\\\\"]\\\")\\nfi\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.8611111044883728,\n \"alphanum_fraction\": 0.8611111044883728,\n \"avg_line_length\": 23,\n \"blob_id\": \"91a90a5b5253a3c94bfff9c6a258b67f658d2be8\",\n \"content_id\": \"2c486c490670b67cd6854c51490477641d70b6fe\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 72,\n \"license_type\": \"no_license\",\n \"max_line_length\": 29,\n \"num_lines\": 3,\n \"path\": \"/src/amf/__init__.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import data\\nfrom fm import ForwardMapping\\nfrom rm import ReverseMapping\\n\"\n },\n {\n \"alpha_fraction\": 0.7781955003738403,\n \"alphanum_fraction\": 0.7819548845291138,\n \"avg_line_length\": 19.461538314819336,\n \"blob_id\": \"e5745ea83081fe921aafb9494a65c6615c46565d\",\n \"content_id\": \"e9f129c08b7512e45d606f8ccfaf1ecdf4784d4e\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 266,\n \"license_type\": \"no_license\",\n \"max_line_length\": 47,\n \"num_lines\": 13,\n \"path\": \"/src/sampling/utils/get_image_match_features.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from image_match.goldberg import ImageSignature\\nimport sys\\nimport numpy as np\\n\\n# input file\\nimageName = sys.argv[1]\\n\\n# generate signature\\ngis = ImageSignature()\\nfeatures = gis.generate_signature(imageName)\\n\\n# save signature\\nnp.save(\\\"image_match_features\\\", features)\\n\"\n },\n {\n \"alpha_fraction\": 0.612515926361084,\n \"alphanum_fraction\": 0.6263416409492493,\n \"avg_line_length\": 27.48186492919922,\n \"blob_id\": \"8b3e6e1918ce334b2a4a0e87e0cd1a80681be8e5\",\n \"content_id\": \"cce442fd6bc75d265c10ec508ed72c6399a6dfba\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5497,\n \"license_type\": \"no_license\",\n \"max_line_length\": 83,\n \"num_lines\": 193,\n \"path\": \"/data/dataset_analysis/plot_alps_slps.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# snippet from lfd.py\\nfrom mpl_toolkits.mplot3d import Axes3D\\nimport matplotlib.pyplot as plt\\n\\nwith open(\\\"1/predictions_best/alp_dataset_1488308839.txt\\\", \\\"r\\\") as alpDatasetFile:\\n alpDataset = alpDatasetFile.readlines()\\n alpPoints = [eval(x) for x in alpDataset]\\n\\n m = \\\"o\\\"\\n c = \\\"r\\\"\\n \\n # plot ALPs\\n fig = plt.figure()\\n ax = fig.add_subplot(111, projection='3d')\\n plotName = \\\"alp_points_best\\\" + \\\".png\\\"\\n alpPoints = zip(*alpPoints)\\n [xs, ys, zs] = alpPoints\\n ax.scatter(xs, ys, zs, c=c, marker=m)\\n ax.set_xlabel(\\\"Max-align-turn\\\")\\n ax.set_ylabel(\\\"Max-cohere-turn\\\")\\n ax.set_zlabel(\\\"Max-separate-turn\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\n fig = plt.figure()\\n plotName = \\\"alp_points_dim1_best\\\" + \\\".png\\\"\\n plt.hist(xs, color=c)\\n plt.xlabel(\\\"Max-align-turn\\\")\\n plt.ylabel(\\\"Number of points in each value range\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\n fig = plt.figure()\\n plotName = \\\"alp_points_dim2_best\\\" + \\\".png\\\"\\n plt.hist(ys, color=c)\\n plt.xlabel(\\\"Max-cohere-turn\\\")\\n plt.ylabel(\\\"Number of points in each value range\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\n fig = plt.figure()\\n plotName = \\\"alp_points_dim3_best\\\" + \\\".png\\\"\\n plt.hist(zs, color=c)\\n plt.xlabel(\\\"Max-separate-turn\\\")\\n plt.ylabel(\\\"Number of points in each value range\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\nwith open(\\\"1/predictions_worst/alp_dataset_1488319453.txt\\\", \\\"r\\\") as alpDatasetFile:\\n alpDataset = alpDatasetFile.readlines()\\n alpPoints = [eval(x) for x in alpDataset]\\n\\n m = \\\"o\\\"\\n c = \\\"r\\\"\\n \\n # plot ALPs\\n fig = plt.figure()\\n ax = fig.add_subplot(111, projection='3d')\\n plotName = \\\"alp_points_worst\\\" + \\\".png\\\"\\n alpPoints = zip(*alpPoints)\\n [xs, ys, zs] = alpPoints\\n ax.scatter(xs, ys, zs, c=c, marker=m)\\n ax.set_xlabel(\\\"Max-align-turn\\\")\\n ax.set_ylabel(\\\"Max-cohere-turn\\\")\\n ax.set_zlabel(\\\"Max-separate-turn\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\n fig = plt.figure()\\n plotName = \\\"alp_points_dim1_worst\\\" + \\\".png\\\"\\n plt.hist(xs, color=c)\\n plt.xlabel(\\\"Max-align-turn\\\")\\n plt.ylabel(\\\"Number of points in each value range\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\n fig = plt.figure()\\n plotName = \\\"alp_points_dim2_worst\\\" + \\\".png\\\"\\n plt.hist(ys, color=c)\\n plt.xlabel(\\\"Max-cohere-turn\\\")\\n plt.ylabel(\\\"Number of points in each value range\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\n fig = plt.figure()\\n plotName = \\\"alp_points_dim3_worst\\\" + \\\".png\\\"\\n plt.hist(zs, color=c)\\n plt.xlabel(\\\"Max-separate-turn\\\")\\n plt.ylabel(\\\"Number of points in each value range\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\nwith open(\\\"1/predictions_best/slp_dataset_1488308839.txt\\\", \\\"r\\\") as slpDatasetFile:\\n slpDataset = slpDatasetFile.readlines()\\n slpPoints = [eval(x) for x in slpDataset]\\n\\n m = \\\"o\\\"\\n c = \\\"r\\\"\\n \\n # plot SLPs\\n fig = plt.figure()\\n plotName = \\\"slp_points_best\\\" + \\\".png\\\"\\n slpPoints = zip(*slpPoints)\\n [xs, ys] = slpPoints\\n plt.plot(xs, ys, c+m)\\n plt.xlabel(\\\"Maximum Area\\\")\\n plt.ylabel(\\\"Maximum Perimeter\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\n fig = plt.figure()\\n plotName = \\\"slp_points_dim1_best\\\" + \\\".png\\\"\\n plt.hist(xs, color=c)\\n plt.xlabel(\\\"Maximum Area\\\")\\n plt.ylabel(\\\"Number of points in each value range\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n \\n fig = plt.figure()\\n plotName = \\\"slp_points_dim2_best\\\" + \\\".png\\\"\\n plt.hist(ys, color=c)\\n plt.xlabel(\\\"Maximum Perimeter\\\")\\n plt.ylabel(\\\"Number of points in each value range\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\nwith open(\\\"1/predictions_worst/slp_dataset_1488319453.txt\\\", \\\"r\\\") as slpDatasetFile:\\n slpDataset = slpDatasetFile.readlines()\\n slpPoints = [eval(x) for x in slpDataset]\\n\\n m = \\\"o\\\"\\n c = \\\"r\\\"\\n \\n # plot SLPs\\n fig = plt.figure()\\n plotName = \\\"slp_points_worst\\\" + \\\".png\\\"\\n slpPoints = zip(*slpPoints)\\n [xs, ys] = slpPoints\\n plt.plot(xs, ys, c+m)\\n plt.xlabel(\\\"Maximum Area\\\")\\n plt.ylabel(\\\"Maximum Perimeter\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\n fig = plt.figure()\\n plotName = \\\"slp_points_dim1_worst\\\" + \\\".png\\\"\\n plt.hist(xs, color=c)\\n plt.xlabel(\\\"Maximum Area\\\")\\n plt.ylabel(\\\"Number of points in each value range\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n \\n fig = plt.figure()\\n plotName = \\\"slp_points_dim2_worst\\\" + \\\".png\\\"\\n plt.hist(ys, color=c)\\n plt.xlabel(\\\"Maximum Perimeter\\\")\\n plt.ylabel(\\\"Number of points in each value range\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\nwith open(\\\"slp_dataset_1488996573.txt\\\", \\\"r\\\") as slpDatasetFile:\\n slpDataset = slpDatasetFile.readlines()\\n slpPoints = [eval(x) for x in slpDataset]\\n\\n m = \\\"o\\\"\\n c = \\\"r\\\"\\n \\n # plot SLPs\\n fig = plt.figure()\\n plotName = \\\"slp_points_guided\\\" + \\\".png\\\"\\n slpPoints = zip(*slpPoints)\\n [xs, ys] = slpPoints\\n plt.plot(xs, ys, c+m)\\n plt.xlabel(\\\"Maximum Area\\\")\\n plt.ylabel(\\\"Maximum Perimeter\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\\n fig = plt.figure()\\n plotName = \\\"slp_points_dim1_guided\\\" + \\\".png\\\"\\n plt.hist(xs, color=c)\\n plt.xlabel(\\\"Maximum Area\\\")\\n plt.ylabel(\\\"Number of points in each value range\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n \\n fig = plt.figure()\\n plotName = \\\"slp_points_dim2_guided\\\" + \\\".png\\\"\\n plt.hist(ys, color=c)\\n plt.xlabel(\\\"Maximum Perimeter\\\")\\n plt.ylabel(\\\"Number of points in each value range\\\")\\n plt.savefig(plotName)\\n plt.close(fig)\\n\"\n },\n {\n \"alpha_fraction\": 0.7403846383094788,\n \"alphanum_fraction\": 0.7403846383094788,\n \"avg_line_length\": 33.66666793823242,\n \"blob_id\": \"9e9e52a4dc407c0845d04708b7e1dbeb1c6a2e3c\",\n \"content_id\": \"486c468e6c5a58f1286525eeb12b49523ad3a940\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 104,\n \"license_type\": \"no_license\",\n \"max_line_length\": 65,\n \"num_lines\": 3,\n \"path\": \"/src/sampling/sampling_raw.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# sample data\\necho \\\"==>sampling data\\\"\\n$SIMULATION_TOOL $ABM/sample_$ABM$SIMULATION_EXTENSION > $ABM.txt\\n\"\n },\n {\n \"alpha_fraction\": 0.587643027305603,\n \"alphanum_fraction\": 0.612356960773468,\n \"avg_line_length\": 25.325302124023438,\n \"blob_id\": \"d859a85c93f92e7398f42515ddbe06ccd838d28b\",\n \"content_id\": \"75b4b605c3ff80e7cd04c5b542a8378370bd666b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2185,\n \"license_type\": \"no_license\",\n \"max_line_length\": 99,\n \"num_lines\": 83,\n \"path\": \"/src/amf/misc.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# misc.py has a bunch of random functions that are used now and then.\\n# This is intended as a set of \\\"helper\\\" tools, and probably do not need to be used by the user\\n# hence, why it is not imported in __init__.py\\n\\nimport time\\nimport math\\nimport numpy\\nimport sys\\n\\ndef distance(v1, v2):\\n if len(v1) != len(v2):\\n raise ValueError, \\\"distance: vectors %s and %s are different lengths.\\\" % (repr(v1), repr(v2))\\n\\n runsum = 0.0\\n for idx in xrange(len(v1)):\\n runsum += (v2[idx] - v1[idx]) ** 2\\n\\n return runsum / len(v1)\\n\\n #return sum( (x2 - x1)**2 for x1, x2 in zip(v1, v2) ) / float(len(v1))\\n\\n# subtract two lists from one another\\ndef list_sub(v1, v2):\\n return tuple( x1 - x2 for x1, x2 in zip(v1, v2) )\\n\\ndef list_add(v1, v2):\\n return tuple( x1 + x2 for x1, x2 in zip(v1, v2) )\\n\\ndef col_average(lists):\\n return [ col / len(lists) for col in col_sum(lists) ]\\n\\ndef average(l):\\n return sum(l) / float(len(l))\\n\\n#def stddev(lists):\\n# return tuple(numpy.std(l) for l in lists)\\n\\ndef col_sum(lists):\\n return col_reduce(lambda x, y: x + y, lists)\\n\\ndef col_max(lists):\\n return col_reduce(max, lists)\\n\\ndef col_min(lists):\\n return col_reduce(min, lists)\\n\\ndef get_col(lists, col):\\n \\\"\\\"\\\" Retrieve a single column from a data set as a list. \\\"\\\"\\\"\\n\\n return tuple( row[col] for row in lists )\\n\\ndef col_reduce(function, matrix):\\n \\\"\\\"\\\" performs a 'reduce' operation on each column of a matrix \\\"\\\"\\\"\\n\\n cur = list(matrix[0])\\n\\n for row in matrix[1:]:\\n for idx, item in enumerate(row):\\n cur[idx] = function(cur[idx], item)\\n\\n return tuple(cur)\\n\\ndef xfrange(start, stop, step):\\n if start > stop and step > 0.0 or start < stop and step < 0.0 or step == 0.0:\\n sys.stderr.write(\\\"Bad start/stop/step! %f %f %f \\\\n\\\" % (start, stop, step))\\n\\n else:\\n cur = 0.0\\n while start + cur * step < stop:\\n yield start + cur * step\\n cur += 1.0\\n \\ndef binary(n):\\n \\\"\\\"\\\" Generate all binary numbers of length n \\\"\\\"\\\"\\n out = []\\n _binary_helper(n, (), out)\\n return tuple(out)\\n\\ndef _binary_helper(n, l, output):\\n if n == 0:\\n output.append(l)\\n else:\\n _binary_helper(n-1, l + (0,), output), _binary_helper(n-1, l + (1,), output)\\n\"\n },\n {\n \"alpha_fraction\": 0.8258426785469055,\n \"alphanum_fraction\": 0.8314606547355652,\n \"avg_line_length\": 24.428571701049805,\n \"blob_id\": \"ade6db38966377eae78579263b22144fdbdb265e\",\n \"content_id\": \"509a662bbe5947b5f0167425788a2dd614a6fd88\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 178,\n \"license_type\": \"no_license\",\n \"max_line_length\": 41,\n \"num_lines\": 7,\n \"path\": \"/src/filter.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# load configuration\\nexport USE_FILTER_DESCRIPTOR_SETTING=1\\nsource configuration.sh\\n\\n# filter configurations per demonstration\\necho \\\"=>filtering configurations\\\"\\npython filter.py\\n\"\n },\n {\n \"alpha_fraction\": 0.7972350120544434,\n \"alphanum_fraction\": 0.8018433451652527,\n \"avg_line_length\": 26.125,\n \"blob_id\": \"481628707a47dff288c11b9e6d909231cc7349a3\",\n \"content_id\": \"45f19fd24842bcc08b169ef136eb4ca63171e586\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 217,\n \"license_type\": \"no_license\",\n \"max_line_length\": 67,\n \"num_lines\": 8,\n \"path\": \"/src/match_sketches.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# use neural network image features for matching\\nexport USE_LSH_DESCRIPTOR_SETTING=1\\n\\n# load configuration\\nsource configuration.sh\\n\\n# execute script\\npython match_sketches.py ../data/demonstrations_all/$ABM/sketched .\\n\"\n },\n {\n \"alpha_fraction\": 0.664454996585846,\n \"alphanum_fraction\": 0.6714060306549072,\n \"avg_line_length\": 27.258928298950195,\n \"blob_id\": \"e4774f1caf353de94b7c81dd6ef02ecbe90bf191\",\n \"content_id\": \"1361346e20e305c02277554221ed63446a8c9e8b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3165,\n \"license_type\": \"no_license\",\n \"max_line_length\": 88,\n \"num_lines\": 112,\n \"path\": \"/src/sampling/eum/predict_eum.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import pandas as pd\\nimport numpy as np\\nimport sys\\n\\nfrom negotiation_model import *\\nfrom bdm_agent import *\\n\\n# seed this for fixed environment\\n# for better replication of results\\nrandom.seed(0)\\n\\n# constants\\nINDEPENDENT_VARIABLES = 2\\nDEPENDENT_VARIABLES = 2\\nSAMPLES = 10\\nTIME_LAPSE = 200\\n\\ndef get_salience_values(timeLapse, numberOfAgents):\\n salienceValues = np.zeros((timeLapse, numberOfAgents))\\n\\n for stepNumber in range(timeLapse):\\n for agentNumber in range(numberOfAgents):\\n salienceValues[stepNumber][agentNumber] = random.random()\\n\\n return salienceValues\\n\\n# Defining the model objects\\nclass BDMActor(NegotiationActor):\\n DecisionClass = BDM_Agent\\n\\nclass NegotiationModel_(NegotiationModel):\\n # Variables for median caching\\n median_computed_last = -1\\n median = -1\\n \\n def find_median(self):\\n if self.median_computed_last != self.schedule.steps:\\n self.median = super().find_median()\\n self.median_computed_last = self.schedule.steps\\n return self.median\\n\\nclass ModelOutput:\\n def __init__(self, model):\\n '''\\n Store data from model run.\\n '''\\n self.agent_vars = model.datacollector.get_agent_vars_dataframe()\\n self.model_vars = model.datacollector.get_model_vars_dataframe()\\n self.log = model.log\\n\\ndef load_data():\\n # Load data \\n bookData = pd.read_csv(\\\"sampling/eum/BDM_ColdWar.csv\\\")\\n bookData.Position = (bookData.Position + 100)/200\\n\\n return bookData\\n\\ndef main_eum(qValue, tValue, bookData, salienceValues):\\n # define agents\\n agents = []\\n\\n for i, row in bookData.iterrows():\\n newAgent = BDMActor(row.Country, row.Capability, row.Position, 1)\\n newAgent.decision_model.Q = qValue\\n newAgent.decision_model.T = tValue\\n newAgent.salience = salienceValues[i]\\n agents.append(newAgent)\\n\\n # instantiate model\\n model = NegotiationModel_(agents)\\n\\n # run model\\n for stepNumber in range(TIME_LAPSE):\\n agentNumber = 0\\n for agent in model.agents:\\n #agent.salience = random.random()\\n agent.salience = salienceValues[stepNumber][agentNumber]\\n agentNumber += 1 \\n model.step()\\n\\n # collect data for next steps\\n dependentValues = []\\n modelOutput = ModelOutput(model)\\n dependentValues.append(np.mean(list(modelOutput.model_vars[\\\"Median\\\"][-SAMPLES-1:])))\\n dependentValues.append(np.mean(list(modelOutput.model_vars[\\\"Mean\\\"][-SAMPLES-1:])))\\n \\n return tuple(dependentValues)\\n \\ndef main():\\n # read data\\n bookData = load_data()\\n\\n # get salience values\\n [rows, columns] = bookData.shape\\n salienceValues = get_salience_values(TIME_LAPSE, rows)\\n\\n # simulate using specific ALPs\\n alpConfigurationList = eval(sys.argv[1])\\n\\n # store corresponding slps\\n slpConfigurationList = []\\n\\n for alpConfiguration in alpConfigurationList:\\n [qValue, tValue] = alpConfiguration\\n\\n # run model using those ALPs\\n dependentValues = main_eum(qValue, tValue, bookData, salienceValues)\\n slpConfigurationList.append(dependentValues)\\n\\n print(slpConfigurationList)\\n \\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.648815393447876,\n \"alphanum_fraction\": 0.6504195332527161,\n \"avg_line_length\": 46.11627960205078,\n \"blob_id\": \"35dd288d1da85d245866b180a4d70d40dbca58ec\",\n \"content_id\": \"49685798ac9fa1fd65d6d86ee36b4a5a04ac1523\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 8104,\n \"license_type\": \"no_license\",\n \"max_line_length\": 282,\n \"num_lines\": 172,\n \"path\": \"/src/filter.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import os\\nimport sampling.utils.featurize_image as fi\\nfrom sampling.utils.process_folder_images import process_folder\\nimport sys\\nimport inspect\\nimport glob\\nimport numpy as np\\n\\ncurrentDir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))\\nparentDir = os.path.dirname(currentDir)\\nsys.path.insert(0,parentDir)\\n\\nimport amf.misc as misc\\n\\ndef normalized_distance(_a, _b):\\n _a = np.array(_a)\\n _b = np.array(_b)\\n\\n b = _b.astype(int)\\n a = _a.astype(int)\\n norm_diff = np.linalg.norm(b - a)\\n norm1 = np.linalg.norm(b)\\n norm2 = np.linalg.norm(a)\\n return norm_diff / (norm1 + norm2)\\n\\nclass FilterConfiguration(object):\\n def __init__(self):\\n # get environment variables\\n self.abm = os.environ[\\\"ABM\\\"]\\n self.demonstrationFolder = os.environ[\\\"DEMONSTRATION_FOLDER\\\"]\\n self.descriptorSetting = os.environ[\\\"DESCRIPTOR_SETTING\\\"]\\n self.numDependent = eval(os.environ[\\\"NUM_DEPENDENT\\\"])\\n self.dataFileName = os.environ[\\\"DATA_FILE\\\"]\\n self.scaleDataFileName = os.environ[\\\"SCALE_DATA_FILE\\\"]\\n self.images = eval(os.environ[\\\"IMAGES\\\"])\\n\\n # read configuration\\n configurationFileName = os.environ[\\\"CONFIGURATION_FILE\\\"]\\n configuration = open(configurationFileName).read()\\n execd = {}\\n exec(configuration, execd)\\n self.numIndependent = execd['NUM_INDEPENDENT']\\n \\n# filter selected configurations based on closeness to demonstrations\\ndef select_demonstration_configuratons(filterConfiguration, scaleData):\\n # load network if needed\\n net = fi.get_network(filterConfiguration.descriptorSetting)\\n\\n # for normalization\\n if(scaleData is not None):\\n [dependentValuesMax, dependentValuesMin] = scaleData\\n \\n # process directories in sorted order\\n sortedPaths = []\\n for dirName, subdirList, fileList in os.walk(filterConfiguration.demonstrationFolder):\\n for sortedSubdirName in sorted(subdirList):\\n sortedPaths.append(dirName + \\\"/\\\" + sortedSubdirName)\\n\\n # get slps for demonstrations\\n demonstrationSlpsList = []\\n demonstrationNameList = []\\n for sortedPath in sortedPaths:\\n for dirName, subdirList, fileList in os.walk(sortedPath):\\n # only use non empty subdirectories\\n if(len(fileList) > 0):\\n # traverse all files in this directory\\n if(bool(filterConfiguration.images) == True):\\n # read dependent values as images\\n dependentValues = [tuple(fi.get_features(net, dirName + \\\"/\\\" + fileName, filterConfiguration.descriptorSetting)) for fileName in fileList]\\n else:\\n # read dependent values as text\\n dependentValues = [tuple([eval(x) for x in open(dirName + \\\"/\\\" + fileName).readlines()[0].replace(\\\"\\\\n\\\", \\\"\\\").split(\\\" \\\")]) for fileName in fileList]\\n\\n dependentValues = tuple(misc.col_average(dependentValues))\\n demonstrationSlpsList.append(dependentValues)\\n demonstrationName = dirName.split(\\\"/\\\")[-1].strip(\\\" \\\")\\n demonstrationNameList.append(demonstrationName)\\n\\n minDistanceAlpsList = []\\n minDistanceSlpsList = []\\n for demonstrationIndex in range(len(demonstrationSlpsList)):\\n demonstrationSlps = demonstrationSlpsList[demonstrationIndex]\\n\\n # normalize demonstration slps\\n if(scaleData is not None):\\n demonstrationSlps = tuple((demonstrationSlps[index] - dependentValuesMin[index])/(dependentValuesMax[index] - dependentValuesMin[index]) if (dependentValuesMax[index] > dependentValuesMin[index]) else dependentValuesMax[index] for index in range(len(demonstrationSlps)))\\n \\n distanceList = []\\n \\n with open(\\\"suggested_slps.txt\\\", \\\"r\\\") as suggestedSlpsFile:\\n lines = suggestedSlpsFile.readlines()\\n \\n for suggestionIndex in range(len(lines)):\\n # get slp corresponding to this demonstration\\n line = lines[suggestionIndex]\\n suggestedSlps = eval(line)[demonstrationIndex]\\n\\n if(bool(filterConfiguration.images == True)):\\n if(eval(os.environ[\\\"IMAGE_FEATURIZATION_HOMOGENEOUS\\\"]) == 0):\\n # this is needed if the two image featurization stages are different\\n folderNames = glob.glob(\\\"sampling/\\\" + filterConfiguration.abm + \\\"/images_*\\\")\\n suggestionFolder = folderNames[suggestionIndex]\\n suggestedSlps = process_folder(suggestionFolder)\\n \\n # normalize suggested slps\\n if(scaleData is not None):\\n suggestedSlps = tuple((suggestedSlps[index] - dependentValuesMin[index])/(dependentValuesMax[index] - dependentValuesMin[index]) if (dependentValuesMax[index] > dependentValuesMin[index]) else dependentValuesMax[index] for index in range(len(suggestedSlps)))\\n \\n # compute distance\\n if(os.environ[\\\"DISTANCE_METHOD\\\"] == \\\"NORMALIZED\\\"):\\n distance = normalized_distance(suggestedSlps, demonstrationSlps)\\n else:\\n # default option\\n distance = misc.distance(suggestedSlps, demonstrationSlps)\\n distanceList.append((distance, suggestionIndex, suggestedSlps))\\n\\n # select entry corresponding to smallest distance\\n minDistanceIndex = min(distanceList)[1]\\n minDistanceSlps = min(distanceList)[2]\\n\\n # read the suggested alps corresponding to that index\\n with open(\\\"suggested_alps.txt\\\", \\\"r\\\") as suggestedAlpsFile:\\n minDistanceAlps = eval(suggestedAlpsFile.readlines()[minDistanceIndex])[demonstrationIndex]\\n \\n # write to file\\n directory = \\\"../data/predictions/\\\" + filterConfiguration.abm + \\\"/\\\" + demonstrationNameList[demonstrationIndex] \\n if not os.path.exists(directory):\\n os.makedirs(directory)\\n with open(directory + \\\"/predicted_alps.txt\\\", \\\"w\\\") as predictionFile:\\n predictionFile.write(str(minDistanceAlps) + \\\" \\\" + str(minDistanceSlps))\\n\\n if(bool(filterConfiguration.images == True)):\\n # if images used, copy images\\n # get all image simulation folders\\n folderNames = glob.glob(\\\"sampling/\\\" + filterConfiguration.abm + \\\"/images_*\\\")\\n minFolder = folderNames[minDistanceIndex]\\n os.system(\\\"cp -r \\\" + minFolder + \\\" \\\" + directory)\\n \\n# filter configurations with command line parameters\\ndef main():\\n filterConfiguration = FilterConfiguration()\\n\\n # edit scale data file name based on descriptor setting\\n scaleDataFileNamePrefix = filterConfiguration.scaleDataFileName.split(\\\".\\\")[0]\\n scaleDataFileNameExtension = filterConfiguration.scaleDataFileName.split(\\\".\\\")[1]\\n scaleDataFileName = scaleDataFileNamePrefix + \\\"_\\\" + filterConfiguration.abm\\n if(bool(filterConfiguration.images) == True):\\n scaleDataFileName += \\\"_\\\" + filterConfiguration.descriptorSetting.lower()\\n scaleDataFileName += \\\".\\\" + scaleDataFileNameExtension\\n\\n # get scale information\\n fileName = \\\"sampling/\\\" + filterConfiguration.abm + \\\"/\\\" + scaleDataFileName\\n\\n # scale data assumed not present by default\\n scaleData = None\\n\\n if(os.path.exists(fileName) == True):\\n with open(fileName, \\\"r\\\") as scaleDataFile:\\n # read in order of writing in map.py\\n maxs = eval(scaleDataFile.readline().strip(\\\"\\\\n\\\"))\\n mins = eval(scaleDataFile.readline().strip(\\\"\\\\n\\\"))\\n\\n dependentMaxs = maxs[filterConfiguration.numIndependent:]\\n dependentMins = mins[filterConfiguration.numIndependent:]\\n dependentValuesMax = {i: dependentMaxs[i] for i in range(len(dependentMaxs))}\\n dependentValuesMin = {i: dependentMins[i] for i in range(len(dependentMins))}\\n scaleData = [dependentValuesMax, dependentValuesMin]\\n\\n select_demonstration_configuratons(filterConfiguration, scaleData)\\n \\n# execute main\\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.593353271484375,\n \"alphanum_fraction\": 0.6056758761405945,\n \"avg_line_length\": 22.286956787109375,\n \"blob_id\": \"81b80e6126f8ea8bfb5aacaaa1f1220f4c320dd7\",\n \"content_id\": \"f57dbc1885b6e61862bd3d9c7a5dad507ac44b03\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2678,\n \"license_type\": \"no_license\",\n \"max_line_length\": 129,\n \"num_lines\": 115,\n \"path\": \"/src/sampling/schelling/sample_schelling_parallel.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from model import SchellingModel\\nimport random\\nimport multiprocessing as mp\\n\\n# only use three of the several ALPs\\n# density (0,1)\\n# minority_pc (0,1)\\n\\n# use one model-level SLP\\n# happy\\n\\n# parallel processing constants\\nBATCH_SIZE = 8\\n\\n# constant ALPs\\nHEIGHT = 20\\nWIDTH = 20\\nHOMOPHILY = 4\\n\\n# constants\\n# max_iters in the example is specified as 1000\\n# so we run for 1000 steps\\nINDEPENDENT_VARIABLES = 2\\nDEPENDENT_VARIABLES = 1\\nRUNS = 10000\\nSAMPLES = 10\\nTIME_LAPSE = 1000\\n\\n# process single run data\\ndef process_run(q, run, density, minorityPc):\\n try:\\n line = main_eum(density, minorityPc)\\n except:\\n line = \\\"\\\"\\n \\n # add newline for clean output\\n line = \\\"\\\\n\\\" + line\\n \\n # write to queue\\n q.put(line)\\n \\ndef main_schelling(density, minorityPc):\\n # instantiate and run model\\n model = SchellingModel(\\n height=HEIGHT,\\n width=WIDTH,\\n density=density,\\n minority_pc=minorityPc,\\n homophily=HOMOPHILY)\\n\\n for i in range(SAMPLES):\\n try:\\n # step\\n model.step()\\n except:\\n # saturated\\n # no empty cells\\n pass\\n\\n # collect data for next steps\\n dependentValues = []\\n\\n for i in range(SAMPLES):\\n # step\\n model.step()\\n\\n # read data\\n data = model.datacollector.get_model_vars_dataframe() \\n dependentValues.append(list(data.happy)[-SAMPLES-1:])\\n\\n # print line corresponding to this execution\\n line = str(density) + \\\" \\\" + str(minorityPc)\\n for dependentValue in dependentValues:\\n line += \\\" \\\" + str(dependentValue).replace(\\\"[\\\", \\\"\\\").replace(\\\"]\\\", \\\"\\\").replace(\\\" \\\", \\\"\\\")\\n\\n return line\\n\\n \\ndef main():\\n # header string\\n print(\\\"i\\\"*INDEPENDENT_VARIABLES + \\\"d\\\"*DEPENDENT_VARIABLES)\\n\\n densityValues = []\\n minorityPcValues = []\\n \\n for run in range(RUNS):\\n # sample random ALPs\\n density = random.random() \\n minorityPc = random.random() \\n\\n densityValues.append(density)\\n minorityPcValues.append(minorityPc)\\n\\n # setup a list of processes that we want to run\\n q = mp.Queue()\\n processes = [mp.Process(target=process_run, args=(q, run, densityValues[run], minorityPcValues[run])) for run in range(RUNS)]\\n batchSize = BATCH_SIZE\\n batches = [processes[i:i+batchSize] for i in range(0, len(processes), batchSize)]\\n\\n outFile = open(\\\"output.txt\\\", \\\"a\\\")\\n \\n for batch in batches:\\n # run processes\\n for p in batch:\\n p.start()\\n for p in batch:\\n line = q.get()\\n outFile.write(line)\\n # exit the completed processes\\n for p in batch:\\n p.join()\\n\\n outFile.close()\\n\\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.5906005501747131,\n \"alphanum_fraction\": 0.6073107123374939,\n \"avg_line_length\": 31.457626342773438,\n \"blob_id\": \"9a7f39e9b6aef59e42587e05febaddc8a1629db8\",\n \"content_id\": \"25299fd63872ecf9830f3e159794c1cad234e2cf\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1915,\n \"license_type\": \"no_license\",\n \"max_line_length\": 106,\n \"num_lines\": 59,\n \"path\": \"/snippets/turbulence/plot_suggested_slps.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import matplotlib.pyplot as plt\\nimport numpy as np\\n\\n#demonstrations = [\\\"a\\\", \\\"b\\\", \\\"c\\\"]\\n#demonstrations = [\\\"_a\\\", \\\"_b\\\", \\\"_c\\\"]\\ndemonstrations = [\\\"a\\\"]\\n#demonstrations = [\\\"_a\\\"]\\n#demonstrations = [\\\"a_filtered\\\"]\\n\\nfolderName = \\\"error_weighted\\\"\\n#folderName = \\\"min_error\\\"\\n\\nwith open(\\\"../../src/sampling/turbulence/scale_data_turbulence_density.txt\\\", \\\"r\\\") as inFile:\\n allLines = inFile.readlines()\\n maxValueDensity = eval(allLines[0])[-1]\\n minValueDensity = eval(allLines[1])[-1]\\n\\nfor index, item in enumerate(demonstrations):\\n allPointsX = []\\n allPointsY = []\\n\\n # scatter\\n \\n plt.figure()\\n with open(folderName + \\\"/suggested_slps_\\\" + item + \\\".txt\\\", \\\"r\\\") as inFile:\\n for line in inFile:\\n point = eval(line)[0]\\n\\n # scale\\n #point = tuple([(point[0]-minValueDensity)/(maxValueDensity-minValueDensity)])\\n\\n plt.scatter(point[0], point[0], c=\\\"b\\\", marker=\\\"o\\\", alpha=0.3)\\n allPointsX.append(point[0])\\n allPointsY.append(point[0])\\n \\n plt.scatter(point[0], point[0], c=\\\"b\\\", marker=\\\"o\\\", alpha=0.3, label=\\\"Suggested SLPs\\\")\\n\\n with open(\\\"demonstration_slps.txt\\\", \\\"r\\\") as inFile:\\n point = eval(inFile.readlines()[index])\\n\\n # scale\\n #point = tuple([(point[0]-minValueDensity)/(maxValueDensity-minValueDensity)])\\n \\n plt.scatter(point[0], point[0], c=\\\"r\\\", marker=\\\"s\\\", alpha=1.0, label=\\\"Demonstration SLPs\\\")\\n\\n plt.scatter(np.mean(allPointsX), np.mean(allPointsY), c=\\\"g\\\", marker=\\\"s\\\", alpha=1.0, label=\\\"Mean SLPs\\\")\\n plt.xlabel(\\\"Density\\\")\\n plt.ylabel(\\\"Density\\\")\\n plt.legend()\\n plt.savefig(folderName + \\\"/suggested_slps_\\\" + item)\\n\\n # histogram\\n \\n plt.figure()\\n plt.hist(allPointsX, bins=50, range=(0,50))\\n plt.xlabel(\\\"Density\\\")\\n plt.ylabel(\\\"Number of Points\\\")\\n plt.gca().set_ylim([0,100])\\n plt.savefig(folderName + \\\"/histogram_suggested_slps_\\\" + item)\\n\"\n },\n {\n \"alpha_fraction\": 0.699999988079071,\n \"alphanum_fraction\": 0.7785714268684387,\n \"avg_line_length\": 22.33333396911621,\n \"blob_id\": \"f7f52b9a07b9f0ffd9d5d242d30cadc8e5c9381a\",\n \"content_id\": \"c2c9a561ebca1d1d5f50d1ef74188a8d3d61f2bc\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 140,\n \"license_type\": \"no_license\",\n \"max_line_length\": 54,\n \"num_lines\": 6,\n \"path\": \"/src/domainconfs/turbulence.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"TRAINING_SIZE = 200\\nVALIDATION_SIZE = 2500\\nRM_GRANULARITY = 300\\nREGRESSION = \\\"sklearn.neighbors.KNeighborsRegressor()\\\"\\n\\nNUM_INDEPENDENT = 2\\n\"\n },\n {\n \"alpha_fraction\": 0.5873016119003296,\n \"alphanum_fraction\": 0.5873016119003296,\n \"avg_line_length\": 41,\n \"blob_id\": \"7b5b921b865a519db7dd55473231f2561b10b310\",\n \"content_id\": \"bcbbfa8916d5a7b31b6eeee281019ab173e38515\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 252,\n \"license_type\": \"no_license\",\n \"max_line_length\": 61,\n \"num_lines\": 6,\n \"path\": \"/data/domaindata/fix_results.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# needed in case two results answers the same name identifier\\nwith open(\\\"results.csv\\\", \\\"r\\\") as inFile:\\n with open(\\\"temp.csv\\\", \\\"w\\\") as outFile:\\n for line in inFile:\\n line = line.replace(\\\"|\\\", \\\"\\\\\\\",\\\\\\\"\\\")\\n outFile.write(line)\\n\"\n },\n {\n \"alpha_fraction\": 0.6573289632797241,\n \"alphanum_fraction\": 0.6801303029060364,\n \"avg_line_length\": 22.25757598876953,\n \"blob_id\": \"56265a9ff22d8615011ce681c7f3c10fa2c4981b\",\n \"content_id\": \"bef667f68ffdf3fceb5aef11dcc07b40212654f0\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1535,\n \"license_type\": \"no_license\",\n \"max_line_length\": 67,\n \"num_lines\": 66,\n \"path\": \"/src/sampling/forest_fire/predict_forest_fire.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from model import ForestFire\\nimport random\\nimport sys\\nimport numpy as np\\n\\n# only one ALP\\n# density (0,1)\\n\\n# constant ALPs\\nHEIGHT = 100\\nWIDTH = 100\\n\\n# constants\\n# max_iters in the example is specified as 1000\\n# so we run for 1000 steps\\nINDEPENDENT_VARIABLES = 1\\nDEPENDENT_VARIABLES = 3\\nRUNS = 10000\\nSAMPLES = 10\\nTIME_LAPSE = 200\\n\\ndef main_forest_fire(density=0.65):\\n # instantiate and run model\\n model = ForestFire(height=HEIGHT, width=WIDTH, density=density)\\n\\n # run time lapse\\n for i in range(TIME_LAPSE):\\n model.step()\\n\\n # collect data for next steps\\n dependentValues = []\\n\\n for i in range(SAMPLES):\\n # step\\n model.step()\\n\\n # read data\\n data = model.datacollector.get_model_vars_dataframe()\\n burnedTrees = np.array(list(data.BurnedOut)[-SAMPLES-1:])\\n fineTrees = np.array(list(data.Fine)[-SAMPLES-1:])\\n initialTrees = burnedTrees + fineTrees\\n dependentValues.append(np.mean(burnedTrees/initialTrees))\\n\\n return tuple(dependentValues)\\n \\ndef main():\\n # fixed randomness\\n # parimarily done to test feedback\\n random.seed(0)\\n \\n # simulate using specific ALPs\\n alpConfigurationList = eval(sys.argv[1])\\n\\n # store corresponding slps\\n slpConfigurationList = []\\n \\n for alpConfiguration in alpConfigurationList:\\n [density] = alpConfiguration\\n \\n # run model using those ALPs\\n dependentValues = main_forest_fire(density)\\n slpConfigurationList.append(dependentValues)\\n\\n print(slpConfigurationList)\\n \\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.6421149373054504,\n \"alphanum_fraction\": 0.6500611305236816,\n \"avg_line_length\": 27.452173233032227,\n \"blob_id\": \"9f0341d6f2f2df8aa43a8dffb7056ede562fc780\",\n \"content_id\": \"890c9096b744269d09268aa6c9619c8fa586c910\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3272,\n \"license_type\": \"no_license\",\n \"max_line_length\": 92,\n \"num_lines\": 115,\n \"path\": \"/src/sampling/eum/sample_eum.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import pandas as pd\\nimport numpy as np\\nimport sys\\n\\nfrom negotiation_model import *\\nfrom bdm_agent import *\\n\\n# seed this for fixed environment\\n# for better replication of results\\nrandom.seed(0)\\n\\n# constants\\nINDEPENDENT_VARIABLES = 2\\nDEPENDENT_VARIABLES = 2\\nRUNS = 10000\\nSAMPLES = 10\\nTIME_LAPSE = 200\\n\\ndef get_salience_values(timeLapse, numberOfAgents):\\n salienceValues = np.zeros((timeLapse, numberOfAgents))\\n\\n for stepNumber in range(timeLapse):\\n for agentNumber in range(numberOfAgents):\\n salienceValues[stepNumber][agentNumber] = random.random()\\n\\n return salienceValues\\n\\n# Defining the model objects\\nclass BDMActor(NegotiationActor):\\n DecisionClass = BDM_Agent\\n\\nclass NegotiationModel_(NegotiationModel):\\n # Variables for median caching\\n median_computed_last = -1\\n median = -1\\n \\n def find_median(self):\\n if self.median_computed_last != self.schedule.steps:\\n self.median = super().find_median()\\n self.median_computed_last = self.schedule.steps\\n return self.median\\n\\nclass ModelOutput:\\n def __init__(self, model):\\n '''\\n Store data from model run.\\n '''\\n self.agent_vars = model.datacollector.get_agent_vars_dataframe()\\n self.model_vars = model.datacollector.get_model_vars_dataframe()\\n self.log = model.log\\n\\ndef load_data():\\n # Load data \\n bookData = pd.read_csv(\\\"BDM_ColdWar.csv\\\")\\n bookData.Position = (bookData.Position + 100)/200\\n\\n return bookData\\n\\ndef main_eum(qValue, tValue, bookData, salienceValues):\\n # define agents\\n agents = []\\n\\n for i, row in bookData.iterrows():\\n newAgent = BDMActor(row.Country, row.Capability, row.Position, 1)\\n newAgent.decision_model.Q = qValue\\n newAgent.decision_model.T = tValue\\n newAgent.salience = salienceValues[i]\\n agents.append(newAgent)\\n\\n # instantiate model\\n model = NegotiationModel_(agents)\\n\\n # run model\\n for stepNumber in range(TIME_LAPSE):\\n agentNumber = 0\\n for agent in model.agents:\\n #agent.salience = random.random()\\n agent.salience = salienceValues[stepNumber][agentNumber]\\n agentNumber += 1 \\n model.step()\\n\\n # collect data for next steps\\n dependentValues = []\\n modelOutput = ModelOutput(model)\\n dependentValues.append(list(modelOutput.model_vars[\\\"Median\\\"][-SAMPLES-1:]))\\n dependentValues.append(list(modelOutput.model_vars[\\\"Mean\\\"][-SAMPLES-1:]))\\n \\n # print line corresponding to this execution\\n line = str(qValue) + \\\" \\\" + str(tValue)\\n for dependentValue in dependentValues:\\n line += \\\" \\\" + str(dependentValue).replace(\\\"[\\\", \\\"\\\").replace(\\\"]\\\", \\\"\\\").replace(\\\" \\\", \\\"\\\")\\n print(line)\\n\\ndef main():\\n # header string\\n print(\\\"i\\\"*INDEPENDENT_VARIABLES + \\\"d\\\"*DEPENDENT_VARIABLES)\\n\\n # read data\\n bookData = load_data()\\n\\n # get salience values\\n [rows, columns] = bookData.shape\\n salienceValues = get_salience_values(TIME_LAPSE, rows)\\n \\n for run in range(RUNS):\\n print(run, file=sys.stderr)\\n \\n # sample random ALPs\\n qValue = random.random()\\n tValue = random.random()\\n \\n # run model using those ALPs\\n main_eum(qValue, tValue, bookData, salienceValues)\\n \\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.5925377607345581,\n \"alphanum_fraction\": 0.5987563133239746,\n \"avg_line_length\": 39.2023811340332,\n \"blob_id\": \"dab20f1fa3ceb3346a5f77d3e106f650a24c62ae\",\n \"content_id\": \"d69d05eab60363b2473cf96ff9331a809e81d73e\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3377,\n \"license_type\": \"no_license\",\n \"max_line_length\": 103,\n \"num_lines\": 84,\n \"path\": \"/src/amf/fm.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# fm.py contains the ForwardMapping frontend.\\n\\nimport data\\nimport sys\\nimport numpy as np\\nimport sklearn\\nfrom sklearn.gaussian_process import GaussianProcessRegressor\\nfrom smt.surrogate_models import RBF, KRG, KPLS, KPLSK\\nimport os\\n\\n# This thing is the Forward Mapping\\nclass ForwardMapping(object):\\n\\n # pass in the regression algorithm that we will be using, along with any initialization parameters\\n def __init__(self, regression_class, num_dependent):\\n \\\"\\\"\\\"\\n Initialized a forward mapping.\\n Pass in the regression class to be used, the initial parameters for initlializing this reg class.\\n Also, pass the the number of dependent variables the data set will have.\\n \\\"\\\"\\\"\\n self._mappers = [eval(regression_class) for x in range(num_dependent)]\\n\\n def train(self, training_set):\\n # remove data entries which contain nan\\n training_set = [x for x in training_set if not np.isnan(x[0]).any()]\\n training_set = [x for x in training_set if not np.isnan(x[1]).any()]\\n training_sets = data.split_dep(training_set)\\n\\n for mapper, tset in zip(self._mappers, training_sets):\\n X = np.array([list(x[0]) for x in tset])\\n Y = np.array([x[1] for x in tset])\\n \\n if(\\\"smt\\\" in str(mapper)):\\n # using surrogate modeling toolbox\\n X = np.array([x[0] for x in X])\\n VThreshold = 0.001\\n\\n XSafe = [X[0]]\\n YSafe = [Y[0]]\\n for index in range(1, len(X)):\\n currentX = X[index]\\n currentY = Y[index]\\n XSafetyCheck = XSafe + [currentX]\\n YSafetyCheck = YSafe + [currentY]\\n\\n # we need to only collect those points that do not case matrix inversion problems\\n # check using svd\\n configurationFileName = os.environ[\\\"CONFIGURATION_FILE\\\"]\\n configuration = open(configurationFileName).read()\\n execd = {}\\n exec(configuration, execd)\\n numIndependent = execd['NUM_INDEPENDENT']\\n numDependent = eval(os.environ[\\\"NUM_DEPENDENT\\\"])\\n XSafetyCheck = np.array(XSafetyCheck).reshape(numIndependent, -1)\\n YSafetyCheck = np.array(YSafetyCheck).reshape(numDependent, -1)\\n XYSafetyCheck = np.concatenate((XSafetyCheck, YSafetyCheck), axis=0)\\n U, s, V = np.linalg.svd(XYSafetyCheck, full_matrices=True)\\n V = np.absolute(V)\\n\\n if(np.min(V) > VThreshold):\\n XSafe.append(currentX)\\n YSafe.append(currentY)\\n \\n # convert to numpy array\\n # train model\\n XSafe = np.array(XSafe)\\n YSafe = np.array(YSafe)\\n mapper.set_training_values(XSafe, YSafe)\\n mapper.train()\\n else:\\n mapper.fit(X, Y)\\n\\n def predict(self, configuration):\\n prediction = []\\n for mapper in self._mappers:\\n if(\\\"smt\\\" in str(mapper)):\\n mapperPrediction = mapper.predict_values(np.array(list(configuration)).reshape(1, -1))\\n mapperPrediction = [x[0] for x in mapperPrediction]\\n else:\\n mapperPrediction = mapper.predict(np.array(list(configuration)).reshape(1, -1)) \\n prediction.append(mapperPrediction)\\n prediction = tuple([x[0] for x in prediction])\\n \\n return prediction\\n\"\n },\n {\n \"alpha_fraction\": 0.7027027010917664,\n \"alphanum_fraction\": 0.8108108043670654,\n \"avg_line_length\": 36,\n \"blob_id\": \"80bd3913abf6b1ca1504165909b145c267f58805\",\n \"content_id\": \"8cb3915a5ec3995ca8a235d648d9cf09a614929c\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 37,\n \"license_type\": \"no_license\",\n \"max_line_length\": 36,\n \"num_lines\": 1,\n \"path\": \"/app/run_public_url.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"ngrok http 5000 -subdomain=swarm-lfd\\n\"\n },\n {\n \"alpha_fraction\": 0.6554731726646423,\n \"alphanum_fraction\": 0.6720453500747681,\n \"avg_line_length\": 25.662790298461914,\n \"blob_id\": \"c00719c19bf8a7489d616f74b4ffb39138f77786\",\n \"content_id\": \"dfe90210275d5a2213fa24eaae6b80a264af5f29\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2293,\n \"license_type\": \"no_license\",\n \"max_line_length\": 89,\n \"num_lines\": 86,\n \"path\": \"/src/sampling/civil_violence/predict_civil_violence.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from model import CivilViolenceModel\\nimport random\\nimport sys\\nimport numpy as np\\nimport os\\n\\n# only use three of the several ALPs\\n# citizen_density (0,1)\\n# cop_density (0,1)\\n# legitimacy (0,1)\\n\\n# use three model-level SLPs\\n# Quiescent\\n# Active\\n# Jailed\\n\\n# constant ALPs\\nHEIGHT = 40\\nWIDTH = 40\\nCITIZEN_VISION = 7\\nCOP_VISION = 7\\nMAX_JAIL_TERM = 1000\\n\\n# constants\\n# max_iters in the example is specified as 1000\\n# so we run for 1000 steps\\nINDEPENDENT_VARIABLES = 3\\nDEPENDENT_VARIABLES = 3\\nSAMPLES = 10\\nTIME_LAPSE = 1000\\n\\ndef main_civil_violence(citizenDensity, copDensity, legitimacy):\\n # instantiate and run model\\n model = CivilViolenceModel(\\n height=HEIGHT,\\n width=WIDTH,\\n citizen_vision=CITIZEN_VISION,\\n cop_vision=COP_VISION,\\n max_jail_term=MAX_JAIL_TERM,\\n citizen_density=citizenDensity,\\n cop_density=copDensity,\\n legitimacy=legitimacy)\\n\\n # run time lapse\\n for i in range(TIME_LAPSE):\\n model.step()\\n\\n # collect data for next steps\\n dependentValues = []\\n\\n for i in range(SAMPLES):\\n # step\\n model.step()\\n\\n # read data\\n data = model.dc.get_model_vars_dataframe() \\n dependentValues.append(np.mean(list(data.Quiescent)[-SAMPLES-1:]))\\n dependentValues.append(np.mean(list(data.Active)[-SAMPLES-1:]))\\n dependentValues.append(np.mean(list(data.Jailed)[-SAMPLES-1:]))\\n\\n return tuple(dependentValues)\\n \\ndef main():\\n # simulate using specific ALPs\\n alpConfigurationList = eval(sys.argv[1])\\n\\n # store corresponding slps\\n slpConfigurationList = []\\n \\n for alpConfiguration in alpConfigurationList:\\n [citizenDensity, copDensity, legitimacy] = alpConfiguration\\n\\n # cop density + citizen density should be less than 1\\n if(citizenDensity + copDensity > 1):\\n # invalid parameters\\n # set to infinity so distance is infinite\\n # this value will be considered as a bad suggestion when filtering\\n dependentValues = [float(\\\"Inf\\\")]*eval(os.environ[\\\"NUM_DEPENDENT\\\"])\\n else:\\n # run model using those ALPs\\n dependentValues = main_civil_violence(citizenDensity, copDensity, legitimacy)\\n slpConfigurationList.append(dependentValues)\\n\\n print(slpConfigurationList)\\n \\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.6349663734436035,\n \"alphanum_fraction\": 0.6551392674446106,\n \"avg_line_length\": 25.024999618530273,\n \"blob_id\": \"180657a6bfc68f62ae32fc65df08057a4dceb821\",\n \"content_id\": \"0dac153578d624ace0ca139bc6150746db256084\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2082,\n \"license_type\": \"no_license\",\n \"max_line_length\": 92,\n \"num_lines\": 80,\n \"path\": \"/src/sampling/civil_violence/sample_civil_violence.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from model import CivilViolenceModel\\nimport random\\n\\n# only use three of the several ALPs\\n# citizen_density (0,1)\\n# cop_density (0,1)\\n# legitimacy (0,1)\\n\\n# use three model-level SLPs\\n# Quiescent\\n# Active\\n# Jailed\\n\\n# constant ALPs\\nHEIGHT = 40\\nWIDTH = 40\\nCITIZEN_VISION = 7\\nCOP_VISION = 7\\nMAX_JAIL_TERM = 1000\\n\\n# constants\\n# max_iters in the example is specified as 1000\\n# so we run for 1000 steps\\nINDEPENDENT_VARIABLES = 3\\nDEPENDENT_VARIABLES = 3\\nRUNS = 10000\\nSAMPLES = 10\\nTIME_LAPSE = 1000\\n\\ndef main_civil_violence(citizenDensity, copDensity, legitimacy):\\n # instantiate and run model\\n model = CivilViolenceModel(\\n height=HEIGHT,\\n width=WIDTH,\\n citizen_vision=CITIZEN_VISION,\\n cop_vision=COP_VISION,\\n max_jail_term=MAX_JAIL_TERM,\\n citizen_density=citizenDensity,\\n cop_density=copDensity,\\n legitimacy=legitimacy)\\n\\n # run time lapse\\n for i in range(TIME_LAPSE):\\n model.step()\\n\\n # collect data for next steps\\n dependentValues = []\\n\\n for i in range(SAMPLES):\\n # step\\n model.step()\\n\\n # read data\\n data = model.dc.get_model_vars_dataframe() \\n dependentValues.append(list(data.Quiescent)[-SAMPLES-1:])\\n dependentValues.append(list(data.Active)[-SAMPLES-1:])\\n dependentValues.append(list(data.Jailed)[-SAMPLES-1:])\\n\\n # print line corresponding to this execution\\n line = str(citizenDensity) + \\\" \\\" + str(copDensity) + \\\" \\\" + str(legitimacy)\\n for dependentValue in dependentValues:\\n line += \\\" \\\" + str(dependentValue).replace(\\\"[\\\", \\\"\\\").replace(\\\"]\\\", \\\"\\\").replace(\\\" \\\", \\\"\\\")\\n print(line)\\n\\n \\ndef main():\\n # header string\\n print(\\\"i\\\"*INDEPENDENT_VARIABLES + \\\"d\\\"*DEPENDENT_VARIABLES)\\n\\n for run in range(RUNS):\\n # sample random ALPs\\n # cop density + citizen density should be less than 1\\n citizenDensity = random.random() \\n copDensity = (1-citizenDensity)*random.random() \\n legitimacy = random.random()\\n \\n # run model using those ALPs\\n main_civil_violence(citizenDensity, copDensity, legitimacy)\\n\\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.6672787070274353,\n \"alphanum_fraction\": 0.6705839037895203,\n \"avg_line_length\": 31.807228088378906,\n \"blob_id\": \"dc44bad69890e9034018b1c7ed7e876a945268d2\",\n \"content_id\": \"4069ce92e26a40606230288b7df6e0d449728504\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2723,\n \"license_type\": \"no_license\",\n \"max_line_length\": 198,\n \"num_lines\": 83,\n \"path\": \"/src/map.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# This script takes each row of a data set and processes it for use with AMF\\nimport sys\\nimport math\\nimport os\\n\\nno_entry = \\\"NOENTRY\\\"\\nSCALE_DATA_FILE = \\\"scale_data.txt\\\"\\n\\ndef map_abm(tokens):\\n \\\"\\\"\\\"\\n The raw data set is a combination of floats followed by a combination of lists and looks like this:\\n FLOAT FLOAT ... LIST LIST ...\\n Non-lists are independent variables\\n Lists are dependent variables\\n It takes the average of the list and returns it with the independent variables.\\n \\\"\\\"\\\"\\n\\n tokenAvg = []\\n\\n for token in tokens:\\n if(type(token) == list):\\n avg = sum(token)/len(token)\\n else:\\n avg = token\\n\\n tokenAvg.append(avg)\\n\\n return tuple(tokenAvg)\\n\\ndef linear_scale(dataset):\\n \\\"\\\"\\\"Linearly scales an entire data set to a range of 0-1\\\"\\\"\\\"\\n\\n maxs = list(dataset[0])\\n mins = list(dataset[0])\\n\\n for datarow in dataset:\\n for idx, item in enumerate(datarow):\\n if item == no_entry: continue\\n\\n maxs[idx] = max(maxs[idx], item)\\n mins[idx] = min(mins[idx], item)\\n\\n scaleDataFileName = SCALE_DATA_FILE\\n abm = os.environ[\\\"ABM\\\"]\\n descriptorSetting = os.environ[\\\"DESCRIPTOR_SETTING\\\"]\\n\\n # edit scale data file name based on descriptor setting\\n scaleDataFileNamePrefix = scaleDataFileName.split(\\\".\\\")[0]\\n scaleDataFileNameExtension = scaleDataFileName.split(\\\".\\\")[1]\\n scaleDataFileName = scaleDataFileNamePrefix + \\\"_\\\" + abm\\n if(len(descriptorSetting) > 0):\\n # valid descriptor\\n scaleDataFileName += \\\"_\\\" + descriptorSetting.lower()\\n scaleDataFileName += \\\".\\\" + scaleDataFileNameExtension\\n \\n with open(scaleDataFileName, \\\"w\\\") as scaleDataFile:\\n scaleDataFile.write(str(maxs))\\n scaleDataFile.write(\\\"\\\\n\\\")\\n scaleDataFile.write(str(mins))\\n\\n return [ [ (no_entry if item == no_entry else maxs[idx] if maxs[idx] == mins[idx] else (item - mins[idx]) / (maxs[idx] - mins[idx])) for idx, item in enumerate(datarow) ] for datarow in dataset ]\\n\\ndef parse(data_file_path):\\n \\\"\\\"\\\"\\n Applies 'map_function' to each row (as a string) of a data set.\\n 'map_function' should takes a list of tokens as the only parameter and return a string as the only parameter.\\n Prints out the new data set.\\n \\\"\\\"\\\"\\n\\n return [ [ float(token) if token.find(',') == -1 else [ float(item) for item in token.split(',') ] for token in line.split() ] for line in open(data_file_path).xreadlines() ]\\n\\ndef dump(dataset):\\n print \\\"\\\\n\\\".join( \\\" \\\".join(str(item) for item in row) for row in dataset)\\n\\ndef process_abm():\\n parsed = parse(sys.argv[1])\\n averaged = map(map_abm, parsed)\\n scaled = linear_scale(averaged)\\n dump(scaled)\\n\\n# average over system level paramters in raw data file\\n# raw data file is provided as argument\\nprocess_abm()\\n\"\n },\n {\n \"alpha_fraction\": 0.6269554495811462,\n \"alphanum_fraction\": 0.6281588673591614,\n \"avg_line_length\": 26.09782600402832,\n \"blob_id\": \"910587596e08507a8f46aa10de2d59ca35161a15\",\n \"content_id\": \"a29d16ac260c7abde4aa19fa97ee8bb0378abaa3\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2493,\n \"license_type\": \"no_license\",\n \"max_line_length\": 130,\n \"num_lines\": 92,\n \"path\": \"/src/sampling/utils/process_training_images.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# libraries used\\nimport sys\\nimport os\\nimport featurize_image as fi\\nimport multiprocessing as mp\\nimport glob\\n\\n# parallel processing constants\\nBATCH_SIZE = 8\\n\\n# process single run data\\ndef process_run(q, run, runConfiguration, line):\\n\\n # unpack run configuration\\n [inFolder, dataFile, numDependent, net, descriptorSetting] = runConfiguration\\n\\n # initialize collection of dependent values\\n depValuesStrings = [\\\"\\\"]*numDependent\\n\\n for imageName in glob.glob(inFolder + \\\"/\\\" + str(run) + \\\"_*.*\\\"):\\n # read image\\n print >> sys.stderr, (\\\"now processing \\\" + imageName)\\n \\n # define dependent values\\n depValues = fi.get_features(net, imageName, descriptorSetting)\\n\\n for index in range(len(depValues)):\\n depValuesStrings[index] += str(depValues[index]) + \\\",\\\"\\n\\n # add value data to line\\n for depValuesString in depValuesStrings:\\n line += \\\" \\\" + depValuesString[:-1]\\n\\n # remove extra \\\",\\\" at end\\n line = line + \\\"\\\\n\\\"\\n\\n # write to queue\\n q.put(line)\\n\\n# read images and extract features as dependent variables\\ndef main():\\n\\n # get environment variables\\n abm = os.environ[\\\"ABM\\\"]\\n descriptorSetting = os.environ[\\\"DESCRIPTOR_SETTING\\\"]\\n numDependent = eval(os.environ[\\\"NUM_DEPENDENT\\\"])\\n dataFileName = os.environ[\\\"DATA_FILE\\\"]\\n\\n dataFile = open(abm + \\\".txt\\\", \\\"r\\\")\\n tempFile = open(\\\"temp.txt\\\", \\\"w\\\")\\n\\n # set up image reading\\n inFolder = abm + \\\"/images\\\"\\n\\n # load feature extractor if needed\\n net = fi.get_network(descriptorSetting)\\n\\n # read data\\n lines = dataFile.readlines()\\n\\n # get number of runs\\n runs = len(lines)\\n \\n # process runs\\n runConfiguration = [inFolder, dataFile, numDependent, net, descriptorSetting]\\n\\n # setup a list of processes that we want to run\\n q = mp.Queue()\\n processes = [mp.Process(target=process_run, args=(q, run, runConfiguration, lines[run].strip(\\\" \\\\n\\\"))) for run in range(runs)] \\n batchSize = BATCH_SIZE\\n batches = [processes[i:i+batchSize] for i in range(0, len(processes), batchSize)]\\n\\n for batch in batches:\\n # run processes\\n for p in batch:\\n p.start()\\n for p in batch:\\n line = q.get()\\n tempFile.write(line)\\n # exit the completed processes\\n for p in batch:\\n p.join()\\n\\n # close files and rename\\n dataFile.close()\\n tempFile.close()\\n\\n os.remove(abm + \\\".txt\\\")\\n os.rename(\\\"temp.txt\\\", abm + \\\".txt\\\")\\n\\nif __name__ == \\\"__main__\\\":\\n main()\\n\"\n },\n {\n \"alpha_fraction\": 0.5741397738456726,\n \"alphanum_fraction\": 0.5817003846168518,\n \"avg_line_length\": 35.206703186035156,\n \"blob_id\": \"9310d151e1411d6b2469b6a34580daca836eca07\",\n \"content_id\": \"6e231729abd51e8557ec5f9e45a0633308bc892a\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 6481,\n \"license_type\": \"no_license\",\n \"max_line_length\": 121,\n \"num_lines\": 179,\n \"path\": \"/src/sampling/utils/outlier_detection.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import numpy as np\\nimport scipy.io as sio\\nfrom sets import Set\\nfrom ltfatpy.sigproc.thresh import thresh\\nfrom sklearn.cluster import KMeans, MeanShift\\nfrom collections import Counter\\n\\nimport sys\\n\\ndef lpod(trainData, testData, trainLabels, testLabels, lNum, tol, lambdaValue):\\n # this is a python port of the original matlab code provided by the authors\\n # of the 2017 paper titled \\\"Efficient Outlier Detection for High-Dimensional Data\\\"\\n \\n # supress invalid division warning\\n # this happens in the matlab code too so we're ok\\n np.seterr(divide='ignore', invalid='ignore')\\n \\n # size of x and y\\n # check for size mismatch\\n [rtrainData,ctrainData] = trainData.shape\\n [rtestData,ctestData] = testData.shape\\n assert rtrainData == rtestData\\n\\n # allocate memory to the maximum size\\n # T is the score matrix of trainData\\n # P is the loading matrix of trainData\\n # U is the score matrix of testData\\n # Q is the loading matrix of testData\\n # B is the matrix of regression coefficient\\n # W is the weight matrix of trainData \\n n = max(ctrainData, ctestData)\\n T = np.zeros((rtrainData, n)) \\n P = np.zeros((ctrainData, n)) \\n U = np.zeros((rtestData, n)) \\n Q = np.zeros((ctestData, n)) \\n B = np.zeros((n, n)) \\n W = P \\n k = 0\\n\\n # iteration loop if residual is larger than specfied\\n iteration = 1\\n\\n while(iteration < lNum and np.linalg.norm(testData) > tol and k < n):\\n # choose the column of x has the largest square of sum as t\\n # choose the column of y has the largest square of sum as u\\n tidx = np.argmax(sum(np.multiply(trainData, trainData)))\\n uidx = np.argmax(sum(np.multiply(testData, testData)))\\n t1 = trainData[:, tidx]\\n u = testData[:, uidx]\\n t = np.zeros((rtrainData, 1))\\n\\n # iteration for outer modeling until convergence\\n while (np.linalg.norm(t1 - t) > tol):\\n # computing the weight vector of trainData\\n # computing the score vector of trainData\\n # obtaining q by soft-threshold\\n w = np.matmul(trainData.transpose(), u) \\n w = w / np.linalg.norm(w)\\n t = t1\\n t1 = np.matmul(trainData, w) \\n q, dummy = thresh(np.matmul(testData.transpose(), t1), lambdaValue, thresh_type=\\\"soft\\\")\\n \\n # computing the loading vector of testData according to t\\n # computing the score vector of testData\\n q = q / np.linalg.norm(q)\\n u = np.matmul(testData, q) \\n \\n # update p based on t\\n # computing the loading vecor of trainData\\n t = t1\\n p = np.matmul(trainData.transpose(), t) / np.matmul(t.transpose(), t)\\n pnorm = np.linalg.norm(p)\\n p = p / pnorm \\n t = t * pnorm\\n w = w * pnorm\\n\\n # regression and residuals\\n t = t.reshape((-1, 1))\\n p = p.reshape((-1, 1))\\n u = u.reshape((-1, 1))\\n q = q.reshape((-1, 1))\\n\\n b = np.matmul(u.transpose(), t) / np.matmul(t.transpose(), t)\\n trainData = trainData - np.matmul(t, p.transpose())\\n testData = testData - (b * np.matmul(t, q.transpose()))\\n\\n # save iteration results to outputs:\\n k += 1\\n T[:, k] = list(t)\\n P[:, k] = list(p)\\n U[:, k] = list(u)\\n Q[:, k] = list(q)\\n W[:, k] = list(w)\\n B[k, k] = b\\n \\n iteration += 1\\n\\n # end of iterations\\n T = T[:, :k]\\n P = P[:, :k]\\n U = U[:, :k]\\n Q = Q[:, :k]\\n W = W[:, :k]\\n B = B[:k, :k]\\n\\n # prediction stage:\\n # predicting real-valued matrix for test_data\\n # transforming outputs into the corresponding label matrix\\n # 0.5 is the default threshold value\\n outputs = np.matmul(np.matmul(np.matmul(P, B), Q.transpose()).transpose(), trainLabels)\\n [numClass, numTesting] = outputs.shape\\n predictionLabels = np.zeros((numClass, numTesting))\\n for i in range(numTesting):\\n for j in range(numClass):\\n if(outputs[j,i] >= 0.5): \\n predictionLabels[j,i] = 1\\n else:\\n predictionLabels[j,i] = 0\\n\\n return predictionLabels\\n\\ndef do_lpod(X, Y):\\n # outlier detection\\n Y = Y.astype(float).transpose()\\n allDataLength = len(Y)\\n trainDataFraction = 0.7\\n trainDataLength = np.floor(trainDataFraction * allDataLength)\\n \\n # create logical index vector\\n # randomize order to select random elements\\n trainIndexes = np.random.choice(int(allDataLength), int(trainDataLength), replace=False)\\n testIndexes = np.setdiff1d(range(allDataLength), trainIndexes)\\n \\n trainData = X[trainIndexes, :]\\n trainLabels = Y[trainIndexes]\\n testData = X[testIndexes, :]\\n testLabels = Y[testIndexes]\\n \\n # hyper parameters\\n # get predictions\\n lNum = 50\\n tol = 1e-10\\n lambdaValue = 0.1\\n predictionLabels = lpod(trainData.transpose(), testData.transpose(), trainLabels, testLabels, lNum, tol, lambdaValue)\\n lpodX = np.concatenate((trainData, testData), axis=0)\\n lpodY = np.concatenate((trainLabels, testLabels), axis=0)\\n\\n return [lpodX, lpodY]\\n\\ndef get_approximate_labels(allConfigurations):\\n # cluster points\\n # outlier is decided as the minority cluster\\n # count population of each cluster\\n # outliers are labeled as 1\\n # inliers are labeled as 0\\n # force 2 clusters as backup in case meanshift finds a single cluster\\n X = np.array(allConfigurations)\\n meanShift = MeanShift(bin_seeding=True)\\n meanShift.fit(X)\\n nClusters = 2\\n kmeans = KMeans(init='k-means++', n_clusters=nClusters)\\n kmeans.fit(X)\\n labelCounts = Counter(meanShift.labels_)\\n\\n if(len(labelCounts.keys()) > 1):\\n # multiple clusters found by meanshift\\n labels = meanShift.labels_\\n else:\\n # single cluster found by meanshift\\n # use kmeans clusters instead\\n labels = kmeans.labels_\\n \\n labelCounts = Counter(labels)\\n minLabel = min(labelCounts, key=labelCounts.get)\\n labels = ['1' if x == minLabel else '0' for x in labels]\\n labels = [eval(x) for x in labels]\\n labels = np.array(labels, ndmin=2)\\n \\n return [X, labels]\\n\"\n },\n {\n \"alpha_fraction\": 0.6721804738044739,\n \"alphanum_fraction\": 0.6827067732810974,\n \"avg_line_length\": 16.972972869873047,\n \"blob_id\": \"feca9bd17df0f29a1e5cb57306497076c663157d\",\n \"content_id\": \"29d5276f4eddeedd6e0e60a5b60961e973ddda45\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Java\",\n \"length_bytes\": 665,\n \"license_type\": \"no_license\",\n \"max_line_length\": 62,\n \"num_lines\": 37,\n \"path\": \"/src/sampling/heatbugs/run_heatbugs.java\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import heatbugs.HeatBugs;\\nimport sim.engine.*;\\n\\npublic class run_heatbugs extends SimState\\n{\\n\\tpublic run_heatbugs(long seed)\\n\\t{\\n\\t\\tsuper(seed);\\t\\n\\t}\\n\\n\\tpublic static void main(String[] args)\\n\\t{\\n\\t\\t// initialize\\n\\t\\tSimState state = new HeatBugs(System.currentTimeMillis());\\n\\t\\t\\n\\t\\t// set parameters\\n\\t\\tHeatBugs heatBugs = (HeatBugs)state;\\n\\t\\theatBugs.setDiffusionConstant(0.0);\\n\\n\\t\\t// run\\n\\t\\tstate.start();\\n\\n\\t\\tdo\\n\\t\\t{\\n\\t\\t\\tif (!state.schedule.step(state)) \\n\\t\\t\\t\\tbreak;\\n\\t\\t}\\n\\t\\twhile(state.schedule.getSteps() < 5000);\\n\\t\\t\\n\\t\\t// save state as image\\n\\t\\t// get some state information\\n\\t\\tSystem.out.println(heatBugs.getRandomMovementProbability());\\n\\n\\t\\tstate.finish();\\n\\t\\tSystem.exit(0);\\n\\t}\\n}\\n\"\n },\n {\n \"alpha_fraction\": 0.6717171669006348,\n \"alphanum_fraction\": 0.6717171669006348,\n \"avg_line_length\": 27.285715103149414,\n \"blob_id\": \"024ca2f638c42bb5954b9c78058c458f022b7583\",\n \"content_id\": \"2d8eb79139ebed4a23ead99f5fa490eaf658bb34\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 198,\n \"license_type\": \"no_license\",\n \"max_line_length\": 74,\n \"num_lines\": 7,\n \"path\": \"/snippets/plot_error.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import numpy as np\\n\\nwith open(\\\"../src/suggested_alps_fm_prediction_error.txt\\\", \\\"r\\\") as inFile:\\n errors = [eval(x) for x in inFile.readlines()]\\n\\n print np.mean(errors)\\n print np.std(errors)\\n\"\n },\n {\n \"alpha_fraction\": 0.62347012758255,\n \"alphanum_fraction\": 0.6335493326187134,\n \"avg_line_length\": 26.50494956970215,\n \"blob_id\": \"4299bab84a5f3e6c72a9b75a73bf215dd333796e\",\n \"content_id\": \"39d4907cc464658a0fd1a679c1ee7b0268a5acea\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2778,\n \"license_type\": \"no_license\",\n \"max_line_length\": 105,\n \"num_lines\": 101,\n \"path\": \"/app/app_server.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import os\\nimport base64\\nimport glob\\n\\nfrom flask import Flask, render_template, request\\nfrom werkzeug import secure_filename\\n\\n\\nUPLOAD_FOLDER = \\\"sketches/\\\"\\nUPLOAD_FILENAME = \\\"demonstration.png\\\"\\nMATCH_FOLDER = \\\"matches/\\\"\\nABM = \\\"flocking\\\"\\n\\napp = Flask(__name__)\\napp.config[\\\"UPLOAD_FOLDER\\\"] = UPLOAD_FOLDER\\n\\n\\n@app.route('/')\\ndef root():\\n return \\\"You are here: /\\\"\\n\\n\\n@app.route(\\\"/index\\\")\\ndef index():\\n return render_template(\\\"index.html\\\")\\n\\n@app.route(\\\"/upload\\\", methods=[\\\"GET\\\", \\\"POST\\\"])\\ndef upload_file():\\n\\n if (request.method == \\\"POST\\\"):\\n # get the base64 encoded image\\n encodedImage = request.form[\\\"encodedImage\\\"]\\n\\n # save to disk\\n with open(\\\"demonstration.jpg\\\", \\\"wb\\\") as outFile:\\n outFile.write(base64.b64decode(encodedImage))\\n\\n # convert to png\\n os.system(\\\"convert demonstration.jpg demonstration.png\\\")\\n os.system(\\\"rm -f demonstration.jpg\\\")\\n \\n # clear old data\\n os.system(\\\"rm -f \\\" + UPLOAD_FOLDER + \\\"*\\\")\\n os.system(\\\"rm -f \\\" + MATCH_FOLDER + \\\"*\\\")\\n\\n # save file in selected location\\n fileName = os.path.join(app.config[\\\"UPLOAD_FOLDER\\\"], UPLOAD_FILENAME)\\n os.system(\\\"mv demonstration.png \\\" + fileName)\\n\\n # initiate matching and call pipeline background process\\n os.system(\\\"rm -rf \\\" + \\\"../data/predictions/\\\" + ABM + \\\"/\\\" + MATCH_FOLDER + \\\"/*\\\")\\n os.system('curl localhost:5001/match')\\n os.system(\\\"./do_pipeline.sh &\\\")\\n \\n # refresh will be a get request\\n return render_template(\\\"redirect.html\\\")\\n\\n\\n@app.route(\\\"/redirect\\\")\\ndef redirect():\\n # get background process status\\n fileExists = os.path.isfile(\\\"../data/predictions/\\\" + ABM + \\\"/\\\" + MATCH_FOLDER + \\\"predicted_alps.txt\\\")\\n\\n return str(fileExists)\\n\\n\\n@app.route('/results')\\ndef results():\\n return render_template(\\\"results.html\\\")\\n\\n\\n@app.route('/alps')\\ndef get_suggested_alps():\\n # get suggested ALPs from file\\n # this will be the response sent\\n with open(\\\"../data/predictions/\\\" + ABM + \\\"/\\\" + MATCH_FOLDER + \\\"predicted_alps.txt\\\") as inFile:\\n prediction = inFile.readlines()[0]\\n suggestedAlps = prediction.split(\\\" (\\\")[0]\\n\\n return suggestedAlps\\n\\n\\n@app.route('/match')\\ndef get_match():\\n # return matched image as json\\n imageName = glob.glob(\\\"matches/*.png\\\")[0]\\n with open(imageName, \\\"rb\\\") as imageFile:\\n imageString = base64.b64encode(imageFile.read())\\n\\n return imageString\\n\\n\\n@app.route('/prediction')\\ndef get_prediction():\\n # return matched image as json\\n imageFolder = glob.glob(\\\"../data/predictions/\\\" + ABM + \\\"/\\\" + MATCH_FOLDER + \\\"/images_*\\\")[0]\\n imageName = imageFolder + \\\"/0_200.png\\\"\\n with open(imageName, \\\"rb\\\") as imageFile:\\n imageString = base64.b64encode(imageFile.read())\\n\\n return imageString\\n\"\n },\n {\n \"alpha_fraction\": 0.729411780834198,\n \"alphanum_fraction\": 0.729411780834198,\n \"avg_line_length\": 16,\n \"blob_id\": \"9b5e514746b13a44a2de2147a8dee5e93c9a49cf\",\n \"content_id\": \"c241871ffcbca08fc11761bb6eb8677ac11fe5ed\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 85,\n \"license_type\": \"no_license\",\n \"max_line_length\": 25,\n \"num_lines\": 5,\n \"path\": \"/src/sampling/sampling.sh\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# generate raw file\\n#./sampling_raw.sh\\n\\n# process raw file\\n./sampling_process_raw.sh\\n\"\n },\n {\n \"alpha_fraction\": 0.6430445909500122,\n \"alphanum_fraction\": 0.654199481010437,\n \"avg_line_length\": 42.485713958740234,\n \"blob_id\": \"5b7dd6773462dab1fca49591af102fe140960d1b\",\n \"content_id\": \"dda6cd16ec97f0010a4b62e76231ba9d1702d50f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1524,\n \"license_type\": \"no_license\",\n \"max_line_length\": 120,\n \"num_lines\": 35,\n \"path\": \"/data/dataset_analysis/simulations/plot_fm_prediction_error.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import matplotlib.pyplot as plt\\nimport numpy as np\\n\\npointSelectionMethods = [\\\"error_weighted\\\", \\\"min_error\\\"]\\ndemonstrations = [\\\"1\\\", \\\"2\\\", \\\"3\\\"]\\n\\n# plot fm prediction error for various demonstrations\\nfor pointSelectionMethod in pointSelectionMethods:\\n # plot unguided and guided data\\n plt.figure()\\n\\n print(pointSelectionMethod)\\n \\n for demonstration in demonstrations:\\n with open(pointSelectionMethod + \\\"/\\\" + demonstration + \\\"/unguided/fm_prediction_error.txt\\\") as unguidedDataFile:\\n unguidedData = [eval(x) for x in unguidedDataFile.readlines()]\\n with open(pointSelectionMethod + \\\"/\\\" + demonstration + \\\"/guided/fm_prediction_error.txt\\\") as guidedDataFile:\\n guidedData = [eval(x) for x in guidedDataFile.readlines()]\\n\\n mean = np.mean(unguidedData)\\n sigma = np.std(unguidedData)\\n plot1 = plt.errorbar(eval(demonstration), mean, sigma, fmt='s', color='b', markersize=20, elinewidth=2)\\n print(mean, sigma)\\n \\n mean = np.mean(guidedData)\\n sigma = np.std(guidedData)\\n plot2 = plt.errorbar(eval(demonstration), mean, sigma, fmt='s', color='g', markersize=20, elinewidth=2)\\n print(mean, sigma)\\n \\n plt.gca().set_ylim(bottom=-0.01)\\n plt.legend([plot1, plot2], [\\\"Unguided Data\\\", \\\"Guided Data\\\"], loc=\\\"best\\\")\\n plt.xticks(range(len(demonstrations)+2), [\\\" \\\", \\\"a\\\", \\\"b\\\", \\\"c\\\", \\\" \\\"])\\n plt.xlabel(\\\"Demonstration (Demo)\\\")\\n plt.ylabel(\\\"FM Prediction Error\\\")\\n plt.savefig(pointSelectionMethod + \\\".png\\\")\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.628926694393158,\n \"alphanum_fraction\": 0.6511780023574829,\n \"avg_line_length\": 23.253969192504883,\n \"blob_id\": \"303c78a0e88e4f0ff01976c0e5b4273178738f70\",\n \"content_id\": \"8be2c577064da046efa5eff6857b099eb04f8b71\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1528,\n \"license_type\": \"no_license\",\n \"max_line_length\": 92,\n \"num_lines\": 63,\n \"path\": \"/src/sampling/forest_fire/sample_forest_fire.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from model import ForestFire\\nimport random\\nimport numpy as np\\n\\n# only one ALP\\n# density (0,1)\\n\\n# constant ALPs\\nHEIGHT = 100\\nWIDTH = 100\\n\\n# constants\\n# max_iters in the example is specified as 1000\\n# so we run for 1000 steps\\nINDEPENDENT_VARIABLES = 1\\nDEPENDENT_VARIABLES = 1\\nRUNS = 10000\\nSAMPLES = 10\\nTIME_LAPSE = 200\\n\\ndef main_forest_fire(density=0.65):\\n # instantiate and run model\\n model = ForestFire(height=HEIGHT, width=WIDTH, density=density)\\n\\n # run time lapse\\n for i in range(TIME_LAPSE):\\n model.step()\\n\\n # collect data for next steps\\n dependentValues = []\\n\\n for i in range(SAMPLES):\\n # step\\n model.step()\\n\\n # read data\\n data = model.datacollector.get_model_vars_dataframe()\\n burnedTrees = np.array(list(data.BurnedOut)[-SAMPLES-1:])\\n fineTrees = np.array(list(data.Fine)[-SAMPLES-1:])\\n initialTrees = burnedTrees + fineTrees\\n dependentValues.append(list(burnedTrees/initialTrees))\\n\\n # print line corresponding to this execution\\n line = str(density)\\n for dependentValue in dependentValues:\\n line += \\\" \\\" + str(dependentValue).replace(\\\"[\\\", \\\"\\\").replace(\\\"]\\\", \\\"\\\").replace(\\\" \\\", \\\"\\\")\\n print(line)\\n \\ndef main():\\n\\n # header string\\n print(\\\"i\\\"*INDEPENDENT_VARIABLES + \\\"d\\\"*DEPENDENT_VARIABLES)\\n\\n for run in range(RUNS):\\n\\n # sample random ALPs\\n # cop density + citizen density should be less than 1 \\n density = random.random()\\n \\n # run model using those ALPs\\n main_forest_fire(density)\\n\\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.7006369233131409,\n \"alphanum_fraction\": 0.7707006335258484,\n \"avg_line_length\": 21.428571701049805,\n \"blob_id\": \"3ca71683ff14d9dfb27a0017fde89a5bc7de12d7\",\n \"content_id\": \"2ed16a83192c7e76496fc97354113e21bbd05dd0\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 157,\n \"license_type\": \"no_license\",\n \"max_line_length\": 54,\n \"num_lines\": 7,\n \"path\": \"/src/domainconfs/aids.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"TRAINING_SIZE = 300\\nVALIDATION_SIZE = 1000\\nRM_GRANULARITY = 55\\nREGRESSION = \\\"sklearn.neighbors.KNeighborsRegressor()\\\"\\n\\nNUM_INDEPENDENT = 2\\nNUM_DEPENDENT = 2\\n\"\n },\n {\n \"alpha_fraction\": 0.6805850863456726,\n \"alphanum_fraction\": 0.6828457713127136,\n \"avg_line_length\": 41.97142791748047,\n \"blob_id\": \"fbffb23a39ddbe55972330d45169fce39ff864ba\",\n \"content_id\": \"03dc45db4a8abb59562161fde903a2043907ebc2\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 7520,\n \"license_type\": \"no_license\",\n \"max_line_length\": 287,\n \"num_lines\": 175,\n \"path\": \"/src/match_sketches.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import os\\nimport sampling.utils.featurize_image as fi\\nimport sys\\nimport glob\\nimport amf.misc\\nimport matplotlib.pyplot as plt\\nfrom sklearn.neighbors import LSHForest\\nimport numpy as np\\nimport cPickle\\nimport random\\n\\ndef initialize_featurization():\\n # edit scale data file name based on descriptor setting\\n scaleDataFileNamePrefix = os.environ[\\\"SCALE_DATA_FILE\\\"].split(\\\".\\\")[0]\\n scaleDataFileNameExtension = os.environ[\\\"SCALE_DATA_FILE\\\"].split(\\\".\\\")[1]\\n scaleDataFileName = scaleDataFileNamePrefix + \\\"_\\\" + os.environ[\\\"ABM\\\"]\\n if(bool(eval(os.environ[\\\"IMAGES\\\"])) == True):\\n scaleDataFileName += \\\"_\\\" + os.environ[\\\"DESCRIPTOR_SETTING\\\"].lower()\\n scaleDataFileName += \\\".\\\" + scaleDataFileNameExtension\\n\\n # get scale information\\n with open(\\\"sampling/\\\" + os.environ[\\\"ABM\\\"] + \\\"/\\\" + scaleDataFileName, \\\"r\\\") as scaleDataFile:\\n # read in order of writing in map.py\\n maxs = eval(scaleDataFile.readline().strip(\\\"\\\\n\\\"))\\n mins = eval(scaleDataFile.readline().strip(\\\"\\\\n\\\"))\\n\\n dependentMaxs = maxs[-eval(os.environ[\\\"NUM_DEPENDENT\\\"]):]\\n dependentMins = mins[-eval(os.environ[\\\"NUM_DEPENDENT\\\"]):]\\n dependentValuesMax = {i: dependentMaxs[i] for i in range(len(dependentMaxs))}\\n dependentValuesMin = {i: dependentMins[i] for i in range(len(dependentMins))}\\n scaleData = [dependentValuesMax, dependentValuesMin]\\n\\n \\n # load feature extractor if needed\\n net = fi.get_network(os.environ[\\\"DESCRIPTOR_SETTING\\\"])\\n\\n return [net, scaleData]\\n\\ndef get_dependent_values_image(net, scaleData, imageName):\\n # define dependent values\\n dependentValuesImage = fi.get_features(net, imageName, os.environ[\\\"DESCRIPTOR_SETTING\\\"])\\n # normalize\\n [dependentValuesMax, dependentValuesMin] = scaleData\\n dependentValuesImage = tuple((dependentValuesImage[index] - dependentValuesMin[index])/(dependentValuesMax[index] - dependentValuesMin[index]) if (dependentValuesMax[index] > dependentValuesMin[index]) else dependentValuesMax[index] for index in range(len(dependentValuesImage)))\\n\\n return dependentValuesImage\\n\\ndef get_dependent_values_folder(net, scaleData, folderName, downsample=1):\\n # initialize collection of dependent values\\n imageNamesFolder = []\\n dependentValuesFolder = []\\n \\n # expect everything in this folder to be an image\\n # read sketches data\\n allFileNames = glob.glob(folderName + \\\"/*.*\\\")\\n allFileNames = random.sample(allFileNames, int(len(allFileNames)/downsample))\\n for imageName in sorted(allFileNames):\\n # featurize image\\n print >> sys.stderr, (\\\"now processing \\\" + imageName) \\n dependentValuesImage = get_dependent_values_image(net, scaleData, imageName)\\n\\n # append to list\\n imageNamesFolder.append(imageName)\\n dependentValuesFolder.append(dependentValuesImage)\\n\\n return [imageNamesFolder, dependentValuesFolder]\\n\\ndef do_matching_lsh(net, scaleData, imageNamesSketchList, dependentValuesSketchList):\\n lshModelFileName = \\\"../../_swarm-lfd-data/\\\" + os.environ[\\\"ABM\\\"] + \\\"/lsh_forest_model\\\" + \\\"_\\\" + os.environ[\\\"DESCRIPTOR_SETTING\\\"].lower() + \\\".p\\\"\\n\\n if(os.path.exists(lshModelFileName) == False):\\n # no model found in directory\\n # train a locality sensitive hashing model on the simulation data\\n [imageNamesSimulationList, dependentValuesSimulationList] = get_dependent_values_folder(net, scaleData, \\\"../../_swarm-lfd-data/\\\" + os.environ[\\\"ABM\\\"] + \\\"/images\\\")\\n lshForest = LSHForest()\\n lshForest.fit(dependentValuesSimulationList)\\n\\n # use disk to cache results\\n with open(lshModelFileName, \\\"wb\\\") as outFile:\\n cPickle.dump(lshForest, outFile)\\n with open(\\\"image_names_simulation.p\\\", \\\"wb\\\") as outFile:\\n cPickle.dump(imageNamesSimulationList, outFile)\\n else:\\n # use existing model\\n with open(lshModelFileName, \\\"rb\\\") as inFile:\\n lshForest = cPickle.load(inFile)\\n with open(\\\"../../_swarm-lfd-data/\\\" + os.environ[\\\"ABM\\\"] + \\\"/image_names_simulation.p\\\", \\\"rb\\\") as inFile:\\n imageNamesSimulationList = cPickle.load(inFile)\\n\\n # now match\\n print >> sys.stderr, \\\"\\\\n\\\"\\n print >> sys.stderr, \\\"now matching\\\"\\n matches = []\\n for imageNameSketch, dependentValuesSketch in zip(imageNamesSketchList, dependentValuesSketchList):\\n minDistance = float('Inf')\\n minDistanceImage = None\\n distances = []\\n\\n # use lsh to get neighbors\\n [distances, indices] = lshForest.kneighbors(np.array(dependentValuesSketch).reshape(1, -1))\\n\\n # use indices to get closest images\\n # read image name and image feature files\\n # then only compare with the few retrieved\\n # this is only useful if LSH features are different from features used in the loop\\n for index in indices[0]:\\n imageNameSimulation = imageNamesSimulationList[index]\\n dependentValuesSimulation = get_dependent_values_image(net, scaleData, imageNameSimulation)\\n\\n # compute distance\\n distance = amf.misc.distance(dependentValuesSketch, dependentValuesSimulation)\\n if(distance < minDistance):\\n minDistance = distance\\n minDistanceImage = imageNameSimulation\\n\\n # pick minimum distance\\n matches.append((imageNameSketch, minDistanceImage))\\n\\n # clear memory\\n del lshForest\\n\\n return matches\\n\\ndef do_matching(net, scaleData, imageNamesSketchList, dependentValuesSketchList):\\n\\n # get raw features\\n # [imageNamesSimulationList, dependentValuesSimulationList] = get_dependent_values_folder(net, scaleData, \\\"../../_swarm-lfd-data/\\\" + os.environ[\\\"ABM\\\"] + \\\"/images\\\", downsample=11000)\\n [imageNamesSimulationList, dependentValuesSimulationList] = get_dependent_values_folder(net, scaleData, \\\"../../_swarm-lfd-data/\\\" + os.environ[\\\"ABM\\\"] + \\\"/images_1000\\\")\\n\\n # now match\\n print >> sys.stderr, \\\"\\\\n\\\"\\n print >> sys.stderr, \\\"now matching\\\"\\n matches = []\\n for imageNameSketch, dependentValuesSketch in zip(imageNamesSketchList, dependentValuesSketchList):\\n minDistance = float('Inf')\\n minDistanceImage = None\\n distances = []\\n\\n for imageNameSimulation, dependentValuesSimulation in zip(imageNamesSimulationList, dependentValuesSimulationList):\\n # compute distance\\n distance = amf.misc.distance(dependentValuesSketch, dependentValuesSimulation)\\n if(distance < minDistance):\\n minDistance = distance\\n minDistanceImage = imageNameSimulation\\n\\n # pick minimum distance\\n matches.append((imageNameSketch, minDistanceImage))\\n\\n return matches\\n\\ndef main():\\n # get sketches folder\\n sketchesFolder = sys.argv[1]\\n \\n # initialize and retrieve features\\n [net, scaleData] = initialize_featurization()\\n [imageNamesSketchList, dependentValuesSketchList] = get_dependent_values_folder(net, scaleData, sketchesFolder)\\n\\n # matches = do_matching_lsh(net, scaleData, imageNamesSketchList, dependentValuesSketchList)\\n matches = do_matching(net, scaleData, imageNamesSketchList, dependentValuesSketchList)\\n\\n # print matches\\n # save matches\\n simulationSaveFolder = sys.argv[2]\\n print >> sys.stderr, \\\"\\\\n\\\"\\n print >> sys.stderr, \\\"now printing matches:\\\"\\n print >> sys.stderr, \\\"\\\\n\\\"\\n for x, y in matches:\\n print >> sys.stderr, x\\n print >> sys.stderr, y\\n print >> sys.stderr, \\\"\\\\n\\\"\\n\\n os.system(\\\"cp \\\" + y + \\\" \\\" + simulationSaveFolder)\\n \\nmain()\\n\"\n },\n {\n \"alpha_fraction\": 0.7051281929016113,\n \"alphanum_fraction\": 0.7692307829856873,\n \"avg_line_length\": 21.285715103149414,\n \"blob_id\": \"1fd3ccaa98a7e798627221abe13a8c2be0c6fee9\",\n \"content_id\": \"e335465a61bbee713df5f15a1d4610f47810a677\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 156,\n \"license_type\": \"no_license\",\n \"max_line_length\": 54,\n \"num_lines\": 7,\n \"path\": \"/src/domainconfs/civil_violence.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"TRAINING_SIZE = 330\\nVALIDATION_SIZE = 100\\nRM_GRANULARITY = 40\\nREGRESSION = \\\"sklearn.neighbors.KNeighborsRegressor()\\\"\\n\\nNUM_INDEPENDENT = 3\\nNUM_DEPENDENT = 3\\n\"\n },\n {\n \"alpha_fraction\": 0.6403161883354187,\n \"alphanum_fraction\": 0.7351778745651245,\n \"avg_line_length\": 22,\n \"blob_id\": \"02ab3a125819d197a0b3c5083757951007e66128\",\n \"content_id\": \"effe63c2db702f02d9b6fa78d4ead36edce8b704\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 253,\n \"license_type\": \"no_license\",\n \"max_line_length\": 54,\n \"num_lines\": 11,\n \"path\": \"/src/domainconfs/wolf_sheep.py\",\n \"repo_name\": \"karanbudhraja/PABMD\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#TRAINING_SIZES = range(40, 440, 40)\\n#VALIDATION_SIZE = 200\\n#RM_GRANULARITIES = range(15, 65, 10)\\n\\nTRAINING_SIZE = 40\\nVALIDATION_SIZE = 10\\nRM_GRANULARITY = 20\\nREGRESSION = \\\"sklearn.neighbors.KNeighborsRegressor()\\\"\\n\\nNUM_INDEPENDENT = 5\\nNUM_DEPENDENT = 2\\n\"\n }\n]"},"num_files":{"kind":"number","value":63,"string":"63"}}},{"rowIdx":2980397,"cells":{"repo_name":{"kind":"string","value":"agamdamaraju/SentimentAnalysis"},"repo_url":{"kind":"string","value":"https://github.com/agamdamaraju/SentimentAnalysis"},"snapshot_id":{"kind":"string","value":"882808e49f2aba8aaa3754e637db37dfbbe22dbc"},"revision_id":{"kind":"string","value":"66b154605d063747557485b58fcebe3bdfe3f017"},"directory_id":{"kind":"string","value":"d84bb130efff2c97bad0c5a0f44f44e50f72850e"},"branch_name":{"kind":"string","value":"refs/heads/main"},"visit_date":{"kind":"timestamp","value":"2023-01-24T07:33:32.928004","string":"2023-01-24T07:33:32.928004"},"revision_date":{"kind":"timestamp","value":"2020-12-03T14:41:06","string":"2020-12-03T14:41:06"},"committer_date":{"kind":"timestamp","value":"2020-12-03T14:41:06","string":"2020-12-03T14:41:06"},"github_id":{"kind":"null"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"null"},"gha_updated_at":{"kind":"null"},"gha_pushed_at":{"kind":"null"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.7746614813804626,"alphanum_fraction":0.7756286263465881,"avg_line_length":33.46666717529297,"blob_id":"2ac342185bad117905486901764b0b8e8ce530e7","content_id":"4cfc2ea4b1c894a07b9e73e2354b9467c855d764","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":1034,"license_type":"no_license","max_line_length":152,"num_lines":30,"path":"/README.md","repo_name":"agamdamaraju/SentimentAnalysis","src_encoding":"UTF-8","text":"# Sentiment Analysis\nA project of sentiment analysis based on twitter dataset.\n\n# Motivation\nSentiment analysis is extremely useful in social media monitoring as it allows us to gain an overview of the wider public opinion behind certain topics.\n\n# About\nThis is purely Python programming based project from scratch, which classifies the overall twitter data based on negative and positive scores. \n\n# Process overview\n- Reading dataset.\n- Data preprocessing\n - Lower case conversion.\n - Punctualtion removal.\n - Tokenization (Organising tweets, retweet count and reply count).\n - Modeling (No Packages are used).\n - Positive score calculation based on predefined positive words.\n - Negative score calculation based on predefined negative words.\n - Net score calculation.\n- Writing calculated data in csv file.\n\n# Editor\nSpyder 4\n\n# Scripts \nPython(.py) [Code file](https://github.com/AgamDamaraju/SentimentAnalysis/) is attached.\n\n# Credits\n- [University of Michigan](https://umich.edu/)\n- [Coursera](https://www.coursera.org)\n"},{"alpha_fraction":0.5751683115959167,"alphanum_fraction":0.5804038643836975,"avg_line_length":29.11627960205078,"blob_id":"b119b896710114fbf5c93a83b56dde98ca483e90","content_id":"392e5b60a4ac7e68b52cc20e5f23862ac3e0867f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2674,"license_type":"no_license","max_line_length":106,"num_lines":86,"path":"/sentiment_analysis.py","repo_name":"agamdamaraju/SentimentAnalysis","src_encoding":"UTF-8","text":"# List of punctuations\r\npunctuation_chars = [\"'\", '\"', \",\", \".\", \"!\", \":\", \";\", '#', '@']\r\n\r\n# Function for removing punctuations from the word\r\ndef strip_punctuation(word):\r\n punct = \",\".join(punctuation_chars) \r\n for i in word:\r\n if i in punct:\r\n word = word.replace(i,\"\")\r\n return word\r\n\r\n# Function for counting positive words\r\ndef get_pos(sentence):\r\n punctWords = sentence.split()\r\n pos_wrds = 0\r\n for i in punctWords:\r\n words = strip_punctuation(i)\r\n words_lower = words.lower()\r\n if words_lower in positive_words:\r\n pos_wrds += 1\r\n return pos_wrds\r\n\r\n# Function for counting negative words\r\ndef get_neg(sentence):\r\n punctWords = sentence.split()\r\n neg_wrds = 0\r\n for i in punctWords:\r\n words = strip_punctuation(i)\r\n words_lower = words.lower()\r\n if words_lower in negative_words:\r\n neg_wrds += 1\r\n return neg_wrds\r\n\r\n# Function for positive words collection\r\npositive_words = []\r\nwith open(\"positive_words.txt\") as pos_f:\r\n for lin in pos_f:\r\n if lin[0] != ';' and lin[0] != '\\n':\r\n positive_words.append(lin.strip())\r\n\r\n# Function for negative words collection\r\nnegative_words = []\r\nwith open(\"negative_words.txt\") as pos_f:\r\n for lin in pos_f:\r\n if lin[0] != ';' and lin[0] != '\\n':\r\n negative_words.append(lin.strip())\r\n\r\n# Collection for tweets, retweets and replies\r\nreT_reP = []\r\ntxt = []\r\nwith open(\"project_twitter_data.csv\") as data_read:\r\n for lin in data_read:\r\n for char in lin:\r\n if char == \"\\n\":\r\n lin = lin.replace(char, \"\")\r\n lst = lin.split(\",\")\r\n reT_reP.append(tuple(lst[-2:]))\r\n txt.append(tuple(lst[:-2]))\r\n retrep = reT_reP[1:] \r\n text = txt[1:] \r\n \r\nretweets = []\r\nreplies = []\r\nfor tup in retrep:\r\n retweets.append(tup[0])\r\n replies.append(tup[1])\r\n\r\n# Evaluation of Positive and Negative scores\r\npositive_score = []\r\nnegative_score = []\r\nfor tup in text: \r\n for txt in tup:\r\n positive_score.append(get_pos(txt))\r\n negative_score.append(get_neg(txt))\r\n\r\n# Evaluation of Net score\r\nnet_score = []\r\nfor ps, ns in zip(positive_score, negative_score):\r\n nt = ps-ns\r\n net_score.append(nt)\r\n\r\n# Writing results in csv format \r\nwith open(\"resulting_data.csv\", \"w\") as data_write:\r\n data_write.write(\"Number of Retweets, Number of Replies, Positive Score, Negative Score, Net Score\\n\")\r\n for rt, rp, ps, ns, nt in zip(retweets, replies, positive_score, negative_score, net_score):\r\n data_write.write(str(rt)+\", \"+str(rp)+\", \"+str(ps)+\", \"+str(ns)+\", \"+str(nt)+\"\\n\")"}],"string":"[\n {\n \"alpha_fraction\": 0.7746614813804626,\n \"alphanum_fraction\": 0.7756286263465881,\n \"avg_line_length\": 33.46666717529297,\n \"blob_id\": \"2ac342185bad117905486901764b0b8e8ce530e7\",\n \"content_id\": \"4cfc2ea4b1c894a07b9e73e2354b9467c855d764\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 1034,\n \"license_type\": \"no_license\",\n \"max_line_length\": 152,\n \"num_lines\": 30,\n \"path\": \"/README.md\",\n \"repo_name\": \"agamdamaraju/SentimentAnalysis\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Sentiment Analysis\\nA project of sentiment analysis based on twitter dataset.\\n\\n# Motivation\\nSentiment analysis is extremely useful in social media monitoring as it allows us to gain an overview of the wider public opinion behind certain topics.\\n\\n# About\\nThis is purely Python programming based project from scratch, which classifies the overall twitter data based on negative and positive scores. \\n\\n# Process overview\\n- Reading dataset.\\n- Data preprocessing\\n - Lower case conversion.\\n - Punctualtion removal.\\n - Tokenization (Organising tweets, retweet count and reply count).\\n - Modeling (No Packages are used).\\n - Positive score calculation based on predefined positive words.\\n - Negative score calculation based on predefined negative words.\\n - Net score calculation.\\n- Writing calculated data in csv file.\\n\\n# Editor\\nSpyder 4\\n\\n# Scripts \\nPython(.py) [Code file](https://github.com/AgamDamaraju/SentimentAnalysis/) is attached.\\n\\n# Credits\\n- [University of Michigan](https://umich.edu/)\\n- [Coursera](https://www.coursera.org)\\n\"\n },\n {\n \"alpha_fraction\": 0.5751683115959167,\n \"alphanum_fraction\": 0.5804038643836975,\n \"avg_line_length\": 29.11627960205078,\n \"blob_id\": \"b119b896710114fbf5c93a83b56dde98ca483e90\",\n \"content_id\": \"392e5b60a4ac7e68b52cc20e5f23862ac3e0867f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2674,\n \"license_type\": \"no_license\",\n \"max_line_length\": 106,\n \"num_lines\": 86,\n \"path\": \"/sentiment_analysis.py\",\n \"repo_name\": \"agamdamaraju/SentimentAnalysis\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# List of punctuations\\r\\npunctuation_chars = [\\\"'\\\", '\\\"', \\\",\\\", \\\".\\\", \\\"!\\\", \\\":\\\", \\\";\\\", '#', '@']\\r\\n\\r\\n# Function for removing punctuations from the word\\r\\ndef strip_punctuation(word):\\r\\n punct = \\\",\\\".join(punctuation_chars) \\r\\n for i in word:\\r\\n if i in punct:\\r\\n word = word.replace(i,\\\"\\\")\\r\\n return word\\r\\n\\r\\n# Function for counting positive words\\r\\ndef get_pos(sentence):\\r\\n punctWords = sentence.split()\\r\\n pos_wrds = 0\\r\\n for i in punctWords:\\r\\n words = strip_punctuation(i)\\r\\n words_lower = words.lower()\\r\\n if words_lower in positive_words:\\r\\n pos_wrds += 1\\r\\n return pos_wrds\\r\\n\\r\\n# Function for counting negative words\\r\\ndef get_neg(sentence):\\r\\n punctWords = sentence.split()\\r\\n neg_wrds = 0\\r\\n for i in punctWords:\\r\\n words = strip_punctuation(i)\\r\\n words_lower = words.lower()\\r\\n if words_lower in negative_words:\\r\\n neg_wrds += 1\\r\\n return neg_wrds\\r\\n\\r\\n# Function for positive words collection\\r\\npositive_words = []\\r\\nwith open(\\\"positive_words.txt\\\") as pos_f:\\r\\n for lin in pos_f:\\r\\n if lin[0] != ';' and lin[0] != '\\\\n':\\r\\n positive_words.append(lin.strip())\\r\\n\\r\\n# Function for negative words collection\\r\\nnegative_words = []\\r\\nwith open(\\\"negative_words.txt\\\") as pos_f:\\r\\n for lin in pos_f:\\r\\n if lin[0] != ';' and lin[0] != '\\\\n':\\r\\n negative_words.append(lin.strip())\\r\\n\\r\\n# Collection for tweets, retweets and replies\\r\\nreT_reP = []\\r\\ntxt = []\\r\\nwith open(\\\"project_twitter_data.csv\\\") as data_read:\\r\\n for lin in data_read:\\r\\n for char in lin:\\r\\n if char == \\\"\\\\n\\\":\\r\\n lin = lin.replace(char, \\\"\\\")\\r\\n lst = lin.split(\\\",\\\")\\r\\n reT_reP.append(tuple(lst[-2:]))\\r\\n txt.append(tuple(lst[:-2]))\\r\\n retrep = reT_reP[1:] \\r\\n text = txt[1:] \\r\\n \\r\\nretweets = []\\r\\nreplies = []\\r\\nfor tup in retrep:\\r\\n retweets.append(tup[0])\\r\\n replies.append(tup[1])\\r\\n\\r\\n# Evaluation of Positive and Negative scores\\r\\npositive_score = []\\r\\nnegative_score = []\\r\\nfor tup in text: \\r\\n for txt in tup:\\r\\n positive_score.append(get_pos(txt))\\r\\n negative_score.append(get_neg(txt))\\r\\n\\r\\n# Evaluation of Net score\\r\\nnet_score = []\\r\\nfor ps, ns in zip(positive_score, negative_score):\\r\\n nt = ps-ns\\r\\n net_score.append(nt)\\r\\n\\r\\n# Writing results in csv format \\r\\nwith open(\\\"resulting_data.csv\\\", \\\"w\\\") as data_write:\\r\\n data_write.write(\\\"Number of Retweets, Number of Replies, Positive Score, Negative Score, Net Score\\\\n\\\")\\r\\n for rt, rp, ps, ns, nt in zip(retweets, replies, positive_score, negative_score, net_score):\\r\\n data_write.write(str(rt)+\\\", \\\"+str(rp)+\\\", \\\"+str(ps)+\\\", \\\"+str(ns)+\\\", \\\"+str(nt)+\\\"\\\\n\\\")\"\n }\n]"},"num_files":{"kind":"number","value":2,"string":"2"}}},{"rowIdx":2980398,"cells":{"repo_name":{"kind":"string","value":"sinramyeon/youtubewidget"},"repo_url":{"kind":"string","value":"https://github.com/sinramyeon/youtubewidget"},"snapshot_id":{"kind":"string","value":"5729e84c10cc2ec75af041a8ae047be88b03a726"},"revision_id":{"kind":"string","value":"fbd765f322f9b3d0082d45d7eca82c384031afa3"},"directory_id":{"kind":"string","value":"a4b05502ed720ecbd5d36068e510e9f4bee729bf"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2021-09-08T07:04:52.267137","string":"2021-09-08T07:04:52.267137"},"revision_date":{"kind":"timestamp","value":"2018-03-08T05:30:11","string":"2018-03-08T05:30:11"},"committer_date":{"kind":"timestamp","value":"2018-03-08T05:30:11","string":"2018-03-08T05:30:11"},"github_id":{"kind":"null"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"null"},"gha_updated_at":{"kind":"null"},"gha_pushed_at":{"kind":"null"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.6171284914016724,"alphanum_fraction":0.6171284914016724,"avg_line_length":23.060606002807617,"blob_id":"b62fcb0075f4691eadd65e79422b26ebb5245563","content_id":"e10d6c707ff5f83f0d9131fb7a317802cc9ba5de","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":794,"license_type":"no_license","max_line_length":87,"num_lines":33,"path":"/youtubeTitle.py","repo_name":"sinramyeon/youtubewidget","src_encoding":"UTF-8","text":"from __future__ import print_function\nfrom pywinauto import Desktop\nimport platform\nimport appscript\n\ndef FormatUrl(url):\n if url and not url.startswith(\"https://\"): \n return \"http://\" + url\n return url\n\ndef GetURL() : \n \n if platform.system() == \"Darwin\" :\n urls = appscript.app('Google Chrome').windows.tabs.URL()\n print(platform)\n \n else :\n urls = appscript.app('Google Chrome').windows.tabs.URL()\n\n chrome_window = Desktop(backend=\"uia\").window(class_name_re='Chrome')\n chrome = chrome_window['Google Chrome']\n\n address_bar_wrapper = chrome_window['Google Chrome'].main.Edit.wrapper_object()\n\n url = address_bar_wrapper.legacy_properties()\n\n print(url['Value'])\n\n\n \n \n def GetTitle() \n: pass\n"},{"alpha_fraction":0.8481012582778931,"alphanum_fraction":0.8481012582778931,"avg_line_length":39,"blob_id":"aef7e7807472eb71fe62dbe210a89a21ad9f8911","content_id":"6336181f911ae6f1d42b7b9af9e9411832516ff4","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":79,"license_type":"no_license","max_line_length":63,"num_lines":2,"path":"/README.md","repo_name":"sinramyeon/youtubewidget","src_encoding":"UTF-8","text":"# youtubewidget\nsimple youtube widget for windows made by python with pywinauto"}],"string":"[\n {\n \"alpha_fraction\": 0.6171284914016724,\n \"alphanum_fraction\": 0.6171284914016724,\n \"avg_line_length\": 23.060606002807617,\n \"blob_id\": \"b62fcb0075f4691eadd65e79422b26ebb5245563\",\n \"content_id\": \"e10d6c707ff5f83f0d9131fb7a317802cc9ba5de\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 794,\n \"license_type\": \"no_license\",\n \"max_line_length\": 87,\n \"num_lines\": 33,\n \"path\": \"/youtubeTitle.py\",\n \"repo_name\": \"sinramyeon/youtubewidget\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from __future__ import print_function\\nfrom pywinauto import Desktop\\nimport platform\\nimport appscript\\n\\ndef FormatUrl(url):\\n if url and not url.startswith(\\\"https://\\\"): \\n return \\\"http://\\\" + url\\n return url\\n\\ndef GetURL() : \\n \\n if platform.system() == \\\"Darwin\\\" :\\n urls = appscript.app('Google Chrome').windows.tabs.URL()\\n print(platform)\\n \\n else :\\n urls = appscript.app('Google Chrome').windows.tabs.URL()\\n\\n chrome_window = Desktop(backend=\\\"uia\\\").window(class_name_re='Chrome')\\n chrome = chrome_window['Google Chrome']\\n\\n address_bar_wrapper = chrome_window['Google Chrome'].main.Edit.wrapper_object()\\n\\n url = address_bar_wrapper.legacy_properties()\\n\\n print(url['Value'])\\n\\n\\n \\n \\n def GetTitle() \\n: pass\\n\"\n },\n {\n \"alpha_fraction\": 0.8481012582778931,\n \"alphanum_fraction\": 0.8481012582778931,\n \"avg_line_length\": 39,\n \"blob_id\": \"aef7e7807472eb71fe62dbe210a89a21ad9f8911\",\n \"content_id\": \"6336181f911ae6f1d42b7b9af9e9411832516ff4\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 79,\n \"license_type\": \"no_license\",\n \"max_line_length\": 63,\n \"num_lines\": 2,\n \"path\": \"/README.md\",\n \"repo_name\": \"sinramyeon/youtubewidget\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# youtubewidget\\nsimple youtube widget for windows made by python with pywinauto\"\n }\n]"},"num_files":{"kind":"number","value":2,"string":"2"}}},{"rowIdx":2980399,"cells":{"repo_name":{"kind":"string","value":"Z-Shuming/NXSpider"},"repo_url":{"kind":"string","value":"https://github.com/Z-Shuming/NXSpider"},"snapshot_id":{"kind":"string","value":"4d0fcf693bf004e22747b8af28bf4c7b27170c75"},"revision_id":{"kind":"string","value":"68e588c0612d0ab2af3a820ff88ca24d698ceeb7"},"directory_id":{"kind":"string","value":"d605fdd342d09e9ad696a3a729e20ee07753c712"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2022-01-12T23:08:27.402606","string":"2022-01-12T23:08:27.402606"},"revision_date":{"kind":"timestamp","value":"2018-12-17T15:36:03","string":"2018-12-17T15:36:03"},"committer_date":{"kind":"timestamp","value":"2018-12-17T15:36:03","string":"2018-12-17T15:36:03"},"github_id":{"kind":"null"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"null"},"gha_updated_at":{"kind":"null"},"gha_pushed_at":{"kind":"null"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.5572916865348816,"alphanum_fraction":0.625,"avg_line_length":13.769230842590332,"blob_id":"027b9a5be43029169a1f49de2046171a372c7ddf","content_id":"19bce644bfaed95012b5d8f501f37dc2ed6af270","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":192,"license_type":"permissive","max_line_length":35,"num_lines":13,"path":"/NXSpider/debug_test/test.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/26.\n# email to LipsonChan@yahoo.com\n#\n\nfrom NXSpider.spider.mp3 import Mp3\n\nm1=Mp3(__parse_recursion__={})\nm=Mp3()\n\nprint('a')\n"},{"alpha_fraction":0.5741106867790222,"alphanum_fraction":0.5839921236038208,"avg_line_length":34.30232620239258,"blob_id":"c5f018043076de0f21cb1cc885ece28ee375d3b6","content_id":"24d33f857c7a2f47650656a4986dada0d36e2c6a","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3036,"license_type":"permissive","max_line_length":119,"num_lines":86,"path":"/NXSpider/bin/show_ctrl.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/6/19.\n# email to LipsonChan@yahoo.com\n#\nimport sys\n\nfrom cement.core.controller import expose\nfrom terminaltables import AsciiTable\n\nfrom NXSpider.bin.base_ctrl import NXSpiderBaseController, py2_decoding\nfrom NXSpider.bin.print_as_table import print_mp3s, print_playlist, print_users, print_albums, print_artists, print_mvs\nfrom NXSpider.common import PYTHON2, log\nfrom NXSpider.spider import api\nfrom NXSpider.spider.api import search_types, PLAYLIST_CLASSES\n\nPRINT_ATTR_FUNC_MAP = {\n 'mp3': ['songs', print_mp3s],\n 'playlist': ['playlists', print_playlist],\n 'user': ['userprofiles', print_users],\n 'artist': ['artists', print_artists],\n 'album': ['albums', print_albums],\n 'mv': ['mvs', print_mvs],\n}\n\n\nclass ShowController(NXSpiderBaseController):\n class Meta:\n label = \"show\"\n stacked_on = 'base'\n description = \"NXSpider\"\n\n @expose(help=\"search [-ar ] [-pl ] \"\n \"[-ur ] [-mp3 ] [-ab ] [-mv ]\")\n def search(self):\n search_key = 'mp3'\n key_num = 0\n for k, v in search_types.items():\n if getattr(self.app.pargs, k, None):\n search_key = k\n key_num += 1\n if key_num > 1:\n log.print_err(\"it could search by only one type\")\n\n # input must be decode in python2\n search_value = getattr(self.app.pargs, search_key)\n search_value = py2_decoding(search_value)\n\n res = api.search(search_value, stype=search_key,\n offset=self.app.pargs.offset or 0,\n limit=self.app.pargs.limit or 50)\n\n if not res:\n log.print_info(\"nothing found!\")\n return\n\n if search_key in PRINT_ATTR_FUNC_MAP:\n func = PRINT_ATTR_FUNC_MAP[search_key][1] # type: function\n value = (res.get(PRINT_ATTR_FUNC_MAP[search_key][0], [])) # type: list\n func(value)\n\n @expose(help=\"show artists ablum, usage: sw-ar-ab -ar [-offset ] [-limit ]\")\n def sw_ar_ab(self):\n if self.param_check(['artist'], sys._getframe().f_code.co_name) is False:\n return\n\n artistid = self.app.pargs.artist # type: list\n artist_detail = api.get_artist_album(artistid,\n offset=self.app.pargs.offset or 0,\n limit=self.app.pargs.limit or 50)\n\n print_albums(artist_detail['hotAlbums'])\n pass\n\n @expose(help=\"show user playlist, usage: sw-ur-pl -ur \")\n def sw_ur_pl(self):\n # todo\n pass\n\n @expose(help=\"show all playlist classes, usage: sw-pl-classes\")\n def sw_pl_classes(self):\n table = AsciiTable([[\"Group\", \"Classes(which can be as a input)\"]])\n table_data = [[k, ', '.join(v)] for k, v in PLAYLIST_CLASSES.items()]\n table.table_data.extend(table_data)\n print(table.table)\n"},{"alpha_fraction":0.6136986017227173,"alphanum_fraction":0.6356164216995239,"avg_line_length":15.590909004211426,"blob_id":"701e757e158767823c9594c3083424df6fd2f639","content_id":"1304c1c7d11f468089209fc287f77c8c06c59d0e","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":365,"license_type":"permissive","max_line_length":42,"num_lines":22,"path":"/NXSpider/common/log.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/19.\n# email to LipsonChan@yahoo.com\n#\n\nfrom colorama import Fore\nfrom colorama import init\ninit(autoreset=True)\n\n\ndef print_err(msg):\n print(Fore.RED + \"ERROR: \" + msg)\n\n\ndef print_warn(msg):\n print(Fore.YELLOW + \"WARNING: \" + msg)\n\n\ndef print_info(msg):\n print(Fore.GREEN + \"INFO: \" + msg)\n"},{"alpha_fraction":0.5047619342803955,"alphanum_fraction":0.5809524059295654,"avg_line_length":16.16666603088379,"blob_id":"4eaae03255274fb51cbfa6f2850b7d21fca58981","content_id":"aeae71ab6357de4e2205afdda64a00e9943a049a","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":105,"license_type":"permissive","max_line_length":33,"num_lines":6,"path":"/NXSpider/spider/__init__.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/19.\n# email to LipsonChan@yahoo.com\n#\n\n\n"},{"alpha_fraction":0.5492264628410339,"alphanum_fraction":0.5724331736564636,"avg_line_length":22.700000762939453,"blob_id":"ebc2a604a7f6c68d5488ad84a11f8c9ae271a78c","content_id":"4c5a603565b92ea36bbf1b9c1daa670330e06022","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1422,"license_type":"permissive","max_line_length":72,"num_lines":60,"path":"/setup.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/5/21.\n# email to LipsonChan@yahoo.com\n#\nimport os\n\nfrom setuptools import setup, find_packages, Command\n\nfrom NXSpider import version\n\n\nclass CleanCommand(Command):\n \"\"\"Custom clean command to tidy up the project root.\"\"\"\n user_options = []\n\n def initialize_options(self):\n pass\n\n def finalize_options(self):\n pass\n\n def run(self):\n os.system('rm -vrf ./build ./dist ./*.pyc ./*.tgz ./*.egg-info')\n\n\nsetup(\n name=version.__title__,\n version=version.__version__,\n author=version.__author__,\n author_email=version.__author_email__,\n url=version.__url__,\n description=version.__description__,\n license=version.__license__,\n packages=find_packages(),\n include_package_data=True,\n install_requires=[\n \"mutagen>=1.40,<2\",\n \"terminaltables>=3,<4\",\n \"cryptography>=2.2,<3\",\n \"requests>=2.19,<3\",\n \"colorama>=0.3,<0.4\",\n \"cement>=2,<3\",\n \"six>=1,<=2\",\n ],\n entry_points={\n \"console_scripts\": [\"nxspider=NXSpider.bin.cli:main\"]\n },\n classifiers=[\n 'Environment :: Console',\n 'Environment :: Web Environment',\n 'Programming Language :: Python :: 2.7',\n 'Programming Language :: Python :: 3.6',\n ],\n keywords=['music', 'netease', 'cli', 'spider'],\n cmdclass={\n 'clean': CleanCommand,\n },\n)\n"},{"alpha_fraction":0.5513054132461548,"alphanum_fraction":0.6836673021316528,"avg_line_length":24.734375,"blob_id":"3ec5d3668e67601c71558775ce639cf292a2b580","content_id":"57c9b08ea340aa270d7ae8179ec0c1c27b83815e","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1647,"license_type":"permissive","max_line_length":270,"num_lines":64,"path":"/NXSpider/spider/common_keys.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/16.\n# email to LipsonChan@yahoo.com\n#\n\nimport json\nimport os\n\nfrom NXSpider.common import encrypt\nfrom NXSpider.common import tools, log\n\ncsrf_dict = {'csrf_token': 'csrf'}\n\n\ndef encrypted_request(obj):\n data = {\n 'params': create_params_by_dict(obj),\n 'encSecKey': encSecKey\n }\n return data\n\n\ndef create_params_by_dict(obj):\n try:\n return create_params_text(json.dumps(obj))\n except Exception as e:\n log.print_err('create params error: %s' % e)\n return None\n\n\ndef create_params_text(text):\n nonce = '0CoJUm6Qyw8W8jud'\n nonce2 = 16 * 'F'\n encText0 = encrypt.aes(text, nonce).decode(\"utf-8\")\n encText = encrypt.aes(encText0, nonce2)\n return encText\n\n\ndef decrpyt_params(text):\n nonce = '0CoJUm6Qyw8W8jud'\n nonce2 = 16 * 'F'\n decText0 = encrypt.aes_decode(text, nonce2)\n decText = encrypt.aes_decode(decText0, nonce)\n return decText\n\n\ndef rsa_encrypt(text, pubKey, modulus):\n text = text[::-1]\n rs = int(tools.hex(text), 16) ** int(pubKey, 16) % int(modulus, 16)\n return format(rs, 'x').zfill(256)\n\n\ndef create_secretKey(size):\n return (\n ''.join(map(lambda xx: (hex(ord(xx))[2:]), os.urandom(size)))\n )[0:16]\n\n\nmodulus = '00e0b509f6259df8642dbc35662901477df22677ec152b5ff68ace615bb7b725152b3ab17a876aea8a5aa76d2e417629ec4ee341f56135fccf695280104e0312ecbda92557c93870114af6c9d05c4f7f0c3685b7a46bee255932575cce10b424d813cfe4875d3e82047b97ddef52741d546b8e289dc6935b3ece0462db0a22b8e7'\npubKey = '010001'\nsecKey = 16 * 'F'\nencSecKey = rsa_encrypt(secKey, pubKey, modulus)\n"},{"alpha_fraction":0.5777778029441833,"alphanum_fraction":0.6083333492279053,"avg_line_length":24.714284896850586,"blob_id":"d038133f9a9151dcfaa23e5a68c544f417367b08","content_id":"7e0b12770a6474b077c6a8b93230dac0978b95e3","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":360,"license_type":"permissive","max_line_length":69,"num_lines":14,"path":"/NXSpider/version.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/6/6.\n# email to LipsonChan@yahoo.com\n#\n\n__title__ = 'NXSpider'\n__version__ = '0.0.5'\n__description__ = 'NetEase X Spider, help u to have a better life!!!'\n__url__ = 'https://github.com/Grass-CLP/NXSpider'\n__author__ = 'LipsonChan'\n__author_email__ = 'LipsonChan@yahoo.com'\n__license__ = 'BSD-3-Clause License'\n"},{"alpha_fraction":0.6242424249649048,"alphanum_fraction":0.6393939256668091,"avg_line_length":19.030303955078125,"blob_id":"9cd1a9375c1b6377c023880130429c51b8828c16","content_id":"252eac29cf6a940c55a0a5f6cdbb531606d0dd7d","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":660,"license_type":"permissive","max_line_length":44,"num_lines":33,"path":"/NXSpider/model/export.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/6/20.\n# email to LipsonChan@yahoo.com\n#\n\nfrom NXSpider.common.config import Config\nif Config().get_no_mongo():\n from NXSpider.model.dict_model import *\nelse:\n from NXSpider.model.mongo_model import *\n\n\nmodel_download_url = 'download_url'\nmodel_is_download = 'downloaded'\nmodel_download_path = 'download_path'\n\n__all__ = [\n 'ConfigModel',\n 'UserModel',\n 'AlbumModel',\n 'PlaylistModel',\n 'Mp4Model',\n 'VideoModel',\n 'ArtistModel',\n 'Mp3Model',\n 'AuthorModel',\n 'update_dynamic_doc',\n 'model_download_url',\n 'model_is_download',\n 'get_one_model_by_key',\n]"},{"alpha_fraction":0.569031298160553,"alphanum_fraction":0.5827612280845642,"avg_line_length":19.484375,"blob_id":"548c7290f0283b25f877b06731245f14bd4f24b6","content_id":"b3a8dea9e934ac6c937dd65fef4949dbad40d883","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1311,"license_type":"permissive","max_line_length":53,"num_lines":64,"path":"/NXSpider/bin/models.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/6/19.\n# email to LipsonChan@yahoo.com\n#\nfrom NXSpider.spider.album import Album\nfrom NXSpider.spider.artist import Artist\nfrom NXSpider.spider.mp3 import Mp3\nfrom NXSpider.spider.mv import MV\nfrom NXSpider.spider.playlist import Playlist\nfrom NXSpider.spider.user import User\n\nno_rec_artist_mo = Artist(\n __parse_recursion__={}\n)\n\nno_rec_album_mo = Album(\n __model_rfilter__={'artist', 'songs'},\n __parse_recursion__={'artists': no_rec_artist_mo}\n)\n\nno_rec_mv_mo = MV(\n __parse_recursion__={'artists': no_rec_artist_mo}\n)\n\ndw_mp3_mo = Mp3(\n __parse_recursion__={\n 'artists': no_rec_artist_mo,\n 'album': no_rec_album_mo,\n 'mv': no_rec_mv_mo,\n }\n)\n\nplaylist_mo = Playlist(\n __model_rfilter__={'artist'},\n __parse_recursion__={\n 'mp3': dw_mp3_mo,\n 'creator': User(),\n }\n)\n\nartist_mo = Artist(\n __model_rfilter__={'artist'},\n __parse_recursion__={\n 'mp3': dw_mp3_mo,\n }\n)\n\nalbum_mo = Album(\n __model_rfilter__={'artist'},\n __parse_recursion__={\n 'mp3': dw_mp3_mo,\n 'artists': no_rec_artist_mo,\n }\n)\n\nartist_album_mo = Artist(\n __model_rfilter__={},\n __parse_recursion__={\n 'albums': album_mo,\n 'artists': no_rec_artist_mo,\n }\n)\n"},{"alpha_fraction":0.535625696182251,"alphanum_fraction":0.5718929171562195,"avg_line_length":25.75286102294922,"blob_id":"76d63045396b147e02afa192bca9e5e0fc41e2b3","content_id":"c809ebc0f4c9702b46590bd142e9bbc7447792e2","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":12299,"license_type":"permissive","max_line_length":141,"num_lines":437,"path":"/NXSpider/spider/api.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/5/24.\n# email to LipsonChan@yahoo.com\n#\nimport re\nfrom collections import OrderedDict\nfrom functools import reduce\n\nimport requests\n\nfrom NXSpider.common import log\nfrom NXSpider.spider.common_keys import encrypted_request\n\nbase_url = \"http://music.163.com\"\nbase_https_url = \"https://music.163.com\"\n\nheaders = {\n 'Referer': 'http://music.163.com/',\n 'Host': 'music.163.com',\n 'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_2) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/55.0.2883.95 Safari/537.36',\n 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8',\n}\n\nTOP_LIST_ALL = {\n 0: ['云音乐新歌榜', '3779629'],\n 1: ['云音乐热歌榜', '3778678'],\n 2: ['网易原创歌曲榜', '2884035'],\n 3: ['云音乐飙升榜', '19723756'],\n 4: ['云音乐电音榜', '10520166'],\n 5: ['UK排行榜周榜', '180106'],\n 6: ['美国Billboard周榜', '60198'],\n 7: ['KTV嗨榜', '21845217'],\n 8: ['iTunes榜', '11641012'],\n 9: ['Hit FM Top榜', '120001'],\n 10: ['日本Oricon周榜', '60131'],\n 11: ['韩国Melon排行榜周榜', '3733003'],\n 12: ['韩国Mnet排行榜周榜', '60255'],\n 13: ['韩国Melon原声周榜', '46772709'],\n 14: ['中国TOP排行榜(港台榜)', '112504'],\n 15: ['中国TOP排行榜(内地榜)', '64016'],\n 16: ['香港电台中文歌曲龙虎榜', '10169002'],\n 17: ['华语金曲榜', '4395559'],\n 18: ['中国嘻哈榜', '1899724'],\n 19: ['法国 NRJ EuroHot 30周榜', '27135204'],\n 20: ['台湾Hito排行榜', '112463'],\n 21: ['Beatport全球电子舞曲榜', '3812895'],\n 22: ['云音乐ACG音乐榜', '71385702'],\n 23: ['云音乐嘻哈榜', '991319590']\n}\n\nPLAYLIST_CLASSES = OrderedDict([\n ('语种', ['华语', '欧美', '日语', '韩语', '粤语', '小语种']),\n ('风格', ['流行', '摇滚', '民谣', '电子', '舞曲', '说唱', '轻音乐', '爵士', '乡村', 'R&B/Soul', '古典', '民族', '英伦', '金属', '朋克', '蓝调', '雷鬼',\n '世界音乐', '拉丁', '另类/独立', 'New Age', '古风', '后摇', 'Bossa Nova']),\n ('场景', ['清晨', '夜晚', '学习', '工作', '午休', '下午茶', '地铁', '驾车', '运动', '旅行', '散步', '酒吧']),\n ('情感', ['怀旧', '清新', '浪漫', '性感', '伤感', '治愈', '放松', '孤独', '感动', '兴奋', '快乐', '安静', '思念']),\n ('主题', ['影视原声', 'ACG', '儿童', '校园', '游戏', '70后', '80后', '90后', '网络歌曲', 'KTV', '经典', '翻唱', '吉他', '钢琴', '器乐', '榜单', '00后'])\n])\n\nALL_CLASSES = reduce(lambda x, y: x + y, [v for k, v in PLAYLIST_CLASSES.items()])\n\nMV_TYPE = ['ALL', 'ZH', 'EA', 'KR', 'JP']\n\n\n# 搜索单曲(1)，歌手(100)，专辑(10)，歌单(1000)，用户(1002) *(type)*\nsearch_types = {\n 'mp3': 1,\n 'artist': 100,\n 'album': 10,\n 'playlist': 1000,\n 'user': 1002,\n 'mv': 1004,\n}\n\n\ndef api_request(url, data=None, method=\"get\", json=True,\n session=None, headers=headers, encrypt=True, https=False):\n \"\"\"\n request and try\n :param https:\n :param encrypt:\n :param url:\n :param data:\n :param method:\n :param json:\n :param session:\n :type session: requests.Session\n :param headers:\n :return:\n \"\"\"\n url = base_https_url + url if https else base_url + url\n request_obj = session or requests\n\n # update cookies\n if isinstance(request_obj, requests.Session):\n for cookie in request_obj.cookies:\n if cookie.name == '__csrf':\n data['csrf_token'] = cookie.value\n break\n\n # encrypt\n if encrypt:\n data = encrypted_request(data)\n\n method = 'get' if not data and method == 'get' else 'post'\n request_method = getattr(request_obj, method, None) or request_obj.get\n try:\n req = request_method(url, data=data, headers=headers, timeout=10)\n req.encoding = \"UTF-8\"\n res = req.json() if json else req.text\n # if session:\n # session.cookies.save()\n return res\n except ValueError as e:\n log.print_err(\"api do not return a valuable json\")\n return {}\n except requests.exceptions.RequestException as e:\n log.print_warn(\"request error: %s\" % url)\n return {}\n\n\ndef get_top_songlist(idx=0, offset=0):\n action = TOP_LIST_ALL[idx][1]\n res = api_request(action, json=False)\n if not res:\n return None\n\n songids = re.findall(r'/song\\?id=(\\d+)', res)\n if not songids:\n return None\n\n songids = list(set(songids))\n details = get_mp3_details(songids, offset=offset)\n return details\n\n\ndef get_mp3_link(song_id):\n # obj = '{\"ids\":[' + str(song_id) + '], br:\"320000\",csrf_token:\"csrf\"}'\n data = {'ids': [song_id], 'br': 320000, 'csrf_token': 'csrf'}\n url = \"/weapi/song/enhance/player/url\"\n res = api_request(url, data)\n if res and res['code'] == 200:\n return res['data'][0]['url']\n\n\ndef get_mp3_links(song_ids):\n # obj = '{\"ids\":[' + str(song_id) + '], br:\"320000\",csrf_token:\"csrf\"}'\n data = {'ids': song_ids, 'br': 320000, 'csrf_token': 'csrf'}\n url = \"/weapi/song/enhance/player/url\"\n res = api_request(url, data)\n if res and res['code'] == 200:\n return {x['id']: x['url'] for x in res['data']}\n\n\ndef get_mp3_details(song_ids, offset=0):\n tmpids = song_ids[offset:]\n tmpids = tmpids[0:100]\n tmpids = list(map(str, tmpids))\n action = '/api/song/detail?ids=[{}]'.format( # NOQA\n ','.join(tmpids))\n\n res = api_request(action)\n if res and res['code'] == 200:\n return {x['id']: x for x in res['songs']}\n\n\ndef get_mv_link(mv_id, r):\n data = {'id': mv_id, 'r': r, 'csrf_token': 'csrf'}\n url = \"/weapi/song/enhance/download/mv/url\"\n\n res = api_request(url, data)\n if res and res['code'] == 200:\n return res['data']['url']\n\n\ndef get_video_link(vid, r):\n data = {'ids': [vid], 'resolution': r, 'csrf_token': ''}\n url = \"/weapi/cloudvideo/playurl\"\n\n res = api_request(url, data)\n if res and res['code'] == 200:\n return res['urls'][0]['url']\n\n\ndef get_video_links(vids, r):\n data = {'ids': vids, 'resolution': r, 'csrf_token': ''}\n url = \"/weapi/cloudvideo/playurl\"\n\n res = api_request(url, data)\n if res and res['code'] == 200:\n return {x['id']: x['url'] for x in res['data']}\n\n\ndef get_mv_detail(mvid):\n data = {'id': mvid, 'csrf_token': 'csrf'}\n url = \"/weapi/v1/mv/detail\"\n\n res = api_request(url, data)\n return res.get('data', None)\n\n\ndef get_video_detail(vid):\n data = {'id': vid, 'csrf_token': 'csrf'}\n url = \"/weapi/cloudvideo/v1/video/detail\"\n\n res = api_request(url, data)\n return res.get('data', None)\n\n\ndef get_playlist_detail(playlist_id):\n url = \"/api/playlist/detail?id={}&upd\" \\\n .format(playlist_id)\n\n res = api_request(url)\n return res.get('result', None)\n\n\ndef get_playlist_detail_v3(id):\n action = '/weapi/v3/playlist/detail'\n data = {'id': id, 'total': 'true', 'csrf_token': 'csrf', 'limit': 1000, 'n': 1000, 'offset': 0}\n\n res = api_request(action, data)\n return res.get('playlist', None)\n\n\ndef get_top_playlists(category='全部', order='hot', offset=0, limit=50):\n \"\"\"\n get playlists, but not detail\n :param category:\n :param order:\n :param offset:\n :param limit:\n :return:\n \"\"\"\n action = u'/api/playlist/list?cat={}&order={}&offset={}&total={}&limit={}'.format( # NOQA\n category, order, offset, 'true' if offset else 'false',\n limit) # NOQA\n\n res = api_request(action)\n return res.get('playlists', None)\n\n\n# may be not useful\ndef get_playlist_classes():\n action = '/weapi/playlist/catalogue'\n res = api_request(action, json=False)\n if res and res['code'] == 200:\n return re\n\n\ndef get_playlist_catelogs():\n path = '/weapi/playlist/catalogue'\n return api_request(path, json=False)\n\n\ndef top_artists(offset=0, limit=100):\n action = '/api/artist/top?offset={}&total=false&limit={}'.format( # NOQA\n offset, limit)\n res = api_request(action)\n return res.get('artists', None)\n\n\ndef get_artists_songs(artist_id):\n action = '/api/artist/{}'.format(artist_id)\n res = api_request(action)\n if res and res['code'] == 200:\n return res\n\n\ndef get_artist_album(artist_id, offset=0, limit=50):\n action = '/api/artist/albums/{}?offset={}&limit={}'.format(\n artist_id, offset, limit)\n res = api_request(action)\n if res and res['code'] == 200:\n return res\n # if res and res['code'] == 200:\n # return res['hotAlbums']\n # return res.get('hotAlbums', None)\n\n\ndef get_album_detail(album_id):\n action = '/api/album/{}'.format(album_id)\n res = api_request(action)\n return res.get('album', None)\n\n\ndef search(s, stype=1, offset=0, total='true', limit=60):\n if isinstance(stype, str):\n if stype in search_types:\n stype = search_types[stype]\n else:\n return None\n\n action = '/api/search/get'\n data = {\n 's': s,\n 'type': stype,\n 'offset': offset,\n 'total': total,\n 'limit': limit\n }\n res = api_request(action, data, encrypt=False)\n return res.get('result', None)\n\n\n# has been change, fuck!\ndef user_playlist(uid, session=None, offset=0, limit=50):\n action = '/weapi/user/playlist'\n data = {'uid': uid, 'csrf_token': 'csrf',\n 'limit': limit, 'offset': offset,\n 'wordwrap': 7}\n\n res = api_request(action, data=data, session=session)\n return res.get('playlist', None)\n\n\ndef user_playlist_old(uid, offset=0, limit=100, session=None):\n action = '/api/user/playlist/?offset={}&limit={}&uid={}'.format( # NOQA\n offset, limit, uid)\n\n res = api_request(action, session=session)\n return res.get('playlist', None)\n\n\ndef my_subcount(session):\n action = '/weapi/subcount'\n res = api_request(action, data={}, session=session)\n\n return res\n\n\ndef my_mvs(session):\n action = '/weapi/mv/sublist'\n data = dict(\n offset=0,\n limit=1000,\n )\n res = api_request(action, data=data, session=session)\n\n return res.get('data', [])\n\n\ndef my_videos(session):\n action = '/weapi/cloudvideo/allvideo/sublist'\n data = dict(\n offset=0,\n limit=1000,\n )\n res = api_request(action, data=data, session=session)\n\n return res.get('data', [])\n\n\ndef hot_mvs():\n # todo, wait to find out how to encrypt and decrypt\n action = '/api/mv/toplist'\n action = '/api/mv/first'\n action = '/api/mv/hot' # well done\n\n data = dict(\n cat=u'内地',\n order='hot',\n offset=0,\n limit=50,\n )\n\n res = api_request(action, data=data)\n return res\n\n\ndef top_mvs(offset=0, limit=50):\n action = '/weapi/mv/toplist'\n data = dict(\n offset=offset,\n limit=limit,\n )\n\n res = api_request(action, data=data)\n return res.get('data', [])\n\n\ndef all_mvs(offset=0, limit=50):\n action = '/weapi/mv/all'\n data = dict(\n offset=offset,\n limit=limit,\n )\n res = api_request(action, data=data)\n return res.get('data', [])\n\n\ndef phone_login(username, password, session):\n \"\"\"\n\n :param username:\n :param password: must be encrypt by md5 hex\n :param session:\n :type session: requests.Session\n :return:\n \"\"\"\n action = '/weapi/login/cellphone'\n data = {\n 'phone': username,\n 'password': password,\n 'rememberLogin': 'true'\n }\n\n res = api_request(action, data=data, session=session)\n if res:\n return res\n\n\ndef login(username, password, session):\n \"\"\"\n :type username: str\n :param username:\n :param password: must be encrypt by md5 hex\n :param session:\n :type session: requests.Session\n :return:\n \"\"\"\n if username.isdigit():\n return phone_login(username, password, session)\n action = '/weapi/login'\n client_token = '1_jVUMqWEPke0/1/Vu56xCmJpo5vP1grjn_SOVVDzOc78w8OKLVZ2JH7IfkjSXqgfmh'\n session.cookies.load()\n data = {\n 'username': username,\n 'password': password,\n 'rememberLogin': 'true',\n 'clientToken': client_token,\n }\n res = api_request(action, data=data, session=session, https=True)\n if res:\n return res\n"},{"alpha_fraction":0.4895104765892029,"alphanum_fraction":0.4979020953178406,"avg_line_length":27.039215087890625,"blob_id":"4653ea5fad0727c600d18d60c33443c5b781692a","content_id":"f065e805a7171a37d66075a72a457f183d3e1cf1","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2860,"license_type":"permissive","max_line_length":94,"num_lines":102,"path":"/NXSpider/spider/mp3.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/23.\n# email to LipsonChan@yahoo.com\n#\n\nimport re\n\nimport os\n\nfrom NXSpider.spider.album import Album\nfrom NXSpider.spider.api import get_mp3_link, get_mv_detail\nfrom NXSpider.spider.artist import Artist\nfrom NXSpider.spider.mv import MV\nfrom NXSpider.spider.base_driver import *\n\n\nclass Mp3(Music163Obj):\n __model_name__ = Mp3Model\n __model_rfilter__ = {\n 'artists', # read from album model\n }\n __parse_recursion__ = {\n 'artists': Artist(),\n 'album': Album(__model_rfilter__={'artist','songs'}),\n 'mv': MV(),\n }\n\n @attr_replace(attr_name='mvid', new_name='mv')\n def replace_mvid(self, obj):\n if obj != 0:\n obj = get_mv_detail(obj)\n return obj\n return None\n\n def pre_save(self, doc, obj):\n \"\"\"\n :param doc:\n :param obj:\n :type doc: Mp3Model\n :return:\n \"\"\"\n # set artists\n if getattr(doc, 'album', False) and getattr(doc['album'], 'artists', False):\n doc.artists = [a for a in doc['album']['artists']]\n\n def download_filename(self, doc):\n \"\"\"\n implement pls\n get a name to save file\n need be complete by child\n :param doc:\n :return:\n \"\"\"\n authors = ','.join([x['name'] for x in doc.artists])\n author = re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', authors.strip())\n mp3_name = re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', doc['name'])\n name = \"%s - %s.mp3\" % (author, mp3_name)\n return name\n\n def download_filename_full(self, doc):\n \"\"\"\n implement pls\n get a path to save file, by relative path\n need be complete by child\n :param doc:\n :type doc: Mp3Model\n :return:\n :rtype: str\n \"\"\"\n # authors = reduce(lambda x, y: x + ',' + y, [x['name'] for x in doc.artists])\n authors = ','.join([x['name'] for x in doc.artists])\n author = re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', authors.strip())\n mp3_name = re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', doc['name'])\n name = os.path.join(author, \"%s - %s.mp3\" % (author, mp3_name))\n return name\n\n def url_load(self, doc):\n \"\"\"\n implement pls\n :param doc:\n :return:\n :rtype: str\n \"\"\"\n try:\n return get_mp3_link(doc['id'])\n except:\n return None\n\n def shortcut_self_path(self, doc):\n \"\"\"\n implement pls, not force\n return self short cut path\n :param doc:\n :return:\n \"\"\"\n result = []\n result.extend([os.path.join(\"artist\", re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', x['name']))\n for x in doc.artists])\n result.append(os.path.join(\"album\", re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', doc.album['name'])))\n return result\n"},{"alpha_fraction":0.7113480567932129,"alphanum_fraction":0.7159177660942078,"avg_line_length":27.54347801208496,"blob_id":"d7c2044b545a4fee68f749e57eceef7a4767b965","content_id":"7da0e8176a8ee2036229790663b64970b35faf22","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"reStructuredText","length_bytes":1313,"license_type":"permissive","max_line_length":78,"num_lines":46,"path":"/CONTRIBUTING.rst","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"============\nContributing\n============\n\nContributions are welcome, and they are greatly appreciated! Every\nlittle bit helps, and credit will always be given.\n\nYou can contribute in many ways:\n\nTypes of Contributions\n----------------------\n\nReport Bugs\n~~~~~~~~~~~\n\nReport bugs at https://github.com/Grass-CLP/NXSpider/issues.\n\nIf you are reporting a bug, please include:\n\n* Your operating system name and version.\n* Any details about your local setup that might be helpful in troubleshooting.\n* Detailed steps to reproduce the bug.\n\nFix Bugs\n~~~~~~~~\n\nLook through the GitHub issues for bugs. Anything tagged with \"bug\"\nis open to whoever wants to implement it.\n\nImplement Features\n~~~~~~~~~~~~~~~~~~\n\nLook through the GitHub issues for features. Anything tagged with\n\"feature\" is open to whoever wants to implement it.\n\nPull Request Guidelines\n-----------------------\n\nBefore you submit a pull request, check that it meets these guidelines:\n\n1. The pull request should include tests.\n2. If the pull request adds functionality, the docs should be updated. Put\n your new functionality into a function with a docstring, and add the\n feature to the list in README.rst.\n3. The pull request should work for Python 2.7, and 3.x, and for PyPy.\n and make sure that the tests pass for all supported Python versions.\n"},{"alpha_fraction":0.5103048086166382,"alphanum_fraction":0.5161103010177612,"avg_line_length":38.25925827026367,"blob_id":"cd2de5da1409d8c29ee08e9b2fba36f01adf56f1","content_id":"e82b810c2886c563597b3771d591134967056261","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":13794,"license_type":"permissive","max_line_length":116,"num_lines":351,"path":"/NXSpider/bin/spider_ctrl.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/6/6.\n# email to LipsonChan@yahoo.com\n#\nimport sys\n\nimport requests\nfrom cement.core.controller import expose\n\nfrom NXSpider.bin.base_ctrl import NXSpiderBaseController, py2_decoding, py2_encoding\nfrom NXSpider.bin.print_as_table import print_playlist, print_albums\nfrom NXSpider.common import log\nfrom NXSpider.common.config import Config\nfrom NXSpider.spider import api\n\n\nclass SpiderController(NXSpiderBaseController):\n class Meta:\n label = \"spider\"\n stacked_on = 'base'\n description = \"NXSpider\"\n arguments = [\n (['-cls', '--cls'],\n dict(help=\"classes or catalogue\")),\n (['-ar', '--artist'],\n dict(help=\"artist\")),\n (['-pl', '--playlist'],\n dict(help=\"playlist\")),\n (['-ur', '--user'],\n dict(help=\"user\")),\n (['-mv', '--mv'],\n dict(help=\"artist\")),\n (['-mp3', '--mp3'],\n dict(help=\"song\")),\n (['-ab', '--album'],\n dict(help=\"album\")),\n (['-dw', '--download'],\n dict(help=\"download files, [], eg. -dw mv,mp3\")),\n (['-offset'],\n dict(help=\"offset index, eg. -offset 50\")),\n (['-limit'],\n dict(help=\"limit size, eg. -limit 50\")),\n (['-lu'],\n dict(help=\"login user or phone number\")),\n (['-lp'],\n dict(help=\"login password\")),\n ]\n\n @expose(help=\"spider mp3, usage: smp3 -mp3 [-dw ]\")\n def smp3s(self):\n from NXSpider.bin.models import dw_mp3_mo\n\n if self.param_check(['mp3'], sys._getframe().f_code.co_name) is False:\n return\n\n download_type = self.parse_download()\n mp3s = self.app.pargs.mp3.split(',') # type: list\n details = api.get_mp3_details(mp3s)\n\n for mid, detail in details.items():\n log.print_info(u\"<{}>\".format(detail['name']))\n dw_mp3_mo.parse_model(detail,\n download_type=download_type,\n file_check=Config().get_file_check(),\n shortcuts_stack=[] if Config().get_shortcut() else None)\n\n log.print_info(\"spider complete!~\")\n pass\n\n @expose(help=\"spider playlist, usage: spls -pl [-dw ] \")\n def spls(self):\n from NXSpider.bin.models import playlist_mo\n\n if self.param_check(['playlist'], sys._getframe().f_code.co_name) is False:\n return\n\n download_type = self.parse_download()\n playlists = self.app.pargs.playlist.split(',') # type: list\n\n for pid in playlists:\n playlist_detail = api.get_playlist_detail(pid)\n if playlist_detail:\n log.print_info(u\"<{}> author：{}\".format(\n playlist_detail['name'],\n playlist_detail['creator']['nickname'],\n ))\n playlist_mo.parse_model(playlist_detail,\n download_type=download_type,\n file_check=Config().get_file_check(),\n shortcuts_stack=[] if Config().get_shortcut() else None)\n\n log.print_info(\"spider complete!~\")\n pass\n\n @expose(\n help=\"spider playlist of a classes, usage: scls-pls -cls \"\n \"[-dw ] [-offset ] [-limit ]\")\n def scls_pls(self):\n from NXSpider.bin.models import playlist_mo\n\n if self.param_check(['cls'], sys._getframe().f_code.co_name) is False:\n return\n\n download_type = self.parse_download()\n class_name = self.app.pargs.cls\n class_name = py2_decoding(class_name)\n\n if class_name != u\"全部\" and py2_encoding(class_name) not in api.ALL_CLASSES:\n log.print_err(\"class name is wrong, pls check by run : nxspider sw-pl-classes\")\n return\n\n playlists = api.get_top_playlists(category=class_name,\n offset=self.app.pargs.offset or 0,\n limit=self.app.pargs.limit or 50) # type: list\n\n log.print_info(\"playlists bellow will be crawled\")\n print_playlist(playlists)\n\n for pl_obj in playlists:\n playlist_detail = api.get_playlist_detail(pl_obj['id'])\n if playlist_detail:\n log.print_info(u\"<{}> author：{}\".format(\n playlist_detail['name'],\n playlist_detail['creator']['nickname'],\n ))\n playlist_mo.parse_model(playlist_detail,\n download_type=download_type,\n file_check=Config().get_file_check(),\n shortcuts_stack=[] if Config().get_shortcut() else None)\n log.print_info(\"spider complete!~\")\n pass\n\n @expose(help=\"spider artist top mp3, usage: sar-top-mp3 -ar [-dw ]\")\n def sar_top_mp3(self):\n from NXSpider.bin.models import artist_mo\n\n if self.param_check(['artist'], sys._getframe().f_code.co_name) is False:\n return\n\n download_type = self.parse_download()\n artists = self.app.pargs.artist.split(',') # type: list\n\n for arid in artists:\n detail = api.get_artists_songs(arid)\n if detail is None:\n continue\n\n artist_detail = detail['artist']\n artist_detail['mp3'] = detail['hotSongs']\n\n log.print_info(u\"<{}>\".format(artist_detail['name']))\n artist_mo.parse_model(artist_detail,\n download_type=download_type,\n file_check=Config().get_file_check(),\n shortcuts_stack=[] if Config().get_shortcut() else None)\n\n log.print_info(\"spider complete!~\")\n pass\n\n @expose(help=\"spider artist albums, usage: sar-albums -ar [-dw ] \"\n \"[-offset ] [-limit ]\")\n def sar_albums(self):\n from NXSpider.bin.models import artist_album_mo\n\n if self.param_check(['artist'], sys._getframe().f_code.co_name) is False:\n return\n\n download_type = self.parse_download()\n artists = self.app.pargs.artist.split(',') # type: list\n\n for arid in artists:\n detail = api.get_artist_album(arid,\n offset=self.app.pargs.offset or 0,\n limit=self.app.pargs.limit or 50)\n if detail is None:\n continue\n\n artist_detail = detail['artist']\n album_details = [api.get_album_detail(d['id']) for d in detail['hotAlbums']]\n album_details = [d for d in album_details if d]\n artist_detail['albums'] = album_details\n\n log.print_info(u\"<{}>\".format(artist_detail['name']))\n log.print_info(\"albums bellow will be crawled\")\n print_albums(artist_detail['albums'])\n\n artist_album_mo.parse_model(artist_detail,\n download_type=download_type,\n file_check=Config().get_file_check(),\n shortcuts_stack=[] if Config().get_shortcut() else None)\n\n log.print_info(\"spider complete!~\")\n pass\n\n @expose(help=\"spider album, usage: sab -ab [-dw ]\")\n def sab(self):\n from NXSpider.bin.models import album_mo\n\n if self.param_check(['album'], sys._getframe().f_code.co_name) is False:\n return\n\n download_type = self.parse_download()\n albums = self.app.pargs.album.split(',') # type: list\n\n for pid in albums:\n album_detail = api.get_album_detail(pid)\n if album_detail is None:\n continue\n\n log.print_info(u\"{} artist：{}\".format(\n \"<\" + album_detail['name'] + \">\",\n album_detail['artist']['name'],\n ))\n album_mo.parse_model(album_detail,\n download_type=download_type,\n file_check=Config().get_file_check(),\n shortcuts_stack=[] if Config().get_shortcut() else None)\n\n log.print_info(\"spider complete!~\")\n pass\n\n @expose(\n help=\"spider a user's playlists, usage: sur-pls -ur \"\n \"[-dw ] [-offset ] [-limit ]\")\n def sur_pls(self):\n from NXSpider.bin.models import playlist_mo\n\n if self.param_check(['user'], sys._getframe().f_code.co_name) is False:\n return\n\n download_type = self.parse_download()\n user_id = self.app.pargs.user\n playlists = api.user_playlist(user_id,\n offset=self.app.pargs.offset or 0,\n limit=self.app.pargs.limit or 50)\n\n log.print_info(\"playlists bellow will be crawled\")\n print_playlist(playlists)\n\n for pl_obj in playlists:\n playlist_detail = api.get_playlist_detail(pl_obj['id'])\n if playlist_detail:\n log.print_info(u\"<{}> author：{}\".format(\n playlist_detail['name'],\n playlist_detail['creator']['nickname'],\n ))\n playlist_mo.parse_model(playlist_detail,\n download_type=download_type,\n file_check=Config().get_file_check(),\n shortcuts_stack=[] if Config().get_shortcut() else None)\n log.print_info(\"spider complete!~\")\n pass\n\n @expose(\n help=\"spider top mvs usage: stop-mv [-offset ] [-limit ]\")\n def stop_mvs(self):\n from NXSpider.bin.models import no_rec_mv_mo\n\n mvs = api.top_mvs(offset=self.app.pargs.offset or 0,\n limit=self.app.pargs.limit or 50)\n mvs = [api.get_mv_detail(d['id']) for d in mvs]\n mvs = [d for d in mvs if d]\n\n for mv in mvs:\n no_rec_mv_mo.parse_model(mv, download_type=['mv'],\n file_check=Config().get_file_check(),\n shortcuts_stack=[] if Config().get_shortcut() else None)\n\n log.print_info(\"spider complete!~\")\n\n pass\n\n @expose(\n help=\"login and spider mv, usage: login-smv -lu [-lp ]\")\n def login_smv(self):\n from NXSpider.bin.models import no_rec_mv_mo\n\n if self.param_check(['lu'], sys._getframe().f_code.co_name) is False:\n return\n\n plaintext_pwd = self.app.pargs.lp or None\n if plaintext_pwd is None:\n import getpass\n plaintext_pwd = getpass.getpass(\"Please input your password:\")\n\n session = requests.session()\n import hashlib\n password = hashlib.md5(plaintext_pwd).hexdigest()\n res = api.login(self.app.pargs.lu, password, session)\n if res.get('code', 0) != 200:\n log.print_err('login failed, msg: {}'.format(res.get('msg', \"none\")))\n exit()\n\n mvs = api.my_mvs(session)\n mvs = [api.get_mv_detail(d['id']) for d in mvs]\n mvs = [d for d in mvs if d]\n\n for mv in mvs:\n no_rec_mv_mo.parse_model(mv, download_type=['mv'],\n file_check=Config().get_file_check(),\n shortcuts_stack=[] if Config().get_shortcut() else None)\n\n log.print_info(\"spider complete!~\")\n\n pass\n\n @expose(\n help=\"login and spider playlists, usage: login-spls -lu [-lp ] [-dw ]\")\n def login_spls(self):\n if self.param_check(['lu', 'lp'], sys._getframe().f_code.co_name) is False:\n return\n\n from NXSpider.bin.models import playlist_mo\n\n plaintext_pwd = self.app.pargs.lp or None\n if plaintext_pwd is None:\n import getpass\n plaintext_pwd = getpass.getpass(\"Please input your password:\")\n\n session = requests.session()\n import hashlib\n password = hashlib.md5(plaintext_pwd).hexdigest()\n res = api.login(self.app.pargs.lu, password, session)\n if res.get('code', 0) != 200:\n log.print_err('login failed, msg: {}'.format(res.get('msg', \"none\")))\n exit()\n\n user_id = res['account']['id']\n download_type = self.parse_download()\n playlists = api.user_playlist(user_id,\n offset=self.app.pargs.offset or 0,\n limit=self.app.pargs.limit or 1000)\n\n log.print_info(\"playlists bellow will be crawled\")\n print_playlist(playlists)\n\n for pl_obj in playlists:\n playlist_detail = api.get_playlist_detail(pl_obj['id'])\n if playlist_detail:\n log.print_info(u\"<{}> author：{}\".format(\n playlist_detail['name'],\n playlist_detail['creator']['nickname'],\n ))\n playlist_mo.parse_model(playlist_detail,\n download_type=download_type,\n file_check=Config().get_file_check(),\n shortcuts_stack=[] if Config().get_shortcut() else None)\n log.print_info(\"spider complete!~\")\n pass\n"},{"alpha_fraction":0.6816326379776001,"alphanum_fraction":0.6938775777816772,"avg_line_length":34.07143020629883,"blob_id":"8ceb505bf02e62f8a8f6989ce6570c4b3d8aa755","content_id":"e07373c047ac915cab80e31caee5c5ec56d82493","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":530,"license_type":"permissive","max_line_length":156,"num_lines":14,"path":"/README_EN.md","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"NXSpider\n=================\n\nNXSpider, NetEase X Spider, A powerful for mp3,mv spider, can help you download mp3 and mv with media tags. Base on python, mongodb, and recursion algorithm\n\n[](LICENSE.md)\n\n\n\n[新手先看这里](SIMPLE_USE.md) | [历史版本在这里](VERSION.md) | [开发详情在这里](DEV.md)\n\n\n==========\ntodo: to be continue"},{"alpha_fraction":0.5470085740089417,"alphanum_fraction":0.6111111044883728,"avg_line_length":14.600000381469727,"blob_id":"4f5921b94f057e383094bdbae9fea67d44788142","content_id":"7bd839cb6197f9a53416f08da81d902af42764b3","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":234,"license_type":"permissive","max_line_length":33,"num_lines":15,"path":"/NXSpider/common/__init__.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/19.\n# email to LipsonChan@yahoo.com\n#\nimport sys\n\nPYTHON3 = False\nPYTHON2 = False\nif sys.version > \"3\":\n PYTHON3 = True\n\nif sys.version[0] == '2':\n PYTHON2 = True\n"},{"alpha_fraction":0.5006303191184998,"alphanum_fraction":0.5146378874778748,"avg_line_length":39.556819915771484,"blob_id":"273e62ee2ef3e909cc0f9e2f8797ad711a6f02f4","content_id":"c4628d2e84e3850e6cb1689090f3fc6e80199fb6","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":7139,"license_type":"permissive","max_line_length":116,"num_lines":176,"path":"/NXSpider/bin/config_ctrl.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/6/5.\n# email to LipsonChan@yahoo.com\n#\nimport json\nimport os\n\nfrom cement.core.controller import CementBaseController, expose\n\nfrom NXSpider.common import log\nfrom NXSpider.common.config import Config, default_path_key, mv_resolutions\nfrom NXSpider.common.constant import default_download_dir\n\n\nclass ConfigController(CementBaseController):\n class Meta:\n label = \"config\"\n stacked_on = 'base'\n description = \"NXSpider config\"\n arguments = [\n (['-mh', '--mhost'],\n dict(help=\"mongo host\")),\n (['-mp', '--mport'],\n dict(help=\"mongo port\")),\n (['-mu', '--muser'],\n dict(help=\"mongo user\")),\n (['-mpw', '--mpassword'],\n dict(help=\"mongo password\")),\n (['-mn', '--mdbname'],\n dict(help=\"mongo db name, default nxspider\")),\n (['-nomongo', '--nomongo'],\n dict(help=\"no mongo mode, true(1) or false(0), default True, while true, mongodb will be never used\")),\n\n (['-path', '--path_download'],\n dict(help=\"download path, default ~/.nxspider/download_files/, eg. -path defualt;path1;path2\")),\n (['-mvr', '--mv_resolution'],\n dict(help=\"mv default resolution, [240, 480, 720, 1080] default 720\")),\n (['-stt', '--shortcut'],\n dict(help=\"build shortcuts, true(1) or false(0), default False\")),\n (['-tag', '--media_tag'],\n dict(help=\"media tag, true(1) or false(0), default True\")),\n (['-tag163', '--media_tag_163'],\n dict(help=\"media tag of 163 comment, true(1) or false(0), default True\")),\n (['-dfc', '--download_file_check'],\n dict(help=\"check file exist in paths set, true(1) or false(0), default True\")),\n ]\n\n @expose(help=\"config mongo server with [-h --host] [-p --port] \"\n \"[-u --user] [-pw --password] [-c --collections]\")\n def config_mongo(self):\n config = Config()\n config_dict = config.config # type: dict\n mongo_key = 'mongo'\n is_config = False\n try:\n if self.app.pargs.mhost is not None:\n config_dict[mongo_key]['host'] = self.app.pargs.mhost\n config_dict['no_mongo'] = False\n is_config = True\n\n if self.app.pargs.mport is not None:\n config_dict[mongo_key]['port'] = int(self.app.pargs.mport)\n is_config = True\n\n if self.app.pargs.muser is not None:\n config_dict[mongo_key]['username'] = self.app.pargs.muser\n is_config = True\n\n if self.app.pargs.mpassword is not None:\n config_dict[mongo_key]['password'] = self.app.pargs.mpassword\n is_config = True\n\n if self.app.pargs.mdbname is not None:\n config_dict[mongo_key]['name'] = self.app.pargs.mdbname\n is_config = True\n\n if self.app.pargs.nomongo is not None:\n config_dict['no_mongo'] = True if self.app.pargs.nomongo.lower() == 'true'\\\n or self.app.pargs.nomongo == '1' else False\n is_config = True\n except:\n log.print_err(\"input error, pls check\")\n raise\n if is_config:\n config.save_config_file()\n log.print_info(\"config success\")\n self.config_show()\n\n @expose(help=\"config spider behavior with [-path ]\"\n \"[-mvr <240,480,720,1080>] [-tag <1 or 0>] [-tag163 <1 or 0>]\"\n \"[-dfc <1 or 0>] [-stt <1 or 0>]\")\n def config_spider(self):\n config = Config()\n config_dict = config.config # type: dict\n is_config = False\n try:\n if self.app.pargs.path_download is not None:\n paths = self.app.pargs.path_download.split(',') # type: list\n if default_path_key in paths:\n index = paths.index(default_path_key)\n paths.remove(default_path_key)\n paths.insert(index, default_download_dir)\n\n final_paths = []\n for p in paths:\n try:\n # some error need pass\n if os.path.isdir(p) is False:\n os.mkdir(p)\n final_paths.append(p)\n except:\n log.print_warn(\"path may be wrong and be deleted: {}\".format(p))\n pass\n\n if not final_paths:\n final_paths.append(default_download_dir)\n\n log.print_info('path will be set as: ' + ','.join(final_paths))\n\n config_dict['download_path'] = final_paths\n is_config = True\n\n if self.app.pargs.mv_resolution is not None:\n r = int(self.app.pargs.mv_resolution)\n if r not in mv_resolutions:\n log.print_warn(\"-mvr resolution config skip, value must be 240,480,720,1080\")\n config_dict['mv_def_resolution'] = r\n is_config = True\n\n if self.app.pargs.media_tag is not None:\n config_dict['media_tag'] = True if self.app.pargs.media_tag.lower() == 'true'\\\n or self.app.pargs.media_tag == '1' else False\n is_config = True\n\n if self.app.pargs.media_tag_163 is not None:\n config_dict['media_tag_163'] = True if self.app.pargs.media_tag_163.lower() == 'true' \\\n or self.app.pargs.media_tag_163 == '1' else False\n is_config = True\n\n if self.app.pargs.shortcut is not None:\n config_dict['shortcut'] = True if self.app.pargs.shortcut.lower() == 'true' \\\n or self.app.pargs.shortcut == '1' else False\n is_config = True\n\n\n except:\n log.print_err(\"input error, pls check\")\n raise\n if is_config:\n config.save_config_file()\n log.print_info(\"config success\")\n self.config_show()\n\n @expose(help=\"check config is valid or not\")\n def config_check(self):\n self.config_show()\n try:\n config = Config()\n if config.config_test():\n log.print_info('config check complete, all is well done!')\n except:\n log.print_err('config check failed, pls re config')\n\n @expose(help=\"clear all config, u need re-config from beginning\")\n def config_clear(self):\n Config().config_reset()\n log.print_info(\"config has been reset, u need re-config from beginning pls\")\n self.config_show()\n\n @expose(help=\"list all config opt\")\n def config_show(self):\n config_dict = Config().config\n log.print_info(\"config will be show fellow:\")\n print(json.dumps(config_dict, ensure_ascii=False, indent=1))\n\n"},{"alpha_fraction":0.7139001488685608,"alphanum_fraction":0.7363922595977783,"avg_line_length":27.87013053894043,"blob_id":"d430fcf244be5988a571684c01a0c2ac3a34bda3","content_id":"71157e5eb0ea60b8df291ecb424356375324a565","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":3963,"license_type":"permissive","max_line_length":109,"num_lines":77,"path":"/DEV.md","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"NXSpider\n=================\n\nNXSpider， NetEase X Spider，一个强大的（网易云音乐）mp3,mv爬虫，可以下载和收集mp3,mv信息,同时附带多媒体标签信息。采用python编写，mongo数据库(非必须)，递归算法核心实现\n\n[](LICENSE.md)\n\n\n\n[新手先看这里](SIMPLE_USE.md) | [历史版本在这里](VERSION.md) | [开发详情在这里](DEV.md)\n\n## 开发及问题(不关心的可以不看了)\n\n### 开发调试\n1. 已安装则采用(下文采用该方式进行说明)：\n```\n$nxspider ...\n```\n2. 未安装+win采用(未安装指没执行setup.py,只用代码运行)：\n```\nbash_python(3) NXSipder/bin/cli.py ...\n```\n3. 未安装+linux采用：\n```\nPYTHONPATH=. python(3) NXSipder/bin/cli.py ...\n```\n4. 注意,win下采用bash_python时，分割符 ```,``` 必须改为 ```:``` eg.\n```\n$bash_python.bat NXSpider/bin/cli.py config-spider -path d:\\netease_dw:default\n```\n\n### 配置说明\n1. ~~config-mongo必须运行爬取前配置!!!，dbname可不指定，默认为nxspider~~\n2. 默认无mongodb模式，需要采集数据则通过配置第一条即可 `nxspider config-mongo ..`\n3. 一旦配置了 `mongodb -mh` 则 `-nomogo` 会自动设置为0，即配置了host就会使用的意思\n2. path **强烈建议**爬取前配置，```,``` 为多个下载路径分隔符。default指`~/.nxspider/download_files/`\n3. 其他配置可选，请查看 `nxspider -h`\n4. ```-tag 1``` 建议保留 `-tag163 1` 根据需求保留\n6. 配置完建议执行 ```nxspider config-check``` 检查配置正确性\n\n### 注意\n- 本项目纯粹是学习开发使用，欢迎大家互相讨论，下载的资料请24小时内删除\n- 涉及侵权以及版权问题欢迎讨论和提出\n\n### 协助开发或2次开发建议\n1. 希望尽可能(yahoo邮件)跟作者(LipsonChan)联系，以及对项目进行加★\n2. 核心代码为NXSpider/bin以及NXSpider/spider/base_driver.py\n3. 主要逻辑为通过api获得对于json数据，采用递归+配置方式，自动下载可下载对象\n4. 如果有任何反馈，希望回复到项目的issue,注明版本号，运行环境及描述清楚问题\n\n### windows + python2问题\n1. 参数输入(目前只有查询)非中文和latin，可能会出现问题\n2. 不推荐配置路径中有非latin字符，及路径最好是英文，没测试过\n3. 查询结果输出无法显示非中文和latin数据，如韩文（日文可以）\n\n### 其他问题\n- 同歌手同名MP3不会被重复下载\n\n### 开发历程\n- 基于spider163项目(不满足且有小bug)，开发mongodb以及可以更多爬取项目\n- 增加更多url，MV下载，尝试使用eapi发现很难发现加密规则\n- 修改项目结构为下载驱动driver和model双层，配合metaclass进行递归式下载，减少后续增量开发工作\n- 为了配合Netease app和桌面软件，修改mp3 mp4 tag信息\n- 艰难的使用了eyed3，windows超级麻烦(现成的dll还要配置，还有字符集问题)，发现不支持mp4，非常难过\n- 发现mutagen，果断弃坑eyed3\n- 发布0.0.1版本，可以通过playlist，ablum，aritst_top_mp3爬取和下载mp3,mv\n- 发现mongodb可能有人不喜欢，新增无数据库版本（默认无需使用），通过配置可切换\n- 新增通过获取最火playlist进行爬取\n- 新增了好多功能，读 [VERSION.md](VERSION.md)\n- api全靠`fiddler`和猜，**一把辛酸一把泪**\n\n### 下阶段开发\n- 通过用户id，爬取该用户所有的歌单 √\n- 通过歌手id，爬取该歌手所有专辑 √\n- 通过排行版，爬取最新n个歌单 √\n- mongodb 可选？不强制（但对于离线则无法加tag，虽然只是个人用） √\n- 通过登录(或非登录)，爬取用户收藏的歌单，mv **不安全(明文账号密码)，而且经常登录api会被限，不开放说明**\n"},{"alpha_fraction":0.4928571283817291,"alphanum_fraction":0.6714285612106323,"avg_line_length":16.625,"blob_id":"997278c5c556dd28c599e926e1520b11ce90cd0a","content_id":"3bc1f70e3849ded624faa36b73cc7843fa56d66c","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Text","length_bytes":140,"license_type":"permissive","max_line_length":20,"num_lines":8,"path":"/requirements.txt","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"six>=1,<=2\ncement>=2,<3\nmutagen>=1.40,<2\nterminaltables>=3,<4\ncryptography>=2.2,<3\nmongoengine>=0.15,<1\nrequests>=2.19,<3\ncolorama>=0.3,<0.4"},{"alpha_fraction":0.5651697516441345,"alphanum_fraction":0.582694411277771,"avg_line_length":23.026315689086914,"blob_id":"bc07e9820d1fa550e673fd1209b4b8a50a92bd9b","content_id":"c488966393e41549013cefe0feb20cfd4156a44c","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":913,"license_type":"permissive","max_line_length":76,"num_lines":38,"path":"/NXSpider/utility/shortcut.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/12/15.\n# email to LipsonChan@yahoo.com\n#\n\nimport os\n\nif os.name == \"nt\":\n def symlink_ms(source, link_name):\n import ctypes\n csl = ctypes.windll.kernel32.CreateSymbolicLinkW\n csl.argtypes = (ctypes.c_wchar_p, ctypes.c_wchar_p, ctypes.c_uint32)\n csl.restype = ctypes.c_ubyte\n flags = 1 if os.path.isdir(source) else 0\n try:\n if csl(link_name, source.replace('/', '\\\\'), flags) == 0:\n raise ctypes.WinError()\n except Exception as e:\n pass\n\n\n os.symlink = symlink_ms\n\n\ndef symlink(source, link_name):\n try:\n os.symlink(source, link_name)\n except Exception as e:\n pass\n\n\n# a = os.getcwd()\n# src = os.path.join(a, \"install.bat\")\n# dst = os.path.join(a, \"test.link\")\n#\n# os.symlink(os.path.join(a, \"install.bat\"), os.path.join(a, \"test.link\"))\n"},{"alpha_fraction":0.5379061102867126,"alphanum_fraction":0.5848375558853149,"avg_line_length":16.3125,"blob_id":"7a82dab413fa9f73c3b00dd08bb34ca098db3103","content_id":"7bf4bb6ce6ae6676b958ec059cce89eaffc3c901","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":277,"license_type":"permissive","max_line_length":41,"num_lines":16,"path":"/NXSpider/spider/artist.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/26.\n# email to LipsonChan@yahoo.com\n#\n\nfrom NXSpider.spider.base_driver import *\n\n\nclass Artist(Music163Obj):\n __model_name__ = ArtistModel\n __model_rfilter__ = {\n 'img1v1Id',\n 'picId',\n }\n"},{"alpha_fraction":0.5031654238700867,"alphanum_fraction":0.5240848064422607,"avg_line_length":26.3157901763916,"blob_id":"43765f2a01ae42e3f33332441c5a6e27d61b515c","content_id":"aba075c3a131165a30a7e42fc5f0a9391c4adf88","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3633,"license_type":"permissive","max_line_length":90,"num_lines":133,"path":"/NXSpider/utility/media_tag.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/5/17.\n# email to LipsonChan@yahoo.com\n#\nimport os\n\nfrom mutagen import id3\nfrom mutagen.id3 import ID3\nfrom mutagen.mp4 import MP4\n\nfrom NXSpider.common.config import Config\nfrom NXSpider.common.encrypt import aes_ecb\nfrom NXSpider.model.export import Mp3Model, Mp4Model\n\nkey_names = ['title', 'artist', 'album', 'album_artist', 'track_num', 'comment']\nencoding_keys = ['title', 'artist', 'album', 'album_artist', 'comment']\nmutagen_idv3_key_map = {\n 'title': 'TIT2',\n 'artist': 'TPE1',\n 'album': 'TALB',\n 'album_artist': 'TPE2',\n 'track_num': 'TRCK',\n 'comment': 'COMM'\n}\n\nmugagen_mp4_key_map = {\n 'title': '\\xa9nam',\n 'artist': '\\xa9ART',\n 'comment': '\\xa9cmt',\n}\naes_code = \"#14ljk_!\\]&0U<'(\"\n\n\n# @tools.ignored(Exception)\ndef attach_mp3_idv3(doc, file):\n \"\"\"\n :type doc: Mp3Model\n :param doc:\n :param file:\n :return:\n \"\"\"\n artists = [('[\"%s\",%d]' % (x['name'], x.id)) for x in doc.artists]\n artists_str = \",\".join(artists)\n\n authors = \",\".join([x['name'] for x in doc.artists])\n data = {\n 'title': doc['name'],\n 'artist': authors,\n 'album': doc['album']['name'],\n 'album_artist': authors,\n 'track_num': str(doc['no']),\n }\n\n if Config().get_media_tag_163():\n comment_plaintext = u'music:{\"musicId\":%d,\"musicName\":\"%s\",\"bitrate\":320000,' \\\n u'\"albumId\":%d,\"album\":\"%s\", \"artist\":[%s]}' \\\n % (doc.id, doc['name'], doc.album.id, doc.album['name'],\n artists_str)\n comment = \"163 key(Don't modify):\" + aes_ecb(comment_plaintext, aes_code).decode()\n data['comment'] = comment\n\n try:\n mp3 = ID3(file, v2_version=3)\n for k, v in data.items():\n if k not in mutagen_idv3_key_map:\n continue\n\n if k == 'comment':\n mp3.add(id3.COMM(lang='XXX', text=v))\n continue\n\n attr_type = getattr(id3, mutagen_idv3_key_map[k], None)\n if attr_type:\n mp3.add(attr_type(text=v))\n\n mp3.save(v2_version=3)\n except Exception as e:\n return False\n return True\n\n\n# @tools.ignored(Exception)\ndef attach_mp4_tag(doc, file):\n \"\"\"\n :type doc: Mp4Model\n :param doc:\n :param file:\n :return:\n \"\"\"\n authors = u\",\".join([x['name'] for x in doc.artists])\n data = {\n 'title': doc['name'],\n 'artist': authors,\n }\n\n if Config().get_media_tag_163():\n comment_plaintext = u'mv:{\"title\":\"%s\",\"mvId\":%d,\"artistId\":%d,' \\\n u'\"artistName\":\"%s\",\"pubTime\":\"%s\",\"bitrate\":%d}' \\\n % (doc['name'], doc.id, doc.artists[0].id,\n doc.artists[0]['name'], doc['publishTime'],\n doc['download_video_r'])\n comment = \"163 key(Don't modify):\" + aes_ecb(comment_plaintext, aes_code).decode()\n data['comment'] = comment\n\n try:\n mp4 = MP4(file)\n for k, v in data.items():\n if k not in mugagen_mp4_key_map:\n continue\n mp4[mugagen_mp4_key_map[k]] = v\n\n mp4.save()\n except Exception as e:\n return False\n return True\n\n\ndef attach_media_tag(doc, file):\n \"\"\"\n attach a media idv3\n :param doc:\n :param file:\n :return:\n \"\"\"\n if not os.path.exists(file):\n return False\n\n if isinstance(doc, Mp3Model):\n return attach_mp3_idv3(doc, file)\n elif isinstance(doc, Mp4Model):\n return attach_mp4_tag(doc, file)\n"},{"alpha_fraction":0.5641250610351562,"alphanum_fraction":0.6272276043891907,"avg_line_length":30.69444465637207,"blob_id":"e804681beff3131412ea21c32748de816d189470","content_id":"31c87e97a223c09fe5a8e2246260cb3c186e3234","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3445,"license_type":"permissive","max_line_length":115,"num_lines":108,"path":"/NXSpider/debug_test/crawl.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/5/11.\n# email to LipsonChan@yahoo.com\n#\nimport hashlib\n\nimport requests\n\nfrom NXSpider.common import tools, log\nfrom NXSpider.common.constant import all_download_type\nfrom NXSpider.spider.api import get_playlist_detail, get_mp3_links, get_album_detail, get_top_playlists, \\\n get_playlist_classes, \\\n top_artists, get_artists_songs, get_artist_album, get_mp3_details, search, phone_login, login, user_playlist, \\\n get_playlist_detail_v3, get_playlist_catelogs, hot_mvs, all_mvs, top_mvs\nfrom NXSpider.spider.playlist import Playlist\n\n# def playlist_by_id(link, download_type=all_download_type,\n# save=True, file_check=True):\n# playlist_detail = get_playlist_detail(link)\n# with tools.ignored(Exception):\n# log.print_info(\"%s author：%s\" % (\n# \"<\" + playlist_detail['name'] + \">\",\n# playlist_detail['creator']['nickname'],\n# ))\n#\n# playlist_mo = Playlist()\n# playlist_mo.parse_model(playlist_detail,\n# save=save, download_type=download_type,\n# file_check=file_check)\n# pass\n\n\n# def playlist_by_page(page, type=u\"全部\"):\n# play_url = constant.play_url.format(type, page * 35)\n# titles = []\n# try:\n# log.print_info(\"begin crawl playlist page: {}, type: {}\".format(page, type))\n# acmsk = {'class': 'msk'}\n# scnb = {'class': 'nb'}\n# dcu = {'class': 'u-cover u-cover-1'}\n# ucm = {'class': 'm-cvrlst f-cb'}\n# data = tools.curl(play_url, constant.header, type=constant.RETURE_HTML)\n# lst = data.find('ul', ucm)\n# for play in lst.find_all('div', dcu):\n# title = tools.encode(play.find('a', acmsk)['title'])\n# link = tools.encode(play.find('a', acmsk)['href']).replace(\"/playlist?id=\", \"\")\n# # playlist_by_id(link)\n# print(\"link: {}, title: {}\".format(link, title))\n# except Exception as e:\n# log.print_err(\"playlist page error：{} type：{} page：{}\".format(e, type, page))\n# log.print_err(e)\n\n\n# def playlist_all_page():\n# cf = \"全部\"\n# for i in range(36):\n# playlist_by_page(i + 1, cf)\n\n# test = get_mv_details([5322493,239037])\ntest = top_mvs()\ntest = all_mvs()\ntest = hot_mvs()\ntest = get_playlist_catelogs()\ntest = get_playlist_classes()\ntest = get_top_playlists()\ntest = user_playlist(48872048)\ntest = get_playlist_detail(92024088)\ntest = get_playlist_detail(2246057871)\ntest = get_playlist_detail(107020750)\ntest = get_playlist_detail_v3(107020750)\n\np = hashlib.md5('000'.encode('utf-8')).hexdigest()\ns = requests.Session()\n\ntest = user_playlist(92024088, session=s)\ntest = get_playlist_detail(107020750)\n\n\ntest = search('周杰伦', stype=100)\ntest = top_artists()\ntest = get_playlist_classes()\ntest = get_mp3_details([412902496,412902496,412902496])\n\ntest = get_artists_songs(9621)\ntest = get_artist_album(9621)\n\n\ntest = top_artists()\n# test = playlist_classes()\ntest = get_top_playlists()\n\ntest = get_album_detail(32324)\n\n# test = get_mp3_link(412902496)\n# playlist_by_id(466225104)\ntests = get_mp3_links([412902496, 326904])\n# http://music.163.com/api/song/detail?ids=[412902496,326904] work\n# http://music.163.com/api/album/32324 don't work\n# http://music.163.com/#/artist?id=9621\n# http://music.163.com/api/artist/9621 don't work\n# playlist_all_page()\n\n\n\n\nprint('a')\n"},{"alpha_fraction":0.5634615421295166,"alphanum_fraction":0.5788461565971375,"avg_line_length":23.186046600341797,"blob_id":"e418f50b90ec3ae27156087529e48aa0ed53a751","content_id":"4cdbcafe5489408d8cfb6b9d805c0596cfa7b2ba","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1040,"license_type":"permissive","max_line_length":60,"num_lines":43,"path":"/NXSpider/spider/playlist.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/20.\n# email to LipsonChan@yahoo.com\n#\nimport os\n\nfrom NXSpider.spider.mp3 import Mp3\nfrom NXSpider.spider.base_driver import *\nfrom NXSpider.spider.user import User\n\n\nclass Playlist(Music163Obj):\n __model_name__ = PlaylistModel\n __model_rfilter__ = {\n # garbage properties\n 'coverImgId_str',\n 'coverImgId_str',\n }\n __parse_recursion__ = {\n 'mp3': Mp3(),\n 'creator': User(),\n }\n\n @attr_replace(attr_name='tracks', new_name='mp3')\n def replace_tracks(self, obj):\n return obj\n\n @attr_replace(attr_name='creator', new_name='creator')\n def replace_creator(self, obj):\n if 'userId' in obj and 'id' not in obj:\n obj['id'] = obj['userId']\n return obj\n\n def shortcut_relative_name(self, doc):\n \"\"\"\n implement pls, not force\n return short cut relative path\n :param doc:\n :return:\n \"\"\"\n return os.path.join(self.__file_type__, doc['name'])\n"},{"alpha_fraction":0.5814639925956726,"alphanum_fraction":0.5897284746170044,"avg_line_length":27.711864471435547,"blob_id":"35db5a3db506d34cba4fb78ae0fd15ce29c7f519","content_id":"f07c1f1fd00e7607fa1c5609301c8cccff46621f","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1694,"license_type":"permissive","max_line_length":89,"num_lines":59,"path":"/NXSpider/bin/base_ctrl.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/6/20.\n# email to LipsonChan@yahoo.com\n#\nimport sys\n\nfrom cement.core.controller import CementBaseController\n\nfrom NXSpider.common import log, PYTHON2\nfrom NXSpider.common.constant import all_download_type\n\n\ndef py2_decoding(string):\n if PYTHON2:\n return string.decode(sys.getfilesystemencoding())\n return string\n\n\ndef py2_encoding(string):\n if PYTHON2:\n return string.encode('utf8')\n\n\nclass NXSpiderBaseController(CementBaseController):\n def parse_download(self):\n \"\"\"\n lost of spider function will parse -dw param, this will do it\n :return:\n \"\"\"\n if self.app.pargs.download is None:\n download_type = []\n else:\n download_type = self.app.pargs.download.split(',') # type: list\n download_type = list(filter(lambda x: x in all_download_type, download_type))\n return download_type\n\n def param_check(self, params, func_name):\n \"\"\"\n this will check param inputted and require is complete or not, and print help\n help will be in expose(help='...'), and got by function name\n :param params:\n :param func_name:\n :return:\n \"\"\"\n help = None\n fun = getattr(self, func_name, None)\n if fun and getattr(fun, '__cement_meta__', None):\n help = fun.__cement_meta__['help']\n\n for p in params:\n param = getattr(self.app.pargs, p, None)\n if param is None:\n log.print_err(\"param {} miss, see help:\".format(p))\n if help:\n print(help)\n return False\n return True\n"},{"alpha_fraction":0.5121457576751709,"alphanum_fraction":0.5206477642059326,"avg_line_length":31.5,"blob_id":"fb53946dc8c617a632c9e4d281f7d5375a9ddd72","content_id":"67e06b7e8ccd3518a1b90b290589bf4b3114fcc6","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2470,"license_type":"permissive","max_line_length":87,"num_lines":76,"path":"/NXSpider/bin/print_as_table.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/6/23.\n# email to LipsonChan@yahoo.com\n#\nfrom terminaltables import AsciiTable\n\n\ndef print_playlist(pls):\n table = AsciiTable([[\"ID\", \"Name\", \"User\", \"PlayCount\"]])\n table_data = [[str(item['id']), item['name'],\n item['creator']['nickname'],\n str(item['playCount']),\n ] for item in pls]\n table.table_data.extend(table_data)\n print(table.table)\n pass\n\n\ndef print_albums(abs):\n from terminaltables import AsciiTable\n table = AsciiTable([[\"ID\", \"Album\", \"Artist\", \"ArtistID\"]])\n table_data = [[str(item['id']), item['name'],\n ','.join([ar['name'] for ar in item['artists']]),\n ','.join([str(ar['id']) for ar in item['artists']]),\n ] for item in abs]\n table.table_data.extend(table_data)\n print(table.table)\n pass\n\n\ndef print_mp3s(mp3s):\n table = AsciiTable([[\"ID\", \"Name\", \"Album\", \"AlbumID\", \"Artist\", \"ArtistID\"]])\n table_data = [[str(item['id']), item['name'],\n item['album']['name'], item['album']['id'],\n ','.join([ar['name'] for ar in item['artists']]),\n ','.join([str(ar['id']) for ar in item['artists']]),\n ] for item in mp3s]\n table.table_data.extend(table_data)\n print(table.table)\n pass\n\n\ndef print_users(users):\n table = AsciiTable([[\"ID\", \"Name\", \"Signature\"]])\n table_data = [[str(item['userId']), item['nickname'],\n item['signature'],\n ] for item in users]\n table.table_data.extend(table_data)\n print(table.table)\n pass\n\n\ndef print_artists(artists):\n table = AsciiTable([[\"ID\", \"Name\", \"AlbumNum\", \"MVNum\"]])\n table_data = [[str(item['id']), item['name'],\n str(item['albumSize']),\n str(item['mvSize'])\n ] for item in artists]\n table.table_data.extend(table_data)\n print(table.table)\n pass\n\n\ndef print_mvs(mvs):\n table = AsciiTable([[\"ID\", \"Name\", \"Artist\", \"ArtistID\", \"Duration\", \"PlayCount\"]])\n table_data = [[str(item['id']), item['name'],\n item['artistName'],\n item['artistId'],\n '%02d:%02d' % divmod(int(item['duration'] / 1000), 60),\n item['playCount'],\n ] for item in mvs]\n table.table_data.extend(table_data)\n print(table.table)\n pass\n"},{"alpha_fraction":0.8409090638160706,"alphanum_fraction":0.8409090638160706,"avg_line_length":44,"blob_id":"f18016d1d2f1a3befd4f983276bfe2c29e51ccdf","content_id":"e33b80f3decd4f6314ca782a57810ba4599a45f1","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":44,"license_type":"permissive","max_line_length":44,"num_lines":1,"path":"/TODO.md","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"readme modify(add shortcut and add password)"},{"alpha_fraction":0.6499999761581421,"alphanum_fraction":0.6558333039283752,"avg_line_length":21.641510009765625,"blob_id":"c28537a0d84c9a8c488b4bd9ae7b490d5b88fd61","content_id":"2e89f6728ad634a408cb5dd6f10a37e24d1533bf","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1200,"license_type":"permissive","max_line_length":74,"num_lines":53,"path":"/NXSpider/bin/cli.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/6/5.\n# email to LipsonChan@yahoo.com\n#\nimport datetime\n\nfrom cement.core.controller import CementBaseController, expose\nfrom cement.core.foundation import CementApp\n\nfrom NXSpider.bin.config_ctrl import ConfigController\nfrom NXSpider import version\nfrom colorama import init\n\nfrom NXSpider.bin.show_ctrl import ShowController\nfrom NXSpider.bin.spider_ctrl import SpiderController\nfrom NXSpider.common.config import Config\n\n\nBANNER = \"\"\"\n{} Application v{}\nCopyright (c) {} {}\n\"\"\".format(version.__title__, version.__version__ ,\n version.__author__, datetime.datetime.now().year)\n\ninit(autoreset=True)\n\n\nclass VersionController(CementBaseController):\n class Meta:\n label = 'base'\n description = ''\n arguments = [\n (['-v', '--version'], dict(action='version', version=BANNER)),\n ]\n\n\nclass App(CementApp):\n class Meta:\n label = \"NXSpider\"\n base_controller = \"base\"\n handlers = [VersionController, ConfigController,\n SpiderController, ShowController]\n\n\ndef main():\n with App() as app:\n app.run()\n\n\nif __name__ == \"__main__\":\n main()\n"},{"alpha_fraction":0.6171428561210632,"alphanum_fraction":0.6628571152687073,"avg_line_length":18.44444465637207,"blob_id":"f46c816fb50f98028b417b1ab5d853cf55857d7b","content_id":"fd4f494187a4ef59c4370e5e325ca2af10a99728","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":175,"license_type":"permissive","max_line_length":41,"num_lines":9,"path":"/NXSpider/debug_test/config.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/20.\n# email to LipsonChan@yahoo.com\n#\nfrom NXSpider.common.config import Config\n\npaths = Config().get_paths()\n"},{"alpha_fraction":0.6215596199035645,"alphanum_fraction":0.6567278504371643,"avg_line_length":26.82978630065918,"blob_id":"35e02726e2ce107392c9a82d8b6571f88488883d","content_id":"37b99b1acbb50bb205229dbc611993566182e44a","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1388,"license_type":"permissive","max_line_length":99,"num_lines":47,"path":"/NXSpider/debug_test/idv3.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/5/16.\n# email to LipsonChan@yahoo.com\n#\nimport locale\n\n# import eyed3\nfrom mutagen.mp4 import MP4\n\nfrom NXSpider.utility.idv3_base import delete_163comment\nfrom NXSpider.utility.modify_idv3 import attach_media_tag_by_path, attach_media_tag_by_conf_path\n\nfrom mutagen.id3 import ID3, TALB, COMM\nfrom mutagen.easyid3 import EasyID3\nTALB\nCOMM\nMP4\n# ['TIT2', 'TPE1', 'TRCK', 'TALB', 'TPOS', 'TSSE', 'APIC:', 'COMM::XXX']\n# 名称 歌手 trackNum 专辑 pos 编码 comments\n# 'TPE2' 'aART' album_artist\n\nattach_media_tag_by_path(u'E:\\\\test')\nattach_media_tag_by_conf_path()\n\n# a = ID3(u'E:\\\\薛之谦 - 丑八怪.mp3', v2_version=3)\n#\n# a = MP4(u'E:\\\\(G)I-DLE - LATATA.mp4')\n#\n# a = eyed3.load(u'E:\\\\薛之谦 - 丑八怪.mp3')\n# a.tag.comments.remove('')\n# a = eyed3.load(u'E:\\\\(G)I-DLE - LATATA.mp3')\n# a = eyed3.load(u'E:\\\\(G)I-DLE - LATATA.mp4')\n# a.tag.title(\"test\")\n# a.tag.comments.remove('')\n\nlocale.getlocale(locale.LC_ALL)\nfile = u\"E:/新建文件夹/Billion - Dancing Alone_wapi - 副本.mp3\"\nfile = u\"D:\\\\Project\\\\python\\\\music.crawl\\\\download_files\\\\mp3\\\\1호선 뮤직\\\\1호선 뮤직 - 열대야.mp3\"\nfile = u\"D:\\\\Project\\\\python\\\\music.crawl\\\\download_files\\\\mp3\\\\Dal★shabet\\\\Dal★shabet - JOKER.mp3\"\ndelete_163comment(file)\nprint('a')\n\nattach_media_tag_by_conf_path()\n\nprint('a')\n"},{"alpha_fraction":0.5739436745643616,"alphanum_fraction":0.5850487351417542,"avg_line_length":27.604650497436523,"blob_id":"b6815bacf89ab5be051803d8579230d0e2a05f31","content_id":"3394e3a705dc9c092f4303b337cab3e3e6f6d43e","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3692,"license_type":"permissive","max_line_length":91,"num_lines":129,"path":"/NXSpider/utility/modify_idv3.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/5/17.\n# email to LipsonChan@yahoo.com\n#\nimport os\nimport shutil\n\nfrom NXSpider.common.config import Config\nfrom NXSpider.model.mongo_model import Mp3Model, Mp4Model\nfrom NXSpider.spider.mp3 import Mp3\nfrom NXSpider.spider.mv import MV\nfrom NXSpider.common import tools, log, constant\nfrom NXSpider.spider.base_driver import Music163Obj\nfrom NXSpider.utility.media_tag import attach_media_tag\n\nmedia_types = {\n 'mp3': [Mp3Model, Mp3],\n 'mp4': [Mp4Model, MV],\n}\npaths = Config().get_paths()\n\n\ndef attach_media_idv3_by_db():\n \"\"\"\n attach media idv3 by paths in config\n :return:\n \"\"\"\n for suffix, type_setting in media_types:\n model, driver = type_setting # type: Mp3Model or Mp4Model, Music163Obj\n objs = model.objects(downloaded=True)\n for obj in objs:\n file_path = driver.download_check(obj, check_file=True)\n if not file_path:\n continue\n\n if is_latin1(file_path):\n res = attach_media_tag(obj, file_path)\n else:\n res = attach_shadow(file_path, suffix, obj)\n\n log.print_info('idv3 attach %s, file: %s'\n % ('success' if res else 'failed', file_path))\n\n\n@tools.ignored(Exception)\ndef load_media_obj(media_type, download_name):\n return media_types[media_type][0].objects(download_file_name=download_name).first()\n\n\ndef create_tmp_dir():\n path = os.path.join(constant.main_dir, 'tmp')\n if not os.path.exists(path):\n os.makedirs(path)\n return path\n\n\ndef delete_tmp_dir():\n path = os.path.join(constant.main_dir, 'tmp')\n if os.path.exists(path):\n os.removedirs(path)\n\n\ndef is_latin1(s):\n try:\n s.encode(\"iso-8859-1\")\n return True\n except UnicodeEncodeError:\n return False\n\n\ndef attach_media_tag_by_path(path):\n \"\"\"\n please run in python3 if your os is windows, cause os.walk has a encoding bug\n :type path: str\n :param path:\n :return:\n \"\"\"\n\n for root, dirs, files in os.walk(path):\n for file in files: # type: str\n suffix_i = file.rfind('.')\n suffix = file[suffix_i + 1:]\n if suffix not in media_types:\n continue\n\n split_txt = ' - '\n split_i = file.find(split_txt)\n if split_i == -1:\n continue\n\n artist = file[:split_i]\n download_file_name = os.path.join(artist, file)\n obj = load_media_obj(suffix, download_file_name)\n file_path = os.path.join(root, file)\n\n # fuck!!!! although i fix python-magic encoding bug of Chinese str in windows,\n # the fuck libmagic.dll doesn't recognize Korean. so, i am really pissed off to\n # change file name and change back\n '''\n if is_latin1(file_path):\n res = attach_media_tag(obj, file_path)\n else:\n res = attach_shadow(file_path, suffix, obj)\n '''\n\n # Dobby is free, thanks to mutagen\n res = attach_media_tag(obj, file_path)\n\n log.print_info('idv3 attach %s, file: %s'\n % ('success' if res else 'failed', file_path))\n\n\ndef attach_shadow(file, suffix, obj):\n tmp_name = 'abc'\n tmp_file = os.path.join(create_tmp_dir(), tmp_name + '.' + suffix)\n shutil.copy(file, tmp_file)\n res = attach_media_tag(obj, tmp_file)\n if res:\n os.remove(file)\n shutil.copy(tmp_file, file)\n os.remove(tmp_file)\n return res\n\n\ndef attach_media_tag_by_conf_path():\n for path in paths:\n attach_media_tag_by_path(path)\n\n\n"},{"alpha_fraction":0.5217254161834717,"alphanum_fraction":0.5456548929214478,"avg_line_length":22.701492309570312,"blob_id":"7be0b87f30c2310ff4afd9df912a7f946d96a6f7","content_id":"1844da35769a2acbc61697b96657038cfe5d963c","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3192,"license_type":"permissive","max_line_length":80,"num_lines":134,"path":"/NXSpider/utility/idv3_base.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/5/16.\n# email to LipsonChan@yahoo.com\n#\nimport os\n\nimport eyed3\nfrom eyed3.id3 import ID3_V2_3\nfrom mutagen.id3 import ID3, Encoding\nfrom mutagen.mp4 import MP4\nfrom mutagen import id3\n\nkey_names = ['title', 'artist', 'album', 'album_artist', 'track_num', 'comment']\nencoding_keys = ['title', 'artist', 'album', 'album_artist', 'comment']\n# ['TIT2', 'TPE1', 'TRCK', 'TALB', 'TPOS', 'TSSE', 'APIC:', 'COMM::XXX']\n# 名称 歌手 trackNum 专辑 pos 编码 comments\n# 'TPE2' 'aART' album_artist\nmutagen_idv3_key_map = {\n 'title': 'TIT2',\n 'artist': 'TPE1',\n 'album': 'TALB',\n 'album_artist': 'TPE2',\n 'track_num': 'TRCK',\n 'commemt': 'COMM'\n}\n\nmugagen_mp4_key_map = {\n 'title': '\\xa9nam',\n 'artist': '\\xa9ART',\n 'comment': '\\xa9cmt',\n}\n\n\ndef delete_163comment(file):\n \"\"\"\n :type file: str\n :param file:\n :return:\n \"\"\"\n if not os.path.exists(file):\n return\n suffix = file[file.rfind('.'):]\n if suffix not in ['.mp3', '.mp4']:\n return\n\n try:\n id3_file = eyed3.load(file)\n id3_file.initTag(version=ID3_V2_3)\n id3_file.tag.comments.remove('', lang=b'XXX')\n id3_file.tag.save()\n except Exception as e:\n raise e\n\n\ndef attach_mp3_idv3(filename, data):\n\n\n try:\n mp3 = ID3(filename, v2_version=3)\n for k, v in data.items():\n if k not in mutagen_idv3_key_map:\n continue\n\n if k == 'comment':\n mp3.add(id3.COMM(lang='XXX', text=v))\n continue\n\n attr_type = getattr(id3, mutagen_idv3_key_map[k], None)\n if attr_type:\n mp3.add(attr_type(text=v))\n\n mp3.save()\n except Exception as e:\n return False\n return True\n\n\ndef attach_mp4_tag(filename, data):\n if not os.path.exists(filename):\n return False\n\n try:\n mp4 = MP4(filename)\n for k, v in data.items:\n if k not in mugagen_mp4_key_map:\n continue\n mp4[mugagen_mp4_key_map[k]] = v\n\n mp4.save()\n except Exception as e:\n return False\n return True\n\n\ndef attach_media_idv3(filename, data, encoding_func=None):\n \"\"\"\n attach data to idv3 of a file\n :param encoding_func: change encoding of values\n :type encoding_func: function | None\n :type filename: str\n :param filename:\n :type data: dict\n :param data:\n :return:\n \"\"\"\n if not os.path.exists(filename):\n return\n suffix = filename[filename.rfind('.'):]\n if suffix not in ['.mp3', '.mp4']:\n return\n\n try:\n id3_file = eyed3.load(filename)\n id3_file.initTag(version=ID3_V2_3)\n\n for k, v in data.items():\n if k not in key_names:\n continue\n\n if encoding_func and k in encoding_keys:\n v = encoding_func(v)\n\n if k == 'comment':\n id3_file.tag.comments.set(v, description=u'', lang=b'XXX')\n continue\n\n setattr(id3_file.tag, k, v)\n id3_file.tag.save()\n except Exception as e:\n return False\n\n return True\n"},{"alpha_fraction":0.5893169641494751,"alphanum_fraction":0.5980735421180725,"avg_line_length":12.75903606414795,"blob_id":"b9312747257c76e58e1a96dd24c26b58a4bc54db","content_id":"00024bb729b062fe31c7b5f136b605884c043357","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1142,"license_type":"permissive","max_line_length":42,"num_lines":83,"path":"/NXSpider/model/dict_model.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/19.\n# email to LipsonChan@yahoo.com\n#\n\n\nclass BaseDoc(dict):\n def save(self):\n pass\n\n def __getattr__(self, item):\n if item not in self:\n return None\n\n value = self[item]\n if isinstance(value, dict):\n value = BaseDoc(value)\n return value\n\n\ndef get_one_model_by_key(model, model_id):\n \"\"\"\n load or create a model by id\n :param model:\n :param model_id:\n :return: doc, is_new\n :rtype: (DynamicDocument, boolean)\n \"\"\"\n\n return model(id=model_id), True\n\n\ndef update_dynamic_doc(doc, data):\n \"\"\"\n :type doc: dict\n :type data: dict\n :param doc:\n :param data:\n :return:\n \"\"\"\n doc.update(data)\n\n\nclass ConfigModel(BaseDoc):\n pass\n\n\nclass AuthorModel(BaseDoc):\n pass\n\n\nclass Mp3Model(BaseDoc):\n pass\n\n\nclass ArtistModel(BaseDoc):\n pass\n\n\nclass Mp4Model(BaseDoc):\n pass\n\n\nclass VideoModel(BaseDoc):\n pass\n\n\nclass PlaylistModel(BaseDoc):\n pass\n\n\nclass AlbumModel(BaseDoc):\n pass\n\n\nclass UserModel(BaseDoc):\n pass\n\n\nclass Downloaded(BaseDoc):\n pass\n"},{"alpha_fraction":0.6626602411270142,"alphanum_fraction":0.71875,"avg_line_length":25.02083396911621,"blob_id":"c9fc4189718da3f0c1194de25751a5e944384b8e","content_id":"18c8965992575f8a0b1a8f57ffc742cfa213ea2f","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":2290,"license_type":"permissive","max_line_length":109,"num_lines":48,"path":"/VERSION.md","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"NXSpider\n=================\n\nNXSpider， NetEase X Spider，一个强大的（网易云音乐）mp3,mv爬虫，可以下载和收集mp3,mv信息,同时附带多媒体标签信息。采用python编写，mongo数据库(非必须)，递归算法核心实现\n\n[](LICENSE.md)\n\n\n\n[新手先看这里](SIMPLE_USE.md) | [历史版本在这里](VERSION.md) | [开发详情在这里](DEV.md)\n\n## Update History\n\n#### 2018-12-17 v0.0.5\n- 增加功能，通过配置，创建歌单，歌手，唱片目录系统（快捷方式）\n- 增加功能，通过歌曲id，爬取歌曲\n- 增加功能（失败），通过登录爬取收藏视频（非MV），但由于接口问题，未能使用\n- 优化说明文档和安装脚本，使安装和更新更容易\n\n#### 2018-06-23 v0.0.4\n- 增加功能，通过登录爬取用户收藏mv\n- 增加功能，通过登录爬取用户所有歌单（包括收藏）\n- 增加功能，通过歌单分类，爬取最火歌单\n- 增加功能，打印歌单分类\n- 增加功能，爬取最火mv\n- 优化代码，优化所有文档，优化输出\n- 解决小bug：mv id存在但获取不到详情时的代码异常\n- 默认mongo非使用，且不安装python依赖包，通过 `pip+req.txt` 进行安装\n- 方便windows用户增加小部分bat\n- 项目功能已满足作者使用，暂时没什么开发欲望了\n\n#### 2018-06-21 v0.0.3\n- 增加功能，通过用户id获得歌单，并爬取歌单信息，mp3,mv\n- 增加功能，通过歌手id爬取所有专辑信息，下载所有专辑mp3,mv\n- 增加 `VERSION.md`\n\n#### 2018-06-20 v0.0.2\n- mongodb改为非必须，默认不需要，并提供相应配置\n- 增加功能，爬取特定分类最火歌单信息并下载mp3,mv\n- 增加 `SIMPLE_USE.md` 提供最简单的入门说明\n- 优化部分代码\n\n#### 2018-06-19 v0.0.1\n- 首次发布, mongodb + 递归 + python(3) + tag\n- 搜索歌单，歌手，专辑，用户等信息\n- 通过歌单，歌手，专辑id，下载mp3, mv\n- 所有信息会保存早mongodb中\n- 高可配置，所有mp3, mv可以通过配置，增加多媒体标签信息(歌手，专辑，唱片，163comment!!!)"},{"alpha_fraction":0.5491606593132019,"alphanum_fraction":0.5755395889282227,"avg_line_length":18.85714340209961,"blob_id":"530e7a9bfccc5dadbabf8c69e909dd2b5dfd553c","content_id":"4b47fcab01e8332da4b185ae8a9d4749d9b4ca5f","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":417,"license_type":"permissive","max_line_length":41,"num_lines":21,"path":"/NXSpider/spider/user.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/25.\n# email to LipsonChan@yahoo.com\n#\nfrom NXSpider.spider.base_driver import *\n\n\nclass User(Music163Obj):\n __model_name__ = UserModel\n __model__rfilter__ = {\n 'userId',\n\n # garbage properties\n 'avatarImgId',\n 'backgroundImgId',\n 'avatarImgIdStr',\n 'backgroundImgIdStr',\n 'avatarImgId_str',\n }\n"},{"alpha_fraction":0.5626780390739441,"alphanum_fraction":0.5769230723381042,"avg_line_length":24.962963104248047,"blob_id":"04697374363e500c53ab8b0b11ee5d1d0825f07c","content_id":"710eb2425b4a0a8a046e80e1606da6b350d00bbe","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":702,"license_type":"permissive","max_line_length":78,"num_lines":27,"path":"/NXSpider/debug_test/change_mp3_to_mp4.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/5/15.\n# email to LipsonChan@yahoo.com\n#\nimport os\nimport shutil\n\npaths = ['mv']\nproject_path = os.path.dirname(os.path.realpath(__file__))\nproject_path = os.path.dirname(project_path)\ndownload_path = os.path.join(project_path, 'download_files')\n\n\ndef rename_file(path):\n for root, dirs, files in os.walk(path):\n for f in files:\n f = f # type: str\n suffix = f[f.rfind('.'):]\n if suffix == '.mp3':\n new_name = f[:f.rfind('.')] + \".mp4\"\n os.rename(os.path.join(root, f), os.path.join(root, new_name))\n\n\nfor p in paths:\n rename_file(os.path.join(download_path, p))\n\n"},{"alpha_fraction":0.439113050699234,"alphanum_fraction":0.6412213444709778,"avg_line_length":51.903846740722656,"blob_id":"4d9428d2ea1a15e718af14fa2680b79d4b04a47c","content_id":"96fe0f6ea777e48aff68fb75b1c0244e67f8bd11","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2751,"license_type":"permissive","max_line_length":314,"num_lines":52,"path":"/NXSpider/debug_test/aes_decode.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/5/15.\n# email to LipsonChan@yahoo.com\n#\nimport base64\n\nfrom NXSpider.common.encrypt import aes_ecb, aes_ecb_decode\n\nBS = 16\npad = lambda s: s + (BS - len(s) % BS) * chr(0)\nunpad = lambda s: s[0:-ord(s[-1])]\n\naes_code = \"#14ljk_!\\]&0U<'(\"\n\nciphertext = \"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\"\nciphertext = \"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\"\nciphertext = \"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\"\nciphertext = \"wUxNl9+oaCX1+TGUoqXxPMW1W97D0rB/UWviTQG1qGf7VPc9xScpSdO78dNI29kCFN9qWTWmEEe2JvA0TgeZBdmmiikLL8RQ6zvuAfKjzTOJhL6Ff1X4hw6syUlw+wqQwRL0WQvY6ysW/OlDmqa/GHaZOirjQv580Xplu9P2wBfxW3T8HrjzPdtS05xs2Pa6JgoErDzGhmlv7ny0yYWkHqwfFtlDDlwUTAAEOXhEFI/yNE5qt4jaQABURo9ZCBoircTF92wi2BSkQHoTzDgarpOnvzmafSOR+46gt3OolILhXrtef/wPpXih9gkLrKY0\"\nciphertext = \"I7ZTqvekU9UjuYios3KeugpKIryhEim4OTsU7t3xaN2ZBcRHRQ0XzpsDmVa76Sh8aJ4JOmBcD37oN9sLBCkqOpoGAVROkWpNfEsYTsLz13E4Q4OIT+BBn9Acvw9CNtviGXBUyDkeec3x4RGuI6T/efUn+MntkU2ZkyyIt7VhcoV7XTFGyFx9A2UB72Tw/vn2iRXeTeS0wuAJqxJN6b9GINrizT8a+QORT7FoyPl7e4Zf8Sg4E51UnhuOGtulNc8n7MnOkbyfJSZWfHD2sHRi6g==\"\nplaintext = 'music:{\"albumId\":2084576,\"alias\":[],\"mMusic\":\"{\\\\\"fid\\\\\":2007708232328038,\\\\\"br\\\\\":160000,' \\\n '\\\\\"size\\\\\":10701396}\",\"album\":\"Eternal Light\",\"musicId\":22712173,\"musicName\":\"Refrain\",\"mvId\":0,' \\\n '\"duration\":532000,\"lMusic\":\"{\\\\\"fid\\\\\":2008807743955814,\\\\\"br\\\\\":96000,\\\\\"size\\\\\":6437585}\",' \\\n '\"bitrate\":320000,\"albumPic\":\"http://p2.music.126.net/fNtMX44fvaGByURP0AbOZQ==/836728348761063.jpg\",' \\\n '\"mp3DocId\":1981319953261764,\"artist\":[[\"Anan Ryoko\",16069]],\"hMusic\":\"{\\\\\"fid\\\\\":1981319953261764,' \\\n '\\\\\"br\\\\\":320000,\\\\\"size\\\\\":21359877}\",\"albumPicDocId\":836728348761063}\"'\nplaintext = 'music:{\"musicId\":28341255,\"musicName\":\"Dancing Alone\",\"albumId\":2774171,\"album\":\"Remove Before Flight\",\"albumPicDocId\":6025323720389239,\"albumPic\":\"http://p1.music.126.net/w0uwPgaqDCJYYZJjzQfa7w==/6025323720389239.jpg\",\"bitrate\":320000,\"mp3DocId\":\"69bb414114182d8e7b8a5214168bc991\",\"duration\":211200}'\nplaintext = 'music:{\"musicId\":28341255,\"musicName\":\"Dancing Alone\",' \\\n '\"albumId\":2774171,\"album\":\"Remove Before Flight\",\"albumPicDocId\":6025323720389239,' \\\n '\"albumPic\":\"http://p1.music.126.net/w0uwPgaqDCJYYZJjzQfa7w==/6025323720389239.jpg\",' \\\n '\"bitrate\":320000,\"mp3DocId\":\"69bb414114182d8e7b8a5214168bc991\",\"duration\":211200}'\nplaintext = 'music:{\"musicId\":28341255,\"musicName\":\"Dancing Alone\",' \\\n '\"albumId\":2774171,\"album\":\"haha\",\"albumPicDocId\":6025323720389239,' \\\n '\"bitrate\":320000,\"mp3DocId\":\"69bb414114182d8e7b8a5214168bc991\",\"duration\":211200}'\n# plaintext = 'music:{\"musicId\":28341255,\"musicName\":\"Dancing Alone\",' \\\n# '\"bitrate\":320000}'\n# plaintext = 'mv:{\"title\":\"LATATA\",\"mvId\":5906035,\"artistId\":127524,' \\\n# '\"artistName\":\"Rainbow\",\"picId\":5991238859751539,\"pubTime\":\"2013-05-17\",' \\\n# '\"bitrate\":240,\"duration\":258000,\"briefIntro\":\"\",\"detailIntro\":\"\"}'\n# plaintext = 'mv:{\"title\":\"LATATA\",\"mvId\":5906035,\"artistId\":14055085,' \\\n# '\"artistName\":\"(G)I-DLE\",\"picId\":109951163279867003,\"pubTime\":\"2018-05-02\",' \\\n# '\"bitrate\":720,\"duration\":221000,\"briefIntro\":\"\",\"detailIntro\":\"\"}'\n\nb = aes_code.encode('utf-8')\na = base64.b64decode(ciphertext)\na = aes_ecb_decode(ciphertext, aes_code)\nc = aes_ecb(plaintext, aes_code)\nd = c.decode()\n\nprint(d)\nprint(d)\n"},{"alpha_fraction":0.5011028051376343,"alphanum_fraction":0.507278323173523,"avg_line_length":26.64634132385254,"blob_id":"6c8d9a50fde0455f8c3b2dcacce2089db3fcd7f6","content_id":"1fda559ba5125955eedbcab03ba9b7182cde17ff","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2267,"license_type":"permissive","max_line_length":87,"num_lines":82,"path":"/NXSpider/spider/mv.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/20.\n# email to LipsonChan@yahoo.com\n#\nimport os\nimport re\n\nfrom NXSpider.common.config import Config\nfrom NXSpider.spider.api import get_mv_link\nfrom NXSpider.spider.artist import Artist\nfrom NXSpider.spider.base_driver import *\n\n\ndef get_target_r(obj, limit_r=Config().get_mv_resolution()):\n max_valid = max([x['br'] for x in obj['brs']])\n return limit_r if max_valid > limit_r else max_valid\n\n\nclass MV(Music163Obj):\n __model_name__ = Mp4Model\n __model_rfilter__ = {\n }\n __parse_recursion__ = {\n 'artists': Artist(),\n }\n\n def download_filename(self, doc):\n \"\"\"\n implement pls\n get a name to save file\n need be complete by child\n :param doc:\n :return:\n \"\"\"\n authors = \",\".join([x['name'] for x in doc.artists])\n author = re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', authors.strip())\n mv_name = re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', doc['name'])\n name = \"%s - %s.mp4\" % (author, mv_name)\n return name\n\n def download_filename_full(self, doc):\n \"\"\"\n implement pls\n get a path to save file, by relative path\n need be complete by child\n :param doc:\n :return:\n :rtype: str\n \"\"\"\n authors = \",\".join([x['name'] for x in doc.artists])\n author = re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', authors.strip())\n mv_name = re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', doc['name'])\n name = os.path.join(author, \"%s - %s.mp4\" % (author, mv_name))\n return name\n\n def url_load(self, doc):\n \"\"\"\n implement pls\n :param doc:\n :return:\n :rtype: str\n \"\"\"\n try:\n target_r = get_target_r(doc, Config().get_mv_resolution())\n doc['download_video_r'] = target_r\n return get_mv_link(doc['id'], target_r)\n except:\n return None\n\n def shortcut_self_path(self, doc):\n \"\"\"\n implement pls, not force\n return self short cut path\n :param doc:\n :return:\n \"\"\"\n result = []\n result.extend([os.path.join(\"artist\", re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', x['name']))\n for x in doc.artists])\n return result\n"},{"alpha_fraction":0.5629921555519104,"alphanum_fraction":0.586614191532135,"avg_line_length":20.16666603088379,"blob_id":"5c3dea037165cd937978439b869b6f853a5d2094","content_id":"00b7c92bb06e4132dbbd5c12886224d6a47cd901","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":508,"license_type":"permissive","max_line_length":52,"num_lines":24,"path":"/NXSpider/spider/album.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/26.\n# email to LipsonChan@yahoo.com\n#\n\nfrom NXSpider.spider.artist import Artist\nfrom NXSpider.spider.base_driver import *\n\n\nclass Album(Music163Obj):\n __model_name__ = AlbumModel\n __model_rfilter__ = {\n # 'song', # TODO need delete???\n 'artist',\n }\n __parse_recursion__ = {\n 'artists': Artist(),\n }\n\n @attr_replace(attr_name='songs', new_name='mp3')\n def replace_song(self, obj):\n return obj\n"},{"alpha_fraction":0.6957504749298096,"alphanum_fraction":0.7034357786178589,"avg_line_length":23.842697143554688,"blob_id":"3211e6d00cef7ff32038c21e5ffda2e4202fc108","content_id":"5b0a6e0dcda0ff66346e7a48dd3460c21e78492a","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":2212,"license_type":"permissive","max_line_length":78,"num_lines":89,"path":"/NXSpider/common/tools.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n\n\"\"\"\nMIT License\n\nCopyright (c) 2017 Cheng YuMeng\n\nPermission is hereby granted, free of charge, to any person obtaining a copy\nof this software and associated documentation files (the \"Software\"), to deal\nin the Software without restriction, including without limitation the rights\nto use, copy, modify, merge, publish, distribute, sublicense, and/or sell\ncopies of the Software, and to permit persons to whom the Software is\nfurnished to do so, subject to the following conditions:\n\nThe above copyright notice and this permission notice shall be included in all\ncopies or substantial portions of the Software.\n\nTHE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\nIMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\nFITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\nAUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\nLIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\nOUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\nSOFTWARE.\n\"\"\"\n\nimport codecs\nimport contextlib\nimport hashlib\nimport locale\n\nfrom NXSpider.common import PYTHON3\n\nRETURN_JSON = \"return json data\"\nRETURE_HTML = \"return html data\"\n\nlang, sys_encoding = locale.getdefaultlocale()\n\n@contextlib.contextmanager\ndef ignored(*exceptions):\n try:\n yield\n except exceptions:\n pass\n\n\ndef encode(s):\n if PYTHON3 is True:\n return codecs.encode(s, \"utf-8\").decode(\"utf-8\")\n else:\n return s.encode(\"utf-8\")\n\n\ndef encode_sys(s):\n s.encode(sys_encoding)\n\n\ndef hex(s):\n if PYTHON3 is True:\n return codecs.encode(bytes(s, encoding=\"utf8\"), 'hex')\n else:\n return s.encode(\"hex\")\n\n\ndef md5(s):\n m = hashlib.md5()\n m.update(s.encode(\"utf-8\"))\n return m.hexdigest()\n\n\ndef is_unicode(obj):\n if PYTHON3:\n return isinstance(obj, str)\n else:\n from __builtin__ import unicode\n return isinstance(obj, unicode)\n\n\ndef input_format(string):\n \"\"\"\n :type string: str\n :param string:\n :return:\n \"\"\"\n if PYTHON3:\n return string\n else:\n return string.encode()\n\n"},{"alpha_fraction":0.5391644835472107,"alphanum_fraction":0.5489556193351746,"avg_line_length":25.413793563842773,"blob_id":"122dc14c3630ef281f91bcdf46677b4d920a59d1","content_id":"154f9f791c283ba1c48c05f04736b6da3361597c","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1532,"license_type":"permissive","max_line_length":71,"num_lines":58,"path":"/NXSpider/spider/video.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/12/12.\n# email to LipsonChan@yahoo.com\n#\n\nimport os\nimport re\n\nfrom NXSpider.common.config import Config\nfrom NXSpider.spider.api import get_video_link\nfrom NXSpider.spider.artist import Artist\nfrom NXSpider.spider.base_driver import *\n\n\ndef get_target_r(obj, limit_r=Config().get_mv_resolution()):\n max_valid = max([x['resolution'] for x in obj['resolutions']])\n return limit_r if max_valid > limit_r else max_valid\n\n\nclass Video(Music163Obj):\n __model_name__ = VideoModel\n __model_rfilter__ = {\n }\n __parse_recursion__ = {\n 'artists': Artist(),\n }\n\n def download_filename_full(self, doc):\n \"\"\"\n implement pls\n get a path to save file, by relative path\n need be complete by child\n :param doc:\n :return:\n :rtype: str\n \"\"\"\n # todo modify\n authors = \",\".join([x['name'] for x in doc.artists])\n author = re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', authors.strip())\n mp3_name = re.sub(\"[\\\\\\\\/:*?\\\"<>|]\", '', doc['name'])\n name = os.path.join(author, \"%s - %s.mp4\" % (author, mp3_name))\n return name\n\n def url_load(self, doc):\n \"\"\"\n implement pls\n :param doc:\n :return:\n :rtype: str\n \"\"\"\n try:\n target_r = get_target_r(doc, Config().get_mv_resolution())\n doc['download_video_r'] = target_r\n return get_video_link(doc['id'], target_r)\n except:\n return None\n"},{"alpha_fraction":0.49380558729171753,"alphanum_fraction":0.5022236108779907,"avg_line_length":30.01477813720703,"blob_id":"a1184863c446c795f067e5b8a110fb09ef6a9da3","content_id":"e80ff9c0e9f50356db461db01d919055b25003ec","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":6296,"license_type":"permissive","max_line_length":104,"num_lines":203,"path":"/NXSpider/common/config.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/19.\n# email to LipsonChan@yahoo.com\n#\nimport json\nimport os\n\nfrom NXSpider.common import log\nfrom NXSpider.common.constant import config_path, default_download_dir\nfrom NXSpider.common.singleton import Singleton\n\ndefault_path_key = 'default'\nmv_resolutions = [240, 480, 720, 1080]\n\n\ndef utf8_data_to_file(f, data):\n if hasattr(data, 'decode'):\n f.write(data.decode('utf-8'))\n else:\n f.write(data)\n\n\nclass Config(Singleton):\n def __init__(self):\n if hasattr(self, '_init'):\n return\n self._init = True\n self.config_file_path = config_path\n\n # the deep level must be less than 2\n self.default_config = {\n 'mongo': {\n 'name': 'nxspider',\n 'host': 'localhost',\n 'port': 27017,\n 'username': 'None',\n 'password': 'None',\n },\n 'download_path': [default_download_dir],\n 'mv_def_resolution': 720,\n 'media_tag': True,\n 'media_tag_163': True,\n 'download_file_check': True,\n 'debug_log': True,\n 'no_mongo': True,\n 'shortcut': False,\n }\n self.config = {}\n\n need_save = False\n try:\n f = open(self.config_file_path, 'r')\n self.config = json.loads(f.read())\n f.close()\n except IOError or ValueError:\n self.__generate_config_file()\n need_save = True\n\n # merge default\n for k, v in self.default_config.items():\n if k not in self.config:\n self.config[k] = v\n need_save = True\n\n paths = self.config['download_path'] # type: list\n if len(paths) == 0:\n paths.append(default_download_dir)\n need_save = True\n\n if need_save:\n self.save_config_file()\n\n def __getitem__(self, key):\n if key in self.config:\n return self.config[key]\n return None\n\n def __generate_config_file(self):\n f = open(self.config_file_path, 'w')\n utf8_data_to_file(f, json.dumps(self.default_config, indent=2))\n f.close()\n\n def save_config_file(self):\n f = open(self.config_file_path, 'w')\n utf8_data_to_file(f, json.dumps(self.config, indent=2))\n f.close()\n\n def get_shortcut(self):\n return self.config['shortcut']\n\n def get_mongo(self):\n return self.config['mongo']\n\n def get_path(self):\n return self.config['download_path'][0]\n\n def get_paths(self):\n return self.config['download_path']\n\n def get_mv_resolution(self):\n return self.config['mv_def_resolution']\n\n def get_media_tag(self):\n return self.config['media_tag']\n\n def get_media_tag_163(self):\n return self.config['media_tag_163']\n\n def get_file_check(self):\n return self.config['download_file_check']\n\n def get_no_mongo(self):\n return self.config['no_mongo']\n\n def save_config_dict(self, obj):\n \"\"\"\n :type obj: dict\n :param obj:\n :return:\n \"\"\"\n for k, v in obj.items():\n if isinstance(v, dict) and isinstance(self.config[k], dict):\n self.config[k].update(v)\n else:\n self.config[k] = v\n\n def config_test(self):\n result = True\n try:\n\n # check mongodb config\n if self.config['no_mongo'] is False:\n log.print_info('check mongodb config')\n mongo = self.config['mongo']\n for k in ['name', 'host', 'port']:\n if k not in mongo:\n log.print_err(\"mongo config error, key mongo.{} is not set yet\".format(k))\n result = False\n\n # try import model, which will connect to server and exit if server config wrong\n import NXSpider.model.mongo_model\n\n for k in ['download_path', 'mv_def_resolution', 'media_tag', 'media_tag_163']:\n if k not in self.config:\n log.print_err(\"config error, key {} is not set yet\".format(k))\n result = False\n\n # check type\n type_check = {\n 'download_path': list,\n 'mv_def_resolution': int,\n 'media_tag': bool,\n 'media_tag_163': bool,\n 'download_file_check': bool,\n 'no_mongo': bool,\n }\n\n need_save = False\n for k, v in type_check.items():\n if not isinstance(self.config[k], v):\n log.print_err(\"config error, {} is not a require type, \"\n \"and is reset to default value: {}\".format(k, self.default_config[k]))\n self.config[k] = self.default_config[k]\n need_save = True\n result = False\n\n # download path check\n final_paths = []\n for p in self.config['download_path']:\n try:\n # some error need pass\n if os.path.isdir(p) is False:\n os.mkdir(p)\n final_paths.append(p)\n except:\n log.print_warn(\"download path may be wrong and be deleted: {}\".format(p))\n need_save = True\n result = False\n pass\n\n # mv resolution check\n if self.config['mv_def_resolution'] not in mv_resolutions:\n log.print_warn(\"mv_def_resolution will be reset to default: {}\"\n .format(self.default_config['mv_def_resolution']))\n self.config['mv_def_resolution'] = self.default_config['mv_def_resolution']\n need_save = True\n result = False\n\n if need_save:\n self.config['download_path'] = final_paths\n self.save_config_file()\n\n return result\n except Exception as e:\n log.print_err(e)\n\n return False\n\n def config_reset(self):\n self.config = self.default_config\n self.save_config_file()\n"},{"alpha_fraction":0.6059637665748596,"alphanum_fraction":0.6448349356651306,"avg_line_length":25.828571319580078,"blob_id":"dd33d4596c0fc2bbcee9abb15c27b21f8b82e59b","content_id":"abbdf977aa3e6baca5a2db2c08a58a98e54d815f","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1878,"license_type":"permissive","max_line_length":80,"num_lines":70,"path":"/NXSpider/common/encrypt.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/19.\n# email to LipsonChan@yahoo.com\n#\n\nimport base64\n\nfrom cryptography.hazmat.primitives.ciphers import (\n Cipher, algorithms, modes\n)\nfrom cryptography.hazmat.backends import default_backend\n\n\ndef aes(text, sec_key):\n backend = default_backend()\n pad = 16 - len(text) % 16\n text_t = text + pad * chr(pad)\n cipher = Cipher(\n algorithms.AES(sec_key.encode('utf-8')),\n modes.CBC(b'0102030405060708'),\n backend=backend\n )\n encryptor = cipher.encryptor()\n ciphertext = encryptor.update(text_t.encode('utf-8')) + encryptor.finalize()\n ciphertext = base64.b64encode(ciphertext)\n return ciphertext\n\n\ndef aes_decode(text, sec_key):\n backend = default_backend()\n ciphertext = base64.b64decode(text)\n cipher = Cipher(\n algorithms.AES(sec_key.encode('utf-8')),\n modes.CBC(b'0102030405060708'),\n backend=backend\n )\n decryptor = cipher.decryptor()\n plaintext = decryptor.update(ciphertext) + decryptor.finalize()\n return plaintext\n\n\ndef aes_ecb_decode(text, sec_key):\n backend = default_backend()\n ciphertext = base64.b64decode(text)\n cipher = Cipher(\n algorithms.AES(sec_key.encode('utf-8')),\n modes.ECB(),\n backend=backend\n )\n decryptor = cipher.decryptor()\n plaintext = decryptor.update(ciphertext) + decryptor.finalize()\n return plaintext\n\n\ndef aes_ecb(text, sec_key):\n backend = default_backend()\n text = text.encode('utf-8')\n pad = 16 - len(text) % 16\n text_t = text + (b'\\0' * pad)\n cipher = Cipher(\n algorithms.AES(sec_key.encode('utf-8')),\n modes.ECB(),\n backend=backend\n )\n encryptor = cipher.encryptor()\n ciphertext = encryptor.update(text_t) + encryptor.finalize()\n ciphertext = base64.b64encode(ciphertext)\n return ciphertext\n"},{"alpha_fraction":0.5271945595741272,"alphanum_fraction":0.5307171940803528,"avg_line_length":29.856521606445312,"blob_id":"a2ee77697e3d80288bcdabc164de0dbd0148d59c","content_id":"777289756164b98bb187586869f34e6d4b7e0594","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":14196,"license_type":"permissive","max_line_length":86,"num_lines":460,"path":"/NXSpider/spider/base_driver.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/25.\n# email to LipsonChan@yahoo.com\n#\nimport codecs\nimport json\nimport os\nfrom inspect import isfunction\n\nimport requests\nimport six\n\nfrom NXSpider.common.config import Config\nfrom NXSpider.common import tools, log\nfrom NXSpider.model.export import model_download_path\nfrom NXSpider.utility.media_tag import attach_media_tag\nfrom NXSpider.model.export import *\n\n\nclass Music163ObjException(Exception):\n pass\n\n\nclass Music163ObjMetaClass(type):\n def __new__(mcs, name, bases, attrs):\n \"\"\"\n :param name:\n :type name: str\n :param bases:\n :param attrs:\n :type attrs: dict\n :return:\n \"\"\"\n if name == \"Music163Obj\":\n return type.__new__(mcs, name, bases, attrs)\n if '__model_name__' not in attrs or attrs['__model_name__'] is None:\n attrs['__model_name__'] = name + \"Model\"\n\n # set download level 1 path\n attrs['__file_type__'] = name.lower()\n\n # create download file dir for download path\n # obj_path = os.path.join(get_download_path(), attrs['__file_type__'])\n # if not os.path.exists(obj_path):\n # os.makedirs(obj_path)\n\n # create a empty dict\n attrs['__attrs_replace_fucs__'] = dict()\n attrs['__attrs_replace_map__'] = dict()\n\n # set replace\n replace_map = dict()\n for k, v in attrs.items():\n if isfunction(v):\n if getattr(v, '__attrs_replace__', False):\n org_name = getattr(v, '__attrs_replace_org__')\n replace_map[org_name] = getattr(v, '__attrs_replace_new__')\n attrs['__attrs_replace_fucs__'][org_name] = v\n attrs['__attrs_replace_map__'] = replace_map\n return type.__new__(mcs, name, bases, attrs)\n\n\ndef attr_replace(attr_name=None, new_name=None):\n \"\"\"\n replace attr value\n decorator for fun: (self, obj)->object\n\n :param new_name:\n :param attr_name:\n :return:\n \"\"\"\n\n def de(f):\n \"\"\"\n :type f: function\n :param f:\n :return:\n \"\"\"\n f.__attrs_replace__ = True\n def_name = f.__name__\n f.__attrs_replace_org__ = attr_name if attr_name else def_name\n f.__attrs_replace_new__ = new_name if new_name else def_name\n return f\n\n return de\n\n\n# class Music163Obj(object, metaclass=Music163ObjMetaClass):\nclass Music163Obj(six.with_metaclass(Music163ObjMetaClass)):\n __model_name__ = None\n __model_rfilter__ = set()\n __file_type__ = 'other'\n __parse_recursion__ = {}\n\n __attrs_replace_fucs__ = dict()\n __attrs_replace_map__ = dict()\n\n # __metaclass__ = Music163ObjMetaClass\n\n def __init__(self, *args, **kwargs):\n \"\"\"\n In some case, something want to change def class setting,\n like recursion or filter\n :param args:\n :param kwargs:\n \"\"\"\n for k, v in kwargs.items():\n if k in ['__parse_recursion__', '__model_rfilter__']:\n setattr(self, k, v)\n\n def download_filename(self, doc):\n \"\"\"\n implement pls\n get a name to save file\n need be complete by child\n :param doc:\n :return:\n :rtype: str\n \"\"\"\n return None\n\n def download_filename_full(self, doc):\n \"\"\"\n implement pls\n get a path to save file, by relative path\n need be complete by child\n :param doc:\n :return:\n :rtype: str\n \"\"\"\n return None\n\n def url_load(self, doc):\n \"\"\"\n implement pls\n :param doc:\n :return:\n :rtype: str\n \"\"\"\n return None\n\n def pre_save(self, doc, obj):\n \"\"\"\n implement pls, not force\n do something before doc save\n get download link here is better\n :param doc:\n :param obj:\n :return:\n \"\"\"\n pass\n\n def shortcut_relative_name(self, doc):\n \"\"\"\n implement pls, not force\n return short cut relative path\n :param doc:\n :return:\n :rtype: str\n \"\"\"\n pass\n\n def shortcut_self_path(self, doc):\n \"\"\"\n implement pls, not force\n return self short cut path\n :param doc:\n :return:\n :rtype: list[str]\n \"\"\"\n return []\n\n @staticmethod\n def request_file(doc):\n \"\"\"\n implement pls\n :param doc:\n :return:\n :rtype: bytes\n \"\"\"\n url = None\n try:\n url = getattr(doc, model_download_url, None)\n if url is None:\n return None\n\n r = requests.get(url)\n if r.status_code != 200:\n return None\n return r.content\n except Exception as e:\n log.print_info(u\"url download failed %s , err: %s\" % (url, e))\n return None\n\n def download_relative_path(self, doc):\n download_file_name = 'download_file_name'\n if download_file_name not in doc:\n doc[download_file_name] = self.download_filename_full(doc)\n filename = doc[download_file_name]\n if filename:\n return os.path.join(self.__file_type__, filename)\n else:\n return None\n\n def download_check(self, doc, check_file=False):\n \"\"\"\n :param doc:\n :param check_file: check dist file to download\n :type doc: DynamicDocument\n :return:\n \"\"\"\n if check_file:\n file_relative_path = self.download_relative_path(doc)\n for path in Config().get_paths():\n file_path = os.path.join(path, file_relative_path)\n if os.path.exists(file_path):\n self.download_log(doc, download_path=file_path)\n return file_path\n self.download_log(doc, downloaded=False)\n return False\n\n return hasattr(doc, model_is_download) and doc[model_is_download]\n\n def create_shortcut(self, doc, shortcuts_stack):\n \"\"\"\n :param doc:\n :param shortcuts_stack:\n :type shortcuts_stack: list[str]\n :return:\n \"\"\"\n from NXSpider.utility.shortcut import symlink\n if not (hasattr(doc, model_is_download) and doc[model_is_download]\n and hasattr(doc, model_download_path)\n and os.path.exists(doc[model_download_path])\n and self.download_filename(doc)):\n return\n\n file_path = doc[model_download_path]\n shortcut_root = os.path.join(Config().get_path(), 'shortcuts')\n file_name = self.download_filename(doc)\n if file_path and file_path is not True:\n if shortcuts_stack:\n path = os.path.join(shortcut_root, *shortcuts_stack)\n os.makedirs(path, exist_ok=True)\n target = os.path.join(path, file_name)\n symlink(file_path, os.path.join(shortcut_root, target))\n\n for k in self.shortcut_self_path(doc):\n path = os.path.join(shortcut_root, k)\n os.makedirs(path, exist_ok=True)\n target = os.path.join(path, file_name)\n symlink(file_path, os.path.join(shortcut_root, target))\n pass\n\n @staticmethod\n def download_log(doc, downloaded=True, download_path=\"\"):\n doc[model_is_download] = downloaded\n doc[model_download_path] = download_path\n\n def download_file_tag(self, filename, doc):\n if self.__file_type__ in ['mp3', 'mp4', 'mv'] \\\n and Config().get_media_tag():\n attach_media_tag(doc, filename)\n\n def download_file(self, doc):\n \"\"\"\n download file from music 163\n :param doc:\n :return:\n \"\"\"\n file_relative_path = self.download_relative_path(doc)\n path = Config().get_path()\n content = self.request_file(doc)\n\n if content is None:\n log.print_err(u\"file download failed : %s\" % file_relative_path)\n return False\n\n try:\n file_name = os.path.join(path, file_relative_path)\n\n # dir make\n dir_name = os.path.dirname(file_name)\n if not os.path.exists(dir_name):\n os.makedirs(dir_name)\n\n # file write\n with open(file_name, \"wb\") as code:\n code.write(content)\n\n self.download_file_tag(file_name, doc)\n\n log.print_info(u\"file download complete: %s\" % file_relative_path)\n self.download_log(doc)\n return True\n except Exception as e:\n log.print_err(\"file save failed : %s, err: %s\" % (file_relative_path, e))\n return False\n\n def debug_save_json(self, obj):\n with tools.ignored(Exception):\n relative_path = os.path.join(self.__file_type__ + '.debug',\n self.__file_type__ + \"_\" +\n str(obj['id']) + '.json')\n file_name = os.path.join(Config().get_path(), relative_path)\n with codecs.open(file_name, \"wb\", encoding='utf8') as code:\n test = json.dumps(obj, ensure_ascii=False)\n code.write(test)\n\n def debug_print(self, obj):\n msg = u\"spider {} complete, id: {}\".format(self.__file_type__, obj['id'])\n if 'name' in obj:\n msg += u\", name: {}\".format(obj['name'])\n\n log.print_info(msg)\n\n def try_download(self, doc, download_type, file_check):\n if self.__file_type__ not in download_type:\n return True\n\n # download file and set download flag\n if download_type is not None \\\n and not self.download_check(doc, check_file=file_check):\n # need download， try url which is set first or get new url and download\n if getattr(doc, model_download_url, None) \\\n and self.download_file(doc):\n return True\n\n # get download link here\n doc[model_download_url] = self.url_load(doc)\n self.download_file(doc)\n else:\n name = self.download_relative_path(doc)\n if name:\n log.print_info(u\"file is exist or is not need to download : %s\"\n % name)\n\n def parse_model(self, crawl_dict, download_type=None,\n file_check=False, save=True, debug=False,\n shortcuts_stack=None):\n \"\"\"\n Get a model from db or create, update and save!!!\n this will replace some attributes into models by load_save_model also.\n by @attr_replace(attr_name, new_name)\n :param debug:\n :param file_check:\n :param download_type:\n :param save: save doc\n :param crawl_dict: must have id attr\n :param shortcuts_stack: stack of shortcuts path\n :type crawl_dict: dict\n :type shortcuts_stack: list[str]\n :return:\n :rtype: DynamicDocument\n :type save: bool\n \"\"\"\n if debug:\n # self.debug_save_json(crawl_dict)\n pass\n\n # get id\n if 'id' not in crawl_dict:\n log.print_err(u\"can not load id by json obj %s\" % json.dumps(crawl_dict))\n return None\n doc_id = crawl_dict['id']\n\n # load a mongo document\n doc, is_new_doc = get_one_model_by_key(self.__model_name__, doc_id)\n if doc is None:\n log.print_err(u\"can not load a doc by obj %s_%d\"\n % (self.__file_type__, doc_id))\n return None\n\n # shortcuts in stack\n if shortcuts_stack is not None and isinstance(shortcuts_stack, list) \\\n and self.shortcut_relative_name(doc):\n shortcuts_stack.append(self.shortcut_relative_name(doc))\n\n # if is_new_doc:\n # replace attr or ignore\n obj = dict()\n for k, v in crawl_dict.items():\n if k in self.__model_rfilter__:\n continue\n obj[k] = v\n\n # replace object\n if k not in self.__attrs_replace_fucs__:\n continue\n\n # change attr\n if isinstance(v, list):\n v = [self.__attrs_replace_fucs__[k](self, x) for x in v]\n else:\n v = self.__attrs_replace_fucs__[k](self, v)\n\n # replace key name\n del obj[k]\n obj[self.__attrs_replace_map__[k]] = v\n\n # recursion replace a attr into a model\n for k, v in self.__parse_recursion__.items():\n if k not in obj:\n continue\n\n if isinstance(obj[k], list):\n obj[k] = [v.parse_model(x, save=save, download_type=download_type,\n file_check=file_check, debug=debug,\n shortcuts_stack=shortcuts_stack)\n for x in obj[k]]\n elif isinstance(obj[k], dict):\n obj[k] = v.parse_model(obj[k], save=save, download_type=download_type,\n file_check=file_check, debug=debug,\n shortcuts_stack=shortcuts_stack)\n\n # update json to doc, this must be after recursion\n update_dynamic_doc(doc, obj)\n\n # modify doc and\n self.pre_save(doc, crawl_dict)\n\n # try download\n self.try_download(doc, download_type, file_check)\n\n if shortcuts_stack:\n if self.shortcut_relative_name(doc):\n shortcuts_stack.pop()\n self.create_shortcut(doc, shortcuts_stack)\n\n # save document\n if save and callable(getattr(doc, 'save', None)):\n doc.save()\n\n if debug:\n self.debug_print(crawl_dict)\n\n return doc\n\n\n__all__ = [\n 'Music163Obj',\n 'attr_replace',\n 'Music163ObjException',\n\n 'ConfigModel',\n 'UserModel',\n 'AlbumModel',\n 'PlaylistModel',\n 'Mp4Model',\n 'VideoModel',\n 'ArtistModel',\n 'Mp3Model',\n 'AuthorModel',\n 'update_dynamic_doc',\n 'model_download_url',\n 'model_is_download',\n 'get_one_model_by_key',\n]\n"},{"alpha_fraction":0.619271457195282,"alphanum_fraction":0.6321974396705627,"avg_line_length":23.314285278320312,"blob_id":"8f211dd3739379317f3b16469266fd74c03aca61","content_id":"a461ba57bf3615e7e28e78cab6a3f5255be7c356","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":851,"license_type":"permissive","max_line_length":72,"num_lines":35,"path":"/NXSpider/debug_test/move.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/5/14.\n# email to LipsonChan@yahoo.com\n#\nimport os\nimport shutil\nimport sys\nfrom glob import glob\n\n# reload(sys)\n# sys.setdefaultencoding('utf8')\n\npaths = ['mv', 'mp3']\nproject_path = os.path.dirname(os.path.realpath(__file__))\nproject_path = os.path.dirname(project_path)\ndownload_path = os.path.join(project_path, 'download_files')\n\n\ndef move_file(path):\n files = os.listdir(path)\n for f in files:\n if os.path.isdir(os.path.join(path, f)):\n continue\n\n author = f[0: f.find(' - ')].strip()\n author_path = os.path.join(path, author)\n if not os.path.exists(author_path):\n os.makedirs(author_path)\n shutil.move(os.path.join(path, f), os.path.join(author_path, f))\n\n\nfor p in paths:\n move_file(os.path.join(download_path, p))\n"},{"alpha_fraction":0.6714165806770325,"alphanum_fraction":0.6974015235900879,"avg_line_length":27.759037017822266,"blob_id":"c194360d84bcbb7b02fae99f9ebac455922339a5","content_id":"3531c5fa50ce0bc70ecfdf790fa506ecc3d56fe4","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":3794,"license_type":"permissive","max_line_length":214,"num_lines":83,"path":"/SIMPLE_USE.md","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"NXSpider\n=================\n\nNXSpider, NetEase X Spider, A powerful for mp3,mv spider, can help you download mp3 and mv with media tags. Base on python, mongodb, and recursion algorithm\n\n[](LICENSE.md)\n\n\n\n[新手先看这里](SIMPLE_USE.md) | [历史版本在这里](VERSION.md) | [开发详情在这里](DEV.md)\n\n这里为非IT同学增加最简单使用方法（请保证电脑能上网），请根据以下流程。注意，所有跳转建议按住ctrl后再点击（或右键新窗口打开）\n\n### 点赞\n- 请给项目点个star\n- 请感谢网易云，感谢上帝，感谢国家，顺便的话感谢作者\n\n## Linux用户\n[摸着自己的良心说，到底是不是IT人员](README.md)\n\n## Mac用户\n待补充\n\n## Windows用户\n\n### 建议\n- 使用python3.6，最好卸载2.7\n- 所有路径（python安装路径，NXSpider路径，下载路径）不要有中文名，空格希望也不要有\n\n### 下载python3(自行跳过)\n- [点击这里下载(32)](https://www.python.org/ftp/python/3.6.5/python-3.6.5.exe) 或 [点击这里下载(64)](https://www.python.org/ftp/python/3.6.5/python-3.6.5-amd64.exe) 或 [自行选择](https://www.python.org/downloads/release/python-365/)\n- 双击你下载的文件，进行安装，跟正常安装软件一样\n- 推荐大概在第一页勾选 ```for all users```\n- 推荐大概在第二页选择自定义安装路径，大概是 ```C:\\Program Files```\n- **强烈要求**大概在第二页勾选 ```Add Python to environment variables```\n\n- 后面一直按继续即可\n- 安装完毕后按住键盘 ```win+r``` 在开始建附近会弹出个运行框，输入 ```cmd``` 回车\n\n- 回车后会弹出个黑色界面，通常我们叫 console(控制台)，输入 ```python``` 并回车\n- 看到以下界面则为安装成功\n\n\n### 下载本项目\n- [点击这里下载](https://github.com/Grass-CLP/NXSpider/archive/master.zip) 或者点击右上角 ```clone or download``` 选择zip\n- 解压缩到你想要的目录下，比如 ```G:/NXSpider``` 并进入目录\n- 双击目录下文件 ```install.bat``` 会弹出个窗，等待结束（希望没看到error）\n\n### 运行本项目\n1. 双击运行目录下 `start.bat` \n2. 在控制台中就可以输入说明中的命令了(可以ctrl+c ctrl+v或右键复制)\n3. 建议先输入 ```nxspider -h```\n3. 先配置你的下载目录，比如目录是 ```G:/media```, 则输入并回车 ```nxspider config-spider -path G:/media```\n4. **推荐**打开你的网易云音乐，点击:本地音乐 -> 选择目录 -> 添加文件夹 -> 选择你的目录 -> 确定\n\n### 通过搜索下载示例\n $ nxspider search -ar \"张学友\"\n $ nxspider sab -ab 19093 -dw mp3,mv \n # -dw mp3,mv表示想要下载mp3和mv，不想下载mv则删除最后的,mv\n \n $ nxspider search -mp3 \"爱我别走\"\n $ nxspider smp3s -mp3 187947 -dw mp3\n \n搜索歌神唱片找到ID：\n\n\n\n爬专辑19093：\n\n\n\n### 暴力下载最火歌单(不截图了)\n $ nxspider scls-pls -cls 全部 -dw mp3,mv\n\n\n### tips\n- 其他命令见 [README.md](README.md)\n- 目前暂时没做很好的退出，想退出直接把控制台(黑窗)关掉就好\n- 目前暂时没做很好的重试，每次输入命令前复制到txt里面，遇到卡住的话，关掉重新打开个重新输入运行就可以\n- 下载后重新进去网易云音乐，看看你的本地音乐，点击**匹配音乐**，是不是充满惊喜\n\n## 最后\n有其他问题请到issue下留言，或者发邮件给我LipsonChan的雅虎邮箱"},{"alpha_fraction":0.6365979313850403,"alphanum_fraction":0.6597937941551208,"avg_line_length":23.25,"blob_id":"6bc60273335c06c1aa5e1aecad9a98f96c6683f3","content_id":"2817592a065e44c071baba87b01fc8b1ec8472bc","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":388,"license_type":"permissive","max_line_length":63,"num_lines":16,"path":"/NXSpider/common/constant.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/5/30.\n# email to LipsonChan@yahoo.com\n#\nimport os\n\n\nall_download_type = ['mp3', 'mv']\nmain_dir = os.path.join(os.path.expanduser('~'), '.nxspider')\nconfig_path = os.path.join(main_dir, 'config.json')\ndefault_download_dir = os.path.join(main_dir, 'download_files')\n\nif os.path.isdir(main_dir) is False:\n os.mkdir(main_dir)\n"},{"alpha_fraction":0.652942419052124,"alphanum_fraction":0.6816385984420776,"avg_line_length":26.485713958740234,"blob_id":"c8dae3781978082a219f0600417114f6f8393fef","content_id":"3fea56e7d126bd3b85739579264105143c1955b6","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":6745,"license_type":"permissive","max_line_length":109,"num_lines":175,"path":"/README.md","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"NXSpider\n=================\n\nNXSpider， NetEase X Spider，一个强大的（网易云音乐）mp3,mv爬虫，可以下载和收集mp3,mv信息,同时附带多媒体标签信息。采用python编写，mongo数据库(非必须)，递归算法核心实现\n\n[](LICENSE.md)\n\n\n\n[新手先看这里](SIMPLE_USE.md) | [历史版本在这里](VERSION.md) | [开发详情在这里](DEV.md)\n\n## 灵感来自\n\n[chengyumeng/spider163](https://github.com/chengyumeng/spider163)\n\n[Binaryify/NeteaseCloudMusicApi](https://github.com/Binaryify/NeteaseCloudMusicApi)\n\n[darknessomi/musicbox](https://github.com/darknessomi/musicbox)\n\n[sqaiyan/netmusic-node](https://github.com/sqaiyan/netmusic-node)\n\n## 功能特性\n1. 采用命令行形式运行，配置在个人账号目录下```~/.nxspider/```\n2. 搜索歌单，歌手，专辑，用户等信息\n3. 通过各种方法，如歌单，歌手，专辑等，下载mp3, mv\n4. 所有信息会保存在mongodb中（optional）\n5. 所有mp3, mv可以通过配置，增加多媒体标签信息(歌手，专辑，唱片，163comment!!!)\n\n## 运行依赖\n1. python3(推荐) or python2, 请配置好path\n2. windows or linux\n3. [mongodb](https://docs.mongodb.com/manual/installation/)(默认非必需)\n\n## Git clone安装\n $ git clone https://github.com/Grass-CLP/NXSpider.git && cd NXSpider\n $ python(3) setup.py install\n $ pip(3) install -r requirememts.txt # optional, if you use mongodb\n\n## 使用指南\n### 简单示例\n\t$ nxspider -h\n\t# 显示帮助信息\n\n\t$ nxspider -v\n\t# 显示软件版本信息\n\n### 配置, <>内容为值，or为可选\n\t$ nxspider config-spider -path \"\"\n\t# 强烈建议第一次前进行配置,强烈建议加上双引号保证空格等问题\n\t# 配置下载文件的路径，第一个路径为存储路径，其他路径会检查文件是否下载 \n\n\t$ nxspider config-mongo -mh -mp -mn \n\t# 配置mongodb服务器(非必需)，必须pip安装 requirements\n\n\t$ nxspider config-check\n\t# 检查配置内容\n\n### 搜索命令\n\t# 以下为搜索命令，其中<>内文具体的文字内容，可以输入中文，韩文，日语等(注意加上双引号)\n\t$ nxspider search -ar \"\" [-offset ] [-limit ]\n\t$ nxspider search -pl \"\" [-offset ] [-limit ]\n\t$ nxspider search -ur \"\" [-offset ] [-limit ]\n\t$ nxspider search -mp3 \"\" [-offset ] [-limit ]\n\t$ nxspider search -ab \"\" [-offset ] [-limit ]\n\t$ nxspider search -mv \"\" [-offset ] [-limit ]\n\n### 爬取命令\n\t$ nxspider sab -ab [-dw ]\n $ nxspider sab -ab 3084625,3107072 -dw mv,mp3\n\t# 根据专辑id(按逗号隔开)爬取歌曲，下载mp3,mv(参数配置),所有[]内为可选内容\n\n $ nxspider smp3s -mp3 [-dw ]\n\t# 根据歌曲id(按逗号隔开)爬取歌曲\n\n\t$ nxspider spls -pl [-dw ]\n\t# 根据歌单id(按逗号隔开)爬取歌曲\n\n\t$ nxspider sar-top-mp3 -ar [-dw ]\n\t# 根据歌手id爬取该歌手top50首\n\n\t$ nxspider sar-albums -ar [-dw ] [-offset ] [-limit ]\n\t# 根据歌手id爬取该歌手第个开始共个专辑，offset默认0，limit默认50\n\n\t$ nxspider sur-pls -ur [-dw ] [-offset ] [-limit ]\n\t# 根据用户id爬取该用户的所有歌单（包括收藏）\n\n\t$ nxspider scls-pls -cls [-dw ] [-offset ] [-limit ]\n\t$ nxspider scls-pls -cls 全部 -dw mv,mp3 -offset 0 -limit 20\n\t# 根据歌单分配，爬取该分类下最火歌单,并下载mp3和mv\n\n\t$ nxspider stop-mvs [-offset ] [-limit ]\n\t# 爬取最火mv\n\n### 其他命令\n\t$ nxspider sw_ur_pl\n\t# 显示所有歌单分类\n\n\t$ nxspider config-mongo -nomongo <1 or 0>\n\t# 配置是否使用mongo存储数据，默认不使用，1为不使用，0为使用\n\n\t$ nxspider config-spider -mvr <240 or 480 or 720 or 1080>\n\t# 配置下载mv最高的分辨率，默认720\n\n $ nxspider config-spider --shortcut <1 or 0>\n\t# 配置是否创建快捷方式（目前只创建歌单，歌手，唱片的快捷方式文件目录），默认0\n\t# 在windows下配置1的话需要右键\"以管理员身份运行\" start.bat\n\n\t$ nxspider config-spider -tag <1 or 0>\n\t# 配置是否为媒体文件增加媒体标签，默认1\n\n\t$ nxspider config-spider -tag163 <1 or 0>\n\t# 配置是否为媒体文件增加网易云音乐app特定信息媒体标签，默认1\n\n\t$ nxspider config-check\n\t# 检查配置正确性，并修复部分错误配置\n\n\t$ nxspider sw-ar-ab -ar [-offset ] [-limit ]\n\t# 通过歌手id打印该歌手的唱片信息\n\n\t$ nxspider login-smv -lu [-lp ]\n\t# 建议不使用-lp(会显示密码)，登录后爬取自己收藏的所有mv\n\n\t$ nxspider login-spls -lu [-lp ] [-dw ]\n\t# 建议不使用-lp(会显示密码)，登录后爬取自己所有歌单（包括最爱），下载mp3和mv（记得别忘输入 -dw mp3,mv）\n\n\t$ nxspider config_clear\n\t# 恢复所有设置到默认，建议执行后进行配置下载目录\n\n\t$ nxspider -h\n\t# 更多命令请参考这个\n\n### 部分命令效果图\n\n\n\n\n\n\n### 小tips\n- 默认配置须关心的为：不使用mongo + 下载路径为 `~/.nxspider/donwload_files/`\n- 使用mongo的话，**强烈要求**执行 `pip(3) install -r requirememts.txt`\n- 配置 ```-tag163 1``` 直接把下载目录添加到PC版某云软件的下载目录下，软件会自动识别已下载\n- 关闭爬取，目前代码是幂等运行，暂时建议直接 ```ctrl + c``` 强制关闭即可\n\n\n### 系统兼容\n| 系统及平台 | 结果 |\n| --- | --- |\n| win7(GBK) + python2 | 字符集问题 |\n| win7(GBK) + python3 | 通过 |\n| win10(GBK) + python2 | 字符集问题 |\n| win10(GBK) + python3 | 通过 |\n| centos7.2(utf8) + python2 | 通过 |\n| centos7.2(utf8) + python3 | 通过 |\n| mac + python | 待补充 |\n\n\n### 效果图示例\n- 已下载mv\n\n\n\n- 已下载MP3\n\n\n\n- 已采集数据\n\n\n\n## 最后\n- 程序关闭，建议直接关闭console或 `ctrl+c`\n- 希望能给我加个★\n- 有其他问题请到issue下留言，或者发邮件给我LipsonChan的雅虎邮箱\n- 提交issue的话建议先阅读 [Contributing](CONTRIBUTING.rst)"},{"alpha_fraction":0.654574990272522,"alphanum_fraction":0.659041702747345,"avg_line_length":26.261993408203125,"blob_id":"fbfb98575aad527e93f7b2ab2fd477c2736426a8","content_id":"5793123e37e81b8f114e745c96a909ac5a59ac3e","detected_licenses":["BSD-3-Clause"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":7388,"license_type":"permissive","max_line_length":72,"num_lines":271,"path":"/NXSpider/model/mongo_model.py","repo_name":"Z-Shuming/NXSpider","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# -*- coding: utf-8 -*-\n#\n# created by Lipson on 2018/4/19.\n# email to LipsonChan@yahoo.com\n#\nfrom datetime import datetime\n\nimport pymongo\nfrom mongoengine import connect, DynamicDocument, Document, signals\nfrom mongoengine.fields import *\nfrom pymongo.errors import ServerSelectionTimeoutError\n\nfrom NXSpider.common import log\nfrom NXSpider.common.config import Config\nmongodb_conf = Config().get_mongo()\n\ntry:\n\n client = pymongo.MongoClient(host=mongodb_conf['host'],\n port=mongodb_conf['port'],\n connectTimeoutMS=3000,\n serverSelectionTimeoutMS=3000)\n test_connect = client.database.test.count()\n del client\nexcept ServerSelectionTimeoutError as e:\n log.print_err(\"mongodb server config error\")\n exit()\n\nmodel_download_url = 'download_url'\nmodel_is_download = 'downloaded'\n\n\ndef field_value(field, value):\n \"\"\"\n Converts a supplied value to the type required by the field.\n If the field requires a EmbeddedDocument the EmbeddedDocument\n is created and updated using the supplied data.\n :param field:\n :param value:\n :return:\n \"\"\"\n if field.__class__ in (ListField, SortedListField):\n # return a list of the field values\n return [\n field_value(field.field, item)\n for item in value]\n\n elif field.__class__ in (\n EmbeddedDocumentField,\n GenericEmbeddedDocumentField,\n ReferenceField,\n GenericReferenceField,\n LazyReferenceField):\n\n if isinstance(value, Document):\n return value\n\n if isinstance(value, dict):\n embedded_doc = field.document_type()\n update_doc(embedded_doc, value)\n return embedded_doc\n return None\n else:\n return value\n\n\ndef update_doc(doc, data):\n \"\"\"\n Update an document to match the supplied dictionary.\n :param doc:\n :param data:\n :return:\n \"\"\"\n\n for key, value in data.items():\n if hasattr(doc, key):\n value = field_value(doc._fields[key], value)\n setattr(doc, key, value)\n else:\n # handle invalid key\n pass\n\n return doc\n\n\ndef update_doc_filter(doc, filter_set, *dicts):\n \"\"\"\n update document by filter and dicts\n :param doc:\n :param filter_set: Set\n :param dicts:\n :return:\n \"\"\"\n obj = {}\n for d in dicts:\n obj.update({k: v for k, v in d.items() if k in filter_set})\n\n return update_doc(doc, obj)\n\n\ndef update_dynamic_doc(doc, data):\n for key, value in data.items():\n if key in doc._fields:\n value = field_value(doc._fields[key], value)\n setattr(doc, key, value)\n\n\ndef update_dynamic_doc_filter(doc, filter_set, *dicts):\n \"\"\"\n update document by filter and dicts\n :param doc:\n :param filter_set: Set\n :param dicts:\n :return:\n \"\"\"\n obj = {}\n for d in dicts:\n obj.update({k: v for k, v in d.items() if k in filter_set})\n\n return update_dynamic_doc(doc, obj)\n\n\ndef update_dynamic_doc_filter(doc, filter_set, *dicts):\n \"\"\"\n update document by filter and dicts\n :param doc:\n :param filter_set: Set\n :param dicts:\n :type dicts: dict\n :type filter_set: list\n :return:\n :rtype: DynamicDocument | None\n \"\"\"\n obj = {}\n for d in dicts:\n obj.update({k: v for k, v in d.items() if k in filter_set})\n\n return update_dynamic_doc(doc, obj)\n\n\ndef update_dynamic_doc_rfilter(doc, rfilter_set, *dicts):\n \"\"\"\n update document by filter and dicts\n :param doc:\n :param rfilter_set: Set\n :param dicts:\n :type dicts: dict\n :type rfilter_set: Set\n :return:\n :rtype: DynamicDocument | None\n \"\"\"\n rfilter_set = {} if rfilter_set is None else rfilter_set\n obj = {}\n for d in dicts:\n obj.update({k: v for k, v in d.items() if k not in rfilter_set})\n\n return update_dynamic_doc(doc, obj)\n\n\ndef get_one_model_by_key(model, model_id):\n \"\"\"\n load or create a model by id\n :type model: DynamicDocument\n :param model:\n :param model_id:\n :return: doc, is_new\n :rtype: (DynamicDocument, boolean)\n \"\"\"\n\n try:\n res = model.objects(id=model_id).first()\n if res is None:\n res = model(id=model_id)\n return res, True\n return res, False\n except Exception as e:\n log.print_err('load a doc err: %s' % e)\n return None, True\n\n\ndef update_dynamic_doc(doc, data):\n for key, value in data.items():\n if key in doc._fields:\n value = field_value(doc._fields[key], value)\n setattr(doc, key, value)\n\n\nclass ConfigModel(DynamicDocument):\n id = StringField(primary_key=True)\n updated_at = DateTimeField(default=datetime.utcnow)\n pass\n\n\nclass AuthorModel(DynamicDocument):\n id = LongField(primary_key=True)\n updated_at = DateTimeField(default=datetime.utcnow)\n pass\n\n\nclass Mp3Model(DynamicDocument):\n id = LongField(primary_key=True)\n artists = ListField(ReferenceField('ArtistModel'))\n album = ReferenceField('AlbumModel')\n mv = ReferenceField('Mp4Model')\n updated_at = DateTimeField(default=datetime.utcnow)\n pass\n\n\nclass ArtistModel(DynamicDocument):\n id = LongField(primary_key=True)\n # mp3s = ListField(ReferenceField(Mp3Model))\n updated_at = DateTimeField(default=datetime.utcnow)\n\n\nclass Mp4Model(DynamicDocument):\n id = LongField(primary_key=True)\n artists = ListField(ReferenceField(ArtistModel))\n updated_at = DateTimeField(default=datetime.utcnow)\n pass\n\n\nclass VideoModel(DynamicDocument):\n id = StringField(primary_key=True)\n artists = ListField(ReferenceField(ArtistModel))\n updated_at = DateTimeField(default=datetime.utcnow)\n pass\n\n\nclass PlaylistModel(DynamicDocument):\n id = LongField(primary_key=True)\n mp3s = ListField(LazyReferenceField(Mp3Model))\n mp4s = ListField(LazyReferenceField(Mp4Model))\n updated_at = DateTimeField(default=datetime.utcnow)\n\n\nclass AlbumModel(DynamicDocument):\n id = LongField(primary_key=True)\n mp3s = ListField(LazyReferenceField(Mp3Model))\n artists = ListField(LazyReferenceField(ArtistModel))\n updated_at = DateTimeField(default=datetime.utcnow)\n\n\nclass UserModel(DynamicDocument):\n id = LongField(primary_key=True)\n # friends = ListField(ReferenceField('self'))\n updated_at = DateTimeField(default=datetime.utcnow)\n\n\nclass Downloaded(DynamicDocument):\n model = StringField(required=True)\n model_id = LongField(required=True)\n updated_at = DateTimeField(default=datetime.utcnow)\n\n\ndef update_timestamp(sender, document, **kwargs):\n document.updated_at = datetime.utcnow()\n\n\nsignals.pre_save.connect(update_timestamp, sender=PlaylistModel)\nsignals.pre_save.connect(update_timestamp, sender=VideoModel)\nsignals.pre_save.connect(update_timestamp, sender=Mp4Model)\nsignals.pre_save.connect(update_timestamp, sender=Mp3Model)\nsignals.pre_save.connect(update_timestamp, sender=AuthorModel)\nsignals.pre_save.connect(update_timestamp, sender=ConfigModel)\nsignals.pre_save.connect(update_timestamp, sender=UserModel)\nsignals.pre_save.connect(update_timestamp, sender=ArtistModel)\nsignals.pre_save.connect(update_timestamp, sender=AlbumModel)\n\nconnect(mongodb_conf['name'], host=mongodb_conf['host'],\n port=mongodb_conf['port'], connectTimeoutMS=3000)\n"}],"string":"[\n {\n \"alpha_fraction\": 0.5572916865348816,\n \"alphanum_fraction\": 0.625,\n \"avg_line_length\": 13.769230842590332,\n \"blob_id\": \"027b9a5be43029169a1f49de2046171a372c7ddf\",\n \"content_id\": \"19bce644bfaed95012b5d8f501f37dc2ed6af270\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 192,\n \"license_type\": \"permissive\",\n \"max_line_length\": 35,\n \"num_lines\": 13,\n \"path\": \"/NXSpider/debug_test/test.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/26.\\n# email to LipsonChan@yahoo.com\\n#\\n\\nfrom NXSpider.spider.mp3 import Mp3\\n\\nm1=Mp3(__parse_recursion__={})\\nm=Mp3()\\n\\nprint('a')\\n\"\n },\n {\n \"alpha_fraction\": 0.5741106867790222,\n \"alphanum_fraction\": 0.5839921236038208,\n \"avg_line_length\": 34.30232620239258,\n \"blob_id\": \"c5f018043076de0f21cb1cc885ece28ee375d3b6\",\n \"content_id\": \"24d33f857c7a2f47650656a4986dada0d36e2c6a\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3036,\n \"license_type\": \"permissive\",\n \"max_line_length\": 119,\n \"num_lines\": 86,\n \"path\": \"/NXSpider/bin/show_ctrl.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/6/19.\\n# email to LipsonChan@yahoo.com\\n#\\nimport sys\\n\\nfrom cement.core.controller import expose\\nfrom terminaltables import AsciiTable\\n\\nfrom NXSpider.bin.base_ctrl import NXSpiderBaseController, py2_decoding\\nfrom NXSpider.bin.print_as_table import print_mp3s, print_playlist, print_users, print_albums, print_artists, print_mvs\\nfrom NXSpider.common import PYTHON2, log\\nfrom NXSpider.spider import api\\nfrom NXSpider.spider.api import search_types, PLAYLIST_CLASSES\\n\\nPRINT_ATTR_FUNC_MAP = {\\n 'mp3': ['songs', print_mp3s],\\n 'playlist': ['playlists', print_playlist],\\n 'user': ['userprofiles', print_users],\\n 'artist': ['artists', print_artists],\\n 'album': ['albums', print_albums],\\n 'mv': ['mvs', print_mvs],\\n}\\n\\n\\nclass ShowController(NXSpiderBaseController):\\n class Meta:\\n label = \\\"show\\\"\\n stacked_on = 'base'\\n description = \\\"NXSpider\\\"\\n\\n @expose(help=\\\"search [-ar ] [-pl ] \\\"\\n \\\"[-ur ] [-mp3 ] [-ab ] [-mv ]\\\")\\n def search(self):\\n search_key = 'mp3'\\n key_num = 0\\n for k, v in search_types.items():\\n if getattr(self.app.pargs, k, None):\\n search_key = k\\n key_num += 1\\n if key_num > 1:\\n log.print_err(\\\"it could search by only one type\\\")\\n\\n # input must be decode in python2\\n search_value = getattr(self.app.pargs, search_key)\\n search_value = py2_decoding(search_value)\\n\\n res = api.search(search_value, stype=search_key,\\n offset=self.app.pargs.offset or 0,\\n limit=self.app.pargs.limit or 50)\\n\\n if not res:\\n log.print_info(\\\"nothing found!\\\")\\n return\\n\\n if search_key in PRINT_ATTR_FUNC_MAP:\\n func = PRINT_ATTR_FUNC_MAP[search_key][1] # type: function\\n value = (res.get(PRINT_ATTR_FUNC_MAP[search_key][0], [])) # type: list\\n func(value)\\n\\n @expose(help=\\\"show artists ablum, usage: sw-ar-ab -ar [-offset ] [-limit ]\\\")\\n def sw_ar_ab(self):\\n if self.param_check(['artist'], sys._getframe().f_code.co_name) is False:\\n return\\n\\n artistid = self.app.pargs.artist # type: list\\n artist_detail = api.get_artist_album(artistid,\\n offset=self.app.pargs.offset or 0,\\n limit=self.app.pargs.limit or 50)\\n\\n print_albums(artist_detail['hotAlbums'])\\n pass\\n\\n @expose(help=\\\"show user playlist, usage: sw-ur-pl -ur \\\")\\n def sw_ur_pl(self):\\n # todo\\n pass\\n\\n @expose(help=\\\"show all playlist classes, usage: sw-pl-classes\\\")\\n def sw_pl_classes(self):\\n table = AsciiTable([[\\\"Group\\\", \\\"Classes(which can be as a input)\\\"]])\\n table_data = [[k, ', '.join(v)] for k, v in PLAYLIST_CLASSES.items()]\\n table.table_data.extend(table_data)\\n print(table.table)\\n\"\n },\n {\n \"alpha_fraction\": 0.6136986017227173,\n \"alphanum_fraction\": 0.6356164216995239,\n \"avg_line_length\": 15.590909004211426,\n \"blob_id\": \"701e757e158767823c9594c3083424df6fd2f639\",\n \"content_id\": \"1304c1c7d11f468089209fc287f77c8c06c59d0e\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 365,\n \"license_type\": \"permissive\",\n \"max_line_length\": 42,\n \"num_lines\": 22,\n \"path\": \"/NXSpider/common/log.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/19.\\n# email to LipsonChan@yahoo.com\\n#\\n\\nfrom colorama import Fore\\nfrom colorama import init\\ninit(autoreset=True)\\n\\n\\ndef print_err(msg):\\n print(Fore.RED + \\\"ERROR: \\\" + msg)\\n\\n\\ndef print_warn(msg):\\n print(Fore.YELLOW + \\\"WARNING: \\\" + msg)\\n\\n\\ndef print_info(msg):\\n print(Fore.GREEN + \\\"INFO: \\\" + msg)\\n\"\n },\n {\n \"alpha_fraction\": 0.5047619342803955,\n \"alphanum_fraction\": 0.5809524059295654,\n \"avg_line_length\": 16.16666603088379,\n \"blob_id\": \"4eaae03255274fb51cbfa6f2850b7d21fca58981\",\n \"content_id\": \"aeae71ab6357de4e2205afdda64a00e9943a049a\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 105,\n \"license_type\": \"permissive\",\n \"max_line_length\": 33,\n \"num_lines\": 6,\n \"path\": \"/NXSpider/spider/__init__.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/19.\\n# email to LipsonChan@yahoo.com\\n#\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5492264628410339,\n \"alphanum_fraction\": 0.5724331736564636,\n \"avg_line_length\": 22.700000762939453,\n \"blob_id\": \"ebc2a604a7f6c68d5488ad84a11f8c9ae271a78c\",\n \"content_id\": \"4c5a603565b92ea36bbf1b9c1daa670330e06022\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1422,\n \"license_type\": \"permissive\",\n \"max_line_length\": 72,\n \"num_lines\": 60,\n \"path\": \"/setup.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/5/21.\\n# email to LipsonChan@yahoo.com\\n#\\nimport os\\n\\nfrom setuptools import setup, find_packages, Command\\n\\nfrom NXSpider import version\\n\\n\\nclass CleanCommand(Command):\\n \\\"\\\"\\\"Custom clean command to tidy up the project root.\\\"\\\"\\\"\\n user_options = []\\n\\n def initialize_options(self):\\n pass\\n\\n def finalize_options(self):\\n pass\\n\\n def run(self):\\n os.system('rm -vrf ./build ./dist ./*.pyc ./*.tgz ./*.egg-info')\\n\\n\\nsetup(\\n name=version.__title__,\\n version=version.__version__,\\n author=version.__author__,\\n author_email=version.__author_email__,\\n url=version.__url__,\\n description=version.__description__,\\n license=version.__license__,\\n packages=find_packages(),\\n include_package_data=True,\\n install_requires=[\\n \\\"mutagen>=1.40,<2\\\",\\n \\\"terminaltables>=3,<4\\\",\\n \\\"cryptography>=2.2,<3\\\",\\n \\\"requests>=2.19,<3\\\",\\n \\\"colorama>=0.3,<0.4\\\",\\n \\\"cement>=2,<3\\\",\\n \\\"six>=1,<=2\\\",\\n ],\\n entry_points={\\n \\\"console_scripts\\\": [\\\"nxspider=NXSpider.bin.cli:main\\\"]\\n },\\n classifiers=[\\n 'Environment :: Console',\\n 'Environment :: Web Environment',\\n 'Programming Language :: Python :: 2.7',\\n 'Programming Language :: Python :: 3.6',\\n ],\\n keywords=['music', 'netease', 'cli', 'spider'],\\n cmdclass={\\n 'clean': CleanCommand,\\n },\\n)\\n\"\n },\n {\n \"alpha_fraction\": 0.5513054132461548,\n \"alphanum_fraction\": 0.6836673021316528,\n \"avg_line_length\": 24.734375,\n \"blob_id\": \"3ec5d3668e67601c71558775ce639cf292a2b580\",\n \"content_id\": \"57c9b08ea340aa270d7ae8179ec0c1c27b83815e\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1647,\n \"license_type\": \"permissive\",\n \"max_line_length\": 270,\n \"num_lines\": 64,\n \"path\": \"/NXSpider/spider/common_keys.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/16.\\n# email to LipsonChan@yahoo.com\\n#\\n\\nimport json\\nimport os\\n\\nfrom NXSpider.common import encrypt\\nfrom NXSpider.common import tools, log\\n\\ncsrf_dict = {'csrf_token': 'csrf'}\\n\\n\\ndef encrypted_request(obj):\\n data = {\\n 'params': create_params_by_dict(obj),\\n 'encSecKey': encSecKey\\n }\\n return data\\n\\n\\ndef create_params_by_dict(obj):\\n try:\\n return create_params_text(json.dumps(obj))\\n except Exception as e:\\n log.print_err('create params error: %s' % e)\\n return None\\n\\n\\ndef create_params_text(text):\\n nonce = '0CoJUm6Qyw8W8jud'\\n nonce2 = 16 * 'F'\\n encText0 = encrypt.aes(text, nonce).decode(\\\"utf-8\\\")\\n encText = encrypt.aes(encText0, nonce2)\\n return encText\\n\\n\\ndef decrpyt_params(text):\\n nonce = '0CoJUm6Qyw8W8jud'\\n nonce2 = 16 * 'F'\\n decText0 = encrypt.aes_decode(text, nonce2)\\n decText = encrypt.aes_decode(decText0, nonce)\\n return decText\\n\\n\\ndef rsa_encrypt(text, pubKey, modulus):\\n text = text[::-1]\\n rs = int(tools.hex(text), 16) ** int(pubKey, 16) % int(modulus, 16)\\n return format(rs, 'x').zfill(256)\\n\\n\\ndef create_secretKey(size):\\n return (\\n ''.join(map(lambda xx: (hex(ord(xx))[2:]), os.urandom(size)))\\n )[0:16]\\n\\n\\nmodulus = '00e0b509f6259df8642dbc35662901477df22677ec152b5ff68ace615bb7b725152b3ab17a876aea8a5aa76d2e417629ec4ee341f56135fccf695280104e0312ecbda92557c93870114af6c9d05c4f7f0c3685b7a46bee255932575cce10b424d813cfe4875d3e82047b97ddef52741d546b8e289dc6935b3ece0462db0a22b8e7'\\npubKey = '010001'\\nsecKey = 16 * 'F'\\nencSecKey = rsa_encrypt(secKey, pubKey, modulus)\\n\"\n },\n {\n \"alpha_fraction\": 0.5777778029441833,\n \"alphanum_fraction\": 0.6083333492279053,\n \"avg_line_length\": 24.714284896850586,\n \"blob_id\": \"d038133f9a9151dcfaa23e5a68c544f417367b08\",\n \"content_id\": \"7e0b12770a6474b077c6a8b93230dac0978b95e3\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 360,\n \"license_type\": \"permissive\",\n \"max_line_length\": 69,\n \"num_lines\": 14,\n \"path\": \"/NXSpider/version.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/6/6.\\n# email to LipsonChan@yahoo.com\\n#\\n\\n__title__ = 'NXSpider'\\n__version__ = '0.0.5'\\n__description__ = 'NetEase X Spider, help u to have a better life!!!'\\n__url__ = 'https://github.com/Grass-CLP/NXSpider'\\n__author__ = 'LipsonChan'\\n__author_email__ = 'LipsonChan@yahoo.com'\\n__license__ = 'BSD-3-Clause License'\\n\"\n },\n {\n \"alpha_fraction\": 0.6242424249649048,\n \"alphanum_fraction\": 0.6393939256668091,\n \"avg_line_length\": 19.030303955078125,\n \"blob_id\": \"9cd1a9375c1b6377c023880130429c51b8828c16\",\n \"content_id\": \"252eac29cf6a940c55a0a5f6cdbb531606d0dd7d\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 660,\n \"license_type\": \"permissive\",\n \"max_line_length\": 44,\n \"num_lines\": 33,\n \"path\": \"/NXSpider/model/export.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/6/20.\\n# email to LipsonChan@yahoo.com\\n#\\n\\nfrom NXSpider.common.config import Config\\nif Config().get_no_mongo():\\n from NXSpider.model.dict_model import *\\nelse:\\n from NXSpider.model.mongo_model import *\\n\\n\\nmodel_download_url = 'download_url'\\nmodel_is_download = 'downloaded'\\nmodel_download_path = 'download_path'\\n\\n__all__ = [\\n 'ConfigModel',\\n 'UserModel',\\n 'AlbumModel',\\n 'PlaylistModel',\\n 'Mp4Model',\\n 'VideoModel',\\n 'ArtistModel',\\n 'Mp3Model',\\n 'AuthorModel',\\n 'update_dynamic_doc',\\n 'model_download_url',\\n 'model_is_download',\\n 'get_one_model_by_key',\\n]\"\n },\n {\n \"alpha_fraction\": 0.569031298160553,\n \"alphanum_fraction\": 0.5827612280845642,\n \"avg_line_length\": 19.484375,\n \"blob_id\": \"548c7290f0283b25f877b06731245f14bd4f24b6\",\n \"content_id\": \"b3a8dea9e934ac6c937dd65fef4949dbad40d883\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1311,\n \"license_type\": \"permissive\",\n \"max_line_length\": 53,\n \"num_lines\": 64,\n \"path\": \"/NXSpider/bin/models.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/6/19.\\n# email to LipsonChan@yahoo.com\\n#\\nfrom NXSpider.spider.album import Album\\nfrom NXSpider.spider.artist import Artist\\nfrom NXSpider.spider.mp3 import Mp3\\nfrom NXSpider.spider.mv import MV\\nfrom NXSpider.spider.playlist import Playlist\\nfrom NXSpider.spider.user import User\\n\\nno_rec_artist_mo = Artist(\\n __parse_recursion__={}\\n)\\n\\nno_rec_album_mo = Album(\\n __model_rfilter__={'artist', 'songs'},\\n __parse_recursion__={'artists': no_rec_artist_mo}\\n)\\n\\nno_rec_mv_mo = MV(\\n __parse_recursion__={'artists': no_rec_artist_mo}\\n)\\n\\ndw_mp3_mo = Mp3(\\n __parse_recursion__={\\n 'artists': no_rec_artist_mo,\\n 'album': no_rec_album_mo,\\n 'mv': no_rec_mv_mo,\\n }\\n)\\n\\nplaylist_mo = Playlist(\\n __model_rfilter__={'artist'},\\n __parse_recursion__={\\n 'mp3': dw_mp3_mo,\\n 'creator': User(),\\n }\\n)\\n\\nartist_mo = Artist(\\n __model_rfilter__={'artist'},\\n __parse_recursion__={\\n 'mp3': dw_mp3_mo,\\n }\\n)\\n\\nalbum_mo = Album(\\n __model_rfilter__={'artist'},\\n __parse_recursion__={\\n 'mp3': dw_mp3_mo,\\n 'artists': no_rec_artist_mo,\\n }\\n)\\n\\nartist_album_mo = Artist(\\n __model_rfilter__={},\\n __parse_recursion__={\\n 'albums': album_mo,\\n 'artists': no_rec_artist_mo,\\n }\\n)\\n\"\n },\n {\n \"alpha_fraction\": 0.535625696182251,\n \"alphanum_fraction\": 0.5718929171562195,\n \"avg_line_length\": 25.75286102294922,\n \"blob_id\": \"76d63045396b147e02afa192bca9e5e0fc41e2b3\",\n \"content_id\": \"c809ebc0f4c9702b46590bd142e9bbc7447792e2\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 12299,\n \"license_type\": \"permissive\",\n \"max_line_length\": 141,\n \"num_lines\": 437,\n \"path\": \"/NXSpider/spider/api.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/5/24.\\n# email to LipsonChan@yahoo.com\\n#\\nimport re\\nfrom collections import OrderedDict\\nfrom functools import reduce\\n\\nimport requests\\n\\nfrom NXSpider.common import log\\nfrom NXSpider.spider.common_keys import encrypted_request\\n\\nbase_url = \\\"http://music.163.com\\\"\\nbase_https_url = \\\"https://music.163.com\\\"\\n\\nheaders = {\\n 'Referer': 'http://music.163.com/',\\n 'Host': 'music.163.com',\\n 'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_2) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/55.0.2883.95 Safari/537.36',\\n 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8',\\n}\\n\\nTOP_LIST_ALL = {\\n 0: ['云音乐新歌榜', '3779629'],\\n 1: ['云音乐热歌榜', '3778678'],\\n 2: ['网易原创歌曲榜', '2884035'],\\n 3: ['云音乐飙升榜', '19723756'],\\n 4: ['云音乐电音榜', '10520166'],\\n 5: ['UK排行榜周榜', '180106'],\\n 6: ['美国Billboard周榜', '60198'],\\n 7: ['KTV嗨榜', '21845217'],\\n 8: ['iTunes榜', '11641012'],\\n 9: ['Hit FM Top榜', '120001'],\\n 10: ['日本Oricon周榜', '60131'],\\n 11: ['韩国Melon排行榜周榜', '3733003'],\\n 12: ['韩国Mnet排行榜周榜', '60255'],\\n 13: ['韩国Melon原声周榜', '46772709'],\\n 14: ['中国TOP排行榜(港台榜)', '112504'],\\n 15: ['中国TOP排行榜(内地榜)', '64016'],\\n 16: ['香港电台中文歌曲龙虎榜', '10169002'],\\n 17: ['华语金曲榜', '4395559'],\\n 18: ['中国嘻哈榜', '1899724'],\\n 19: ['法国 NRJ EuroHot 30周榜', '27135204'],\\n 20: ['台湾Hito排行榜', '112463'],\\n 21: ['Beatport全球电子舞曲榜', '3812895'],\\n 22: ['云音乐ACG音乐榜', '71385702'],\\n 23: ['云音乐嘻哈榜', '991319590']\\n}\\n\\nPLAYLIST_CLASSES = OrderedDict([\\n ('语种', ['华语', '欧美', '日语', '韩语', '粤语', '小语种']),\\n ('风格', ['流行', '摇滚', '民谣', '电子', '舞曲', '说唱', '轻音乐', '爵士', '乡村', 'R&B/Soul', '古典', '民族', '英伦', '金属', '朋克', '蓝调', '雷鬼',\\n '世界音乐', '拉丁', '另类/独立', 'New Age', '古风', '后摇', 'Bossa Nova']),\\n ('场景', ['清晨', '夜晚', '学习', '工作', '午休', '下午茶', '地铁', '驾车', '运动', '旅行', '散步', '酒吧']),\\n ('情感', ['怀旧', '清新', '浪漫', '性感', '伤感', '治愈', '放松', '孤独', '感动', '兴奋', '快乐', '安静', '思念']),\\n ('主题', ['影视原声', 'ACG', '儿童', '校园', '游戏', '70后', '80后', '90后', '网络歌曲', 'KTV', '经典', '翻唱', '吉他', '钢琴', '器乐', '榜单', '00后'])\\n])\\n\\nALL_CLASSES = reduce(lambda x, y: x + y, [v for k, v in PLAYLIST_CLASSES.items()])\\n\\nMV_TYPE = ['ALL', 'ZH', 'EA', 'KR', 'JP']\\n\\n\\n# 搜索单曲(1)，歌手(100)，专辑(10)，歌单(1000)，用户(1002) *(type)*\\nsearch_types = {\\n 'mp3': 1,\\n 'artist': 100,\\n 'album': 10,\\n 'playlist': 1000,\\n 'user': 1002,\\n 'mv': 1004,\\n}\\n\\n\\ndef api_request(url, data=None, method=\\\"get\\\", json=True,\\n session=None, headers=headers, encrypt=True, https=False):\\n \\\"\\\"\\\"\\n request and try\\n :param https:\\n :param encrypt:\\n :param url:\\n :param data:\\n :param method:\\n :param json:\\n :param session:\\n :type session: requests.Session\\n :param headers:\\n :return:\\n \\\"\\\"\\\"\\n url = base_https_url + url if https else base_url + url\\n request_obj = session or requests\\n\\n # update cookies\\n if isinstance(request_obj, requests.Session):\\n for cookie in request_obj.cookies:\\n if cookie.name == '__csrf':\\n data['csrf_token'] = cookie.value\\n break\\n\\n # encrypt\\n if encrypt:\\n data = encrypted_request(data)\\n\\n method = 'get' if not data and method == 'get' else 'post'\\n request_method = getattr(request_obj, method, None) or request_obj.get\\n try:\\n req = request_method(url, data=data, headers=headers, timeout=10)\\n req.encoding = \\\"UTF-8\\\"\\n res = req.json() if json else req.text\\n # if session:\\n # session.cookies.save()\\n return res\\n except ValueError as e:\\n log.print_err(\\\"api do not return a valuable json\\\")\\n return {}\\n except requests.exceptions.RequestException as e:\\n log.print_warn(\\\"request error: %s\\\" % url)\\n return {}\\n\\n\\ndef get_top_songlist(idx=0, offset=0):\\n action = TOP_LIST_ALL[idx][1]\\n res = api_request(action, json=False)\\n if not res:\\n return None\\n\\n songids = re.findall(r'/song\\\\?id=(\\\\d+)', res)\\n if not songids:\\n return None\\n\\n songids = list(set(songids))\\n details = get_mp3_details(songids, offset=offset)\\n return details\\n\\n\\ndef get_mp3_link(song_id):\\n # obj = '{\\\"ids\\\":[' + str(song_id) + '], br:\\\"320000\\\",csrf_token:\\\"csrf\\\"}'\\n data = {'ids': [song_id], 'br': 320000, 'csrf_token': 'csrf'}\\n url = \\\"/weapi/song/enhance/player/url\\\"\\n res = api_request(url, data)\\n if res and res['code'] == 200:\\n return res['data'][0]['url']\\n\\n\\ndef get_mp3_links(song_ids):\\n # obj = '{\\\"ids\\\":[' + str(song_id) + '], br:\\\"320000\\\",csrf_token:\\\"csrf\\\"}'\\n data = {'ids': song_ids, 'br': 320000, 'csrf_token': 'csrf'}\\n url = \\\"/weapi/song/enhance/player/url\\\"\\n res = api_request(url, data)\\n if res and res['code'] == 200:\\n return {x['id']: x['url'] for x in res['data']}\\n\\n\\ndef get_mp3_details(song_ids, offset=0):\\n tmpids = song_ids[offset:]\\n tmpids = tmpids[0:100]\\n tmpids = list(map(str, tmpids))\\n action = '/api/song/detail?ids=[{}]'.format( # NOQA\\n ','.join(tmpids))\\n\\n res = api_request(action)\\n if res and res['code'] == 200:\\n return {x['id']: x for x in res['songs']}\\n\\n\\ndef get_mv_link(mv_id, r):\\n data = {'id': mv_id, 'r': r, 'csrf_token': 'csrf'}\\n url = \\\"/weapi/song/enhance/download/mv/url\\\"\\n\\n res = api_request(url, data)\\n if res and res['code'] == 200:\\n return res['data']['url']\\n\\n\\ndef get_video_link(vid, r):\\n data = {'ids': [vid], 'resolution': r, 'csrf_token': ''}\\n url = \\\"/weapi/cloudvideo/playurl\\\"\\n\\n res = api_request(url, data)\\n if res and res['code'] == 200:\\n return res['urls'][0]['url']\\n\\n\\ndef get_video_links(vids, r):\\n data = {'ids': vids, 'resolution': r, 'csrf_token': ''}\\n url = \\\"/weapi/cloudvideo/playurl\\\"\\n\\n res = api_request(url, data)\\n if res and res['code'] == 200:\\n return {x['id']: x['url'] for x in res['data']}\\n\\n\\ndef get_mv_detail(mvid):\\n data = {'id': mvid, 'csrf_token': 'csrf'}\\n url = \\\"/weapi/v1/mv/detail\\\"\\n\\n res = api_request(url, data)\\n return res.get('data', None)\\n\\n\\ndef get_video_detail(vid):\\n data = {'id': vid, 'csrf_token': 'csrf'}\\n url = \\\"/weapi/cloudvideo/v1/video/detail\\\"\\n\\n res = api_request(url, data)\\n return res.get('data', None)\\n\\n\\ndef get_playlist_detail(playlist_id):\\n url = \\\"/api/playlist/detail?id={}&upd\\\" \\\\\\n .format(playlist_id)\\n\\n res = api_request(url)\\n return res.get('result', None)\\n\\n\\ndef get_playlist_detail_v3(id):\\n action = '/weapi/v3/playlist/detail'\\n data = {'id': id, 'total': 'true', 'csrf_token': 'csrf', 'limit': 1000, 'n': 1000, 'offset': 0}\\n\\n res = api_request(action, data)\\n return res.get('playlist', None)\\n\\n\\ndef get_top_playlists(category='全部', order='hot', offset=0, limit=50):\\n \\\"\\\"\\\"\\n get playlists, but not detail\\n :param category:\\n :param order:\\n :param offset:\\n :param limit:\\n :return:\\n \\\"\\\"\\\"\\n action = u'/api/playlist/list?cat={}&order={}&offset={}&total={}&limit={}'.format( # NOQA\\n category, order, offset, 'true' if offset else 'false',\\n limit) # NOQA\\n\\n res = api_request(action)\\n return res.get('playlists', None)\\n\\n\\n# may be not useful\\ndef get_playlist_classes():\\n action = '/weapi/playlist/catalogue'\\n res = api_request(action, json=False)\\n if res and res['code'] == 200:\\n return re\\n\\n\\ndef get_playlist_catelogs():\\n path = '/weapi/playlist/catalogue'\\n return api_request(path, json=False)\\n\\n\\ndef top_artists(offset=0, limit=100):\\n action = '/api/artist/top?offset={}&total=false&limit={}'.format( # NOQA\\n offset, limit)\\n res = api_request(action)\\n return res.get('artists', None)\\n\\n\\ndef get_artists_songs(artist_id):\\n action = '/api/artist/{}'.format(artist_id)\\n res = api_request(action)\\n if res and res['code'] == 200:\\n return res\\n\\n\\ndef get_artist_album(artist_id, offset=0, limit=50):\\n action = '/api/artist/albums/{}?offset={}&limit={}'.format(\\n artist_id, offset, limit)\\n res = api_request(action)\\n if res and res['code'] == 200:\\n return res\\n # if res and res['code'] == 200:\\n # return res['hotAlbums']\\n # return res.get('hotAlbums', None)\\n\\n\\ndef get_album_detail(album_id):\\n action = '/api/album/{}'.format(album_id)\\n res = api_request(action)\\n return res.get('album', None)\\n\\n\\ndef search(s, stype=1, offset=0, total='true', limit=60):\\n if isinstance(stype, str):\\n if stype in search_types:\\n stype = search_types[stype]\\n else:\\n return None\\n\\n action = '/api/search/get'\\n data = {\\n 's': s,\\n 'type': stype,\\n 'offset': offset,\\n 'total': total,\\n 'limit': limit\\n }\\n res = api_request(action, data, encrypt=False)\\n return res.get('result', None)\\n\\n\\n# has been change, fuck!\\ndef user_playlist(uid, session=None, offset=0, limit=50):\\n action = '/weapi/user/playlist'\\n data = {'uid': uid, 'csrf_token': 'csrf',\\n 'limit': limit, 'offset': offset,\\n 'wordwrap': 7}\\n\\n res = api_request(action, data=data, session=session)\\n return res.get('playlist', None)\\n\\n\\ndef user_playlist_old(uid, offset=0, limit=100, session=None):\\n action = '/api/user/playlist/?offset={}&limit={}&uid={}'.format( # NOQA\\n offset, limit, uid)\\n\\n res = api_request(action, session=session)\\n return res.get('playlist', None)\\n\\n\\ndef my_subcount(session):\\n action = '/weapi/subcount'\\n res = api_request(action, data={}, session=session)\\n\\n return res\\n\\n\\ndef my_mvs(session):\\n action = '/weapi/mv/sublist'\\n data = dict(\\n offset=0,\\n limit=1000,\\n )\\n res = api_request(action, data=data, session=session)\\n\\n return res.get('data', [])\\n\\n\\ndef my_videos(session):\\n action = '/weapi/cloudvideo/allvideo/sublist'\\n data = dict(\\n offset=0,\\n limit=1000,\\n )\\n res = api_request(action, data=data, session=session)\\n\\n return res.get('data', [])\\n\\n\\ndef hot_mvs():\\n # todo, wait to find out how to encrypt and decrypt\\n action = '/api/mv/toplist'\\n action = '/api/mv/first'\\n action = '/api/mv/hot' # well done\\n\\n data = dict(\\n cat=u'内地',\\n order='hot',\\n offset=0,\\n limit=50,\\n )\\n\\n res = api_request(action, data=data)\\n return res\\n\\n\\ndef top_mvs(offset=0, limit=50):\\n action = '/weapi/mv/toplist'\\n data = dict(\\n offset=offset,\\n limit=limit,\\n )\\n\\n res = api_request(action, data=data)\\n return res.get('data', [])\\n\\n\\ndef all_mvs(offset=0, limit=50):\\n action = '/weapi/mv/all'\\n data = dict(\\n offset=offset,\\n limit=limit,\\n )\\n res = api_request(action, data=data)\\n return res.get('data', [])\\n\\n\\ndef phone_login(username, password, session):\\n \\\"\\\"\\\"\\n\\n :param username:\\n :param password: must be encrypt by md5 hex\\n :param session:\\n :type session: requests.Session\\n :return:\\n \\\"\\\"\\\"\\n action = '/weapi/login/cellphone'\\n data = {\\n 'phone': username,\\n 'password': password,\\n 'rememberLogin': 'true'\\n }\\n\\n res = api_request(action, data=data, session=session)\\n if res:\\n return res\\n\\n\\ndef login(username, password, session):\\n \\\"\\\"\\\"\\n :type username: str\\n :param username:\\n :param password: must be encrypt by md5 hex\\n :param session:\\n :type session: requests.Session\\n :return:\\n \\\"\\\"\\\"\\n if username.isdigit():\\n return phone_login(username, password, session)\\n action = '/weapi/login'\\n client_token = '1_jVUMqWEPke0/1/Vu56xCmJpo5vP1grjn_SOVVDzOc78w8OKLVZ2JH7IfkjSXqgfmh'\\n session.cookies.load()\\n data = {\\n 'username': username,\\n 'password': password,\\n 'rememberLogin': 'true',\\n 'clientToken': client_token,\\n }\\n res = api_request(action, data=data, session=session, https=True)\\n if res:\\n return res\\n\"\n },\n {\n \"alpha_fraction\": 0.4895104765892029,\n \"alphanum_fraction\": 0.4979020953178406,\n \"avg_line_length\": 27.039215087890625,\n \"blob_id\": \"4653ea5fad0727c600d18d60c33443c5b781692a\",\n \"content_id\": \"f065e805a7171a37d66075a72a457f183d3e1cf1\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2860,\n \"license_type\": \"permissive\",\n \"max_line_length\": 94,\n \"num_lines\": 102,\n \"path\": \"/NXSpider/spider/mp3.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/23.\\n# email to LipsonChan@yahoo.com\\n#\\n\\nimport re\\n\\nimport os\\n\\nfrom NXSpider.spider.album import Album\\nfrom NXSpider.spider.api import get_mp3_link, get_mv_detail\\nfrom NXSpider.spider.artist import Artist\\nfrom NXSpider.spider.mv import MV\\nfrom NXSpider.spider.base_driver import *\\n\\n\\nclass Mp3(Music163Obj):\\n __model_name__ = Mp3Model\\n __model_rfilter__ = {\\n 'artists', # read from album model\\n }\\n __parse_recursion__ = {\\n 'artists': Artist(),\\n 'album': Album(__model_rfilter__={'artist','songs'}),\\n 'mv': MV(),\\n }\\n\\n @attr_replace(attr_name='mvid', new_name='mv')\\n def replace_mvid(self, obj):\\n if obj != 0:\\n obj = get_mv_detail(obj)\\n return obj\\n return None\\n\\n def pre_save(self, doc, obj):\\n \\\"\\\"\\\"\\n :param doc:\\n :param obj:\\n :type doc: Mp3Model\\n :return:\\n \\\"\\\"\\\"\\n # set artists\\n if getattr(doc, 'album', False) and getattr(doc['album'], 'artists', False):\\n doc.artists = [a for a in doc['album']['artists']]\\n\\n def download_filename(self, doc):\\n \\\"\\\"\\\"\\n implement pls\\n get a name to save file\\n need be complete by child\\n :param doc:\\n :return:\\n \\\"\\\"\\\"\\n authors = ','.join([x['name'] for x in doc.artists])\\n author = re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', authors.strip())\\n mp3_name = re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', doc['name'])\\n name = \\\"%s - %s.mp3\\\" % (author, mp3_name)\\n return name\\n\\n def download_filename_full(self, doc):\\n \\\"\\\"\\\"\\n implement pls\\n get a path to save file, by relative path\\n need be complete by child\\n :param doc:\\n :type doc: Mp3Model\\n :return:\\n :rtype: str\\n \\\"\\\"\\\"\\n # authors = reduce(lambda x, y: x + ',' + y, [x['name'] for x in doc.artists])\\n authors = ','.join([x['name'] for x in doc.artists])\\n author = re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', authors.strip())\\n mp3_name = re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', doc['name'])\\n name = os.path.join(author, \\\"%s - %s.mp3\\\" % (author, mp3_name))\\n return name\\n\\n def url_load(self, doc):\\n \\\"\\\"\\\"\\n implement pls\\n :param doc:\\n :return:\\n :rtype: str\\n \\\"\\\"\\\"\\n try:\\n return get_mp3_link(doc['id'])\\n except:\\n return None\\n\\n def shortcut_self_path(self, doc):\\n \\\"\\\"\\\"\\n implement pls, not force\\n return self short cut path\\n :param doc:\\n :return:\\n \\\"\\\"\\\"\\n result = []\\n result.extend([os.path.join(\\\"artist\\\", re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', x['name']))\\n for x in doc.artists])\\n result.append(os.path.join(\\\"album\\\", re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', doc.album['name'])))\\n return result\\n\"\n },\n {\n \"alpha_fraction\": 0.7113480567932129,\n \"alphanum_fraction\": 0.7159177660942078,\n \"avg_line_length\": 27.54347801208496,\n \"blob_id\": \"d7c2044b545a4fee68f749e57eceef7a4767b965\",\n \"content_id\": \"7da0e8176a8ee2036229790663b64970b35faf22\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"reStructuredText\",\n \"length_bytes\": 1313,\n \"license_type\": \"permissive\",\n \"max_line_length\": 78,\n \"num_lines\": 46,\n \"path\": \"/CONTRIBUTING.rst\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"============\\nContributing\\n============\\n\\nContributions are welcome, and they are greatly appreciated! Every\\nlittle bit helps, and credit will always be given.\\n\\nYou can contribute in many ways:\\n\\nTypes of Contributions\\n----------------------\\n\\nReport Bugs\\n~~~~~~~~~~~\\n\\nReport bugs at https://github.com/Grass-CLP/NXSpider/issues.\\n\\nIf you are reporting a bug, please include:\\n\\n* Your operating system name and version.\\n* Any details about your local setup that might be helpful in troubleshooting.\\n* Detailed steps to reproduce the bug.\\n\\nFix Bugs\\n~~~~~~~~\\n\\nLook through the GitHub issues for bugs. Anything tagged with \\\"bug\\\"\\nis open to whoever wants to implement it.\\n\\nImplement Features\\n~~~~~~~~~~~~~~~~~~\\n\\nLook through the GitHub issues for features. Anything tagged with\\n\\\"feature\\\" is open to whoever wants to implement it.\\n\\nPull Request Guidelines\\n-----------------------\\n\\nBefore you submit a pull request, check that it meets these guidelines:\\n\\n1. The pull request should include tests.\\n2. If the pull request adds functionality, the docs should be updated. Put\\n your new functionality into a function with a docstring, and add the\\n feature to the list in README.rst.\\n3. The pull request should work for Python 2.7, and 3.x, and for PyPy.\\n and make sure that the tests pass for all supported Python versions.\\n\"\n },\n {\n \"alpha_fraction\": 0.5103048086166382,\n \"alphanum_fraction\": 0.5161103010177612,\n \"avg_line_length\": 38.25925827026367,\n \"blob_id\": \"cd2de5da1409d8c29ee08e9b2fba36f01adf56f1\",\n \"content_id\": \"e82b810c2886c563597b3771d591134967056261\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 13794,\n \"license_type\": \"permissive\",\n \"max_line_length\": 116,\n \"num_lines\": 351,\n \"path\": \"/NXSpider/bin/spider_ctrl.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/6/6.\\n# email to LipsonChan@yahoo.com\\n#\\nimport sys\\n\\nimport requests\\nfrom cement.core.controller import expose\\n\\nfrom NXSpider.bin.base_ctrl import NXSpiderBaseController, py2_decoding, py2_encoding\\nfrom NXSpider.bin.print_as_table import print_playlist, print_albums\\nfrom NXSpider.common import log\\nfrom NXSpider.common.config import Config\\nfrom NXSpider.spider import api\\n\\n\\nclass SpiderController(NXSpiderBaseController):\\n class Meta:\\n label = \\\"spider\\\"\\n stacked_on = 'base'\\n description = \\\"NXSpider\\\"\\n arguments = [\\n (['-cls', '--cls'],\\n dict(help=\\\"classes or catalogue\\\")),\\n (['-ar', '--artist'],\\n dict(help=\\\"artist\\\")),\\n (['-pl', '--playlist'],\\n dict(help=\\\"playlist\\\")),\\n (['-ur', '--user'],\\n dict(help=\\\"user\\\")),\\n (['-mv', '--mv'],\\n dict(help=\\\"artist\\\")),\\n (['-mp3', '--mp3'],\\n dict(help=\\\"song\\\")),\\n (['-ab', '--album'],\\n dict(help=\\\"album\\\")),\\n (['-dw', '--download'],\\n dict(help=\\\"download files, [], eg. -dw mv,mp3\\\")),\\n (['-offset'],\\n dict(help=\\\"offset index, eg. -offset 50\\\")),\\n (['-limit'],\\n dict(help=\\\"limit size, eg. -limit 50\\\")),\\n (['-lu'],\\n dict(help=\\\"login user or phone number\\\")),\\n (['-lp'],\\n dict(help=\\\"login password\\\")),\\n ]\\n\\n @expose(help=\\\"spider mp3, usage: smp3 -mp3 [-dw ]\\\")\\n def smp3s(self):\\n from NXSpider.bin.models import dw_mp3_mo\\n\\n if self.param_check(['mp3'], sys._getframe().f_code.co_name) is False:\\n return\\n\\n download_type = self.parse_download()\\n mp3s = self.app.pargs.mp3.split(',') # type: list\\n details = api.get_mp3_details(mp3s)\\n\\n for mid, detail in details.items():\\n log.print_info(u\\\"<{}>\\\".format(detail['name']))\\n dw_mp3_mo.parse_model(detail,\\n download_type=download_type,\\n file_check=Config().get_file_check(),\\n shortcuts_stack=[] if Config().get_shortcut() else None)\\n\\n log.print_info(\\\"spider complete!~\\\")\\n pass\\n\\n @expose(help=\\\"spider playlist, usage: spls -pl [-dw ] \\\")\\n def spls(self):\\n from NXSpider.bin.models import playlist_mo\\n\\n if self.param_check(['playlist'], sys._getframe().f_code.co_name) is False:\\n return\\n\\n download_type = self.parse_download()\\n playlists = self.app.pargs.playlist.split(',') # type: list\\n\\n for pid in playlists:\\n playlist_detail = api.get_playlist_detail(pid)\\n if playlist_detail:\\n log.print_info(u\\\"<{}> author：{}\\\".format(\\n playlist_detail['name'],\\n playlist_detail['creator']['nickname'],\\n ))\\n playlist_mo.parse_model(playlist_detail,\\n download_type=download_type,\\n file_check=Config().get_file_check(),\\n shortcuts_stack=[] if Config().get_shortcut() else None)\\n\\n log.print_info(\\\"spider complete!~\\\")\\n pass\\n\\n @expose(\\n help=\\\"spider playlist of a classes, usage: scls-pls -cls \\\"\\n \\\"[-dw ] [-offset ] [-limit ]\\\")\\n def scls_pls(self):\\n from NXSpider.bin.models import playlist_mo\\n\\n if self.param_check(['cls'], sys._getframe().f_code.co_name) is False:\\n return\\n\\n download_type = self.parse_download()\\n class_name = self.app.pargs.cls\\n class_name = py2_decoding(class_name)\\n\\n if class_name != u\\\"全部\\\" and py2_encoding(class_name) not in api.ALL_CLASSES:\\n log.print_err(\\\"class name is wrong, pls check by run : nxspider sw-pl-classes\\\")\\n return\\n\\n playlists = api.get_top_playlists(category=class_name,\\n offset=self.app.pargs.offset or 0,\\n limit=self.app.pargs.limit or 50) # type: list\\n\\n log.print_info(\\\"playlists bellow will be crawled\\\")\\n print_playlist(playlists)\\n\\n for pl_obj in playlists:\\n playlist_detail = api.get_playlist_detail(pl_obj['id'])\\n if playlist_detail:\\n log.print_info(u\\\"<{}> author：{}\\\".format(\\n playlist_detail['name'],\\n playlist_detail['creator']['nickname'],\\n ))\\n playlist_mo.parse_model(playlist_detail,\\n download_type=download_type,\\n file_check=Config().get_file_check(),\\n shortcuts_stack=[] if Config().get_shortcut() else None)\\n log.print_info(\\\"spider complete!~\\\")\\n pass\\n\\n @expose(help=\\\"spider artist top mp3, usage: sar-top-mp3 -ar [-dw ]\\\")\\n def sar_top_mp3(self):\\n from NXSpider.bin.models import artist_mo\\n\\n if self.param_check(['artist'], sys._getframe().f_code.co_name) is False:\\n return\\n\\n download_type = self.parse_download()\\n artists = self.app.pargs.artist.split(',') # type: list\\n\\n for arid in artists:\\n detail = api.get_artists_songs(arid)\\n if detail is None:\\n continue\\n\\n artist_detail = detail['artist']\\n artist_detail['mp3'] = detail['hotSongs']\\n\\n log.print_info(u\\\"<{}>\\\".format(artist_detail['name']))\\n artist_mo.parse_model(artist_detail,\\n download_type=download_type,\\n file_check=Config().get_file_check(),\\n shortcuts_stack=[] if Config().get_shortcut() else None)\\n\\n log.print_info(\\\"spider complete!~\\\")\\n pass\\n\\n @expose(help=\\\"spider artist albums, usage: sar-albums -ar [-dw ] \\\"\\n \\\"[-offset ] [-limit ]\\\")\\n def sar_albums(self):\\n from NXSpider.bin.models import artist_album_mo\\n\\n if self.param_check(['artist'], sys._getframe().f_code.co_name) is False:\\n return\\n\\n download_type = self.parse_download()\\n artists = self.app.pargs.artist.split(',') # type: list\\n\\n for arid in artists:\\n detail = api.get_artist_album(arid,\\n offset=self.app.pargs.offset or 0,\\n limit=self.app.pargs.limit or 50)\\n if detail is None:\\n continue\\n\\n artist_detail = detail['artist']\\n album_details = [api.get_album_detail(d['id']) for d in detail['hotAlbums']]\\n album_details = [d for d in album_details if d]\\n artist_detail['albums'] = album_details\\n\\n log.print_info(u\\\"<{}>\\\".format(artist_detail['name']))\\n log.print_info(\\\"albums bellow will be crawled\\\")\\n print_albums(artist_detail['albums'])\\n\\n artist_album_mo.parse_model(artist_detail,\\n download_type=download_type,\\n file_check=Config().get_file_check(),\\n shortcuts_stack=[] if Config().get_shortcut() else None)\\n\\n log.print_info(\\\"spider complete!~\\\")\\n pass\\n\\n @expose(help=\\\"spider album, usage: sab -ab [-dw ]\\\")\\n def sab(self):\\n from NXSpider.bin.models import album_mo\\n\\n if self.param_check(['album'], sys._getframe().f_code.co_name) is False:\\n return\\n\\n download_type = self.parse_download()\\n albums = self.app.pargs.album.split(',') # type: list\\n\\n for pid in albums:\\n album_detail = api.get_album_detail(pid)\\n if album_detail is None:\\n continue\\n\\n log.print_info(u\\\"{} artist：{}\\\".format(\\n \\\"<\\\" + album_detail['name'] + \\\">\\\",\\n album_detail['artist']['name'],\\n ))\\n album_mo.parse_model(album_detail,\\n download_type=download_type,\\n file_check=Config().get_file_check(),\\n shortcuts_stack=[] if Config().get_shortcut() else None)\\n\\n log.print_info(\\\"spider complete!~\\\")\\n pass\\n\\n @expose(\\n help=\\\"spider a user's playlists, usage: sur-pls -ur \\\"\\n \\\"[-dw ] [-offset ] [-limit ]\\\")\\n def sur_pls(self):\\n from NXSpider.bin.models import playlist_mo\\n\\n if self.param_check(['user'], sys._getframe().f_code.co_name) is False:\\n return\\n\\n download_type = self.parse_download()\\n user_id = self.app.pargs.user\\n playlists = api.user_playlist(user_id,\\n offset=self.app.pargs.offset or 0,\\n limit=self.app.pargs.limit or 50)\\n\\n log.print_info(\\\"playlists bellow will be crawled\\\")\\n print_playlist(playlists)\\n\\n for pl_obj in playlists:\\n playlist_detail = api.get_playlist_detail(pl_obj['id'])\\n if playlist_detail:\\n log.print_info(u\\\"<{}> author：{}\\\".format(\\n playlist_detail['name'],\\n playlist_detail['creator']['nickname'],\\n ))\\n playlist_mo.parse_model(playlist_detail,\\n download_type=download_type,\\n file_check=Config().get_file_check(),\\n shortcuts_stack=[] if Config().get_shortcut() else None)\\n log.print_info(\\\"spider complete!~\\\")\\n pass\\n\\n @expose(\\n help=\\\"spider top mvs usage: stop-mv [-offset ] [-limit ]\\\")\\n def stop_mvs(self):\\n from NXSpider.bin.models import no_rec_mv_mo\\n\\n mvs = api.top_mvs(offset=self.app.pargs.offset or 0,\\n limit=self.app.pargs.limit or 50)\\n mvs = [api.get_mv_detail(d['id']) for d in mvs]\\n mvs = [d for d in mvs if d]\\n\\n for mv in mvs:\\n no_rec_mv_mo.parse_model(mv, download_type=['mv'],\\n file_check=Config().get_file_check(),\\n shortcuts_stack=[] if Config().get_shortcut() else None)\\n\\n log.print_info(\\\"spider complete!~\\\")\\n\\n pass\\n\\n @expose(\\n help=\\\"login and spider mv, usage: login-smv -lu [-lp ]\\\")\\n def login_smv(self):\\n from NXSpider.bin.models import no_rec_mv_mo\\n\\n if self.param_check(['lu'], sys._getframe().f_code.co_name) is False:\\n return\\n\\n plaintext_pwd = self.app.pargs.lp or None\\n if plaintext_pwd is None:\\n import getpass\\n plaintext_pwd = getpass.getpass(\\\"Please input your password:\\\")\\n\\n session = requests.session()\\n import hashlib\\n password = hashlib.md5(plaintext_pwd).hexdigest()\\n res = api.login(self.app.pargs.lu, password, session)\\n if res.get('code', 0) != 200:\\n log.print_err('login failed, msg: {}'.format(res.get('msg', \\\"none\\\")))\\n exit()\\n\\n mvs = api.my_mvs(session)\\n mvs = [api.get_mv_detail(d['id']) for d in mvs]\\n mvs = [d for d in mvs if d]\\n\\n for mv in mvs:\\n no_rec_mv_mo.parse_model(mv, download_type=['mv'],\\n file_check=Config().get_file_check(),\\n shortcuts_stack=[] if Config().get_shortcut() else None)\\n\\n log.print_info(\\\"spider complete!~\\\")\\n\\n pass\\n\\n @expose(\\n help=\\\"login and spider playlists, usage: login-spls -lu [-lp ] [-dw ]\\\")\\n def login_spls(self):\\n if self.param_check(['lu', 'lp'], sys._getframe().f_code.co_name) is False:\\n return\\n\\n from NXSpider.bin.models import playlist_mo\\n\\n plaintext_pwd = self.app.pargs.lp or None\\n if plaintext_pwd is None:\\n import getpass\\n plaintext_pwd = getpass.getpass(\\\"Please input your password:\\\")\\n\\n session = requests.session()\\n import hashlib\\n password = hashlib.md5(plaintext_pwd).hexdigest()\\n res = api.login(self.app.pargs.lu, password, session)\\n if res.get('code', 0) != 200:\\n log.print_err('login failed, msg: {}'.format(res.get('msg', \\\"none\\\")))\\n exit()\\n\\n user_id = res['account']['id']\\n download_type = self.parse_download()\\n playlists = api.user_playlist(user_id,\\n offset=self.app.pargs.offset or 0,\\n limit=self.app.pargs.limit or 1000)\\n\\n log.print_info(\\\"playlists bellow will be crawled\\\")\\n print_playlist(playlists)\\n\\n for pl_obj in playlists:\\n playlist_detail = api.get_playlist_detail(pl_obj['id'])\\n if playlist_detail:\\n log.print_info(u\\\"<{}> author：{}\\\".format(\\n playlist_detail['name'],\\n playlist_detail['creator']['nickname'],\\n ))\\n playlist_mo.parse_model(playlist_detail,\\n download_type=download_type,\\n file_check=Config().get_file_check(),\\n shortcuts_stack=[] if Config().get_shortcut() else None)\\n log.print_info(\\\"spider complete!~\\\")\\n pass\\n\"\n },\n {\n \"alpha_fraction\": 0.6816326379776001,\n \"alphanum_fraction\": 0.6938775777816772,\n \"avg_line_length\": 34.07143020629883,\n \"blob_id\": \"8ceb505bf02e62f8a8f6989ce6570c4b3d8aa755\",\n \"content_id\": \"e07373c047ac915cab80e31caee5c5ec56d82493\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 530,\n \"license_type\": \"permissive\",\n \"max_line_length\": 156,\n \"num_lines\": 14,\n \"path\": \"/README_EN.md\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"NXSpider\\n=================\\n\\nNXSpider, NetEase X Spider, A powerful for mp3,mv spider, can help you download mp3 and mv with media tags. Base on python, mongodb, and recursion algorithm\\n\\n[](LICENSE.md)\\n\\n\\n\\n[新手先看这里](SIMPLE_USE.md) | [历史版本在这里](VERSION.md) | [开发详情在这里](DEV.md)\\n\\n\\n==========\\ntodo: to be continue\"\n },\n {\n \"alpha_fraction\": 0.5470085740089417,\n \"alphanum_fraction\": 0.6111111044883728,\n \"avg_line_length\": 14.600000381469727,\n \"blob_id\": \"4f5921b94f057e383094bdbae9fea67d44788142\",\n \"content_id\": \"7bd839cb6197f9a53416f08da81d902af42764b3\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 234,\n \"license_type\": \"permissive\",\n \"max_line_length\": 33,\n \"num_lines\": 15,\n \"path\": \"/NXSpider/common/__init__.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/19.\\n# email to LipsonChan@yahoo.com\\n#\\nimport sys\\n\\nPYTHON3 = False\\nPYTHON2 = False\\nif sys.version > \\\"3\\\":\\n PYTHON3 = True\\n\\nif sys.version[0] == '2':\\n PYTHON2 = True\\n\"\n },\n {\n \"alpha_fraction\": 0.5006303191184998,\n \"alphanum_fraction\": 0.5146378874778748,\n \"avg_line_length\": 39.556819915771484,\n \"blob_id\": \"273e62ee2ef3e909cc0f9e2f8797ad711a6f02f4\",\n \"content_id\": \"c4628d2e84e3850e6cb1689090f3fc6e80199fb6\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 7139,\n \"license_type\": \"permissive\",\n \"max_line_length\": 116,\n \"num_lines\": 176,\n \"path\": \"/NXSpider/bin/config_ctrl.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/6/5.\\n# email to LipsonChan@yahoo.com\\n#\\nimport json\\nimport os\\n\\nfrom cement.core.controller import CementBaseController, expose\\n\\nfrom NXSpider.common import log\\nfrom NXSpider.common.config import Config, default_path_key, mv_resolutions\\nfrom NXSpider.common.constant import default_download_dir\\n\\n\\nclass ConfigController(CementBaseController):\\n class Meta:\\n label = \\\"config\\\"\\n stacked_on = 'base'\\n description = \\\"NXSpider config\\\"\\n arguments = [\\n (['-mh', '--mhost'],\\n dict(help=\\\"mongo host\\\")),\\n (['-mp', '--mport'],\\n dict(help=\\\"mongo port\\\")),\\n (['-mu', '--muser'],\\n dict(help=\\\"mongo user\\\")),\\n (['-mpw', '--mpassword'],\\n dict(help=\\\"mongo password\\\")),\\n (['-mn', '--mdbname'],\\n dict(help=\\\"mongo db name, default nxspider\\\")),\\n (['-nomongo', '--nomongo'],\\n dict(help=\\\"no mongo mode, true(1) or false(0), default True, while true, mongodb will be never used\\\")),\\n\\n (['-path', '--path_download'],\\n dict(help=\\\"download path, default ~/.nxspider/download_files/, eg. -path defualt;path1;path2\\\")),\\n (['-mvr', '--mv_resolution'],\\n dict(help=\\\"mv default resolution, [240, 480, 720, 1080] default 720\\\")),\\n (['-stt', '--shortcut'],\\n dict(help=\\\"build shortcuts, true(1) or false(0), default False\\\")),\\n (['-tag', '--media_tag'],\\n dict(help=\\\"media tag, true(1) or false(0), default True\\\")),\\n (['-tag163', '--media_tag_163'],\\n dict(help=\\\"media tag of 163 comment, true(1) or false(0), default True\\\")),\\n (['-dfc', '--download_file_check'],\\n dict(help=\\\"check file exist in paths set, true(1) or false(0), default True\\\")),\\n ]\\n\\n @expose(help=\\\"config mongo server with [-h --host] [-p --port] \\\"\\n \\\"[-u --user] [-pw --password] [-c --collections]\\\")\\n def config_mongo(self):\\n config = Config()\\n config_dict = config.config # type: dict\\n mongo_key = 'mongo'\\n is_config = False\\n try:\\n if self.app.pargs.mhost is not None:\\n config_dict[mongo_key]['host'] = self.app.pargs.mhost\\n config_dict['no_mongo'] = False\\n is_config = True\\n\\n if self.app.pargs.mport is not None:\\n config_dict[mongo_key]['port'] = int(self.app.pargs.mport)\\n is_config = True\\n\\n if self.app.pargs.muser is not None:\\n config_dict[mongo_key]['username'] = self.app.pargs.muser\\n is_config = True\\n\\n if self.app.pargs.mpassword is not None:\\n config_dict[mongo_key]['password'] = self.app.pargs.mpassword\\n is_config = True\\n\\n if self.app.pargs.mdbname is not None:\\n config_dict[mongo_key]['name'] = self.app.pargs.mdbname\\n is_config = True\\n\\n if self.app.pargs.nomongo is not None:\\n config_dict['no_mongo'] = True if self.app.pargs.nomongo.lower() == 'true'\\\\\\n or self.app.pargs.nomongo == '1' else False\\n is_config = True\\n except:\\n log.print_err(\\\"input error, pls check\\\")\\n raise\\n if is_config:\\n config.save_config_file()\\n log.print_info(\\\"config success\\\")\\n self.config_show()\\n\\n @expose(help=\\\"config spider behavior with [-path ]\\\"\\n \\\"[-mvr <240,480,720,1080>] [-tag <1 or 0>] [-tag163 <1 or 0>]\\\"\\n \\\"[-dfc <1 or 0>] [-stt <1 or 0>]\\\")\\n def config_spider(self):\\n config = Config()\\n config_dict = config.config # type: dict\\n is_config = False\\n try:\\n if self.app.pargs.path_download is not None:\\n paths = self.app.pargs.path_download.split(',') # type: list\\n if default_path_key in paths:\\n index = paths.index(default_path_key)\\n paths.remove(default_path_key)\\n paths.insert(index, default_download_dir)\\n\\n final_paths = []\\n for p in paths:\\n try:\\n # some error need pass\\n if os.path.isdir(p) is False:\\n os.mkdir(p)\\n final_paths.append(p)\\n except:\\n log.print_warn(\\\"path may be wrong and be deleted: {}\\\".format(p))\\n pass\\n\\n if not final_paths:\\n final_paths.append(default_download_dir)\\n\\n log.print_info('path will be set as: ' + ','.join(final_paths))\\n\\n config_dict['download_path'] = final_paths\\n is_config = True\\n\\n if self.app.pargs.mv_resolution is not None:\\n r = int(self.app.pargs.mv_resolution)\\n if r not in mv_resolutions:\\n log.print_warn(\\\"-mvr resolution config skip, value must be 240,480,720,1080\\\")\\n config_dict['mv_def_resolution'] = r\\n is_config = True\\n\\n if self.app.pargs.media_tag is not None:\\n config_dict['media_tag'] = True if self.app.pargs.media_tag.lower() == 'true'\\\\\\n or self.app.pargs.media_tag == '1' else False\\n is_config = True\\n\\n if self.app.pargs.media_tag_163 is not None:\\n config_dict['media_tag_163'] = True if self.app.pargs.media_tag_163.lower() == 'true' \\\\\\n or self.app.pargs.media_tag_163 == '1' else False\\n is_config = True\\n\\n if self.app.pargs.shortcut is not None:\\n config_dict['shortcut'] = True if self.app.pargs.shortcut.lower() == 'true' \\\\\\n or self.app.pargs.shortcut == '1' else False\\n is_config = True\\n\\n\\n except:\\n log.print_err(\\\"input error, pls check\\\")\\n raise\\n if is_config:\\n config.save_config_file()\\n log.print_info(\\\"config success\\\")\\n self.config_show()\\n\\n @expose(help=\\\"check config is valid or not\\\")\\n def config_check(self):\\n self.config_show()\\n try:\\n config = Config()\\n if config.config_test():\\n log.print_info('config check complete, all is well done!')\\n except:\\n log.print_err('config check failed, pls re config')\\n\\n @expose(help=\\\"clear all config, u need re-config from beginning\\\")\\n def config_clear(self):\\n Config().config_reset()\\n log.print_info(\\\"config has been reset, u need re-config from beginning pls\\\")\\n self.config_show()\\n\\n @expose(help=\\\"list all config opt\\\")\\n def config_show(self):\\n config_dict = Config().config\\n log.print_info(\\\"config will be show fellow:\\\")\\n print(json.dumps(config_dict, ensure_ascii=False, indent=1))\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.7139001488685608,\n \"alphanum_fraction\": 0.7363922595977783,\n \"avg_line_length\": 27.87013053894043,\n \"blob_id\": \"d430fcf244be5988a571684c01a0c2ac3a34bda3\",\n \"content_id\": \"71157e5eb0ea60b8df291ecb424356375324a565\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 3963,\n \"license_type\": \"permissive\",\n \"max_line_length\": 109,\n \"num_lines\": 77,\n \"path\": \"/DEV.md\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"NXSpider\\n=================\\n\\nNXSpider， NetEase X Spider，一个强大的（网易云音乐）mp3,mv爬虫，可以下载和收集mp3,mv信息,同时附带多媒体标签信息。采用python编写，mongo数据库(非必须)，递归算法核心实现\\n\\n[](LICENSE.md)\\n\\n\\n\\n[新手先看这里](SIMPLE_USE.md) | [历史版本在这里](VERSION.md) | [开发详情在这里](DEV.md)\\n\\n## 开发及问题(不关心的可以不看了)\\n\\n### 开发调试\\n1. 已安装则采用(下文采用该方式进行说明)：\\n```\\n$nxspider ...\\n```\\n2. 未安装+win采用(未安装指没执行setup.py,只用代码运行)：\\n```\\nbash_python(3) NXSipder/bin/cli.py ...\\n```\\n3. 未安装+linux采用：\\n```\\nPYTHONPATH=. python(3) NXSipder/bin/cli.py ...\\n```\\n4. 注意,win下采用bash_python时，分割符 ```,``` 必须改为 ```:``` eg.\\n```\\n$bash_python.bat NXSpider/bin/cli.py config-spider -path d:\\\\netease_dw:default\\n```\\n\\n### 配置说明\\n1. ~~config-mongo必须运行爬取前配置!!!，dbname可不指定，默认为nxspider~~\\n2. 默认无mongodb模式，需要采集数据则通过配置第一条即可 `nxspider config-mongo ..`\\n3. 一旦配置了 `mongodb -mh` 则 `-nomogo` 会自动设置为0，即配置了host就会使用的意思\\n2. path **强烈建议**爬取前配置，```,``` 为多个下载路径分隔符。default指`~/.nxspider/download_files/`\\n3. 其他配置可选，请查看 `nxspider -h`\\n4. ```-tag 1``` 建议保留 `-tag163 1` 根据需求保留\\n6. 配置完建议执行 ```nxspider config-check``` 检查配置正确性\\n\\n### 注意\\n- 本项目纯粹是学习开发使用，欢迎大家互相讨论，下载的资料请24小时内删除\\n- 涉及侵权以及版权问题欢迎讨论和提出\\n\\n### 协助开发或2次开发建议\\n1. 希望尽可能(yahoo邮件)跟作者(LipsonChan)联系，以及对项目进行加★\\n2. 核心代码为NXSpider/bin以及NXSpider/spider/base_driver.py\\n3. 主要逻辑为通过api获得对于json数据，采用递归+配置方式，自动下载可下载对象\\n4. 如果有任何反馈，希望回复到项目的issue,注明版本号，运行环境及描述清楚问题\\n\\n### windows + python2问题\\n1. 参数输入(目前只有查询)非中文和latin，可能会出现问题\\n2. 不推荐配置路径中有非latin字符，及路径最好是英文，没测试过\\n3. 查询结果输出无法显示非中文和latin数据，如韩文（日文可以）\\n\\n### 其他问题\\n- 同歌手同名MP3不会被重复下载\\n\\n### 开发历程\\n- 基于spider163项目(不满足且有小bug)，开发mongodb以及可以更多爬取项目\\n- 增加更多url，MV下载，尝试使用eapi发现很难发现加密规则\\n- 修改项目结构为下载驱动driver和model双层，配合metaclass进行递归式下载，减少后续增量开发工作\\n- 为了配合Netease app和桌面软件，修改mp3 mp4 tag信息\\n- 艰难的使用了eyed3，windows超级麻烦(现成的dll还要配置，还有字符集问题)，发现不支持mp4，非常难过\\n- 发现mutagen，果断弃坑eyed3\\n- 发布0.0.1版本，可以通过playlist，ablum，aritst_top_mp3爬取和下载mp3,mv\\n- 发现mongodb可能有人不喜欢，新增无数据库版本（默认无需使用），通过配置可切换\\n- 新增通过获取最火playlist进行爬取\\n- 新增了好多功能，读 [VERSION.md](VERSION.md)\\n- api全靠`fiddler`和猜，**一把辛酸一把泪**\\n\\n### 下阶段开发\\n- 通过用户id，爬取该用户所有的歌单 √\\n- 通过歌手id，爬取该歌手所有专辑 √\\n- 通过排行版，爬取最新n个歌单 √\\n- mongodb 可选？不强制（但对于离线则无法加tag，虽然只是个人用） √\\n- 通过登录(或非登录)，爬取用户收藏的歌单，mv **不安全(明文账号密码)，而且经常登录api会被限，不开放说明**\\n\"\n },\n {\n \"alpha_fraction\": 0.4928571283817291,\n \"alphanum_fraction\": 0.6714285612106323,\n \"avg_line_length\": 16.625,\n \"blob_id\": \"997278c5c556dd28c599e926e1520b11ce90cd0a\",\n \"content_id\": \"3bc1f70e3849ded624faa36b73cc7843fa56d66c\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Text\",\n \"length_bytes\": 140,\n \"license_type\": \"permissive\",\n \"max_line_length\": 20,\n \"num_lines\": 8,\n \"path\": \"/requirements.txt\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"six>=1,<=2\\ncement>=2,<3\\nmutagen>=1.40,<2\\nterminaltables>=3,<4\\ncryptography>=2.2,<3\\nmongoengine>=0.15,<1\\nrequests>=2.19,<3\\ncolorama>=0.3,<0.4\"\n },\n {\n \"alpha_fraction\": 0.5651697516441345,\n \"alphanum_fraction\": 0.582694411277771,\n \"avg_line_length\": 23.026315689086914,\n \"blob_id\": \"bc07e9820d1fa550e673fd1209b4b8a50a92bd9b\",\n \"content_id\": \"c488966393e41549013cefe0feb20cfd4156a44c\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 913,\n \"license_type\": \"permissive\",\n \"max_line_length\": 76,\n \"num_lines\": 38,\n \"path\": \"/NXSpider/utility/shortcut.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/12/15.\\n# email to LipsonChan@yahoo.com\\n#\\n\\nimport os\\n\\nif os.name == \\\"nt\\\":\\n def symlink_ms(source, link_name):\\n import ctypes\\n csl = ctypes.windll.kernel32.CreateSymbolicLinkW\\n csl.argtypes = (ctypes.c_wchar_p, ctypes.c_wchar_p, ctypes.c_uint32)\\n csl.restype = ctypes.c_ubyte\\n flags = 1 if os.path.isdir(source) else 0\\n try:\\n if csl(link_name, source.replace('/', '\\\\\\\\'), flags) == 0:\\n raise ctypes.WinError()\\n except Exception as e:\\n pass\\n\\n\\n os.symlink = symlink_ms\\n\\n\\ndef symlink(source, link_name):\\n try:\\n os.symlink(source, link_name)\\n except Exception as e:\\n pass\\n\\n\\n# a = os.getcwd()\\n# src = os.path.join(a, \\\"install.bat\\\")\\n# dst = os.path.join(a, \\\"test.link\\\")\\n#\\n# os.symlink(os.path.join(a, \\\"install.bat\\\"), os.path.join(a, \\\"test.link\\\"))\\n\"\n },\n {\n \"alpha_fraction\": 0.5379061102867126,\n \"alphanum_fraction\": 0.5848375558853149,\n \"avg_line_length\": 16.3125,\n \"blob_id\": \"7a82dab413fa9f73c3b00dd08bb34ca098db3103\",\n \"content_id\": \"7bf4bb6ce6ae6676b958ec059cce89eaffc3c901\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 277,\n \"license_type\": \"permissive\",\n \"max_line_length\": 41,\n \"num_lines\": 16,\n \"path\": \"/NXSpider/spider/artist.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/26.\\n# email to LipsonChan@yahoo.com\\n#\\n\\nfrom NXSpider.spider.base_driver import *\\n\\n\\nclass Artist(Music163Obj):\\n __model_name__ = ArtistModel\\n __model_rfilter__ = {\\n 'img1v1Id',\\n 'picId',\\n }\\n\"\n },\n {\n \"alpha_fraction\": 0.5031654238700867,\n \"alphanum_fraction\": 0.5240848064422607,\n \"avg_line_length\": 26.3157901763916,\n \"blob_id\": \"43765f2a01ae42e3f33332441c5a6e27d61b515c\",\n \"content_id\": \"aba075c3a131165a30a7e42fc5f0a9391c4adf88\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3633,\n \"license_type\": \"permissive\",\n \"max_line_length\": 90,\n \"num_lines\": 133,\n \"path\": \"/NXSpider/utility/media_tag.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/5/17.\\n# email to LipsonChan@yahoo.com\\n#\\nimport os\\n\\nfrom mutagen import id3\\nfrom mutagen.id3 import ID3\\nfrom mutagen.mp4 import MP4\\n\\nfrom NXSpider.common.config import Config\\nfrom NXSpider.common.encrypt import aes_ecb\\nfrom NXSpider.model.export import Mp3Model, Mp4Model\\n\\nkey_names = ['title', 'artist', 'album', 'album_artist', 'track_num', 'comment']\\nencoding_keys = ['title', 'artist', 'album', 'album_artist', 'comment']\\nmutagen_idv3_key_map = {\\n 'title': 'TIT2',\\n 'artist': 'TPE1',\\n 'album': 'TALB',\\n 'album_artist': 'TPE2',\\n 'track_num': 'TRCK',\\n 'comment': 'COMM'\\n}\\n\\nmugagen_mp4_key_map = {\\n 'title': '\\\\xa9nam',\\n 'artist': '\\\\xa9ART',\\n 'comment': '\\\\xa9cmt',\\n}\\naes_code = \\\"#14ljk_!\\\\]&0U<'(\\\"\\n\\n\\n# @tools.ignored(Exception)\\ndef attach_mp3_idv3(doc, file):\\n \\\"\\\"\\\"\\n :type doc: Mp3Model\\n :param doc:\\n :param file:\\n :return:\\n \\\"\\\"\\\"\\n artists = [('[\\\"%s\\\",%d]' % (x['name'], x.id)) for x in doc.artists]\\n artists_str = \\\",\\\".join(artists)\\n\\n authors = \\\",\\\".join([x['name'] for x in doc.artists])\\n data = {\\n 'title': doc['name'],\\n 'artist': authors,\\n 'album': doc['album']['name'],\\n 'album_artist': authors,\\n 'track_num': str(doc['no']),\\n }\\n\\n if Config().get_media_tag_163():\\n comment_plaintext = u'music:{\\\"musicId\\\":%d,\\\"musicName\\\":\\\"%s\\\",\\\"bitrate\\\":320000,' \\\\\\n u'\\\"albumId\\\":%d,\\\"album\\\":\\\"%s\\\", \\\"artist\\\":[%s]}' \\\\\\n % (doc.id, doc['name'], doc.album.id, doc.album['name'],\\n artists_str)\\n comment = \\\"163 key(Don't modify):\\\" + aes_ecb(comment_plaintext, aes_code).decode()\\n data['comment'] = comment\\n\\n try:\\n mp3 = ID3(file, v2_version=3)\\n for k, v in data.items():\\n if k not in mutagen_idv3_key_map:\\n continue\\n\\n if k == 'comment':\\n mp3.add(id3.COMM(lang='XXX', text=v))\\n continue\\n\\n attr_type = getattr(id3, mutagen_idv3_key_map[k], None)\\n if attr_type:\\n mp3.add(attr_type(text=v))\\n\\n mp3.save(v2_version=3)\\n except Exception as e:\\n return False\\n return True\\n\\n\\n# @tools.ignored(Exception)\\ndef attach_mp4_tag(doc, file):\\n \\\"\\\"\\\"\\n :type doc: Mp4Model\\n :param doc:\\n :param file:\\n :return:\\n \\\"\\\"\\\"\\n authors = u\\\",\\\".join([x['name'] for x in doc.artists])\\n data = {\\n 'title': doc['name'],\\n 'artist': authors,\\n }\\n\\n if Config().get_media_tag_163():\\n comment_plaintext = u'mv:{\\\"title\\\":\\\"%s\\\",\\\"mvId\\\":%d,\\\"artistId\\\":%d,' \\\\\\n u'\\\"artistName\\\":\\\"%s\\\",\\\"pubTime\\\":\\\"%s\\\",\\\"bitrate\\\":%d}' \\\\\\n % (doc['name'], doc.id, doc.artists[0].id,\\n doc.artists[0]['name'], doc['publishTime'],\\n doc['download_video_r'])\\n comment = \\\"163 key(Don't modify):\\\" + aes_ecb(comment_plaintext, aes_code).decode()\\n data['comment'] = comment\\n\\n try:\\n mp4 = MP4(file)\\n for k, v in data.items():\\n if k not in mugagen_mp4_key_map:\\n continue\\n mp4[mugagen_mp4_key_map[k]] = v\\n\\n mp4.save()\\n except Exception as e:\\n return False\\n return True\\n\\n\\ndef attach_media_tag(doc, file):\\n \\\"\\\"\\\"\\n attach a media idv3\\n :param doc:\\n :param file:\\n :return:\\n \\\"\\\"\\\"\\n if not os.path.exists(file):\\n return False\\n\\n if isinstance(doc, Mp3Model):\\n return attach_mp3_idv3(doc, file)\\n elif isinstance(doc, Mp4Model):\\n return attach_mp4_tag(doc, file)\\n\"\n },\n {\n \"alpha_fraction\": 0.5641250610351562,\n \"alphanum_fraction\": 0.6272276043891907,\n \"avg_line_length\": 30.69444465637207,\n \"blob_id\": \"e804681beff3131412ea21c32748de816d189470\",\n \"content_id\": \"31c87e97a223c09fe5a8e2246260cb3c186e3234\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3445,\n \"license_type\": \"permissive\",\n \"max_line_length\": 115,\n \"num_lines\": 108,\n \"path\": \"/NXSpider/debug_test/crawl.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/5/11.\\n# email to LipsonChan@yahoo.com\\n#\\nimport hashlib\\n\\nimport requests\\n\\nfrom NXSpider.common import tools, log\\nfrom NXSpider.common.constant import all_download_type\\nfrom NXSpider.spider.api import get_playlist_detail, get_mp3_links, get_album_detail, get_top_playlists, \\\\\\n get_playlist_classes, \\\\\\n top_artists, get_artists_songs, get_artist_album, get_mp3_details, search, phone_login, login, user_playlist, \\\\\\n get_playlist_detail_v3, get_playlist_catelogs, hot_mvs, all_mvs, top_mvs\\nfrom NXSpider.spider.playlist import Playlist\\n\\n# def playlist_by_id(link, download_type=all_download_type,\\n# save=True, file_check=True):\\n# playlist_detail = get_playlist_detail(link)\\n# with tools.ignored(Exception):\\n# log.print_info(\\\"%s author：%s\\\" % (\\n# \\\"<\\\" + playlist_detail['name'] + \\\">\\\",\\n# playlist_detail['creator']['nickname'],\\n# ))\\n#\\n# playlist_mo = Playlist()\\n# playlist_mo.parse_model(playlist_detail,\\n# save=save, download_type=download_type,\\n# file_check=file_check)\\n# pass\\n\\n\\n# def playlist_by_page(page, type=u\\\"全部\\\"):\\n# play_url = constant.play_url.format(type, page * 35)\\n# titles = []\\n# try:\\n# log.print_info(\\\"begin crawl playlist page: {}, type: {}\\\".format(page, type))\\n# acmsk = {'class': 'msk'}\\n# scnb = {'class': 'nb'}\\n# dcu = {'class': 'u-cover u-cover-1'}\\n# ucm = {'class': 'm-cvrlst f-cb'}\\n# data = tools.curl(play_url, constant.header, type=constant.RETURE_HTML)\\n# lst = data.find('ul', ucm)\\n# for play in lst.find_all('div', dcu):\\n# title = tools.encode(play.find('a', acmsk)['title'])\\n# link = tools.encode(play.find('a', acmsk)['href']).replace(\\\"/playlist?id=\\\", \\\"\\\")\\n# # playlist_by_id(link)\\n# print(\\\"link: {}, title: {}\\\".format(link, title))\\n# except Exception as e:\\n# log.print_err(\\\"playlist page error：{} type：{} page：{}\\\".format(e, type, page))\\n# log.print_err(e)\\n\\n\\n# def playlist_all_page():\\n# cf = \\\"全部\\\"\\n# for i in range(36):\\n# playlist_by_page(i + 1, cf)\\n\\n# test = get_mv_details([5322493,239037])\\ntest = top_mvs()\\ntest = all_mvs()\\ntest = hot_mvs()\\ntest = get_playlist_catelogs()\\ntest = get_playlist_classes()\\ntest = get_top_playlists()\\ntest = user_playlist(48872048)\\ntest = get_playlist_detail(92024088)\\ntest = get_playlist_detail(2246057871)\\ntest = get_playlist_detail(107020750)\\ntest = get_playlist_detail_v3(107020750)\\n\\np = hashlib.md5('000'.encode('utf-8')).hexdigest()\\ns = requests.Session()\\n\\ntest = user_playlist(92024088, session=s)\\ntest = get_playlist_detail(107020750)\\n\\n\\ntest = search('周杰伦', stype=100)\\ntest = top_artists()\\ntest = get_playlist_classes()\\ntest = get_mp3_details([412902496,412902496,412902496])\\n\\ntest = get_artists_songs(9621)\\ntest = get_artist_album(9621)\\n\\n\\ntest = top_artists()\\n# test = playlist_classes()\\ntest = get_top_playlists()\\n\\ntest = get_album_detail(32324)\\n\\n# test = get_mp3_link(412902496)\\n# playlist_by_id(466225104)\\ntests = get_mp3_links([412902496, 326904])\\n# http://music.163.com/api/song/detail?ids=[412902496,326904] work\\n# http://music.163.com/api/album/32324 don't work\\n# http://music.163.com/#/artist?id=9621\\n# http://music.163.com/api/artist/9621 don't work\\n# playlist_all_page()\\n\\n\\n\\n\\nprint('a')\\n\"\n },\n {\n \"alpha_fraction\": 0.5634615421295166,\n \"alphanum_fraction\": 0.5788461565971375,\n \"avg_line_length\": 23.186046600341797,\n \"blob_id\": \"e418f50b90ec3ae27156087529e48aa0ed53a751\",\n \"content_id\": \"4cdbcafe5489408d8cfb6b9d805c0596cfa7b2ba\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1040,\n \"license_type\": \"permissive\",\n \"max_line_length\": 60,\n \"num_lines\": 43,\n \"path\": \"/NXSpider/spider/playlist.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/20.\\n# email to LipsonChan@yahoo.com\\n#\\nimport os\\n\\nfrom NXSpider.spider.mp3 import Mp3\\nfrom NXSpider.spider.base_driver import *\\nfrom NXSpider.spider.user import User\\n\\n\\nclass Playlist(Music163Obj):\\n __model_name__ = PlaylistModel\\n __model_rfilter__ = {\\n # garbage properties\\n 'coverImgId_str',\\n 'coverImgId_str',\\n }\\n __parse_recursion__ = {\\n 'mp3': Mp3(),\\n 'creator': User(),\\n }\\n\\n @attr_replace(attr_name='tracks', new_name='mp3')\\n def replace_tracks(self, obj):\\n return obj\\n\\n @attr_replace(attr_name='creator', new_name='creator')\\n def replace_creator(self, obj):\\n if 'userId' in obj and 'id' not in obj:\\n obj['id'] = obj['userId']\\n return obj\\n\\n def shortcut_relative_name(self, doc):\\n \\\"\\\"\\\"\\n implement pls, not force\\n return short cut relative path\\n :param doc:\\n :return:\\n \\\"\\\"\\\"\\n return os.path.join(self.__file_type__, doc['name'])\\n\"\n },\n {\n \"alpha_fraction\": 0.5814639925956726,\n \"alphanum_fraction\": 0.5897284746170044,\n \"avg_line_length\": 27.711864471435547,\n \"blob_id\": \"35db5a3db506d34cba4fb78ae0fd15ce29c7f519\",\n \"content_id\": \"f07c1f1fd00e7607fa1c5609301c8cccff46621f\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1694,\n \"license_type\": \"permissive\",\n \"max_line_length\": 89,\n \"num_lines\": 59,\n \"path\": \"/NXSpider/bin/base_ctrl.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/6/20.\\n# email to LipsonChan@yahoo.com\\n#\\nimport sys\\n\\nfrom cement.core.controller import CementBaseController\\n\\nfrom NXSpider.common import log, PYTHON2\\nfrom NXSpider.common.constant import all_download_type\\n\\n\\ndef py2_decoding(string):\\n if PYTHON2:\\n return string.decode(sys.getfilesystemencoding())\\n return string\\n\\n\\ndef py2_encoding(string):\\n if PYTHON2:\\n return string.encode('utf8')\\n\\n\\nclass NXSpiderBaseController(CementBaseController):\\n def parse_download(self):\\n \\\"\\\"\\\"\\n lost of spider function will parse -dw param, this will do it\\n :return:\\n \\\"\\\"\\\"\\n if self.app.pargs.download is None:\\n download_type = []\\n else:\\n download_type = self.app.pargs.download.split(',') # type: list\\n download_type = list(filter(lambda x: x in all_download_type, download_type))\\n return download_type\\n\\n def param_check(self, params, func_name):\\n \\\"\\\"\\\"\\n this will check param inputted and require is complete or not, and print help\\n help will be in expose(help='...'), and got by function name\\n :param params:\\n :param func_name:\\n :return:\\n \\\"\\\"\\\"\\n help = None\\n fun = getattr(self, func_name, None)\\n if fun and getattr(fun, '__cement_meta__', None):\\n help = fun.__cement_meta__['help']\\n\\n for p in params:\\n param = getattr(self.app.pargs, p, None)\\n if param is None:\\n log.print_err(\\\"param {} miss, see help:\\\".format(p))\\n if help:\\n print(help)\\n return False\\n return True\\n\"\n },\n {\n \"alpha_fraction\": 0.5121457576751709,\n \"alphanum_fraction\": 0.5206477642059326,\n \"avg_line_length\": 31.5,\n \"blob_id\": \"fb53946dc8c617a632c9e4d281f7d5375a9ddd72\",\n \"content_id\": \"67e06b7e8ccd3518a1b90b290589bf4b3114fcc6\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2470,\n \"license_type\": \"permissive\",\n \"max_line_length\": 87,\n \"num_lines\": 76,\n \"path\": \"/NXSpider/bin/print_as_table.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/6/23.\\n# email to LipsonChan@yahoo.com\\n#\\nfrom terminaltables import AsciiTable\\n\\n\\ndef print_playlist(pls):\\n table = AsciiTable([[\\\"ID\\\", \\\"Name\\\", \\\"User\\\", \\\"PlayCount\\\"]])\\n table_data = [[str(item['id']), item['name'],\\n item['creator']['nickname'],\\n str(item['playCount']),\\n ] for item in pls]\\n table.table_data.extend(table_data)\\n print(table.table)\\n pass\\n\\n\\ndef print_albums(abs):\\n from terminaltables import AsciiTable\\n table = AsciiTable([[\\\"ID\\\", \\\"Album\\\", \\\"Artist\\\", \\\"ArtistID\\\"]])\\n table_data = [[str(item['id']), item['name'],\\n ','.join([ar['name'] for ar in item['artists']]),\\n ','.join([str(ar['id']) for ar in item['artists']]),\\n ] for item in abs]\\n table.table_data.extend(table_data)\\n print(table.table)\\n pass\\n\\n\\ndef print_mp3s(mp3s):\\n table = AsciiTable([[\\\"ID\\\", \\\"Name\\\", \\\"Album\\\", \\\"AlbumID\\\", \\\"Artist\\\", \\\"ArtistID\\\"]])\\n table_data = [[str(item['id']), item['name'],\\n item['album']['name'], item['album']['id'],\\n ','.join([ar['name'] for ar in item['artists']]),\\n ','.join([str(ar['id']) for ar in item['artists']]),\\n ] for item in mp3s]\\n table.table_data.extend(table_data)\\n print(table.table)\\n pass\\n\\n\\ndef print_users(users):\\n table = AsciiTable([[\\\"ID\\\", \\\"Name\\\", \\\"Signature\\\"]])\\n table_data = [[str(item['userId']), item['nickname'],\\n item['signature'],\\n ] for item in users]\\n table.table_data.extend(table_data)\\n print(table.table)\\n pass\\n\\n\\ndef print_artists(artists):\\n table = AsciiTable([[\\\"ID\\\", \\\"Name\\\", \\\"AlbumNum\\\", \\\"MVNum\\\"]])\\n table_data = [[str(item['id']), item['name'],\\n str(item['albumSize']),\\n str(item['mvSize'])\\n ] for item in artists]\\n table.table_data.extend(table_data)\\n print(table.table)\\n pass\\n\\n\\ndef print_mvs(mvs):\\n table = AsciiTable([[\\\"ID\\\", \\\"Name\\\", \\\"Artist\\\", \\\"ArtistID\\\", \\\"Duration\\\", \\\"PlayCount\\\"]])\\n table_data = [[str(item['id']), item['name'],\\n item['artistName'],\\n item['artistId'],\\n '%02d:%02d' % divmod(int(item['duration'] / 1000), 60),\\n item['playCount'],\\n ] for item in mvs]\\n table.table_data.extend(table_data)\\n print(table.table)\\n pass\\n\"\n },\n {\n \"alpha_fraction\": 0.8409090638160706,\n \"alphanum_fraction\": 0.8409090638160706,\n \"avg_line_length\": 44,\n \"blob_id\": \"f18016d1d2f1a3befd4f983276bfe2c29e51ccdf\",\n \"content_id\": \"e33b80f3decd4f6314ca782a57810ba4599a45f1\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 44,\n \"license_type\": \"permissive\",\n \"max_line_length\": 44,\n \"num_lines\": 1,\n \"path\": \"/TODO.md\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"readme modify(add shortcut and add password)\"\n },\n {\n \"alpha_fraction\": 0.6499999761581421,\n \"alphanum_fraction\": 0.6558333039283752,\n \"avg_line_length\": 21.641510009765625,\n \"blob_id\": \"c28537a0d84c9a8c488b4bd9ae7b490d5b88fd61\",\n \"content_id\": \"2e89f6728ad634a408cb5dd6f10a37e24d1533bf\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1200,\n \"license_type\": \"permissive\",\n \"max_line_length\": 74,\n \"num_lines\": 53,\n \"path\": \"/NXSpider/bin/cli.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/6/5.\\n# email to LipsonChan@yahoo.com\\n#\\nimport datetime\\n\\nfrom cement.core.controller import CementBaseController, expose\\nfrom cement.core.foundation import CementApp\\n\\nfrom NXSpider.bin.config_ctrl import ConfigController\\nfrom NXSpider import version\\nfrom colorama import init\\n\\nfrom NXSpider.bin.show_ctrl import ShowController\\nfrom NXSpider.bin.spider_ctrl import SpiderController\\nfrom NXSpider.common.config import Config\\n\\n\\nBANNER = \\\"\\\"\\\"\\n{} Application v{}\\nCopyright (c) {} {}\\n\\\"\\\"\\\".format(version.__title__, version.__version__ ,\\n version.__author__, datetime.datetime.now().year)\\n\\ninit(autoreset=True)\\n\\n\\nclass VersionController(CementBaseController):\\n class Meta:\\n label = 'base'\\n description = ''\\n arguments = [\\n (['-v', '--version'], dict(action='version', version=BANNER)),\\n ]\\n\\n\\nclass App(CementApp):\\n class Meta:\\n label = \\\"NXSpider\\\"\\n base_controller = \\\"base\\\"\\n handlers = [VersionController, ConfigController,\\n SpiderController, ShowController]\\n\\n\\ndef main():\\n with App() as app:\\n app.run()\\n\\n\\nif __name__ == \\\"__main__\\\":\\n main()\\n\"\n },\n {\n \"alpha_fraction\": 0.6171428561210632,\n \"alphanum_fraction\": 0.6628571152687073,\n \"avg_line_length\": 18.44444465637207,\n \"blob_id\": \"f46c816fb50f98028b417b1ab5d853cf55857d7b\",\n \"content_id\": \"fd4f494187a4ef59c4370e5e325ca2af10a99728\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 175,\n \"license_type\": \"permissive\",\n \"max_line_length\": 41,\n \"num_lines\": 9,\n \"path\": \"/NXSpider/debug_test/config.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/20.\\n# email to LipsonChan@yahoo.com\\n#\\nfrom NXSpider.common.config import Config\\n\\npaths = Config().get_paths()\\n\"\n },\n {\n \"alpha_fraction\": 0.6215596199035645,\n \"alphanum_fraction\": 0.6567278504371643,\n \"avg_line_length\": 26.82978630065918,\n \"blob_id\": \"35e02726e2ce107392c9a82d8b6571f88488883d\",\n \"content_id\": \"37b99b1acbb50bb205229dbc611993566182e44a\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1388,\n \"license_type\": \"permissive\",\n \"max_line_length\": 99,\n \"num_lines\": 47,\n \"path\": \"/NXSpider/debug_test/idv3.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/5/16.\\n# email to LipsonChan@yahoo.com\\n#\\nimport locale\\n\\n# import eyed3\\nfrom mutagen.mp4 import MP4\\n\\nfrom NXSpider.utility.idv3_base import delete_163comment\\nfrom NXSpider.utility.modify_idv3 import attach_media_tag_by_path, attach_media_tag_by_conf_path\\n\\nfrom mutagen.id3 import ID3, TALB, COMM\\nfrom mutagen.easyid3 import EasyID3\\nTALB\\nCOMM\\nMP4\\n# ['TIT2', 'TPE1', 'TRCK', 'TALB', 'TPOS', 'TSSE', 'APIC:', 'COMM::XXX']\\n# 名称 歌手 trackNum 专辑 pos 编码 comments\\n# 'TPE2' 'aART' album_artist\\n\\nattach_media_tag_by_path(u'E:\\\\\\\\test')\\nattach_media_tag_by_conf_path()\\n\\n# a = ID3(u'E:\\\\\\\\薛之谦 - 丑八怪.mp3', v2_version=3)\\n#\\n# a = MP4(u'E:\\\\\\\\(G)I-DLE - LATATA.mp4')\\n#\\n# a = eyed3.load(u'E:\\\\\\\\薛之谦 - 丑八怪.mp3')\\n# a.tag.comments.remove('')\\n# a = eyed3.load(u'E:\\\\\\\\(G)I-DLE - LATATA.mp3')\\n# a = eyed3.load(u'E:\\\\\\\\(G)I-DLE - LATATA.mp4')\\n# a.tag.title(\\\"test\\\")\\n# a.tag.comments.remove('')\\n\\nlocale.getlocale(locale.LC_ALL)\\nfile = u\\\"E:/新建文件夹/Billion - Dancing Alone_wapi - 副本.mp3\\\"\\nfile = u\\\"D:\\\\\\\\Project\\\\\\\\python\\\\\\\\music.crawl\\\\\\\\download_files\\\\\\\\mp3\\\\\\\\1호선 뮤직\\\\\\\\1호선 뮤직 - 열대야.mp3\\\"\\nfile = u\\\"D:\\\\\\\\Project\\\\\\\\python\\\\\\\\music.crawl\\\\\\\\download_files\\\\\\\\mp3\\\\\\\\Dal★shabet\\\\\\\\Dal★shabet - JOKER.mp3\\\"\\ndelete_163comment(file)\\nprint('a')\\n\\nattach_media_tag_by_conf_path()\\n\\nprint('a')\\n\"\n },\n {\n \"alpha_fraction\": 0.5739436745643616,\n \"alphanum_fraction\": 0.5850487351417542,\n \"avg_line_length\": 27.604650497436523,\n \"blob_id\": \"b6815bacf89ab5be051803d8579230d0e2a05f31\",\n \"content_id\": \"3394e3a705dc9c092f4303b337cab3e3e6f6d43e\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3692,\n \"license_type\": \"permissive\",\n \"max_line_length\": 91,\n \"num_lines\": 129,\n \"path\": \"/NXSpider/utility/modify_idv3.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/5/17.\\n# email to LipsonChan@yahoo.com\\n#\\nimport os\\nimport shutil\\n\\nfrom NXSpider.common.config import Config\\nfrom NXSpider.model.mongo_model import Mp3Model, Mp4Model\\nfrom NXSpider.spider.mp3 import Mp3\\nfrom NXSpider.spider.mv import MV\\nfrom NXSpider.common import tools, log, constant\\nfrom NXSpider.spider.base_driver import Music163Obj\\nfrom NXSpider.utility.media_tag import attach_media_tag\\n\\nmedia_types = {\\n 'mp3': [Mp3Model, Mp3],\\n 'mp4': [Mp4Model, MV],\\n}\\npaths = Config().get_paths()\\n\\n\\ndef attach_media_idv3_by_db():\\n \\\"\\\"\\\"\\n attach media idv3 by paths in config\\n :return:\\n \\\"\\\"\\\"\\n for suffix, type_setting in media_types:\\n model, driver = type_setting # type: Mp3Model or Mp4Model, Music163Obj\\n objs = model.objects(downloaded=True)\\n for obj in objs:\\n file_path = driver.download_check(obj, check_file=True)\\n if not file_path:\\n continue\\n\\n if is_latin1(file_path):\\n res = attach_media_tag(obj, file_path)\\n else:\\n res = attach_shadow(file_path, suffix, obj)\\n\\n log.print_info('idv3 attach %s, file: %s'\\n % ('success' if res else 'failed', file_path))\\n\\n\\n@tools.ignored(Exception)\\ndef load_media_obj(media_type, download_name):\\n return media_types[media_type][0].objects(download_file_name=download_name).first()\\n\\n\\ndef create_tmp_dir():\\n path = os.path.join(constant.main_dir, 'tmp')\\n if not os.path.exists(path):\\n os.makedirs(path)\\n return path\\n\\n\\ndef delete_tmp_dir():\\n path = os.path.join(constant.main_dir, 'tmp')\\n if os.path.exists(path):\\n os.removedirs(path)\\n\\n\\ndef is_latin1(s):\\n try:\\n s.encode(\\\"iso-8859-1\\\")\\n return True\\n except UnicodeEncodeError:\\n return False\\n\\n\\ndef attach_media_tag_by_path(path):\\n \\\"\\\"\\\"\\n please run in python3 if your os is windows, cause os.walk has a encoding bug\\n :type path: str\\n :param path:\\n :return:\\n \\\"\\\"\\\"\\n\\n for root, dirs, files in os.walk(path):\\n for file in files: # type: str\\n suffix_i = file.rfind('.')\\n suffix = file[suffix_i + 1:]\\n if suffix not in media_types:\\n continue\\n\\n split_txt = ' - '\\n split_i = file.find(split_txt)\\n if split_i == -1:\\n continue\\n\\n artist = file[:split_i]\\n download_file_name = os.path.join(artist, file)\\n obj = load_media_obj(suffix, download_file_name)\\n file_path = os.path.join(root, file)\\n\\n # fuck!!!! although i fix python-magic encoding bug of Chinese str in windows,\\n # the fuck libmagic.dll doesn't recognize Korean. so, i am really pissed off to\\n # change file name and change back\\n '''\\n if is_latin1(file_path):\\n res = attach_media_tag(obj, file_path)\\n else:\\n res = attach_shadow(file_path, suffix, obj)\\n '''\\n\\n # Dobby is free, thanks to mutagen\\n res = attach_media_tag(obj, file_path)\\n\\n log.print_info('idv3 attach %s, file: %s'\\n % ('success' if res else 'failed', file_path))\\n\\n\\ndef attach_shadow(file, suffix, obj):\\n tmp_name = 'abc'\\n tmp_file = os.path.join(create_tmp_dir(), tmp_name + '.' + suffix)\\n shutil.copy(file, tmp_file)\\n res = attach_media_tag(obj, tmp_file)\\n if res:\\n os.remove(file)\\n shutil.copy(tmp_file, file)\\n os.remove(tmp_file)\\n return res\\n\\n\\ndef attach_media_tag_by_conf_path():\\n for path in paths:\\n attach_media_tag_by_path(path)\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5217254161834717,\n \"alphanum_fraction\": 0.5456548929214478,\n \"avg_line_length\": 22.701492309570312,\n \"blob_id\": \"7be0b87f30c2310ff4afd9df912a7f946d96a6f7\",\n \"content_id\": \"1844da35769a2acbc61697b96657038cfe5d963c\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3192,\n \"license_type\": \"permissive\",\n \"max_line_length\": 80,\n \"num_lines\": 134,\n \"path\": \"/NXSpider/utility/idv3_base.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/5/16.\\n# email to LipsonChan@yahoo.com\\n#\\nimport os\\n\\nimport eyed3\\nfrom eyed3.id3 import ID3_V2_3\\nfrom mutagen.id3 import ID3, Encoding\\nfrom mutagen.mp4 import MP4\\nfrom mutagen import id3\\n\\nkey_names = ['title', 'artist', 'album', 'album_artist', 'track_num', 'comment']\\nencoding_keys = ['title', 'artist', 'album', 'album_artist', 'comment']\\n# ['TIT2', 'TPE1', 'TRCK', 'TALB', 'TPOS', 'TSSE', 'APIC:', 'COMM::XXX']\\n# 名称 歌手 trackNum 专辑 pos 编码 comments\\n# 'TPE2' 'aART' album_artist\\nmutagen_idv3_key_map = {\\n 'title': 'TIT2',\\n 'artist': 'TPE1',\\n 'album': 'TALB',\\n 'album_artist': 'TPE2',\\n 'track_num': 'TRCK',\\n 'commemt': 'COMM'\\n}\\n\\nmugagen_mp4_key_map = {\\n 'title': '\\\\xa9nam',\\n 'artist': '\\\\xa9ART',\\n 'comment': '\\\\xa9cmt',\\n}\\n\\n\\ndef delete_163comment(file):\\n \\\"\\\"\\\"\\n :type file: str\\n :param file:\\n :return:\\n \\\"\\\"\\\"\\n if not os.path.exists(file):\\n return\\n suffix = file[file.rfind('.'):]\\n if suffix not in ['.mp3', '.mp4']:\\n return\\n\\n try:\\n id3_file = eyed3.load(file)\\n id3_file.initTag(version=ID3_V2_3)\\n id3_file.tag.comments.remove('', lang=b'XXX')\\n id3_file.tag.save()\\n except Exception as e:\\n raise e\\n\\n\\ndef attach_mp3_idv3(filename, data):\\n\\n\\n try:\\n mp3 = ID3(filename, v2_version=3)\\n for k, v in data.items():\\n if k not in mutagen_idv3_key_map:\\n continue\\n\\n if k == 'comment':\\n mp3.add(id3.COMM(lang='XXX', text=v))\\n continue\\n\\n attr_type = getattr(id3, mutagen_idv3_key_map[k], None)\\n if attr_type:\\n mp3.add(attr_type(text=v))\\n\\n mp3.save()\\n except Exception as e:\\n return False\\n return True\\n\\n\\ndef attach_mp4_tag(filename, data):\\n if not os.path.exists(filename):\\n return False\\n\\n try:\\n mp4 = MP4(filename)\\n for k, v in data.items:\\n if k not in mugagen_mp4_key_map:\\n continue\\n mp4[mugagen_mp4_key_map[k]] = v\\n\\n mp4.save()\\n except Exception as e:\\n return False\\n return True\\n\\n\\ndef attach_media_idv3(filename, data, encoding_func=None):\\n \\\"\\\"\\\"\\n attach data to idv3 of a file\\n :param encoding_func: change encoding of values\\n :type encoding_func: function | None\\n :type filename: str\\n :param filename:\\n :type data: dict\\n :param data:\\n :return:\\n \\\"\\\"\\\"\\n if not os.path.exists(filename):\\n return\\n suffix = filename[filename.rfind('.'):]\\n if suffix not in ['.mp3', '.mp4']:\\n return\\n\\n try:\\n id3_file = eyed3.load(filename)\\n id3_file.initTag(version=ID3_V2_3)\\n\\n for k, v in data.items():\\n if k not in key_names:\\n continue\\n\\n if encoding_func and k in encoding_keys:\\n v = encoding_func(v)\\n\\n if k == 'comment':\\n id3_file.tag.comments.set(v, description=u'', lang=b'XXX')\\n continue\\n\\n setattr(id3_file.tag, k, v)\\n id3_file.tag.save()\\n except Exception as e:\\n return False\\n\\n return True\\n\"\n },\n {\n \"alpha_fraction\": 0.5893169641494751,\n \"alphanum_fraction\": 0.5980735421180725,\n \"avg_line_length\": 12.75903606414795,\n \"blob_id\": \"b9312747257c76e58e1a96dd24c26b58a4bc54db\",\n \"content_id\": \"00024bb729b062fe31c7b5f136b605884c043357\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1142,\n \"license_type\": \"permissive\",\n \"max_line_length\": 42,\n \"num_lines\": 83,\n \"path\": \"/NXSpider/model/dict_model.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/19.\\n# email to LipsonChan@yahoo.com\\n#\\n\\n\\nclass BaseDoc(dict):\\n def save(self):\\n pass\\n\\n def __getattr__(self, item):\\n if item not in self:\\n return None\\n\\n value = self[item]\\n if isinstance(value, dict):\\n value = BaseDoc(value)\\n return value\\n\\n\\ndef get_one_model_by_key(model, model_id):\\n \\\"\\\"\\\"\\n load or create a model by id\\n :param model:\\n :param model_id:\\n :return: doc, is_new\\n :rtype: (DynamicDocument, boolean)\\n \\\"\\\"\\\"\\n\\n return model(id=model_id), True\\n\\n\\ndef update_dynamic_doc(doc, data):\\n \\\"\\\"\\\"\\n :type doc: dict\\n :type data: dict\\n :param doc:\\n :param data:\\n :return:\\n \\\"\\\"\\\"\\n doc.update(data)\\n\\n\\nclass ConfigModel(BaseDoc):\\n pass\\n\\n\\nclass AuthorModel(BaseDoc):\\n pass\\n\\n\\nclass Mp3Model(BaseDoc):\\n pass\\n\\n\\nclass ArtistModel(BaseDoc):\\n pass\\n\\n\\nclass Mp4Model(BaseDoc):\\n pass\\n\\n\\nclass VideoModel(BaseDoc):\\n pass\\n\\n\\nclass PlaylistModel(BaseDoc):\\n pass\\n\\n\\nclass AlbumModel(BaseDoc):\\n pass\\n\\n\\nclass UserModel(BaseDoc):\\n pass\\n\\n\\nclass Downloaded(BaseDoc):\\n pass\\n\"\n },\n {\n \"alpha_fraction\": 0.6626602411270142,\n \"alphanum_fraction\": 0.71875,\n \"avg_line_length\": 25.02083396911621,\n \"blob_id\": \"c9fc4189718da3f0c1194de25751a5e944384b8e\",\n \"content_id\": \"18c8965992575f8a0b1a8f57ffc742cfa213ea2f\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 2290,\n \"license_type\": \"permissive\",\n \"max_line_length\": 109,\n \"num_lines\": 48,\n \"path\": \"/VERSION.md\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"NXSpider\\n=================\\n\\nNXSpider， NetEase X Spider，一个强大的（网易云音乐）mp3,mv爬虫，可以下载和收集mp3,mv信息,同时附带多媒体标签信息。采用python编写，mongo数据库(非必须)，递归算法核心实现\\n\\n[](LICENSE.md)\\n\\n\\n\\n[新手先看这里](SIMPLE_USE.md) | [历史版本在这里](VERSION.md) | [开发详情在这里](DEV.md)\\n\\n## Update History\\n\\n#### 2018-12-17 v0.0.5\\n- 增加功能，通过配置，创建歌单，歌手，唱片目录系统（快捷方式）\\n- 增加功能，通过歌曲id，爬取歌曲\\n- 增加功能（失败），通过登录爬取收藏视频（非MV），但由于接口问题，未能使用\\n- 优化说明文档和安装脚本，使安装和更新更容易\\n\\n#### 2018-06-23 v0.0.4\\n- 增加功能，通过登录爬取用户收藏mv\\n- 增加功能，通过登录爬取用户所有歌单（包括收藏）\\n- 增加功能，通过歌单分类，爬取最火歌单\\n- 增加功能，打印歌单分类\\n- 增加功能，爬取最火mv\\n- 优化代码，优化所有文档，优化输出\\n- 解决小bug：mv id存在但获取不到详情时的代码异常\\n- 默认mongo非使用，且不安装python依赖包，通过 `pip+req.txt` 进行安装\\n- 方便windows用户增加小部分bat\\n- 项目功能已满足作者使用，暂时没什么开发欲望了\\n\\n#### 2018-06-21 v0.0.3\\n- 增加功能，通过用户id获得歌单，并爬取歌单信息，mp3,mv\\n- 增加功能，通过歌手id爬取所有专辑信息，下载所有专辑mp3,mv\\n- 增加 `VERSION.md`\\n\\n#### 2018-06-20 v0.0.2\\n- mongodb改为非必须，默认不需要，并提供相应配置\\n- 增加功能，爬取特定分类最火歌单信息并下载mp3,mv\\n- 增加 `SIMPLE_USE.md` 提供最简单的入门说明\\n- 优化部分代码\\n\\n#### 2018-06-19 v0.0.1\\n- 首次发布, mongodb + 递归 + python(3) + tag\\n- 搜索歌单，歌手，专辑，用户等信息\\n- 通过歌单，歌手，专辑id，下载mp3, mv\\n- 所有信息会保存早mongodb中\\n- 高可配置，所有mp3, mv可以通过配置，增加多媒体标签信息(歌手，专辑，唱片，163comment!!!)\"\n },\n {\n \"alpha_fraction\": 0.5491606593132019,\n \"alphanum_fraction\": 0.5755395889282227,\n \"avg_line_length\": 18.85714340209961,\n \"blob_id\": \"530e7a9bfccc5dadbabf8c69e909dd2b5dfd553c\",\n \"content_id\": \"4b47fcab01e8332da4b185ae8a9d4749d9b4ca5f\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 417,\n \"license_type\": \"permissive\",\n \"max_line_length\": 41,\n \"num_lines\": 21,\n \"path\": \"/NXSpider/spider/user.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/25.\\n# email to LipsonChan@yahoo.com\\n#\\nfrom NXSpider.spider.base_driver import *\\n\\n\\nclass User(Music163Obj):\\n __model_name__ = UserModel\\n __model__rfilter__ = {\\n 'userId',\\n\\n # garbage properties\\n 'avatarImgId',\\n 'backgroundImgId',\\n 'avatarImgIdStr',\\n 'backgroundImgIdStr',\\n 'avatarImgId_str',\\n }\\n\"\n },\n {\n \"alpha_fraction\": 0.5626780390739441,\n \"alphanum_fraction\": 0.5769230723381042,\n \"avg_line_length\": 24.962963104248047,\n \"blob_id\": \"04697374363e500c53ab8b0b11ee5d1d0825f07c\",\n \"content_id\": \"710eb2425b4a0a8a046e80e1606da6b350d00bbe\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 702,\n \"license_type\": \"permissive\",\n \"max_line_length\": 78,\n \"num_lines\": 27,\n \"path\": \"/NXSpider/debug_test/change_mp3_to_mp4.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/5/15.\\n# email to LipsonChan@yahoo.com\\n#\\nimport os\\nimport shutil\\n\\npaths = ['mv']\\nproject_path = os.path.dirname(os.path.realpath(__file__))\\nproject_path = os.path.dirname(project_path)\\ndownload_path = os.path.join(project_path, 'download_files')\\n\\n\\ndef rename_file(path):\\n for root, dirs, files in os.walk(path):\\n for f in files:\\n f = f # type: str\\n suffix = f[f.rfind('.'):]\\n if suffix == '.mp3':\\n new_name = f[:f.rfind('.')] + \\\".mp4\\\"\\n os.rename(os.path.join(root, f), os.path.join(root, new_name))\\n\\n\\nfor p in paths:\\n rename_file(os.path.join(download_path, p))\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.439113050699234,\n \"alphanum_fraction\": 0.6412213444709778,\n \"avg_line_length\": 51.903846740722656,\n \"blob_id\": \"4d9428d2ea1a15e718af14fa2680b79d4b04a47c\",\n \"content_id\": \"96fe0f6ea777e48aff68fb75b1c0244e67f8bd11\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2751,\n \"license_type\": \"permissive\",\n \"max_line_length\": 314,\n \"num_lines\": 52,\n \"path\": \"/NXSpider/debug_test/aes_decode.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/5/15.\\n# email to LipsonChan@yahoo.com\\n#\\nimport base64\\n\\nfrom NXSpider.common.encrypt import aes_ecb, aes_ecb_decode\\n\\nBS = 16\\npad = lambda s: s + (BS - len(s) % BS) * chr(0)\\nunpad = lambda s: s[0:-ord(s[-1])]\\n\\naes_code = \\\"#14ljk_!\\\\]&0U<'(\\\"\\n\\nciphertext = \\\"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\\\"\\nciphertext = \\\"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\\\"\\nciphertext = \\\"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\\\"\\nciphertext = \\\"wUxNl9+oaCX1+TGUoqXxPMW1W97D0rB/UWviTQG1qGf7VPc9xScpSdO78dNI29kCFN9qWTWmEEe2JvA0TgeZBdmmiikLL8RQ6zvuAfKjzTOJhL6Ff1X4hw6syUlw+wqQwRL0WQvY6ysW/OlDmqa/GHaZOirjQv580Xplu9P2wBfxW3T8HrjzPdtS05xs2Pa6JgoErDzGhmlv7ny0yYWkHqwfFtlDDlwUTAAEOXhEFI/yNE5qt4jaQABURo9ZCBoircTF92wi2BSkQHoTzDgarpOnvzmafSOR+46gt3OolILhXrtef/wPpXih9gkLrKY0\\\"\\nciphertext = \\\"I7ZTqvekU9UjuYios3KeugpKIryhEim4OTsU7t3xaN2ZBcRHRQ0XzpsDmVa76Sh8aJ4JOmBcD37oN9sLBCkqOpoGAVROkWpNfEsYTsLz13E4Q4OIT+BBn9Acvw9CNtviGXBUyDkeec3x4RGuI6T/efUn+MntkU2ZkyyIt7VhcoV7XTFGyFx9A2UB72Tw/vn2iRXeTeS0wuAJqxJN6b9GINrizT8a+QORT7FoyPl7e4Zf8Sg4E51UnhuOGtulNc8n7MnOkbyfJSZWfHD2sHRi6g==\\\"\\nplaintext = 'music:{\\\"albumId\\\":2084576,\\\"alias\\\":[],\\\"mMusic\\\":\\\"{\\\\\\\\\\\"fid\\\\\\\\\\\":2007708232328038,\\\\\\\\\\\"br\\\\\\\\\\\":160000,' \\\\\\n '\\\\\\\\\\\"size\\\\\\\\\\\":10701396}\\\",\\\"album\\\":\\\"Eternal Light\\\",\\\"musicId\\\":22712173,\\\"musicName\\\":\\\"Refrain\\\",\\\"mvId\\\":0,' \\\\\\n '\\\"duration\\\":532000,\\\"lMusic\\\":\\\"{\\\\\\\\\\\"fid\\\\\\\\\\\":2008807743955814,\\\\\\\\\\\"br\\\\\\\\\\\":96000,\\\\\\\\\\\"size\\\\\\\\\\\":6437585}\\\",' \\\\\\n '\\\"bitrate\\\":320000,\\\"albumPic\\\":\\\"http://p2.music.126.net/fNtMX44fvaGByURP0AbOZQ==/836728348761063.jpg\\\",' \\\\\\n '\\\"mp3DocId\\\":1981319953261764,\\\"artist\\\":[[\\\"Anan Ryoko\\\",16069]],\\\"hMusic\\\":\\\"{\\\\\\\\\\\"fid\\\\\\\\\\\":1981319953261764,' \\\\\\n '\\\\\\\\\\\"br\\\\\\\\\\\":320000,\\\\\\\\\\\"size\\\\\\\\\\\":21359877}\\\",\\\"albumPicDocId\\\":836728348761063}\\\"'\\nplaintext = 'music:{\\\"musicId\\\":28341255,\\\"musicName\\\":\\\"Dancing Alone\\\",\\\"albumId\\\":2774171,\\\"album\\\":\\\"Remove Before Flight\\\",\\\"albumPicDocId\\\":6025323720389239,\\\"albumPic\\\":\\\"http://p1.music.126.net/w0uwPgaqDCJYYZJjzQfa7w==/6025323720389239.jpg\\\",\\\"bitrate\\\":320000,\\\"mp3DocId\\\":\\\"69bb414114182d8e7b8a5214168bc991\\\",\\\"duration\\\":211200}'\\nplaintext = 'music:{\\\"musicId\\\":28341255,\\\"musicName\\\":\\\"Dancing Alone\\\",' \\\\\\n '\\\"albumId\\\":2774171,\\\"album\\\":\\\"Remove Before Flight\\\",\\\"albumPicDocId\\\":6025323720389239,' \\\\\\n '\\\"albumPic\\\":\\\"http://p1.music.126.net/w0uwPgaqDCJYYZJjzQfa7w==/6025323720389239.jpg\\\",' \\\\\\n '\\\"bitrate\\\":320000,\\\"mp3DocId\\\":\\\"69bb414114182d8e7b8a5214168bc991\\\",\\\"duration\\\":211200}'\\nplaintext = 'music:{\\\"musicId\\\":28341255,\\\"musicName\\\":\\\"Dancing Alone\\\",' \\\\\\n '\\\"albumId\\\":2774171,\\\"album\\\":\\\"haha\\\",\\\"albumPicDocId\\\":6025323720389239,' \\\\\\n '\\\"bitrate\\\":320000,\\\"mp3DocId\\\":\\\"69bb414114182d8e7b8a5214168bc991\\\",\\\"duration\\\":211200}'\\n# plaintext = 'music:{\\\"musicId\\\":28341255,\\\"musicName\\\":\\\"Dancing Alone\\\",' \\\\\\n# '\\\"bitrate\\\":320000}'\\n# plaintext = 'mv:{\\\"title\\\":\\\"LATATA\\\",\\\"mvId\\\":5906035,\\\"artistId\\\":127524,' \\\\\\n# '\\\"artistName\\\":\\\"Rainbow\\\",\\\"picId\\\":5991238859751539,\\\"pubTime\\\":\\\"2013-05-17\\\",' \\\\\\n# '\\\"bitrate\\\":240,\\\"duration\\\":258000,\\\"briefIntro\\\":\\\"\\\",\\\"detailIntro\\\":\\\"\\\"}'\\n# plaintext = 'mv:{\\\"title\\\":\\\"LATATA\\\",\\\"mvId\\\":5906035,\\\"artistId\\\":14055085,' \\\\\\n# '\\\"artistName\\\":\\\"(G)I-DLE\\\",\\\"picId\\\":109951163279867003,\\\"pubTime\\\":\\\"2018-05-02\\\",' \\\\\\n# '\\\"bitrate\\\":720,\\\"duration\\\":221000,\\\"briefIntro\\\":\\\"\\\",\\\"detailIntro\\\":\\\"\\\"}'\\n\\nb = aes_code.encode('utf-8')\\na = base64.b64decode(ciphertext)\\na = aes_ecb_decode(ciphertext, aes_code)\\nc = aes_ecb(plaintext, aes_code)\\nd = c.decode()\\n\\nprint(d)\\nprint(d)\\n\"\n },\n {\n \"alpha_fraction\": 0.5011028051376343,\n \"alphanum_fraction\": 0.507278323173523,\n \"avg_line_length\": 26.64634132385254,\n \"blob_id\": \"6c8d9a50fde0455f8c3b2dcacce2089db3fcd7f6\",\n \"content_id\": \"1fda559ba5125955eedbcab03ba9b7182cde17ff\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2267,\n \"license_type\": \"permissive\",\n \"max_line_length\": 87,\n \"num_lines\": 82,\n \"path\": \"/NXSpider/spider/mv.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/20.\\n# email to LipsonChan@yahoo.com\\n#\\nimport os\\nimport re\\n\\nfrom NXSpider.common.config import Config\\nfrom NXSpider.spider.api import get_mv_link\\nfrom NXSpider.spider.artist import Artist\\nfrom NXSpider.spider.base_driver import *\\n\\n\\ndef get_target_r(obj, limit_r=Config().get_mv_resolution()):\\n max_valid = max([x['br'] for x in obj['brs']])\\n return limit_r if max_valid > limit_r else max_valid\\n\\n\\nclass MV(Music163Obj):\\n __model_name__ = Mp4Model\\n __model_rfilter__ = {\\n }\\n __parse_recursion__ = {\\n 'artists': Artist(),\\n }\\n\\n def download_filename(self, doc):\\n \\\"\\\"\\\"\\n implement pls\\n get a name to save file\\n need be complete by child\\n :param doc:\\n :return:\\n \\\"\\\"\\\"\\n authors = \\\",\\\".join([x['name'] for x in doc.artists])\\n author = re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', authors.strip())\\n mv_name = re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', doc['name'])\\n name = \\\"%s - %s.mp4\\\" % (author, mv_name)\\n return name\\n\\n def download_filename_full(self, doc):\\n \\\"\\\"\\\"\\n implement pls\\n get a path to save file, by relative path\\n need be complete by child\\n :param doc:\\n :return:\\n :rtype: str\\n \\\"\\\"\\\"\\n authors = \\\",\\\".join([x['name'] for x in doc.artists])\\n author = re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', authors.strip())\\n mv_name = re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', doc['name'])\\n name = os.path.join(author, \\\"%s - %s.mp4\\\" % (author, mv_name))\\n return name\\n\\n def url_load(self, doc):\\n \\\"\\\"\\\"\\n implement pls\\n :param doc:\\n :return:\\n :rtype: str\\n \\\"\\\"\\\"\\n try:\\n target_r = get_target_r(doc, Config().get_mv_resolution())\\n doc['download_video_r'] = target_r\\n return get_mv_link(doc['id'], target_r)\\n except:\\n return None\\n\\n def shortcut_self_path(self, doc):\\n \\\"\\\"\\\"\\n implement pls, not force\\n return self short cut path\\n :param doc:\\n :return:\\n \\\"\\\"\\\"\\n result = []\\n result.extend([os.path.join(\\\"artist\\\", re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', x['name']))\\n for x in doc.artists])\\n return result\\n\"\n },\n {\n \"alpha_fraction\": 0.5629921555519104,\n \"alphanum_fraction\": 0.586614191532135,\n \"avg_line_length\": 20.16666603088379,\n \"blob_id\": \"5c3dea037165cd937978439b869b6f853a5d2094\",\n \"content_id\": \"00b7c92bb06e4132dbbd5c12886224d6a47cd901\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 508,\n \"license_type\": \"permissive\",\n \"max_line_length\": 52,\n \"num_lines\": 24,\n \"path\": \"/NXSpider/spider/album.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/26.\\n# email to LipsonChan@yahoo.com\\n#\\n\\nfrom NXSpider.spider.artist import Artist\\nfrom NXSpider.spider.base_driver import *\\n\\n\\nclass Album(Music163Obj):\\n __model_name__ = AlbumModel\\n __model_rfilter__ = {\\n # 'song', # TODO need delete???\\n 'artist',\\n }\\n __parse_recursion__ = {\\n 'artists': Artist(),\\n }\\n\\n @attr_replace(attr_name='songs', new_name='mp3')\\n def replace_song(self, obj):\\n return obj\\n\"\n },\n {\n \"alpha_fraction\": 0.6957504749298096,\n \"alphanum_fraction\": 0.7034357786178589,\n \"avg_line_length\": 23.842697143554688,\n \"blob_id\": \"3211e6d00cef7ff32038c21e5ffda2e4202fc108\",\n \"content_id\": \"5b0a6e0dcda0ff66346e7a48dd3460c21e78492a\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 2212,\n \"license_type\": \"permissive\",\n \"max_line_length\": 78,\n \"num_lines\": 89,\n \"path\": \"/NXSpider/common/tools.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n\\n\\\"\\\"\\\"\\nMIT License\\n\\nCopyright (c) 2017 Cheng YuMeng\\n\\nPermission is hereby granted, free of charge, to any person obtaining a copy\\nof this software and associated documentation files (the \\\"Software\\\"), to deal\\nin the Software without restriction, including without limitation the rights\\nto use, copy, modify, merge, publish, distribute, sublicense, and/or sell\\ncopies of the Software, and to permit persons to whom the Software is\\nfurnished to do so, subject to the following conditions:\\n\\nThe above copyright notice and this permission notice shall be included in all\\ncopies or substantial portions of the Software.\\n\\nTHE SOFTWARE IS PROVIDED \\\"AS IS\\\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\\nIMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\\nFITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\\nAUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\\nLIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\\nOUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\\nSOFTWARE.\\n\\\"\\\"\\\"\\n\\nimport codecs\\nimport contextlib\\nimport hashlib\\nimport locale\\n\\nfrom NXSpider.common import PYTHON3\\n\\nRETURN_JSON = \\\"return json data\\\"\\nRETURE_HTML = \\\"return html data\\\"\\n\\nlang, sys_encoding = locale.getdefaultlocale()\\n\\n@contextlib.contextmanager\\ndef ignored(*exceptions):\\n try:\\n yield\\n except exceptions:\\n pass\\n\\n\\ndef encode(s):\\n if PYTHON3 is True:\\n return codecs.encode(s, \\\"utf-8\\\").decode(\\\"utf-8\\\")\\n else:\\n return s.encode(\\\"utf-8\\\")\\n\\n\\ndef encode_sys(s):\\n s.encode(sys_encoding)\\n\\n\\ndef hex(s):\\n if PYTHON3 is True:\\n return codecs.encode(bytes(s, encoding=\\\"utf8\\\"), 'hex')\\n else:\\n return s.encode(\\\"hex\\\")\\n\\n\\ndef md5(s):\\n m = hashlib.md5()\\n m.update(s.encode(\\\"utf-8\\\"))\\n return m.hexdigest()\\n\\n\\ndef is_unicode(obj):\\n if PYTHON3:\\n return isinstance(obj, str)\\n else:\\n from __builtin__ import unicode\\n return isinstance(obj, unicode)\\n\\n\\ndef input_format(string):\\n \\\"\\\"\\\"\\n :type string: str\\n :param string:\\n :return:\\n \\\"\\\"\\\"\\n if PYTHON3:\\n return string\\n else:\\n return string.encode()\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.5391644835472107,\n \"alphanum_fraction\": 0.5489556193351746,\n \"avg_line_length\": 25.413793563842773,\n \"blob_id\": \"122dc14c3630ef281f91bcdf46677b4d920a59d1\",\n \"content_id\": \"154f9f791c283ba1c48c05f04736b6da3361597c\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1532,\n \"license_type\": \"permissive\",\n \"max_line_length\": 71,\n \"num_lines\": 58,\n \"path\": \"/NXSpider/spider/video.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/12/12.\\n# email to LipsonChan@yahoo.com\\n#\\n\\nimport os\\nimport re\\n\\nfrom NXSpider.common.config import Config\\nfrom NXSpider.spider.api import get_video_link\\nfrom NXSpider.spider.artist import Artist\\nfrom NXSpider.spider.base_driver import *\\n\\n\\ndef get_target_r(obj, limit_r=Config().get_mv_resolution()):\\n max_valid = max([x['resolution'] for x in obj['resolutions']])\\n return limit_r if max_valid > limit_r else max_valid\\n\\n\\nclass Video(Music163Obj):\\n __model_name__ = VideoModel\\n __model_rfilter__ = {\\n }\\n __parse_recursion__ = {\\n 'artists': Artist(),\\n }\\n\\n def download_filename_full(self, doc):\\n \\\"\\\"\\\"\\n implement pls\\n get a path to save file, by relative path\\n need be complete by child\\n :param doc:\\n :return:\\n :rtype: str\\n \\\"\\\"\\\"\\n # todo modify\\n authors = \\\",\\\".join([x['name'] for x in doc.artists])\\n author = re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', authors.strip())\\n mp3_name = re.sub(\\\"[\\\\\\\\\\\\\\\\/:*?\\\\\\\"<>|]\\\", '', doc['name'])\\n name = os.path.join(author, \\\"%s - %s.mp4\\\" % (author, mp3_name))\\n return name\\n\\n def url_load(self, doc):\\n \\\"\\\"\\\"\\n implement pls\\n :param doc:\\n :return:\\n :rtype: str\\n \\\"\\\"\\\"\\n try:\\n target_r = get_target_r(doc, Config().get_mv_resolution())\\n doc['download_video_r'] = target_r\\n return get_video_link(doc['id'], target_r)\\n except:\\n return None\\n\"\n },\n {\n \"alpha_fraction\": 0.49380558729171753,\n \"alphanum_fraction\": 0.5022236108779907,\n \"avg_line_length\": 30.01477813720703,\n \"blob_id\": \"a1184863c446c795f067e5b8a110fb09ef6a9da3\",\n \"content_id\": \"e80ff9c0e9f50356db461db01d919055b25003ec\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 6296,\n \"license_type\": \"permissive\",\n \"max_line_length\": 104,\n \"num_lines\": 203,\n \"path\": \"/NXSpider/common/config.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/19.\\n# email to LipsonChan@yahoo.com\\n#\\nimport json\\nimport os\\n\\nfrom NXSpider.common import log\\nfrom NXSpider.common.constant import config_path, default_download_dir\\nfrom NXSpider.common.singleton import Singleton\\n\\ndefault_path_key = 'default'\\nmv_resolutions = [240, 480, 720, 1080]\\n\\n\\ndef utf8_data_to_file(f, data):\\n if hasattr(data, 'decode'):\\n f.write(data.decode('utf-8'))\\n else:\\n f.write(data)\\n\\n\\nclass Config(Singleton):\\n def __init__(self):\\n if hasattr(self, '_init'):\\n return\\n self._init = True\\n self.config_file_path = config_path\\n\\n # the deep level must be less than 2\\n self.default_config = {\\n 'mongo': {\\n 'name': 'nxspider',\\n 'host': 'localhost',\\n 'port': 27017,\\n 'username': 'None',\\n 'password': 'None',\\n },\\n 'download_path': [default_download_dir],\\n 'mv_def_resolution': 720,\\n 'media_tag': True,\\n 'media_tag_163': True,\\n 'download_file_check': True,\\n 'debug_log': True,\\n 'no_mongo': True,\\n 'shortcut': False,\\n }\\n self.config = {}\\n\\n need_save = False\\n try:\\n f = open(self.config_file_path, 'r')\\n self.config = json.loads(f.read())\\n f.close()\\n except IOError or ValueError:\\n self.__generate_config_file()\\n need_save = True\\n\\n # merge default\\n for k, v in self.default_config.items():\\n if k not in self.config:\\n self.config[k] = v\\n need_save = True\\n\\n paths = self.config['download_path'] # type: list\\n if len(paths) == 0:\\n paths.append(default_download_dir)\\n need_save = True\\n\\n if need_save:\\n self.save_config_file()\\n\\n def __getitem__(self, key):\\n if key in self.config:\\n return self.config[key]\\n return None\\n\\n def __generate_config_file(self):\\n f = open(self.config_file_path, 'w')\\n utf8_data_to_file(f, json.dumps(self.default_config, indent=2))\\n f.close()\\n\\n def save_config_file(self):\\n f = open(self.config_file_path, 'w')\\n utf8_data_to_file(f, json.dumps(self.config, indent=2))\\n f.close()\\n\\n def get_shortcut(self):\\n return self.config['shortcut']\\n\\n def get_mongo(self):\\n return self.config['mongo']\\n\\n def get_path(self):\\n return self.config['download_path'][0]\\n\\n def get_paths(self):\\n return self.config['download_path']\\n\\n def get_mv_resolution(self):\\n return self.config['mv_def_resolution']\\n\\n def get_media_tag(self):\\n return self.config['media_tag']\\n\\n def get_media_tag_163(self):\\n return self.config['media_tag_163']\\n\\n def get_file_check(self):\\n return self.config['download_file_check']\\n\\n def get_no_mongo(self):\\n return self.config['no_mongo']\\n\\n def save_config_dict(self, obj):\\n \\\"\\\"\\\"\\n :type obj: dict\\n :param obj:\\n :return:\\n \\\"\\\"\\\"\\n for k, v in obj.items():\\n if isinstance(v, dict) and isinstance(self.config[k], dict):\\n self.config[k].update(v)\\n else:\\n self.config[k] = v\\n\\n def config_test(self):\\n result = True\\n try:\\n\\n # check mongodb config\\n if self.config['no_mongo'] is False:\\n log.print_info('check mongodb config')\\n mongo = self.config['mongo']\\n for k in ['name', 'host', 'port']:\\n if k not in mongo:\\n log.print_err(\\\"mongo config error, key mongo.{} is not set yet\\\".format(k))\\n result = False\\n\\n # try import model, which will connect to server and exit if server config wrong\\n import NXSpider.model.mongo_model\\n\\n for k in ['download_path', 'mv_def_resolution', 'media_tag', 'media_tag_163']:\\n if k not in self.config:\\n log.print_err(\\\"config error, key {} is not set yet\\\".format(k))\\n result = False\\n\\n # check type\\n type_check = {\\n 'download_path': list,\\n 'mv_def_resolution': int,\\n 'media_tag': bool,\\n 'media_tag_163': bool,\\n 'download_file_check': bool,\\n 'no_mongo': bool,\\n }\\n\\n need_save = False\\n for k, v in type_check.items():\\n if not isinstance(self.config[k], v):\\n log.print_err(\\\"config error, {} is not a require type, \\\"\\n \\\"and is reset to default value: {}\\\".format(k, self.default_config[k]))\\n self.config[k] = self.default_config[k]\\n need_save = True\\n result = False\\n\\n # download path check\\n final_paths = []\\n for p in self.config['download_path']:\\n try:\\n # some error need pass\\n if os.path.isdir(p) is False:\\n os.mkdir(p)\\n final_paths.append(p)\\n except:\\n log.print_warn(\\\"download path may be wrong and be deleted: {}\\\".format(p))\\n need_save = True\\n result = False\\n pass\\n\\n # mv resolution check\\n if self.config['mv_def_resolution'] not in mv_resolutions:\\n log.print_warn(\\\"mv_def_resolution will be reset to default: {}\\\"\\n .format(self.default_config['mv_def_resolution']))\\n self.config['mv_def_resolution'] = self.default_config['mv_def_resolution']\\n need_save = True\\n result = False\\n\\n if need_save:\\n self.config['download_path'] = final_paths\\n self.save_config_file()\\n\\n return result\\n except Exception as e:\\n log.print_err(e)\\n\\n return False\\n\\n def config_reset(self):\\n self.config = self.default_config\\n self.save_config_file()\\n\"\n },\n {\n \"alpha_fraction\": 0.6059637665748596,\n \"alphanum_fraction\": 0.6448349356651306,\n \"avg_line_length\": 25.828571319580078,\n \"blob_id\": \"dd33d4596c0fc2bbcee9abb15c27b21f8b82e59b\",\n \"content_id\": \"abbdf977aa3e6baca5a2db2c08a58a98e54d815f\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1878,\n \"license_type\": \"permissive\",\n \"max_line_length\": 80,\n \"num_lines\": 70,\n \"path\": \"/NXSpider/common/encrypt.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/19.\\n# email to LipsonChan@yahoo.com\\n#\\n\\nimport base64\\n\\nfrom cryptography.hazmat.primitives.ciphers import (\\n Cipher, algorithms, modes\\n)\\nfrom cryptography.hazmat.backends import default_backend\\n\\n\\ndef aes(text, sec_key):\\n backend = default_backend()\\n pad = 16 - len(text) % 16\\n text_t = text + pad * chr(pad)\\n cipher = Cipher(\\n algorithms.AES(sec_key.encode('utf-8')),\\n modes.CBC(b'0102030405060708'),\\n backend=backend\\n )\\n encryptor = cipher.encryptor()\\n ciphertext = encryptor.update(text_t.encode('utf-8')) + encryptor.finalize()\\n ciphertext = base64.b64encode(ciphertext)\\n return ciphertext\\n\\n\\ndef aes_decode(text, sec_key):\\n backend = default_backend()\\n ciphertext = base64.b64decode(text)\\n cipher = Cipher(\\n algorithms.AES(sec_key.encode('utf-8')),\\n modes.CBC(b'0102030405060708'),\\n backend=backend\\n )\\n decryptor = cipher.decryptor()\\n plaintext = decryptor.update(ciphertext) + decryptor.finalize()\\n return plaintext\\n\\n\\ndef aes_ecb_decode(text, sec_key):\\n backend = default_backend()\\n ciphertext = base64.b64decode(text)\\n cipher = Cipher(\\n algorithms.AES(sec_key.encode('utf-8')),\\n modes.ECB(),\\n backend=backend\\n )\\n decryptor = cipher.decryptor()\\n plaintext = decryptor.update(ciphertext) + decryptor.finalize()\\n return plaintext\\n\\n\\ndef aes_ecb(text, sec_key):\\n backend = default_backend()\\n text = text.encode('utf-8')\\n pad = 16 - len(text) % 16\\n text_t = text + (b'\\\\0' * pad)\\n cipher = Cipher(\\n algorithms.AES(sec_key.encode('utf-8')),\\n modes.ECB(),\\n backend=backend\\n )\\n encryptor = cipher.encryptor()\\n ciphertext = encryptor.update(text_t) + encryptor.finalize()\\n ciphertext = base64.b64encode(ciphertext)\\n return ciphertext\\n\"\n },\n {\n \"alpha_fraction\": 0.5271945595741272,\n \"alphanum_fraction\": 0.5307171940803528,\n \"avg_line_length\": 29.856521606445312,\n \"blob_id\": \"a2ee77697e3d80288bcdabc164de0dbd0148d59c\",\n \"content_id\": \"777289756164b98bb187586869f34e6d4b7e0594\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 14196,\n \"license_type\": \"permissive\",\n \"max_line_length\": 86,\n \"num_lines\": 460,\n \"path\": \"/NXSpider/spider/base_driver.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/4/25.\\n# email to LipsonChan@yahoo.com\\n#\\nimport codecs\\nimport json\\nimport os\\nfrom inspect import isfunction\\n\\nimport requests\\nimport six\\n\\nfrom NXSpider.common.config import Config\\nfrom NXSpider.common import tools, log\\nfrom NXSpider.model.export import model_download_path\\nfrom NXSpider.utility.media_tag import attach_media_tag\\nfrom NXSpider.model.export import *\\n\\n\\nclass Music163ObjException(Exception):\\n pass\\n\\n\\nclass Music163ObjMetaClass(type):\\n def __new__(mcs, name, bases, attrs):\\n \\\"\\\"\\\"\\n :param name:\\n :type name: str\\n :param bases:\\n :param attrs:\\n :type attrs: dict\\n :return:\\n \\\"\\\"\\\"\\n if name == \\\"Music163Obj\\\":\\n return type.__new__(mcs, name, bases, attrs)\\n if '__model_name__' not in attrs or attrs['__model_name__'] is None:\\n attrs['__model_name__'] = name + \\\"Model\\\"\\n\\n # set download level 1 path\\n attrs['__file_type__'] = name.lower()\\n\\n # create download file dir for download path\\n # obj_path = os.path.join(get_download_path(), attrs['__file_type__'])\\n # if not os.path.exists(obj_path):\\n # os.makedirs(obj_path)\\n\\n # create a empty dict\\n attrs['__attrs_replace_fucs__'] = dict()\\n attrs['__attrs_replace_map__'] = dict()\\n\\n # set replace\\n replace_map = dict()\\n for k, v in attrs.items():\\n if isfunction(v):\\n if getattr(v, '__attrs_replace__', False):\\n org_name = getattr(v, '__attrs_replace_org__')\\n replace_map[org_name] = getattr(v, '__attrs_replace_new__')\\n attrs['__attrs_replace_fucs__'][org_name] = v\\n attrs['__attrs_replace_map__'] = replace_map\\n return type.__new__(mcs, name, bases, attrs)\\n\\n\\ndef attr_replace(attr_name=None, new_name=None):\\n \\\"\\\"\\\"\\n replace attr value\\n decorator for fun: (self, obj)->object\\n\\n :param new_name:\\n :param attr_name:\\n :return:\\n \\\"\\\"\\\"\\n\\n def de(f):\\n \\\"\\\"\\\"\\n :type f: function\\n :param f:\\n :return:\\n \\\"\\\"\\\"\\n f.__attrs_replace__ = True\\n def_name = f.__name__\\n f.__attrs_replace_org__ = attr_name if attr_name else def_name\\n f.__attrs_replace_new__ = new_name if new_name else def_name\\n return f\\n\\n return de\\n\\n\\n# class Music163Obj(object, metaclass=Music163ObjMetaClass):\\nclass Music163Obj(six.with_metaclass(Music163ObjMetaClass)):\\n __model_name__ = None\\n __model_rfilter__ = set()\\n __file_type__ = 'other'\\n __parse_recursion__ = {}\\n\\n __attrs_replace_fucs__ = dict()\\n __attrs_replace_map__ = dict()\\n\\n # __metaclass__ = Music163ObjMetaClass\\n\\n def __init__(self, *args, **kwargs):\\n \\\"\\\"\\\"\\n In some case, something want to change def class setting,\\n like recursion or filter\\n :param args:\\n :param kwargs:\\n \\\"\\\"\\\"\\n for k, v in kwargs.items():\\n if k in ['__parse_recursion__', '__model_rfilter__']:\\n setattr(self, k, v)\\n\\n def download_filename(self, doc):\\n \\\"\\\"\\\"\\n implement pls\\n get a name to save file\\n need be complete by child\\n :param doc:\\n :return:\\n :rtype: str\\n \\\"\\\"\\\"\\n return None\\n\\n def download_filename_full(self, doc):\\n \\\"\\\"\\\"\\n implement pls\\n get a path to save file, by relative path\\n need be complete by child\\n :param doc:\\n :return:\\n :rtype: str\\n \\\"\\\"\\\"\\n return None\\n\\n def url_load(self, doc):\\n \\\"\\\"\\\"\\n implement pls\\n :param doc:\\n :return:\\n :rtype: str\\n \\\"\\\"\\\"\\n return None\\n\\n def pre_save(self, doc, obj):\\n \\\"\\\"\\\"\\n implement pls, not force\\n do something before doc save\\n get download link here is better\\n :param doc:\\n :param obj:\\n :return:\\n \\\"\\\"\\\"\\n pass\\n\\n def shortcut_relative_name(self, doc):\\n \\\"\\\"\\\"\\n implement pls, not force\\n return short cut relative path\\n :param doc:\\n :return:\\n :rtype: str\\n \\\"\\\"\\\"\\n pass\\n\\n def shortcut_self_path(self, doc):\\n \\\"\\\"\\\"\\n implement pls, not force\\n return self short cut path\\n :param doc:\\n :return:\\n :rtype: list[str]\\n \\\"\\\"\\\"\\n return []\\n\\n @staticmethod\\n def request_file(doc):\\n \\\"\\\"\\\"\\n implement pls\\n :param doc:\\n :return:\\n :rtype: bytes\\n \\\"\\\"\\\"\\n url = None\\n try:\\n url = getattr(doc, model_download_url, None)\\n if url is None:\\n return None\\n\\n r = requests.get(url)\\n if r.status_code != 200:\\n return None\\n return r.content\\n except Exception as e:\\n log.print_info(u\\\"url download failed %s , err: %s\\\" % (url, e))\\n return None\\n\\n def download_relative_path(self, doc):\\n download_file_name = 'download_file_name'\\n if download_file_name not in doc:\\n doc[download_file_name] = self.download_filename_full(doc)\\n filename = doc[download_file_name]\\n if filename:\\n return os.path.join(self.__file_type__, filename)\\n else:\\n return None\\n\\n def download_check(self, doc, check_file=False):\\n \\\"\\\"\\\"\\n :param doc:\\n :param check_file: check dist file to download\\n :type doc: DynamicDocument\\n :return:\\n \\\"\\\"\\\"\\n if check_file:\\n file_relative_path = self.download_relative_path(doc)\\n for path in Config().get_paths():\\n file_path = os.path.join(path, file_relative_path)\\n if os.path.exists(file_path):\\n self.download_log(doc, download_path=file_path)\\n return file_path\\n self.download_log(doc, downloaded=False)\\n return False\\n\\n return hasattr(doc, model_is_download) and doc[model_is_download]\\n\\n def create_shortcut(self, doc, shortcuts_stack):\\n \\\"\\\"\\\"\\n :param doc:\\n :param shortcuts_stack:\\n :type shortcuts_stack: list[str]\\n :return:\\n \\\"\\\"\\\"\\n from NXSpider.utility.shortcut import symlink\\n if not (hasattr(doc, model_is_download) and doc[model_is_download]\\n and hasattr(doc, model_download_path)\\n and os.path.exists(doc[model_download_path])\\n and self.download_filename(doc)):\\n return\\n\\n file_path = doc[model_download_path]\\n shortcut_root = os.path.join(Config().get_path(), 'shortcuts')\\n file_name = self.download_filename(doc)\\n if file_path and file_path is not True:\\n if shortcuts_stack:\\n path = os.path.join(shortcut_root, *shortcuts_stack)\\n os.makedirs(path, exist_ok=True)\\n target = os.path.join(path, file_name)\\n symlink(file_path, os.path.join(shortcut_root, target))\\n\\n for k in self.shortcut_self_path(doc):\\n path = os.path.join(shortcut_root, k)\\n os.makedirs(path, exist_ok=True)\\n target = os.path.join(path, file_name)\\n symlink(file_path, os.path.join(shortcut_root, target))\\n pass\\n\\n @staticmethod\\n def download_log(doc, downloaded=True, download_path=\\\"\\\"):\\n doc[model_is_download] = downloaded\\n doc[model_download_path] = download_path\\n\\n def download_file_tag(self, filename, doc):\\n if self.__file_type__ in ['mp3', 'mp4', 'mv'] \\\\\\n and Config().get_media_tag():\\n attach_media_tag(doc, filename)\\n\\n def download_file(self, doc):\\n \\\"\\\"\\\"\\n download file from music 163\\n :param doc:\\n :return:\\n \\\"\\\"\\\"\\n file_relative_path = self.download_relative_path(doc)\\n path = Config().get_path()\\n content = self.request_file(doc)\\n\\n if content is None:\\n log.print_err(u\\\"file download failed : %s\\\" % file_relative_path)\\n return False\\n\\n try:\\n file_name = os.path.join(path, file_relative_path)\\n\\n # dir make\\n dir_name = os.path.dirname(file_name)\\n if not os.path.exists(dir_name):\\n os.makedirs(dir_name)\\n\\n # file write\\n with open(file_name, \\\"wb\\\") as code:\\n code.write(content)\\n\\n self.download_file_tag(file_name, doc)\\n\\n log.print_info(u\\\"file download complete: %s\\\" % file_relative_path)\\n self.download_log(doc)\\n return True\\n except Exception as e:\\n log.print_err(\\\"file save failed : %s, err: %s\\\" % (file_relative_path, e))\\n return False\\n\\n def debug_save_json(self, obj):\\n with tools.ignored(Exception):\\n relative_path = os.path.join(self.__file_type__ + '.debug',\\n self.__file_type__ + \\\"_\\\" +\\n str(obj['id']) + '.json')\\n file_name = os.path.join(Config().get_path(), relative_path)\\n with codecs.open(file_name, \\\"wb\\\", encoding='utf8') as code:\\n test = json.dumps(obj, ensure_ascii=False)\\n code.write(test)\\n\\n def debug_print(self, obj):\\n msg = u\\\"spider {} complete, id: {}\\\".format(self.__file_type__, obj['id'])\\n if 'name' in obj:\\n msg += u\\\", name: {}\\\".format(obj['name'])\\n\\n log.print_info(msg)\\n\\n def try_download(self, doc, download_type, file_check):\\n if self.__file_type__ not in download_type:\\n return True\\n\\n # download file and set download flag\\n if download_type is not None \\\\\\n and not self.download_check(doc, check_file=file_check):\\n # need download， try url which is set first or get new url and download\\n if getattr(doc, model_download_url, None) \\\\\\n and self.download_file(doc):\\n return True\\n\\n # get download link here\\n doc[model_download_url] = self.url_load(doc)\\n self.download_file(doc)\\n else:\\n name = self.download_relative_path(doc)\\n if name:\\n log.print_info(u\\\"file is exist or is not need to download : %s\\\"\\n % name)\\n\\n def parse_model(self, crawl_dict, download_type=None,\\n file_check=False, save=True, debug=False,\\n shortcuts_stack=None):\\n \\\"\\\"\\\"\\n Get a model from db or create, update and save!!!\\n this will replace some attributes into models by load_save_model also.\\n by @attr_replace(attr_name, new_name)\\n :param debug:\\n :param file_check:\\n :param download_type:\\n :param save: save doc\\n :param crawl_dict: must have id attr\\n :param shortcuts_stack: stack of shortcuts path\\n :type crawl_dict: dict\\n :type shortcuts_stack: list[str]\\n :return:\\n :rtype: DynamicDocument\\n :type save: bool\\n \\\"\\\"\\\"\\n if debug:\\n # self.debug_save_json(crawl_dict)\\n pass\\n\\n # get id\\n if 'id' not in crawl_dict:\\n log.print_err(u\\\"can not load id by json obj %s\\\" % json.dumps(crawl_dict))\\n return None\\n doc_id = crawl_dict['id']\\n\\n # load a mongo document\\n doc, is_new_doc = get_one_model_by_key(self.__model_name__, doc_id)\\n if doc is None:\\n log.print_err(u\\\"can not load a doc by obj %s_%d\\\"\\n % (self.__file_type__, doc_id))\\n return None\\n\\n # shortcuts in stack\\n if shortcuts_stack is not None and isinstance(shortcuts_stack, list) \\\\\\n and self.shortcut_relative_name(doc):\\n shortcuts_stack.append(self.shortcut_relative_name(doc))\\n\\n # if is_new_doc:\\n # replace attr or ignore\\n obj = dict()\\n for k, v in crawl_dict.items():\\n if k in self.__model_rfilter__:\\n continue\\n obj[k] = v\\n\\n # replace object\\n if k not in self.__attrs_replace_fucs__:\\n continue\\n\\n # change attr\\n if isinstance(v, list):\\n v = [self.__attrs_replace_fucs__[k](self, x) for x in v]\\n else:\\n v = self.__attrs_replace_fucs__[k](self, v)\\n\\n # replace key name\\n del obj[k]\\n obj[self.__attrs_replace_map__[k]] = v\\n\\n # recursion replace a attr into a model\\n for k, v in self.__parse_recursion__.items():\\n if k not in obj:\\n continue\\n\\n if isinstance(obj[k], list):\\n obj[k] = [v.parse_model(x, save=save, download_type=download_type,\\n file_check=file_check, debug=debug,\\n shortcuts_stack=shortcuts_stack)\\n for x in obj[k]]\\n elif isinstance(obj[k], dict):\\n obj[k] = v.parse_model(obj[k], save=save, download_type=download_type,\\n file_check=file_check, debug=debug,\\n shortcuts_stack=shortcuts_stack)\\n\\n # update json to doc, this must be after recursion\\n update_dynamic_doc(doc, obj)\\n\\n # modify doc and\\n self.pre_save(doc, crawl_dict)\\n\\n # try download\\n self.try_download(doc, download_type, file_check)\\n\\n if shortcuts_stack:\\n if self.shortcut_relative_name(doc):\\n shortcuts_stack.pop()\\n self.create_shortcut(doc, shortcuts_stack)\\n\\n # save document\\n if save and callable(getattr(doc, 'save', None)):\\n doc.save()\\n\\n if debug:\\n self.debug_print(crawl_dict)\\n\\n return doc\\n\\n\\n__all__ = [\\n 'Music163Obj',\\n 'attr_replace',\\n 'Music163ObjException',\\n\\n 'ConfigModel',\\n 'UserModel',\\n 'AlbumModel',\\n 'PlaylistModel',\\n 'Mp4Model',\\n 'VideoModel',\\n 'ArtistModel',\\n 'Mp3Model',\\n 'AuthorModel',\\n 'update_dynamic_doc',\\n 'model_download_url',\\n 'model_is_download',\\n 'get_one_model_by_key',\\n]\\n\"\n },\n {\n \"alpha_fraction\": 0.619271457195282,\n \"alphanum_fraction\": 0.6321974396705627,\n \"avg_line_length\": 23.314285278320312,\n \"blob_id\": \"8f211dd3739379317f3b16469266fd74c03aca61\",\n \"content_id\": \"a461ba57bf3615e7e28e78cab6a3f5255be7c356\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 851,\n \"license_type\": \"permissive\",\n \"max_line_length\": 72,\n \"num_lines\": 35,\n \"path\": \"/NXSpider/debug_test/move.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/5/14.\\n# email to LipsonChan@yahoo.com\\n#\\nimport os\\nimport shutil\\nimport sys\\nfrom glob import glob\\n\\n# reload(sys)\\n# sys.setdefaultencoding('utf8')\\n\\npaths = ['mv', 'mp3']\\nproject_path = os.path.dirname(os.path.realpath(__file__))\\nproject_path = os.path.dirname(project_path)\\ndownload_path = os.path.join(project_path, 'download_files')\\n\\n\\ndef move_file(path):\\n files = os.listdir(path)\\n for f in files:\\n if os.path.isdir(os.path.join(path, f)):\\n continue\\n\\n author = f[0: f.find(' - ')].strip()\\n author_path = os.path.join(path, author)\\n if not os.path.exists(author_path):\\n os.makedirs(author_path)\\n shutil.move(os.path.join(path, f), os.path.join(author_path, f))\\n\\n\\nfor p in paths:\\n move_file(os.path.join(download_path, p))\\n\"\n },\n {\n \"alpha_fraction\": 0.6714165806770325,\n \"alphanum_fraction\": 0.6974015235900879,\n \"avg_line_length\": 27.759037017822266,\n \"blob_id\": \"c194360d84bcbb7b02fae99f9ebac455922339a5\",\n \"content_id\": \"3531c5fa50ce0bc70ecfdf790fa506ecc3d56fe4\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 3794,\n \"license_type\": \"permissive\",\n \"max_line_length\": 214,\n \"num_lines\": 83,\n \"path\": \"/SIMPLE_USE.md\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"NXSpider\\n=================\\n\\nNXSpider, NetEase X Spider, A powerful for mp3,mv spider, can help you download mp3 and mv with media tags. Base on python, mongodb, and recursion algorithm\\n\\n[](LICENSE.md)\\n\\n\\n\\n[新手先看这里](SIMPLE_USE.md) | [历史版本在这里](VERSION.md) | [开发详情在这里](DEV.md)\\n\\n这里为非IT同学增加最简单使用方法（请保证电脑能上网），请根据以下流程。注意，所有跳转建议按住ctrl后再点击（或右键新窗口打开）\\n\\n### 点赞\\n- 请给项目点个star\\n- 请感谢网易云，感谢上帝，感谢国家，顺便的话感谢作者\\n\\n## Linux用户\\n[摸着自己的良心说，到底是不是IT人员](README.md)\\n\\n## Mac用户\\n待补充\\n\\n## Windows用户\\n\\n### 建议\\n- 使用python3.6，最好卸载2.7\\n- 所有路径（python安装路径，NXSpider路径，下载路径）不要有中文名，空格希望也不要有\\n\\n### 下载python3(自行跳过)\\n- [点击这里下载(32)](https://www.python.org/ftp/python/3.6.5/python-3.6.5.exe) 或 [点击这里下载(64)](https://www.python.org/ftp/python/3.6.5/python-3.6.5-amd64.exe) 或 [自行选择](https://www.python.org/downloads/release/python-365/)\\n- 双击你下载的文件，进行安装，跟正常安装软件一样\\n- 推荐大概在第一页勾选 ```for all users```\\n- 推荐大概在第二页选择自定义安装路径，大概是 ```C:\\\\Program Files```\\n- **强烈要求**大概在第二页勾选 ```Add Python to environment variables```\\n\\n- 后面一直按继续即可\\n- 安装完毕后按住键盘 ```win+r``` 在开始建附近会弹出个运行框，输入 ```cmd``` 回车\\n\\n- 回车后会弹出个黑色界面，通常我们叫 console(控制台)，输入 ```python``` 并回车\\n- 看到以下界面则为安装成功\\n\\n\\n### 下载本项目\\n- [点击这里下载](https://github.com/Grass-CLP/NXSpider/archive/master.zip) 或者点击右上角 ```clone or download``` 选择zip\\n- 解压缩到你想要的目录下，比如 ```G:/NXSpider``` 并进入目录\\n- 双击目录下文件 ```install.bat``` 会弹出个窗，等待结束（希望没看到error）\\n\\n### 运行本项目\\n1. 双击运行目录下 `start.bat` \\n2. 在控制台中就可以输入说明中的命令了(可以ctrl+c ctrl+v或右键复制)\\n3. 建议先输入 ```nxspider -h```\\n3. 先配置你的下载目录，比如目录是 ```G:/media```, 则输入并回车 ```nxspider config-spider -path G:/media```\\n4. **推荐**打开你的网易云音乐，点击:本地音乐 -> 选择目录 -> 添加文件夹 -> 选择你的目录 -> 确定\\n\\n### 通过搜索下载示例\\n $ nxspider search -ar \\\"张学友\\\"\\n $ nxspider sab -ab 19093 -dw mp3,mv \\n # -dw mp3,mv表示想要下载mp3和mv，不想下载mv则删除最后的,mv\\n \\n $ nxspider search -mp3 \\\"爱我别走\\\"\\n $ nxspider smp3s -mp3 187947 -dw mp3\\n \\n搜索歌神唱片找到ID：\\n\\n\\n\\n爬专辑19093：\\n\\n\\n\\n### 暴力下载最火歌单(不截图了)\\n $ nxspider scls-pls -cls 全部 -dw mp3,mv\\n\\n\\n### tips\\n- 其他命令见 [README.md](README.md)\\n- 目前暂时没做很好的退出，想退出直接把控制台(黑窗)关掉就好\\n- 目前暂时没做很好的重试，每次输入命令前复制到txt里面，遇到卡住的话，关掉重新打开个重新输入运行就可以\\n- 下载后重新进去网易云音乐，看看你的本地音乐，点击**匹配音乐**，是不是充满惊喜\\n\\n## 最后\\n有其他问题请到issue下留言，或者发邮件给我LipsonChan的雅虎邮箱\"\n },\n {\n \"alpha_fraction\": 0.6365979313850403,\n \"alphanum_fraction\": 0.6597937941551208,\n \"avg_line_length\": 23.25,\n \"blob_id\": \"6bc60273335c06c1aa5e1aecad9a98f96c6683f3\",\n \"content_id\": \"2817592a065e44c071baba87b01fc8b1ec8472bc\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 388,\n \"license_type\": \"permissive\",\n \"max_line_length\": 63,\n \"num_lines\": 16,\n \"path\": \"/NXSpider/common/constant.py\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# -*- coding: utf-8 -*-\\n#\\n# created by Lipson on 2018/5/30.\\n# email to LipsonChan@yahoo.com\\n#\\nimport os\\n\\n\\nall_download_type = ['mp3', 'mv']\\nmain_dir = os.path.join(os.path.expanduser('~'), '.nxspider')\\nconfig_path = os.path.join(main_dir, 'config.json')\\ndefault_download_dir = os.path.join(main_dir, 'download_files')\\n\\nif os.path.isdir(main_dir) is False:\\n os.mkdir(main_dir)\\n\"\n },\n {\n \"alpha_fraction\": 0.652942419052124,\n \"alphanum_fraction\": 0.6816385984420776,\n \"avg_line_length\": 26.485713958740234,\n \"blob_id\": \"c8dae3781978082a219f0600417114f6f8393fef\",\n \"content_id\": \"3fea56e7d126bd3b85739579264105143c1955b6\",\n \"detected_licenses\": [\n \"BSD-3-Clause\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 6745,\n \"license_type\": \"permissive\",\n \"max_line_length\": 109,\n \"num_lines\": 175,\n \"path\": \"/README.md\",\n \"repo_name\": \"Z-Shuming/NXSpider\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"NXSpider\\n=================\\n\\nNXSpider， NetEase X Spider，一个强大的（网易云音乐）mp3,mv爬虫，可以下载和收集mp3,mv信息,同时附带多媒体标签信息。采用python编写，mongo数据库(非必须)，递归算法核心实现\\n\\n[](LICENSE.md)\\n\\n\\n\\n[新手先看这里](SIMPLE_USE.md) | [历史版本在这里](VERSION.md) | [开发详情在这里](DEV.md)\\n\\n## 灵感来自\\n\\n[chengyumeng/spider163](https://github.com/chengyumeng/spider163)\\n\\n[Binaryify/NeteaseCloudMusicApi](https://github.com/Binaryify/NeteaseCloudMusicApi)\\n\\n[darknessomi/musicbox](https://github.com/darknessomi/musicbox)\\n\\n[sqaiyan/netmusic-node](https://github.com/sqaiyan/netmusic-node)\\n\\n## 功能特性\\n1. 采用命令行形式运行，配置在个人账号目录下```~/.nxspider/```\\n2. 搜索歌单，歌手，专辑，用户等信息\\n3. 通过各种方法，如歌单，歌手，专辑等，下载mp3, mv\\n4. 所有信息会保存在mongodb中（optional）\\n5. 所有mp3, mv可以通过配置，增加多媒体标签信息(歌手，专辑，唱片，163comment!!!)\\n\\n## 运行依赖\\n1. python3(推荐) or python2, 请配置好path\\n2. windows or linux\\n3. [mongodb](https://docs.mongodb.com/manual/installation/)(默认非必需)\\n\\n## Git clone安装\\n $ git clone https://github.com/Grass-CLP/NXSpider.git && cd NXSpider\\n $ python(3) setup.py install\\n $ pip(3) install -r requirememts.txt # optional, if you use mongodb\\n\\n## 使用指南\\n### 简单示例\\n\\t$ nxspider -h\\n\\t# 显示帮助信息\\n\\n\\t$ nxspider -v\\n\\t# 显示软件版本信息\\n\\n### 配置, <>内容为值，or为可选\\n\\t$ nxspider config-spider -path \\\"\\\"\\n\\t# 强烈建议第一次前进行配置,强烈建议加上双引号保证空格等问题\\n\\t# 配置下载文件的路径，第一个路径为存储路径，其他路径会检查文件是否下载 \\n\\n\\t$ nxspider config-mongo -mh -mp -mn \\n\\t# 配置mongodb服务器(非必需)，必须pip安装 requirements\\n\\n\\t$ nxspider config-check\\n\\t# 检查配置内容\\n\\n### 搜索命令\\n\\t# 以下为搜索命令，其中<>内文具体的文字内容，可以输入中文，韩文，日语等(注意加上双引号)\\n\\t$ nxspider search -ar \\\"\\\" [-offset ] [-limit ]\\n\\t$ nxspider search -pl \\\"\\\" [-offset ] [-limit ]\\n\\t$ nxspider search -ur \\\"\\\" [-offset ] [-limit ]\\n\\t$ nxspider search -mp3 \\\"\\\" [-offset ] [-limit ]\\n\\t$ nxspider search -ab \\\"\\\" [-offset ] [-limit ]\\n\\t$ nxspider search -mv \\\"\\\" [-offset
\n\t\t\n\t\t
\n\t\n\t\t\t![](\"static/images/ingredients.png\")
\n\t\t\t
\n\t\n\t\t\t\n\t\t\t\tShown in this picture:\n\t\t\t\t
\n\t\t\t\t
\n\t\t\n\t\t\t\t
Leafy Greens: kale, spinach, collard greens
\n\t\t\t\n\t\t\n\t
\n\t{% if selected == True %}\n\t\n\t\t
\n\t{% endif %}\n\t{% include 'footer.html' %}\n\n"},{"alpha_fraction":0.7721518874168396,"alphanum_fraction":0.7721518874168396,"avg_line_length":53.3125,"blob_id":"56fda2074a8a62d0f3bdcb7497cf9524a0400b39","content_id":"7bf98de085620afcbfbe45ae0f2c7dc2993aae65","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":870,"license_type":"no_license","max_line_length":183,"num_lines":16,"path":"/README.md","repo_name":"bjp12/recipe-book","src_encoding":"UTF-8","text":"# Recipe Book\n\nMy friends and I love to cook and people are always asking us for our recipes! \nI created this recipe book as a way to easily access our recipes without having to scroll through boring side stories and process pictures. \n\nSome features implemented so far:\n * Build a grocery list by clicking on items you need to go get. Once you click build, this list will be copied to your clipboard so that you can easily text or email it to yourself.\n * Use the menu builder to try out different options for appetizers, entrées, desserts and drinks to design an entire meal with our recipes.\n * Use the search bar to find specific recipes.\n * Find recipes using ingredients you already have!\n \nComing Soon:\n * Process pictures and step by step instructions when you search for a specific recipe\n * More recipes!\n \nLet me know what else you would like to see!\n"},{"alpha_fraction":0.5531841516494751,"alphanum_fraction":0.5612736940383911,"avg_line_length":31.63483238220215,"blob_id":"b03f58fa2ffa6e422b6929ab2a88eb627d6748c2","content_id":"a0f1cb5106791c0a106035e5df8d1bec751720a2","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":5812,"license_type":"no_license","max_line_length":104,"num_lines":178,"path":"/recipe book/homepage.py","repo_name":"bjp12/recipe-book","src_encoding":"UTF-8","text":"import json\nfrom flask import Flask\nfrom flask import render_template, request\nimport csv\nimport pandas as pd\nimport re\n\napp = Flask(__name__)\ndatadict = {}\ningredient_list = []\ndata_pd = pd.read_csv(\"static/Data.csv\")\nmeal_col = data_pd[\"Meal\"]\ncusine_col = data_pd[\"Cuisine\"]\nprep_col = data_pd[\"Prep Time\"]\n# ingredient_list = [row.split(\",\") for row in ingredient_col]\n# ingredient_list = set(x.strip() for l in ingredient_list for x in l)\n\nwith open(\"static/Data.csv\", 'r+', encoding=\"utf-8\") as f:\n data = csv.DictReader(f)\n for row in data:\n name = row['Name']\n ingredients = row['Ingredients']\n for i in ingredients.split(\",\"):\n i = i.strip()\n if i not in ingredient_list:\n ingredient_list.append(i)\n ingredient_list.sort()\n recipe = re.findall(r'\\d.*',row['Recipe'])\n prep = row['Prep Time']\n fullName = row['Full Name']\n datadict[name] = (ingredients, recipe, prep, fullName)\n\nmeals = {i:[] for i in meal_col.unique()}\ncuisines = {i: [] for i in cusine_col.unique()}\ntimes = {i: [] for i in prep_col.unique()}\nfor item in datadict:\n item_row = data_pd.loc[data_pd['Name'] == item]\n name = item_row['Name'].item()\n meals[item_row['Meal'].item()].append(name)\n cuisines[item_row['Cuisine'].item()].append(name)\n times[item_row['Prep Time'].item()].append(name)\n\n\nappetizers = data_pd.loc[data_pd[\"Meal\"] == \"appetizer\"][\"Name\"].tolist()\nentrees = data_pd.loc[data_pd[\"Meal\"] == \"entrée\"][\"Name\"].tolist()\ndesserts = data_pd.loc[data_pd[\"Meal\"] == \"dessert\"][\"Name\"].tolist()\ndrinks = data_pd.loc[data_pd[\"Meal\"] == \"drink\"][\"Name\"].tolist()\n\n@app.route('/')\ndef home():\n return render_template('index.html')\n\n@app.route('/ingredients', methods=[\"GET\",\"POST\"])\ndef ingredients():\n if request.method == 'POST':\n result = request.form.getlist('ingredients')\n select = set()\n for item in datadict:\n for x in result:\n chosen = [i.strip() for i in datadict[item][0].split(',')]\n if x in chosen:\n select.add(item)\n\n return render_template('ingredients.html', \n ilist = ingredient_list, \n data = datadict, \n selected = True,\n selected_list = result,\n filters = [(\"Results\", select)],\n )\n return render_template('ingredients.html', ilist = ingredient_list, data = datadict, selected=False)\n\n@app.route('/meal-builder', methods=[\"GET\", \"POST\"])\ndef meal_builder():\n presets = {\n \"american\":[\"quinoa\", \"mexican-pizza\", \"oreo-truffles\", \"mojito\"],\n \"italian\":[\"quinoa\", \"ravioli\", \"cake\", \"mojito\"],\n \"mexican\":[\"quinoa\", \"spinach-puffs\", \"chocchipcookies\", \"mojito\"]\n }\n\n builder = {\n \"appetizer\": appetizers,\n \"entree\": entrees,\n \"dessert\": desserts,\n \"drink\": drinks\n }\n\n if request.method == 'POST':\n result = request.form.get('change')\n if result in presets.keys():\n preset = presets[result]\n appetizers.remove(preset[0])\n appetizers.insert(0, preset[0])\n entrees.remove(preset[1])\n entrees.insert(0, preset[1])\n desserts.remove(preset[2])\n desserts.insert(0, preset[2])\n drinks.remove(preset[3])\n drinks.insert(0, preset[3])\n else:\n col = builder[result]\n first = col[0]\n col.remove(col[0])\n col.append(first)\n\n return render_template('meal-builder.html', \n data = datadict,\n appetizers = appetizers,\n entrees = entrees,\n desserts = desserts,\n drinks = drinks\n )\n\n else:\n return render_template('meal-builder.html', \n data=datadict,\n appetizers = appetizers,\n entrees = entrees,\n desserts = desserts,\n drinks = drinks\n )\n\n@app.route('/recipes')\n@app.route('/recipes-time')\ndef recipes_time():\n my_order = [\"breakfast\", \"appetizer\",\"entrée\", \"dessert\", \"drink\"]\n order = {key: i for i, key in enumerate(my_order)}\n my_filters = sorted([(k,meals[k]) for k in meals], key=lambda d:order[d[0]])\n return render_template('recipes.html', \n data = datadict,\n \tfilters = my_filters\n )\n\n@app.route('/recipes-cuisine')\ndef recipes_cuisine():\n my_filters = [(k,cuisines[k]) for k in cuisines]\n return render_template('recipes.html',\n data = datadict, \n \tfilters = my_filters\n )\n\n@app.route('/recipes-prep')\ndef recipes_prep():\n my_order = [\"10-20 min\", \"20-30 min\", \"30-40 min\", \"50-60 min\", \"1-2 hrs\"]\n order = {key: i for i, key in enumerate(my_order)}\n my_filters = sorted([(k,times[k]) for k in times], key=lambda d:order[d[0]])\n return render_template('recipes.html', \n data = datadict,\n filters = my_filters\n )\n\n@app.route('/search', methods=[\"GET\", \"POST\"])\ndef search():\n if request.method == 'POST':\n result = request.form.get('search')\n data = data_pd.loc[data_pd[\"Full Name\"].str.lower().str.contains(result.lower(), regex=False)]\n name = data['Name']\n print(\"count: \", name.shape[0])\n if name.shape[0] == 1:\n return render_template('search.html',\n data = data,\n datadict = datadict[name.item()],\n item = result\n )\n else:\n return render_template('search-plural.html',\n data = data,\n names = name.tolist(),\n datadict = '',\n item = result\n )\n\n@app.route('/aboutus')\ndef about():\n return render_template('aboutus.html')\n\nif __name__ == '__main__':\n app.run(host='0.0.0.0', port=5004, debug=True)\n\n"}],"string":"[\n {\n \"alpha_fraction\": 0.5807398557662964,\n \"alphanum_fraction\": 0.594245433807373,\n \"avg_line_length\": 27.88135528564453,\n \"blob_id\": \"499ea62c45a96d493e62d092bcf549445eedd7f7\",\n \"content_id\": \"496eb6d8e80f237dcb8a9400a7c0d8dc8cd51cb3\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"HTML\",\n \"length_bytes\": 1703,\n \"license_type\": \"no_license\",\n \"max_line_length\": 133,\n \"num_lines\": 59,\n \"path\": \"/recipe book/templates/ingredients.html\",\n \"repo_name\": \"bjp12/recipe-book\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\n\\n\\n\\t\n\t\t\t
\n\t\tSelected Ingredients:
\n\t\t\t- \n\t\t\t{% for x in selected_list %}\n\t\t\t\t
- {{ x }} \n\t\t\t{% endfor %}\n\t\t\t
\n\t\t\t{% include 'cards.html' %}\n\t\t
\n\t\\n\\t\\t\\n\\t\\t
\\n\\t\\n\\t\\t\\t![](\\\"static/images/ingredients.png\\\")
\\n\\t\\t\\t
\\n\\t\\n\\t\\t\\t\\n\\t\\t\\t\\tShown in this picture:\\n\\t\\t\\t\\t
\\n\\t\\t\\t\\t
\\n\\t\\t\\n\\t\\t\\t\\t
Leafy Greens: kale, spinach, collard greens
\\n\\t\\t\\t\\n\\t\\t\\n\\t
\\n\\t{% if selected == True %}\\n\\t\\n\\t\\t
\\n\\t{% endif %}\\n\\t{% include 'footer.html' %}\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.7721518874168396,\n \"alphanum_fraction\": 0.7721518874168396,\n \"avg_line_length\": 53.3125,\n \"blob_id\": \"56fda2074a8a62d0f3bdcb7497cf9524a0400b39\",\n \"content_id\": \"7bf98de085620afcbfbe45ae0f2c7dc2993aae65\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 870,\n \"license_type\": \"no_license\",\n \"max_line_length\": 183,\n \"num_lines\": 16,\n \"path\": \"/README.md\",\n \"repo_name\": \"bjp12/recipe-book\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Recipe Book\\n\\nMy friends and I love to cook and people are always asking us for our recipes! \\nI created this recipe book as a way to easily access our recipes without having to scroll through boring side stories and process pictures. \\n\\nSome features implemented so far:\\n * Build a grocery list by clicking on items you need to go get. Once you click build, this list will be copied to your clipboard so that you can easily text or email it to yourself.\\n * Use the menu builder to try out different options for appetizers, entrées, desserts and drinks to design an entire meal with our recipes.\\n * Use the search bar to find specific recipes.\\n * Find recipes using ingredients you already have!\\n \\nComing Soon:\\n * Process pictures and step by step instructions when you search for a specific recipe\\n * More recipes!\\n \\nLet me know what else you would like to see!\\n\"\n },\n {\n \"alpha_fraction\": 0.5531841516494751,\n \"alphanum_fraction\": 0.5612736940383911,\n \"avg_line_length\": 31.63483238220215,\n \"blob_id\": \"b03f58fa2ffa6e422b6929ab2a88eb627d6748c2\",\n \"content_id\": \"a0f1cb5106791c0a106035e5df8d1bec751720a2\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 5812,\n \"license_type\": \"no_license\",\n \"max_line_length\": 104,\n \"num_lines\": 178,\n \"path\": \"/recipe book/homepage.py\",\n \"repo_name\": \"bjp12/recipe-book\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import json\\nfrom flask import Flask\\nfrom flask import render_template, request\\nimport csv\\nimport pandas as pd\\nimport re\\n\\napp = Flask(__name__)\\ndatadict = {}\\ningredient_list = []\\ndata_pd = pd.read_csv(\\\"static/Data.csv\\\")\\nmeal_col = data_pd[\\\"Meal\\\"]\\ncusine_col = data_pd[\\\"Cuisine\\\"]\\nprep_col = data_pd[\\\"Prep Time\\\"]\\n# ingredient_list = [row.split(\\\",\\\") for row in ingredient_col]\\n# ingredient_list = set(x.strip() for l in ingredient_list for x in l)\\n\\nwith open(\\\"static/Data.csv\\\", 'r+', encoding=\\\"utf-8\\\") as f:\\n data = csv.DictReader(f)\\n for row in data:\\n name = row['Name']\\n ingredients = row['Ingredients']\\n for i in ingredients.split(\\\",\\\"):\\n i = i.strip()\\n if i not in ingredient_list:\\n ingredient_list.append(i)\\n ingredient_list.sort()\\n recipe = re.findall(r'\\\\d.*',row['Recipe'])\\n prep = row['Prep Time']\\n fullName = row['Full Name']\\n datadict[name] = (ingredients, recipe, prep, fullName)\\n\\nmeals = {i:[] for i in meal_col.unique()}\\ncuisines = {i: [] for i in cusine_col.unique()}\\ntimes = {i: [] for i in prep_col.unique()}\\nfor item in datadict:\\n item_row = data_pd.loc[data_pd['Name'] == item]\\n name = item_row['Name'].item()\\n meals[item_row['Meal'].item()].append(name)\\n cuisines[item_row['Cuisine'].item()].append(name)\\n times[item_row['Prep Time'].item()].append(name)\\n\\n\\nappetizers = data_pd.loc[data_pd[\\\"Meal\\\"] == \\\"appetizer\\\"][\\\"Name\\\"].tolist()\\nentrees = data_pd.loc[data_pd[\\\"Meal\\\"] == \\\"entrée\\\"][\\\"Name\\\"].tolist()\\ndesserts = data_pd.loc[data_pd[\\\"Meal\\\"] == \\\"dessert\\\"][\\\"Name\\\"].tolist()\\ndrinks = data_pd.loc[data_pd[\\\"Meal\\\"] == \\\"drink\\\"][\\\"Name\\\"].tolist()\\n\\n@app.route('/')\\ndef home():\\n return render_template('index.html')\\n\\n@app.route('/ingredients', methods=[\\\"GET\\\",\\\"POST\\\"])\\ndef ingredients():\\n if request.method == 'POST':\\n result = request.form.getlist('ingredients')\\n select = set()\\n for item in datadict:\\n for x in result:\\n chosen = [i.strip() for i in datadict[item][0].split(',')]\\n if x in chosen:\\n select.add(item)\\n\\n return render_template('ingredients.html', \\n ilist = ingredient_list, \\n data = datadict, \\n selected = True,\\n selected_list = result,\\n filters = [(\\\"Results\\\", select)],\\n )\\n return render_template('ingredients.html', ilist = ingredient_list, data = datadict, selected=False)\\n\\n@app.route('/meal-builder', methods=[\\\"GET\\\", \\\"POST\\\"])\\ndef meal_builder():\\n presets = {\\n \\\"american\\\":[\\\"quinoa\\\", \\\"mexican-pizza\\\", \\\"oreo-truffles\\\", \\\"mojito\\\"],\\n \\\"italian\\\":[\\\"quinoa\\\", \\\"ravioli\\\", \\\"cake\\\", \\\"mojito\\\"],\\n \\\"mexican\\\":[\\\"quinoa\\\", \\\"spinach-puffs\\\", \\\"chocchipcookies\\\", \\\"mojito\\\"]\\n }\\n\\n builder = {\\n \\\"appetizer\\\": appetizers,\\n \\\"entree\\\": entrees,\\n \\\"dessert\\\": desserts,\\n \\\"drink\\\": drinks\\n }\\n\\n if request.method == 'POST':\\n result = request.form.get('change')\\n if result in presets.keys():\\n preset = presets[result]\\n appetizers.remove(preset[0])\\n appetizers.insert(0, preset[0])\\n entrees.remove(preset[1])\\n entrees.insert(0, preset[1])\\n desserts.remove(preset[2])\\n desserts.insert(0, preset[2])\\n drinks.remove(preset[3])\\n drinks.insert(0, preset[3])\\n else:\\n col = builder[result]\\n first = col[0]\\n col.remove(col[0])\\n col.append(first)\\n\\n return render_template('meal-builder.html', \\n data = datadict,\\n appetizers = appetizers,\\n entrees = entrees,\\n desserts = desserts,\\n drinks = drinks\\n )\\n\\n else:\\n return render_template('meal-builder.html', \\n data=datadict,\\n appetizers = appetizers,\\n entrees = entrees,\\n desserts = desserts,\\n drinks = drinks\\n )\\n\\n@app.route('/recipes')\\n@app.route('/recipes-time')\\ndef recipes_time():\\n my_order = [\\\"breakfast\\\", \\\"appetizer\\\",\\\"entrée\\\", \\\"dessert\\\", \\\"drink\\\"]\\n order = {key: i for i, key in enumerate(my_order)}\\n my_filters = sorted([(k,meals[k]) for k in meals], key=lambda d:order[d[0]])\\n return render_template('recipes.html', \\n data = datadict,\\n \\tfilters = my_filters\\n )\\n\\n@app.route('/recipes-cuisine')\\ndef recipes_cuisine():\\n my_filters = [(k,cuisines[k]) for k in cuisines]\\n return render_template('recipes.html',\\n data = datadict, \\n \\tfilters = my_filters\\n )\\n\\n@app.route('/recipes-prep')\\ndef recipes_prep():\\n my_order = [\\\"10-20 min\\\", \\\"20-30 min\\\", \\\"30-40 min\\\", \\\"50-60 min\\\", \\\"1-2 hrs\\\"]\\n order = {key: i for i, key in enumerate(my_order)}\\n my_filters = sorted([(k,times[k]) for k in times], key=lambda d:order[d[0]])\\n return render_template('recipes.html', \\n data = datadict,\\n filters = my_filters\\n )\\n\\n@app.route('/search', methods=[\\\"GET\\\", \\\"POST\\\"])\\ndef search():\\n if request.method == 'POST':\\n result = request.form.get('search')\\n data = data_pd.loc[data_pd[\\\"Full Name\\\"].str.lower().str.contains(result.lower(), regex=False)]\\n name = data['Name']\\n print(\\\"count: \\\", name.shape[0])\\n if name.shape[0] == 1:\\n return render_template('search.html',\\n data = data,\\n datadict = datadict[name.item()],\\n item = result\\n )\\n else:\\n return render_template('search-plural.html',\\n data = data,\\n names = name.tolist(),\\n datadict = '',\\n item = result\\n )\\n\\n@app.route('/aboutus')\\ndef about():\\n return render_template('aboutus.html')\\n\\nif __name__ == '__main__':\\n app.run(host='0.0.0.0', port=5004, debug=True)\\n\\n\"\n }\n]"},"num_files":{"kind":"number","value":3,"string":"3"}}},{"rowIdx":2980385,"cells":{"repo_name":{"kind":"string","value":"polinagal/pyqt_mysql_gui_job_offer"},"repo_url":{"kind":"string","value":"https://github.com/polinagal/pyqt_mysql_gui_job_offer"},"snapshot_id":{"kind":"string","value":"f776920bb57e7a09494f0b210261dc4732e9b859"},"revision_id":{"kind":"string","value":"f19a1bea3ce7f10f18ffebca24068321b40b34ae"},"directory_id":{"kind":"string","value":"63617f36d1c2dbd303d646e6c373914eb4762433"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2021-01-21T06:55:34.122390","string":"2021-01-21T06:55:34.122390"},"revision_date":{"kind":"timestamp","value":"2015-04-05T12:39:13","string":"2015-04-05T12:39:13"},"committer_date":{"kind":"timestamp","value":"2015-04-05T12:39:13","string":"2015-04-05T12:39:13"},"github_id":{"kind":"null"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"null"},"gha_updated_at":{"kind":"null"},"gha_pushed_at":{"kind":"null"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.7571884989738464,"alphanum_fraction":0.7699680328369141,"avg_line_length":32,"blob_id":"51f13036efb0dc73298ffb49c1eab597cb817eaa","content_id":"2837bf3de1472e773a1d621782b051183fce0ff9","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":924,"license_type":"no_license","max_line_length":66,"num_lines":19,"path":"/hello_world.py","repo_name":"polinagal/pyqt_mysql_gui_job_offer","src_encoding":"UTF-8","text":"# -*- coding: utf-8 -*-\n#\nimport sys\nfrom PyQt4.QtGui import * # компоненты интерфейса\n\n# Каждое приложение должно создать объект QApplication\n# sys.argv - список аргументов командной строки\napplication = QApplication(sys.argv)\n\n# QWidget - базовый класс для всех объектов интерфейса\n# пользователя; если использовать для виджета конструктор\n# без родителя, такой виджет станет окном\nwidget = QWidget()\n\nwidget.resize(620, 640) # изменить размеры виджета\nwidget.setWindowTitle(\"Hop, Hey!\") # установить заголовок\nwidget.show() # отобразить окно на экране\n\nsys.exit(application.exec_()) # запуск основного цикла приложения"},{"alpha_fraction":0.6379974484443665,"alphanum_fraction":0.6444159150123596,"avg_line_length":31.5,"blob_id":"a14b4f782c1538e8d59ed9cabdcdfd1ce2dfa411","content_id":"f2af16834c78fe908c10832c68fb29c16b681716","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":928,"license_type":"no_license","max_line_length":71,"num_lines":24,"path":"/mainform.py","repo_name":"polinagal/pyqt_mysql_gui_job_offer","src_encoding":"UTF-8","text":"from PyQt4 import QtCore, QtGui, uic # подключает основные модули PyQt\n \n# прототип главной формы\nclass MainForm(QtGui.QDialog):\n \n # конструктор\n def __init__(self):\n super(MainForm, self).__init__()\n \n # динамически загружает визуальное представление формы\n uic.loadUi(\"mainform.ui\", self)\n \n # связывает событие нажатия на кнопку с методом\n # old style (before pyqt 4.5):\n # self.connect(self.pushButton, QtCore.SIGNAL(\"clicked()\"),\n # self.setTextEdit)\n # new style (qt 4.5):\n self.pushButton.clicked.connect(self.setLabelText)\n\n # def setTextEdit(self):\n # self.textEdit.setText(self.plainTextEdit.text()) # не пашет\n\n def setLabelText(self):\n self.label.setText(\"New label text\")"},{"alpha_fraction":0.689686119556427,"alphanum_fraction":0.6986547112464905,"avg_line_length":31.735294342041016,"blob_id":"4f09be579d3310cc4f934790e74e7260c105b382","content_id":"18003926178e80d3f30a4a8b79e878fd3a8d4bfd","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1115,"license_type":"no_license","max_line_length":101,"num_lines":34,"path":"/hello_mysql.py","repo_name":"polinagal/pyqt_mysql_gui_job_offer","src_encoding":"UTF-8","text":"__author__ = 'pif'\n\nimport sys\nfrom PyQt4 import QtSql\nfrom PyQt4 import QtCore\nfrom PyQt4 import QtGui\n\ninstance = QtGui.QApplication(sys.argv)\ndb = QtSql.QSqlDatabase.addDatabase(\"QMYSQL\")\ndb.setHostName(\"localhost\") # TODO localhost?\ndb.setDatabaseName(\"spring_sql\")\ndb.setUserName('root')\ndb.setPassword('')\nok = db.open()\n\nquery = QtSql.QSqlQuery()\nquery.setForwardOnly(True) # speed optimization\nquery.exec('select * from client')\nwhile query.next():\n for field_number in range(0, query.record().count()):\n value = query.value(field_number)\n if isinstance(value, QtCore.QPyNullVariant):\n value = 'null'\n print(value, end='\\t')\n print()\n # name = query.value(0) if not isinstance(query.value(0), QtCore.QPyNullVariant) else 'null'\n # sex = query.value(1) if not isinstance(query.value(1), QtCore.QPyNullVariant) else 'null'\n # sex_lover = query.value(2) if not isinstance(query.value(2), QtCore.QPyNullVariant) else 'null'\n # print(name, sex, sex_lover)\n\nprint(db.tables())\n\ndb.close()\nQtSql.QSqlDatabase.removeDatabase(\"QMYSQL\") #TODO seriously? \"QMYSQL\"?\n\n\n"},{"alpha_fraction":0.7990115284919739,"alphanum_fraction":0.8006590008735657,"avg_line_length":45.769229888916016,"blob_id":"e1f684214d5344a13c2f16d3195b2d0f94e330e6","content_id":"d87d790e02ea924c59ed78b25bd62b5a0814706f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":1085,"license_type":"no_license","max_line_length":185,"num_lines":13,"path":"/README.MD","repo_name":"polinagal/pyqt_mysql_gui_job_offer","src_encoding":"UTF-8","text":"Необходимо реализовать приложение, которое из прикрепленной БД будет выводить информацию о пациентах в таблице. \nСтолбцы: \n- ФИО\n- дата рождения + возраст\n- пол\n- информация о действующем полисе ОМС при наличии\n- информация о действующем документе (паспорт, свидетельство о рождении и т.д.)\n \nТакже нужен фильтр по основным полям, уже более детальный (отдельно фамилия, имя, отчество, серия документа, серия полиса и т.д.) с возможностью поиска одновременно по нескольким полям.\n \nРеализация должна быть на PyQt4. Используемая СУБД MySQL.\n\n(БД не добавлена в репозиторий из соображений конфиденциальности)."},{"alpha_fraction":0.71100914478302,"alphanum_fraction":0.713302731513977,"avg_line_length":35.41666793823242,"blob_id":"91d30ca8d5a4ad7cc4ad2b3849e056c00eac69f4","content_id":"1e31f753c0c748b231aaccb5c76e1e7e4d3b1e04","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":604,"license_type":"no_license","max_line_length":75,"num_lines":12,"path":"/hello_button.py","repo_name":"polinagal/pyqt_mysql_gui_job_offer","src_encoding":"UTF-8","text":"import sys\nfrom PyQt4 import QtCore, QtGui # подключает основные модули PyQt\nimport mainform # подключает модуль описания формы\n\ndef main():\n app = QtGui.QApplication(sys.argv) # создаёт основной объект программы\n form = mainform.MainForm() # создаёт объект формы\n form.show() # даёт команду на отображение объекта формы и содержимого\n app.exec() # запускает приложение\n\nif __name__ == \"__main__\":\n sys.exit(main())"},{"alpha_fraction":0.5195333957672119,"alphanum_fraction":0.5211611390113831,"avg_line_length":34.79611587524414,"blob_id":"a75bb20ed91ba735624cd4a8ab18fc0e6b8df7e9","content_id":"3531698ad249a788265cb9e2cfadfd9a92491b0b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":3926,"license_type":"no_license","max_line_length":130,"num_lines":103,"path":"/main_dirty.py","repo_name":"polinagal/pyqt_mysql_gui_job_offer","src_encoding":"UTF-8","text":"__author__ = 'pif'\n\nimport sys\nfrom PyQt4 import QtGui\nfrom PyQt4 import QtSql\n\nimport view\n\n\nclass Main():\n\n def __init__(self):\n app = QtGui.QApplication(sys.argv)\n\n self.db = self.open_db()\n\n self.view = view.View()\n self.view.showTableButton.clicked.connect(self.query) # TODO сделать проброс сигнала из view, а не вторгаться беспардонно\n self.view.show()\n\n app.exec()\n\n def open_db(self):\n db = QtSql.QSqlDatabase.addDatabase('QMYSQL')\n db.setHostName('localhost')\n db.setDatabaseName('spring_sql')\n db.setUserName('root')\n db.setPassword('')\n ok = db.open() # TODO что там, если не ок? Кинуть исключение?\n return db\n\n def query(self):\n # query = QtSql.QSqlQuery()\n # query.setForwardOnly(True) # speed optimization\n # query.exec('select * from cats')\n\n common_filter, \\\n lastname_filter, \\\n firstname_filter, \\\n patrname_filter,\\\n gender_filter,\\\n age_from, \\\n age_to, \\\n birthdate_from,\\\n birthdate_to, \\\n document_serial_filter, \\\n document_number_filter \\\n = self.view.get_fields_text()\n\n model = QtSql.QSqlQueryModel()\n model.setQuery(\n \"select \"\n \"CONCAT \"\n \"(\"\n \"client.lastName, ' ', \"\n \"client.firstName, ' ', \"\n \"client.patrName\"\n \") \"\n \"as 'ФИО', \" #TODO зависим от кодировки исходника, вынести в конфиг\n \"client.birthDate as 'дата рождения', \"\n \"TIMESTAMPDIFF(YEAR,client.birthDate,CURDATE()) as 'возраст', \"\n \"case client.sex when 1 then 'м' when 2 then 'ж' end as 'пол', \"\n \"CONCAT\"\n \"(\"\n \"clientdocument.serial, ' ', \"\n \"clientdocument.number\"\n \") \"\n \"as 'документ (серия, номер)' \"\n \"from client INNER JOIN clientdocument \"\n \"ON client.id = clientdocument.client_id \"\n \"where \"\n \"(\"\n \"client.lastName like '%\" + lastname_filter + \"%' and \"\n \"client.firstName like '%\" + firstname_filter + \"%' and \"\n \"client.patrName like '%\" + patrname_filter + \"%' and \"\n \"clientdocument.serial like '%\" + document_serial_filter + \"%' and \"\n \"clientdocument.number like '%\" + document_number_filter + \"%' and \"\n \"case client.sex when 1 then 'м' when 2 then 'ж' end like '%\" + gender_filter + \"%' and \"\n \"client.birthDate >= STR_TO_DATE('\" + birthdate_from + \"', '%Y-%m-%d') and \"\n \"client.birthDate <= STR_TO_DATE('\" + birthdate_to + \"', '%Y-%m-%d') and \"\n \"TIMESTAMPDIFF(YEAR,client.birthDate,CURDATE()) >= \" + str(age_from) + \" and \"\n \"TIMESTAMPDIFF(YEAR,client.birthDate,CURDATE()) <= \" + str(age_to) +\n \") \"\n \"and \"\n \"(\"\n \"client.lastName like '%\" + common_filter + \"%' or \"\n \"client.firstName like '%\" + common_filter + \"%' or \"\n \"client.patrName like '%\" + common_filter + \"%' or \"\n \"clientdocument.serial like '%\" + common_filter + \"%' or \"\n \"clientdocument.number like '%\" + common_filter + \"%'\"\n \") \"\n \"; \")\n #TODO maybe model.setQuery(query)?\n #TODO обработать случаи NULL переменных в sql-запросе\n\n self.view.set_model(model)\n\n def __del__(self): #TODO это может не вызваться, надо чистить ресурсы явно\n self.db.close()\n QtSql.QSqlDatabase.removeDatabase('QMYSQL')\n\nif __name__ == '__main__':\n Main()"},{"alpha_fraction":0.5257548689842224,"alphanum_fraction":0.527531087398529,"avg_line_length":27.897436141967773,"blob_id":"72092661c23244581b04a20b7d7f324dd2e7c066","content_id":"9809627b4b209fa31ce598ebaa4a9e341d13fc7a","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1182,"license_type":"no_license","max_line_length":66,"num_lines":39,"path":"/view_dirty.py","repo_name":"polinagal/pyqt_mysql_gui_job_offer","src_encoding":"UTF-8","text":"__author__ = 'pif'\n\nfrom PyQt4 import QtGui\nfrom PyQt4 import uic\n\n\nclass View(QtGui.QMainWindow):\n\n def __init__(self):\n super(View, self).__init__()\n\n # загружаем интерфейс\n uic.loadUi(\"view.ui\", self)\n\n # подправляем то, что не получилось сделать в Qt Designer\n self.gender.addItem(\"\")\n self.gender.addItem(\"м\")\n self.gender.addItem(\"ж\")\n\n def get_fields_text(self):\n return (\n self.common_filter.text(),\n self.lastName.text(),\n self.firstName.text(),\n self.patrName.text(),\n self.gender.currentText(),\n self.age_from.value(),\n self.age_to.value(),\n self.birthDate_from.date().toString('yyyy-MM-dd'),\n self.birthDate_to.date().toString('yyyy-MM-dd'),\n self.document_serial.text(),\n self.document_number.text(),\n )\n\n def set_model(self, model):\n self.tableView.setModel(model)\n # v = QtGui.QTableView(self.tableWidget)\n # v.setModel(model)\n # v.show()"},{"alpha_fraction":0.6364796161651611,"alphanum_fraction":0.6492347121238708,"avg_line_length":31.54166603088379,"blob_id":"236bc9bb4d9c8fffd9604b61abc9d16a77d67760","content_id":"f6bcc80138892b60f82633f10e330b81c1597515","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":784,"license_type":"no_license","max_line_length":103,"num_lines":24,"path":"/mysql_qtgui.py","repo_name":"polinagal/pyqt_mysql_gui_job_offer","src_encoding":"UTF-8","text":"__author__ = 'pif'\n\nimport sys\nfrom PyQt4 import QtSql\nfrom PyQt4 import QtCore\nfrom PyQt4 import QtGui\nimport view\n\napp = QtGui.QApplication(sys.argv)\napp.exec()\n\n# while query.next():\n# for field_number in range(0, query.record().count()):\n# value = query.value(field_number)\n# if isinstance(value, QtCore.QPyNullVariant):\n# value = 'null'\n# print(value, end='\\t')\n# print()\n# # name = query.value(0) if not isinstance(query.value(0), QtCore.QPyNullVariant) else 'null'\n# # sex = query.value(1) if not isinstance(query.value(1), QtCore.QPyNullVariant) else 'null'\n# # sex_lover = query.value(2) if not isinstance(query.value(2), QtCore.QPyNullVariant) else 'null'\n# # print(name, sex, sex_lover)\n#\n# print(db.tables())\n\n\n\n"}],"string":"[\n {\n \"alpha_fraction\": 0.7571884989738464,\n \"alphanum_fraction\": 0.7699680328369141,\n \"avg_line_length\": 32,\n \"blob_id\": \"51f13036efb0dc73298ffb49c1eab597cb817eaa\",\n \"content_id\": \"2837bf3de1472e773a1d621782b051183fce0ff9\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 924,\n \"license_type\": \"no_license\",\n \"max_line_length\": 66,\n \"num_lines\": 19,\n \"path\": \"/hello_world.py\",\n \"repo_name\": \"polinagal/pyqt_mysql_gui_job_offer\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# -*- coding: utf-8 -*-\\n#\\nimport sys\\nfrom PyQt4.QtGui import * # компоненты интерфейса\\n\\n# Каждое приложение должно создать объект QApplication\\n# sys.argv - список аргументов командной строки\\napplication = QApplication(sys.argv)\\n\\n# QWidget - базовый класс для всех объектов интерфейса\\n# пользователя; если использовать для виджета конструктор\\n# без родителя, такой виджет станет окном\\nwidget = QWidget()\\n\\nwidget.resize(620, 640) # изменить размеры виджета\\nwidget.setWindowTitle(\\\"Hop, Hey!\\\") # установить заголовок\\nwidget.show() # отобразить окно на экране\\n\\nsys.exit(application.exec_()) # запуск основного цикла приложения\"\n },\n {\n \"alpha_fraction\": 0.6379974484443665,\n \"alphanum_fraction\": 0.6444159150123596,\n \"avg_line_length\": 31.5,\n \"blob_id\": \"a14b4f782c1538e8d59ed9cabdcdfd1ce2dfa411\",\n \"content_id\": \"f2af16834c78fe908c10832c68fb29c16b681716\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 928,\n \"license_type\": \"no_license\",\n \"max_line_length\": 71,\n \"num_lines\": 24,\n \"path\": \"/mainform.py\",\n \"repo_name\": \"polinagal/pyqt_mysql_gui_job_offer\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"from PyQt4 import QtCore, QtGui, uic # подключает основные модули PyQt\\n \\n# прототип главной формы\\nclass MainForm(QtGui.QDialog):\\n \\n # конструктор\\n def __init__(self):\\n super(MainForm, self).__init__()\\n \\n # динамически загружает визуальное представление формы\\n uic.loadUi(\\\"mainform.ui\\\", self)\\n \\n # связывает событие нажатия на кнопку с методом\\n # old style (before pyqt 4.5):\\n # self.connect(self.pushButton, QtCore.SIGNAL(\\\"clicked()\\\"),\\n # self.setTextEdit)\\n # new style (qt 4.5):\\n self.pushButton.clicked.connect(self.setLabelText)\\n\\n # def setTextEdit(self):\\n # self.textEdit.setText(self.plainTextEdit.text()) # не пашет\\n\\n def setLabelText(self):\\n self.label.setText(\\\"New label text\\\")\"\n },\n {\n \"alpha_fraction\": 0.689686119556427,\n \"alphanum_fraction\": 0.6986547112464905,\n \"avg_line_length\": 31.735294342041016,\n \"blob_id\": \"4f09be579d3310cc4f934790e74e7260c105b382\",\n \"content_id\": \"18003926178e80d3f30a4a8b79e878fd3a8d4bfd\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1115,\n \"license_type\": \"no_license\",\n \"max_line_length\": 101,\n \"num_lines\": 34,\n \"path\": \"/hello_mysql.py\",\n \"repo_name\": \"polinagal/pyqt_mysql_gui_job_offer\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"__author__ = 'pif'\\n\\nimport sys\\nfrom PyQt4 import QtSql\\nfrom PyQt4 import QtCore\\nfrom PyQt4 import QtGui\\n\\ninstance = QtGui.QApplication(sys.argv)\\ndb = QtSql.QSqlDatabase.addDatabase(\\\"QMYSQL\\\")\\ndb.setHostName(\\\"localhost\\\") # TODO localhost?\\ndb.setDatabaseName(\\\"spring_sql\\\")\\ndb.setUserName('root')\\ndb.setPassword('')\\nok = db.open()\\n\\nquery = QtSql.QSqlQuery()\\nquery.setForwardOnly(True) # speed optimization\\nquery.exec('select * from client')\\nwhile query.next():\\n for field_number in range(0, query.record().count()):\\n value = query.value(field_number)\\n if isinstance(value, QtCore.QPyNullVariant):\\n value = 'null'\\n print(value, end='\\\\t')\\n print()\\n # name = query.value(0) if not isinstance(query.value(0), QtCore.QPyNullVariant) else 'null'\\n # sex = query.value(1) if not isinstance(query.value(1), QtCore.QPyNullVariant) else 'null'\\n # sex_lover = query.value(2) if not isinstance(query.value(2), QtCore.QPyNullVariant) else 'null'\\n # print(name, sex, sex_lover)\\n\\nprint(db.tables())\\n\\ndb.close()\\nQtSql.QSqlDatabase.removeDatabase(\\\"QMYSQL\\\") #TODO seriously? \\\"QMYSQL\\\"?\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.7990115284919739,\n \"alphanum_fraction\": 0.8006590008735657,\n \"avg_line_length\": 45.769229888916016,\n \"blob_id\": \"e1f684214d5344a13c2f16d3195b2d0f94e330e6\",\n \"content_id\": \"d87d790e02ea924c59ed78b25bd62b5a0814706f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 1085,\n \"license_type\": \"no_license\",\n \"max_line_length\": 185,\n \"num_lines\": 13,\n \"path\": \"/README.MD\",\n \"repo_name\": \"polinagal/pyqt_mysql_gui_job_offer\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"Необходимо реализовать приложение, которое из прикрепленной БД будет выводить информацию о пациентах в таблице. \\nСтолбцы: \\n- ФИО\\n- дата рождения + возраст\\n- пол\\n- информация о действующем полисе ОМС при наличии\\n- информация о действующем документе (паспорт, свидетельство о рождении и т.д.)\\n \\nТакже нужен фильтр по основным полям, уже более детальный (отдельно фамилия, имя, отчество, серия документа, серия полиса и т.д.) с возможностью поиска одновременно по нескольким полям.\\n \\nРеализация должна быть на PyQt4. Используемая СУБД MySQL.\\n\\n(БД не добавлена в репозиторий из соображений конфиденциальности).\"\n },\n {\n \"alpha_fraction\": 0.71100914478302,\n \"alphanum_fraction\": 0.713302731513977,\n \"avg_line_length\": 35.41666793823242,\n \"blob_id\": \"91d30ca8d5a4ad7cc4ad2b3849e056c00eac69f4\",\n \"content_id\": \"1e31f753c0c748b231aaccb5c76e1e7e4d3b1e04\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 604,\n \"license_type\": \"no_license\",\n \"max_line_length\": 75,\n \"num_lines\": 12,\n \"path\": \"/hello_button.py\",\n \"repo_name\": \"polinagal/pyqt_mysql_gui_job_offer\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import sys\\nfrom PyQt4 import QtCore, QtGui # подключает основные модули PyQt\\nimport mainform # подключает модуль описания формы\\n\\ndef main():\\n app = QtGui.QApplication(sys.argv) # создаёт основной объект программы\\n form = mainform.MainForm() # создаёт объект формы\\n form.show() # даёт команду на отображение объекта формы и содержимого\\n app.exec() # запускает приложение\\n\\nif __name__ == \\\"__main__\\\":\\n sys.exit(main())\"\n },\n {\n \"alpha_fraction\": 0.5195333957672119,\n \"alphanum_fraction\": 0.5211611390113831,\n \"avg_line_length\": 34.79611587524414,\n \"blob_id\": \"a75bb20ed91ba735624cd4a8ab18fc0e6b8df7e9\",\n \"content_id\": \"3531698ad249a788265cb9e2cfadfd9a92491b0b\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 3926,\n \"license_type\": \"no_license\",\n \"max_line_length\": 130,\n \"num_lines\": 103,\n \"path\": \"/main_dirty.py\",\n \"repo_name\": \"polinagal/pyqt_mysql_gui_job_offer\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"__author__ = 'pif'\\n\\nimport sys\\nfrom PyQt4 import QtGui\\nfrom PyQt4 import QtSql\\n\\nimport view\\n\\n\\nclass Main():\\n\\n def __init__(self):\\n app = QtGui.QApplication(sys.argv)\\n\\n self.db = self.open_db()\\n\\n self.view = view.View()\\n self.view.showTableButton.clicked.connect(self.query) # TODO сделать проброс сигнала из view, а не вторгаться беспардонно\\n self.view.show()\\n\\n app.exec()\\n\\n def open_db(self):\\n db = QtSql.QSqlDatabase.addDatabase('QMYSQL')\\n db.setHostName('localhost')\\n db.setDatabaseName('spring_sql')\\n db.setUserName('root')\\n db.setPassword('')\\n ok = db.open() # TODO что там, если не ок? Кинуть исключение?\\n return db\\n\\n def query(self):\\n # query = QtSql.QSqlQuery()\\n # query.setForwardOnly(True) # speed optimization\\n # query.exec('select * from cats')\\n\\n common_filter, \\\\\\n lastname_filter, \\\\\\n firstname_filter, \\\\\\n patrname_filter,\\\\\\n gender_filter,\\\\\\n age_from, \\\\\\n age_to, \\\\\\n birthdate_from,\\\\\\n birthdate_to, \\\\\\n document_serial_filter, \\\\\\n document_number_filter \\\\\\n = self.view.get_fields_text()\\n\\n model = QtSql.QSqlQueryModel()\\n model.setQuery(\\n \\\"select \\\"\\n \\\"CONCAT \\\"\\n \\\"(\\\"\\n \\\"client.lastName, ' ', \\\"\\n \\\"client.firstName, ' ', \\\"\\n \\\"client.patrName\\\"\\n \\\") \\\"\\n \\\"as 'ФИО', \\\" #TODO зависим от кодировки исходника, вынести в конфиг\\n \\\"client.birthDate as 'дата рождения', \\\"\\n \\\"TIMESTAMPDIFF(YEAR,client.birthDate,CURDATE()) as 'возраст', \\\"\\n \\\"case client.sex when 1 then 'м' when 2 then 'ж' end as 'пол', \\\"\\n \\\"CONCAT\\\"\\n \\\"(\\\"\\n \\\"clientdocument.serial, ' ', \\\"\\n \\\"clientdocument.number\\\"\\n \\\") \\\"\\n \\\"as 'документ (серия, номер)' \\\"\\n \\\"from client INNER JOIN clientdocument \\\"\\n \\\"ON client.id = clientdocument.client_id \\\"\\n \\\"where \\\"\\n \\\"(\\\"\\n \\\"client.lastName like '%\\\" + lastname_filter + \\\"%' and \\\"\\n \\\"client.firstName like '%\\\" + firstname_filter + \\\"%' and \\\"\\n \\\"client.patrName like '%\\\" + patrname_filter + \\\"%' and \\\"\\n \\\"clientdocument.serial like '%\\\" + document_serial_filter + \\\"%' and \\\"\\n \\\"clientdocument.number like '%\\\" + document_number_filter + \\\"%' and \\\"\\n \\\"case client.sex when 1 then 'м' when 2 then 'ж' end like '%\\\" + gender_filter + \\\"%' and \\\"\\n \\\"client.birthDate >= STR_TO_DATE('\\\" + birthdate_from + \\\"', '%Y-%m-%d') and \\\"\\n \\\"client.birthDate <= STR_TO_DATE('\\\" + birthdate_to + \\\"', '%Y-%m-%d') and \\\"\\n \\\"TIMESTAMPDIFF(YEAR,client.birthDate,CURDATE()) >= \\\" + str(age_from) + \\\" and \\\"\\n \\\"TIMESTAMPDIFF(YEAR,client.birthDate,CURDATE()) <= \\\" + str(age_to) +\\n \\\") \\\"\\n \\\"and \\\"\\n \\\"(\\\"\\n \\\"client.lastName like '%\\\" + common_filter + \\\"%' or \\\"\\n \\\"client.firstName like '%\\\" + common_filter + \\\"%' or \\\"\\n \\\"client.patrName like '%\\\" + common_filter + \\\"%' or \\\"\\n \\\"clientdocument.serial like '%\\\" + common_filter + \\\"%' or \\\"\\n \\\"clientdocument.number like '%\\\" + common_filter + \\\"%'\\\"\\n \\\") \\\"\\n \\\"; \\\")\\n #TODO maybe model.setQuery(query)?\\n #TODO обработать случаи NULL переменных в sql-запросе\\n\\n self.view.set_model(model)\\n\\n def __del__(self): #TODO это может не вызваться, надо чистить ресурсы явно\\n self.db.close()\\n QtSql.QSqlDatabase.removeDatabase('QMYSQL')\\n\\nif __name__ == '__main__':\\n Main()\"\n },\n {\n \"alpha_fraction\": 0.5257548689842224,\n \"alphanum_fraction\": 0.527531087398529,\n \"avg_line_length\": 27.897436141967773,\n \"blob_id\": \"72092661c23244581b04a20b7d7f324dd2e7c066\",\n \"content_id\": \"9809627b4b209fa31ce598ebaa4a9e341d13fc7a\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1182,\n \"license_type\": \"no_license\",\n \"max_line_length\": 66,\n \"num_lines\": 39,\n \"path\": \"/view_dirty.py\",\n \"repo_name\": \"polinagal/pyqt_mysql_gui_job_offer\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"__author__ = 'pif'\\n\\nfrom PyQt4 import QtGui\\nfrom PyQt4 import uic\\n\\n\\nclass View(QtGui.QMainWindow):\\n\\n def __init__(self):\\n super(View, self).__init__()\\n\\n # загружаем интерфейс\\n uic.loadUi(\\\"view.ui\\\", self)\\n\\n # подправляем то, что не получилось сделать в Qt Designer\\n self.gender.addItem(\\\"\\\")\\n self.gender.addItem(\\\"м\\\")\\n self.gender.addItem(\\\"ж\\\")\\n\\n def get_fields_text(self):\\n return (\\n self.common_filter.text(),\\n self.lastName.text(),\\n self.firstName.text(),\\n self.patrName.text(),\\n self.gender.currentText(),\\n self.age_from.value(),\\n self.age_to.value(),\\n self.birthDate_from.date().toString('yyyy-MM-dd'),\\n self.birthDate_to.date().toString('yyyy-MM-dd'),\\n self.document_serial.text(),\\n self.document_number.text(),\\n )\\n\\n def set_model(self, model):\\n self.tableView.setModel(model)\\n # v = QtGui.QTableView(self.tableWidget)\\n # v.setModel(model)\\n # v.show()\"\n },\n {\n \"alpha_fraction\": 0.6364796161651611,\n \"alphanum_fraction\": 0.6492347121238708,\n \"avg_line_length\": 31.54166603088379,\n \"blob_id\": \"236bc9bb4d9c8fffd9604b61abc9d16a77d67760\",\n \"content_id\": \"f6bcc80138892b60f82633f10e330b81c1597515\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 784,\n \"license_type\": \"no_license\",\n \"max_line_length\": 103,\n \"num_lines\": 24,\n \"path\": \"/mysql_qtgui.py\",\n \"repo_name\": \"polinagal/pyqt_mysql_gui_job_offer\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"__author__ = 'pif'\\n\\nimport sys\\nfrom PyQt4 import QtSql\\nfrom PyQt4 import QtCore\\nfrom PyQt4 import QtGui\\nimport view\\n\\napp = QtGui.QApplication(sys.argv)\\napp.exec()\\n\\n# while query.next():\\n# for field_number in range(0, query.record().count()):\\n# value = query.value(field_number)\\n# if isinstance(value, QtCore.QPyNullVariant):\\n# value = 'null'\\n# print(value, end='\\\\t')\\n# print()\\n# # name = query.value(0) if not isinstance(query.value(0), QtCore.QPyNullVariant) else 'null'\\n# # sex = query.value(1) if not isinstance(query.value(1), QtCore.QPyNullVariant) else 'null'\\n# # sex_lover = query.value(2) if not isinstance(query.value(2), QtCore.QPyNullVariant) else 'null'\\n# # print(name, sex, sex_lover)\\n#\\n# print(db.tables())\\n\\n\\n\\n\"\n }\n]"},"num_files":{"kind":"number","value":8,"string":"8"}}},{"rowIdx":2980386,"cells":{"repo_name":{"kind":"string","value":"SomEOnESmith/task_04"},"repo_url":{"kind":"string","value":"https://github.com/SomEOnESmith/task_04"},"snapshot_id":{"kind":"string","value":"3f35db740270b67d40a29297ac8c6f9af0ebae63"},"revision_id":{"kind":"string","value":"549ce8e51ff543b7b3206807efe807b351f63f3e"},"directory_id":{"kind":"string","value":"a7d613d043290b8c58b138657410896f1752018a"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2020-07-08T11:49:32.042272","string":"2020-07-08T11:49:32.042272"},"revision_date":{"kind":"timestamp","value":"2019-08-22T13:28:12","string":"2019-08-22T13:28:12"},"committer_date":{"kind":"timestamp","value":"2019-08-22T13:28:12","string":"2019-08-22T13:28:12"},"github_id":{"kind":"number","value":203663994,"string":"203,663,994"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"timestamp","value":"2019-08-21T21:02:29","string":"2019-08-21T21:02:29"},"gha_updated_at":{"kind":"timestamp","value":"2019-06-17T16:25:09","string":"2019-06-17T16:25:09"},"gha_pushed_at":{"kind":"timestamp","value":"2019-08-06T01:45:36","string":"2019-08-06T01:45:36"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.5193199515342712,"alphanum_fraction":0.5486862659454346,"avg_line_length":22.962963104248047,"blob_id":"b94edc40b397bf1cc75f7718f2d0d6757affd92d","content_id":"5180abc9be6319b8c0d91041851dbbecb9f54375","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":647,"license_type":"no_license","max_line_length":47,"num_lines":27,"path":"/restaurants/migrations/0002_auto_20190822_1322.py","repo_name":"SomEOnESmith/task_04","src_encoding":"UTF-8","text":"# Generated by Django 2.1.5 on 2019-08-22 13:22\n\nfrom django.db import migrations, models\n\n\nclass Migration(migrations.Migration):\n\n dependencies = [\n ('restaurants', '0001_initial'),\n ]\n\n operations = [\n migrations.RemoveField(\n model_name='restaurant',\n name='closeing_time',\n ),\n migrations.AddField(\n model_name='restaurant',\n name='closing_time',\n field=models.TimeField(null=True),\n ),\n migrations.AlterField(\n model_name='restaurant',\n name='opening_time',\n field=models.TimeField(),\n ),\n ]\n"}],"string":"[\n {\n \"alpha_fraction\": 0.5193199515342712,\n \"alphanum_fraction\": 0.5486862659454346,\n \"avg_line_length\": 22.962963104248047,\n \"blob_id\": \"b94edc40b397bf1cc75f7718f2d0d6757affd92d\",\n \"content_id\": \"5180abc9be6319b8c0d91041851dbbecb9f54375\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 647,\n \"license_type\": \"no_license\",\n \"max_line_length\": 47,\n \"num_lines\": 27,\n \"path\": \"/restaurants/migrations/0002_auto_20190822_1322.py\",\n \"repo_name\": \"SomEOnESmith/task_04\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Generated by Django 2.1.5 on 2019-08-22 13:22\\n\\nfrom django.db import migrations, models\\n\\n\\nclass Migration(migrations.Migration):\\n\\n dependencies = [\\n ('restaurants', '0001_initial'),\\n ]\\n\\n operations = [\\n migrations.RemoveField(\\n model_name='restaurant',\\n name='closeing_time',\\n ),\\n migrations.AddField(\\n model_name='restaurant',\\n name='closing_time',\\n field=models.TimeField(null=True),\\n ),\\n migrations.AlterField(\\n model_name='restaurant',\\n name='opening_time',\\n field=models.TimeField(),\\n ),\\n ]\\n\"\n }\n]"},"num_files":{"kind":"number","value":1,"string":"1"}}},{"rowIdx":2980387,"cells":{"repo_name":{"kind":"string","value":"evargasv/particle-filter"},"repo_url":{"kind":"string","value":"https://github.com/evargasv/particle-filter"},"snapshot_id":{"kind":"string","value":"07e632351333083628c9526c79c2b77ba81e34b0"},"revision_id":{"kind":"string","value":"38115f2f7a3b6e0769b08fc776502b0b6210e43e"},"directory_id":{"kind":"string","value":"31acb57949acb5d29c03dc5f1360d8e424551cd4"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2022-04-17T15:18:18.537322","string":"2022-04-17T15:18:18.537322"},"revision_date":{"kind":"timestamp","value":"2020-04-13T19:25:16","string":"2020-04-13T19:25:16"},"committer_date":{"kind":"timestamp","value":"2020-04-13T19:25:16","string":"2020-04-13T19:25:16"},"github_id":{"kind":"number","value":255417707,"string":"255,417,707"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"null"},"gha_updated_at":{"kind":"null"},"gha_pushed_at":{"kind":"null"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.8352112770080566,"alphanum_fraction":0.8352112770080566,"avg_line_length":140.60000610351562,"blob_id":"3c75ac9f8ef489b6c1a117322956c51e47d3663d","content_id":"05d3338a3153b525aede7c63ff4857cbc0dfae1f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":710,"license_type":"no_license","max_line_length":549,"num_lines":5,"path":"/README.md","repo_name":"evargasv/particle-filter","src_encoding":"UTF-8","text":"# Particle Filter\n\nParticle Filter or Montecarlo Localization (MCL) algorithm implementation to localise a two dimensional robot (turtlebot) in a given map.\n\nThe MCL localization is an implementation of the Markovian localisation problem where the involved probability density functions are represented through samples (particles) and the Bayes filter is implemented through the Particle Filter. Markov localisation addresses the problem of state estimation from sensor data. Markov localization is a probabilistic algorithm: Instead of maintaining a single hypothesis as to where in the world a robot might be, Markov localization maintains a probability distribution over the space of all such hypotheses.\n\n\n"},{"alpha_fraction":0.4799180328845978,"alphanum_fraction":0.49467211961746216,"avg_line_length":36.53845977783203,"blob_id":"8dc908ed5e9b806d8871eda6f1e193419c50c767","content_id":"3cd3afc29c99c5662dc86ee1c6f269fd74fed0f7","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":7320,"license_type":"no_license","max_line_length":142,"num_lines":195,"path":"/src/particle_filter.py","repo_name":"evargasv/particle-filter","src_encoding":"UTF-8","text":"#!/usr/bin/python\n# -*- coding: utf-8 -*-\n\nimport numpy as np\nfrom functions import angle_wrap\n\n#===============================================================================\nclass ParticleFilter(object):\n '''\n Class to hold the whole particle filter.\n\n p_wei: weights of particles in array of shape (N,)\n p_ang: angle in radians of each particle with respect of world axis, shape (N,)\n p_xy : position in the world frame of the particles, shape (2,N)\n '''\n\n #===========================================================================\n def __init__(self, room_map, num, odom_lin_sigma, odom_ang_sigma,\n meas_rng_noise, meas_ang_noise):\n '''\n Initializes the particle filter\n room_map : an array of lines in the form [x1 y1 x2 y2]\n num : number of particles to use\n odom_lin_sigma: odometry linear noise\n odom_ang_sigma: odometry angular noise\n meas_rng_noise: measurement linear noise\n meas_ang_noise: measurement angular noise\n '''\n\n # Copy parameters\n self.map = room_map\n self.num = num\n self.odom_lin_sigma = odom_lin_sigma\n self.odom_ang_sigma = odom_ang_sigma\n self.meas_rng_noise = meas_rng_noise\n self.meas_ang_noise = meas_ang_noise\n\n # Map\n map_xmin = np.min(self.map[:, 0])\n map_xmax = np.max(self.map[:, 0])\n map_ymin = np.min(self.map[:, 1])\n map_ymax = np.max(self.map[:, 1])\n\n # Particle initialization\n self.p_wei = 1.0 / num * np.ones(num)\n self.p_ang = 2 * np.pi * np.random.rand(num)\n self.p_xy = np.vstack(( map_xmin + (map_xmax - map_xmin) * np.random.rand(num),\n map_ymin + (map_ymax - map_ymin) * np.random.rand(num) ))\n\n #===========================================================================\n def predict(self, odom):\n '''\n Moves particles with the given odometry.\n odom: incremental odometry [delta_x delta_y delta_yaw] in the vehicle frame\n '''\n\n # Add Gaussian noise to odometry measures\n lin_noise = np.random.randn(self.num,2) * self.odom_lin_sigma\n ang_noise = angle_wrap( np.random.randn(self.num) * self.odom_ang_sigma + odom[2] )\n\n # Increment particle positions in correct frame\n odom_noise = np.tile(odom[0:2], (self.num, 1))\n odom_noise += lin_noise\n\n self.p_xy[0,:] += odom_noise[:,0] * np.cos(self.p_ang) - odom_noise[:,1] * np.sin(self.p_ang)\n self.p_xy[1,:] += odom_noise[:,0] * np.sin(self.p_ang) + odom_noise[:,1] * np.cos(self.p_ang)\n\n # Increment angley\n self.p_ang += ang_noise\n self.p_ang = angle_wrap(self.p_ang)\n\n #===========================================================================\n def weight(self, lines):\n '''\n Look for the lines seen from the robot and compare them to the given map.\n Lines expressed as [x1 y1 x2 y2].\n '''\n\n # Constant values for all weightings\n val_rng = 1.0 / (self.meas_rng_noise * np.sqrt(2 * np.pi))\n val_ang = 1.0 / (self.meas_ang_noise * np.sqrt(2 * np.pi))\n\n # origin of the map\n orig = np.array( [0, 0, 0] )\n\n # map and measured lines array for polar coordinates\n expected_lines = np.zeros( (self.map.shape[0], 2) )\n measured_lines = np.zeros( (lines.shape[0], 2) )\n\n # Loop over particles\n for i in range(self.num):\n\n # Position of the particle\n odom = np.array( [self.p_xy[0,i], self.p_xy[1,i], self.p_ang[i] ] )\n # Weight of each measured line\n lines_wei = np.zeros( lines.shape[0] )\n\n # Transform map lines to local frame and to [range theta]\n for j in range(self.map.shape[0]):\n # Transform to [range theta]\n expected_lines[j,:] = self.get_polar_line( self.map[j,:], odom )\n\n # Transform measured lines to [range theta] and weight them\n for j in range(lines.shape[0]):\n\n # Transform measured lines\n measured_lines[j,:] = self.get_polar_line( lines[j,:], orig )\n map_wei = np.zeros( expected_lines.shape[0] )\n\n # Select the map line that best corresponds to the measured line. The weight of them is calculated and the maximum is selected\n for k in range(expected_lines.shape[0]):\n\n # Weight them\n range_wei = val_rng * np.exp((-( measured_lines[j,0] - expected_lines[k,0] )**2.0)/(2.0*(val_rng**2.0)))\n angle_wei = val_ang * np.exp((-( measured_lines[j,1] - expected_lines[k,1] )**2.0)/(2.0*(val_ang**2.0)))\n map_wei[k] = range_wei * angle_wei\n\n # Best associated line on the map\n lines_wei[j] = map_wei.max()\n\n # OPTIONAL question\n # make sure segments correspond, if not put weight to zero\n #\n #\n\n # Take best weighting (best associated lines)\n self.p_wei[i] *= np.sum( lines_wei )\n\n # Normalize weights\n self.p_wei /= np.sum(self.p_wei)\n\n\n #===========================================================================\n\n def resample(self):\n '''\n Systematic resampling of the particles.\n '''\n # Look for particles to replicate\n p_idx = np.zeros(self.num, dtype=np.uint8)\n r = np.random.rand() * (1.0/self.num)\n c = self.p_wei[0]\n j = 0\n\n for i in range(self.num):\n u = r + ( i * (1.0/self.num) )\n while u > c:\n j = (j + 1) % self.num\n c = c + self.p_wei[j]\n p_idx[i] = int(j);\n\n # Pick chosen particles\n self.p_ang = self.p_ang[(p_idx)]\n self.p_wei = self.p_wei[(p_idx)]\n self.p_xy = self.p_xy[:,p_idx]\n #===========================================================================\n def get_mean_particle(self):\n '''\n Gets mean particle.\n '''\n # Weighted mean\n weig = np.vstack((self.p_wei, self.p_wei))\n mean = np.sum(self.p_xy * weig, axis=1) / np.sum(self.p_wei)\n\n ang = np.arctan2( np.sum(self.p_wei * np.sin(self.p_ang)) / np.sum(self.p_wei),\n np.sum(self.p_wei * np.cos(self.p_ang)) / np.sum(self.p_wei) )\n\n return np.array([mean[0], mean[1], ang])\n\n #===========================================================================\n def get_polar_line(self, line, odom):\n '''\n Transforms a line from [x1 y1 x2 y2] from the world frame to the\n vehicle frame using odomotrey [x y ang].\n Returns [range theta]\n '''\n # Line points\n x1 = line[0]\n y1 = line[1]\n x2 = line[2]\n y2 = line[3]\n\n # Compute line (a, b, c) and range\n line = np.array([y1-y2, x2-x1, x1*y2-x2*y1])\n pt = np.array([odom[0], odom[1], 1])\n dist = np.dot(pt, line) / np.linalg.norm(line[:2])\n\n # Compute angle\n if dist > 0:\n ang = np.arctan2(line[1], line[0])\n else:\n ang = np.arctan2(-line[1], -line[0])\n\n # Return in the vehicle frame\n return np.array([np.abs(dist), angle_wrap(ang - odom[2])])\n"}],"string":"[\n {\n \"alpha_fraction\": 0.8352112770080566,\n \"alphanum_fraction\": 0.8352112770080566,\n \"avg_line_length\": 140.60000610351562,\n \"blob_id\": \"3c75ac9f8ef489b6c1a117322956c51e47d3663d\",\n \"content_id\": \"05d3338a3153b525aede7c63ff4857cbc0dfae1f\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 710,\n \"license_type\": \"no_license\",\n \"max_line_length\": 549,\n \"num_lines\": 5,\n \"path\": \"/README.md\",\n \"repo_name\": \"evargasv/particle-filter\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Particle Filter\\n\\nParticle Filter or Montecarlo Localization (MCL) algorithm implementation to localise a two dimensional robot (turtlebot) in a given map.\\n\\nThe MCL localization is an implementation of the Markovian localisation problem where the involved probability density functions are represented through samples (particles) and the Bayes filter is implemented through the Particle Filter. Markov localisation addresses the problem of state estimation from sensor data. Markov localization is a probabilistic algorithm: Instead of maintaining a single hypothesis as to where in the world a robot might be, Markov localization maintains a probability distribution over the space of all such hypotheses.\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.4799180328845978,\n \"alphanum_fraction\": 0.49467211961746216,\n \"avg_line_length\": 36.53845977783203,\n \"blob_id\": \"8dc908ed5e9b806d8871eda6f1e193419c50c767\",\n \"content_id\": \"3cd3afc29c99c5662dc86ee1c6f269fd74fed0f7\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 7320,\n \"license_type\": \"no_license\",\n \"max_line_length\": 142,\n \"num_lines\": 195,\n \"path\": \"/src/particle_filter.py\",\n \"repo_name\": \"evargasv/particle-filter\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/python\\n# -*- coding: utf-8 -*-\\n\\nimport numpy as np\\nfrom functions import angle_wrap\\n\\n#===============================================================================\\nclass ParticleFilter(object):\\n '''\\n Class to hold the whole particle filter.\\n\\n p_wei: weights of particles in array of shape (N,)\\n p_ang: angle in radians of each particle with respect of world axis, shape (N,)\\n p_xy : position in the world frame of the particles, shape (2,N)\\n '''\\n\\n #===========================================================================\\n def __init__(self, room_map, num, odom_lin_sigma, odom_ang_sigma,\\n meas_rng_noise, meas_ang_noise):\\n '''\\n Initializes the particle filter\\n room_map : an array of lines in the form [x1 y1 x2 y2]\\n num : number of particles to use\\n odom_lin_sigma: odometry linear noise\\n odom_ang_sigma: odometry angular noise\\n meas_rng_noise: measurement linear noise\\n meas_ang_noise: measurement angular noise\\n '''\\n\\n # Copy parameters\\n self.map = room_map\\n self.num = num\\n self.odom_lin_sigma = odom_lin_sigma\\n self.odom_ang_sigma = odom_ang_sigma\\n self.meas_rng_noise = meas_rng_noise\\n self.meas_ang_noise = meas_ang_noise\\n\\n # Map\\n map_xmin = np.min(self.map[:, 0])\\n map_xmax = np.max(self.map[:, 0])\\n map_ymin = np.min(self.map[:, 1])\\n map_ymax = np.max(self.map[:, 1])\\n\\n # Particle initialization\\n self.p_wei = 1.0 / num * np.ones(num)\\n self.p_ang = 2 * np.pi * np.random.rand(num)\\n self.p_xy = np.vstack(( map_xmin + (map_xmax - map_xmin) * np.random.rand(num),\\n map_ymin + (map_ymax - map_ymin) * np.random.rand(num) ))\\n\\n #===========================================================================\\n def predict(self, odom):\\n '''\\n Moves particles with the given odometry.\\n odom: incremental odometry [delta_x delta_y delta_yaw] in the vehicle frame\\n '''\\n\\n # Add Gaussian noise to odometry measures\\n lin_noise = np.random.randn(self.num,2) * self.odom_lin_sigma\\n ang_noise = angle_wrap( np.random.randn(self.num) * self.odom_ang_sigma + odom[2] )\\n\\n # Increment particle positions in correct frame\\n odom_noise = np.tile(odom[0:2], (self.num, 1))\\n odom_noise += lin_noise\\n\\n self.p_xy[0,:] += odom_noise[:,0] * np.cos(self.p_ang) - odom_noise[:,1] * np.sin(self.p_ang)\\n self.p_xy[1,:] += odom_noise[:,0] * np.sin(self.p_ang) + odom_noise[:,1] * np.cos(self.p_ang)\\n\\n # Increment angley\\n self.p_ang += ang_noise\\n self.p_ang = angle_wrap(self.p_ang)\\n\\n #===========================================================================\\n def weight(self, lines):\\n '''\\n Look for the lines seen from the robot and compare them to the given map.\\n Lines expressed as [x1 y1 x2 y2].\\n '''\\n\\n # Constant values for all weightings\\n val_rng = 1.0 / (self.meas_rng_noise * np.sqrt(2 * np.pi))\\n val_ang = 1.0 / (self.meas_ang_noise * np.sqrt(2 * np.pi))\\n\\n # origin of the map\\n orig = np.array( [0, 0, 0] )\\n\\n # map and measured lines array for polar coordinates\\n expected_lines = np.zeros( (self.map.shape[0], 2) )\\n measured_lines = np.zeros( (lines.shape[0], 2) )\\n\\n # Loop over particles\\n for i in range(self.num):\\n\\n # Position of the particle\\n odom = np.array( [self.p_xy[0,i], self.p_xy[1,i], self.p_ang[i] ] )\\n # Weight of each measured line\\n lines_wei = np.zeros( lines.shape[0] )\\n\\n # Transform map lines to local frame and to [range theta]\\n for j in range(self.map.shape[0]):\\n # Transform to [range theta]\\n expected_lines[j,:] = self.get_polar_line( self.map[j,:], odom )\\n\\n # Transform measured lines to [range theta] and weight them\\n for j in range(lines.shape[0]):\\n\\n # Transform measured lines\\n measured_lines[j,:] = self.get_polar_line( lines[j,:], orig )\\n map_wei = np.zeros( expected_lines.shape[0] )\\n\\n # Select the map line that best corresponds to the measured line. The weight of them is calculated and the maximum is selected\\n for k in range(expected_lines.shape[0]):\\n\\n # Weight them\\n range_wei = val_rng * np.exp((-( measured_lines[j,0] - expected_lines[k,0] )**2.0)/(2.0*(val_rng**2.0)))\\n angle_wei = val_ang * np.exp((-( measured_lines[j,1] - expected_lines[k,1] )**2.0)/(2.0*(val_ang**2.0)))\\n map_wei[k] = range_wei * angle_wei\\n\\n # Best associated line on the map\\n lines_wei[j] = map_wei.max()\\n\\n # OPTIONAL question\\n # make sure segments correspond, if not put weight to zero\\n #\\n #\\n\\n # Take best weighting (best associated lines)\\n self.p_wei[i] *= np.sum( lines_wei )\\n\\n # Normalize weights\\n self.p_wei /= np.sum(self.p_wei)\\n\\n\\n #===========================================================================\\n\\n def resample(self):\\n '''\\n Systematic resampling of the particles.\\n '''\\n # Look for particles to replicate\\n p_idx = np.zeros(self.num, dtype=np.uint8)\\n r = np.random.rand() * (1.0/self.num)\\n c = self.p_wei[0]\\n j = 0\\n\\n for i in range(self.num):\\n u = r + ( i * (1.0/self.num) )\\n while u > c:\\n j = (j + 1) % self.num\\n c = c + self.p_wei[j]\\n p_idx[i] = int(j);\\n\\n # Pick chosen particles\\n self.p_ang = self.p_ang[(p_idx)]\\n self.p_wei = self.p_wei[(p_idx)]\\n self.p_xy = self.p_xy[:,p_idx]\\n #===========================================================================\\n def get_mean_particle(self):\\n '''\\n Gets mean particle.\\n '''\\n # Weighted mean\\n weig = np.vstack((self.p_wei, self.p_wei))\\n mean = np.sum(self.p_xy * weig, axis=1) / np.sum(self.p_wei)\\n\\n ang = np.arctan2( np.sum(self.p_wei * np.sin(self.p_ang)) / np.sum(self.p_wei),\\n np.sum(self.p_wei * np.cos(self.p_ang)) / np.sum(self.p_wei) )\\n\\n return np.array([mean[0], mean[1], ang])\\n\\n #===========================================================================\\n def get_polar_line(self, line, odom):\\n '''\\n Transforms a line from [x1 y1 x2 y2] from the world frame to the\\n vehicle frame using odomotrey [x y ang].\\n Returns [range theta]\\n '''\\n # Line points\\n x1 = line[0]\\n y1 = line[1]\\n x2 = line[2]\\n y2 = line[3]\\n\\n # Compute line (a, b, c) and range\\n line = np.array([y1-y2, x2-x1, x1*y2-x2*y1])\\n pt = np.array([odom[0], odom[1], 1])\\n dist = np.dot(pt, line) / np.linalg.norm(line[:2])\\n\\n # Compute angle\\n if dist > 0:\\n ang = np.arctan2(line[1], line[0])\\n else:\\n ang = np.arctan2(-line[1], -line[0])\\n\\n # Return in the vehicle frame\\n return np.array([np.abs(dist), angle_wrap(ang - odom[2])])\\n\"\n }\n]"},"num_files":{"kind":"number","value":2,"string":"2"}}},{"rowIdx":2980388,"cells":{"repo_name":{"kind":"string","value":"LoveGalaxy/PyTorch_Tutorial"},"repo_url":{"kind":"string","value":"https://github.com/LoveGalaxy/PyTorch_Tutorial"},"snapshot_id":{"kind":"string","value":"3ed7879bd177bb9c47f8a39a67f0126b81ce0290"},"revision_id":{"kind":"string","value":"482434770cd51ade8c1f7e91f0be67daa8cbcfc2"},"directory_id":{"kind":"string","value":"0d40503c0e101bb881c918ff1e1f429497860acf"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2022-04-29T19:08:01.432004","string":"2022-04-29T19:08:01.432004"},"revision_date":{"kind":"timestamp","value":"2018-06-07T18:24:20","string":"2018-06-07T18:24:20"},"committer_date":{"kind":"timestamp","value":"2018-06-07T18:24:20","string":"2018-06-07T18:24:20"},"github_id":{"kind":"null"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"null"},"gha_updated_at":{"kind":"null"},"gha_pushed_at":{"kind":"null"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.5605888962745667,"alphanum_fraction":0.5979614853858948,"avg_line_length":20.0238094329834,"blob_id":"1be8d3d0fd002c0638e86f9e1a6b63d7db2a9761","content_id":"50c3f7f2a302bb1602e7c50494443f3ca20779b8","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1035,"license_type":"no_license","max_line_length":54,"num_lines":42,"path":"/Part 2： Pytorch简单例子/1 Use Tnensor to implement a NN.py","repo_name":"LoveGalaxy/PyTorch_Tutorial","src_encoding":"UTF-8","text":"import torch\n\n# numpy已经为我们提供了强大的计算能力，但是为了能让我们的数据\n# 能够在GPU上进行加速，我们需要将计算的类型转化为Tensor\n# 设置torch tensor类型，设置计算设备\ndtype = torch.float\ndevice = torch.device(\"cpu\")\n\nN = 64\nD_in = 1000\nH = 100\nD_out = 10\n\nx = torch.randn(N, D_in, device=device, dtype=dtype)\ny = torch.randn(N, D_out, device=device, dtype=dtype)\n\nw1 = torch.randn(D_in, H, device=device, dtype=dtype)\nw2 = torch.randn(H, D_out, device=device, dtype=dtype)\n\nlr = 1e-6\n\nfor t in range(500):\n # 前馈计算\n h = x.mm(w1)\n h_relu = h.clamp(min=0)\n y_pred = h_relu.mm(w2)\n\n # 计算损失\n loss = (y_pred - y).pow(2).sum().item()\n print(t, loss)\n\n # bp算法\n grad_y_pred = 2.0 * (y_pred - y)\n grad_w2 = h_relu.t().mm(grad_y_pred)\n grad_h_relu = grad_y_pred.mm(w2.t())\n grad_h = grad_h_relu.clone()\n grad_h[h < 0] = 0\n grad_w1 = x.t().mm(grad_h)\n\n # 更新网络权值\n w1 -= lr * grad_w1\n w2 -= lr * grad_w2\n"},{"alpha_fraction":0.6621924042701721,"alphanum_fraction":0.6845637559890747,"avg_line_length":22.526315689086914,"blob_id":"58575110fac4ac08a65005af0b9a64a4c02c07d9","content_id":"2741ad8cabc91ebc103044396a87a2e40c2fbc63","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":791,"license_type":"no_license","max_line_length":38,"num_lines":19,"path":"/Part 1： 60分钟快速入门/README.md","repo_name":"LoveGalaxy/PyTorch_Tutorial","src_encoding":"UTF-8","text":"### 第一部分 60min 快速入门\n\n### Tutorial.py\n#### 这是官方文档 60min 快速入门 PyTorch 的第一部分，\n#### 主要介绍了 PyTorch 的一些基本语法\n\n### Autograd.py\n#### 这是官方文档 60min 快速入门 PyTorch 的第二部分，\n#### 主要介绍了 PyTorch 中的自动求导机制\n\n### NeuralNetwork.py\n#### 这是官方文档 60min 快速入门 PyTorch 的第三部分，\n#### 主要实现了一个简单的LeNet网络，介绍了用 Pytorch\n#### 搭建卷积神经网络的主要步骤\n\n### Simple Example.ipynb\n#### 这是官方文档 60min 快速入门 PyTorch 的第四、五部分\n#### 第四部分主要是用卷积神经网络做了一个图像分类器\n#### 第五部分主要是介绍了在多显卡的时候如何使用 Pytorch\n"},{"alpha_fraction":0.5052770376205444,"alphanum_fraction":0.5474933981895447,"avg_line_length":16.738094329833984,"blob_id":"106d5d2fb9b42fc61790c1b6ba7ddd8e6fe32aa9","content_id":"1e5c778723c766a63b893613a4d69897cb6be4d2","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":912,"license_type":"no_license","max_line_length":39,"num_lines":42,"path":"/Part 2： Pytorch简单例子/0 Use numpy to implement a NN.py","repo_name":"LoveGalaxy/PyTorch_Tutorial","src_encoding":"UTF-8","text":"import numpy as np\n\n# N 是 batch 大小\n# D_in, D_out 是神经网络输入和输出的维度\n# H 是神经网络隐藏层的维度\nN = 64\nD_in = 1000\nH = 100\nD_out = 10\n\n# 生成随机数据\nx = np.random.randn(N, D_in)\ny = np.random.randn(N, D_out)\n\n# 初始化神经网络参数权值\nw1 = np.random.rand(D_in, H)\nw2 = np.random.rand(H, D_out)\n\n# 设置学习速率\nlr = 1e-6\n\nfor t in range(500):\n # 前馈计算\n h = x.dot(w1)\n h_relu = np.maximum(h, 0)\n y_pred = h_relu.dot(w2)\n\n # 计算误差（loss），采用的是平方和\n loss = np.square(y_pred - y).sum()\n print(t, loss)\n\n # 反馈计算\n grad_y_pred = 2.0 * (y_pred - y)\n grad_w2 = h_relu.T.dot(grad_y_pred)\n grad_h_relu = grad_y_pred.dot(w2.T)\n grad_h = grad_h_relu.copy()\n grad_h[h < 0] = 0\n grad_w1 = x.T.dot(grad_h)\n\n # ，更新权值\n w1 -= lr * grad_w1\n w2 -= lr * grad_w2\n\n\n\n\n\n\n\n\n\n\n\n\n\n"},{"alpha_fraction":0.6407643556594849,"alphanum_fraction":0.6611465215682983,"avg_line_length":17.66666603088379,"blob_id":"c134b4954a8f6b164996b57f3470544e81224cb3","content_id":"700d17506b0fb9a0074a2ebbdd6de8120014ed36","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1065,"license_type":"no_license","max_line_length":55,"num_lines":42,"path":"/Part 2： Pytorch简单例子/6 Pytorch：Use optim.py","repo_name":"LoveGalaxy/PyTorch_Tutorial","src_encoding":"UTF-8","text":"import torch\n\nN, D_in, H, D_out = 64, 1000, 100, 10\n\nx = torch.randn(N, D_in)\ny = torch.randn(N, D_out)\n\nmodel = torch.nn.Sequential(\n torch.nn.Linear(D_in, H),\n torch.nn.ReLU(),\n torch.nn.Linear(H, D_out)\n)\n\nloss_fn = torch.nn.MSELoss(size_average=False)\n\nlr = 1e-4\n\n# 几种常见的优化器, SGD, Momentum, RMSprop, Adam\n\"\"\"\nSGD 是最普通的优化器, 也可以说没有加速效果, \n而 Momentum 是 SGD 的改良版, 它加入了动量原则. \n后面的 RMSprop 又是 Momentum 的升级版. \n而 Adam 又是 RMSprop 的升级版. \n不过从这个结果中我们看到, Adam 的效果似乎比 RMSprop 要差一点. \n所以说并不是越先进的优化器, 结果越佳. \n我们在自己的试验中可以尝试不同的优化器, 找到那个最适合你数据/网络的优化器.\n\"\"\"\noptimizer = torch.optim.Adam(model.parameters(), lr=lr)\n\n\nfor t in range(500):\n y_pred = model(x)\n\n loss = loss_fn(y_pred, y)\n print(t, loss)\n\n optimizer.zero_grad()\n\n loss.backward()\n\n # 使用优化器来更新梯度\n optimizer.step()\n\n"},{"alpha_fraction":0.7283372282981873,"alphanum_fraction":0.733021080493927,"avg_line_length":16.079999923706055,"blob_id":"97f817d4301cc4c0e59baa50e1c3c7d48d90cdd4","content_id":"c800116e66afc8f0941b93108a8394c586d77073","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":969,"license_type":"no_license","max_line_length":48,"num_lines":25,"path":"/README.md","repo_name":"LoveGalaxy/PyTorch_Tutorial","src_encoding":"UTF-8","text":"# PyTorch_Tutorial\nAbout PyTorch\n\n## 这是一个中文的 Pytorch Tutorial 项目\n\n### 本来想用英文来写，但是英文的资料已经很多了\n### 我的母语是中文，那我就写一个中文的入门代码\n### 现在的代码主要还是以官方的例子为主，我现在\n### 把介绍和注释改为中文，同时我也做了一些相关\n### 的实验，把这些实验得到的调参经验，以及学习\n### 到得深度学习知识加入注释，方便大家理解学习\n### 我会尽可能多的加入代码以及中文的注释\n\n## 阅读大纲\n\n### 第一部分 60min 快速入门\n\n### \n### 第二部分 Pytorch简单例子\n\n### 第三部分 迁移学习\n#### 目前我还不是很清楚迁移学习的具体知识，去补习一下，再来接着更新，估计会暂时跳票一段时间\n\n#### 本人知识也有限，错误之处希望得到大家的指正\n#### 我会尽量每天都更新这个库，任何您的意见我都会认真考虑\n"},{"alpha_fraction":0.6771568059921265,"alphanum_fraction":0.6966902017593384,"avg_line_length":25.897809982299805,"blob_id":"15e1e83328809d17b28d152182f7b0fc562dde44","content_id":"69fb35690355d9f050176f834ff500df7a0f6a88","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":4566,"license_type":"no_license","max_line_length":88,"num_lines":137,"path":"/Part 1： 60分钟快速入门/0 Tutorial.py","repo_name":"LoveGalaxy/PyTorch_Tutorial","src_encoding":"UTF-8","text":"import torch\n\n# python version is 3.6.4\n# torch version is 0.4.0\n# print(torch.__version__)\n\n# create a 5*3 matrix, uninitialized\n# 创建一个没有初始化的 5*3 的矩阵\nx = torch.empty(5, 3)\nprint(x)\n\n# create a 5*3 randomly initialized matrix\n# 创建一个随机初始化的 5*3 矩阵\nx = torch.rand(5, 3)\nprint(x)\n\n# create a matrix filled zeros and of dtype long\n# 创建一个由 0 组成的，数据类型为 long 的矩阵\nx = torch.zeros(5, 3, dtype=torch.long)\nprint(x)\n\n# construct a tensor directly from data\n# 直接从数据中创建 tensor 对象\nx = torch.tensor([5.5, 3])\nprint(x)\n\n# create a tensor basing on existing tensor\n# these methods will reuse properties of the input tensor\n# 首先创建一个全为1的矩阵\n# 然后创建一个和原矩阵一样大,由随机数生成的矩阵\nx = x.new_ones(5, 3, dtype=torch.double)\nprint(x)\nx = torch.randn_like(x, dtype=torch.float)\nprint(x)\n\n# get its size\n(i, j) = x.size() # torch.Size is in fact a tuple, so it supports all tuple operations.\nprint(i, j)\nprint(x.shape) # .shape == .size()\n\n# There are multiple syntaxes for operations\n# In the following example, we will take a look at the addition operation\n# 接下来，以加法为例介绍tensor类的运算操作\n\n\n# Addition: syntax 1\n# 加法： 第一种形式\ny = torch.rand(5, 3)\nprint(x + y)\n\n# Addition: syntax 2\n# 加法： 第二种形式\nprint(torch.add(x, y))\n\n# Addition: providing an output tensor as argument\n# 加法： 参数返回结果\nresult = torch.empty(5, 3)\ntorch.add(x, y, out=result)\nprint(result)\n\n# Addtion in-place\n# 加法：原地加法\n# Any operation that mutates a tensor in-place is post-fixed with an _.\n# For example: x.copy_(y), x.t_(), will change x\ny.add_(x)\nprint(y)\n\n# You can use standard NumPy-like indexing with all bells and whistles!\n# 可以像 numpy 一样索引\nprint(x[:, 1])\n\n# Resizing: If you want to resize/reshape tensor, you can use torch.view\n# 可以是用torch.view方法，重建tensor对象的维度\nx = torch.randn(4, 4, 3)\ny = x.view(3, 16)\nz = x.view(-1, 8) # the size -1 is inferred from other dimensions\nprint(x.size(), y.size(), z.size())\n# 这个方法是顺序整理\n# 例如上面代码中x生成了一个 4 维 3*3 的矩阵\n# 转换为 3 维，长度为 16 的向量时，都是顺序排列\nx = torch.randn(4, 4, 3)\ny = x.view(3, 4, 4)\n# 再考虑一下这个代码，我们把 3*4*4 的张量看做一个长方体\n# 事实上，通过 view 的变换的到得长方体并不是旋转过后的长方体\n# 而是按照顺序重排列之后的长方体\n\n# If you have a one element tensor, use .item() to get the value as a Python number\n# 如果你的 tensor 张量只有一个元素，使用 .item() 获取里面的值\nx = torch.randn(1)\nprint(x)\nprint(x.item())\n\n# The Torch Tensor and NumPy array will share their underlying memory locations,\n# and changing one will change the other.\n# torch tensor 可以和 numpy array 直接转换，但他们会共享内存，\n# 这意味着如果你改变了其中之一，另一个也会随之改变\na = torch.ones(5)\nprint(a)\nb = a.numpy()\nprint(b)\n# 改变了a中元素值，b中的值也随之改变\na.add_(1)\nprint(a)\nprint(b)\n\n# changing the np array changed the Torch Tensor automatically\n# numpy 转 tensor， 接着改变 numpy array， torch tensor 随之改变\n# 注意：不在文件头部 import 模块是不符合 PEP 8 规范的\nimport numpy as np\na = np.ones(5)\nb = torch.from_numpy(a)\nnp.add(a, 1, out=a)\nprint(a)\nprint(b)\n\n# All the Tensors on the CPU except a CharTensor support converting to NumPy and back.\n# 除了 CharTensor 外，别的 tensor 都支持与 numpy array 进行转换\n\n# Tensors can be moved onto any device using the .to method.\n# 用 .to 方法更换计算设备\n\n# let us run this cell only if CUDA is available\n# We will use ``torch.device`` objects to move tensors in and out of GPU\nif torch.cuda.is_available():\n device = torch.device(\"cuda\") # a CUDA device object\n y = torch.ones_like(x, device=device) # directly create a tensor on GPU\n x = x.to(device) # or just use strings ``.to(\"cuda\")``\n z = x + y\n print(z)\n print(z.to(\"cpu\", torch.double)) # ``.to`` can also change dtype together!\n\n# 有的电脑有CUDA，有的没有，为了使代码有更好的移植性，我们可以额这么来写\ndevice = \"cuda\" if torch.cuda.is_available() else \"cpu\"\nx = torch.rand(5, 3, device=device) # rand 0, 1均匀分布\ny = torch.randn(5, 3, device=device) # randn 均值0，方差1的正态分布\nz = x + y\nprint(z)\n\n"}],"string":"[\n {\n \"alpha_fraction\": 0.5605888962745667,\n \"alphanum_fraction\": 0.5979614853858948,\n \"avg_line_length\": 20.0238094329834,\n \"blob_id\": \"1be8d3d0fd002c0638e86f9e1a6b63d7db2a9761\",\n \"content_id\": \"50c3f7f2a302bb1602e7c50494443f3ca20779b8\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1035,\n \"license_type\": \"no_license\",\n \"max_line_length\": 54,\n \"num_lines\": 42,\n \"path\": \"/Part 2： Pytorch简单例子/1 Use Tnensor to implement a NN.py\",\n \"repo_name\": \"LoveGalaxy/PyTorch_Tutorial\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import torch\\n\\n# numpy已经为我们提供了强大的计算能力，但是为了能让我们的数据\\n# 能够在GPU上进行加速，我们需要将计算的类型转化为Tensor\\n# 设置torch tensor类型，设置计算设备\\ndtype = torch.float\\ndevice = torch.device(\\\"cpu\\\")\\n\\nN = 64\\nD_in = 1000\\nH = 100\\nD_out = 10\\n\\nx = torch.randn(N, D_in, device=device, dtype=dtype)\\ny = torch.randn(N, D_out, device=device, dtype=dtype)\\n\\nw1 = torch.randn(D_in, H, device=device, dtype=dtype)\\nw2 = torch.randn(H, D_out, device=device, dtype=dtype)\\n\\nlr = 1e-6\\n\\nfor t in range(500):\\n # 前馈计算\\n h = x.mm(w1)\\n h_relu = h.clamp(min=0)\\n y_pred = h_relu.mm(w2)\\n\\n # 计算损失\\n loss = (y_pred - y).pow(2).sum().item()\\n print(t, loss)\\n\\n # bp算法\\n grad_y_pred = 2.0 * (y_pred - y)\\n grad_w2 = h_relu.t().mm(grad_y_pred)\\n grad_h_relu = grad_y_pred.mm(w2.t())\\n grad_h = grad_h_relu.clone()\\n grad_h[h < 0] = 0\\n grad_w1 = x.t().mm(grad_h)\\n\\n # 更新网络权值\\n w1 -= lr * grad_w1\\n w2 -= lr * grad_w2\\n\"\n },\n {\n \"alpha_fraction\": 0.6621924042701721,\n \"alphanum_fraction\": 0.6845637559890747,\n \"avg_line_length\": 22.526315689086914,\n \"blob_id\": \"58575110fac4ac08a65005af0b9a64a4c02c07d9\",\n \"content_id\": \"2741ad8cabc91ebc103044396a87a2e40c2fbc63\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 791,\n \"license_type\": \"no_license\",\n \"max_line_length\": 38,\n \"num_lines\": 19,\n \"path\": \"/Part 1： 60分钟快速入门/README.md\",\n \"repo_name\": \"LoveGalaxy/PyTorch_Tutorial\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"### 第一部分 60min 快速入门\\n\\n### Tutorial.py\\n#### 这是官方文档 60min 快速入门 PyTorch 的第一部分，\\n#### 主要介绍了 PyTorch 的一些基本语法\\n\\n### Autograd.py\\n#### 这是官方文档 60min 快速入门 PyTorch 的第二部分，\\n#### 主要介绍了 PyTorch 中的自动求导机制\\n\\n### NeuralNetwork.py\\n#### 这是官方文档 60min 快速入门 PyTorch 的第三部分，\\n#### 主要实现了一个简单的LeNet网络，介绍了用 Pytorch\\n#### 搭建卷积神经网络的主要步骤\\n\\n### Simple Example.ipynb\\n#### 这是官方文档 60min 快速入门 PyTorch 的第四、五部分\\n#### 第四部分主要是用卷积神经网络做了一个图像分类器\\n#### 第五部分主要是介绍了在多显卡的时候如何使用 Pytorch\\n\"\n },\n {\n \"alpha_fraction\": 0.5052770376205444,\n \"alphanum_fraction\": 0.5474933981895447,\n \"avg_line_length\": 16.738094329833984,\n \"blob_id\": \"106d5d2fb9b42fc61790c1b6ba7ddd8e6fe32aa9\",\n \"content_id\": \"1e5c778723c766a63b893613a4d69897cb6be4d2\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 912,\n \"license_type\": \"no_license\",\n \"max_line_length\": 39,\n \"num_lines\": 42,\n \"path\": \"/Part 2： Pytorch简单例子/0 Use numpy to implement a NN.py\",\n \"repo_name\": \"LoveGalaxy/PyTorch_Tutorial\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import numpy as np\\n\\n# N 是 batch 大小\\n# D_in, D_out 是神经网络输入和输出的维度\\n# H 是神经网络隐藏层的维度\\nN = 64\\nD_in = 1000\\nH = 100\\nD_out = 10\\n\\n# 生成随机数据\\nx = np.random.randn(N, D_in)\\ny = np.random.randn(N, D_out)\\n\\n# 初始化神经网络参数权值\\nw1 = np.random.rand(D_in, H)\\nw2 = np.random.rand(H, D_out)\\n\\n# 设置学习速率\\nlr = 1e-6\\n\\nfor t in range(500):\\n # 前馈计算\\n h = x.dot(w1)\\n h_relu = np.maximum(h, 0)\\n y_pred = h_relu.dot(w2)\\n\\n # 计算误差（loss），采用的是平方和\\n loss = np.square(y_pred - y).sum()\\n print(t, loss)\\n\\n # 反馈计算\\n grad_y_pred = 2.0 * (y_pred - y)\\n grad_w2 = h_relu.T.dot(grad_y_pred)\\n grad_h_relu = grad_y_pred.dot(w2.T)\\n grad_h = grad_h_relu.copy()\\n grad_h[h < 0] = 0\\n grad_w1 = x.T.dot(grad_h)\\n\\n # ，更新权值\\n w1 -= lr * grad_w1\\n w2 -= lr * grad_w2\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.6407643556594849,\n \"alphanum_fraction\": 0.6611465215682983,\n \"avg_line_length\": 17.66666603088379,\n \"blob_id\": \"c134b4954a8f6b164996b57f3470544e81224cb3\",\n \"content_id\": \"700d17506b0fb9a0074a2ebbdd6de8120014ed36\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1065,\n \"license_type\": \"no_license\",\n \"max_line_length\": 55,\n \"num_lines\": 42,\n \"path\": \"/Part 2： Pytorch简单例子/6 Pytorch：Use optim.py\",\n \"repo_name\": \"LoveGalaxy/PyTorch_Tutorial\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import torch\\n\\nN, D_in, H, D_out = 64, 1000, 100, 10\\n\\nx = torch.randn(N, D_in)\\ny = torch.randn(N, D_out)\\n\\nmodel = torch.nn.Sequential(\\n torch.nn.Linear(D_in, H),\\n torch.nn.ReLU(),\\n torch.nn.Linear(H, D_out)\\n)\\n\\nloss_fn = torch.nn.MSELoss(size_average=False)\\n\\nlr = 1e-4\\n\\n# 几种常见的优化器, SGD, Momentum, RMSprop, Adam\\n\\\"\\\"\\\"\\nSGD 是最普通的优化器, 也可以说没有加速效果, \\n而 Momentum 是 SGD 的改良版, 它加入了动量原则. \\n后面的 RMSprop 又是 Momentum 的升级版. \\n而 Adam 又是 RMSprop 的升级版. \\n不过从这个结果中我们看到, Adam 的效果似乎比 RMSprop 要差一点. \\n所以说并不是越先进的优化器, 结果越佳. \\n我们在自己的试验中可以尝试不同的优化器, 找到那个最适合你数据/网络的优化器.\\n\\\"\\\"\\\"\\noptimizer = torch.optim.Adam(model.parameters(), lr=lr)\\n\\n\\nfor t in range(500):\\n y_pred = model(x)\\n\\n loss = loss_fn(y_pred, y)\\n print(t, loss)\\n\\n optimizer.zero_grad()\\n\\n loss.backward()\\n\\n # 使用优化器来更新梯度\\n optimizer.step()\\n\\n\"\n },\n {\n \"alpha_fraction\": 0.7283372282981873,\n \"alphanum_fraction\": 0.733021080493927,\n \"avg_line_length\": 16.079999923706055,\n \"blob_id\": \"97f817d4301cc4c0e59baa50e1c3c7d48d90cdd4\",\n \"content_id\": \"c800116e66afc8f0941b93108a8394c586d77073\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 969,\n \"license_type\": \"no_license\",\n \"max_line_length\": 48,\n \"num_lines\": 25,\n \"path\": \"/README.md\",\n \"repo_name\": \"LoveGalaxy/PyTorch_Tutorial\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# PyTorch_Tutorial\\nAbout PyTorch\\n\\n## 这是一个中文的 Pytorch Tutorial 项目\\n\\n### 本来想用英文来写，但是英文的资料已经很多了\\n### 我的母语是中文，那我就写一个中文的入门代码\\n### 现在的代码主要还是以官方的例子为主，我现在\\n### 把介绍和注释改为中文，同时我也做了一些相关\\n### 的实验，把这些实验得到的调参经验，以及学习\\n### 到得深度学习知识加入注释，方便大家理解学习\\n### 我会尽可能多的加入代码以及中文的注释\\n\\n## 阅读大纲\\n\\n### 第一部分 60min 快速入门\\n\\n### \\n### 第二部分 Pytorch简单例子\\n\\n### 第三部分 迁移学习\\n#### 目前我还不是很清楚迁移学习的具体知识，去补习一下，再来接着更新，估计会暂时跳票一段时间\\n\\n#### 本人知识也有限，错误之处希望得到大家的指正\\n#### 我会尽量每天都更新这个库，任何您的意见我都会认真考虑\\n\"\n },\n {\n \"alpha_fraction\": 0.6771568059921265,\n \"alphanum_fraction\": 0.6966902017593384,\n \"avg_line_length\": 25.897809982299805,\n \"blob_id\": \"15e1e83328809d17b28d152182f7b0fc562dde44\",\n \"content_id\": \"69fb35690355d9f050176f834ff500df7a0f6a88\",\n \"detected_licenses\": [],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 4566,\n \"license_type\": \"no_license\",\n \"max_line_length\": 88,\n \"num_lines\": 137,\n \"path\": \"/Part 1： 60分钟快速入门/0 Tutorial.py\",\n \"repo_name\": \"LoveGalaxy/PyTorch_Tutorial\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import torch\\n\\n# python version is 3.6.4\\n# torch version is 0.4.0\\n# print(torch.__version__)\\n\\n# create a 5*3 matrix, uninitialized\\n# 创建一个没有初始化的 5*3 的矩阵\\nx = torch.empty(5, 3)\\nprint(x)\\n\\n# create a 5*3 randomly initialized matrix\\n# 创建一个随机初始化的 5*3 矩阵\\nx = torch.rand(5, 3)\\nprint(x)\\n\\n# create a matrix filled zeros and of dtype long\\n# 创建一个由 0 组成的，数据类型为 long 的矩阵\\nx = torch.zeros(5, 3, dtype=torch.long)\\nprint(x)\\n\\n# construct a tensor directly from data\\n# 直接从数据中创建 tensor 对象\\nx = torch.tensor([5.5, 3])\\nprint(x)\\n\\n# create a tensor basing on existing tensor\\n# these methods will reuse properties of the input tensor\\n# 首先创建一个全为1的矩阵\\n# 然后创建一个和原矩阵一样大,由随机数生成的矩阵\\nx = x.new_ones(5, 3, dtype=torch.double)\\nprint(x)\\nx = torch.randn_like(x, dtype=torch.float)\\nprint(x)\\n\\n# get its size\\n(i, j) = x.size() # torch.Size is in fact a tuple, so it supports all tuple operations.\\nprint(i, j)\\nprint(x.shape) # .shape == .size()\\n\\n# There are multiple syntaxes for operations\\n# In the following example, we will take a look at the addition operation\\n# 接下来，以加法为例介绍tensor类的运算操作\\n\\n\\n# Addition: syntax 1\\n# 加法： 第一种形式\\ny = torch.rand(5, 3)\\nprint(x + y)\\n\\n# Addition: syntax 2\\n# 加法： 第二种形式\\nprint(torch.add(x, y))\\n\\n# Addition: providing an output tensor as argument\\n# 加法： 参数返回结果\\nresult = torch.empty(5, 3)\\ntorch.add(x, y, out=result)\\nprint(result)\\n\\n# Addtion in-place\\n# 加法：原地加法\\n# Any operation that mutates a tensor in-place is post-fixed with an _.\\n# For example: x.copy_(y), x.t_(), will change x\\ny.add_(x)\\nprint(y)\\n\\n# You can use standard NumPy-like indexing with all bells and whistles!\\n# 可以像 numpy 一样索引\\nprint(x[:, 1])\\n\\n# Resizing: If you want to resize/reshape tensor, you can use torch.view\\n# 可以是用torch.view方法，重建tensor对象的维度\\nx = torch.randn(4, 4, 3)\\ny = x.view(3, 16)\\nz = x.view(-1, 8) # the size -1 is inferred from other dimensions\\nprint(x.size(), y.size(), z.size())\\n# 这个方法是顺序整理\\n# 例如上面代码中x生成了一个 4 维 3*3 的矩阵\\n# 转换为 3 维，长度为 16 的向量时，都是顺序排列\\nx = torch.randn(4, 4, 3)\\ny = x.view(3, 4, 4)\\n# 再考虑一下这个代码，我们把 3*4*4 的张量看做一个长方体\\n# 事实上，通过 view 的变换的到得长方体并不是旋转过后的长方体\\n# 而是按照顺序重排列之后的长方体\\n\\n# If you have a one element tensor, use .item() to get the value as a Python number\\n# 如果你的 tensor 张量只有一个元素，使用 .item() 获取里面的值\\nx = torch.randn(1)\\nprint(x)\\nprint(x.item())\\n\\n# The Torch Tensor and NumPy array will share their underlying memory locations,\\n# and changing one will change the other.\\n# torch tensor 可以和 numpy array 直接转换，但他们会共享内存，\\n# 这意味着如果你改变了其中之一，另一个也会随之改变\\na = torch.ones(5)\\nprint(a)\\nb = a.numpy()\\nprint(b)\\n# 改变了a中元素值，b中的值也随之改变\\na.add_(1)\\nprint(a)\\nprint(b)\\n\\n# changing the np array changed the Torch Tensor automatically\\n# numpy 转 tensor， 接着改变 numpy array， torch tensor 随之改变\\n# 注意：不在文件头部 import 模块是不符合 PEP 8 规范的\\nimport numpy as np\\na = np.ones(5)\\nb = torch.from_numpy(a)\\nnp.add(a, 1, out=a)\\nprint(a)\\nprint(b)\\n\\n# All the Tensors on the CPU except a CharTensor support converting to NumPy and back.\\n# 除了 CharTensor 外，别的 tensor 都支持与 numpy array 进行转换\\n\\n# Tensors can be moved onto any device using the .to method.\\n# 用 .to 方法更换计算设备\\n\\n# let us run this cell only if CUDA is available\\n# We will use ``torch.device`` objects to move tensors in and out of GPU\\nif torch.cuda.is_available():\\n device = torch.device(\\\"cuda\\\") # a CUDA device object\\n y = torch.ones_like(x, device=device) # directly create a tensor on GPU\\n x = x.to(device) # or just use strings ``.to(\\\"cuda\\\")``\\n z = x + y\\n print(z)\\n print(z.to(\\\"cpu\\\", torch.double)) # ``.to`` can also change dtype together!\\n\\n# 有的电脑有CUDA，有的没有，为了使代码有更好的移植性，我们可以额这么来写\\ndevice = \\\"cuda\\\" if torch.cuda.is_available() else \\\"cpu\\\"\\nx = torch.rand(5, 3, device=device) # rand 0, 1均匀分布\\ny = torch.randn(5, 3, device=device) # randn 均值0，方差1的正态分布\\nz = x + y\\nprint(z)\\n\\n\"\n }\n]"},"num_files":{"kind":"number","value":6,"string":"6"}}},{"rowIdx":2980389,"cells":{"repo_name":{"kind":"string","value":"Thelin90/G.D.C.K"},"repo_url":{"kind":"string","value":"https://github.com/Thelin90/G.D.C.K"},"snapshot_id":{"kind":"string","value":"3ec4a1d6fcd4d4695a846a118b49dd68d61e5113"},"revision_id":{"kind":"string","value":"dcaf437d119b874654fa58dbd560a4545a555864"},"directory_id":{"kind":"string","value":"74e008a03e04c55afb7a1bc0dfdc8c09ae56648e"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2022-11-20T01:37:38.097678","string":"2022-11-20T01:37:38.097678"},"revision_date":{"kind":"timestamp","value":"2019-06-04T11:00:11","string":"2019-06-04T11:00:11"},"committer_date":{"kind":"timestamp","value":"2019-06-04T11:00:11","string":"2019-06-04T11:00:11"},"github_id":{"kind":"number","value":134684092,"string":"134,684,092"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"string","value":"Apache-2.0"},"gha_created_at":{"kind":"timestamp","value":"2018-05-24T08:18:58","string":"2018-05-24T08:18:58"},"gha_updated_at":{"kind":"timestamp","value":"2019-06-04T11:00:19","string":"2019-06-04T11:00:19"},"gha_pushed_at":{"kind":"timestamp","value":"2022-11-11T07:32:39","string":"2022-11-11T07:32:39"},"gha_language":{"kind":"string","value":"Python"},"files":{"kind":"list like","value":[{"alpha_fraction":0.6452810168266296,"alphanum_fraction":0.6930010318756104,"avg_line_length":29.419355392456055,"blob_id":"5fb256f9581199a758db732b0ffe6192d2cb8029","content_id":"8c04384f6998befc03967637e97af16c61f459eb","detected_licenses":["Apache-2.0","MIT"],"is_generated":false,"is_vendor":false,"language":"Dockerfile","length_bytes":1886,"license_type":"permissive","max_line_length":113,"num_lines":62,"path":"/Dockerfile","repo_name":"Thelin90/G.D.C.K","src_encoding":"UTF-8","text":"\nFROM ubuntu:latest\n\n# PATH\nENV PATH /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin\n# Spark\nENV SPARK_VERSION 2.3.1\nENV SPARK_HOME /usr/local/spark\nENV SPARK_LOG_DIR /var/log/spark\nENV SPARK_PID_DIR /var/run/spark\nENV PATH $PATH:$SPARK_HOME/bin\nENV PYSPARK_PYTHON /usr/bin/python3.6\nENV PYSPARK_DRIVER_PYTHON=/usr/bin/python3.6\nENV PYTHONUNBUFFERED 1\n# Java\nENV JAVA_HOME /usr/lib/jvm/java-8-openjdk-amd64/\n# Python\nENV alias python=/usr/bin/python3.6\nENV PYTHONPATH /etc/app/src/\n\n# Install curl\nRUN apt-get update && apt-get install -y curl\n\n# Install Python 3.6, 2.7 is standard to ubuntu:latest\nRUN apt-get update && \\\n apt-get install -y python3.6 && \\\n apt-get install -y python3-pip\n\n# Install Java, had an issue found this: https://stackoverflow.com/questions/46795907/setting-java-home-in-docker\nRUN apt-get update && apt-get install -y openjdk-8-jdk && \\\n apt-get install -y ant && apt-get clean && \\\n rm -rf /var/lib/apt/lists/ && \\\n rm -rf /var/cache/oracle-jdk8-installer;\n\n# Set workspace\nWORKDIR /etc/app\n\n# Add all the project files to the\nADD . /etc/app\n\n# Download Spark\nRUN curl -L http://www.us.apache.org/dist/spark/spark-${SPARK_VERSION}/spark-${SPARK_VERSION}-bin-hadoop2.7.tgz \\\n | tar -xzp -C /usr/local/ && \\\n ln -s spark-${SPARK_VERSION}-bin-hadoop2.7 ${SPARK_HOME}\n\n# Make run.sh executable\nRUN chmod +x /etc/app/scripts/run.sh && chmod +x /etc/app/src/data/input_data\n\n# Give give -rw-r--r-- to python files\nRUN chmod 0644 /etc/app/src/etl.py && \\\n chmod 0644 /etc/app/src/geoip.py && \\\n chmod 0644 /etc/app/src/data/input_data\n\nEXPOSE 8080 8081 6066 \\\n 7077 4040 7001 \\\n 7002 7003 7004 \\\n 7005 7006\n\nENTRYPOINT [\"./scripts/run.sh\"]\n\n# Replace the Entrypoint running the run.sh with this\n# to keep the container alive, tobe able to debug the container\n# ENTRYPOINT [\"tail\", \"-f\", \"/dev/null\"]"},{"alpha_fraction":0.6243016719818115,"alphanum_fraction":0.6298882961273193,"avg_line_length":24.571428298950195,"blob_id":"fbc9ac73f2a3a05852f4d095dbb3677678f7e00f","content_id":"830a13247f6aa8446bdfdd8d31e0cd805aaa47e5","detected_licenses":["Apache-2.0","MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":716,"license_type":"permissive","max_line_length":60,"num_lines":28,"path":"/test/fakesparksession.py","repo_name":"Thelin90/G.D.C.K","src_encoding":"UTF-8","text":"import logging\nfrom pyspark.sql import SparkSession\n\nimport unittest\n\nclass PySparkTest(unittest.TestCase):\n\n @classmethod\n def suppress_py4j_logging(cls):\n logger = logging.getLogger(\"py4j\")\n logger.setLevel(logging.WARN)\n\n @classmethod\n def create_testing_pyspark_session(cls):\n return (SparkSession.builder\n .master(\"local[2]\")\n .appName(\"my-local-testing-pyspark-context\")\n .enableHiveSupport()\n .getOrCreate())\n\n @classmethod\n def setUpClass(cls):\n cls.suppress_py4j_logging()\n cls.spark = cls.create_testing_pyspark_session()\n\n @classmethod\n def tearDownClass(cls):\n cls.spark.stop()\n"},{"alpha_fraction":0.6188467144966125,"alphanum_fraction":0.6244725584983826,"avg_line_length":26.346153259277344,"blob_id":"fe50ca6fa3975efbc46fe872e8a4f478f363f661","content_id":"2990b0de72ecb9ee2f0e7c79c8862f54f8331071","detected_licenses":["Apache-2.0","MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1422,"license_type":"permissive","max_line_length":109,"num_lines":52,"path":"/src/geoip.py","repo_name":"Thelin90/G.D.C.K","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n# coding=utf-8\n\n\"\"\"\nPython script to transform ip to country and city, has been modified\n\"\"\"\n\nimport socket\nimport pygeoip\nimport os\n\n\"\"\" http://geolite.maxmind.com/download/geoip/database/GeoLiteCity.dat.gz \"\"\"\ncwd = os.getcwd()\nrawdata = pygeoip.GeoIP(cwd + \"/src/geoconvertdat/GeoLiteCity.dat\")\n\n\ndef ipquery(ip):\n \"\"\"Function to parse IP to country, city\n http://www.linuxx.eu/2014/05/geolocate-ip-with-python.html\n Have been slightly modified\n\n In Python 3, all strings are sequences of Unicode characters. There is a bytes type that holds raw bytes.\n\n Args:\n ip: The actual IP used to parse the country and city from\n\n Returns: A str value of \"country-city\"\n\n \"\"\"\n\n try:\n socket.inet_aton(str(ip))\n data = rawdata.record_by_name(ip)\n\n if type(data) is dict:\n\n country = type(data['country_name']) is str and data['country_name'] or type(\n data['country_name']) and data['country_name'] or \"NotTraceable\"\n\n city = type(data['city']) and data['city'] or type(\n data['city']) is str and data['city'] or \"NotTraceable\"\n\n if isinstance(country, bytes):\n country = country.decode()\n\n if isinstance(city, bytes):\n city = city.decode()\n\n return country + \"-\" + city\n\n except socket.error:\n return \"NotTraceable-NotTraceable\"\n"},{"alpha_fraction":0.6662476658821106,"alphanum_fraction":0.6900062561035156,"avg_line_length":30.947792053222656,"blob_id":"cbaf446cbe48dcdca50cb6d0cafdbe5f6f885dd4","content_id":"a92ebbdd08c0de53aebac4b300408dca466c3be5","detected_licenses":["Apache-2.0","MIT"],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":7955,"license_type":"permissive","max_line_length":567,"num_lines":249,"path":"/README.md","repo_name":"Thelin90/G.D.C.K","src_encoding":"UTF-8","text":"# Overview\n\n## Trello Feature Board\n\n* [Board](https://trello.com/b/qigu0NnW/gdck)\n\n\n\nGarden of Data Creation Kit `(An edit to the famous G.E.C.K in the Bethesda Fallout game series`) is a project that processes large amounts of data, currently TSV files and performs extraction, transformation and loading (ETL).\n\n## Project Design\n\nThe design of this project is that it should be fast and easy to deploy and run. Below is a description of how the application should be deployment ready.\n\n\n\nMicroservices is an architectural style that structures an application as a collection of loosely coupled services. Therefore enrich business capabilities, such as continuous delivery/deployment of scale.\n\n## Data Processing\n\nApache Spark is the chosen tool used within this project. Spark is quick and very responsive tool to perform data processing with. It provides an analytic engine for large-scale data processing. It is a general distributed in-memory computing framework implemented in scala. Hence spark operates on distributed data collections. However it does not do distributed storage. Spark is proven to be much faster than the popular Apache Hadoop framework. Apache Spark and Hadoop have different purposes but works good together. A high level overview of Apache Spark below: \n\n\n\nHadoop is essentially a distributed data infrastructure. It provides distributes big data collections across multiple nodes within a cluster of commodity servers. A high level overview of Apache Hadoop below: \n\n\n\n\n## PySpark\n\nPySpark have some core concepts. Such as resilient distributed datasets and dataframes.\n\nThis project spefically take advantage of the dataframes.\n\nIt has derived from the resilient distributed datasets concept. Dataframes will provide a higher level of abstraction. Which implicitly allow a query language way of typing to manipulate data. The abstraction represents data and a schema. Which enable better interaction with front end applications.\n\n## Dataset\nThe dataset has the (default) values divided by tab in the given dataset provided (input data): field header (date, time, user_id, url, IP, user_agent_string). See attached source file input_data.\n\n## Setup\n\nClone directly from the source code:\n\n```bash\ngit clone https://github.com/Thelin90/G.D.C.K.git\n```\n\nDocker needs to be installed on the machine. It can also run locally without docker.\n\n### Requirements\n\n* Docker environment\n* Python ^3.6.5\n* Java ^8\n* Spark ^2.3.x `(mininum 2.3.0)`\n\nAssuming that Python, Docker and Java is already setup.\n\n### Setup Apache-Spark\n\nStart with downloading Spark (note that depending on your IDE, you need to specify your Spark location):\n\n- https://spark.apache.org/downloads.html\n\nSet your SPARK_HOME in `.bashrc`\n```bash\nSPARK_HOME='path-to-spark'\n```\n\nThen source the file\n\n```bash\nsource ~/.bashrc\n```\n\n#### PyCharm Example\n\nMark as source:\n\n```bash\nFile -> Settings -> Project Structure -> add root content '+' -> 'path-to-spark'\n```\n\nSet environment variables:\n\n```bash\nRun -> Edit Configuration -> Environment Variables -> add new environment variables\n```\n```bash\nNAME\t\t\t\tVALUE\nPYSPARK_PYTHON\t\t\t'path-to-python'\nPYSPARK_DRIVER_PYTHON\t\t'path-to-python'\n```\n\n### Testing\n\nRun the following command to run tests:\n\n```bash\npython3 -W ignore:ResourceWarning -m unittest test/unittest-etl-process.py\n```\n\n### Manual Run\n\nRemember to set PYTHONPATH in `~/.bashrc`\n\n```bash\nPYTHONPATH=path-to-proj/PySparkDocker/src/\n```\n\nThen `source ~/.bashrc`\n\n\nRun the script manually without using docker.\n\n```bash\nspark-submit src/etl.py\n```\n\n### Run ETL Docker\n\n```bash\ndocker build -t etl-cities-countries .\ndocker run etl-cities-countries\n```\n\n## Result\n\n* Top 5 Countries based on number of events\n* Top 5 Cities based on number of events \n* Top 5 Browsers based on number of unique users\n* Top 5 Operating systems based on number of unique users.\n\n```bash\ndate and time column is becomming one timestamp...\nThe user_agent_string is becomming os and browser...\nConverting IP adress to city and country... \nPrinting Transformed Dataframe Schema\n |-- eventID: long (nullable = false)\n |-- timestamp: timestamp (nullable = true)\n |-- user_id: string (nullable = true)\n |-- url: string (nullable = true)\n |-- os: string (nullable = true)\n |-- browser: string (nullable = true)\n |-- country: string (nullable = true)\n |-- city: string (nullable = true)\n\nPerform load\nTop 5 countries based on number of events\n+--------------+------+\n| country| count|\n+--------------+------+\n|United Kingdom|135831|\n| Ireland| 18522|\n| Sweden| 12143|\n| Norway| 6908|\n| Netherlands| 2131|\n+--------------+------+\nonly showing top 5 rows\n\nTop 5 cities based on number of events\n+------------+-----+\n| city|count|\n+------------+-----+\n|NotTraceable|15336|\n| Dublin| 5750|\n| London| 4827|\n| Manchester| 3736|\n| Bristol| 2801|\n+------------+-----+\nonly showing top 5 rows\n\nTop 5 Browsers based on number of unique users\n+--------------------+-----------------------+\n| browser|count(DISTINCT user_id)|\n+--------------------+-----------------------+\n| Safari 7.0| 27117|\n| Safari 8.0| 17941|\n|Chrome 37.0.2062.124| 5054|\n| Safari 7.0.6| 2804|\n| Safari 7.1| 2486|\n+--------------------+-----------------------+\nonly showing top 5 rows\n\nTop 5 Operating systems based on number of unique users\n+----------------+-----------------------+\n| os|count(DISTINCT user_id)|\n+----------------+-----------------------+\n| IPad iOS 8.0.2| 11126|\n| IPad iOS 7.1.2| 10786|\n| Windows 7| 6722|\n|iPhone iOS 8.0.2| 4651|\n|iPhone iOS 7.1.2| 4496|\n+----------------+-----------------------+\nonly showing top 5 rows\n\nSpark application ends\n```\n\n### Spark UI\n\nTo see the spark jobs process do the following while running the docker image:\n\n```bash\ndocker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' etl-cities-countries\n```\n\nThis will give the IP of the docker container, type following in the web browser:\n\n`container-ip:4040`\n\nThis will give the following output:\n\n\n\n## License\n\nMIT License\n\nCopyright (c) [2018] [Simon Thelin]\n\nPermission is hereby granted, free of charge, to any person obtaining a copy\nof this software and associated documentation files (the \"Software\"), to deal\nin the Software without restriction, including without limitation the rights\nto use, copy, modify, merge, publish, distribute, sublicense, and/or sell\ncopies of the Software, and to permit persons to whom the Software is\nfurnished to do so, subject to the following conditions:\n\nThe above copyright notice and this permission notice shall be included in all\ncopies or substantial portions of the Software.\n\nTHE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\nIMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\nFITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\nAUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\nLIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\nOUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\nSOFTWARE.\n\n[Back To The Top](#read-me-template)\n\n---\n\n## Author Info\n\n* [Simon Thelin](https://www.linkedin.com/in/simon-thelin-3ba96986/)\n\n[Back To The Top](#read-me-template)\n"},{"alpha_fraction":0.752010703086853,"alphanum_fraction":0.752010703086853,"avg_line_length":31.434782028198242,"blob_id":"27b855b3bb9675ebc846937a224d9e3a18fd83a7","content_id":"b54dcb1d8491cdd045112f839504a41df3eecf8b","detected_licenses":["Apache-2.0","MIT"],"is_generated":false,"is_vendor":false,"language":"Markdown","length_bytes":746,"license_type":"permissive","max_line_length":151,"num_lines":23,"path":"/docs/PULL_REQUEST_TEMPLATE.md","repo_name":"Thelin90/G.D.C.K","src_encoding":"UTF-8","text":"# Pull Request Template\n\nPlease read this template to ensure your pull request adheres to the given guidelines.\n\n## Guidelines\n- Use the following format for header: `TD/xxxx/G.D.C.K-text-what-has-been-done` (the repo owner will fill in the xxxx to fit the actual ticket number)\n- Link additions should be added to the bottom of the relevant category.\n- New features or improvements to the existing features are welcome.\n\n`NOTE! ACHTUNG!`\n- Search previous suggestions before making a new one, as yours may be a duplicate.\n- Do not use `#` for commenting the PR commit message, only use `-`.\n\nGiven example:\n\n```\nTD-xxxx/G.D.C.K-some-fixes-to-something\n\n- New feature A\n - additional link\n- Improvement to feature B\n```\nThanks for contributing!\n"},{"alpha_fraction":0.7404580116271973,"alphanum_fraction":0.7557252049446106,"avg_line_length":25.399999618530273,"blob_id":"0564913906bfe9bbe39a4b0f0d74e9bf5a68a34e","content_id":"66cf32f9fd237e9e0b70d16627505d966992ad72","detected_licenses":["Apache-2.0","MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":131,"license_type":"permissive","max_line_length":54,"num_lines":5,"path":"/test/__init__.py","repo_name":"Thelin90/G.D.C.K","src_encoding":"UTF-8","text":"# For relative imports to work in Python 3.6\nfrom os import path\nimport sys\n\nsys.path.append(path.dirname(path.realpath(__file__)))"},{"alpha_fraction":0.4475524425506592,"alphanum_fraction":0.6853147149085999,"avg_line_length":14.88888931274414,"blob_id":"15252489b4854128f28486d0498c5e67c5ad7a97","content_id":"5a051799a5149e6329ec8141c1ecf4a3e8f96c93","detected_licenses":["Apache-2.0","MIT"],"is_generated":false,"is_vendor":false,"language":"Text","length_bytes":143,"license_type":"permissive","max_line_length":22,"num_lines":9,"path":"/scripts/requirements.txt","repo_name":"Thelin90/G.D.C.K","src_encoding":"UTF-8","text":"httpagentparser==1.8.1\nnumpy==1.15.0\npandas==0.23.4\npy4j==0.10.7\npygeoip==0.3.2\npyspark==2.3.2\npython-dateutil==2.7.3\npytz==2018.4\nsix==1.11.0\n"},{"alpha_fraction":0.6843373775482178,"alphanum_fraction":0.6867470145225525,"avg_line_length":30.923076629638672,"blob_id":"506e9683b54f192e1920cb3cc55bde603338da42","content_id":"001abec6b008146ba3b8da8c113370323d9ade9b","detected_licenses":["Apache-2.0","MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":415,"license_type":"permissive","max_line_length":114,"num_lines":13,"path":"/test/unittest-etl-process.py","repo_name":"Thelin90/G.D.C.K","src_encoding":"UTF-8","text":"#!/usr/bin/env python3\nfrom .fakesparksession import PySparkTest\nfrom etl import extract\n\n\nclass SimpleTest(PySparkTest):\n \"\"\"\n Will perform test on the functions used within the ETL process with a fake spark session\n \"\"\"\n def test_basic(self):\n expected_df = extract(self.spark)\n\n self.assertEquals(expected_df.schema.names, [\"date\", \"time\", \"user_id\", \"url\", \"ip\", \"user_agent_string\"])\n"},{"alpha_fraction":0.7422680258750916,"alphanum_fraction":0.7628865838050842,"avg_line_length":31.33333396911621,"blob_id":"1dfe90110602e682728026fb086f76f4d9bd628c","content_id":"c88a7019a69e12ee4353d654b01372f8f7354c8c","detected_licenses":["Apache-2.0","MIT"],"is_generated":false,"is_vendor":false,"language":"Shell","length_bytes":97,"license_type":"permissive","max_line_length":52,"num_lines":3,"path":"/scripts/run.sh","repo_name":"Thelin90/G.D.C.K","src_encoding":"UTF-8","text":"#!/usr/bin/env bash\npython3.6 -m pip install -r scripts/requirements.txt\nspark-submit src/etl.py\n"},{"alpha_fraction":0.645996630191803,"alphanum_fraction":0.6500644683837891,"avg_line_length":28.214492797851562,"blob_id":"dd9e91c9780ba5166792bdb4c6beeee7a19e508d","content_id":"fa967ba6ab9f466180def161e33d3eaee34b6411","detected_licenses":["Apache-2.0","MIT"],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":10079,"license_type":"permissive","max_line_length":116,"num_lines":345,"path":"/src/etl.py","repo_name":"Thelin90/G.D.C.K","src_encoding":"UTF-8","text":"#!/usr/bin/env python\n\n\"\"\"\nPython script to initiate and perform the ETL process\n\"\"\"\n\nimport os\nimport pyspark\nimport httpagentparser\nfrom pyspark.sql.functions import struct\nfrom pyspark.sql.functions import *\nfrom os.path import abspath\nfrom pyspark.sql import SparkSession\nfrom pyspark.sql.types import TimestampType, StringType\nfrom geoip import ipquery\n\n\ndef addMonoticallyIncreasingID(_df, _colname):\n \"\"\"Adds monotically increasing ID column to Spark Dataframe\n\n Args:\n _df: Spark Dataframe\n _colname: Name of the col containing the monotoically increasing ID\n\n Returns: Spark Dataframe with a column containing a monotoically increasing ID\n \"\"\"\n if isinstance(_colname, str):\n return _df.withColumn(_colname, monotonically_increasing_id())\n else:\n raise ValueError\n\n\ndef castToTimeStamp(_df, colname):\n \"\"\" Function casts column values to TimestampType\n Args:\n _df: Spark Dataframe\n colname: Name of column\n\n Returns\n \"\"\"\n coldatatype = _df.schema[colname].dataType\n if isinstance(coldatatype, TimestampType):\n raise ValueError\n else:\n print(\"Column \" + colname + \" with datatype \" + str(coldatatype) + \" typecasting to TimestampType\")\n return _df.withColumn(colname, _df[colname].cast(TimestampType()))\n\n\ndef renameColumn(_df, _oldcolname, _newcolname):\n \"\"\"Rename a column name\n Args:\n _df:\n _oldcolname:\n _newcolname:\n\n Returns:\n \"\"\"\n if isinstance(_oldcolname, str) and isinstance(_newcolname, str):\n return _df.withColumnRenamed(_oldcolname, _newcolname)\n else:\n raise ValueError(\"Column names must be of typ str\")\n\n\ndef concatColName(_df, _newcol, _oldcols, separator):\n \"\"\"\n Args:\n _df: Spark Dataframe\n newcol: Name of the new column\n oldcols: The names of the old columns\n separator: The separator value, examples: `;` `-` `.` `,`\n\n Returns\n \"\"\"\n if isinstance(_newcol, str) and isinstance(_oldcols, list):\n return _df.withColumn(_newcol, concat_ws(separator, _df[_oldcols[0]], _df[_oldcols[1]]))\n else:\n raise ValueError\n\n\ndef splitCol(_df, _split, _colnames):\n \"\"\"Creates 3 new columns from 1 original column\n\n TODO: N new columns from 1 original column\n\n Args:\n _dataframe: The dataframe that needs to split columns\n _split: The value to split on, ex: \"-\", \",\", \"*\"\n _colnames: The new names for the new columns\n\n Returns:\n\n \"\"\"\n split_col = pyspark.sql.functions.split(_df[_colnames[0]], _split)\n ret_df = _df.withColumn(_colnames[1], split_col.getItem(0))\n return ret_df.withColumn(_colnames[2], split_col.getItem(1))\n\n\ndef dropColumns(_df, colnames):\n \"\"\"Drops N number of undesired columns\n\n Args:\n _df: Spark Dataframe\n columnames: List containing column names to be dropped\n\n Returns: Spark Dataframe\n\n \"\"\"\n if isinstance(colnames, list):\n return _df.drop(*colnames)\n else:\n raise ValueError(\"colnames is not a list\")\n\n\ndef mergeDataframes(_df0, _df1, _colnames, _orderbycols):\n \"\"\"Merges two Spark Dataframes to one\n\n Args:\n _df0: Spark Dataframe 0\n _df1: Spark Dataframe 1 to be joined with Spark Datafram 0\n colnames: list of column names to be set\n orderbycol: list containing id value to join by\n\n Returns:\n\n \"\"\"\n print(\"Merge Spark Dataframes...\")\n ret_df = _df0.join(_df1, _df0[_orderbycols[0]] == _df1[_orderbycols[1]])\n ret_df = ret_df.orderBy(_orderbycols[0], ascending=True)\n ret_df = ret_df.select(*_colnames)\n\n return ret_df\n\n\ndef udfCreator(_df, _functioncall, _newcolname, _oldcolname, datatype):\n \"\"\"\n Args:\n _df: Spark Dataframe\n _functioncall: The function to be called\n _newcolname: The name of the new column containing the extracted values\n _oldcolname: The column values from a specific column to be sent, example: ip addresses\n datatype: Which datatype to work with, example: StringType()\n\n Returns: Spark Dataframe with one new additional column containing values from the UDF call\n \"\"\"\n if isinstance(_newcolname, str) and isinstance(_oldcolname, str):\n print(\"Making UDF call to function: \" + str(_functioncall))\n _udf = udf(_functioncall, datatype)\n return _df.withColumn(_newcolname, _udf(_df[_oldcolname]))\n else:\n raise Exception(\"Column names must be of str\")\n\n\ndef getOsBrowser(value):\n \"\"\"Calls the httpagentparser and retrieves the os and browser information\n\n Args:\n value: Each column value of user_agent_string\n\n Returns: The browser and os as a string\n\n \"\"\"\n if isinstance(value, str):\n return str(httpagentparser.simple_detect(value)[0] + \"-\" + httpagentparser.simple_detect(value)[1])\n else:\n raise ValueError\n\n\ndef getCityAndCountry(_ip):\n \"\"\"Function to make call to ipquery, may seem redundant but it makes the code more readable when doing UDF calls\n\n Args:\n _ip: IP address value\n\n Returns: A value of country-city from the IP\n \"\"\"\n return ipquery(_ip)\n\n\ndef load(_df):\n \"\"\"Load function to print the result and to save the dataframe for api calls\n\n TODO: Save dataframe/dataframes to postgres\n\n Args:\n _df: The final dataframe\n\n Returns: Nothing\n\n \"\"\"\n # Peform load process\n\n print(\"Displaying the top 5 rows\")\n _df.show(5)\n\n print(\"Top 5 countries based on number of events\")\n _df.groupBy(\"country\").count().orderBy(\"count\", ascending=False) \\\n .show(5)\n\n print(\"Top 5 cities based on number of events\")\n _df.groupBy(\"city\").count().orderBy(\"count\", ascending=False) \\\n .show(5)\n\n print(\"Top 5 Browsers based on number of unique users\")\n\n _df.groupBy(\"browser\").agg(countDistinct(\"user_id\")) \\\n .orderBy(\"count(DISTINCT user_id)\", ascending=False) \\\n .show(5)\n\n print(\"Top 5 Operating systems based on number of unique users\")\n _df.groupBy(\"os\").agg(countDistinct(\"user_id\")) \\\n .orderBy(\"count(DISTINCT user_id)\", ascending=False) \\\n .show(5)\n\n\ndef transform(_df, _spark):\n \"\"\"This function handles the ransformation of the dataset (biggest part)\n\n Args:\n _df: Initial, unhandled dataframe straight from extraction\n _spark: sparksession\n\n Returns: Final and structured dataframe\n\n \"\"\"\n print(\"Spark Dataframe is being cleaned....\")\n\n print(\"date and time column is becomming one timestamp...\")\n newcolname = \"timestamp\"\n oldcolnames = [\"date\", \"time\"]\n separator = \" \"\n _df = concatColName(_df, newcolname, oldcolnames, separator)\n _df = dropColumns(_df, oldcolnames)\n _df = castToTimeStamp(_df, newcolname)\n\n print(\"The user_agent_string is becomming os and browser...\")\n # Getting the browser and OS from user_agent_string (amazingly fast! wow!)\n newcolname = \"getOsBrowser\"\n oldcolname = \"user_agent_string\"\n _df = udfCreator(_df, getOsBrowser, newcolname, oldcolname, StringType())\n\n # Cleaning Os Browser result\n print(\"Cleaning OS, Browser result\")\n separator = \"-\"\n newcolnames = [\"getOsBrowser\", \"os\", \"browser\"]\n dropcols = [\"getOsBrowser\", \"user_agent_string\"]\n _df = splitCol(_df, separator, newcolnames)\n _df = dropColumns(_df, dropcols)\n\n # Cleaning IP addresses\n print(\"Cleaning IP addresses\")\n separator = \",\"\n newcolnames = [\"ip\", \"ip1\", \"ip2\"]\n dropcols = ['ip']\n _df = splitCol(_df, separator, newcolnames)\n _df = dropColumns(_df, dropcols)\n\n # Adding eventID to the dataframe, so that we can join other dataframes\n newcolname = \"eventID\"\n _df = addMonoticallyIncreasingID(_df, newcolname)\n\n print(\"Converting IP adress to city and country... \")\n newcolname = \"getCityAndCountry\"\n oldcolname = \"ip1\"\n _df = udfCreator(_df, getCityAndCountry, newcolname, oldcolname, StringType())\n\n # Modify ip dataframe for countries and cities of the first ip column\n separator = \"-\"\n newcolnames = [\"getCityAndCountry\", \"country\", \"city\"]\n dropcols = [\"getCityAndCountry\", \"eventID\", \"timestamp\", \"user_id\", \"url\", \"os\", \"browser\", \"ip2\"]\n newcolname = \"ip\"\n oldcolname = \"ip1\"\n _ip = splitCol(_df, separator, newcolnames)\n _ip = dropColumns(_ip, dropcols)\n _ip = renameColumn(_ip, oldcolname, newcolname)\n\n # create a monotonically increasing id\n newcolname = \"id\"\n _ip = addMonoticallyIncreasingID(_ip, newcolname)\n\n # Merge countries and cities to org Spark Dataframes\n newcolnames = [\"eventID\", \"timestamp\", \"user_id\", \"url\", \"os\", \"browser\", \"country\", \"city\", \"ip\"]\n orderbycols = [\"eventID\", \"id\"]\n ret_df = mergeDataframes(\n _df,\n _ip,\n newcolnames,\n orderbycols,\n )\n\n return ret_df\n\n\ndef extract(_spark):\n \"\"\"Extracting the tsv file into a DataFrame\n\n Args:\n _spark: The actual spark session\n\n Returns: Initial dataframe before transform\n\n \"\"\"\n cwd = os.getcwd()\n # Initial read of the given TSV file\n _df = _spark.read.option(\"delimiter\", \"\\t\")\\\n .csv(cwd + \"/src/data/input_data\")\\\n .toDF(\"date\", \"time\", \"user_id\", \"url\", \"ip\", \"user_agent_string\")\n\n _spark.sparkContext.setLogLevel(\"WARN\")\n\n return _df\n\n\ndef main(_spark):\n # Perform extraction\n print(\"Perform extraction\")\n df = extract(spark)\n\n # Perform transformation\n print(\"Perform transformation\")\n df = transform(df, spark)\n\n print(\"Printing Transformed Dataframe Schema\")\n df.printSchema()\n\n # Load the data, do some printing, make it searchable for the API, maybe a postgres\n print(\"Perform load\")\n load(df)\n\n print(\"Spark application ends\")\n\n # Stop spark application\n spark.stop()\n\n\nif __name__ == \"__main__\":\n # Initial setup of spark project\n warehouse_location = abspath('spark-warehouse')\n spark = SparkSession \\\n .builder \\\n .appName(\"CitiesCountriesTest\") \\\n .config(\"spark.sql.warehouse.dir\", warehouse_location) \\\n .enableHiveSupport() \\\n .getOrCreate()\n\n main(spark)\n"}],"string":"[\n {\n \"alpha_fraction\": 0.6452810168266296,\n \"alphanum_fraction\": 0.6930010318756104,\n \"avg_line_length\": 29.419355392456055,\n \"blob_id\": \"5fb256f9581199a758db732b0ffe6192d2cb8029\",\n \"content_id\": \"8c04384f6998befc03967637e97af16c61f459eb\",\n \"detected_licenses\": [\n \"Apache-2.0\",\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Dockerfile\",\n \"length_bytes\": 1886,\n \"license_type\": \"permissive\",\n \"max_line_length\": 113,\n \"num_lines\": 62,\n \"path\": \"/Dockerfile\",\n \"repo_name\": \"Thelin90/G.D.C.K\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"\\nFROM ubuntu:latest\\n\\n# PATH\\nENV PATH /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin\\n# Spark\\nENV SPARK_VERSION 2.3.1\\nENV SPARK_HOME /usr/local/spark\\nENV SPARK_LOG_DIR /var/log/spark\\nENV SPARK_PID_DIR /var/run/spark\\nENV PATH $PATH:$SPARK_HOME/bin\\nENV PYSPARK_PYTHON /usr/bin/python3.6\\nENV PYSPARK_DRIVER_PYTHON=/usr/bin/python3.6\\nENV PYTHONUNBUFFERED 1\\n# Java\\nENV JAVA_HOME /usr/lib/jvm/java-8-openjdk-amd64/\\n# Python\\nENV alias python=/usr/bin/python3.6\\nENV PYTHONPATH /etc/app/src/\\n\\n# Install curl\\nRUN apt-get update && apt-get install -y curl\\n\\n# Install Python 3.6, 2.7 is standard to ubuntu:latest\\nRUN apt-get update && \\\\\\n apt-get install -y python3.6 && \\\\\\n apt-get install -y python3-pip\\n\\n# Install Java, had an issue found this: https://stackoverflow.com/questions/46795907/setting-java-home-in-docker\\nRUN apt-get update && apt-get install -y openjdk-8-jdk && \\\\\\n apt-get install -y ant && apt-get clean && \\\\\\n rm -rf /var/lib/apt/lists/ && \\\\\\n rm -rf /var/cache/oracle-jdk8-installer;\\n\\n# Set workspace\\nWORKDIR /etc/app\\n\\n# Add all the project files to the\\nADD . /etc/app\\n\\n# Download Spark\\nRUN curl -L http://www.us.apache.org/dist/spark/spark-${SPARK_VERSION}/spark-${SPARK_VERSION}-bin-hadoop2.7.tgz \\\\\\n | tar -xzp -C /usr/local/ && \\\\\\n ln -s spark-${SPARK_VERSION}-bin-hadoop2.7 ${SPARK_HOME}\\n\\n# Make run.sh executable\\nRUN chmod +x /etc/app/scripts/run.sh && chmod +x /etc/app/src/data/input_data\\n\\n# Give give -rw-r--r-- to python files\\nRUN chmod 0644 /etc/app/src/etl.py && \\\\\\n chmod 0644 /etc/app/src/geoip.py && \\\\\\n chmod 0644 /etc/app/src/data/input_data\\n\\nEXPOSE 8080 8081 6066 \\\\\\n 7077 4040 7001 \\\\\\n 7002 7003 7004 \\\\\\n 7005 7006\\n\\nENTRYPOINT [\\\"./scripts/run.sh\\\"]\\n\\n# Replace the Entrypoint running the run.sh with this\\n# to keep the container alive, tobe able to debug the container\\n# ENTRYPOINT [\\\"tail\\\", \\\"-f\\\", \\\"/dev/null\\\"]\"\n },\n {\n \"alpha_fraction\": 0.6243016719818115,\n \"alphanum_fraction\": 0.6298882961273193,\n \"avg_line_length\": 24.571428298950195,\n \"blob_id\": \"fbc9ac73f2a3a05852f4d095dbb3677678f7e00f\",\n \"content_id\": \"830a13247f6aa8446bdfdd8d31e0cd805aaa47e5\",\n \"detected_licenses\": [\n \"Apache-2.0\",\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 716,\n \"license_type\": \"permissive\",\n \"max_line_length\": 60,\n \"num_lines\": 28,\n \"path\": \"/test/fakesparksession.py\",\n \"repo_name\": \"Thelin90/G.D.C.K\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"import logging\\nfrom pyspark.sql import SparkSession\\n\\nimport unittest\\n\\nclass PySparkTest(unittest.TestCase):\\n\\n @classmethod\\n def suppress_py4j_logging(cls):\\n logger = logging.getLogger(\\\"py4j\\\")\\n logger.setLevel(logging.WARN)\\n\\n @classmethod\\n def create_testing_pyspark_session(cls):\\n return (SparkSession.builder\\n .master(\\\"local[2]\\\")\\n .appName(\\\"my-local-testing-pyspark-context\\\")\\n .enableHiveSupport()\\n .getOrCreate())\\n\\n @classmethod\\n def setUpClass(cls):\\n cls.suppress_py4j_logging()\\n cls.spark = cls.create_testing_pyspark_session()\\n\\n @classmethod\\n def tearDownClass(cls):\\n cls.spark.stop()\\n\"\n },\n {\n \"alpha_fraction\": 0.6188467144966125,\n \"alphanum_fraction\": 0.6244725584983826,\n \"avg_line_length\": 26.346153259277344,\n \"blob_id\": \"fe50ca6fa3975efbc46fe872e8a4f478f363f661\",\n \"content_id\": \"2990b0de72ecb9ee2f0e7c79c8862f54f8331071\",\n \"detected_licenses\": [\n \"Apache-2.0\",\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 1422,\n \"license_type\": \"permissive\",\n \"max_line_length\": 109,\n \"num_lines\": 52,\n \"path\": \"/src/geoip.py\",\n \"repo_name\": \"Thelin90/G.D.C.K\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n# coding=utf-8\\n\\n\\\"\\\"\\\"\\nPython script to transform ip to country and city, has been modified\\n\\\"\\\"\\\"\\n\\nimport socket\\nimport pygeoip\\nimport os\\n\\n\\\"\\\"\\\" http://geolite.maxmind.com/download/geoip/database/GeoLiteCity.dat.gz \\\"\\\"\\\"\\ncwd = os.getcwd()\\nrawdata = pygeoip.GeoIP(cwd + \\\"/src/geoconvertdat/GeoLiteCity.dat\\\")\\n\\n\\ndef ipquery(ip):\\n \\\"\\\"\\\"Function to parse IP to country, city\\n http://www.linuxx.eu/2014/05/geolocate-ip-with-python.html\\n Have been slightly modified\\n\\n In Python 3, all strings are sequences of Unicode characters. There is a bytes type that holds raw bytes.\\n\\n Args:\\n ip: The actual IP used to parse the country and city from\\n\\n Returns: A str value of \\\"country-city\\\"\\n\\n \\\"\\\"\\\"\\n\\n try:\\n socket.inet_aton(str(ip))\\n data = rawdata.record_by_name(ip)\\n\\n if type(data) is dict:\\n\\n country = type(data['country_name']) is str and data['country_name'] or type(\\n data['country_name']) and data['country_name'] or \\\"NotTraceable\\\"\\n\\n city = type(data['city']) and data['city'] or type(\\n data['city']) is str and data['city'] or \\\"NotTraceable\\\"\\n\\n if isinstance(country, bytes):\\n country = country.decode()\\n\\n if isinstance(city, bytes):\\n city = city.decode()\\n\\n return country + \\\"-\\\" + city\\n\\n except socket.error:\\n return \\\"NotTraceable-NotTraceable\\\"\\n\"\n },\n {\n \"alpha_fraction\": 0.6662476658821106,\n \"alphanum_fraction\": 0.6900062561035156,\n \"avg_line_length\": 30.947792053222656,\n \"blob_id\": \"cbaf446cbe48dcdca50cb6d0cafdbe5f6f885dd4\",\n \"content_id\": \"a92ebbdd08c0de53aebac4b300408dca466c3be5\",\n \"detected_licenses\": [\n \"Apache-2.0\",\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 7955,\n \"license_type\": \"permissive\",\n \"max_line_length\": 567,\n \"num_lines\": 249,\n \"path\": \"/README.md\",\n \"repo_name\": \"Thelin90/G.D.C.K\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Overview\\n\\n## Trello Feature Board\\n\\n* [Board](https://trello.com/b/qigu0NnW/gdck)\\n\\n\\n\\nGarden of Data Creation Kit `(An edit to the famous G.E.C.K in the Bethesda Fallout game series`) is a project that processes large amounts of data, currently TSV files and performs extraction, transformation and loading (ETL).\\n\\n## Project Design\\n\\nThe design of this project is that it should be fast and easy to deploy and run. Below is a description of how the application should be deployment ready.\\n\\n\\n\\nMicroservices is an architectural style that structures an application as a collection of loosely coupled services. Therefore enrich business capabilities, such as continuous delivery/deployment of scale.\\n\\n## Data Processing\\n\\nApache Spark is the chosen tool used within this project. Spark is quick and very responsive tool to perform data processing with. It provides an analytic engine for large-scale data processing. It is a general distributed in-memory computing framework implemented in scala. Hence spark operates on distributed data collections. However it does not do distributed storage. Spark is proven to be much faster than the popular Apache Hadoop framework. Apache Spark and Hadoop have different purposes but works good together. A high level overview of Apache Spark below: \\n\\n\\n\\nHadoop is essentially a distributed data infrastructure. It provides distributes big data collections across multiple nodes within a cluster of commodity servers. A high level overview of Apache Hadoop below: \\n\\n\\n\\n\\n## PySpark\\n\\nPySpark have some core concepts. Such as resilient distributed datasets and dataframes.\\n\\nThis project spefically take advantage of the dataframes.\\n\\nIt has derived from the resilient distributed datasets concept. Dataframes will provide a higher level of abstraction. Which implicitly allow a query language way of typing to manipulate data. The abstraction represents data and a schema. Which enable better interaction with front end applications.\\n\\n## Dataset\\nThe dataset has the (default) values divided by tab in the given dataset provided (input data): field header (date, time, user_id, url, IP, user_agent_string). See attached source file input_data.\\n\\n## Setup\\n\\nClone directly from the source code:\\n\\n```bash\\ngit clone https://github.com/Thelin90/G.D.C.K.git\\n```\\n\\nDocker needs to be installed on the machine. It can also run locally without docker.\\n\\n### Requirements\\n\\n* Docker environment\\n* Python ^3.6.5\\n* Java ^8\\n* Spark ^2.3.x `(mininum 2.3.0)`\\n\\nAssuming that Python, Docker and Java is already setup.\\n\\n### Setup Apache-Spark\\n\\nStart with downloading Spark (note that depending on your IDE, you need to specify your Spark location):\\n\\n- https://spark.apache.org/downloads.html\\n\\nSet your SPARK_HOME in `.bashrc`\\n```bash\\nSPARK_HOME='path-to-spark'\\n```\\n\\nThen source the file\\n\\n```bash\\nsource ~/.bashrc\\n```\\n\\n#### PyCharm Example\\n\\nMark as source:\\n\\n```bash\\nFile -> Settings -> Project Structure -> add root content '+' -> 'path-to-spark'\\n```\\n\\nSet environment variables:\\n\\n```bash\\nRun -> Edit Configuration -> Environment Variables -> add new environment variables\\n```\\n```bash\\nNAME\\t\\t\\t\\tVALUE\\nPYSPARK_PYTHON\\t\\t\\t'path-to-python'\\nPYSPARK_DRIVER_PYTHON\\t\\t'path-to-python'\\n```\\n\\n### Testing\\n\\nRun the following command to run tests:\\n\\n```bash\\npython3 -W ignore:ResourceWarning -m unittest test/unittest-etl-process.py\\n```\\n\\n### Manual Run\\n\\nRemember to set PYTHONPATH in `~/.bashrc`\\n\\n```bash\\nPYTHONPATH=path-to-proj/PySparkDocker/src/\\n```\\n\\nThen `source ~/.bashrc`\\n\\n\\nRun the script manually without using docker.\\n\\n```bash\\nspark-submit src/etl.py\\n```\\n\\n### Run ETL Docker\\n\\n```bash\\ndocker build -t etl-cities-countries .\\ndocker run etl-cities-countries\\n```\\n\\n## Result\\n\\n* Top 5 Countries based on number of events\\n* Top 5 Cities based on number of events \\n* Top 5 Browsers based on number of unique users\\n* Top 5 Operating systems based on number of unique users.\\n\\n```bash\\ndate and time column is becomming one timestamp...\\nThe user_agent_string is becomming os and browser...\\nConverting IP adress to city and country... \\nPrinting Transformed Dataframe Schema\\n |-- eventID: long (nullable = false)\\n |-- timestamp: timestamp (nullable = true)\\n |-- user_id: string (nullable = true)\\n |-- url: string (nullable = true)\\n |-- os: string (nullable = true)\\n |-- browser: string (nullable = true)\\n |-- country: string (nullable = true)\\n |-- city: string (nullable = true)\\n\\nPerform load\\nTop 5 countries based on number of events\\n+--------------+------+\\n| country| count|\\n+--------------+------+\\n|United Kingdom|135831|\\n| Ireland| 18522|\\n| Sweden| 12143|\\n| Norway| 6908|\\n| Netherlands| 2131|\\n+--------------+------+\\nonly showing top 5 rows\\n\\nTop 5 cities based on number of events\\n+------------+-----+\\n| city|count|\\n+------------+-----+\\n|NotTraceable|15336|\\n| Dublin| 5750|\\n| London| 4827|\\n| Manchester| 3736|\\n| Bristol| 2801|\\n+------------+-----+\\nonly showing top 5 rows\\n\\nTop 5 Browsers based on number of unique users\\n+--------------------+-----------------------+\\n| browser|count(DISTINCT user_id)|\\n+--------------------+-----------------------+\\n| Safari 7.0| 27117|\\n| Safari 8.0| 17941|\\n|Chrome 37.0.2062.124| 5054|\\n| Safari 7.0.6| 2804|\\n| Safari 7.1| 2486|\\n+--------------------+-----------------------+\\nonly showing top 5 rows\\n\\nTop 5 Operating systems based on number of unique users\\n+----------------+-----------------------+\\n| os|count(DISTINCT user_id)|\\n+----------------+-----------------------+\\n| IPad iOS 8.0.2| 11126|\\n| IPad iOS 7.1.2| 10786|\\n| Windows 7| 6722|\\n|iPhone iOS 8.0.2| 4651|\\n|iPhone iOS 7.1.2| 4496|\\n+----------------+-----------------------+\\nonly showing top 5 rows\\n\\nSpark application ends\\n```\\n\\n### Spark UI\\n\\nTo see the spark jobs process do the following while running the docker image:\\n\\n```bash\\ndocker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' etl-cities-countries\\n```\\n\\nThis will give the IP of the docker container, type following in the web browser:\\n\\n`container-ip:4040`\\n\\nThis will give the following output:\\n\\n\\n\\n## License\\n\\nMIT License\\n\\nCopyright (c) [2018] [Simon Thelin]\\n\\nPermission is hereby granted, free of charge, to any person obtaining a copy\\nof this software and associated documentation files (the \\\"Software\\\"), to deal\\nin the Software without restriction, including without limitation the rights\\nto use, copy, modify, merge, publish, distribute, sublicense, and/or sell\\ncopies of the Software, and to permit persons to whom the Software is\\nfurnished to do so, subject to the following conditions:\\n\\nThe above copyright notice and this permission notice shall be included in all\\ncopies or substantial portions of the Software.\\n\\nTHE SOFTWARE IS PROVIDED \\\"AS IS\\\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\\nIMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\\nFITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\\nAUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\\nLIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\\nOUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\\nSOFTWARE.\\n\\n[Back To The Top](#read-me-template)\\n\\n---\\n\\n## Author Info\\n\\n* [Simon Thelin](https://www.linkedin.com/in/simon-thelin-3ba96986/)\\n\\n[Back To The Top](#read-me-template)\\n\"\n },\n {\n \"alpha_fraction\": 0.752010703086853,\n \"alphanum_fraction\": 0.752010703086853,\n \"avg_line_length\": 31.434782028198242,\n \"blob_id\": \"27b855b3bb9675ebc846937a224d9e3a18fd83a7\",\n \"content_id\": \"b54dcb1d8491cdd045112f839504a41df3eecf8b\",\n \"detected_licenses\": [\n \"Apache-2.0\",\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Markdown\",\n \"length_bytes\": 746,\n \"license_type\": \"permissive\",\n \"max_line_length\": 151,\n \"num_lines\": 23,\n \"path\": \"/docs/PULL_REQUEST_TEMPLATE.md\",\n \"repo_name\": \"Thelin90/G.D.C.K\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# Pull Request Template\\n\\nPlease read this template to ensure your pull request adheres to the given guidelines.\\n\\n## Guidelines\\n- Use the following format for header: `TD/xxxx/G.D.C.K-text-what-has-been-done` (the repo owner will fill in the xxxx to fit the actual ticket number)\\n- Link additions should be added to the bottom of the relevant category.\\n- New features or improvements to the existing features are welcome.\\n\\n`NOTE! ACHTUNG!`\\n- Search previous suggestions before making a new one, as yours may be a duplicate.\\n- Do not use `#` for commenting the PR commit message, only use `-`.\\n\\nGiven example:\\n\\n```\\nTD-xxxx/G.D.C.K-some-fixes-to-something\\n\\n- New feature A\\n - additional link\\n- Improvement to feature B\\n```\\nThanks for contributing!\\n\"\n },\n {\n \"alpha_fraction\": 0.7404580116271973,\n \"alphanum_fraction\": 0.7557252049446106,\n \"avg_line_length\": 25.399999618530273,\n \"blob_id\": \"0564913906bfe9bbe39a4b0f0d74e9bf5a68a34e\",\n \"content_id\": \"66cf32f9fd237e9e0b70d16627505d966992ad72\",\n \"detected_licenses\": [\n \"Apache-2.0\",\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 131,\n \"license_type\": \"permissive\",\n \"max_line_length\": 54,\n \"num_lines\": 5,\n \"path\": \"/test/__init__.py\",\n \"repo_name\": \"Thelin90/G.D.C.K\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"# For relative imports to work in Python 3.6\\nfrom os import path\\nimport sys\\n\\nsys.path.append(path.dirname(path.realpath(__file__)))\"\n },\n {\n \"alpha_fraction\": 0.4475524425506592,\n \"alphanum_fraction\": 0.6853147149085999,\n \"avg_line_length\": 14.88888931274414,\n \"blob_id\": \"15252489b4854128f28486d0498c5e67c5ad7a97\",\n \"content_id\": \"5a051799a5149e6329ec8141c1ecf4a3e8f96c93\",\n \"detected_licenses\": [\n \"Apache-2.0\",\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Text\",\n \"length_bytes\": 143,\n \"license_type\": \"permissive\",\n \"max_line_length\": 22,\n \"num_lines\": 9,\n \"path\": \"/scripts/requirements.txt\",\n \"repo_name\": \"Thelin90/G.D.C.K\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"httpagentparser==1.8.1\\nnumpy==1.15.0\\npandas==0.23.4\\npy4j==0.10.7\\npygeoip==0.3.2\\npyspark==2.3.2\\npython-dateutil==2.7.3\\npytz==2018.4\\nsix==1.11.0\\n\"\n },\n {\n \"alpha_fraction\": 0.6843373775482178,\n \"alphanum_fraction\": 0.6867470145225525,\n \"avg_line_length\": 30.923076629638672,\n \"blob_id\": \"506e9683b54f192e1920cb3cc55bde603338da42\",\n \"content_id\": \"001abec6b008146ba3b8da8c113370323d9ade9b\",\n \"detected_licenses\": [\n \"Apache-2.0\",\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 415,\n \"license_type\": \"permissive\",\n \"max_line_length\": 114,\n \"num_lines\": 13,\n \"path\": \"/test/unittest-etl-process.py\",\n \"repo_name\": \"Thelin90/G.D.C.K\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python3\\nfrom .fakesparksession import PySparkTest\\nfrom etl import extract\\n\\n\\nclass SimpleTest(PySparkTest):\\n \\\"\\\"\\\"\\n Will perform test on the functions used within the ETL process with a fake spark session\\n \\\"\\\"\\\"\\n def test_basic(self):\\n expected_df = extract(self.spark)\\n\\n self.assertEquals(expected_df.schema.names, [\\\"date\\\", \\\"time\\\", \\\"user_id\\\", \\\"url\\\", \\\"ip\\\", \\\"user_agent_string\\\"])\\n\"\n },\n {\n \"alpha_fraction\": 0.7422680258750916,\n \"alphanum_fraction\": 0.7628865838050842,\n \"avg_line_length\": 31.33333396911621,\n \"blob_id\": \"1dfe90110602e682728026fb086f76f4d9bd628c\",\n \"content_id\": \"c88a7019a69e12ee4353d654b01372f8f7354c8c\",\n \"detected_licenses\": [\n \"Apache-2.0\",\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Shell\",\n \"length_bytes\": 97,\n \"license_type\": \"permissive\",\n \"max_line_length\": 52,\n \"num_lines\": 3,\n \"path\": \"/scripts/run.sh\",\n \"repo_name\": \"Thelin90/G.D.C.K\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env bash\\npython3.6 -m pip install -r scripts/requirements.txt\\nspark-submit src/etl.py\\n\"\n },\n {\n \"alpha_fraction\": 0.645996630191803,\n \"alphanum_fraction\": 0.6500644683837891,\n \"avg_line_length\": 28.214492797851562,\n \"blob_id\": \"dd9e91c9780ba5166792bdb4c6beeee7a19e508d\",\n \"content_id\": \"fa967ba6ab9f466180def161e33d3eaee34b6411\",\n \"detected_licenses\": [\n \"Apache-2.0\",\n \"MIT\"\n ],\n \"is_generated\": false,\n \"is_vendor\": false,\n \"language\": \"Python\",\n \"length_bytes\": 10079,\n \"license_type\": \"permissive\",\n \"max_line_length\": 116,\n \"num_lines\": 345,\n \"path\": \"/src/etl.py\",\n \"repo_name\": \"Thelin90/G.D.C.K\",\n \"src_encoding\": \"UTF-8\",\n \"text\": \"#!/usr/bin/env python\\n\\n\\\"\\\"\\\"\\nPython script to initiate and perform the ETL process\\n\\\"\\\"\\\"\\n\\nimport os\\nimport pyspark\\nimport httpagentparser\\nfrom pyspark.sql.functions import struct\\nfrom pyspark.sql.functions import *\\nfrom os.path import abspath\\nfrom pyspark.sql import SparkSession\\nfrom pyspark.sql.types import TimestampType, StringType\\nfrom geoip import ipquery\\n\\n\\ndef addMonoticallyIncreasingID(_df, _colname):\\n \\\"\\\"\\\"Adds monotically increasing ID column to Spark Dataframe\\n\\n Args:\\n _df: Spark Dataframe\\n _colname: Name of the col containing the monotoically increasing ID\\n\\n Returns: Spark Dataframe with a column containing a monotoically increasing ID\\n \\\"\\\"\\\"\\n if isinstance(_colname, str):\\n return _df.withColumn(_colname, monotonically_increasing_id())\\n else:\\n raise ValueError\\n\\n\\ndef castToTimeStamp(_df, colname):\\n \\\"\\\"\\\" Function casts column values to TimestampType\\n Args:\\n _df: Spark Dataframe\\n colname: Name of column\\n\\n Returns\\n \\\"\\\"\\\"\\n coldatatype = _df.schema[colname].dataType\\n if isinstance(coldatatype, TimestampType):\\n raise ValueError\\n else:\\n print(\\\"Column \\\" + colname + \\\" with datatype \\\" + str(coldatatype) + \\\" typecasting to TimestampType\\\")\\n return _df.withColumn(colname, _df[colname].cast(TimestampType()))\\n\\n\\ndef renameColumn(_df, _oldcolname, _newcolname):\\n \\\"\\\"\\\"Rename a column name\\n Args:\\n _df:\\n _oldcolname:\\n _newcolname:\\n\\n Returns:\\n \\\"\\\"\\\"\\n if isinstance(_oldcolname, str) and isinstance(_newcolname, str):\\n return _df.withColumnRenamed(_oldcolname, _newcolname)\\n else:\\n raise ValueError(\\\"Column names must be of typ str\\\")\\n\\n\\ndef concatColName(_df, _newcol, _oldcols, separator):\\n \\\"\\\"\\\"\\n Args:\\n _df: Spark Dataframe\\n newcol: Name of the new column\\n oldcols: The names of the old columns\\n separator: The separator value, examples: `;` `-` `.` `,`\\n\\n Returns\\n \\\"\\\"\\\"\\n if isinstance(_newcol, str) and isinstance(_oldcols, list):\\n return _df.withColumn(_newcol, concat_ws(separator, _df[_oldcols[0]], _df[_oldcols[1]]))\\n else:\\n raise ValueError\\n\\n\\ndef splitCol(_df, _split, _colnames):\\n \\\"\\\"\\\"Creates 3 new columns from 1 original column\\n\\n TODO: N new columns from 1 original column\\n\\n Args:\\n _dataframe: The dataframe that needs to split columns\\n _split: The value to split on, ex: \\\"-\\\", \\\",\\\", \\\"*\\\"\\n _colnames: The new names for the new columns\\n\\n Returns:\\n\\n \\\"\\\"\\\"\\n split_col = pyspark.sql.functions.split(_df[_colnames[0]], _split)\\n ret_df = _df.withColumn(_colnames[1], split_col.getItem(0))\\n return ret_df.withColumn(_colnames[2], split_col.getItem(1))\\n\\n\\ndef dropColumns(_df, colnames):\\n \\\"\\\"\\\"Drops N number of undesired columns\\n\\n Args:\\n _df: Spark Dataframe\\n columnames: List containing column names to be dropped\\n\\n Returns: Spark Dataframe\\n\\n \\\"\\\"\\\"\\n if isinstance(colnames, list):\\n return _df.drop(*colnames)\\n else:\\n raise ValueError(\\\"colnames is not a list\\\")\\n\\n\\ndef mergeDataframes(_df0, _df1, _colnames, _orderbycols):\\n \\\"\\\"\\\"Merges two Spark Dataframes to one\\n\\n Args:\\n _df0: Spark Dataframe 0\\n _df1: Spark Dataframe 1 to be joined with Spark Datafram 0\\n colnames: list of column names to be set\\n orderbycol: list containing id value to join by\\n\\n Returns:\\n\\n \\\"\\\"\\\"\\n print(\\\"Merge Spark Dataframes...\\\")\\n ret_df = _df0.join(_df1, _df0[_orderbycols[0]] == _df1[_orderbycols[1]])\\n ret_df = ret_df.orderBy(_orderbycols[0], ascending=True)\\n ret_df = ret_df.select(*_colnames)\\n\\n return ret_df\\n\\n\\ndef udfCreator(_df, _functioncall, _newcolname, _oldcolname, datatype):\\n \\\"\\\"\\\"\\n Args:\\n _df: Spark Dataframe\\n _functioncall: The function to be called\\n _newcolname: The name of the new column containing the extracted values\\n _oldcolname: The column values from a specific column to be sent, example: ip addresses\\n datatype: Which datatype to work with, example: StringType()\\n\\n Returns: Spark Dataframe with one new additional column containing values from the UDF call\\n \\\"\\\"\\\"\\n if isinstance(_newcolname, str) and isinstance(_oldcolname, str):\\n print(\\\"Making UDF call to function: \\\" + str(_functioncall))\\n _udf = udf(_functioncall, datatype)\\n return _df.withColumn(_newcolname, _udf(_df[_oldcolname]))\\n else:\\n raise Exception(\\\"Column names must be of str\\\")\\n\\n\\ndef getOsBrowser(value):\\n \\\"\\\"\\\"Calls the httpagentparser and retrieves the os and browser information\\n\\n Args:\\n value: Each column value of user_agent_string\\n\\n Returns: The browser and os as a string\\n\\n \\\"\\\"\\\"\\n if isinstance(value, str):\\n return str(httpagentparser.simple_detect(value)[0] + \\\"-\\\" + httpagentparser.simple_detect(value)[1])\\n else:\\n raise ValueError\\n\\n\\ndef getCityAndCountry(_ip):\\n \\\"\\\"\\\"Function to make call to ipquery, may seem redundant but it makes the code more readable when doing UDF calls\\n\\n Args:\\n _ip: IP address value\\n\\n Returns: A value of country-city from the IP\\n \\\"\\\"\\\"\\n return ipquery(_ip)\\n\\n\\ndef load(_df):\\n \\\"\\\"\\\"Load function to print the result and to save the dataframe for api calls\\n\\n TODO: Save dataframe/dataframes to postgres\\n\\n Args:\\n _df: The final dataframe\\n\\n Returns: Nothing\\n\\n \\\"\\\"\\\"\\n # Peform load process\\n\\n print(\\\"Displaying the top 5 rows\\\")\\n _df.show(5)\\n\\n print(\\\"Top 5 countries based on number of events\\\")\\n _df.groupBy(\\\"country\\\").count().orderBy(\\\"count\\\", ascending=False) \\\\\\n .show(5)\\n\\n print(\\\"Top 5 cities based on number of events\\\")\\n _df.groupBy(\\\"city\\\").count().orderBy(\\\"count\\\", ascending=False) \\\\\\n .show(5)\\n\\n print(\\\"Top 5 Browsers based on number of unique users\\\")\\n\\n _df.groupBy(\\\"browser\\\").agg(countDistinct(\\\"user_id\\\")) \\\\\\n .orderBy(\\\"count(DISTINCT user_id)\\\", ascending=False) \\\\\\n .show(5)\\n\\n print(\\\"Top 5 Operating systems based on number of unique users\\\")\\n _df.groupBy(\\\"os\\\").agg(countDistinct(\\\"user_id\\\")) \\\\\\n .orderBy(\\\"count(DISTINCT user_id)\\\", ascending=False) \\\\\\n .show(5)\\n\\n\\ndef transform(_df, _spark):\\n \\\"\\\"\\\"This function handles the ransformation of the dataset (biggest part)\\n\\n Args:\\n _df: Initial, unhandled dataframe straight from extraction\\n _spark: sparksession\\n\\n Returns: Final and structured dataframe\\n\\n \\\"\\\"\\\"\\n print(\\\"Spark Dataframe is being cleaned....\\\")\\n\\n print(\\\"date and time column is becomming one timestamp...\\\")\\n newcolname = \\\"timestamp\\\"\\n oldcolnames = [\\\"date\\\", \\\"time\\\"]\\n separator = \\\" \\\"\\n _df = concatColName(_df, newcolname, oldcolnames, separator)\\n _df = dropColumns(_df, oldcolnames)\\n _df = castToTimeStamp(_df, newcolname)\\n\\n print(\\\"The user_agent_string is becomming os and browser...\\\")\\n # Getting the browser and OS from user_agent_string (amazingly fast! wow!)\\n newcolname = \\\"getOsBrowser\\\"\\n oldcolname = \\\"user_agent_string\\\"\\n _df = udfCreator(_df, getOsBrowser, newcolname, oldcolname, StringType())\\n\\n # Cleaning Os Browser result\\n print(\\\"Cleaning OS, Browser result\\\")\\n separator = \\\"-\\\"\\n newcolnames = [\\\"getOsBrowser\\\", \\\"os\\\", \\\"browser\\\"]\\n dropcols = [\\\"getOsBrowser\\\", \\\"user_agent_string\\\"]\\n _df = splitCol(_df, separator, newcolnames)\\n _df = dropColumns(_df, dropcols)\\n\\n # Cleaning IP addresses\\n print(\\\"Cleaning IP addresses\\\")\\n separator = \\\",\\\"\\n newcolnames = [\\\"ip\\\", \\\"ip1\\\", \\\"ip2\\\"]\\n dropcols = ['ip']\\n _df = splitCol(_df, separator, newcolnames)\\n _df = dropColumns(_df, dropcols)\\n\\n # Adding eventID to the dataframe, so that we can join other dataframes\\n newcolname = \\\"eventID\\\"\\n _df = addMonoticallyIncreasingID(_df, newcolname)\\n\\n print(\\\"Converting IP adress to city and country... \\\")\\n newcolname = \\\"getCityAndCountry\\\"\\n oldcolname = \\\"ip1\\\"\\n _df = udfCreator(_df, getCityAndCountry, newcolname, oldcolname, StringType())\\n\\n # Modify ip dataframe for countries and cities of the first ip column\\n separator = \\\"-\\\"\\n newcolnames = [\\\"getCityAndCountry\\\", \\\"country\\\", \\\"city\\\"]\\n dropcols = [\\\"getCityAndCountry\\\", \\\"eventID\\\", \\\"timestamp\\\", \\\"user_id\\\", \\\"url\\\", \\\"os\\\", \\\"browser\\\", \\\"ip2\\\"]\\n newcolname = \\\"ip\\\"\\n oldcolname = \\\"ip1\\\"\\n _ip = splitCol(_df, separator, newcolnames)\\n _ip = dropColumns(_ip, dropcols)\\n _ip = renameColumn(_ip, oldcolname, newcolname)\\n\\n # create a monotonically increasing id\\n newcolname = \\\"id\\\"\\n _ip = addMonoticallyIncreasingID(_ip, newcolname)\\n\\n # Merge countries and cities to org Spark Dataframes\\n newcolnames = [\\\"eventID\\\", \\\"timestamp\\\", \\\"user_id\\\", \\\"url\\\", \\\"os\\\", \\\"browser\\\", \\\"country\\\", \\\"city\\\", \\\"ip\\\"]\\n orderbycols = [\\\"eventID\\\", \\\"id\\\"]\\n ret_df = mergeDataframes(\\n _df,\\n _ip,\\n newcolnames,\\n orderbycols,\\n )\\n\\n return ret_df\\n\\n\\ndef extract(_spark):\\n \\\"\\\"\\\"Extracting the tsv file into a DataFrame\\n\\n Args:\\n _spark: The actual spark session\\n\\n Returns: Initial dataframe before transform\\n\\n \\\"\\\"\\\"\\n cwd = os.getcwd()\\n # Initial read of the given TSV file\\n _df = _spark.read.option(\\\"delimiter\\\", \\\"\\\\t\\\")\\\\\\n .csv(cwd + \\\"/src/data/input_data\\\")\\\\\\n .toDF(\\\"date\\\", \\\"time\\\", \\\"user_id\\\", \\\"url\\\", \\\"ip\\\", \\\"user_agent_string\\\")\\n\\n _spark.sparkContext.setLogLevel(\\\"WARN\\\")\\n\\n return _df\\n\\n\\ndef main(_spark):\\n # Perform extraction\\n print(\\\"Perform extraction\\\")\\n df = extract(spark)\\n\\n # Perform transformation\\n print(\\\"Perform transformation\\\")\\n df = transform(df, spark)\\n\\n print(\\\"Printing Transformed Dataframe Schema\\\")\\n df.printSchema()\\n\\n # Load the data, do some printing, make it searchable for the API, maybe a postgres\\n print(\\\"Perform load\\\")\\n load(df)\\n\\n print(\\\"Spark application ends\\\")\\n\\n # Stop spark application\\n spark.stop()\\n\\n\\nif __name__ == \\\"__main__\\\":\\n # Initial setup of spark project\\n warehouse_location = abspath('spark-warehouse')\\n spark = SparkSession \\\\\\n .builder \\\\\\n .appName(\\\"CitiesCountriesTest\\\") \\\\\\n .config(\\\"spark.sql.warehouse.dir\\\", warehouse_location) \\\\\\n .enableHiveSupport() \\\\\\n .getOrCreate()\\n\\n main(spark)\\n\"\n }\n]"},"num_files":{"kind":"number","value":10,"string":"10"}}},{"rowIdx":2980390,"cells":{"repo_name":{"kind":"string","value":"Arkkav/git-learing"},"repo_url":{"kind":"string","value":"https://github.com/Arkkav/git-learing"},"snapshot_id":{"kind":"string","value":"d0b748d6304a1357a88ac227b76387b03e2fd4a4"},"revision_id":{"kind":"string","value":"4888dbe55f1ab77f2c3972ed12d9fe03f54b991e"},"directory_id":{"kind":"string","value":"fc4ac2cb918c9d8ca4f71d45657a5991ea8c54ba"},"branch_name":{"kind":"string","value":"refs/heads/master"},"visit_date":{"kind":"timestamp","value":"2021-01-07T07:30:13.035094","string":"2021-01-07T07:30:13.035094"},"revision_date":{"kind":"timestamp","value":"2020-03-31T08:11:40","string":"2020-03-31T08:11:40"},"committer_date":{"kind":"timestamp","value":"2020-03-31T08:11:40","string":"2020-03-31T08:11:40"},"github_id":{"kind":"number","value":241614959,"string":"241,614,959"},"star_events_count":{"kind":"number","value":0,"string":"0"},"fork_events_count":{"kind":"number","value":0,"string":"0"},"gha_license_id":{"kind":"null"},"gha_created_at":{"kind":"null"},"gha_updated_at":{"kind":"null"},"gha_pushed_at":{"kind":"null"},"gha_language":{"kind":"null"},"files":{"kind":"list like","value":[{"alpha_fraction":0.7599999904632568,"alphanum_fraction":0.7599999904632568,"avg_line_length":15.777777671813965,"blob_id":"4b3878dbf7e6f7ca78759c26d05edfe5ea5fb12d","content_id":"674b63226989a3e5758f6090c591cd1405228dfe","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":159,"license_type":"no_license","max_line_length":36,"num_lines":9,"path":"/Django_projects/vk_oauth/vk_oauth/apps/articles/views.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"from django.http import HttpResponse\n\n\ndef index(request):\n\treturn HttpResponse(\"Привет, мир!\")\n\n\ndef test(request):\n\treturn HttpResponse(\"Test page\")"},{"alpha_fraction":0.6798780560493469,"alphanum_fraction":0.6829268336296082,"avg_line_length":22.5,"blob_id":"855fff72efeea51d23aca1e346a5a3815e793dba","content_id":"eec7a8f21692a18311c71ffc0b3102879c53fa3d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":418,"license_type":"no_license","max_line_length":50,"num_lines":14,"path":"/Python_stepik_2/1.5.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"class MoneyBox:\n\n\tdef __init__(self, capacity):\n\t\t# конструктор с аргументом – вместимость копилки\n\t\tself.capacity = capacity\n\t\tself.count = 0\n\n\tdef can_add(self, v):\n\t\t# True, если можно добавить v монет, False иначе\n\t\t\treturn self.count + v <= self.capacity\n\n\tdef add(self, v):\n\t\t# положить v монет в копилку\n\t\tself.count += v"},{"alpha_fraction":0.4110429584980011,"alphanum_fraction":0.4519427418708801,"avg_line_length":24.789474487304688,"blob_id":"b64f815a42415b83cbb2d9cb49d13c89ba9e52f9","content_id":"6216bec9c6c429d744dd7dee7fc93470e3e7d527","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":543,"license_type":"no_license","max_line_length":73,"num_lines":19,"path":"/MyPr1/ex3.7.4.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"a = ('\\n'.join(input().lower() for i in range(int(input())))).split('\\n')\nx = 0\ny = 0\nfor s in a:\n\tif s[0:2] == 'юг':\n\t\ty -= int(s[3:])\n\telif s[0:5] == 'запад':\n\t\tx -= int(s[6:])\n\telif s[0:6] == 'восток':\n\t\tx += int(s[7:])\n\telif s[0:5] == 'север':\n\t\ty += int(s[6:])\nprint(str(x) + ' ' + str(y))\n\n# s = {'север': 0, 'запад': 0, 'юг': 0, 'восток': 0}\n# for i in range(int(input())):\n# k = input().split()\n# s[k[0]] += int(k[1])\n# print(s['восток'] - s['запад'], s['север'] - s['юг'])"},{"alpha_fraction":0.6132075190544128,"alphanum_fraction":0.6320754885673523,"avg_line_length":16.83333396911621,"blob_id":"aafb89885ca199c9b477a8768556461600af3b65","content_id":"69ef3b2aa248b2fbc2f70f0b1236e878d72792df","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":106,"license_type":"no_license","max_line_length":39,"num_lines":6,"path":"/Python_stepik_2/3.3.9.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import re\nimport sys\n\nfor line in sys.stdin:\n\tline = line.rstrip()\n\tprint(re.sub(r\"(\\w)\\1+\", r'\\1', line))"},{"alpha_fraction":0.540145993232727,"alphanum_fraction":0.5839415788650513,"avg_line_length":21.83333396911621,"blob_id":"8a18f08225cde2f00b7e8b3282e28a98358f6075","content_id":"ab9914576b0697636d9645953a329b02c8f00d76","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":137,"license_type":"no_license","max_line_length":62,"num_lines":6,"path":"/Python_stepik_2/test.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"a = set([1, 2, 3])\nb = set([1, 3, 2])\n\nprint(a == b, type(a), sorted(b))\n\ns = \"abba com mother bill mother com abba dog abba mother com\"\n"},{"alpha_fraction":0.5813953280448914,"alphanum_fraction":0.6511628031730652,"avg_line_length":31.125,"blob_id":"4d30309fd74e9fd1f215536dba5bf6f235cae44f","content_id":"e2fdc7c55b697444b689fad3df860ade54260594","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":258,"license_type":"no_license","max_line_length":82,"num_lines":8,"path":"/MyPr1/ex17.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import requests\n\nr = requests.get('https://stepic.org/media/attachments/course67/3.6.3/699991.txt')\nb = r.text.strip()\nwhile b[0:2] !='We':\n r = requests.get('https://stepic.org/media/attachments/course67/3.6.3/' + b)\n b = r.text.strip()\n print(b)\n\n"},{"alpha_fraction":0.5072045922279358,"alphanum_fraction":0.5072045922279358,"avg_line_length":19.47058868408203,"blob_id":"8cce65453e5ec2ce4f038b0a6cf68e7864f72aa4","content_id":"c8562ace5d78ea7961d8806aed73796c3142b889","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":347,"license_type":"no_license","max_line_length":47,"num_lines":17,"path":"/MyPr1/ex19.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"a, b, c, d = input(), input(), input(), input()\ng = {}\nfor i in range(len(a)):\n g[a[i]] = b[i]\n\ndef cypher(a, d):\n b = ''\n for i in range(len(a)):\n b += d[a[i]]\n return b\ndef uncode(d):\n uncode = {}\n for key, value in d.items():\n uncode[value] = key\n return uncode\nprint(cypher(c, g))\nprint(cypher(d, uncode(g)))"},{"alpha_fraction":0.41081079840660095,"alphanum_fraction":0.4216216206550598,"avg_line_length":22.125,"blob_id":"e35a367eab4f8d809beb1114231571f9ad791909","content_id":"64157a156220be3636d99a1d2aeb86b1688f872d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":185,"license_type":"no_license","max_line_length":50,"num_lines":8,"path":"/MyPr1/ex10.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"\nlst = [i.lower() for i in input().split()]\ns = {}\nfor i in lst:\n if s.get(i) == None:\n s[i] = 1\n else:\n s[i] += 1\nprint(*[i + ' ' + str(s[i]) for i in s], sep='\\n')"},{"alpha_fraction":0.6003372669219971,"alphanum_fraction":0.6003372669219971,"avg_line_length":22.760000228881836,"blob_id":"8f4afefea8a30bf4fe64ec43b19ba0d04d8d5c29","content_id":"610afc42a15a1c74842f2e3f794d2dbaacb8aa13","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":624,"license_type":"no_license","max_line_length":102,"num_lines":25,"path":"/Python_stepik_2/3.5.2.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import json\n\n\ndef count_descendant(Vertex): # для класса Vertex возвращает всех потомков []\n\tch = [Vertex]\n\tfor i in js:\n\t\tif Vertex in i[\"parents\"]:\n\t\t\tch.append(i[\"name\"])\n\t\t\tch.extend(count_descendant(i[\"name\"]))\n\treturn ch\n\n\njs = json.loads(input())\nprint(*sorted([i[\"name\"] + ' : ' + str(len(set(count_descendant(i[\"name\"])))) for i in js]), sep='\\n')\n\n\n\n# import json\n#\n# cls = {c['name']: c['parents'] for c in json.loads(input())}\n#\n# isbase = lambda b, d: b == d or any(isbase(b, c) for c in cls[d])\n#\n# for p in sorted(cls):\n# print(p, ':', len({c for c in cls if isbase(p, c)}))"},{"alpha_fraction":0.5525835752487183,"alphanum_fraction":0.5610942244529724,"avg_line_length":21.243244171142578,"blob_id":"c01f5e9efb826bd08cadeae8fb6293a4299f3928","content_id":"9f939bcde84169a0a33803f3395e4e3c06dd4c29","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1737,"license_type":"no_license","max_line_length":83,"num_lines":74,"path":"/Python_stepik_2/2.1.2.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"n = int(input())\nVertex = {} # вершина : множество прямых предков\n\n\ndef get_direct_ancestor(a, vertex): # возвращает множество всех предков\n\n\tstack = []\n\tfor i in vertex[a]:\n\t\tstack.append(i)\n\t\tif i:\n\t\t\tstack.extend(get_direct_ancestor(i, vertex))\n\treturn set(stack)\n\n\nfor i in range(n):\n\tVertexBuff = input().split()\n\tVertex[VertexBuff[0]] = set()\n\tif len(VertexBuff) != 1:\n\t\t[Vertex[VertexBuff[0]].add(i) for i in VertexBuff[2:]]\nq = int(input())\nex = [input() for i in range(q)]\nre = []\nfor i in range(len(ex) - 1, 0, -1):\n\tif ex[i] not in Vertex:\n\t\tcontinue\n\tanc = get_direct_ancestor(ex[i], Vertex)\n\tfor j in range(i):\n\t\tif ex[i] in re:\n\t\t\tcontinue\n\t\tif ex[j] == ex[i] or ex[j] in anc:\n\t\t\tre.append(ex[i])\nfor i in range(len(re) - 1, -1, -1):\n\tprint(re[i])\n\n# \tчто это за генератор?\n# def checkdup(d):\n# return cls[d] is None or any(map(checkdup, cls[d]))\n#\n# cls = {d: set(b[1:]) for _ in range(int(input())) for d, *b in [input().split()]}\n#\n# for _ in range(int(input())):\n# c = input()\n# if checkdup(c):\n# print(c)\n# cls[c] = None\n\n\n# # код преподавателя\n# n = int(input())\n# classes = {}\n# for i in range(n):\n# line = input()\n# parts = line.split(\" : \")\n# cls = parts[0]\n# if len(parts) == 1:\n# classes[cls] = []\n# else:\n# classes[cls] = parts[1].split(\" \")\n#\n#\n# def check(src, dest):\n# if src == dest:\n# return True\n# return any([check(child, dest) for child in classes[src]])\n#\n#\n# m = int(input())\n# used = []\n#\n# for i in range(m):\n# cls = input()\n# if any([check(cls, used_one) for used_one in used]):\n# print(cls)\n# used.append(cls)"},{"alpha_fraction":0.3731343150138855,"alphanum_fraction":0.45895522832870483,"avg_line_length":32.54166793823242,"blob_id":"8d45181b2b54373826f4c7fad51fd1cc2c095e86","content_id":"c60dad311c2946929cf166fe94f10669a2d78117","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":814,"license_type":"no_license","max_line_length":161,"num_lines":24,"path":"/MyPr1/ex3.7.5.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"with open(r'C:\\Users\\arkka\\PycharmProjects\\MyPr1\\dataset_3380_5.txt', 'r') as file1:\n\td = {'1' : [0, 0], '2' : [0, 0], '3' : [0, 0], '4' : [0, 0], '5' : [0, 0], '6' : [0, 0], '7' : [0, 0], '8' : [0, 0], '9' : [0, 0], '10' : [0, 0], '11' : [0, 0]}\n\ta = file1.read().strip().split('\\n')\n\ta = [i.split() for i in a]\nfor ai in a:\n\td[ai[0]][0] += int(ai[2])\n\td[ai[0]][1] += 1\nfor key in d:\n\tif d[key] == [0, 0]:\n\t\td[key] = '-'\n\telse:\n\t\td[key] = str(d[key][0] / d[key][1])\nprint(*[i + ' ' + d[i] for i in d], sep='\\n')\n\n# d = {i: [] for i in range(1,12)}\n# with open(r'D:\\Новая папка\\dataset_3380_5.txt','r', encoding='utf-8') as f1:\n# for i in f1:\n# d[int(i.split()[0])].append(float(i.split()[2]))\n#\n# for i in range(1,12):\n# if d[i]:\n# print(i, sum(d[i])/len(d[i]))\n# else:\n# print(i, '-')"},{"alpha_fraction":0.6646825671195984,"alphanum_fraction":0.6726190447807312,"avg_line_length":23.047618865966797,"blob_id":"580ec1d93ba093f7874df3dd57cfba997dee2469","content_id":"e0c2899ee67f19c0085a2c0f2db0c77da087b1f0","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":506,"license_type":"no_license","max_line_length":49,"num_lines":21,"path":"/Python_stepik_2/2.1.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"try:\n\tfoo()\nexcept (AssertionError, ZeroDivisionError) as e:\n\tprint(type(e).__name__)\nexcept\tArithmeticError as e:\n\tprint('ArithmeticError')\n\n\n\n\n# except Exception as e:\n\t# print(str(type(e))[8:-2])\n\t# print(type(e).__name__)\n# \tif 'ZeroDivisionError' in str(type(e)):\n# \t\tprint(\"ZeroDivisionError\")\n# \telif 'ArithmeticError' in str(type(e)):\n# \t\tprint(\"ArithmeticError\")\n# except AssertionError as e:\n# \tprint(type(e).__name__)\n# print(type(AssertionError()))\n# print(str(type(AssertionError()))[8:-2])"},{"alpha_fraction":0.6084848642349243,"alphanum_fraction":0.6278787851333618,"avg_line_length":29.592592239379883,"blob_id":"996d0e463cd70fd5ec339de74ff800e8897c13b3","content_id":"6206dc923a6a316974b4d2dd061dc33130338ddc","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1044,"license_type":"no_license","max_line_length":106,"num_lines":27,"path":"/Interview_Tasks/habr_1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"# Есть файл, в котором содержаться слова разделённые пробелом.\n# Например: \"abba com mother bill mother com abba dog abba mother com\".\n# Нужно найти и вывести тройку слов, которые чаще всего встречаются вместе (порядок не имеет значения).\n# То есть в моём примере тройки слов это \"abba com mother\", \"com mother bill\", \"mother bill mother\" и т.д.\n# Тут правильным ответом должно быть \"abba com mother\" (частота — 3 раза).\n# https://habr.com/en/post/439576/\n\ns = \"abba com mother bill mother com abba dog abba mother com\".split()\na = []\nd = dict()\nmax_i = 0\nmax_s = ''\nfor i in range(1,len(s)-1):\n\tb = sorted([s[i - 1], s[i], s[i + 1]])\n\ta.append(b)\n\tc = str(b)\n\tif b in a:\n\t\tif c in d:\n\t\t\td[c] += 1\n\t\telse:\n\t\t\td[c] = 1\n\t\tif d[c] > max_i:\n\t\t\tmax_i = d[c]\n\t\t\tmax_s = s[i - 1] + ' ' + s[i] + ' ' + s[i + 1]\n\nprint(d)\nprint(max_s)"},{"alpha_fraction":0.7368420958518982,"alphanum_fraction":0.7368420958518982,"avg_line_length":24.66666603088379,"blob_id":"38aa4010566c6fbd377706c19ad8f48309279acd","content_id":"9aef216ae3c16cf94ddfb72d5ffd4ef1de97cd77","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":76,"license_type":"no_license","max_line_length":32,"num_lines":3,"path":"/MyPr1/codewars1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"def number_to_string(num):\n\treturn str(num)\nprint(number_to_string(input()))"},{"alpha_fraction":0.45652174949645996,"alphanum_fraction":0.489130437374115,"avg_line_length":22.08333396911621,"blob_id":"60d9e47cd68d7b50be0b8e461cad6dcb72a286e3","content_id":"7029cc47bb7d02a0a41ce363c0427fb5c75c8bc8","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":276,"license_type":"no_license","max_line_length":58,"num_lines":12,"path":"/MyPr1/example1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"numbers = [int(i) for i in input().split()]\nsums = ''\nl = len(numbers)\nif l == 1:\n sums = str(numbers[0])\nelse:\n j = 0\n while j < l - 1:\n sums += str(numbers[j - 1] + numbers[j + 1]) + ' '\n j += 1\n sums += str(numbers[l - 2] + numbers[0])\nprint(sums)"},{"alpha_fraction":0.6374622583389282,"alphanum_fraction":0.6374622583389282,"avg_line_length":22.5,"blob_id":"c82d11a570760586cd5505d197a92c2dd9d60bea","content_id":"1795de39c5aecd22d5a29f8390c815f72f8bbd1f","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":331,"license_type":"no_license","max_line_length":75,"num_lines":14,"path":"/Python_stepik_2/2.2.2.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import simplecrypt\n\nwith open(\"encrypted.bin\", \"rb\") as inp, open(\"passwords.txt\", \"r\") as pas:\n\tencrypted = inp.read()\n\t# a = pas.read().strip().split('\\n')\n\tfor i in pas:\n\t\ta = i.strip()\n\t\ttry:\n\t\t\ts = simplecrypt.decrypt(a, encrypted)\n\t\texcept simplecrypt.DecryptionException:\n\t\t\tprint('Error')\n\t\telse:\n\t\t\tprint(a)\n\t\t\tprint(s)\n\n\n"},{"alpha_fraction":0.48759007453918457,"alphanum_fraction":0.5572457909584045,"avg_line_length":20.55172348022461,"blob_id":"83e9480b9a6f6ceb6efa7fb275f7576a5cca2b56","content_id":"9bdb3361b56da7a89e8de076809112b3b57c1a21","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":1411,"license_type":"no_license","max_line_length":64,"num_lines":58,"path":"/Python_stepik_2/1.3.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"def closest_mod_5(x):\n\treturn (x if x % 5 == 0 else x + (5 - x % 5))\nprint(closest_mod_5(int(input())))\n\n# closest_mod_5 = lambda x: (x + 4) // 5 * 5\n\n# a=1\n# b=2\n# lst = [5,6,7,8]\n# dct = {\"1\":11,\"2\":22}\n# print(a, b, lst, dct)\n#\n# def pr(aa, bb, cc, *llst, **ddct):\n# print(aa) #=a\n# print(bb) #=b\n# print(cc) #возьмётся из списка\n# print(llst) #остатки списка в виде кортежа\n# print(lst) #глобальная переменная\n# for k in ddct:\n# print(k, ddct[k])\n# print(ddct) #словарь в к-й попали dct, r=8,t=6\n# print(*llst) #элементы кортежа\n# print(*ddct) #ключи словаря\n# print(*dct) #глобальная переменная\n# pr(a, b, *lst, **dct, r=8,t=6)\n# 1 2 [5, 6, 7, 8] {'1': 11, '2': 22}\n# 1\n# 2\n# 5\n# (6, 7, 8)\n# [5, 6, 7, 8]\n# 1 11\n# 2 22\n# r 8\n# t 6\n# {'1': 11, '2': 22, 'r': 8, 't': 6}\n# 6 7 8\n# 1 2 r t\n# 1 2\n\nКод проверки времени выполнения функций\nimport time\ndef one(a, b, c):\n\treturn a + b + c\n\n\ndef two(*args):\n\treturn args[0] + args[1] + args[2]\ncount = 9000000\n_startTime = time.time()\nfor num in range(count):\n\tone(1, 2, 3)\nprint(\"Positional test:\", time.time() - _startTime)\n\n_startTime = time.time()\nfor num in range(count):\n\ttwo(1, 2, 3)\nprint(\"Positional args as list test:\", time.time() - _startTime)"},{"alpha_fraction":0.4267834722995758,"alphanum_fraction":0.46057572960853577,"avg_line_length":24,"blob_id":"bd284837e79d5efee2efbfea4ab10d0b29403308","content_id":"1d245b720e7145422cada3b137b9ab314de46526","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":799,"license_type":"no_license","max_line_length":84,"num_lines":32,"path":"/MyPr1/ex13.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"# s = ''\n# a = []\n# line = ''\n# with open(r'C:\\Users\\arkka\\PycharmProjects\\MyPr1\\dataset_3363_3.txt', 'r') as inf:\n# for line in inf:\n# line += line.strip().lower() + ' '\n# for ltr in line:\n# if ltr != ' ':\n# s += ltr\n# else:\n# a.append(s)\n# s = ''\n# s = {}\n# max = 0\n# max_w = ''\n# for i in a:\n# if s.get(i) == None:\n# s[i] = 1\n# else:\n# s[i] += 1\n# if max < s[i]:\n# max = s[i]\n# max_w = i\n#\n# print(max_w, s[max_w], sep=' ')\n\nwith open(r'C:\\Users\\arkka\\PycharmProjects\\MyPr1\\dataset_3363_3.txt', 'w') as inf:\n inf.write('12\\n')\n inf.write('1234\\n')\nwith open(r'C:\\Users\\arkka\\PycharmProjects\\MyPr1\\dataset_3363_3.txt', 'r') as inf:\n line = inf.read()\nprint(line)"},{"alpha_fraction":0.6388888955116272,"alphanum_fraction":0.6419752836227417,"avg_line_length":17.05555534362793,"blob_id":"645c605853ca2260fb5b6d03c272c9edd2db91d1","content_id":"387366f1260bc2528fc049675effd4821597f6c4","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":324,"license_type":"no_license","max_line_length":41,"num_lines":18,"path":"/Python_stepik_2/3.3.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import re\nimport sys\n\npattern = r\"cat\"\nfor line in sys.stdin:\n\tline = line.rstrip()\n\tmatch_object = re.findall(pattern, line)\n\tif len(match_object) >= 2:\n\t\tprint(line)\n\n#catcatsghdtscat\n# import re\n# import sys\n#\n# for line in sys.stdin:\n# line = line.strip()\n# if re.search(r\"cat.*cat\", line):\n# print(line)"},{"alpha_fraction":0.40112993121147156,"alphanum_fraction":0.4067796468734741,"avg_line_length":13.75,"blob_id":"fde79baf9072f84fb3ea4642d06117bd565e0577","content_id":"b7630c4a4aa2a060ee60210b01eba1b6e9ff9acb","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":177,"license_type":"no_license","max_line_length":20,"num_lines":12,"path":"/MyPr1/ex11.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"def f(x):\n return x + 1\n\nn = int(input())\nb = {}\nfor i in range(n):\n a = int(input())\n if a in b:\n print(b[a])\n else:\n b[a] = f(a)\n print(b[a])\n"},{"alpha_fraction":0.7093023061752319,"alphanum_fraction":0.7093023061752319,"avg_line_length":18.22222137451172,"blob_id":"70031e6e790ca23c3bd652b6d459d0e33adc016e","content_id":"a92e5966fbcc3aa1fe1075550a55c2a9227150ec","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":172,"license_type":"no_license","max_line_length":40,"num_lines":9,"path":"/Python_stepik_2/3.3.2.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import re\nimport sys\n\npattern = r\"\\bcat\\b\"\nfor line in sys.stdin:\n\tline = line.rstrip()\n\tmatch_object = re.search(pattern, line)\n\tif match_object is not None:\n\t\tprint(line)"},{"alpha_fraction":0.6648044586181641,"alphanum_fraction":0.6703910827636719,"avg_line_length":18.55555534362793,"blob_id":"4b94f35ae6290fa404558cf9c042a767b450986b","content_id":"0999e7eb1f35623347f73c00f6a7f13fc82813df","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":179,"license_type":"no_license","max_line_length":40,"num_lines":9,"path":"/Python_stepik_2/3.3.5.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import re\nimport sys\n\npattern = r\"\\b(.+)\\1\\b\"\nfor line in sys.stdin:\n\tline = line.rstrip()\n\tmatch_object = re.search(pattern, line)\n\tif match_object is not None:\n\t\tprint(line)\n\n\n\n"},{"alpha_fraction":0.6766917109489441,"alphanum_fraction":0.7368420958518982,"avg_line_length":32.5,"blob_id":"eacd28b3222d72abd6ee9b5a5cadf68259965f66","content_id":"44768accd9dc247f4d282df7bf7d3c275058889d","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":133,"license_type":"no_license","max_line_length":79,"num_lines":4,"path":"/MyPr1/ex16.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import requests\nr = requests.get('https://stepic.org/media/attachments/course67/3.6.2/983.txt')\nb = r.text.splitlines()\nprint(len(b))"},{"alpha_fraction":0.5966101884841919,"alphanum_fraction":0.6203389763832092,"avg_line_length":13.047618865966797,"blob_id":"3c264c182393a726f6a223676a35d8b91de4efc9","content_id":"9f3d578f29667b01761d0bde965d46e8a652cc44","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":295,"license_type":"no_license","max_line_length":61,"num_lines":21,"path":"/Python_stepik_2/2.3.2.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import math\nimport itertools\n\ndef primes():\n\tcount = 2\n\n\twhile True:\n\t\tisprime = True\n\n\t\tfor x in range(2, int(math.sqrt(count) + 1)):\n\t\t\tif count % x == 0:\n\t\t\t\tisprime = False\n\t\t\t\tbreak\n\n\t\tif isprime:\n\t\t\tyield count\n\n\t\tcount += 1\n\n\nprint(list(itertools.takewhile(lambda x: x <= 31, primes())))\n"},{"alpha_fraction":0.3937007784843445,"alphanum_fraction":0.4724409580230713,"avg_line_length":13.222222328186035,"blob_id":"82932a0f7775c5193fdb3f8ad6545894af82bb8c","content_id":"498df92f3b6d3bfe1f7eb7dcc61a5815f93bbe01","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":127,"license_type":"no_license","max_line_length":20,"num_lines":9,"path":"/MyPr1/ex3.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"sum1 = 0\nsum2 = 0\nwhile True:\n a = int(input())\n sum1 += a\n sum2 += a ** 2\n if sum1 == 0:\n break\nprint(sum2)"},{"alpha_fraction":0.7222222089767456,"alphanum_fraction":0.7222222089767456,"avg_line_length":25.75,"blob_id":"073cdce3d5ffc02e889f98a63ce3d4b5a6430941","content_id":"f56217f6bd0f4d3391d252bb8323384fb3f6bf24","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":108,"license_type":"no_license","max_line_length":35,"num_lines":4,"path":"/Python_stepik_2/1.6.3.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"class LoggableList(list, Loggable):\n\tdef append(self, object):\n\t\tsuper().append(object)\n\t\tself.log(object)\n\n"},{"alpha_fraction":0.6722221970558167,"alphanum_fraction":0.6722221970558167,"avg_line_length":29.16666603088379,"blob_id":"a38ba225262f080ba9dbfd617405fe24c81268ef","content_id":"4c7fc8b5b23840f34b7397aead5d42590b045e09","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":180,"license_type":"no_license","max_line_length":81,"num_lines":6,"path":"/Python_stepik_2/2.2.1.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import datetime\n\nyyyy, mm, dd = input().split()\ndelta = int(input())\nd = datetime.date(int(yyyy), int(mm), int(dd)) + datetime.timedelta(days = delta)\nprint(d.year, d.month, d.day)"},{"alpha_fraction":0.5845070481300354,"alphanum_fraction":0.6338028311729431,"avg_line_length":16.875,"blob_id":"8a085c15551c7f991a014512138e557f5d65d3e3","content_id":"1671f6a7d877cc5e4af713b7bf04bd9ce837d23c","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":142,"license_type":"no_license","max_line_length":42,"num_lines":8,"path":"/Python_stepik_2/3.3.10.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import re\nimport sys\n\nfor line in sys.stdin:\n\tline = line.rstrip()\n\ts = re.fullmatch(r\"(1(01*0)*1|0)*\", line)\n\tif s is not None:\n\t\tprint(s[0])"},{"alpha_fraction":0.549199104309082,"alphanum_fraction":0.5606407523155212,"avg_line_length":24.735294342041016,"blob_id":"d8fa33e64296e1986a8a021b0cdc7866e6410383","content_id":"d3d55582c8c2c6f4daba139e217bae296276038b","detected_licenses":[],"is_generated":false,"is_vendor":false,"language":"Python","length_bytes":874,"license_type":"no_license","max_line_length":112,"num_lines":34,"path":"/Python_stepik_2/3.4.2.py","repo_name":"Arkkav/git-learing","src_encoding":"UTF-8","text":"import re\nimport requests\n\nhtml_file = input()\n# html_file = 'http://pastebin.com/raw/7543p0ns'\n\nlink_pattern = re.compile(r'\\n\\t\\t\\t
\\n\\t\\tSelected Ingredients:
\\n\\t\\t\\t- \\n\\t\\t\\t{% for x in selected_list %}\\n\\t\\t\\t\\t
- {{ x }} \\n\\t\\t\\t{% endfor %}\\n\\t\\t\\t
\\n\\t\\t\\t{% include 'cards.html' %}\\n\\t\\t
\\n\\t