it is working now

src/generate_queries_alternative.py

@@ -7,6 +7,7 @@ import chromadb
 import spacy
 import numpy as np
 import os
+from src.sparql_query_wikibase import wikibase_properties_id, classes_wikibase
 #os.environ["TOKENIZERS_PARALLELISM"] = "false"
 
 import spacy
@@ -20,63 +21,7 @@ def get_nlp():
         return spacy.load("en_core_web_sm")
 nlp = get_nlp()
 
-wikibase_properties_id = {'instance of': 'P2',
- 'reference URL': 'P24',
- 'start time': 'P15',
- 'end time': 'P16',
- 'occupation title': 'P25',
- 'educated at': 'P9',
- 'employer': 'P10',
- 'work location': 'P7',
- 'award received': 'P18',
- 'point in time': 'P28',
- 'exact match': 'P23',
- 'date of birth': 'P3',
- 'place of birth': 'P4',
- 'date of death': 'P5',
- 'country of citizenship': 'P6',
- 'occupation': 'P19',
- 'sex or gender': 'P8',
- 'official website': 'P17',
- 'perfumes': 'P27',
- 'who wears it': 'P26',
- 'inception': 'P11',
- 'headquarters location': 'P12',
- 'parent organization': 'P13',
- 'founded by': 'P14',
- 'owned by': 'P22',
- 'industry': 'P20',
- 'country': 'P30',
- 'total revenue': 'P21',
- 'designer employed': 'P29',
- 'country of origin': 'P30',
- 'fashion collection': 'P31',
- 'fashion season': 'P32',
- 'fashion show location': 'P33',
- 'description of fashion collection': 'P34',
- 'image of fashion collection': 'P35',
- 'editor of fashion collection description': 'P36',
- 'date of fashion collection': 'P37',
- 'fashion show category': 'P38',
- 'fashion house X fashion collection': 'P39'}
-
-classes_wikibase = {'fashion designer': 'Q5',
- 'fashion house': 'Q1',
- 'business': 'Q9',
- 'academic institution': 'Q2',
- 'geographic location': 'Q4',
- 'fashion award': 'Q8',
- 'gender': 'Q6',
- 'occupation': 'Q7',
- 'human': 'Q36',
- 'organization': 'Q3',
- 'brand': 'Q38',
- 'lifestyle brand': 'Q3417',
- 'privately held company': 'Q1729',
- 'fashion season': 'Q8199',
- 'fashion show category': 'Q8200',
- 'fashion season collection': 'Q8201',
- 'fashion journalist': 'Q8207'}
+
 
 questions_queries_all = [{ "question": education_most_popular_question, "query": education_most_popular_query},
                           { "question": how_many_designers_per_fashion_house_question, "query": how_many_designers_per_fashion_house_query},
@@ -91,19 +36,12 @@ questions_queries_all = [{ "question": education_most_popular_question, "query":
                            {"question": collections_jeans_question, "query": collections_jeans_query},
                            {"question": creative_directors_school_question, "query": creative_directors_school_query}, 
                            {"question": fashion_houses_thematic_collection_question, "query": fashion_houses_thematic_collection_query},
-                           {"question": fashion_house_directors_question.substitute({ "x": "Chanel"}), "query": fashion_house_directors_query},
-                          { "question": designer_fashion_house_directors_question.substitute({ "x": "Alexander McQueen"}), "query": designer_fashion_house_directors_query},
-                          { "question": country_designer_question.substitute({ "x": "Italy"}), "query": country_designer_query},
-                          { "question": designer_order_fashion_collection_question.substitute({ "x": "Alexander McQueen"}), "query": designer_order_fashion_collection_query},
-                          { "question": designer_fashion_director_question2.substitute({ "x": "Alexander McQueen"}), "query": designer_fashion_director_query2},
-                          { "question": year_designers_birth_question.substitute({ "x": "1970"}), "query": year_designers_birth_query}
-                        #    {"question": fashion_house_directors_question.substitute({ "x": f"{"Chanel"}"}), "query": fashion_house_directors_query.substitute({ "x": f"'{"Chanel"}'"})},
-                        #   { "question": designer_fashion_house_directors_question.substitute({ "x": f"{"Alexander McQueen"}"}), "query": designer_fashion_house_directors_query.substitute({ "x": f"'{"Alexander McQueen"}'"})},
-                        #   {"question": country_designer_question.substitute({ "x": f"{"Italy"}"}), "query": country_designer_query.substitute({ "x": f"'{"Italy"}'"})},
-                        #   {  "question": designer_order_fashion_collection_question.substitute({ "x": f"{"Alexander McQueen"}"}), "query": designer_order_fashion_collection_query.substitute({ "x": f"'{"Alexander McQueen"}'"})},
-                        #   {"question": designer_fashion_director_question2.substitute({ "x": f"{"Alexander McQueen"}"}), "query": designer_fashion_director_query2.substitute({ "x": f"'{"Alexander McQueen"}'"})},
-                        #   { "question": year_designers_birth_question.substitute({ "x": 1970}), "query": year_designers_birth_query.substitute({ "x": 1970})},
-                        #   { "question": fashion_collection_images_question.substitute({ "x": f"{"Chanel"}"}), "query": fashion_collection_images_query.substitute({ "x": f"'{"Chanel"}'"})}
+                            {"question": fashion_house_directors_question, "query": fashion_house_directors_query},
+                          { "question": designer_fashion_house_directors_question, "query": designer_fashion_house_directors_query},
+                          { "question": country_designer_question, "query": country_designer_query},
+                          { "question": designer_order_fashion_collection_question, "query": designer_order_fashion_collection_query},
+                          { "question": designer_fashion_director_question2, "query": designer_fashion_director_query2},
+                          { "question": year_designers_birth_question, "query": year_designers_birth_query}
                            ]
 
 if os.path.exists("web_app/query_log.json"):
@@ -206,8 +144,7 @@ def similarity_question(question, questions_queries_dictionary, collection, n_re
 
     # Store each unique document in the vector embedding database
     for i, d in enumerate(masked_documents):
-        response = get_embeddings(d)
-        embedding = response[0]  # Extract the first (and only) embedding from the nested list
+        embedding = get_embeddings(d)[0]
         
         # Check if embedding is unique
         is_duplicate = any(np.allclose(embedding, np.array(e), atol=1e-6) for e in unique_embeddings.values())
@@ -222,8 +159,7 @@ def similarity_question(question, questions_queries_dictionary, collection, n_re
     
     # Compute the embedding for the input question
     masked_question = mask_entities(question, nlp)
-    response = get_embeddings(d)
-    query_embedding = response[0]  # Extract embedding
+    query_embedding = get_embeddings(d)[0]
 
     results = collection.query(
         query_embeddings=[query_embedding],  # Ensure correct format
@@ -265,9 +201,7 @@ def similarity_question_no_masking(question, questions_queries_dictionary, colle
 
     # Store each unique document in the vector embedding database
     for i, d in enumerate(original_documents):
-        response = get_embeddings(d)
-        embedding = response[0]  # Extract the first (and only) embedding from the nested list
-        
+        embedding = get_embeddings(d)[0]
         # Check if embedding is unique
         is_duplicate = any(np.allclose(embedding, np.array(e), atol=1e-6) for e in unique_embeddings.values())
         if not is_duplicate:
@@ -281,8 +215,7 @@ def similarity_question_no_masking(question, questions_queries_dictionary, colle
     
     # Compute the embedding for the input question
     
-    response = get_embeddings(question)
-    query_embedding = response[0]  # Extract embedding
+    query_embedding = get_embeddings(question)[0]
 
     results = collection.query(
         query_embeddings=[query_embedding],  # Ensure correct format

src/questions_queries.py

@@ -5,8 +5,8 @@ from src.sparql_query_wikibase import *
 
 
 
-fashion_collection_images_question = Template("Images of collections produced by $x")
-fashion_collection_images_query = Template("""
+fashion_collection_images_question = "Images of collections produced by Chanel"
+fashion_collection_images_query = """
 PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
 PREFIX wb: <https://fashionwiki.wikibase.cloud/entity/>
 PREFIX pq: <https://fashionwiki.wikibase.cloud/prop/qualifier/>  
@@ -17,9 +17,9 @@ PREFIX prov: <http://www.w3.org/ns/prov#>
 
 SELECT ?statement ?fashion_collectionLabel ?image_url ?reference_URL 
   ?fashion_seasonLabel (YEAR(?fashion_year) AS ?year) ?categoryLabel WHERE {
-    { ?fashion_house  rdfs:label $x@en.}  # Match label
+    { ?fashion_house  rdfs:label "Chanel"@en.}  # Match label
   UNION
-  { ?fashion_house  skos:altLabel $x@en. }  # Match alias   
+  { ?fashion_house  skos:altLabel "Chanel"@en. }  # Match alias   
                                            
   
   ?fashion_house p:P31 ?statement.
@@ -39,10 +39,10 @@ SELECT ?statement ?fashion_collectionLabel ?image_url ?reference_URL
   }
 
   SERVICE wikibase:label { bd:serviceParam wikibase:language "en". } }
-                                           """)
+                                           """
 
-year_designers_birth_question = Template("Which designers were born in $x?")
-year_designers_birth_query = Template("""
+year_designers_birth_question = "Which designers were born in 1963?"
+year_designers_birth_query = """
 PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
 PREFIX wb: <https://fashionwiki.wikibase.cloud/entity/>
 PREFIX ps: <https://fashionwiki.wikibase.cloud/prop/statement/> 
@@ -53,13 +53,13 @@ PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
 
 SELECT ?designerLabel ?birthdate  WHERE {
   ?designer wbt:P3 ?birthdate .
-  FILTER (YEAR(?birthdate) = $x)
+  FILTER (YEAR(?birthdate) = 1963)
 
   SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
-} """)
+} """
 
-fashion_house_directors_question = Template("Which designers have been creative directors of $x ?")
-fashion_house_directors_query = Template("""
+fashion_house_directors_question = "Which designers have been creative directors of Chanel ?"
+fashion_house_directors_query = """
 PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
 PREFIX wb: <https://fashionwiki.wikibase.cloud/entity/>
 PREFIX pq: <https://fashionwiki.wikibase.cloud/prop/qualifier/>  
@@ -74,9 +74,9 @@ SELECT
 WHERE {
   # Fashion house identified by brand name
   ?fashion_house wbt:P2 wb:Q1.                                       
-    { ?fashion_house rdfs:label $x@en.}  # Match label
+    { ?fashion_house rdfs:label "Chanel"@en.}  # Match label
   UNION
-  { ?fashion_house skos:altLabel $x@en. }  # Match alias   
+  { ?fashion_house skos:altLabel "Chanel"@en. }  # Match alias   
   
   # Link through designer employed (P29) property for the fashion house  - the property for designer employment is P10
   {
@@ -139,12 +139,12 @@ WHERE {
   FILTER(?reference_URL_source != "")
   # Retrieve labels
   SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
-}   ORDER BY ?start_year """)
+}   ORDER BY ?start_year """
 
 
 
-designer_fashion_house_directors_question = Template("In which fashion house did $x work as a creative director?")
-designer_fashion_house_directors_query = Template("""  
+designer_fashion_house_directors_question = "In which fashion house did Alessandro Michele work as a creative director?"
+designer_fashion_house_directors_query = """  
 PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
 PREFIX wb: <https://fashionwiki.wikibase.cloud/entity/>
 PREFIX pq: <https://fashionwiki.wikibase.cloud/prop/qualifier/>  
@@ -160,9 +160,9 @@ SELECT ?fashion_houseLabel
 WHERE {
    # Designer linked to fashion house through roles or founded_by
     ?fashion_designer wbt:P2 wb:Q5.                                              
-     { ?fashion_designer  rdfs:label $x@en.}  # Match label
+     { ?fashion_designer  rdfs:label "Alessandro Michele"@en.}  # Match label
   UNION
-  { ?fashion_designer  skos:altLabel $x@en. }  # Match alias   
+  { ?fashion_designer  skos:altLabel "Alessandro Michele"@en. }  # Match alias   
    
    # Link through employer (P10)
    ?fashion_designer p:P10 ?statement.
@@ -200,11 +200,11 @@ WHERE {
 
    SERVICE wikibase:label { bd:serviceParam wikibase:language "en". } 
 }
-""")
+"""
 
 
-designer_fashion_director_question2 = Template("In which fashion house did $x work as a creative director?")
-designer_fashion_director_query2 = Template("""PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
+designer_fashion_director_question2 = "In which fashion house did Alessandro Michele work as a creative director?"
+designer_fashion_director_query2 = """PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
 PREFIX wb: <https://fashionwiki.wikibase.cloud/entity/>
 PREFIX pq: <https://fashionwiki.wikibase.cloud/prop/qualifier/>  
 PREFIX pr: <https://fashionwiki.wikibase.cloud/prop/reference/>
@@ -217,9 +217,9 @@ SELECT
 WHERE {
   # Fashion designer identified by name
   ?creative_director wbt:P2 wb:Q5.                                       
-    { ?creative_director  rdfs:label $x@en.}  # Match label
+    { ?creative_director  rdfs:label "Alessandro Michele"@en.}  # Match label
   UNION
-  { ?creative_director  skos:altLabel $x@en. }  # Match alias   
+  { ?creative_director  skos:altLabel "Alessandro Michele"@en. }  # Match alias   
 
   # Link through designer employed (P29) property for the fashion house  - the property for designer employment is P30
   {
@@ -256,10 +256,10 @@ WHERE {
 
   # Retrieve labels
   SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
-}   ORDER BY ?start_year """)
+}   ORDER BY ?start_year """
 
-country_designer_question = Template("Which designers are from $x ?")
-country_designer_query = Template("""PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
+country_designer_question = "Which designers are from Italy ?"
+country_designer_query = """PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
 PREFIX wb: <https://fashionwiki.wikibase.cloud/entity/>
 PREFIX pq: <https://fashionwiki.wikibase.cloud/prop/qualifier/>  
 PREFIX pr: <https://fashionwiki.wikibase.cloud/prop/reference/>
@@ -276,9 +276,9 @@ SELECT ?fashion_designerLabel   (SAMPLE(?statement) AS ?statement)
   # Access the statement for the country relationship
   ?fashion_designer p:P6 ?statement.
   ?statement ps:P6 ?country.
-    { ?country  rdfs:label $x@en.}  # Match label
+    { ?country  rdfs:label "Italy"@en.}  # Match label
   UNION
-  { ?country skos:altLabel $x@en. }  # Match alias   
+  { ?country skos:altLabel "Italy"@en. }  # Match alias   
 
   # Retrieve references from the statement
   OPTIONAL {
@@ -290,9 +290,9 @@ SELECT ?fashion_designerLabel   (SAMPLE(?statement) AS ?statement)
   SERVICE wikibase:label { bd:serviceParam wikibase:language "en". } 
 } 
 GROUP BY ?fashion_designerLabel 
-ORDER BY ?fashion_designerLabel """)
+ORDER BY ?fashion_designerLabel """
 
-country_designer_query = Template("""PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
+country_designer_query = """PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
 PREFIX wb: <https://fashionwiki.wikibase.cloud/entity/>
 PREFIX pq: <https://fashionwiki.wikibase.cloud/prop/qualifier/>  
 PREFIX pr: <https://fashionwiki.wikibase.cloud/prop/reference/>
@@ -306,9 +306,9 @@ SELECT ?fashion_designerLabel (SAMPLE(?reference_URL) AS ?reference_URL) {
   ?fashion_designer wbt:P2 wb:Q5.
   ?fashion_designer wbt:P6 ?country.
   ?fashion_designer p:P6 ?statement.
-    { ?country  rdfs:label $x@en.}  # Match label
+    { ?country  rdfs:label "Italy"@en.}  # Match label
   UNION
-  { ?country  skos:altLabel $x@en. }  # Match alias   
+  { ?country  skos:altLabel "Italy"@en. }  # Match alias   
  
 
   # Retrieve references from the statement
@@ -321,11 +321,11 @@ SELECT ?fashion_designerLabel (SAMPLE(?reference_URL) AS ?reference_URL) {
   SERVICE wikibase:label { bd:serviceParam wikibase:language "en". } 
 } 
 GROUP BY ?fashion_designerLabel 
-ORDER BY ?fashion_designerLabel  """)
+ORDER BY ?fashion_designerLabel  """
 
 
-designer_order_fashion_collection_question = Template("Which fashion collections did $x work on?")
-designer_order_fashion_collection_query = Template("""PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
+designer_order_fashion_collection_question = "Which fashion collections did Alessandro Michele work on?"
+designer_order_fashion_collection_query = """PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
 PREFIX wb: <https://fashionwiki.wikibase.cloud/entity/>
 PREFIX pq: <https://fashionwiki.wikibase.cloud/prop/qualifier/>  
 PREFIX pr: <https://fashionwiki.wikibase.cloud/prop/reference/>
@@ -340,9 +340,9 @@ SELECT ?fashion_houseLabel
       (YEAR(?fashion_year) as ?year ) ?fashion_seasonLabel ?categoryLabel WHERE {
    # Designer linked to fashion house through roles or founded_by
    ?fashion_designer wbt:P2 wb:Q5. 
-  { ?fashion_designer  rdfs:label $x@en.}  # Match label
+  { ?fashion_designer  rdfs:label "Alessandro Michele"@en.}  # Match label
   UNION
-  { ?fashion_designer  skos:altLabel $x@en. }  # Match alias                                               
+  { ?fashion_designer  skos:altLabel "Alessandro Michele"@en. }  # Match alias                                               
                                             
    
    # Link through "creative director" or other roles
@@ -384,7 +384,7 @@ FILTER(
    }
 
    SERVICE wikibase:label { bd:serviceParam wikibase:language "en". } 
-} ORDER BY ?year """)
+} ORDER BY ?year """
 
 
 education_most_popular_question = "What are the most popular educational institutions for fashion designers?"
@@ -683,20 +683,13 @@ questions_queries_dict = [ { "question": education_most_popular_question, "query
                            {"question": collections_sustainability_theme_question, "query": collections_sustainability_theme_query},
                            ]
 
-# questions_queries_dict_entity = [{"question": fashion_house_directors_question.substitute({ "x": f"{"Chanel"}"}), "query": fashion_house_directors_query},
-#                           { "question": designer_fashion_house_directors_question.substitute({ "x": f"{"Alexander McQueen"}"}), "query": designer_fashion_house_directors_query},
-#                           { "question": country_designer_question.substitute({ "x": f"{"Italy"}"}), "query": country_designer_query},
-#                           { "question": designer_order_fashion_collection_question.substitute({ "x": f"{"Alexander McQueen"}"}), "query": designer_order_fashion_collection_query},
-#                           { "question": designer_fashion_director_question2.substitute({ "x": f"{"Alexander McQueen"}"}), "query": designer_fashion_director_query2},
-#                           { "question": year_designers_birth_question.substitute({ "x": f"{"1970"}"}), "query": year_designers_birth_query}]
-
 
-questions_queries_dict_entity = [{"question": fashion_house_directors_question.substitute({ "x": "Chanel"}), "query": fashion_house_directors_query},
-                          { "question": designer_fashion_house_directors_question.substitute({ "x": "Alexander McQueen"}), "query": designer_fashion_house_directors_query},
-                          { "question": country_designer_question.substitute({ "x": "Italy"}), "query": country_designer_query},
-                          { "question": designer_order_fashion_collection_question.substitute({ "x": "Alexander McQueen"}), "query": designer_order_fashion_collection_query},
-                          { "question": designer_fashion_director_question2.substitute({ "x": "Alexander McQueen"}), "query": designer_fashion_director_query2},
-                          { "question": year_designers_birth_question.substitute({ "x": "1970"}), "query": year_designers_birth_query}]
+questions_queries_dict_entity = [{"question": fashion_house_directors_question, "query": fashion_house_directors_query},
+                          { "question": designer_fashion_house_directors_question, "query": designer_fashion_house_directors_query},
+                          { "question": country_designer_question, "query": country_designer_query},
+                          { "question": designer_order_fashion_collection_question, "query": designer_order_fashion_collection_query},
+                          { "question": designer_fashion_director_question2, "query": designer_fashion_director_query2},
+                          { "question": year_designers_birth_question, "query": year_designers_birth_query}]
 
 
 # Designers

src/sparql_query_wikibase.py

@@ -10,7 +10,70 @@ from wikibaseintegrator.wbi_helpers import get_user_agent
 import pandas as pd
 from string import Template
 queries = False
-from src.wikibase_helpers import wikibase_properties_id, classes_wikibase, config, wikibase_api_url
+wikibase_api_url = 'https://fashionwiki.wikibase.cloud/w/api.php'
+config = {
+"SPARQL_ENDPOINT_URL": "https://fashionwiki.wikibase.cloud/query/sparql",
+'USER_AGENT':  'YourBotName/1.0 (https://yourwebsite.org/bot-info)',
+'WIKIBASE_URL': wikibase_api_url,
+}
+
+wikibase_properties_id = {'instance of': 'P2',
+ 'reference URL': 'P24',
+ 'start time': 'P15',
+ 'end time': 'P16',
+ 'occupation title': 'P25',
+ 'educated at': 'P9',
+ 'employer': 'P10',
+ 'work location': 'P7',
+ 'award received': 'P18',
+ 'point in time': 'P28',
+ 'exact match': 'P23',
+ 'date of birth': 'P3',
+ 'place of birth': 'P4',
+ 'date of death': 'P5',
+ 'country of citizenship': 'P6',
+ 'occupation': 'P19',
+ 'sex or gender': 'P8',
+ 'official website': 'P17',
+ 'perfumes': 'P27',
+ 'who wears it': 'P26',
+ 'inception': 'P11',
+ 'headquarters location': 'P12',
+ 'parent organization': 'P13',
+ 'founded by': 'P14',
+ 'owned by': 'P22',
+ 'industry': 'P20',
+ 'country': 'P30',
+ 'total revenue': 'P21',
+ 'designer employed': 'P29',
+ 'country of origin': 'P30',
+ 'fashion collection': 'P31',
+ 'fashion season': 'P32',
+ 'fashion show location': 'P33',
+ 'description of fashion collection': 'P34',
+ 'image of fashion collection': 'P35',
+ 'editor of fashion collection description': 'P36',
+ 'date of fashion collection': 'P37',
+ 'fashion show category': 'P38',
+ 'fashion house X fashion collection': 'P39'}
+
+classes_wikibase = {'fashion designer': 'Q5',
+ 'fashion house': 'Q1',
+ 'business': 'Q9',
+ 'academic institution': 'Q2',
+ 'geographic location': 'Q4',
+ 'fashion award': 'Q8',
+ 'gender': 'Q6',
+ 'occupation': 'Q7',
+ 'human': 'Q36',
+ 'organization': 'Q3',
+ 'brand': 'Q38',
+ 'lifestyle brand': 'Q3417',
+ 'privately held company': 'Q1729',
+ 'fashion season': 'Q8199',
+ 'fashion show category': 'Q8200',
+ 'fashion season collection': 'Q8201',
+ 'fashion journalist': 'Q8207'}
 
 
 def execute_sparql_query(query: str, prefix: str | None = None, endpoint: str | None = None, user_agent: str | None = None, max_retries: int = 1000, retry_after: int = 60) -> dict:
@@ -65,111 +128,3 @@ def get_results_to_df( query):
     df = df.map(lambda x: x['value'] if pd.notnull(x) else None)
     return df
 
-if queries:
-    query_fashion_designers_template = Template("""
-    PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
-    PREFIX wb: <https://fashionwiki.wikibase.cloud/entity/>
-    PREFIX pq: <https://fashionwiki.wikibase.cloud/prop/qualifier/>  
-    PREFIX ps: <https://fashionwiki.wikibase.cloud/prop/statement/> 
-    PREFIX p: <https://fashionwiki.wikibase.cloud/prop/>            
-
-    SELECT ?fashionDesignerLabel ?fashionDesigner WHERE {
-    ?fashionDesigner wbt:$instance_of wb:$fashion_designer.
-
-    SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
-    } ORDER BY ?fashionDesignerLabel
-    """)
-    query_fashion_designers = query_fashion_designers_template.substitute(
-        {
-            "instance_of": wikibase_properties_id["instance of"],
-            "fashion_designer": classes_wikibase["fashion designer"],
-
-        }
-    )
-
-    query_fashion_houses_template = Template("""
-    PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
-    PREFIX wb: <https://fashionwiki.wikibase.cloud/entity/>
-    PREFIX pq: <https://fashionwiki.wikibase.cloud/prop/qualifier/>  
-    PREFIX ps: <https://fashionwiki.wikibase.cloud/prop/statement/> 
-    PREFIX p: <https://fashionwiki.wikibase.cloud/prop/>            
-
-    SELECT ?fashionHouseLabel ?fashionHouse   WHERE {
-    ?fashionHouse wbt:$instance_of wb:$fashion_house.
-
-    SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
-    } ORDER BY ?fashionHouseLabel
-    """)
-    #query_fashion_designers = query_fashion_designers_template.substitute(wikidata_properties_id["occupation"], fashion_designer = classes_wikidata["fashion designer"], grand_couturier = classes_wikidata["grand couturier"])
-    query_fashion_house= query_fashion_houses_template.substitute(
-        {
-            "instance_of": wikibase_properties_id["instance of"],
-            "fashion_house": classes_wikibase["fashion house"],
-
-        }
-    )
-
-
-
-    query_school_template = Template("""
-    PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
-    PREFIX wb: <https://fashionwiki.wikibase.cloud/entity/>
-    PREFIX pq: <https://fashionwiki.wikibase.cloud/prop/qualifier/>  
-    PREFIX ps: <https://fashionwiki.wikibase.cloud/prop/statement/> 
-    PREFIX p: <https://fashionwiki.wikibase.cloud/prop/>            
-
-    SELECT ?fashionSchoolLabel  WHERE {
-    ?fashionSchool wbt:$instance_of wb:$academic_institution.
-
-    SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
-    } ORDER BY ?fashionSchoolLabel
-    """)
-    #query_fashion_designers = query_fashion_designers_template.substitute(wikidata_properties_id["occupation"], fashion_designer = classes_wikidata["fashion designer"], grand_couturier = classes_wikidata["grand couturier"])
-    query_school = query_school_template.substitute(
-        {
-            "instance_of": wikibase_properties_id["instance of"],
-            "academic_institution": classes_wikibase["academic institution"],
-        })
-
-    query_award_template = Template("""
-    PREFIX wbt: <https://fashionwiki.wikibase.cloud/prop/direct/>
-    PREFIX wb: <https://fashionwiki.wikibase.cloud/entity/>
-    PREFIX pq: <https://fashionwiki.wikibase.cloud/prop/qualifier/>  
-    PREFIX ps: <https://fashionwiki.wikibase.cloud/prop/statement/> 
-    PREFIX p: <https://fashionwiki.wikibase.cloud/prop/>            
-
-    SELECT ?fashionAwardLabel  WHERE {
-    ?fashionAward wbt:$instance_of wb:$fashion_award.
-
-    SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
-    } ORDER BY ?fashionAwardLabel
-    """)
-    #query_fashion_designers = query_fashion_designers_template.substitute(wikidata_properties_id["occupation"], fashion_designer = classes_wikidata["fashion designer"], grand_couturier = classes_wikidata["grand couturier"])
-    query_award = query_award_template.substitute(
-        {
-            "instance_of": wikibase_properties_id["instance of"],
-            "fashion_award": classes_wikibase["fashion award"],
-        })
-
-
-
-def get_fashion_designers_wikibase(output_file):
-    df_designers = get_results_to_df(query_fashion_designers)
-    df_designers.to_csv(output_file, index=False)
-    return get_results_to_df(query_fashion_designers)
-
-
-def get_fashion_houses_wikibase(output_file):
-    df_fashion_houses = get_results_to_df(query_fashion_house)
-    df_fashion_houses.to_csv(output_file, index=False)
-    return get_results_to_df(query_fashion_house)
-
-def get_schools_wikibase(output_file):
-    df_schools = get_results_to_df(query_school)
-    df_schools.to_csv(output_file, index=False)
-    return get_results_to_df(query_school)
-
-def get_awards_wikibase(output_file):
-    df_awards = get_results_to_df(query_award)
-    df_awards.to_csv(output_file, index=False)
-    return get_results_to_df(query_award)

src/visual_qa.py

@@ -9,7 +9,7 @@ os.environ["TOKENIZERS_PARALLELISM"] = "false"
 from datasets import load_dataset
 import chromadb
 
-from datetime import datetime
+from datetime import datetime   
 
 def initialize_collection(collection_name="clip_image_embeddings"):
     # Initialize ChromaDB client (PersistentClient stores embeddings between runs)
@@ -73,7 +73,7 @@ device = torch.device("mps" if torch.backends.mps.is_available() else "cpu")
 model = CLIPModel.from_pretrained(model_name).to(device)
 processor = CLIPProcessor.from_pretrained(model_name)
 
-def main_text_retrieve_images(text, result_query=None, n_retrieved=3):
+def main_text_retrieve_images(text, result_query=None, n_retrieved=5):
 
     
 
@@ -86,6 +86,7 @@ def main_text_retrieve_images(text, result_query=None, n_retrieved=3):
 
     # Convert to pandas DataFrame
     df_emb = full_dataset.to_pandas()
+    #print(df_emb.head())  # Debugging statement
     df_emb = df_emb.drop_duplicates(subset='image_urls')
     df_emb['fashion_clip_image'] = df_emb['fashion_clip_image'].apply(lambda x: x[0] if type(x) == list else None)
     df_emb['image_url'] = df_emb['image_urls'].apply(lambda x: x[0] if x else None)
@@ -112,14 +113,14 @@ def main_text_retrieve_images(text, result_query=None, n_retrieved=3):
         query_embeddings=text_features[0],
         n_results=n_retrieved
     )
-    #print(results)
+
     result_doc = pd.DataFrame(results['documents'][0], columns=["image_url"])
     df_result = df_emb.merge(result_doc[['image_url']], on='image_url', how='inner')
     # Remove columns fashion_clip_image, image_urls, and description
     df_result = df_result.drop(columns=['fashion_clip_image', 'description', 'editor', 'publish_date', 'image_urls'])
     return df_result.to_dict(orient='records')
 
-if __name__ == "__main__":
-    text = "dress"
-    result = main_text_retrieve_images(text)
-    print(result)
+# if __name__ == "__main__":
+#     text = "dress"
+#     result = main_text_retrieve_images(text)
+#     print(result)

src/wikibase_helpers.py

@@ -1,171 +0,0 @@
-import requests
-import logging
-from wikibaseintegrator import WikibaseIntegrator, datatypes,  wbi_helpers
-from wikibaseintegrator.wbi_config import config
-from wikibaseintegrator.wbi_exceptions import MWApiError
-from wikibaseintegrator.wbi_helpers import mediawiki_api_call_helper, SearchError
-from wikibaseintegrator import WikibaseIntegrator, wbi_login
-login_wikibase = wbi_login.Login(user="Traopia", password="diqfiz-wodnI4-jafwax", mediawiki_api_url="https://fashionwiki.wikibase.cloud/w/api.php")
-
-wikibase_api_url = 'https://fashionwiki.wikibase.cloud/w/api.php'
-config = {
-"SPARQL_ENDPOINT_URL": "https://fashionwiki.wikibase.cloud/query/sparql",
-'USER_AGENT':  'YourBotName/1.0 (https://yourwebsite.org/bot-info)',
-'WIKIBASE_URL': wikibase_api_url,
-}
-
-
-
-
-# List of valid language codes (can be expanded)
-VALID_LANGUAGE_CODES = ['en']
-
-def get_property_id_by_label(property_label, api_url):
-    """
-    Resolve the property label to its corresponding property ID from Wikibase.
-    
-    Args:
-        property_label (str): The label of the property to search.
-        api_url (str): The API URL of the target Wikibase or Wikidata.
-    
-    Returns:
-        str: The property ID if found, otherwise None.
-    """
-    url = f'{api_url}/w/api.php?action=wbsearchentities&search={property_label}&language=en&type=property&format=json'
-    response = requests.get(url)
-    
-    if response.status_code == 200:
-        search_results = response.json()
-        if 'search' in search_results and search_results['search']:
-            # Return the first matching property ID
-            return search_results['search'][0]['id']
-        else:
-            logging.info(f"No property found for label: {property_label}")
-            return None
-    else:
-        logging.error(f"Failed to search for property by label in the target Wikibase. HTTP Status Code: {response.status_code}")
-        return None
-    
-
-def get_entity_id_by_label(search_string,wiki, dict_result=False) -> list:
-    """
-    Performs a search for entities in the Wikibase instance using labels and aliases.
-    You can have more information on the parameters in the MediaWiki API help (https://www.wikidata.org/w/api.php?action=help&modules=wbsearchentities)
-
-    :param search_string: A string which should be searched for in the Wikibase instance (labels and aliases)
-    :param wiki: The wiki to search in. It can be "wikidata" or "wikibase"
-    :param dict_result: If True, the result will be a list of dictionaries with the keys 'id', 'label', 'match', 'description' and 'aliases'. If False, the result will be a list of strings with the entity IDs.
-    :return: A list of dictionaries or strings with the search results
-    """
-
-    if wiki == "wikibase":
-        login = login_wikibase
-        mediawiki_api_url = wikibase_api_url
-
-    language = "en"
-    strict_language = False
-
-    params = {
-        'action': 'wbsearchentities',
-        'search': search_string,
-        'language': language,
-        'type': "item",
-        'limit': 50,
-        'format': 'json',
-    }
-
-    if strict_language:
-        params.update({'strict_language': ''})
-
-    cont_count = 0
-    results = []
-
-    while True:
-        params.update({'continue': cont_count})
-        search_results = mediawiki_api_call_helper(data=params, login = login, mediawiki_api_url=mediawiki_api_url, user_agent = config['USER_AGENT'])
-        if search_results['success'] != 1:
-            raise SearchError('Wikibase API wbsearchentities failed')
-
-        for i in search_results['search']:
-            if dict_result:
-                description = i['description'] if 'description' in i else None
-                aliases = i['aliases'] if 'aliases' in i else None
-                results.append({
-                    'id': i['id'],
-                    'label': i['label'],
-                    'match': i['match'],
-                    'description': description,
-                    'aliases': aliases
-                })
-            else:
-                results.append(i['id'])
-
-        if 'search-continue' not in search_results:
-            break
-        cont_count = search_results['search-continue']
-        if cont_count >= 50:
-            break
-    return results
-
-
-wikibase_properties_id = {"instance of": get_property_id_by_label("instance of", wikibase_api_url),
-                    "reference URL": get_property_id_by_label("reference URL", wikibase_api_url),
-                    "start time": get_property_id_by_label("start time", wikibase_api_url),
-                    "end time": get_property_id_by_label("end time", wikibase_api_url),
-                    "occupation title": get_property_id_by_label("occupation title", wikibase_api_url),
-                    "educated at": get_property_id_by_label("educated at", wikibase_api_url),
-                    "employer": get_property_id_by_label("employer", wikibase_api_url),
-                    "work location": get_property_id_by_label("work location", wikibase_api_url),
-                    "award received": get_property_id_by_label("award received", wikibase_api_url),
-                    "point in time": get_property_id_by_label("point in time", wikibase_api_url),
-                    "exact match": get_property_id_by_label("exact match", wikibase_api_url),
-                    "date of birth": get_property_id_by_label("date of birth", wikibase_api_url),
-                    "place of birth": get_property_id_by_label("place of birth", wikibase_api_url),
-                    "date of death": get_property_id_by_label("date of death", wikibase_api_url),
-                    "country of citizenship": get_property_id_by_label("country of citizenship", wikibase_api_url),
-                    "occupation": get_property_id_by_label("occupation", wikibase_api_url),
-                    "sex or gender": get_property_id_by_label("sex or gender", wikibase_api_url),
-                    "official website": get_property_id_by_label("official website", wikibase_api_url),
-                    "perfumes": get_property_id_by_label("perfumes", wikibase_api_url),
-                    "who wears it": get_property_id_by_label("who wears it", wikibase_api_url),
-                    "inception": get_property_id_by_label("inception", wikibase_api_url),
-                    "headquarters location": get_property_id_by_label("headquarters location", wikibase_api_url),
-                    "parent organization": get_property_id_by_label("parent organization", wikibase_api_url),
-                    "founded by": get_property_id_by_label("founded by", wikibase_api_url),
-                    "owned by": get_property_id_by_label("owned by", wikibase_api_url),
-                    "industry": get_property_id_by_label("industry", wikibase_api_url),
-                    "country": get_property_id_by_label("country", wikibase_api_url),
-                    "total revenue": get_property_id_by_label("total revenue", wikibase_api_url),
-                    "designer employed": get_property_id_by_label("designer employed", wikibase_api_url),
-                    "country of origin": get_property_id_by_label("country of origin", wikibase_api_url),
-                    "fashion collection": get_property_id_by_label("fashion collection", wikibase_api_url),
-                    "fashion season": get_property_id_by_label("fashion season", wikibase_api_url),
-                    "fashion show location": get_property_id_by_label("fashion show location", wikibase_api_url),
-                    "description of fashion collection": get_property_id_by_label("description of fashion collection", wikibase_api_url),
-                    "image of fashion collection": get_property_id_by_label("image of fashion collection", wikibase_api_url),
-                    "editor of fashion collection description": get_property_id_by_label("editor of fashion collection description", wikibase_api_url),
-                    "date of fashion collection": get_property_id_by_label("date of fashion collection", wikibase_api_url),
-                    "fashion show category": get_property_id_by_label("fashion show category", wikibase_api_url),
-                    "fashion house X fashion collection": get_property_id_by_label("fashion house X fashion collection", wikibase_api_url),
-                    "designer of collection": get_property_id_by_label("designer of collection", wikibase_api_url)}
-
-
-classes_wikibase = {"fashion designer": get_entity_id_by_label("fashion designer", "wikibase")[0],
-                        "fashion house": get_entity_id_by_label("fashion house", "wikibase")[0],
-                        "business": get_entity_id_by_label("business", "wikibase")[0],
-                        "academic institution": get_entity_id_by_label("academic institution", "wikibase")[0],
-                        "geographic location": get_entity_id_by_label("geographic location", "wikibase")[0],
-                        "fashion award": get_entity_id_by_label("fashion award","wikibase")[0],
-                        "gender":get_entity_id_by_label("gender","wikibase")[0] , 
-                        "occupation": get_entity_id_by_label("occupation","wikibase")[0],
-                        "human": get_entity_id_by_label("human","wikibase")[0],
-                        "organization": get_entity_id_by_label("organization","wikibase")[0],
-                        "brand": get_entity_id_by_label("brand","wikibase")[0],
-                        "lifestyle brand": get_entity_id_by_label("lifestyle brand","wikibase")[0],
-                        "privately held company": get_entity_id_by_label("privately held company","wikibase")[0],
-                        "fashion award": get_entity_id_by_label("fashion award", "wikibase")[0],
-                        "fashion season": get_entity_id_by_label("fashion season", "wikibase")[0],
-                        "fashion show category": get_entity_id_by_label("fashion show category", "wikibase")[0],
-                        "fashion season collection": get_entity_id_by_label("fashion season collection", "wikibase")[0],
-                        "fashion journalist": get_entity_id_by_label("fashion journalist", "wikibase")[0],
-                        }