code
stringlengths 31
1.05M
| apis
list | extract_api
stringlengths 97
1.91M
|
|---|---|---|
'''
@FileName : data_parser.py
@EditTime : 2021-11-29 13:59:47
@Author : <NAME>
@Email : <EMAIL>
@Description :
'''
from __future__ import absolute_import
from __future__ import print_function
from __future__ import division
import sys
import os
import os.path as osp
import platform
import json
from collections import namedtuple
import cv2
import numpy as np
import torch
from torch.utils.data import Dataset
Keypoints = namedtuple('Keypoints',
['keypoints', 'gender_gt', 'gender_pd'])
Keypoints.__new__.__defaults__ = (None,) * len(Keypoints._fields)
def read_keypoints(keypoint_fn, num_people, num_joint):
if not os.path.exists(keypoint_fn):
keypoints = [np.zeros((num_joint, 3))] * num_people # keypoints may not exist
flags = np.zeros((num_people,))
valid = 0
return keypoints, flags, valid
with open(keypoint_fn) as keypoint_file:
data = json.load(keypoint_file)
valid = 1
keypoints = []
flags = np.zeros((len(data['people'])))
for idx, person_data in enumerate(data['people']):
if person_data is None:
body_keypoints = np.zeros((num_joint, 3),
dtype=np.float32)
else:
flags[idx] = 1
body_keypoints = np.array(person_data['pose_keypoints_2d'],
dtype=np.float32)
body_keypoints = body_keypoints.reshape([-1, 3])
keypoints.append(body_keypoints)
return keypoints[:num_people], flags[:num_people], valid
def read_joints(keypoint_fn, use_hands=True, use_face=True,
use_face_contour=False):
"""
load 3D annotation
"""
with open(keypoint_fn) as keypoint_file:
data = json.load(keypoint_file)
keypoints = []
gender_pd = []
gender_gt = []
for idx, person_data in enumerate(data['people']):
try:
body_keypoints = np.array(person_data['pose_keypoints_3d'],
dtype=np.float32)
body_keypoints = body_keypoints.reshape([-1, 4])
if use_hands:
left_hand_keyp = np.array(
person_data['hand_left_keypoints_3d'],
dtype=np.float32).reshape([-1, 4])
right_hand_keyp = np.array(
person_data['hand_right_keypoints_3d'],
dtype=np.float32).reshape([-1, 4])
body_keypoints = np.concatenate(
[body_keypoints, left_hand_keyp, right_hand_keyp], axis=0)
if use_face:
# TODO: Make parameters, 17 is the offset for the eye brows,
# etc. 51 is the total number of FLAME compatible landmarks
face_keypoints = np.array(
person_data['face_keypoints_3d'],
dtype=np.float32).reshape([-1, 4])[17: 17 + 51, :]
contour_keyps = np.array(
[], dtype=body_keypoints.dtype).reshape(0, 4)
if use_face_contour:
contour_keyps = np.array(
person_data['face_keypoints_3d'],
dtype=np.float32).reshape([-1, 4])[:17, :]
body_keypoints = np.concatenate(
[body_keypoints, face_keypoints, contour_keyps], axis=0)
keypoints.append(body_keypoints)
except:
keypoints = None
if 'gender_pd' in person_data:
gender_pd.append(person_data['gender_pd'])
if 'gender_gt' in person_data:
gender_gt.append(person_data['gender_gt'])
return Keypoints(keypoints=keypoints, gender_pd=gender_pd,
gender_gt=gender_gt)
def smpl_to_annotation(model_type='smpl', use_hands=False, use_face=False,
use_face_contour=False, pose_format='coco17'):
if pose_format == 'halpe':
if model_type == 'smplhalpe':
# Halpe to SMPL
return np.array([0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25],
dtype=np.int32)
else:
raise ValueError('Unknown model type: {}'.format(model_type))
class FittingData(Dataset):
NUM_BODY_JOINTS = 17
NUM_HAND_JOINTS = 20
def __init__(self, data_folder, img_folder='images',
keyp_folder='keypoints',
use_hands=False,
use_face=False,
dtype=torch.float32,
model_type='smplx',
joints_to_ign=None,
use_face_contour=False,
pose_format='coco17',
use_3d=False,
use_hip=True,
frames=1,
num_people=1,
**kwargs):
super(FittingData, self).__init__()
self.use_hands = use_hands
self.use_face = use_face
self.model_type = model_type
self.dtype = dtype
self.use_3d = use_3d
self.use_hip = use_hip
self.joints_to_ign = joints_to_ign
self.use_face_contour = use_face_contour
self.pose_format = pose_format
if self.pose_format == 'halpe':
self.NUM_BODY_JOINTS = 26
self.num_joints = (self.NUM_BODY_JOINTS +
2 * self.NUM_HAND_JOINTS * use_hands)
self.data_folder = data_folder
self.img_folder = osp.join(data_folder, img_folder)
self.keyp_folder = osp.join(data_folder, keyp_folder)
img_serials = sorted(os.listdir(self.img_folder))
self.img_paths = []
for i_s in img_serials:
i_s_dir = osp.join(self.img_folder, i_s)
img_cameras = sorted(os.listdir(i_s_dir))
this_serials = []
for i_cam in img_cameras:
i_c_dir = osp.join(i_s_dir, i_cam)
cam_imgs = [osp.join(i_s, i_cam, img_fn)
for img_fn in os.listdir(i_c_dir)
if img_fn.endswith('.png') or
img_fn.endswith('.jpg') and
not img_fn.startswith('.')]
cam_imgs = sorted(cam_imgs)
this_serials.append(cam_imgs)
self.img_paths.append(this_serials)
self.cnt = 0
self.serial_cnt = 0
self.max_frames = frames
self.min_frames = 13
self.num_people = num_people
# if len(cam_imgs) < frames:
# self.frames = len(cam_imgs)
# else:
self.frames = frames
def get_model2data(self):
# Map SMPL to Halpe
return np.array([0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25],
dtype=np.int32)
def get_left_shoulder(self):
return 2
def get_right_shoulder(self):
return 5
def get_joint_weights(self):
# The weights for the joint terms in the optimization
optim_weights = np.ones(self.num_joints + 2 * self.use_hands +
self.use_face * 51 +
17 * self.use_face_contour,
dtype=np.float32)
# Neck, Left and right hip
# These joints are ignored because SMPL has no neck joint and the
# annotation of the hips is ambiguous.
# if self.joints_to_ign is not None and -1 not in self.joints_to_ign:
# optim_weights[self.joints_to_ign] = 0.
# return torch.tensor(optim_weights, dtype=self.dtype)
if (self.pose_format != 'lsp14' and self.pose_format != 'halpe') or not self.use_hip:
optim_weights[11] = 0.
optim_weights[12] = 0.
return torch.tensor(optim_weights, dtype=self.dtype)
def __len__(self):
return len(self.img_paths)
def __getitem__(self, idx):
img_path = self.img_paths[idx]
return self.read_item(img_path)
def read_item(self, img_paths):
"""Load keypoints according to img name"""
keypoints = []
total_flags = []
count = 0
for imgs in img_paths:
cam_keps = []
cam_flag = []
for img in imgs:
if platform.system() == 'Windows':
seq_name, cam_name, f_name = img.split('\\')
else:
seq_name, cam_name, f_name = img.split('/')
index = f_name.split('.')[0]
keypoint_fn = osp.join(self.keyp_folder, seq_name, cam_name, '%s_keypoints.json' %index)
keypoints_, flags, valid = read_keypoints(keypoint_fn, self.num_people, self.NUM_BODY_JOINTS)
count += valid
cam_flag.append(flags)
cam_keps.append(keypoints_)
keypoints.append(cam_keps)
total_flags.append(cam_flag)
total_flags = np.array(total_flags, dtype=np.int)
total_flags = np.max(total_flags, axis=0)
camparam = os.path.join(self.data_folder, 'camparams', seq_name, 'camparams.txt')
output_dict = { 'camparam': camparam,
'img_path': img_paths,
'keypoints': keypoints,
'flags':total_flags,
'count':count}
return output_dict
def __iter__(self):
return self
def __next__(self):
return self.next()
def next(self):
if self.serial_cnt >= len(self.img_paths):
raise StopIteration
img_path = self.img_paths[self.serial_cnt]
img_paths = []
for cam in img_path:
if self.cnt+self.max_frames > len(cam):
if len(cam) - self.cnt < self.min_frames:
img_paths.append(cam[-self.min_frames:])
else:
img_paths.append(cam[self.cnt:])
else:
img_paths.append(cam[self.cnt:self.cnt+self.max_frames]) #
self.frames = len(img_paths[0])
if self.cnt + self.max_frames >= len(cam):
self.cnt = 0
self.serial_cnt += 1
else:
self.cnt += self.frames
return self.read_item(img_paths)
|
[
"os.path.exists",
"collections.namedtuple",
"os.listdir",
"numpy.ones",
"os.path.join",
"numpy.max",
"numpy.array",
"numpy.zeros",
"torch.tensor",
"platform.system",
"numpy.concatenate",
"json.load"
] |
[((452, 516), 'collections.namedtuple', 'namedtuple', (['"""Keypoints"""', "['keypoints', 'gender_gt', 'gender_pd']"], {}), "('Keypoints', ['keypoints', 'gender_gt', 'gender_pd'])\n", (462, 516), False, 'from collections import namedtuple\n'), ((676, 703), 'os.path.exists', 'os.path.exists', (['keypoint_fn'], {}), '(keypoint_fn)\n', (690, 703), False, 'import os\n'), ((807, 830), 'numpy.zeros', 'np.zeros', (['(num_people,)'], {}), '((num_people,))\n', (815, 830), True, 'import numpy as np\n'), ((949, 973), 'json.load', 'json.load', (['keypoint_file'], {}), '(keypoint_file)\n', (958, 973), False, 'import json\n'), ((1782, 1806), 'json.load', 'json.load', (['keypoint_file'], {}), '(keypoint_file)\n', (1791, 1806), False, 'import json\n'), ((5495, 5528), 'os.path.join', 'osp.join', (['data_folder', 'img_folder'], {}), '(data_folder, img_folder)\n', (5503, 5528), True, 'import os.path as osp\n'), ((5556, 5590), 'os.path.join', 'osp.join', (['data_folder', 'keyp_folder'], {}), '(data_folder, keyp_folder)\n', (5564, 5590), True, 'import os.path as osp\n'), ((6710, 6834), 'numpy.array', 'np.array', (['[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, \n 21, 22, 23, 24, 25]'], {'dtype': 'np.int32'}), '([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18,\n 19, 20, 21, 22, 23, 24, 25], dtype=np.int32)\n', (6718, 6834), True, 'import numpy as np\n'), ((7057, 7174), 'numpy.ones', 'np.ones', (['(self.num_joints + 2 * self.use_hands + self.use_face * 51 + 17 * self.\n use_face_contour)'], {'dtype': 'np.float32'}), '(self.num_joints + 2 * self.use_hands + self.use_face * 51 + 17 *\n self.use_face_contour, dtype=np.float32)\n', (7064, 7174), True, 'import numpy as np\n'), ((7806, 7851), 'torch.tensor', 'torch.tensor', (['optim_weights'], {'dtype': 'self.dtype'}), '(optim_weights, dtype=self.dtype)\n', (7818, 7851), False, 'import torch\n'), ((8976, 9011), 'numpy.array', 'np.array', (['total_flags'], {'dtype': 'np.int'}), '(total_flags, dtype=np.int)\n', (8984, 9011), True, 'import numpy as np\n'), ((9034, 9061), 'numpy.max', 'np.max', (['total_flags'], {'axis': '(0)'}), '(total_flags, axis=0)\n', (9040, 9061), True, 'import numpy as np\n'), ((9082, 9152), 'os.path.join', 'os.path.join', (['self.data_folder', '"""camparams"""', 'seq_name', '"""camparams.txt"""'], {}), "(self.data_folder, 'camparams', seq_name, 'camparams.txt')\n", (9094, 9152), False, 'import os\n'), ((1167, 1209), 'numpy.zeros', 'np.zeros', (['(num_joint, 3)'], {'dtype': 'np.float32'}), '((num_joint, 3), dtype=np.float32)\n', (1175, 1209), True, 'import numpy as np\n'), ((1316, 1376), 'numpy.array', 'np.array', (["person_data['pose_keypoints_2d']"], {'dtype': 'np.float32'}), "(person_data['pose_keypoints_2d'], dtype=np.float32)\n", (1324, 1376), True, 'import numpy as np\n'), ((1963, 2023), 'numpy.array', 'np.array', (["person_data['pose_keypoints_3d']"], {'dtype': 'np.float32'}), "(person_data['pose_keypoints_3d'], dtype=np.float32)\n", (1971, 2023), True, 'import numpy as np\n'), ((4028, 4152), 'numpy.array', 'np.array', (['[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, \n 21, 22, 23, 24, 25]'], {'dtype': 'np.int32'}), '([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18,\n 19, 20, 21, 22, 23, 24, 25], dtype=np.int32)\n', (4036, 4152), True, 'import numpy as np\n'), ((5621, 5648), 'os.listdir', 'os.listdir', (['self.img_folder'], {}), '(self.img_folder)\n', (5631, 5648), False, 'import os\n'), ((5732, 5762), 'os.path.join', 'osp.join', (['self.img_folder', 'i_s'], {}), '(self.img_folder, i_s)\n', (5740, 5762), True, 'import os.path as osp\n'), ((726, 750), 'numpy.zeros', 'np.zeros', (['(num_joint, 3)'], {}), '((num_joint, 3))\n', (734, 750), True, 'import numpy as np\n'), ((2497, 2570), 'numpy.concatenate', 'np.concatenate', (['[body_keypoints, left_hand_keyp, right_hand_keyp]'], {'axis': '(0)'}), '([body_keypoints, left_hand_keyp, right_hand_keyp], axis=0)\n', (2511, 2570), True, 'import numpy as np\n'), ((3289, 3360), 'numpy.concatenate', 'np.concatenate', (['[body_keypoints, face_keypoints, contour_keyps]'], {'axis': '(0)'}), '([body_keypoints, face_keypoints, contour_keyps], axis=0)\n', (3303, 3360), True, 'import numpy as np\n'), ((5796, 5815), 'os.listdir', 'os.listdir', (['i_s_dir'], {}), '(i_s_dir)\n', (5806, 5815), False, 'import os\n'), ((5911, 5935), 'os.path.join', 'osp.join', (['i_s_dir', 'i_cam'], {}), '(i_s_dir, i_cam)\n', (5919, 5935), True, 'import os.path as osp\n'), ((8572, 8647), 'os.path.join', 'osp.join', (['self.keyp_folder', 'seq_name', 'cam_name', "('%s_keypoints.json' % index)"], {}), "(self.keyp_folder, seq_name, cam_name, '%s_keypoints.json' % index)\n", (8580, 8647), True, 'import os.path as osp\n'), ((5964, 5992), 'os.path.join', 'osp.join', (['i_s', 'i_cam', 'img_fn'], {}), '(i_s, i_cam, img_fn)\n', (5972, 5992), True, 'import os.path as osp\n'), ((8314, 8331), 'platform.system', 'platform.system', ([], {}), '()\n', (8329, 8331), False, 'import platform\n'), ((2180, 2245), 'numpy.array', 'np.array', (["person_data['hand_left_keypoints_3d']"], {'dtype': 'np.float32'}), "(person_data['hand_left_keypoints_3d'], dtype=np.float32)\n", (2188, 2245), True, 'import numpy as np\n'), ((2338, 2404), 'numpy.array', 'np.array', (["person_data['hand_right_keypoints_3d']"], {'dtype': 'np.float32'}), "(person_data['hand_right_keypoints_3d'], dtype=np.float32)\n", (2346, 2404), True, 'import numpy as np\n'), ((2971, 3011), 'numpy.array', 'np.array', (['[]'], {'dtype': 'body_keypoints.dtype'}), '([], dtype=body_keypoints.dtype)\n', (2979, 3011), True, 'import numpy as np\n'), ((6035, 6054), 'os.listdir', 'os.listdir', (['i_c_dir'], {}), '(i_c_dir)\n', (6045, 6054), False, 'import os\n'), ((2803, 2863), 'numpy.array', 'np.array', (["person_data['face_keypoints_3d']"], {'dtype': 'np.float32'}), "(person_data['face_keypoints_3d'], dtype=np.float32)\n", (2811, 2863), True, 'import numpy as np\n'), ((3120, 3180), 'numpy.array', 'np.array', (["person_data['face_keypoints_3d']"], {'dtype': 'np.float32'}), "(person_data['face_keypoints_3d'], dtype=np.float32)\n", (3128, 3180), True, 'import numpy as np\n')]
|
import numpy as np
import matplotlib.pyplot as plt
import math
def log(list_name):
for i in range(len(list_name)):
list_name[i] = math.log10(list_name[i])
print(list_name[i])
return list_name
size = 4
x = np.arange(size)
video_file = [11132, 21164, 34452, 45208] # 每帧视频文件大小（byte）
video_file = log(video_file)
data_to_cloud = [127, 248, 365, 488] # 每帧所有edge上传的文件大小（byte）（2,2,3,4个摄像头）
data_to_cloud = log(data_to_cloud)
total_width, n = 0.8, 3
width = total_width / n
x = x - (total_width - width) / 2
plt.xlabel('Total Camera Numbers', fontsize=20)
plt.ylabel('Communication Cost (lg(Byte))', fontsize=20)
plt.bar(x-0.45*width, video_file, fc='#036564', width=0.75*width, label='Input Data to Cloud (Cloud)')
# plt.bar(x-0.45*width, data_to_cam, fc='#033649', width=0.75*width, bottom=video_file, label='Feedback (Cloud)')
plt.bar(x+0.45*width, data_to_cloud, fc='#764D39', width=0.75*width, label='Input Data to Cloud (EATP)')
# plt.bar(x+0.45*width, data_to_cam, fc='#250807', width=0.75*width, bottom=data_to_cloud, label='Feedback (EaOT)')
plt.xticks(x, (2, 4, 6, 8), fontsize=18)
plt.yticks(fontsize=18)
plt.legend(loc='center', bbox_to_anchor=(0.62, 0.11), fontsize=17)
plt.show()
|
[
"matplotlib.pyplot.xticks",
"matplotlib.pyplot.ylabel",
"matplotlib.pyplot.legend",
"matplotlib.pyplot.xlabel",
"matplotlib.pyplot.bar",
"matplotlib.pyplot.yticks",
"math.log10",
"numpy.arange",
"matplotlib.pyplot.show"
] |
[((232, 247), 'numpy.arange', 'np.arange', (['size'], {}), '(size)\n', (241, 247), True, 'import numpy as np\n'), ((533, 580), 'matplotlib.pyplot.xlabel', 'plt.xlabel', (['"""Total Camera Numbers"""'], {'fontsize': '(20)'}), "('Total Camera Numbers', fontsize=20)\n", (543, 580), True, 'import matplotlib.pyplot as plt\n'), ((581, 637), 'matplotlib.pyplot.ylabel', 'plt.ylabel', (['"""Communication Cost (lg(Byte))"""'], {'fontsize': '(20)'}), "('Communication Cost (lg(Byte))', fontsize=20)\n", (591, 637), True, 'import matplotlib.pyplot as plt\n'), ((638, 750), 'matplotlib.pyplot.bar', 'plt.bar', (['(x - 0.45 * width)', 'video_file'], {'fc': '"""#036564"""', 'width': '(0.75 * width)', 'label': '"""Input Data to Cloud (Cloud)"""'}), "(x - 0.45 * width, video_file, fc='#036564', width=0.75 * width,\n label='Input Data to Cloud (Cloud)')\n", (645, 750), True, 'import matplotlib.pyplot as plt\n'), ((855, 969), 'matplotlib.pyplot.bar', 'plt.bar', (['(x + 0.45 * width)', 'data_to_cloud'], {'fc': '"""#764D39"""', 'width': '(0.75 * width)', 'label': '"""Input Data to Cloud (EATP)"""'}), "(x + 0.45 * width, data_to_cloud, fc='#764D39', width=0.75 * width,\n label='Input Data to Cloud (EATP)')\n", (862, 969), True, 'import matplotlib.pyplot as plt\n'), ((1076, 1116), 'matplotlib.pyplot.xticks', 'plt.xticks', (['x', '(2, 4, 6, 8)'], {'fontsize': '(18)'}), '(x, (2, 4, 6, 8), fontsize=18)\n', (1086, 1116), True, 'import matplotlib.pyplot as plt\n'), ((1117, 1140), 'matplotlib.pyplot.yticks', 'plt.yticks', ([], {'fontsize': '(18)'}), '(fontsize=18)\n', (1127, 1140), True, 'import matplotlib.pyplot as plt\n'), ((1141, 1207), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': '"""center"""', 'bbox_to_anchor': '(0.62, 0.11)', 'fontsize': '(17)'}), "(loc='center', bbox_to_anchor=(0.62, 0.11), fontsize=17)\n", (1151, 1207), True, 'import matplotlib.pyplot as plt\n'), ((1208, 1218), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (1216, 1218), True, 'import matplotlib.pyplot as plt\n'), ((144, 168), 'math.log10', 'math.log10', (['list_name[i]'], {}), '(list_name[i])\n', (154, 168), False, 'import math\n')]
|
import numpy
from chainer import cuda
import chainer.serializers as S
import chainer.links as L
from nltk.corpus import stopwords
from context_models import CbowContext, BiLstmContext
from defs import IN_TO_OUT_UNITS_RATIO, NEGATIVE_SAMPLING_NUM
class ModelReader(object):
'''
Reads a pre-trained model using a config file
'''
def __init__(self, config_file):
self.gpu = -1 # todo support gpu
print('Reading config file: ' + config_file)
params = self.read_config_file(config_file)
print('Config: ', params)
self.w, self.word2index, self.index2word, self.model = self.read_model(params)
def read_config_file(self, filename):
params = {}
config_path = filename[:filename.rfind('/')+1]
params['config_path'] = config_path
with open(filename, 'r') as f:
for line in f:
if not line.startswith('#'):
[param, val] = line.strip().split()
params[param] = val
return params
def read_model(self, params, train=False):
if 'model_type' in params:
model_type = params['model_type']
else:
model_type = 'lstm_context'
if model_type == 'lstm_context':
return self.read_lstm_model(params, train)
elif model_type == 'bow_context':
return self.read_bow_model(params)
else:
raise Exception("Unknown model type: " + model_type)
def read_lstm_model(self, params, train):
assert train == False # reading a model to continue training is currently not supported
words_file = params['config_path'] + params['words_file']
model_file = params['config_path'] + params['model_file']
unit = int(params['unit'])
deep = (params['deep'] == 'yes')
drop_ratio = float(params['drop_ratio'])
#read and normalize target word embeddings
w, word2index, index2word = self.read_words(words_file)
s = numpy.sqrt((w * w).sum(1))
s[s==0.] = 1.
w /= s.reshape((s.shape[0], 1)) # normalize
context_word_units = unit
lstm_hidden_units = IN_TO_OUT_UNITS_RATIO*unit
target_word_units = IN_TO_OUT_UNITS_RATIO*unit
cs = [1 for _ in range(len(word2index))] # dummy word counts - not used for eval
loss_func = L.NegativeSampling(target_word_units, cs, NEGATIVE_SAMPLING_NUM) # dummy loss func - not used for eval
model = BiLstmContext(deep, self.gpu, word2index, context_word_units, lstm_hidden_units, target_word_units, loss_func, train, drop_ratio)
S.load_npz(model_file, model,strict=False)
return w, word2index, index2word, model
def read_bow_model(self, params):
words_file = params['config_path'] + params['words_file']
contexts_file = params['config_path'] + params['contexts_file'] if 'contexts_file' in params else None
window_size = int(params['window_size'])
use_stopwords = params['stopwords']
if 'word_counts_file' in params:
word_counts_file = params['config_path'] + params['word_counts_file']
else:
word_counts_file = None
if use_stopwords == 'yes':
stop = set(stopwords.words('english') + ['.',',','(',')','[',']',':','"',"'","'s","-",';','?','!','|','%','/','\\'])
else:
stop = set()
word_counts = self.read_word_counts(word_counts_file) if word_counts_file is not None else None
# read and normalize target words embeddings
w, word2index, index2word = self.read_words(words_file)
s = numpy.sqrt((w * w).sum(1))
s[s==0.] = 1.
w /= s.reshape((s.shape[0], 1)) # normalize
# read and normalize context words embeddings (if using different embeddings for context words)
if contexts_file is not None:
c, _, _ = self.read_words(words_file) # assuming words and contexts vocabs are identical
s = numpy.sqrt((c * c).sum(1))
s[s==0.] = 1.
c /= s.reshape((s.shape[0], 1)) # normalize
else:
c = None
model = CbowContext(w, c, word2index, stop, window_size, word_counts)
return w, word2index, index2word, model
def read_words(self, filename):
with open(filename, 'r') as f:
ss = f.readline().split()
n_vocab, n_units = int(ss[0]), int(ss[1])
word2index = {}
index2word = []
w = numpy.empty((n_vocab, n_units), dtype=numpy.float32)
for i, line in enumerate(f):
ss = line.split()
assert len(ss) == n_units + 1
word = ss[0]
word2index[word] = i
index2word.append(word)
w[i] = numpy.array([float(s) for s in ss[1:]], dtype=numpy.float32)
return w, word2index, index2word
def read_word_counts(self, filename):
counts = {}
with open(filename) as f:
for line in f:
if len(line) > 0:
tokens = line.split('\t')
word = tokens[0].strip()
count = int(tokens[1].strip())
counts[word] = count
return counts
|
[
"nltk.corpus.stopwords.words",
"context_models.CbowContext",
"numpy.empty",
"context_models.BiLstmContext",
"chainer.links.NegativeSampling",
"chainer.serializers.load_npz"
] |
[((2480, 2544), 'chainer.links.NegativeSampling', 'L.NegativeSampling', (['target_word_units', 'cs', 'NEGATIVE_SAMPLING_NUM'], {}), '(target_word_units, cs, NEGATIVE_SAMPLING_NUM)\n', (2498, 2544), True, 'import chainer.links as L\n'), ((2608, 2741), 'context_models.BiLstmContext', 'BiLstmContext', (['deep', 'self.gpu', 'word2index', 'context_word_units', 'lstm_hidden_units', 'target_word_units', 'loss_func', 'train', 'drop_ratio'], {}), '(deep, self.gpu, word2index, context_word_units,\n lstm_hidden_units, target_word_units, loss_func, train, drop_ratio)\n', (2621, 2741), False, 'from context_models import CbowContext, BiLstmContext\n'), ((2746, 2789), 'chainer.serializers.load_npz', 'S.load_npz', (['model_file', 'model'], {'strict': '(False)'}), '(model_file, model, strict=False)\n', (2756, 2789), True, 'import chainer.serializers as S\n'), ((4353, 4414), 'context_models.CbowContext', 'CbowContext', (['w', 'c', 'word2index', 'stop', 'window_size', 'word_counts'], {}), '(w, c, word2index, stop, window_size, word_counts)\n', (4364, 4414), False, 'from context_models import CbowContext, BiLstmContext\n'), ((4713, 4765), 'numpy.empty', 'numpy.empty', (['(n_vocab, n_units)'], {'dtype': 'numpy.float32'}), '((n_vocab, n_units), dtype=numpy.float32)\n', (4724, 4765), False, 'import numpy\n'), ((3413, 3439), 'nltk.corpus.stopwords.words', 'stopwords.words', (['"""english"""'], {}), "('english')\n", (3428, 3439), False, 'from nltk.corpus import stopwords\n')]
|
# -*- coding: utf-8 -*-
# vim: tabstop=4 shiftwidth=4 softtabstop=4
#
# Copyright (C) 2012-2016 GEM Foundation
#
# OpenQuake is free software: you can redistribute it and/or modify it
# under the terms of the GNU Affero General Public License as published
# by the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# OpenQuake is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU Affero General Public License for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with OpenQuake. If not, see <http://www.gnu.org/licenses/>.
"""
Module exports :class:`ZhaoEtAl2006Asc`, :class:`ZhaoEtAl2006SInter`,
:class:`ZhaoEtAl2006SSlab`, :class:`ZhaoEtAl2006SInterNSHMP2008` and
:class:`ZhaoEtAl2006SSlabNSHMP2014`
"""
from __future__ import division
import numpy as np
# standard acceleration of gravity in m/s**2
from scipy.constants import g
import copy
from openquake.hazardlib.gsim.base import GMPE, CoeffsTable
from openquake.hazardlib import const
from openquake.hazardlib.imt import PGA, PGV, SA
class ZhaoEtAl2006Asc(GMPE):
"""
Implements GMPE developed by <NAME> et al. and published as
"Attenuation Relations of Strong Ground Motion in Japan Using Site
Classification Based on Predominant Period" (2006, Bulletin of the
Seismological Society of America, Volume 96, No. 3, pages 898-913).
This class implements the equations for 'Active Shallow Crust'
(that's why the class name ends with 'Asc').
"""
#: Supported tectonic region type is active shallow crust, this means
#: that factors SI, SS and SSL are assumed 0 in equation 1, p. 901.
DEFINED_FOR_TECTONIC_REGION_TYPE = const.TRT.ACTIVE_SHALLOW_CRUST
#: Supported intensity measure types are spectral acceleration,
#: and peak ground acceleration, see paragraph 'Development of Base Model'
#: p. 901.
DEFINED_FOR_INTENSITY_MEASURE_TYPES = set([
PGA,
SA
])
#: Supported intensity measure component is geometric mean
#: of two horizontal components :
#: attr:`~openquake.hazardlib.const.IMC.AVERAGE_HORIZONTAL`, see paragraph
#: 'Development of Base Model', p. 901.
DEFINED_FOR_INTENSITY_MEASURE_COMPONENT = const.IMC.AVERAGE_HORIZONTAL
#: Supported standard deviation types are inter-event, intra-event
#: and total, see equation 3, p. 902.
DEFINED_FOR_STANDARD_DEVIATION_TYPES = set([
const.StdDev.TOTAL,
const.StdDev.INTER_EVENT,
const.StdDev.INTRA_EVENT
])
#: Required site parameters is Vs30.
#: See table 2, p. 901.
REQUIRES_SITES_PARAMETERS = set(('vs30', ))
#: Required rupture parameters are magnitude, rake, and focal depth.
#: See paragraph 'Development of Base Model', p. 901.
REQUIRES_RUPTURE_PARAMETERS = set(('mag', 'rake', 'hypo_depth'))
#: Required distance measure is Rrup.
#: See paragraph 'Development of Base Model', p. 902.
REQUIRES_DISTANCES = set(('rrup', ))
def get_mean_and_stddevs(self, sites, rup, dists, imt, stddev_types):
"""
See :meth:`superclass method
<.base.GroundShakingIntensityModel.get_mean_and_stddevs>`
for spec of input and result values.
"""
# extracting dictionary of coefficients specific to required
# intensity measure type.
C = self.COEFFS_ASC[imt]
# mean value as given by equation 1, p. 901, without considering the
# interface and intraslab terms (that is SI, SS, SSL = 0) and the
# inter and intra event terms, plus the magnitude-squared term
# correction factor (equation 5 p. 909).
mean = self._compute_magnitude_term(C, rup.mag) +\
self._compute_distance_term(C, rup.mag, dists.rrup) +\
self._compute_focal_depth_term(C, rup.hypo_depth) +\
self._compute_faulting_style_term(C, rup.rake) +\
self._compute_site_class_term(C, sites.vs30) +\
self._compute_magnitude_squared_term(P=0.0, M=6.3, Q=C['QC'],
W=C['WC'], mag=rup.mag)
# convert from cm/s**2 to g
mean = np.log(np.exp(mean) * 1e-2 / g)
stddevs = self._get_stddevs(C['sigma'], C['tauC'], stddev_types,
num_sites=len(sites.vs30))
return mean, stddevs
def _get_stddevs(self, sigma, tau, stddev_types, num_sites):
"""
Return standard deviations as defined in equation 3 p. 902.
"""
stddevs = []
for stddev_type in stddev_types:
assert stddev_type in self.DEFINED_FOR_STANDARD_DEVIATION_TYPES
if stddev_type == const.StdDev.TOTAL:
sigma_t = np.sqrt(sigma ** 2 + tau ** 2)
stddevs.append(sigma_t + np.zeros(num_sites))
elif stddev_type == const.StdDev.INTRA_EVENT:
stddevs.append(sigma + np.zeros(num_sites))
elif stddev_type == const.StdDev.INTER_EVENT:
stddevs.append(tau + np.zeros(num_sites))
return stddevs
def _compute_magnitude_term(self, C, mag):
"""
Compute first term in equation 1, p. 901.
"""
return C['a'] * mag
def _compute_distance_term(self, C, mag, rrup):
"""
Compute second and third terms in equation 1, p. 901.
"""
term1 = C['b'] * rrup
term2 = - np.log(rrup + C['c'] * np.exp(C['d'] * mag))
return term1 + term2
def _compute_focal_depth_term(self, C, hypo_depth):
"""
Compute fourth term in equation 1, p. 901.
"""
# p. 901. "(i.e, depth is capped at 125 km)".
focal_depth = hypo_depth
if focal_depth > 125.0:
focal_depth = 125.0
# p. 902. "We used the value of 15 km for the
# depth coefficient hc ...".
hc = 15.0
# p. 901. "When h is larger than hc, the depth terms takes
# effect ...". The next sentence specifies h>=hc.
return float(focal_depth >= hc) * C['e'] * (focal_depth - hc)
def _compute_faulting_style_term(self, C, rake):
"""
Compute fifth term in equation 1, p. 901.
"""
# p. 900. "The differentiation in focal mechanism was
# based on a rake angle criterion, with a rake of +/- 45
# as demarcation between dip-slip and strike-slip."
return float(rake > 45.0 and rake < 135.0) * C['FR']
def _compute_site_class_term(self, C, vs30):
"""
Compute nine-th term in equation 1, p. 901.
"""
# map vs30 value to site class, see table 2, p. 901.
site_term = np.zeros(len(vs30))
# hard rock
site_term[vs30 > 1100.0] = C['CH']
# rock
site_term[(vs30 > 600) & (vs30 <= 1100)] = C['C1']
# hard soil
site_term[(vs30 > 300) & (vs30 <= 600)] = C['C2']
# medium soil
site_term[(vs30 > 200) & (vs30 <= 300)] = C['C3']
# soft soil
site_term[vs30 <= 200] = C['C4']
return site_term
def _compute_magnitude_squared_term(self, P, M, Q, W, mag):
"""
Compute magnitude squared term, equation 5, p. 909.
"""
return P * (mag - M) + Q * (mag - M) ** 2 + W
#: Coefficient table obtained by joining table 4 (except columns for
#: SI, SS, SSL), table 5 (both at p. 903) and table 6 (only columns for
#: QC WC TauC), p. 907.
COEFFS_ASC = CoeffsTable(sa_damping=5, table="""\
IMT a b c d e FR CH C1 C2 C3 C4 sigma QC WC tauC
pga 1.101 -0.00564 0.0055 1.080 0.01412 0.251 0.293 1.111 1.344 1.355 1.420 0.604 0.0 0.0 0.303
0.05 1.076 -0.00671 0.0075 1.060 0.01463 0.251 0.939 1.684 1.793 1.747 1.814 0.640 0.0 0.0 0.326
0.10 1.118 -0.00787 0.0090 1.083 0.01423 0.240 1.499 2.061 2.135 2.031 2.082 0.694 0.0 0.0 0.342
0.15 1.134 -0.00722 0.0100 1.053 0.01509 0.251 1.462 1.916 2.168 2.052 2.113 0.702 0.0 0.0 0.331
0.20 1.147 -0.00659 0.0120 1.014 0.01462 0.260 1.280 1.669 2.085 2.001 2.030 0.692 0.0 0.0 0.312
0.25 1.149 -0.00590 0.0140 0.966 0.01459 0.269 1.121 1.468 1.942 1.941 1.937 0.682 0.0 0.0 0.298
0.30 1.163 -0.00520 0.0150 0.934 0.01458 0.259 0.852 1.172 1.683 1.808 1.770 0.670 0.0 0.0 0.300
0.40 1.200 -0.00422 0.0100 0.959 0.01257 0.248 0.365 0.655 1.127 1.482 1.397 0.659 0.0 0.0 0.346
0.50 1.250 -0.00338 0.0060 1.008 0.01114 0.247 -0.207 0.071 0.515 0.934 0.955 0.653 -0.0126 0.0116 0.338
0.60 1.293 -0.00282 0.0030 1.088 0.01019 0.233 -0.705 -0.429 -0.003 0.394 0.559 0.653 -0.0329 0.0202 0.349
0.70 1.336 -0.00258 0.0025 1.084 0.00979 0.220 -1.144 -0.866 -0.449 -0.111 0.188 0.652 -0.0501 0.0274 0.351
0.80 1.386 -0.00242 0.0022 1.088 0.00944 0.232 -1.609 -1.325 -0.928 -0.620 -0.246 0.647 -0.0650 0.0336 0.356
0.90 1.433 -0.00232 0.0020 1.109 0.00972 0.220 -2.023 -1.732 -1.349 -1.066 -0.643 0.653 -0.0781 0.0391 0.348
1.00 1.479 -0.00220 0.0020 1.115 0.01005 0.211 -2.451 -2.152 -1.776 -1.523 -1.084 0.657 -0.0899 0.0440 0.338
1.25 1.551 -0.00207 0.0020 1.083 0.01003 0.251 -3.243 -2.923 -2.542 -2.327 -1.936 0.660 -0.1148 0.0545 0.313
1.50 1.621 -0.00224 0.0020 1.091 0.00928 0.248 -3.888 -3.548 -3.169 -2.979 -2.661 0.664 -0.1351 0.0630 0.306
2.00 1.694 -0.00201 0.0025 1.055 0.00833 0.263 -4.783 -4.410 -4.039 -3.871 -3.640 0.669 -0.1672 0.0764 0.283
2.50 1.748 -0.00187 0.0028 1.052 0.00776 0.262 -5.444 -5.049 -4.698 -4.496 -4.341 0.671 -0.1921 0.0869 0.287
3.00 1.759 -0.00147 0.0032 1.025 0.00644 0.307 -5.839 -5.431 -5.089 -4.893 -4.758 0.667 -0.2124 0.0954 0.278
4.00 1.826 -0.00195 0.0040 1.044 0.00590 0.353 -6.598 -6.181 -5.882 -5.698 -5.588 0.647 -0.2445 0.1088 0.273
5.00 1.825 -0.00237 0.0050 1.065 0.00510 0.248 -6.752 -6.347 -6.051 -5.873 -5.798 0.643 -0.2694 0.1193 0.275
""")
class ZhaoEtAl2006SInter(ZhaoEtAl2006Asc):
"""
Implements GMPE developed by <NAME> et al and published as
"Attenuation Relations of Strong Ground Motion in Japan Using Site
Classification Based on Predominant Period" (2006, Bulletin of the
Seismological Society of America, Volume 96, No. 3, pages
898-913). This class implements the equations for 'Subduction
Interface' (that's why the class name ends with 'SInter'). This
class extends the
:class:`openquake.hazardlib.gsim.zhao_2006.ZhaoEtAl2006Asc`
because the equation for subduction interface is obtained from the
equation for active shallow crust, by removing the faulting style
term and adding a subduction interface term.
"""
#: Supported tectonic region type is subduction interface, this means
#: that factors FR, SS and SSL are assumed 0 in equation 1, p. 901.
DEFINED_FOR_TECTONIC_REGION_TYPE = const.TRT.SUBDUCTION_INTERFACE
#: Required rupture parameters are magnitude and focal depth.
REQUIRES_RUPTURE_PARAMETERS = set(('mag', 'hypo_depth'))
def get_mean_and_stddevs(self, sites, rup, dists, imt, stddev_types):
"""
See :meth:`superclass method
<.base.GroundShakingIntensityModel.get_mean_and_stddevs>`
for spec of input and result values.
"""
# extracting dictionary of coefficients specific to required
# intensity measure type.
C = self.COEFFS_ASC[imt]
C_SINTER = self.COEFFS_SINTER[imt]
# mean value as given by equation 1, p. 901, without considering the
# faulting style and intraslab terms (that is FR, SS, SSL = 0) and the
# inter and intra event terms, plus the magnitude-squared term
# correction factor (equation 5 p. 909)
mean = self._compute_magnitude_term(C, rup.mag) +\
self._compute_distance_term(C, rup.mag, dists.rrup) +\
self._compute_focal_depth_term(C, rup.hypo_depth) +\
self._compute_site_class_term(C, sites.vs30) + \
self._compute_magnitude_squared_term(P=0.0, M=6.3,
Q=C_SINTER['QI'],
W=C_SINTER['WI'],
mag=rup.mag) +\
C_SINTER['SI']
# convert from cm/s**2 to g
mean = np.log(np.exp(mean) * 1e-2 / g)
stddevs = self._get_stddevs(C['sigma'], C_SINTER['tauI'], stddev_types,
num_sites=len(sites.vs30))
return mean, stddevs
#: Coefficient table containing subduction interface coefficients,
#: taken from table 4, p. 903 (only column SI), and table 6, p. 907
#: (only columns QI, WI, TauI)
COEFFS_SINTER = CoeffsTable(sa_damping=5, table="""\
IMT SI QI WI tauI
pga 0.000 0.0 0.0 0.308
0.05 0.000 0.0 0.0 0.343
0.10 0.000 0.0 0.0 0.403
0.15 0.000 -0.0138 0.0286 0.367
0.20 0.000 -0.0256 0.0352 0.328
0.25 0.000 -0.0348 0.0403 0.289
0.30 0.000 -0.0423 0.0445 0.280
0.40 -0.041 -0.0541 0.0511 0.271
0.50 -0.053 -0.0632 0.0562 0.277
0.60 -0.103 -0.0707 0.0604 0.296
0.70 -0.146 -0.0771 0.0639 0.313
0.80 -0.164 -0.0825 0.0670 0.329
0.90 -0.206 -0.0874 0.0697 0.324
1.00 -0.239 -0.0917 0.0721 0.328
1.25 -0.256 -0.1009 0.0772 0.339
1.50 -0.306 -0.1083 0.0814 0.352
2.00 -0.321 -0.1202 0.0880 0.360
2.50 -0.337 -0.1293 0.0931 0.356
3.00 -0.331 -0.1368 0.0972 0.338
4.00 -0.390 -0.1486 0.1038 0.307
5.00 -0.498 -0.1578 0.1090 0.272
""")
class ZhaoEtAl2006SSlab(ZhaoEtAl2006Asc):
"""
Implements GMPE developed by <NAME> et al and published as
"Attenuation Relations of Strong Ground Motion in Japan Using Site
Classification Based on Predominant Period" (2006, Bulletin of the
Seismological Society of America, Volume 96, No. 3, pages
898-913). This class implements the equations for 'Subduction
Slab'. (that's why the class name ends with 'SSlab'). This class
extends the
:class:`openquake.hazardlib.gsim.zhao_2006.ZhaoEtAl2006Asc`
because the equation for subduction slab is obtained from the
equation for active shallow crust, by removing the faulting style
term and adding subduction slab terms.
"""
#: Supported tectonic region type is subduction interface, this means
#: that factors FR, SS and SSL are assumed 0 in equation 1, p. 901.
DEFINED_FOR_TECTONIC_REGION_TYPE = const.TRT.SUBDUCTION_INTRASLAB
#: Required rupture parameters are magnitude and focal depth.
REQUIRES_RUPTURE_PARAMETERS = set(('mag', 'hypo_depth'))
def get_mean_and_stddevs(self, sites, rup, dists, imt, stddev_types):
"""
See :meth:`superclass method
<.base.GroundShakingIntensityModel.get_mean_and_stddevs>`
for spec of input and result values.
"""
# extracting dictionary of coefficients specific to required
# intensity measure type.
C = self.COEFFS_ASC[imt]
C_SSLAB = self.COEFFS_SSLAB[imt]
# to avoid singularity at 0.0 (in the calculation of the
# slab correction term), replace 0 values with 0.1
d = dists.rrup
d[d == 0.0] = 0.1
# mean value as given by equation 1, p. 901, without considering the
# faulting style and intraslab terms (that is FR, SS, SSL = 0) and the
# inter and intra event terms, plus the magnitude-squared term
# correction factor (equation 5 p. 909)
mean = self._compute_magnitude_term(C, rup.mag) +\
self._compute_distance_term(C, rup.mag, d) +\
self._compute_focal_depth_term(C, rup.hypo_depth) +\
self._compute_site_class_term(C, sites.vs30) +\
self._compute_magnitude_squared_term(P=C_SSLAB['PS'], M=6.5,
Q=C_SSLAB['QS'],
W=C_SSLAB['WS'],
mag=rup.mag) +\
C_SSLAB['SS'] + self._compute_slab_correction_term(C_SSLAB, d)
# convert from cm/s**2 to g
mean = np.log(np.exp(mean) * 1e-2 / g)
stddevs = self._get_stddevs(C['sigma'], C_SSLAB['tauS'], stddev_types,
num_sites=len(sites.vs30))
return mean, stddevs
def _compute_slab_correction_term(self, C, rrup):
"""
Compute path modification term for slab events, that is
the 8-th term in equation 1, p. 901.
"""
slab_term = C['SSL'] * np.log(rrup)
return slab_term
#: Coefficient table containing subduction slab coefficients taken from
#: table 4, p. 903 (only columns for SS and SSL), and table 6, p. 907
#: (only columns for PS, QS, WS, TauS)
COEFFS_SSLAB = CoeffsTable(sa_damping=5, table="""\
IMT SS SSL PS QS WS tauS
pga 2.607 -0.528 0.1392 0.1584 -0.0529 0.321
0.05 2.764 -0.551 0.1636 0.1932 -0.0841 0.378
0.10 2.156 -0.420 0.1690 0.2057 -0.0877 0.420
0.15 2.161 -0.431 0.1669 0.1984 -0.0773 0.372
0.20 1.901 -0.372 0.1631 0.1856 -0.0644 0.324
0.25 1.814 -0.360 0.1588 0.1714 -0.0515 0.294
0.30 2.181 -0.450 0.1544 0.1573 -0.0395 0.284
0.40 2.432 -0.506 0.1460 0.1309 -0.0183 0.278
0.50 2.629 -0.554 0.1381 0.1078 -0.0008 0.272
0.60 2.702 -0.575 0.1307 0.0878 0.0136 0.285
0.70 2.654 -0.572 0.1239 0.0705 0.0254 0.290
0.80 2.480 -0.540 0.1176 0.0556 0.0352 0.299
0.90 2.332 -0.522 0.1116 0.0426 0.0432 0.289
1.00 2.233 -0.509 0.1060 0.0314 0.0498 0.286
1.25 2.029 -0.469 0.0933 0.0093 0.0612 0.277
1.50 1.589 -0.379 0.0821 -0.0062 0.0674 0.282
2.00 0.966 -0.248 0.0628 -0.0235 0.0692 0.300
2.50 0.789 -0.221 0.0465 -0.0287 0.0622 0.292
3.00 1.037 -0.263 0.0322 -0.0261 0.0496 0.274
4.00 0.561 -0.169 0.0083 -0.0065 0.0150 0.281
5.00 0.225 -0.120 -0.0117 0.0246 -0.0268 0.296
""")
class ZhaoEtAl2006SInterNSHMP2008(ZhaoEtAl2006SInter):
"""
Extend :class:`ZhaoEtAl2006SInter` and fix hypocentral depth at 20 km
as defined the by National Seismic Hazard Mapping Project for the 2008 US
hazard model.
The calculation of the total standard deviation is done considering the
inter-event standard deviation as defined in table 5, page 903 of Zhao's
paper.
The class implement the equation as coded in ``subroutine zhao`` in
``hazSUBXnga.f`` Fotran code available at:
http://earthquake.usgs.gov/hazards/products/conterminous/2008/software/
"""
def get_mean_and_stddevs(self, sites, rup, dists, imt, stddev_types):
"""
See :meth:`superclass method
<.base.GroundShakingIntensityModel.get_mean_and_stddevs>`
for spec of input and result values.
Call super class method with hypocentral depth fixed at 20 km
"""
# create new rupture context to avoid changing the original one
new_rup = copy.deepcopy(rup)
new_rup.hypo_depth = 20.
mean, stddevs = super(ZhaoEtAl2006SInterNSHMP2008, self). \
get_mean_and_stddevs(sites, new_rup, dists, imt, stddev_types)
return mean, stddevs
COEFFS_SINTER = CoeffsTable(sa_damping=5, table="""\
IMT SI QI WI tauI
pga 0.000 0.0 0.0 0.3976
0.05 0.000 0.0 0.0 0.4437
0.10 0.000 0.0 0.0 0.4903
0.15 0.000 -0.0138 0.0286 0.4603
0.20 0.000 -0.0256 0.0352 0.4233
0.25 0.000 -0.0348 0.0403 0.3908
0.30 0.000 -0.0423 0.0445 0.3790
0.40 -0.041 -0.0541 0.0511 0.3897
0.50 -0.053 -0.0632 0.0562 0.3890
0.60 -0.103 -0.0707 0.0604 0.4014
0.70 -0.146 -0.0771 0.0639 0.4079
0.80 -0.164 -0.0825 0.0670 0.4183
0.90 -0.206 -0.0874 0.0697 0.4106
1.00 -0.239 -0.0917 0.0721 0.4101
1.25 -0.256 -0.1009 0.0772 0.4021
1.50 -0.306 -0.1083 0.0814 0.4076
2.00 -0.321 -0.1202 0.0880 0.4138
2.50 -0.337 -0.1293 0.0931 0.4108
3.00 -0.331 -0.1368 0.0972 0.3961
4.00 -0.390 -0.1486 0.1038 0.3821
5.00 -0.498 -0.1578 0.1090 0.3766
""")
class ZhaoEtAl2006SSlabNSHMP2014(ZhaoEtAl2006SSlab):
"""
For the 2014 US National Seismic Hazard Maps the magnitude of Zhao et al.
(2006) for the subduction inslab events is capped at magnitude Mw 7.8
"""
def get_mean_and_stddevs(self, sites, rup, dists, imt, stddev_types):
"""
See :meth:`superclass method
<.base.GroundShakingIntensityModel.get_mean_and_stddevs>`
for spec of input and result values.
"""
# extracting dictionary of coefficients specific to required
# intensity measure type.
C = self.COEFFS_ASC[imt]
C_SSLAB = self.COEFFS_SSLAB[imt]
# to avoid singularity at 0.0 (in the calculation of the
# slab correction term), replace 0 values with 0.1
d = dists.rrup
d[d == 0.0] = 0.1
if rup.mag > 7.8:
rup_mag = 7.8
else:
rup_mag = rup.mag
# mean value as given by equation 1, p. 901, without considering the
# faulting style and intraslab terms (that is FR, SS, SSL = 0) and the
# inter and intra event terms, plus the magnitude-squared term
# correction factor (equation 5 p. 909)
mean = self._compute_magnitude_term(C, rup_mag) +\
self._compute_distance_term(C, rup_mag, d) +\
self._compute_focal_depth_term(C, rup.hypo_depth) +\
self._compute_site_class_term(C, sites.vs30) +\
self._compute_magnitude_squared_term(P=C_SSLAB['PS'], M=6.5,
Q=C_SSLAB['QS'],
W=C_SSLAB['WS'],
mag=rup_mag) +\
C_SSLAB['SS'] + self._compute_slab_correction_term(C_SSLAB, d)
# convert from cm/s**2 to g
mean = np.log(np.exp(mean) * 1e-2 / g)
stddevs = self._get_stddevs(C['sigma'], C_SSLAB['tauS'], stddev_types,
num_sites=len(sites.vs30))
return mean, stddevs
# Coefficient table taken from <NAME>'s "White paper on
# Proposed Ground-motion Prediction Equations (GMPEs) for 2015
# National Seismic Hazard Maps" (2012, page 16).
# Values were interpolated to include all listed periods.
# MF is the linear multiplicative factor.
COEFFS_SITE_FACTORS = CoeffsTable(sa_damping=5, table="""\
IMT MF
pga 0.50
pgv 1.00
0.05 0.44
0.10 0.44
0.15 0.53
0.20 0.60
0.25 0.72
0.30 0.81
0.40 1.00
0.50 1.01
0.60 1.02
0.70 1.02
0.80 1.03
0.90 1.04
1.00 1.04
1.25 1.19
1.50 1.31
2.00 1.51
2.50 1.34
3.00 1.21
4.00 1.09
5.00 1.00
""")
class ZhaoEtAl2006SInterCascadia(ZhaoEtAl2006SInter):
"""
Implements the interface GMPE developed by <NAME> et al modified
by the Japan/Cascadia site factors as proposed by <NAME>.
(2012). White paper on proposed ground-motion prediction equations
(GMPEs) for 2015 National Seismic Hazard Maps Final Version,
Nov. 2012, 50 pp. This class extends the
:class:`openquake.hazardlib.gsim.zhao_2006.ZhaoEtAl2006Asc`
because the equation for subduction interface is obtained from the
equation for active shallow crust, by removing the faulting style
term and adding a subduction interface term.
"""
def get_mean_and_stddevs(self, sites, rup, dists, imt, stddev_types):
"""
See :meth:`superclass method
<.base.GroundShakingIntensityModel.get_mean_and_stddevs>`
for spec of input and result values.
"""
# extracting dictionary of coefficients specific to required
# intensity measure type.
C = self.COEFFS_ASC[imt]
C_SINTER = self.COEFFS_SINTER[imt]
C_SF = COEFFS_SITE_FACTORS[imt]
# mean value as given by equation 1, p. 901, without considering the
# faulting style and intraslab terms (that is FR, SS, SSL = 0) and the
# inter and intra event terms, plus the magnitude-squared term
# correction factor (equation 5 p. 909)
mean = self._compute_magnitude_term(C, rup.mag) +\
self._compute_distance_term(C, rup.mag, dists.rrup) +\
self._compute_focal_depth_term(C, rup.hypo_depth) +\
self._compute_site_class_term(C, sites.vs30) + \
self._compute_magnitude_squared_term(P=0.0, M=6.3,
Q=C_SINTER['QI'],
W=C_SINTER['WI'],
mag=rup.mag) +\
C_SINTER['SI']
# multiply by site factor to "convert" Japan values to Cascadia values
# then convert from cm/s**2 to g
mean = np.log((np.exp(mean) * C_SF["MF"]) * 1e-2 / g)
stddevs = self._get_stddevs(C['sigma'], C_SINTER['tauI'], stddev_types,
num_sites=len(sites.vs30))
return mean, stddevs
class ZhaoEtAl2006SSlabCascadia(ZhaoEtAl2006SSlab):
"""
Implements GMPE developed by <NAME> et al modified
by the Japan/Cascadia site factors as proposed by <NAME>.
(2012). White paper on proposed ground-motion prediction equations
(GMPEs) for 2015 National Seismic Hazard Maps Final Version,
Nov. 2012, 50 pp. This class extends the
:class:`openquake.hazardlib.gsim.zhao_2006.ZhaoEtAl2006Asc`
because the equation for subduction slab is obtained from the
equation for active shallow crust, by removing the faulting style
term and adding subduction slab terms.
"""
def get_mean_and_stddevs(self, sites, rup, dists, imt, stddev_types):
"""
See :meth:`superclass method
<.base.GroundShakingIntensityModel.get_mean_and_stddevs>`
for spec of input and result values.
"""
# extracting dictionary of coefficients specific to required
# intensity measure type.
C = self.COEFFS_ASC[imt]
C_SSLAB = self.COEFFS_SSLAB[imt]
C_SF = COEFFS_SITE_FACTORS[imt]
# to avoid singularity at 0.0 (in the calculation of the
# slab correction term), replace 0 values with 0.1
d = dists.rrup
d[d == 0.0] = 0.1
# mean value as given by equation 1, p. 901, without considering the
# faulting style and intraslab terms (that is FR, SS, SSL = 0) and the
# inter and intra event terms, plus the magnitude-squared term
# correction factor (equation 5 p. 909)
mean = self._compute_magnitude_term(C, rup.mag) +\
self._compute_distance_term(C, rup.mag, d) +\
self._compute_focal_depth_term(C, rup.hypo_depth) +\
self._compute_site_class_term(C, sites.vs30) +\
self._compute_magnitude_squared_term(P=C_SSLAB['PS'], M=6.5,
Q=C_SSLAB['QS'],
W=C_SSLAB['WS'],
mag=rup.mag) +\
C_SSLAB['SS'] + self._compute_slab_correction_term(C_SSLAB, d)
# multiply by site factor to "convert" Japan values to Cascadia values
# then convert from cm/s**2 to g
mean = np.log((np.exp(mean) * C_SF["MF"]) * 1e-2 / g)
stddevs = self._get_stddevs(C['sigma'], C_SSLAB['tauS'], stddev_types,
num_sites=len(sites.vs30))
return mean, stddevs
|
[
"numpy.sqrt",
"numpy.log",
"numpy.exp",
"openquake.hazardlib.gsim.base.CoeffsTable",
"numpy.zeros",
"copy.deepcopy"
] |
[((23450, 23869), 'openquake.hazardlib.gsim.base.CoeffsTable', 'CoeffsTable', ([], {'sa_damping': '(5)', 'table': '""" IMT MF\n pga 0.50\n pgv 1.00\n 0.05 0.44\n 0.10 0.44\n 0.15 0.53\n 0.20 0.60\n 0.25 0.72\n 0.30 0.81\n 0.40 1.00\n 0.50 1.01\n 0.60 1.02\n 0.70 1.02\n 0.80 1.03\n 0.90 1.04\n 1.00 1.04\n 1.25 1.19\n 1.50 1.31\n 2.00 1.51\n 2.50 1.34\n 3.00 1.21\n 4.00 1.09\n 5.00 1.00\n """'}), '(sa_damping=5, table=\n """ IMT MF\n pga 0.50\n pgv 1.00\n 0.05 0.44\n 0.10 0.44\n 0.15 0.53\n 0.20 0.60\n 0.25 0.72\n 0.30 0.81\n 0.40 1.00\n 0.50 1.01\n 0.60 1.02\n 0.70 1.02\n 0.80 1.03\n 0.90 1.04\n 1.00 1.04\n 1.25 1.19\n 1.50 1.31\n 2.00 1.51\n 2.50 1.34\n 3.00 1.21\n 4.00 1.09\n 5.00 1.00\n """\n )\n', (23461, 23869), False, 'from openquake.hazardlib.gsim.base import GMPE, CoeffsTable\n'), ((7631, 10389), 'openquake.hazardlib.gsim.base.CoeffsTable', 'CoeffsTable', ([], {'sa_damping': '(5)', 'table': '""" IMT a b c d e FR CH C1 C2 C3 C4 sigma QC WC tauC\n pga 1.101 -0.00564 0.0055 1.080 0.01412 0.251 0.293 1.111 1.344 1.355 1.420 0.604 0.0 0.0 0.303\n 0.05 1.076 -0.00671 0.0075 1.060 0.01463 0.251 0.939 1.684 1.793 1.747 1.814 0.640 0.0 0.0 0.326\n 0.10 1.118 -0.00787 0.0090 1.083 0.01423 0.240 1.499 2.061 2.135 2.031 2.082 0.694 0.0 0.0 0.342\n 0.15 1.134 -0.00722 0.0100 1.053 0.01509 0.251 1.462 1.916 2.168 2.052 2.113 0.702 0.0 0.0 0.331\n 0.20 1.147 -0.00659 0.0120 1.014 0.01462 0.260 1.280 1.669 2.085 2.001 2.030 0.692 0.0 0.0 0.312\n 0.25 1.149 -0.00590 0.0140 0.966 0.01459 0.269 1.121 1.468 1.942 1.941 1.937 0.682 0.0 0.0 0.298\n 0.30 1.163 -0.00520 0.0150 0.934 0.01458 0.259 0.852 1.172 1.683 1.808 1.770 0.670 0.0 0.0 0.300\n 0.40 1.200 -0.00422 0.0100 0.959 0.01257 0.248 0.365 0.655 1.127 1.482 1.397 0.659 0.0 0.0 0.346\n 0.50 1.250 -0.00338 0.0060 1.008 0.01114 0.247 -0.207 0.071 0.515 0.934 0.955 0.653 -0.0126 0.0116 0.338\n 0.60 1.293 -0.00282 0.0030 1.088 0.01019 0.233 -0.705 -0.429 -0.003 0.394 0.559 0.653 -0.0329 0.0202 0.349\n 0.70 1.336 -0.00258 0.0025 1.084 0.00979 0.220 -1.144 -0.866 -0.449 -0.111 0.188 0.652 -0.0501 0.0274 0.351\n 0.80 1.386 -0.00242 0.0022 1.088 0.00944 0.232 -1.609 -1.325 -0.928 -0.620 -0.246 0.647 -0.0650 0.0336 0.356\n 0.90 1.433 -0.00232 0.0020 1.109 0.00972 0.220 -2.023 -1.732 -1.349 -1.066 -0.643 0.653 -0.0781 0.0391 0.348\n 1.00 1.479 -0.00220 0.0020 1.115 0.01005 0.211 -2.451 -2.152 -1.776 -1.523 -1.084 0.657 -0.0899 0.0440 0.338\n 1.25 1.551 -0.00207 0.0020 1.083 0.01003 0.251 -3.243 -2.923 -2.542 -2.327 -1.936 0.660 -0.1148 0.0545 0.313\n 1.50 1.621 -0.00224 0.0020 1.091 0.00928 0.248 -3.888 -3.548 -3.169 -2.979 -2.661 0.664 -0.1351 0.0630 0.306\n 2.00 1.694 -0.00201 0.0025 1.055 0.00833 0.263 -4.783 -4.410 -4.039 -3.871 -3.640 0.669 -0.1672 0.0764 0.283\n 2.50 1.748 -0.00187 0.0028 1.052 0.00776 0.262 -5.444 -5.049 -4.698 -4.496 -4.341 0.671 -0.1921 0.0869 0.287\n 3.00 1.759 -0.00147 0.0032 1.025 0.00644 0.307 -5.839 -5.431 -5.089 -4.893 -4.758 0.667 -0.2124 0.0954 0.278\n 4.00 1.826 -0.00195 0.0040 1.044 0.00590 0.353 -6.598 -6.181 -5.882 -5.698 -5.588 0.647 -0.2445 0.1088 0.273\n 5.00 1.825 -0.00237 0.0050 1.065 0.00510 0.248 -6.752 -6.347 -6.051 -5.873 -5.798 0.643 -0.2694 0.1193 0.275\n """'}), '(sa_damping=5, table=\n """ IMT a b c d e FR CH C1 C2 C3 C4 sigma QC WC tauC\n pga 1.101 -0.00564 0.0055 1.080 0.01412 0.251 0.293 1.111 1.344 1.355 1.420 0.604 0.0 0.0 0.303\n 0.05 1.076 -0.00671 0.0075 1.060 0.01463 0.251 0.939 1.684 1.793 1.747 1.814 0.640 0.0 0.0 0.326\n 0.10 1.118 -0.00787 0.0090 1.083 0.01423 0.240 1.499 2.061 2.135 2.031 2.082 0.694 0.0 0.0 0.342\n 0.15 1.134 -0.00722 0.0100 1.053 0.01509 0.251 1.462 1.916 2.168 2.052 2.113 0.702 0.0 0.0 0.331\n 0.20 1.147 -0.00659 0.0120 1.014 0.01462 0.260 1.280 1.669 2.085 2.001 2.030 0.692 0.0 0.0 0.312\n 0.25 1.149 -0.00590 0.0140 0.966 0.01459 0.269 1.121 1.468 1.942 1.941 1.937 0.682 0.0 0.0 0.298\n 0.30 1.163 -0.00520 0.0150 0.934 0.01458 0.259 0.852 1.172 1.683 1.808 1.770 0.670 0.0 0.0 0.300\n 0.40 1.200 -0.00422 0.0100 0.959 0.01257 0.248 0.365 0.655 1.127 1.482 1.397 0.659 0.0 0.0 0.346\n 0.50 1.250 -0.00338 0.0060 1.008 0.01114 0.247 -0.207 0.071 0.515 0.934 0.955 0.653 -0.0126 0.0116 0.338\n 0.60 1.293 -0.00282 0.0030 1.088 0.01019 0.233 -0.705 -0.429 -0.003 0.394 0.559 0.653 -0.0329 0.0202 0.349\n 0.70 1.336 -0.00258 0.0025 1.084 0.00979 0.220 -1.144 -0.866 -0.449 -0.111 0.188 0.652 -0.0501 0.0274 0.351\n 0.80 1.386 -0.00242 0.0022 1.088 0.00944 0.232 -1.609 -1.325 -0.928 -0.620 -0.246 0.647 -0.0650 0.0336 0.356\n 0.90 1.433 -0.00232 0.0020 1.109 0.00972 0.220 -2.023 -1.732 -1.349 -1.066 -0.643 0.653 -0.0781 0.0391 0.348\n 1.00 1.479 -0.00220 0.0020 1.115 0.01005 0.211 -2.451 -2.152 -1.776 -1.523 -1.084 0.657 -0.0899 0.0440 0.338\n 1.25 1.551 -0.00207 0.0020 1.083 0.01003 0.251 -3.243 -2.923 -2.542 -2.327 -1.936 0.660 -0.1148 0.0545 0.313\n 1.50 1.621 -0.00224 0.0020 1.091 0.00928 0.248 -3.888 -3.548 -3.169 -2.979 -2.661 0.664 -0.1351 0.0630 0.306\n 2.00 1.694 -0.00201 0.0025 1.055 0.00833 0.263 -4.783 -4.410 -4.039 -3.871 -3.640 0.669 -0.1672 0.0764 0.283\n 2.50 1.748 -0.00187 0.0028 1.052 0.00776 0.262 -5.444 -5.049 -4.698 -4.496 -4.341 0.671 -0.1921 0.0869 0.287\n 3.00 1.759 -0.00147 0.0032 1.025 0.00644 0.307 -5.839 -5.431 -5.089 -4.893 -4.758 0.667 -0.2124 0.0954 0.278\n 4.00 1.826 -0.00195 0.0040 1.044 0.00590 0.353 -6.598 -6.181 -5.882 -5.698 -5.588 0.647 -0.2445 0.1088 0.273\n 5.00 1.825 -0.00237 0.0050 1.065 0.00510 0.248 -6.752 -6.347 -6.051 -5.873 -5.798 0.643 -0.2694 0.1193 0.275\n """\n )\n', (7642, 10389), False, 'from openquake.hazardlib.gsim.base import GMPE, CoeffsTable\n'), ((13164, 14188), 'openquake.hazardlib.gsim.base.CoeffsTable', 'CoeffsTable', ([], {'sa_damping': '(5)', 'table': '""" IMT SI QI WI tauI\n pga 0.000 0.0 0.0 0.308\n 0.05 0.000 0.0 0.0 0.343\n 0.10 0.000 0.0 0.0 0.403\n 0.15 0.000 -0.0138 0.0286 0.367\n 0.20 0.000 -0.0256 0.0352 0.328\n 0.25 0.000 -0.0348 0.0403 0.289\n 0.30 0.000 -0.0423 0.0445 0.280\n 0.40 -0.041 -0.0541 0.0511 0.271\n 0.50 -0.053 -0.0632 0.0562 0.277\n 0.60 -0.103 -0.0707 0.0604 0.296\n 0.70 -0.146 -0.0771 0.0639 0.313\n 0.80 -0.164 -0.0825 0.0670 0.329\n 0.90 -0.206 -0.0874 0.0697 0.324\n 1.00 -0.239 -0.0917 0.0721 0.328\n 1.25 -0.256 -0.1009 0.0772 0.339\n 1.50 -0.306 -0.1083 0.0814 0.352\n 2.00 -0.321 -0.1202 0.0880 0.360\n 2.50 -0.337 -0.1293 0.0931 0.356\n 3.00 -0.331 -0.1368 0.0972 0.338\n 4.00 -0.390 -0.1486 0.1038 0.307\n 5.00 -0.498 -0.1578 0.1090 0.272\n """'}), '(sa_damping=5, table=\n """ IMT SI QI WI tauI\n pga 0.000 0.0 0.0 0.308\n 0.05 0.000 0.0 0.0 0.343\n 0.10 0.000 0.0 0.0 0.403\n 0.15 0.000 -0.0138 0.0286 0.367\n 0.20 0.000 -0.0256 0.0352 0.328\n 0.25 0.000 -0.0348 0.0403 0.289\n 0.30 0.000 -0.0423 0.0445 0.280\n 0.40 -0.041 -0.0541 0.0511 0.271\n 0.50 -0.053 -0.0632 0.0562 0.277\n 0.60 -0.103 -0.0707 0.0604 0.296\n 0.70 -0.146 -0.0771 0.0639 0.313\n 0.80 -0.164 -0.0825 0.0670 0.329\n 0.90 -0.206 -0.0874 0.0697 0.324\n 1.00 -0.239 -0.0917 0.0721 0.328\n 1.25 -0.256 -0.1009 0.0772 0.339\n 1.50 -0.306 -0.1083 0.0814 0.352\n 2.00 -0.321 -0.1202 0.0880 0.360\n 2.50 -0.337 -0.1293 0.0931 0.356\n 3.00 -0.331 -0.1368 0.0972 0.338\n 4.00 -0.390 -0.1486 0.1038 0.307\n 5.00 -0.498 -0.1578 0.1090 0.272\n """\n )\n', (13175, 14188), False, 'from openquake.hazardlib.gsim.base import GMPE, CoeffsTable\n'), ((17435, 18833), 'openquake.hazardlib.gsim.base.CoeffsTable', 'CoeffsTable', ([], {'sa_damping': '(5)', 'table': '""" IMT SS SSL PS QS WS tauS\n pga 2.607 -0.528 0.1392 0.1584 -0.0529 0.321\n 0.05 2.764 -0.551 0.1636 0.1932 -0.0841 0.378\n 0.10 2.156 -0.420 0.1690 0.2057 -0.0877 0.420\n 0.15 2.161 -0.431 0.1669 0.1984 -0.0773 0.372\n 0.20 1.901 -0.372 0.1631 0.1856 -0.0644 0.324\n 0.25 1.814 -0.360 0.1588 0.1714 -0.0515 0.294\n 0.30 2.181 -0.450 0.1544 0.1573 -0.0395 0.284\n 0.40 2.432 -0.506 0.1460 0.1309 -0.0183 0.278\n 0.50 2.629 -0.554 0.1381 0.1078 -0.0008 0.272\n 0.60 2.702 -0.575 0.1307 0.0878 0.0136 0.285\n 0.70 2.654 -0.572 0.1239 0.0705 0.0254 0.290\n 0.80 2.480 -0.540 0.1176 0.0556 0.0352 0.299\n 0.90 2.332 -0.522 0.1116 0.0426 0.0432 0.289\n 1.00 2.233 -0.509 0.1060 0.0314 0.0498 0.286\n 1.25 2.029 -0.469 0.0933 0.0093 0.0612 0.277\n 1.50 1.589 -0.379 0.0821 -0.0062 0.0674 0.282\n 2.00 0.966 -0.248 0.0628 -0.0235 0.0692 0.300\n 2.50 0.789 -0.221 0.0465 -0.0287 0.0622 0.292\n 3.00 1.037 -0.263 0.0322 -0.0261 0.0496 0.274\n 4.00 0.561 -0.169 0.0083 -0.0065 0.0150 0.281\n 5.00 0.225 -0.120 -0.0117 0.0246 -0.0268 0.296\n """'}), '(sa_damping=5, table=\n """ IMT SS SSL PS QS WS tauS\n pga 2.607 -0.528 0.1392 0.1584 -0.0529 0.321\n 0.05 2.764 -0.551 0.1636 0.1932 -0.0841 0.378\n 0.10 2.156 -0.420 0.1690 0.2057 -0.0877 0.420\n 0.15 2.161 -0.431 0.1669 0.1984 -0.0773 0.372\n 0.20 1.901 -0.372 0.1631 0.1856 -0.0644 0.324\n 0.25 1.814 -0.360 0.1588 0.1714 -0.0515 0.294\n 0.30 2.181 -0.450 0.1544 0.1573 -0.0395 0.284\n 0.40 2.432 -0.506 0.1460 0.1309 -0.0183 0.278\n 0.50 2.629 -0.554 0.1381 0.1078 -0.0008 0.272\n 0.60 2.702 -0.575 0.1307 0.0878 0.0136 0.285\n 0.70 2.654 -0.572 0.1239 0.0705 0.0254 0.290\n 0.80 2.480 -0.540 0.1176 0.0556 0.0352 0.299\n 0.90 2.332 -0.522 0.1116 0.0426 0.0432 0.289\n 1.00 2.233 -0.509 0.1060 0.0314 0.0498 0.286\n 1.25 2.029 -0.469 0.0933 0.0093 0.0612 0.277\n 1.50 1.589 -0.379 0.0821 -0.0062 0.0674 0.282\n 2.00 0.966 -0.248 0.0628 -0.0235 0.0692 0.300\n 2.50 0.789 -0.221 0.0465 -0.0287 0.0622 0.292\n 3.00 1.037 -0.263 0.0322 -0.0261 0.0496 0.274\n 4.00 0.561 -0.169 0.0083 -0.0065 0.0150 0.281\n 5.00 0.225 -0.120 -0.0117 0.0246 -0.0268 0.296\n """\n )\n', (17446, 18833), False, 'from openquake.hazardlib.gsim.base import GMPE, CoeffsTable\n'), ((20084, 21129), 'openquake.hazardlib.gsim.base.CoeffsTable', 'CoeffsTable', ([], {'sa_damping': '(5)', 'table': '""" IMT SI QI WI tauI\n pga 0.000 0.0 0.0 0.3976\n 0.05 0.000 0.0 0.0 0.4437\n 0.10 0.000 0.0 0.0 0.4903\n 0.15 0.000 -0.0138 0.0286 0.4603\n 0.20 0.000 -0.0256 0.0352 0.4233\n 0.25 0.000 -0.0348 0.0403 0.3908\n 0.30 0.000 -0.0423 0.0445 0.3790\n 0.40 -0.041 -0.0541 0.0511 0.3897\n 0.50 -0.053 -0.0632 0.0562 0.3890\n 0.60 -0.103 -0.0707 0.0604 0.4014\n 0.70 -0.146 -0.0771 0.0639 0.4079\n 0.80 -0.164 -0.0825 0.0670 0.4183\n 0.90 -0.206 -0.0874 0.0697 0.4106\n 1.00 -0.239 -0.0917 0.0721 0.4101\n 1.25 -0.256 -0.1009 0.0772 0.4021\n 1.50 -0.306 -0.1083 0.0814 0.4076\n 2.00 -0.321 -0.1202 0.0880 0.4138\n 2.50 -0.337 -0.1293 0.0931 0.4108\n 3.00 -0.331 -0.1368 0.0972 0.3961\n 4.00 -0.390 -0.1486 0.1038 0.3821\n 5.00 -0.498 -0.1578 0.1090 0.3766\n """'}), '(sa_damping=5, table=\n """ IMT SI QI WI tauI\n pga 0.000 0.0 0.0 0.3976\n 0.05 0.000 0.0 0.0 0.4437\n 0.10 0.000 0.0 0.0 0.4903\n 0.15 0.000 -0.0138 0.0286 0.4603\n 0.20 0.000 -0.0256 0.0352 0.4233\n 0.25 0.000 -0.0348 0.0403 0.3908\n 0.30 0.000 -0.0423 0.0445 0.3790\n 0.40 -0.041 -0.0541 0.0511 0.3897\n 0.50 -0.053 -0.0632 0.0562 0.3890\n 0.60 -0.103 -0.0707 0.0604 0.4014\n 0.70 -0.146 -0.0771 0.0639 0.4079\n 0.80 -0.164 -0.0825 0.0670 0.4183\n 0.90 -0.206 -0.0874 0.0697 0.4106\n 1.00 -0.239 -0.0917 0.0721 0.4101\n 1.25 -0.256 -0.1009 0.0772 0.4021\n 1.50 -0.306 -0.1083 0.0814 0.4076\n 2.00 -0.321 -0.1202 0.0880 0.4138\n 2.50 -0.337 -0.1293 0.0931 0.4108\n 3.00 -0.331 -0.1368 0.0972 0.3961\n 4.00 -0.390 -0.1486 0.1038 0.3821\n 5.00 -0.498 -0.1578 0.1090 0.3766\n """\n )\n', (20095, 21129), False, 'from openquake.hazardlib.gsim.base import GMPE, CoeffsTable\n'), ((19837, 19855), 'copy.deepcopy', 'copy.deepcopy', (['rup'], {}), '(rup)\n', (19850, 19855), False, 'import copy\n'), ((17183, 17195), 'numpy.log', 'np.log', (['rrup'], {}), '(rrup)\n', (17189, 17195), True, 'import numpy as np\n'), ((4892, 4922), 'numpy.sqrt', 'np.sqrt', (['(sigma ** 2 + tau ** 2)'], {}), '(sigma ** 2 + tau ** 2)\n', (4899, 4922), True, 'import numpy as np\n'), ((4328, 4340), 'numpy.exp', 'np.exp', (['mean'], {}), '(mean)\n', (4334, 4340), True, 'import numpy as np\n'), ((12766, 12778), 'numpy.exp', 'np.exp', (['mean'], {}), '(mean)\n', (12772, 12778), True, 'import numpy as np\n'), ((16766, 16778), 'numpy.exp', 'np.exp', (['mean'], {}), '(mean)\n', (16772, 16778), True, 'import numpy as np\n'), ((22960, 22972), 'numpy.exp', 'np.exp', (['mean'], {}), '(mean)\n', (22966, 22972), True, 'import numpy as np\n'), ((4964, 4983), 'numpy.zeros', 'np.zeros', (['num_sites'], {}), '(num_sites)\n', (4972, 4983), True, 'import numpy as np\n'), ((5602, 5622), 'numpy.exp', 'np.exp', (["(C['d'] * mag)"], {}), "(C['d'] * mag)\n", (5608, 5622), True, 'import numpy as np\n'), ((25926, 25938), 'numpy.exp', 'np.exp', (['mean'], {}), '(mean)\n', (25932, 25938), True, 'import numpy as np\n'), ((28395, 28407), 'numpy.exp', 'np.exp', (['mean'], {}), '(mean)\n', (28401, 28407), True, 'import numpy as np\n'), ((5082, 5101), 'numpy.zeros', 'np.zeros', (['num_sites'], {}), '(num_sites)\n', (5090, 5101), True, 'import numpy as np\n'), ((5198, 5217), 'numpy.zeros', 'np.zeros', (['num_sites'], {}), '(num_sites)\n', (5206, 5217), True, 'import numpy as np\n')]
|
# PRISM CONVERSION FROM ASCII GRIDS -- TASMIN / TASMAX
# header info
# ncols 2015
# nrows 1320
# xllcorner -2301787.7731349
# yllcorner 108069.7858797
# cellsize 2000
# NODATA_value -9999
import rasterio, glob, os
from rasterio import Affine
import numpy as np
from pathos import multiprocessing as mp
# input_path = '/Data/Base_Data/Climate/AK_CAN_2km/historical/singleBand/pr'
# #'/Data/Base_Data/Climate/AK_CAN_2km/historical/singleBand/prism/AK_2KM_PRISM/Temperature/2km/older'
# output_path = '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2'
# groups = ['min_temp', 'max_temp']
# # # STEP 1 -- CONVERT TO GTIFF FROM ASC AND TXT
# list the data we want
variables = [ 'tmin', 'tmax' ]
input_path_ak = '/Data/Base_Data/Climate/AK_CAN_2km/historical/singleBand/prism/AK_2KM_PRISM/Temperature/2km/older'
input_path_can = '/Data/Base_Data/Climate/AK_CAN_2km/historical/singleBand/prism/AK_CAN_2km_PRISM/CAN_originals/older'
for variable in variables:
for ak_test, input_path in zip( [True,False], [input_path_ak,input_path_can] ):
output_path = os.path.join( '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2', variable,'raw_converted' )
if not os.path.exists( output_path ):
os.makedirs( output_path )
if ak_test:
input_path = input_path_ak
if variable == 'tmin':
v = 'min_temp'
elif variable == 'tmax':
v = 'max_temp'
else:
NotImplemented( 'only tmax / tmin currently supported' )
files = glob.glob( os.path.join( input_path, v, '*'+variable+'*.txt' ) )
else:
input_path = input_path_can
files = glob.glob( os.path.join( input_path, '*'+variable+'*.asc' ) )
ext = files[0].split('.')[1]
output_filenames = [ os.path.join( output_path, os.path.basename( fn ).replace( '.'+ext, '.tif' ) ) for fn in files ]
crs = {'init':'epsg:4326'}
args = [ (i,j,crs) for i,j in zip(files, output_filenames) ]
def bounds_to_extent( bounds ):
'''
take input rasterio bounds object and return an extent
'''
l,b,r,t = bounds
return [ (l,b), (r,b), (r,t), (l,t), (l,b) ]
def convert_to_gtiff( fn, output_filename, crs={'init':'epsg:3338'} ):
'''
convert the ascii rasters from PRISM to gtiff
'''
print( fn )
rst = rasterio.open( fn )
arr = rst.read( 1 ) # get the first and only band
meta = rst.meta
meta.update( compress='lzw', driver='GTiff', crs=crs )
# drop the transform to overcome rasterio warnings
if 'transform' in meta.keys():
meta.pop( 'transform' )
# write them out
with rasterio.open( output_filename, 'w', **meta ) as out:
out.write( arr, 1 )
return output_filename
if __name__ == '__main__':
pool = mp.Pool( 32 )
pool.map( lambda x: convert_to_gtiff( *x ), args )
pool.close()
pool.join()
# # # STEP 2 -- MERGE IT WITH GDAL TOOLS
# list the data
caw = sorted( glob.glob( os.path.join( '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2',variable,'raw_converted', 'caw*.tif' ) ) )
ak = sorted( glob.glob( os.path.join( '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2',variable,'raw_converted', 'ak_*.tif' ) ) )
grouped = zip( ak, caw )
# merge these files:
# log = open( '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2/batch_run.bat', 'w' )
for ak,ca in grouped:
out = ak.replace( 'ak_', 'akcan_')
ca_out = ca.replace( '.tif', '_3338.tif' )
os.system( 'gdalwarp -overwrite -r near -t_srs EPSG:3338 -s_srs EPSG:4326 -ot Float32 ' + ca + ' ' + ca_out )
ca_scale = ca_out.replace( '.tif', '_scaled.tif' )
os.system( 'gdal_calc.py --overwrite -A ' + ca_out + ' --outfile=' + ca_scale + ' --calc="A*(0.1)" --NoDataValue=-9999 --type=Float32' )
os.system( 'gdal_merge.py -init -9999 -n -9999 -a_nodata -9999 -ot Float32 -o ' + out + ' ' + ak + ' ' + ca_scale )
final = ca.replace( '.tif', '_merged.tif' ).replace( 'raw_converted', 'merged' ).replace( 'caw_', 'akcan_' )
if not os.path.exists( os.path.dirname(final) ):
os.makedirs(os.path.dirname(final))
os.system( 'gdal_translate -co "COMPRESS=LZW" ' + out + ' ' + final )
# # DUE TO SOME WEIRDNESS WITH VIRTUALENV AND GDAL_MERGE.PY I am writing this out to a text file and running it when not in virtualenv
# out = ak.replace( 'ak_', 'akcan_')
# ca_out = ca.replace( '.tif', '_3338.tif' )
# log.write( 'gdalwarp -overwrite -r near -t_srs EPSG:3338 -s_srs EPSG:4326 -ot Float32 ' + ca + ' ' + ca_out + '\n' )
# ca_scale = ca_out.replace( '.tif', '_scaled.tif' )
# log.write( 'gdal_calc.py --overwrite -A ' + ca_out + ' --outfile=' + ca_scale + ' --calc="A*(0.1)" --NoDataValue=-9999 --type=Float32' + '\n' )
# log.write( 'gdal_merge.py -init -9999 -n -9999 -a_nodata -9999 -ot Float32 -o ' + out + ' ' + ak + ' ' + ca_scale + '\n' )
# final = ca.replace( '.tif', '_merged.tif' )
# log.write( 'gdal_translate -co "COMPRESS=LZW" ' + out + ' ' + final + '\n' )
# # # STEP 3 -- INTERPOLATE / REGRID / MASK to match existing SNAP resources
def coordinates( fn=None, meta=None, numpy_array=None, input_crs=None, to_latlong=False ):
'''
take a raster file as input and return the centroid coords for each
of the grid cells as a pair of numpy 2d arrays (longitude, latitude)
'''
import rasterio
import numpy as np
from affine import Affine
from pyproj import Proj, transform
if fn:
# Read raster
with rasterio.open( fn ) as r:
T0 = r.affine # upper-left pixel corner affine transform
p1 = Proj( r.crs )
A = r.read( 1 ) # pixel values
elif (meta is not None) & (numpy_array is not None):
A = numpy_array
if input_crs != None:
p1 = Proj( input_crs )
T0 = meta[ 'affine' ]
else:
p1 = None
T0 = meta[ 'affine' ]
else:
BaseException( 'check inputs' )
# All rows and columns
cols, rows = np.meshgrid(np.arange(A.shape[1]), np.arange(A.shape[0]))
# Get affine transform for pixel centres
T1 = T0 * Affine.translation( 0.5, 0.5 )
# Function to convert pixel row/column index (from 0) to easting/northing at centre
rc2en = lambda r, c: ( c, r ) * T1
# All eastings and northings (there is probably a faster way to do this)
eastings, northings = np.vectorize(rc2en, otypes=[np.float, np.float])(rows, cols)
if to_latlong == False:
return eastings, northings
elif (to_latlong == True) & (input_crs != None):
# Project all longitudes, latitudes
longs, lats = transform(p1, p1.to_latlong(), eastings, northings)
return longs, lats
else:
BaseException( 'cant reproject to latlong without an input_crs' )
def xyz_to_grid( x, y, z, grid, method='cubic', output_dtype=np.float32, *args, **kwargs ):
'''
interpolate points to a grid. simple wrapper around
scipy.interpolate.griddata. Points and grid must be
in the same coordinate system
x = 1-D np.array of x coordinates / x,y,z must be same length
y = 1-D np.array of y coordinates / x,y,z must be same length
z = 1-D np.array of z coordinates / x,y,z must be same length
grid = tuple of meshgrid as made using numpy.meshgrid()
order (xi, yi)
method = one of 'cubic', 'near', 'linear'
'''
from scipy.interpolate import griddata
zi = griddata( (x, y), z, grid, method=method )
zi = np.flipud( zi ).astype( output_dtype )
return zi
import glob, os, rasterio
import numpy as np
from rasterio.warp import reproject, RESAMPLING
import pandas as pd
# ORIG_AK_RAW = '/Data/Base_Data/Climate/AK_CAN_2km/historical/singleBand/prism/AK_2KM_PRISM/Temperature/2km/older'
# TEMPLATE:
template_fn = '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/akcan_template/tas_mean_C_AR5_CCSM4_rcp26_01_2006.tif'
rst = rasterio.open( template_fn )
mask = rst.read_masks()
input_dir = os.path.join( '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2', variable, 'merged' )
files = glob.glob( os.path.join( input_dir, 'akcan_*.tif' ) )
# * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
# make an empty raster to hold the new output
for fn in files:
template = rasterio.open( template_fn )
meta = template.meta
meta.update( compress='lzw' )
output_filename = os.path.join( '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2', variable, os.path.basename( fn ).replace( 'merged', 'final' ) )
with rasterio.open( output_filename, 'w', **meta ) as out:
out.write( np.empty_like( template.read( 1 ) ), 1 )
# run it with gdalwarp
command = 'gdalwarp ' + fn + ' ' + output_filename
os.system( command )
# * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
# for fn in files:
# cur = rasterio.open( fn )
# # GET THE COORDS AND THE DATA IN A WAY THAT IS INTEPOLATABLE
# lons, lats = coordinates( fn )
# df = pd.DataFrame( {'lons':lons.ravel(), 'lats':lats.ravel(), 'dat':cur.read(1).ravel() } )
# new_df = df[df.dat != -9999]
# new_grid = xyz_to_grid( new_df.lons.tolist(), new_df.lats.tolist(), new_df.dat.tolist(), (lons,lats), method='cubic', output_dtype=np.float32 )
# # new_grid[ np.isnan( new_grid ) ] = -9999
# output_arr = np.empty_like( rst.read( 1 ) )
# # now reproject the new_grid to the extent/resolution/crs of the ALF data -- idiotic crs we use here
# reproject( new_grid, output_arr, src_transform=cur.affine, src_crs={ 'init':'epsg:3338' }, src_nodata=None, \
# dst_transform=rst.affine, dst_crs=rst.crs,\
# dst_nodata=None, resampling=RESAMPLING.cubic_spline, SOURCE_EXTRA=1000 )
# output_arr[ mask == 0 ] = rst.nodata
# new_path = os.path.join( '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2', variable, 'prepped' )
# output_filename = os.path.join( new_path, os.path.basename( fn ) )
# meta = rst.meta
# meta.update( compress='lzw' )
# meta.pop( 'transform' )
# with rasterio.open( output_filename, 'w', **meta ) as out:
# out.write( output_arr, 1 )
# # # # END PREP
# THIS IS HOW IT WAS CONVERTED FROM THE TXT/ASC FORMATS
# for group in groups:
# # list the data we want to convert to GTiff
# # remember that month 14 is the annual average
# files = glob.glob( os.path.join( input_path, group, '*.txt' ) )
# for fn in files:
# # print fn
# rst = rasterio.open( fn )
# arr = rst.read( 1 ) # get the first and only band
# output_filename = os.path.join( output_path, os.path.basename(fn).replace( '.txt', '.tif' ) )
# meta = rst.meta
# meta.update( compress='lzw', driver='GTiff', crs={'init':'epsg:3338'} )
# # drop the transform to overcome rasterio warnings
# if 'transform' in meta.keys():
# meta.pop( 'transform' )
# # write them out
# with rasterio.open( output_filename, 'w', **meta ) as out:
# out.write( arr, 1 )
|
[
"os.path.exists",
"os.makedirs",
"pathos.multiprocessing.Pool",
"numpy.flipud",
"rasterio.open",
"scipy.interpolate.griddata",
"os.path.join",
"os.path.dirname",
"affine.Affine.translation",
"os.path.basename",
"pyproj.Proj",
"os.system",
"numpy.vectorize",
"numpy.arange"
] |
[((7696, 7722), 'rasterio.open', 'rasterio.open', (['template_fn'], {}), '(template_fn)\n', (7709, 7722), False, 'import rasterio\n'), ((7764, 7883), 'os.path.join', 'os.path.join', (['"""/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2"""', 'variable', '"""merged"""'], {}), "(\n '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2'\n , variable, 'merged')\n", (7776, 7883), False, 'import glob, os, rasterio\n'), ((1110, 1236), 'os.path.join', 'os.path.join', (['"""/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2"""', 'variable', '"""raw_converted"""'], {}), "(\n '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2'\n , variable, 'raw_converted')\n", (1122, 1236), False, 'import glob, os, rasterio\n'), ((3456, 3573), 'os.system', 'os.system', (["(\n 'gdalwarp -overwrite -r near -t_srs EPSG:3338 -s_srs EPSG:4326 -ot Float32 '\n + ca + ' ' + ca_out)"], {}), "(\n 'gdalwarp -overwrite -r near -t_srs EPSG:3338 -s_srs EPSG:4326 -ot Float32 '\n + ca + ' ' + ca_out)\n", (3465, 3573), False, 'import glob, os, rasterio\n'), ((3621, 3759), 'os.system', 'os.system', (['(\'gdal_calc.py --overwrite -A \' + ca_out + \' --outfile=\' + ca_scale +\n \' --calc="A*(0.1)" --NoDataValue=-9999 --type=Float32\')'], {}), '(\'gdal_calc.py --overwrite -A \' + ca_out + \' --outfile=\' +\n ca_scale + \' --calc="A*(0.1)" --NoDataValue=-9999 --type=Float32\')\n', (3630, 3759), False, 'import glob, os, rasterio\n'), ((3760, 3882), 'os.system', 'os.system', (["('gdal_merge.py -init -9999 -n -9999 -a_nodata -9999 -ot Float32 -o ' + out +\n ' ' + ak + ' ' + ca_scale)"], {}), "(\n 'gdal_merge.py -init -9999 -n -9999 -a_nodata -9999 -ot Float32 -o ' +\n out + ' ' + ak + ' ' + ca_scale)\n", (3769, 3882), False, 'import glob, os, rasterio\n'), ((4079, 4146), 'os.system', 'os.system', (['(\'gdal_translate -co "COMPRESS=LZW" \' + out + \' \' + final)'], {}), '(\'gdal_translate -co "COMPRESS=LZW" \' + out + \' \' + final)\n', (4088, 4146), False, 'import glob, os, rasterio\n'), ((7202, 7242), 'scipy.interpolate.griddata', 'griddata', (['(x, y)', 'z', 'grid'], {'method': 'method'}), '((x, y), z, grid, method=method)\n', (7210, 7242), False, 'from scipy.interpolate import griddata\n'), ((7896, 7934), 'os.path.join', 'os.path.join', (['input_dir', '"""akcan_*.tif"""'], {}), "(input_dir, 'akcan_*.tif')\n", (7908, 7934), False, 'import glob, os, rasterio\n'), ((8134, 8160), 'rasterio.open', 'rasterio.open', (['template_fn'], {}), '(template_fn)\n', (8147, 8160), False, 'import rasterio\n'), ((8590, 8608), 'os.system', 'os.system', (['command'], {}), '(command)\n', (8599, 8608), False, 'import glob, os, rasterio\n'), ((1238, 1265), 'os.path.exists', 'os.path.exists', (['output_path'], {}), '(output_path)\n', (1252, 1265), False, 'import glob, os, rasterio\n'), ((1272, 1296), 'os.makedirs', 'os.makedirs', (['output_path'], {}), '(output_path)\n', (1283, 1296), False, 'import glob, os, rasterio\n'), ((2274, 2291), 'rasterio.open', 'rasterio.open', (['fn'], {}), '(fn)\n', (2287, 2291), False, 'import rasterio\n'), ((2712, 2723), 'pathos.multiprocessing.Pool', 'mp.Pool', (['(32)'], {}), '(32)\n', (2719, 2723), True, 'from pathos import multiprocessing as mp\n'), ((2897, 3035), 'os.path.join', 'os.path.join', (['"""/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2"""', 'variable', '"""raw_converted"""', '"""caw*.tif"""'], {}), "(\n '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2'\n , variable, 'raw_converted', 'caw*.tif')\n", (2909, 3035), False, 'import glob, os, rasterio\n'), ((3055, 3193), 'os.path.join', 'os.path.join', (['"""/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2"""', 'variable', '"""raw_converted"""', '"""ak_*.tif"""'], {}), "(\n '/workspace/Shared/Tech_Projects/EPSCoR_Southcentral/project_data/prism_v2'\n , variable, 'raw_converted', 'ak_*.tif')\n", (3067, 3193), False, 'import glob, os, rasterio\n'), ((5863, 5884), 'numpy.arange', 'np.arange', (['A.shape[1]'], {}), '(A.shape[1])\n', (5872, 5884), True, 'import numpy as np\n'), ((5886, 5907), 'numpy.arange', 'np.arange', (['A.shape[0]'], {}), '(A.shape[0])\n', (5895, 5907), True, 'import numpy as np\n'), ((5964, 5992), 'affine.Affine.translation', 'Affine.translation', (['(0.5)', '(0.5)'], {}), '(0.5, 0.5)\n', (5982, 5992), False, 'from affine import Affine\n'), ((6217, 6265), 'numpy.vectorize', 'np.vectorize', (['rc2en'], {'otypes': '[np.float, np.float]'}), '(rc2en, otypes=[np.float, np.float])\n', (6229, 6265), True, 'import numpy as np\n'), ((8400, 8443), 'rasterio.open', 'rasterio.open', (['output_filename', '"""w"""'], {}), "(output_filename, 'w', **meta)\n", (8413, 8443), False, 'import rasterio\n'), ((1529, 1582), 'os.path.join', 'os.path.join', (['input_path', 'v', "('*' + variable + '*.txt')"], {}), "(input_path, v, '*' + variable + '*.txt')\n", (1541, 1582), False, 'import glob, os, rasterio\n'), ((1644, 1694), 'os.path.join', 'os.path.join', (['input_path', "('*' + variable + '*.asc')"], {}), "(input_path, '*' + variable + '*.asc')\n", (1656, 1694), False, 'import glob, os, rasterio\n'), ((2568, 2611), 'rasterio.open', 'rasterio.open', (['output_filename', '"""w"""'], {}), "(output_filename, 'w', **meta)\n", (2581, 2611), False, 'import rasterio\n'), ((4012, 4034), 'os.path.dirname', 'os.path.dirname', (['final'], {}), '(final)\n', (4027, 4034), False, 'import glob, os, rasterio\n'), ((4053, 4075), 'os.path.dirname', 'os.path.dirname', (['final'], {}), '(final)\n', (4068, 4075), False, 'import glob, os, rasterio\n'), ((5418, 5435), 'rasterio.open', 'rasterio.open', (['fn'], {}), '(fn)\n', (5431, 5435), False, 'import rasterio\n'), ((5515, 5526), 'pyproj.Proj', 'Proj', (['r.crs'], {}), '(r.crs)\n', (5519, 5526), False, 'from pyproj import Proj, transform\n'), ((7252, 7265), 'numpy.flipud', 'np.flipud', (['zi'], {}), '(zi)\n', (7261, 7265), True, 'import numpy as np\n'), ((5674, 5689), 'pyproj.Proj', 'Proj', (['input_crs'], {}), '(input_crs)\n', (5678, 5689), False, 'from pyproj import Proj, transform\n'), ((8339, 8359), 'os.path.basename', 'os.path.basename', (['fn'], {}), '(fn)\n', (8355, 8359), False, 'import glob, os, rasterio\n'), ((1777, 1797), 'os.path.basename', 'os.path.basename', (['fn'], {}), '(fn)\n', (1793, 1797), False, 'import glob, os, rasterio\n')]
|
#!/usr/bin/py2
import cv2
import imutils
import numpy as np
from solver import Solver
from Recognizer import OCR
from skimage.segmentation import clear_border
from imutils.perspective import four_point_transform
class Sudoku(object):
def __init__(self, image):
self.image = image
self.gray = None
def initialize_image(self):
self.image = cv2.imread(self.image)
self.image = imutils.resize(self.image, width=600)
return
def fetch_rectangle(self):
self.gray = cv2.cvtColor(self.image, cv2.COLOR_BGR2GRAY)
blurred = cv2.GaussianBlur(self.gray, (7, 7), 3)
thresh = cv2.adaptiveThreshold(blurred, 255,cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY, 11, 2)
thresh = cv2.bitwise_not(thresh)
cnts = cv2.findContours(thresh.copy(), cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)
cnts = imutils.grab_contours(cnts)
cnts = sorted(cnts, key=cv2.contourArea, reverse=True)
puzzleCnt = None
for c in cnts:
peri = cv2.arcLength(c, True)
approx = cv2.approxPolyDP(c, 0.02 * peri, True)
if len(approx) == 4:
puzzleCnt = approx
break
return puzzleCnt
def extract_sudoku_board(self,board):
original_image = four_point_transform(self.image, board.reshape(4, 2))
gray_image = four_point_transform(self.gray, board.reshape(4, 2))
return gray_image
def split_board(self,board):
return board.shape[1] // 9, board.shape[0] // 9
def extract_digit(self,cell):
thresh = cv2.threshold(cell, 0, 255,
cv2.THRESH_BINARY_INV | cv2.THRESH_OTSU)[1]
thresh = clear_border(thresh)
cnts = cv2.findContours(thresh.copy(), cv2.RETR_EXTERNAL,
cv2.CHAIN_APPROX_SIMPLE)
cnts = imutils.grab_contours(cnts)
if len(cnts) == 0:
return None
c = max(cnts, key=cv2.contourArea)
mask = np.zeros(thresh.shape, dtype="uint8")
cv2.drawContours(mask, [c], -1, 255, -1)
(h, w) = thresh.shape
percentFilled = cv2.countNonZero(mask) / float(w * h)
if percentFilled < 0.03:
return None
digit = cv2.bitwise_and(thresh, thresh, mask=mask)
return digit
def process_cells(self,stepX,stepY,board):
ocr = OCR()
sudoku_array = np.zeros((9, 9), dtype="int")
cellLocs = []
boolean = True
for y in range(0, 9):
row = []
for x in range(0, 9):
startX = x * stepX
startY = y * stepY
endX = (x + 1) * stepX
endY = (y + 1) * stepY
row.append((startX, startY, endX, endY))
cell = board[startY:endY, startX:endX]
digit = self.extract_digit(cell)
if (digit is not None):
cv2.imwrite("img-"+str(y)+str(x)+".png",digit)
sudoku_array[y][x] = ocr.prediction(digit)
return sudoku_array
def solve(self):
self.initialize_image()
board = self.fetch_rectangle()
if board is None:
return
board = self.extract_sudoku_board(board)
x,y = self.split_board(board)
final_board = self.process_cells(x,y,board)
return final_board
def manipulate(board):
canShow = True
while (canShow):
decision = raw_input("\nWanna make any corrections to the board? Press y if yes:-")
if (decision and decision[0].lower() == "y"):
canShow = False
break
values = raw_input("\nEnter Row,Column and Value. Ex: 2,2,3: ").split(",")
try:
row,col,val = list(map(int,values[:3]))
check = lambda x: x>0 and x<10
if (all([check(i) for i in [row,col,val]])):
board[row-1][col-1] = val
print("\nUpdated Board\n")
print(board)
else:
print("\nInvalid input")
except:
print("\nInvalid input")
return board
sudoku_board = Sudoku("Images/sudoku.jpg").solve()
print(sudoku_board)
updated_board = manipulate(sudoku_board)
Solver().solution(updated_board)
|
[
"Recognizer.OCR",
"cv2.drawContours",
"cv2.countNonZero",
"cv2.threshold",
"solver.Solver",
"cv2.arcLength",
"cv2.bitwise_and",
"skimage.segmentation.clear_border",
"imutils.resize",
"cv2.adaptiveThreshold",
"imutils.grab_contours",
"numpy.zeros",
"cv2.approxPolyDP",
"cv2.cvtColor",
"cv2.bitwise_not",
"cv2.GaussianBlur",
"cv2.imread"
] |
[((354, 376), 'cv2.imread', 'cv2.imread', (['self.image'], {}), '(self.image)\n', (364, 376), False, 'import cv2\n'), ((392, 429), 'imutils.resize', 'imutils.resize', (['self.image'], {'width': '(600)'}), '(self.image, width=600)\n', (406, 429), False, 'import imutils\n'), ((486, 530), 'cv2.cvtColor', 'cv2.cvtColor', (['self.image', 'cv2.COLOR_BGR2GRAY'], {}), '(self.image, cv2.COLOR_BGR2GRAY)\n', (498, 530), False, 'import cv2\n'), ((543, 581), 'cv2.GaussianBlur', 'cv2.GaussianBlur', (['self.gray', '(7, 7)', '(3)'], {}), '(self.gray, (7, 7), 3)\n', (559, 581), False, 'import cv2\n'), ((594, 692), 'cv2.adaptiveThreshold', 'cv2.adaptiveThreshold', (['blurred', '(255)', 'cv2.ADAPTIVE_THRESH_GAUSSIAN_C', 'cv2.THRESH_BINARY', '(11)', '(2)'], {}), '(blurred, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.\n THRESH_BINARY, 11, 2)\n', (615, 692), False, 'import cv2\n'), ((698, 721), 'cv2.bitwise_not', 'cv2.bitwise_not', (['thresh'], {}), '(thresh)\n', (713, 721), False, 'import cv2\n'), ((817, 844), 'imutils.grab_contours', 'imutils.grab_contours', (['cnts'], {}), '(cnts)\n', (838, 844), False, 'import imutils\n'), ((1521, 1541), 'skimage.segmentation.clear_border', 'clear_border', (['thresh'], {}), '(thresh)\n', (1533, 1541), False, 'from skimage.segmentation import clear_border\n'), ((1641, 1668), 'imutils.grab_contours', 'imutils.grab_contours', (['cnts'], {}), '(cnts)\n', (1662, 1668), False, 'import imutils\n'), ((1753, 1790), 'numpy.zeros', 'np.zeros', (['thresh.shape'], {'dtype': '"""uint8"""'}), "(thresh.shape, dtype='uint8')\n", (1761, 1790), True, 'import numpy as np\n'), ((1793, 1833), 'cv2.drawContours', 'cv2.drawContours', (['mask', '[c]', '(-1)', '(255)', '(-1)'], {}), '(mask, [c], -1, 255, -1)\n', (1809, 1833), False, 'import cv2\n'), ((1969, 2011), 'cv2.bitwise_and', 'cv2.bitwise_and', (['thresh', 'thresh'], {'mask': 'mask'}), '(thresh, thresh, mask=mask)\n', (1984, 2011), False, 'import cv2\n'), ((2082, 2087), 'Recognizer.OCR', 'OCR', ([], {}), '()\n', (2085, 2087), False, 'from Recognizer import OCR\n'), ((2106, 2135), 'numpy.zeros', 'np.zeros', (['(9, 9)'], {'dtype': '"""int"""'}), "((9, 9), dtype='int')\n", (2114, 2135), True, 'import numpy as np\n'), ((3596, 3604), 'solver.Solver', 'Solver', ([], {}), '()\n', (3602, 3604), False, 'from solver import Solver\n'), ((952, 974), 'cv2.arcLength', 'cv2.arcLength', (['c', '(True)'], {}), '(c, True)\n', (965, 974), False, 'import cv2\n'), ((987, 1025), 'cv2.approxPolyDP', 'cv2.approxPolyDP', (['c', '(0.02 * peri)', '(True)'], {}), '(c, 0.02 * peri, True)\n', (1003, 1025), False, 'import cv2\n'), ((1435, 1503), 'cv2.threshold', 'cv2.threshold', (['cell', '(0)', '(255)', '(cv2.THRESH_BINARY_INV | cv2.THRESH_OTSU)'], {}), '(cell, 0, 255, cv2.THRESH_BINARY_INV | cv2.THRESH_OTSU)\n', (1448, 1503), False, 'import cv2\n'), ((1877, 1899), 'cv2.countNonZero', 'cv2.countNonZero', (['mask'], {}), '(mask)\n', (1893, 1899), False, 'import cv2\n')]
|
import numpy as np
from public_tool.form_index import form_index
from XGB_HMM.form_B_matrix_by_XGB import form_B_matrix_by_XGB
from XGB_HMM.predict import self_pred
def pred_proba_XGB(A, model, pi, O, allow_flag, lengths):
# 对dataset形成pred_proba，注意这里的dataset是solve_on_raw_data后的结果，即附带allow_flag的数据
# output:
# pred_proba：数组类型
n_states = len(pi)
pred_proba = np.zeros((O.shape[0], n_states))
for i in range(len(lengths)):
begin_index, end_index = form_index(lengths, i)
now_O = O[begin_index:end_index, :]
now_allow_flag = allow_flag[begin_index:end_index]
now_pred_proba = np.zeros((now_O.shape[0], n_states))
now_allow_B = form_B_matrix_by_XGB(model, now_O[now_allow_flag == 1], pi)
_, now_allow_pred_proba, _ = self_pred(now_allow_B, [now_allow_B.shape[0]], A, pi)
now_pred_proba[now_allow_flag == 1] = now_allow_pred_proba
pred_proba[begin_index:end_index] = now_pred_proba
return pred_proba
|
[
"XGB_HMM.form_B_matrix_by_XGB.form_B_matrix_by_XGB",
"numpy.zeros",
"XGB_HMM.predict.self_pred",
"public_tool.form_index.form_index"
] |
[((397, 429), 'numpy.zeros', 'np.zeros', (['(O.shape[0], n_states)'], {}), '((O.shape[0], n_states))\n', (405, 429), True, 'import numpy as np\n'), ((501, 523), 'public_tool.form_index.form_index', 'form_index', (['lengths', 'i'], {}), '(lengths, i)\n', (511, 523), False, 'from public_tool.form_index import form_index\n'), ((657, 693), 'numpy.zeros', 'np.zeros', (['(now_O.shape[0], n_states)'], {}), '((now_O.shape[0], n_states))\n', (665, 693), True, 'import numpy as np\n'), ((719, 778), 'XGB_HMM.form_B_matrix_by_XGB.form_B_matrix_by_XGB', 'form_B_matrix_by_XGB', (['model', 'now_O[now_allow_flag == 1]', 'pi'], {}), '(model, now_O[now_allow_flag == 1], pi)\n', (739, 778), False, 'from XGB_HMM.form_B_matrix_by_XGB import form_B_matrix_by_XGB\n'), ((817, 870), 'XGB_HMM.predict.self_pred', 'self_pred', (['now_allow_B', '[now_allow_B.shape[0]]', 'A', 'pi'], {}), '(now_allow_B, [now_allow_B.shape[0]], A, pi)\n', (826, 870), False, 'from XGB_HMM.predict import self_pred\n')]
|
"""
Software that detects each of the yellow shapes on the video frames and
classifies the shapes into classes: circle, rectangle, triangle.
USAGE: python3 shape_detection.py <video path> <output video path>
"""
import sys
import cv2
import imutils
import numpy as np
from tqdm import tqdm
BOX_COLORS = {
"triangle": (255, 0, 0),
"rectangle": (0, 255, 0),
"circle" : (0, 0, 255)
}
def get_contours(image: np.ndarray) -> (np.ndarray, np.ndarray):
"""
Gets edge and yellow contours from an image.
Parameters
----------
image: np.ndarray
Target image.
Returns
-------
edges_filled: np.ndarray
Detected edges in an image as a boolean 2D map.
yellow_contours: np.ndarray
Detected yellow contours in an image.
"""
# get edges
edges = cv2.Canny(image, 10, 255)
kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (3, 3))
edges_thresh = cv2.morphologyEx(edges, cv2.MORPH_CLOSE, kernel)
edges_filled = np.zeros_like(edges_thresh)
edges_contours = cv2.findContours(edges_thresh.copy(), cv2.RETR_EXTERNAL,
cv2.CHAIN_APPROX_SIMPLE)
edges_contours = imutils.grab_contours(edges_contours)
for cont in edges_contours:
cv2.drawContours(edges_filled, [cont], 0, 255, -1)
# select yellow color
hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)
yellow_lower = np.array([30, 10, 10])
yellow_upper = np.array([90, 255, 255])
mask_yellow = cv2.inRange(hsv, yellow_lower, yellow_upper)
yellow_output = cv2.bitwise_and(image, image, mask=mask_yellow)
gray = cv2.cvtColor(yellow_output, cv2.COLOR_BGR2GRAY)
thresh = cv2.threshold(gray, 60, 255, cv2.THRESH_BINARY)[1]
yellow_contours = cv2.findContours(thresh.copy(), cv2.RETR_EXTERNAL,
cv2.CHAIN_APPROX_SIMPLE)
yellow_contours = imutils.grab_contours(yellow_contours)
return edges_filled, yellow_contours
def detect(tcontour: np.ndarray) -> str:
"""
Detects shape by a contour.
Parameters
----------
tcontour: np.ndarray
Target contour.
Returns
-------
shape: str
Detected shape of a contour.
"""
shape = "unidentified"
peri = cv2.arcLength(tcontour, True)
approx = cv2.approxPolyDP(tcontour, 0.04 * peri, True)
if len(approx) == 3:
shape = "triangle"
elif len(approx) == 4:
shape = "rectangle"
else:
shape = "circle"
return shape
if __name__ == "__main__":
if len(sys.argv) != 3:
print(f'USAGE: python3 {sys.argv[0]} <video path> <output video path>')
sys.exit()
VIDEO_PATH = sys.argv[1]
OUTPUT_PATH = sys.argv[2]
# open input and output videos
cap = cv2.VideoCapture(VIDEO_PATH)
width = int(cap.get(3))
height = int(cap.get(4))
fps = cap.get(5)
frame_count = int(cap.get(7))
out = cv2.VideoWriter(OUTPUT_PATH, -1, fps, (width, height))
if not cap.isOpened() or not out.isOpened():
raise RuntimeError("video file cannot be opened")
print(f'\nVideo "{VIDEO_PATH}" opened for processing\n')
for i in tqdm(range(frame_count), desc="Processing video"):
ret, frame = cap.read()
if ret is True:
# preprocess frame
resized = imutils.resize(frame, width=300)
ratio = frame.shape[0] / float(resized.shape[0])
blurred = cv2.GaussianBlur(resized, (3, 3), 0)
edges_map, contours = get_contours(blurred)
for contour in contours:
if cv2.contourArea(contour) > 50:
# get contour's center
M = cv2.moments(contour)
rel_cX = int(M["m10"] / M["m00"])
rel_cY = int(M["m01"] / M["m00"])
if edges_map[rel_cY, rel_cX]:
SHAPE = detect(contour)
# get exact coordinates
contour = contour.astype("float")
contour *= ratio
cX = int((M["m10"] / M["m00"]) * ratio)
cY = int((M["m01"] / M["m00"]) * ratio)
contour = contour.astype("int")
# draw contour and shape name
cv2.drawContours(frame, [contour], -1, BOX_COLORS[SHAPE], 2)
cv2.putText(frame, SHAPE, (cX, cY), cv2.FONT_HERSHEY_SIMPLEX,
0.5, (255, 255, 255), 2)
out.write(frame)
else:
break
cap.release()
out.release()
print(f'\nVideo "{OUTPUT_PATH}" successfully saved')
|
[
"numpy.array",
"cv2.approxPolyDP",
"sys.exit",
"cv2.threshold",
"cv2.arcLength",
"numpy.zeros_like",
"cv2.VideoWriter",
"cv2.contourArea",
"imutils.grab_contours",
"cv2.drawContours",
"cv2.putText",
"cv2.morphologyEx",
"cv2.cvtColor",
"cv2.moments",
"cv2.Canny",
"cv2.GaussianBlur",
"cv2.inRange",
"cv2.bitwise_and",
"imutils.resize",
"cv2.VideoCapture",
"cv2.getStructuringElement"
] |
[((821, 846), 'cv2.Canny', 'cv2.Canny', (['image', '(10)', '(255)'], {}), '(image, 10, 255)\n', (830, 846), False, 'import cv2\n'), ((860, 912), 'cv2.getStructuringElement', 'cv2.getStructuringElement', (['cv2.MORPH_ELLIPSE', '(3, 3)'], {}), '(cv2.MORPH_ELLIPSE, (3, 3))\n', (885, 912), False, 'import cv2\n'), ((932, 980), 'cv2.morphologyEx', 'cv2.morphologyEx', (['edges', 'cv2.MORPH_CLOSE', 'kernel'], {}), '(edges, cv2.MORPH_CLOSE, kernel)\n', (948, 980), False, 'import cv2\n'), ((1000, 1027), 'numpy.zeros_like', 'np.zeros_like', (['edges_thresh'], {}), '(edges_thresh)\n', (1013, 1027), True, 'import numpy as np\n'), ((1190, 1227), 'imutils.grab_contours', 'imutils.grab_contours', (['edges_contours'], {}), '(edges_contours)\n', (1211, 1227), False, 'import imutils\n'), ((1356, 1394), 'cv2.cvtColor', 'cv2.cvtColor', (['image', 'cv2.COLOR_BGR2HSV'], {}), '(image, cv2.COLOR_BGR2HSV)\n', (1368, 1394), False, 'import cv2\n'), ((1414, 1436), 'numpy.array', 'np.array', (['[30, 10, 10]'], {}), '([30, 10, 10])\n', (1422, 1436), True, 'import numpy as np\n'), ((1456, 1480), 'numpy.array', 'np.array', (['[90, 255, 255]'], {}), '([90, 255, 255])\n', (1464, 1480), True, 'import numpy as np\n'), ((1499, 1543), 'cv2.inRange', 'cv2.inRange', (['hsv', 'yellow_lower', 'yellow_upper'], {}), '(hsv, yellow_lower, yellow_upper)\n', (1510, 1543), False, 'import cv2\n'), ((1564, 1611), 'cv2.bitwise_and', 'cv2.bitwise_and', (['image', 'image'], {'mask': 'mask_yellow'}), '(image, image, mask=mask_yellow)\n', (1579, 1611), False, 'import cv2\n'), ((1623, 1670), 'cv2.cvtColor', 'cv2.cvtColor', (['yellow_output', 'cv2.COLOR_BGR2GRAY'], {}), '(yellow_output, cv2.COLOR_BGR2GRAY)\n', (1635, 1670), False, 'import cv2\n'), ((1894, 1932), 'imutils.grab_contours', 'imutils.grab_contours', (['yellow_contours'], {}), '(yellow_contours)\n', (1915, 1932), False, 'import imutils\n'), ((2259, 2288), 'cv2.arcLength', 'cv2.arcLength', (['tcontour', '(True)'], {}), '(tcontour, True)\n', (2272, 2288), False, 'import cv2\n'), ((2302, 2347), 'cv2.approxPolyDP', 'cv2.approxPolyDP', (['tcontour', '(0.04 * peri)', '(True)'], {}), '(tcontour, 0.04 * peri, True)\n', (2318, 2347), False, 'import cv2\n'), ((2767, 2795), 'cv2.VideoCapture', 'cv2.VideoCapture', (['VIDEO_PATH'], {}), '(VIDEO_PATH)\n', (2783, 2795), False, 'import cv2\n'), ((2918, 2972), 'cv2.VideoWriter', 'cv2.VideoWriter', (['OUTPUT_PATH', '(-1)', 'fps', '(width, height)'], {}), '(OUTPUT_PATH, -1, fps, (width, height))\n', (2933, 2972), False, 'import cv2\n'), ((1268, 1318), 'cv2.drawContours', 'cv2.drawContours', (['edges_filled', '[cont]', '(0)', '(255)', '(-1)'], {}), '(edges_filled, [cont], 0, 255, -1)\n', (1284, 1318), False, 'import cv2\n'), ((1684, 1731), 'cv2.threshold', 'cv2.threshold', (['gray', '(60)', '(255)', 'cv2.THRESH_BINARY'], {}), '(gray, 60, 255, cv2.THRESH_BINARY)\n', (1697, 1731), False, 'import cv2\n'), ((2651, 2661), 'sys.exit', 'sys.exit', ([], {}), '()\n', (2659, 2661), False, 'import sys\n'), ((3316, 3348), 'imutils.resize', 'imutils.resize', (['frame'], {'width': '(300)'}), '(frame, width=300)\n', (3330, 3348), False, 'import imutils\n'), ((3432, 3468), 'cv2.GaussianBlur', 'cv2.GaussianBlur', (['resized', '(3, 3)', '(0)'], {}), '(resized, (3, 3), 0)\n', (3448, 3468), False, 'import cv2\n'), ((3581, 3605), 'cv2.contourArea', 'cv2.contourArea', (['contour'], {}), '(contour)\n', (3596, 3605), False, 'import cv2\n'), ((3680, 3700), 'cv2.moments', 'cv2.moments', (['contour'], {}), '(contour)\n', (3691, 3700), False, 'import cv2\n'), ((4318, 4378), 'cv2.drawContours', 'cv2.drawContours', (['frame', '[contour]', '(-1)', 'BOX_COLORS[SHAPE]', '(2)'], {}), '(frame, [contour], -1, BOX_COLORS[SHAPE], 2)\n', (4334, 4378), False, 'import cv2\n'), ((4403, 4494), 'cv2.putText', 'cv2.putText', (['frame', 'SHAPE', '(cX, cY)', 'cv2.FONT_HERSHEY_SIMPLEX', '(0.5)', '(255, 255, 255)', '(2)'], {}), '(frame, SHAPE, (cX, cY), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, \n 255, 255), 2)\n', (4414, 4494), False, 'import cv2\n')]
|
# -*- coding: utf-8 -*-
# Copyright 2017, IBM.
#
# This source code is licensed under the Apache License, Version 2.0 found in
# the LICENSE.txt file in the root directory of this source tree.
# pylint: disable=invalid-name,missing-docstring
from test.python.common import QiskitTestCase
import json
import unittest
import numpy as np
from numpy.linalg import norm
import qiskit
import qiskit._compiler
from qiskit import ClassicalRegister
from qiskit import QuantumCircuit
from qiskit import QuantumJob
from qiskit import QuantumRegister
from qiskit.backends.local.qasm_simulator_cpp import (QasmSimulatorCpp,
cx_error_matrix,
x90_error_matrix)
class TestLocalQasmSimulatorCpp(QiskitTestCase):
"""
Test job_processor module.
"""
def setUp(self):
self.seed = 88
self.qasm_filename = self._get_resource_path('qasm/example.qasm')
with open(self.qasm_filename, 'r') as qasm_file:
self.qasm_text = qasm_file.read()
self.qasm_ast = qiskit.qasm.Qasm(data=self.qasm_text).parse()
self.qasm_be = qiskit.unroll.CircuitBackend(['u1', 'u2', 'u3', 'id', 'cx'])
self.qasm_circ = qiskit.unroll.Unroller(self.qasm_ast, self.qasm_be).execute()
qr = QuantumRegister(2, 'q')
cr = ClassicalRegister(2, 'c')
qc = QuantumCircuit(qr, cr)
qc.h(qr[0])
qc.measure(qr[0], cr[0])
self.qc = qc
# create qobj
compiled_circuit1 = qiskit._compiler.compile_circuit(self.qc, format='json')
compiled_circuit2 = qiskit._compiler.compile_circuit(self.qasm_circ, format='json')
self.qobj = {'id': 'test_qobj',
'config': {
'max_credits': 3,
'shots': 2000,
'backend_name': 'local_qasm_simulator_cpp',
'seed': 1111
},
'circuits': [
{
'name': 'test_circuit1',
'compiled_circuit': compiled_circuit1,
'basis_gates': 'u1,u2,u3,cx,id',
'layout': None,
},
{
'name': 'test_circuit2',
'compiled_circuit': compiled_circuit2,
'basis_gates': 'u1,u2,u3,cx,id',
'layout': None,
}
]}
# Simulator backend
try:
self.backend = QasmSimulatorCpp()
except FileNotFoundError as fnferr:
raise unittest.SkipTest(
'cannot find {} in path'.format(fnferr))
self.q_job = QuantumJob(self.qobj,
backend=self.backend,
preformatted=True)
def test_x90_coherent_error_matrix(self):
X90 = np.array([[1, -1j], [-1j, 1]]) / np.sqrt(2)
U = x90_error_matrix(0., 0.).dot(X90)
target = X90
self.assertAlmostEqual(norm(U - target), 0.0, places=10,
msg="identity error matrix")
U = x90_error_matrix(np.pi / 2., 0.).dot(X90)
target = -1j * np.array([[0, 1], [1, 0]])
self.assertAlmostEqual(norm(U - target), 0.0, places=10)
U = x90_error_matrix(0., np.pi / 2.).dot(X90)
target = np.array([[1., -1], [1, 1.]]) / np.sqrt(2.)
self.assertAlmostEqual(norm(U - target), 0.0, places=10)
U = x90_error_matrix(np.pi / 2, np.pi / 2.).dot(X90)
target = np.array([[0., -1], [1, 0.]])
self.assertAlmostEqual(norm(U - target), 0.0, places=10)
U = x90_error_matrix(0.02, -0.03)
self.assertAlmostEqual(norm(U.dot(U.conj().T) - np.eye(2)), 0.0,
places=10, msg="Test error matrix is unitary")
def test_cx_coherent_error_matrix(self):
CX = np.array([[1, 0, 0, 0], [0, 0, 0, 1], [0, 0, 1, 0], [0, 1, 0, 0]])
U = cx_error_matrix(0., 0.).dot(CX)
target = CX
self.assertAlmostEqual(norm(U - target), 0.0, places=10,
msg="identity error matrix")
U = cx_error_matrix(np.pi / 2., 0.).dot(CX)
target = np.array([[1, 0, 1j, 0],
[0, -1j, 0, 1],
[1j, 0, 1, 0],
[0, 1, 0, -1j]]) / np.sqrt(2)
self.assertAlmostEqual(norm(U - target), 0.0, places=10)
U = cx_error_matrix(0.03, -0.04)
self.assertAlmostEqual(norm(U.dot(U.conj().T) - np.eye(4)), 0.0,
places=10, msg="Test error matrix is unitary")
def test_run_qobj(self):
result = self.backend.run(self.q_job).result()
shots = self.qobj['config']['shots']
threshold = 0.04 * shots
counts = result.get_counts('test_circuit2')
target = {'100 100': shots / 8, '011 011': shots / 8,
'101 101': shots / 8, '111 111': shots / 8,
'000 000': shots / 8, '010 010': shots / 8,
'110 110': shots / 8, '001 001': shots / 8}
self.assertDictAlmostEqual(counts, target, threshold)
def test_qobj_measure_opt(self):
filename = self._get_resource_path('qobj/cpp_measure_opt.json')
with open(filename, 'r') as file:
q_job = QuantumJob(json.load(file),
backend=self.backend,
preformatted=True)
result = self.backend.run(q_job).result()
shots = q_job.qobj['config']['shots']
expected_data = {
'measure (opt)': {
'deterministic': True,
'counts': {'00': shots},
'statevector': np.array([1, 0, 0, 0])},
'x0 measure (opt)': {
'deterministic': True,
'counts': {'01': shots},
'statevector': np.array([0, 1, 0, 0])},
'x1 measure (opt)': {
'deterministic': True,
'counts': {'10': shots},
'statevector': np.array([0, 0, 1, 0])},
'x0 x1 measure (opt)': {
'deterministic': True,
'counts': {'11': shots},
'statevector': np.array([0, 0, 0, 1])},
'y0 measure (opt)': {
'deterministic': True,
'counts': {'01': shots},
'statevector': np.array([0, 1j, 0, 0])},
'y1 measure (opt)': {
'deterministic': True,
'counts': {'10': shots},
'statevector': np.array([0, 0, 1j, 0])},
'y0 y1 measure (opt)': {
'deterministic': True,
'counts': {'11': shots},
'statevector': np.array([0, 0, 0, -1j])},
'h0 measure (opt)': {
'deterministic': False,
'counts': {'00': shots / 2, '01': shots / 2},
'statevector': np.array([1 / np.sqrt(2), 1 / np.sqrt(2), 0, 0])},
'h1 measure (opt)': {
'deterministic': False,
'counts': {'00': shots / 2, '10': shots / 2},
'statevector': np.array([1 / np.sqrt(2), 0, 1 / np.sqrt(2), 0])},
'h0 h1 measure (opt)': {
'deterministic': False,
'counts': {'00': shots / 4, '01': shots / 4,
'10': shots / 4, '11': shots / 4},
'statevector': np.array([0.5, 0.5, 0.5, 0.5])},
'bell measure (opt)': {
'deterministic': False,
'counts': {'00': shots / 2, '11': shots / 2},
'statevector': np.array([1 / np.sqrt(2), 0, 0, 1 / np.sqrt(2)])}
}
for name in expected_data:
# Check counts:
counts = result.get_counts(name)
expected_counts = expected_data[name]['counts']
if expected_data[name].get('deterministic', False):
self.assertEqual(counts, expected_counts,
msg=name + ' counts')
else:
threshold = 0.04 * shots
self.assertDictAlmostEqual(counts, expected_counts,
threshold, msg=name + 'counts')
# Check snapshot
snapshots = result.get_snapshots(name)
self.assertEqual(set(snapshots), {'0'},
msg=name + ' snapshot keys')
self.assertEqual(len(snapshots['0']), 3,
msg=name + ' snapshot length')
state = snapshots['0']['statevector'][0]
expected_state = expected_data[name]['statevector']
fidelity = np.abs(expected_state.dot(state.conj())) ** 2
self.assertAlmostEqual(fidelity, 1.0, places=10,
msg=name + ' snapshot fidelity')
rho = snapshots['0']['density_matrix']
self.assertAlmostEqual(np.trace(rho), 1)
prob = snapshots['0']['probabilities']
self.assertAlmostEqual(np.sum(prob), 1)
def test_qobj_measure_opt_flag(self):
filename = self._get_resource_path('qobj/cpp_measure_opt_flag.json')
with open(filename, 'r') as file:
q_job = QuantumJob(json.load(file),
backend=self.backend,
preformatted=True)
result = self.backend.run(q_job).result()
shots = q_job.qobj['config']['shots']
sampled_measurements = {
'measure (sampled)': True,
'trivial (sampled)': True,
'reset1 (shots)': False,
'reset2 (shots)': False,
'reset3 (shots)': False,
'gate1 (shots)': False,
'gate2 (shots)': False,
'gate3 (shots)': False,
'gate4 (shots)': False
}
for name in sampled_measurements:
snapshots = result.get_snapshots(name)
# Check snapshot keys
self.assertEqual(set(snapshots), {'0'},
msg=name + ' snapshot keys')
# Check number of snapshots
# there should be 1 for measurement sampling optimization
# and there should be >1 for each shot beign simulated.
num_snapshots = len(snapshots['0'].get('statevector', []))
if sampled_measurements[name] is True:
self.assertEqual(num_snapshots, 1,
msg=name + ' snapshot length')
else:
self.assertEqual(num_snapshots, shots,
msg=name + ' snapshot length')
def test_qobj_reset(self):
filename = self._get_resource_path('qobj/cpp_reset.json')
with open(filename, 'r') as file:
q_job = QuantumJob(json.load(file),
backend=self.backend,
preformatted=True)
result = self.backend.run(q_job).result()
expected_data = {
'reset': {'statevector': np.array([1, 0])},
'x reset': {'statevector': np.array([1, 0])},
'y reset': {'statevector': np.array([1, 0])},
'h reset': {'statevector': np.array([1, 0])}
}
for name in expected_data:
# Check snapshot is |0> state
snapshots = result.get_snapshots(name)
self.assertEqual(set(snapshots), {'0'},
msg=name + ' snapshot keys')
self.assertEqual(len(snapshots['0']), 1,
msg=name + ' snapshot length')
state = snapshots['0']['statevector'][0]
expected_state = expected_data[name]['statevector']
fidelity = np.abs(expected_state.dot(state.conj())) ** 2
self.assertAlmostEqual(fidelity, 1.0, places=10,
msg=name + ' snapshot fidelity')
def test_qobj_save_load(self):
filename = self._get_resource_path('qobj/cpp_save_load.json')
with open(filename, 'r') as file:
q_job = QuantumJob(json.load(file),
backend=self.backend,
preformatted=True)
result = self.backend.run(q_job).result()
snapshots = result.get_snapshots('save_command')
self.assertEqual(set(snapshots), {'0', '1', '10', '11'},
msg='snapshot keys')
state0 = snapshots['0']['statevector'][0]
state10 = snapshots['10']['statevector'][0]
state1 = snapshots['1']['statevector'][0]
state11 = snapshots['11']['statevector'][0]
expected_state0 = np.array([1, 0])
expected_state10 = np.array([1 / np.sqrt(2), 1 / np.sqrt(2)])
fidelity0 = np.abs(expected_state0.dot(state0.conj())) ** 2
fidelity1 = np.abs(expected_state0.dot(state1.conj())) ** 2
fidelity10 = np.abs(expected_state10.dot(state10.conj())) ** 2
fidelity11 = np.abs(expected_state10.dot(state11.conj())) ** 2
self.assertAlmostEqual(fidelity0, 1.0, places=10, msg='snapshot 0')
self.assertAlmostEqual(fidelity10, 1.0, places=10, msg='snapshot 0')
self.assertAlmostEqual(fidelity1, 1.0, places=10, msg='snapshot 0')
self.assertAlmostEqual(fidelity11, 1.0, places=10, msg='snapshot 0')
def test_qobj_single_qubit_gates(self):
filename = self._get_resource_path('qobj/cpp_single_qubit_gates.json')
with open(filename, 'r') as file:
q_job = QuantumJob(json.load(file),
backend=self.backend,
preformatted=True)
result = self.backend.run(q_job).result()
expected_data = {
'snapshot': {
'statevector': np.array([1, 0])},
'id(U)': {
'statevector': np.array([1, 0])},
'id(u3)': {
'statevector': np.array([1, 0])},
'id(u1)': {
'statevector': np.array([1, 0])},
'id(direct)': {
'statevector': np.array([1, 0])},
'x(U)': {
'statevector': np.array([0, 1])},
'x(u3)': {
'statevector': np.array([0, 1])},
'x(direct)': {
'statevector': np.array([0, 1])},
'y(U)': {
'statevector': np.array([0, 1j])},
'y(u3)': {
'statevector': np.array([0, 1j])},
'y(direct)': {
'statevector': np.array([0, 1j])},
'h(U)': {
'statevector': np.array([1 / np.sqrt(2), 1 / np.sqrt(2)])},
'h(u3)': {
'statevector': np.array([1 / np.sqrt(2), 1 / np.sqrt(2)])},
'h(u2)': {
'statevector': np.array([1 / np.sqrt(2), 1 / np.sqrt(2)])},
'h(direct)': {
'statevector': np.array([1 / np.sqrt(2), 1 / np.sqrt(2)])},
'h(direct) z(U)': {
'statevector': np.array([1 / np.sqrt(2), -1 / np.sqrt(2)])},
'h(direct) z(u3)': {
'statevector': np.array([1 / np.sqrt(2), -1 / np.sqrt(2)])},
'h(direct) z(u1)': {
'statevector': np.array([1 / np.sqrt(2), -1 / np.sqrt(2)])},
'h(direct) z(direct)': {
'statevector': np.array([1 / np.sqrt(2), -1 / np.sqrt(2)])},
'h(direct) s(U)': {
'statevector': np.array([1 / np.sqrt(2), 1j / np.sqrt(2)])},
'h(direct) s(u3)': {
'statevector': np.array([1 / np.sqrt(2), 1j / np.sqrt(2)])},
'h(direct) s(u1)': {
'statevector': np.array([1 / np.sqrt(2), 1j / np.sqrt(2)])},
'h(direct) s(direct)': {
'statevector': np.array([1 / np.sqrt(2), 1j / np.sqrt(2)])},
'h(direct) sdg(U)': {
'statevector': np.array([1 / np.sqrt(2), -1j / np.sqrt(2)])},
'h(direct) sdg(u3)': {
'statevector': np.array([1 / np.sqrt(2), -1j / np.sqrt(2)])},
'h(direct) sdg(u1)': {
'statevector': np.array([1 / np.sqrt(2), -1j / np.sqrt(2)])},
'h(direct) sdg(direct)': {
'statevector': np.array([1 / np.sqrt(2), -1j / np.sqrt(2)])},
'h(direct) t(U)': {
'statevector': np.array([1 / np.sqrt(2), 0.5 + 0.5j])},
'h(direct) t(u3)': {
'statevector': np.array([1 / np.sqrt(2), 0.5 + 0.5j])},
'h(direct) t(u1)': {
'statevector': np.array([1 / np.sqrt(2), 0.5 + 0.5j])},
'h(direct) t(direct)': {
'statevector': np.array([1 / np.sqrt(2), 0.5 + 0.5j])},
'h(direct) tdg(U)': {
'statevector': np.array([1 / np.sqrt(2), 0.5 - 0.5j])},
'h(direct) tdg(u3)': {
'statevector': np.array([1 / np.sqrt(2), 0.5 - 0.5j])},
'h(direct) tdg(u1)': {
'statevector': np.array([1 / np.sqrt(2), 0.5 - 0.5j])},
'h(direct) tdg(direct)': {
'statevector': np.array([1 / np.sqrt(2), 0.5 - 0.5j])}
}
for name in expected_data:
# Check snapshot
snapshots = result.get_snapshots(name)
self.assertEqual(set(snapshots), {'0'},
msg=name + ' snapshot keys')
self.assertEqual(len(snapshots['0']), 1,
msg=name + ' snapshot length')
state = snapshots['0']['statevector'][0]
expected_state = expected_data[name]['statevector']
inner_product = expected_state.dot(state.conj())
self.assertAlmostEqual(inner_product, 1.0, places=10,
msg=name + ' snapshot fidelity')
def test_qobj_two_qubit_gates(self):
filename = self._get_resource_path('qobj/cpp_two_qubit_gates.json')
with open(filename, 'r') as file:
q_job = QuantumJob(json.load(file),
backend=self.backend,
preformatted=True)
result = self.backend.run(q_job).result()
expected_data = {
'h0 CX01': {
'statevector': np.array([1 / np.sqrt(2), 0, 0, 1 / np.sqrt(2)])},
'h0 CX10': {
'statevector': np.array([1 / np.sqrt(2), 1 / np.sqrt(2), 0, 0])},
'h1 CX01': {
'statevector': np.array([1 / np.sqrt(2), 0, 1 / np.sqrt(2), 0])},
'h1 CX10': {
'statevector': np.array([1 / np.sqrt(2), 0, 0, 1 / np.sqrt(2)])},
'h0 cx01': {
'statevector': np.array([1 / np.sqrt(2), 0, 0, 1 / np.sqrt(2)])},
'h0 cx10': {
'statevector': np.array([1 / np.sqrt(2), 1 / np.sqrt(2), 0, 0])},
'h1 cx01': {
'statevector': np.array([1 / np.sqrt(2), 0, 1 / np.sqrt(2), 0])},
'h1 cx10': {
'statevector': np.array([1 / np.sqrt(2), 0, 0, 1 / np.sqrt(2)])},
'h0 cz01': {
'statevector': np.array([1 / np.sqrt(2), 1 / np.sqrt(2), 0, 0])},
'h0 cz10': {
'statevector': np.array([1 / np.sqrt(2), 1 / np.sqrt(2), 0, 0])},
'h1 cz01': {
'statevector': np.array([1 / np.sqrt(2), 0, 1 / np.sqrt(2), 0])},
'h1 cz10': {
'statevector': np.array([1 / np.sqrt(2), 0, 1 / np.sqrt(2), 0])},
'h0 h1 cz01': {'statevector': np.array([0.5, 0.5, 0.5, -0.5])},
'h0 h1 cz10': {'statevector': np.array([0.5, 0.5, 0.5, -0.5])},
'h0 rzz01': {
'statevector': np.array([1 / np.sqrt(2), 1j / np.sqrt(2), 0, 0])},
'h0 rzz10': {
'statevector': np.array([1 / np.sqrt(2), 1j / np.sqrt(2), 0, 0])},
'h1 rzz01': {
'statevector': np.array([1 / np.sqrt(2), 0, 1j / np.sqrt(2), 0])},
'h1 rzz10': {
'statevector': np.array([1 / np.sqrt(2), 0, 1j / np.sqrt(2), 0])},
'h0 h1 rzz01': {'statevector': np.array([0.5, 0.5j, 0.5j, 0.5])},
'h0 h1 rzz10': {'statevector': np.array([0.5, 0.5j, 0.5j, 0.5])}
}
for name in expected_data:
# Check snapshot
snapshots = result.get_snapshots(name)
self.assertEqual(set(snapshots), {'0'},
msg=name + ' snapshot keys')
self.assertEqual(len(snapshots['0']), 1,
msg=name + ' snapshot length')
state = snapshots['0']['statevector'][0]
expected_state = expected_data[name]['statevector']
fidelity = np.abs(expected_state.dot(state.conj())) ** 2
self.assertAlmostEqual(fidelity, 1.0, places=10,
msg=name + ' snapshot fidelity')
def test_conditionals(self):
filename = self._get_resource_path('qobj/cpp_conditionals.json')
with open(filename, 'r') as file:
q_job = QuantumJob(json.load(file),
backend=self.backend,
preformatted=True)
result = self.backend.run(q_job).result()
expected_data = {
'single creg (c0=0)': {
'statevector': np.array([1, 0, 0, 0])},
'single creg (c0=1)': {
'statevector': np.array([0, 0, 0, 1])},
'two creg (c1=0)': {
'statevector': np.array([1, 0, 0, 0])},
'two creg (c1=1)': {
'statevector': np.array([0, 0, 0, 1])}
}
for name in expected_data:
# Check snapshot
snapshots = result.get_snapshots(name)
self.assertEqual(set(snapshots), {'0'},
msg=name + ' snapshot keys')
self.assertEqual(len(snapshots['0']), 1,
msg=name + ' snapshot length')
state = snapshots['0']['statevector'][0]
expected_state = expected_data[name]['statevector']
fidelity = np.abs(expected_state.dot(state.conj())) ** 2
self.assertAlmostEqual(fidelity, 1.0, places=10,
msg=name + ' snapshot fidelity')
if __name__ == '__main__':
unittest.main(verbosity=2)
|
[
"numpy.trace",
"numpy.sqrt",
"numpy.array",
"qiskit.qasm.Qasm",
"numpy.linalg.norm",
"unittest.main",
"qiskit.QuantumJob",
"qiskit.backends.local.qasm_simulator_cpp.x90_error_matrix",
"qiskit.QuantumCircuit",
"qiskit.backends.local.qasm_simulator_cpp.cx_error_matrix",
"numpy.eye",
"qiskit._compiler.compile_circuit",
"qiskit.unroll.CircuitBackend",
"qiskit.ClassicalRegister",
"qiskit.backends.local.qasm_simulator_cpp.QasmSimulatorCpp",
"json.load",
"numpy.sum",
"qiskit.unroll.Unroller",
"qiskit.QuantumRegister"
] |
[((22521, 22547), 'unittest.main', 'unittest.main', ([], {'verbosity': '(2)'}), '(verbosity=2)\n', (22534, 22547), False, 'import unittest\n'), ((1345, 1368), 'qiskit.QuantumRegister', 'QuantumRegister', (['(2)', '"""q"""'], {}), "(2, 'q')\n", (1360, 1368), False, 'from qiskit import QuantumRegister\n'), ((1382, 1407), 'qiskit.ClassicalRegister', 'ClassicalRegister', (['(2)', '"""c"""'], {}), "(2, 'c')\n", (1399, 1407), False, 'from qiskit import ClassicalRegister\n'), ((1421, 1443), 'qiskit.QuantumCircuit', 'QuantumCircuit', (['qr', 'cr'], {}), '(qr, cr)\n', (1435, 1443), False, 'from qiskit import QuantumCircuit\n'), ((1568, 1624), 'qiskit._compiler.compile_circuit', 'qiskit._compiler.compile_circuit', (['self.qc'], {'format': '"""json"""'}), "(self.qc, format='json')\n", (1600, 1624), False, 'import qiskit\n'), ((1653, 1716), 'qiskit._compiler.compile_circuit', 'qiskit._compiler.compile_circuit', (['self.qasm_circ'], {'format': '"""json"""'}), "(self.qasm_circ, format='json')\n", (1685, 1716), False, 'import qiskit\n'), ((2877, 2939), 'qiskit.QuantumJob', 'QuantumJob', (['self.qobj'], {'backend': 'self.backend', 'preformatted': '(True)'}), '(self.qobj, backend=self.backend, preformatted=True)\n', (2887, 2939), False, 'from qiskit import QuantumJob\n'), ((3728, 3759), 'numpy.array', 'np.array', (['[[0.0, -1], [1, 0.0]]'], {}), '([[0.0, -1], [1, 0.0]])\n', (3736, 3759), True, 'import numpy as np\n'), ((3835, 3864), 'qiskit.backends.local.qasm_simulator_cpp.x90_error_matrix', 'x90_error_matrix', (['(0.02)', '(-0.03)'], {}), '(0.02, -0.03)\n', (3851, 3864), False, 'from qiskit.backends.local.qasm_simulator_cpp import QasmSimulatorCpp, cx_error_matrix, x90_error_matrix\n'), ((4075, 4141), 'numpy.array', 'np.array', (['[[1, 0, 0, 0], [0, 0, 0, 1], [0, 0, 1, 0], [0, 1, 0, 0]]'], {}), '([[1, 0, 0, 0], [0, 0, 0, 1], [0, 0, 1, 0], [0, 1, 0, 0]])\n', (4083, 4141), True, 'import numpy as np\n'), ((4644, 4672), 'qiskit.backends.local.qasm_simulator_cpp.cx_error_matrix', 'cx_error_matrix', (['(0.03)', '(-0.04)'], {}), '(0.03, -0.04)\n', (4659, 4672), False, 'from qiskit.backends.local.qasm_simulator_cpp import QasmSimulatorCpp, cx_error_matrix, x90_error_matrix\n'), ((12868, 12884), 'numpy.array', 'np.array', (['[1, 0]'], {}), '([1, 0])\n', (12876, 12884), True, 'import numpy as np\n'), ((1180, 1240), 'qiskit.unroll.CircuitBackend', 'qiskit.unroll.CircuitBackend', (["['u1', 'u2', 'u3', 'id', 'cx']"], {}), "(['u1', 'u2', 'u3', 'id', 'cx'])\n", (1208, 1240), False, 'import qiskit\n'), ((2698, 2716), 'qiskit.backends.local.qasm_simulator_cpp.QasmSimulatorCpp', 'QasmSimulatorCpp', ([], {}), '()\n', (2714, 2716), False, 'from qiskit.backends.local.qasm_simulator_cpp import QasmSimulatorCpp, cx_error_matrix, x90_error_matrix\n'), ((3065, 3099), 'numpy.array', 'np.array', (['[[1, -1.0j], [-1.0j, 1]]'], {}), '([[1, -1.0j], [-1.0j, 1]])\n', (3073, 3099), True, 'import numpy as np\n'), ((3098, 3108), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (3105, 3108), True, 'import numpy as np\n'), ((3207, 3223), 'numpy.linalg.norm', 'norm', (['(U - target)'], {}), '(U - target)\n', (3211, 3223), False, 'from numpy.linalg import norm\n'), ((3378, 3404), 'numpy.array', 'np.array', (['[[0, 1], [1, 0]]'], {}), '([[0, 1], [1, 0]])\n', (3386, 3404), True, 'import numpy as np\n'), ((3436, 3452), 'numpy.linalg.norm', 'norm', (['(U - target)'], {}), '(U - target)\n', (3440, 3452), False, 'from numpy.linalg import norm\n'), ((3541, 3572), 'numpy.array', 'np.array', (['[[1.0, -1], [1, 1.0]]'], {}), '([[1.0, -1], [1, 1.0]])\n', (3549, 3572), True, 'import numpy as np\n'), ((3573, 3585), 'numpy.sqrt', 'np.sqrt', (['(2.0)'], {}), '(2.0)\n', (3580, 3585), True, 'import numpy as np\n'), ((3616, 3632), 'numpy.linalg.norm', 'norm', (['(U - target)'], {}), '(U - target)\n', (3620, 3632), False, 'from numpy.linalg import norm\n'), ((3789, 3805), 'numpy.linalg.norm', 'norm', (['(U - target)'], {}), '(U - target)\n', (3793, 3805), False, 'from numpy.linalg import norm\n'), ((4237, 4253), 'numpy.linalg.norm', 'norm', (['(U - target)'], {}), '(U - target)\n', (4241, 4253), False, 'from numpy.linalg import norm\n'), ((4400, 4485), 'numpy.array', 'np.array', (['[[1, 0, 1.0j, 0], [0, -1.0j, 0, 1], [1.0j, 0, 1, 0], [0, 1, 0, -1.0j]]'], {}), '([[1, 0, 1.0j, 0], [0, -1.0j, 0, 1], [1.0j, 0, 1, 0], [0, 1, 0, -1.0j]]\n )\n', (4408, 4485), True, 'import numpy as np\n'), ((4556, 4566), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (4563, 4566), True, 'import numpy as np\n'), ((4598, 4614), 'numpy.linalg.norm', 'norm', (['(U - target)'], {}), '(U - target)\n', (4602, 4614), False, 'from numpy.linalg import norm\n'), ((3121, 3147), 'qiskit.backends.local.qasm_simulator_cpp.x90_error_matrix', 'x90_error_matrix', (['(0.0)', '(0.0)'], {}), '(0.0, 0.0)\n', (3137, 3147), False, 'from qiskit.backends.local.qasm_simulator_cpp import QasmSimulatorCpp, cx_error_matrix, x90_error_matrix\n'), ((3313, 3347), 'qiskit.backends.local.qasm_simulator_cpp.x90_error_matrix', 'x90_error_matrix', (['(np.pi / 2.0)', '(0.0)'], {}), '(np.pi / 2.0, 0.0)\n', (3329, 3347), False, 'from qiskit.backends.local.qasm_simulator_cpp import QasmSimulatorCpp, cx_error_matrix, x90_error_matrix\n'), ((3482, 3516), 'qiskit.backends.local.qasm_simulator_cpp.x90_error_matrix', 'x90_error_matrix', (['(0.0)', '(np.pi / 2.0)'], {}), '(0.0, np.pi / 2.0)\n', (3498, 3516), False, 'from qiskit.backends.local.qasm_simulator_cpp import QasmSimulatorCpp, cx_error_matrix, x90_error_matrix\n'), ((3662, 3702), 'qiskit.backends.local.qasm_simulator_cpp.x90_error_matrix', 'x90_error_matrix', (['(np.pi / 2)', '(np.pi / 2.0)'], {}), '(np.pi / 2, np.pi / 2.0)\n', (3678, 3702), False, 'from qiskit.backends.local.qasm_simulator_cpp import QasmSimulatorCpp, cx_error_matrix, x90_error_matrix\n'), ((4154, 4179), 'qiskit.backends.local.qasm_simulator_cpp.cx_error_matrix', 'cx_error_matrix', (['(0.0)', '(0.0)'], {}), '(0.0, 0.0)\n', (4169, 4179), False, 'from qiskit.backends.local.qasm_simulator_cpp import QasmSimulatorCpp, cx_error_matrix, x90_error_matrix\n'), ((4343, 4376), 'qiskit.backends.local.qasm_simulator_cpp.cx_error_matrix', 'cx_error_matrix', (['(np.pi / 2.0)', '(0.0)'], {}), '(np.pi / 2.0, 0.0)\n', (4358, 4376), False, 'from qiskit.backends.local.qasm_simulator_cpp import QasmSimulatorCpp, cx_error_matrix, x90_error_matrix\n'), ((5532, 5547), 'json.load', 'json.load', (['file'], {}), '(file)\n', (5541, 5547), False, 'import json\n'), ((5916, 5938), 'numpy.array', 'np.array', (['[1, 0, 0, 0]'], {}), '([1, 0, 0, 0])\n', (5924, 5938), True, 'import numpy as np\n'), ((6086, 6108), 'numpy.array', 'np.array', (['[0, 1, 0, 0]'], {}), '([0, 1, 0, 0])\n', (6094, 6108), True, 'import numpy as np\n'), ((6256, 6278), 'numpy.array', 'np.array', (['[0, 0, 1, 0]'], {}), '([0, 0, 1, 0])\n', (6264, 6278), True, 'import numpy as np\n'), ((6429, 6451), 'numpy.array', 'np.array', (['[0, 0, 0, 1]'], {}), '([0, 0, 0, 1])\n', (6437, 6451), True, 'import numpy as np\n'), ((6599, 6624), 'numpy.array', 'np.array', (['[0, 1.0j, 0, 0]'], {}), '([0, 1.0j, 0, 0])\n', (6607, 6624), True, 'import numpy as np\n'), ((6770, 6795), 'numpy.array', 'np.array', (['[0, 0, 1.0j, 0]'], {}), '([0, 0, 1.0j, 0])\n', (6778, 6795), True, 'import numpy as np\n'), ((6944, 6970), 'numpy.array', 'np.array', (['[0, 0, 0, -1.0j]'], {}), '([0, 0, 0, -1.0j])\n', (6952, 6970), True, 'import numpy as np\n'), ((7638, 7668), 'numpy.array', 'np.array', (['[0.5, 0.5, 0.5, 0.5]'], {}), '([0.5, 0.5, 0.5, 0.5])\n', (7646, 7668), True, 'import numpy as np\n'), ((9152, 9165), 'numpy.trace', 'np.trace', (['rho'], {}), '(rho)\n', (9160, 9165), True, 'import numpy as np\n'), ((9256, 9268), 'numpy.sum', 'np.sum', (['prob'], {}), '(prob)\n', (9262, 9268), True, 'import numpy as np\n'), ((9466, 9481), 'json.load', 'json.load', (['file'], {}), '(file)\n', (9475, 9481), False, 'import json\n'), ((11018, 11033), 'json.load', 'json.load', (['file'], {}), '(file)\n', (11027, 11033), False, 'import json\n'), ((11251, 11267), 'numpy.array', 'np.array', (['[1, 0]'], {}), '([1, 0])\n', (11259, 11267), True, 'import numpy as np\n'), ((11309, 11325), 'numpy.array', 'np.array', (['[1, 0]'], {}), '([1, 0])\n', (11317, 11325), True, 'import numpy as np\n'), ((11367, 11383), 'numpy.array', 'np.array', (['[1, 0]'], {}), '([1, 0])\n', (11375, 11383), True, 'import numpy as np\n'), ((11425, 11441), 'numpy.array', 'np.array', (['[1, 0]'], {}), '([1, 0])\n', (11433, 11441), True, 'import numpy as np\n'), ((12298, 12313), 'json.load', 'json.load', (['file'], {}), '(file)\n', (12307, 12313), False, 'import json\n'), ((13737, 13752), 'json.load', 'json.load', (['file'], {}), '(file)\n', (13746, 13752), False, 'import json\n'), ((13990, 14006), 'numpy.array', 'np.array', (['[1, 0]'], {}), '([1, 0])\n', (13998, 14006), True, 'import numpy as np\n'), ((14063, 14079), 'numpy.array', 'np.array', (['[1, 0]'], {}), '([1, 0])\n', (14071, 14079), True, 'import numpy as np\n'), ((14137, 14153), 'numpy.array', 'np.array', (['[1, 0]'], {}), '([1, 0])\n', (14145, 14153), True, 'import numpy as np\n'), ((14211, 14227), 'numpy.array', 'np.array', (['[1, 0]'], {}), '([1, 0])\n', (14219, 14227), True, 'import numpy as np\n'), ((14289, 14305), 'numpy.array', 'np.array', (['[1, 0]'], {}), '([1, 0])\n', (14297, 14305), True, 'import numpy as np\n'), ((14361, 14377), 'numpy.array', 'np.array', (['[0, 1]'], {}), '([0, 1])\n', (14369, 14377), True, 'import numpy as np\n'), ((14434, 14450), 'numpy.array', 'np.array', (['[0, 1]'], {}), '([0, 1])\n', (14442, 14450), True, 'import numpy as np\n'), ((14511, 14527), 'numpy.array', 'np.array', (['[0, 1]'], {}), '([0, 1])\n', (14519, 14527), True, 'import numpy as np\n'), ((14583, 14602), 'numpy.array', 'np.array', (['[0, 1.0j]'], {}), '([0, 1.0j])\n', (14591, 14602), True, 'import numpy as np\n'), ((14657, 14676), 'numpy.array', 'np.array', (['[0, 1.0j]'], {}), '([0, 1.0j])\n', (14665, 14676), True, 'import numpy as np\n'), ((14735, 14754), 'numpy.array', 'np.array', (['[0, 1.0j]'], {}), '([0, 1.0j])\n', (14743, 14754), True, 'import numpy as np\n'), ((18200, 18215), 'json.load', 'json.load', (['file'], {}), '(file)\n', (18209, 18215), False, 'import json\n'), ((19722, 19753), 'numpy.array', 'np.array', (['[0.5, 0.5, 0.5, -0.5]'], {}), '([0.5, 0.5, 0.5, -0.5])\n', (19730, 19753), True, 'import numpy as np\n'), ((19798, 19829), 'numpy.array', 'np.array', (['[0.5, 0.5, 0.5, -0.5]'], {}), '([0.5, 0.5, 0.5, -0.5])\n', (19806, 19829), True, 'import numpy as np\n'), ((20311, 20343), 'numpy.array', 'np.array', (['[0.5, 0.5j, 0.5j, 0.5]'], {}), '([0.5, 0.5j, 0.5j, 0.5])\n', (20319, 20343), True, 'import numpy as np\n'), ((20389, 20421), 'numpy.array', 'np.array', (['[0.5, 0.5j, 0.5j, 0.5]'], {}), '([0.5, 0.5j, 0.5j, 0.5])\n', (20397, 20421), True, 'import numpy as np\n'), ((21267, 21282), 'json.load', 'json.load', (['file'], {}), '(file)\n', (21276, 21282), False, 'import json\n'), ((21530, 21552), 'numpy.array', 'np.array', (['[1, 0, 0, 0]'], {}), '([1, 0, 0, 0])\n', (21538, 21552), True, 'import numpy as np\n'), ((21622, 21644), 'numpy.array', 'np.array', (['[0, 0, 0, 1]'], {}), '([0, 0, 0, 1])\n', (21630, 21644), True, 'import numpy as np\n'), ((21711, 21733), 'numpy.array', 'np.array', (['[1, 0, 0, 0]'], {}), '([1, 0, 0, 0])\n', (21719, 21733), True, 'import numpy as np\n'), ((21800, 21822), 'numpy.array', 'np.array', (['[0, 0, 0, 1]'], {}), '([0, 0, 0, 1])\n', (21808, 21822), True, 'import numpy as np\n'), ((1107, 1144), 'qiskit.qasm.Qasm', 'qiskit.qasm.Qasm', ([], {'data': 'self.qasm_text'}), '(data=self.qasm_text)\n', (1123, 1144), False, 'import qiskit\n'), ((1270, 1321), 'qiskit.unroll.Unroller', 'qiskit.unroll.Unroller', (['self.qasm_ast', 'self.qasm_be'], {}), '(self.qasm_ast, self.qasm_be)\n', (1292, 1321), False, 'import qiskit\n'), ((3921, 3930), 'numpy.eye', 'np.eye', (['(2)'], {}), '(2)\n', (3927, 3930), True, 'import numpy as np\n'), ((4729, 4738), 'numpy.eye', 'np.eye', (['(4)'], {}), '(4)\n', (4735, 4738), True, 'import numpy as np\n'), ((12926, 12936), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (12933, 12936), True, 'import numpy as np\n'), ((12942, 12952), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (12949, 12952), True, 'import numpy as np\n'), ((7152, 7162), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (7159, 7162), True, 'import numpy as np\n'), ((7168, 7178), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (7175, 7178), True, 'import numpy as np\n'), ((7370, 7380), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (7377, 7380), True, 'import numpy as np\n'), ((7389, 7399), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (7396, 7399), True, 'import numpy as np\n'), ((7854, 7864), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (7861, 7864), True, 'import numpy as np\n'), ((7876, 7886), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (7883, 7886), True, 'import numpy as np\n'), ((14822, 14832), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (14829, 14832), True, 'import numpy as np\n'), ((14838, 14848), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (14845, 14848), True, 'import numpy as np\n'), ((14921, 14931), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (14928, 14931), True, 'import numpy as np\n'), ((14937, 14947), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (14944, 14947), True, 'import numpy as np\n'), ((15020, 15030), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15027, 15030), True, 'import numpy as np\n'), ((15036, 15046), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15043, 15046), True, 'import numpy as np\n'), ((15123, 15133), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15130, 15133), True, 'import numpy as np\n'), ((15139, 15149), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15146, 15149), True, 'import numpy as np\n'), ((15231, 15241), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15238, 15241), True, 'import numpy as np\n'), ((15248, 15258), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15255, 15258), True, 'import numpy as np\n'), ((15341, 15351), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15348, 15351), True, 'import numpy as np\n'), ((15358, 15368), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15365, 15368), True, 'import numpy as np\n'), ((15451, 15461), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15458, 15461), True, 'import numpy as np\n'), ((15468, 15478), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15475, 15478), True, 'import numpy as np\n'), ((15565, 15575), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15572, 15575), True, 'import numpy as np\n'), ((15582, 15592), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15589, 15592), True, 'import numpy as np\n'), ((15674, 15684), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15681, 15684), True, 'import numpy as np\n'), ((15691, 15701), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15698, 15701), True, 'import numpy as np\n'), ((15784, 15794), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15791, 15794), True, 'import numpy as np\n'), ((15801, 15811), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15808, 15811), True, 'import numpy as np\n'), ((15894, 15904), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15901, 15904), True, 'import numpy as np\n'), ((15911, 15921), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (15918, 15921), True, 'import numpy as np\n'), ((16008, 16018), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16015, 16018), True, 'import numpy as np\n'), ((16025, 16035), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16032, 16035), True, 'import numpy as np\n'), ((16119, 16129), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16126, 16129), True, 'import numpy as np\n'), ((16137, 16147), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16144, 16147), True, 'import numpy as np\n'), ((16232, 16242), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16239, 16242), True, 'import numpy as np\n'), ((16250, 16260), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16257, 16260), True, 'import numpy as np\n'), ((16345, 16355), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16352, 16355), True, 'import numpy as np\n'), ((16363, 16373), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16370, 16373), True, 'import numpy as np\n'), ((16462, 16472), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16469, 16472), True, 'import numpy as np\n'), ((16480, 16490), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16487, 16490), True, 'import numpy as np\n'), ((16572, 16582), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16579, 16582), True, 'import numpy as np\n'), ((16677, 16687), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16684, 16687), True, 'import numpy as np\n'), ((16782, 16792), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16789, 16792), True, 'import numpy as np\n'), ((16891, 16901), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (16898, 16901), True, 'import numpy as np\n'), ((16997, 17007), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (17004, 17007), True, 'import numpy as np\n'), ((17104, 17114), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (17111, 17114), True, 'import numpy as np\n'), ((17211, 17221), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (17218, 17221), True, 'import numpy as np\n'), ((17322, 17332), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (17329, 17332), True, 'import numpy as np\n'), ((18466, 18476), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (18473, 18476), True, 'import numpy as np\n'), ((18488, 18498), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (18495, 18498), True, 'import numpy as np\n'), ((18573, 18583), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (18580, 18583), True, 'import numpy as np\n'), ((18589, 18599), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (18596, 18599), True, 'import numpy as np\n'), ((18680, 18690), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (18687, 18690), True, 'import numpy as np\n'), ((18699, 18709), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (18706, 18709), True, 'import numpy as np\n'), ((18787, 18797), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (18794, 18797), True, 'import numpy as np\n'), ((18809, 18819), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (18816, 18819), True, 'import numpy as np\n'), ((18894, 18904), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (18901, 18904), True, 'import numpy as np\n'), ((18916, 18926), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (18923, 18926), True, 'import numpy as np\n'), ((19001, 19011), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19008, 19011), True, 'import numpy as np\n'), ((19017, 19027), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19024, 19027), True, 'import numpy as np\n'), ((19108, 19118), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19115, 19118), True, 'import numpy as np\n'), ((19127, 19137), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19134, 19137), True, 'import numpy as np\n'), ((19215, 19225), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19222, 19225), True, 'import numpy as np\n'), ((19237, 19247), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19244, 19247), True, 'import numpy as np\n'), ((19322, 19332), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19329, 19332), True, 'import numpy as np\n'), ((19338, 19348), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19345, 19348), True, 'import numpy as np\n'), ((19429, 19439), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19436, 19439), True, 'import numpy as np\n'), ((19445, 19455), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19452, 19455), True, 'import numpy as np\n'), ((19536, 19546), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19543, 19546), True, 'import numpy as np\n'), ((19555, 19565), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19562, 19565), True, 'import numpy as np\n'), ((19643, 19653), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19650, 19653), True, 'import numpy as np\n'), ((19662, 19672), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19669, 19672), True, 'import numpy as np\n'), ((19903, 19913), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19910, 19913), True, 'import numpy as np\n'), ((19920, 19930), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (19927, 19930), True, 'import numpy as np\n'), ((20012, 20022), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (20019, 20022), True, 'import numpy as np\n'), ((20029, 20039), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (20036, 20039), True, 'import numpy as np\n'), ((20121, 20131), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (20128, 20131), True, 'import numpy as np\n'), ((20141, 20151), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (20148, 20151), True, 'import numpy as np\n'), ((20230, 20240), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (20237, 20240), True, 'import numpy as np\n'), ((20250, 20260), 'numpy.sqrt', 'np.sqrt', (['(2)'], {}), '(2)\n', (20257, 20260), True, 'import numpy as np\n')]
|
import json
import random
from typing import NamedTuple, Any
import numpy
from numpy.testing import assert_array_almost_equal, assert_almost_equal
import torch
import pytest
from flaky import flaky
from allennlp.common.checks import ConfigurationError
from allennlp.common.testing import AllenNlpTestCase
from allennlp.common.util import sanitize
from allennlp.nn import util
from allennlp.models import load_archive
class TestNnUtil(AllenNlpTestCase):
def test_get_sequence_lengths_from_binary_mask(self):
binary_mask = torch.tensor(
[
[True, True, True, False, False, False],
[True, True, False, False, False, False],
[True, True, True, True, True, True],
[True, False, False, False, False, False],
]
)
lengths = util.get_lengths_from_binary_sequence_mask(binary_mask)
numpy.testing.assert_array_equal(lengths.numpy(), numpy.array([3, 2, 6, 1]))
def test_get_mask_from_sequence_lengths(self):
sequence_lengths = torch.LongTensor([4, 3, 1, 4, 2])
mask = util.get_mask_from_sequence_lengths(sequence_lengths, 5).data.numpy()
assert_almost_equal(
mask,
[[1, 1, 1, 1, 0], [1, 1, 1, 0, 0], [1, 0, 0, 0, 0], [1, 1, 1, 1, 0], [1, 1, 0, 0, 0]],
)
def test_get_sequence_lengths_converts_to_long_tensor_and_avoids_variable_overflow(self):
# Tests the following weird behaviour in Pytorch 0.1.12
# doesn't happen for our sequence masks:
#
# mask = torch.ones([260]).bool()
# mask.sum() # equals 260.
# var_mask = t.a.V(mask)
# var_mask.sum() # equals 4, due to 8 bit precision - the sum overflows.
binary_mask = torch.ones(2, 260).bool()
lengths = util.get_lengths_from_binary_sequence_mask(binary_mask)
numpy.testing.assert_array_equal(lengths.data.numpy(), numpy.array([260, 260]))
def test_clamp_tensor(self):
# Test on uncoalesced sparse tensor
i = torch.LongTensor([[0, 1, 1, 0], [2, 0, 2, 2]])
v = torch.FloatTensor([3, 4, -5, 3])
tensor = torch.sparse.FloatTensor(i, v, torch.Size([2, 3]))
clamped_tensor = util.clamp_tensor(tensor, minimum=-3, maximum=3).to_dense()
assert_almost_equal(clamped_tensor, [[0, 0, 3], [3, 0, -3]])
# Test on coalesced sparse tensor
i = torch.LongTensor([[0, 1, 1], [2, 0, 2]])
v = torch.FloatTensor([3, 4, -5])
tensor = torch.sparse.FloatTensor(i, v, torch.Size([2, 3]))
clamped_tensor = util.clamp_tensor(tensor, minimum=-3, maximum=3).to_dense()
assert_almost_equal(clamped_tensor, [[0, 0, 3], [3, 0, -3]])
# Test on dense tensor
tensor = torch.tensor([[5, -4, 3], [-3, 0, -30]])
clamped_tensor = util.clamp_tensor(tensor, minimum=-3, maximum=3)
assert_almost_equal(clamped_tensor, [[3, -3, 3], [-3, 0, -3]])
def test_sort_tensor_by_length(self):
tensor = torch.rand([5, 7, 9])
tensor[0, 3:, :] = 0
tensor[1, 4:, :] = 0
tensor[2, 1:, :] = 0
tensor[3, 5:, :] = 0
sequence_lengths = torch.LongTensor([3, 4, 1, 5, 7])
sorted_tensor, sorted_lengths, reverse_indices, _ = util.sort_batch_by_length(
tensor, sequence_lengths
)
# Test sorted indices are padded correctly.
numpy.testing.assert_array_equal(sorted_tensor[1, 5:, :].data.numpy(), 0.0)
numpy.testing.assert_array_equal(sorted_tensor[2, 4:, :].data.numpy(), 0.0)
numpy.testing.assert_array_equal(sorted_tensor[3, 3:, :].data.numpy(), 0.0)
numpy.testing.assert_array_equal(sorted_tensor[4, 1:, :].data.numpy(), 0.0)
assert sorted_lengths.data.equal(torch.LongTensor([7, 5, 4, 3, 1]))
# Test restoration indices correctly recover the original tensor.
assert sorted_tensor.index_select(0, reverse_indices).data.equal(tensor.data)
def test_get_final_encoder_states(self):
encoder_outputs = torch.Tensor(
[
[[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12]],
[[13, 14, 15, 16], [17, 18, 19, 20], [21, 22, 23, 24]],
]
)
mask = torch.tensor([[True, True, True], [True, True, False]])
final_states = util.get_final_encoder_states(encoder_outputs, mask, bidirectional=False)
assert_almost_equal(final_states.data.numpy(), [[9, 10, 11, 12], [17, 18, 19, 20]])
final_states = util.get_final_encoder_states(encoder_outputs, mask, bidirectional=True)
assert_almost_equal(final_states.data.numpy(), [[9, 10, 3, 4], [17, 18, 15, 16]])
def test_masked_softmax_no_mask(self):
# Testing the general unmasked 1D case.
vector_1d = torch.FloatTensor([[1.0, 2.0, 3.0]])
vector_1d_softmaxed = util.masked_softmax(vector_1d, None).data.numpy()
assert_array_almost_equal(
vector_1d_softmaxed, numpy.array([[0.090031, 0.244728, 0.665241]])
)
assert_almost_equal(1.0, numpy.sum(vector_1d_softmaxed), decimal=6)
vector_1d = torch.FloatTensor([[1.0, 2.0, 5.0]])
vector_1d_softmaxed = util.masked_softmax(vector_1d, None).data.numpy()
assert_array_almost_equal(vector_1d_softmaxed, numpy.array([[0.017148, 0.046613, 0.93624]]))
# Testing the unmasked 1D case where the input is all 0s.
vector_zero = torch.FloatTensor([[0.0, 0.0, 0.0]])
vector_zero_softmaxed = util.masked_softmax(vector_zero, None).data.numpy()
assert_array_almost_equal(
vector_zero_softmaxed, numpy.array([[0.33333334, 0.33333334, 0.33333334]])
)
# Testing the general unmasked batched case.
matrix = torch.FloatTensor([[1.0, 2.0, 5.0], [1.0, 2.0, 3.0]])
masked_matrix_softmaxed = util.masked_softmax(matrix, None).data.numpy()
assert_array_almost_equal(
masked_matrix_softmaxed,
numpy.array(
[[0.01714783, 0.04661262, 0.93623955], [0.09003057, 0.24472847, 0.66524096]]
),
)
# Testing the unmasked batched case where one of the inputs are all 0s.
matrix = torch.FloatTensor([[1.0, 2.0, 5.0], [0.0, 0.0, 0.0]])
masked_matrix_softmaxed = util.masked_softmax(matrix, None).data.numpy()
assert_array_almost_equal(
masked_matrix_softmaxed,
numpy.array(
[[0.01714783, 0.04661262, 0.93623955], [0.33333334, 0.33333334, 0.33333334]]
),
)
def test_masked_softmax_masked(self):
# Testing the general masked 1D case.
vector_1d = torch.FloatTensor([[1.0, 2.0, 5.0]])
mask_1d = torch.tensor([[True, False, True]])
vector_1d_softmaxed = util.masked_softmax(vector_1d, mask_1d).data.numpy()
assert_array_almost_equal(vector_1d_softmaxed, numpy.array([[0.01798621, 0.0, 0.98201382]]))
vector_1d = torch.FloatTensor([[0.0, 2.0, 3.0, 4.0]])
mask_1d = torch.tensor([[True, False, True, True]])
vector_1d_softmaxed = util.masked_softmax(vector_1d, mask_1d).data.numpy()
assert_array_almost_equal(
vector_1d_softmaxed, numpy.array([[0.01321289, 0.0, 0.26538793, 0.72139918]])
)
# Testing the masked 1D case where the input is all 0s and the mask
# is not all 0s.
vector_1d = torch.FloatTensor([[0.0, 0.0, 0.0, 0.0]])
mask_1d = torch.tensor([[False, False, False, True]])
vector_1d_softmaxed = util.masked_softmax(vector_1d, mask_1d).data.numpy()
assert_array_almost_equal(vector_1d_softmaxed, numpy.array([[0, 0, 0, 1]]))
# Testing the masked 1D case where the input is not all 0s
# and the mask is all 0s.
vector_1d = torch.FloatTensor([[0.0, 2.0, 3.0, 4.0]])
mask_1d = torch.tensor([[False, False, False, False]])
vector_1d_softmaxed = util.masked_softmax(vector_1d, mask_1d).data.numpy()
assert_array_almost_equal(vector_1d_softmaxed, numpy.array([[0.0, 0.0, 0.0, 0.0]]))
# Testing the masked 1D case where the input is all 0s and
# the mask is all 0s.
vector_1d = torch.FloatTensor([[0.0, 0.0, 0.0, 0.0]])
mask_1d = torch.tensor([[False, False, False, False]])
vector_1d_softmaxed = util.masked_softmax(vector_1d, mask_1d).data.numpy()
assert_array_almost_equal(vector_1d_softmaxed, numpy.array([[0.0, 0.0, 0.0, 0.0]]))
# Testing the masked 1D case where there are large elements in the
# padding.
vector_1d = torch.FloatTensor([[1.0, 1.0, 1e5]])
mask_1d = torch.tensor([[True, True, False]])
vector_1d_softmaxed = util.masked_softmax(vector_1d, mask_1d).data.numpy()
assert_array_almost_equal(vector_1d_softmaxed, numpy.array([[0.5, 0.5, 0]]))
# Testing the general masked batched case.
matrix = torch.FloatTensor([[1.0, 2.0, 5.0], [1.0, 2.0, 3.0]])
mask = torch.tensor([[True, False, True], [True, True, True]])
masked_matrix_softmaxed = util.masked_softmax(matrix, mask).data.numpy()
assert_array_almost_equal(
masked_matrix_softmaxed,
numpy.array([[0.01798621, 0.0, 0.98201382], [0.090031, 0.244728, 0.665241]]),
)
# Testing the masked batch case where one of the inputs is all 0s but
# none of the masks are all 0.
matrix = torch.FloatTensor([[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]])
mask = torch.tensor([[True, False, True], [True, True, True]])
masked_matrix_softmaxed = util.masked_softmax(matrix, mask).data.numpy()
assert_array_almost_equal(
masked_matrix_softmaxed, numpy.array([[0.5, 0.0, 0.5], [0.090031, 0.244728, 0.665241]])
)
# Testing the masked batch case where one of the inputs is all 0s and
# one of the masks are all 0.
matrix = torch.FloatTensor([[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]])
mask = torch.tensor([[True, False, True], [False, False, False]])
masked_matrix_softmaxed = util.masked_softmax(matrix, mask).data.numpy()
assert_array_almost_equal(
masked_matrix_softmaxed, numpy.array([[0.5, 0.0, 0.5], [0.0, 0.0, 0.0]])
)
matrix = torch.FloatTensor([[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]])
mask = torch.tensor([[False, False, False], [True, False, True]])
masked_matrix_softmaxed = util.masked_softmax(matrix, mask).data.numpy()
assert_array_almost_equal(
masked_matrix_softmaxed, numpy.array([[0.0, 0.0, 0.0], [0.11920292, 0.0, 0.88079708]])
)
def test_masked_softmax_memory_efficient_masked(self):
# Testing the general masked 1D case.
vector_1d = torch.FloatTensor([[1.0, 2.0, 5.0]])
mask_1d = torch.tensor([[True, False, True]])
vector_1d_softmaxed = util.masked_softmax(
vector_1d, mask_1d, memory_efficient=True
).data.numpy()
assert_array_almost_equal(vector_1d_softmaxed, numpy.array([[0.01798621, 0.0, 0.98201382]]))
vector_1d = torch.FloatTensor([[0.0, 2.0, 3.0, 4.0]])
mask_1d = torch.tensor([[True, False, True, True]])
vector_1d_softmaxed = util.masked_softmax(
vector_1d, mask_1d, memory_efficient=True
).data.numpy()
assert_array_almost_equal(
vector_1d_softmaxed, numpy.array([[0.01321289, 0.0, 0.26538793, 0.72139918]])
)
# Testing the masked 1D case where the input is all 0s and the mask
# is not all 0s.
vector_1d = torch.FloatTensor([[0.0, 0.0, 0.0, 0.0]])
mask_1d = torch.tensor([[False, False, False, True]])
vector_1d_softmaxed = util.masked_softmax(
vector_1d, mask_1d, memory_efficient=True
).data.numpy()
assert_array_almost_equal(vector_1d_softmaxed, numpy.array([[0, 0, 0, 1]]))
# Testing the masked 1D case where the input is not all 0s
# and the mask is all 0s.
vector_1d = torch.FloatTensor([[0.0, 2.0, 3.0, 4.0]])
mask_1d = torch.tensor([[False, False, False, False]])
vector_1d_softmaxed = util.masked_softmax(
vector_1d, mask_1d, memory_efficient=True
).data.numpy()
assert_array_almost_equal(vector_1d_softmaxed, numpy.array([[0.25, 0.25, 0.25, 0.25]]))
# Testing the masked 1D case where the input is all 0s and
# the mask is all 0s.
vector_1d = torch.FloatTensor([[0.0, 0.0, 0.0, 0.0]])
mask_1d = torch.tensor([[False, False, False, False]])
vector_1d_softmaxed = util.masked_softmax(
vector_1d, mask_1d, memory_efficient=True
).data.numpy()
assert_array_almost_equal(vector_1d_softmaxed, numpy.array([[0.25, 0.25, 0.25, 0.25]]))
# Testing the masked 1D case where there are large elements in the
# padding.
vector_1d = torch.FloatTensor([[1.0, 1.0, 1e5]])
mask_1d = torch.tensor([[True, True, False]])
vector_1d_softmaxed = util.masked_softmax(
vector_1d, mask_1d, memory_efficient=True
).data.numpy()
assert_array_almost_equal(vector_1d_softmaxed, numpy.array([[0.5, 0.5, 0]]))
# Testing the general masked batched case.
matrix = torch.FloatTensor([[1.0, 2.0, 5.0], [1.0, 2.0, 3.0]])
mask = torch.tensor([[True, False, True], [True, True, True]])
masked_matrix_softmaxed = util.masked_softmax(
matrix, mask, memory_efficient=True
).data.numpy()
assert_array_almost_equal(
masked_matrix_softmaxed,
numpy.array([[0.01798621, 0.0, 0.98201382], [0.090031, 0.244728, 0.665241]]),
)
# Testing the masked batch case where one of the inputs is all 0s but
# none of the masks are all 0.
matrix = torch.FloatTensor([[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]])
mask = torch.tensor([[True, False, True], [True, True, True]])
masked_matrix_softmaxed = util.masked_softmax(
matrix, mask, memory_efficient=True
).data.numpy()
assert_array_almost_equal(
masked_matrix_softmaxed, numpy.array([[0.5, 0.0, 0.5], [0.090031, 0.244728, 0.665241]])
)
# Testing the masked batch case where one of the inputs is all 0s and
# one of the masks are all 0.
matrix = torch.FloatTensor([[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]])
mask = torch.tensor([[True, False, True], [False, False, False]])
masked_matrix_softmaxed = util.masked_softmax(
matrix, mask, memory_efficient=True
).data.numpy()
assert_array_almost_equal(
masked_matrix_softmaxed,
numpy.array([[0.5, 0.0, 0.5], [0.33333333, 0.33333333, 0.33333333]]),
)
matrix = torch.FloatTensor([[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]])
mask = torch.tensor([[False, False, False], [True, False, True]])
masked_matrix_softmaxed = util.masked_softmax(
matrix, mask, memory_efficient=True
).data.numpy()
assert_array_almost_equal(
masked_matrix_softmaxed,
numpy.array([[0.33333333, 0.33333333, 0.33333333], [0.11920292, 0.0, 0.88079708]]),
)
def test_masked_log_softmax_masked(self):
# Tests replicated from test_softmax_masked - we test that exponentiated,
# the log softmax contains the correct elements (masked elements should be == 1).
# Testing the general masked 1D case.
vector_1d = torch.FloatTensor([[1.0, 2.0, 5.0]])
mask_1d = torch.tensor([[True, False, True]])
vector_1d_softmaxed = util.masked_log_softmax(vector_1d, mask_1d).data.numpy()
assert_array_almost_equal(
numpy.exp(vector_1d_softmaxed), numpy.array([[0.01798621, 0.0, 0.98201382]])
)
vector_1d = torch.FloatTensor([[0.0, 2.0, 3.0, 4.0]])
mask_1d = torch.tensor([[True, False, True, True]])
vector_1d_softmaxed = util.masked_log_softmax(vector_1d, mask_1d).data.numpy()
assert_array_almost_equal(
numpy.exp(vector_1d_softmaxed), numpy.array([[0.01321289, 0.0, 0.26538793, 0.72139918]])
)
# Testing the masked 1D case where the input is all 0s and the mask
# is not all 0s.
vector_1d = torch.FloatTensor([[0.0, 0.0, 0.0, 0.0]])
mask_1d = torch.tensor([[False, False, False, True]])
vector_1d_softmaxed = util.masked_log_softmax(vector_1d, mask_1d).data.numpy()
assert_array_almost_equal(
numpy.exp(vector_1d_softmaxed), numpy.array([[0.0, 0.0, 0.0, 1.0]])
)
# Testing the masked 1D case where the input is not all 0s
# and the mask is all 0s. The output here will be arbitrary, but it should not be nan.
vector_1d = torch.FloatTensor([[0.0, 2.0, 3.0, 4.0]])
mask_1d = torch.tensor([[False, False, False, False]])
vector_1d_softmaxed = util.masked_log_softmax(vector_1d, mask_1d).data.numpy()
assert not numpy.isnan(vector_1d_softmaxed).any()
def test_masked_max(self):
# Testing the general masked 1D case.
vector_1d = torch.FloatTensor([1.0, 12.0, 5.0])
mask_1d = torch.tensor([True, False, True])
vector_1d_maxed = util.masked_max(vector_1d, mask_1d, dim=0).data.numpy()
assert_array_almost_equal(vector_1d_maxed, 5.0)
# Testing if all masks are zero, the output will be arbitrary, but it should not be nan.
vector_1d = torch.FloatTensor([1.0, 12.0, 5.0])
mask_1d = torch.tensor([False, False, False])
vector_1d_maxed = util.masked_max(vector_1d, mask_1d, dim=0).data.numpy()
assert not numpy.isnan(vector_1d_maxed).any()
# Testing batch value and batch masks
matrix = torch.FloatTensor([[1.0, 12.0, 5.0], [-1.0, -2.0, 3.0]])
mask = torch.tensor([[True, False, True], [True, True, False]])
matrix_maxed = util.masked_max(matrix, mask, dim=-1).data.numpy()
assert_array_almost_equal(matrix_maxed, numpy.array([5.0, -1.0]))
# Testing keepdim for batch value and batch masks
matrix = torch.FloatTensor([[1.0, 12.0, 5.0], [-1.0, -2.0, 3.0]])
mask = torch.tensor([[True, False, True], [True, True, False]])
matrix_maxed = util.masked_max(matrix, mask, dim=-1, keepdim=True).data.numpy()
assert_array_almost_equal(matrix_maxed, numpy.array([[5.0], [-1.0]]))
# Testing broadcast
matrix = torch.FloatTensor(
[[[1.0, 2.0], [12.0, 3.0], [5.0, -1.0]], [[-1.0, -3.0], [-2.0, -0.5], [3.0, 8.0]]]
)
mask = torch.tensor([[True, False, True], [True, True, False]]).unsqueeze(-1)
matrix_maxed = util.masked_max(matrix, mask, dim=1).data.numpy()
assert_array_almost_equal(matrix_maxed, numpy.array([[5.0, 2.0], [-1.0, -0.5]]))
def test_masked_mean(self):
# Testing the general masked 1D case.
vector_1d = torch.FloatTensor([1.0, 12.0, 5.0])
mask_1d = torch.tensor([True, False, True])
vector_1d_mean = util.masked_mean(vector_1d, mask_1d, dim=0).data.numpy()
assert_array_almost_equal(vector_1d_mean, 3.0)
# Testing if all masks are zero, the output will be arbitrary, but it should not be nan.
vector_1d = torch.FloatTensor([1.0, 12.0, 5.0])
mask_1d = torch.tensor([False, False, False])
vector_1d_mean = util.masked_mean(vector_1d, mask_1d, dim=0).data.numpy()
assert not numpy.isnan(vector_1d_mean).any()
# Testing batch value and batch masks
matrix = torch.FloatTensor([[1.0, 12.0, 5.0], [-1.0, -2.0, 3.0]])
mask = torch.tensor([[True, False, True], [True, True, False]])
matrix_mean = util.masked_mean(matrix, mask, dim=-1).data.numpy()
assert_array_almost_equal(matrix_mean, numpy.array([3.0, -1.5]))
# Testing keepdim for batch value and batch masks
matrix = torch.FloatTensor([[1.0, 12.0, 5.0], [-1.0, -2.0, 3.0]])
mask = torch.tensor([[True, False, True], [True, True, False]])
matrix_mean = util.masked_mean(matrix, mask, dim=-1, keepdim=True).data.numpy()
assert_array_almost_equal(matrix_mean, numpy.array([[3.0], [-1.5]]))
# Testing broadcast
matrix = torch.FloatTensor(
[[[1.0, 2.0], [12.0, 3.0], [5.0, -1.0]], [[-1.0, -3.0], [-2.0, -0.5], [3.0, 8.0]]]
)
mask = torch.tensor([[True, False, True], [True, True, False]]).unsqueeze(-1)
matrix_mean = util.masked_mean(matrix, mask, dim=1).data.numpy()
assert_array_almost_equal(matrix_mean, numpy.array([[3.0, 0.5], [-1.5, -1.75]]))
def test_masked_flip(self):
tensor = torch.FloatTensor(
[[[6, 6, 6], [1, 1, 1], [2, 2, 2]], [[3, 3, 3], [4, 4, 4], [5, 5, 5]]]
)
solution = [[[6, 6, 6], [0, 0, 0]], [[4, 4, 4], [3, 3, 3]]]
response = util.masked_flip(tensor, [1, 2])
assert_almost_equal(response, solution)
tensor = torch.FloatTensor(
[
[[6, 6, 6], [1, 1, 1], [2, 2, 2], [0, 0, 0]],
[[3, 3, 3], [4, 4, 4], [5, 5, 5], [1, 2, 3]],
]
)
solution = [
[[2, 2, 2], [1, 1, 1], [6, 6, 6], [0, 0, 0]],
[[1, 2, 3], [5, 5, 5], [4, 4, 4], [3, 3, 3]],
]
response = util.masked_flip(tensor, [3, 4])
assert_almost_equal(response, solution)
tensor = torch.FloatTensor(
[
[[6, 6, 6], [1, 1, 1], [2, 2, 2], [0, 0, 0]],
[[3, 3, 3], [4, 4, 4], [5, 5, 5], [1, 2, 3]],
[[1, 1, 1], [2, 2, 2], [0, 0, 0], [0, 0, 0]],
]
)
solution = [
[[2, 2, 2], [1, 1, 1], [6, 6, 6], [0, 0, 0]],
[[1, 2, 3], [5, 5, 5], [4, 4, 4], [3, 3, 3]],
[[2, 2, 2], [1, 1, 1], [0, 0, 0], [0, 0, 0]],
]
response = util.masked_flip(tensor, [3, 4, 2])
assert_almost_equal(response, solution)
def test_get_text_field_mask_returns_a_correct_mask(self):
text_field_tensors = {
"indexer_name": {
"tokens": torch.LongTensor([[3, 4, 5, 0, 0], [1, 2, 0, 0, 0]]),
"token_characters": torch.LongTensor(
[
[[1, 2], [3, 0], [2, 0], [0, 0], [0, 0]],
[[5, 0], [4, 6], [0, 0], [0, 0], [0, 0]],
]
),
}
}
assert_almost_equal(
util.get_text_field_mask(text_field_tensors).long().numpy(),
[[1, 1, 1, 0, 0], [1, 1, 0, 0, 0]],
)
def test_get_text_field_mask_returns_a_correct_mask_character_only_input(self):
text_field_tensors = {
"indexer_name": {
"token_characters": torch.LongTensor(
[
[[1, 2, 3], [3, 0, 1], [2, 1, 0], [0, 0, 0]],
[[5, 5, 5], [4, 6, 0], [0, 0, 0], [0, 0, 0]],
]
)
}
}
assert_almost_equal(
util.get_text_field_mask(text_field_tensors).long().numpy(),
[[1, 1, 1, 0], [1, 1, 0, 0]],
)
def test_get_text_field_mask_returns_a_correct_mask_list_field(self):
text_field_tensors = {
"indexer_name": {
"list_tokens": torch.LongTensor(
[
[[1, 2], [3, 0], [2, 0], [0, 0], [0, 0]],
[[5, 0], [4, 6], [0, 0], [0, 0], [0, 0]],
]
)
}
}
actual_mask = (
util.get_text_field_mask(text_field_tensors, num_wrapping_dims=1).long().numpy()
)
expected_mask = (text_field_tensors["indexer_name"]["list_tokens"].numpy() > 0).astype(
"int32"
)
assert_almost_equal(actual_mask, expected_mask)
def test_get_text_field_mask_returns_mask_key(self):
text_field_tensors = {
"indexer_name": {
"tokens": torch.LongTensor([[3, 4, 5, 0, 0], [1, 2, 0, 0, 0]]),
"mask": torch.tensor([[False, False, True]]),
}
}
assert_almost_equal(
util.get_text_field_mask(text_field_tensors).long().numpy(), [[0, 0, 1]]
)
def test_weighted_sum_works_on_simple_input(self):
batch_size = 1
sentence_length = 5
embedding_dim = 4
sentence_array = numpy.random.rand(batch_size, sentence_length, embedding_dim)
sentence_tensor = torch.from_numpy(sentence_array).float()
attention_tensor = torch.FloatTensor([[0.3, 0.4, 0.1, 0, 1.2]])
aggregated_array = util.weighted_sum(sentence_tensor, attention_tensor).data.numpy()
assert aggregated_array.shape == (batch_size, embedding_dim)
expected_array = (
0.3 * sentence_array[0, 0]
+ 0.4 * sentence_array[0, 1]
+ 0.1 * sentence_array[0, 2]
+ 0.0 * sentence_array[0, 3]
+ 1.2 * sentence_array[0, 4]
)
numpy.testing.assert_almost_equal(aggregated_array, [expected_array], decimal=5)
def test_weighted_sum_handles_higher_order_input(self):
batch_size = 1
length_1 = 5
length_2 = 6
length_3 = 2
embedding_dim = 4
sentence_array = numpy.random.rand(batch_size, length_1, length_2, length_3, embedding_dim)
attention_array = numpy.random.rand(batch_size, length_1, length_2, length_3)
sentence_tensor = torch.from_numpy(sentence_array).float()
attention_tensor = torch.from_numpy(attention_array).float()
aggregated_array = util.weighted_sum(sentence_tensor, attention_tensor).data.numpy()
assert aggregated_array.shape == (batch_size, length_1, length_2, embedding_dim)
expected_array = (
attention_array[0, 3, 2, 0] * sentence_array[0, 3, 2, 0]
+ attention_array[0, 3, 2, 1] * sentence_array[0, 3, 2, 1]
)
numpy.testing.assert_almost_equal(aggregated_array[0, 3, 2], expected_array, decimal=5)
def test_weighted_sum_handles_uneven_higher_order_input(self):
batch_size = 1
length_1 = 5
length_2 = 6
length_3 = 2
embedding_dim = 4
sentence_array = numpy.random.rand(batch_size, length_3, embedding_dim)
attention_array = numpy.random.rand(batch_size, length_1, length_2, length_3)
sentence_tensor = torch.from_numpy(sentence_array).float()
attention_tensor = torch.from_numpy(attention_array).float()
aggregated_array = util.weighted_sum(sentence_tensor, attention_tensor).data.numpy()
assert aggregated_array.shape == (batch_size, length_1, length_2, embedding_dim)
for i in range(length_1):
for j in range(length_2):
expected_array = (
attention_array[0, i, j, 0] * sentence_array[0, 0]
+ attention_array[0, i, j, 1] * sentence_array[0, 1]
)
numpy.testing.assert_almost_equal(
aggregated_array[0, i, j], expected_array, decimal=5
)
def test_weighted_sum_handles_3d_attention_with_3d_matrix(self):
batch_size = 1
length_1 = 5
length_2 = 2
embedding_dim = 4
sentence_array = numpy.random.rand(batch_size, length_2, embedding_dim)
attention_array = numpy.random.rand(batch_size, length_1, length_2)
sentence_tensor = torch.from_numpy(sentence_array).float()
attention_tensor = torch.from_numpy(attention_array).float()
aggregated_array = util.weighted_sum(sentence_tensor, attention_tensor).data.numpy()
assert aggregated_array.shape == (batch_size, length_1, embedding_dim)
for i in range(length_1):
expected_array = (
attention_array[0, i, 0] * sentence_array[0, 0]
+ attention_array[0, i, 1] * sentence_array[0, 1]
)
numpy.testing.assert_almost_equal(aggregated_array[0, i], expected_array, decimal=5)
def test_viterbi_decode(self):
# Test Viterbi decoding is equal to greedy decoding with no pairwise potentials.
sequence_logits = torch.nn.functional.softmax(torch.rand([5, 9]), dim=-1)
transition_matrix = torch.zeros([9, 9])
indices, _ = util.viterbi_decode(sequence_logits.data, transition_matrix)
_, argmax_indices = torch.max(sequence_logits, 1)
assert indices == argmax_indices.data.squeeze().tolist()
# Test Viterbi decoding works with start and end transitions
sequence_logits = torch.nn.functional.softmax(torch.rand([5, 9]), dim=-1)
transition_matrix = torch.zeros([9, 9])
allowed_start_transitions = torch.zeros([9])
# Force start tag to be an 8
allowed_start_transitions[:8] = float("-inf")
allowed_end_transitions = torch.zeros([9])
# Force end tag to be a 0
allowed_end_transitions[1:] = float("-inf")
indices, _ = util.viterbi_decode(
sequence_logits.data,
transition_matrix,
allowed_end_transitions=allowed_end_transitions,
allowed_start_transitions=allowed_start_transitions,
)
assert indices[0] == 8
assert indices[-1] == 0
# Test that pairwise potentials affect the sequence correctly and that
# viterbi_decode can handle -inf values.
sequence_logits = torch.FloatTensor(
[
[0, 0, 0, 3, 5],
[0, 0, 0, 3, 4],
[0, 0, 0, 3, 4],
[0, 0, 0, 3, 4],
[0, 0, 0, 3, 4],
[0, 0, 0, 3, 4],
]
)
# The same tags shouldn't appear sequentially.
transition_matrix = torch.zeros([5, 5])
for i in range(5):
transition_matrix[i, i] = float("-inf")
indices, _ = util.viterbi_decode(sequence_logits, transition_matrix)
assert indices == [4, 3, 4, 3, 4, 3]
# Test that unbalanced pairwise potentials break ties
# between paths with equal unary potentials.
sequence_logits = torch.FloatTensor(
[
[0, 0, 0, 4, 4],
[0, 0, 0, 4, 4],
[0, 0, 0, 4, 4],
[0, 0, 0, 4, 4],
[0, 0, 0, 4, 4],
[0, 0, 0, 4, 4],
]
)
# The 5th tag has a penalty for appearing sequentially
# or for transitioning to the 4th tag, making the best
# path uniquely to take the 4th tag only.
transition_matrix = torch.zeros([5, 5])
transition_matrix[4, 4] = -10
transition_matrix[4, 3] = -10
transition_matrix[3, 4] = -10
indices, _ = util.viterbi_decode(sequence_logits, transition_matrix)
assert indices == [3, 3, 3, 3, 3, 3]
sequence_logits = torch.FloatTensor([[1, 0, 0, 4], [1, 0, 6, 2], [0, 3, 0, 4]])
# Best path would normally be [3, 2, 3] but we add a
# potential from 2 -> 1, making [3, 2, 1] the best path.
transition_matrix = torch.zeros([4, 4])
transition_matrix[0, 0] = 1
transition_matrix[2, 1] = 5
indices, value = util.viterbi_decode(sequence_logits, transition_matrix)
assert indices == [3, 2, 1]
assert value.numpy() == 18
# Test that providing evidence results in paths containing specified tags.
sequence_logits = torch.FloatTensor(
[
[0, 0, 0, 7, 7],
[0, 0, 0, 7, 7],
[0, 0, 0, 7, 7],
[0, 0, 0, 7, 7],
[0, 0, 0, 7, 7],
[0, 0, 0, 7, 7],
]
)
# The 5th tag has a penalty for appearing sequentially
# or for transitioning to the 4th tag, making the best
# path to take the 4th tag for every label.
transition_matrix = torch.zeros([5, 5])
transition_matrix[4, 4] = -10
transition_matrix[4, 3] = -2
transition_matrix[3, 4] = -2
# The 1st, 4th and 5th sequence elements are observed - they should be
# equal to 2, 0 and 4. The last tag should be equal to 3, because although
# the penalty for transitioning to the 4th tag is -2, the unary potential
# is 7, which is greater than the combination for any of the other labels.
observations = [2, -1, -1, 0, 4, -1]
indices, _ = util.viterbi_decode(sequence_logits, transition_matrix, observations)
assert indices == [2, 3, 3, 0, 4, 3]
def test_viterbi_decode_top_k(self):
# Test cases taken from: https://gist.github.com/PetrochukM/afaa3613a99a8e7213d2efdd02ae4762
# Test Viterbi decoding is equal to greedy decoding with no pairwise potentials.
sequence_logits = torch.autograd.Variable(torch.rand([5, 9]))
transition_matrix = torch.zeros([9, 9])
indices, _ = util.viterbi_decode(sequence_logits.data, transition_matrix, top_k=5)
_, argmax_indices = torch.max(sequence_logits, 1)
assert indices[0] == argmax_indices.data.squeeze().tolist()
# Test that pairwise potentials effect the sequence correctly and that
# viterbi_decode can handle -inf values.
sequence_logits = torch.FloatTensor(
[
[0, 0, 0, 3, 4],
[0, 0, 0, 3, 4],
[0, 0, 0, 3, 4],
[0, 0, 0, 3, 4],
[0, 0, 0, 3, 4],
[0, 0, 0, 3, 4],
]
)
# The same tags shouldn't appear sequentially.
transition_matrix = torch.zeros([5, 5])
for i in range(5):
transition_matrix[i, i] = float("-inf")
indices, _ = util.viterbi_decode(sequence_logits, transition_matrix, top_k=5)
assert indices[0] == [3, 4, 3, 4, 3, 4]
# Test that unbalanced pairwise potentials break ties
# between paths with equal unary potentials.
sequence_logits = torch.FloatTensor(
[
[0, 0, 0, 4, 4],
[0, 0, 0, 4, 4],
[0, 0, 0, 4, 4],
[0, 0, 0, 4, 4],
[0, 0, 0, 4, 4],
[0, 0, 0, 4, 0],
]
)
# The 5th tag has a penalty for appearing sequentially
# or for transitioning to the 4th tag, making the best
# path uniquely to take the 4th tag only.
transition_matrix = torch.zeros([5, 5])
transition_matrix[4, 4] = -10
transition_matrix[4, 3] = -10
indices, _ = util.viterbi_decode(sequence_logits, transition_matrix, top_k=5)
assert indices[0] == [3, 3, 3, 3, 3, 3]
sequence_logits = torch.FloatTensor([[1, 0, 0, 4], [1, 0, 6, 2], [0, 3, 0, 4]])
# Best path would normally be [3, 2, 3] but we add a
# potential from 2 -> 1, making [3, 2, 1] the best path.
transition_matrix = torch.zeros([4, 4])
transition_matrix[0, 0] = 1
transition_matrix[2, 1] = 5
indices, value = util.viterbi_decode(sequence_logits, transition_matrix, top_k=5)
assert indices[0] == [3, 2, 1]
assert value[0] == 18
def _brute_decode(
tag_sequence: torch.Tensor, transition_matrix: torch.Tensor, top_k: int = 5
) -> Any:
"""
Top-k decoder that uses brute search instead of the Viterbi Decode dynamic programing algorithm
"""
# Create all possible sequences
sequences = [[]] # type: ignore
for i in range(len(tag_sequence)):
new_sequences = [] # type: ignore
for j in range(len(tag_sequence[i])):
for sequence in sequences:
new_sequences.append(sequence[:] + [j])
sequences = new_sequences
# Score
scored_sequences = [] # type: ignore
for sequence in sequences:
emission_score = sum(tag_sequence[i, j] for i, j in enumerate(sequence))
transition_score = sum(
transition_matrix[sequence[i - 1], sequence[i]] for i in range(1, len(sequence))
)
score = emission_score + transition_score
scored_sequences.append((score, sequence))
# Get the top k scores / paths
top_k_sequences = sorted(scored_sequences, key=lambda r: r[0], reverse=True)[:top_k]
scores, paths = zip(*top_k_sequences)
return paths, scores # type: ignore
# Run 100 randomly generated parameters and compare the outputs.
for i in range(100):
num_tags = random.randint(1, 5)
seq_len = random.randint(1, 5)
k = random.randint(1, 5)
sequence_logits = torch.rand([seq_len, num_tags])
transition_matrix = torch.rand([num_tags, num_tags])
viterbi_paths_v1, viterbi_scores_v1 = util.viterbi_decode(
sequence_logits, transition_matrix, top_k=k
)
viterbi_path_brute, viterbi_score_brute = _brute_decode(
sequence_logits, transition_matrix, top_k=k
)
numpy.testing.assert_almost_equal(
list(viterbi_score_brute), viterbi_scores_v1.tolist(), decimal=3
)
numpy.testing.assert_equal(sanitize(viterbi_paths_v1), viterbi_path_brute)
def test_sequence_cross_entropy_with_logits_masks_loss_correctly(self):
# test weight masking by checking that a tensor with non-zero values in
# masked positions returns the same loss as a tensor with zeros in those
# positions.
tensor = torch.rand([5, 7, 4])
tensor[0, 3:, :] = 0
tensor[1, 4:, :] = 0
tensor[2, 2:, :] = 0
tensor[3, :, :] = 0
weights = (tensor != 0.0)[:, :, 0].long().squeeze(-1)
tensor2 = tensor.clone()
tensor2[0, 3:, :] = 2
tensor2[1, 4:, :] = 13
tensor2[2, 2:, :] = 234
tensor2[3, :, :] = 65
targets = torch.LongTensor(numpy.random.randint(0, 3, [5, 7]))
targets *= weights
loss = util.sequence_cross_entropy_with_logits(tensor, targets, weights)
loss2 = util.sequence_cross_entropy_with_logits(tensor2, targets, weights)
assert loss.data.numpy() == loss2.data.numpy()
def test_sequence_cross_entropy_with_logits_smooths_labels_correctly(self):
tensor = torch.rand([1, 3, 4])
targets = torch.LongTensor(numpy.random.randint(0, 3, [1, 3]))
weights = torch.ones([2, 3])
loss = util.sequence_cross_entropy_with_logits(
tensor, targets, weights, label_smoothing=0.1
)
correct_loss = 0.0
for prediction, label in zip(tensor.squeeze(0), targets.squeeze(0)):
prediction = torch.nn.functional.log_softmax(prediction, dim=-1)
correct_loss += prediction[label] * 0.9
# incorrect elements
correct_loss += prediction.sum() * 0.1 / 4
# Average over sequence.
correct_loss = -correct_loss / 3
numpy.testing.assert_array_almost_equal(loss.data.numpy(), correct_loss.data.numpy())
def test_sequence_cross_entropy_with_logits_averages_batch_correctly(self):
# test batch average is the same as dividing the batch averaged
# loss by the number of batches containing any non-padded tokens.
tensor = torch.rand([5, 7, 4])
tensor[0, 3:, :] = 0
tensor[1, 4:, :] = 0
tensor[2, 2:, :] = 0
tensor[3, :, :] = 0
weights = (tensor != 0.0)[:, :, 0].long().squeeze(-1)
targets = torch.LongTensor(numpy.random.randint(0, 3, [5, 7]))
targets *= weights
loss = util.sequence_cross_entropy_with_logits(tensor, targets, weights)
vector_loss = util.sequence_cross_entropy_with_logits(
tensor, targets, weights, average=None
)
# Batch has one completely padded row, so divide by 4.
assert loss.data.numpy() == vector_loss.sum().item() / 4
@flaky(max_runs=3, min_passes=1)
def test_sequence_cross_entropy_with_logits_averages_token_correctly(self):
# test token average is the same as multiplying the per-batch loss
# with the per-batch weights and dividing by the total weight
tensor = torch.rand([5, 7, 4])
tensor[0, 3:, :] = 0
tensor[1, 4:, :] = 0
tensor[2, 2:, :] = 0
tensor[3, :, :] = 0
weights = (tensor != 0.0)[:, :, 0].long().squeeze(-1)
targets = torch.LongTensor(numpy.random.randint(0, 3, [5, 7]))
targets *= weights
loss = util.sequence_cross_entropy_with_logits(tensor, targets, weights, average="token")
vector_loss = util.sequence_cross_entropy_with_logits(
tensor, targets, weights, average=None
)
total_token_loss = (vector_loss * weights.float().sum(dim=-1)).sum()
average_token_loss = (total_token_loss / weights.float().sum()).detach()
assert_almost_equal(loss.detach().item(), average_token_loss.item(), decimal=5)
def test_sequence_cross_entropy_with_logits_gamma_correctly(self):
batch = 1
length = 3
classes = 4
gamma = abs(numpy.random.randn()) # [0, +inf)
tensor = torch.rand([batch, length, classes])
targets = torch.LongTensor(numpy.random.randint(0, classes, [batch, length]))
weights = torch.ones([batch, length])
loss = util.sequence_cross_entropy_with_logits(tensor, targets, weights, gamma=gamma)
correct_loss = 0.0
for logit, label in zip(tensor.squeeze(0), targets.squeeze(0)):
p = torch.nn.functional.softmax(logit, dim=-1)
pt = p[label]
ft = (1 - pt) ** gamma
correct_loss += -pt.log() * ft
# Average over sequence.
correct_loss = correct_loss / length
numpy.testing.assert_array_almost_equal(loss.data.numpy(), correct_loss.data.numpy())
def test_sequence_cross_entropy_with_logits_alpha_float_correctly(self):
batch = 1
length = 3
classes = 2 # alpha float for binary class only
alpha = (
numpy.random.rand() if numpy.random.rand() > 0.5 else (1.0 - numpy.random.rand())
) # [0, 1]
tensor = torch.rand([batch, length, classes])
targets = torch.LongTensor(numpy.random.randint(0, classes, [batch, length]))
weights = torch.ones([batch, length])
loss = util.sequence_cross_entropy_with_logits(tensor, targets, weights, alpha=alpha)
correct_loss = 0.0
for logit, label in zip(tensor.squeeze(0), targets.squeeze(0)):
logp = torch.nn.functional.log_softmax(logit, dim=-1)
logpt = logp[label]
if label:
at = alpha
else:
at = 1 - alpha
correct_loss += -logpt * at
# Average over sequence.
correct_loss = correct_loss / length
numpy.testing.assert_array_almost_equal(loss.data.numpy(), correct_loss.data.numpy())
def test_sequence_cross_entropy_with_logits_alpha_single_float_correctly(self):
batch = 1
length = 3
classes = 2 # alpha float for binary class only
alpha = (
numpy.random.rand() if numpy.random.rand() > 0.5 else (1.0 - numpy.random.rand())
) # [0, 1]
alpha = torch.tensor(alpha)
tensor = torch.rand([batch, length, classes])
targets = torch.LongTensor(numpy.random.randint(0, classes, [batch, length]))
weights = torch.ones([batch, length])
loss = util.sequence_cross_entropy_with_logits(tensor, targets, weights, alpha=alpha)
correct_loss = 0.0
for logit, label in zip(tensor.squeeze(0), targets.squeeze(0)):
logp = torch.nn.functional.log_softmax(logit, dim=-1)
logpt = logp[label]
if label:
at = alpha
else:
at = 1 - alpha
correct_loss += -logpt * at
# Average over sequence.
correct_loss = correct_loss / length
numpy.testing.assert_array_almost_equal(loss.data.numpy(), correct_loss.data.numpy())
def test_sequence_cross_entropy_with_logits_alpha_list_correctly(self):
batch = 1
length = 3
classes = 4 # alpha float for binary class only
alpha = abs(numpy.random.randn(classes)) # [0, +inf)
tensor = torch.rand([batch, length, classes])
targets = torch.LongTensor(numpy.random.randint(0, classes, [batch, length]))
weights = torch.ones([batch, length])
loss = util.sequence_cross_entropy_with_logits(tensor, targets, weights, alpha=alpha)
correct_loss = 0.0
for logit, label in zip(tensor.squeeze(0), targets.squeeze(0)):
logp = torch.nn.functional.log_softmax(logit, dim=-1)
logpt = logp[label]
at = alpha[label]
correct_loss += -logpt * at
# Average over sequence.
correct_loss = correct_loss / length
numpy.testing.assert_array_almost_equal(loss.data.numpy(), correct_loss.data.numpy())
def test_replace_masked_values_replaces_masked_values_with_finite_value(self):
tensor = torch.FloatTensor([[[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12]]])
mask = torch.tensor([[True, True, False]])
replaced = util.replace_masked_values(tensor, mask.unsqueeze(-1), 2).data.numpy()
assert_almost_equal(replaced, [[[1, 2, 3, 4], [5, 6, 7, 8], [2, 2, 2, 2]]])
def test_logsumexp(self):
# First a simple example where we add probabilities in log space.
tensor = torch.FloatTensor([[0.4, 0.1, 0.2]])
log_tensor = tensor.log()
log_summed = util.logsumexp(log_tensor, dim=-1, keepdim=False)
assert_almost_equal(log_summed.exp().data.numpy(), [0.7])
log_summed = util.logsumexp(log_tensor, dim=-1, keepdim=True)
assert_almost_equal(log_summed.exp().data.numpy(), [[0.7]])
# Then some more atypical examples, and making sure this will work with how we handle
# log masks.
tensor = torch.FloatTensor([[float("-inf"), 20.0]])
assert_almost_equal(util.logsumexp(tensor).data.numpy(), [20.0])
tensor = torch.FloatTensor([[-200.0, 20.0]])
assert_almost_equal(util.logsumexp(tensor).data.numpy(), [20.0])
tensor = torch.FloatTensor([[20.0, 20.0], [-200.0, 200.0]])
assert_almost_equal(util.logsumexp(tensor, dim=0).data.numpy(), [20.0, 200.0])
def test_flatten_and_batch_shift_indices(self):
indices = numpy.array(
[[[1, 2, 3, 4], [5, 6, 7, 8], [9, 9, 9, 9]], [[2, 1, 0, 7], [7, 7, 2, 3], [0, 0, 4, 2]]]
)
indices = torch.tensor(indices, dtype=torch.long)
shifted_indices = util.flatten_and_batch_shift_indices(indices, 10)
numpy.testing.assert_array_equal(
shifted_indices.data.numpy(),
numpy.array(
[1, 2, 3, 4, 5, 6, 7, 8, 9, 9, 9, 9, 12, 11, 10, 17, 17, 17, 12, 13, 10, 10, 14, 12]
),
)
def test_batched_index_select(self):
indices = numpy.array([[[1, 2], [3, 4]], [[5, 6], [7, 8]]])
# Each element is a vector of its index.
targets = torch.ones([2, 10, 3]).cumsum(1) - 1
# Make the second batch double its index so they're different.
targets[1, :, :] *= 2
indices = torch.tensor(indices, dtype=torch.long)
selected = util.batched_index_select(targets, indices)
assert list(selected.size()) == [2, 2, 2, 3]
ones = numpy.ones([3])
numpy.testing.assert_array_equal(selected[0, 0, 0, :].data.numpy(), ones)
numpy.testing.assert_array_equal(selected[0, 0, 1, :].data.numpy(), ones * 2)
numpy.testing.assert_array_equal(selected[0, 1, 0, :].data.numpy(), ones * 3)
numpy.testing.assert_array_equal(selected[0, 1, 1, :].data.numpy(), ones * 4)
numpy.testing.assert_array_equal(selected[1, 0, 0, :].data.numpy(), ones * 10)
numpy.testing.assert_array_equal(selected[1, 0, 1, :].data.numpy(), ones * 12)
numpy.testing.assert_array_equal(selected[1, 1, 0, :].data.numpy(), ones * 14)
numpy.testing.assert_array_equal(selected[1, 1, 1, :].data.numpy(), ones * 16)
indices = numpy.array([[[1, 11], [3, 4]], [[5, 6], [7, 8]]])
indices = torch.tensor(indices, dtype=torch.long)
with pytest.raises(ConfigurationError):
util.batched_index_select(targets, indices)
indices = numpy.array([[[1, -1], [3, 4]], [[5, 6], [7, 8]]])
indices = torch.tensor(indices, dtype=torch.long)
with pytest.raises(ConfigurationError):
util.batched_index_select(targets, indices)
def test_batched_span_select(self):
# Each element is a vector of its index.
targets = torch.ones([3, 12, 2]).cumsum(1) - 1
spans = torch.LongTensor(
[
[[0, 0], [1, 2], [5, 8], [10, 10]],
[[i, i] for i in range(3, -1, -1)],
[[0, 3], [1, 4], [2, 5], [10, 11]],
]
)
selected, mask = util.batched_span_select(targets, spans)
selected = torch.where(mask.unsqueeze(-1), selected, torch.empty_like(selected).fill_(-1))
numpy.testing.assert_array_equal(
selected,
[
[
[[0, 0], [-1, -1], [-1, -1], [-1, -1]],
[[2, 2], [1, 1], [-1, -1], [-1, -1]],
[[8, 8], [7, 7], [6, 6], [5, 5]],
[[10, 10], [-1, -1], [-1, -1], [-1, -1]],
],
[[[i, i], [-1, -1], [-1, -1], [-1, -1]] for i in range(3, -1, -1)],
[
[[3, 3], [2, 2], [1, 1], [0, 0]],
[[4, 4], [3, 3], [2, 2], [1, 1]],
[[5, 5], [4, 4], [3, 3], [2, 2]],
[[11, 11], [10, 10], [-1, -1], [-1, -1]],
],
],
)
def test_flattened_index_select(self):
indices = numpy.array([[1, 2], [3, 4]])
targets = torch.ones([2, 6, 3]).cumsum(1) - 1
# Make the second batch double its index so they're different.
targets[1, :, :] *= 2
indices = torch.tensor(indices, dtype=torch.long)
selected = util.flattened_index_select(targets, indices)
assert list(selected.size()) == [2, 2, 2, 3]
ones = numpy.ones([3])
numpy.testing.assert_array_equal(selected[0, 0, 0, :].data.numpy(), ones)
numpy.testing.assert_array_equal(selected[0, 0, 1, :].data.numpy(), ones * 2)
numpy.testing.assert_array_equal(selected[0, 1, 0, :].data.numpy(), ones * 3)
numpy.testing.assert_array_equal(selected[0, 1, 1, :].data.numpy(), ones * 4)
numpy.testing.assert_array_equal(selected[1, 0, 0, :].data.numpy(), ones * 2)
numpy.testing.assert_array_equal(selected[1, 0, 1, :].data.numpy(), ones * 4)
numpy.testing.assert_array_equal(selected[1, 1, 0, :].data.numpy(), ones * 6)
numpy.testing.assert_array_equal(selected[1, 1, 1, :].data.numpy(), ones * 8)
# Check we only accept 2D indices.
with pytest.raises(ConfigurationError):
util.flattened_index_select(targets, torch.ones([3, 4, 5]))
def test_bucket_values(self):
indices = torch.LongTensor([1, 2, 7, 1, 56, 900])
bucketed_distances = util.bucket_values(indices)
numpy.testing.assert_array_equal(
bucketed_distances.numpy(), numpy.array([1, 2, 5, 1, 8, 9])
)
def test_add_sentence_boundary_token_ids_handles_2D_input(self):
tensor = torch.from_numpy(numpy.array([[1, 2, 3], [4, 5, 0]]))
mask = tensor > 0
bos = 9
eos = 10
new_tensor, new_mask = util.add_sentence_boundary_token_ids(tensor, mask, bos, eos)
expected_new_tensor = numpy.array([[9, 1, 2, 3, 10], [9, 4, 5, 10, 0]])
assert (new_tensor.data.numpy() == expected_new_tensor).all()
assert (new_mask.data.numpy() == (expected_new_tensor > 0)).all()
def test_add_sentence_boundary_token_ids_handles_3D_input(self):
tensor = torch.from_numpy(
numpy.array(
[
[[1, 2, 3, 4], [5, 5, 5, 5], [6, 8, 1, 2]],
[[4, 3, 2, 1], [8, 7, 6, 5], [0, 0, 0, 0]],
]
)
)
mask = (tensor > 0).sum(dim=-1) > 0
bos = torch.from_numpy(numpy.array([9, 9, 9, 9]))
eos = torch.from_numpy(numpy.array([10, 10, 10, 10]))
new_tensor, new_mask = util.add_sentence_boundary_token_ids(tensor, mask, bos, eos)
expected_new_tensor = numpy.array(
[
[[9, 9, 9, 9], [1, 2, 3, 4], [5, 5, 5, 5], [6, 8, 1, 2], [10, 10, 10, 10]],
[[9, 9, 9, 9], [4, 3, 2, 1], [8, 7, 6, 5], [10, 10, 10, 10], [0, 0, 0, 0]],
]
)
assert (new_tensor.data.numpy() == expected_new_tensor).all()
assert (new_mask.data.numpy() == ((expected_new_tensor > 0).sum(axis=-1) > 0)).all()
def test_remove_sentence_boundaries(self):
tensor = torch.from_numpy(numpy.random.rand(3, 5, 7))
mask = torch.from_numpy(
# The mask with two elements is to test the corner case
# of an empty sequence, so here we are removing boundaries
# from "<S> </S>"
numpy.array([[1, 1, 0, 0, 0], [1, 1, 1, 1, 1], [1, 1, 1, 1, 0]])
).bool()
new_tensor, new_mask = util.remove_sentence_boundaries(tensor, mask)
expected_new_tensor = torch.zeros(3, 3, 7)
expected_new_tensor[1, 0:3, :] = tensor[1, 1:4, :]
expected_new_tensor[2, 0:2, :] = tensor[2, 1:3, :]
assert_array_almost_equal(new_tensor.data.numpy(), expected_new_tensor.data.numpy())
expected_new_mask = torch.from_numpy(numpy.array([[0, 0, 0], [1, 1, 1], [1, 1, 0]])).bool()
assert (new_mask.data.numpy() == expected_new_mask.data.numpy()).all()
def test_add_positional_features(self):
# This is hard to test, so we check that we get the same result as the
# original tensorflow implementation:
# https://github.com/tensorflow/tensor2tensor/blob/master/tensor2tensor/layers/common_attention.py#L270
tensor2tensor_result = numpy.asarray(
[
[0.00000000e00, 0.00000000e00, 1.00000000e00, 1.00000000e00],
[8.41470957e-01, 9.99999902e-05, 5.40302277e-01, 1.00000000e00],
[9.09297407e-01, 1.99999980e-04, -4.16146845e-01, 1.00000000e00],
]
)
tensor = torch.zeros([2, 3, 4])
result = util.add_positional_features(tensor, min_timescale=1.0, max_timescale=1.0e4)
numpy.testing.assert_almost_equal(result[0].detach().cpu().numpy(), tensor2tensor_result)
numpy.testing.assert_almost_equal(result[1].detach().cpu().numpy(), tensor2tensor_result)
# Check case with odd number of dimensions.
tensor2tensor_result = numpy.asarray(
[
[
0.00000000e00,
0.00000000e00,
0.00000000e00,
1.00000000e00,
1.00000000e00,
1.00000000e00,
0.00000000e00,
],
[
8.41470957e-01,
9.99983307e-03,
9.99999902e-05,
5.40302277e-01,
9.99949992e-01,
1.00000000e00,
0.00000000e00,
],
[
9.09297407e-01,
1.99986659e-02,
1.99999980e-04,
-4.16146815e-01,
9.99800026e-01,
1.00000000e00,
0.00000000e00,
],
]
)
tensor = torch.zeros([2, 3, 7])
result = util.add_positional_features(tensor, min_timescale=1.0, max_timescale=1.0e4)
numpy.testing.assert_almost_equal(result[0].detach().cpu().numpy(), tensor2tensor_result)
numpy.testing.assert_almost_equal(result[1].detach().cpu().numpy(), tensor2tensor_result)
def test_combine_tensors_and_multiply(self):
tensors = [torch.Tensor([[[2, 3]]]), torch.Tensor([[[5, 5]]])]
weight = torch.Tensor([4, 5])
combination = "x"
assert_almost_equal(
util.combine_tensors_and_multiply(combination, tensors, weight), [[8 + 15]]
)
combination = "y"
assert_almost_equal(
util.combine_tensors_and_multiply(combination, tensors, weight), [[20 + 25]]
)
combination = "x,y"
weight2 = torch.Tensor([4, 5, 4, 5])
assert_almost_equal(
util.combine_tensors_and_multiply(combination, tensors, weight2), [[8 + 20 + 15 + 25]]
)
combination = "x-y"
assert_almost_equal(
util.combine_tensors_and_multiply(combination, tensors, weight), [[-3 * 4 + -2 * 5]]
)
combination = "y-x"
assert_almost_equal(
util.combine_tensors_and_multiply(combination, tensors, weight), [[3 * 4 + 2 * 5]]
)
combination = "y+x"
assert_almost_equal(
util.combine_tensors_and_multiply(combination, tensors, weight), [[7 * 4 + 8 * 5]]
)
combination = "y*x"
assert_almost_equal(
util.combine_tensors_and_multiply(combination, tensors, weight), [[10 * 4 + 15 * 5]]
)
combination = "y/x"
assert_almost_equal(
util.combine_tensors_and_multiply(combination, tensors, weight),
[[(5 / 2) * 4 + (5 / 3) * 5]],
decimal=4,
)
combination = "x/y"
assert_almost_equal(
util.combine_tensors_and_multiply(combination, tensors, weight),
[[(2 / 5) * 4 + (3 / 5) * 5]],
decimal=4,
)
with pytest.raises(ConfigurationError):
util.combine_tensors_and_multiply("x+y+y", tensors, weight)
with pytest.raises(ConfigurationError):
util.combine_tensors_and_multiply("x%y", tensors, weight)
def test_combine_tensors_and_multiply_with_same_batch_size_and_embedding_dim(self):
# This test just makes sure we handle some potential edge cases where the lengths of all
# dimensions are the same, making sure that the multiplication with the weight vector
# happens along the right dimension (it should be the last one).
tensors = [torch.Tensor([[[5, 5], [4, 4]], [[2, 3], [1, 1]]])] # (2, 2, 2)
weight = torch.Tensor([4, 5]) # (2,)
combination = "x"
assert_almost_equal(
util.combine_tensors_and_multiply(combination, tensors, weight),
[[20 + 25, 16 + 20], [8 + 15, 4 + 5]],
)
tensors = [
torch.Tensor([[[5, 5], [2, 2]], [[4, 4], [3, 3]]]),
torch.Tensor([[[2, 3]], [[1, 1]]]),
]
weight = torch.Tensor([4, 5])
combination = "x*y"
assert_almost_equal(
util.combine_tensors_and_multiply(combination, tensors, weight),
[
[5 * 2 * 4 + 5 * 3 * 5, 2 * 2 * 4 + 2 * 3 * 5],
[4 * 1 * 4 + 4 * 1 * 5, 3 * 1 * 4 + 3 * 1 * 5],
],
)
def test_combine_tensors_and_multiply_with_batch_size_one(self):
seq_len_1 = 10
seq_len_2 = 5
embedding_dim = 8
combination = "x,y,x*y"
t1 = torch.randn(1, seq_len_1, embedding_dim)
t2 = torch.randn(1, seq_len_2, embedding_dim)
combined_dim = util.get_combined_dim(combination, [embedding_dim, embedding_dim])
weight = torch.Tensor(combined_dim)
result = util.combine_tensors_and_multiply(
combination, [t1.unsqueeze(2), t2.unsqueeze(1)], weight
)
assert_almost_equal(result.size(), [1, seq_len_1, seq_len_2])
def test_combine_tensors_and_multiply_with_batch_size_one_and_seq_len_one(self):
seq_len_1 = 10
seq_len_2 = 1
embedding_dim = 8
combination = "x,y,x*y"
t1 = torch.randn(1, seq_len_1, embedding_dim)
t2 = torch.randn(1, seq_len_2, embedding_dim)
combined_dim = util.get_combined_dim(combination, [embedding_dim, embedding_dim])
weight = torch.Tensor(combined_dim)
result = util.combine_tensors_and_multiply(
combination, [t1.unsqueeze(2), t2.unsqueeze(1)], weight
)
assert_almost_equal(result.size(), [1, seq_len_1, seq_len_2])
def test_has_tensor(self):
has_tensor = util.has_tensor
tensor = torch.tensor([1, 2, 3])
assert has_tensor(["a", 10, tensor])
assert not has_tensor(["a", 10])
assert has_tensor(("a", 10, tensor))
assert not has_tensor(("a", 10))
assert has_tensor({"a": tensor, "b": 1})
assert not has_tensor({"a": 10, "b": 1})
assert has_tensor(tensor)
assert not has_tensor(3)
assert has_tensor({"x": [0, {"inside": {"double_inside": [3, [10, tensor]]}}]})
def test_combine_initial_dims(self):
tensor = torch.randn(4, 10, 20, 17, 5)
tensor2d = util.combine_initial_dims(tensor)
assert list(tensor2d.size()) == [4 * 10 * 20 * 17, 5]
def test_uncombine_initial_dims(self):
embedding2d = torch.randn(4 * 10 * 20 * 17 * 5, 12)
embedding = util.uncombine_initial_dims(embedding2d, torch.Size((4, 10, 20, 17, 5)))
assert list(embedding.size()) == [4, 10, 20, 17, 5, 12]
def test_inspect_model_parameters(self):
model_archive = str(
self.FIXTURES_ROOT / "decomposable_attention" / "serialization" / "model.tar.gz"
)
parameters_inspection = str(
self.FIXTURES_ROOT / "decomposable_attention" / "parameters_inspection.json"
)
model = load_archive(model_archive).model
with open(parameters_inspection) as file:
parameters_inspection_dict = json.load(file)
assert parameters_inspection_dict == util.inspect_parameters(model)
def test_move_to_device(self):
# We're faking the tensor here so that we can test the calls to .cuda() without actually
# needing a GPU.
class FakeTensor(torch.Tensor):
def __init__(self):
self._device = None
def cuda(self, device):
self._device = device
return self
class A(NamedTuple):
a: int
b: torch.Tensor
structured_obj = {
"a": [A(1, FakeTensor()), A(2, FakeTensor())],
"b": FakeTensor(),
"c": (1, FakeTensor()),
}
new_device = 4
moved_obj = util.move_to_device(structured_obj, new_device)
assert moved_obj["a"][0].a == 1
assert moved_obj["a"][0].b._device == new_device
assert moved_obj["a"][1].b._device == new_device
assert moved_obj["b"]._device == new_device
assert moved_obj["c"][0] == 1
assert moved_obj["c"][1]._device == new_device
def test_extend_layer(self):
lin_layer = torch.nn.Linear(10, 5)
new_dim = 8
old_weights = lin_layer.weight.data.clone()
old_bias = lin_layer.bias.data.clone()
util.extend_layer(lin_layer, new_dim)
assert lin_layer.weight.data.shape == (8, 10)
assert lin_layer.bias.data.shape == (8,)
assert (lin_layer.weight.data[:5] == old_weights).all()
assert (lin_layer.bias.data[:5] == old_bias).all()
assert lin_layer.out_features == new_dim
def test_masked_topk_selects_top_scored_items_and_respects_masking(self):
items = torch.randn([3, 4, 5]).clamp(min=0.0, max=1.0)
items[0, :2, :] = 1
items[1, 2:, :] = 1
items[2, 2:, :] = 1
scores = items.sum(-1)
mask = torch.ones([3, 4]).bool()
mask[1, 0] = 0
mask[1, 3] = 0
pruned_scores, pruned_mask, pruned_indices = util.masked_topk(scores, mask, 2)
# Second element in the batch would have indices 2, 3, but
# 3 and 0 are masked, so instead it has 1, 2.
numpy.testing.assert_array_equal(
pruned_indices.data.numpy(), numpy.array([[0, 1], [1, 2], [2, 3]])
)
numpy.testing.assert_array_equal(pruned_mask.data.numpy(), numpy.ones([3, 2]))
# scores should be the result of index_selecting the pruned_indices.
correct_scores = util.batched_index_select(scores.unsqueeze(-1), pruned_indices).squeeze(-1)
self.assert_array_equal_with_mask(correct_scores, pruned_scores, pruned_mask)
def test_masked_topk_works_for_completely_masked_rows(self):
items = torch.randn([3, 4, 5]).clamp(min=0.0, max=1.0)
items[0, :2, :] = 1
items[1, 2:, :] = 1
items[2, 2:, :] = 1
scores = items.sum(-1)
mask = torch.ones([3, 4]).bool()
mask[1, 0] = 0
mask[1, 3] = 0
mask[2, :] = 0 # fully masked last batch element.
pruned_scores, pruned_mask, pruned_indices = util.masked_topk(scores, mask, 2)
# We can't check the last row here, because it's completely masked.
# Instead we'll check that the scores for these elements are very small.
numpy.testing.assert_array_equal(
pruned_indices[:2].data.numpy(), numpy.array([[0, 1], [1, 2]])
)
numpy.testing.assert_array_equal(
pruned_mask.data.numpy(), numpy.array([[1, 1], [1, 1], [0, 0]])
)
# scores should be the result of index_selecting the pruned_indices.
correct_scores = util.batched_index_select(scores.unsqueeze(-1), pruned_indices).squeeze(-1)
self.assert_array_equal_with_mask(correct_scores, pruned_scores, pruned_mask)
def test_masked_topk_selects_top_scored_items_and_respects_masking_different_num_items(self):
items = torch.randn([3, 4, 5]).clamp(min=0.0, max=1.0)
items[0, 0, :] = 1.5
items[0, 1, :] = 2
items[0, 3, :] = 1
items[1, 1:3, :] = 1
items[2, 0, :] = 1
items[2, 1, :] = 2
items[2, 2, :] = 1.5
scores = items.sum(-1)
mask = torch.ones([3, 4]).bool()
mask[1, 3] = 0
k = torch.tensor([3, 2, 1], dtype=torch.long)
pruned_scores, pruned_mask, pruned_indices = util.masked_topk(scores, mask, k)
# Second element in the batch would have indices 2, 3, but
# 3 and 0 are masked, so instead it has 1, 2.
numpy.testing.assert_array_equal(
pruned_indices.data.numpy(), numpy.array([[0, 1, 3], [1, 2, 2], [1, 2, 2]])
)
numpy.testing.assert_array_equal(
pruned_mask.data.numpy(), numpy.array([[1, 1, 1], [1, 1, 0], [1, 0, 0]])
)
# scores should be the result of index_selecting the pruned_indices.
correct_scores = util.batched_index_select(scores.unsqueeze(-1), pruned_indices).squeeze(-1)
self.assert_array_equal_with_mask(correct_scores, pruned_scores, pruned_mask)
def test_masked_topk_works_for_row_with_no_items_requested(self):
# Case where `num_items_to_keep` is a tensor rather than an int. Make sure it does the right
# thing when no items are requested for one of the rows.
items = torch.randn([3, 4, 5]).clamp(min=0.0, max=1.0)
items[0, :3, :] = 1
items[1, 2:, :] = 1
items[2, 2:, :] = 1
scores = items.sum(-1)
mask = torch.ones([3, 4]).bool()
mask[1, 0] = 0
mask[1, 3] = 0
k = torch.tensor([3, 2, 0], dtype=torch.long)
pruned_scores, pruned_mask, pruned_indices = util.masked_topk(scores, mask, k)
# First element just picks top three entries. Second would pick entries 2 and 3, but 0 and 3
# are masked, so it takes 1 and 2 (repeating the second index). The third element is
# entirely masked and just repeats the largest index with a top-3 score.
numpy.testing.assert_array_equal(
pruned_indices.data.numpy(), numpy.array([[0, 1, 2], [1, 2, 2], [3, 3, 3]])
)
numpy.testing.assert_array_equal(
pruned_mask.data.numpy(), numpy.array([[1, 1, 1], [1, 1, 0], [0, 0, 0]])
)
# scores should be the result of index_selecting the pruned_indices.
correct_scores = util.batched_index_select(scores.unsqueeze(-1), pruned_indices).squeeze(-1)
self.assert_array_equal_with_mask(correct_scores, pruned_scores, pruned_mask)
def test_masked_topk_works_for_multiple_dimensions(self):
# fmt: off
items = torch.FloatTensor([ # (3, 2, 5)
[[4, 2, 9, 9, 7], [-4, -2, -9, -9, -7]],
[[5, 4, 1, 8, 8], [9, 1, 7, 4, 1]],
[[9, 8, 9, 6, 0], [2, 2, 2, 2, 2]],
]).unsqueeze(-1).expand(3, 2, 5, 4)
mask = torch.tensor([
[[False, False, False, False, False], [True, True, True, True, True]],
[[True, True, True, True, False], [False, True, True, True, True]],
[[True, False, True, True, True], [False, True, False, True, True]],
]).unsqueeze(-1).expand(3, 2, 5, 4)
# This is the same as just specifying a scalar int, but we want to test this behavior
k = torch.ones(3, 5, 4, dtype=torch.long)
k[1, 3, :] = 2
target_items = torch.FloatTensor([
[[-4, -2, -9, -9, -7], [0, 0, 0, 0, 0]],
[[5, 4, 7, 8, 1], [0, 0, 0, 4, 0]],
[[9, 2, 9, 6, 2], [0, 0, 0, 0, 0]],
]).unsqueeze(-1).expand(3, 2, 5, 4)
target_mask = torch.ones(3, 2, 5, 4, dtype=torch.bool)
target_mask[:, 1, :, :] = 0
target_mask[1, 1, 3, :] = 1
target_indices = torch.LongTensor([
[[1, 1, 1, 1, 1], [0, 0, 0, 0, 0]],
[[0, 0, 1, 0, 1], [0, 0, 0, 1, 0]],
[[0, 1, 0, 0, 1], [0, 0, 0, 0, 0]],
]).unsqueeze(-1).expand(3, 2, 5, 4)
# fmt: on
pruned_items, pruned_mask, pruned_indices = util.masked_topk(items, mask, k, dim=1)
numpy.testing.assert_array_equal(pruned_mask.data.numpy(), target_mask.data.numpy())
self.assert_array_equal_with_mask(pruned_items, target_items, pruned_mask)
self.assert_array_equal_with_mask(pruned_indices, target_indices, pruned_mask)
def assert_array_equal_with_mask(self, a, b, mask):
numpy.testing.assert_array_equal((a * mask).data.numpy(), (b * mask).data.numpy())
def test_tensors_equal(self):
# Basic
assert util.tensors_equal(torch.tensor([1]), torch.tensor([1]))
assert not util.tensors_equal(torch.tensor([1]), torch.tensor([2]))
# Bool
assert util.tensors_equal(torch.tensor([True]), torch.tensor([True]))
# Cross dtype
assert util.tensors_equal(torch.tensor([1]), torch.tensor([1.0]))
assert util.tensors_equal(torch.tensor([1]), torch.tensor([True]))
# Containers
assert util.tensors_equal([torch.tensor([1])], [torch.tensor([1])])
assert not util.tensors_equal([torch.tensor([1])], [torch.tensor([2])])
assert util.tensors_equal({"key": torch.tensor([1])}, {"key": torch.tensor([1])})
|
[
"allennlp.nn.util.bucket_values",
"allennlp.nn.util.masked_topk",
"numpy.random.rand",
"torch.LongTensor",
"allennlp.nn.util.get_text_field_mask",
"torch.max",
"allennlp.nn.util.get_combined_dim",
"torch.from_numpy",
"allennlp.nn.util.add_sentence_boundary_token_ids",
"numpy.array",
"allennlp.nn.util.logsumexp",
"allennlp.nn.util.inspect_parameters",
"allennlp.nn.util.clamp_tensor",
"allennlp.nn.util.masked_max",
"allennlp.nn.util.flattened_index_select",
"torch.nn.functional.softmax",
"allennlp.nn.util.sequence_cross_entropy_with_logits",
"numpy.testing.assert_array_almost_equal",
"allennlp.nn.util.viterbi_decode",
"allennlp.nn.util.combine_initial_dims",
"flaky.flaky",
"numpy.asarray",
"allennlp.nn.util.get_mask_from_sequence_lengths",
"numpy.exp",
"numpy.testing.assert_almost_equal",
"allennlp.nn.util.weighted_sum",
"allennlp.nn.util.get_final_encoder_states",
"allennlp.nn.util.masked_mean",
"torch.randn",
"random.randint",
"numpy.ones",
"allennlp.nn.util.flatten_and_batch_shift_indices",
"torch.Tensor",
"torch.empty_like",
"allennlp.nn.util.remove_sentence_boundaries",
"pytest.raises",
"numpy.isnan",
"torch.nn.functional.log_softmax",
"allennlp.nn.util.masked_log_softmax",
"torch.Size",
"numpy.random.randn",
"allennlp.nn.util.get_lengths_from_binary_sequence_mask",
"allennlp.common.util.sanitize",
"allennlp.models.load_archive",
"allennlp.nn.util.sort_batch_by_length",
"allennlp.nn.util.masked_softmax",
"allennlp.nn.util.move_to_device",
"allennlp.nn.util.combine_tensors_and_multiply",
"torch.FloatTensor",
"json.load",
"torch.tensor",
"allennlp.nn.util.add_positional_features",
"allennlp.nn.util.extend_layer",
"allennlp.nn.util.masked_flip",
"numpy.sum",
"torch.nn.Linear",
"allennlp.nn.util.batched_index_select",
"numpy.random.randint",
"allennlp.nn.util.batched_span_select",
"torch.zeros",
"torch.rand",
"torch.ones"
] |
[((40211, 40242), 'flaky.flaky', 'flaky', ([], {'max_runs': '(3)', 'min_passes': '(1)'}), '(max_runs=3, min_passes=1)\n', (40216, 40242), False, 'from flaky import flaky\n'), ((537, 725), 'torch.tensor', 'torch.tensor', (['[[True, True, True, False, False, False], [True, True, False, False, False,\n False], [True, True, True, True, True, True], [True, False, False, \n False, False, False]]'], {}), '([[True, True, True, False, False, False], [True, True, False, \n False, False, False], [True, True, True, True, True, True], [True, \n False, False, False, False, False]])\n', (549, 725), False, 'import torch\n'), ((835, 890), 'allennlp.nn.util.get_lengths_from_binary_sequence_mask', 'util.get_lengths_from_binary_sequence_mask', (['binary_mask'], {}), '(binary_mask)\n', (877, 890), False, 'from allennlp.nn import util\n'), ((1055, 1088), 'torch.LongTensor', 'torch.LongTensor', (['[4, 3, 1, 4, 2]'], {}), '([4, 3, 1, 4, 2])\n', (1071, 1088), False, 'import torch\n'), ((1182, 1299), 'numpy.testing.assert_almost_equal', 'assert_almost_equal', (['mask', '[[1, 1, 1, 1, 0], [1, 1, 1, 0, 0], [1, 0, 0, 0, 0], [1, 1, 1, 1, 0], [1, 1,\n 0, 0, 0]]'], {}), '(mask, [[1, 1, 1, 1, 0], [1, 1, 1, 0, 0], [1, 0, 0, 0, 0\n ], [1, 1, 1, 1, 0], [1, 1, 0, 0, 0]])\n', (1201, 1299), False, 'from numpy.testing import assert_array_almost_equal, assert_almost_equal\n'), ((1805, 1860), 'allennlp.nn.util.get_lengths_from_binary_sequence_mask', 'util.get_lengths_from_binary_sequence_mask', (['binary_mask'], {}), '(binary_mask)\n', (1847, 1860), False, 'from allennlp.nn import util\n'), ((2039, 2085), 'torch.LongTensor', 'torch.LongTensor', (['[[0, 1, 1, 0], [2, 0, 2, 2]]'], {}), '([[0, 1, 1, 0], [2, 0, 2, 2]])\n', (2055, 2085), False, 'import torch\n'), ((2098, 2130), 'torch.FloatTensor', 'torch.FloatTensor', (['[3, 4, -5, 3]'], {}), '([3, 4, -5, 3])\n', (2115, 2130), False, 'import torch\n'), ((2292, 2352), 'numpy.testing.assert_almost_equal', 'assert_almost_equal', (['clamped_tensor', '[[0, 0, 3], [3, 0, -3]]'], {}), '(clamped_tensor, [[0, 0, 3], [3, 0, -3]])\n', (2311, 2352), False, 'from numpy.testing import assert_array_almost_equal, assert_almost_equal\n'), ((2408, 2448), 'torch.LongTensor', 'torch.LongTensor', (['[[0, 1, 1], [2, 0, 2]]'], {}), '([[0, 1, 1], [2, 0, 2]])\n', (2424, 2448), False, 'import torch\n'), ((2461, 2490), 'torch.FloatTensor', 'torch.FloatTensor', (['[3, 4, -5]'], {}), '([3, 4, -5])\n', (2478, 2490), False, 'import torch\n'), ((2652, 2712), 'numpy.testing.assert_almost_equal', 'assert_almost_equal', (['clamped_tensor', '[[0, 0, 3], [3, 0, -3]]'], {}), '(clamped_tensor, [[0, 0, 3], [3, 0, -3]])\n', (2671, 2712), False, 'from numpy.testing import assert_array_almost_equal, assert_almost_equal\n'), ((2762, 2802), 'torch.tensor', 'torch.tensor', (['[[5, -4, 3], [-3, 0, -30]]'], {}), '([[5, -4, 3], [-3, 0, -30]])\n', (2774, 2802), False, 'import torch\n'), ((2828, 2876), 'allennlp.nn.util.clamp_tensor', 'util.clamp_tensor', (['tensor'], {'minimum': '(-3)', 'maximum': '(3)'}), '(tensor, minimum=-3, maximum=3)\n', (2845, 2876), False, 'from allennlp.nn import util\n'), ((2885, 2947), 'numpy.testing.assert_almost_equal', 'assert_almost_equal', (['clamped_tensor', '[[3, -3, 3], [-3, 0, -3]]'], {}), '(clamped_tensor, [[3, -3, 3], [-3, 0, -3]])\n', (2904, 2947), False, 'from numpy.testing import assert_array_almost_equal, assert_almost_equal\n'), ((3008, 3029), 'torch.rand', 'torch.rand', (['[5, 7, 9]'], {}), '([5, 7, 9])\n', (3018, 3029), False, 'import torch\n'), ((3174, 3207), 'torch.LongTensor', 'torch.LongTensor', (['[3, 4, 1, 5, 7]'], {}), '([3, 4, 1, 5, 7])\n', (3190, 3207), False, 'import torch\n'), ((3268, 3319), 'allennlp.nn.util.sort_batch_by_length', 'util.sort_batch_by_length', (['tensor', 'sequence_lengths'], {}), '(tensor, sequence_lengths)\n', (3293, 3319), False, 'from allennlp.nn import util\n'), ((4041, 4163), 'torch.Tensor', 'torch.Tensor', (['[[[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12]], [[13, 14, 15, 16], [17, 18,\n 19, 20], [21, 22, 23, 24]]]'], {}), '([[[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12]], [[13, 14, 15, \n 16], [17, 18, 19, 20], [21, 22, 23, 24]]])\n', (4053, 4163), False, 'import torch\n'), ((4243, 4298), 'torch.tensor', 'torch.tensor', (['[[True, True, True], [True, True, False]]'], {}), '([[True, True, True], [True, True, False]])\n', (4255, 4298), False, 'import torch\n'), ((4322, 4395), 'allennlp.nn.util.get_final_encoder_states', 'util.get_final_encoder_states', (['encoder_outputs', 'mask'], {'bidirectional': '(False)'}), '(encoder_outputs, mask, bidirectional=False)\n', (4351, 4395), False, 'from allennlp.nn import util\n'), ((4511, 4583), 'allennlp.nn.util.get_final_encoder_states', 'util.get_final_encoder_states', (['encoder_outputs', 'mask'], {'bidirectional': '(True)'}), '(encoder_outputs, mask, bidirectional=True)\n', (4540, 4583), False, 'from allennlp.nn import util\n'), ((4786, 4822), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 2.0, 3.0]]'], {}), '([[1.0, 2.0, 3.0]])\n', (4803, 4822), False, 'import torch\n'), ((5124, 5160), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 2.0, 5.0]]'], {}), '([[1.0, 2.0, 5.0]])\n', (5141, 5160), False, 'import torch\n'), ((5431, 5467), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 0.0, 0.0]]'], {}), '([[0.0, 0.0, 0.0]])\n', (5448, 5467), False, 'import torch\n'), ((5755, 5808), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 2.0, 5.0], [1.0, 2.0, 3.0]]'], {}), '([[1.0, 2.0, 5.0], [1.0, 2.0, 3.0]])\n', (5772, 5808), False, 'import torch\n'), ((6203, 6256), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 2.0, 5.0], [0.0, 0.0, 0.0]]'], {}), '([[1.0, 2.0, 5.0], [0.0, 0.0, 0.0]])\n', (6220, 6256), False, 'import torch\n'), ((6662, 6698), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 2.0, 5.0]]'], {}), '([[1.0, 2.0, 5.0]])\n', (6679, 6698), False, 'import torch\n'), ((6717, 6752), 'torch.tensor', 'torch.tensor', (['[[True, False, True]]'], {}), '([[True, False, True]])\n', (6729, 6752), False, 'import torch\n'), ((6958, 6999), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 2.0, 3.0, 4.0]]'], {}), '([[0.0, 2.0, 3.0, 4.0]])\n', (6975, 6999), False, 'import torch\n'), ((7018, 7059), 'torch.tensor', 'torch.tensor', (['[[True, False, True, True]]'], {}), '([[True, False, True, True]])\n', (7030, 7059), False, 'import torch\n'), ((7400, 7441), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 0.0, 0.0, 0.0]]'], {}), '([[0.0, 0.0, 0.0, 0.0]])\n', (7417, 7441), False, 'import torch\n'), ((7460, 7503), 'torch.tensor', 'torch.tensor', (['[[False, False, False, True]]'], {}), '([[False, False, False, True]])\n', (7472, 7503), False, 'import torch\n'), ((7793, 7834), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 2.0, 3.0, 4.0]]'], {}), '([[0.0, 2.0, 3.0, 4.0]])\n', (7810, 7834), False, 'import torch\n'), ((7853, 7897), 'torch.tensor', 'torch.tensor', (['[[False, False, False, False]]'], {}), '([[False, False, False, False]])\n', (7865, 7897), False, 'import torch\n'), ((8191, 8232), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 0.0, 0.0, 0.0]]'], {}), '([[0.0, 0.0, 0.0, 0.0]])\n', (8208, 8232), False, 'import torch\n'), ((8251, 8295), 'torch.tensor', 'torch.tensor', (['[[False, False, False, False]]'], {}), '([[False, False, False, False]])\n', (8263, 8295), False, 'import torch\n'), ((8586, 8627), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 1.0, 100000.0]]'], {}), '([[1.0, 1.0, 100000.0]])\n', (8603, 8627), False, 'import torch\n'), ((8641, 8676), 'torch.tensor', 'torch.tensor', (['[[True, True, False]]'], {}), '([[True, True, False]])\n', (8653, 8676), False, 'import torch\n'), ((8914, 8967), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 2.0, 5.0], [1.0, 2.0, 3.0]]'], {}), '([[1.0, 2.0, 5.0], [1.0, 2.0, 3.0]])\n', (8931, 8967), False, 'import torch\n'), ((8983, 9038), 'torch.tensor', 'torch.tensor', (['[[True, False, True], [True, True, True]]'], {}), '([[True, False, True], [True, True, True]])\n', (8995, 9038), False, 'import torch\n'), ((9427, 9480), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]]'], {}), '([[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]])\n', (9444, 9480), False, 'import torch\n'), ((9496, 9551), 'torch.tensor', 'torch.tensor', (['[[True, False, True], [True, True, True]]'], {}), '([[True, False, True], [True, True, True]])\n', (9508, 9551), False, 'import torch\n'), ((9912, 9965), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]]'], {}), '([[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]])\n', (9929, 9965), False, 'import torch\n'), ((9981, 10039), 'torch.tensor', 'torch.tensor', (['[[True, False, True], [False, False, False]]'], {}), '([[True, False, True], [False, False, False]])\n', (9993, 10039), False, 'import torch\n'), ((10269, 10322), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]]'], {}), '([[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]])\n', (10286, 10322), False, 'import torch\n'), ((10338, 10396), 'torch.tensor', 'torch.tensor', (['[[False, False, False], [True, False, True]]'], {}), '([[False, False, False], [True, False, True]])\n', (10350, 10396), False, 'import torch\n'), ((10748, 10784), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 2.0, 5.0]]'], {}), '([[1.0, 2.0, 5.0]])\n', (10765, 10784), False, 'import torch\n'), ((10803, 10838), 'torch.tensor', 'torch.tensor', (['[[True, False, True]]'], {}), '([[True, False, True]])\n', (10815, 10838), False, 'import torch\n'), ((11089, 11130), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 2.0, 3.0, 4.0]]'], {}), '([[0.0, 2.0, 3.0, 4.0]])\n', (11106, 11130), False, 'import torch\n'), ((11149, 11190), 'torch.tensor', 'torch.tensor', (['[[True, False, True, True]]'], {}), '([[True, False, True, True]])\n', (11161, 11190), False, 'import torch\n'), ((11576, 11617), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 0.0, 0.0, 0.0]]'], {}), '([[0.0, 0.0, 0.0, 0.0]])\n', (11593, 11617), False, 'import torch\n'), ((11636, 11679), 'torch.tensor', 'torch.tensor', (['[[False, False, False, True]]'], {}), '([[False, False, False, True]])\n', (11648, 11679), False, 'import torch\n'), ((12014, 12055), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 2.0, 3.0, 4.0]]'], {}), '([[0.0, 2.0, 3.0, 4.0]])\n', (12031, 12055), False, 'import torch\n'), ((12074, 12118), 'torch.tensor', 'torch.tensor', (['[[False, False, False, False]]'], {}), '([[False, False, False, False]])\n', (12086, 12118), False, 'import torch\n'), ((12461, 12502), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 0.0, 0.0, 0.0]]'], {}), '([[0.0, 0.0, 0.0, 0.0]])\n', (12478, 12502), False, 'import torch\n'), ((12521, 12565), 'torch.tensor', 'torch.tensor', (['[[False, False, False, False]]'], {}), '([[False, False, False, False]])\n', (12533, 12565), False, 'import torch\n'), ((12905, 12946), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 1.0, 100000.0]]'], {}), '([[1.0, 1.0, 100000.0]])\n', (12922, 12946), False, 'import torch\n'), ((12960, 12995), 'torch.tensor', 'torch.tensor', (['[[True, True, False]]'], {}), '([[True, True, False]])\n', (12972, 12995), False, 'import torch\n'), ((13278, 13331), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 2.0, 5.0], [1.0, 2.0, 3.0]]'], {}), '([[1.0, 2.0, 5.0], [1.0, 2.0, 3.0]])\n', (13295, 13331), False, 'import torch\n'), ((13347, 13402), 'torch.tensor', 'torch.tensor', (['[[True, False, True], [True, True, True]]'], {}), '([[True, False, True], [True, True, True]])\n', (13359, 13402), False, 'import torch\n'), ((13836, 13889), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]]'], {}), '([[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]])\n', (13853, 13889), False, 'import torch\n'), ((13905, 13960), 'torch.tensor', 'torch.tensor', (['[[True, False, True], [True, True, True]]'], {}), '([[True, False, True], [True, True, True]])\n', (13917, 13960), False, 'import torch\n'), ((14366, 14419), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]]'], {}), '([[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]])\n', (14383, 14419), False, 'import torch\n'), ((14435, 14493), 'torch.tensor', 'torch.tensor', (['[[True, False, True], [False, False, False]]'], {}), '([[True, False, True], [False, False, False]])\n', (14447, 14493), False, 'import torch\n'), ((14802, 14855), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]]'], {}), '([[0.0, 0.0, 0.0], [1.0, 2.0, 3.0]])\n', (14819, 14855), False, 'import torch\n'), ((14871, 14929), 'torch.tensor', 'torch.tensor', (['[[False, False, False], [True, False, True]]'], {}), '([[False, False, False], [True, False, True]])\n', (14883, 14929), False, 'import torch\n'), ((15520, 15556), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 2.0, 5.0]]'], {}), '([[1.0, 2.0, 5.0]])\n', (15537, 15556), False, 'import torch\n'), ((15575, 15610), 'torch.tensor', 'torch.tensor', (['[[True, False, True]]'], {}), '([[True, False, True]])\n', (15587, 15610), False, 'import torch\n'), ((15853, 15894), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 2.0, 3.0, 4.0]]'], {}), '([[0.0, 2.0, 3.0, 4.0]])\n', (15870, 15894), False, 'import torch\n'), ((15913, 15954), 'torch.tensor', 'torch.tensor', (['[[True, False, True, True]]'], {}), '([[True, False, True, True]])\n', (15925, 15954), False, 'import torch\n'), ((16310, 16351), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 0.0, 0.0, 0.0]]'], {}), '([[0.0, 0.0, 0.0, 0.0]])\n', (16327, 16351), False, 'import torch\n'), ((16370, 16413), 'torch.tensor', 'torch.tensor', (['[[False, False, False, True]]'], {}), '([[False, False, False, True]])\n', (16382, 16413), False, 'import torch\n'), ((16810, 16851), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.0, 2.0, 3.0, 4.0]]'], {}), '([[0.0, 2.0, 3.0, 4.0]])\n', (16827, 16851), False, 'import torch\n'), ((16870, 16914), 'torch.tensor', 'torch.tensor', (['[[False, False, False, False]]'], {}), '([[False, False, False, False]])\n', (16882, 16914), False, 'import torch\n'), ((17158, 17193), 'torch.FloatTensor', 'torch.FloatTensor', (['[1.0, 12.0, 5.0]'], {}), '([1.0, 12.0, 5.0])\n', (17175, 17193), False, 'import torch\n'), ((17212, 17245), 'torch.tensor', 'torch.tensor', (['[True, False, True]'], {}), '([True, False, True])\n', (17224, 17245), False, 'import torch\n'), ((17336, 17383), 'numpy.testing.assert_array_almost_equal', 'assert_array_almost_equal', (['vector_1d_maxed', '(5.0)'], {}), '(vector_1d_maxed, 5.0)\n', (17361, 17383), False, 'from numpy.testing import assert_array_almost_equal, assert_almost_equal\n'), ((17502, 17537), 'torch.FloatTensor', 'torch.FloatTensor', (['[1.0, 12.0, 5.0]'], {}), '([1.0, 12.0, 5.0])\n', (17519, 17537), False, 'import torch\n'), ((17556, 17591), 'torch.tensor', 'torch.tensor', (['[False, False, False]'], {}), '([False, False, False])\n', (17568, 17591), False, 'import torch\n'), ((17792, 17848), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 12.0, 5.0], [-1.0, -2.0, 3.0]]'], {}), '([[1.0, 12.0, 5.0], [-1.0, -2.0, 3.0]])\n', (17809, 17848), False, 'import torch\n'), ((17864, 17920), 'torch.tensor', 'torch.tensor', (['[[True, False, True], [True, True, False]]'], {}), '([[True, False, True], [True, True, False]])\n', (17876, 17920), False, 'import torch\n'), ((18145, 18201), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 12.0, 5.0], [-1.0, -2.0, 3.0]]'], {}), '([[1.0, 12.0, 5.0], [-1.0, -2.0, 3.0]])\n', (18162, 18201), False, 'import torch\n'), ((18217, 18273), 'torch.tensor', 'torch.tensor', (['[[True, False, True], [True, True, False]]'], {}), '([[True, False, True], [True, True, False]])\n', (18229, 18273), False, 'import torch\n'), ((18486, 18592), 'torch.FloatTensor', 'torch.FloatTensor', (['[[[1.0, 2.0], [12.0, 3.0], [5.0, -1.0]], [[-1.0, -3.0], [-2.0, -0.5], [3.0,\n 8.0]]]'], {}), '([[[1.0, 2.0], [12.0, 3.0], [5.0, -1.0]], [[-1.0, -3.0], [\n -2.0, -0.5], [3.0, 8.0]]])\n', (18503, 18592), False, 'import torch\n'), ((18957, 18992), 'torch.FloatTensor', 'torch.FloatTensor', (['[1.0, 12.0, 5.0]'], {}), '([1.0, 12.0, 5.0])\n', (18974, 18992), False, 'import torch\n'), ((19011, 19044), 'torch.tensor', 'torch.tensor', (['[True, False, True]'], {}), '([True, False, True])\n', (19023, 19044), False, 'import torch\n'), ((19135, 19181), 'numpy.testing.assert_array_almost_equal', 'assert_array_almost_equal', (['vector_1d_mean', '(3.0)'], {}), '(vector_1d_mean, 3.0)\n', (19160, 19181), False, 'from numpy.testing import assert_array_almost_equal, assert_almost_equal\n'), ((19300, 19335), 'torch.FloatTensor', 'torch.FloatTensor', (['[1.0, 12.0, 5.0]'], {}), '([1.0, 12.0, 5.0])\n', (19317, 19335), False, 'import torch\n'), ((19354, 19389), 'torch.tensor', 'torch.tensor', (['[False, False, False]'], {}), '([False, False, False])\n', (19366, 19389), False, 'import torch\n'), ((19589, 19645), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 12.0, 5.0], [-1.0, -2.0, 3.0]]'], {}), '([[1.0, 12.0, 5.0], [-1.0, -2.0, 3.0]])\n', (19606, 19645), False, 'import torch\n'), ((19661, 19717), 'torch.tensor', 'torch.tensor', (['[[True, False, True], [True, True, False]]'], {}), '([[True, False, True], [True, True, False]])\n', (19673, 19717), False, 'import torch\n'), ((19941, 19997), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1.0, 12.0, 5.0], [-1.0, -2.0, 3.0]]'], {}), '([[1.0, 12.0, 5.0], [-1.0, -2.0, 3.0]])\n', (19958, 19997), False, 'import torch\n'), ((20013, 20069), 'torch.tensor', 'torch.tensor', (['[[True, False, True], [True, True, False]]'], {}), '([[True, False, True], [True, True, False]])\n', (20025, 20069), False, 'import torch\n'), ((20281, 20387), 'torch.FloatTensor', 'torch.FloatTensor', (['[[[1.0, 2.0], [12.0, 3.0], [5.0, -1.0]], [[-1.0, -3.0], [-2.0, -0.5], [3.0,\n 8.0]]]'], {}), '([[[1.0, 2.0], [12.0, 3.0], [5.0, -1.0]], [[-1.0, -3.0], [\n -2.0, -0.5], [3.0, 8.0]]])\n', (20298, 20387), False, 'import torch\n'), ((20703, 20796), 'torch.FloatTensor', 'torch.FloatTensor', (['[[[6, 6, 6], [1, 1, 1], [2, 2, 2]], [[3, 3, 3], [4, 4, 4], [5, 5, 5]]]'], {}), '([[[6, 6, 6], [1, 1, 1], [2, 2, 2]], [[3, 3, 3], [4, 4, 4],\n [5, 5, 5]]])\n', (20720, 20796), False, 'import torch\n'), ((20902, 20934), 'allennlp.nn.util.masked_flip', 'util.masked_flip', (['tensor', '[1, 2]'], {}), '(tensor, [1, 2])\n', (20918, 20934), False, 'from allennlp.nn import util\n'), ((20943, 20982), 'numpy.testing.assert_almost_equal', 'assert_almost_equal', (['response', 'solution'], {}), '(response, solution)\n', (20962, 20982), False, 'from numpy.testing import assert_array_almost_equal, assert_almost_equal\n'), ((21001, 21116), 'torch.FloatTensor', 'torch.FloatTensor', (['[[[6, 6, 6], [1, 1, 1], [2, 2, 2], [0, 0, 0]], [[3, 3, 3], [4, 4, 4], [5, 5,\n 5], [1, 2, 3]]]'], {}), '([[[6, 6, 6], [1, 1, 1], [2, 2, 2], [0, 0, 0]], [[3, 3, 3],\n [4, 4, 4], [5, 5, 5], [1, 2, 3]]])\n', (21018, 21116), False, 'import torch\n'), ((21348, 21380), 'allennlp.nn.util.masked_flip', 'util.masked_flip', (['tensor', '[3, 4]'], {}), '(tensor, [3, 4])\n', (21364, 21380), False, 'from allennlp.nn import util\n'), ((21389, 21428), 'numpy.testing.assert_almost_equal', 'assert_almost_equal', (['response', 'solution'], {}), '(response, solution)\n', (21408, 21428), False, 'from numpy.testing import assert_array_almost_equal, assert_almost_equal\n'), ((21447, 21612), 'torch.FloatTensor', 'torch.FloatTensor', (['[[[6, 6, 6], [1, 1, 1], [2, 2, 2], [0, 0, 0]], [[3, 3, 3], [4, 4, 4], [5, 5,\n 5], [1, 2, 3]], [[1, 1, 1], [2, 2, 2], [0, 0, 0], [0, 0, 0]]]'], {}), '([[[6, 6, 6], [1, 1, 1], [2, 2, 2], [0, 0, 0]], [[3, 3, 3],\n [4, 4, 4], [5, 5, 5], [1, 2, 3]], [[1, 1, 1], [2, 2, 2], [0, 0, 0], [0,\n 0, 0]]])\n', (21464, 21612), False, 'import torch\n'), ((21914, 21949), 'allennlp.nn.util.masked_flip', 'util.masked_flip', (['tensor', '[3, 4, 2]'], {}), '(tensor, [3, 4, 2])\n', (21930, 21949), False, 'from allennlp.nn import util\n'), ((21958, 21997), 'numpy.testing.assert_almost_equal', 'assert_almost_equal', (['response', 'solution'], {}), '(response, solution)\n', (21977, 21997), False, 'from numpy.testing import assert_array_almost_equal, assert_almost_equal\n'), ((23880, 23927), 'numpy.testing.assert_almost_equal', 'assert_almost_equal', (['actual_mask', 'expected_mask'], {}), '(actual_mask, expected_mask)\n', (23899, 23927), False, 'from numpy.testing import assert_array_almost_equal, assert_almost_equal\n'), ((24495, 24556), 'numpy.random.rand', 'numpy.random.rand', (['batch_size', 'sentence_length', 'embedding_dim'], {}), '(batch_size, sentence_length, embedding_dim)\n', (24512, 24556), False, 'import numpy\n'), ((24651, 24695), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.3, 0.4, 0.1, 0, 1.2]]'], {}), '([[0.3, 0.4, 0.1, 0, 1.2]])\n', (24668, 24695), False, 'import torch\n'), ((25106, 25191), 'numpy.testing.assert_almost_equal', 'numpy.testing.assert_almost_equal', (['aggregated_array', '[expected_array]'], {'decimal': '(5)'}), '(aggregated_array, [expected_array], decimal=5\n )\n', (25139, 25191), False, 'import numpy\n'), ((25385, 25459), 'numpy.random.rand', 'numpy.random.rand', (['batch_size', 'length_1', 'length_2', 'length_3', 'embedding_dim'], {}), '(batch_size, length_1, length_2, length_3, embedding_dim)\n', (25402, 25459), False, 'import numpy\n'), ((25486, 25545), 'numpy.random.rand', 'numpy.random.rand', (['batch_size', 'length_1', 'length_2', 'length_3'], {}), '(batch_size, length_1, length_2, length_3)\n', (25503, 25545), False, 'import numpy\n'), ((26049, 26140), 'numpy.testing.assert_almost_equal', 'numpy.testing.assert_almost_equal', (['aggregated_array[0, 3, 2]', 'expected_array'], {'decimal': '(5)'}), '(aggregated_array[0, 3, 2], expected_array,\n decimal=5)\n', (26082, 26140), False, 'import numpy\n'), ((26342, 26396), 'numpy.random.rand', 'numpy.random.rand', (['batch_size', 'length_3', 'embedding_dim'], {}), '(batch_size, length_3, embedding_dim)\n', (26359, 26396), False, 'import numpy\n'), ((26423, 26482), 'numpy.random.rand', 'numpy.random.rand', (['batch_size', 'length_1', 'length_2', 'length_3'], {}), '(batch_size, length_1, length_2, length_3)\n', (26440, 26482), False, 'import numpy\n'), ((27398, 27452), 'numpy.random.rand', 'numpy.random.rand', (['batch_size', 'length_2', 'embedding_dim'], {}), '(batch_size, length_2, embedding_dim)\n', (27415, 27452), False, 'import numpy\n'), ((27479, 27528), 'numpy.random.rand', 'numpy.random.rand', (['batch_size', 'length_1', 'length_2'], {}), '(batch_size, length_1, length_2)\n', (27496, 27528), False, 'import numpy\n'), ((28378, 28397), 'torch.zeros', 'torch.zeros', (['[9, 9]'], {}), '([9, 9])\n', (28389, 28397), False, 'import torch\n'), ((28419, 28479), 'allennlp.nn.util.viterbi_decode', 'util.viterbi_decode', (['sequence_logits.data', 'transition_matrix'], {}), '(sequence_logits.data, transition_matrix)\n', (28438, 28479), False, 'from allennlp.nn import util\n'), ((28508, 28537), 'torch.max', 'torch.max', (['sequence_logits', '(1)'], {}), '(sequence_logits, 1)\n', (28517, 28537), False, 'import torch\n'), ((28783, 28802), 'torch.zeros', 'torch.zeros', (['[9, 9]'], {}), '([9, 9])\n', (28794, 28802), False, 'import torch\n'), ((28839, 28855), 'torch.zeros', 'torch.zeros', (['[9]'], {}), '([9])\n', (28850, 28855), False, 'import torch\n'), ((28981, 28997), 'torch.zeros', 'torch.zeros', (['[9]'], {}), '([9])\n', (28992, 28997), False, 'import torch\n'), ((29105, 29275), 'allennlp.nn.util.viterbi_decode', 'util.viterbi_decode', (['sequence_logits.data', 'transition_matrix'], {'allowed_end_transitions': 'allowed_end_transitions', 'allowed_start_transitions': 'allowed_start_transitions'}), '(sequence_logits.data, transition_matrix,\n allowed_end_transitions=allowed_end_transitions,\n allowed_start_transitions=allowed_start_transitions)\n', (29124, 29275), False, 'from allennlp.nn import util\n'), ((29545, 29670), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0, 0, 0, 3, 5], [0, 0, 0, 3, 4], [0, 0, 0, 3, 4], [0, 0, 0, 3, 4], [0, 0,\n 0, 3, 4], [0, 0, 0, 3, 4]]'], {}), '([[0, 0, 0, 3, 5], [0, 0, 0, 3, 4], [0, 0, 0, 3, 4], [0, 0,\n 0, 3, 4], [0, 0, 0, 3, 4], [0, 0, 0, 3, 4]])\n', (29562, 29670), False, 'import torch\n'), ((29883, 29902), 'torch.zeros', 'torch.zeros', (['[5, 5]'], {}), '([5, 5])\n', (29894, 29902), False, 'import torch\n'), ((30003, 30058), 'allennlp.nn.util.viterbi_decode', 'util.viterbi_decode', (['sequence_logits', 'transition_matrix'], {}), '(sequence_logits, transition_matrix)\n', (30022, 30058), False, 'from allennlp.nn import util\n'), ((30246, 30371), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0, 0, 0, 4, 4], [0, 0, 0, 4, 4], [0, 0, 0, 4, 4], [0, 0, 0, 4, 4], [0, 0,\n 0, 4, 4], [0, 0, 0, 4, 4]]'], {}), '([[0, 0, 0, 4, 4], [0, 0, 0, 4, 4], [0, 0, 0, 4, 4], [0, 0,\n 0, 4, 4], [0, 0, 0, 4, 4], [0, 0, 0, 4, 4]])\n', (30263, 30371), False, 'import torch\n'), ((30705, 30724), 'torch.zeros', 'torch.zeros', (['[5, 5]'], {}), '([5, 5])\n', (30716, 30724), False, 'import torch\n'), ((30860, 30915), 'allennlp.nn.util.viterbi_decode', 'util.viterbi_decode', (['sequence_logits', 'transition_matrix'], {}), '(sequence_logits, transition_matrix)\n', (30879, 30915), False, 'from allennlp.nn import util\n'), ((30988, 31049), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1, 0, 0, 4], [1, 0, 6, 2], [0, 3, 0, 4]]'], {}), '([[1, 0, 0, 4], [1, 0, 6, 2], [0, 3, 0, 4]])\n', (31005, 31049), False, 'import torch\n'), ((31204, 31223), 'torch.zeros', 'torch.zeros', (['[4, 4]'], {}), '([4, 4])\n', (31215, 31223), False, 'import torch\n'), ((31321, 31376), 'allennlp.nn.util.viterbi_decode', 'util.viterbi_decode', (['sequence_logits', 'transition_matrix'], {}), '(sequence_logits, transition_matrix)\n', (31340, 31376), False, 'from allennlp.nn import util\n'), ((31558, 31683), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0, 0, 0, 7, 7], [0, 0, 0, 7, 7], [0, 0, 0, 7, 7], [0, 0, 0, 7, 7], [0, 0,\n 0, 7, 7], [0, 0, 0, 7, 7]]'], {}), '([[0, 0, 0, 7, 7], [0, 0, 0, 7, 7], [0, 0, 0, 7, 7], [0, 0,\n 0, 7, 7], [0, 0, 0, 7, 7], [0, 0, 0, 7, 7]])\n', (31575, 31683), False, 'import torch\n'), ((32019, 32038), 'torch.zeros', 'torch.zeros', (['[5, 5]'], {}), '([5, 5])\n', (32030, 32038), False, 'import torch\n'), ((32544, 32613), 'allennlp.nn.util.viterbi_decode', 'util.viterbi_decode', (['sequence_logits', 'transition_matrix', 'observations'], {}), '(sequence_logits, transition_matrix, observations)\n', (32563, 32613), False, 'from allennlp.nn import util\n'), ((32990, 33009), 'torch.zeros', 'torch.zeros', (['[9, 9]'], {}), '([9, 9])\n', (33001, 33009), False, 'import torch\n'), ((33032, 33101), 'allennlp.nn.util.viterbi_decode', 'util.viterbi_decode', (['sequence_logits.data', 'transition_matrix'], {'top_k': '(5)'}), '(sequence_logits.data, transition_matrix, top_k=5)\n', (33051, 33101), False, 'from allennlp.nn import util\n'), ((33131, 33160), 'torch.max', 'torch.max', (['sequence_logits', '(1)'], {}), '(sequence_logits, 1)\n', (33140, 33160), False, 'import torch\n'), ((33384, 33509), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0, 0, 0, 3, 4], [0, 0, 0, 3, 4], [0, 0, 0, 3, 4], [0, 0, 0, 3, 4], [0, 0,\n 0, 3, 4], [0, 0, 0, 3, 4]]'], {}), '([[0, 0, 0, 3, 4], [0, 0, 0, 3, 4], [0, 0, 0, 3, 4], [0, 0,\n 0, 3, 4], [0, 0, 0, 3, 4], [0, 0, 0, 3, 4]])\n', (33401, 33509), False, 'import torch\n'), ((33722, 33741), 'torch.zeros', 'torch.zeros', (['[5, 5]'], {}), '([5, 5])\n', (33733, 33741), False, 'import torch\n'), ((33842, 33906), 'allennlp.nn.util.viterbi_decode', 'util.viterbi_decode', (['sequence_logits', 'transition_matrix'], {'top_k': '(5)'}), '(sequence_logits, transition_matrix, top_k=5)\n', (33861, 33906), False, 'from allennlp.nn import util\n'), ((34097, 34222), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0, 0, 0, 4, 4], [0, 0, 0, 4, 4], [0, 0, 0, 4, 4], [0, 0, 0, 4, 4], [0, 0,\n 0, 4, 4], [0, 0, 0, 4, 0]]'], {}), '([[0, 0, 0, 4, 4], [0, 0, 0, 4, 4], [0, 0, 0, 4, 4], [0, 0,\n 0, 4, 4], [0, 0, 0, 4, 4], [0, 0, 0, 4, 0]])\n', (34114, 34222), False, 'import torch\n'), ((34556, 34575), 'torch.zeros', 'torch.zeros', (['[5, 5]'], {}), '([5, 5])\n', (34567, 34575), False, 'import torch\n'), ((34673, 34737), 'allennlp.nn.util.viterbi_decode', 'util.viterbi_decode', (['sequence_logits', 'transition_matrix'], {'top_k': '(5)'}), '(sequence_logits, transition_matrix, top_k=5)\n', (34692, 34737), False, 'from allennlp.nn import util\n'), ((34813, 34874), 'torch.FloatTensor', 'torch.FloatTensor', (['[[1, 0, 0, 4], [1, 0, 6, 2], [0, 3, 0, 4]]'], {}), '([[1, 0, 0, 4], [1, 0, 6, 2], [0, 3, 0, 4]])\n', (34830, 34874), False, 'import torch\n'), ((35029, 35048), 'torch.zeros', 'torch.zeros', (['[4, 4]'], {}), '([4, 4])\n', (35040, 35048), False, 'import torch\n'), ((35146, 35210), 'allennlp.nn.util.viterbi_decode', 'util.viterbi_decode', (['sequence_logits', 'transition_matrix'], {'top_k': '(5)'}), '(sequence_logits, transition_matrix, top_k=5)\n', (35165, 35210), False, 'from allennlp.nn import util\n'), ((37813, 37834), 'torch.rand', 'torch.rand', (['[5, 7, 4]'], {}), '([5, 7, 4])\n', (37823, 37834), False, 'import torch\n'), ((38282, 38347), 'allennlp.nn.util.sequence_cross_entropy_with_logits', 'util.sequence_cross_entropy_with_logits', (['tensor', 'targets', 'weights'], {}), '(tensor, targets, weights)\n', (38321, 38347), False, 'from allennlp.nn import util\n'), ((38364, 38430), 'allennlp.nn.util.sequence_cross_entropy_with_logits', 'util.sequence_cross_entropy_with_logits', (['tensor2', 'targets', 'weights'], {}), '(tensor2, targets, weights)\n', (38403, 38430), False, 'from allennlp.nn import util\n'), ((38584, 38605), 'torch.rand', 'torch.rand', (['[1, 3, 4]'], {}), '([1, 3, 4])\n', (38594, 38605), False, 'import torch\n'), ((38696, 38714), 'torch.ones', 'torch.ones', (['[2, 3]'], {}), '([2, 3])\n', (38706, 38714), False, 'import torch\n'), ((38730, 38820), 'allennlp.nn.util.sequence_cross_entropy_with_logits', 'util.sequence_cross_entropy_with_logits', (['tensor', 'targets', 'weights'], {'label_smoothing': '(0.1)'}), '(tensor, targets, weights,\n label_smoothing=0.1)\n', (38769, 38820), False, 'from allennlp.nn import util\n'), ((39573, 39594), 'torch.rand', 'torch.rand', (['[5, 7, 4]'], {}), '([5, 7, 4])\n', (39583, 39594), False, 'import torch\n'), ((39886, 39951), 'allennlp.nn.util.sequence_cross_entropy_with_logits', 'util.sequence_cross_entropy_with_logits', (['tensor', 'targets', 'weights'], {}), '(tensor, targets, weights)\n', (39925, 39951), False, 'from allennlp.nn import util\n'), ((39975, 40054), 'allennlp.nn.util.sequence_cross_entropy_with_logits', 'util.sequence_cross_entropy_with_logits', (['tensor', 'targets', 'weights'], {'average': 'None'}), '(tensor, targets, weights, average=None)\n', (40014, 40054), False, 'from allennlp.nn import util\n'), ((40485, 40506), 'torch.rand', 'torch.rand', (['[5, 7, 4]'], {}), '([5, 7, 4])\n', (40495, 40506), False, 'import torch\n'), ((40798, 40885), 'allennlp.nn.util.sequence_cross_entropy_with_logits', 'util.sequence_cross_entropy_with_logits', (['tensor', 'targets', 'weights'], {'average': '"""token"""'}), "(tensor, targets, weights, average=\n 'token')\n", (40837, 40885), False, 'from allennlp.nn import util\n'), ((40904, 40983), 'allennlp.nn.util.sequence_cross_entropy_with_logits', 'util.sequence_cross_entropy_with_logits', (['tensor', 'targets', 'weights'], {'average': 'None'}), '(tensor, targets, weights, average=None)\n', (40943, 40983), False, 'from allennlp.nn import util\n'), ((41454, 41490), 'torch.rand', 'torch.rand', (['[batch, length, classes]'], {}), '([batch, length, classes])\n', (41464, 41490), False, 'import torch\n'), ((41595, 41622), 'torch.ones', 'torch.ones', (['[batch, length]'], {}), '([batch, length])\n', (41605, 41622), False, 'import torch\n'), ((41639, 41717), 'allennlp.nn.util.sequence_cross_entropy_with_logits', 'util.sequence_cross_entropy_with_logits', (['tensor', 'targets', 'weights'], {'gamma': 'gamma'}), '(tensor, targets, weights, gamma=gamma)\n', (41678, 41717), False, 'from allennlp.nn import util\n'), ((42475, 42511), 'torch.rand', 'torch.rand', (['[batch, length, classes]'], {}), '([batch, length, classes])\n', (42485, 42511), False, 'import torch\n'), ((42616, 42643), 'torch.ones', 'torch.ones', (['[batch, length]'], {}), '([batch, length])\n', (42626, 42643), False, 'import torch\n'), ((42660, 42738), 'allennlp.nn.util.sequence_cross_entropy_with_logits', 'util.sequence_cross_entropy_with_logits', (['tensor', 'targets', 'weights'], {'alpha': 'alpha'}), '(tensor, targets, weights, alpha=alpha)\n', (42699, 42738), False, 'from allennlp.nn import util\n'), ((43574, 43593), 'torch.tensor', 'torch.tensor', (['alpha'], {}), '(alpha)\n', (43586, 43593), False, 'import torch\n'), ((43612, 43648), 'torch.rand', 'torch.rand', (['[batch, length, classes]'], {}), '([batch, length, classes])\n', (43622, 43648), False, 'import torch\n'), ((43753, 43780), 'torch.ones', 'torch.ones', (['[batch, length]'], {}), '([batch, length])\n', (43763, 43780), False, 'import torch\n'), ((43797, 43875), 'allennlp.nn.util.sequence_cross_entropy_with_logits', 'util.sequence_cross_entropy_with_logits', (['tensor', 'targets', 'weights'], {'alpha': 'alpha'}), '(tensor, targets, weights, alpha=alpha)\n', (43836, 43875), False, 'from allennlp.nn import util\n'), ((44635, 44671), 'torch.rand', 'torch.rand', (['[batch, length, classes]'], {}), '([batch, length, classes])\n', (44645, 44671), False, 'import torch\n'), ((44776, 44803), 'torch.ones', 'torch.ones', (['[batch, length]'], {}), '([batch, length])\n', (44786, 44803), False, 'import torch\n'), ((44820, 44898), 'allennlp.nn.util.sequence_cross_entropy_with_logits', 'util.sequence_cross_entropy_with_logits', (['tensor', 'targets', 'weights'], {'alpha': 'alpha'}), '(tensor, targets, weights, alpha=alpha)\n', (44859, 44898), False, 'from allennlp.nn import util\n'), ((45440, 45506), 'torch.FloatTensor', 'torch.FloatTensor', (['[[[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12]]]'], {}), '([[[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12]]])\n', (45457, 45506), False, 'import torch\n'), ((45522, 45557), 'torch.tensor', 'torch.tensor', (['[[True, True, False]]'], {}), '([[True, True, False]])\n', (45534, 45557), False, 'import torch\n'), ((45656, 45731), 'numpy.testing.assert_almost_equal', 'assert_almost_equal', (['replaced', '[[[1, 2, 3, 4], [5, 6, 7, 8], [2, 2, 2, 2]]]'], {}), '(replaced, [[[1, 2, 3, 4], [5, 6, 7, 8], [2, 2, 2, 2]]])\n', (45675, 45731), False, 'from numpy.testing import assert_array_almost_equal, assert_almost_equal\n'), ((45854, 45890), 'torch.FloatTensor', 'torch.FloatTensor', (['[[0.4, 0.1, 0.2]]'], {}), '([[0.4, 0.1, 0.2]])\n', (45871, 45890), False, 'import torch\n'), ((45946, 45995), 'allennlp.nn.util.logsumexp', 'util.logsumexp', (['log_tensor'], {'dim': '(-1)', 'keepdim': '(False)'}), '(log_tensor, dim=-1, keepdim=False)\n', (45960, 45995), False, 'from allennlp.nn import util\n'), ((46083, 46131), 'allennlp.nn.util.logsumexp', 'util.logsumexp', (['log_tensor'], {'dim': '(-1)', 'keepdim': '(True)'}), '(log_tensor, dim=-1, keepdim=True)\n', (46097, 46131), False, 'from allennlp.nn import util\n'), ((46466, 46501), 'torch.FloatTensor', 'torch.FloatTensor', (['[[-200.0, 20.0]]'], {}), '([[-200.0, 20.0]])\n', (46483, 46501), False, 'import torch\n'), ((46592, 46642), 'torch.FloatTensor', 'torch.FloatTensor', (['[[20.0, 20.0], [-200.0, 200.0]]'], {}), '([[20.0, 20.0], [-200.0, 200.0]])\n', (46609, 46642), False, 'import torch\n'), ((46801, 46906), 'numpy.array', 'numpy.array', (['[[[1, 2, 3, 4], [5, 6, 7, 8], [9, 9, 9, 9]], [[2, 1, 0, 7], [7, 7, 2, 3], [\n 0, 0, 4, 2]]]'], {}), '([[[1, 2, 3, 4], [5, 6, 7, 8], [9, 9, 9, 9]], [[2, 1, 0, 7], [7,\n 7, 2, 3], [0, 0, 4, 2]]])\n', (46812, 46906), False, 'import numpy\n'), ((46943, 46982), 'torch.tensor', 'torch.tensor', (['indices'], {'dtype': 'torch.long'}), '(indices, dtype=torch.long)\n', (46955, 46982), False, 'import torch\n'), ((47009, 47058), 'allennlp.nn.util.flatten_and_batch_shift_indices', 'util.flatten_and_batch_shift_indices', (['indices', '(10)'], {}), '(indices, 10)\n', (47045, 47058), False, 'from allennlp.nn import util\n'), ((47354, 47403), 'numpy.array', 'numpy.array', (['[[[1, 2], [3, 4]], [[5, 6], [7, 8]]]'], {}), '([[[1, 2], [3, 4]], [[5, 6], [7, 8]]])\n', (47365, 47403), False, 'import numpy\n'), ((47627, 47666), 'torch.tensor', 'torch.tensor', (['indices'], {'dtype': 'torch.long'}), '(indices, dtype=torch.long)\n', (47639, 47666), False, 'import torch\n'), ((47686, 47729), 'allennlp.nn.util.batched_index_select', 'util.batched_index_select', (['targets', 'indices'], {}), '(targets, indices)\n', (47711, 47729), False, 'from allennlp.nn import util\n'), ((47799, 47814), 'numpy.ones', 'numpy.ones', (['[3]'], {}), '([3])\n', (47809, 47814), False, 'import numpy\n'), ((48523, 48573), 'numpy.array', 'numpy.array', (['[[[1, 11], [3, 4]], [[5, 6], [7, 8]]]'], {}), '([[[1, 11], [3, 4]], [[5, 6], [7, 8]]])\n', (48534, 48573), False, 'import numpy\n'), ((48592, 48631), 'torch.tensor', 'torch.tensor', (['indices'], {'dtype': 'torch.long'}), '(indices, dtype=torch.long)\n', (48604, 48631), False, 'import torch\n'), ((48755, 48805), 'numpy.array', 'numpy.array', (['[[[1, -1], [3, 4]], [[5, 6], [7, 8]]]'], {}), '([[[1, -1], [3, 4]], [[5, 6], [7, 8]]])\n', (48766, 48805), False, 'import numpy\n'), ((48824, 48863), 'torch.tensor', 'torch.tensor', (['indices'], {'dtype': 'torch.long'}), '(indices, dtype=torch.long)\n', (48836, 48863), False, 'import torch\n'), ((49366, 49406), 'allennlp.nn.util.batched_span_select', 'util.batched_span_select', (['targets', 'spans'], {}), '(targets, spans)\n', (49390, 49406), False, 'from allennlp.nn import util\n'), ((50289, 50318), 'numpy.array', 'numpy.array', (['[[1, 2], [3, 4]]'], {}), '([[1, 2], [3, 4]])\n', (50300, 50318), False, 'import numpy\n'), ((50492, 50531), 'torch.tensor', 'torch.tensor', (['indices'], {'dtype': 'torch.long'}), '(indices, dtype=torch.long)\n', (50504, 50531), False, 'import torch\n'), ((50552, 50597), 'allennlp.nn.util.flattened_index_select', 'util.flattened_index_select', (['targets', 'indices'], {}), '(targets, indices)\n', (50579, 50597), False, 'from allennlp.nn import util\n'), ((50668, 50683), 'numpy.ones', 'numpy.ones', (['[3]'], {}), '([3])\n', (50678, 50683), False, 'import numpy\n'), ((51586, 51625), 'torch.LongTensor', 'torch.LongTensor', (['[1, 2, 7, 1, 56, 900]'], {}), '([1, 2, 7, 1, 56, 900])\n', (51602, 51625), False, 'import torch\n'), ((51655, 51682), 'allennlp.nn.util.bucket_values', 'util.bucket_values', (['indices'], {}), '(indices)\n', (51673, 51682), False, 'from allennlp.nn import util\n'), ((52038, 52098), 'allennlp.nn.util.add_sentence_boundary_token_ids', 'util.add_sentence_boundary_token_ids', (['tensor', 'mask', 'bos', 'eos'], {}), '(tensor, mask, bos, eos)\n', (52074, 52098), False, 'from allennlp.nn import util\n'), ((52129, 52178), 'numpy.array', 'numpy.array', (['[[9, 1, 2, 3, 10], [9, 4, 5, 10, 0]]'], {}), '([[9, 1, 2, 3, 10], [9, 4, 5, 10, 0]])\n', (52140, 52178), False, 'import numpy\n'), ((52836, 52896), 'allennlp.nn.util.add_sentence_boundary_token_ids', 'util.add_sentence_boundary_token_ids', (['tensor', 'mask', 'bos', 'eos'], {}), '(tensor, mask, bos, eos)\n', (52872, 52896), False, 'from allennlp.nn import util\n'), ((52927, 53102), 'numpy.array', 'numpy.array', (['[[[9, 9, 9, 9], [1, 2, 3, 4], [5, 5, 5, 5], [6, 8, 1, 2], [10, 10, 10, 10]],\n [[9, 9, 9, 9], [4, 3, 2, 1], [8, 7, 6, 5], [10, 10, 10, 10], [0, 0, 0, 0]]]'], {}), '([[[9, 9, 9, 9], [1, 2, 3, 4], [5, 5, 5, 5], [6, 8, 1, 2], [10, \n 10, 10, 10]], [[9, 9, 9, 9], [4, 3, 2, 1], [8, 7, 6, 5], [10, 10, 10, \n 10], [0, 0, 0, 0]]])\n', (52938, 53102), False, 'import numpy\n'), ((53763, 53808), 'allennlp.nn.util.remove_sentence_boundaries', 'util.remove_sentence_boundaries', (['tensor', 'mask'], {}), '(tensor, mask)\n', (53794, 53808), False, 'from allennlp.nn import util\n'), ((53840, 53860), 'torch.zeros', 'torch.zeros', (['(3)', '(3)', '(7)'], {}), '(3, 3, 7)\n', (53851, 53860), False, 'import torch\n'), ((54565, 54705), 'numpy.asarray', 'numpy.asarray', (['[[0.0, 0.0, 1.0, 1.0], [0.841470957, 9.99999902e-05, 0.540302277, 1.0], [\n 0.909297407, 0.00019999998, -0.416146845, 1.0]]'], {}), '([[0.0, 0.0, 1.0, 1.0], [0.841470957, 9.99999902e-05, \n 0.540302277, 1.0], [0.909297407, 0.00019999998, -0.416146845, 1.0]])\n', (54578, 54705), False, 'import numpy\n'), ((54877, 54899), 'torch.zeros', 'torch.zeros', (['[2, 3, 4]'], {}), '([2, 3, 4])\n', (54888, 54899), False, 'import torch\n'), ((54917, 54995), 'allennlp.nn.util.add_positional_features', 'util.add_positional_features', (['tensor'], {'min_timescale': '(1.0)', 'max_timescale': '(10000.0)'}), '(tensor, min_timescale=1.0, max_timescale=10000.0)\n', (54945, 54995), False, 'from allennlp.nn import util\n'), ((55274, 55504), 'numpy.asarray', 'numpy.asarray', (['[[0.0, 0.0, 0.0, 1.0, 1.0, 1.0, 0.0], [0.841470957, 0.00999983307, \n 9.99999902e-05, 0.540302277, 0.999949992, 1.0, 0.0], [0.909297407, \n 0.0199986659, 0.00019999998, -0.416146815, 0.999800026, 1.0, 0.0]]'], {}), '([[0.0, 0.0, 0.0, 1.0, 1.0, 1.0, 0.0], [0.841470957, \n 0.00999983307, 9.99999902e-05, 0.540302277, 0.999949992, 1.0, 0.0], [\n 0.909297407, 0.0199986659, 0.00019999998, -0.416146815, 0.999800026, \n 1.0, 0.0]])\n', (55287, 55504), False, 'import numpy\n'), ((56202, 56224), 'torch.zeros', 'torch.zeros', (['[2, 3, 7]'], {}), '([2, 3, 7])\n', (56213, 56224), False, 'import torch\n'), ((56242, 56320), 'allennlp.nn.util.add_positional_features', 'util.add_positional_features', (['tensor'], {'min_timescale': '(1.0)', 'max_timescale': '(10000.0)'}), '(tensor, min_timescale=1.0, max_timescale=10000.0)\n', (56270, 56320), False, 'from allennlp.nn import util\n'), ((56653, 56673), 'torch.Tensor', 'torch.Tensor', (['[4, 5]'], {}), '([4, 5])\n', (56665, 56673), False, 'import torch\n'), ((57030, 57056), 'torch.Tensor', 'torch.Tensor', (['[4, 5, 4, 5]'], {}), '([4, 5, 4, 5])\n', (57042, 57056), False, 'import torch\n'), ((58967, 58987), 'torch.Tensor', 'torch.Tensor', (['[4, 5]'], {}), '([4, 5])\n', (58979, 58987), False, 'import torch\n'), ((59350, 59370), 'torch.Tensor', 'torch.Tensor', (['[4, 5]'], {}), '([4, 5])\n', (59362, 59370), False, 'import torch\n'), ((59859, 59899), 'torch.randn', 'torch.randn', (['(1)', 'seq_len_1', 'embedding_dim'], {}), '(1, seq_len_1, embedding_dim)\n', (59870, 59899), False, 'import torch\n'), ((59913, 59953), 'torch.randn', 'torch.randn', (['(1)', 'seq_len_2', 'embedding_dim'], {}), '(1, seq_len_2, embedding_dim)\n', (59924, 59953), False, 'import torch\n'), ((59977, 60043), 'allennlp.nn.util.get_combined_dim', 'util.get_combined_dim', (['combination', '[embedding_dim, embedding_dim]'], {}), '(combination, [embedding_dim, embedding_dim])\n', (59998, 60043), False, 'from allennlp.nn import util\n'), ((60061, 60087), 'torch.Tensor', 'torch.Tensor', (['combined_dim'], {}), '(combined_dim)\n', (60073, 60087), False, 'import torch\n'), ((60493, 60533), 'torch.randn', 'torch.randn', (['(1)', 'seq_len_1', 'embedding_dim'], {}), '(1, seq_len_1, embedding_dim)\n', (60504, 60533), False, 'import torch\n'), ((60547, 60587), 'torch.randn', 'torch.randn', (['(1)', 'seq_len_2', 'embedding_dim'], {}), '(1, seq_len_2, embedding_dim)\n', (60558, 60587), False, 'import torch\n'), ((60611, 60677), 'allennlp.nn.util.get_combined_dim', 'util.get_combined_dim', (['combination', '[embedding_dim, embedding_dim]'], {}), '(combination, [embedding_dim, embedding_dim])\n', (60632, 60677), False, 'from allennlp.nn import util\n'), ((60695, 60721), 'torch.Tensor', 'torch.Tensor', (['combined_dim'], {}), '(combined_dim)\n', (60707, 60721), False, 'import torch\n'), ((61011, 61034), 'torch.tensor', 'torch.tensor', (['[1, 2, 3]'], {}), '([1, 2, 3])\n', (61023, 61034), False, 'import torch\n'), ((61524, 61553), 'torch.randn', 'torch.randn', (['(4)', '(10)', '(20)', '(17)', '(5)'], {}), '(4, 10, 20, 17, 5)\n', (61535, 61553), False, 'import torch\n'), ((61574, 61607), 'allennlp.nn.util.combine_initial_dims', 'util.combine_initial_dims', (['tensor'], {}), '(tensor)\n', (61599, 61607), False, 'from allennlp.nn import util\n'), ((61736, 61773), 'torch.randn', 'torch.randn', (['(4 * 10 * 20 * 17 * 5)', '(12)'], {}), '(4 * 10 * 20 * 17 * 5, 12)\n', (61747, 61773), False, 'import torch\n'), ((63132, 63179), 'allennlp.nn.util.move_to_device', 'util.move_to_device', (['structured_obj', 'new_device'], {}), '(structured_obj, new_device)\n', (63151, 63179), False, 'from allennlp.nn import util\n'), ((63533, 63555), 'torch.nn.Linear', 'torch.nn.Linear', (['(10)', '(5)'], {}), '(10, 5)\n', (63548, 63555), False, 'import torch\n'), ((63685, 63722), 'allennlp.nn.util.extend_layer', 'util.extend_layer', (['lin_layer', 'new_dim'], {}), '(lin_layer, new_dim)\n', (63702, 63722), False, 'from allennlp.nn import util\n'), ((64401, 64434), 'allennlp.nn.util.masked_topk', 'util.masked_topk', (['scores', 'mask', '(2)'], {}), '(scores, mask, 2)\n', (64417, 64434), False, 'from allennlp.nn import util\n'), ((65486, 65519), 'allennlp.nn.util.masked_topk', 'util.masked_topk', (['scores', 'mask', '(2)'], {}), '(scores, mask, 2)\n', (65502, 65519), False, 'from allennlp.nn import util\n'), ((66665, 66706), 'torch.tensor', 'torch.tensor', (['[3, 2, 1]'], {'dtype': 'torch.long'}), '([3, 2, 1], dtype=torch.long)\n', (66677, 66706), False, 'import torch\n'), ((66761, 66794), 'allennlp.nn.util.masked_topk', 'util.masked_topk', (['scores', 'mask', 'k'], {}), '(scores, mask, k)\n', (66777, 66794), False, 'from allennlp.nn import util\n'), ((67977, 68018), 'torch.tensor', 'torch.tensor', (['[3, 2, 0]'], {'dtype': 'torch.long'}), '([3, 2, 0], dtype=torch.long)\n', (67989, 68018), False, 'import torch\n'), ((68073, 68106), 'allennlp.nn.util.masked_topk', 'util.masked_topk', (['scores', 'mask', 'k'], {}), '(scores, mask, k)\n', (68089, 68106), False, 'from allennlp.nn import util\n'), ((69675, 69712), 'torch.ones', 'torch.ones', (['(3)', '(5)', '(4)'], {'dtype': 'torch.long'}), '(3, 5, 4, dtype=torch.long)\n', (69685, 69712), False, 'import torch\n'), ((69996, 70036), 'torch.ones', 'torch.ones', (['(3)', '(2)', '(5)', '(4)'], {'dtype': 'torch.bool'}), '(3, 2, 5, 4, dtype=torch.bool)\n', (70006, 70036), False, 'import torch\n'), ((70413, 70452), 'allennlp.nn.util.masked_topk', 'util.masked_topk', (['items', 'mask', 'k'], {'dim': '(1)'}), '(items, mask, k, dim=1)\n', (70429, 70452), False, 'from allennlp.nn import util\n'), ((949, 974), 'numpy.array', 'numpy.array', (['[3, 2, 6, 1]'], {}), '([3, 2, 6, 1])\n', (960, 974), False, 'import numpy\n'), ((1924, 1947), 'numpy.array', 'numpy.array', (['[260, 260]'], {}), '([260, 260])\n', (1935, 1947), False, 'import numpy\n'), ((2179, 2197), 'torch.Size', 'torch.Size', (['[2, 3]'], {}), '([2, 3])\n', (2189, 2197), False, 'import torch\n'), ((2539, 2557), 'torch.Size', 'torch.Size', (['[2, 3]'], {}), '([2, 3])\n', (2549, 2557), False, 'import torch\n'), ((3773, 3806), 'torch.LongTensor', 'torch.LongTensor', (['[7, 5, 4, 3, 1]'], {}), '([7, 5, 4, 3, 1])\n', (3789, 3806), False, 'import torch\n'), ((4971, 5016), 'numpy.array', 'numpy.array', (['[[0.090031, 0.244728, 0.665241]]'], {}), '([[0.090031, 0.244728, 0.665241]])\n', (4982, 5016), False, 'import numpy\n'), ((5060, 5090), 'numpy.sum', 'numpy.sum', (['vector_1d_softmaxed'], {}), '(vector_1d_softmaxed)\n', (5069, 5090), False, 'import numpy\n'), ((5296, 5340), 'numpy.array', 'numpy.array', (['[[0.017148, 0.046613, 0.93624]]'], {}), '([[0.017148, 0.046613, 0.93624]])\n', (5307, 5340), False, 'import numpy\n'), ((5622, 5673), 'numpy.array', 'numpy.array', (['[[0.33333334, 0.33333334, 0.33333334]]'], {}), '([[0.33333334, 0.33333334, 0.33333334]])\n', (5633, 5673), False, 'import numpy\n'), ((5974, 6067), 'numpy.array', 'numpy.array', (['[[0.01714783, 0.04661262, 0.93623955], [0.09003057, 0.24472847, 0.66524096]]'], {}), '([[0.01714783, 0.04661262, 0.93623955], [0.09003057, 0.24472847,\n 0.66524096]])\n', (5985, 6067), False, 'import numpy\n'), ((6422, 6515), 'numpy.array', 'numpy.array', (['[[0.01714783, 0.04661262, 0.93623955], [0.33333334, 0.33333334, 0.33333334]]'], {}), '([[0.01714783, 0.04661262, 0.93623955], [0.33333334, 0.33333334,\n 0.33333334]])\n', (6433, 6515), False, 'import numpy\n'), ((6891, 6935), 'numpy.array', 'numpy.array', (['[[0.01798621, 0.0, 0.98201382]]'], {}), '([[0.01798621, 0.0, 0.98201382]])\n', (6902, 6935), False, 'import numpy\n'), ((7211, 7267), 'numpy.array', 'numpy.array', (['[[0.01321289, 0.0, 0.26538793, 0.72139918]]'], {}), '([[0.01321289, 0.0, 0.26538793, 0.72139918]])\n', (7222, 7267), False, 'import numpy\n'), ((7642, 7669), 'numpy.array', 'numpy.array', (['[[0, 0, 0, 1]]'], {}), '([[0, 0, 0, 1]])\n', (7653, 7669), False, 'import numpy\n'), ((8036, 8071), 'numpy.array', 'numpy.array', (['[[0.0, 0.0, 0.0, 0.0]]'], {}), '([[0.0, 0.0, 0.0, 0.0]])\n', (8047, 8071), False, 'import numpy\n'), ((8434, 8469), 'numpy.array', 'numpy.array', (['[[0.0, 0.0, 0.0, 0.0]]'], {}), '([[0.0, 0.0, 0.0, 0.0]])\n', (8445, 8469), False, 'import numpy\n'), ((8815, 8843), 'numpy.array', 'numpy.array', (['[[0.5, 0.5, 0]]'], {}), '([[0.5, 0.5, 0]])\n', (8826, 8843), False, 'import numpy\n'), ((9204, 9280), 'numpy.array', 'numpy.array', (['[[0.01798621, 0.0, 0.98201382], [0.090031, 0.244728, 0.665241]]'], {}), '([[0.01798621, 0.0, 0.98201382], [0.090031, 0.244728, 0.665241]])\n', (9215, 9280), False, 'import numpy\n'), ((9705, 9767), 'numpy.array', 'numpy.array', (['[[0.5, 0.0, 0.5], [0.090031, 0.244728, 0.665241]]'], {}), '([[0.5, 0.0, 0.5], [0.090031, 0.244728, 0.665241]])\n', (9716, 9767), False, 'import numpy\n'), ((10193, 10240), 'numpy.array', 'numpy.array', (['[[0.5, 0.0, 0.5], [0.0, 0.0, 0.0]]'], {}), '([[0.5, 0.0, 0.5], [0.0, 0.0, 0.0]])\n', (10204, 10240), False, 'import numpy\n'), ((10550, 10611), 'numpy.array', 'numpy.array', (['[[0.0, 0.0, 0.0], [0.11920292, 0.0, 0.88079708]]'], {}), '([[0.0, 0.0, 0.0], [0.11920292, 0.0, 0.88079708]])\n', (10561, 10611), False, 'import numpy\n'), ((11022, 11066), 'numpy.array', 'numpy.array', (['[[0.01798621, 0.0, 0.98201382]]'], {}), '([[0.01798621, 0.0, 0.98201382]])\n', (11033, 11066), False, 'import numpy\n'), ((11387, 11443), 'numpy.array', 'numpy.array', (['[[0.01321289, 0.0, 0.26538793, 0.72139918]]'], {}), '([[0.01321289, 0.0, 0.26538793, 0.72139918]])\n', (11398, 11443), False, 'import numpy\n'), ((11863, 11890), 'numpy.array', 'numpy.array', (['[[0, 0, 0, 1]]'], {}), '([[0, 0, 0, 1]])\n', (11874, 11890), False, 'import numpy\n'), ((12302, 12341), 'numpy.array', 'numpy.array', (['[[0.25, 0.25, 0.25, 0.25]]'], {}), '([[0.25, 0.25, 0.25, 0.25]])\n', (12313, 12341), False, 'import numpy\n'), ((12749, 12788), 'numpy.array', 'numpy.array', (['[[0.25, 0.25, 0.25, 0.25]]'], {}), '([[0.25, 0.25, 0.25, 0.25]])\n', (12760, 12788), False, 'import numpy\n'), ((13179, 13207), 'numpy.array', 'numpy.array', (['[[0.5, 0.5, 0]]'], {}), '([[0.5, 0.5, 0]])\n', (13190, 13207), False, 'import numpy\n'), ((13613, 13689), 'numpy.array', 'numpy.array', (['[[0.01798621, 0.0, 0.98201382], [0.090031, 0.244728, 0.665241]]'], {}), '([[0.01798621, 0.0, 0.98201382], [0.090031, 0.244728, 0.665241]])\n', (13624, 13689), False, 'import numpy\n'), ((14159, 14221), 'numpy.array', 'numpy.array', (['[[0.5, 0.0, 0.5], [0.090031, 0.244728, 0.665241]]'], {}), '([[0.5, 0.0, 0.5], [0.090031, 0.244728, 0.665241]])\n', (14170, 14221), False, 'import numpy\n'), ((14704, 14772), 'numpy.array', 'numpy.array', (['[[0.5, 0.0, 0.5], [0.33333333, 0.33333333, 0.33333333]]'], {}), '([[0.5, 0.0, 0.5], [0.33333333, 0.33333333, 0.33333333]])\n', (14715, 14772), False, 'import numpy\n'), ((15140, 15227), 'numpy.array', 'numpy.array', (['[[0.33333333, 0.33333333, 0.33333333], [0.11920292, 0.0, 0.88079708]]'], {}), '([[0.33333333, 0.33333333, 0.33333333], [0.11920292, 0.0, \n 0.88079708]])\n', (15151, 15227), False, 'import numpy\n'), ((15745, 15775), 'numpy.exp', 'numpy.exp', (['vector_1d_softmaxed'], {}), '(vector_1d_softmaxed)\n', (15754, 15775), False, 'import numpy\n'), ((15777, 15821), 'numpy.array', 'numpy.array', (['[[0.01798621, 0.0, 0.98201382]]'], {}), '([[0.01798621, 0.0, 0.98201382]])\n', (15788, 15821), False, 'import numpy\n'), ((16089, 16119), 'numpy.exp', 'numpy.exp', (['vector_1d_softmaxed'], {}), '(vector_1d_softmaxed)\n', (16098, 16119), False, 'import numpy\n'), ((16121, 16177), 'numpy.array', 'numpy.array', (['[[0.01321289, 0.0, 0.26538793, 0.72139918]]'], {}), '([[0.01321289, 0.0, 0.26538793, 0.72139918]])\n', (16132, 16177), False, 'import numpy\n'), ((16548, 16578), 'numpy.exp', 'numpy.exp', (['vector_1d_softmaxed'], {}), '(vector_1d_softmaxed)\n', (16557, 16578), False, 'import numpy\n'), ((16580, 16615), 'numpy.array', 'numpy.array', (['[[0.0, 0.0, 0.0, 1.0]]'], {}), '([[0.0, 0.0, 0.0, 1.0]])\n', (16591, 16615), False, 'import numpy\n'), ((18043, 18067), 'numpy.array', 'numpy.array', (['[5.0, -1.0]'], {}), '([5.0, -1.0])\n', (18054, 18067), False, 'import numpy\n'), ((18410, 18438), 'numpy.array', 'numpy.array', (['[[5.0], [-1.0]]'], {}), '([[5.0], [-1.0]])\n', (18421, 18438), False, 'import numpy\n'), ((18817, 18856), 'numpy.array', 'numpy.array', (['[[5.0, 2.0], [-1.0, -0.5]]'], {}), '([[5.0, 2.0], [-1.0, -0.5]])\n', (18828, 18856), False, 'import numpy\n'), ((19839, 19863), 'numpy.array', 'numpy.array', (['[3.0, -1.5]'], {}), '([3.0, -1.5])\n', (19850, 19863), False, 'import numpy\n'), ((20205, 20233), 'numpy.array', 'numpy.array', (['[[3.0], [-1.5]]'], {}), '([[3.0], [-1.5]])\n', (20216, 20233), False, 'import numpy\n'), ((20611, 20651), 'numpy.array', 'numpy.array', (['[[3.0, 0.5], [-1.5, -1.75]]'], {}), '([[3.0, 0.5], [-1.5, -1.75]])\n', (20622, 20651), False, 'import numpy\n'), ((28058, 28146), 'numpy.testing.assert_almost_equal', 'numpy.testing.assert_almost_equal', (['aggregated_array[0, i]', 'expected_array'], {'decimal': '(5)'}), '(aggregated_array[0, i], expected_array,\n decimal=5)\n', (28091, 28146), False, 'import numpy\n'), ((28322, 28340), 'torch.rand', 'torch.rand', (['[5, 9]'], {}), '([5, 9])\n', (28332, 28340), False, 'import torch\n'), ((28727, 28745), 'torch.rand', 'torch.rand', (['[5, 9]'], {}), '([5, 9])\n', (28737, 28745), False, 'import torch\n'), ((32942, 32960), 'torch.rand', 'torch.rand', (['[5, 9]'], {}), '([5, 9])\n', (32952, 32960), False, 'import torch\n'), ((36791, 36811), 'random.randint', 'random.randint', (['(1)', '(5)'], {}), '(1, 5)\n', (36805, 36811), False, 'import random\n'), ((36834, 36854), 'random.randint', 'random.randint', (['(1)', '(5)'], {}), '(1, 5)\n', (36848, 36854), False, 'import random\n'), ((36871, 36891), 'random.randint', 'random.randint', (['(1)', '(5)'], {}), '(1, 5)\n', (36885, 36891), False, 'import random\n'), ((36922, 36953), 'torch.rand', 'torch.rand', (['[seq_len, num_tags]'], {}), '([seq_len, num_tags])\n', (36932, 36953), False, 'import torch\n'), ((36986, 37018), 'torch.rand', 'torch.rand', (['[num_tags, num_tags]'], {}), '([num_tags, num_tags])\n', (36996, 37018), False, 'import torch\n'), ((37069, 37133), 'allennlp.nn.util.viterbi_decode', 'util.viterbi_decode', (['sequence_logits', 'transition_matrix'], {'top_k': 'k'}), '(sequence_logits, transition_matrix, top_k=k)\n', (37088, 37133), False, 'from allennlp.nn import util\n'), ((38203, 38237), 'numpy.random.randint', 'numpy.random.randint', (['(0)', '(3)', '[5, 7]'], {}), '(0, 3, [5, 7])\n', (38223, 38237), False, 'import numpy\n'), ((38641, 38675), 'numpy.random.randint', 'numpy.random.randint', (['(0)', '(3)', '[1, 3]'], {}), '(0, 3, [1, 3])\n', (38661, 38675), False, 'import numpy\n'), ((38969, 39020), 'torch.nn.functional.log_softmax', 'torch.nn.functional.log_softmax', (['prediction'], {'dim': '(-1)'}), '(prediction, dim=-1)\n', (39000, 39020), False, 'import torch\n'), ((39807, 39841), 'numpy.random.randint', 'numpy.random.randint', (['(0)', '(3)', '[5, 7]'], {}), '(0, 3, [5, 7])\n', (39827, 39841), False, 'import numpy\n'), ((40719, 40753), 'numpy.random.randint', 'numpy.random.randint', (['(0)', '(3)', '[5, 7]'], {}), '(0, 3, [5, 7])\n', (40739, 40753), False, 'import numpy\n'), ((41401, 41421), 'numpy.random.randn', 'numpy.random.randn', ([], {}), '()\n', (41419, 41421), False, 'import numpy\n'), ((41526, 41575), 'numpy.random.randint', 'numpy.random.randint', (['(0)', 'classes', '[batch, length]'], {}), '(0, classes, [batch, length])\n', (41546, 41575), False, 'import numpy\n'), ((41834, 41876), 'torch.nn.functional.softmax', 'torch.nn.functional.softmax', (['logit'], {'dim': '(-1)'}), '(logit, dim=-1)\n', (41861, 41876), False, 'import torch\n'), ((42355, 42374), 'numpy.random.rand', 'numpy.random.rand', ([], {}), '()\n', (42372, 42374), False, 'import numpy\n'), ((42547, 42596), 'numpy.random.randint', 'numpy.random.randint', (['(0)', 'classes', '[batch, length]'], {}), '(0, classes, [batch, length])\n', (42567, 42596), False, 'import numpy\n'), ((42858, 42904), 'torch.nn.functional.log_softmax', 'torch.nn.functional.log_softmax', (['logit'], {'dim': '(-1)'}), '(logit, dim=-1)\n', (42889, 42904), False, 'import torch\n'), ((43456, 43475), 'numpy.random.rand', 'numpy.random.rand', ([], {}), '()\n', (43473, 43475), False, 'import numpy\n'), ((43684, 43733), 'numpy.random.randint', 'numpy.random.randint', (['(0)', 'classes', '[batch, length]'], {}), '(0, classes, [batch, length])\n', (43704, 43733), False, 'import numpy\n'), ((43995, 44041), 'torch.nn.functional.log_softmax', 'torch.nn.functional.log_softmax', (['logit'], {'dim': '(-1)'}), '(logit, dim=-1)\n', (44026, 44041), False, 'import torch\n'), ((44575, 44602), 'numpy.random.randn', 'numpy.random.randn', (['classes'], {}), '(classes)\n', (44593, 44602), False, 'import numpy\n'), ((44707, 44756), 'numpy.random.randint', 'numpy.random.randint', (['(0)', 'classes', '[batch, length]'], {}), '(0, classes, [batch, length])\n', (44727, 44756), False, 'import numpy\n'), ((45018, 45064), 'torch.nn.functional.log_softmax', 'torch.nn.functional.log_softmax', (['logit'], {'dim': '(-1)'}), '(logit, dim=-1)\n', (45049, 45064), False, 'import torch\n'), ((47155, 47256), 'numpy.array', 'numpy.array', (['[1, 2, 3, 4, 5, 6, 7, 8, 9, 9, 9, 9, 12, 11, 10, 17, 17, 17, 12, 13, 10, 10,\n 14, 12]'], {}), '([1, 2, 3, 4, 5, 6, 7, 8, 9, 9, 9, 9, 12, 11, 10, 17, 17, 17, 12,\n 13, 10, 10, 14, 12])\n', (47166, 47256), False, 'import numpy\n'), ((48645, 48678), 'pytest.raises', 'pytest.raises', (['ConfigurationError'], {}), '(ConfigurationError)\n', (48658, 48678), False, 'import pytest\n'), ((48692, 48735), 'allennlp.nn.util.batched_index_select', 'util.batched_index_select', (['targets', 'indices'], {}), '(targets, indices)\n', (48717, 48735), False, 'from allennlp.nn import util\n'), ((48877, 48910), 'pytest.raises', 'pytest.raises', (['ConfigurationError'], {}), '(ConfigurationError)\n', (48890, 48910), False, 'import pytest\n'), ((48924, 48967), 'allennlp.nn.util.batched_index_select', 'util.batched_index_select', (['targets', 'indices'], {}), '(targets, indices)\n', (48949, 48967), False, 'from allennlp.nn import util\n'), ((51426, 51459), 'pytest.raises', 'pytest.raises', (['ConfigurationError'], {}), '(ConfigurationError)\n', (51439, 51459), False, 'import pytest\n'), ((51765, 51796), 'numpy.array', 'numpy.array', (['[1, 2, 5, 1, 8, 9]'], {}), '([1, 2, 5, 1, 8, 9])\n', (51776, 51796), False, 'import numpy\n'), ((51911, 51946), 'numpy.array', 'numpy.array', (['[[1, 2, 3], [4, 5, 0]]'], {}), '([[1, 2, 3], [4, 5, 0]])\n', (51922, 51946), False, 'import numpy\n'), ((52440, 52545), 'numpy.array', 'numpy.array', (['[[[1, 2, 3, 4], [5, 5, 5, 5], [6, 8, 1, 2]], [[4, 3, 2, 1], [8, 7, 6, 5], [\n 0, 0, 0, 0]]]'], {}), '([[[1, 2, 3, 4], [5, 5, 5, 5], [6, 8, 1, 2]], [[4, 3, 2, 1], [8,\n 7, 6, 5], [0, 0, 0, 0]]])\n', (52451, 52545), False, 'import numpy\n'), ((52716, 52741), 'numpy.array', 'numpy.array', (['[9, 9, 9, 9]'], {}), '([9, 9, 9, 9])\n', (52727, 52741), False, 'import numpy\n'), ((52774, 52803), 'numpy.array', 'numpy.array', (['[10, 10, 10, 10]'], {}), '([10, 10, 10, 10])\n', (52785, 52803), False, 'import numpy\n'), ((53407, 53433), 'numpy.random.rand', 'numpy.random.rand', (['(3)', '(5)', '(7)'], {}), '(3, 5, 7)\n', (53424, 53433), False, 'import numpy\n'), ((56584, 56608), 'torch.Tensor', 'torch.Tensor', (['[[[2, 3]]]'], {}), '([[[2, 3]]])\n', (56596, 56608), False, 'import torch\n'), ((56610, 56634), 'torch.Tensor', 'torch.Tensor', (['[[[5, 5]]]'], {}), '([[[5, 5]]])\n', (56622, 56634), False, 'import torch\n'), ((56742, 56805), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['combination', 'tensors', 'weight'], {}), '(combination, tensors, weight)\n', (56775, 56805), False, 'from allennlp.nn import util\n'), ((56896, 56959), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['combination', 'tensors', 'weight'], {}), '(combination, tensors, weight)\n', (56929, 56959), False, 'from allennlp.nn import util\n'), ((57098, 57162), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['combination', 'tensors', 'weight2'], {}), '(combination, tensors, weight2)\n', (57131, 57162), False, 'from allennlp.nn import util\n'), ((57265, 57328), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['combination', 'tensors', 'weight'], {}), '(combination, tensors, weight)\n', (57298, 57328), False, 'from allennlp.nn import util\n'), ((57430, 57493), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['combination', 'tensors', 'weight'], {}), '(combination, tensors, weight)\n', (57463, 57493), False, 'from allennlp.nn import util\n'), ((57593, 57656), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['combination', 'tensors', 'weight'], {}), '(combination, tensors, weight)\n', (57626, 57656), False, 'from allennlp.nn import util\n'), ((57756, 57819), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['combination', 'tensors', 'weight'], {}), '(combination, tensors, weight)\n', (57789, 57819), False, 'from allennlp.nn import util\n'), ((57921, 57984), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['combination', 'tensors', 'weight'], {}), '(combination, tensors, weight)\n', (57954, 57984), False, 'from allennlp.nn import util\n'), ((58132, 58195), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['combination', 'tensors', 'weight'], {}), '(combination, tensors, weight)\n', (58165, 58195), False, 'from allennlp.nn import util\n'), ((58287, 58320), 'pytest.raises', 'pytest.raises', (['ConfigurationError'], {}), '(ConfigurationError)\n', (58300, 58320), False, 'import pytest\n'), ((58334, 58393), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['"""x+y+y"""', 'tensors', 'weight'], {}), "('x+y+y', tensors, weight)\n", (58367, 58393), False, 'from allennlp.nn import util\n'), ((58408, 58441), 'pytest.raises', 'pytest.raises', (['ConfigurationError'], {}), '(ConfigurationError)\n', (58421, 58441), False, 'import pytest\n'), ((58455, 58512), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['"""x%y"""', 'tensors', 'weight'], {}), "('x%y', tensors, weight)\n", (58488, 58512), False, 'from allennlp.nn import util\n'), ((58885, 58935), 'torch.Tensor', 'torch.Tensor', (['[[[5, 5], [4, 4]], [[2, 3], [1, 1]]]'], {}), '([[[5, 5], [4, 4]], [[2, 3], [1, 1]]])\n', (58897, 58935), False, 'import torch\n'), ((59064, 59127), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['combination', 'tensors', 'weight'], {}), '(combination, tensors, weight)\n', (59097, 59127), False, 'from allennlp.nn import util\n'), ((59223, 59273), 'torch.Tensor', 'torch.Tensor', (['[[[5, 5], [2, 2]], [[4, 4], [3, 3]]]'], {}), '([[[5, 5], [2, 2]], [[4, 4], [3, 3]]])\n', (59235, 59273), False, 'import torch\n'), ((59287, 59321), 'torch.Tensor', 'torch.Tensor', (['[[[2, 3]], [[1, 1]]]'], {}), '([[[2, 3]], [[1, 1]]])\n', (59299, 59321), False, 'import torch\n'), ((59440, 59503), 'allennlp.nn.util.combine_tensors_and_multiply', 'util.combine_tensors_and_multiply', (['combination', 'tensors', 'weight'], {}), '(combination, tensors, weight)\n', (59473, 59503), False, 'from allennlp.nn import util\n'), ((61836, 61866), 'torch.Size', 'torch.Size', (['(4, 10, 20, 17, 5)'], {}), '((4, 10, 20, 17, 5))\n', (61846, 61866), False, 'import torch\n'), ((62262, 62289), 'allennlp.models.load_archive', 'load_archive', (['model_archive'], {}), '(model_archive)\n', (62274, 62289), False, 'from allennlp.models import load_archive\n'), ((62387, 62402), 'json.load', 'json.load', (['file'], {}), '(file)\n', (62396, 62402), False, 'import json\n'), ((62448, 62478), 'allennlp.nn.util.inspect_parameters', 'util.inspect_parameters', (['model'], {}), '(model)\n', (62471, 62478), False, 'from allennlp.nn import util\n'), ((64640, 64677), 'numpy.array', 'numpy.array', (['[[0, 1], [1, 2], [2, 3]]'], {}), '([[0, 1], [1, 2], [2, 3]])\n', (64651, 64677), False, 'import numpy\n'), ((64755, 64773), 'numpy.ones', 'numpy.ones', (['[3, 2]'], {}), '([3, 2])\n', (64765, 64773), False, 'import numpy\n'), ((65765, 65794), 'numpy.array', 'numpy.array', (['[[0, 1], [1, 2]]'], {}), '([[0, 1], [1, 2]])\n', (65776, 65794), False, 'import numpy\n'), ((65885, 65922), 'numpy.array', 'numpy.array', (['[[1, 1], [1, 1], [0, 0]]'], {}), '([[1, 1], [1, 1], [0, 0]])\n', (65896, 65922), False, 'import numpy\n'), ((67000, 67046), 'numpy.array', 'numpy.array', (['[[0, 1, 3], [1, 2, 2], [1, 2, 2]]'], {}), '([[0, 1, 3], [1, 2, 2], [1, 2, 2]])\n', (67011, 67046), False, 'import numpy\n'), ((67137, 67183), 'numpy.array', 'numpy.array', (['[[1, 1, 1], [1, 1, 0], [1, 0, 0]]'], {}), '([[1, 1, 1], [1, 1, 0], [1, 0, 0]])\n', (67148, 67183), False, 'import numpy\n'), ((68466, 68512), 'numpy.array', 'numpy.array', (['[[0, 1, 2], [1, 2, 2], [3, 3, 3]]'], {}), '([[0, 1, 2], [1, 2, 2], [3, 3, 3]])\n', (68477, 68512), False, 'import numpy\n'), ((68603, 68649), 'numpy.array', 'numpy.array', (['[[1, 1, 1], [1, 1, 0], [0, 0, 0]]'], {}), '([[1, 1, 1], [1, 1, 0], [0, 0, 0]])\n', (68614, 68649), False, 'import numpy\n'), ((70950, 70967), 'torch.tensor', 'torch.tensor', (['[1]'], {}), '([1])\n', (70962, 70967), False, 'import torch\n'), ((70969, 70986), 'torch.tensor', 'torch.tensor', (['[1]'], {}), '([1])\n', (70981, 70986), False, 'import torch\n'), ((71114, 71134), 'torch.tensor', 'torch.tensor', (['[True]'], {}), '([True])\n', (71126, 71134), False, 'import torch\n'), ((71136, 71156), 'torch.tensor', 'torch.tensor', (['[True]'], {}), '([True])\n', (71148, 71156), False, 'import torch\n'), ((71215, 71232), 'torch.tensor', 'torch.tensor', (['[1]'], {}), '([1])\n', (71227, 71232), False, 'import torch\n'), ((71234, 71253), 'torch.tensor', 'torch.tensor', (['[1.0]'], {}), '([1.0])\n', (71246, 71253), False, 'import torch\n'), ((71289, 71306), 'torch.tensor', 'torch.tensor', (['[1]'], {}), '([1])\n', (71301, 71306), False, 'import torch\n'), ((71308, 71328), 'torch.tensor', 'torch.tensor', (['[True]'], {}), '([True])\n', (71320, 71328), False, 'import torch\n'), ((1761, 1779), 'torch.ones', 'torch.ones', (['(2)', '(260)'], {}), '(2, 260)\n', (1771, 1779), False, 'import torch\n'), ((2224, 2272), 'allennlp.nn.util.clamp_tensor', 'util.clamp_tensor', (['tensor'], {'minimum': '(-3)', 'maximum': '(3)'}), '(tensor, minimum=-3, maximum=3)\n', (2241, 2272), False, 'from allennlp.nn import util\n'), ((2584, 2632), 'allennlp.nn.util.clamp_tensor', 'util.clamp_tensor', (['tensor'], {'minimum': '(-3)', 'maximum': '(3)'}), '(tensor, minimum=-3, maximum=3)\n', (2601, 2632), False, 'from allennlp.nn import util\n'), ((18625, 18681), 'torch.tensor', 'torch.tensor', (['[[True, False, True], [True, True, False]]'], {}), '([[True, False, True], [True, True, False]])\n', (18637, 18681), False, 'import torch\n'), ((20420, 20476), 'torch.tensor', 'torch.tensor', (['[[True, False, True], [True, True, False]]'], {}), '([[True, False, True], [True, True, False]])\n', (20432, 20476), False, 'import torch\n'), ((22149, 22201), 'torch.LongTensor', 'torch.LongTensor', (['[[3, 4, 5, 0, 0], [1, 2, 0, 0, 0]]'], {}), '([[3, 4, 5, 0, 0], [1, 2, 0, 0, 0]])\n', (22165, 22201), False, 'import torch\n'), ((22239, 22345), 'torch.LongTensor', 'torch.LongTensor', (['[[[1, 2], [3, 0], [2, 0], [0, 0], [0, 0]], [[5, 0], [4, 6], [0, 0], [0, 0],\n [0, 0]]]'], {}), '([[[1, 2], [3, 0], [2, 0], [0, 0], [0, 0]], [[5, 0], [4, 6],\n [0, 0], [0, 0], [0, 0]]])\n', (22255, 22345), False, 'import torch\n'), ((22818, 22932), 'torch.LongTensor', 'torch.LongTensor', (['[[[1, 2, 3], [3, 0, 1], [2, 1, 0], [0, 0, 0]], [[5, 5, 5], [4, 6, 0], [0, 0,\n 0], [0, 0, 0]]]'], {}), '([[[1, 2, 3], [3, 0, 1], [2, 1, 0], [0, 0, 0]], [[5, 5, 5],\n [4, 6, 0], [0, 0, 0], [0, 0, 0]]])\n', (22834, 22932), False, 'import torch\n'), ((23383, 23489), 'torch.LongTensor', 'torch.LongTensor', (['[[[1, 2], [3, 0], [2, 0], [0, 0], [0, 0]], [[5, 0], [4, 6], [0, 0], [0, 0],\n [0, 0]]]'], {}), '([[[1, 2], [3, 0], [2, 0], [0, 0], [0, 0]], [[5, 0], [4, 6],\n [0, 0], [0, 0], [0, 0]]])\n', (23399, 23489), False, 'import torch\n'), ((24073, 24125), 'torch.LongTensor', 'torch.LongTensor', (['[[3, 4, 5, 0, 0], [1, 2, 0, 0, 0]]'], {}), '([[3, 4, 5, 0, 0], [1, 2, 0, 0, 0]])\n', (24089, 24125), False, 'import torch\n'), ((24151, 24187), 'torch.tensor', 'torch.tensor', (['[[False, False, True]]'], {}), '([[False, False, True]])\n', (24163, 24187), False, 'import torch\n'), ((24583, 24615), 'torch.from_numpy', 'torch.from_numpy', (['sentence_array'], {}), '(sentence_array)\n', (24599, 24615), False, 'import torch\n'), ((25572, 25604), 'torch.from_numpy', 'torch.from_numpy', (['sentence_array'], {}), '(sentence_array)\n', (25588, 25604), False, 'import torch\n'), ((25640, 25673), 'torch.from_numpy', 'torch.from_numpy', (['attention_array'], {}), '(attention_array)\n', (25656, 25673), False, 'import torch\n'), ((26509, 26541), 'torch.from_numpy', 'torch.from_numpy', (['sentence_array'], {}), '(sentence_array)\n', (26525, 26541), False, 'import torch\n'), ((26577, 26610), 'torch.from_numpy', 'torch.from_numpy', (['attention_array'], {}), '(attention_array)\n', (26593, 26610), False, 'import torch\n'), ((27086, 27177), 'numpy.testing.assert_almost_equal', 'numpy.testing.assert_almost_equal', (['aggregated_array[0, i, j]', 'expected_array'], {'decimal': '(5)'}), '(aggregated_array[0, i, j], expected_array,\n decimal=5)\n', (27119, 27177), False, 'import numpy\n'), ((27555, 27587), 'torch.from_numpy', 'torch.from_numpy', (['sentence_array'], {}), '(sentence_array)\n', (27571, 27587), False, 'import torch\n'), ((27623, 27656), 'torch.from_numpy', 'torch.from_numpy', (['attention_array'], {}), '(attention_array)\n', (27639, 27656), False, 'import torch\n'), ((37488, 37514), 'allennlp.common.util.sanitize', 'sanitize', (['viterbi_paths_v1'], {}), '(viterbi_paths_v1)\n', (37496, 37514), False, 'from allennlp.common.util import sanitize\n'), ((42378, 42397), 'numpy.random.rand', 'numpy.random.rand', ([], {}), '()\n', (42395, 42397), False, 'import numpy\n'), ((42416, 42435), 'numpy.random.rand', 'numpy.random.rand', ([], {}), '()\n', (42433, 42435), False, 'import numpy\n'), ((43479, 43498), 'numpy.random.rand', 'numpy.random.rand', ([], {}), '()\n', (43496, 43498), False, 'import numpy\n'), ((43517, 43536), 'numpy.random.rand', 'numpy.random.rand', ([], {}), '()\n', (43534, 43536), False, 'import numpy\n'), ((51510, 51531), 'torch.ones', 'torch.ones', (['[3, 4, 5]'], {}), '([3, 4, 5])\n', (51520, 51531), False, 'import torch\n'), ((64096, 64118), 'torch.randn', 'torch.randn', (['[3, 4, 5]'], {}), '([3, 4, 5])\n', (64107, 64118), False, 'import torch\n'), ((64275, 64293), 'torch.ones', 'torch.ones', (['[3, 4]'], {}), '([3, 4])\n', (64285, 64293), False, 'import torch\n'), ((65122, 65144), 'torch.randn', 'torch.randn', (['[3, 4, 5]'], {}), '([3, 4, 5])\n', (65133, 65144), False, 'import torch\n'), ((65301, 65319), 'torch.ones', 'torch.ones', (['[3, 4]'], {}), '([3, 4])\n', (65311, 65319), False, 'import torch\n'), ((66313, 66335), 'torch.randn', 'torch.randn', (['[3, 4, 5]'], {}), '([3, 4, 5])\n', (66324, 66335), False, 'import torch\n'), ((66603, 66621), 'torch.ones', 'torch.ones', (['[3, 4]'], {}), '([3, 4])\n', (66613, 66621), False, 'import torch\n'), ((67713, 67735), 'torch.randn', 'torch.randn', (['[3, 4, 5]'], {}), '([3, 4, 5])\n', (67724, 67735), False, 'import torch\n'), ((67892, 67910), 'torch.ones', 'torch.ones', (['[3, 4]'], {}), '([3, 4])\n', (67902, 67910), False, 'import torch\n'), ((71026, 71043), 'torch.tensor', 'torch.tensor', (['[1]'], {}), '([1])\n', (71038, 71043), False, 'import torch\n'), ((71045, 71062), 'torch.tensor', 'torch.tensor', (['[2]'], {}), '([2])\n', (71057, 71062), False, 'import torch\n'), ((71387, 71404), 'torch.tensor', 'torch.tensor', (['[1]'], {}), '([1])\n', (71399, 71404), False, 'import torch\n'), ((71408, 71425), 'torch.tensor', 'torch.tensor', (['[1]'], {}), '([1])\n', (71420, 71425), False, 'import torch\n'), ((71550, 71567), 'torch.tensor', 'torch.tensor', (['[1]'], {}), '([1])\n', (71562, 71567), False, 'import torch\n'), ((71578, 71595), 'torch.tensor', 'torch.tensor', (['[1]'], {}), '([1])\n', (71590, 71595), False, 'import torch\n'), ((1104, 1160), 'allennlp.nn.util.get_mask_from_sequence_lengths', 'util.get_mask_from_sequence_lengths', (['sequence_lengths', '(5)'], {}), '(sequence_lengths, 5)\n', (1139, 1160), False, 'from allennlp.nn import util\n'), ((4853, 4889), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'None'], {}), '(vector_1d, None)\n', (4872, 4889), False, 'from allennlp.nn import util\n'), ((5191, 5227), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'None'], {}), '(vector_1d, None)\n', (5210, 5227), False, 'from allennlp.nn import util\n'), ((5500, 5538), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_zero', 'None'], {}), '(vector_zero, None)\n', (5519, 5538), False, 'from allennlp.nn import util\n'), ((5843, 5876), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['matrix', 'None'], {}), '(matrix, None)\n', (5862, 5876), False, 'from allennlp.nn import util\n'), ((6291, 6324), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['matrix', 'None'], {}), '(matrix, None)\n', (6310, 6324), False, 'from allennlp.nn import util\n'), ((6783, 6822), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'mask_1d'], {}), '(vector_1d, mask_1d)\n', (6802, 6822), False, 'from allennlp.nn import util\n'), ((7090, 7129), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'mask_1d'], {}), '(vector_1d, mask_1d)\n', (7109, 7129), False, 'from allennlp.nn import util\n'), ((7534, 7573), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'mask_1d'], {}), '(vector_1d, mask_1d)\n', (7553, 7573), False, 'from allennlp.nn import util\n'), ((7928, 7967), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'mask_1d'], {}), '(vector_1d, mask_1d)\n', (7947, 7967), False, 'from allennlp.nn import util\n'), ((8326, 8365), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'mask_1d'], {}), '(vector_1d, mask_1d)\n', (8345, 8365), False, 'from allennlp.nn import util\n'), ((8707, 8746), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'mask_1d'], {}), '(vector_1d, mask_1d)\n', (8726, 8746), False, 'from allennlp.nn import util\n'), ((9073, 9106), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['matrix', 'mask'], {}), '(matrix, mask)\n', (9092, 9106), False, 'from allennlp.nn import util\n'), ((9586, 9619), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['matrix', 'mask'], {}), '(matrix, mask)\n', (9605, 9619), False, 'from allennlp.nn import util\n'), ((10074, 10107), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['matrix', 'mask'], {}), '(matrix, mask)\n', (10093, 10107), False, 'from allennlp.nn import util\n'), ((10431, 10464), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['matrix', 'mask'], {}), '(matrix, mask)\n', (10450, 10464), False, 'from allennlp.nn import util\n'), ((10869, 10931), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'mask_1d'], {'memory_efficient': '(True)'}), '(vector_1d, mask_1d, memory_efficient=True)\n', (10888, 10931), False, 'from allennlp.nn import util\n'), ((11221, 11283), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'mask_1d'], {'memory_efficient': '(True)'}), '(vector_1d, mask_1d, memory_efficient=True)\n', (11240, 11283), False, 'from allennlp.nn import util\n'), ((11710, 11772), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'mask_1d'], {'memory_efficient': '(True)'}), '(vector_1d, mask_1d, memory_efficient=True)\n', (11729, 11772), False, 'from allennlp.nn import util\n'), ((12149, 12211), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'mask_1d'], {'memory_efficient': '(True)'}), '(vector_1d, mask_1d, memory_efficient=True)\n', (12168, 12211), False, 'from allennlp.nn import util\n'), ((12596, 12658), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'mask_1d'], {'memory_efficient': '(True)'}), '(vector_1d, mask_1d, memory_efficient=True)\n', (12615, 12658), False, 'from allennlp.nn import util\n'), ((13026, 13088), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['vector_1d', 'mask_1d'], {'memory_efficient': '(True)'}), '(vector_1d, mask_1d, memory_efficient=True)\n', (13045, 13088), False, 'from allennlp.nn import util\n'), ((13437, 13493), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['matrix', 'mask'], {'memory_efficient': '(True)'}), '(matrix, mask, memory_efficient=True)\n', (13456, 13493), False, 'from allennlp.nn import util\n'), ((13995, 14051), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['matrix', 'mask'], {'memory_efficient': '(True)'}), '(matrix, mask, memory_efficient=True)\n', (14014, 14051), False, 'from allennlp.nn import util\n'), ((14528, 14584), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['matrix', 'mask'], {'memory_efficient': '(True)'}), '(matrix, mask, memory_efficient=True)\n', (14547, 14584), False, 'from allennlp.nn import util\n'), ((14964, 15020), 'allennlp.nn.util.masked_softmax', 'util.masked_softmax', (['matrix', 'mask'], {'memory_efficient': '(True)'}), '(matrix, mask, memory_efficient=True)\n', (14983, 15020), False, 'from allennlp.nn import util\n'), ((15641, 15684), 'allennlp.nn.util.masked_log_softmax', 'util.masked_log_softmax', (['vector_1d', 'mask_1d'], {}), '(vector_1d, mask_1d)\n', (15664, 15684), False, 'from allennlp.nn import util\n'), ((15985, 16028), 'allennlp.nn.util.masked_log_softmax', 'util.masked_log_softmax', (['vector_1d', 'mask_1d'], {}), '(vector_1d, mask_1d)\n', (16008, 16028), False, 'from allennlp.nn import util\n'), ((16444, 16487), 'allennlp.nn.util.masked_log_softmax', 'util.masked_log_softmax', (['vector_1d', 'mask_1d'], {}), '(vector_1d, mask_1d)\n', (16467, 16487), False, 'from allennlp.nn import util\n'), ((16945, 16988), 'allennlp.nn.util.masked_log_softmax', 'util.masked_log_softmax', (['vector_1d', 'mask_1d'], {}), '(vector_1d, mask_1d)\n', (16968, 16988), False, 'from allennlp.nn import util\n'), ((17021, 17053), 'numpy.isnan', 'numpy.isnan', (['vector_1d_softmaxed'], {}), '(vector_1d_softmaxed)\n', (17032, 17053), False, 'import numpy\n'), ((17272, 17314), 'allennlp.nn.util.masked_max', 'util.masked_max', (['vector_1d', 'mask_1d'], {'dim': '(0)'}), '(vector_1d, mask_1d, dim=0)\n', (17287, 17314), False, 'from allennlp.nn import util\n'), ((17618, 17660), 'allennlp.nn.util.masked_max', 'util.masked_max', (['vector_1d', 'mask_1d'], {'dim': '(0)'}), '(vector_1d, mask_1d, dim=0)\n', (17633, 17660), False, 'from allennlp.nn import util\n'), ((17693, 17721), 'numpy.isnan', 'numpy.isnan', (['vector_1d_maxed'], {}), '(vector_1d_maxed)\n', (17704, 17721), False, 'import numpy\n'), ((17944, 17981), 'allennlp.nn.util.masked_max', 'util.masked_max', (['matrix', 'mask'], {'dim': '(-1)'}), '(matrix, mask, dim=-1)\n', (17959, 17981), False, 'from allennlp.nn import util\n'), ((18297, 18348), 'allennlp.nn.util.masked_max', 'util.masked_max', (['matrix', 'mask'], {'dim': '(-1)', 'keepdim': '(True)'}), '(matrix, mask, dim=-1, keepdim=True)\n', (18312, 18348), False, 'from allennlp.nn import util\n'), ((18719, 18755), 'allennlp.nn.util.masked_max', 'util.masked_max', (['matrix', 'mask'], {'dim': '(1)'}), '(matrix, mask, dim=1)\n', (18734, 18755), False, 'from allennlp.nn import util\n'), ((19070, 19113), 'allennlp.nn.util.masked_mean', 'util.masked_mean', (['vector_1d', 'mask_1d'], {'dim': '(0)'}), '(vector_1d, mask_1d, dim=0)\n', (19086, 19113), False, 'from allennlp.nn import util\n'), ((19415, 19458), 'allennlp.nn.util.masked_mean', 'util.masked_mean', (['vector_1d', 'mask_1d'], {'dim': '(0)'}), '(vector_1d, mask_1d, dim=0)\n', (19431, 19458), False, 'from allennlp.nn import util\n'), ((19491, 19518), 'numpy.isnan', 'numpy.isnan', (['vector_1d_mean'], {}), '(vector_1d_mean)\n', (19502, 19518), False, 'import numpy\n'), ((19740, 19778), 'allennlp.nn.util.masked_mean', 'util.masked_mean', (['matrix', 'mask'], {'dim': '(-1)'}), '(matrix, mask, dim=-1)\n', (19756, 19778), False, 'from allennlp.nn import util\n'), ((20092, 20144), 'allennlp.nn.util.masked_mean', 'util.masked_mean', (['matrix', 'mask'], {'dim': '(-1)', 'keepdim': '(True)'}), '(matrix, mask, dim=-1, keepdim=True)\n', (20108, 20144), False, 'from allennlp.nn import util\n'), ((20513, 20550), 'allennlp.nn.util.masked_mean', 'util.masked_mean', (['matrix', 'mask'], {'dim': '(1)'}), '(matrix, mask, dim=1)\n', (20529, 20550), False, 'from allennlp.nn import util\n'), ((24723, 24775), 'allennlp.nn.util.weighted_sum', 'util.weighted_sum', (['sentence_tensor', 'attention_tensor'], {}), '(sentence_tensor, attention_tensor)\n', (24740, 24775), False, 'from allennlp.nn import util\n'), ((25709, 25761), 'allennlp.nn.util.weighted_sum', 'util.weighted_sum', (['sentence_tensor', 'attention_tensor'], {}), '(sentence_tensor, attention_tensor)\n', (25726, 25761), False, 'from allennlp.nn import util\n'), ((26646, 26698), 'allennlp.nn.util.weighted_sum', 'util.weighted_sum', (['sentence_tensor', 'attention_tensor'], {}), '(sentence_tensor, attention_tensor)\n', (26663, 26698), False, 'from allennlp.nn import util\n'), ((27692, 27744), 'allennlp.nn.util.weighted_sum', 'util.weighted_sum', (['sentence_tensor', 'attention_tensor'], {}), '(sentence_tensor, attention_tensor)\n', (27709, 27744), False, 'from allennlp.nn import util\n'), ((47471, 47493), 'torch.ones', 'torch.ones', (['[2, 10, 3]'], {}), '([2, 10, 3])\n', (47481, 47493), False, 'import torch\n'), ((49076, 49098), 'torch.ones', 'torch.ones', (['[3, 12, 2]'], {}), '([3, 12, 2])\n', (49086, 49098), False, 'import torch\n'), ((49469, 49495), 'torch.empty_like', 'torch.empty_like', (['selected'], {}), '(selected)\n', (49485, 49495), False, 'import torch\n'), ((50337, 50358), 'torch.ones', 'torch.ones', (['[2, 6, 3]'], {}), '([2, 6, 3])\n', (50347, 50358), False, 'import torch\n'), ((53650, 53714), 'numpy.array', 'numpy.array', (['[[1, 1, 0, 0, 0], [1, 1, 1, 1, 1], [1, 1, 1, 1, 0]]'], {}), '([[1, 1, 0, 0, 0], [1, 1, 1, 1, 1], [1, 1, 1, 1, 0]])\n', (53661, 53714), False, 'import numpy\n'), ((54118, 54164), 'numpy.array', 'numpy.array', (['[[0, 0, 0], [1, 1, 1], [1, 1, 0]]'], {}), '([[0, 0, 0], [1, 1, 1], [1, 1, 0]])\n', (54129, 54164), False, 'import numpy\n'), ((71467, 71484), 'torch.tensor', 'torch.tensor', (['[1]'], {}), '([1])\n', (71479, 71484), False, 'import torch\n'), ((71488, 71505), 'torch.tensor', 'torch.tensor', (['[2]'], {}), '([2])\n', (71500, 71505), False, 'import torch\n'), ((23655, 23720), 'allennlp.nn.util.get_text_field_mask', 'util.get_text_field_mask', (['text_field_tensors'], {'num_wrapping_dims': '(1)'}), '(text_field_tensors, num_wrapping_dims=1)\n', (23679, 23720), False, 'from allennlp.nn import util\n'), ((46404, 46426), 'allennlp.nn.util.logsumexp', 'util.logsumexp', (['tensor'], {}), '(tensor)\n', (46418, 46426), False, 'from allennlp.nn import util\n'), ((46530, 46552), 'allennlp.nn.util.logsumexp', 'util.logsumexp', (['tensor'], {}), '(tensor)\n', (46544, 46552), False, 'from allennlp.nn import util\n'), ((46671, 46700), 'allennlp.nn.util.logsumexp', 'util.logsumexp', (['tensor'], {'dim': '(0)'}), '(tensor, dim=0)\n', (46685, 46700), False, 'from allennlp.nn import util\n'), ((69023, 69160), 'torch.FloatTensor', 'torch.FloatTensor', (['[[[4, 2, 9, 9, 7], [-4, -2, -9, -9, -7]], [[5, 4, 1, 8, 8], [9, 1, 7, 4, 1]\n ], [[9, 8, 9, 6, 0], [2, 2, 2, 2, 2]]]'], {}), '([[[4, 2, 9, 9, 7], [-4, -2, -9, -9, -7]], [[5, 4, 1, 8, 8\n ], [9, 1, 7, 4, 1]], [[9, 8, 9, 6, 0], [2, 2, 2, 2, 2]]])\n', (69040, 69160), False, 'import torch\n'), ((69265, 69501), 'torch.tensor', 'torch.tensor', (['[[[False, False, False, False, False], [True, True, True, True, True]], [[\n True, True, True, True, False], [False, True, True, True, True]], [[\n True, False, True, True, True], [False, True, False, True, True]]]'], {}), '([[[False, False, False, False, False], [True, True, True, True,\n True]], [[True, True, True, True, False], [False, True, True, True, \n True]], [[True, False, True, True, True], [False, True, False, True, \n True]]])\n', (69277, 69501), False, 'import torch\n'), ((69760, 69897), 'torch.FloatTensor', 'torch.FloatTensor', (['[[[-4, -2, -9, -9, -7], [0, 0, 0, 0, 0]], [[5, 4, 7, 8, 1], [0, 0, 0, 4, 0]\n ], [[9, 2, 9, 6, 2], [0, 0, 0, 0, 0]]]'], {}), '([[[-4, -2, -9, -9, -7], [0, 0, 0, 0, 0]], [[5, 4, 7, 8, 1\n ], [0, 0, 0, 4, 0]], [[9, 2, 9, 6, 2], [0, 0, 0, 0, 0]]])\n', (69777, 69897), False, 'import torch\n'), ((70135, 70265), 'torch.LongTensor', 'torch.LongTensor', (['[[[1, 1, 1, 1, 1], [0, 0, 0, 0, 0]], [[0, 0, 1, 0, 1], [0, 0, 0, 1, 0]], [[\n 0, 1, 0, 0, 1], [0, 0, 0, 0, 0]]]'], {}), '([[[1, 1, 1, 1, 1], [0, 0, 0, 0, 0]], [[0, 0, 1, 0, 1], [0,\n 0, 0, 1, 0]], [[0, 1, 0, 0, 1], [0, 0, 0, 0, 0]]])\n', (70151, 70265), False, 'import torch\n'), ((22517, 22561), 'allennlp.nn.util.get_text_field_mask', 'util.get_text_field_mask', (['text_field_tensors'], {}), '(text_field_tensors)\n', (22541, 22561), False, 'from allennlp.nn import util\n'), ((23103, 23147), 'allennlp.nn.util.get_text_field_mask', 'util.get_text_field_mask', (['text_field_tensors'], {}), '(text_field_tensors)\n', (23127, 23147), False, 'from allennlp.nn import util\n'), ((24254, 24298), 'allennlp.nn.util.get_text_field_mask', 'util.get_text_field_mask', (['text_field_tensors'], {}), '(text_field_tensors)\n', (24278, 24298), False, 'from allennlp.nn import util\n')]
|
"""PyMC3-ArviZ conversion code."""
import logging
import warnings
from typing import ( # pylint: disable=unused-import
TYPE_CHECKING,
Any,
Dict,
Iterable,
List,
Mapping,
Optional,
Tuple,
Union,
)
import numpy as np
import xarray as xr
from aesara.graph.basic import Constant
from aesara.tensor.sharedvar import SharedVariable
from aesara.tensor.subtensor import AdvancedIncSubtensor
from arviz import InferenceData, concat, rcParams
from arviz.data.base import CoordSpec, DimSpec
from arviz.data.base import dict_to_dataset as _dict_to_dataset
from arviz.data.base import generate_dims_coords, make_attrs, requires
import pymc3
from pymc3.aesaraf import extract_obs_data
from pymc3.distributions import logpt
from pymc3.model import modelcontext
from pymc3.util import get_default_varnames
if TYPE_CHECKING:
from typing import Set # pylint: disable=ungrouped-imports
from pymc3.backends.base import MultiTrace # pylint: disable=invalid-name
from pymc3.model import Model
___all__ = [""]
_log = logging.getLogger("pymc3")
# random variable object ...
Var = Any # pylint: disable=invalid-name
class _DefaultTrace:
"""
Utility for collecting samples into a dictionary.
Name comes from its similarity to ``defaultdict``:
entries are lazily created.
Parameters
----------
samples : int
The number of samples that will be collected, per variable,
into the trace.
Attributes
----------
trace_dict : Dict[str, np.ndarray]
A dictionary constituting a trace. Should be extracted
after a procedure has filled the `_DefaultTrace` using the
`insert()` method
"""
trace_dict: Dict[str, np.ndarray] = {}
_len: Optional[int] = None
def __init__(self, samples: int):
self._len = samples
self.trace_dict = {}
def insert(self, k: str, v, idx: int):
"""
Insert `v` as the value of the `idx`th sample for the variable `k`.
Parameters
----------
k: str
Name of the variable.
v: anything that can go into a numpy array (including a numpy array)
The value of the `idx`th sample from variable `k`
ids: int
The index of the sample we are inserting into the trace.
"""
value_shape = np.shape(v)
# initialize if necessary
if k not in self.trace_dict:
array_shape = (self._len,) + value_shape
self.trace_dict[k] = np.empty(array_shape, dtype=np.array(v).dtype)
# do the actual insertion
if value_shape == ():
self.trace_dict[k][idx] = v
else:
self.trace_dict[k][idx, :] = v
def dict_to_dataset(
data,
library=None,
coords=None,
dims=None,
attrs=None,
default_dims=None,
skip_event_dims=None,
index_origin=None,
):
"""Temporal workaround for dict_to_dataset.
Once ArviZ>0.11.2 release is available, only two changes are needed for everything to work.
1) this should be deleted, 2) dict_to_dataset should be imported as is from arviz, no underscore,
also remove unnecessary imports
"""
if default_dims is None:
return _dict_to_dataset(
data, library=library, coords=coords, dims=dims, skip_event_dims=skip_event_dims
)
else:
out_data = {}
for name, vals in data.items():
vals = np.atleast_1d(vals)
val_dims = dims.get(name)
val_dims, coords = generate_dims_coords(vals.shape, name, dims=val_dims, coords=coords)
coords = {key: xr.IndexVariable((key,), data=coords[key]) for key in val_dims}
out_data[name] = xr.DataArray(vals, dims=val_dims, coords=coords)
return xr.Dataset(data_vars=out_data, attrs=make_attrs(library=library))
class InferenceDataConverter: # pylint: disable=too-many-instance-attributes
"""Encapsulate InferenceData specific logic."""
model = None # type: Optional[Model]
nchains = None # type: int
ndraws = None # type: int
posterior_predictive = None # Type: Optional[Mapping[str, np.ndarray]]
predictions = None # Type: Optional[Mapping[str, np.ndarray]]
prior = None # Type: Optional[Mapping[str, np.ndarray]]
def __init__(
self,
*,
trace=None,
prior=None,
posterior_predictive=None,
log_likelihood=True,
predictions=None,
coords: Optional[CoordSpec] = None,
dims: Optional[DimSpec] = None,
model=None,
save_warmup: Optional[bool] = None,
density_dist_obs: bool = True,
index_origin: Optional[int] = None,
):
self.save_warmup = rcParams["data.save_warmup"] if save_warmup is None else save_warmup
self.trace = trace
# this permits us to get the model from command-line argument or from with model:
self.model = modelcontext(model)
self.attrs = None
if trace is not None:
self.nchains = trace.nchains if hasattr(trace, "nchains") else 1
if hasattr(trace.report, "n_draws") and trace.report.n_draws is not None:
self.ndraws = trace.report.n_draws
self.attrs = {
"sampling_time": trace.report.t_sampling,
"tuning_steps": trace.report.n_tune,
}
else:
self.ndraws = len(trace)
if self.save_warmup:
warnings.warn(
"Warmup samples will be stored in posterior group and will not be"
" excluded from stats and diagnostics."
" Do not slice the trace manually before conversion",
UserWarning,
)
self.ntune = len(self.trace) - self.ndraws
self.posterior_trace, self.warmup_trace = self.split_trace()
else:
self.nchains = self.ndraws = 0
self.prior = prior
self.posterior_predictive = posterior_predictive
self.log_likelihood = log_likelihood
self.predictions = predictions
self.index_origin = rcParams["data.index_origin"] if index_origin is None else index_origin
def arbitrary_element(dct: Dict[Any, np.ndarray]) -> np.ndarray:
return next(iter(dct.values()))
if trace is None:
# if you have a posterior_predictive built with keep_dims,
# you'll lose here, but there's nothing I can do about that.
self.nchains = 1
get_from = None
if predictions is not None:
get_from = predictions
elif posterior_predictive is not None:
get_from = posterior_predictive
elif prior is not None:
get_from = prior
if get_from is None:
# pylint: disable=line-too-long
raise ValueError(
"When constructing InferenceData must have at least"
" one of trace, prior, posterior_predictive or predictions."
)
aelem = arbitrary_element(get_from)
self.ndraws = aelem.shape[0]
self.coords = {} if coords is None else coords
if hasattr(self.model, "coords"):
self.coords = {**self.model.coords, **self.coords}
self.coords = {key: value for key, value in self.coords.items() if value is not None}
self.dims = {} if dims is None else dims
if hasattr(self.model, "RV_dims"):
model_dims = {
var_name: [dim for dim in dims if dim is not None]
for var_name, dims in self.model.RV_dims.items()
}
self.dims = {**model_dims, **self.dims}
self.density_dist_obs = density_dist_obs
self.observations = self.find_observations()
def find_observations(self) -> Optional[Dict[str, Var]]:
"""If there are observations available, return them as a dictionary."""
if self.model is None:
return None
observations = {}
for obs in self.model.observed_RVs:
aux_obs = getattr(obs.tag, "observations", None)
if aux_obs is not None:
try:
obs_data = extract_obs_data(aux_obs)
observations[obs.name] = obs_data
except TypeError:
warnings.warn(f"Could not extract data from symbolic observation {obs}")
else:
warnings.warn(f"No data for observation {obs}")
return observations
def split_trace(self) -> Tuple[Union[None, "MultiTrace"], Union[None, "MultiTrace"]]:
"""Split MultiTrace object into posterior and warmup.
Returns
-------
trace_posterior: MultiTrace or None
The slice of the trace corresponding to the posterior. If the posterior
trace is empty, None is returned
trace_warmup: MultiTrace or None
The slice of the trace corresponding to the warmup. If the warmup trace is
empty or ``save_warmup=False``, None is returned
"""
trace_posterior = None
trace_warmup = None
if self.save_warmup and self.ntune > 0:
trace_warmup = self.trace[: self.ntune]
if self.ndraws > 0:
trace_posterior = self.trace[self.ntune :]
return trace_posterior, trace_warmup
def log_likelihood_vals_point(self, point, var, log_like_fun):
"""Compute log likelihood for each observed point."""
# TODO: This is a cheap hack; we should filter-out the correct
# variables some other way
point = {i.name: point[i.name] for i in log_like_fun.f.maker.inputs if i.name in point}
log_like_val = np.atleast_1d(log_like_fun(point))
if isinstance(var.owner.op, AdvancedIncSubtensor):
try:
obs_data = extract_obs_data(var.tag.observations)
except TypeError:
warnings.warn(f"Could not extract data from symbolic observation {var}")
mask = obs_data.mask
if np.ndim(mask) > np.ndim(log_like_val):
mask = np.any(mask, axis=-1)
log_like_val = np.where(mask, np.nan, log_like_val)
return log_like_val
def _extract_log_likelihood(self, trace):
"""Compute log likelihood of each observation."""
if self.trace is None:
return None
if self.model is None:
return None
if self.log_likelihood is True:
cached = [(var, self.model.fn(logpt(var))) for var in self.model.observed_RVs]
else:
cached = [
(var, self.model.fn(logpt(var)))
for var in self.model.observed_RVs
if var.name in self.log_likelihood
]
log_likelihood_dict = _DefaultTrace(len(trace.chains))
for var, log_like_fun in cached:
for k, chain in enumerate(trace.chains):
log_like_chain = [
self.log_likelihood_vals_point(point, var, log_like_fun)
for point in trace.points([chain])
]
log_likelihood_dict.insert(var.name, np.stack(log_like_chain), k)
return log_likelihood_dict.trace_dict
@requires("trace")
def posterior_to_xarray(self):
"""Convert the posterior to an xarray dataset."""
var_names = get_default_varnames(self.trace.varnames, include_transformed=False)
data = {}
data_warmup = {}
for var_name in var_names:
if self.warmup_trace:
data_warmup[var_name] = np.array(
self.warmup_trace.get_values(var_name, combine=False, squeeze=False)
)
if self.posterior_trace:
data[var_name] = np.array(
self.posterior_trace.get_values(var_name, combine=False, squeeze=False)
)
return (
dict_to_dataset(
data,
library=pymc3,
coords=self.coords,
dims=self.dims,
attrs=self.attrs,
index_origin=self.index_origin,
),
dict_to_dataset(
data_warmup,
library=pymc3,
coords=self.coords,
dims=self.dims,
attrs=self.attrs,
index_origin=self.index_origin,
),
)
@requires("trace")
def sample_stats_to_xarray(self):
"""Extract sample_stats from PyMC3 trace."""
data = {}
rename_key = {
"model_logp": "lp",
"mean_tree_accept": "acceptance_rate",
"depth": "tree_depth",
"tree_size": "n_steps",
}
data = {}
data_warmup = {}
for stat in self.trace.stat_names:
name = rename_key.get(stat, stat)
if name == "tune":
continue
if self.warmup_trace:
data_warmup[name] = np.array(
self.warmup_trace.get_sampler_stats(stat, combine=False)
)
if self.posterior_trace:
data[name] = np.array(self.posterior_trace.get_sampler_stats(stat, combine=False))
return (
dict_to_dataset(
data,
library=pymc3,
dims=None,
coords=self.coords,
attrs=self.attrs,
index_origin=self.index_origin,
),
dict_to_dataset(
data_warmup,
library=pymc3,
dims=None,
coords=self.coords,
attrs=self.attrs,
index_origin=self.index_origin,
),
)
@requires("trace")
@requires("model")
def log_likelihood_to_xarray(self):
"""Extract log likelihood and log_p data from PyMC3 trace."""
if self.predictions or not self.log_likelihood:
return None
data_warmup = {}
data = {}
warn_msg = (
"Could not compute log_likelihood, it will be omitted. "
"Check your model object or set log_likelihood=False"
)
if self.posterior_trace:
try:
data = self._extract_log_likelihood(self.posterior_trace)
except TypeError:
warnings.warn(warn_msg)
if self.warmup_trace:
try:
data_warmup = self._extract_log_likelihood(self.warmup_trace)
except TypeError:
warnings.warn(warn_msg)
return (
dict_to_dataset(
data,
library=pymc3,
dims=self.dims,
coords=self.coords,
skip_event_dims=True,
index_origin=self.index_origin,
),
dict_to_dataset(
data_warmup,
library=pymc3,
dims=self.dims,
coords=self.coords,
skip_event_dims=True,
index_origin=self.index_origin,
),
)
def translate_posterior_predictive_dict_to_xarray(self, dct) -> xr.Dataset:
"""Take Dict of variables to numpy ndarrays (samples) and translate into dataset."""
data = {}
for k, ary in dct.items():
shape = ary.shape
if shape[0] == self.nchains and shape[1] == self.ndraws:
data[k] = ary
elif shape[0] == self.nchains * self.ndraws:
data[k] = ary.reshape((self.nchains, self.ndraws, *shape[1:]))
else:
data[k] = np.expand_dims(ary, 0)
# pylint: disable=line-too-long
_log.warning(
"posterior predictive variable %s's shape not compatible with number of chains and draws. "
"This can mean that some draws or even whole chains are not represented.",
k,
)
return dict_to_dataset(
data, library=pymc3, coords=self.coords, dims=self.dims, index_origin=self.index_origin
)
@requires(["posterior_predictive"])
def posterior_predictive_to_xarray(self):
"""Convert posterior_predictive samples to xarray."""
return self.translate_posterior_predictive_dict_to_xarray(self.posterior_predictive)
@requires(["predictions"])
def predictions_to_xarray(self):
"""Convert predictions (out of sample predictions) to xarray."""
return self.translate_posterior_predictive_dict_to_xarray(self.predictions)
def priors_to_xarray(self):
"""Convert prior samples (and if possible prior predictive too) to xarray."""
if self.prior is None:
return {"prior": None, "prior_predictive": None}
if self.observations is not None:
prior_predictive_vars = list(self.observations.keys())
prior_vars = [key for key in self.prior.keys() if key not in prior_predictive_vars]
else:
prior_vars = list(self.prior.keys())
prior_predictive_vars = None
priors_dict = {}
for group, var_names in zip(
("prior", "prior_predictive"), (prior_vars, prior_predictive_vars)
):
priors_dict[group] = (
None
if var_names is None
else dict_to_dataset(
{k: np.expand_dims(self.prior[k], 0) for k in var_names},
library=pymc3,
coords=self.coords,
dims=self.dims,
index_origin=self.index_origin,
)
)
return priors_dict
@requires("observations")
@requires("model")
def observed_data_to_xarray(self):
"""Convert observed data to xarray."""
if self.predictions:
return None
return dict_to_dataset(
self.observations,
library=pymc3,
coords=self.coords,
dims=self.dims,
default_dims=[],
index_origin=self.index_origin,
)
@requires(["trace", "predictions"])
@requires("model")
def constant_data_to_xarray(self):
"""Convert constant data to xarray."""
# For constant data, we are concerned only with deterministics and
# data. The constant data vars must be either pm.Data
# (TensorSharedVariable) or pm.Deterministic
constant_data_vars = {} # type: Dict[str, Var]
def is_data(name, var) -> bool:
assert self.model is not None
return (
var not in self.model.deterministics
and var not in self.model.observed_RVs
and var not in self.model.free_RVs
and var not in self.model.potentials
and (self.observations is None or name not in self.observations)
and isinstance(var, (Constant, SharedVariable))
)
# I don't know how to find pm.Data, except that they are named
# variables that aren't observed or free RVs, nor are they
# deterministics, and then we eliminate observations.
for name, var in self.model.named_vars.items():
if is_data(name, var):
constant_data_vars[name] = var
if not constant_data_vars:
return None
constant_data = {}
for name, vals in constant_data_vars.items():
if hasattr(vals, "get_value"):
vals = vals.get_value()
elif hasattr(vals, "data"):
vals = vals.data
constant_data[name] = vals
return dict_to_dataset(
constant_data,
library=pymc3,
coords=self.coords,
dims=self.dims,
default_dims=[],
index_origin=self.index_origin,
)
def to_inference_data(self):
"""Convert all available data to an InferenceData object.
Note that if groups can not be created (e.g., there is no `trace`, so
the `posterior` and `sample_stats` can not be extracted), then the InferenceData
will not have those groups.
"""
id_dict = {
"posterior": self.posterior_to_xarray(),
"sample_stats": self.sample_stats_to_xarray(),
"log_likelihood": self.log_likelihood_to_xarray(),
"posterior_predictive": self.posterior_predictive_to_xarray(),
"predictions": self.predictions_to_xarray(),
**self.priors_to_xarray(),
"observed_data": self.observed_data_to_xarray(),
}
if self.predictions:
id_dict["predictions_constant_data"] = self.constant_data_to_xarray()
else:
id_dict["constant_data"] = self.constant_data_to_xarray()
return InferenceData(save_warmup=self.save_warmup, **id_dict)
def to_inference_data(
trace: Optional["MultiTrace"] = None,
*,
prior: Optional[Dict[str, Any]] = None,
posterior_predictive: Optional[Dict[str, Any]] = None,
log_likelihood: Union[bool, Iterable[str]] = True,
coords: Optional[CoordSpec] = None,
dims: Optional[DimSpec] = None,
model: Optional["Model"] = None,
save_warmup: Optional[bool] = None,
density_dist_obs: bool = True,
) -> InferenceData:
"""Convert pymc3 data into an InferenceData object.
All three of them are optional arguments, but at least one of ``trace``,
``prior`` and ``posterior_predictive`` must be present.
For a usage example read the
:ref:`Creating InferenceData section on from_pymc3 <creating_InferenceData>`
Parameters
----------
trace : MultiTrace, optional
Trace generated from MCMC sampling. Output of
:func:`~pymc3.sampling.sample`.
prior : dict, optional
Dictionary with the variable names as keys, and values numpy arrays
containing prior and prior predictive samples.
posterior_predictive : dict, optional
Dictionary with the variable names as keys, and values numpy arrays
containing posterior predictive samples.
log_likelihood : bool or array_like of str, optional
List of variables to calculate `log_likelihood`. Defaults to True which calculates
`log_likelihood` for all observed variables. If set to False, log_likelihood is skipped.
coords : dict of {str: array-like}, optional
Map of coordinate names to coordinate values
dims : dict of {str: list of str}, optional
Map of variable names to the coordinate names to use to index its dimensions.
model : Model, optional
Model used to generate ``trace``. It is not necessary to pass ``model`` if in
``with`` context.
save_warmup : bool, optional
Save warmup iterations InferenceData object. If not defined, use default
defined by the rcParams.
density_dist_obs : bool, default True
Store variables passed with ``observed`` arg to
:class:`~pymc.distributions.DensityDist` in the generated InferenceData.
Returns
-------
arviz.InferenceData
"""
if isinstance(trace, InferenceData):
return trace
return InferenceDataConverter(
trace=trace,
prior=prior,
posterior_predictive=posterior_predictive,
log_likelihood=log_likelihood,
coords=coords,
dims=dims,
model=model,
save_warmup=save_warmup,
density_dist_obs=density_dist_obs,
).to_inference_data()
### Later I could have this return ``None`` if the ``idata_orig`` argument is supplied. But
### perhaps we should have an inplace argument?
def predictions_to_inference_data(
predictions,
posterior_trace: Optional["MultiTrace"] = None,
model: Optional["Model"] = None,
coords: Optional[CoordSpec] = None,
dims: Optional[DimSpec] = None,
idata_orig: Optional[InferenceData] = None,
inplace: bool = False,
) -> InferenceData:
"""Translate out-of-sample predictions into ``InferenceData``.
Parameters
----------
predictions: Dict[str, np.ndarray]
The predictions are the return value of :func:`~pymc3.sample_posterior_predictive`,
a dictionary of strings (variable names) to numpy ndarrays (draws).
posterior_trace: MultiTrace
This should be a trace that has been thinned appropriately for
``pymc3.sample_posterior_predictive``. Specifically, any variable whose shape is
a deterministic function of the shape of any predictor (explanatory, independent, etc.)
variables must be *removed* from this trace.
model: Model
The pymc3 model. It can be ommited if within a model context.
coords: Dict[str, array-like[Any]]
Coordinates for the variables. Map from coordinate names to coordinate values.
dims: Dict[str, array-like[str]]
Map from variable name to ordered set of coordinate names.
idata_orig: InferenceData, optional
If supplied, then modify this inference data in place, adding ``predictions`` and
(if available) ``predictions_constant_data`` groups. If this is not supplied, make a
fresh InferenceData
inplace: boolean, optional
If idata_orig is supplied and inplace is True, merge the predictions into idata_orig,
rather than returning a fresh InferenceData object.
Returns
-------
InferenceData:
May be modified ``idata_orig``.
"""
if inplace and not idata_orig:
raise ValueError(
"Do not pass True for inplace unless passing" "an existing InferenceData as idata_orig"
)
new_idata = InferenceDataConverter(
trace=posterior_trace,
predictions=predictions,
model=model,
coords=coords,
dims=dims,
log_likelihood=False,
).to_inference_data()
if idata_orig is None:
return new_idata
elif inplace:
concat([idata_orig, new_idata], dim=None, inplace=True)
return idata_orig
else:
# if we are not returning in place, then merge the old groups into the new inference
# data and return that.
concat([new_idata, idata_orig], dim=None, copy=True, inplace=True)
return new_idata
|
[
"logging.getLogger",
"xarray.IndexVariable",
"arviz.data.base.dict_to_dataset",
"numpy.array",
"pymc3.aesaraf.extract_obs_data",
"numpy.where",
"numpy.ndim",
"numpy.stack",
"warnings.warn",
"pymc3.util.get_default_varnames",
"numpy.any",
"numpy.shape",
"arviz.data.base.generate_dims_coords",
"numpy.atleast_1d",
"pymc3.distributions.logpt",
"pymc3.model.modelcontext",
"arviz.InferenceData",
"xarray.DataArray",
"arviz.concat",
"arviz.data.base.make_attrs",
"numpy.expand_dims",
"arviz.data.base.requires"
] |
[((1054, 1080), 'logging.getLogger', 'logging.getLogger', (['"""pymc3"""'], {}), "('pymc3')\n", (1071, 1080), False, 'import logging\n'), ((11415, 11432), 'arviz.data.base.requires', 'requires', (['"""trace"""'], {}), "('trace')\n", (11423, 11432), False, 'from arviz.data.base import generate_dims_coords, make_attrs, requires\n'), ((12608, 12625), 'arviz.data.base.requires', 'requires', (['"""trace"""'], {}), "('trace')\n", (12616, 12625), False, 'from arviz.data.base import generate_dims_coords, make_attrs, requires\n'), ((13946, 13963), 'arviz.data.base.requires', 'requires', (['"""trace"""'], {}), "('trace')\n", (13954, 13963), False, 'from arviz.data.base import generate_dims_coords, make_attrs, requires\n'), ((13969, 13986), 'arviz.data.base.requires', 'requires', (['"""model"""'], {}), "('model')\n", (13977, 13986), False, 'from arviz.data.base import generate_dims_coords, make_attrs, requires\n'), ((16344, 16378), 'arviz.data.base.requires', 'requires', (["['posterior_predictive']"], {}), "(['posterior_predictive'])\n", (16352, 16378), False, 'from arviz.data.base import generate_dims_coords, make_attrs, requires\n'), ((16586, 16611), 'arviz.data.base.requires', 'requires', (["['predictions']"], {}), "(['predictions'])\n", (16594, 16611), False, 'from arviz.data.base import generate_dims_coords, make_attrs, requires\n'), ((17916, 17940), 'arviz.data.base.requires', 'requires', (['"""observations"""'], {}), "('observations')\n", (17924, 17940), False, 'from arviz.data.base import generate_dims_coords, make_attrs, requires\n'), ((17946, 17963), 'arviz.data.base.requires', 'requires', (['"""model"""'], {}), "('model')\n", (17954, 17963), False, 'from arviz.data.base import generate_dims_coords, make_attrs, requires\n'), ((18342, 18376), 'arviz.data.base.requires', 'requires', (["['trace', 'predictions']"], {}), "(['trace', 'predictions'])\n", (18350, 18376), False, 'from arviz.data.base import generate_dims_coords, make_attrs, requires\n'), ((18382, 18399), 'arviz.data.base.requires', 'requires', (['"""model"""'], {}), "('model')\n", (18390, 18399), False, 'from arviz.data.base import generate_dims_coords, make_attrs, requires\n'), ((2352, 2363), 'numpy.shape', 'np.shape', (['v'], {}), '(v)\n', (2360, 2363), True, 'import numpy as np\n'), ((3240, 3342), 'arviz.data.base.dict_to_dataset', '_dict_to_dataset', (['data'], {'library': 'library', 'coords': 'coords', 'dims': 'dims', 'skip_event_dims': 'skip_event_dims'}), '(data, library=library, coords=coords, dims=dims,\n skip_event_dims=skip_event_dims)\n', (3256, 3342), True, 'from arviz.data.base import dict_to_dataset as _dict_to_dataset\n'), ((4950, 4969), 'pymc3.model.modelcontext', 'modelcontext', (['model'], {}), '(model)\n', (4962, 4969), False, 'from pymc3.model import modelcontext\n'), ((11546, 11614), 'pymc3.util.get_default_varnames', 'get_default_varnames', (['self.trace.varnames'], {'include_transformed': '(False)'}), '(self.trace.varnames, include_transformed=False)\n', (11566, 11614), False, 'from pymc3.util import get_default_varnames\n'), ((21077, 21131), 'arviz.InferenceData', 'InferenceData', ([], {'save_warmup': 'self.save_warmup'}), '(save_warmup=self.save_warmup, **id_dict)\n', (21090, 21131), False, 'from arviz import InferenceData, concat, rcParams\n'), ((3452, 3471), 'numpy.atleast_1d', 'np.atleast_1d', (['vals'], {}), '(vals)\n', (3465, 3471), True, 'import numpy as np\n'), ((3541, 3609), 'arviz.data.base.generate_dims_coords', 'generate_dims_coords', (['vals.shape', 'name'], {'dims': 'val_dims', 'coords': 'coords'}), '(vals.shape, name, dims=val_dims, coords=coords)\n', (3561, 3609), False, 'from arviz.data.base import generate_dims_coords, make_attrs, requires\n'), ((3730, 3778), 'xarray.DataArray', 'xr.DataArray', (['vals'], {'dims': 'val_dims', 'coords': 'coords'}), '(vals, dims=val_dims, coords=coords)\n', (3742, 3778), True, 'import xarray as xr\n'), ((10325, 10361), 'numpy.where', 'np.where', (['mask', 'np.nan', 'log_like_val'], {}), '(mask, np.nan, log_like_val)\n', (10333, 10361), True, 'import numpy as np\n'), ((26181, 26236), 'arviz.concat', 'concat', (['[idata_orig, new_idata]'], {'dim': 'None', 'inplace': '(True)'}), '([idata_orig, new_idata], dim=None, inplace=True)\n', (26187, 26236), False, 'from arviz import InferenceData, concat, rcParams\n'), ((26406, 26472), 'arviz.concat', 'concat', (['[new_idata, idata_orig]'], {'dim': 'None', 'copy': '(True)', 'inplace': '(True)'}), '([new_idata, idata_orig], dim=None, copy=True, inplace=True)\n', (26412, 26472), False, 'from arviz import InferenceData, concat, rcParams\n'), ((3637, 3679), 'xarray.IndexVariable', 'xr.IndexVariable', (['(key,)'], {'data': 'coords[key]'}), '((key,), data=coords[key])\n', (3653, 3679), True, 'import xarray as xr\n'), ((3831, 3858), 'arviz.data.base.make_attrs', 'make_attrs', ([], {'library': 'library'}), '(library=library)\n', (3841, 3858), False, 'from arviz.data.base import generate_dims_coords, make_attrs, requires\n'), ((8589, 8636), 'warnings.warn', 'warnings.warn', (['f"""No data for observation {obs}"""'], {}), "(f'No data for observation {obs}')\n", (8602, 8636), False, 'import warnings\n'), ((10007, 10045), 'pymc3.aesaraf.extract_obs_data', 'extract_obs_data', (['var.tag.observations'], {}), '(var.tag.observations)\n', (10023, 10045), False, 'from pymc3.aesaraf import extract_obs_data\n'), ((10214, 10227), 'numpy.ndim', 'np.ndim', (['mask'], {}), '(mask)\n', (10221, 10227), True, 'import numpy as np\n'), ((10230, 10251), 'numpy.ndim', 'np.ndim', (['log_like_val'], {}), '(log_like_val)\n', (10237, 10251), True, 'import numpy as np\n'), ((10276, 10297), 'numpy.any', 'np.any', (['mask'], {'axis': '(-1)'}), '(mask, axis=-1)\n', (10282, 10297), True, 'import numpy as np\n'), ((5525, 5716), 'warnings.warn', 'warnings.warn', (['"""Warmup samples will be stored in posterior group and will not be excluded from stats and diagnostics. Do not slice the trace manually before conversion"""', 'UserWarning'], {}), "(\n 'Warmup samples will be stored in posterior group and will not be excluded from stats and diagnostics. Do not slice the trace manually before conversion'\n , UserWarning)\n", (5538, 5716), False, 'import warnings\n'), ((8348, 8373), 'pymc3.aesaraf.extract_obs_data', 'extract_obs_data', (['aux_obs'], {}), '(aux_obs)\n', (8364, 8373), False, 'from pymc3.aesaraf import extract_obs_data\n'), ((10092, 10164), 'warnings.warn', 'warnings.warn', (['f"""Could not extract data from symbolic observation {var}"""'], {}), "(f'Could not extract data from symbolic observation {var}')\n", (10105, 10164), False, 'import warnings\n'), ((11334, 11358), 'numpy.stack', 'np.stack', (['log_like_chain'], {}), '(log_like_chain)\n', (11342, 11358), True, 'import numpy as np\n'), ((14556, 14579), 'warnings.warn', 'warnings.warn', (['warn_msg'], {}), '(warn_msg)\n', (14569, 14579), False, 'import warnings\n'), ((14751, 14774), 'warnings.warn', 'warnings.warn', (['warn_msg'], {}), '(warn_msg)\n', (14764, 14774), False, 'import warnings\n'), ((15847, 15869), 'numpy.expand_dims', 'np.expand_dims', (['ary', '(0)'], {}), '(ary, 0)\n', (15861, 15869), True, 'import numpy as np\n'), ((2550, 2561), 'numpy.array', 'np.array', (['v'], {}), '(v)\n', (2558, 2561), True, 'import numpy as np\n'), ((8482, 8554), 'warnings.warn', 'warnings.warn', (['f"""Could not extract data from symbolic observation {obs}"""'], {}), "(f'Could not extract data from symbolic observation {obs}')\n", (8495, 8554), False, 'import warnings\n'), ((10688, 10698), 'pymc3.distributions.logpt', 'logpt', (['var'], {}), '(var)\n', (10693, 10698), False, 'from pymc3.distributions import logpt\n'), ((10810, 10820), 'pymc3.distributions.logpt', 'logpt', (['var'], {}), '(var)\n', (10815, 10820), False, 'from pymc3.distributions import logpt\n'), ((17634, 17666), 'numpy.expand_dims', 'np.expand_dims', (['self.prior[k]', '(0)'], {}), '(self.prior[k], 0)\n', (17648, 17666), True, 'import numpy as np\n')]
|
from __future__ import print_function
import json
import os
import requests
from datetime import datetime
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
DEMO_UID = 0
PREDICTION_RESPONSE_KEY_QUERY_ID = "query_id"
PREDICTION_RESPONSE_KEY_OUTPUT = "output"
PREDICTION_RESPONSE_KEY_USED_DEFAULT = "default"
PREDICTION_ERROR_RESPONSE_KEY_ERROR = "error"
PREDICTION_ERROR_RESPONSE_KEY_CAUSE = "cause"
classes = [
'airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse',
'ship', 'truck'
]
positive_class = classes.index('airplane')
negative_class = classes.index('bird')
def recover_pixels(x):
return np.transpose(x.reshape(3, 32, 32), (1, 2, 0))
def show_example_images(images, labels, num_rows):
imgs_per_row = 6
num_images = imgs_per_row * num_rows
idxs = np.random.randint(0, len(labels), num_images)
f, axes = plt.subplots(
nrows=num_rows,
ncols=imgs_per_row,
figsize=(1.5 * imgs_per_row, 1.5 * num_rows))
f.tight_layout()
for i, idx in enumerate(idxs):
image = recover_pixels(images[idx])
label = labels[idx]
cur_ax = axes[i / imgs_per_row][i % imgs_per_row]
cur_ax.imshow(image.astype(np.ubyte), interpolation="nearest")
cur_ax.axis('off')
if label == 0:
title = classes[negative_class]
else:
title = classes[positive_class]
cur_ax.set_title(title)
def load_cifar(cifar_location, cifar_filename="cifar_train.data", norm=True):
cifar_path = os.path.join(cifar_location, cifar_filename)
# print("Source file: %s" % cifar_path)
df = pd.read_csv(cifar_path, sep=",", header=None)
data = df.values
print("Number of image files: %d" % len(data))
y = data[:, 0]
X = data[:, 1:]
Z = X
if norm:
mu = np.mean(X.T, 0)
sigma = np.var(X.T, 0)
Z = (X.T - mu) / np.array([np.sqrt(z) if z > 0 else 1. for z in sigma])
Z = Z.T
return Z, y
def filter_data(X, y):
X_train, y_train = [], []
for (example, label) in zip(X, y):
if label == positive_class:
X_train.append(example)
y_train.append(1.0)
elif label == negative_class:
X_train.append(example)
y_train.append(0.0)
X_train = np.array(X_train)
y_train = np.array(y_train)
return X_train, y_train
def cifar_update(host, app, uid, x, y, print_result=False):
url = "http://%s:1337/%s/update" % (host, app)
req_json = json.dumps({
'uid': uid,
'input': list(x),
'label': float(y),
# These updates aren't coming from predictions made by a particular
# model, so we can ignore the model name and version fields.
'model_name': 'NA',
'model_version': 1
})
headers = {'Content-type': 'application/json'}
start = datetime.now()
r = requests.post(url, headers=headers, data=req_json)
end = datetime.now()
latency = (end - start).total_seconds() * 1000.0
if print_result:
print("'%s', %f ms" % (r.text, latency))
def parse_pred(p):
json_prediction = json.loads(p)
if PREDICTION_RESPONSE_KEY_OUTPUT in json_prediction:
# Prediction was successful, return parsed data
qid = int(json_prediction[PREDICTION_RESPONSE_KEY_QUERY_ID])
pred = int(json_prediction[PREDICTION_RESPONSE_KEY_OUTPUT])
return qid, pred
elif PREDICTION_ERROR_RESPONSE_KEY_ERROR in json_prediction:
# Prediction is an error, log the issue
error_name = str(json_prediction[PREDICTION_ERROR_RESPONSE_KEY_ERROR])
print(error_name)
error_cause = str(json_prediction[PREDICTION_ERROR_RESPONSE_KEY_CAUSE])
print("Error executing prediction!")
print("{}: {}".format(error_name, error_cause))
return None
def cifar_prediction(host, app, uid, x):
url = "http://%s/%s/predict" % (host, app)
req_json = json.dumps({'uid': uid, 'input': list(x)})
headers = {'Content-type': 'application/json'}
start = datetime.now()
r = requests.post(url, headers=headers, data=req_json)
end = datetime.now()
latency = (end - start).total_seconds() * 1000.0
parsed_prediction = parse_pred(r.text)
if parsed_prediction:
qid, pred = parsed_prediction
if pred == -1.0:
pred = 0.0
assert pred == 1.0 or pred == 0.0
return (pred, latency)
else:
return None
def run_iteration(host, app, uid, test_x, test_y):
correct = 0
false_pos = 0
false_neg = 0
latencies = []
true_pos = 0
true_neg = 0
total = 100
for i in range(total):
example_num = np.random.randint(0, len(test_y))
correct_y = float(test_y[example_num])
prediction = cifar_prediction(host, app, uid, test_x[example_num])
if not prediction:
continue
pred_y, latency = prediction
if correct_y == pred_y:
if correct_y == 0:
true_neg += 1
elif correct_y == 1:
true_pos += 1
correct += 1
elif correct_y == 0 and pred_y == 1:
false_pos += 1
elif correct_y == 1 and pred_y == 0:
false_neg += 1
else:
print("predicted: {p}, correct: {c}".format(p=pred_y, c=correct_y))
latencies.append(latency)
total = float(total)
return (float(correct) / total, float(false_pos) / total,
float(false_neg) / total, float(true_pos) / total,
float(true_neg) / total, np.mean(latencies))
def run_serving_workload(host, app, test_x, test_y):
fig, (ax_acc) = plt.subplots(1, 1, sharex=True)
ax_acc.set_ylabel("application accuracy")
ax_acc.set_xlabel("iterations")
ax_acc.set_ylim(0, 1.0)
xs = []
accs = []
lats = []
j = 0
uid = DEMO_UID
while True:
correct, fp, fn, tp, tn, mean_lat, = run_iteration(
host, app, uid, test_x, test_y)
xs.append(j)
accs.append(correct)
lats.append(mean_lat)
j += 1
ax_acc.set_xlim(0, j + 1)
ax_acc.plot(xs, accs, 'b')
fig.tight_layout()
fig.canvas.draw()
def run_serving_workload_show_latency(host, app, test_x, test_y):
fig, (ax_acc, ax_lat) = plt.subplots(2, 1, sharex=True)
ax_acc.set_ylabel("accuracy")
ax_lat.set_xlabel("time")
ax_lat.set_ylabel("latency")
ax_acc.set_ylim(0, 1.0)
xs = []
accs = []
lats = []
j = 0
uid = DEMO_UID
while True:
correct, fp, fn, tp, tn, mean_lat, = run_iteration(
host, app, uid, test_x, test_y)
xs.append(j)
accs.append(correct)
lats.append(mean_lat)
j += 1
ax_acc.set_xlim(0, j + 1)
ax_lat.set_xlim(0, j + 1)
ax_acc.plot(xs, accs, 'b')
ax_lat.plot(xs, lats, 'r')
ax_lat.set_ylim(0, 300)
fig.canvas.draw()
print(("Accuracy: {cor}, false positives: {fp}, "
"false negatives: {fn}, true positives: {tp}, "
"true negatives: {tn}").format(
cor=correct, fp=fp, fn=fn, tp=tp, tn=tn))
print("Mean latency: {lat} ms".format(lat=mean_lat))
def enable_feedback(host, app, test_x, test_y, num_updates):
uid = DEMO_UID
for i in range(num_updates):
example_num = np.random.randint(0, len(test_y))
cifar_update(host, app, uid, test_x[example_num],
float(test_y[example_num]))
|
[
"numpy.mean",
"json.loads",
"requests.post",
"numpy.sqrt",
"pandas.read_csv",
"os.path.join",
"numpy.array",
"datetime.datetime.now",
"matplotlib.pyplot.subplots",
"numpy.var"
] |
[((883, 981), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {'nrows': 'num_rows', 'ncols': 'imgs_per_row', 'figsize': '(1.5 * imgs_per_row, 1.5 * num_rows)'}), '(nrows=num_rows, ncols=imgs_per_row, figsize=(1.5 *\n imgs_per_row, 1.5 * num_rows))\n', (895, 981), True, 'import matplotlib.pyplot as plt\n'), ((1542, 1586), 'os.path.join', 'os.path.join', (['cifar_location', 'cifar_filename'], {}), '(cifar_location, cifar_filename)\n', (1554, 1586), False, 'import os\n'), ((1640, 1685), 'pandas.read_csv', 'pd.read_csv', (['cifar_path'], {'sep': '""","""', 'header': 'None'}), "(cifar_path, sep=',', header=None)\n", (1651, 1685), True, 'import pandas as pd\n'), ((2310, 2327), 'numpy.array', 'np.array', (['X_train'], {}), '(X_train)\n', (2318, 2327), True, 'import numpy as np\n'), ((2342, 2359), 'numpy.array', 'np.array', (['y_train'], {}), '(y_train)\n', (2350, 2359), True, 'import numpy as np\n'), ((2872, 2886), 'datetime.datetime.now', 'datetime.now', ([], {}), '()\n', (2884, 2886), False, 'from datetime import datetime\n'), ((2895, 2945), 'requests.post', 'requests.post', (['url'], {'headers': 'headers', 'data': 'req_json'}), '(url, headers=headers, data=req_json)\n', (2908, 2945), False, 'import requests\n'), ((2956, 2970), 'datetime.datetime.now', 'datetime.now', ([], {}), '()\n', (2968, 2970), False, 'from datetime import datetime\n'), ((3137, 3150), 'json.loads', 'json.loads', (['p'], {}), '(p)\n', (3147, 3150), False, 'import json\n'), ((4058, 4072), 'datetime.datetime.now', 'datetime.now', ([], {}), '()\n', (4070, 4072), False, 'from datetime import datetime\n'), ((4081, 4131), 'requests.post', 'requests.post', (['url'], {'headers': 'headers', 'data': 'req_json'}), '(url, headers=headers, data=req_json)\n', (4094, 4131), False, 'import requests\n'), ((4142, 4156), 'datetime.datetime.now', 'datetime.now', ([], {}), '()\n', (4154, 4156), False, 'from datetime import datetime\n'), ((5667, 5698), 'matplotlib.pyplot.subplots', 'plt.subplots', (['(1)', '(1)'], {'sharex': '(True)'}), '(1, 1, sharex=True)\n', (5679, 5698), True, 'import matplotlib.pyplot as plt\n'), ((6311, 6342), 'matplotlib.pyplot.subplots', 'plt.subplots', (['(2)', '(1)'], {'sharex': '(True)'}), '(2, 1, sharex=True)\n', (6323, 6342), True, 'import matplotlib.pyplot as plt\n'), ((1833, 1848), 'numpy.mean', 'np.mean', (['X.T', '(0)'], {}), '(X.T, 0)\n', (1840, 1848), True, 'import numpy as np\n'), ((1865, 1879), 'numpy.var', 'np.var', (['X.T', '(0)'], {}), '(X.T, 0)\n', (1871, 1879), True, 'import numpy as np\n'), ((5572, 5590), 'numpy.mean', 'np.mean', (['latencies'], {}), '(latencies)\n', (5579, 5590), True, 'import numpy as np\n'), ((1915, 1925), 'numpy.sqrt', 'np.sqrt', (['z'], {}), '(z)\n', (1922, 1925), True, 'import numpy as np\n')]
|
"""
Module contains tools for processing files into DataFrames or other objects
"""
from __future__ import annotations
from collections import abc
import csv
import sys
from textwrap import fill
from typing import Any
import warnings
import numpy as np
import pandas._libs.lib as lib
from pandas._libs.parsers import STR_NA_VALUES
from pandas._typing import (
ArrayLike,
DtypeArg,
FilePathOrBuffer,
StorageOptions,
)
from pandas.errors import (
AbstractMethodError,
ParserWarning,
)
from pandas.util._decorators import (
Appender,
deprecate_nonkeyword_arguments,
)
from pandas.util._validators import validate_bool_kwarg
from pandas.core.dtypes.common import (
is_file_like,
is_float,
is_integer,
is_list_like,
)
from pandas.core import generic
from pandas.core.frame import DataFrame
from pandas.core.indexes.api import RangeIndex
from pandas.io.common import validate_header_arg
from pandas.io.parsers.base_parser import (
ParserBase,
is_index_col,
parser_defaults,
)
from pandas.io.parsers.c_parser_wrapper import CParserWrapper
from pandas.io.parsers.python_parser import (
FixedWidthFieldParser,
PythonParser,
)
_doc_read_csv_and_table = (
r"""
{summary}
Also supports optionally iterating or breaking of the file
into chunks.
Additional help can be found in the online docs for
`IO Tools <https://pandas.pydata.org/pandas-docs/stable/user_guide/io.html>`_.
Parameters
----------
filepath_or_buffer : str, path object or file-like object
Any valid string path is acceptable. The string could be a URL. Valid
URL schemes include http, ftp, s3, gs, and file. For file URLs, a host is
expected. A local file could be: file://localhost/path/to/table.csv.
If you want to pass in a path object, pandas accepts any ``os.PathLike``.
By file-like object, we refer to objects with a ``read()`` method, such as
a file handle (e.g. via builtin ``open`` function) or ``StringIO``.
sep : str, default {_default_sep}
Delimiter to use. If sep is None, the C engine cannot automatically detect
the separator, but the Python parsing engine can, meaning the latter will
be used and automatically detect the separator by Python's builtin sniffer
tool, ``csv.Sniffer``. In addition, separators longer than 1 character and
different from ``'\s+'`` will be interpreted as regular expressions and
will also force the use of the Python parsing engine. Note that regex
delimiters are prone to ignoring quoted data. Regex example: ``'\r\t'``.
delimiter : str, default ``None``
Alias for sep.
header : int, list of int, default 'infer'
Row number(s) to use as the column names, and the start of the
data. Default behavior is to infer the column names: if no names
are passed the behavior is identical to ``header=0`` and column
names are inferred from the first line of the file, if column
names are passed explicitly then the behavior is identical to
``header=None``. Explicitly pass ``header=0`` to be able to
replace existing names. The header can be a list of integers that
specify row locations for a multi-index on the columns
e.g. [0,1,3]. Intervening rows that are not specified will be
skipped (e.g. 2 in this example is skipped). Note that this
parameter ignores commented lines and empty lines if
``skip_blank_lines=True``, so ``header=0`` denotes the first line of
data rather than the first line of the file.
names : array-like, optional
List of column names to use. If the file contains a header row,
then you should explicitly pass ``header=0`` to override the column names.
Duplicates in this list are not allowed.
index_col : int, str, sequence of int / str, or False, default ``None``
Column(s) to use as the row labels of the ``DataFrame``, either given as
string name or column index. If a sequence of int / str is given, a
MultiIndex is used.
Note: ``index_col=False`` can be used to force pandas to *not* use the first
column as the index, e.g. when you have a malformed file with delimiters at
the end of each line.
usecols : list-like or callable, optional
Return a subset of the columns. If list-like, all elements must either
be positional (i.e. integer indices into the document columns) or strings
that correspond to column names provided either by the user in `names` or
inferred from the document header row(s). For example, a valid list-like
`usecols` parameter would be ``[0, 1, 2]`` or ``['foo', 'bar', 'baz']``.
Element order is ignored, so ``usecols=[0, 1]`` is the same as ``[1, 0]``.
To instantiate a DataFrame from ``data`` with element order preserved use
``pd.read_csv(data, usecols=['foo', 'bar'])[['foo', 'bar']]`` for columns
in ``['foo', 'bar']`` order or
``pd.read_csv(data, usecols=['foo', 'bar'])[['bar', 'foo']]``
for ``['bar', 'foo']`` order.
If callable, the callable function will be evaluated against the column
names, returning names where the callable function evaluates to True. An
example of a valid callable argument would be ``lambda x: x.upper() in
['AAA', 'BBB', 'DDD']``. Using this parameter results in much faster
parsing time and lower memory usage.
squeeze : bool, default False
If the parsed data only contains one column then return a Series.
prefix : str, optional
Prefix to add to column numbers when no header, e.g. 'X' for X0, X1, ...
mangle_dupe_cols : bool, default True
Duplicate columns will be specified as 'X', 'X.1', ...'X.N', rather than
'X'...'X'. Passing in False will cause data to be overwritten if there
are duplicate names in the columns.
dtype : Type name or dict of column -> type, optional
Data type for data or columns. E.g. {{'a': np.float64, 'b': np.int32,
'c': 'Int64'}}
Use `str` or `object` together with suitable `na_values` settings
to preserve and not interpret dtype.
If converters are specified, they will be applied INSTEAD
of dtype conversion.
engine : {{'c', 'python'}}, optional
Parser engine to use. The C engine is faster while the python engine is
currently more feature-complete.
converters : dict, optional
Dict of functions for converting values in certain columns. Keys can either
be integers or column labels.
true_values : list, optional
Values to consider as True.
false_values : list, optional
Values to consider as False.
skipinitialspace : bool, default False
Skip spaces after delimiter.
skiprows : list-like, int or callable, optional
Line numbers to skip (0-indexed) or number of lines to skip (int)
at the start of the file.
If callable, the callable function will be evaluated against the row
indices, returning True if the row should be skipped and False otherwise.
An example of a valid callable argument would be ``lambda x: x in [0, 2]``.
skipfooter : int, default 0
Number of lines at bottom of file to skip (Unsupported with engine='c').
nrows : int, optional
Number of rows of file to read. Useful for reading pieces of large files.
na_values : scalar, str, list-like, or dict, optional
Additional strings to recognize as NA/NaN. If dict passed, specific
per-column NA values. By default the following values are interpreted as
NaN: '"""
+ fill("', '".join(sorted(STR_NA_VALUES)), 70, subsequent_indent=" ")
+ """'.
keep_default_na : bool, default True
Whether or not to include the default NaN values when parsing the data.
Depending on whether `na_values` is passed in, the behavior is as follows:
* If `keep_default_na` is True, and `na_values` are specified, `na_values`
is appended to the default NaN values used for parsing.
* If `keep_default_na` is True, and `na_values` are not specified, only
the default NaN values are used for parsing.
* If `keep_default_na` is False, and `na_values` are specified, only
the NaN values specified `na_values` are used for parsing.
* If `keep_default_na` is False, and `na_values` are not specified, no
strings will be parsed as NaN.
Note that if `na_filter` is passed in as False, the `keep_default_na` and
`na_values` parameters will be ignored.
na_filter : bool, default True
Detect missing value markers (empty strings and the value of na_values). In
data without any NAs, passing na_filter=False can improve the performance
of reading a large file.
verbose : bool, default False
Indicate number of NA values placed in non-numeric columns.
skip_blank_lines : bool, default True
If True, skip over blank lines rather than interpreting as NaN values.
parse_dates : bool or list of int or names or list of lists or dict, \
default False
The behavior is as follows:
* boolean. If True -> try parsing the index.
* list of int or names. e.g. If [1, 2, 3] -> try parsing columns 1, 2, 3
each as a separate date column.
* list of lists. e.g. If [[1, 3]] -> combine columns 1 and 3 and parse as
a single date column.
* dict, e.g. {{'foo' : [1, 3]}} -> parse columns 1, 3 as date and call
result 'foo'
If a column or index cannot be represented as an array of datetimes,
say because of an unparsable value or a mixture of timezones, the column
or index will be returned unaltered as an object data type. For
non-standard datetime parsing, use ``pd.to_datetime`` after
``pd.read_csv``. To parse an index or column with a mixture of timezones,
specify ``date_parser`` to be a partially-applied
:func:`pandas.to_datetime` with ``utc=True``. See
:ref:`io.csv.mixed_timezones` for more.
Note: A fast-path exists for iso8601-formatted dates.
infer_datetime_format : bool, default False
If True and `parse_dates` is enabled, pandas will attempt to infer the
format of the datetime strings in the columns, and if it can be inferred,
switch to a faster method of parsing them. In some cases this can increase
the parsing speed by 5-10x.
keep_date_col : bool, default False
If True and `parse_dates` specifies combining multiple columns then
keep the original columns.
date_parser : function, optional
Function to use for converting a sequence of string columns to an array of
datetime instances. The default uses ``dateutil.parser.parser`` to do the
conversion. Pandas will try to call `date_parser` in three different ways,
advancing to the next if an exception occurs: 1) Pass one or more arrays
(as defined by `parse_dates`) as arguments; 2) concatenate (row-wise) the
string values from the columns defined by `parse_dates` into a single array
and pass that; and 3) call `date_parser` once for each row using one or
more strings (corresponding to the columns defined by `parse_dates`) as
arguments.
dayfirst : bool, default False
DD/MM format dates, international and European format.
cache_dates : bool, default True
If True, use a cache of unique, converted dates to apply the datetime
conversion. May produce significant speed-up when parsing duplicate
date strings, especially ones with timezone offsets.
.. versionadded:: 0.25.0
iterator : bool, default False
Return TextFileReader object for iteration or getting chunks with
``get_chunk()``.
.. versionchanged:: 1.2
``TextFileReader`` is a context manager.
chunksize : int, optional
Return TextFileReader object for iteration.
See the `IO Tools docs
<https://pandas.pydata.org/pandas-docs/stable/io.html#io-chunking>`_
for more information on ``iterator`` and ``chunksize``.
.. versionchanged:: 1.2
``TextFileReader`` is a context manager.
compression : {{'infer', 'gzip', 'bz2', 'zip', 'xz', None}}, default 'infer'
For on-the-fly decompression of on-disk data. If 'infer' and
`filepath_or_buffer` is path-like, then detect compression from the
following extensions: '.gz', '.bz2', '.zip', or '.xz' (otherwise no
decompression). If using 'zip', the ZIP file must contain only one data
file to be read in. Set to None for no decompression.
thousands : str, optional
Thousands separator.
decimal : str, default '.'
Character to recognize as decimal point (e.g. use ',' for European data).
lineterminator : str (length 1), optional
Character to break file into lines. Only valid with C parser.
quotechar : str (length 1), optional
The character used to denote the start and end of a quoted item. Quoted
items can include the delimiter and it will be ignored.
quoting : int or csv.QUOTE_* instance, default 0
Control field quoting behavior per ``csv.QUOTE_*`` constants. Use one of
QUOTE_MINIMAL (0), QUOTE_ALL (1), QUOTE_NONNUMERIC (2) or QUOTE_NONE (3).
doublequote : bool, default ``True``
When quotechar is specified and quoting is not ``QUOTE_NONE``, indicate
whether or not to interpret two consecutive quotechar elements INSIDE a
field as a single ``quotechar`` element.
escapechar : str (length 1), optional
One-character string used to escape other characters.
comment : str, optional
Indicates remainder of line should not be parsed. If found at the beginning
of a line, the line will be ignored altogether. This parameter must be a
single character. Like empty lines (as long as ``skip_blank_lines=True``),
fully commented lines are ignored by the parameter `header` but not by
`skiprows`. For example, if ``comment='#'``, parsing
``#empty\\na,b,c\\n1,2,3`` with ``header=0`` will result in 'a,b,c' being
treated as the header.
encoding : str, optional
Encoding to use for UTF when reading/writing (ex. 'utf-8'). `List of Python
standard encodings
<https://docs.python.org/3/library/codecs.html#standard-encodings>`_ .
.. versionchanged:: 1.2
When ``encoding`` is ``None``, ``errors="replace"`` is passed to
``open()``. Otherwise, ``errors="strict"`` is passed to ``open()``.
This behavior was previously only the case for ``engine="python"``.
.. versionchanged:: 1.3.0
``encoding_errors`` is a new argument. ``encoding`` has no longer an
influence on how encoding errors are handled.
encoding_errors : str, optional, default "strict"
How encoding errors are treated. `List of possible values
<https://docs.python.org/3/library/codecs.html#error-handlers>`_ .
.. versionadded:: 1.3.0
dialect : str or csv.Dialect, optional
If provided, this parameter will override values (default or not) for the
following parameters: `delimiter`, `doublequote`, `escapechar`,
`skipinitialspace`, `quotechar`, and `quoting`. If it is necessary to
override values, a ParserWarning will be issued. See csv.Dialect
documentation for more details.
error_bad_lines : bool, default ``None``
Lines with too many fields (e.g. a csv line with too many commas) will by
default cause an exception to be raised, and no DataFrame will be returned.
If False, then these "bad lines" will be dropped from the DataFrame that is
returned.
.. deprecated:: 1.3.0
The ``on_bad_lines`` parameter should be used instead to specify behavior upon
encountering a bad line instead.
warn_bad_lines : bool, default ``None``
If error_bad_lines is False, and warn_bad_lines is True, a warning for each
"bad line" will be output.
.. deprecated:: 1.3.0
The ``on_bad_lines`` parameter should be used instead to specify behavior upon
encountering a bad line instead.
on_bad_lines : {{'error', 'warn', 'skip'}}, default 'error'
Specifies what to do upon encountering a bad line (a line with too many fields).
Allowed values are :
- 'error', raise an Exception when a bad line is encountered.
- 'warn', raise a warning when a bad line is encountered and skip that line.
- 'skip', skip bad lines without raising or warning when they are encountered.
.. versionadded:: 1.3.0
delim_whitespace : bool, default False
Specifies whether or not whitespace (e.g. ``' '`` or ``'\t'``) will be
used as the sep. Equivalent to setting ``sep='\\s+'``. If this option
is set to True, nothing should be passed in for the ``delimiter``
parameter.
low_memory : bool, default True
Internally process the file in chunks, resulting in lower memory use
while parsing, but possibly mixed type inference. To ensure no mixed
types either set False, or specify the type with the `dtype` parameter.
Note that the entire file is read into a single DataFrame regardless,
use the `chunksize` or `iterator` parameter to return the data in chunks.
(Only valid with C parser).
memory_map : bool, default False
If a filepath is provided for `filepath_or_buffer`, map the file object
directly onto memory and access the data directly from there. Using this
option can improve performance because there is no longer any I/O overhead.
float_precision : str, optional
Specifies which converter the C engine should use for floating-point
values. The options are ``None`` or 'high' for the ordinary converter,
'legacy' for the original lower precision pandas converter, and
'round_trip' for the round-trip converter.
.. versionchanged:: 1.2
{storage_options}
.. versionadded:: 1.2
Returns
-------
DataFrame or TextParser
A comma-separated values (csv) file is returned as two-dimensional
data structure with labeled axes.
See Also
--------
DataFrame.to_csv : Write DataFrame to a comma-separated values (csv) file.
read_csv : Read a comma-separated values (csv) file into DataFrame.
read_fwf : Read a table of fixed-width formatted lines into DataFrame.
Examples
--------
>>> pd.{func_name}('data.csv') # doctest: +SKIP
"""
)
_c_parser_defaults = {
"delim_whitespace": False,
"na_filter": True,
"low_memory": True,
"memory_map": False,
"error_bad_lines": None,
"warn_bad_lines": None,
"float_precision": None,
}
_fwf_defaults = {"colspecs": "infer", "infer_nrows": 100, "widths": None}
_c_unsupported = {"skipfooter"}
_python_unsupported = {"low_memory", "float_precision"}
_deprecated_defaults: dict[str, Any] = {"error_bad_lines": None, "warn_bad_lines": None}
_deprecated_args: set[str] = {"error_bad_lines", "warn_bad_lines"}
def validate_integer(name, val, min_val=0):
"""
Checks whether the 'name' parameter for parsing is either
an integer OR float that can SAFELY be cast to an integer
without losing accuracy. Raises a ValueError if that is
not the case.
Parameters
----------
name : str
Parameter name (used for error reporting)
val : int or float
The value to check
min_val : int
Minimum allowed value (val < min_val will result in a ValueError)
"""
msg = f"'{name:s}' must be an integer >={min_val:d}"
if val is not None:
if is_float(val):
if int(val) != val:
raise ValueError(msg)
val = int(val)
elif not (is_integer(val) and val >= min_val):
raise ValueError(msg)
return val
def _validate_names(names):
"""
Raise ValueError if the `names` parameter contains duplicates or has an
invalid data type.
Parameters
----------
names : array-like or None
An array containing a list of the names used for the output DataFrame.
Raises
------
ValueError
If names are not unique or are not ordered (e.g. set).
"""
if names is not None:
if len(names) != len(set(names)):
raise ValueError("Duplicate names are not allowed.")
if not (
is_list_like(names, allow_sets=False) or isinstance(names, abc.KeysView)
):
raise ValueError("Names should be an ordered collection.")
def _read(filepath_or_buffer: FilePathOrBuffer, kwds):
"""Generic reader of line files."""
if kwds.get("date_parser", None) is not None:
if isinstance(kwds["parse_dates"], bool):
kwds["parse_dates"] = True
# Extract some of the arguments (pass chunksize on).
iterator = kwds.get("iterator", False)
chunksize = validate_integer("chunksize", kwds.get("chunksize", None), 1)
nrows = kwds.get("nrows", None)
# Check for duplicates in names.
_validate_names(kwds.get("names", None))
# Create the parser.
parser = TextFileReader(filepath_or_buffer, **kwds)
if chunksize or iterator:
return parser
with parser:
return parser.read(nrows)
@deprecate_nonkeyword_arguments(
version=None, allowed_args=["filepath_or_buffer"], stacklevel=3
)
@Appender(
_doc_read_csv_and_table.format(
func_name="read_csv",
summary="Read a comma-separated values (csv) file into DataFrame.",
_default_sep="','",
storage_options=generic._shared_docs["storage_options"],
)
)
def read_csv(
filepath_or_buffer: FilePathOrBuffer,
sep=lib.no_default,
delimiter=None,
# Column and Index Locations and Names
header="infer",
names=lib.no_default,
index_col=None,
usecols=None,
squeeze=False,
prefix=lib.no_default,
mangle_dupe_cols=True,
# General Parsing Configuration
dtype: DtypeArg | None = None,
engine=None,
converters=None,
true_values=None,
false_values=None,
skipinitialspace=False,
skiprows=None,
skipfooter=0,
nrows=None,
# NA and Missing Data Handling
na_values=None,
keep_default_na=True,
na_filter=True,
verbose=False,
skip_blank_lines=True,
# Datetime Handling
parse_dates=False,
infer_datetime_format=False,
keep_date_col=False,
date_parser=None,
dayfirst=False,
cache_dates=True,
# Iteration
iterator=False,
chunksize=None,
# Quoting, Compression, and File Format
compression="infer",
thousands=None,
decimal: str = ".",
lineterminator=None,
quotechar='"',
quoting=csv.QUOTE_MINIMAL,
doublequote=True,
escapechar=None,
comment=None,
encoding=None,
encoding_errors: str | None = "strict",
dialect=None,
# Error Handling
error_bad_lines=None,
warn_bad_lines=None,
# TODO (2.0): set on_bad_lines to "error".
# See _refine_defaults_read comment for why we do this.
on_bad_lines=None,
# Internal
delim_whitespace=False,
low_memory=_c_parser_defaults["low_memory"],
memory_map=False,
float_precision=None,
storage_options: StorageOptions = None,
):
# locals() should never be modified
kwds = locals().copy()
del kwds["filepath_or_buffer"]
del kwds["sep"]
kwds_defaults = _refine_defaults_read(
dialect,
delimiter,
delim_whitespace,
engine,
sep,
error_bad_lines,
warn_bad_lines,
on_bad_lines,
names,
prefix,
defaults={"delimiter": ","},
)
kwds.update(kwds_defaults)
return _read(filepath_or_buffer, kwds)
@deprecate_nonkeyword_arguments(
version=None, allowed_args=["filepath_or_buffer"], stacklevel=3
)
@Appender(
_doc_read_csv_and_table.format(
func_name="read_table",
summary="Read general delimited file into DataFrame.",
_default_sep=r"'\\t' (tab-stop)",
storage_options=generic._shared_docs["storage_options"],
)
)
def read_table(
filepath_or_buffer: FilePathOrBuffer,
sep=lib.no_default,
delimiter=None,
# Column and Index Locations and Names
header="infer",
names=lib.no_default,
index_col=None,
usecols=None,
squeeze=False,
prefix=lib.no_default,
mangle_dupe_cols=True,
# General Parsing Configuration
dtype: DtypeArg | None = None,
engine=None,
converters=None,
true_values=None,
false_values=None,
skipinitialspace=False,
skiprows=None,
skipfooter=0,
nrows=None,
# NA and Missing Data Handling
na_values=None,
keep_default_na=True,
na_filter=True,
verbose=False,
skip_blank_lines=True,
# Datetime Handling
parse_dates=False,
infer_datetime_format=False,
keep_date_col=False,
date_parser=None,
dayfirst=False,
cache_dates=True,
# Iteration
iterator=False,
chunksize=None,
# Quoting, Compression, and File Format
compression="infer",
thousands=None,
decimal: str = ".",
lineterminator=None,
quotechar='"',
quoting=csv.QUOTE_MINIMAL,
doublequote=True,
escapechar=None,
comment=None,
encoding=None,
dialect=None,
# Error Handling
error_bad_lines=None,
warn_bad_lines=None,
# TODO (2.0): set on_bad_lines to "error".
# See _refine_defaults_read comment for why we do this.
on_bad_lines=None,
encoding_errors: str | None = "strict",
# Internal
delim_whitespace=False,
low_memory=_c_parser_defaults["low_memory"],
memory_map=False,
float_precision=None,
):
# locals() should never be modified
kwds = locals().copy()
del kwds["filepath_or_buffer"]
del kwds["sep"]
kwds_defaults = _refine_defaults_read(
dialect,
delimiter,
delim_whitespace,
engine,
sep,
error_bad_lines,
warn_bad_lines,
on_bad_lines,
names,
prefix,
defaults={"delimiter": "\t"},
)
kwds.update(kwds_defaults)
return _read(filepath_or_buffer, kwds)
def read_fwf(
filepath_or_buffer: FilePathOrBuffer,
colspecs="infer",
widths=None,
infer_nrows=100,
**kwds,
):
r"""
Read a table of fixed-width formatted lines into DataFrame.
Also supports optionally iterating or breaking of the file
into chunks.
Additional help can be found in the `online docs for IO Tools
<https://pandas.pydata.org/pandas-docs/stable/user_guide/io.html>`_.
Parameters
----------
filepath_or_buffer : str, path object or file-like object
Any valid string path is acceptable. The string could be a URL. Valid
URL schemes include http, ftp, s3, and file. For file URLs, a host is
expected. A local file could be:
``file://localhost/path/to/table.csv``.
If you want to pass in a path object, pandas accepts any
``os.PathLike``.
By file-like object, we refer to objects with a ``read()`` method,
such as a file handle (e.g. via builtin ``open`` function)
or ``StringIO``.
colspecs : list of tuple (int, int) or 'infer'. optional
A list of tuples giving the extents of the fixed-width
fields of each line as half-open intervals (i.e., [from, to[ ).
String value 'infer' can be used to instruct the parser to try
detecting the column specifications from the first 100 rows of
the data which are not being skipped via skiprows (default='infer').
widths : list of int, optional
A list of field widths which can be used instead of 'colspecs' if
the intervals are contiguous.
infer_nrows : int, default 100
The number of rows to consider when letting the parser determine the
`colspecs`.
**kwds : optional
Optional keyword arguments can be passed to ``TextFileReader``.
Returns
-------
DataFrame or TextParser
A comma-separated values (csv) file is returned as two-dimensional
data structure with labeled axes.
See Also
--------
DataFrame.to_csv : Write DataFrame to a comma-separated values (csv) file.
read_csv : Read a comma-separated values (csv) file into DataFrame.
Examples
--------
>>> pd.read_fwf('data.csv') # doctest: +SKIP
"""
# Check input arguments.
if colspecs is None and widths is None:
raise ValueError("Must specify either colspecs or widths")
elif colspecs not in (None, "infer") and widths is not None:
raise ValueError("You must specify only one of 'widths' and 'colspecs'")
# Compute 'colspecs' from 'widths', if specified.
if widths is not None:
colspecs, col = [], 0
for w in widths:
colspecs.append((col, col + w))
col += w
kwds["colspecs"] = colspecs
kwds["infer_nrows"] = infer_nrows
kwds["engine"] = "python-fwf"
return _read(filepath_or_buffer, kwds)
class TextFileReader(abc.Iterator):
"""
Passed dialect overrides any of the related parser options
"""
def __init__(self, f, engine=None, **kwds):
self.f = f
if engine is not None:
engine_specified = True
else:
engine = "python"
engine_specified = False
self.engine = engine
self._engine_specified = kwds.get("engine_specified", engine_specified)
_validate_skipfooter(kwds)
dialect = _extract_dialect(kwds)
if dialect is not None:
kwds = _merge_with_dialect_properties(dialect, kwds)
if kwds.get("header", "infer") == "infer":
kwds["header"] = 0 if kwds.get("names") is None else None
self.orig_options = kwds
# miscellanea
self._currow = 0
options = self._get_options_with_defaults(engine)
options["storage_options"] = kwds.get("storage_options", None)
self.chunksize = options.pop("chunksize", None)
self.nrows = options.pop("nrows", None)
self.squeeze = options.pop("squeeze", False)
self._check_file_or_buffer(f, engine)
self.options, self.engine = self._clean_options(options, engine)
if "has_index_names" in kwds:
self.options["has_index_names"] = kwds["has_index_names"]
self._engine = self._make_engine(self.engine)
def close(self):
self._engine.close()
def _get_options_with_defaults(self, engine):
kwds = self.orig_options
options = {}
default: object | None
for argname, default in parser_defaults.items():
value = kwds.get(argname, default)
# see gh-12935
if argname == "mangle_dupe_cols" and not value:
raise ValueError("Setting mangle_dupe_cols=False is not supported yet")
else:
options[argname] = value
for argname, default in _c_parser_defaults.items():
if argname in kwds:
value = kwds[argname]
if engine != "c" and value != default:
if "python" in engine and argname not in _python_unsupported:
pass
elif value == _deprecated_defaults.get(argname, default):
pass
else:
raise ValueError(
f"The {repr(argname)} option is not supported with the "
f"{repr(engine)} engine"
)
else:
value = _deprecated_defaults.get(argname, default)
options[argname] = value
if engine == "python-fwf":
for argname, default in _fwf_defaults.items():
options[argname] = kwds.get(argname, default)
return options
def _check_file_or_buffer(self, f, engine):
# see gh-16530
if is_file_like(f) and engine != "c" and not hasattr(f, "__next__"):
# The C engine doesn't need the file-like to have the "__next__"
# attribute. However, the Python engine explicitly calls
# "__next__(...)" when iterating through such an object, meaning it
# needs to have that attribute
raise ValueError(
"The 'python' engine cannot iterate through this file buffer."
)
def _clean_options(self, options, engine):
result = options.copy()
fallback_reason = None
# C engine not supported yet
if engine == "c":
if options["skipfooter"] > 0:
fallback_reason = "the 'c' engine does not support skipfooter"
engine = "python"
sep = options["delimiter"]
delim_whitespace = options["delim_whitespace"]
if sep is None and not delim_whitespace:
if engine == "c":
fallback_reason = (
"the 'c' engine does not support "
"sep=None with delim_whitespace=False"
)
engine = "python"
elif sep is not None and len(sep) > 1:
if engine == "c" and sep == r"\s+":
result["delim_whitespace"] = True
del result["delimiter"]
elif engine not in ("python", "python-fwf"):
# wait until regex engine integrated
fallback_reason = (
"the 'c' engine does not support "
"regex separators (separators > 1 char and "
r"different from '\s+' are interpreted as regex)"
)
engine = "python"
elif delim_whitespace:
if "python" in engine:
result["delimiter"] = r"\s+"
elif sep is not None:
encodeable = True
encoding = sys.getfilesystemencoding() or "utf-8"
try:
if len(sep.encode(encoding)) > 1:
encodeable = False
except UnicodeDecodeError:
encodeable = False
if not encodeable and engine not in ("python", "python-fwf"):
fallback_reason = (
f"the separator encoded in {encoding} "
"is > 1 char long, and the 'c' engine "
"does not support such separators"
)
engine = "python"
quotechar = options["quotechar"]
if quotechar is not None and isinstance(quotechar, (str, bytes)):
if (
len(quotechar) == 1
and ord(quotechar) > 127
and engine not in ("python", "python-fwf")
):
fallback_reason = (
"ord(quotechar) > 127, meaning the "
"quotechar is larger than one byte, "
"and the 'c' engine does not support such quotechars"
)
engine = "python"
if fallback_reason and self._engine_specified:
raise ValueError(fallback_reason)
if engine == "c":
for arg in _c_unsupported:
del result[arg]
if "python" in engine:
for arg in _python_unsupported:
if fallback_reason and result[arg] != _c_parser_defaults[arg]:
raise ValueError(
"Falling back to the 'python' engine because "
f"{fallback_reason}, but this causes {repr(arg)} to be "
"ignored as it is not supported by the 'python' engine."
)
del result[arg]
if fallback_reason:
warnings.warn(
(
"Falling back to the 'python' engine because "
f"{fallback_reason}; you can avoid this warning by specifying "
"engine='python'."
),
ParserWarning,
stacklevel=5,
)
index_col = options["index_col"]
names = options["names"]
converters = options["converters"]
na_values = options["na_values"]
skiprows = options["skiprows"]
validate_header_arg(options["header"])
for arg in _deprecated_args:
parser_default = _c_parser_defaults[arg]
depr_default = _deprecated_defaults[arg]
if result.get(arg, depr_default) != depr_default:
msg = (
f"The {arg} argument has been deprecated and will be "
"removed in a future version.\n\n"
)
warnings.warn(msg, FutureWarning, stacklevel=7)
else:
result[arg] = parser_default
if index_col is True:
raise ValueError("The value of index_col couldn't be 'True'")
if is_index_col(index_col):
if not isinstance(index_col, (list, tuple, np.ndarray)):
index_col = [index_col]
result["index_col"] = index_col
names = list(names) if names is not None else names
# type conversion-related
if converters is not None:
if not isinstance(converters, dict):
raise TypeError(
"Type converters must be a dict or subclass, "
f"input was a {type(converters).__name__}"
)
else:
converters = {}
# Converting values to NA
keep_default_na = options["keep_default_na"]
na_values, na_fvalues = _clean_na_values(na_values, keep_default_na)
# handle skiprows; this is internally handled by the
# c-engine, so only need for python parsers
if engine != "c":
if is_integer(skiprows):
skiprows = list(range(skiprows))
if skiprows is None:
skiprows = set()
elif not callable(skiprows):
skiprows = set(skiprows)
# put stuff back
result["names"] = names
result["converters"] = converters
result["na_values"] = na_values
result["na_fvalues"] = na_fvalues
result["skiprows"] = skiprows
return result, engine
def __next__(self):
try:
return self.get_chunk()
except StopIteration:
self.close()
raise
def _make_engine(self, engine="c"):
mapping: dict[str, type[ParserBase]] = {
"c": CParserWrapper,
"python": PythonParser,
"python-fwf": FixedWidthFieldParser,
}
if engine not in mapping:
raise ValueError(
f"Unknown engine: {engine} (valid options are {mapping.keys()})"
)
# error: Too many arguments for "ParserBase"
return mapping[engine](self.f, **self.options) # type: ignore[call-arg]
def _failover_to_python(self):
raise AbstractMethodError(self)
def read(self, nrows=None):
nrows = validate_integer("nrows", nrows)
index, columns, col_dict = self._engine.read(nrows)
if index is None:
if col_dict:
# Any column is actually fine:
new_rows = len(next(iter(col_dict.values())))
index = RangeIndex(self._currow, self._currow + new_rows)
else:
new_rows = 0
else:
new_rows = len(index)
df = DataFrame(col_dict, columns=columns, index=index)
self._currow += new_rows
if self.squeeze and len(df.columns) == 1:
return df[df.columns[0]].copy()
return df
def get_chunk(self, size=None):
if size is None:
size = self.chunksize
if self.nrows is not None:
if self._currow >= self.nrows:
raise StopIteration
size = min(size, self.nrows - self._currow)
return self.read(nrows=size)
def __enter__(self):
return self
def __exit__(self, exc_type, exc_value, traceback):
self.close()
def TextParser(*args, **kwds):
"""
Converts lists of lists/tuples into DataFrames with proper type inference
and optional (e.g. string to datetime) conversion. Also enables iterating
lazily over chunks of large files
Parameters
----------
data : file-like object or list
delimiter : separator character to use
dialect : str or csv.Dialect instance, optional
Ignored if delimiter is longer than 1 character
names : sequence, default
header : int, default 0
Row to use to parse column labels. Defaults to the first row. Prior
rows will be discarded
index_col : int or list, optional
Column or columns to use as the (possibly hierarchical) index
has_index_names: bool, default False
True if the cols defined in index_col have an index name and are
not in the header.
na_values : scalar, str, list-like, or dict, optional
Additional strings to recognize as NA/NaN.
keep_default_na : bool, default True
thousands : str, optional
Thousands separator
comment : str, optional
Comment out remainder of line
parse_dates : bool, default False
keep_date_col : bool, default False
date_parser : function, optional
skiprows : list of integers
Row numbers to skip
skipfooter : int
Number of line at bottom of file to skip
converters : dict, optional
Dict of functions for converting values in certain columns. Keys can
either be integers or column labels, values are functions that take one
input argument, the cell (not column) content, and return the
transformed content.
encoding : str, optional
Encoding to use for UTF when reading/writing (ex. 'utf-8')
squeeze : bool, default False
returns Series if only one column.
infer_datetime_format: bool, default False
If True and `parse_dates` is True for a column, try to infer the
datetime format based on the first datetime string. If the format
can be inferred, there often will be a large parsing speed-up.
float_precision : str, optional
Specifies which converter the C engine should use for floating-point
values. The options are `None` or `high` for the ordinary converter,
`legacy` for the original lower precision pandas converter, and
`round_trip` for the round-trip converter.
.. versionchanged:: 1.2
"""
kwds["engine"] = "python"
return TextFileReader(*args, **kwds)
def _clean_na_values(na_values, keep_default_na=True):
na_fvalues: set | dict
if na_values is None:
if keep_default_na:
na_values = STR_NA_VALUES
else:
na_values = set()
na_fvalues = set()
elif isinstance(na_values, dict):
old_na_values = na_values.copy()
na_values = {} # Prevent aliasing.
# Convert the values in the na_values dictionary
# into array-likes for further use. This is also
# where we append the default NaN values, provided
# that `keep_default_na=True`.
for k, v in old_na_values.items():
if not is_list_like(v):
v = [v]
if keep_default_na:
v = set(v) | STR_NA_VALUES
na_values[k] = v
na_fvalues = {k: _floatify_na_values(v) for k, v in na_values.items()}
else:
if not is_list_like(na_values):
na_values = [na_values]
na_values = _stringify_na_values(na_values)
if keep_default_na:
na_values = na_values | STR_NA_VALUES
na_fvalues = _floatify_na_values(na_values)
return na_values, na_fvalues
def _floatify_na_values(na_values):
# create float versions of the na_values
result = set()
for v in na_values:
try:
v = float(v)
if not np.isnan(v):
result.add(v)
except (TypeError, ValueError, OverflowError):
pass
return result
def _stringify_na_values(na_values):
""" return a stringified and numeric for these values """
result: list[int | str | float] = []
for x in na_values:
result.append(str(x))
result.append(x)
try:
v = float(x)
# we are like 999 here
if v == int(v):
v = int(v)
result.append(f"{v}.0")
result.append(str(v))
result.append(v)
except (TypeError, ValueError, OverflowError):
pass
try:
result.append(int(x))
except (TypeError, ValueError, OverflowError):
pass
return set(result)
def _refine_defaults_read(
dialect: str | csv.Dialect,
delimiter: str | object,
delim_whitespace: bool,
engine: str,
sep: str | object,
error_bad_lines: bool | None,
warn_bad_lines: bool | None,
on_bad_lines: str | None,
names: ArrayLike | None | object,
prefix: str | None | object,
defaults: dict[str, Any],
):
"""Validate/refine default values of input parameters of read_csv, read_table.
Parameters
----------
dialect : str or csv.Dialect
If provided, this parameter will override values (default or not) for the
following parameters: `delimiter`, `doublequote`, `escapechar`,
`skipinitialspace`, `quotechar`, and `quoting`. If it is necessary to
override values, a ParserWarning will be issued. See csv.Dialect
documentation for more details.
delimiter : str or object
Alias for sep.
delim_whitespace : bool
Specifies whether or not whitespace (e.g. ``' '`` or ``'\t'``) will be
used as the sep. Equivalent to setting ``sep='\\s+'``. If this option
is set to True, nothing should be passed in for the ``delimiter``
parameter.
engine : {{'c', 'python'}}
Parser engine to use. The C engine is faster while the python engine is
currently more feature-complete.
sep : str or object
A delimiter provided by the user (str) or a sentinel value, i.e.
pandas._libs.lib.no_default.
error_bad_lines : str or None
Whether to error on a bad line or not.
warn_bad_lines : str or None
Whether to warn on a bad line or not.
on_bad_lines : str or None
An option for handling bad lines or a sentinel value(None).
names : array-like, optional
List of column names to use. If the file contains a header row,
then you should explicitly pass ``header=0`` to override the column names.
Duplicates in this list are not allowed.
prefix : str, optional
Prefix to add to column numbers when no header, e.g. 'X' for X0, X1, ...
defaults: dict
Default values of input parameters.
Returns
-------
kwds : dict
Input parameters with correct values.
Raises
------
ValueError :
If a delimiter was specified with ``sep`` (or ``delimiter``) and
``delim_whitespace=True``.
If on_bad_lines is specified(not ``None``) and ``error_bad_lines``/
``warn_bad_lines`` is True.
"""
# fix types for sep, delimiter to Union(str, Any)
delim_default = defaults["delimiter"]
kwds: dict[str, Any] = {}
# gh-23761
#
# When a dialect is passed, it overrides any of the overlapping
# parameters passed in directly. We don't want to warn if the
# default parameters were passed in (since it probably means
# that the user didn't pass them in explicitly in the first place).
#
# "delimiter" is the annoying corner case because we alias it to
# "sep" before doing comparison to the dialect values later on.
# Thus, we need a flag to indicate that we need to "override"
# the comparison to dialect values by checking if default values
# for BOTH "delimiter" and "sep" were provided.
if dialect is not None:
kwds["sep_override"] = delimiter is None and (
sep is lib.no_default or sep == delim_default
)
if delimiter and (sep is not lib.no_default):
raise ValueError("Specified a sep and a delimiter; you can only specify one.")
if names is not lib.no_default and prefix is not lib.no_default:
raise ValueError("Specified named and prefix; you can only specify one.")
kwds["names"] = None if names is lib.no_default else names
kwds["prefix"] = None if prefix is lib.no_default else prefix
# Alias sep -> delimiter.
if delimiter is None:
delimiter = sep
if delim_whitespace and (delimiter is not lib.no_default):
raise ValueError(
"Specified a delimiter with both sep and "
"delim_whitespace=True; you can only specify one."
)
if delimiter is lib.no_default:
# assign default separator value
kwds["delimiter"] = delim_default
else:
kwds["delimiter"] = delimiter
if engine is not None:
kwds["engine_specified"] = True
else:
kwds["engine"] = "c"
kwds["engine_specified"] = False
# Ensure that on_bad_lines and error_bad_lines/warn_bad_lines
# aren't specified at the same time. If so, raise. Otherwise,
# alias on_bad_lines to "error" if error/warn_bad_lines not set
# and on_bad_lines is not set. on_bad_lines is defaulted to None
# so we can tell if it is set (this is why this hack exists).
if on_bad_lines is not None:
if error_bad_lines is not None or warn_bad_lines is not None:
raise ValueError(
"Both on_bad_lines and error_bad_lines/warn_bad_lines are set. "
"Please only set on_bad_lines."
)
if on_bad_lines == "error":
kwds["on_bad_lines"] = ParserBase.BadLineHandleMethod.ERROR
elif on_bad_lines == "warn":
kwds["on_bad_lines"] = ParserBase.BadLineHandleMethod.WARN
elif on_bad_lines == "skip":
kwds["on_bad_lines"] = ParserBase.BadLineHandleMethod.SKIP
else:
raise ValueError(f"Argument {on_bad_lines} is invalid for on_bad_lines")
else:
if error_bad_lines is not None:
# Must check is_bool, because other stuff(e.g. non-empty lists) eval to true
validate_bool_kwarg(error_bad_lines, "error_bad_lines")
if error_bad_lines:
kwds["on_bad_lines"] = ParserBase.BadLineHandleMethod.ERROR
else:
if warn_bad_lines is not None:
# This is the case where error_bad_lines is False
# We can only warn/skip if error_bad_lines is False
# None doesn't work because backwards-compatibility reasons
validate_bool_kwarg(warn_bad_lines, "warn_bad_lines")
if warn_bad_lines:
kwds["on_bad_lines"] = ParserBase.BadLineHandleMethod.WARN
else:
kwds["on_bad_lines"] = ParserBase.BadLineHandleMethod.SKIP
else:
# Backwards compat, when only error_bad_lines = false, we warn
kwds["on_bad_lines"] = ParserBase.BadLineHandleMethod.WARN
else:
# Everything None -> Error
kwds["on_bad_lines"] = ParserBase.BadLineHandleMethod.ERROR
return kwds
def _extract_dialect(kwds: dict[str, Any]) -> csv.Dialect | None:
"""
Extract concrete csv dialect instance.
Returns
-------
csv.Dialect or None
"""
if kwds.get("dialect") is None:
return None
dialect = kwds["dialect"]
if dialect in csv.list_dialects():
dialect = csv.get_dialect(dialect)
_validate_dialect(dialect)
return dialect
MANDATORY_DIALECT_ATTRS = (
"delimiter",
"doublequote",
"escapechar",
"skipinitialspace",
"quotechar",
"quoting",
)
def _validate_dialect(dialect: csv.Dialect) -> None:
"""
Validate csv dialect instance.
Raises
------
ValueError
If incorrect dialect is provided.
"""
for param in MANDATORY_DIALECT_ATTRS:
if not hasattr(dialect, param):
raise ValueError(f"Invalid dialect {dialect} provided")
def _merge_with_dialect_properties(
dialect: csv.Dialect,
defaults: dict[str, Any],
) -> dict[str, Any]:
"""
Merge default kwargs in TextFileReader with dialect parameters.
Parameters
----------
dialect : csv.Dialect
Concrete csv dialect. See csv.Dialect documentation for more details.
defaults : dict
Keyword arguments passed to TextFileReader.
Returns
-------
kwds : dict
Updated keyword arguments, merged with dialect parameters.
"""
kwds = defaults.copy()
for param in MANDATORY_DIALECT_ATTRS:
dialect_val = getattr(dialect, param)
parser_default = parser_defaults[param]
provided = kwds.get(param, parser_default)
# Messages for conflicting values between the dialect
# instance and the actual parameters provided.
conflict_msgs = []
# Don't warn if the default parameter was passed in,
# even if it conflicts with the dialect (gh-23761).
if provided != parser_default and provided != dialect_val:
msg = (
f"Conflicting values for '{param}': '{provided}' was "
f"provided, but the dialect specifies '{dialect_val}'. "
"Using the dialect-specified value."
)
# Annoying corner case for not warning about
# conflicts between dialect and delimiter parameter.
# Refer to the outer "_read_" function for more info.
if not (param == "delimiter" and kwds.pop("sep_override", False)):
conflict_msgs.append(msg)
if conflict_msgs:
warnings.warn("\n\n".join(conflict_msgs), ParserWarning, stacklevel=2)
kwds[param] = dialect_val
return kwds
def _validate_skipfooter(kwds: dict[str, Any]) -> None:
"""
Check whether skipfooter is compatible with other kwargs in TextFileReader.
Parameters
----------
kwds : dict
Keyword arguments passed to TextFileReader.
Raises
------
ValueError
If skipfooter is not compatible with other parameters.
"""
if kwds.get("skipfooter"):
if kwds.get("iterator") or kwds.get("chunksize"):
raise ValueError("'skipfooter' not supported for iteration")
if kwds.get("nrows"):
raise ValueError("'skipfooter' not supported with 'nrows'")
|
[
"csv.get_dialect",
"pandas.errors.AbstractMethodError",
"pandas.io.common.validate_header_arg",
"pandas.core.dtypes.common.is_file_like",
"pandas.core.dtypes.common.is_list_like",
"sys.getfilesystemencoding",
"pandas.io.parsers.base_parser.is_index_col",
"pandas.core.dtypes.common.is_float",
"csv.list_dialects",
"pandas.io.parsers.base_parser.parser_defaults.items",
"pandas.util._decorators.deprecate_nonkeyword_arguments",
"pandas.core.dtypes.common.is_integer",
"pandas.util._validators.validate_bool_kwarg",
"numpy.isnan",
"warnings.warn",
"pandas.core.frame.DataFrame",
"pandas.core.indexes.api.RangeIndex"
] |
[((20583, 20683), 'pandas.util._decorators.deprecate_nonkeyword_arguments', 'deprecate_nonkeyword_arguments', ([], {'version': 'None', 'allowed_args': "['filepath_or_buffer']", 'stacklevel': '(3)'}), "(version=None, allowed_args=[\n 'filepath_or_buffer'], stacklevel=3)\n", (20613, 20683), False, 'from pandas.util._decorators import Appender, deprecate_nonkeyword_arguments\n'), ((23045, 23145), 'pandas.util._decorators.deprecate_nonkeyword_arguments', 'deprecate_nonkeyword_arguments', ([], {'version': 'None', 'allowed_args': "['filepath_or_buffer']", 'stacklevel': '(3)'}), "(version=None, allowed_args=[\n 'filepath_or_buffer'], stacklevel=3)\n", (23075, 23145), False, 'from pandas.util._decorators import Appender, deprecate_nonkeyword_arguments\n'), ((18938, 18951), 'pandas.core.dtypes.common.is_float', 'is_float', (['val'], {}), '(val)\n', (18946, 18951), False, 'from pandas.core.dtypes.common import is_file_like, is_float, is_integer, is_list_like\n'), ((29966, 29989), 'pandas.io.parsers.base_parser.parser_defaults.items', 'parser_defaults.items', ([], {}), '()\n', (29987, 29989), False, 'from pandas.io.parsers.base_parser import ParserBase, is_index_col, parser_defaults\n'), ((35587, 35625), 'pandas.io.common.validate_header_arg', 'validate_header_arg', (["options['header']"], {}), "(options['header'])\n", (35606, 35625), False, 'from pandas.io.common import validate_header_arg\n'), ((36247, 36270), 'pandas.io.parsers.base_parser.is_index_col', 'is_index_col', (['index_col'], {}), '(index_col)\n', (36259, 36270), False, 'from pandas.io.parsers.base_parser import ParserBase, is_index_col, parser_defaults\n'), ((38322, 38347), 'pandas.errors.AbstractMethodError', 'AbstractMethodError', (['self'], {}), '(self)\n', (38341, 38347), False, 'from pandas.errors import AbstractMethodError, ParserWarning\n'), ((38834, 38883), 'pandas.core.frame.DataFrame', 'DataFrame', (['col_dict'], {'columns': 'columns', 'index': 'index'}), '(col_dict, columns=columns, index=index)\n', (38843, 38883), False, 'from pandas.core.frame import DataFrame\n'), ((51130, 51149), 'csv.list_dialects', 'csv.list_dialects', ([], {}), '()\n', (51147, 51149), False, 'import csv\n'), ((51169, 51193), 'csv.get_dialect', 'csv.get_dialect', (['dialect'], {}), '(dialect)\n', (51184, 51193), False, 'import csv\n'), ((31296, 31311), 'pandas.core.dtypes.common.is_file_like', 'is_file_like', (['f'], {}), '(f)\n', (31308, 31311), False, 'from pandas.core.dtypes.common import is_file_like, is_float, is_integer, is_list_like\n'), ((35063, 35240), 'warnings.warn', 'warnings.warn', (['f"""Falling back to the \'python\' engine because {fallback_reason}; you can avoid this warning by specifying engine=\'python\'."""', 'ParserWarning'], {'stacklevel': '(5)'}), '(\n f"Falling back to the \'python\' engine because {fallback_reason}; you can avoid this warning by specifying engine=\'python\'."\n , ParserWarning, stacklevel=5)\n', (35076, 35240), False, 'import warnings\n'), ((37144, 37164), 'pandas.core.dtypes.common.is_integer', 'is_integer', (['skiprows'], {}), '(skiprows)\n', (37154, 37164), False, 'from pandas.core.dtypes.common import is_file_like, is_float, is_integer, is_list_like\n'), ((49767, 49822), 'pandas.util._validators.validate_bool_kwarg', 'validate_bool_kwarg', (['error_bad_lines', '"""error_bad_lines"""'], {}), "(error_bad_lines, 'error_bad_lines')\n", (49786, 49822), False, 'from pandas.util._validators import validate_bool_kwarg\n'), ((19704, 19741), 'pandas.core.dtypes.common.is_list_like', 'is_list_like', (['names'], {'allow_sets': '(False)'}), '(names, allow_sets=False)\n', (19716, 19741), False, 'from pandas.core.dtypes.common import is_file_like, is_float, is_integer, is_list_like\n'), ((36020, 36067), 'warnings.warn', 'warnings.warn', (['msg', 'FutureWarning'], {'stacklevel': '(7)'}), '(msg, FutureWarning, stacklevel=7)\n', (36033, 36067), False, 'import warnings\n'), ((38675, 38724), 'pandas.core.indexes.api.RangeIndex', 'RangeIndex', (['self._currow', '(self._currow + new_rows)'], {}), '(self._currow, self._currow + new_rows)\n', (38685, 38724), False, 'from pandas.core.indexes.api import RangeIndex\n'), ((42891, 42914), 'pandas.core.dtypes.common.is_list_like', 'is_list_like', (['na_values'], {}), '(na_values)\n', (42903, 42914), False, 'from pandas.core.dtypes.common import is_file_like, is_float, is_integer, is_list_like\n'), ((43352, 43363), 'numpy.isnan', 'np.isnan', (['v'], {}), '(v)\n', (43360, 43363), True, 'import numpy as np\n'), ((19068, 19083), 'pandas.core.dtypes.common.is_integer', 'is_integer', (['val'], {}), '(val)\n', (19078, 19083), False, 'from pandas.core.dtypes.common import is_file_like, is_float, is_integer, is_list_like\n'), ((42640, 42655), 'pandas.core.dtypes.common.is_list_like', 'is_list_like', (['v'], {}), '(v)\n', (42652, 42655), False, 'from pandas.core.dtypes.common import is_file_like, is_float, is_integer, is_list_like\n'), ((50238, 50291), 'pandas.util._validators.validate_bool_kwarg', 'validate_bool_kwarg', (['warn_bad_lines', '"""warn_bad_lines"""'], {}), "(warn_bad_lines, 'warn_bad_lines')\n", (50257, 50291), False, 'from pandas.util._validators import validate_bool_kwarg\n'), ((33225, 33252), 'sys.getfilesystemencoding', 'sys.getfilesystemencoding', ([], {}), '()\n', (33250, 33252), False, 'import sys\n')]
|
# This file is part of the pyMOR project (http://www.pymor.org).
# Copyright 2013-2019 pyMOR developers and contributors. All rights reserved.
# License: BSD 2-Clause License (http://opensource.org/licenses/BSD-2-Clause)
import numpy as np
import scipy.linalg as spla
from pymor.algorithms.arnoldi import arnoldi
from pymor.algorithms.gram_schmidt import gram_schmidt, gram_schmidt_biorth
from pymor.core.interfaces import BasicInterface
from pymor.models.iosys import LTIModel, SecondOrderModel, LinearDelayModel
from pymor.operators.constructions import LincombOperator
from pymor.reductors.basic import LTIPGReductor, SOLTIPGReductor, DelayLTIPGReductor
class GenericBHIReductor(BasicInterface):
r"""Generic bitangential Hermite interpolation reductor.
This is a generic reductor for reducing any linear
:class:`~pymor.models.iosys.InputStateOutputModel` with
the transfer function which can be written in the generalized
coprime factorization :math:`\mathcal{C}(s) \mathcal{K}(s)^{-1}
\mathcal{B}(s)` as in [BG09]_.
The interpolation here is limited to only up to the first
derivative.
Hence, interpolation points are assumed to be pairwise distinct.
Parameters
----------
fom
Model.
"""
PGReductor = None
def __init__(self, fom):
self.fom = fom
self._product = None
def _B_apply(self, s, V):
raise NotImplementedError
def _C_apply_adjoint(self, s, V):
raise NotImplementedError
def _K_apply_inverse(self, s, V):
raise NotImplementedError
def _K_apply_inverse_adjoint(self, s, V):
raise NotImplementedError
def reduce(self, sigma, b, c, projection='orth'):
"""Bitangential Hermite interpolation.
Parameters
----------
sigma
Interpolation points (closed under conjugation), list of
length `r`.
b
Right tangential directions, |VectorArray| of length `r`
from `self.fom.input_space`.
c
Left tangential directions, |VectorArray| of length `r` from
`self.fom.output_space`.
projection
Projection method:
- `'orth'`: projection matrices are orthogonalized with
respect to the Euclidean inner product
- `'biorth'`: projection matrices are biorthogolized with
respect to the E product
Returns
-------
rom
Reduced model.
"""
r = len(sigma)
assert b in self.fom.input_space and len(b) == r
assert c in self.fom.output_space and len(c) == r
assert projection in ('orth', 'biorth')
# rescale tangential directions (to avoid overflow or underflow)
if b.dim > 1:
b.scal(1 / b.l2_norm())
else:
b = self.fom.input_space.ones(r)
if c.dim > 1:
c.scal(1 / c.l2_norm())
else:
c = self.fom.output_space.ones(r)
# compute projection matrices
self.V = self.fom.state_space.empty(reserve=r)
self.W = self.fom.state_space.empty(reserve=r)
for i in range(r):
if sigma[i].imag == 0:
Bb = self._B_apply(sigma[i].real, b.real[i])
self.V.append(self._K_apply_inverse(sigma[i].real, Bb))
CTc = self._C_apply_adjoint(sigma[i].real, c.real[i])
self.W.append(self._K_apply_inverse_adjoint(sigma[i].real, CTc))
elif sigma[i].imag > 0:
Bb = self._B_apply(sigma[i], b[i])
v = self._K_apply_inverse(sigma[i], Bb)
self.V.append(v.real)
self.V.append(v.imag)
CTc = self._C_apply_adjoint(sigma[i], c[i].conj())
w = self._K_apply_inverse_adjoint(sigma[i], CTc)
self.W.append(w.real)
self.W.append(w.imag)
if projection == 'orth':
self.V = gram_schmidt(self.V, atol=0, rtol=0)
self.W = gram_schmidt(self.W, atol=0, rtol=0)
elif projection == 'biorth':
self.V, self.W = gram_schmidt_biorth(self.V, self.W, product=self._product)
self.pg_reductor = self.PGReductor(self.fom, self.W, self.V, projection == 'biorth')
rom = self.pg_reductor.reduce()
return rom
def reconstruct(self, u):
"""Reconstruct high-dimensional vector from reduced vector `u`."""
return self.RB[:u.dim].lincomb(u.to_numpy())
class LTI_BHIReductor(GenericBHIReductor):
"""Bitangential Hermite interpolation for |LTIModels|.
Parameters
----------
fom
|LTIModel|.
"""
PGReductor = LTIPGReductor
def __init__(self, fom):
assert isinstance(fom, LTIModel)
self.fom = fom
self._product = fom.E
def _B_apply(self, s, V):
return self.fom.B.apply(V)
def _C_apply_adjoint(self, s, V):
return self.fom.C.apply_adjoint(V)
def _K_apply_inverse(self, s, V):
sEmA = s * self.fom.E - self.fom.A
return sEmA.apply_inverse(V)
def _K_apply_inverse_adjoint(self, s, V):
sEmA = s * self.fom.E - self.fom.A
return sEmA.apply_inverse_adjoint(V)
def reduce(self, sigma, b, c, projection='orth', use_arnoldi=False):
"""Bitangential Hermite interpolation.
Parameters
----------
sigma
Interpolation points (closed under conjugation), list of
length `r`.
b
Right tangential directions, |VectorArray| of length `r`
from `self.fom.input_space`.
c
Left tangential directions, |VectorArray| of length `r` from
`self.fom.output_space`.
projection
Projection method:
- `'orth'`: projection matrices are orthogonalized with
respect to the Euclidean inner product
- `'biorth'`: projection matrices are biorthogolized with
respect to the E product
use_arnoldi
Should the Arnoldi process be used for rational
interpolation. Available only for SISO systems. Otherwise,
it is ignored.
Returns
-------
rom
Reduced model.
"""
if use_arnoldi and self.fom.input_dim == 1 and self.fom.output_dim == 1:
return self.reduce_arnoldi(sigma, b, c)
else:
return super().reduce(sigma, b, c, projection=projection)
def reduce_arnoldi(self, sigma, b, c):
"""Bitangential Hermite interpolation for SISO |LTIModels|.
Parameters
----------
sigma
Interpolation points (closed under conjugation), list of
length `r`.
b
Right tangential directions, |VectorArray| of length `r`
from `self.fom.B.source`.
c
Left tangential directions, |VectorArray| of length `r` from
`self.fom.C.range`.
Returns
-------
rom
Reduced |LTIModel| model.
"""
fom = self.fom
assert fom.input_dim == 1 and fom.output_dim == 1
r = len(sigma)
assert b in fom.B.source and len(b) == r
assert c in fom.C.range and len(c) == r
self.V = arnoldi(fom.A, fom.E, fom.B, sigma)
self.W = arnoldi(fom.A, fom.E, fom.C, sigma, trans=True)
rom = super(GenericBHIReductor, self).reduce()
return rom
class SO_BHIReductor(GenericBHIReductor):
"""Bitangential Hermite interpolation for second-order systems.
Parameters
----------
fom
:class:`~pymor.models.iosys.SecondOrderModel`.
"""
PGReductor = SOLTIPGReductor
def __init__(self, fom):
assert isinstance(fom, SecondOrderModel)
self.fom = fom
self._product = fom.M
def _B_apply(self, s, V):
return self.fom.B.apply(V)
def _C_apply_adjoint(self, s, V):
x = self.fom.Cp.apply_adjoint(V)
y = self.fom.Cv.apply_adjoint(V)
return x + y * s.conjugate()
def _K_apply_inverse(self, s, V):
s2MpsEpK = s**2 * self.fom.M + s * self.fom.E + self.fom.K
return s2MpsEpK.apply_inverse(V)
def _K_apply_inverse_adjoint(self, s, V):
s2MpsEpK = s**2 * self.fom.M + s * self.fom.E + self.fom.K
return s2MpsEpK.apply_inverse_adjoint(V)
class DelayBHIReductor(GenericBHIReductor):
"""Bitangential Hermite interpolation for delay systems.
Parameters
----------
fom
:class:`~pymor.models.iosys.LinearDelayModel`.
"""
PGReductor = DelayLTIPGReductor
def __init__(self, fom):
assert isinstance(fom, LinearDelayModel)
self.fom = fom
self._product = fom.E
def _B_apply(self, s, V):
return self.fom.B.apply(V)
def _C_apply_adjoint(self, s, V):
return self.fom.C.apply_adjoint(V)
def _K_apply_inverse(self, s, V):
Ks = LincombOperator((self.fom.E, self.fom.A) + self.fom.Ad,
(s, -1) + tuple(-np.exp(-taui * s) for taui in self.fom.tau))
return Ks.apply_inverse(V)
def _K_apply_inverse_adjoint(self, s, V):
Ks = LincombOperator((self.fom.E, self.fom.A) + self.fom.Ad,
(s, -1) + tuple(-np.exp(-taui * s) for taui in self.fom.tau))
return Ks.apply_inverse_adjoint(V)
class TFInterpReductor(BasicInterface):
"""Loewner bitangential Hermite interpolation reductor.
See [BG12]_.
Parameters
----------
fom
Model with `eval_tf` and `eval_dtf` methods.
"""
def __init__(self, fom):
self.fom = fom
def reduce(self, sigma, b, c):
"""Realization-independent tangential Hermite interpolation.
Parameters
----------
sigma
Interpolation points (closed under conjugation), list of
length `r`.
b
Right tangential directions, |NumPy array| of shape
`(fom.input_dim, r)`.
c
Left tangential directions, |NumPy array| of shape
`(fom.output_dim, r)`.
Returns
-------
lti
|LTIModel| interpolating the transfer function of `fom`.
"""
fom = self.fom
r = len(sigma)
assert isinstance(b, np.ndarray) and b.shape == (fom.input_dim, r)
assert isinstance(c, np.ndarray) and c.shape == (fom.output_dim, r)
# rescale tangential directions (to avoid overflow or underflow)
if b.shape[0] > 1:
for i in range(r):
b[:, i] /= spla.norm(b[:, i])
else:
b = np.ones((1, r))
if c.shape[0] > 1:
for i in range(r):
c[:, i] /= spla.norm(c[:, i])
else:
c = np.ones((1, r))
# matrices of the interpolatory LTI system
Er = np.empty((r, r), dtype=complex)
Ar = np.empty((r, r), dtype=complex)
Br = np.empty((r, fom.input_dim), dtype=complex)
Cr = np.empty((fom.output_dim, r), dtype=complex)
Hs = [fom.eval_tf(s) for s in sigma]
dHs = [fom.eval_dtf(s) for s in sigma]
for i in range(r):
for j in range(r):
if i != j:
Er[i, j] = -c[:, i].dot((Hs[i] - Hs[j]).dot(b[:, j])) / (sigma[i] - sigma[j])
Ar[i, j] = -c[:, i].dot((sigma[i] * Hs[i] - sigma[j] * Hs[j])).dot(b[:, j]) / (sigma[i] - sigma[j])
else:
Er[i, i] = -c[:, i].dot(dHs[i].dot(b[:, i]))
Ar[i, i] = -c[:, i].dot((Hs[i] + sigma[i] * dHs[i]).dot(b[:, i]))
Br[i, :] = Hs[i].T.dot(c[:, i])
Cr[:, i] = Hs[i].dot(b[:, i])
# transform the system to have real matrices
T = np.zeros((r, r), dtype=complex)
for i in range(r):
if sigma[i].imag == 0:
T[i, i] = 1
else:
indices = np.nonzero(np.isclose(sigma[i + 1:], sigma[i].conjugate()))[0]
if len(indices) > 0:
j = i + 1 + indices[0]
T[i, i] = 1
T[i, j] = 1
T[j, i] = -1j
T[j, j] = 1j
Er = (T.dot(Er).dot(T.conj().T)).real
Ar = (T.dot(Ar).dot(T.conj().T)).real
Br = (T.dot(Br)).real
Cr = (Cr.dot(T.conj().T)).real
return LTIModel.from_matrices(Ar, Br, Cr, D=None, E=Er, cont_time=fom.cont_time)
|
[
"pymor.algorithms.gram_schmidt.gram_schmidt_biorth",
"pymor.algorithms.gram_schmidt.gram_schmidt",
"numpy.ones",
"numpy.exp",
"numpy.zeros",
"numpy.empty",
"scipy.linalg.norm",
"pymor.algorithms.arnoldi.arnoldi",
"pymor.models.iosys.LTIModel.from_matrices"
] |
[((7315, 7350), 'pymor.algorithms.arnoldi.arnoldi', 'arnoldi', (['fom.A', 'fom.E', 'fom.B', 'sigma'], {}), '(fom.A, fom.E, fom.B, sigma)\n', (7322, 7350), False, 'from pymor.algorithms.arnoldi import arnoldi\n'), ((7368, 7415), 'pymor.algorithms.arnoldi.arnoldi', 'arnoldi', (['fom.A', 'fom.E', 'fom.C', 'sigma'], {'trans': '(True)'}), '(fom.A, fom.E, fom.C, sigma, trans=True)\n', (7375, 7415), False, 'from pymor.algorithms.arnoldi import arnoldi\n'), ((10916, 10947), 'numpy.empty', 'np.empty', (['(r, r)'], {'dtype': 'complex'}), '((r, r), dtype=complex)\n', (10924, 10947), True, 'import numpy as np\n'), ((10961, 10992), 'numpy.empty', 'np.empty', (['(r, r)'], {'dtype': 'complex'}), '((r, r), dtype=complex)\n', (10969, 10992), True, 'import numpy as np\n'), ((11006, 11049), 'numpy.empty', 'np.empty', (['(r, fom.input_dim)'], {'dtype': 'complex'}), '((r, fom.input_dim), dtype=complex)\n', (11014, 11049), True, 'import numpy as np\n'), ((11063, 11107), 'numpy.empty', 'np.empty', (['(fom.output_dim, r)'], {'dtype': 'complex'}), '((fom.output_dim, r), dtype=complex)\n', (11071, 11107), True, 'import numpy as np\n'), ((11830, 11861), 'numpy.zeros', 'np.zeros', (['(r, r)'], {'dtype': 'complex'}), '((r, r), dtype=complex)\n', (11838, 11861), True, 'import numpy as np\n'), ((12447, 12520), 'pymor.models.iosys.LTIModel.from_matrices', 'LTIModel.from_matrices', (['Ar', 'Br', 'Cr'], {'D': 'None', 'E': 'Er', 'cont_time': 'fom.cont_time'}), '(Ar, Br, Cr, D=None, E=Er, cont_time=fom.cont_time)\n', (12469, 12520), False, 'from pymor.models.iosys import LTIModel, SecondOrderModel, LinearDelayModel\n'), ((3982, 4018), 'pymor.algorithms.gram_schmidt.gram_schmidt', 'gram_schmidt', (['self.V'], {'atol': '(0)', 'rtol': '(0)'}), '(self.V, atol=0, rtol=0)\n', (3994, 4018), False, 'from pymor.algorithms.gram_schmidt import gram_schmidt, gram_schmidt_biorth\n'), ((4040, 4076), 'pymor.algorithms.gram_schmidt.gram_schmidt', 'gram_schmidt', (['self.W'], {'atol': '(0)', 'rtol': '(0)'}), '(self.W, atol=0, rtol=0)\n', (4052, 4076), False, 'from pymor.algorithms.gram_schmidt import gram_schmidt, gram_schmidt_biorth\n'), ((10685, 10700), 'numpy.ones', 'np.ones', (['(1, r)'], {}), '((1, r))\n', (10692, 10700), True, 'import numpy as np\n'), ((10835, 10850), 'numpy.ones', 'np.ones', (['(1, r)'], {}), '((1, r))\n', (10842, 10850), True, 'import numpy as np\n'), ((4143, 4201), 'pymor.algorithms.gram_schmidt.gram_schmidt_biorth', 'gram_schmidt_biorth', (['self.V', 'self.W'], {'product': 'self._product'}), '(self.V, self.W, product=self._product)\n', (4162, 4201), False, 'from pymor.algorithms.gram_schmidt import gram_schmidt, gram_schmidt_biorth\n'), ((10636, 10654), 'scipy.linalg.norm', 'spla.norm', (['b[:, i]'], {}), '(b[:, i])\n', (10645, 10654), True, 'import scipy.linalg as spla\n'), ((10786, 10804), 'scipy.linalg.norm', 'spla.norm', (['c[:, i]'], {}), '(c[:, i])\n', (10795, 10804), True, 'import scipy.linalg as spla\n'), ((9085, 9102), 'numpy.exp', 'np.exp', (['(-taui * s)'], {}), '(-taui * s)\n', (9091, 9102), True, 'import numpy as np\n'), ((9327, 9344), 'numpy.exp', 'np.exp', (['(-taui * s)'], {}), '(-taui * s)\n', (9333, 9344), True, 'import numpy as np\n')]
|
#!/usr/bin/env python
import random
import numpy as np
import tensorflow as tf
import cv2
import matplotlib.pyplot as plt
seed = 0
random.seed(seed)
np.random.seed(seed)
tf.random.set_seed(seed)
(x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()
x_train = x_train[..., None]
x_test = x_test[..., None]
datagen = tf.keras.preprocessing.image.ImageDataGenerator(
rotation_range=15,
width_shift_range=0.1,
height_shift_range=0.1,
zoom_range=0.1)
#validation_split=0.2
#flow = datagen.flow(x_train, y_train, batch_size=16, subset="training")
#flow = datagen.flow(x_train, y_train, batch_size=16, subset="validation")
flow = datagen.flow(x_train, y_train, batch_size=16)
plt.figure(figsize=(19.2, 10.8))
for i in range(16):
x, y = flow.next()
for j in range(16):
plt.subplot(16, 16, i*16+j+1)
plt.imshow(x[j, ..., 0])
plt.xticks([]), plt.yticks([]), plt.title(y[j], x=-0.2, y=0.6)
plt.show()
|
[
"matplotlib.pyplot.imshow",
"tensorflow.random.set_seed",
"matplotlib.pyplot.xticks",
"tensorflow.keras.datasets.mnist.load_data",
"tensorflow.keras.preprocessing.image.ImageDataGenerator",
"random.seed",
"matplotlib.pyplot.figure",
"matplotlib.pyplot.yticks",
"numpy.random.seed",
"matplotlib.pyplot.title",
"matplotlib.pyplot.subplot",
"matplotlib.pyplot.show"
] |
[((135, 152), 'random.seed', 'random.seed', (['seed'], {}), '(seed)\n', (146, 152), False, 'import random\n'), ((153, 173), 'numpy.random.seed', 'np.random.seed', (['seed'], {}), '(seed)\n', (167, 173), True, 'import numpy as np\n'), ((174, 198), 'tensorflow.random.set_seed', 'tf.random.set_seed', (['seed'], {}), '(seed)\n', (192, 198), True, 'import tensorflow as tf\n'), ((239, 274), 'tensorflow.keras.datasets.mnist.load_data', 'tf.keras.datasets.mnist.load_data', ([], {}), '()\n', (272, 274), True, 'import tensorflow as tf\n'), ((342, 475), 'tensorflow.keras.preprocessing.image.ImageDataGenerator', 'tf.keras.preprocessing.image.ImageDataGenerator', ([], {'rotation_range': '(15)', 'width_shift_range': '(0.1)', 'height_shift_range': '(0.1)', 'zoom_range': '(0.1)'}), '(rotation_range=15,\n width_shift_range=0.1, height_shift_range=0.1, zoom_range=0.1)\n', (389, 475), True, 'import tensorflow as tf\n'), ((718, 750), 'matplotlib.pyplot.figure', 'plt.figure', ([], {'figsize': '(19.2, 10.8)'}), '(figsize=(19.2, 10.8))\n', (728, 750), True, 'import matplotlib.pyplot as plt\n'), ((960, 970), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (968, 970), True, 'import matplotlib.pyplot as plt\n'), ((826, 861), 'matplotlib.pyplot.subplot', 'plt.subplot', (['(16)', '(16)', '(i * 16 + j + 1)'], {}), '(16, 16, i * 16 + j + 1)\n', (837, 861), True, 'import matplotlib.pyplot as plt\n'), ((864, 888), 'matplotlib.pyplot.imshow', 'plt.imshow', (['x[j, ..., 0]'], {}), '(x[j, ..., 0])\n', (874, 888), True, 'import matplotlib.pyplot as plt\n'), ((897, 911), 'matplotlib.pyplot.xticks', 'plt.xticks', (['[]'], {}), '([])\n', (907, 911), True, 'import matplotlib.pyplot as plt\n'), ((913, 927), 'matplotlib.pyplot.yticks', 'plt.yticks', (['[]'], {}), '([])\n', (923, 927), True, 'import matplotlib.pyplot as plt\n'), ((929, 959), 'matplotlib.pyplot.title', 'plt.title', (['y[j]'], {'x': '(-0.2)', 'y': '(0.6)'}), '(y[j], x=-0.2, y=0.6)\n', (938, 959), True, 'import matplotlib.pyplot as plt\n')]
|
import os, sys
import numpy as np
from sklearn.cluster import KMeans
import matplotlib.pyplot as plt
percentages = [0.01, 0.1, 0.2, 0.4, 0.5, 0.6]
for percentage in percentages:
data = []
save_path = '../logs/SOM_weights_MNIST_noise_{}.npy'.format(percentage)
wts = np.load(save_path).reshape(-1, 784)
print ("============{}============".format(wts.shape))
kmeans = KMeans(n_clusters=10).fit(wts)
centers = kmeans.cluster_centers_
for i in range(2):
for j in range(5):
plt.subplot(2, 5, i*5 + j + 1)
plt.imshow(centers[i*5+j].reshape(28, 28).T)
if (i == 0) and (j == 0): plt.title("MNIST Noise {}".format(percentage))
plt.show()
|
[
"sklearn.cluster.KMeans",
"numpy.load",
"matplotlib.pyplot.subplot",
"matplotlib.pyplot.show"
] |
[((647, 657), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (655, 657), True, 'import matplotlib.pyplot as plt\n'), ((272, 290), 'numpy.load', 'np.load', (['save_path'], {}), '(save_path)\n', (279, 290), True, 'import numpy as np\n'), ((374, 395), 'sklearn.cluster.KMeans', 'KMeans', ([], {'n_clusters': '(10)'}), '(n_clusters=10)\n', (380, 395), False, 'from sklearn.cluster import KMeans\n'), ((489, 521), 'matplotlib.pyplot.subplot', 'plt.subplot', (['(2)', '(5)', '(i * 5 + j + 1)'], {}), '(2, 5, i * 5 + j + 1)\n', (500, 521), True, 'import matplotlib.pyplot as plt\n')]
|
import matplotlib
matplotlib.use('Agg')
from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
from matplotlib.figure import Figure
from matplotlib.pyplot import gcf
from flask import Flask, render_template, request, flash, redirect
import pandas as pd
import librosa
import ffmpeg
import librosa.display
import numpy as np
import matplotlib.pyplot as plt
import io
import os
import base64
from tensorflow.keras.models import Model, load_model
from tensorflow.keras.layers import Flatten, Dense, Dropout
from tensorflow.keras.optimizers import Adam
from tensorflow.keras.applications import MobileNetV2
from keras.preprocessing.image import img_to_array
from PIL import Image
THIS_DIR = os.path.dirname(os.path.realpath(__file__))
BIRD_DATA = os.path.join(THIS_DIR, 'bird_data.xlsx')
def fig2img(fig):
''' Transforms matplotlib figure to image '''
fig.canvas.draw()
w,h = fig.canvas.get_width_height()
buf = np.fromstring(fig.canvas.tostring_argb(), dtype=np.uint8)
buf.shape = (w,h,4)
buf = np.roll(buf,3,axis = 2)
w, h, d = buf.shape
return Image.frombytes("RGB",(w,h),buf.tostring())
def create_spectrogram(file):
''' loads audio file and creates spectrogram '''
signal, sr = librosa.load(file,duration=10)
fig = gcf()
DPI = fig.get_dpi()
fig = plt.figure()
fig.set_size_inches(224/float(DPI),224/float(DPI))
ax = plt.Axes(fig, [0., 0., 1., 1.])
ax.set_axis_off()
fig.add_axes(ax)
S = librosa.feature.melspectrogram(y=signal,sr=sr,
n_fft=1024,
hop_length=1024,
n_mels=128,
htk=True,
fmin=1400,
fmax=sr/2)
librosa.display.specshow(librosa.power_to_db(S**2,ref=np.max), fmin=1400,y_axis='linear')
image = fig2img(fig)
image = img_to_array(image)
image = np.array([image])
return image, fig
def predict(model, image):
''' makes prediction out of the spectrogram '''
net = MobileNetV2(include_top=False,
weights='imagenet',
input_tensor=None,
input_shape=(224,224,3))
x = net.output
x = Flatten()(x)
x = Dropout(0.5)(x)
output_layer = Dense(5, activation='softmax')(x)
loaded_model = Model(inputs=net.input, outputs=output_layer)
loaded_model.load_weights(model)
loaded_model.compile(optimizer=Adam(),
loss='categorical_crossentropy', metrics=['accuracy'])
pred = loaded_model.predict(image)
return pred
def get_bird_data(bird):
df = pd.read_excel(BIRD_DATA)
df = df[df['species']==bird].reset_index(drop=True)
name = df['name'][0]
en_name = df['en_name'][0]
desc = df['desc'][0]
return name, en_name, desc
def create_bird_path(bird):
img_path = '/static/images/'
bird = bird.lower()
img_file = bird + '.jpg'
bird_path = img_path + img_file
return bird_path
def create_result(pred, classes):
''' creates results (bird class and probability) '''
top = np.argsort(pred[0])[:-2:-1]
result = {'bird': '', 'probability': ''}
result['bird'] = classes[top[0]]
result['probability'] = int(round(pred[0][top[0]],2)*100)
return result
|
[
"keras.preprocessing.image.img_to_array",
"numpy.argsort",
"numpy.array",
"tensorflow.keras.layers.Dense",
"pandas.read_excel",
"librosa.load",
"tensorflow.keras.models.Model",
"tensorflow.keras.applications.MobileNetV2",
"matplotlib.use",
"matplotlib.pyplot.gcf",
"tensorflow.keras.layers.Dropout",
"librosa.power_to_db",
"tensorflow.keras.layers.Flatten",
"librosa.feature.melspectrogram",
"numpy.roll",
"matplotlib.pyplot.Axes",
"os.path.join",
"os.path.realpath",
"tensorflow.keras.optimizers.Adam",
"matplotlib.pyplot.figure"
] |
[((18, 39), 'matplotlib.use', 'matplotlib.use', (['"""Agg"""'], {}), "('Agg')\n", (32, 39), False, 'import matplotlib\n'), ((767, 807), 'os.path.join', 'os.path.join', (['THIS_DIR', '"""bird_data.xlsx"""'], {}), "(THIS_DIR, 'bird_data.xlsx')\n", (779, 807), False, 'import os\n'), ((727, 753), 'os.path.realpath', 'os.path.realpath', (['__file__'], {}), '(__file__)\n', (743, 753), False, 'import os\n'), ((1043, 1066), 'numpy.roll', 'np.roll', (['buf', '(3)'], {'axis': '(2)'}), '(buf, 3, axis=2)\n', (1050, 1066), True, 'import numpy as np\n'), ((1247, 1278), 'librosa.load', 'librosa.load', (['file'], {'duration': '(10)'}), '(file, duration=10)\n', (1259, 1278), False, 'import librosa\n'), ((1291, 1296), 'matplotlib.pyplot.gcf', 'gcf', ([], {}), '()\n', (1294, 1296), False, 'from matplotlib.pyplot import gcf\n'), ((1331, 1343), 'matplotlib.pyplot.figure', 'plt.figure', ([], {}), '()\n', (1341, 1343), True, 'import matplotlib.pyplot as plt\n'), ((1413, 1448), 'matplotlib.pyplot.Axes', 'plt.Axes', (['fig', '[0.0, 0.0, 1.0, 1.0]'], {}), '(fig, [0.0, 0.0, 1.0, 1.0])\n', (1421, 1448), True, 'import matplotlib.pyplot as plt\n'), ((1497, 1623), 'librosa.feature.melspectrogram', 'librosa.feature.melspectrogram', ([], {'y': 'signal', 'sr': 'sr', 'n_fft': '(1024)', 'hop_length': '(1024)', 'n_mels': '(128)', 'htk': '(True)', 'fmin': '(1400)', 'fmax': '(sr / 2)'}), '(y=signal, sr=sr, n_fft=1024, hop_length=1024,\n n_mels=128, htk=True, fmin=1400, fmax=sr / 2)\n', (1527, 1623), False, 'import librosa\n'), ((2022, 2041), 'keras.preprocessing.image.img_to_array', 'img_to_array', (['image'], {}), '(image)\n', (2034, 2041), False, 'from keras.preprocessing.image import img_to_array\n'), ((2054, 2071), 'numpy.array', 'np.array', (['[image]'], {}), '([image])\n', (2062, 2071), True, 'import numpy as np\n'), ((2185, 2285), 'tensorflow.keras.applications.MobileNetV2', 'MobileNetV2', ([], {'include_top': '(False)', 'weights': '"""imagenet"""', 'input_tensor': 'None', 'input_shape': '(224, 224, 3)'}), "(include_top=False, weights='imagenet', input_tensor=None,\n input_shape=(224, 224, 3))\n", (2196, 2285), False, 'from tensorflow.keras.applications import MobileNetV2\n'), ((2500, 2545), 'tensorflow.keras.models.Model', 'Model', ([], {'inputs': 'net.input', 'outputs': 'output_layer'}), '(inputs=net.input, outputs=output_layer)\n', (2505, 2545), False, 'from tensorflow.keras.models import Model, load_model\n'), ((2793, 2817), 'pandas.read_excel', 'pd.read_excel', (['BIRD_DATA'], {}), '(BIRD_DATA)\n', (2806, 2817), True, 'import pandas as pd\n'), ((1915, 1954), 'librosa.power_to_db', 'librosa.power_to_db', (['(S ** 2)'], {'ref': 'np.max'}), '(S ** 2, ref=np.max)\n', (1934, 1954), False, 'import librosa\n'), ((2391, 2400), 'tensorflow.keras.layers.Flatten', 'Flatten', ([], {}), '()\n', (2398, 2400), False, 'from tensorflow.keras.layers import Flatten, Dense, Dropout\n'), ((2412, 2424), 'tensorflow.keras.layers.Dropout', 'Dropout', (['(0.5)'], {}), '(0.5)\n', (2419, 2424), False, 'from tensorflow.keras.layers import Flatten, Dense, Dropout\n'), ((2447, 2477), 'tensorflow.keras.layers.Dense', 'Dense', (['(5)'], {'activation': '"""softmax"""'}), "(5, activation='softmax')\n", (2452, 2477), False, 'from tensorflow.keras.layers import Flatten, Dense, Dropout\n'), ((3263, 3282), 'numpy.argsort', 'np.argsort', (['pred[0]'], {}), '(pred[0])\n', (3273, 3282), True, 'import numpy as np\n'), ((2618, 2624), 'tensorflow.keras.optimizers.Adam', 'Adam', ([], {}), '()\n', (2622, 2624), False, 'from tensorflow.keras.optimizers import Adam\n')]
|
import os
from pathlib import Path
import cv2
import numpy as np
import pandas as pd
from pandas import DataFrame
from sklearn.model_selection import train_test_split
def create_info_csv(mvtec_dir: Path) -> DataFrame:
df = pd.DataFrame({})
for data_type in ["train", "test"]:
for p in mvtec_dir.glob(f"*/{data_type}/*/*.png"):
raw_stem = p.stem
defect = p.parents[0].name
data_type = p.parents[1].name
category = p.parents[2].name
df = df.append(
{
"raw_img_path": str(p),
"raw_stem": raw_stem,
"defect": defect,
"data_type": data_type,
"category": category,
},
ignore_index=True,
)
for category in df["category"].unique():
category_df = df.query("data_type=='train' & category==@category")
_, val_index = train_test_split(
category_df.index.tolist(),
train_size=0.8,
test_size=0.2,
random_state=5,
shuffle=True,
)
df.loc[val_index, "data_type"] = "val"
df["stem"] = df.apply(
lambda x: f"{x.category}_{x.data_type}_{x.defect}_{x.raw_stem}",
axis=1,
)
df["raw_mask_path"] = df.apply(
lambda x: f"{mvtec_dir}/{x.category}/ground_truth/{x.defect}/{x.raw_stem}_mask.png",
axis=1,
)
return df
def move_images_and_masks(df: DataFrame) -> None:
os.makedirs("/data/images", exist_ok=True)
os.makedirs("/data/masks", exist_ok=True)
for i in df.index:
raw_img_path, raw_mask_path, stem = df.loc[i, ["raw_img_path", "raw_mask_path", "stem"]]
if os.path.exists(raw_mask_path):
os.rename(raw_mask_path, f"/data/masks/{stem}.png")
else:
# create masks for train images
img = cv2.imread(raw_img_path)
mask = np.zeros(img.shape)
cv2.imwrite(f"/data/masks/{stem}.png", mask)
os.rename(raw_img_path, f"/data/images/{stem}.png")
df.drop(columns=["raw_stem", "raw_img_path", "raw_mask_path"])
df.to_csv("/data/info.csv", index=False)
if __name__ == "__main__":
mvtec_dir = Path("/data/MVTec")
df = create_info_csv(mvtec_dir)
move_images_and_masks(df)
|
[
"os.path.exists",
"cv2.imwrite",
"os.makedirs",
"pathlib.Path",
"os.rename",
"numpy.zeros",
"pandas.DataFrame",
"cv2.imread"
] |
[((231, 247), 'pandas.DataFrame', 'pd.DataFrame', (['{}'], {}), '({})\n', (243, 247), True, 'import pandas as pd\n'), ((1542, 1584), 'os.makedirs', 'os.makedirs', (['"""/data/images"""'], {'exist_ok': '(True)'}), "('/data/images', exist_ok=True)\n", (1553, 1584), False, 'import os\n'), ((1589, 1630), 'os.makedirs', 'os.makedirs', (['"""/data/masks"""'], {'exist_ok': '(True)'}), "('/data/masks', exist_ok=True)\n", (1600, 1630), False, 'import os\n'), ((2276, 2295), 'pathlib.Path', 'Path', (['"""/data/MVTec"""'], {}), "('/data/MVTec')\n", (2280, 2295), False, 'from pathlib import Path\n'), ((1764, 1793), 'os.path.exists', 'os.path.exists', (['raw_mask_path'], {}), '(raw_mask_path)\n', (1778, 1793), False, 'import os\n'), ((2065, 2116), 'os.rename', 'os.rename', (['raw_img_path', 'f"""/data/images/{stem}.png"""'], {}), "(raw_img_path, f'/data/images/{stem}.png')\n", (2074, 2116), False, 'import os\n'), ((1807, 1858), 'os.rename', 'os.rename', (['raw_mask_path', 'f"""/data/masks/{stem}.png"""'], {}), "(raw_mask_path, f'/data/masks/{stem}.png')\n", (1816, 1858), False, 'import os\n'), ((1935, 1959), 'cv2.imread', 'cv2.imread', (['raw_img_path'], {}), '(raw_img_path)\n', (1945, 1959), False, 'import cv2\n'), ((1979, 1998), 'numpy.zeros', 'np.zeros', (['img.shape'], {}), '(img.shape)\n', (1987, 1998), True, 'import numpy as np\n'), ((2011, 2055), 'cv2.imwrite', 'cv2.imwrite', (['f"""/data/masks/{stem}.png"""', 'mask'], {}), "(f'/data/masks/{stem}.png', mask)\n", (2022, 2055), False, 'import cv2\n')]
|
import tensorflow as tf
# numpy 是个科学计算的工具包，这里通过Numpy生成模拟数据
from numpy.random import RandomState
# 训练数据batch的大小
batch_size = 8
# 定义神经网络的参数，这里还是沿用3.4.2 小结中给出的神经网络结构
w1 = tf.Variable(tf.random_normal([2, 3], stddev=1, seed=1))
w2 = tf.Variable(tf.random_normal([3, 1], stddev=1, seed=1))
# 在shape的维度上使用None可以方便使用不打的batch大小，在训练时需要把数据
# 分成比较小的batch，但是在测试时，可以一次性使用全部数据，当数据集比较小时这样比较
# 方便测试，但是数据集比较大时放入一个batch会导致内存溢出
x = tf.placeholder(tf.float32, shape=(None, 2), name="x-input")
y_ = tf.placeholder(tf.float32, shape=(None, 1), name='y-input')
# 定义神经网络向前传播的过程 x w1 w2 两层神经
a = tf.matmul(x, w1)
y = tf.matmul(a, w2)
# 定义损失函数和反向传播的算法
# tf.clip_by_value 因为 log 会产生 none (如 log-3 ), 用它来限定不出现none
# 替代方法 cross_entropy = -tf.reduce_sum(y_*tf.log(y_conv + 1e-10))
cross_entropy = -tf.reduce_mean(y_ * tf.log(tf.clip_by_value(y, 1e-10, 1.0)))
train_step = tf.train.AdamOptimizer(0.001).minimize(cross_entropy)
# 通过随机数生成一个模拟数据集
rdm = RandomState(1)
X = rdm.rand(128, 2)
# Y 为对数据集数据 进行 结果收集分类 和大于1 为1 小于 1为0
# 定义规则来给样本的标签。在这里所有x1 + x2 < 1 的样本都被认为是正样本（比如零件合格）
# 而其他为负样本（比如样本不合格）。和TensorFlow 游乐场中的表示法不大一样的地方是，
# 这里的0表示负样本，1 表示正样本。大部分解决分类问题的神经网络都采用
# 0 和 1 的表示方法
Y = [[int(x1 + x2) < 1] for (x1, x2) in X]
# 创建一个会话运行TensorFlow程序
with tf.Session() as sess:
# 初始化变量
init_op = tf.global_variables_initializer()
sess.run(init_op)
# 在训练之前神经网络参数
print("w1:", sess.run(w1))
print("w2:", sess.run(w2))
print("\n")
'''
训练之前神经网络参数的值
w1: [[-0.81131822 1.48459876 0.06532937]
[-2.44270396 0.0992484 0.59122431]]
w2: [[-0.81131822]
[ 1.48459876]
[ 0.06532937]]
'''
# 设定训练的轮数
STEPS = 5000
for i in range(STEPS):
start = (i * batch_size) % 128
end = (i * batch_size) % 128 + batch_size
# 通过选取样本训练神经网络并更新参数
sess.run(train_step, feed_dict={x: X[start:end], y_: Y[start:end]})
if i % 1000 == 0:
# 每隔一段时间计算在所有数据上的交叉熵并输出
total_cross_entropy = sess.run(cross_entropy, feed_dict={x: X, y_: Y})
print("After %d training steps(s), cross entropy on all data is %g" % (i, total_cross_entropy))
'''
输出结果
After 0 training steps(s), cross entropy on all data is 0.0674925
After 1000 training steps(s), cross entropy on all data is 0.0163385
After 2000 training steps(s), cross entropy on all data is 0.00907547
After 3000 training steps(s), cross entropy on all data is 0.00714436
After 4000 training steps(s), cross entropy on all data is 0.00578471
通过这个结果可以发现随着训练的进行，交叉熵是逐渐减小的。交叉熵越小说明预测的结果和真实的结果差距越小
'''
print("\n")
print("w1:", sess.run(w1))
print("w2:", sess.run(w2))
'''
w1: [[-1.9618274 2.58235407 1.68203783]
[-3.4681716 1.06982327 2.11788988]]
w2: [[-1.8247149 ]
[ 2.68546653]
[ 1.41819501]]
可以发现这两个参数的取值已经发生了编发，这个变化是训练的结果
它使得这个神经网络能根号的拟合提供的训练数据
'''
'''
1、定义神经网络的结构和前向传播的输出结果
2、定义损失函数以及选择反向传播的优化算法
3、生成会话（tf.Session）并且在训练数据上反复运行反向传播优化算法
'''
|
[
"tensorflow.random_normal",
"tensorflow.Session",
"tensorflow.placeholder",
"tensorflow.global_variables_initializer",
"tensorflow.matmul",
"tensorflow.clip_by_value",
"tensorflow.train.AdamOptimizer",
"numpy.random.RandomState"
] |
[((416, 475), 'tensorflow.placeholder', 'tf.placeholder', (['tf.float32'], {'shape': '(None, 2)', 'name': '"""x-input"""'}), "(tf.float32, shape=(None, 2), name='x-input')\n", (430, 475), True, 'import tensorflow as tf\n'), ((481, 540), 'tensorflow.placeholder', 'tf.placeholder', (['tf.float32'], {'shape': '(None, 1)', 'name': '"""y-input"""'}), "(tf.float32, shape=(None, 1), name='y-input')\n", (495, 540), True, 'import tensorflow as tf\n'), ((577, 593), 'tensorflow.matmul', 'tf.matmul', (['x', 'w1'], {}), '(x, w1)\n', (586, 593), True, 'import tensorflow as tf\n'), ((598, 614), 'tensorflow.matmul', 'tf.matmul', (['a', 'w2'], {}), '(a, w2)\n', (607, 614), True, 'import tensorflow as tf\n'), ((927, 941), 'numpy.random.RandomState', 'RandomState', (['(1)'], {}), '(1)\n', (938, 941), False, 'from numpy.random import RandomState\n'), ((182, 224), 'tensorflow.random_normal', 'tf.random_normal', (['[2, 3]'], {'stddev': '(1)', 'seed': '(1)'}), '([2, 3], stddev=1, seed=1)\n', (198, 224), True, 'import tensorflow as tf\n'), ((243, 285), 'tensorflow.random_normal', 'tf.random_normal', (['[3, 1]'], {'stddev': '(1)', 'seed': '(1)'}), '([3, 1], stddev=1, seed=1)\n', (259, 285), True, 'import tensorflow as tf\n'), ((1225, 1237), 'tensorflow.Session', 'tf.Session', ([], {}), '()\n', (1235, 1237), True, 'import tensorflow as tf\n'), ((1273, 1306), 'tensorflow.global_variables_initializer', 'tf.global_variables_initializer', ([], {}), '()\n', (1304, 1306), True, 'import tensorflow as tf\n'), ((849, 878), 'tensorflow.train.AdamOptimizer', 'tf.train.AdamOptimizer', (['(0.001)'], {}), '(0.001)\n', (871, 878), True, 'import tensorflow as tf\n'), ((802, 833), 'tensorflow.clip_by_value', 'tf.clip_by_value', (['y', '(1e-10)', '(1.0)'], {}), '(y, 1e-10, 1.0)\n', (818, 833), True, 'import tensorflow as tf\n')]
|
"""Retokenization helpers
This module provides helpers for projecting span annotations from one tokenization to another.
Notes:
* Code is ported from https://github.com/nyu-mll/jiant/blob/master/jiant/utils/retokenize.py
* Please keep this code as a standalone utility; don't make this module depend on jiant modules.
"""
from typing import Iterable, Sequence, Tuple, Union
from Levenshtein.StringMatcher import StringMatcher
from nltk.tokenize.util import string_span_tokenize
import numpy as np
_DTYPE = np.int32
def _mat_from_blocks_dense(mb, n_chars_src, n_chars_tgt):
M = np.zeros((n_chars_src, n_chars_tgt), dtype=_DTYPE)
for i in range(len(mb)):
b = mb[i] # current block
# Fill in-between this block and last block
if i > 0:
lb = mb[i - 1] # last block
s0 = lb[0] + lb[2] # top
e0 = b[0] # bottom
s1 = lb[1] + lb[2] # left
e1 = b[1] # right
M[s0:e0, s1:e1] = 1
# Fill matching region on diagonal
M[b[0]: b[0] + b[2], b[1]: b[1] + b[2]] = 2 * \
np.identity(b[2], dtype=_DTYPE)
return M
def _mat_from_spans_dense(spans: Sequence[Tuple[int, int]], n_chars: int) -> np.ndarray:
"""Construct a token-to-char matrix from a list of char spans."""
M = np.zeros((len(spans), n_chars), dtype=_DTYPE)
for i, s in enumerate(spans):
M[i, s[0]: s[1]] = 1
return M
def token_to_char(text: str, sep=" ") -> np.ndarray:
"""Takes a string, space tokenizes the string, and returns a mapping from tokens to chars.
Examples:
>>> token_to_char("testing 1, 2, 3")
# produces a (m) token by (M) char matrix:
t e s t i n g 1 , 2 , 3
testing [[1 1 1 1 1 1 1 0 0 0 0 0 0 0 0]
1, [0 0 0 0 0 0 0 0 1 1 0 0 0 0 0]
2, [0 0 0 0 0 0 0 0 0 0 0 1 1 0 0]
3 [0 0 0 0 0 0 0 0 0 0 0 0 0 0 1]]
Args:
text (str): string to tokenize and build the token to char mapping.
Returns:
np.ndarray mapping from (m) tokens to (M) chars.
"""
spans = string_span_tokenize(text, sep=sep)
return _mat_from_spans_dense(tuple(spans), len(text))
def _mat_from_blocks(
mb: Sequence[Tuple[int, int, int]], n_chars_src: int, n_chars_tgt: int
) -> np.ndarray:
"""Construct a char-to-char matrix from a list of matching blocks.
mb is a sequence of (s1, s2, n_char) tuples, where s1 and s2 are the start indices in the
first and second sequence, and n_char is the length of the block.
Args:
mb: list of triples of non-overlapping matching subsequences in source str and target.
n_chars_src (int): number of chars in the source string.
n_chars_tgt (int): number of chars in the target string.
Returns:
np.ndarray adjacency matrix mapping chars in the source str to chars in the target str.
"""
return _mat_from_blocks_dense(mb, n_chars_src, n_chars_tgt)
def char_to_char(source: str, target: str) -> np.ndarray:
"""Find the character adjacency matrix mapping source string chars to target string chars.
Uses StringMatcher from Levenshtein package to find non-overlapping matching subsequences in
input strings. Uses the result to create a character adjacency matrix from source to target.
(https://docs.python.org/2/library/difflib.html#difflib.SequenceMatcher.get_matching_blocks)
Args:
source (str): string of source chars.
target (str): string of target chars.
Returns:
np.ndarray adjacency matrix mapping chars in the source str to chars in the target str.
"""
sm = StringMatcher(seq1=source, seq2=target)
mb = sm.get_matching_blocks()
return _mat_from_blocks(mb, len(source), len(target))
class TokenAligner(object):
"""Align two similiar tokenizations.
Args:
source (Union[Iterable[str], str]): Source text tokens or string.
target (Union[Iterable[str], str]): Target text tokens or string.
Usage:
ta = TokenAligner(source_tokens, target_tokens)
target_span = ta.project_span(*source_span)
Uses Levenshtein distance to align two similar tokenizations, and provide facilities to project
indices and spans from the source tokenization to the target.
Let source contain m tokens and M chars, and target contain n tokens and N chars. The token
alignment is treated as a (sparse) m x n adjacency matrix T representing the bipartite graph
between the source and target tokens.
This is constructed by performing a character-level alignment using Levenshtein distance to
obtain a (M x N) character adjacency matrix C. We then construct token-to-character matricies
U (m x M) and V (n x N) and construct T as:
T = (U C V')
where V' denotes the transpose.
Spans of non-aligned bytes are assumed to contain a many-to-many alignment of all chars in that
range. This can lead to unwanted alignments if, for example, two consecutive tokens are mapped
to escape sequences:
source: ["'s", "["]
target: ["'", "s", "["]
In the above case, "'s'" may be spuriously aligned to "'" while "[" has no character match
with "s" or "[", and so is aligned to both. To make a correct alignment more likely, ensure
that the characters in target correspond as closely as possible to those in source. For example,
the following will align correctly:
source: ["'s", "["]
target: ["'", "s", "["]
"""
def __init__(self, source: Union[Iterable[str], str], target: Union[Iterable[str], str]):
# Coerce source and target to space-delimited string.
if not isinstance(source, str):
source = " ".join(source)
if not isinstance(target, str):
target = " ".join(target)
# (m X M) source token idx to source char idx
self.U = token_to_char(source)
# (n x N) target token idx to target char idx
self.V = token_to_char(target)
# (M x N) source char idx to target char idx
self.C = char_to_char(source, target)
# Token transfer matrix from (m) tokens in source to (n) tokens in the target. Mat value at
# index i, j measures the character overlap btwn the ith source token and jth target token.
self.source_token_idx_to_target_token_idx = self.U.dot(
self.C).dot(self.V.T)
self.source_token_idx_to_target_char_idx = self.U.dot(self.C)
self.source_char_idx_to_target_token_idx = self.C.dot(self.V.T)
def project_token_idxs(self, idxs: Union[int, Sequence[int]]) -> Sequence[int]:
"""Project source token index(s) to target token indices.
Takes a list of token indices in the source token sequence, and returns the corresponding
tokens in the target sequence.
Args:
idxs (Union[int, Sequence[int]]): source token index(s) to get related target indices.
Examples:
>>> source_tokens = ['abc', 'def', 'ghi', 'jkl']
>>> target_tokens = ['abc', 'd', 'ef', 'ghi', 'jkl']
>>> ta = TokenAligner(source_tokens, target_tokens)
>>> print(ta.project_token_idxs([1, 2]))
[1 2 3]
Returns:
List[int] representing the target indices associated with the provided source indices.
"""
if isinstance(idxs, int):
idxs = [idxs]
# column indices
return self.source_token_idx_to_target_token_idx[idxs].nonzero()[1]
@staticmethod
def _project_span(mat, start, end, inclusive):
if inclusive:
end = end + 1
target_matches = mat[start:end].nonzero()[1].tolist()
if len(target_matches) == 0:
raise ValueError(
f"Project {(start, end)} into empty span in target sequence")
output_start, output_end = min(target_matches), max(target_matches)
if not inclusive:
output_end = output_end + 1
return (output_start, output_end)
def project_token_span(self, start, end, inclusive=False) -> Tuple[int, int]:
"""Project a span from source to target token sequence.
Notes:
When param inclusive=False, the end index is interpreted as exclusive,
and the end of the span returned by the function will also be exclusive.
When param inclusive=True, both start and end indexes are interpreted as inclusive,
and the span returned by the function will also be inclusive.
Examples:
>>> source_tokens = ['abc', 'def', 'ghi', 'jkl']
>>> target_tokens = ['abc', 'd', 'ef', 'ghi', 'jkl']
>>> ta = TokenAligner(source_tokens, target_tokens)
>>> start, end = 0, 2
>>> print(ta.project_token_span(start, end))
(0, 3)
Raise:
When target span is empty
Returns:
Tuple[int, int] representing the target span corresponding to the source span.
"""
return self._project_span(
mat=self.source_token_idx_to_target_token_idx, start=start, end=end, inclusive=inclusive
)
def project_token_to_char_span(self, start, end, inclusive=False) -> Tuple[int, int]:
"""Project a span from source to target token sequence.
Notes:
When param inclusive=False, the end index is interpreted as exclusive,
and the end of the span returned by the function will also be exclusive.
When param inclusive=True, both start and end indexes are interpreted as inclusive,
and the span returned by the function will also be inclusive.
Examples:
>>> source_tokens = ['abc', 'def', 'ghi', 'jkl']
>>> target_str = 'abc d ef ghi jkl'
>>> ta = TokenAligner(source_tokens, target_str)
>>> start, end = 0, 2
>>> print(ta.project_token_to_char_span(start, end))
(0, 8)
Raise:
When target span is empty
Returns:
Tuple[int, int] representing the target span corresponding to the source span.
"""
return self._project_span(
mat=self.source_token_idx_to_target_char_idx, start=start, end=end, inclusive=inclusive
)
def project_char_to_token_span(self, start, end, inclusive=False) -> Tuple[int, int]:
"""Project a span from source to target token sequence.
Notes:
When param inclusive=False, the end index is interpreted as exclusive,
and the end of the span returned by the function will also be exclusive.
When param inclusive=True, both start and end indexes are interpreted as inclusive,
and the span returned by the function will also be inclusive.
Examples:
>>> source_str = 'abc def ghi jkl'
>>> target_tokens = ['abc', 'd', 'ef', 'ghi', 'jkl']
>>> ta = TokenAligner(source_str, target_tokens)
>>> start, end = 0, 4
>>> print(ta.project_char_to_token_span(start, end))
(0, 1)
Raise:
When target span is empty
Returns:
Tuple[int, int] representing the target span corresponding to the source span.
"""
return self._project_span(
mat=self.source_char_idx_to_target_token_idx, start=start, end=end, inclusive=inclusive
)
|
[
"numpy.identity",
"numpy.zeros",
"Levenshtein.StringMatcher.StringMatcher",
"nltk.tokenize.util.string_span_tokenize"
] |
[((597, 647), 'numpy.zeros', 'np.zeros', (['(n_chars_src, n_chars_tgt)'], {'dtype': '_DTYPE'}), '((n_chars_src, n_chars_tgt), dtype=_DTYPE)\n', (605, 647), True, 'import numpy as np\n'), ((2133, 2168), 'nltk.tokenize.util.string_span_tokenize', 'string_span_tokenize', (['text'], {'sep': 'sep'}), '(text, sep=sep)\n', (2153, 2168), False, 'from nltk.tokenize.util import string_span_tokenize\n'), ((3676, 3715), 'Levenshtein.StringMatcher.StringMatcher', 'StringMatcher', ([], {'seq1': 'source', 'seq2': 'target'}), '(seq1=source, seq2=target)\n', (3689, 3715), False, 'from Levenshtein.StringMatcher import StringMatcher\n'), ((1106, 1137), 'numpy.identity', 'np.identity', (['b[2]'], {'dtype': '_DTYPE'}), '(b[2], dtype=_DTYPE)\n', (1117, 1137), True, 'import numpy as np\n')]
|
# -*- coding: utf-8 -*-
"""
:math:`IC_TC_P` Colour Encoding
===============================
Defines the :math:`IC_TC_P` colour encoding related transformations:
- :func:`colour.RGB_to_ICtCp`
- :func:`colour.ICtCp_to_RGB`
- :func:`colour.XYZ_to_ICtCp`
- :func:`colour.ICtCp_to_XYZ`
References
----------
- :cite:`Dolby2016a` : Dolby. (2016). WHAT IS ICtCp? - INTRODUCTION.
https://www.dolby.com/us/en/technologies/dolby-vision/ICtCp-white-paper.pdf
- :cite:`InternationalTelecommunicationUnion2018` : International
Telecommunication Union. (2018). Recommendation ITU-R BT.2100-2 - Image
parameter values for high dynamic range television for use in production
and international programme exchange.
https://www.itu.int/dms_pubrec/itu-r/rec/bt/\
R-REC-BT.2100-2-201807-I!!PDF-E.pdf
- :cite:`Lu2016c` : <NAME>., <NAME>., <NAME>., <NAME>., <NAME>., Pytlarz,
J., <NAME>., <NAME>., & <NAME>. (2016). ITP Colour Space and Its
Compression Performance for High Dynamic Range and Wide Colour Gamut Video
Distribution. ZTE Communications, 14(1), 32-38.
"""
import numpy as np
from colour.algebra import vector_dot
from colour.colorimetry import CCS_ILLUMINANTS
from colour.models.rgb import RGB_COLOURSPACES, RGB_to_XYZ, XYZ_to_RGB
from colour.models.rgb.transfer_functions import (
eotf_ST2084, eotf_inverse_ST2084, oetf_HLG_BT2100, oetf_inverse_HLG_BT2100)
from colour.utilities import (domain_range_scale, from_range_1, to_domain_1,
validate_method)
__author__ = 'Colour Developers'
__copyright__ = 'Copyright (C) 2013-2021 - Colour Developers'
__license__ = 'New BSD License - https://opensource.org/licenses/BSD-3-Clause'
__maintainer__ = 'Colour Developers'
__email__ = '<EMAIL>'
__status__ = 'Production'
__all__ = [
'MATRIX_ICTCP_RGB_TO_LMS', 'MATRIX_ICTCP_LMS_TO_RGB',
'MATRIX_ICTCP_LMS_P_TO_ICTCP', 'MATRIX_ICTCP_ICTCP_TO_LMS_P',
'MATRIX_ICTCP_LMS_P_TO_ICTCP_HLG_BT2100_2',
'MATRIX_ICTCP_ICTCP_TO_LMS_P_HLG_BT2100_2', 'RGB_to_ICtCp', 'ICtCp_to_RGB',
'XYZ_to_ICtCp', 'ICtCp_to_XYZ'
]
MATRIX_ICTCP_RGB_TO_LMS = np.array([
[1688, 2146, 262],
[683, 2951, 462],
[99, 309, 3688],
]) / 4096
"""
*ITU-R BT.2020* colourspace to normalised cone responses matrix.
MATRIX_ICTCP_RGB_TO_LMS : array_like, (3, 3)
"""
MATRIX_ICTCP_LMS_TO_RGB = np.linalg.inv(MATRIX_ICTCP_RGB_TO_LMS)
"""
:math:`IC_TC_P` colourspace normalised cone responses to *ITU-R BT.2020*
colourspace matrix.
MATRIX_ICTCP_LMS_TO_RGB : array_like, (3, 3)
"""
MATRIX_ICTCP_LMS_P_TO_ICTCP = np.array([
[2048, 2048, 0],
[6610, -13613, 7003],
[17933, -17390, -543],
]) / 4096
"""
:math:`LMS_p` *SMPTE ST 2084:2014* encoded normalised cone responses to
:math:`IC_TC_P` colour encoding matrix.
MATRIX_ICTCP_LMS_P_TO_ICTCP : array_like, (3, 3)
"""
MATRIX_ICTCP_ICTCP_TO_LMS_P = np.linalg.inv(MATRIX_ICTCP_LMS_P_TO_ICTCP)
"""
:math:`IC_TC_P` colour encoding to :math:`LMS_p` *SMPTE ST 2084:2014* encoded
normalised cone responses matrix.
MATRIX_ICTCP_ICTCP_TO_LMS_P : array_like, (3, 3)
"""
MATRIX_ICTCP_LMS_P_TO_ICTCP_HLG_BT2100_2 = np.array([
[2048, 2048, 0],
[3625, -7465, 3840],
[9500, -9212, -288],
]) / 4096
"""
:math:`LMS_p` *SMPTE ST 2084:2014* encoded normalised cone responses to
:math:`IC_TC_P` colour encoding matrix as given in *ITU-R BT.2100-2*.
MATRIX_ICTCP_LMS_P_TO_ICTCP_HLG_BT2100_2 : array_like, (3, 3)
"""
MATRIX_ICTCP_ICTCP_TO_LMS_P_HLG_BT2100_2 = np.linalg.inv(
MATRIX_ICTCP_LMS_P_TO_ICTCP_HLG_BT2100_2)
"""
:math:`IC_TC_P` colour encoding to :math:`LMS_p` *SMPTE ST 2084:2014* encoded
normalised cone responses matrix as given in *ITU-R BT.2100-2*.
MATRIX_ICTCP_ICTCP_TO_LMS_P_HLG_BT2100_2 : array_like, (3, 3)
"""
def RGB_to_ICtCp(RGB, method='Dolby 2016', L_p=10000):
"""
Converts from *ITU-R BT.2020* colourspace to :math:`IC_TC_P` colour
encoding.
Parameters
----------
RGB : array_like
*ITU-R BT.2020* colourspace array.
method : unicode, optional
**{'Dolby 2016', 'ITU-R BT.2100-1 HLG', 'ITU-R BT.2100-1 PQ',
'ITU-R BT.2100-2 HLG', 'ITU-R BT.2100-2 PQ'}**,
Computation method. *Recommendation ITU-R BT.2100* defines multiple
variants of the :math:`IC_TC_P` colour encoding:
- *ITU-R BT.2100-1*
- *SMPTE ST 2084:2014* inverse electro-optical transfer
function (EOTF / EOCF) and the :math:`IC_TC_P` matrix from
:cite:`Dolby2016a`: *Dolby 2016*, *ITU-R BT.2100-1 PQ*,
*ITU-R BT.2100-2 PQ* methods.
- *Recommendation ITU-R BT.2100* *Reference HLG* opto-electrical
transfer function (OETF / OECF) and the :math:`IC_TC_P` matrix
from :cite:`Dolby2016a`: *ITU-R BT.2100-1 HLG* method.
- *ITU-R BT.2100-2*
- *SMPTE ST 2084:2014* inverse electro-optical transfer
function (EOTF / EOCF) and the :math:`IC_TC_P` matrix from
:cite:`Dolby2016a`: *Dolby 2016*, *ITU-R BT.2100-1 PQ*,
*ITU-R BT.2100-2 PQ* methods.
- *Recommendation ITU-R BT.2100* *Reference HLG* opto-electrical
transfer function (OETF / OECF) and a custom :math:`IC_TC_P`
matrix from :cite:`InternationalTelecommunicationUnion2018`:
*ITU-R BT.2100-2 HLG* method.
L_p : numeric, optional
Display peak luminance :math:`cd/m^2` for *SMPTE ST 2084:2014*
non-linear encoding. This parameter should stay at its default
:math:`10000 cd/m^2` value for practical applications. It is exposed so
that the definition can be used as a fitting function.
Returns
-------
ndarray
:math:`IC_TC_P` colour encoding array.
Warnings
--------
The underlying *SMPTE ST 2084:2014* transfer function is an absolute
transfer function.
Notes
-----
- The *ITU-R BT.2100-1 PQ* and *ITU-R BT.2100-2 PQ* methods are aliases
for the *Dolby 2016* method.
- The underlying *SMPTE ST 2084:2014* transfer function is an absolute
transfer function, thus the domain and range values for the *Reference*
and *1* scales are only indicative that the data is not affected by
scale transformations. The effective domain of *SMPTE ST 2084:2014*
inverse electro-optical transfer function (EOTF / EOCF) is
[0.0001, 10000].
+------------+-----------------------+------------------+
| **Domain** | **Scale - Reference** | **Scale - 1** |
+============+=======================+==================+
| ``RGB`` | ``UN`` | ``UN`` |
+------------+-----------------------+------------------+
+------------+-----------------------+------------------+
| **Range** | **Scale - Reference** | **Scale - 1** |
+============+=======================+==================+
| ``ICtCp`` | ``I`` : [0, 1] | ``I`` : [0, 1] |
| | | |
| | ``CT`` : [-1, 1] | ``CT`` : [-1, 1] |
| | | |
| | ``CP`` : [-1, 1] | ``CP`` : [-1, 1] |
+------------+-----------------------+------------------+
References
----------
:cite:`Dolby2016a`, :cite:`Lu2016c`
Examples
--------
>>> RGB = np.array([0.45620519, 0.03081071, 0.04091952])
>>> RGB_to_ICtCp(RGB) # doctest: +ELLIPSIS
array([ 0.0735136..., 0.0047525..., 0.0935159...])
>>> RGB_to_ICtCp(RGB, method='ITU-R BT.2100-2 HLG') # doctest: +ELLIPSIS
array([ 0.6256789..., -0.0198449..., 0.3591125...])
"""
RGB = to_domain_1(RGB)
method = validate_method(method, [
'Dolby 2016', 'ITU-R BT.2100-1 HLG', 'ITU-R BT.2100-1 PQ',
'ITU-R BT.2100-2 HLG', 'ITU-R BT.2100-2 PQ'
])
is_hlg_method = 'hlg' in method
is_BT2100_2_method = '2100-2' in method
LMS = vector_dot(MATRIX_ICTCP_RGB_TO_LMS, RGB)
with domain_range_scale('ignore'):
LMS_p = (oetf_HLG_BT2100(LMS)
if is_hlg_method else eotf_inverse_ST2084(LMS, L_p))
ICtCp = (vector_dot(MATRIX_ICTCP_LMS_P_TO_ICTCP_HLG_BT2100_2, LMS_p)
if (is_hlg_method and is_BT2100_2_method) else vector_dot(
MATRIX_ICTCP_LMS_P_TO_ICTCP, LMS_p))
return from_range_1(ICtCp)
def ICtCp_to_RGB(ICtCp, method='Dolby 2016', L_p=10000):
"""
Converts from :math:`IC_TC_P` colour encoding to *ITU-R BT.2020*
colourspace.
Parameters
----------
ICtCp : array_like
:math:`IC_TC_P` colour encoding array.
method : unicode, optional
**{'Dolby 2016', 'ITU-R BT.2100-1 HLG', 'ITU-R BT.2100-1 PQ',
'ITU-R BT.2100-2 HLG', 'ITU-R BT.2100-2 PQ'}**,
Computation method. *Recommendation ITU-R BT.2100* defines multiple
variants of the :math:`IC_TC_P` colour encoding:
- *ITU-R BT.2100-1*
- *SMPTE ST 2084:2014* inverse electro-optical transfer
function (EOTF / EOCF) and the :math:`IC_TC_P` matrix from
:cite:`Dolby2016a`: *Dolby 2016*, *ITU-R BT.2100-1 PQ*,
*ITU-R BT.2100-2 PQ* methods.
- *Recommendation ITU-R BT.2100* *Reference HLG* opto-electrical
transfer function (OETF / OECF) and the :math:`IC_TC_P` matrix
from :cite:`Dolby2016a`: *ITU-R BT.2100-1 HLG* method.
- *ITU-R BT.2100-2*
- *SMPTE ST 2084:2014* inverse electro-optical transfer
function (EOTF / EOCF) and the :math:`IC_TC_P` matrix from
:cite:`Dolby2016a`: *Dolby 2016*, *ITU-R BT.2100-1 PQ*,
*ITU-R BT.2100-2 PQ* methods.
- *Recommendation ITU-R BT.2100* *Reference HLG* opto-electrical
transfer function (OETF / OECF) and a custom :math:`IC_TC_P`
matrix from :cite:`InternationalTelecommunicationUnion2018`:
*ITU-R BT.2100-2 HLG* method.
L_p : numeric, optional
Display peak luminance :math:`cd/m^2` for *SMPTE ST 2084:2014*
non-linear encoding. This parameter should stay at its default
:math:`10000 cd/m^2` value for practical applications. It is exposed so
that the definition can be used as a fitting function.
Returns
-------
ndarray
*ITU-R BT.2020* colourspace array.
Warnings
--------
The underlying *SMPTE ST 2084:2014* transfer function is an absolute
transfer function.
Notes
-----
- The *ITU-R BT.2100-1 PQ* and *ITU-R BT.2100-2 PQ* methods are aliases
for the *Dolby 2016* method.
- The underlying *SMPTE ST 2084:2014* transfer function is an absolute
transfer function, thus the domain and range values for the *Reference*
and *1* scales are only indicative that the data is not affected by
scale transformations.
+------------+-----------------------+------------------+
| **Domain** | **Scale - Reference** | **Scale - 1** |
+============+=======================+==================+
| ``ICtCp`` | ``I`` : [0, 1] | ``I`` : [0, 1] |
| | | |
| | ``CT`` : [-1, 1] | ``CT`` : [-1, 1] |
| | | |
| | ``CP`` : [-1, 1] | ``CP`` : [-1, 1] |
+------------+-----------------------+------------------+
+------------+-----------------------+------------------+
| **Range** | **Scale - Reference** | **Scale - 1** |
+============+=======================+==================+
| ``RGB`` | ``UN`` | ``UN`` |
+------------+-----------------------+------------------+
References
----------
:cite:`Dolby2016a`, :cite:`Lu2016c`
Examples
--------
>>> ICtCp = np.array([0.07351364, 0.00475253, 0.09351596])
>>> ICtCp_to_RGB(ICtCp) # doctest: +ELLIPSIS
array([ 0.4562052..., 0.0308107..., 0.0409195...])
>>> ICtCp = np.array([0.62567899, -0.01984490, 0.35911259])
>>> ICtCp_to_RGB(ICtCp, method='ITU-R BT.2100-2 HLG') # doctest: +ELLIPSIS
array([ 0.4562052..., 0.0308107..., 0.0409195...])
"""
ICtCp = to_domain_1(ICtCp)
method = validate_method(method, [
'Dolby 2016', 'ITU-R BT.2100-1 HLG', 'ITU-R BT.2100-1 PQ',
'ITU-R BT.2100-2 HLG', 'ITU-R BT.2100-2 PQ'
])
is_hlg_method = 'hlg' in method
is_BT2100_2_method = '2100-2' in method
LMS_p = (vector_dot(MATRIX_ICTCP_ICTCP_TO_LMS_P_HLG_BT2100_2, ICtCp)
if (is_hlg_method and is_BT2100_2_method) else vector_dot(
MATRIX_ICTCP_ICTCP_TO_LMS_P, ICtCp))
with domain_range_scale('ignore'):
LMS = (oetf_inverse_HLG_BT2100(LMS_p)
if is_hlg_method else eotf_ST2084(LMS_p, L_p))
RGB = vector_dot(MATRIX_ICTCP_LMS_TO_RGB, LMS)
return from_range_1(RGB)
def XYZ_to_ICtCp(XYZ,
illuminant=CCS_ILLUMINANTS[
'CIE 1931 2 Degree Standard Observer']['D65'],
chromatic_adaptation_transform='CAT02',
method='Dolby 2016',
L_p=10000):
"""
Converts from *CIE XYZ* tristimulus values to :math:`IC_TC_P` colour
encoding.
Parameters
----------
XYZ : array_like
*CIE XYZ* tristimulus values.
illuminant : array_like, optional
Source illuminant chromaticity coordinates.
chromatic_adaptation_transform : unicode, optional
**{'CAT02', 'XYZ Scaling', '<NAME>', 'Bradford', 'Sharp',
'Fairchild', 'CMCCAT97', 'CMCCAT2000', 'CAT02 Brill 2008', 'CAT16',
'Bianco 2010', 'Bianco PC 2010'}**,
*Chromatic adaptation* transform.
method : unicode, optional
**{'Dolby 2016', 'ITU-R BT.2100-1 HLG', 'ITU-R BT.2100-1 PQ',
'ITU-R BT.2100-2 HLG', 'ITU-R BT.2100-2 PQ'}**,
Computation method. *Recommendation ITU-R BT.2100* defines multiple
variants of the :math:`IC_TC_P` colour encoding:
- *ITU-R BT.2100-1*
- *SMPTE ST 2084:2014* inverse electro-optical transfer
function (EOTF / EOCF) and the :math:`IC_TC_P` matrix from
:cite:`Dolby2016a`: *Dolby 2016*, *ITU-R BT.2100-1 PQ*,
*ITU-R BT.2100-2 PQ* methods.
- *Recommendation ITU-R BT.2100* *Reference HLG* opto-electrical
transfer function (OETF / OECF) and the :math:`IC_TC_P` matrix
from :cite:`Dolby2016a`: *ITU-R BT.2100-1 HLG* method.
- *ITU-R BT.2100-2*
- *SMPTE ST 2084:2014* inverse electro-optical transfer
function (EOTF / EOCF) and the :math:`IC_TC_P` matrix from
:cite:`Dolby2016a`: *Dolby 2016*, *ITU-R BT.2100-1 PQ*,
*ITU-R BT.2100-2 PQ* methods.
- *Recommendation ITU-R BT.2100* *Reference HLG* opto-electrical
transfer function (OETF / OECF) and a custom :math:`IC_TC_P`
matrix from :cite:`InternationalTelecommunicationUnion2018`:
*ITU-R BT.2100-2 HLG* method.
L_p : numeric, optional
Display peak luminance :math:`cd/m^2` for *SMPTE ST 2084:2014*
non-linear encoding. This parameter should stay at its default
:math:`10000 cd/m^2` value for practical applications. It is exposed so
that the definition can be used as a fitting function.
Returns
-------
ndarray
:math:`IC_TC_P` colour encoding array.
Warnings
--------
The underlying *SMPTE ST 2084:2014* transfer function is an absolute
transfer function.
Notes
-----
- The underlying *SMPTE ST 2084:2014* transfer function is an absolute
transfer function, thus the domain and range values for the *Reference*
- The *ITU-R BT.2100-1 PQ* and *ITU-R BT.2100-2 PQ* methods are aliases
for the *Dolby 2016* method.
and *1* scales are only indicative that the data is not affected by
scale transformations. The effective domain of *SMPTE ST 2084:2014*
inverse electro-optical transfer function (EOTF / EOCF) is
[0.0001, 10000].
+------------+-----------------------+------------------+
| **Domain** | **Scale - Reference** | **Scale - 1** |
+============+=======================+==================+
| ``XYZ`` | ``UN`` | ``UN`` |
+------------+-----------------------+------------------+
+------------+-----------------------+------------------+
| **Range** | **Scale - Reference** | **Scale - 1** |
+============+=======================+==================+
| ``ICtCp`` | ``I`` : [0, 1] | ``I`` : [0, 1] |
| | | |
| | ``CT`` : [-1, 1] | ``CT`` : [-1, 1] |
| | | |
| | ``CP`` : [-1, 1] | ``CP`` : [-1, 1] |
+------------+-----------------------+------------------+
References
----------
:cite:`Dolby2016a`, :cite:`Lu2016c`
Examples
--------
>>> XYZ = np.array([0.20654008, 0.12197225, 0.05136952])
>>> XYZ_to_ICtCp(XYZ) # doctest: +ELLIPSIS
array([ 0.0685809..., -0.0028384..., 0.0602098...])
>>> XYZ_to_ICtCp(XYZ, method='ITU-R BT.2100-2 HLG') # doctest: +ELLIPSIS
array([ 0.5924279..., -0.0374073..., 0.2512267...])
"""
BT2020 = RGB_COLOURSPACES['ITU-R BT.2020']
RGB = XYZ_to_RGB(
XYZ,
illuminant,
BT2020.whitepoint,
BT2020.matrix_XYZ_to_RGB,
chromatic_adaptation_transform,
)
return RGB_to_ICtCp(RGB, method, L_p)
def ICtCp_to_XYZ(ICtCp,
illuminant=CCS_ILLUMINANTS[
'CIE 1931 2 Degree Standard Observer']['D65'],
chromatic_adaptation_transform='CAT02',
method='Dolby 2016',
L_p=10000):
"""
Converts from :math:`IC_TC_P` colour encoding to *CIE XYZ* tristimulus
values.
Parameters
----------
ICtCp : array_like
:math:`IC_TC_P` colour encoding array.
illuminant : array_like, optional
Source illuminant chromaticity coordinates.
chromatic_adaptation_transform : unicode, optional
**{'CAT02', 'XYZ Scaling', '<NAME>', 'Bradford', 'Sharp',
'Fairchild', 'CMCCAT97', 'CMCCAT2000', 'CAT02 Brill 2008', 'CAT16',
'Bianco 2010', 'Bianco PC 2010'}**,
*Chromatic adaptation* transform.
method : unicode, optional
**{'Dolby 2016', 'ITU-R BT.2100-1 HLG', 'ITU-R BT.2100-1 PQ',
'ITU-R BT.2100-2 HLG', 'ITU-R BT.2100-2 PQ'}**,
Computation method. *Recommendation ITU-R BT.2100* defines multiple
variants of the :math:`IC_TC_P` colour encoding:
- *ITU-R BT.2100-1*
- *SMPTE ST 2084:2014* inverse electro-optical transfer
function (EOTF / EOCF) and the :math:`IC_TC_P` matrix from
:cite:`Dolby2016a`: *Dolby 2016*, *ITU-R BT.2100-1 PQ*,
*ITU-R BT.2100-2 PQ* methods.
- *Recommendation ITU-R BT.2100* *Reference HLG* opto-electrical
transfer function (OETF / OECF) and the :math:`IC_TC_P` matrix
from :cite:`Dolby2016a`: *ITU-R BT.2100-1 HLG* method.
- *ITU-R BT.2100-2*
- *SMPTE ST 2084:2014* inverse electro-optical transfer
function (EOTF / EOCF) and the :math:`IC_TC_P` matrix from
:cite:`Dolby2016a`: *Dolby 2016*, *ITU-R BT.2100-1 PQ*,
*ITU-R BT.2100-2 PQ* methods.
- *Recommendation ITU-R BT.2100* *Reference HLG* opto-electrical
transfer function (OETF / OECF) and a custom :math:`IC_TC_P`
matrix from :cite:`InternationalTelecommunicationUnion2018`:
*ITU-R BT.2100-2 HLG* method.
L_p : numeric, optional
Display peak luminance :math:`cd/m^2` for *SMPTE ST 2084:2014*
non-linear encoding. This parameter should stay at its default
:math:`10000 cd/m^2` value for practical applications. It is exposed so
that the definition can be used as a fitting function.
Returns
-------
ndarray
*CIE XYZ* tristimulus values.
Warnings
--------
The underlying *SMPTE ST 2084:2014* transfer function is an absolute
transfer function.
Notes
-----
- The *ITU-R BT.2100-1 PQ* and *ITU-R BT.2100-2 PQ* methods are aliases
for the *Dolby 2016* method.
- The underlying *SMPTE ST 2084:2014* transfer function is an absolute
transfer function, thus the domain and range values for the *Reference*
and *1* scales are only indicative that the data is not affected by
scale transformations.
+------------+-----------------------+------------------+
| **Domain** | **Scale - Reference** | **Scale - 1** |
+============+=======================+==================+
| ``ICtCp`` | ``I`` : [0, 1] | ``I`` : [0, 1] |
| | | |
| | ``CT`` : [-1, 1] | ``CT`` : [-1, 1] |
| | | |
| | ``CP`` : [-1, 1] | ``CP`` : [-1, 1] |
+------------+-----------------------+------------------+
+------------+-----------------------+------------------+
| **Range** | **Scale - Reference** | **Scale - 1** |
+============+=======================+==================+
| ``XYZ`` | ``UN`` | ``UN`` |
+------------+-----------------------+------------------+
References
----------
:cite:`Dolby2016a`, :cite:`Lu2016c`
Examples
--------
>>> ICtCp = np.array([0.06858097, -0.00283842, 0.06020983])
>>> ICtCp_to_XYZ(ICtCp) # doctest: +ELLIPSIS
array([ 0.2065400..., 0.1219722..., 0.0513695...])
>>> ICtCp = np.array([0.59242792, -0.03740730, 0.25122675])
>>> ICtCp_to_XYZ(ICtCp, method='ITU-R BT.2100-2 HLG') # doctest: +ELLIPSIS
array([ 0.2065400..., 0.1219722..., 0.0513695...])
"""
RGB = ICtCp_to_RGB(ICtCp, method, L_p)
BT2020 = RGB_COLOURSPACES['ITU-R BT.2020']
XYZ = RGB_to_XYZ(
RGB,
BT2020.whitepoint,
illuminant,
BT2020.matrix_RGB_to_XYZ,
chromatic_adaptation_transform,
)
return XYZ
|
[
"colour.algebra.vector_dot",
"colour.utilities.to_domain_1",
"colour.utilities.from_range_1",
"colour.models.rgb.RGB_to_XYZ",
"colour.models.rgb.transfer_functions.oetf_HLG_BT2100",
"colour.models.rgb.XYZ_to_RGB",
"colour.utilities.domain_range_scale",
"colour.models.rgb.transfer_functions.eotf_inverse_ST2084",
"colour.utilities.validate_method",
"numpy.array",
"numpy.linalg.inv",
"colour.models.rgb.transfer_functions.eotf_ST2084",
"colour.models.rgb.transfer_functions.oetf_inverse_HLG_BT2100"
] |
[((2348, 2386), 'numpy.linalg.inv', 'np.linalg.inv', (['MATRIX_ICTCP_RGB_TO_LMS'], {}), '(MATRIX_ICTCP_RGB_TO_LMS)\n', (2361, 2386), True, 'import numpy as np\n'), ((2861, 2903), 'numpy.linalg.inv', 'np.linalg.inv', (['MATRIX_ICTCP_LMS_P_TO_ICTCP'], {}), '(MATRIX_ICTCP_LMS_P_TO_ICTCP)\n', (2874, 2903), True, 'import numpy as np\n'), ((3467, 3522), 'numpy.linalg.inv', 'np.linalg.inv', (['MATRIX_ICTCP_LMS_P_TO_ICTCP_HLG_BT2100_2'], {}), '(MATRIX_ICTCP_LMS_P_TO_ICTCP_HLG_BT2100_2)\n', (3480, 3522), True, 'import numpy as np\n'), ((2115, 2179), 'numpy.array', 'np.array', (['[[1688, 2146, 262], [683, 2951, 462], [99, 309, 3688]]'], {}), '([[1688, 2146, 262], [683, 2951, 462], [99, 309, 3688]])\n', (2123, 2179), True, 'import numpy as np\n'), ((2565, 2637), 'numpy.array', 'np.array', (['[[2048, 2048, 0], [6610, -13613, 7003], [17933, -17390, -543]]'], {}), '([[2048, 2048, 0], [6610, -13613, 7003], [17933, -17390, -543]])\n', (2573, 2637), True, 'import numpy as np\n'), ((3118, 3187), 'numpy.array', 'np.array', (['[[2048, 2048, 0], [3625, -7465, 3840], [9500, -9212, -288]]'], {}), '([[2048, 2048, 0], [3625, -7465, 3840], [9500, -9212, -288]])\n', (3126, 3187), True, 'import numpy as np\n'), ((7721, 7737), 'colour.utilities.to_domain_1', 'to_domain_1', (['RGB'], {}), '(RGB)\n', (7732, 7737), False, 'from colour.utilities import domain_range_scale, from_range_1, to_domain_1, validate_method\n'), ((7751, 7884), 'colour.utilities.validate_method', 'validate_method', (['method', "['Dolby 2016', 'ITU-R BT.2100-1 HLG', 'ITU-R BT.2100-1 PQ',\n 'ITU-R BT.2100-2 HLG', 'ITU-R BT.2100-2 PQ']"], {}), "(method, ['Dolby 2016', 'ITU-R BT.2100-1 HLG',\n 'ITU-R BT.2100-1 PQ', 'ITU-R BT.2100-2 HLG', 'ITU-R BT.2100-2 PQ'])\n", (7766, 7884), False, 'from colour.utilities import domain_range_scale, from_range_1, to_domain_1, validate_method\n'), ((7995, 8035), 'colour.algebra.vector_dot', 'vector_dot', (['MATRIX_ICTCP_RGB_TO_LMS', 'RGB'], {}), '(MATRIX_ICTCP_RGB_TO_LMS, RGB)\n', (8005, 8035), False, 'from colour.algebra import vector_dot\n'), ((8396, 8415), 'colour.utilities.from_range_1', 'from_range_1', (['ICtCp'], {}), '(ICtCp)\n', (8408, 8415), False, 'from colour.utilities import domain_range_scale, from_range_1, to_domain_1, validate_method\n'), ((12336, 12354), 'colour.utilities.to_domain_1', 'to_domain_1', (['ICtCp'], {}), '(ICtCp)\n', (12347, 12354), False, 'from colour.utilities import domain_range_scale, from_range_1, to_domain_1, validate_method\n'), ((12368, 12501), 'colour.utilities.validate_method', 'validate_method', (['method', "['Dolby 2016', 'ITU-R BT.2100-1 HLG', 'ITU-R BT.2100-1 PQ',\n 'ITU-R BT.2100-2 HLG', 'ITU-R BT.2100-2 PQ']"], {}), "(method, ['Dolby 2016', 'ITU-R BT.2100-1 HLG',\n 'ITU-R BT.2100-1 PQ', 'ITU-R BT.2100-2 HLG', 'ITU-R BT.2100-2 PQ'])\n", (12383, 12501), False, 'from colour.utilities import domain_range_scale, from_range_1, to_domain_1, validate_method\n'), ((12960, 13000), 'colour.algebra.vector_dot', 'vector_dot', (['MATRIX_ICTCP_LMS_TO_RGB', 'LMS'], {}), '(MATRIX_ICTCP_LMS_TO_RGB, LMS)\n', (12970, 13000), False, 'from colour.algebra import vector_dot\n'), ((13013, 13030), 'colour.utilities.from_range_1', 'from_range_1', (['RGB'], {}), '(RGB)\n', (13025, 13030), False, 'from colour.utilities import domain_range_scale, from_range_1, to_domain_1, validate_method\n'), ((17633, 17741), 'colour.models.rgb.XYZ_to_RGB', 'XYZ_to_RGB', (['XYZ', 'illuminant', 'BT2020.whitepoint', 'BT2020.matrix_XYZ_to_RGB', 'chromatic_adaptation_transform'], {}), '(XYZ, illuminant, BT2020.whitepoint, BT2020.matrix_XYZ_to_RGB,\n chromatic_adaptation_transform)\n', (17643, 17741), False, 'from colour.models.rgb import RGB_COLOURSPACES, RGB_to_XYZ, XYZ_to_RGB\n'), ((22412, 22520), 'colour.models.rgb.RGB_to_XYZ', 'RGB_to_XYZ', (['RGB', 'BT2020.whitepoint', 'illuminant', 'BT2020.matrix_RGB_to_XYZ', 'chromatic_adaptation_transform'], {}), '(RGB, BT2020.whitepoint, illuminant, BT2020.matrix_RGB_to_XYZ,\n chromatic_adaptation_transform)\n', (22422, 22520), False, 'from colour.models.rgb import RGB_COLOURSPACES, RGB_to_XYZ, XYZ_to_RGB\n'), ((8046, 8074), 'colour.utilities.domain_range_scale', 'domain_range_scale', (['"""ignore"""'], {}), "('ignore')\n", (8064, 8074), False, 'from colour.utilities import domain_range_scale, from_range_1, to_domain_1, validate_method\n'), ((8198, 8257), 'colour.algebra.vector_dot', 'vector_dot', (['MATRIX_ICTCP_LMS_P_TO_ICTCP_HLG_BT2100_2', 'LMS_p'], {}), '(MATRIX_ICTCP_LMS_P_TO_ICTCP_HLG_BT2100_2, LMS_p)\n', (8208, 8257), False, 'from colour.algebra import vector_dot\n'), ((8318, 8364), 'colour.algebra.vector_dot', 'vector_dot', (['MATRIX_ICTCP_LMS_P_TO_ICTCP', 'LMS_p'], {}), '(MATRIX_ICTCP_LMS_P_TO_ICTCP, LMS_p)\n', (8328, 8364), False, 'from colour.algebra import vector_dot\n'), ((12615, 12674), 'colour.algebra.vector_dot', 'vector_dot', (['MATRIX_ICTCP_ICTCP_TO_LMS_P_HLG_BT2100_2', 'ICtCp'], {}), '(MATRIX_ICTCP_ICTCP_TO_LMS_P_HLG_BT2100_2, ICtCp)\n', (12625, 12674), False, 'from colour.algebra import vector_dot\n'), ((12735, 12781), 'colour.algebra.vector_dot', 'vector_dot', (['MATRIX_ICTCP_ICTCP_TO_LMS_P', 'ICtCp'], {}), '(MATRIX_ICTCP_ICTCP_TO_LMS_P, ICtCp)\n', (12745, 12781), False, 'from colour.algebra import vector_dot\n'), ((12811, 12839), 'colour.utilities.domain_range_scale', 'domain_range_scale', (['"""ignore"""'], {}), "('ignore')\n", (12829, 12839), False, 'from colour.utilities import domain_range_scale, from_range_1, to_domain_1, validate_method\n'), ((8093, 8113), 'colour.models.rgb.transfer_functions.oetf_HLG_BT2100', 'oetf_HLG_BT2100', (['LMS'], {}), '(LMS)\n', (8108, 8113), False, 'from colour.models.rgb.transfer_functions import eotf_ST2084, eotf_inverse_ST2084, oetf_HLG_BT2100, oetf_inverse_HLG_BT2100\n'), ((8153, 8182), 'colour.models.rgb.transfer_functions.eotf_inverse_ST2084', 'eotf_inverse_ST2084', (['LMS', 'L_p'], {}), '(LMS, L_p)\n', (8172, 8182), False, 'from colour.models.rgb.transfer_functions import eotf_ST2084, eotf_inverse_ST2084, oetf_HLG_BT2100, oetf_inverse_HLG_BT2100\n'), ((12856, 12886), 'colour.models.rgb.transfer_functions.oetf_inverse_HLG_BT2100', 'oetf_inverse_HLG_BT2100', (['LMS_p'], {}), '(LMS_p)\n', (12879, 12886), False, 'from colour.models.rgb.transfer_functions import eotf_ST2084, eotf_inverse_ST2084, oetf_HLG_BT2100, oetf_inverse_HLG_BT2100\n'), ((12924, 12947), 'colour.models.rgb.transfer_functions.eotf_ST2084', 'eotf_ST2084', (['LMS_p', 'L_p'], {}), '(LMS_p, L_p)\n', (12935, 12947), False, 'from colour.models.rgb.transfer_functions import eotf_ST2084, eotf_inverse_ST2084, oetf_HLG_BT2100, oetf_inverse_HLG_BT2100\n')]
|
#!/usr/bin/env python3
import csv
import numpy
thr_sig=5.0
def sigmoid(x):
return 1.0/(1.0+numpy.exp(-(x-thr_sig)))
if __name__=="__main__":
#parameters
time_pitch=1.0 #ms
save_pitch=10
save_pitch_weight=1000
simlen_sec=900.0
simlen=int(simlen_sec*1000.0/time_pitch)
tauL=10.0 #ms
phi=80.0/1000.0
phi_input=80.0/1000.0
alpha_som=0.5
alpha_dnd=0.5
beta_som=0.0
beta_dnd=0.0
gamma=1.0
c0=70.0
eta_som=0.2
eta_dnd=0.2
taudeltaW=1.0*1000.0 #ms
tau_mean=60.0*1000.0
eta_Wdecay=1e-7
Wnoise_amp=5e-3/numpy.sqrt(time_pitch)
som_input_num=50
dnd_input_num=som_input_num+0
group1_num=10
input_src_num=4
tau_input=10.0 #ms
input_amp=0.1/numpy.sqrt(time_pitch)
noise_amp=0.1/numpy.sqrt(time_pitch)
Winit=5.0
Wmin=0.0
E0=0.05
#variables
x=0.0
y=0.0
Ex=E0
Ey=E0
input_src=numpy.zeros(input_src_num)
som_input_current=numpy.zeros(som_input_num)
dnd_input_current=numpy.zeros(dnd_input_num)
som_inputPSC=numpy.zeros(som_input_num)
dnd_inputPSC=numpy.zeros(dnd_input_num)
deltaWsom=numpy.zeros(som_input_num)
deltaWdnd=numpy.zeros(dnd_input_num)
Wsom=Winit*(numpy.random.rand(som_input_num))
Wdnd=Winit*(numpy.random.rand(dnd_input_num))
#save
f_activity=open("activity.csv", "w")
csv_activity=csv.writer(f_activity, delimiter=",")
f_Wsom=open("Wsom.csv", "w")
csv_Wsom=csv.writer(f_Wsom, delimiter=",")
f_Wdnd=open("Wdnd.csv", "w")
csv_Wdnd=csv.writer(f_Wdnd, delimiter=",")
f_som_input=open("som_input.csv", "w")
csv_som_input=csv.writer(f_som_input, delimiter=",")
f_dnd_input=open("dnd_input.csv", "w")
csv_dnd_input=csv.writer(f_dnd_input, delimiter=",")
som_src=numpy.zeros([som_input_num, input_src_num])
som_src[:group1_num, 0]=1.0
som_src[group1_num:, 2]=1.0
dnd_src=numpy.zeros([dnd_input_num, input_src_num])
dnd_src[:group1_num,1]=1.0
dnd_src[group1_num:,3]=1.0
#simulation
for t in range(simlen):
time_sec=float(t)*time_pitch/1000.0
if time_sec==int(time_sec):
print(time_sec,"sec")
#source signal
input_src=input_src+time_pitch*(-input_src/tau_input+input_amp*numpy.random.randn(input_src_num))
#inputs
som_input_current+=time_pitch*(-som_input_current/tauL+som_src@input_src+noise_amp*numpy.random.randn(som_input_num))
dnd_input_current+=time_pitch*(-dnd_input_current/tauL+dnd_src@input_src+noise_amp*numpy.random.randn(dnd_input_num))
som_input=phi_input*sigmoid(som_input_current)
dnd_input=phi_input*sigmoid(dnd_input_current)
som_inputPSC+=time_pitch*(-som_inputPSC/tauL+som_input)
dnd_inputPSC+=time_pitch*(-dnd_inputPSC/tauL+dnd_input)
#dynamics
xprev=x+0.0
yprev=y+0.0
Isom=Wsom@som_inputPSC
Idnd=Wdnd@dnd_inputPSC
x=sigmoid(Isom+beta_som*yprev)
y=sigmoid(Idnd+beta_dnd*xprev)
z=(1.0+gamma*y)*phi*x
#plasticity
#som
Wsom+=time_pitch*(eta_som*deltaWsom+Wnoise_amp*numpy.random.randn(som_input_num)-eta_Wdecay*Wsom)
Wsom[Wsom<Wmin]=Wmin
theta_som=c0*Ex*Ex
deltaWsom+=time_pitch*(-deltaWsom+((1.0-alpha_som)*x*(x-theta_som)+alpha_som*x*y)*(1.0-x)*som_inputPSC)/taudeltaW
#dnd
Wdnd+=time_pitch*(eta_dnd*deltaWdnd+Wnoise_amp*numpy.random.randn(dnd_input_num)-eta_Wdecay*Wdnd)
Wdnd[Wdnd<Wmin]=Wmin
theta_dnd=c0*Ey*Ey
deltaWdnd+=time_pitch*(-deltaWdnd+((1.0-alpha_dnd)*y*(y-theta_dnd)+alpha_dnd*x*y)*(1.0-y)*dnd_inputPSC)/taudeltaW
Ex+=time_pitch*(-Ex+x)/tau_mean
Ey+=time_pitch*(-Ey+y)/tau_mean
if t%save_pitch==0:
csv_activity.writerow([time_sec, x, y, z]); f_activity.flush();
csv_som_input.writerow(numpy.hstack([time_sec, som_input])); f_som_input.flush();
csv_dnd_input.writerow(numpy.hstack([time_sec, dnd_input])); f_dnd_input.flush();
if t%save_pitch_weight==0:
csv_Wsom.writerow(numpy.hstack([time_sec, Wsom])); f_Wsom.flush();
csv_Wdnd.writerow(numpy.hstack([time_sec, Wdnd])); f_Wdnd.flush();
|
[
"numpy.sqrt",
"numpy.random.rand",
"numpy.hstack",
"csv.writer",
"numpy.exp",
"numpy.zeros",
"numpy.random.randn"
] |
[((917, 943), 'numpy.zeros', 'numpy.zeros', (['input_src_num'], {}), '(input_src_num)\n', (928, 943), False, 'import numpy\n'), ((966, 992), 'numpy.zeros', 'numpy.zeros', (['som_input_num'], {}), '(som_input_num)\n', (977, 992), False, 'import numpy\n'), ((1015, 1041), 'numpy.zeros', 'numpy.zeros', (['dnd_input_num'], {}), '(dnd_input_num)\n', (1026, 1041), False, 'import numpy\n'), ((1059, 1085), 'numpy.zeros', 'numpy.zeros', (['som_input_num'], {}), '(som_input_num)\n', (1070, 1085), False, 'import numpy\n'), ((1103, 1129), 'numpy.zeros', 'numpy.zeros', (['dnd_input_num'], {}), '(dnd_input_num)\n', (1114, 1129), False, 'import numpy\n'), ((1144, 1170), 'numpy.zeros', 'numpy.zeros', (['som_input_num'], {}), '(som_input_num)\n', (1155, 1170), False, 'import numpy\n'), ((1185, 1211), 'numpy.zeros', 'numpy.zeros', (['dnd_input_num'], {}), '(dnd_input_num)\n', (1196, 1211), False, 'import numpy\n'), ((1381, 1418), 'csv.writer', 'csv.writer', (['f_activity'], {'delimiter': '""","""'}), "(f_activity, delimiter=',')\n", (1391, 1418), False, 'import csv\n'), ((1466, 1499), 'csv.writer', 'csv.writer', (['f_Wsom'], {'delimiter': '""","""'}), "(f_Wsom, delimiter=',')\n", (1476, 1499), False, 'import csv\n'), ((1546, 1579), 'csv.writer', 'csv.writer', (['f_Wdnd'], {'delimiter': '""","""'}), "(f_Wdnd, delimiter=',')\n", (1556, 1579), False, 'import csv\n'), ((1642, 1680), 'csv.writer', 'csv.writer', (['f_som_input'], {'delimiter': '""","""'}), "(f_som_input, delimiter=',')\n", (1652, 1680), False, 'import csv\n'), ((1742, 1780), 'csv.writer', 'csv.writer', (['f_dnd_input'], {'delimiter': '""","""'}), "(f_dnd_input, delimiter=',')\n", (1752, 1780), False, 'import csv\n'), ((1794, 1837), 'numpy.zeros', 'numpy.zeros', (['[som_input_num, input_src_num]'], {}), '([som_input_num, input_src_num])\n', (1805, 1837), False, 'import numpy\n'), ((1915, 1958), 'numpy.zeros', 'numpy.zeros', (['[dnd_input_num, input_src_num]'], {}), '([dnd_input_num, input_src_num])\n', (1926, 1958), False, 'import numpy\n'), ((585, 607), 'numpy.sqrt', 'numpy.sqrt', (['time_pitch'], {}), '(time_pitch)\n', (595, 607), False, 'import numpy\n'), ((743, 765), 'numpy.sqrt', 'numpy.sqrt', (['time_pitch'], {}), '(time_pitch)\n', (753, 765), False, 'import numpy\n'), ((784, 806), 'numpy.sqrt', 'numpy.sqrt', (['time_pitch'], {}), '(time_pitch)\n', (794, 806), False, 'import numpy\n'), ((1228, 1260), 'numpy.random.rand', 'numpy.random.rand', (['som_input_num'], {}), '(som_input_num)\n', (1245, 1260), False, 'import numpy\n'), ((1278, 1310), 'numpy.random.rand', 'numpy.random.rand', (['dnd_input_num'], {}), '(dnd_input_num)\n', (1295, 1310), False, 'import numpy\n'), ((97, 122), 'numpy.exp', 'numpy.exp', (['(-(x - thr_sig))'], {}), '(-(x - thr_sig))\n', (106, 122), False, 'import numpy\n'), ((3919, 3954), 'numpy.hstack', 'numpy.hstack', (['[time_sec, som_input]'], {}), '([time_sec, som_input])\n', (3931, 3954), False, 'import numpy\n'), ((4013, 4048), 'numpy.hstack', 'numpy.hstack', (['[time_sec, dnd_input]'], {}), '([time_sec, dnd_input])\n', (4025, 4048), False, 'import numpy\n'), ((4137, 4167), 'numpy.hstack', 'numpy.hstack', (['[time_sec, Wsom]'], {}), '([time_sec, Wsom])\n', (4149, 4167), False, 'import numpy\n'), ((4216, 4246), 'numpy.hstack', 'numpy.hstack', (['[time_sec, Wdnd]'], {}), '([time_sec, Wdnd])\n', (4228, 4246), False, 'import numpy\n'), ((2426, 2459), 'numpy.random.randn', 'numpy.random.randn', (['som_input_num'], {}), '(som_input_num)\n', (2444, 2459), False, 'import numpy\n'), ((2552, 2585), 'numpy.random.randn', 'numpy.random.randn', (['dnd_input_num'], {}), '(dnd_input_num)\n', (2570, 2585), False, 'import numpy\n'), ((2275, 2308), 'numpy.random.randn', 'numpy.random.randn', (['input_src_num'], {}), '(input_src_num)\n', (2293, 2308), False, 'import numpy\n'), ((3169, 3202), 'numpy.random.randn', 'numpy.random.randn', (['som_input_num'], {}), '(som_input_num)\n', (3187, 3202), False, 'import numpy\n'), ((3468, 3501), 'numpy.random.randn', 'numpy.random.randn', (['dnd_input_num'], {}), '(dnd_input_num)\n', (3486, 3501), False, 'import numpy\n')]
|
# Copyright 2020 Makani Technologies LLC
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""View functions to interact with web clients."""
import atexit
import json
import logging
import os
import re
import string
import time
from django import http
from django import shortcuts
from django import template
from django.core import urlresolvers
from makani.analysis.checks import log_util
from makani.avionics.network import message_type as aio_message_type
from makani.avionics.network import network_config
from makani.gs.monitor2.apps.layout import autogen
from makani.gs.monitor2.apps.layout import base as layout_base
from makani.gs.monitor2.apps.layout import layout_util
from makani.gs.monitor2.apps.layout import loader
from makani.gs.monitor2.apps.layout import memory as layout_memory
from makani.gs.monitor2.apps.layout import stoplights
from makani.gs.monitor2.apps.layout import widgets
from makani.gs.monitor2.apps.receiver import receiver_manager
from makani.gs.monitor2.apps.receiver import views as receiver_views
from makani.gs.monitor2.project import settings
from makani.lib.bazel import bazel_util
from makani.lib.python import c_helpers
from makani.lib.python import debug_util
from makani.lib.python import struct_tree
from makani.lib.python.h5_utils import h5_io
import numpy
MESSAGE_TYPE_HELPER = c_helpers.EnumHelper('MessageType', aio_message_type)
CONFIG_FILES = {
'plot_defs': os.path.join(settings.MONITOR_PATH, 'configs/plot_defs.json'),
}
def Home(request):
"""Get the response for the home page."""
layout_names = loader.LayoutLoader().Names()
layout_names.sort()
all_layouts = [
{'name': layout,
'url': urlresolvers.reverse(
'view_aio_layout', args=[loader.LayoutLoader().ModuleName(layout)])}
for layout in layout_names]
context = {
'layouts': all_layouts,
'canvas_cols': settings.CSS_GRID_COLUMNS,
}
_CreateAndAddClientIdToContext(context)
template_name = 'home.html'
return shortcuts.render(request, template_name, context,
context_instance=template.RequestContext(request))
def _ListFiles(path_arg):
"""List files under a local path."""
path_template = string.Template(path_arg)
prefix_path = path_template.substitute(os.environ)
sub_paths = os.listdir(prefix_path)
return prefix_path, sub_paths
def _GetFullFilePath(prefix_path, sub_path):
return os.path.join(prefix_path, sub_path)
def SelectAllLogs(request):
"""Select all logs in the last visited directory."""
current_path = request.session['current_path']
try:
prefix_path, sub_paths = _ListFiles(current_path)
except OSError:
return http.HttpResponse('Cannot list directory "%s"!' % current_path)
file_list = []
for sub_path in sorted(sub_paths):
# Construct the full path.
if sub_path.endswith('.h5') and not sub_path.startswith('format'):
full_path = _GetFullFilePath(prefix_path, sub_path)
if not os.path.isdir(full_path):
file_list.append(full_path)
return http.HttpResponse(';\n'.join(file_list))
def Console(request, command, args):
"""Take commandlines from the client and respond with console outputs.
Args:
request: The HTML resquest object.
command: The command to be run. Only 'ls' is permitted for now.
args: The string of arguments to the command.
Returns:
The HttpResponse telling the output of the command.
"""
if command != 'ls':
message = 'Command "%s" is not allowed.' % command
return http.HttpResponse(message)
arg_template = string.Template(args)
arg_path = arg_template.safe_substitute(
{'MAKANI_HOME': bazel_util.GetWorkspaceRoot()})
try:
prefix_path, sub_paths = _ListFiles(arg_path)
request.session['current_path'] = arg_path
except OSError:
return http.HttpResponse('Cannot list directory "%s"!' % arg_path)
file_list = []
for sub_path in sorted(sub_paths):
# Construct the full path.
full_path = _GetFullFilePath(prefix_path, sub_path)
if os.path.isdir(full_path):
# If this is a directory, add the javascript to allow users to click
# into it.
file_list.append(
'<a href="javascript:void(0)" onclick="onListFiles(\'%s\')">%s</a>'
% (full_path, sub_path))
elif sub_path.endswith('.h5') and not sub_path.startswith('format'):
# If this is an HDF5 file, add the javascript to allow users to
# visualize it.
file_list.append(
'<a href="javascript:void(0)" onclick="onAddLog(\'%s\')">%s</a>'
% (full_path, sub_path))
else:
file_list.append(sub_path)
text = '<br>'.join(file_list)
return http.HttpResponse(text)
def _GetMinMessageFrequency():
"""Get the minimum frequency across all message types."""
config = network_config.NetworkConfig(settings.NETWORK_YAML)
return min(m.frequency_hz for m in config.all_messages if m.frequency_hz > 0)
def _TryToEnforceAioReceiver(client_id):
"""Ensure that the client is subscribed to the AioReceiver."""
# TODO: Investigate always running the AioReceiver.
message_receiver = receiver_manager.ReceiverManager.GetReceiver(client_id)
if not message_receiver:
if receiver_manager.ReceiverManager.CheckAndStartAioReceiver(
client_id, receiver_views.CreateAioReceiver):
# A new AioReceiver is started.
# Get the longest period for all messages, and multiply it by two to
# make sure we do not miss any message.
time.sleep(2.0 / _GetMinMessageFrequency())
return receiver_manager.ReceiverManager.GetReceiver(client_id)
else:
return message_receiver
def ViewMessageType(request, client_id, message_type,
template_name='monitor.html'):
"""View information within a message by automatically generating a layout.
Args:
request: An HttpRequest from the client.
client_id: The ID of the client's browser tab.
message_type: The Enum name of a message type.
template_name: The HTML template used to render the layout.
Returns:
An HttpResponse in the format of a serialized JSON object.
"""
configs = _LoadConfigs()
_TryToEnforceAioReceiver(client_id)
resp = _GetMessage(request, client_id, message_type)
resp = resp.Data(convert_to_basic_types=True) if resp else {}
configs['scenarios'] = autogen.GenerateScenario(resp, message_type)
context = _PrepareContext(configs)
new_client_id = _CreateAndAddClientIdToContext(context)
context['periodic_url'] = '/dashboard/periodic/msg_enum/%s/%s' % (
new_client_id, message_type)
context['content_width'] = settings.CSS_GRID_COLUMNS
context['order_horizontally'] = True
return shortcuts.render(request, template_name, context,
context_instance=template.RequestContext(request))
def UpdateMessageOptions(unused_request, client_id):
"""Detect what messages have been received and update the client.
Args:
unused_request: An HttpRequest from the client.
client_id: The ID of the client's browser tab.
Returns:
An HttpResponse about a dictionary of {message_enum: message_short_name}
"""
message_receiver = _TryToEnforceAioReceiver(client_id)
info = message_receiver.GetReceivedMessageTypes() if message_receiver else []
return http.HttpResponse(json.dumps(info))
def ViewAioLayout(request, layout_name):
"""Open a monitor layout that get data from AIO.
Args:
request: An HttpRequest from the client.
layout_name: Name of the layout associated with the client.
Returns:
An HttpResponse in the format of a serialized JSON object.
"""
context = {'receiver_type': 'aio'}
return _ViewLayout(request, layout_name, context)
def BrowseLog(request, path):
"""Browse the log by expanding the field at `path`.
Args:
request: An HttpRequest from the client.
path: A path pointing to one field in the log.
Returns:
An HttpResponse serializing a list of names for child fields.
"""
# The log structure may differ across logs, we always use the first log to
# construct the log structure.
log_path = request.session['log_paths'][0]
log_data = struct_tree.StructTree(log_path, fail_silently=True, readonly=True)
try:
skeleton = log_data.Skeleton(path, depth=1)
except h5_io.H5IndexError:
return http.HttpResponse('{}')
parent_path = path
d3_data = struct_tree.DictToD3Tree(skeleton, '.', parent_path)
if 'children' in d3_data:
# The first layer is a placeholder. Starts from the second layer.
return http.HttpResponse(json.dumps(d3_data['children']))
else:
return http.HttpResponse('{}')
def ViewLogStructure(request, paths, template_name='log_structure.html'):
"""View structure of an HDF5 log at given log path.
Args:
request: An HttpRequest from the client.
paths: Paths to the local log files.
template_name: The HTML template used to render the layout.
Returns:
An HttpResponse that renders the log structure.
"""
# `context` includes variables used to render the HTML.
context = {
'graph_width': 6000,
'graph_height': 6000,
'frame_width': 200,
'frame_height': 540,
'canvas_cols': 12,
}
log_paths = []
for path in paths.split(';'):
path = path.strip()
if not path:
continue
path_template = string.Template(path)
log_path = path_template.substitute(os.environ)
basename = os.path.basename(log_path)
if basename.startswith('(') and basename.endswith(')'):
dirname = os.path.dirname(log_path)
regex_pattern = re.compile(basename[1:-1]+'$')
filenames = os.listdir(dirname)
matched_files = [f for f in filenames if regex_pattern.match(f)]
log_paths += [os.path.join(dirname, f) for f in matched_files]
else:
log_paths.append(log_path)
if not log_paths:
context['errors'] = 'Cannot find log data'
else:
# Use the first log to index fields.
log_data = struct_tree.StructTree(
log_paths[0], fail_silently=True, readonly=True)
log_skeleton = log_data.Skeleton(depth=1)
d3_data = struct_tree.DictToD3Tree(log_skeleton, '/')
d3_data['expand_url'] = urlresolvers.reverse('browse_log', args=[''])
request.session['log_paths'] = log_paths
context['skeleton'] = json.dumps(d3_data)
order_horizontally = True
configs = _LoadConfigs()
scenarios = layout_base.AssembleLayout([
('Signals', [
widgets.DictLinesWidget('series', None, interactive=True,
use_markers=True),
]),
], desired_view_cols=1, order_horizontally=order_horizontally)
layout_names = loader.LayoutLoader().ModuleNames()
layout_names.sort()
configs['scenarios'] = scenarios
context.update(_PrepareContext(configs))
context['layout_names'] = layout_names
context['content_width'] = settings.CSS_GRID_COLUMNS - 2
context['order_horizontally'] = order_horizontally
_CreateAndAddClientIdToContext(context)
return shortcuts.render(request, template_name, context,
context_instance=template.RequestContext(request))
def PeriodicDataPoll(request, client_id, layout_name):
"""Compute realtime data and respond to periodic polling from a client layout.
Args:
request: An HttpRequest from the client.
client_id: The ID of the client's browser tab.
layout_name: Name of the layout associated with the client.
Returns:
An HttpResponse in the format of a serialized JSON object.
"""
aggregated_message = _GetMessage(request, client_id)
if not aggregated_message:
aggregated_message = struct_tree.StructTree(
{}, fail_silently=True, readonly=True)
layout = loader.LayoutLoader().GetLayoutByModuleName(layout_name)
tab_memory = layout_memory.GetMemory(client_id, False)
if tab_memory is not None:
# Load the persistent memory.
layout.Import(tab_memory)
else:
layout.Initialize()
tab_memory = layout_memory.GetMemory(client_id, True)
# Start the AIO receiver in case the server has restarted.
_TryToEnforceAioReceiver(client_id)
try:
data = layout.Filter(aggregated_message)
except Exception: # pylint: disable=broad-except
# layout.Filter may introduce any kind of exception.
logging.error('PeriodicDataPoll encountered an error:\n%s',
debug_util.FormatTraceback())
layout.Export(tab_memory)
return http.HttpResponse('{}')
# Save the persistent memory.
layout.Export(tab_memory)
resp = data.Json()
if settings.DEBUG:
resp['__message__'] = '\n-----------------------------\n'.join(
'Error in indicator "%s":\n%s' % (k, v)
for k, v in layout.ErrorReport())
resp_str = json.dumps(resp)
layout.ClearErrors()
return http.HttpResponse(resp_str)
def _DownSample(data, length):
window_size = max(1, len(data)/length)
if window_size > 1:
data = data[:len(data) / window_size * window_size]
return numpy.mean(data.reshape(-1, window_size), 1), window_size
else:
return data, 1
def GetLogData(request, mode, fields):
"""Get values of data fields within a log file."""
log_paths = request.session['log_paths']
fields = [f.strip() for f in fields.split('\n') if f.strip()]
field_labels = layout_util.GetDistinguishableNames(
fields, '.', ['kAioNode', 'kMessageType'])
if mode == 'merge':
series = ConcatenateLogData(log_paths, field_labels)
else: # By default, mode = 'compare'
series = CompareLogData(log_paths, field_labels)
resp = {'series': series}
return http.HttpResponse(json.dumps(resp))
def _StringReplace(subject, translate):
for s, t in translate:
subject = subject.replace(s, t)
return subject
def GetMessageSnapshot(request, client_id, title):
aggregated_message = _GetMessage(request, client_id)
result = aggregated_message.Data(True)
response = http.HttpResponse(content_type='text/plain')
response['Content-Disposition'] = (
'attachment; filename=snapshot_%s.json' % title)
response.write(json.dumps(result, indent=2))
return response
def GetRawLogData(request, fields):
"""Get values of data fields within a log file."""
log_paths = request.session['log_paths']
fields = [f.strip() for f in fields.split('\n') if f.strip()]
field_labels = layout_util.GetDistinguishableNames(
fields, '.', ['kAioNode', 'kMessageType'])
result = {}
# Remove special characters so variables can be parsed and loaded into Matlab.
bad_chars = ['.', ',', '-', '+', '(', ')', '[', ']', '{', '}', ':',
'kMessageType', 'kAioNode', 'messages', 'message']
replacement = list(zip(bad_chars, ['_'] * len(bad_chars)))
replacement = [('[:]', ''), (':,', ''), (' ', '')] + replacement
for log_path in log_paths:
base_name = os.path.basename(log_path)
log_name = 'log_' + _StringReplace(base_name[:base_name.find('.')],
replacement)
log_data = struct_tree.StructTree(
log_path, fail_silently=True, readonly=True)
result[log_name] = {}
for field, legend_label in field_labels.iteritems():
data, timestamps = log_util.GetOrderedDedupDataAndTimeByField(
log_data, field, rebase=False)
result[log_name][_StringReplace(legend_label, replacement)] = {
'values': data.tolist() if data is not None else None,
'timestamps': timestamps.tolist() if timestamps is not None else None,
'status': 'success' if data is not None else 'missing',
}
response = http.HttpResponse(content_type='text/plain')
response['Content-Disposition'] = 'attachment; filename=makani_log_data.json'
response.write(json.dumps(result, indent=2))
return response
def ConcatenateLogData(log_paths, field_labels):
"""Get series of data, each corresponding to field values in all logs."""
series = {}
base_timeline = float('inf')
for log_path in log_paths:
log_data = struct_tree.StructTree(
log_path, fail_silently=True, readonly=True)
for field, legend_label in field_labels.iteritems():
data, timestamps = log_util.GetOrderedDedupDataAndTimeByField(
log_data, field, rebase=False)
if data is None or timestamps is None:
continue
base_timeline = min(base_timeline, float(timestamps[0]))
if legend_label not in series:
series[legend_label] = {'x': timestamps, 'y': data}
else:
series[legend_label]['x'] = numpy.concatenate(
(series[legend_label]['x'], timestamps))
series[legend_label]['y'] = numpy.concatenate(
(series[legend_label]['y'], data))
result = {}
for field, legend_label in field_labels.iteritems():
timestamps, _ = _DownSample(
series[legend_label]['x'], settings.MAX_DATA_POINTS_PER_LOG_FIELD)
data, downsample_rate = _DownSample(
series[legend_label]['y'], settings.MAX_DATA_POINTS_PER_LOG_FIELD)
if downsample_rate > 1:
legend_label += '(/%d)' % downsample_rate
result[legend_label] = {'x': (timestamps - base_timeline).tolist(),
'y': data.tolist()}
return result
def CompareLogData(log_paths, field_labels):
"""Get series of data, each corresponding to field values within a log."""
series = {}
base_timeline = float('inf')
for log_path in log_paths:
log_data = struct_tree.StructTree(
log_path, fail_silently=True, readonly=True)
log_name = os.path.basename(log_path)
if '.' in log_name:
log_name = log_name[:log_name.rfind('.')]
for field, legend_label in field_labels.iteritems():
data, timestamps = log_util.GetOrderedDedupDataAndTimeByField(
log_data, field, rebase=True)
if data is None or timestamps is None:
continue
data, _ = _DownSample(data, settings.MAX_DATA_POINTS_PER_LOG_FIELD)
timestamps, downsample_rate = _DownSample(
timestamps, settings.MAX_DATA_POINTS_PER_LOG_FIELD)
base_timeline = min(base_timeline, float(timestamps[0]))
short_name = '%s.%s' % (log_name, legend_label)
if downsample_rate > 1:
short_name += '(/%d)' % downsample_rate
series[short_name] = {'x': timestamps,
'y': data.tolist()}
for short_name in series:
series[short_name]['x'] = (series[short_name]['x'] - base_timeline).tolist()
return series
def PeriodicMessagePoll(request, client_id, message_type=None):
"""Retrieve realtime data and respond to periodic polling from a message view.
Args:
request: An HttpRequest from the client.
client_id: The ID of the client's browser tab.
message_type: The Enum name of a message type.
Returns:
An HttpResponse in the format of a serialized JSON object.
"""
resp = _GetMessage(request, client_id, message_type)
if not resp:
resp = {}
else:
resp = resp.Data(convert_to_basic_types=True)
resp_str = json.dumps(resp)
return http.HttpResponse(resp_str)
def _LoadConfigs():
"""Load default layout configuration parameters."""
configs = {}
for cf, filename in CONFIG_FILES.iteritems():
with open(filename, 'r') as fp:
configs[cf] = json.load(fp)
if 'plot_defs' not in configs:
logging.Error('Missing definitions for plotting javascripts.')
return configs
def _PrepareContext(configs):
"""Prepare the context to render the layout."""
context = {}
fig_templates = set()
canvas_cols = configs['scenarios']['canvas']['grid_width']
context['canvas_cols'] = canvas_cols
row_height_px = configs['scenarios']['canvas']['row_height_px']
ui_objs = []
max_cols = canvas_cols
for stripe in configs['scenarios']['views']:
for view in stripe['stripe']:
view['canvas_cols'] = int(
float(view['grid_width']) / stripe['grid_width'] * canvas_cols + 0.5)
for indicator in view['indicators']:
ui_obj = indicator
if 'rows' not in ui_obj:
ui_obj['height'] = 'auto'
else:
rows = ui_obj['rows']
ui_obj['height'] = str(rows * row_height_px) + 'px'
if 'cols' not in ui_obj:
ui_obj['cols'] = max_cols
# TODO: Change `id` to 'indicator_id', and 'selector'
# to 'dom_selector'.
ui_obj['id'] = 'ui_obj_%s' % len(ui_objs)
ui_obj['selector'] = '#%s' % (ui_obj['id'])
ui_objs.append(ui_obj)
fig_templates.add(ui_obj['template'])
context['fig_templates'] = fig_templates
context['plot_defs'] = configs['plot_defs']
context['views'] = configs['scenarios']['views']
context['ui_objs_str'] = json.dumps(ui_objs)
context['stoplight_error'] = stoplights.STOPLIGHT_ERROR
context['stoplight_warning'] = stoplights.STOPLIGHT_WARNING
context['stoplight_normal'] = stoplights.STOPLIGHT_NORMAL
context['stoplight_unavailable'] = stoplights.STOPLIGHT_UNAVAILABLE
context['stoplight_any'] = stoplights.STOPLIGHT_ANY
return context
def _GetMessage(unused_request, client_id, message_type=None):
"""Get a message from the receiver."""
message_receiver = receiver_manager.ReceiverManager.GetReceiver(client_id)
resp = struct_tree.StructTree({}, fail_silently=True, readonly=True)
if message_receiver:
if message_type is not None:
message_enum = MESSAGE_TYPE_HELPER.Value(message_type)
else:
message_enum = None
resp = message_receiver.GetLatest(message_enum)
return resp
def _CreateAndAddClientIdToContext(context):
client_id = receiver_manager.ReceiverManager.GetNewClientId()
context['client_id'] = client_id
return client_id
def _ViewLayout(request, layout_name, extra_context=None):
"""Get a monitor layout according to `layout_name`."""
layout = loader.LayoutLoader().GetLayoutByModuleName(layout_name)
if layout is None:
return http.HttpResponseRedirect(urlresolvers.reverse('home'))
layout.Initialize()
configs = _LoadConfigs()
configs['scenarios'] = layout.Scenario()
context = _PrepareContext(configs)
client_id = _CreateAndAddClientIdToContext(context)
# Initialize the layout.
layout.Export(layout_memory.GetMemory(client_id, True))
# Add polling URL.
context['periodic_url'] = '/dashboard/periodic/layout/%s/%s' % (client_id,
layout_name)
context['layout_name'] = layout_name
context['content_width'] = settings.CSS_GRID_COLUMNS
context['order_horizontally'] = layout.OrderHorizontally()
context['default_font_size'] = layout.DefaultFontSize()
context['sim_mode'] = settings.POPULATE_MESSAGES_FROM_SIM
if extra_context:
context.update(extra_context)
template_name = 'monitor.html'
return shortcuts.render(request, template_name, context,
context_instance=template.RequestContext(request))
|
[
"logging.Error",
"re.compile",
"makani.lib.python.debug_util.FormatTraceback",
"makani.gs.monitor2.apps.layout.autogen.GenerateScenario",
"django.core.urlresolvers.reverse",
"makani.analysis.checks.log_util.GetOrderedDedupDataAndTimeByField",
"makani.lib.python.c_helpers.EnumHelper",
"os.listdir",
"makani.gs.monitor2.apps.receiver.receiver_manager.ReceiverManager.GetReceiver",
"django.http.HttpResponse",
"json.dumps",
"makani.gs.monitor2.apps.layout.widgets.DictLinesWidget",
"makani.gs.monitor2.apps.receiver.receiver_manager.ReceiverManager.CheckAndStartAioReceiver",
"os.path.isdir",
"numpy.concatenate",
"makani.lib.python.struct_tree.DictToD3Tree",
"makani.gs.monitor2.apps.layout.memory.GetMemory",
"os.path.dirname",
"makani.avionics.network.network_config.NetworkConfig",
"makani.lib.python.struct_tree.StructTree",
"string.Template",
"makani.gs.monitor2.apps.receiver.receiver_manager.ReceiverManager.GetNewClientId",
"makani.gs.monitor2.apps.layout.layout_util.GetDistinguishableNames",
"makani.lib.bazel.bazel_util.GetWorkspaceRoot",
"os.path.join",
"django.template.RequestContext",
"os.path.basename",
"json.load",
"makani.gs.monitor2.apps.layout.loader.LayoutLoader"
] |
[((1826, 1879), 'makani.lib.python.c_helpers.EnumHelper', 'c_helpers.EnumHelper', (['"""MessageType"""', 'aio_message_type'], {}), "('MessageType', aio_message_type)\n", (1846, 1879), False, 'from makani.lib.python import c_helpers\n'), ((1915, 1976), 'os.path.join', 'os.path.join', (['settings.MONITOR_PATH', '"""configs/plot_defs.json"""'], {}), "(settings.MONITOR_PATH, 'configs/plot_defs.json')\n", (1927, 1976), False, 'import os\n'), ((2696, 2721), 'string.Template', 'string.Template', (['path_arg'], {}), '(path_arg)\n', (2711, 2721), False, 'import string\n'), ((2789, 2812), 'os.listdir', 'os.listdir', (['prefix_path'], {}), '(prefix_path)\n', (2799, 2812), False, 'import os\n'), ((2901, 2936), 'os.path.join', 'os.path.join', (['prefix_path', 'sub_path'], {}), '(prefix_path, sub_path)\n', (2913, 2936), False, 'import os\n'), ((4051, 4072), 'string.Template', 'string.Template', (['args'], {}), '(args)\n', (4066, 4072), False, 'import string\n'), ((5151, 5174), 'django.http.HttpResponse', 'http.HttpResponse', (['text'], {}), '(text)\n', (5168, 5174), False, 'from django import http\n'), ((5279, 5330), 'makani.avionics.network.network_config.NetworkConfig', 'network_config.NetworkConfig', (['settings.NETWORK_YAML'], {}), '(settings.NETWORK_YAML)\n', (5307, 5330), False, 'from makani.avionics.network import network_config\n'), ((5594, 5649), 'makani.gs.monitor2.apps.receiver.receiver_manager.ReceiverManager.GetReceiver', 'receiver_manager.ReceiverManager.GetReceiver', (['client_id'], {}), '(client_id)\n', (5638, 5649), False, 'from makani.gs.monitor2.apps.receiver import receiver_manager\n'), ((6804, 6848), 'makani.gs.monitor2.apps.layout.autogen.GenerateScenario', 'autogen.GenerateScenario', (['resp', 'message_type'], {}), '(resp, message_type)\n', (6828, 6848), False, 'from makani.gs.monitor2.apps.layout import autogen\n'), ((8616, 8683), 'makani.lib.python.struct_tree.StructTree', 'struct_tree.StructTree', (['log_path'], {'fail_silently': '(True)', 'readonly': '(True)'}), '(log_path, fail_silently=True, readonly=True)\n', (8638, 8683), False, 'from makani.lib.python import struct_tree\n'), ((8836, 8888), 'makani.lib.python.struct_tree.DictToD3Tree', 'struct_tree.DictToD3Tree', (['skeleton', '"""."""', 'parent_path'], {}), "(skeleton, '.', parent_path)\n", (8860, 8888), False, 'from makani.lib.python import struct_tree\n'), ((12211, 12252), 'makani.gs.monitor2.apps.layout.memory.GetMemory', 'layout_memory.GetMemory', (['client_id', '(False)'], {}), '(client_id, False)\n', (12234, 12252), True, 'from makani.gs.monitor2.apps.layout import memory as layout_memory\n'), ((13150, 13166), 'json.dumps', 'json.dumps', (['resp'], {}), '(resp)\n', (13160, 13166), False, 'import json\n'), ((13199, 13226), 'django.http.HttpResponse', 'http.HttpResponse', (['resp_str'], {}), '(resp_str)\n', (13216, 13226), False, 'from django import http\n'), ((13694, 13772), 'makani.gs.monitor2.apps.layout.layout_util.GetDistinguishableNames', 'layout_util.GetDistinguishableNames', (['fields', '"""."""', "['kAioNode', 'kMessageType']"], {}), "(fields, '.', ['kAioNode', 'kMessageType'])\n", (13729, 13772), False, 'from makani.gs.monitor2.apps.layout import layout_util\n'), ((14308, 14352), 'django.http.HttpResponse', 'http.HttpResponse', ([], {'content_type': '"""text/plain"""'}), "(content_type='text/plain')\n", (14325, 14352), False, 'from django import http\n'), ((14726, 14804), 'makani.gs.monitor2.apps.layout.layout_util.GetDistinguishableNames', 'layout_util.GetDistinguishableNames', (['fields', '"""."""', "['kAioNode', 'kMessageType']"], {}), "(fields, '.', ['kAioNode', 'kMessageType'])\n", (14761, 14804), False, 'from makani.gs.monitor2.apps.layout import layout_util\n'), ((15957, 16001), 'django.http.HttpResponse', 'http.HttpResponse', ([], {'content_type': '"""text/plain"""'}), "(content_type='text/plain')\n", (15974, 16001), False, 'from django import http\n'), ((19326, 19342), 'json.dumps', 'json.dumps', (['resp'], {}), '(resp)\n', (19336, 19342), False, 'import json\n'), ((19352, 19379), 'django.http.HttpResponse', 'http.HttpResponse', (['resp_str'], {}), '(resp_str)\n', (19369, 19379), False, 'from django import http\n'), ((20984, 21003), 'json.dumps', 'json.dumps', (['ui_objs'], {}), '(ui_objs)\n', (20994, 21003), False, 'import json\n'), ((21452, 21507), 'makani.gs.monitor2.apps.receiver.receiver_manager.ReceiverManager.GetReceiver', 'receiver_manager.ReceiverManager.GetReceiver', (['client_id'], {}), '(client_id)\n', (21496, 21507), False, 'from makani.gs.monitor2.apps.receiver import receiver_manager\n'), ((21517, 21578), 'makani.lib.python.struct_tree.StructTree', 'struct_tree.StructTree', (['{}'], {'fail_silently': '(True)', 'readonly': '(True)'}), '({}, fail_silently=True, readonly=True)\n', (21539, 21578), False, 'from makani.lib.python import struct_tree\n'), ((21859, 21908), 'makani.gs.monitor2.apps.receiver.receiver_manager.ReceiverManager.GetNewClientId', 'receiver_manager.ReceiverManager.GetNewClientId', ([], {}), '()\n', (21906, 21908), False, 'from makani.gs.monitor2.apps.receiver import receiver_manager\n'), ((4006, 4032), 'django.http.HttpResponse', 'http.HttpResponse', (['message'], {}), '(message)\n', (4023, 4032), False, 'from django import http\n'), ((4513, 4537), 'os.path.isdir', 'os.path.isdir', (['full_path'], {}), '(full_path)\n', (4526, 4537), False, 'import os\n'), ((5684, 5790), 'makani.gs.monitor2.apps.receiver.receiver_manager.ReceiverManager.CheckAndStartAioReceiver', 'receiver_manager.ReceiverManager.CheckAndStartAioReceiver', (['client_id', 'receiver_views.CreateAioReceiver'], {}), '(client_id,\n receiver_views.CreateAioReceiver)\n', (5741, 5790), False, 'from makani.gs.monitor2.apps.receiver import receiver_manager\n'), ((6017, 6072), 'makani.gs.monitor2.apps.receiver.receiver_manager.ReceiverManager.GetReceiver', 'receiver_manager.ReceiverManager.GetReceiver', (['client_id'], {}), '(client_id)\n', (6061, 6072), False, 'from makani.gs.monitor2.apps.receiver import receiver_manager\n'), ((7772, 7788), 'json.dumps', 'json.dumps', (['info'], {}), '(info)\n', (7782, 7788), False, 'import json\n'), ((9068, 9091), 'django.http.HttpResponse', 'http.HttpResponse', (['"""{}"""'], {}), "('{}')\n", (9085, 9091), False, 'from django import http\n'), ((9787, 9808), 'string.Template', 'string.Template', (['path'], {}), '(path)\n', (9802, 9808), False, 'import string\n'), ((9876, 9902), 'os.path.basename', 'os.path.basename', (['log_path'], {}), '(log_path)\n', (9892, 9902), False, 'import os\n'), ((10411, 10482), 'makani.lib.python.struct_tree.StructTree', 'struct_tree.StructTree', (['log_paths[0]'], {'fail_silently': '(True)', 'readonly': '(True)'}), '(log_paths[0], fail_silently=True, readonly=True)\n', (10433, 10482), False, 'from makani.lib.python import struct_tree\n'), ((10552, 10595), 'makani.lib.python.struct_tree.DictToD3Tree', 'struct_tree.DictToD3Tree', (['log_skeleton', '"""/"""'], {}), "(log_skeleton, '/')\n", (10576, 10595), False, 'from makani.lib.python import struct_tree\n'), ((10624, 10669), 'django.core.urlresolvers.reverse', 'urlresolvers.reverse', (['"""browse_log"""'], {'args': "['']"}), "('browse_log', args=[''])\n", (10644, 10669), False, 'from django.core import urlresolvers\n'), ((10741, 10760), 'json.dumps', 'json.dumps', (['d3_data'], {}), '(d3_data)\n', (10751, 10760), False, 'import json\n'), ((12057, 12118), 'makani.lib.python.struct_tree.StructTree', 'struct_tree.StructTree', (['{}'], {'fail_silently': '(True)', 'readonly': '(True)'}), '({}, fail_silently=True, readonly=True)\n', (12079, 12118), False, 'from makani.lib.python import struct_tree\n'), ((12395, 12435), 'makani.gs.monitor2.apps.layout.memory.GetMemory', 'layout_memory.GetMemory', (['client_id', '(True)'], {}), '(client_id, True)\n', (12418, 12435), True, 'from makani.gs.monitor2.apps.layout import memory as layout_memory\n'), ((14008, 14024), 'json.dumps', 'json.dumps', (['resp'], {}), '(resp)\n', (14018, 14024), False, 'import json\n'), ((14463, 14491), 'json.dumps', 'json.dumps', (['result'], {'indent': '(2)'}), '(result, indent=2)\n', (14473, 14491), False, 'import json\n'), ((15217, 15243), 'os.path.basename', 'os.path.basename', (['log_path'], {}), '(log_path)\n', (15233, 15243), False, 'import os\n'), ((15383, 15450), 'makani.lib.python.struct_tree.StructTree', 'struct_tree.StructTree', (['log_path'], {'fail_silently': '(True)', 'readonly': '(True)'}), '(log_path, fail_silently=True, readonly=True)\n', (15405, 15450), False, 'from makani.lib.python import struct_tree\n'), ((16099, 16127), 'json.dumps', 'json.dumps', (['result'], {'indent': '(2)'}), '(result, indent=2)\n', (16109, 16127), False, 'import json\n'), ((16363, 16430), 'makani.lib.python.struct_tree.StructTree', 'struct_tree.StructTree', (['log_path'], {'fail_silently': '(True)', 'readonly': '(True)'}), '(log_path, fail_silently=True, readonly=True)\n', (16385, 16430), False, 'from makani.lib.python import struct_tree\n'), ((17770, 17837), 'makani.lib.python.struct_tree.StructTree', 'struct_tree.StructTree', (['log_path'], {'fail_silently': '(True)', 'readonly': '(True)'}), '(log_path, fail_silently=True, readonly=True)\n', (17792, 17837), False, 'from makani.lib.python import struct_tree\n'), ((17862, 17888), 'os.path.basename', 'os.path.basename', (['log_path'], {}), '(log_path)\n', (17878, 17888), False, 'import os\n'), ((19626, 19688), 'logging.Error', 'logging.Error', (['"""Missing definitions for plotting javascripts."""'], {}), "('Missing definitions for plotting javascripts.')\n", (19639, 19688), False, 'import logging\n'), ((22463, 22503), 'makani.gs.monitor2.apps.layout.memory.GetMemory', 'layout_memory.GetMemory', (['client_id', '(True)'], {}), '(client_id, True)\n', (22486, 22503), True, 'from makani.gs.monitor2.apps.layout import memory as layout_memory\n'), ((2063, 2084), 'makani.gs.monitor2.apps.layout.loader.LayoutLoader', 'loader.LayoutLoader', ([], {}), '()\n', (2082, 2084), False, 'from makani.gs.monitor2.apps.layout import loader\n'), ((2577, 2609), 'django.template.RequestContext', 'template.RequestContext', (['request'], {}), '(request)\n', (2600, 2609), False, 'from django import template\n'), ((3161, 3224), 'django.http.HttpResponse', 'http.HttpResponse', (['(\'Cannot list directory "%s"!\' % current_path)'], {}), '(\'Cannot list directory "%s"!\' % current_path)\n', (3178, 3224), False, 'from django import http\n'), ((4138, 4167), 'makani.lib.bazel.bazel_util.GetWorkspaceRoot', 'bazel_util.GetWorkspaceRoot', ([], {}), '()\n', (4165, 4167), False, 'from makani.lib.bazel import bazel_util\n'), ((4304, 4363), 'django.http.HttpResponse', 'http.HttpResponse', (['(\'Cannot list directory "%s"!\' % arg_path)'], {}), '(\'Cannot list directory "%s"!\' % arg_path)\n', (4321, 4363), False, 'from django import http\n'), ((7244, 7276), 'django.template.RequestContext', 'template.RequestContext', (['request'], {}), '(request)\n', (7267, 7276), False, 'from django import template\n'), ((8779, 8802), 'django.http.HttpResponse', 'http.HttpResponse', (['"""{}"""'], {}), "('{}')\n", (8796, 8802), False, 'from django import http\n'), ((9016, 9047), 'json.dumps', 'json.dumps', (["d3_data['children']"], {}), "(d3_data['children'])\n", (9026, 9047), False, 'import json\n'), ((9979, 10004), 'os.path.dirname', 'os.path.dirname', (['log_path'], {}), '(log_path)\n', (9994, 10004), False, 'import os\n'), ((10027, 10059), 're.compile', 're.compile', (["(basename[1:-1] + '$')"], {}), "(basename[1:-1] + '$')\n", (10037, 10059), False, 'import re\n'), ((10076, 10095), 'os.listdir', 'os.listdir', (['dirname'], {}), '(dirname)\n', (10086, 10095), False, 'import os\n'), ((11093, 11114), 'makani.gs.monitor2.apps.layout.loader.LayoutLoader', 'loader.LayoutLoader', ([], {}), '()\n', (11112, 11114), False, 'from makani.gs.monitor2.apps.layout import loader\n'), ((11526, 11558), 'django.template.RequestContext', 'template.RequestContext', (['request'], {}), '(request)\n', (11549, 11558), False, 'from django import template\n'), ((12139, 12160), 'makani.gs.monitor2.apps.layout.loader.LayoutLoader', 'loader.LayoutLoader', ([], {}), '()\n', (12158, 12160), False, 'from makani.gs.monitor2.apps.layout import loader\n'), ((12852, 12875), 'django.http.HttpResponse', 'http.HttpResponse', (['"""{}"""'], {}), "('{}')\n", (12869, 12875), False, 'from django import http\n'), ((15568, 15641), 'makani.analysis.checks.log_util.GetOrderedDedupDataAndTimeByField', 'log_util.GetOrderedDedupDataAndTimeByField', (['log_data', 'field'], {'rebase': '(False)'}), '(log_data, field, rebase=False)\n', (15610, 15641), False, 'from makani.analysis.checks import log_util\n'), ((16522, 16595), 'makani.analysis.checks.log_util.GetOrderedDedupDataAndTimeByField', 'log_util.GetOrderedDedupDataAndTimeByField', (['log_data', 'field'], {'rebase': '(False)'}), '(log_data, field, rebase=False)\n', (16564, 16595), False, 'from makani.analysis.checks import log_util\n'), ((18043, 18115), 'makani.analysis.checks.log_util.GetOrderedDedupDataAndTimeByField', 'log_util.GetOrderedDedupDataAndTimeByField', (['log_data', 'field'], {'rebase': '(True)'}), '(log_data, field, rebase=True)\n', (18085, 18115), False, 'from makani.analysis.checks import log_util\n'), ((19575, 19588), 'json.load', 'json.load', (['fp'], {}), '(fp)\n', (19584, 19588), False, 'import json\n'), ((22092, 22113), 'makani.gs.monitor2.apps.layout.loader.LayoutLoader', 'loader.LayoutLoader', ([], {}), '()\n', (22111, 22113), False, 'from makani.gs.monitor2.apps.layout import loader\n'), ((22207, 22235), 'django.core.urlresolvers.reverse', 'urlresolvers.reverse', (['"""home"""'], {}), "('home')\n", (22227, 22235), False, 'from django.core import urlresolvers\n'), ((23146, 23178), 'django.template.RequestContext', 'template.RequestContext', (['request'], {}), '(request)\n', (23169, 23178), False, 'from django import template\n'), ((3453, 3477), 'os.path.isdir', 'os.path.isdir', (['full_path'], {}), '(full_path)\n', (3466, 3477), False, 'import os\n'), ((10187, 10211), 'os.path.join', 'os.path.join', (['dirname', 'f'], {}), '(dirname, f)\n', (10199, 10211), False, 'import os\n'), ((12780, 12808), 'makani.lib.python.debug_util.FormatTraceback', 'debug_util.FormatTraceback', ([], {}), '()\n', (12806, 12808), False, 'from makani.lib.python import debug_util\n'), ((16877, 16935), 'numpy.concatenate', 'numpy.concatenate', (["(series[legend_label]['x'], timestamps)"], {}), "((series[legend_label]['x'], timestamps))\n", (16894, 16935), False, 'import numpy\n'), ((16985, 17037), 'numpy.concatenate', 'numpy.concatenate', (["(series[legend_label]['y'], data)"], {}), "((series[legend_label]['y'], data))\n", (17002, 17037), False, 'import numpy\n'), ((10890, 10965), 'makani.gs.monitor2.apps.layout.widgets.DictLinesWidget', 'widgets.DictLinesWidget', (['"""series"""', 'None'], {'interactive': '(True)', 'use_markers': '(True)'}), "('series', None, interactive=True, use_markers=True)\n", (10913, 10965), False, 'from makani.gs.monitor2.apps.layout import widgets\n'), ((2228, 2249), 'makani.gs.monitor2.apps.layout.loader.LayoutLoader', 'loader.LayoutLoader', ([], {}), '()\n', (2247, 2249), False, 'from makani.gs.monitor2.apps.layout import loader\n')]
|
# coding=utf-8
# Copyright 2018 The TF-Agents Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Tests for tf_agents.bandits.agents.neural_linucb_agent."""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import os
from absl.testing import parameterized
import numpy as np
import tensorflow as tf # pylint: disable=g-explicit-tensorflow-version-import
import tensorflow_probability as tfp
from tf_agents.bandits.agents import neural_linucb_agent
from tf_agents.bandits.agents import utils as bandit_utils
from tf_agents.bandits.drivers import driver_utils
from tf_agents.bandits.networks import global_and_arm_feature_network
from tf_agents.bandits.policies import policy_utilities
from tf_agents.bandits.specs import utils as bandit_spec_utils
from tf_agents.networks import network
from tf_agents.specs import tensor_spec
from tf_agents.trajectories import policy_step
from tf_agents.trajectories import time_step
from tf_agents.utils import common
from tensorflow.python.framework import test_util # pylint: disable=g-direct-tensorflow-import # TF internal
tfd = tfp.distributions
class DummyNet(network.Network):
def __init__(self, observation_spec, encoding_dim=10):
super(DummyNet, self).__init__(
observation_spec, state_spec=(), name='DummyNet')
context_dim = observation_spec.shape[0]
# Store custom layers that can be serialized through the Checkpointable API.
self._dummy_layers = [
tf.keras.layers.Dense(
encoding_dim,
kernel_initializer=tf.compat.v1.initializers.constant(
np.ones([context_dim, encoding_dim])),
bias_initializer=tf.compat.v1.initializers.constant(
np.zeros([encoding_dim])))
]
def call(self, inputs, step_type=None, network_state=()):
del step_type
inputs = tf.cast(inputs, tf.float32)
for layer in self._dummy_layers:
inputs = layer(inputs)
return inputs, network_state
def test_cases():
return parameterized.named_parameters(
{
'testcase_name': '_batch1_contextdim10',
'batch_size': 1,
'context_dim': 10,
}, {
'testcase_name': '_batch4_contextdim5',
'batch_size': 4,
'context_dim': 5,
})
def _get_initial_and_final_steps(batch_size, context_dim):
observation = np.array(range(batch_size * context_dim)).reshape(
[batch_size, context_dim])
reward = np.random.uniform(0.0, 1.0, [batch_size])
initial_step = time_step.TimeStep(
tf.constant(
time_step.StepType.FIRST, dtype=tf.int32, shape=[batch_size],
name='step_type'),
tf.constant(0.0, dtype=tf.float32, shape=[batch_size], name='reward'),
tf.constant(1.0, dtype=tf.float32, shape=[batch_size], name='discount'),
tf.constant(observation, dtype=tf.float32,
shape=[batch_size, context_dim], name='observation'))
final_step = time_step.TimeStep(
tf.constant(
time_step.StepType.LAST, dtype=tf.int32, shape=[batch_size],
name='step_type'),
tf.constant(reward, dtype=tf.float32, shape=[batch_size], name='reward'),
tf.constant(1.0, dtype=tf.float32, shape=[batch_size], name='discount'),
tf.constant(observation + 100.0, dtype=tf.float32,
shape=[batch_size, context_dim], name='observation'))
return initial_step, final_step
def _get_initial_and_final_steps_with_action_mask(batch_size,
context_dim,
num_actions=None):
observation = np.array(range(batch_size * context_dim)).reshape(
[batch_size, context_dim])
observation = tf.constant(observation, dtype=tf.float32)
mask = 1 - tf.eye(batch_size, num_columns=num_actions, dtype=tf.int32)
reward = np.random.uniform(0.0, 1.0, [batch_size])
initial_step = time_step.TimeStep(
tf.constant(
time_step.StepType.FIRST,
dtype=tf.int32,
shape=[batch_size],
name='step_type'),
tf.constant(0.0, dtype=tf.float32, shape=[batch_size], name='reward'),
tf.constant(1.0, dtype=tf.float32, shape=[batch_size], name='discount'),
(observation, mask))
final_step = time_step.TimeStep(
tf.constant(
time_step.StepType.LAST,
dtype=tf.int32,
shape=[batch_size],
name='step_type'),
tf.constant(reward, dtype=tf.float32, shape=[batch_size], name='reward'),
tf.constant(1.0, dtype=tf.float32, shape=[batch_size], name='discount'),
(observation + 100.0, mask))
return initial_step, final_step
def _get_action_step(action):
return policy_step.PolicyStep(
action=tf.convert_to_tensor(action),
info=policy_utilities.PolicyInfo())
def _get_experience(initial_step, action_step, final_step):
single_experience = driver_utils.trajectory_for_bandit(
initial_step, action_step, final_step)
# Adds a 'time' dimension.
return tf.nest.map_structure(
lambda x: tf.expand_dims(tf.convert_to_tensor(x), 1),
single_experience)
@test_util.run_all_in_graph_and_eager_modes
class NeuralLinUCBAgentTest(tf.test.TestCase, parameterized.TestCase):
def setUp(self):
super(NeuralLinUCBAgentTest, self).setUp()
tf.compat.v1.enable_resource_variables()
@test_cases()
def testInitializeAgentNumTrainSteps0(self, batch_size, context_dim):
num_actions = 5
observation_spec = tensor_spec.TensorSpec([context_dim], tf.float32)
time_step_spec = time_step.time_step_spec(observation_spec)
action_spec = tensor_spec.BoundedTensorSpec(
dtype=tf.int32, shape=(), minimum=0, maximum=num_actions - 1)
encoder = DummyNet(observation_spec)
agent = neural_linucb_agent.NeuralLinUCBAgent(
time_step_spec=time_step_spec,
action_spec=action_spec,
encoding_network=encoder,
encoding_network_num_train_steps=0,
encoding_dim=10,
optimizer=None)
self.evaluate(agent.initialize())
@test_cases()
def testInitializeAgentNumTrainSteps10(self, batch_size, context_dim):
num_actions = 5
observation_spec = tensor_spec.TensorSpec([context_dim], tf.float32)
time_step_spec = time_step.time_step_spec(observation_spec)
action_spec = tensor_spec.BoundedTensorSpec(
dtype=tf.int32, shape=(), minimum=0, maximum=num_actions - 1)
encoder = DummyNet(observation_spec)
agent = neural_linucb_agent.NeuralLinUCBAgent(
time_step_spec=time_step_spec,
action_spec=action_spec,
encoding_network=encoder,
encoding_network_num_train_steps=10,
encoding_dim=10,
optimizer=None)
self.evaluate(agent.initialize())
@test_cases()
def testNeuralLinUCBUpdateNumTrainSteps0(self, batch_size=1, context_dim=10):
"""Check NeuralLinUCBAgent updates when behaving like LinUCB."""
# Construct a `Trajectory` for the given action, observation, reward.
num_actions = 5
initial_step, final_step = _get_initial_and_final_steps(
batch_size, context_dim)
action = np.random.randint(num_actions, size=batch_size, dtype=np.int32)
action_step = _get_action_step(action)
experience = _get_experience(initial_step, action_step, final_step)
# Construct an agent and perform the update.
observation_spec = tensor_spec.TensorSpec([context_dim], tf.float32)
time_step_spec = time_step.time_step_spec(observation_spec)
action_spec = tensor_spec.BoundedTensorSpec(
dtype=tf.int32, shape=(), minimum=0, maximum=num_actions - 1)
encoder = DummyNet(observation_spec)
encoding_dim = 10
agent = neural_linucb_agent.NeuralLinUCBAgent(
time_step_spec=time_step_spec,
action_spec=action_spec,
encoding_network=encoder,
encoding_network_num_train_steps=0,
encoding_dim=encoding_dim,
optimizer=tf.compat.v1.train.AdamOptimizer(learning_rate=1e-2))
loss_info = agent.train(experience)
self.evaluate(agent.initialize())
self.evaluate(tf.compat.v1.global_variables_initializer())
self.evaluate(loss_info)
final_a = self.evaluate(agent.cov_matrix)
final_b = self.evaluate(agent.data_vector)
# Compute the expected updated estimates.
observations_list = tf.dynamic_partition(
data=tf.reshape(tf.cast(experience.observation, tf.float64),
[batch_size, context_dim]),
partitions=tf.convert_to_tensor(action),
num_partitions=num_actions)
rewards_list = tf.dynamic_partition(
data=tf.reshape(tf.cast(experience.reward, tf.float64), [batch_size]),
partitions=tf.convert_to_tensor(action),
num_partitions=num_actions)
expected_a_updated_list = []
expected_b_updated_list = []
for _, (observations_for_arm, rewards_for_arm) in enumerate(zip(
observations_list, rewards_list)):
encoded_observations_for_arm, _ = encoder(observations_for_arm)
encoded_observations_for_arm = tf.cast(
encoded_observations_for_arm, dtype=tf.float64)
num_samples_for_arm_current = tf.cast(
tf.shape(rewards_for_arm)[0], tf.float64)
num_samples_for_arm_total = num_samples_for_arm_current
# pylint: disable=cell-var-from-loop
def true_fn():
a_new = tf.matmul(
encoded_observations_for_arm,
encoded_observations_for_arm,
transpose_a=True)
b_new = bandit_utils.sum_reward_weighted_observations(
rewards_for_arm, encoded_observations_for_arm)
return a_new, b_new
def false_fn():
return (tf.zeros([encoding_dim, encoding_dim], dtype=tf.float64),
tf.zeros([encoding_dim], dtype=tf.float64))
a_new, b_new = tf.cond(
tf.squeeze(num_samples_for_arm_total) > 0,
true_fn,
false_fn)
expected_a_updated_list.append(self.evaluate(a_new))
expected_b_updated_list.append(self.evaluate(b_new))
# Check that the actual updated estimates match the expectations.
self.assertAllClose(expected_a_updated_list, final_a)
self.assertAllClose(expected_b_updated_list, final_b)
@test_cases()
def testNeuralLinUCBUpdateDistributed(self, batch_size=1, context_dim=10):
"""Same as above but with distributed LinUCB updates."""
# Construct a `Trajectory` for the given action, observation, reward.
num_actions = 5
initial_step, final_step = _get_initial_and_final_steps(
batch_size, context_dim)
action = np.random.randint(num_actions, size=batch_size, dtype=np.int32)
action_step = _get_action_step(action)
experience = _get_experience(initial_step, action_step, final_step)
# Construct an agent and perform the update.
observation_spec = tensor_spec.TensorSpec([context_dim], tf.float32)
time_step_spec = time_step.time_step_spec(observation_spec)
action_spec = tensor_spec.BoundedTensorSpec(
dtype=tf.int32, shape=(), minimum=0, maximum=num_actions - 1)
encoder = DummyNet(observation_spec)
encoding_dim = 10
agent = neural_linucb_agent.NeuralLinUCBAgent(
time_step_spec=time_step_spec,
action_spec=action_spec,
encoding_network=encoder,
encoding_network_num_train_steps=0,
encoding_dim=encoding_dim,
optimizer=tf.compat.v1.train.AdamOptimizer(learning_rate=1e-2))
self.evaluate(agent.initialize())
self.evaluate(tf.compat.v1.global_variables_initializer())
# Call the distributed LinUCB training instead of agent.train().
train_fn = common.function_in_tf1()(
agent.compute_loss_using_linucb_distributed)
reward = tf.cast(experience.reward, agent._dtype)
loss_info = train_fn(
experience.observation, action, reward, weights=None)
self.evaluate(loss_info)
final_a = self.evaluate(agent.cov_matrix)
final_b = self.evaluate(agent.data_vector)
# Compute the expected updated estimates.
observations_list = tf.dynamic_partition(
data=tf.reshape(tf.cast(experience.observation, tf.float64),
[batch_size, context_dim]),
partitions=tf.convert_to_tensor(action),
num_partitions=num_actions)
rewards_list = tf.dynamic_partition(
data=tf.reshape(tf.cast(experience.reward, tf.float64), [batch_size]),
partitions=tf.convert_to_tensor(action),
num_partitions=num_actions)
expected_a_updated_list = []
expected_b_updated_list = []
for _, (observations_for_arm, rewards_for_arm) in enumerate(zip(
observations_list, rewards_list)):
encoded_observations_for_arm, _ = encoder(observations_for_arm)
encoded_observations_for_arm = tf.cast(
encoded_observations_for_arm, dtype=tf.float64)
num_samples_for_arm_current = tf.cast(
tf.shape(rewards_for_arm)[0], tf.float64)
num_samples_for_arm_total = num_samples_for_arm_current
# pylint: disable=cell-var-from-loop
def true_fn():
a_new = tf.matmul(
encoded_observations_for_arm,
encoded_observations_for_arm,
transpose_a=True)
b_new = bandit_utils.sum_reward_weighted_observations(
rewards_for_arm, encoded_observations_for_arm)
return a_new, b_new
def false_fn():
return (tf.zeros([encoding_dim, encoding_dim], dtype=tf.float64),
tf.zeros([encoding_dim], dtype=tf.float64))
a_new, b_new = tf.cond(
tf.squeeze(num_samples_for_arm_total) > 0,
true_fn,
false_fn)
expected_a_updated_list.append(self.evaluate(a_new))
expected_b_updated_list.append(self.evaluate(b_new))
# Check that the actual updated estimates match the expectations.
self.assertAllClose(expected_a_updated_list, final_a)
self.assertAllClose(expected_b_updated_list, final_b)
@test_cases()
def testNeuralLinUCBUpdateNumTrainSteps10(self, batch_size=1, context_dim=10):
"""Check NeuralLinUCBAgent updates when behaving like eps-greedy."""
# Construct a `Trajectory` for the given action, observation, reward.
num_actions = 5
initial_step, final_step = _get_initial_and_final_steps(
batch_size, context_dim)
action = np.random.randint(num_actions, size=batch_size, dtype=np.int32)
action_step = _get_action_step(action)
experience = _get_experience(initial_step, action_step, final_step)
# Construct an agent and perform the update.
observation_spec = tensor_spec.TensorSpec([context_dim], tf.float32)
time_step_spec = time_step.time_step_spec(observation_spec)
action_spec = tensor_spec.BoundedTensorSpec(
dtype=tf.int32, shape=(), minimum=0, maximum=num_actions - 1)
encoder = DummyNet(observation_spec)
encoding_dim = 10
variable_collection = neural_linucb_agent.NeuralLinUCBVariableCollection(
num_actions, encoding_dim)
agent = neural_linucb_agent.NeuralLinUCBAgent(
time_step_spec=time_step_spec,
action_spec=action_spec,
encoding_network=encoder,
encoding_network_num_train_steps=10,
encoding_dim=encoding_dim,
variable_collection=variable_collection,
optimizer=tf.compat.v1.train.AdamOptimizer(learning_rate=0.001))
loss_info, _ = agent.train(experience)
self.evaluate(agent.initialize())
self.evaluate(tf.compat.v1.global_variables_initializer())
loss_value = self.evaluate(loss_info)
self.assertGreater(loss_value, 0.0)
@test_cases()
def testNeuralLinUCBUpdateNumTrainSteps10MaskedActions(
self, batch_size=1, context_dim=10):
"""Check updates when behaving like eps-greedy and using masked actions."""
# Construct a `Trajectory` for the given action, observation, reward.
num_actions = 5
initial_step, final_step = _get_initial_and_final_steps_with_action_mask(
batch_size, context_dim, num_actions)
action = np.random.randint(num_actions, size=batch_size, dtype=np.int32)
action_step = _get_action_step(action)
experience = _get_experience(initial_step, action_step, final_step)
# Construct an agent and perform the update.
observation_spec = (tensor_spec.TensorSpec([context_dim], tf.float32),
tensor_spec.TensorSpec([num_actions], tf.int32))
time_step_spec = time_step.time_step_spec(observation_spec)
action_spec = tensor_spec.BoundedTensorSpec(
dtype=tf.int32, shape=(), minimum=0, maximum=num_actions - 1)
encoder = DummyNet(observation_spec[0])
encoding_dim = 10
agent = neural_linucb_agent.NeuralLinUCBAgent(
time_step_spec=time_step_spec,
action_spec=action_spec,
encoding_network=encoder,
encoding_network_num_train_steps=10,
encoding_dim=encoding_dim,
optimizer=tf.compat.v1.train.AdamOptimizer(learning_rate=0.001),
observation_and_action_constraint_splitter=lambda x: (x[0], x[1]))
loss_info, _ = agent.train(experience)
self.evaluate(agent.initialize())
self.evaluate(tf.compat.v1.global_variables_initializer())
loss_value = self.evaluate(loss_info)
self.assertGreater(loss_value, 0.0)
def testInitializeRestoreVariableCollection(self):
if not tf.executing_eagerly():
self.skipTest('Test only works in eager mode.')
num_actions = 5
encoding_dim = 7
variable_collection = neural_linucb_agent.NeuralLinUCBVariableCollection(
num_actions=num_actions, encoding_dim=encoding_dim)
self.evaluate(tf.compat.v1.global_variables_initializer())
self.evaluate(variable_collection.num_samples_list)
checkpoint = tf.train.Checkpoint(variable_collection=variable_collection)
checkpoint_dir = self.get_temp_dir()
checkpoint_prefix = os.path.join(checkpoint_dir, 'checkpoint')
checkpoint.save(file_prefix=checkpoint_prefix)
variable_collection.actions_from_reward_layer.assign(False)
latest_checkpoint = tf.train.latest_checkpoint(checkpoint_dir)
checkpoint_load_status = checkpoint.restore(latest_checkpoint)
self.evaluate(checkpoint_load_status.initialize_or_restore())
self.assertEqual(
self.evaluate(variable_collection.actions_from_reward_layer), True)
def testTrainPerArmAgentWithMask(self):
num_actions = 5
obs_spec = bandit_spec_utils.create_per_arm_observation_spec(
2, 3, num_actions, add_action_mask=True)
time_step_spec = time_step.time_step_spec(obs_spec)
action_spec = tensor_spec.BoundedTensorSpec(
dtype=tf.int32, shape=(), minimum=0, maximum=num_actions - 1)
encoding_dim = 10
encoder = (
global_and_arm_feature_network.create_feed_forward_common_tower_network(
obs_spec[0], (4, 3), (3, 4), (4, 2), encoding_dim))
agent = neural_linucb_agent.NeuralLinUCBAgent(
time_step_spec=time_step_spec,
action_spec=action_spec,
encoding_network=encoder,
encoding_network_num_train_steps=10,
encoding_dim=encoding_dim,
observation_and_action_constraint_splitter=lambda x: (x[0], x[1]),
accepts_per_arm_features=True,
optimizer=tf.compat.v1.train.AdamOptimizer(learning_rate=0.001))
observations = ({
bandit_spec_utils.GLOBAL_FEATURE_KEY:
tf.constant([[1, 2], [3, 4]], dtype=tf.float32),
bandit_spec_utils.PER_ARM_FEATURE_KEY:
tf.cast(
tf.reshape(tf.range(30), shape=[2, 5, 3]), dtype=tf.float32)
}, tf.ones(shape=(2, num_actions), dtype=tf.int32))
actions = np.array([0, 3], dtype=np.int32)
rewards = np.array([0.5, 3.0], dtype=np.float32)
initial_step = time_step.TimeStep(
tf.constant(
time_step.StepType.FIRST,
dtype=tf.int32,
shape=[2],
name='step_type'),
tf.constant(0.0, dtype=tf.float32, shape=[2], name='reward'),
tf.constant(1.0, dtype=tf.float32, shape=[2], name='discount'),
observations)
final_step = time_step.TimeStep(
tf.constant(
time_step.StepType.LAST,
dtype=tf.int32,
shape=[2],
name='step_type'),
tf.constant(rewards, dtype=tf.float32, name='reward'),
tf.constant(1.0, dtype=tf.float32, shape=[2], name='discount'),
observations)
action_step = policy_step.PolicyStep(
action=tf.convert_to_tensor(actions),
info=policy_utilities.PerArmPolicyInfo(
chosen_arm_features=np.array([[1, 2, 3], [3, 2, 1]],
dtype=np.float32)))
experience = _get_experience(initial_step, action_step, final_step)
loss_info, _ = agent.train(experience, None)
self.evaluate(tf.compat.v1.initialize_all_variables())
loss_value = self.evaluate(loss_info)
self.assertGreater(loss_value, 0.0)
def testTrainPerArmAgentVariableActions(self):
num_actions = 5
obs_spec = bandit_spec_utils.create_per_arm_observation_spec(
2, 3, num_actions, add_num_actions_feature=True)
time_step_spec = time_step.time_step_spec(obs_spec)
action_spec = tensor_spec.BoundedTensorSpec(
dtype=tf.int32, shape=(), minimum=0, maximum=num_actions - 1)
encoding_dim = 10
encoder = (
global_and_arm_feature_network.create_feed_forward_common_tower_network(
obs_spec, (4, 3), (3, 4), (4, 2), encoding_dim))
agent = neural_linucb_agent.NeuralLinUCBAgent(
time_step_spec=time_step_spec,
action_spec=action_spec,
encoding_network=encoder,
encoding_network_num_train_steps=10,
encoding_dim=encoding_dim,
accepts_per_arm_features=True,
optimizer=tf.compat.v1.train.AdamOptimizer(learning_rate=0.001))
observations = {
bandit_spec_utils.GLOBAL_FEATURE_KEY:
tf.constant([[1, 2], [3, 4]], dtype=tf.float32),
bandit_spec_utils.PER_ARM_FEATURE_KEY:
tf.cast(
tf.reshape(tf.range(30), shape=[2, 5, 3]), dtype=tf.float32),
bandit_spec_utils.NUM_ACTIONS_FEATURE_KEY:
tf.constant([3, 4], dtype=tf.int32)
}
actions = np.array([0, 3], dtype=np.int32)
rewards = np.array([0.5, 3.0], dtype=np.float32)
initial_step = time_step.TimeStep(
tf.constant(
time_step.StepType.FIRST,
dtype=tf.int32,
shape=[2],
name='step_type'),
tf.constant(0.0, dtype=tf.float32, shape=[2], name='reward'),
tf.constant(1.0, dtype=tf.float32, shape=[2], name='discount'),
observations)
final_step = time_step.TimeStep(
tf.constant(
time_step.StepType.LAST,
dtype=tf.int32,
shape=[2],
name='step_type'),
tf.constant(rewards, dtype=tf.float32, name='reward'),
tf.constant(1.0, dtype=tf.float32, shape=[2], name='discount'),
observations)
action_step = policy_step.PolicyStep(
action=tf.convert_to_tensor(actions),
info=policy_utilities.PerArmPolicyInfo(
chosen_arm_features=np.array([[1, 2, 3], [3, 2, 1]],
dtype=np.float32)))
experience = _get_experience(initial_step, action_step, final_step)
loss_info, _ = agent.train(experience, None)
self.evaluate(tf.compat.v1.initialize_all_variables())
loss_value = self.evaluate(loss_info)
self.assertGreater(loss_value, 0.0)
if __name__ == '__main__':
tf.test.main()
|
[
"tensorflow.train.Checkpoint",
"tensorflow.shape",
"tensorflow.compat.v1.train.AdamOptimizer",
"tf_agents.specs.tensor_spec.BoundedTensorSpec",
"numpy.array",
"tf_agents.trajectories.time_step.time_step_spec",
"tensorflow.cast",
"tensorflow.compat.v1.global_variables_initializer",
"tensorflow.eye",
"tf_agents.bandits.agents.neural_linucb_agent.NeuralLinUCBAgent",
"tensorflow.executing_eagerly",
"tf_agents.bandits.specs.utils.create_per_arm_observation_spec",
"tf_agents.bandits.drivers.driver_utils.trajectory_for_bandit",
"tf_agents.bandits.policies.policy_utilities.PolicyInfo",
"tensorflow.matmul",
"tf_agents.bandits.networks.global_and_arm_feature_network.create_feed_forward_common_tower_network",
"tensorflow.convert_to_tensor",
"tensorflow.zeros",
"tf_agents.specs.tensor_spec.TensorSpec",
"numpy.ones",
"tf_agents.bandits.agents.utils.sum_reward_weighted_observations",
"tensorflow.compat.v1.enable_resource_variables",
"tensorflow.range",
"tensorflow.squeeze",
"tensorflow.train.latest_checkpoint",
"tf_agents.bandits.agents.neural_linucb_agent.NeuralLinUCBVariableCollection",
"tensorflow.ones",
"os.path.join",
"absl.testing.parameterized.named_parameters",
"tensorflow.test.main",
"numpy.random.randint",
"numpy.zeros",
"tensorflow.constant",
"numpy.random.uniform",
"tensorflow.compat.v1.initialize_all_variables",
"tf_agents.utils.common.function_in_tf1"
] |
[((2541, 2735), 'absl.testing.parameterized.named_parameters', 'parameterized.named_parameters', (["{'testcase_name': '_batch1_contextdim10', 'batch_size': 1, 'context_dim': 10}", "{'testcase_name': '_batch4_contextdim5', 'batch_size': 4, 'context_dim': 5}"], {}), "({'testcase_name': '_batch1_contextdim10',\n 'batch_size': 1, 'context_dim': 10}, {'testcase_name':\n '_batch4_contextdim5', 'batch_size': 4, 'context_dim': 5})\n", (2571, 2735), False, 'from absl.testing import parameterized\n'), ((2985, 3026), 'numpy.random.uniform', 'np.random.uniform', (['(0.0)', '(1.0)', '[batch_size]'], {}), '(0.0, 1.0, [batch_size])\n', (3002, 3026), True, 'import numpy as np\n'), ((4249, 4291), 'tensorflow.constant', 'tf.constant', (['observation'], {'dtype': 'tf.float32'}), '(observation, dtype=tf.float32)\n', (4260, 4291), True, 'import tensorflow as tf\n'), ((4376, 4417), 'numpy.random.uniform', 'np.random.uniform', (['(0.0)', '(1.0)', '[batch_size]'], {}), '(0.0, 1.0, [batch_size])\n', (4393, 4417), True, 'import numpy as np\n'), ((5414, 5487), 'tf_agents.bandits.drivers.driver_utils.trajectory_for_bandit', 'driver_utils.trajectory_for_bandit', (['initial_step', 'action_step', 'final_step'], {}), '(initial_step, action_step, final_step)\n', (5448, 5487), False, 'from tf_agents.bandits.drivers import driver_utils\n'), ((23937, 23951), 'tensorflow.test.main', 'tf.test.main', ([], {}), '()\n', (23949, 23951), True, 'import tensorflow as tf\n'), ((2385, 2412), 'tensorflow.cast', 'tf.cast', (['inputs', 'tf.float32'], {}), '(inputs, tf.float32)\n', (2392, 2412), True, 'import tensorflow as tf\n'), ((3070, 3165), 'tensorflow.constant', 'tf.constant', (['time_step.StepType.FIRST'], {'dtype': 'tf.int32', 'shape': '[batch_size]', 'name': '"""step_type"""'}), "(time_step.StepType.FIRST, dtype=tf.int32, shape=[batch_size],\n name='step_type')\n", (3081, 3165), True, 'import tensorflow as tf\n'), ((3190, 3259), 'tensorflow.constant', 'tf.constant', (['(0.0)'], {'dtype': 'tf.float32', 'shape': '[batch_size]', 'name': '"""reward"""'}), "(0.0, dtype=tf.float32, shape=[batch_size], name='reward')\n", (3201, 3259), True, 'import tensorflow as tf\n'), ((3267, 3338), 'tensorflow.constant', 'tf.constant', (['(1.0)'], {'dtype': 'tf.float32', 'shape': '[batch_size]', 'name': '"""discount"""'}), "(1.0, dtype=tf.float32, shape=[batch_size], name='discount')\n", (3278, 3338), True, 'import tensorflow as tf\n'), ((3346, 3445), 'tensorflow.constant', 'tf.constant', (['observation'], {'dtype': 'tf.float32', 'shape': '[batch_size, context_dim]', 'name': '"""observation"""'}), "(observation, dtype=tf.float32, shape=[batch_size, context_dim],\n name='observation')\n", (3357, 3445), True, 'import tensorflow as tf\n'), ((3502, 3596), 'tensorflow.constant', 'tf.constant', (['time_step.StepType.LAST'], {'dtype': 'tf.int32', 'shape': '[batch_size]', 'name': '"""step_type"""'}), "(time_step.StepType.LAST, dtype=tf.int32, shape=[batch_size],\n name='step_type')\n", (3513, 3596), True, 'import tensorflow as tf\n'), ((3621, 3693), 'tensorflow.constant', 'tf.constant', (['reward'], {'dtype': 'tf.float32', 'shape': '[batch_size]', 'name': '"""reward"""'}), "(reward, dtype=tf.float32, shape=[batch_size], name='reward')\n", (3632, 3693), True, 'import tensorflow as tf\n'), ((3701, 3772), 'tensorflow.constant', 'tf.constant', (['(1.0)'], {'dtype': 'tf.float32', 'shape': '[batch_size]', 'name': '"""discount"""'}), "(1.0, dtype=tf.float32, shape=[batch_size], name='discount')\n", (3712, 3772), True, 'import tensorflow as tf\n'), ((3780, 3887), 'tensorflow.constant', 'tf.constant', (['(observation + 100.0)'], {'dtype': 'tf.float32', 'shape': '[batch_size, context_dim]', 'name': '"""observation"""'}), "(observation + 100.0, dtype=tf.float32, shape=[batch_size,\n context_dim], name='observation')\n", (3791, 3887), True, 'import tensorflow as tf\n'), ((4305, 4364), 'tensorflow.eye', 'tf.eye', (['batch_size'], {'num_columns': 'num_actions', 'dtype': 'tf.int32'}), '(batch_size, num_columns=num_actions, dtype=tf.int32)\n', (4311, 4364), True, 'import tensorflow as tf\n'), ((4461, 4556), 'tensorflow.constant', 'tf.constant', (['time_step.StepType.FIRST'], {'dtype': 'tf.int32', 'shape': '[batch_size]', 'name': '"""step_type"""'}), "(time_step.StepType.FIRST, dtype=tf.int32, shape=[batch_size],\n name='step_type')\n", (4472, 4556), True, 'import tensorflow as tf\n'), ((4601, 4670), 'tensorflow.constant', 'tf.constant', (['(0.0)'], {'dtype': 'tf.float32', 'shape': '[batch_size]', 'name': '"""reward"""'}), "(0.0, dtype=tf.float32, shape=[batch_size], name='reward')\n", (4612, 4670), True, 'import tensorflow as tf\n'), ((4678, 4749), 'tensorflow.constant', 'tf.constant', (['(1.0)'], {'dtype': 'tf.float32', 'shape': '[batch_size]', 'name': '"""discount"""'}), "(1.0, dtype=tf.float32, shape=[batch_size], name='discount')\n", (4689, 4749), True, 'import tensorflow as tf\n'), ((4819, 4913), 'tensorflow.constant', 'tf.constant', (['time_step.StepType.LAST'], {'dtype': 'tf.int32', 'shape': '[batch_size]', 'name': '"""step_type"""'}), "(time_step.StepType.LAST, dtype=tf.int32, shape=[batch_size],\n name='step_type')\n", (4830, 4913), True, 'import tensorflow as tf\n'), ((4958, 5030), 'tensorflow.constant', 'tf.constant', (['reward'], {'dtype': 'tf.float32', 'shape': '[batch_size]', 'name': '"""reward"""'}), "(reward, dtype=tf.float32, shape=[batch_size], name='reward')\n", (4969, 5030), True, 'import tensorflow as tf\n'), ((5038, 5109), 'tensorflow.constant', 'tf.constant', (['(1.0)'], {'dtype': 'tf.float32', 'shape': '[batch_size]', 'name': '"""discount"""'}), "(1.0, dtype=tf.float32, shape=[batch_size], name='discount')\n", (5049, 5109), True, 'import tensorflow as tf\n'), ((5829, 5869), 'tensorflow.compat.v1.enable_resource_variables', 'tf.compat.v1.enable_resource_variables', ([], {}), '()\n', (5867, 5869), True, 'import tensorflow as tf\n'), ((6002, 6051), 'tf_agents.specs.tensor_spec.TensorSpec', 'tensor_spec.TensorSpec', (['[context_dim]', 'tf.float32'], {}), '([context_dim], tf.float32)\n', (6024, 6051), False, 'from tf_agents.specs import tensor_spec\n'), ((6073, 6115), 'tf_agents.trajectories.time_step.time_step_spec', 'time_step.time_step_spec', (['observation_spec'], {}), '(observation_spec)\n', (6097, 6115), False, 'from tf_agents.trajectories import time_step\n'), ((6134, 6230), 'tf_agents.specs.tensor_spec.BoundedTensorSpec', 'tensor_spec.BoundedTensorSpec', ([], {'dtype': 'tf.int32', 'shape': '()', 'minimum': '(0)', 'maximum': '(num_actions - 1)'}), '(dtype=tf.int32, shape=(), minimum=0, maximum=\n num_actions - 1)\n', (6163, 6230), False, 'from tf_agents.specs import tensor_spec\n'), ((6289, 6485), 'tf_agents.bandits.agents.neural_linucb_agent.NeuralLinUCBAgent', 'neural_linucb_agent.NeuralLinUCBAgent', ([], {'time_step_spec': 'time_step_spec', 'action_spec': 'action_spec', 'encoding_network': 'encoder', 'encoding_network_num_train_steps': '(0)', 'encoding_dim': '(10)', 'optimizer': 'None'}), '(time_step_spec=time_step_spec,\n action_spec=action_spec, encoding_network=encoder,\n encoding_network_num_train_steps=0, encoding_dim=10, optimizer=None)\n', (6326, 6485), False, 'from tf_agents.bandits.agents import neural_linucb_agent\n'), ((6698, 6747), 'tf_agents.specs.tensor_spec.TensorSpec', 'tensor_spec.TensorSpec', (['[context_dim]', 'tf.float32'], {}), '([context_dim], tf.float32)\n', (6720, 6747), False, 'from tf_agents.specs import tensor_spec\n'), ((6769, 6811), 'tf_agents.trajectories.time_step.time_step_spec', 'time_step.time_step_spec', (['observation_spec'], {}), '(observation_spec)\n', (6793, 6811), False, 'from tf_agents.trajectories import time_step\n'), ((6830, 6926), 'tf_agents.specs.tensor_spec.BoundedTensorSpec', 'tensor_spec.BoundedTensorSpec', ([], {'dtype': 'tf.int32', 'shape': '()', 'minimum': '(0)', 'maximum': '(num_actions - 1)'}), '(dtype=tf.int32, shape=(), minimum=0, maximum=\n num_actions - 1)\n', (6859, 6926), False, 'from tf_agents.specs import tensor_spec\n'), ((6985, 7182), 'tf_agents.bandits.agents.neural_linucb_agent.NeuralLinUCBAgent', 'neural_linucb_agent.NeuralLinUCBAgent', ([], {'time_step_spec': 'time_step_spec', 'action_spec': 'action_spec', 'encoding_network': 'encoder', 'encoding_network_num_train_steps': '(10)', 'encoding_dim': '(10)', 'optimizer': 'None'}), '(time_step_spec=time_step_spec,\n action_spec=action_spec, encoding_network=encoder,\n encoding_network_num_train_steps=10, encoding_dim=10, optimizer=None)\n', (7022, 7182), False, 'from tf_agents.bandits.agents import neural_linucb_agent\n'), ((7630, 7693), 'numpy.random.randint', 'np.random.randint', (['num_actions'], {'size': 'batch_size', 'dtype': 'np.int32'}), '(num_actions, size=batch_size, dtype=np.int32)\n', (7647, 7693), True, 'import numpy as np\n'), ((7882, 7931), 'tf_agents.specs.tensor_spec.TensorSpec', 'tensor_spec.TensorSpec', (['[context_dim]', 'tf.float32'], {}), '([context_dim], tf.float32)\n', (7904, 7931), False, 'from tf_agents.specs import tensor_spec\n'), ((7953, 7995), 'tf_agents.trajectories.time_step.time_step_spec', 'time_step.time_step_spec', (['observation_spec'], {}), '(observation_spec)\n', (7977, 7995), False, 'from tf_agents.trajectories import time_step\n'), ((8014, 8110), 'tf_agents.specs.tensor_spec.BoundedTensorSpec', 'tensor_spec.BoundedTensorSpec', ([], {'dtype': 'tf.int32', 'shape': '()', 'minimum': '(0)', 'maximum': '(num_actions - 1)'}), '(dtype=tf.int32, shape=(), minimum=0, maximum=\n num_actions - 1)\n', (8043, 8110), False, 'from tf_agents.specs import tensor_spec\n'), ((11064, 11127), 'numpy.random.randint', 'np.random.randint', (['num_actions'], {'size': 'batch_size', 'dtype': 'np.int32'}), '(num_actions, size=batch_size, dtype=np.int32)\n', (11081, 11127), True, 'import numpy as np\n'), ((11316, 11365), 'tf_agents.specs.tensor_spec.TensorSpec', 'tensor_spec.TensorSpec', (['[context_dim]', 'tf.float32'], {}), '([context_dim], tf.float32)\n', (11338, 11365), False, 'from tf_agents.specs import tensor_spec\n'), ((11387, 11429), 'tf_agents.trajectories.time_step.time_step_spec', 'time_step.time_step_spec', (['observation_spec'], {}), '(observation_spec)\n', (11411, 11429), False, 'from tf_agents.trajectories import time_step\n'), ((11448, 11544), 'tf_agents.specs.tensor_spec.BoundedTensorSpec', 'tensor_spec.BoundedTensorSpec', ([], {'dtype': 'tf.int32', 'shape': '()', 'minimum': '(0)', 'maximum': '(num_actions - 1)'}), '(dtype=tf.int32, shape=(), minimum=0, maximum=\n num_actions - 1)\n', (11477, 11544), False, 'from tf_agents.specs import tensor_spec\n'), ((12198, 12238), 'tensorflow.cast', 'tf.cast', (['experience.reward', 'agent._dtype'], {}), '(experience.reward, agent._dtype)\n', (12205, 12238), True, 'import tensorflow as tf\n'), ((14779, 14842), 'numpy.random.randint', 'np.random.randint', (['num_actions'], {'size': 'batch_size', 'dtype': 'np.int32'}), '(num_actions, size=batch_size, dtype=np.int32)\n', (14796, 14842), True, 'import numpy as np\n'), ((15031, 15080), 'tf_agents.specs.tensor_spec.TensorSpec', 'tensor_spec.TensorSpec', (['[context_dim]', 'tf.float32'], {}), '([context_dim], tf.float32)\n', (15053, 15080), False, 'from tf_agents.specs import tensor_spec\n'), ((15102, 15144), 'tf_agents.trajectories.time_step.time_step_spec', 'time_step.time_step_spec', (['observation_spec'], {}), '(observation_spec)\n', (15126, 15144), False, 'from tf_agents.trajectories import time_step\n'), ((15163, 15259), 'tf_agents.specs.tensor_spec.BoundedTensorSpec', 'tensor_spec.BoundedTensorSpec', ([], {'dtype': 'tf.int32', 'shape': '()', 'minimum': '(0)', 'maximum': '(num_actions - 1)'}), '(dtype=tf.int32, shape=(), minimum=0, maximum=\n num_actions - 1)\n', (15192, 15259), False, 'from tf_agents.specs import tensor_spec\n'), ((15353, 15430), 'tf_agents.bandits.agents.neural_linucb_agent.NeuralLinUCBVariableCollection', 'neural_linucb_agent.NeuralLinUCBVariableCollection', (['num_actions', 'encoding_dim'], {}), '(num_actions, encoding_dim)\n', (15403, 15430), False, 'from tf_agents.bandits.agents import neural_linucb_agent\n'), ((16456, 16519), 'numpy.random.randint', 'np.random.randint', (['num_actions'], {'size': 'batch_size', 'dtype': 'np.int32'}), '(num_actions, size=batch_size, dtype=np.int32)\n', (16473, 16519), True, 'import numpy as np\n'), ((16854, 16896), 'tf_agents.trajectories.time_step.time_step_spec', 'time_step.time_step_spec', (['observation_spec'], {}), '(observation_spec)\n', (16878, 16896), False, 'from tf_agents.trajectories import time_step\n'), ((16915, 17011), 'tf_agents.specs.tensor_spec.BoundedTensorSpec', 'tensor_spec.BoundedTensorSpec', ([], {'dtype': 'tf.int32', 'shape': '()', 'minimum': '(0)', 'maximum': '(num_actions - 1)'}), '(dtype=tf.int32, shape=(), minimum=0, maximum=\n num_actions - 1)\n', (16944, 17011), False, 'from tf_agents.specs import tensor_spec\n'), ((17904, 18010), 'tf_agents.bandits.agents.neural_linucb_agent.NeuralLinUCBVariableCollection', 'neural_linucb_agent.NeuralLinUCBVariableCollection', ([], {'num_actions': 'num_actions', 'encoding_dim': 'encoding_dim'}), '(num_actions=num_actions,\n encoding_dim=encoding_dim)\n', (17954, 18010), False, 'from tf_agents.bandits.agents import neural_linucb_agent\n'), ((18152, 18212), 'tensorflow.train.Checkpoint', 'tf.train.Checkpoint', ([], {'variable_collection': 'variable_collection'}), '(variable_collection=variable_collection)\n', (18171, 18212), True, 'import tensorflow as tf\n'), ((18278, 18320), 'os.path.join', 'os.path.join', (['checkpoint_dir', '"""checkpoint"""'], {}), "(checkpoint_dir, 'checkpoint')\n", (18290, 18320), False, 'import os\n'), ((18462, 18504), 'tensorflow.train.latest_checkpoint', 'tf.train.latest_checkpoint', (['checkpoint_dir'], {}), '(checkpoint_dir)\n', (18488, 18504), True, 'import tensorflow as tf\n'), ((18814, 18908), 'tf_agents.bandits.specs.utils.create_per_arm_observation_spec', 'bandit_spec_utils.create_per_arm_observation_spec', (['(2)', '(3)', 'num_actions'], {'add_action_mask': '(True)'}), '(2, 3, num_actions,\n add_action_mask=True)\n', (18863, 18908), True, 'from tf_agents.bandits.specs import utils as bandit_spec_utils\n'), ((18935, 18969), 'tf_agents.trajectories.time_step.time_step_spec', 'time_step.time_step_spec', (['obs_spec'], {}), '(obs_spec)\n', (18959, 18969), False, 'from tf_agents.trajectories import time_step\n'), ((18988, 19084), 'tf_agents.specs.tensor_spec.BoundedTensorSpec', 'tensor_spec.BoundedTensorSpec', ([], {'dtype': 'tf.int32', 'shape': '()', 'minimum': '(0)', 'maximum': '(num_actions - 1)'}), '(dtype=tf.int32, shape=(), minimum=0, maximum=\n num_actions - 1)\n', (19017, 19084), False, 'from tf_agents.specs import tensor_spec\n'), ((19135, 19262), 'tf_agents.bandits.networks.global_and_arm_feature_network.create_feed_forward_common_tower_network', 'global_and_arm_feature_network.create_feed_forward_common_tower_network', (['obs_spec[0]', '(4, 3)', '(3, 4)', '(4, 2)', 'encoding_dim'], {}), '(\n obs_spec[0], (4, 3), (3, 4), (4, 2), encoding_dim)\n', (19206, 19262), False, 'from tf_agents.bandits.networks import global_and_arm_feature_network\n'), ((20040, 20072), 'numpy.array', 'np.array', (['[0, 3]'], {'dtype': 'np.int32'}), '([0, 3], dtype=np.int32)\n', (20048, 20072), True, 'import numpy as np\n'), ((20087, 20125), 'numpy.array', 'np.array', (['[0.5, 3.0]'], {'dtype': 'np.float32'}), '([0.5, 3.0], dtype=np.float32)\n', (20095, 20125), True, 'import numpy as np\n'), ((21413, 21515), 'tf_agents.bandits.specs.utils.create_per_arm_observation_spec', 'bandit_spec_utils.create_per_arm_observation_spec', (['(2)', '(3)', 'num_actions'], {'add_num_actions_feature': '(True)'}), '(2, 3, num_actions,\n add_num_actions_feature=True)\n', (21462, 21515), True, 'from tf_agents.bandits.specs import utils as bandit_spec_utils\n'), ((21542, 21576), 'tf_agents.trajectories.time_step.time_step_spec', 'time_step.time_step_spec', (['obs_spec'], {}), '(obs_spec)\n', (21566, 21576), False, 'from tf_agents.trajectories import time_step\n'), ((21595, 21691), 'tf_agents.specs.tensor_spec.BoundedTensorSpec', 'tensor_spec.BoundedTensorSpec', ([], {'dtype': 'tf.int32', 'shape': '()', 'minimum': '(0)', 'maximum': '(num_actions - 1)'}), '(dtype=tf.int32, shape=(), minimum=0, maximum=\n num_actions - 1)\n', (21624, 21691), False, 'from tf_agents.specs import tensor_spec\n'), ((21742, 21866), 'tf_agents.bandits.networks.global_and_arm_feature_network.create_feed_forward_common_tower_network', 'global_and_arm_feature_network.create_feed_forward_common_tower_network', (['obs_spec', '(4, 3)', '(3, 4)', '(4, 2)', 'encoding_dim'], {}), '(\n obs_spec, (4, 3), (3, 4), (4, 2), encoding_dim)\n', (21813, 21866), False, 'from tf_agents.bandits.networks import global_and_arm_feature_network\n'), ((22618, 22650), 'numpy.array', 'np.array', (['[0, 3]'], {'dtype': 'np.int32'}), '([0, 3], dtype=np.int32)\n', (22626, 22650), True, 'import numpy as np\n'), ((22665, 22703), 'numpy.array', 'np.array', (['[0.5, 3.0]'], {'dtype': 'np.float32'}), '([0.5, 3.0], dtype=np.float32)\n', (22673, 22703), True, 'import numpy as np\n'), ((5258, 5286), 'tensorflow.convert_to_tensor', 'tf.convert_to_tensor', (['action'], {}), '(action)\n', (5278, 5286), True, 'import tensorflow as tf\n'), ((5299, 5328), 'tf_agents.bandits.policies.policy_utilities.PolicyInfo', 'policy_utilities.PolicyInfo', ([], {}), '()\n', (5326, 5328), False, 'from tf_agents.bandits.policies import policy_utilities\n'), ((8583, 8626), 'tensorflow.compat.v1.global_variables_initializer', 'tf.compat.v1.global_variables_initializer', ([], {}), '()\n', (8624, 8626), True, 'import tensorflow as tf\n'), ((9540, 9595), 'tensorflow.cast', 'tf.cast', (['encoded_observations_for_arm'], {'dtype': 'tf.float64'}), '(encoded_observations_for_arm, dtype=tf.float64)\n', (9547, 9595), True, 'import tensorflow as tf\n'), ((11977, 12020), 'tensorflow.compat.v1.global_variables_initializer', 'tf.compat.v1.global_variables_initializer', ([], {}), '()\n', (12018, 12020), True, 'import tensorflow as tf\n'), ((12106, 12130), 'tf_agents.utils.common.function_in_tf1', 'common.function_in_tf1', ([], {}), '()\n', (12128, 12130), False, 'from tf_agents.utils import common\n'), ((13239, 13294), 'tensorflow.cast', 'tf.cast', (['encoded_observations_for_arm'], {'dtype': 'tf.float64'}), '(encoded_observations_for_arm, dtype=tf.float64)\n', (13246, 13294), True, 'import tensorflow as tf\n'), ((15899, 15942), 'tensorflow.compat.v1.global_variables_initializer', 'tf.compat.v1.global_variables_initializer', ([], {}), '()\n', (15940, 15942), True, 'import tensorflow as tf\n'), ((16709, 16758), 'tf_agents.specs.tensor_spec.TensorSpec', 'tensor_spec.TensorSpec', (['[context_dim]', 'tf.float32'], {}), '([context_dim], tf.float32)\n', (16731, 16758), False, 'from tf_agents.specs import tensor_spec\n'), ((16784, 16831), 'tf_agents.specs.tensor_spec.TensorSpec', 'tensor_spec.TensorSpec', (['[num_actions]', 'tf.int32'], {}), '([num_actions], tf.int32)\n', (16806, 16831), False, 'from tf_agents.specs import tensor_spec\n'), ((17567, 17610), 'tensorflow.compat.v1.global_variables_initializer', 'tf.compat.v1.global_variables_initializer', ([], {}), '()\n', (17608, 17610), True, 'import tensorflow as tf\n'), ((17759, 17781), 'tensorflow.executing_eagerly', 'tf.executing_eagerly', ([], {}), '()\n', (17779, 17781), True, 'import tensorflow as tf\n'), ((18034, 18077), 'tensorflow.compat.v1.global_variables_initializer', 'tf.compat.v1.global_variables_initializer', ([], {}), '()\n', (18075, 18077), True, 'import tensorflow as tf\n'), ((19977, 20024), 'tensorflow.ones', 'tf.ones', ([], {'shape': '(2, num_actions)', 'dtype': 'tf.int32'}), '(shape=(2, num_actions), dtype=tf.int32)\n', (19984, 20024), True, 'import tensorflow as tf\n'), ((20173, 20260), 'tensorflow.constant', 'tf.constant', (['time_step.StepType.FIRST'], {'dtype': 'tf.int32', 'shape': '[2]', 'name': '"""step_type"""'}), "(time_step.StepType.FIRST, dtype=tf.int32, shape=[2], name=\n 'step_type')\n", (20184, 20260), True, 'import tensorflow as tf\n'), ((20314, 20374), 'tensorflow.constant', 'tf.constant', (['(0.0)'], {'dtype': 'tf.float32', 'shape': '[2]', 'name': '"""reward"""'}), "(0.0, dtype=tf.float32, shape=[2], name='reward')\n", (20325, 20374), True, 'import tensorflow as tf\n'), ((20384, 20446), 'tensorflow.constant', 'tf.constant', (['(1.0)'], {'dtype': 'tf.float32', 'shape': '[2]', 'name': '"""discount"""'}), "(1.0, dtype=tf.float32, shape=[2], name='discount')\n", (20395, 20446), True, 'import tensorflow as tf\n'), ((20515, 20601), 'tensorflow.constant', 'tf.constant', (['time_step.StepType.LAST'], {'dtype': 'tf.int32', 'shape': '[2]', 'name': '"""step_type"""'}), "(time_step.StepType.LAST, dtype=tf.int32, shape=[2], name=\n 'step_type')\n", (20526, 20601), True, 'import tensorflow as tf\n'), ((20655, 20708), 'tensorflow.constant', 'tf.constant', (['rewards'], {'dtype': 'tf.float32', 'name': '"""reward"""'}), "(rewards, dtype=tf.float32, name='reward')\n", (20666, 20708), True, 'import tensorflow as tf\n'), ((20718, 20780), 'tensorflow.constant', 'tf.constant', (['(1.0)'], {'dtype': 'tf.float32', 'shape': '[2]', 'name': '"""discount"""'}), "(1.0, dtype=tf.float32, shape=[2], name='discount')\n", (20729, 20780), True, 'import tensorflow as tf\n'), ((21205, 21244), 'tensorflow.compat.v1.initialize_all_variables', 'tf.compat.v1.initialize_all_variables', ([], {}), '()\n', (21242, 21244), True, 'import tensorflow as tf\n'), ((22304, 22351), 'tensorflow.constant', 'tf.constant', (['[[1, 2], [3, 4]]'], {'dtype': 'tf.float32'}), '([[1, 2], [3, 4]], dtype=tf.float32)\n', (22315, 22351), True, 'import tensorflow as tf\n'), ((22562, 22597), 'tensorflow.constant', 'tf.constant', (['[3, 4]'], {'dtype': 'tf.int32'}), '([3, 4], dtype=tf.int32)\n', (22573, 22597), True, 'import tensorflow as tf\n'), ((22751, 22838), 'tensorflow.constant', 'tf.constant', (['time_step.StepType.FIRST'], {'dtype': 'tf.int32', 'shape': '[2]', 'name': '"""step_type"""'}), "(time_step.StepType.FIRST, dtype=tf.int32, shape=[2], name=\n 'step_type')\n", (22762, 22838), True, 'import tensorflow as tf\n'), ((22892, 22952), 'tensorflow.constant', 'tf.constant', (['(0.0)'], {'dtype': 'tf.float32', 'shape': '[2]', 'name': '"""reward"""'}), "(0.0, dtype=tf.float32, shape=[2], name='reward')\n", (22903, 22952), True, 'import tensorflow as tf\n'), ((22962, 23024), 'tensorflow.constant', 'tf.constant', (['(1.0)'], {'dtype': 'tf.float32', 'shape': '[2]', 'name': '"""discount"""'}), "(1.0, dtype=tf.float32, shape=[2], name='discount')\n", (22973, 23024), True, 'import tensorflow as tf\n'), ((23093, 23179), 'tensorflow.constant', 'tf.constant', (['time_step.StepType.LAST'], {'dtype': 'tf.int32', 'shape': '[2]', 'name': '"""step_type"""'}), "(time_step.StepType.LAST, dtype=tf.int32, shape=[2], name=\n 'step_type')\n", (23104, 23179), True, 'import tensorflow as tf\n'), ((23233, 23286), 'tensorflow.constant', 'tf.constant', (['rewards'], {'dtype': 'tf.float32', 'name': '"""reward"""'}), "(rewards, dtype=tf.float32, name='reward')\n", (23244, 23286), True, 'import tensorflow as tf\n'), ((23296, 23358), 'tensorflow.constant', 'tf.constant', (['(1.0)'], {'dtype': 'tf.float32', 'shape': '[2]', 'name': '"""discount"""'}), "(1.0, dtype=tf.float32, shape=[2], name='discount')\n", (23307, 23358), True, 'import tensorflow as tf\n'), ((23783, 23822), 'tensorflow.compat.v1.initialize_all_variables', 'tf.compat.v1.initialize_all_variables', ([], {}), '()\n', (23820, 23822), True, 'import tensorflow as tf\n'), ((5587, 5610), 'tensorflow.convert_to_tensor', 'tf.convert_to_tensor', (['x'], {}), '(x)\n', (5607, 5610), True, 'import tensorflow as tf\n'), ((8432, 8484), 'tensorflow.compat.v1.train.AdamOptimizer', 'tf.compat.v1.train.AdamOptimizer', ([], {'learning_rate': '(0.01)'}), '(learning_rate=0.01)\n', (8464, 8484), True, 'import tensorflow as tf\n'), ((8983, 9011), 'tensorflow.convert_to_tensor', 'tf.convert_to_tensor', (['action'], {}), '(action)\n', (9003, 9011), True, 'import tensorflow as tf\n'), ((9188, 9216), 'tensorflow.convert_to_tensor', 'tf.convert_to_tensor', (['action'], {}), '(action)\n', (9208, 9216), True, 'import tensorflow as tf\n'), ((9848, 9939), 'tensorflow.matmul', 'tf.matmul', (['encoded_observations_for_arm', 'encoded_observations_for_arm'], {'transpose_a': '(True)'}), '(encoded_observations_for_arm, encoded_observations_for_arm,\n transpose_a=True)\n', (9857, 9939), True, 'import tensorflow as tf\n'), ((9989, 10085), 'tf_agents.bandits.agents.utils.sum_reward_weighted_observations', 'bandit_utils.sum_reward_weighted_observations', (['rewards_for_arm', 'encoded_observations_for_arm'], {}), '(rewards_for_arm,\n encoded_observations_for_arm)\n', (10034, 10085), True, 'from tf_agents.bandits.agents import utils as bandit_utils\n'), ((11866, 11918), 'tensorflow.compat.v1.train.AdamOptimizer', 'tf.compat.v1.train.AdamOptimizer', ([], {'learning_rate': '(0.01)'}), '(learning_rate=0.01)\n', (11898, 11918), True, 'import tensorflow as tf\n'), ((12682, 12710), 'tensorflow.convert_to_tensor', 'tf.convert_to_tensor', (['action'], {}), '(action)\n', (12702, 12710), True, 'import tensorflow as tf\n'), ((12887, 12915), 'tensorflow.convert_to_tensor', 'tf.convert_to_tensor', (['action'], {}), '(action)\n', (12907, 12915), True, 'import tensorflow as tf\n'), ((13547, 13638), 'tensorflow.matmul', 'tf.matmul', (['encoded_observations_for_arm', 'encoded_observations_for_arm'], {'transpose_a': '(True)'}), '(encoded_observations_for_arm, encoded_observations_for_arm,\n transpose_a=True)\n', (13556, 13638), True, 'import tensorflow as tf\n'), ((13688, 13784), 'tf_agents.bandits.agents.utils.sum_reward_weighted_observations', 'bandit_utils.sum_reward_weighted_observations', (['rewards_for_arm', 'encoded_observations_for_arm'], {}), '(rewards_for_arm,\n encoded_observations_for_arm)\n', (13733, 13784), True, 'from tf_agents.bandits.agents import utils as bandit_utils\n'), ((15744, 15797), 'tensorflow.compat.v1.train.AdamOptimizer', 'tf.compat.v1.train.AdamOptimizer', ([], {'learning_rate': '(0.001)'}), '(learning_rate=0.001)\n', (15776, 15797), True, 'import tensorflow as tf\n'), ((17337, 17390), 'tensorflow.compat.v1.train.AdamOptimizer', 'tf.compat.v1.train.AdamOptimizer', ([], {'learning_rate': '(0.001)'}), '(learning_rate=0.001)\n', (17369, 17390), True, 'import tensorflow as tf\n'), ((19641, 19694), 'tensorflow.compat.v1.train.AdamOptimizer', 'tf.compat.v1.train.AdamOptimizer', ([], {'learning_rate': '(0.001)'}), '(learning_rate=0.001)\n', (19673, 19694), True, 'import tensorflow as tf\n'), ((19776, 19823), 'tensorflow.constant', 'tf.constant', (['[[1, 2], [3, 4]]'], {'dtype': 'tf.float32'}), '([[1, 2], [3, 4]], dtype=tf.float32)\n', (19787, 19823), True, 'import tensorflow as tf\n'), ((20861, 20890), 'tensorflow.convert_to_tensor', 'tf.convert_to_tensor', (['actions'], {}), '(actions)\n', (20881, 20890), True, 'import tensorflow as tf\n'), ((22170, 22223), 'tensorflow.compat.v1.train.AdamOptimizer', 'tf.compat.v1.train.AdamOptimizer', ([], {'learning_rate': '(0.001)'}), '(learning_rate=0.001)\n', (22202, 22223), True, 'import tensorflow as tf\n'), ((23439, 23468), 'tensorflow.convert_to_tensor', 'tf.convert_to_tensor', (['actions'], {}), '(actions)\n', (23459, 23468), True, 'import tensorflow as tf\n'), ((8867, 8910), 'tensorflow.cast', 'tf.cast', (['experience.observation', 'tf.float64'], {}), '(experience.observation, tf.float64)\n', (8874, 8910), True, 'import tensorflow as tf\n'), ((9114, 9152), 'tensorflow.cast', 'tf.cast', (['experience.reward', 'tf.float64'], {}), '(experience.reward, tf.float64)\n', (9121, 9152), True, 'import tensorflow as tf\n'), ((9663, 9688), 'tensorflow.shape', 'tf.shape', (['rewards_for_arm'], {}), '(rewards_for_arm)\n', (9671, 9688), True, 'import tensorflow as tf\n'), ((10161, 10217), 'tensorflow.zeros', 'tf.zeros', (['[encoding_dim, encoding_dim]'], {'dtype': 'tf.float64'}), '([encoding_dim, encoding_dim], dtype=tf.float64)\n', (10169, 10217), True, 'import tensorflow as tf\n'), ((10235, 10277), 'tensorflow.zeros', 'tf.zeros', (['[encoding_dim]'], {'dtype': 'tf.float64'}), '([encoding_dim], dtype=tf.float64)\n', (10243, 10277), True, 'import tensorflow as tf\n'), ((10319, 10356), 'tensorflow.squeeze', 'tf.squeeze', (['num_samples_for_arm_total'], {}), '(num_samples_for_arm_total)\n', (10329, 10356), True, 'import tensorflow as tf\n'), ((12566, 12609), 'tensorflow.cast', 'tf.cast', (['experience.observation', 'tf.float64'], {}), '(experience.observation, tf.float64)\n', (12573, 12609), True, 'import tensorflow as tf\n'), ((12813, 12851), 'tensorflow.cast', 'tf.cast', (['experience.reward', 'tf.float64'], {}), '(experience.reward, tf.float64)\n', (12820, 12851), True, 'import tensorflow as tf\n'), ((13362, 13387), 'tensorflow.shape', 'tf.shape', (['rewards_for_arm'], {}), '(rewards_for_arm)\n', (13370, 13387), True, 'import tensorflow as tf\n'), ((13860, 13916), 'tensorflow.zeros', 'tf.zeros', (['[encoding_dim, encoding_dim]'], {'dtype': 'tf.float64'}), '([encoding_dim, encoding_dim], dtype=tf.float64)\n', (13868, 13916), True, 'import tensorflow as tf\n'), ((13934, 13976), 'tensorflow.zeros', 'tf.zeros', (['[encoding_dim]'], {'dtype': 'tf.float64'}), '([encoding_dim], dtype=tf.float64)\n', (13942, 13976), True, 'import tensorflow as tf\n'), ((14018, 14055), 'tensorflow.squeeze', 'tf.squeeze', (['num_samples_for_arm_total'], {}), '(num_samples_for_arm_total)\n', (14028, 14055), True, 'import tensorflow as tf\n'), ((22448, 22460), 'tensorflow.range', 'tf.range', (['(30)'], {}), '(30)\n', (22456, 22460), True, 'import tensorflow as tf\n'), ((2140, 2176), 'numpy.ones', 'np.ones', (['[context_dim, encoding_dim]'], {}), '([context_dim, encoding_dim])\n', (2147, 2176), True, 'import numpy as np\n'), ((2260, 2284), 'numpy.zeros', 'np.zeros', (['[encoding_dim]'], {}), '([encoding_dim])\n', (2268, 2284), True, 'import numpy as np\n'), ((19920, 19932), 'tensorflow.range', 'tf.range', (['(30)'], {}), '(30)\n', (19928, 19932), True, 'import tensorflow as tf\n'), ((20972, 21022), 'numpy.array', 'np.array', (['[[1, 2, 3], [3, 2, 1]]'], {'dtype': 'np.float32'}), '([[1, 2, 3], [3, 2, 1]], dtype=np.float32)\n', (20980, 21022), True, 'import numpy as np\n'), ((23550, 23600), 'numpy.array', 'np.array', (['[[1, 2, 3], [3, 2, 1]]'], {'dtype': 'np.float32'}), '([[1, 2, 3], [3, 2, 1]], dtype=np.float32)\n', (23558, 23600), True, 'import numpy as np\n')]
|
from __future__ import print_function, absolute_import
from distutils import sysconfig
from distutils import version
from distutils.core import Extension
import glob
import io
import multiprocessing
import os
import re
import subprocess
import sys
import warnings
from textwrap import fill
PY3 = (sys.version_info[0] >= 3)
try:
from subprocess import check_output
except ImportError:
# check_output is not available in Python 2.6
def check_output(*popenargs, **kwargs):
"""
Run command with arguments and return its output as a byte
string.
Backported from Python 2.7 as it's implemented as pure python
on stdlib.
"""
process = subprocess.Popen(
stdout=subprocess.PIPE, *popenargs, **kwargs)
output, unused_err = process.communicate()
retcode = process.poll()
if retcode:
cmd = kwargs.get("args")
if cmd is None:
cmd = popenargs[0]
error = subprocess.CalledProcessError(retcode, cmd)
error.output = output
raise error
return output
if sys.platform != 'win32':
if sys.version_info[0] < 3:
from commands import getstatusoutput
else:
from subprocess import getstatusoutput
if PY3:
import configparser
else:
import ConfigParser as configparser
# matplotlib build options, which can be altered using setup.cfg
options = {
'display_status': True,
'verbose': False,
'backend': None,
'basedirlist': None
}
setup_cfg = os.environ.get('MPLSETUPCFG', 'setup.cfg')
if os.path.exists(setup_cfg):
config = configparser.SafeConfigParser()
config.read(setup_cfg)
try:
options['display_status'] = not config.getboolean("status", "suppress")
except:
pass
try:
options['backend'] = config.get("rc_options", "backend")
except:
pass
try:
options['basedirlist'] = [
x.strip() for x in
config.get("directories", "basedirlist").split(',')]
except:
pass
else:
config = None
def get_win32_compiler():
"""
Determine the compiler being used on win32.
"""
# Used to determine mingw32 or msvc
# This is pretty bad logic, someone know a better way?
for v in sys.argv:
if 'mingw32' in v:
return 'mingw32'
return 'msvc'
win32_compiler = get_win32_compiler()
def extract_versions():
"""
Extracts version values from the main matplotlib __init__.py and
returns them as a dictionary.
"""
with open('lib/matplotlib/__init__.py') as fd:
for line in fd.readlines():
if (line.startswith('__version__')):
exec(line.strip())
return locals()
def has_include_file(include_dirs, filename):
"""
Returns `True` if `filename` can be found in one of the
directories in `include_dirs`.
"""
for dir in include_dirs:
if os.path.exists(os.path.join(dir, filename)):
return True
return False
def check_include_file(include_dirs, filename, package):
"""
Raises an exception if the given include file can not be found.
"""
if sys.platform == 'win32':
include_dirs.extend(os.getenv('INCLUDE', '.').split(';'))
if not has_include_file(include_dirs, filename):
raise CheckFailed(
"The C/C++ header for %s (%s) could not be found. You "
"may need to install the development package." %
(package, filename))
def get_base_dirs():
"""
Returns a list of standard base directories on this platform.
"""
if options['basedirlist']:
return options['basedirlist']
basedir_map = {
'win32': ['win32_static',],
'darwin': ['/usr/local/', '/usr', '/usr/X11', '/opt/local'],
'sunos5': [os.getenv('MPLIB_BASE') or '/usr/local',],
'gnu0': ['/usr'],
'aix5': ['/usr/local'],
}
return basedir_map.get(sys.platform, ['/usr/local', '/usr'])
def is_min_version(found, minversion):
"""
Returns `True` if `found` is at least as high a version as
`minversion`.
"""
expected_version = version.LooseVersion(minversion)
found_version = version.LooseVersion(found)
return found_version >= expected_version
# Define the display functions only if display_status is True.
if options['display_status']:
def print_line(char='='):
print(char * 76)
def print_status(package, status):
initial_indent = "%22s: " % package
indent = ' ' * 24
print(fill(str(status), width=76,
initial_indent=initial_indent,
subsequent_indent=indent))
def print_message(message):
indent = ' ' * 24 + "* "
print(fill(str(message), width=76,
initial_indent=indent,
subsequent_indent=indent))
def print_raw(section):
print(section)
else:
def print_line(*args, **kwargs):
pass
print_status = print_message = print_raw = print_line
# Remove the -Wstrict-prototypesoption, is it's not valid for C++
customize_compiler = sysconfig.customize_compiler
def my_customize_compiler(compiler):
retval = customize_compiler(compiler)
try:
compiler.compiler_so.remove('-Wstrict-prototypes')
except (ValueError, AttributeError):
pass
return retval
sysconfig.customize_compiler = my_customize_compiler
def make_extension(name, files, *args, **kwargs):
"""
Make a new extension. Automatically sets include_dirs and
library_dirs to the base directories appropriate for this
platform.
`name` is the name of the extension.
`files` is a list of source files.
Any additional arguments are passed to the
`distutils.core.Extension` constructor.
"""
ext = DelayedExtension(name, files, *args, **kwargs)
for dir in get_base_dirs():
include_dir = os.path.join(dir, 'include')
if os.path.exists(include_dir):
ext.include_dirs.append(include_dir)
for lib in ('lib', 'lib64'):
lib_dir = os.path.join(dir, lib)
if os.path.exists(lib_dir):
ext.library_dirs.append(lib_dir)
ext.include_dirs.append('.')
return ext
class PkgConfig(object):
"""
This is a class for communicating with pkg-config.
"""
def __init__(self):
"""
Determines whether pkg-config exists on this machine.
"""
if sys.platform == 'win32':
self.has_pkgconfig = False
else:
self.set_pkgconfig_path()
status, output = getstatusoutput("pkg-config --help")
self.has_pkgconfig = (status == 0)
def set_pkgconfig_path(self):
pkgconfig_path = sysconfig.get_config_var('LIBDIR')
if pkgconfig_path is None:
return
pkgconfig_path = os.path.join(pkgconfig_path, 'pkgconfig')
if not os.path.isdir(pkgconfig_path):
return
try:
os.environ['PKG_CONFIG_PATH'] += ':' + pkgconfig_path
except KeyError:
os.environ['PKG_CONFIG_PATH'] = pkgconfig_path
def setup_extension(self, ext, package, default_include_dirs=[],
default_library_dirs=[], default_libraries=[],
alt_exec=None):
"""
Add parameters to the given `ext` for the given `package`.
"""
flag_map = {
'-I': 'include_dirs', '-L': 'library_dirs', '-l': 'libraries'}
executable = alt_exec
if self.has_pkgconfig:
executable = 'pkg-config {0}'.format(package)
use_defaults = True
if executable is not None:
command = "{0} --libs --cflags ".format(executable)
try:
output = check_output(command, shell=True,
stderr=subprocess.STDOUT)
except subprocess.CalledProcessError:
pass
else:
output = output.decode(sys.getfilesystemencoding())
use_defaults = False
for token in output.split():
attr = flag_map.get(token[:2])
if attr is not None:
getattr(ext, attr).insert(0, token[2:])
if use_defaults:
basedirs = get_base_dirs()
for base in basedirs:
for include in default_include_dirs:
dir = os.path.join(base, include)
if os.path.exists(dir):
ext.include_dirs.append(dir)
for lib in default_library_dirs:
dir = os.path.join(base, lib)
if os.path.exists(dir):
ext.library_dirs.append(dir)
ext.libraries.extend(default_libraries)
return True
return False
def get_version(self, package):
"""
Get the version of the package from pkg-config.
"""
if not self.has_pkgconfig:
return None
status, output = getstatusoutput(
"pkg-config %s --modversion" % (package))
if status == 0:
return output
return None
# The PkgConfig class should be used through this singleton
pkg_config = PkgConfig()
class CheckFailed(Exception):
"""
Exception thrown when a `SetupPackage.check` method fails.
"""
pass
class SetupPackage(object):
optional = False
def check(self):
"""
Checks whether the dependencies are met. Should raise a
`CheckFailed` exception if the dependency could not be met,
otherwise return a string indicating a version number or some
other message indicating what was found.
"""
pass
def get_packages(self):
"""
Get a list of package names to add to the configuration.
These are added to the `packages` list passed to
`distutils.setup`.
"""
return []
def get_namespace_packages(self):
"""
Get a list of namespace package names to add to the configuration.
These are added to the `namespace_packages` list passed to
`distutils.setup`.
"""
return []
def get_py_modules(self):
"""
Get a list of top-level modules to add to the configuration.
These are added to the `py_modules` list passed to
`distutils.setup`.
"""
return []
def get_package_data(self):
"""
Get a package data dictionary to add to the configuration.
These are merged into to the `package_data` list passed to
`distutils.setup`.
"""
return {}
def get_extension(self):
"""
Get a list of C extensions (`distutils.core.Extension`
objects) to add to the configuration. These are added to the
`extensions` list passed to `distutils.setup`.
"""
return None
def get_install_requires(self):
"""
Get a list of Python packages that we require.
pip/easy_install will attempt to download and install this
package if it is not installed.
"""
return []
def get_setup_requires(self):
"""
Get a list of Python packages that we require at build time.
pip/easy_install will attempt to download and install this
package if it is not installed.
"""
return []
def _check_for_pkg_config(self, package, include_file, min_version=None,
version=None):
"""
A convenience function for writing checks for a
pkg_config-defined dependency.
`package` is the pkg_config package name.
`include_file` is a top-level include file we expect to find.
`min_version` is the minimum version required.
`version` will override the found version if this package
requires an alternate method for that.
"""
if version is None:
version = pkg_config.get_version(package)
if version is None:
raise CheckFailed(
"pkg-config information for '%s' could not be found." %
package)
if min_version == 'PATCH':
raise CheckFailed(
"Requires patches that have not been merged upstream.")
if min_version:
if (not is_min_version(version, min_version)):
raise CheckFailed(
"Requires %s %s or later. Found %s." %
(package, min_version, version))
ext = self.get_extension()
if ext is None:
ext = make_extension('test', [])
pkg_config.setup_extension(ext, package)
check_include_file(ext.include_dirs, include_file, package)
return 'version %s' % version
class OptionalPackage(SetupPackage):
optional = True
force = False
config_category = "packages"
def get_config(self):
"""
Look at `setup.cfg` and return one of ["auto", True, False] indicating
if the package is at default state ("auto"), forced by the user (True)
or opted-out (False).
"""
try:
return config.getboolean(self.config_category, self.name)
except:
return "auto"
def check(self):
"""
Do not override this method!
For custom dependency checks override self.check_requirements().
Two things are checked: Configuration file and requirements.
"""
# Check configuration file
conf = self.get_config()
# Default "auto" state or install forced by user
if conf in [True, 'auto']:
message = "installing"
# Set non-optional if user sets `True` in config
if conf is True:
self.optional = False
# Configuration opt-out by user
else:
# Some backend extensions (e.g. Agg) need to be built for certain
# other GUI backends (e.g. TkAgg) even when manually disabled
if self.force is True:
message = "installing forced (config override)"
else:
raise CheckFailed("skipping due to configuration")
# Check requirements and add extra information (if any) to message.
# If requirements are not met a CheckFailed should be raised in there.
additional_info = self.check_requirements()
if additional_info:
message += ", " + additional_info
# No CheckFailed raised until now, return install message.
return message
def check_requirements(self):
"""
Override this method to do custom dependency checks.
- Raise CheckFailed() if requirements are not met.
- Return message with additional information, or an empty string
(or None) for no additional information.
"""
return ""
class OptionalBackendPackage(OptionalPackage):
config_category = "gui_support"
class Platform(SetupPackage):
name = "platform"
def check(self):
return sys.platform
class Python(SetupPackage):
name = "python"
def check(self):
major, minor1, minor2, s, tmp = sys.version_info
if major < 2:
raise CheckFailed(
"Requires Python 2.6 or later")
elif major == 2 and minor1 < 6:
raise CheckFailed(
"Requires Python 2.6 or later (in the 2.x series)")
elif major == 3 and minor1 < 1:
raise CheckFailed(
"Requires Python 3.1 or later (in the 3.x series)")
return sys.version
class Matplotlib(SetupPackage):
name = "matplotlib"
def check(self):
return extract_versions()['__version__']
def get_packages(self):
return [
'matplotlib',
'matplotlib.backends',
'matplotlib.backends.qt_editor',
'matplotlib.compat',
'matplotlib.projections',
'matplotlib.axes',
'matplotlib.sphinxext',
'matplotlib.style',
'matplotlib.testing',
'matplotlib.testing.jpl_units',
'matplotlib.tri',
]
def get_py_modules(self):
return ['pylab']
def get_package_data(self):
return {
'matplotlib':
[
'mpl-data/fonts/afm/*.afm',
'mpl-data/fonts/pdfcorefonts/*.afm',
'mpl-data/fonts/pdfcorefonts/*.txt',
'mpl-data/fonts/ttf/*.ttf',
'mpl-data/fonts/ttf/LICENSE_STIX',
'mpl-data/fonts/ttf/COPYRIGHT.TXT',
'mpl-data/fonts/ttf/README.TXT',
'mpl-data/fonts/ttf/RELEASENOTES.TXT',
'mpl-data/images/*.xpm',
'mpl-data/images/*.svg',
'mpl-data/images/*.gif',
'mpl-data/images/*.png',
'mpl-data/images/*.ppm',
'mpl-data/example/*.npy',
'mpl-data/matplotlibrc',
'backends/web_backend/*.*',
'backends/web_backend/jquery/js/*',
'backends/web_backend/jquery/css/themes/base/*.*',
'backends/web_backend/jquery/css/themes/base/images/*',
'backends/web_backend/css/*.*',
'backends/Matplotlib.nib/*',
'style/stylelib/*.mplstyle',
]}
class SampleData(OptionalPackage):
"""
This handles the sample data that ships with matplotlib. It is
technically optional, though most often will be desired.
"""
name = "sample_data"
def get_package_data(self):
return {
'matplotlib':
[
'mpl-data/sample_data/*.*',
'mpl-data/sample_data/axes_grid/*.*',
]}
class Toolkits(OptionalPackage):
name = "toolkits"
def get_packages(self):
return [
'mpl_toolkits',
'mpl_toolkits.mplot3d',
'mpl_toolkits.axes_grid',
'mpl_toolkits.axes_grid1',
'mpl_toolkits.axisartist',
]
def get_namespace_packages(self):
return ['mpl_toolkits']
class Tests(OptionalPackage):
name = "tests"
nose_min_version = '0.11.1'
def check(self):
super(Tests, self).check()
msgs = []
msg_template = ('{package} is required to run the matplotlib test '
'suite. pip/easy_install may attempt to install it '
'after matplotlib.')
bad_nose = msg_template.format(
package='nose %s or later' % self.nose_min_version
)
try:
import nose
if is_min_version(nose.__version__, self.nose_min_version):
msgs += ['using nose version %s' % nose.__version__]
else:
msgs += [bad_nose]
except ImportError:
msgs += [bad_nose]
if sys.version_info >= (3, 3):
msgs += ['using unittest.mock']
else:
try:
import mock
msgs += ['using mock %s' % mock.__version__]
except ImportError:
msgs += [msg_template.format(package='mock')]
return ' / '.join(msgs)
def get_packages(self):
return [
'matplotlib.tests',
]
def get_package_data(self):
baseline_images = [
'tests/baseline_images/%s/*' % x
for x in os.listdir('lib/matplotlib/tests/baseline_images')]
return {
'matplotlib':
baseline_images +
[
'tests/mpltest.ttf',
'tests/test_rcparams.rc'
]}
def get_install_requires(self):
requires = ['nose>=%s' % self.nose_min_version]
if not sys.version_info >= (3, 3):
requires += ['mock']
return requires
class DelayedExtension(Extension, object):
"""
A distutils Extension subclass where some of its members
may have delayed computation until reaching the build phase.
This is so we can, for example, get the Numpy include dirs
after pip has installed Numpy for us if it wasn't already
on the system.
"""
def __init__(self, *args, **kwargs):
super(DelayedExtension, self).__init__(*args, **kwargs)
self._finalized = False
self._hooks = {}
def add_hook(self, member, func):
"""
Add a hook to dynamically compute a member.
Parameters
----------
member : string
The name of the member
func : callable
The function to call to get dynamically-computed values
for the member.
"""
self._hooks[member] = func
def finalize(self):
self._finalized = True
class DelayedMember(property):
def __init__(self, name):
self._name = name
def __get__(self, obj, objtype=None):
result = getattr(obj, '_' + self._name, [])
if obj._finalized:
if self._name in obj._hooks:
result = obj._hooks[self._name]() + result
return result
def __set__(self, obj, value):
setattr(obj, '_' + self._name, value)
include_dirs = DelayedMember('include_dirs')
class Numpy(SetupPackage):
name = "numpy"
@staticmethod
def include_dirs_hook():
if sys.version_info[0] >= 3:
import builtins
if hasattr(builtins, '__NUMPY_SETUP__'):
del builtins.__NUMPY_SETUP__
import imp
import numpy
imp.reload(numpy)
else:
import __builtin__
if hasattr(__builtin__, '__NUMPY_SETUP__'):
del __builtin__.__NUMPY_SETUP__
import numpy
reload(numpy)
ext = Extension('test', [])
ext.include_dirs.append(numpy.get_include())
if not has_include_file(
ext.include_dirs, os.path.join("numpy", "arrayobject.h")):
warnings.warn(
"The C headers for numpy could not be found. "
"You may need to install the development package")
return [numpy.get_include()]
def check(self):
min_version = extract_versions()['__version__numpy__']
try:
import numpy
except ImportError:
return 'not found. pip may install it below.'
if not is_min_version(numpy.__version__, min_version):
raise SystemExit(
"Requires numpy %s or later to build. (Found %s)" %
(min_version, numpy.__version__))
return 'version %s' % numpy.__version__
def add_flags(self, ext):
# Ensure that PY_ARRAY_UNIQUE_SYMBOL is uniquely defined for
# each extension
array_api_name = 'MPL_' + ext.name.replace('.', '_') + '_ARRAY_API'
ext.define_macros.append(('PY_ARRAY_UNIQUE_SYMBOL', array_api_name))
ext.add_hook('include_dirs', self.include_dirs_hook)
def get_setup_requires(self):
return ['numpy>=1.5']
def get_install_requires(self):
return ['numpy>=1.5']
class CXX(SetupPackage):
name = 'pycxx'
def check(self):
if PY3:
# There is no version of PyCXX in the wild that will work
# with Python 3.x
self.__class__.found_external = False
return ("Official versions of PyCXX are not compatible with "
"Python 3.x. Using local copy")
self.__class__.found_external = True
old_stdout = sys.stdout
if PY3:
sys.stdout = io.StringIO()
else:
sys.stdout = io.BytesIO()
try:
import CXX
except ImportError:
self.__class__.found_external = False
return "Couldn't import. Using local copy."
finally:
sys.stdout = old_stdout
try:
return self._check_for_pkg_config(
'PyCXX', 'CXX/Extensions.hxx', min_version='6.2.4')
except CheckFailed as e:
# It's ok to just proceed here, since the `import CXX`
# worked above, and PyCXX (at least upstream) ensures that
# its header files are on the default distutils include
# path (either in a standard C place such as /usr/include,
# or in /usr/include/pythonX.Y.
return 'Using system CXX (version unknown, no pkg-config info)'
def add_flags(self, ext):
if self.found_external and not 'sdist' in sys.argv:
support_dir = os.path.normpath(
os.path.join(
sys.prefix,
'share',
'python%d.%d' % (
sys.version_info[0], sys.version_info[1]),
'CXX'))
if not os.path.exists(support_dir):
# On Fedora 17, these files are installed in /usr/share/CXX
support_dir = '/usr/src/CXX'
ext.sources.extend([
os.path.join(support_dir, x) for x in
['cxxsupport.cxx', 'cxx_extensions.cxx',
'IndirectPythonInterface.cxx',
'cxxextensions.c']])
pkg_config.setup_extension(ext, 'PyCXX')
else:
ext.include_dirs.append('extern')
ext.sources.extend(glob.glob('extern/CXX/*.cxx'))
ext.sources.extend(glob.glob('extern/CXX/*.c'))
ext.define_macros.append(('PYCXX_ISO_CPP_LIB', '1'))
if PY3:
ext.define_macros.append(('PYCXX_PYTHON_2TO3', '1'))
if not (sys.platform == 'win32' and win32_compiler == 'msvc'):
ext.libraries.append('stdc++')
ext.libraries.append('m')
class LibAgg(SetupPackage):
name = 'libagg'
def check(self):
self.__class__.found_external = True
try:
return self._check_for_pkg_config(
'libagg', 'agg2/agg_basics.h', min_version='PATCH')
except CheckFailed as e:
self.__class__.found_external = False
return str(e) + ' Using local copy.'
def add_flags(self, ext):
if self.found_external:
pkg_config.setup_extension(ext, 'libagg')
else:
ext.include_dirs.append('extern/agg24/include')
agg_sources = [
'agg_bezier_arc.cpp',
'agg_curves.cpp',
'agg_image_filters.cpp',
'agg_trans_affine.cpp',
'agg_vcgen_contour.cpp',
'agg_vcgen_dash.cpp',
'agg_vcgen_stroke.cpp',
'agg_vpgen_segmentator.cpp'
]
ext.sources.extend(
os.path.join('extern', 'agg24', 'src', x) for x in agg_sources)
class FreeType(SetupPackage):
name = "freetype"
def check(self):
if sys.platform == 'win32':
return "Unknown version"
status, output = getstatusoutput("freetype-config --version")
if status == 0:
version = output
else:
version = None
return self._check_for_pkg_config(
'freetype2', 'ft2build.h',
min_version='2.4', version=version)
def add_flags(self, ext):
pkg_config.setup_extension(
ext, 'freetype2',
default_include_dirs=[
'freetype2', 'lib/freetype2/include',
'lib/freetype2/include/freetype2'],
default_library_dirs=[
'freetype2/lib'],
default_libraries=['freetype', 'z'],
alt_exec='freetype-config')
def get_extension(self):
ext = make_extension('freetype2', [])
self.add_flags(ext)
return ext
class FT2Font(SetupPackage):
name = 'ft2font'
def get_extension(self):
sources = [
'src/ft2font.cpp',
'src/mplutils.cpp'
]
ext = make_extension('matplotlib.ft2font', sources)
FreeType().add_flags(ext)
Numpy().add_flags(ext)
CXX().add_flags(ext)
return ext
class Png(SetupPackage):
name = "png"
def check(self):
try:
return self._check_for_pkg_config(
'libpng', 'png.h',
min_version='1.2')
except CheckFailed as e:
self.__class__.found_external = False
return str(e) + ' Using unknown version.'
def get_extension(self):
sources = [
'src/_png.cpp', 'src/mplutils.cpp'
]
ext = make_extension('matplotlib._png', sources)
pkg_config.setup_extension(
ext, 'libpng', default_libraries=['png', 'z'])
Numpy().add_flags(ext)
CXX().add_flags(ext)
return ext
class Qhull(SetupPackage):
name = "qhull"
def check(self):
self.__class__.found_external = True
try:
return self._check_for_pkg_config(
'qhull', 'qhull/qhull_a.h', min_version='2003.1')
except CheckFailed as e:
self.__class__.found_pkgconfig = False
# Qhull may not be in the pkg-config system but may still be
# present on this system, so check if the header files can be
# found.
include_dirs = [
os.path.join(x, 'include', 'qhull') for x in get_base_dirs()]
if has_include_file(include_dirs, 'qhull_a.h'):
return 'Using system Qhull (version unknown, no pkg-config info)'
else:
self.__class__.found_external = False
return str(e) + ' Using local copy.'
def add_flags(self, ext):
if self.found_external:
pkg_config.setup_extension(ext, 'qhull',
default_libraries=['qhull'])
else:
ext.include_dirs.append('extern')
ext.sources.extend(glob.glob('extern/qhull/*.c'))
class TTConv(SetupPackage):
name = "ttconv"
def get_extension(self):
sources = [
'src/_ttconv.cpp',
'extern/ttconv/pprdrv_tt.cpp',
'extern/ttconv/pprdrv_tt2.cpp',
'extern/ttconv/ttutil.cpp'
]
ext = make_extension('matplotlib.ttconv', sources)
Numpy().add_flags(ext)
CXX().add_flags(ext)
ext.include_dirs.append('extern')
return ext
class Path(SetupPackage):
name = "path"
def get_extension(self):
sources = [
'src/_path.cpp',
'src/path_cleanup.cpp',
'src/agg_py_transforms.cpp'
]
ext = make_extension('matplotlib._path', sources)
Numpy().add_flags(ext)
LibAgg().add_flags(ext)
CXX().add_flags(ext)
return ext
class Image(SetupPackage):
name = "image"
def get_extension(self):
sources = [
'src/_image.cpp', 'src/mplutils.cpp'
]
ext = make_extension('matplotlib._image', sources)
Numpy().add_flags(ext)
LibAgg().add_flags(ext)
CXX().add_flags(ext)
return ext
class Contour(SetupPackage):
name = "contour"
def get_extension(self):
sources = [
"src/cntr.c"
]
ext = make_extension('matplotlib._cntr', sources)
Numpy().add_flags(ext)
return ext
class Delaunay(SetupPackage):
name = "delaunay"
def get_packages(self):
return ['matplotlib.delaunay']
def get_extension(self):
sources = ["_delaunay.cpp", "VoronoiDiagramGenerator.cpp",
"delaunay_utils.cpp", "natneighbors.cpp"]
sources = [os.path.join('lib/matplotlib/delaunay', s) for s in sources]
ext = make_extension('matplotlib._delaunay', sources)
Numpy().add_flags(ext)
return ext
class QhullWrap(SetupPackage):
name = "qhull_wrap"
def get_extension(self):
sources = ['src/qhull_wrap.c']
ext = make_extension('matplotlib._qhull', sources,
define_macros=[('MPL_DEVNULL', os.devnull)])
Numpy().add_flags(ext)
Qhull().add_flags(ext)
return ext
class Tri(SetupPackage):
name = "tri"
def get_extension(self):
sources = [
"lib/matplotlib/tri/_tri.cpp",
"src/mplutils.cpp"
]
ext = make_extension('matplotlib._tri', sources)
Numpy().add_flags(ext)
CXX().add_flags(ext)
return ext
class Six(SetupPackage):
name = "six"
min_version = "1.3"
def check(self):
try:
import six
except ImportError:
return (
"six was not found.")
if not is_min_version(six.__version__, self.min_version):
raise CheckFailed(
"Requires six %s or later. Found %s." %
(self.min_version, six.__version__))
return "using six version %s" % six.__version__
def get_install_requires(self):
return ['six>={0}'.format(self.min_version)]
class Dateutil(SetupPackage):
name = "dateutil"
def __init__(self, version=None):
self.version = version
def check(self):
try:
import dateutil
except ImportError:
# dateutil 2.1 has a file encoding bug that breaks installation on
# python 3.3
# https://github.com/matplotlib/matplotlib/issues/2373
# hack around the problem by installing the the (working) v2.0
major, minor1, _, _, _ = sys.version_info
if self.version is None and (major, minor1) == (3, 3):
self.version = '!=2.1'
return (
"dateutil was not found. It is required for date axis "
"support. pip/easy_install may attempt to install it "
"after matplotlib.")
return "using dateutil version %s" % dateutil.__version__
def get_install_requires(self):
dateutil = 'python-dateutil'
if self.version is not None:
dateutil += self.version
return [dateutil]
class Tornado(OptionalPackage):
name = "tornado"
def check(self):
try:
import tornado
except ImportError:
return (
"tornado was not found. It is required for the WebAgg "
"backend. pip/easy_install may attempt to install it "
"after matplotlib.")
return "using tornado version %s" % tornado.version
class Pyparsing(SetupPackage):
name = "pyparsing"
def is_ok(self):
# pyparsing 2.0.0 bug, but it may be patched in distributions
try:
import pyparsing
f = pyparsing.Forward()
f <<= pyparsing.Literal('a')
return f is not None
except (ImportError, TypeError):
return False
def check(self):
try:
import pyparsing
except ImportError:
return (
"pyparsing was not found. It is required for mathtext "
"support. pip/easy_install may attempt to install it "
"after matplotlib.")
required = [1, 5, 6]
if [int(x) for x in pyparsing.__version__.split('.')] < required:
return (
"matplotlib requires pyparsing >= {0}".format(
'.'.join(str(x) for x in required)))
if not self.is_ok():
return (
"Your pyparsing contains a bug that will be monkey-patched by "
"matplotlib. For best results, upgrade to pyparsing 2.0.1 or "
"later.")
return "using pyparsing version %s" % pyparsing.__version__
def get_install_requires(self):
if self.is_ok():
return ['pyparsing>=1.5.6']
else:
return ['pyparsing>=1.5.6,!=2.0.0']
class BackendAgg(OptionalBackendPackage):
name = "agg"
def get_extension(self):
sources = [
"src/mplutils.cpp",
"src/agg_py_transforms.cpp",
"src/_backend_agg.cpp"
]
ext = make_extension('matplotlib.backends._backend_agg', sources)
Numpy().add_flags(ext)
LibAgg().add_flags(ext)
FreeType().add_flags(ext)
CXX().add_flags(ext)
return ext
class BackendTkAgg(OptionalBackendPackage):
name = "tkagg"
def __init__(self):
self.tcl_tk_cache = None
def check_requirements(self):
try:
if PY3:
import tkinter as Tkinter
else:
import Tkinter
except ImportError:
raise CheckFailed('TKAgg requires Tkinter.')
except RuntimeError:
raise CheckFailed('Tkinter present but import failed.')
else:
if Tkinter.TkVersion < 8.3:
raise CheckFailed("Tcl/Tk v8.3 or later required.")
ext = self.get_extension()
check_include_file(ext.include_dirs, "tk.h", "Tk")
try:
tk_v = Tkinter.__version__.split()[-2]
except (AttributeError, IndexError):
# Tkinter.__version__ has been removed in python 3
tk_v = 'version not identified'
BackendAgg.force = True
return "version %s" % tk_v
def get_extension(self):
sources = [
'src/agg_py_transforms.cpp',
'src/_tkagg.cpp'
]
ext = make_extension('matplotlib.backends._tkagg', sources)
self.add_flags(ext)
Numpy().add_flags(ext)
LibAgg().add_flags(ext)
CXX().add_flags(ext)
return ext
def query_tcltk(self):
"""
Tries to open a Tk window in order to query the Tk object
about its library paths. This should never be called more
than once by the same process, as Tk intricacies may cause the
Python interpreter to hang. The function also has a workaround
if no X server is running (useful for autobuild systems).
"""
# Use cached values if they exist, which ensures this function
# only executes once
if self.tcl_tk_cache is not None:
return self.tcl_tk_cache
# By this point, we already know that Tkinter imports correctly
if PY3:
import tkinter as Tkinter
else:
import Tkinter
tcl_lib_dir = ''
tk_lib_dir = ''
# First try to open a Tk window (requires a running X server)
try:
tk = Tkinter.Tk()
except Tkinter.TclError:
# Next, start Tcl interpreter without opening a Tk window
# (no need for X server) This feature is available in
# python version 2.4 and up
try:
tcl = Tkinter.Tcl()
except AttributeError: # Python version not high enough
pass
except Tkinter.TclError: # Something went wrong while opening Tcl
pass
else:
tcl_lib_dir = str(tcl.getvar('tcl_library'))
# Guess Tk location based on Tcl location
(head, tail) = os.path.split(tcl_lib_dir)
tail = tail.replace('Tcl', 'Tk').replace('tcl', 'tk')
tk_lib_dir = os.path.join(head, tail)
if not os.path.exists(tk_lib_dir):
tk_lib_dir = tcl_lib_dir.replace(
'Tcl', 'Tk').replace('tcl', 'tk')
else:
# Obtain Tcl and Tk locations from Tk widget
tk.withdraw()
tcl_lib_dir = str(tk.getvar('tcl_library'))
tk_lib_dir = str(tk.getvar('tk_library'))
tk.destroy()
# Save directories and version string to cache
self.tcl_tk_cache = tcl_lib_dir, tk_lib_dir, str(Tkinter.TkVersion)[:3]
return self.tcl_tk_cache
def parse_tcl_config(self, tcl_lib_dir, tk_lib_dir):
try:
if PY3:
import tkinter as Tkinter
else:
import Tkinter
except ImportError:
return None
tcl_poss = [tcl_lib_dir,
os.path.normpath(os.path.join(tcl_lib_dir, '..')),
"/usr/lib/tcl" + str(Tkinter.TclVersion),
"/usr/lib"]
tk_poss = [tk_lib_dir,
os.path.normpath(os.path.join(tk_lib_dir, '..')),
"/usr/lib/tk" + str(Tkinter.TkVersion),
"/usr/lib"]
for ptcl, ptk in zip(tcl_poss, tk_poss):
tcl_config = os.path.join(ptcl, "tclConfig.sh")
tk_config = os.path.join(ptk, "tkConfig.sh")
if (os.path.exists(tcl_config) and os.path.exists(tk_config)):
break
if not (os.path.exists(tcl_config) and os.path.exists(tk_config)):
return None
def get_var(file, varname):
p = subprocess.Popen(
'. %s ; eval echo ${%s}' % (file, varname),
shell=True,
executable="/bin/sh",
stdout=subprocess.PIPE)
result = p.communicate()[0]
return result.decode('ascii')
tcl_lib_dir = get_var(
tcl_config, 'TCL_LIB_SPEC').split()[0][2:].strip()
tcl_inc_dir = get_var(
tcl_config, 'TCL_INCLUDE_SPEC')[2:].strip()
tcl_lib = get_var(tcl_config, 'TCL_LIB_FLAG')[2:].strip()
tk_lib_dir = get_var(tk_config, 'TK_LIB_SPEC').split()[0][2:].strip()
tk_inc_dir = get_var(tk_config, 'TK_INCLUDE_SPEC').strip()
if tk_inc_dir == '':
tk_inc_dir = tcl_inc_dir
else:
tk_inc_dir = tk_inc_dir[2:]
tk_lib = get_var(tk_config, 'TK_LIB_FLAG')[2:].strip()
if not os.path.exists(os.path.join(tk_inc_dir, 'tk.h')):
return None
return (tcl_lib_dir, tcl_inc_dir, tcl_lib,
tk_lib_dir, tk_inc_dir, tk_lib)
def guess_tcl_config(self, tcl_lib_dir, tk_lib_dir, tk_ver):
if not (os.path.exists(tcl_lib_dir) and os.path.exists(tk_lib_dir)):
return None
tcl_lib = os.path.normpath(os.path.join(tcl_lib_dir, '../'))
tk_lib = os.path.normpath(os.path.join(tk_lib_dir, '../'))
tcl_inc = os.path.normpath(
os.path.join(tcl_lib_dir,
'../../include/tcl' + tk_ver))
if not os.path.exists(tcl_inc):
tcl_inc = os.path.normpath(
os.path.join(tcl_lib_dir,
'../../include'))
tk_inc = os.path.normpath(os.path.join(
tk_lib_dir,
'../../include/tk' + tk_ver))
if not os.path.exists(tk_inc):
tk_inc = os.path.normpath(os.path.join(
tk_lib_dir,
'../../include'))
if not os.path.exists(os.path.join(tk_inc, 'tk.h')):
tk_inc = tcl_inc
if not os.path.exists(tcl_inc):
# this is a hack for suse linux, which is broken
if (sys.platform.startswith('linux') and
os.path.exists('/usr/include/tcl.h') and
os.path.exists('/usr/include/tk.h')):
tcl_inc = '/usr/include'
tk_inc = '/usr/include'
if not os.path.exists(os.path.join(tk_inc, 'tk.h')):
return None
return tcl_lib, tcl_inc, 'tcl' + tk_ver, tk_lib, tk_inc, 'tk' + tk_ver
def hardcoded_tcl_config(self):
tcl_inc = "/usr/local/include"
tk_inc = "/usr/local/include"
tcl_lib = "/usr/local/lib"
tk_lib = "/usr/local/lib"
return tcl_lib, tcl_inc, 'tcl', tk_lib, tk_inc, 'tk'
def add_flags(self, ext):
if sys.platform == 'win32':
major, minor1, minor2, s, tmp = sys.version_info
if sys.version_info[0:2] < (3, 4):
ext.include_dirs.extend(['win32_static/include/tcl85'])
ext.libraries.extend(['tk85', 'tcl85'])
else:
ext.include_dirs.extend(['win32_static/include/tcl86'])
ext.libraries.extend(['tk86t', 'tcl86t'])
ext.library_dirs.extend([os.path.join(sys.prefix, 'dlls')])
elif sys.platform == 'darwin':
# this config section lifted directly from Imaging - thanks to
# the effbot!
# First test for a MacOSX/darwin framework install
from os.path import join, exists
framework_dirs = [
join(os.getenv('HOME'), '/Library/Frameworks'),
'/Library/Frameworks',
'/System/Library/Frameworks/',
]
# Find the directory that contains the Tcl.framework and
# Tk.framework bundles.
tk_framework_found = 0
for F in framework_dirs:
# both Tcl.framework and Tk.framework should be present
for fw in 'Tcl', 'Tk':
if not exists(join(F, fw + '.framework')):
break
else:
# ok, F is now directory with both frameworks. Continure
# building
tk_framework_found = 1
break
if tk_framework_found:
# For 8.4a2, we must add -I options that point inside
# the Tcl and Tk frameworks. In later release we
# should hopefully be able to pass the -F option to
# gcc, which specifies a framework lookup path.
tk_include_dirs = [
join(F, fw + '.framework', H)
for fw in ('Tcl', 'Tk')
for H in ('Headers', 'Versions/Current/PrivateHeaders')
]
# For 8.4a2, the X11 headers are not included. Rather
# than include a complicated search, this is a
# hard-coded path. It could bail out if X11 libs are
# not found...
# tk_include_dirs.append('/usr/X11R6/include')
frameworks = ['-framework', 'Tcl', '-framework', 'Tk']
ext.include_dirs.extend(tk_include_dirs)
ext.extra_link_args.extend(frameworks)
ext.extra_compile_args.extend(frameworks)
# you're still here? ok we'll try it this way...
else:
# There are 3 methods to try, in decreasing order of "smartness"
#
# 1. Parse the tclConfig.sh and tkConfig.sh files that have
# all the information we need
#
# 2. Guess the include and lib dirs based on the location of
# Tkinter's 'tcl_library' and 'tk_library' variables.
#
# 3. Use some hardcoded locations that seem to work on a lot
# of distros.
# Query Tcl/Tk system for library paths and version string
try:
tcl_lib_dir, tk_lib_dir, tk_ver = self.query_tcltk()
except:
tk_ver = ''
result = self.hardcoded_tcl_config()
else:
result = self.parse_tcl_config(tcl_lib_dir, tk_lib_dir)
if result is None:
result = self.guess_tcl_config(
tcl_lib_dir, tk_lib_dir, tk_ver)
if result is None:
result = self.hardcoded_tcl_config()
# Add final versions of directories and libraries to ext lists
(tcl_lib_dir, tcl_inc_dir, tcl_lib,
tk_lib_dir, tk_inc_dir, tk_lib) = result
ext.include_dirs.extend([tcl_inc_dir, tk_inc_dir])
ext.library_dirs.extend([tcl_lib_dir, tk_lib_dir])
ext.libraries.extend([tcl_lib, tk_lib])
class BackendGtk(OptionalBackendPackage):
name = "gtk"
def check_requirements(self):
try:
import gtk
except ImportError:
raise CheckFailed("Requires pygtk")
except RuntimeError:
raise CheckFailed('pygtk present, but import failed.')
else:
version = (2, 2, 0)
if gtk.pygtk_version < version:
raise CheckFailed(
"Requires pygtk %d.%d.%d or later. "
"Found %d.%d.%d" % (version + gtk.pygtk_version))
ext = self.get_extension()
self.add_flags(ext)
check_include_file(ext.include_dirs,
os.path.join("gtk", "gtk.h"),
'gtk')
check_include_file(ext.include_dirs,
os.path.join("pygtk", "pygtk.h"),
'pygtk')
return 'Gtk: %s pygtk: %s' % (
".".join(str(x) for x in gtk.gtk_version),
".".join(str(x) for x in gtk.pygtk_version))
def get_package_data(self):
return {'matplotlib': ['mpl-data/*.glade']}
def get_extension(self):
sources = [
'src/_backend_gdk.c'
]
ext = make_extension('matplotlib.backends._backend_gdk', sources)
self.add_flags(ext)
Numpy().add_flags(ext)
return ext
def add_flags(self, ext):
if sys.platform == 'win32':
def getoutput(s):
ret = os.popen(s).read().strip()
return ret
if 'PKG_CONFIG_PATH' not in os.environ:
# If Gtk+ is installed, pkg-config is required to be installed
os.environ['PKG_CONFIG_PATH'] = 'C:\\GTK\\lib\\pkgconfig'
# popen broken on my win32 plaform so I can't use pkgconfig
ext.library_dirs.extend(
['C:/GTK/bin', 'C:/GTK/lib'])
ext.include_dirs.extend(
['win32_static/include/pygtk-2.0',
'C:/GTK/include',
'C:/GTK/include/gobject',
'C:/GTK/include/gext',
'C:/GTK/include/glib',
'C:/GTK/include/pango',
'C:/GTK/include/atk',
'C:/GTK/include/X11',
'C:/GTK/include/cairo',
'C:/GTK/include/gdk',
'C:/GTK/include/gdk-pixbuf',
'C:/GTK/include/gtk',
])
pygtkIncludes = getoutput(
'pkg-config --cflags-only-I pygtk-2.0').split()
gtkIncludes = getoutput(
'pkg-config --cflags-only-I gtk+-2.0').split()
includes = pygtkIncludes + gtkIncludes
ext.include_dirs.extend([include[2:] for include in includes])
pygtkLinker = getoutput('pkg-config --libs pygtk-2.0').split()
gtkLinker = getoutput('pkg-config --libs gtk+-2.0').split()
linkerFlags = pygtkLinker + gtkLinker
ext.libraries.extend(
[flag[2:] for flag in linkerFlags if flag.startswith('-l')])
ext.library_dirs.extend(
[flag[2:] for flag in linkerFlags if flag.startswith('-L')])
ext.extra_link_args.extend(
[flag for flag in linkerFlags if not
(flag.startswith('-l') or flag.startswith('-L'))])
# visual studio doesn't need the math library
if (sys.platform == 'win32' and
win32_compiler == 'msvc' and
'm' in ext.libraries):
ext.libraries.remove('m')
elif sys.platform != 'win32':
pkg_config.setup_extension(ext, 'pygtk-2.0')
pkg_config.setup_extension(ext, 'gtk+-2.0')
class BackendGtkAgg(BackendGtk):
name = "gtkagg"
def check(self):
try:
return super(BackendGtkAgg, self).check()
except:
raise
else:
BackendAgg.force = True
def get_package_data(self):
return {'matplotlib': ['mpl-data/*.glade']}
def get_extension(self):
sources = [
'src/agg_py_transforms.cpp',
'src/_gtkagg.cpp',
'src/mplutils.cpp'
]
ext = make_extension('matplotlib.backends._gtkagg', sources)
self.add_flags(ext)
LibAgg().add_flags(ext)
CXX().add_flags(ext)
Numpy().add_flags(ext)
return ext
def backend_gtk3agg_internal_check(x):
try:
import gi
except ImportError:
return (False, "Requires pygobject to be installed.")
try:
gi.require_version("Gtk", "3.0")
except ValueError:
return (False, "Requires gtk3 development files to be installed.")
except AttributeError:
return (False, "pygobject version too old.")
try:
from gi.repository import Gtk, Gdk, GObject
except (ImportError, RuntimeError):
return (False, "Requires pygobject to be installed.")
return (True, "version %s.%s.%s" % (
Gtk.get_major_version(),
Gtk.get_micro_version(),
Gtk.get_minor_version()))
class BackendGtk3Agg(OptionalBackendPackage):
name = "gtk3agg"
def check_requirements(self):
if 'TRAVIS' in os.environ:
raise CheckFailed("Can't build with Travis")
if PY3:
raise CheckFailed("gtk3agg backend does not work on Python 3")
# This check needs to be performed out-of-process, because
# importing gi and then importing regular old pygtk afterward
# segfaults the interpreter.
try:
p = multiprocessing.Pool()
except:
return "unknown (can not use multiprocessing to determine)"
try:
success, msg = p.map(backend_gtk3agg_internal_check, [0])[0]
except:
success = False
msg = "Could not determine"
finally:
p.close()
p.join()
if success:
BackendAgg.force = True
return msg
else:
raise CheckFailed(msg)
def get_package_data(self):
return {'matplotlib': ['mpl-data/*.glade']}
def backend_gtk3cairo_internal_check(x):
try:
import cairocffi
except ImportError:
try:
import cairo
except ImportError:
return (False, "Requires cairocffi or pycairo to be installed.")
try:
import gi
except ImportError:
return (False, "Requires pygobject to be installed.")
try:
gi.require_version("Gtk", "3.0")
except ValueError:
return (False, "Requires gtk3 development files to be installed.")
except AttributeError:
return (False, "pygobject version too old.")
try:
from gi.repository import Gtk, Gdk, GObject
except (RuntimeError, ImportError):
return (False, "Requires pygobject to be installed.")
return (True, "version %s.%s.%s" % (
Gtk.get_major_version(),
Gtk.get_micro_version(),
Gtk.get_minor_version()))
class BackendGtk3Cairo(OptionalBackendPackage):
name = "gtk3cairo"
def check_requirements(self):
if 'TRAVIS' in os.environ:
raise CheckFailed("Can't build with Travis")
# This check needs to be performed out-of-process, because
# importing gi and then importing regular old pygtk afterward
# segfaults the interpreter.
try:
p = multiprocessing.Pool()
except:
return "unknown (can not use multiprocessing to determine)"
success, msg = p.map(backend_gtk3cairo_internal_check, [0])[0]
p.close()
p.join()
if success:
BackendAgg.force = True
return msg
else:
raise CheckFailed(msg)
def get_package_data(self):
return {'matplotlib': ['mpl-data/*.glade']}
class BackendWxAgg(OptionalBackendPackage):
name = "wxagg"
def check_requirements(self):
try:
import wxversion
except ImportError:
raise CheckFailed("requires wxPython")
try:
_wx_ensure_failed = wxversion.AlreadyImportedError
except AttributeError:
_wx_ensure_failed = wxversion.VersionError
try:
wxversion.ensureMinimal('2.8')
except _wx_ensure_failed:
pass
try:
import wx
backend_version = wx.VERSION_STRING
except ImportError:
raise CheckFailed("requires wxPython")
# Extra version check in case wxversion lacks AlreadyImportedError;
# then VersionError might have been raised and ignored when
# there really *is* a problem with the version.
major, minor = [int(n) for n in backend_version.split('.')[:2]]
if major < 2 or (major < 3 and minor < 8):
raise CheckFailed(
"Requires wxPython 2.8, found %s" % backend_version)
BackendAgg.force = True
return "version %s" % backend_version
class BackendMacOSX(OptionalBackendPackage):
name = 'macosx'
def check_requirements(self):
if sys.platform != 'darwin':
raise CheckFailed("Mac OS-X only")
return 'darwin'
def get_extension(self):
sources = [
'src/_macosx.m',
'src/agg_py_transforms.cpp',
'src/path_cleanup.cpp'
]
ext = make_extension('matplotlib.backends._macosx', sources)
Numpy().add_flags(ext)
LibAgg().add_flags(ext)
CXX().add_flags(ext)
ext.extra_link_args.extend(['-framework', 'Cocoa'])
return ext
class Windowing(OptionalBackendPackage):
"""
Builds the windowing extension.
"""
name = "windowing"
def check_requirements(self):
if sys.platform != 'win32':
raise CheckFailed("Microsoft Windows only")
config = self.get_config()
if config is False:
raise CheckFailed("skipping due to configuration")
return "installing"
def get_extension(self):
sources = [
"src/_windowing.cpp"
]
ext = make_extension('matplotlib._windowing', sources)
ext.include_dirs.extend(['C:/include'])
ext.libraries.extend(['user32'])
ext.library_dirs.extend(['C:/lib'])
ext.extra_link_args.append("-mwindows")
return ext
class BackendQtBase(OptionalBackendPackage):
def convert_qt_version(self, version):
version = '%x' % version
temp = []
while len(version) > 0:
version, chunk = version[:-2], version[-2:]
temp.insert(0, str(int(chunk, 16)))
return '.'.join(temp)
def check_requirements(self):
'''
If PyQt4/PyQt5 is already imported, importing PyQt5/PyQt4 will fail
so we need to test in a subprocess (as for Gtk3).
'''
try:
p = multiprocessing.Pool()
except:
# Can't do multiprocessing, fall back to normal approach ( this will fail if importing both PyQt4 and PyQt5 )
try:
# Try in-process
msg = self.callback(self)
except RuntimeError:
raise CheckFailed("Could not import: are PyQt4 & PyQt5 both installed?")
except:
# Raise any other exceptions
raise
else:
# Multiprocessing OK
try:
msg = p.map(self.callback, [self])[0]
except:
# If we hit an error on multiprocessing raise it
raise
finally:
# Tidy up multiprocessing
p.close()
p.join()
return msg
def backend_qt4_internal_check(self):
try:
from PyQt4 import QtCore
except ImportError:
raise CheckFailed("PyQt4 not found")
try:
qt_version = QtCore.QT_VERSION
pyqt_version_str = QtCore.QT_VERSION_STR
except AttributeError:
raise CheckFailed('PyQt4 not correctly imported')
else:
BackendAgg.force = True
return ("Qt: %s, PyQt: %s" % (self.convert_qt_version(qt_version), pyqt_version_str))
class BackendQt4(BackendQtBase):
name = "qt4agg"
def __init__(self, *args, **kwargs):
BackendQtBase.__init__(self, *args, **kwargs)
self.callback = backend_qt4_internal_check
def backend_qt5_internal_check(self):
try:
from PyQt5 import QtCore
except ImportError:
raise CheckFailed("PyQt5 not found")
try:
qt_version = QtCore.QT_VERSION
pyqt_version_str = QtCore.QT_VERSION_STR
except AttributeError:
raise CheckFailed('PyQt5 not correctly imported')
else:
BackendAgg.force = True
return ("Qt: %s, PyQt: %s" % (self.convert_qt_version(qt_version), pyqt_version_str))
class BackendQt5(BackendQtBase):
name = "qt5agg"
def __init__(self, *args, **kwargs):
BackendQtBase.__init__(self, *args, **kwargs)
self.callback = backend_qt5_internal_check
def backend_pyside_internal_check(self):
try:
from PySide import __version__
from PySide import QtCore
except ImportError:
raise CheckFailed("PySide not found")
else:
BackendAgg.force = True
return ("Qt: %s, PySide: %s" %
(QtCore.__version__, __version__))
class BackendPySide(BackendQtBase):
name = "pyside"
def __init__(self, *args, **kwargs):
BackendQtBase.__init__(self, *args, **kwargs)
self.callback = backend_pyside_internal_check
class BackendCairo(OptionalBackendPackage):
name = "cairo"
def check_requirements(self):
try:
import cairocffi
except ImportError:
try:
import cairo
except ImportError:
raise CheckFailed("cairocffi or pycairo not found")
else:
return "pycairo version %s" % cairo.version
else:
return "cairocffi version %s" % cairocffi.version
class DviPng(SetupPackage):
name = "dvipng"
optional = True
def check(self):
try:
output = check_output('dvipng -version', shell=True,
stderr=subprocess.STDOUT)
return "version %s" % output.splitlines()[1].decode().split()[-1]
except (IndexError, ValueError, subprocess.CalledProcessError):
raise CheckFailed()
class Ghostscript(SetupPackage):
name = "ghostscript"
optional = True
def check(self):
try:
if sys.platform == 'win32':
command = 'gswin32c --version'
try:
output = check_output(command, shell=True,
stderr=subprocess.STDOUT)
except subprocess.CalledProcessError:
command = 'gswin64c --version'
output = check_output(command, shell=True,
stderr=subprocess.STDOUT)
else:
command = 'gs --version'
output = check_output(command, shell=True,
stderr=subprocess.STDOUT)
return "version %s" % output.decode()[:-1]
except (IndexError, ValueError, subprocess.CalledProcessError):
raise CheckFailed()
class LaTeX(SetupPackage):
name = "latex"
optional = True
def check(self):
try:
output = check_output('latex -version', shell=True,
stderr=subprocess.STDOUT)
line = output.splitlines()[0].decode()
pattern = '(3\.1\d+)|(MiKTeX \d+.\d+)'
match = re.search(pattern, line)
return "version %s" % match.group(0)
except (IndexError, ValueError, AttributeError, subprocess.CalledProcessError):
raise CheckFailed()
class PdfToPs(SetupPackage):
name = "pdftops"
optional = True
def check(self):
try:
output = check_output('pdftops -v', shell=True,
stderr=subprocess.STDOUT)
for line in output.splitlines():
line = line.decode()
if 'version' in line:
return "version %s" % line.split()[2]
except (IndexError, ValueError, subprocess.CalledProcessError):
pass
raise CheckFailed()
|
[
"io.BytesIO",
"sys.platform.startswith",
"gi.repository.Gtk.get_micro_version",
"re.search",
"os.path.exists",
"os.listdir",
"sys.getfilesystemencoding",
"pyparsing.Forward",
"CXX",
"subprocess.Popen",
"subprocess.CalledProcessError",
"os.path.split",
"os.path.isdir",
"os.popen",
"numpy.get_include",
"warnings.warn",
"io.StringIO",
"pyparsing.__version__.split",
"pyparsing.Literal",
"distutils.sysconfig.get_config_var",
"subprocess.getstatusoutput",
"subprocess.check_output",
"glob.glob",
"Tkinter.__version__.split",
"ConfigParser.SafeConfigParser",
"Tkinter.Tk",
"imp.reload",
"gi.require_version",
"wxversion.ensureMinimal",
"gi.repository.Gtk.get_minor_version",
"distutils.core.Extension",
"gi.repository.Gtk.get_major_version",
"os.getenv",
"Tkinter.Tcl",
"os.environ.get",
"os.path.join",
"multiprocessing.Pool",
"distutils.version.LooseVersion"
] |
[((1564, 1606), 'os.environ.get', 'os.environ.get', (['"""MPLSETUPCFG"""', '"""setup.cfg"""'], {}), "('MPLSETUPCFG', 'setup.cfg')\n", (1578, 1606), False, 'import os\n'), ((1610, 1635), 'os.path.exists', 'os.path.exists', (['setup_cfg'], {}), '(setup_cfg)\n', (1624, 1635), False, 'import os\n'), ((1650, 1681), 'ConfigParser.SafeConfigParser', 'configparser.SafeConfigParser', ([], {}), '()\n', (1679, 1681), True, 'import ConfigParser as configparser\n'), ((4201, 4233), 'distutils.version.LooseVersion', 'version.LooseVersion', (['minversion'], {}), '(minversion)\n', (4221, 4233), False, 'from distutils import version\n'), ((4254, 4281), 'distutils.version.LooseVersion', 'version.LooseVersion', (['found'], {}), '(found)\n', (4274, 4281), False, 'from distutils import version\n'), ((5971, 5999), 'os.path.join', 'os.path.join', (['dir', '"""include"""'], {}), "(dir, 'include')\n", (5983, 5999), False, 'import os\n'), ((6011, 6038), 'os.path.exists', 'os.path.exists', (['include_dir'], {}), '(include_dir)\n', (6025, 6038), False, 'import os\n'), ((6817, 6851), 'distutils.sysconfig.get_config_var', 'sysconfig.get_config_var', (['"""LIBDIR"""'], {}), "('LIBDIR')\n", (6841, 6851), False, 'from distutils import sysconfig\n'), ((6932, 6973), 'os.path.join', 'os.path.join', (['pkgconfig_path', '"""pkgconfig"""'], {}), "(pkgconfig_path, 'pkgconfig')\n", (6944, 6973), False, 'import os\n'), ((9155, 9210), 'subprocess.getstatusoutput', 'getstatusoutput', (["('pkg-config %s --modversion' % package)"], {}), "('pkg-config %s --modversion' % package)\n", (9170, 9210), False, 'from subprocess import getstatusoutput\n'), ((22115, 22136), 'distutils.core.Extension', 'Extension', (['"""test"""', '[]'], {}), "('test', [])\n", (22124, 22136), False, 'from distutils.core import Extension\n'), ((27288, 27332), 'subprocess.getstatusoutput', 'getstatusoutput', (['"""freetype-config --version"""'], {}), "('freetype-config --version')\n", (27303, 27332), False, 'from subprocess import getstatusoutput\n'), ((52850, 52882), 'gi.require_version', 'gi.require_version', (['"""Gtk"""', '"""3.0"""'], {}), "('Gtk', '3.0')\n", (52868, 52882), False, 'import gi\n'), ((54790, 54822), 'gi.require_version', 'gi.require_version', (['"""Gtk"""', '"""3.0"""'], {}), "('Gtk', '3.0')\n", (54808, 54822), False, 'import gi\n'), ((702, 764), 'subprocess.Popen', 'subprocess.Popen', (['*popenargs'], {'stdout': 'subprocess.PIPE'}), '(*popenargs, stdout=subprocess.PIPE, **kwargs)\n', (718, 764), False, 'import subprocess\n'), ((2990, 3017), 'os.path.join', 'os.path.join', (['dir', 'filename'], {}), '(dir, filename)\n', (3002, 3017), False, 'import os\n'), ((6148, 6170), 'os.path.join', 'os.path.join', (['dir', 'lib'], {}), '(dir, lib)\n', (6160, 6170), False, 'import os\n'), ((6186, 6209), 'os.path.exists', 'os.path.exists', (['lib_dir'], {}), '(lib_dir)\n', (6200, 6209), False, 'import os\n'), ((6673, 6709), 'subprocess.getstatusoutput', 'getstatusoutput', (['"""pkg-config --help"""'], {}), "('pkg-config --help')\n", (6688, 6709), False, 'from subprocess import getstatusoutput\n'), ((6989, 7018), 'os.path.isdir', 'os.path.isdir', (['pkgconfig_path'], {}), '(pkgconfig_path)\n', (7002, 7018), False, 'import os\n'), ((21882, 21899), 'imp.reload', 'imp.reload', (['numpy'], {}), '(numpy)\n', (21892, 21899), False, 'import imp\n'), ((22169, 22188), 'numpy.get_include', 'numpy.get_include', ([], {}), '()\n', (22186, 22188), False, 'import numpy\n'), ((22310, 22428), 'warnings.warn', 'warnings.warn', (['"""The C headers for numpy could not be found. You may need to install the development package"""'], {}), "(\n 'The C headers for numpy could not be found. You may need to install the development package'\n )\n", (22323, 22428), False, 'import warnings\n'), ((22472, 22491), 'numpy.get_include', 'numpy.get_include', ([], {}), '()\n', (22489, 22491), False, 'import numpy\n'), ((23916, 23929), 'io.StringIO', 'io.StringIO', ([], {}), '()\n', (23927, 23929), False, 'import io\n'), ((23969, 23981), 'io.BytesIO', 'io.BytesIO', ([], {}), '()\n', (23979, 23981), False, 'import io\n'), ((31990, 32032), 'os.path.join', 'os.path.join', (['"""lib/matplotlib/delaunay"""', 's'], {}), "('lib/matplotlib/delaunay', s)\n", (32002, 32032), False, 'import os\n'), ((35066, 35085), 'pyparsing.Forward', 'pyparsing.Forward', ([], {}), '()\n', (35083, 35085), False, 'import pyparsing\n'), ((35104, 35126), 'pyparsing.Literal', 'pyparsing.Literal', (['"""a"""'], {}), "('a')\n", (35121, 35126), False, 'import pyparsing\n'), ((38884, 38896), 'Tkinter.Tk', 'Tkinter.Tk', ([], {}), '()\n', (38894, 38896), False, 'import Tkinter\n'), ((40932, 40966), 'os.path.join', 'os.path.join', (['ptcl', '"""tclConfig.sh"""'], {}), "(ptcl, 'tclConfig.sh')\n", (40944, 40966), False, 'import os\n'), ((40991, 41023), 'os.path.join', 'os.path.join', (['ptk', '"""tkConfig.sh"""'], {}), "(ptk, 'tkConfig.sh')\n", (41003, 41023), False, 'import os\n'), ((41273, 41395), 'subprocess.Popen', 'subprocess.Popen', (["('. %s ; eval echo ${%s}' % (file, varname))"], {'shell': '(True)', 'executable': '"""/bin/sh"""', 'stdout': 'subprocess.PIPE'}), "('. %s ; eval echo ${%s}' % (file, varname), shell=True,\n executable='/bin/sh', stdout=subprocess.PIPE)\n", (41289, 41395), False, 'import subprocess\n'), ((42509, 42541), 'os.path.join', 'os.path.join', (['tcl_lib_dir', '"""../"""'], {}), "(tcl_lib_dir, '../')\n", (42521, 42541), False, 'import os\n'), ((42577, 42608), 'os.path.join', 'os.path.join', (['tk_lib_dir', '"""../"""'], {}), "(tk_lib_dir, '../')\n", (42589, 42608), False, 'import os\n'), ((42659, 42714), 'os.path.join', 'os.path.join', (['tcl_lib_dir', "('../../include/tcl' + tk_ver)"], {}), "(tcl_lib_dir, '../../include/tcl' + tk_ver)\n", (42671, 42714), False, 'import os\n'), ((42756, 42779), 'os.path.exists', 'os.path.exists', (['tcl_inc'], {}), '(tcl_inc)\n', (42770, 42779), False, 'import os\n'), ((42945, 42998), 'os.path.join', 'os.path.join', (['tk_lib_dir', "('../../include/tk' + tk_ver)"], {}), "(tk_lib_dir, '../../include/tk' + tk_ver)\n", (42957, 42998), False, 'import os\n'), ((43040, 43062), 'os.path.exists', 'os.path.exists', (['tk_inc'], {}), '(tk_inc)\n', (43054, 43062), False, 'import os\n'), ((43285, 43308), 'os.path.exists', 'os.path.exists', (['tcl_inc'], {}), '(tcl_inc)\n', (43299, 43308), False, 'import os\n'), ((48846, 48874), 'os.path.join', 'os.path.join', (['"""gtk"""', '"""gtk.h"""'], {}), "('gtk', 'gtk.h')\n", (48858, 48874), False, 'import os\n'), ((48982, 49014), 'os.path.join', 'os.path.join', (['"""pygtk"""', '"""pygtk.h"""'], {}), "('pygtk', 'pygtk.h')\n", (48994, 49014), False, 'import os\n'), ((53859, 53881), 'multiprocessing.Pool', 'multiprocessing.Pool', ([], {}), '()\n', (53879, 53881), False, 'import multiprocessing\n'), ((55711, 55733), 'multiprocessing.Pool', 'multiprocessing.Pool', ([], {}), '()\n', (55731, 55733), False, 'import multiprocessing\n'), ((56552, 56582), 'wxversion.ensureMinimal', 'wxversion.ensureMinimal', (['"""2.8"""'], {}), "('2.8')\n", (56575, 56582), False, 'import wxversion\n'), ((59211, 59233), 'multiprocessing.Pool', 'multiprocessing.Pool', ([], {}), '()\n', (59231, 59233), False, 'import multiprocessing\n'), ((62512, 62581), 'subprocess.check_output', 'check_output', (['"""dvipng -version"""'], {'shell': '(True)', 'stderr': 'subprocess.STDOUT'}), "('dvipng -version', shell=True, stderr=subprocess.STDOUT)\n", (62524, 62581), False, 'from subprocess import check_output\n'), ((63853, 63921), 'subprocess.check_output', 'check_output', (['"""latex -version"""'], {'shell': '(True)', 'stderr': 'subprocess.STDOUT'}), "('latex -version', shell=True, stderr=subprocess.STDOUT)\n", (63865, 63921), False, 'from subprocess import check_output\n'), ((64078, 64102), 're.search', 're.search', (['pattern', 'line'], {}), '(pattern, line)\n', (64087, 64102), False, 'import re\n'), ((64400, 64464), 'subprocess.check_output', 'check_output', (['"""pdftops -v"""'], {'shell': '(True)', 'stderr': 'subprocess.STDOUT'}), "('pdftops -v', shell=True, stderr=subprocess.STDOUT)\n", (64412, 64464), False, 'from subprocess import check_output\n'), ((1002, 1045), 'subprocess.CalledProcessError', 'subprocess.CalledProcessError', (['retcode', 'cmd'], {}), '(retcode, cmd)\n', (1031, 1045), False, 'import subprocess\n'), ((3864, 3887), 'os.getenv', 'os.getenv', (['"""MPLIB_BASE"""'], {}), "('MPLIB_BASE')\n", (3873, 3887), False, 'import os\n'), ((7863, 7922), 'subprocess.check_output', 'check_output', (['command'], {'shell': '(True)', 'stderr': 'subprocess.STDOUT'}), '(command, shell=True, stderr=subprocess.STDOUT)\n', (7875, 7922), False, 'from subprocess import check_output\n'), ((19708, 19758), 'os.listdir', 'os.listdir', (['"""lib/matplotlib/tests/baseline_images"""'], {}), "('lib/matplotlib/tests/baseline_images')\n", (19718, 19758), False, 'import os\n'), ((22257, 22295), 'os.path.join', 'os.path.join', (['"""numpy"""', '"""arrayobject.h"""'], {}), "('numpy', 'arrayobject.h')\n", (22269, 22295), False, 'import os\n'), ((24920, 25025), 'os.path.join', 'os.path.join', (['sys.prefix', '"""share"""', "('python%d.%d' % (sys.version_info[0], sys.version_info[1]))", '"""CXX"""'], {}), "(sys.prefix, 'share', 'python%d.%d' % (sys.version_info[0], sys\n .version_info[1]), 'CXX')\n", (24932, 25025), False, 'import os\n'), ((25162, 25189), 'os.path.exists', 'os.path.exists', (['support_dir'], {}), '(support_dir)\n', (25176, 25189), False, 'import os\n'), ((25686, 25715), 'glob.glob', 'glob.glob', (['"""extern/CXX/*.cxx"""'], {}), "('extern/CXX/*.cxx')\n", (25695, 25715), False, 'import glob\n'), ((25748, 25775), 'glob.glob', 'glob.glob', (['"""extern/CXX/*.c"""'], {}), "('extern/CXX/*.c')\n", (25757, 25775), False, 'import glob\n'), ((28387, 28392), 'CXX', 'CXX', ([], {}), '()\n', (28390, 28392), False, 'import CXX\n'), ((29062, 29067), 'CXX', 'CXX', ([], {}), '()\n', (29065, 29067), False, 'import CXX\n'), ((30244, 30273), 'glob.glob', 'glob.glob', (['"""extern/qhull/*.c"""'], {}), "('extern/qhull/*.c')\n", (30253, 30273), False, 'import glob\n'), ((30644, 30649), 'CXX', 'CXX', ([], {}), '()\n', (30647, 30649), False, 'import CXX\n'), ((31071, 31076), 'CXX', 'CXX', ([], {}), '()\n', (31074, 31076), False, 'import CXX\n'), ((31402, 31407), 'CXX', 'CXX', ([], {}), '()\n', (31405, 31407), False, 'import CXX\n'), ((32781, 32786), 'CXX', 'CXX', ([], {}), '()\n', (32784, 32786), False, 'import CXX\n'), ((36646, 36651), 'CXX', 'CXX', ([], {}), '()\n', (36649, 36651), False, 'import CXX\n'), ((37400, 37427), 'Tkinter.__version__.split', 'Tkinter.__version__.split', ([], {}), '()\n', (37425, 37427), False, 'import Tkinter\n'), ((37955, 37960), 'CXX', 'CXX', ([], {}), '()\n', (37958, 37960), False, 'import CXX\n'), ((40539, 40570), 'os.path.join', 'os.path.join', (['tcl_lib_dir', '""".."""'], {}), "(tcl_lib_dir, '..')\n", (40551, 40570), False, 'import os\n'), ((40735, 40765), 'os.path.join', 'os.path.join', (['tk_lib_dir', '""".."""'], {}), "(tk_lib_dir, '..')\n", (40747, 40765), False, 'import os\n'), ((41040, 41066), 'os.path.exists', 'os.path.exists', (['tcl_config'], {}), '(tcl_config)\n', (41054, 41066), False, 'import os\n'), ((41071, 41096), 'os.path.exists', 'os.path.exists', (['tk_config'], {}), '(tk_config)\n', (41085, 41096), False, 'import os\n'), ((41137, 41163), 'os.path.exists', 'os.path.exists', (['tcl_config'], {}), '(tcl_config)\n', (41151, 41163), False, 'import os\n'), ((41168, 41193), 'os.path.exists', 'os.path.exists', (['tk_config'], {}), '(tk_config)\n', (41182, 41193), False, 'import os\n'), ((42147, 42179), 'os.path.join', 'os.path.join', (['tk_inc_dir', '"""tk.h"""'], {}), "(tk_inc_dir, 'tk.h')\n", (42159, 42179), False, 'import os\n'), ((42388, 42415), 'os.path.exists', 'os.path.exists', (['tcl_lib_dir'], {}), '(tcl_lib_dir)\n', (42402, 42415), False, 'import os\n'), ((42420, 42446), 'os.path.exists', 'os.path.exists', (['tk_lib_dir'], {}), '(tk_lib_dir)\n', (42434, 42446), False, 'import os\n'), ((42837, 42879), 'os.path.join', 'os.path.join', (['tcl_lib_dir', '"""../../include"""'], {}), "(tcl_lib_dir, '../../include')\n", (42849, 42879), False, 'import os\n'), ((43102, 43143), 'os.path.join', 'os.path.join', (['tk_lib_dir', '"""../../include"""'], {}), "(tk_lib_dir, '../../include')\n", (43114, 43143), False, 'import os\n'), ((43209, 43237), 'os.path.join', 'os.path.join', (['tk_inc', '"""tk.h"""'], {}), "(tk_inc, 'tk.h')\n", (43221, 43237), False, 'import os\n'), ((43387, 43419), 'sys.platform.startswith', 'sys.platform.startswith', (['"""linux"""'], {}), "('linux')\n", (43410, 43419), False, 'import sys\n'), ((43440, 43476), 'os.path.exists', 'os.path.exists', (['"""/usr/include/tcl.h"""'], {}), "('/usr/include/tcl.h')\n", (43454, 43476), False, 'import os\n'), ((43497, 43532), 'os.path.exists', 'os.path.exists', (['"""/usr/include/tk.h"""'], {}), "('/usr/include/tk.h')\n", (43511, 43532), False, 'import os\n'), ((43647, 43675), 'os.path.join', 'os.path.join', (['tk_inc', '"""tk.h"""'], {}), "(tk_inc, 'tk.h')\n", (43659, 43675), False, 'import os\n'), ((52607, 52612), 'CXX', 'CXX', ([], {}), '()\n', (52610, 52612), False, 'import CXX\n'), ((53275, 53298), 'gi.repository.Gtk.get_major_version', 'Gtk.get_major_version', ([], {}), '()\n', (53296, 53298), False, 'from gi.repository import Gtk, Gdk, GObject\n'), ((53308, 53331), 'gi.repository.Gtk.get_micro_version', 'Gtk.get_micro_version', ([], {}), '()\n', (53329, 53331), False, 'from gi.repository import Gtk, Gdk, GObject\n'), ((53341, 53364), 'gi.repository.Gtk.get_minor_version', 'Gtk.get_minor_version', ([], {}), '()\n', (53362, 53364), False, 'from gi.repository import Gtk, Gdk, GObject\n'), ((55215, 55238), 'gi.repository.Gtk.get_major_version', 'Gtk.get_major_version', ([], {}), '()\n', (55236, 55238), False, 'from gi.repository import Gtk, Gdk, GObject\n'), ((55248, 55271), 'gi.repository.Gtk.get_micro_version', 'Gtk.get_micro_version', ([], {}), '()\n', (55269, 55271), False, 'from gi.repository import Gtk, Gdk, GObject\n'), ((55281, 55304), 'gi.repository.Gtk.get_minor_version', 'Gtk.get_minor_version', ([], {}), '()\n', (55302, 55304), False, 'from gi.repository import Gtk, Gdk, GObject\n'), ((57822, 57827), 'CXX', 'CXX', ([], {}), '()\n', (57825, 57827), False, 'import CXX\n'), ((63472, 63531), 'subprocess.check_output', 'check_output', (['command'], {'shell': '(True)', 'stderr': 'subprocess.STDOUT'}), '(command, shell=True, stderr=subprocess.STDOUT)\n', (63484, 63531), False, 'from subprocess import check_output\n'), ((3264, 3289), 'os.getenv', 'os.getenv', (['"""INCLUDE"""', '"""."""'], {}), "('INCLUDE', '.')\n", (3273, 3289), False, 'import os\n'), ((8089, 8116), 'sys.getfilesystemencoding', 'sys.getfilesystemencoding', ([], {}), '()\n', (8114, 8116), False, 'import sys\n'), ((8534, 8561), 'os.path.join', 'os.path.join', (['base', 'include'], {}), '(base, include)\n', (8546, 8561), False, 'import os\n'), ((8585, 8604), 'os.path.exists', 'os.path.exists', (['dir'], {}), '(dir)\n', (8599, 8604), False, 'import os\n'), ((8734, 8757), 'os.path.join', 'os.path.join', (['base', 'lib'], {}), '(base, lib)\n', (8746, 8757), False, 'import os\n'), ((8781, 8800), 'os.path.exists', 'os.path.exists', (['dir'], {}), '(dir)\n', (8795, 8800), False, 'import os\n'), ((25361, 25389), 'os.path.join', 'os.path.join', (['support_dir', 'x'], {}), '(support_dir, x)\n', (25373, 25389), False, 'import os\n'), ((27049, 27090), 'os.path.join', 'os.path.join', (['"""extern"""', '"""agg24"""', '"""src"""', 'x'], {}), "('extern', 'agg24', 'src', x)\n", (27061, 27090), False, 'import os\n'), ((29640, 29675), 'os.path.join', 'os.path.join', (['x', '"""include"""', '"""qhull"""'], {}), "(x, 'include', 'qhull')\n", (29652, 29675), False, 'import os\n'), ((35577, 35609), 'pyparsing.__version__.split', 'pyparsing.__version__.split', (['"""."""'], {}), "('.')\n", (35604, 35609), False, 'import pyparsing\n'), ((39145, 39158), 'Tkinter.Tcl', 'Tkinter.Tcl', ([], {}), '()\n', (39156, 39158), False, 'import Tkinter\n'), ((39519, 39545), 'os.path.split', 'os.path.split', (['tcl_lib_dir'], {}), '(tcl_lib_dir)\n', (39532, 39545), False, 'import os\n'), ((39645, 39669), 'os.path.join', 'os.path.join', (['head', 'tail'], {}), '(head, tail)\n', (39657, 39669), False, 'import os\n'), ((44514, 44546), 'os.path.join', 'os.path.join', (['sys.prefix', '"""dlls"""'], {}), "(sys.prefix, 'dlls')\n", (44526, 44546), False, 'import os\n'), ((63050, 63109), 'subprocess.check_output', 'check_output', (['command'], {'shell': '(True)', 'stderr': 'subprocess.STDOUT'}), '(command, shell=True, stderr=subprocess.STDOUT)\n', (63062, 63109), False, 'from subprocess import check_output\n'), ((39693, 39719), 'os.path.exists', 'os.path.exists', (['tk_lib_dir'], {}), '(tk_lib_dir)\n', (39707, 39719), False, 'import os\n'), ((44851, 44868), 'os.getenv', 'os.getenv', (['"""HOME"""'], {}), "('HOME')\n", (44860, 44868), False, 'import os\n'), ((45934, 45963), 'os.path.join', 'join', (['F', "(fw + '.framework')", 'H'], {}), "(F, fw + '.framework', H)\n", (45938, 45963), False, 'from os.path import join, exists\n'), ((63286, 63345), 'subprocess.check_output', 'check_output', (['command'], {'shell': '(True)', 'stderr': 'subprocess.STDOUT'}), '(command, shell=True, stderr=subprocess.STDOUT)\n', (63298, 63345), False, 'from subprocess import check_output\n'), ((45317, 45343), 'os.path.join', 'join', (['F', "(fw + '.framework')"], {}), "(F, fw + '.framework')\n", (45321, 45343), False, 'from os.path import join, exists\n'), ((49657, 49668), 'os.popen', 'os.popen', (['s'], {}), '(s)\n', (49665, 49668), False, 'import os\n')]
|
# Copyright 2021 Sony Group Corporation.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
'''
MUSDB18 data-iterator code for MSS.
'''
import random
import numpy as np
import musdb
from nnabla.utils.data_source import DataSource
class Compose():
"""Composes several augmentation transforms.
Args:
augmentations: list of augmentations to compose.
"""
def __init__(self, transforms):
self.transforms = transforms
def __call__(self, audio):
for t in self.transforms:
audio = t(audio)
return audio
def _augment_gain(audio, low=0.75, high=1.25):
"""Applies a random gain between `low` and `high`"""
g = random.uniform(low, high)
return audio * g
def _augment_channelswap(audio):
"""Swap channels of stereo signals with a probability of p=0.5"""
if audio.shape[0] == 2 and random.random() < 0.5:
return np.flip(audio, 0)
else:
return audio
def load_datasources(parser, args):
"""Loads the specified dataset from commandline arguments
Returns:
train_dataset, validation_dataset
"""
parser.add_argument('--is-wav', action='store_true', default=True,
help='loads wav instead of STEMS')
parser.add_argument('--samples-per-track', type=int, default=64)
parser.add_argument(
'--source-augmentations', type=str, nargs='+',
default=['gain', 'channelswap']
)
args = parser.parse_args()
source_augmentations = Compose(
[globals()['_augment_' + aug] for aug in args.source_augmentations]
)
train_dataset = MUSDBDataSource(
source_augmentations=source_augmentations, random_track_mix=True, args=args)
return train_dataset, args
class MUSDBDataSource(DataSource):
def __init__(
self,
args,
download=False,
samples_per_track=64,
source_augmentations=lambda audio: audio,
random_track_mix=False,
dtype=np.float32,
seed=42,
rng=None
):
"""
MUSDB18 nnabla.utils.data_source that samples from the MUSDB tracks
using track and excerpts with replacement.
Parameters
----------
args : additional arguments used to add further control for
the musdb dataset initialization function.
download : boolean
automatically download 7s preview version of MUS
samples_per_track : int
sets the number of samples, yielded from each track per epoch.
Defaults to 64
source_augmentations : list[callables]
provide list of augmentation function that take a multi-channel
audio file of shape (src, samples) as input and output. Defaults to
no-augmentations (input = output)
random_track_mix : boolean
randomly mixes sources from different tracks to assemble a
custom mix. This augmenation is only applied for the train subset.
seed : int
control randomness of dataset iterations
dtype : numeric type
data type of torch output tuple x and y
"""
super(MUSDBDataSource, self).__init__(shuffle=True)
if rng is None:
rng = np.random.RandomState(seed)
self.rng = rng
random.seed(seed)
self.args = args
self.download = args.root is None
self.samples_per_track = samples_per_track
self.source_augmentations = source_augmentations
self.random_track_mix = random_track_mix
self.mus = musdb.DB(
root=args.root,
is_wav=args.is_wav,
split=None,
subsets='train',
download=download
)
print(f"Finished loading dataset with {len(self.mus.tracks)} tracks.")
self.sample_rate = 44100 # musdb has fixed sample rate
self.dtype = dtype
self._size = len(self.mus.tracks) * self.samples_per_track
self._variables = ('mixture', 'target')
self.reset()
def _get_data(self, position):
index = self._indexes[position]
audio_sources = []
target_ind = None
# select track
track = self.mus.tracks[index // self.samples_per_track]
# at training time we assemble a custom mix
if self.args.seq_dur:
for k, source in enumerate(self.mus.setup['sources']):
# memorize index of target source
if source == self.args.target:
target_ind = k
# select a random track
if self.random_track_mix:
track = random.choice(self.mus.tracks)
# set the excerpt duration
track.chunk_duration = self.args.seq_dur
# set random start index
track.chunk_start = random.uniform(
0, track.duration - self.args.seq_dur
)
# load source audio and apply time domain source_augmentations
audio = track.sources[source].audio.T
audio = self.source_augmentations(audio)
audio_sources.append(audio)
# create stem tensor of shape (source, channel, samples)
stems = np.stack(audio_sources, axis=0)
# # apply linear mix over source index=0
x = np.sum(stems, axis=0)
# get the target stem
if target_ind is not None:
y = stems[target_ind]
# assuming vocal/accompaniment scenario if target!=source
else:
vocind = list(self.mus.setup['sources'].keys()).index('vocals')
# apply time domain subtraction
y = x - stems[vocind]
# for validation and test, we deterministically yield the full musdb track
else:
# get the non-linear source mix straight from musdb
x = track.audio.T
y = track.targets[self.args.target].audio.T
return x, y
def reset(self):
if self._shuffle:
self._indexes = self.rng.permutation(self._size)
else:
self._indexes = np.arange(self._size)
super(MUSDBDataSource, self).reset()
|
[
"numpy.flip",
"random.uniform",
"random.choice",
"numpy.arange",
"random.seed",
"numpy.stack",
"musdb.DB",
"numpy.sum",
"random.random",
"numpy.random.RandomState"
] |
[((1174, 1199), 'random.uniform', 'random.uniform', (['low', 'high'], {}), '(low, high)\n', (1188, 1199), False, 'import random\n'), ((1395, 1412), 'numpy.flip', 'np.flip', (['audio', '(0)'], {}), '(audio, 0)\n', (1402, 1412), True, 'import numpy as np\n'), ((3811, 3828), 'random.seed', 'random.seed', (['seed'], {}), '(seed)\n', (3822, 3828), False, 'import random\n'), ((4072, 4168), 'musdb.DB', 'musdb.DB', ([], {'root': 'args.root', 'is_wav': 'args.is_wav', 'split': 'None', 'subsets': '"""train"""', 'download': 'download'}), "(root=args.root, is_wav=args.is_wav, split=None, subsets='train',\n download=download)\n", (4080, 4168), False, 'import musdb\n'), ((1357, 1372), 'random.random', 'random.random', ([], {}), '()\n', (1370, 1372), False, 'import random\n'), ((3751, 3778), 'numpy.random.RandomState', 'np.random.RandomState', (['seed'], {}), '(seed)\n', (3772, 3778), True, 'import numpy as np\n'), ((5783, 5814), 'numpy.stack', 'np.stack', (['audio_sources'], {'axis': '(0)'}), '(audio_sources, axis=0)\n', (5791, 5814), True, 'import numpy as np\n'), ((5884, 5905), 'numpy.sum', 'np.sum', (['stems'], {'axis': '(0)'}), '(stems, axis=0)\n', (5890, 5905), True, 'import numpy as np\n'), ((6691, 6712), 'numpy.arange', 'np.arange', (['self._size'], {}), '(self._size)\n', (6700, 6712), True, 'import numpy as np\n'), ((5366, 5419), 'random.uniform', 'random.uniform', (['(0)', '(track.duration - self.args.seq_dur)'], {}), '(0, track.duration - self.args.seq_dur)\n', (5380, 5419), False, 'import random\n'), ((5156, 5186), 'random.choice', 'random.choice', (['self.mus.tracks'], {}), '(self.mus.tracks)\n', (5169, 5186), False, 'import random\n')]
|
import unittest
import numpy as np
from revpy import fare_transformation
class FareTransformationTest(unittest.TestCase):
def setUp(self):
# example data from page 13 of research paper
# "Optimization of Mixed Fare Structures: Theory and Applications"
# by <NAME> al. (2010)
self.fares = np.array([1200, 1000, 800, 600, 400, 200])
self.demands = np.array([31.2, 10.9, 14.8, 19.9, 26.9, 36.3])
def test_faretrafo_zero_demand(self):
demands = np.zeros(self.fares.shape)
adjusted_fares, adjusted_demand = \
fare_transformation.calc_fare_transformation(self.fares, demands)
np.testing.assert_equal([1200, np.nan, np.nan, np.nan, np.nan, np.nan],
adjusted_fares)
np.testing.assert_equal([0, np.nan, np.nan, np.nan, np.nan, np.nan],
adjusted_demand)
def test_example1(self):
# test example from above mentioned paper
adjusted_fares, adjusted_demand = \
fare_transformation.calc_fare_transformation(self.fares,
self.demands)
np.testing.assert_almost_equal(adjusted_fares, [1200, 427, 231, 28,
np.nan, np.nan], 0)
def test_example2(self):
# example containing some zero demands
demands = np.array([0, 15, 0, 30, 2, 60])
adjusted_fares, adjusted_demand = \
fare_transformation.calc_fare_transformation(self.fares, demands)
np.testing.assert_almost_equal(adjusted_fares, [1200, 1000, np.nan,
400, np.nan, np.nan, ])
def test_efficient_strategies(self):
fares = np.array([69.5, 59.5, 48.5, 37.5, 29.])
demands = np.array([3, 1, 0, 0, 10])
Q = demands.cumsum()
TR = Q*fares
__, __, __, __, eff_indices = \
fare_transformation.efficient_strategies(Q, TR, fares[0])
self.assertEqual(eff_indices.tolist(), [0, 1, 4])
|
[
"revpy.fare_transformation.calc_fare_transformation",
"revpy.fare_transformation.efficient_strategies",
"numpy.testing.assert_equal",
"numpy.array",
"numpy.zeros",
"numpy.testing.assert_almost_equal"
] |
[((329, 371), 'numpy.array', 'np.array', (['[1200, 1000, 800, 600, 400, 200]'], {}), '([1200, 1000, 800, 600, 400, 200])\n', (337, 371), True, 'import numpy as np\n'), ((395, 441), 'numpy.array', 'np.array', (['[31.2, 10.9, 14.8, 19.9, 26.9, 36.3]'], {}), '([31.2, 10.9, 14.8, 19.9, 26.9, 36.3])\n', (403, 441), True, 'import numpy as np\n'), ((503, 529), 'numpy.zeros', 'np.zeros', (['self.fares.shape'], {}), '(self.fares.shape)\n', (511, 529), True, 'import numpy as np\n'), ((587, 652), 'revpy.fare_transformation.calc_fare_transformation', 'fare_transformation.calc_fare_transformation', (['self.fares', 'demands'], {}), '(self.fares, demands)\n', (631, 652), False, 'from revpy import fare_transformation\n'), ((662, 753), 'numpy.testing.assert_equal', 'np.testing.assert_equal', (['[1200, np.nan, np.nan, np.nan, np.nan, np.nan]', 'adjusted_fares'], {}), '([1200, np.nan, np.nan, np.nan, np.nan, np.nan],\n adjusted_fares)\n', (685, 753), True, 'import numpy as np\n'), ((790, 879), 'numpy.testing.assert_equal', 'np.testing.assert_equal', (['[0, np.nan, np.nan, np.nan, np.nan, np.nan]', 'adjusted_demand'], {}), '([0, np.nan, np.nan, np.nan, np.nan, np.nan],\n adjusted_demand)\n', (813, 879), True, 'import numpy as np\n'), ((1045, 1115), 'revpy.fare_transformation.calc_fare_transformation', 'fare_transformation.calc_fare_transformation', (['self.fares', 'self.demands'], {}), '(self.fares, self.demands)\n', (1089, 1115), False, 'from revpy import fare_transformation\n'), ((1182, 1273), 'numpy.testing.assert_almost_equal', 'np.testing.assert_almost_equal', (['adjusted_fares', '[1200, 427, 231, 28, np.nan, np.nan]', '(0)'], {}), '(adjusted_fares, [1200, 427, 231, 28, np.nan,\n np.nan], 0)\n', (1212, 1273), True, 'import numpy as np\n'), ((1421, 1452), 'numpy.array', 'np.array', (['[0, 15, 0, 30, 2, 60]'], {}), '([0, 15, 0, 30, 2, 60])\n', (1429, 1452), True, 'import numpy as np\n'), ((1510, 1575), 'revpy.fare_transformation.calc_fare_transformation', 'fare_transformation.calc_fare_transformation', (['self.fares', 'demands'], {}), '(self.fares, demands)\n', (1554, 1575), False, 'from revpy import fare_transformation\n'), ((1585, 1679), 'numpy.testing.assert_almost_equal', 'np.testing.assert_almost_equal', (['adjusted_fares', '[1200, 1000, np.nan, 400, np.nan, np.nan]'], {}), '(adjusted_fares, [1200, 1000, np.nan, 400, np\n .nan, np.nan])\n', (1615, 1679), True, 'import numpy as np\n'), ((1791, 1831), 'numpy.array', 'np.array', (['[69.5, 59.5, 48.5, 37.5, 29.0]'], {}), '([69.5, 59.5, 48.5, 37.5, 29.0])\n', (1799, 1831), True, 'import numpy as np\n'), ((1853, 1879), 'numpy.array', 'np.array', (['[3, 1, 0, 0, 10]'], {}), '([3, 1, 0, 0, 10])\n', (1861, 1879), True, 'import numpy as np\n'), ((1983, 2040), 'revpy.fare_transformation.efficient_strategies', 'fare_transformation.efficient_strategies', (['Q', 'TR', 'fares[0]'], {}), '(Q, TR, fares[0])\n', (2023, 2040), False, 'from revpy import fare_transformation\n')]
|
from __future__ import (absolute_import, division, print_function,
unicode_literals)
from matplotlib.externals import six
from matplotlib.tri import Triangulation
import _tri as _tri
import numpy as np
class TriFinder(object):
"""
Abstract base class for classes used to find the triangles of a
Triangulation in which (x,y) points lie.
Rather than instantiate an object of a class derived from TriFinder, it is
usually better to use the function
:func:`matplotlib.tri.Triangulation.get_trifinder`.
Derived classes implement __call__(x,y) where x,y are array_like point
coordinates of the same shape.
"""
def __init__(self, triangulation):
if not isinstance(triangulation, Triangulation):
raise ValueError('Expected a Triangulation object')
self._triangulation = triangulation
class TrapezoidMapTriFinder(TriFinder):
"""
:class:`~matplotlib.tri.TriFinder` class implemented using the trapezoid
map algorithm from the book "Computational Geometry, Algorithms and
Applications", second edition, by <NAME>, <NAME>, <NAME>
and <NAME>.
The triangulation must be valid, i.e. it must not have duplicate points,
triangles formed from colinear points, or overlapping triangles. The
algorithm has some tolerance to triangles formed from colinear points, but
this should not be relied upon.
"""
def __init__(self, triangulation):
TriFinder.__init__(self, triangulation)
self._cpp_trifinder = _tri.TrapezoidMapTriFinder(
triangulation.get_cpp_triangulation())
self._initialize()
def __call__(self, x, y):
"""
Return an array containing the indices of the triangles in which the
specified x,y points lie, or -1 for points that do not lie within a
triangle.
*x*, *y* are array_like x and y coordinates of the same shape and any
number of dimensions.
Returns integer array with the same shape and *x* and *y*.
"""
x = np.asarray(x, dtype=np.float64)
y = np.asarray(y, dtype=np.float64)
if x.shape != y.shape:
raise ValueError("x and y must be array-like with the same shape")
# C++ does the heavy lifting, and expects 1D arrays.
indices = self._cpp_trifinder.find_many(x.ravel(), y.ravel())
indices.shape = x.shape
return indices
def _get_tree_stats(self):
"""
Return a python list containing the statistics about the node tree:
0: number of nodes (tree size)
1: number of unique nodes
2: number of trapezoids (tree leaf nodes)
3: number of unique trapezoids
4: maximum parent count (max number of times a node is repeated in
tree)
5: maximum depth of tree (one more than the maximum number of
comparisons needed to search through the tree)
6: mean of all trapezoid depths (one more than the average number
of comparisons needed to search through the tree)
"""
return self._cpp_trifinder.get_tree_stats()
def _initialize(self):
"""
Initialize the underlying C++ object. Can be called multiple times if,
for example, the triangulation is modified.
"""
self._cpp_trifinder.initialize()
def _print_tree(self):
"""
Print a text representation of the node tree, which is useful for
debugging purposes.
"""
self._cpp_trifinder.print_tree()
|
[
"numpy.asarray"
] |
[((2063, 2094), 'numpy.asarray', 'np.asarray', (['x'], {'dtype': 'np.float64'}), '(x, dtype=np.float64)\n', (2073, 2094), True, 'import numpy as np\n'), ((2107, 2138), 'numpy.asarray', 'np.asarray', (['y'], {'dtype': 'np.float64'}), '(y, dtype=np.float64)\n', (2117, 2138), True, 'import numpy as np\n')]
|
""" Generate BpForms for all of the proteins in PRO, verify
them, and calculate their properties
:Author: <NAME> <<EMAIL>>
:Date: 2019-06-24
:Copyright: 2019, Karr Lab
:License: MIT
"""
from Bio import SeqIO
from Bio.Seq import Seq
from Bio.SeqRecord import SeqRecord
from matplotlib import pyplot
from xml.etree import ElementTree
import bpforms
import copy
import csv
import matplotlib
import numpy
import os
import pickle
import re
import requests
import requests_cache
IN_URL = 'https://proconsortium.org/download/current/pro_nonreasoned.obo'
IN_OBO_FILENAME = os.path.join('examples', 'pro_nonreasoned.obo')
IN_PKL_FILENAME = os.path.join('examples', 'pro_nonreasoned.pkl')
IN_TSV_FILELANE = os.path.join('examples', 'pro_input.in.tsv') # from <NAME>
IN_MONOMERS_FILENAME = os.path.join('examples', 'pro.monomers.csv')
UNIPROT_SEQ_ENDPOINT = 'https://www.uniprot.org/uniprot/{}.fasta'
UNIPROT_XML_ENDPOINT = 'https://www.uniprot.org/uniprot/{}.xml'
OUT_PICKLE_FILENAME = os.path.join('examples', 'pro_input.out.pkl')
OUT_PICKLE_FILENAME_2 = os.path.join('examples', 'pro_input.out.2.pkl')
OUT_TSV_FILENAME = os.path.join('examples', 'pro_input.out.tsv')
OUT_FASTA_FILENAME = os.path.join('examples', 'pro_input.fasta')
OUT_FIG_FILENAME = os.path.join('examples', 'pro_input.svg')
OUT_STRUCTURE_DIRNAME = os.path.join('examples', 'pro_input_structure')
OUT_VIZ_DIRNAME = os.path.join('examples', 'pro_input_viz')
cache_name = os.path.join('examples', 'pro')
session = requests_cache.core.CachedSession(cache_name, backend='sqlite', expire_after=None)
session.mount('https://www.uniprot.org/', requests.adapters.HTTPAdapter(max_retries=5))
AA_CHARS_TO_CODES = {
'Ala': 'A',
'Arg': 'R',
'Asn': 'N',
'Asp': 'D',
'Cys': 'C',
'Glu': 'E',
'Gln': 'Q',
'Gly': 'G',
'His': 'H',
'Ile': 'I',
'Leu': 'L',
'Lys': 'K',
'Met': 'M',
'Phe': 'F',
'Pro': 'P',
'Ser': 'S',
'Thr': 'T',
'Trp': 'W',
'Tyr': 'Y',
'Val': 'V',
}
def run(in_obo_filename=IN_OBO_FILENAME, in_pkl_filename=IN_PKL_FILENAME, in_tsv_filename=IN_TSV_FILELANE,
in_monomers_filename=IN_MONOMERS_FILENAME,
max_num_proteins=None,
out_pickle_filename=OUT_PICKLE_FILENAME, out_pickle_filename_2=OUT_PICKLE_FILENAME_2,
out_tsv_filename=OUT_TSV_FILENAME, out_fasta_filename=OUT_FASTA_FILENAME,
out_fig_filename=OUT_FIG_FILENAME, out_structure_dirname=OUT_STRUCTURE_DIRNAME,
out_viz_dirname=OUT_VIZ_DIRNAME):
""" Download PRO ontology, generate proteoforms, and encode with BpForms
Args:
in_obo_filename (:obj:`str`, optional): path to save/read PRO ontology in OBO format
in_pkl_filename (:obj:`str`, optional): path to save/read parsed content of PRO ontology
in_tsv_filename (:obj:`str`, optional): path to read PRO entries in TSV format
in_monomers_filename (:obj:`str`, optional): path to list of ids of monomeric forms used
by PRO and their alphabet code in tab-separated format
max_num_proteins (:obj:`int`, optional): maximum number of proteins to analyze
out_pickle_filename (:obj:`str`, optional): path to save results in pickle format
out_pickle_filename_2 (:obj:`str`, optional): path to save results in pickle format
out_tsv_filename (:obj:`str`, optional): path to save results in tab-separated format
out_fasta_filename (:obj:`str`, optional): path to save results in FASTA format
out_fig_filename (:obj:`str`, optional): path to save plot of results
out_structure_dirname (:obj:`str`, optional): path to save preoteoforms in CML format
out_viz_dirname (:obj:`str`, optional): path to save preoteoforms im SVG format
Returns:
:obj:`list` of :obj:`dict`: proteoforms encoded with BpForms
"""
# get the PRO ontology and extract the modified proteins from the ontology
# proteins = get_pro_from_obo(obo_filename=in_obo_filename, pkl_filename=in_pkl_filename, max_num_proteins=max_num_proteins)
proteins = get_pro_from_tsv(in_tsv_filename, max_num_proteins=max_num_proteins)
# parse the modified proteins and retrieve their sequences
if not os.path.isfile(out_pickle_filename):
# parse the modified proteins and retrieve their sequences
parsed_proteins = []
for i_protein, protein in enumerate(proteins):
if i_protein % 100 == 0:
print('Parsing protein {} of {}'.format(i_protein + 1, len(proteins)))
parsed_proteins.append(parse_protein(protein))
# save the parsed proteins in pickle format
with open(out_pickle_filename, 'wb') as file:
pickle.dump(parsed_proteins, file)
else:
# load saved parsed proteins in pickle format
with open(out_pickle_filename, 'rb') as file:
parsed_proteins = pickle.load(file)
# read list of monomers
monomers = {}
with open(in_monomers_filename, 'r') as file:
reader = csv.DictReader(file, dialect='excel')
for row in reader:
monomers[row['PRO id']] = {
'mod': bpforms.protein_alphabet.monomers.get(row['BpForms code'], None),
'origin': [],
}
if row['Base monomer']:
monomers[row['PRO id']]['origin'] = row['Base monomer'].split(', ')
# generate list of modified monomeric forms
for protein in parsed_proteins:
for modification in protein['modifications']:
if modification['monomer'] not in monomers:
monomers[modification['monomer']] = {
'mod': None,
'origin': [],
}
# print list of unmapped monomers
unmapped_monomers = []
for monomer, code in monomers.items():
if not code['mod']:
unmapped_monomers.append(monomer)
unmapped_monomers.sort()
if unmapped_monomers:
print('Several PRO monomeric forms have not been mapped to BpForms monomeric forms:\n {}'.format(
'\n '.join(unmapped_monomers)))
# check for inconsistencies between residue and modified monomeric form
monomer_codes = {}
for code, monomer in bpforms.protein_alphabet.monomers.items():
monomer_codes[monomer] = code
for protein in parsed_proteins:
for modification in protein.get('modifications', []):
if modification['residue'] and modification['monomer']:
monomer = monomers.get(modification['monomer'], None)
if (monomer['mod'] and monomer['mod'].get_canonical_code(monomer_codes) != modification['residue']) \
or (monomer['origin'] and modification['residue'] not in monomer['origin']):
codes = set(monomer['origin'])
if monomer['mod']:
codes.add(monomer['mod'].get_canonical_code(monomer_codes))
msg = 'Modified monomeric form {} potentially inconsistent with residue {} != {}'.format(
modification['monomer'], modification['residue'],
', '.join(codes))
print(protein['id'] + ': ' + msg)
# generate BpForms for each protein
if not os.path.isdir(out_structure_dirname):
os.mkdir(out_structure_dirname)
if not os.path.isdir(out_viz_dirname):
os.mkdir(out_viz_dirname)
if not os.path.isfile(out_pickle_filename_2):
for i_protein, protein in enumerate(parsed_proteins):
if i_protein % 100 == 0:
print('Generating BpForms {} of {}'.format(i_protein + 1, len(parsed_proteins)))
protein['modified_seq'] = None
if not protein['uniprot_id']:
continue
if not protein['seq']:
continue
if protein['pro_errors']:
continue
processed_form = gen_bpform(protein, monomers, monomer_codes, apply_modifications=False)
protein['processed_seq'] = str(processed_form)
if not processed_form.validate():
processed_formula = processed_form.get_formula()
protein['processed_formula'] = str(processed_formula)
protein['processed_mol_wt'] = processed_form.get_mol_wt()
protein['processed_charge'] = processed_form.get_charge()
if not protein['modifications']:
continue
modified_form = gen_bpform(protein, monomers, monomer_codes, include_annotations=False)
protein['modified_seq'] = str(modified_form)
modified_form = gen_bpform(protein, monomers, monomer_codes)
if not modified_form.validate():
modified_formula = modified_form.get_formula()
protein['modified_full_seq'] = str(modified_form)
protein['modified_formula'] = str(modified_formula)
protein['modified_mol_wt'] = modified_form.get_mol_wt()
protein['modified_charge'] = modified_form.get_charge()
protein['modifications_formula'] = str(modified_formula - processed_formula)
protein['modifications_mol_wt'] = protein['modified_mol_wt'] - protein['processed_mol_wt']
protein['modifications_charge'] = protein['modified_charge'] - protein['processed_charge']
# with open(os.path.join(out_structure_dirname, protein['id'] + '.cml'), 'w') as file:
# file.write(modified_form.export('cml'))
form = gen_bpform(protein, monomers, monomer_codes,
apply_processing=False, include_annotations=True)
seq_features = []
if protein['processing']:
seq_features.append({
'label': 'Processed',
'color': '#cccccc',
'positions': [],
})
last = 0
for p in protein['processing']:
seq_features[0]['positions'].append([last + 1, p['start'] - 1])
last = p['end']
seq_features[0]['positions'].append([
protein['processing'][-1]['end'] + 1,
len(form.seq),
])
if protein['processing'][0]['start'] == 1:
seq_features[0]['positions'].pop(0)
if protein['processing'][-1]['end'] == len(form.seq):
seq_features[0]['positions'].pop(len(seq_features[0]['positions']) - 1)
with open(os.path.join(out_viz_dirname, protein['id'] + '.svg'), 'w') as file:
file.write(form.get_genomic_image(seq_features, width=910))
if modified_form.get_canonical_seq(monomer_codes) != protein['processed_seq']:
protein['pro_errors'].append('Modified sequence for {} not compatible with the processed sequence'.format(
protein['id']))
# save the parsed proteins in pickle format
with open(out_pickle_filename_2, 'wb') as file:
pickle.dump(parsed_proteins, file)
else:
with open(out_pickle_filename_2, 'rb') as file:
parsed_proteins = pickle.load(file)
# save the proteoforms in TSV format
with open(out_tsv_filename, 'w') as file:
writer = csv.writer(file, dialect='excel-tab')
writer.writerow(['PRO id', 'UniProt id', 'Organism',
'Unmodified sequence (IUBMB)',
'Processing', 'Deletions', 'Processsed sequence (IUBMB)', 'Processsed formula',
'Processsed molecular weight', 'Processsed charge',
'Modifications', 'Crosslinks', 'Modified sequence (abbreviated BpForms)', 'Modified sequence (BpForms)',
'Is modified sequence concrete', 'Modified formula', 'Modified molecular weight', 'Modified charge',
'Modifications formula', 'Modifications molecular weight', 'Modifications charge',
'PRO issues', 'Monomeric form issues'])
for parsed_protein in parsed_proteins:
if parsed_protein.get('pro_errors', None):
pro_errors = '. '.join(parsed_protein['pro_errors']) + '.'
else:
pro_errors = None
if parsed_protein.get('modified_errors', None):
modified_errors = '. '.join(parsed_protein['modified_errors']) + '.'
else:
modified_errors = None
writer.writerow([
parsed_protein['id'],
parsed_protein.get('uniprot_id', None),
parsed_protein.get('organism', None),
parsed_protein.get('seq', None),
', '.join('{}-{}'.format(p['start'], p['end']) for p in parsed_protein['processing']),
', '.join('{}-{}'.format(deletion[0], deletion[1]) for deletion in parsed_protein.get('deletions', [])),
parsed_protein.get('processed_seq', None),
parsed_protein.get('processed_formula', None),
parsed_protein.get('processed_mol_wt', None),
parsed_protein.get('processed_charge', None),
', '.join('{} --> {} ({})'.format(m['residue'] or '?', m['monomer'], ', '.join(str(p) for p in m['positions']))
for m in parsed_protein['modifications']),
', '.join('{}{}-{}{}'.format(xlink[0][1], xlink[0][0], xlink[1][1], xlink[1][0])
for xlink in parsed_protein.get('crosslinks', [])),
parsed_protein.get('modified_seq', None),
parsed_protein.get('modified_full_seq', None),
parsed_protein.get('modified_concrete', False),
parsed_protein.get('modified_formula', None),
parsed_protein.get('modified_mol_wt', None),
parsed_protein.get('modified_charge', None),
parsed_protein.get('modifications_formula', None),
parsed_protein.get('modifications_mol_wt', None),
parsed_protein.get('modifications_charge', None),
pro_errors,
modified_errors,
])
# save the proteoforms in FASTA format
seqs = (SeqRecord(id='{} | {}'.format(protein['id'], protein['uniprot_id']),
seq=Seq(protein['modified_seq']),
description='')
for protein in parsed_proteins
if protein['modified_seq'])
SeqIO.write(seqs, out_fasta_filename, "fasta")
# analyze frequency of modifications
plot_modifications(parsed_proteins, fig_filename=out_fig_filename)
# return proteins
return proteins, parsed_proteins
def get_pro_from_obo(obo_filename=IN_OBO_FILENAME, pkl_filename=IN_PKL_FILENAME, max_num_proteins=None):
""" Get the PRO ontology and extract the modified proteins from the ontology
Args:
obo_filename (:obj:`str`, optional): filename to save PRO in OBO format
pkl_filename (:obj:`str`, optional): filename to save/read PRO from pickled file
max_num_proteins (:obj:`int`, optional): maximum number of proteins to analyze
Returns:
:obj:`list` of :obj:`dict`: list of PRO ontology terms for modified proteins
"""
# download PRO
if not os.path.isfile(obo_filename):
response = requests.get(IN_URL)
response.raise_for_status()
with open(obo_filename, 'wb') as file:
file.write(response.content)
# parse PRO or read from cache
if not os.path.isfile(pkl_filename):
# parse PRO
proteins = []
protein = None
with open(obo_filename, 'r') as file:
for line in file:
line = line.rstrip('\n')
if line.startswith('['):
if line.startswith('[Term]'):
if max_num_proteins is not None and len(proteins) >= max_num_proteins:
break
protein = {}
else:
protein = None
elif line and protein is not None:
key, _, value = line.partition(': ')
if key not in protein:
protein[key] = []
protein[key].append(value)
if key == 'comment' and value.startswith('Category=organism-modification.'):
proteins.append(protein)
# save PRO in pickle format
with open(pkl_filename, 'wb') as file:
pickle.dump(proteins, file)
else:
# load PRO from pickle format
with open(pkl_filename, 'rb') as file:
proteins = pickle.load(file)
if max_num_proteins is not None and max_num_proteins < len(proteins):
proteins = proteins[0:max_num_proteins]
# return PRO
return proteins
def get_pro_from_tsv(filename, max_num_proteins=None):
""" Extract PRO entries from TSV file
Args:
obo_filename (:obj:`str`, optional): filename to save PRO in OBO format
max_num_proteins (:obj:`int`, optional): maximum number of proteins to analyze
Returns:
:obj:`list` of :obj:`dict`: list of PRO ontology terms for modified proteins
"""
proteins = []
with open(filename, 'r') as file:
reader = csv.DictReader(file, fieldnames=('id', 'category', 'synonym_type', 'seq'), dialect='excel-tab')
for row in reader:
proteins.append({
'id': [row['id']],
'category': [row['category']],
'synonym': ['"{}" {} PRO-proteoform-std'.format(row['seq'], row['synonym_type'])],
})
if max_num_proteins is not None and len(proteins) >= max_num_proteins:
break
return proteins
def parse_protein(protein):
""" Parse the modification information from a term for a modified protein
Args:
protein (:obj:`dict`): term for a modified protein
Returns:
:obj:`dict` with PRO id, UniProt id, processing start position, processing end position, unmodified sequence, and modifications
"""
assert len(protein['id']) == 1
id = protein['id'][0]
errors = []
seq_synonyms = []
for synonym in protein.get('synonym', []):
if synonym.startswith('"UniProtKB:') and ' PRO-proteoform-std' in synonym:
seq_synonyms.append(synonym)
if not seq_synonyms:
errors.append('No synonym which defines a modified sequence')
return {
'id': id,
'uniprot_id': None,
'processing': [],
'modifications': [],
'seq': None,
'pro_errors': errors,
}
elif len(seq_synonyms) > 1:
errors.append('Multiple synonyms which define modified sequences')
synonym = seq_synonyms[0]
uniprot_id, _, processing_modifications_type = synonym.partition(', ')
uniprot_id = uniprot_id.partition(':')[2]
organism_name = None
response = session.get(UNIPROT_XML_ENDPOINT.format(uniprot_id))
response.raise_for_status()
if response.content:
xml_root = ElementTree.fromstring(response.content)
entry = xml_root.find('{http://uniprot.org/uniprot}entry')
organism = entry.find('{http://uniprot.org/uniprot}organism')
names = organism.findall('{http://uniprot.org/uniprot}name')
for name in names:
if name.get('type') == 'scientific':
organism_name = name.text
break
response = session.get(UNIPROT_SEQ_ENDPOINT.format(uniprot_id))
response.raise_for_status()
seq = response.content.decode('utf-8').partition('\n')[2].replace('\n', '')
if not seq:
errors.append('No sequence for UniProt entry; entry may be deprecated')
processing_modifications = processing_modifications_type.partition('"')[0]
processing = []
while True:
match = re.match(r'^(\?|\d+)\-(\?|\d+)(, |$)', processing_modifications)
if match:
if match.group(1) == '?':
start = None
errors.append('Unknown processing start position')
else:
start = int(float(match.group(1)))
if start <= 0 or start > len(seq):
errors.append('Start position must be within sequence')
if match.group(2) == '?':
end = None
errors.append('Unknown processing end position')
else:
end = int(float(match.group(2)))
if end <= 0 or end > len(seq):
errors.append('End position must be within sequence')
if start and end and start > end:
errors.append('End position must be after start position')
processing.append({
'start': start,
'end': end,
})
processing_modifications = processing_modifications[len(match.group(0)):]
else:
break
if processing_modifications.startswith('which') \
or processing_modifications.startswith('with') \
or 'MOD:00046 OR Thr-163, MOD:00047' in processing_modifications:
modifications_str = []
errors.append('Unable to parse sequence')
elif processing_modifications:
modifications_str = processing_modifications.split('|')
else:
modifications_str = []
modifications = []
for modification in modifications_str:
if modification and modification[0] == '(' and modification[-1] == ')':
modification = modification[1:-1]
if ' or ' in modification or ' and/or ' in modification:
errors.append('Boolean logic not supported')
continue
if ', ' in modification:
residue_positions, _, monomer = modification.partition(', ')
residue_codes = set()
positions = []
for residue_position in residue_positions.split('/'):
residue_chars, _, position = residue_position.partition('-')
residue_code = AA_CHARS_TO_CODES[residue_chars]
position = int(float(position))
if position > len(seq):
errors.append('Position {} is greater than the sequence length {}'.format(position, len(seq)))
elif seq[position - 1] != residue_code:
errors.append('Position {} != {}'.format(position, residue_code))
residue_codes.add(residue_code)
positions.append(position)
if len(residue_codes) != 1 and monomer != 'PR:000026291':
residue_code = None
errors.append('Residues {{{}}} annotated with the same modification {}'.format(
', '.join(residue_codes), monomer))
else:
residue_code = None
positions = []
monomer = modification
if monomer == 'PR:000026291':
for residue_code in residue_codes:
modifications.append({
'residue': residue_code,
'positions': [p for p in positions if seq[p - 1] == residue_code],
'monomer': monomer,
})
else:
modifications.append({
'residue': residue_code,
'positions': positions,
'monomer': monomer,
})
return {
'id': id,
'uniprot_id': uniprot_id,
'organism': organism_name,
'processing': processing,
'modifications': modifications,
'seq': seq,
'pro_errors': errors,
}
def gen_bpform(protein, pro_ids_to_bpform_monomers, monomer_codes,
apply_processing=True, apply_modifications=True, include_annotations=True):
""" Generate BpForm for a modified protein in PRO
Args:
protein (:obj:`dict`): term for modified protein
pro_ids_to_bpform_monomers (:obj:`dict`): dictionary which maps ids of monomeric forms
used by PRO to monomeric forms in the BpForms protein alphabet
monomer_codes (:obj:`dict`): dictionary that maps monomers to their codes in the alphabet
apply_processing (:obj:`bool`, optional): if :obj:`True`, include processing in proteoform
apply_modifications (:obj:`bool`, optional): if :obj:`True`, include modifications in proteoform
include_annotations (:obj:`bool`, optional): if :obj:`True`, include metadata about modified monomers
Returns:
:obj:`bpforms.ProteinForm`: BpForm for a term in PRO
"""
form = bpforms.ProteinForm()
monomers = bpforms.protein_alphabet.monomers
# generate BpForm for unmodified sequence
for base in protein['seq']:
form.seq.append(monomers[base])
# apply processing
modifications = copy.deepcopy(protein['modifications'])
seq = protein['seq']
if apply_processing and protein['processing']:
procesed_seq = []
seq = ''
for processing in protein['processing']:
procesed_seq.extend(form.seq[processing['start']-1:processing['end']])
seq += protein['seq'][processing['start']-1:processing['end']]
form.seq = procesed_seq
for modification in modifications:
modification['processed_positions'] = []
for position in modification['positions']:
seq_len = 0
processed_position = None
for processing in protein['processing']:
if position >= processing['start'] and position <= processing['end']:
processed_position = seq_len + position - processing['start'] + 1
break
seq_len += processing['end'] - processing['start'] + 1
if processed_position is not None:
modification['processed_positions'].append(processed_position)
else:
for modification in modifications:
modification['processed_positions'] = modification['positions']
# apply modifications
if apply_modifications:
concrete = True
protein['modified_errors'] = []
for modification in modifications:
monomer = pro_ids_to_bpform_monomers[modification['monomer']]['mod']
origin = pro_ids_to_bpform_monomers[modification['monomer']]['origin']
if modification['monomer'].startswith('CHEBI:'):
mod_ns = 'chebi'
elif modification['monomer'].startswith('MOD:'):
mod_ns = 'mod'
elif modification['monomer'].startswith('PR:'):
mod_ns = 'pr'
elif modification['monomer'].startswith('UniCarbKB:'):
mod_ns = 'unicarbkb'
else:
raise ValueError('Unsupported identifier {}'.format(modification['monomer']))
if modification['monomer'] == 'PR:000026291':
if include_annotations:
monomer = bpforms.Monomer().from_dict(
monomers[modification['residue']].to_dict(
alphabet=bpforms.protein_alphabet),
alphabet=bpforms.protein_alphabet)
else:
monomer = bpforms.Monomer()
monomer.id = None
monomer.name = None
monomer.synonyms = []
monomer.identifiers = [bpforms.Identifier('pr', modification['monomer'])]
monomer.comments = None
elif modification['monomer'].startswith('CHEBI:'):
if include_annotations:
monomer = bpforms.Monomer().from_dict(
monomers[modification['residue']].to_dict(
alphabet=bpforms.protein_alphabet),
alphabet=bpforms.protein_alphabet)
else:
monomer = bpforms.Monomer()
monomer.id = None
monomer.name = None
monomer.synonyms = []
monomer.identifiers = [bpforms.Identifier('chebi', modification['monomer'])]
monomer.comments = None
elif monomer is None:
concrete = False
monomer = bpforms.Monomer(
identifiers=[bpforms.Identifier(mod_ns, modification['monomer'])])
if modification['positions']:
for position in modification['processed_positions']:
if form.seq[position - 1] == monomers[seq[position - 1]]:
if monomer not in bpforms.protein_alphabet.monomers.values():
monomer.base_monomers = [form.seq[position - 1]]
form.seq[position - 1] = monomer
else:
protein['modified_errors'].append('Unable to set monomeric form at position {}'.format(
position))
elif modification['residue']:
concrete = False
if include_annotations:
monomer2 = bpforms.Monomer().from_dict(
monomer.to_dict(
alphabet=bpforms.protein_alphabet),
alphabet=bpforms.protein_alphabet)
else:
monomer2 = bpforms.Monomer()
monomer2.id = None
monomer2.name = None
monomer2.synonyms = []
monomer2.identifiers = [bpforms.Identifier(mod_ns, modification['monomer'])]
monomer2.base_monomers = [bpforms.protein_alphabet.monomers.get(modification['positions'])]
monomer2.start_position = seq.find(modification['residue']) + 1
monomer2.end_position = seq.rfind(modification['residue']) + 1
set_monomer = False
for i_monomer in range(monomer2.start_position, monomer2.end_position + 1):
if form.seq[i_monomer - 1] == monomers[seq[i_monomer - 1]]:
set_monomer = True
form.seq[i_monomer - 1] = monomer2
break
if not set_monomer:
protein['modified_errors'].append('Unable to set monomeric form')
else:
concrete = False
canonical_code = monomer.get_canonical_code(monomer_codes)
if include_annotations:
monomer2 = bpforms.Monomer().from_dict(
monomer.to_dict(
alphabet=bpforms.protein_alphabet),
alphabet=bpforms.protein_alphabet)
else:
monomer2 = bpforms.Monomer()
monomer2.id = None
monomer2.name = None
monomer2.synonyms = []
monomer2.identifiers = [bpforms.Identifier(mod_ns, modification['monomer'])]
monomer2.monomers_position = [
bpforms.protein_alphabet.monomers.get(code) for code in origin]
if canonical_code and canonical_code != '?':
start_position = seq.find(canonical_code) + 1
end_position = seq.rfind(canonical_code) + 1
if start_position == 0:
protein['modified_errors'].append('Sequence does not contain residue {} for modification {}'.format(
canonical_code, modification['monomer']))
else:
monomer2.start_position = start_position
monomer2.end_position = end_position
elif origin:
start_position = float('inf')
end_position = -float('inf')
for base in origin:
start_pos = seq.find(base) + 1
if start_pos > 0:
start_position = min(start_position, start_pos)
end_pos = seq.rfind(base) + 1
if end_pos > 0:
end_position = max(end_position, end_pos)
if numpy.isinf(start_position):
protein['modified_errors'].append('Sequence does not contain residues {} for modification {}'.format(
', '.join(origin), modification['monomer']))
else:
monomer2.start_position = start_position
monomer2.end_position = end_position
else:
monomer2.start_position = 1
monomer2.end_position = len(seq)
if monomer2.start_position:
set_monomer = False
for i_monomer in range(monomer2.start_position, monomer2.end_position + 1):
if form.seq[i_monomer - 1] == monomers[seq[i_monomer - 1]]:
monomer2.base_monomers = [bpforms.protein_alphabet.monomers.get(seq[i_monomer - 1])]
form.seq[i_monomer - 1] = monomer2
set_monomer = True
break
if not set_monomer:
protein['modified_errors'].append('Unable to set monomeric form')
# crosslinks
if protein['processing']:
xlinks = []
seq_len = 0
protein['crosslinks'] = []
protein['deletions'] = []
for left, right in zip(protein['processing'][0:-1], protein['processing'][1:]):
seq_len += left['end'] - left['start'] + 1
i_left = seq_len
i_right = i_left + 1
if left['end'] + 1 == right['start']:
protein['crosslinks'].append(((left['end'], protein['seq'][left['end']-1]),
(right['start'], protein['seq'][right['start'] - 1])))
else:
protein['deletions'].append((left['end'] + 1, right['start'] - 1))
if left['end'] + 1 != right['start']:
continue
#err = False
# if protein['seq'][left['end'] - 1] != 'C' or form.seq[i_left - 1] != bpforms.protein_alphabet.monomers.C:
# err = True
# protein['modified_errors'].append('Disulfide bond site {}{} != C'.format(
# protein['seq'][left['end'] - 1], left['end']))
# if protein['seq'][right['start'] - 1] != 'C' or form.seq[i_right - 1] != bpforms.protein_alphabet.monomers.C:
# err = True
# protein['modified_errors'].append('Disulfide bond site {}{} != C'.format(
# protein['seq'][right['start'] - 1], right['start']))
#
# if err:
# continue
concrete = False
i_left = '{}-{}'.format(seq_len - (left['end'] - left['start'] + 1) + 1, seq_len)
i_right = '{}-{}'.format(seq_len + 1, seq_len + (right['end'] - right['start'] + 1))
if apply_modifications:
form.crosslinks.add(bpforms.Bond(
#l_bond_atoms=[bpforms.Atom(bpforms.Monomer, 'S', position=11, monomer=i_left)],
#r_bond_atoms=[bpforms.Atom(bpforms.Monomer, 'S', position=11, monomer=i_right)],
#l_displaced_atoms=[bpforms.Atom(bpforms.Monomer, 'H', position=11, monomer=i_left)],
#r_displaced_atoms=[bpforms.Atom(bpforms.Monomer, 'H', position=11, monomer=i_right)],
comments='The polymer contains a disulfide bond between the ranges {} and {}'.format(i_left, i_right),
))
# validate
if apply_modifications:
protein['modified_concrete'] = concrete
protein['modified_errors'].extend(form.validate())
# return proteoform represented with BpForms
return form
def plot_modifications(proteins, organism='Homo sapiens', fig_filename=OUT_FIG_FILENAME):
""" Plot a summary of the modifications in PRO
Args:
proteins (:obj:`list` of :obj:`dict`): entries in PRO ontology
organism (:obj:`str`, optional): organism to analyze
fig_filename (:obj:`str`, optional): path to save analysis
"""
code_freq = {}
canonical_code_freq = {}
for protein in proteins:
if (organism is None or protein.get('organism', None) == organism) and protein.get('modified_seq', None):
for modification in protein['modifications']:
if modification['residue'] and modification['monomer']:
n_mods = max(1, len(modification['positions']))
if modification['residue'] not in canonical_code_freq:
canonical_code_freq[modification['residue']] = 0
if modification['monomer'] not in code_freq:
code_freq[modification['monomer']] = 0
canonical_code_freq[modification['residue']] += n_mods
code_freq[modification['monomer']] += n_mods
pyplot.style.use('ggplot')
fig, axes = pyplot.subplots(nrows=1, ncols=2, gridspec_kw={'width_ratios': [1, 4]})
fig.set_size_inches(9.3, 1.5)
plot_codes(canonical_code_freq,
'Frequency of modifications',
axes[0], ignore_canonical=False)
plot_codes(code_freq,
'Frequency of modified monomeric forms',
axes[1], ignore_canonical=True)
fig.savefig(fig_filename, transparent=True,
bbox_inches=matplotlib.transforms.Bbox([[0.69, -0.5], [8.35, 1.5]]))
pyplot.close(fig)
def plot_codes(code_freq, title, axis, ignore_canonical=False):
id_freqs = []
for code, count in code_freq.items():
if ignore_canonical and code in ['A', 'C', 'G', 'U']:
continue
id_freqs.append((code, count))
id_freqs.sort()
y_pos = numpy.arange(len(id_freqs))
freq = numpy.array([id_freq[-1] for id_freq in id_freqs])
freq = freq / numpy.sum(freq) * 100.
x_tick_labels = {id: y_pos for y_pos, (id, _) in enumerate(id_freqs)}
axis.bar(y_pos, freq, align='center', alpha=0.5)
axis.set_xticks(y_pos)
axis.set_xticklabels(x_tick_labels, rotation=270, fontsize=6, fontfamily='Raleway')
axis.set_ylabel('Frequency (%)', fontdict={
'fontsize': 10,
'fontweight': 'regular',
'fontfamily': 'Raleway',
})
axis.set_title(title, fontdict={
'fontsize': 10,
'fontweight': 'regular',
'fontfamily': 'Raleway',
})
axis.set_xlim((-0.75, len(id_freqs) - 0.25))
|
[
"csv.DictReader",
"Bio.Seq.Seq",
"numpy.array",
"bpforms.protein_alphabet.monomers.values",
"copy.deepcopy",
"requests_cache.core.CachedSession",
"matplotlib.pyplot.style.use",
"bpforms.protein_alphabet.monomers.items",
"matplotlib.pyplot.close",
"os.path.isdir",
"os.mkdir",
"Bio.SeqIO.write",
"xml.etree.ElementTree.fromstring",
"numpy.isinf",
"bpforms.protein_alphabet.monomers.get",
"bpforms.ProteinForm",
"requests.adapters.HTTPAdapter",
"re.match",
"csv.writer",
"requests.get",
"pickle.load",
"os.path.isfile",
"matplotlib.transforms.Bbox",
"bpforms.Monomer",
"bpforms.Identifier",
"pickle.dump",
"os.path.join",
"numpy.sum",
"matplotlib.pyplot.subplots"
] |
[((568, 615), 'os.path.join', 'os.path.join', (['"""examples"""', '"""pro_nonreasoned.obo"""'], {}), "('examples', 'pro_nonreasoned.obo')\n", (580, 615), False, 'import os\n'), ((634, 681), 'os.path.join', 'os.path.join', (['"""examples"""', '"""pro_nonreasoned.pkl"""'], {}), "('examples', 'pro_nonreasoned.pkl')\n", (646, 681), False, 'import os\n'), ((700, 744), 'os.path.join', 'os.path.join', (['"""examples"""', '"""pro_input.in.tsv"""'], {}), "('examples', 'pro_input.in.tsv')\n", (712, 744), False, 'import os\n'), ((783, 827), 'os.path.join', 'os.path.join', (['"""examples"""', '"""pro.monomers.csv"""'], {}), "('examples', 'pro.monomers.csv')\n", (795, 827), False, 'import os\n'), ((981, 1026), 'os.path.join', 'os.path.join', (['"""examples"""', '"""pro_input.out.pkl"""'], {}), "('examples', 'pro_input.out.pkl')\n", (993, 1026), False, 'import os\n'), ((1051, 1098), 'os.path.join', 'os.path.join', (['"""examples"""', '"""pro_input.out.2.pkl"""'], {}), "('examples', 'pro_input.out.2.pkl')\n", (1063, 1098), False, 'import os\n'), ((1118, 1163), 'os.path.join', 'os.path.join', (['"""examples"""', '"""pro_input.out.tsv"""'], {}), "('examples', 'pro_input.out.tsv')\n", (1130, 1163), False, 'import os\n'), ((1185, 1228), 'os.path.join', 'os.path.join', (['"""examples"""', '"""pro_input.fasta"""'], {}), "('examples', 'pro_input.fasta')\n", (1197, 1228), False, 'import os\n'), ((1248, 1289), 'os.path.join', 'os.path.join', (['"""examples"""', '"""pro_input.svg"""'], {}), "('examples', 'pro_input.svg')\n", (1260, 1289), False, 'import os\n'), ((1314, 1361), 'os.path.join', 'os.path.join', (['"""examples"""', '"""pro_input_structure"""'], {}), "('examples', 'pro_input_structure')\n", (1326, 1361), False, 'import os\n'), ((1380, 1421), 'os.path.join', 'os.path.join', (['"""examples"""', '"""pro_input_viz"""'], {}), "('examples', 'pro_input_viz')\n", (1392, 1421), False, 'import os\n'), ((1436, 1467), 'os.path.join', 'os.path.join', (['"""examples"""', '"""pro"""'], {}), "('examples', 'pro')\n", (1448, 1467), False, 'import os\n'), ((1478, 1564), 'requests_cache.core.CachedSession', 'requests_cache.core.CachedSession', (['cache_name'], {'backend': '"""sqlite"""', 'expire_after': 'None'}), "(cache_name, backend='sqlite',\n expire_after=None)\n", (1511, 1564), False, 'import requests_cache\n'), ((1603, 1647), 'requests.adapters.HTTPAdapter', 'requests.adapters.HTTPAdapter', ([], {'max_retries': '(5)'}), '(max_retries=5)\n', (1632, 1647), False, 'import requests\n'), ((6201, 6242), 'bpforms.protein_alphabet.monomers.items', 'bpforms.protein_alphabet.monomers.items', ([], {}), '()\n', (6240, 6242), False, 'import bpforms\n'), ((14665, 14711), 'Bio.SeqIO.write', 'SeqIO.write', (['seqs', 'out_fasta_filename', '"""fasta"""'], {}), "(seqs, out_fasta_filename, 'fasta')\n", (14676, 14711), False, 'from Bio import SeqIO\n'), ((24825, 24846), 'bpforms.ProteinForm', 'bpforms.ProteinForm', ([], {}), '()\n', (24844, 24846), False, 'import bpforms\n'), ((25059, 25098), 'copy.deepcopy', 'copy.deepcopy', (["protein['modifications']"], {}), "(protein['modifications'])\n", (25072, 25098), False, 'import copy\n'), ((37373, 37399), 'matplotlib.pyplot.style.use', 'pyplot.style.use', (['"""ggplot"""'], {}), "('ggplot')\n", (37389, 37399), False, 'from matplotlib import pyplot\n'), ((37416, 37487), 'matplotlib.pyplot.subplots', 'pyplot.subplots', ([], {'nrows': '(1)', 'ncols': '(2)', 'gridspec_kw': "{'width_ratios': [1, 4]}"}), "(nrows=1, ncols=2, gridspec_kw={'width_ratios': [1, 4]})\n", (37431, 37487), False, 'from matplotlib import pyplot\n'), ((37917, 37934), 'matplotlib.pyplot.close', 'pyplot.close', (['fig'], {}), '(fig)\n', (37929, 37934), False, 'from matplotlib import pyplot\n'), ((38255, 38305), 'numpy.array', 'numpy.array', (['[id_freq[-1] for id_freq in id_freqs]'], {}), '([id_freq[-1] for id_freq in id_freqs])\n', (38266, 38305), False, 'import numpy\n'), ((4189, 4224), 'os.path.isfile', 'os.path.isfile', (['out_pickle_filename'], {}), '(out_pickle_filename)\n', (4203, 4224), False, 'import os\n'), ((4994, 5031), 'csv.DictReader', 'csv.DictReader', (['file'], {'dialect': '"""excel"""'}), "(file, dialect='excel')\n", (5008, 5031), False, 'import csv\n'), ((7244, 7280), 'os.path.isdir', 'os.path.isdir', (['out_structure_dirname'], {}), '(out_structure_dirname)\n', (7257, 7280), False, 'import os\n'), ((7290, 7321), 'os.mkdir', 'os.mkdir', (['out_structure_dirname'], {}), '(out_structure_dirname)\n', (7298, 7321), False, 'import os\n'), ((7334, 7364), 'os.path.isdir', 'os.path.isdir', (['out_viz_dirname'], {}), '(out_viz_dirname)\n', (7347, 7364), False, 'import os\n'), ((7374, 7399), 'os.mkdir', 'os.mkdir', (['out_viz_dirname'], {}), '(out_viz_dirname)\n', (7382, 7399), False, 'import os\n'), ((7412, 7449), 'os.path.isfile', 'os.path.isfile', (['out_pickle_filename_2'], {}), '(out_pickle_filename_2)\n', (7426, 7449), False, 'import os\n'), ((11446, 11483), 'csv.writer', 'csv.writer', (['file'], {'dialect': '"""excel-tab"""'}), "(file, dialect='excel-tab')\n", (11456, 11483), False, 'import csv\n'), ((15477, 15505), 'os.path.isfile', 'os.path.isfile', (['obo_filename'], {}), '(obo_filename)\n', (15491, 15505), False, 'import os\n'), ((15526, 15546), 'requests.get', 'requests.get', (['IN_URL'], {}), '(IN_URL)\n', (15538, 15546), False, 'import requests\n'), ((15718, 15746), 'os.path.isfile', 'os.path.isfile', (['pkl_filename'], {}), '(pkl_filename)\n', (15732, 15746), False, 'import os\n'), ((17525, 17624), 'csv.DictReader', 'csv.DictReader', (['file'], {'fieldnames': "('id', 'category', 'synonym_type', 'seq')", 'dialect': '"""excel-tab"""'}), "(file, fieldnames=('id', 'category', 'synonym_type', 'seq'),\n dialect='excel-tab')\n", (17539, 17624), False, 'import csv\n'), ((19335, 19375), 'xml.etree.ElementTree.fromstring', 'ElementTree.fromstring', (['response.content'], {}), '(response.content)\n', (19357, 19375), False, 'from xml.etree import ElementTree\n'), ((20131, 20199), 're.match', 're.match', (['"""^(\\\\?|\\\\d+)\\\\-(\\\\?|\\\\d+)(, |$)"""', 'processing_modifications'], {}), "('^(\\\\?|\\\\d+)\\\\-(\\\\?|\\\\d+)(, |$)', processing_modifications)\n", (20139, 20199), False, 'import re\n'), ((4679, 4713), 'pickle.dump', 'pickle.dump', (['parsed_proteins', 'file'], {}), '(parsed_proteins, file)\n', (4690, 4713), False, 'import pickle\n'), ((4862, 4879), 'pickle.load', 'pickle.load', (['file'], {}), '(file)\n', (4873, 4879), False, 'import pickle\n'), ((11192, 11226), 'pickle.dump', 'pickle.dump', (['parsed_proteins', 'file'], {}), '(parsed_proteins, file)\n', (11203, 11226), False, 'import pickle\n'), ((11323, 11340), 'pickle.load', 'pickle.load', (['file'], {}), '(file)\n', (11334, 11340), False, 'import pickle\n'), ((16735, 16762), 'pickle.dump', 'pickle.dump', (['proteins', 'file'], {}), '(proteins, file)\n', (16746, 16762), False, 'import pickle\n'), ((16881, 16898), 'pickle.load', 'pickle.load', (['file'], {}), '(file)\n', (16892, 16898), False, 'import pickle\n'), ((37856, 37911), 'matplotlib.transforms.Bbox', 'matplotlib.transforms.Bbox', (['[[0.69, -0.5], [8.35, 1.5]]'], {}), '([[0.69, -0.5], [8.35, 1.5]])\n', (37882, 37911), False, 'import matplotlib\n'), ((38324, 38339), 'numpy.sum', 'numpy.sum', (['freq'], {}), '(freq)\n', (38333, 38339), False, 'import numpy\n'), ((5122, 5186), 'bpforms.protein_alphabet.monomers.get', 'bpforms.protein_alphabet.monomers.get', (["row['BpForms code']", 'None'], {}), "(row['BpForms code'], None)\n", (5159, 5186), False, 'import bpforms\n'), ((14510, 14538), 'Bio.Seq.Seq', 'Seq', (["protein['modified_seq']"], {}), "(protein['modified_seq'])\n", (14513, 14538), False, 'from Bio.Seq import Seq\n'), ((27504, 27521), 'bpforms.Monomer', 'bpforms.Monomer', ([], {}), '()\n', (27519, 27521), False, 'import bpforms\n'), ((27669, 27718), 'bpforms.Identifier', 'bpforms.Identifier', (['"""pr"""', "modification['monomer']"], {}), "('pr', modification['monomer'])\n", (27687, 27718), False, 'import bpforms\n'), ((10671, 10724), 'os.path.join', 'os.path.join', (['out_viz_dirname', "(protein['id'] + '.svg')"], {}), "(out_viz_dirname, protein['id'] + '.svg')\n", (10683, 10724), False, 'import os\n'), ((28165, 28182), 'bpforms.Monomer', 'bpforms.Monomer', ([], {}), '()\n', (28180, 28182), False, 'import bpforms\n'), ((28330, 28382), 'bpforms.Identifier', 'bpforms.Identifier', (['"""chebi"""', "modification['monomer']"], {}), "('chebi', modification['monomer'])\n", (28348, 28382), False, 'import bpforms\n'), ((29604, 29621), 'bpforms.Monomer', 'bpforms.Monomer', ([], {}), '()\n', (29619, 29621), False, 'import bpforms\n'), ((29773, 29824), 'bpforms.Identifier', 'bpforms.Identifier', (['mod_ns', "modification['monomer']"], {}), "(mod_ns, modification['monomer'])\n", (29791, 29824), False, 'import bpforms\n'), ((29868, 29932), 'bpforms.protein_alphabet.monomers.get', 'bpforms.protein_alphabet.monomers.get', (["modification['positions']"], {}), "(modification['positions'])\n", (29905, 29932), False, 'import bpforms\n'), ((31000, 31017), 'bpforms.Monomer', 'bpforms.Monomer', ([], {}), '()\n', (31015, 31017), False, 'import bpforms\n'), ((31169, 31220), 'bpforms.Identifier', 'bpforms.Identifier', (['mod_ns', "modification['monomer']"], {}), "(mod_ns, modification['monomer'])\n", (31187, 31220), False, 'import bpforms\n'), ((31289, 31332), 'bpforms.protein_alphabet.monomers.get', 'bpforms.protein_alphabet.monomers.get', (['code'], {}), '(code)\n', (31326, 31332), False, 'import bpforms\n'), ((27233, 27250), 'bpforms.Monomer', 'bpforms.Monomer', ([], {}), '()\n', (27248, 27250), False, 'import bpforms\n'), ((28855, 28897), 'bpforms.protein_alphabet.monomers.values', 'bpforms.protein_alphabet.monomers.values', ([], {}), '()\n', (28895, 28897), False, 'import bpforms\n'), ((32468, 32495), 'numpy.isinf', 'numpy.isinf', (['start_position'], {}), '(start_position)\n', (32479, 32495), False, 'import numpy\n'), ((27894, 27911), 'bpforms.Monomer', 'bpforms.Monomer', ([], {}), '()\n', (27909, 27911), False, 'import bpforms\n'), ((29358, 29375), 'bpforms.Monomer', 'bpforms.Monomer', ([], {}), '()\n', (29373, 29375), False, 'import bpforms\n'), ((30754, 30771), 'bpforms.Monomer', 'bpforms.Monomer', ([], {}), '()\n', (30769, 30771), False, 'import bpforms\n'), ((28569, 28620), 'bpforms.Identifier', 'bpforms.Identifier', (['mod_ns', "modification['monomer']"], {}), "(mod_ns, modification['monomer'])\n", (28587, 28620), False, 'import bpforms\n'), ((33291, 33348), 'bpforms.protein_alphabet.monomers.get', 'bpforms.protein_alphabet.monomers.get', (['seq[i_monomer - 1]'], {}), '(seq[i_monomer - 1])\n', (33328, 33348), False, 'import bpforms\n')]
|
import argparse
import contextlib
import csv
import logging
import os
import random
import subprocess
import tempfile
from typing import Callable, Dict, Iterable, List
import numpy as np
import ray
from ray.experimental.raysort import constants
from ray.experimental.raysort import logging_utils
from ray.experimental.raysort import sortlib
from ray.experimental.raysort import tracing_utils
from ray.experimental.raysort.types import (
BlockInfo,
ByteCount,
RecordCount,
PartId,
PartInfo,
Path,
)
Args = argparse.Namespace
# ------------------------------------------------------------
# Parse Arguments
# ------------------------------------------------------------
def get_args(*args, **kwargs):
parser = argparse.ArgumentParser()
parser.add_argument(
"--ray_address",
default="auto",
type=str,
help="if set to None, will launch a local Ray cluster",
)
parser.add_argument(
"--total_data_size",
default=1 * 1000 * 1024 * 1024 * 1024,
type=ByteCount,
help="total data size in bytes",
)
parser.add_argument(
"--num_mappers",
default=256,
type=int,
help="number of map tasks",
)
parser.add_argument(
"--num_mappers_per_round",
default=16,
type=int,
help="number of map tasks per first-stage merge tasks",
)
parser.add_argument(
"--num_reducers",
default=16,
type=int,
help="number of second-stage reduce tasks",
)
parser.add_argument(
"--num_concurrent_rounds",
default=4,
type=int,
help="max number of rounds of map/merge tasks in flight",
)
parser.add_argument(
"--reducer_input_chunk",
default=100 * 1024 * 1024,
type=ByteCount,
help="bytes to read from each file in reduce tasks",
)
parser.add_argument(
"--skip_sorting",
default=False,
action="store_true",
help="if set, no sorting is actually performed",
)
parser.add_argument(
"--skip_input",
default=False,
action="store_true",
help="if set, mappers will not read data from disk",
)
parser.add_argument(
"--skip_output",
default=False,
action="store_true",
help="if set, reducers will not write out results to disk",
)
# Which tasks to run?
tasks_group = parser.add_argument_group(
"tasks to run", "if no task is specified, will run all tasks"
)
tasks = ["generate_input", "sort", "validate_output"]
for task in tasks:
tasks_group.add_argument(f"--{task}", action="store_true")
args = parser.parse_args(*args, **kwargs)
# Derive additional arguments.
args.input_part_size = ByteCount(args.total_data_size / args.num_mappers)
assert args.num_mappers % args.num_mappers_per_round == 0
args.num_rounds = int(args.num_mappers / args.num_mappers_per_round)
args.mount_points = _get_mount_points()
# If no tasks are specified, run all tasks.
args_dict = vars(args)
if not any(args_dict[task] for task in tasks):
for task in tasks:
args_dict[task] = True
return args
def _get_mount_points():
default_ret = [tempfile.gettempdir()]
mnt = "/mnt"
if os.path.exists(mnt):
ret = [os.path.join(mnt, d) for d in os.listdir(mnt)]
if len(ret) > 0:
return ret
return default_ret
# ------------------------------------------------------------
# Generate Input
# ------------------------------------------------------------
def _part_info(args: Args, part_id: PartId, kind="input") -> PartInfo:
node = ray.worker.global_worker.node_ip_address
mnt = random.choice(args.mount_points)
filepath = _get_part_path(mnt, part_id, kind)
return PartInfo(part_id, node, filepath)
def _get_part_path(mnt: Path, part_id: PartId, kind="input") -> Path:
assert kind in {"input", "output", "temp"}
dir_fmt = constants.DATA_DIR_FMT[kind]
dirpath = dir_fmt.format(mnt=mnt)
os.makedirs(dirpath, exist_ok=True)
filename_fmt = constants.FILENAME_FMT[kind]
filename = filename_fmt.format(part_id=part_id)
filepath = os.path.join(dirpath, filename)
return filepath
@ray.remote
def generate_part(
args: Args, part_id: PartId, size: RecordCount, offset: RecordCount
) -> PartInfo:
logging_utils.init()
pinfo = _part_info(args, part_id)
subprocess.run(
[constants.GENSORT_PATH, f"-b{offset}", f"{size}", pinfo.path], check=True
)
logging.info(f"Generated input {pinfo}")
return pinfo
def generate_input(args: Args):
if args.skip_input:
return
size = constants.bytes_to_records(args.input_part_size)
offset = 0
tasks = []
for part_id in range(args.num_mappers):
tasks.append(generate_part.remote(args, part_id, size, offset))
offset += size
assert offset == constants.bytes_to_records(args.total_data_size), args
logging.info(f"Generating {len(tasks)} partitions")
parts = ray.get(tasks)
with open(constants.INPUT_MANIFEST_FILE, "w") as fout:
writer = csv.writer(fout)
writer.writerows(parts)
# ------------------------------------------------------------
# Sort
# ------------------------------------------------------------
def _load_manifest(args: Args, path: Path) -> List[PartInfo]:
if args.skip_input:
return [PartInfo(i, None, None) for i in range(args.num_mappers)]
with open(path) as fin:
reader = csv.reader(fin)
return [PartInfo(int(part_id), node, path) for part_id, node, path in reader]
def _load_partition(args: Args, path: Path) -> np.ndarray:
if args.skip_input:
return np.frombuffer(
np.random.bytes(args.input_part_size), dtype=np.uint8
).copy()
return np.fromfile(path, dtype=np.uint8)
def _dummy_sort_and_partition(
part: np.ndarray, boundaries: List[int]
) -> List[BlockInfo]:
N = len(boundaries)
offset = 0
size = int(np.ceil(part.size / N))
blocks = []
for _ in range(N):
blocks.append((offset, size))
offset += size
return blocks
@ray.remote
@tracing_utils.timeit("map")
def mapper(
args: Args, mapper_id: PartId, boundaries: List[int], path: Path
) -> List[np.ndarray]:
logging_utils.init()
part = _load_partition(args, path)
sort_fn = (
_dummy_sort_and_partition if args.skip_sorting else sortlib.sort_and_partition
)
blocks = sort_fn(part, boundaries)
return [part[offset : offset + size] for offset, size in blocks]
def _dummy_merge(
num_blocks: int, _n: int, get_block: Callable[[int, int], np.ndarray]
) -> Iterable[np.ndarray]:
blocks = [((i, 0), get_block(i, 0)) for i in range(num_blocks)]
while len(blocks) > 0:
(m, d), block = blocks.pop(random.randrange(len(blocks)))
yield block
d_ = d + 1
block = get_block(m, d_)
if block is None:
continue
blocks.append(((m, d_), block))
def _merge_impl(
args: Args,
M: int,
pinfo: PartInfo,
get_block: Callable[[int, int], np.ndarray],
skip_output=False,
):
merge_fn = _dummy_merge if args.skip_sorting else sortlib.merge_partitions
merger = merge_fn(M, get_block)
if skip_output:
for datachunk in merger:
del datachunk
else:
with open(pinfo.path, "wb") as fout:
for datachunk in merger:
fout.write(datachunk)
return pinfo
# See worker_placement_groups() for why `num_cpus=0`.
@ray.remote(num_cpus=0, resources={"worker": 1})
@tracing_utils.timeit("merge")
def merge_mapper_blocks(
args: Args, reducer_id: PartId, mapper_id: PartId, *blocks: List[np.ndarray]
) -> PartInfo:
part_id = constants.merge_part_ids(reducer_id, mapper_id)
pinfo = _part_info(args, part_id, kind="temp")
M = len(blocks)
def get_block(i, d):
if i >= M or d > 0:
return None
return blocks[i]
return _merge_impl(args, M, pinfo, get_block)
# See worker_placement_groups() for why `num_cpus=0`.
@ray.remote(num_cpus=0, resources={"worker": 1})
@tracing_utils.timeit("reduce")
def final_merge(
args: Args, reducer_id: PartId, *merged_parts: List[PartInfo]
) -> PartInfo:
M = len(merged_parts)
def _load_block_chunk(pinfo: PartInfo, d: int) -> np.ndarray:
return np.fromfile(
pinfo.path,
dtype=np.uint8,
count=args.reducer_input_chunk,
offset=d * args.reducer_input_chunk,
)
def get_block(i, d):
ret = _load_block_chunk(merged_parts[i], d)
if ret.size == 0:
return None
return ret
pinfo = _part_info(args, reducer_id, "output")
return _merge_impl(args, M, pinfo, get_block, args.skip_output)
def _node_res(node: str) -> Dict[str, float]:
return {"resources": {f"node:{node}": 1e-3}}
@contextlib.contextmanager
def worker_placement_groups(args: Args) -> List[ray.PlacementGroupID]:
"""
Returns one placement group per node with a `worker` resource. To run
tasks in the placement group, use
`@ray.remote(num_cpus=0, resources={"worker": 1})`. Ray does not
automatically reserve CPU resources, so tasks must specify `num_cpus=0`
in order to run in a placement group.
"""
pgs = [ray.util.placement_group([{"worker": 1}]) for _ in range(args.num_reducers)]
ray.get([pg.ready() for pg in pgs])
try:
yield pgs
finally:
for pg in pgs:
ray.util.remove_placement_group(pg)
@tracing_utils.timeit("sort", report_time=True)
def sort_main(args: Args):
parts = _load_manifest(args, constants.INPUT_MANIFEST_FILE)
assert len(parts) == args.num_mappers
boundaries = sortlib.get_boundaries(args.num_reducers)
mapper_opt = {
"num_returns": args.num_reducers,
"num_cpus": os.cpu_count() / args.num_concurrent_rounds,
} # Load balance across worker nodes by setting `num_cpus`.
merge_results = np.empty((args.num_rounds, args.num_reducers), dtype=object)
part_id = 0
with worker_placement_groups(args) as pgs:
for round in range(args.num_rounds):
# Limit the number of in-flight rounds.
num_extra_rounds = round - args.num_concurrent_rounds + 1
if num_extra_rounds > 0:
ray.wait(
[f for f in merge_results.flatten() if f is not None],
num_returns=num_extra_rounds * args.num_reducers,
)
# Submit map tasks.
mapper_results = np.empty(
(args.num_mappers_per_round, args.num_reducers), dtype=object
)
for _ in range(args.num_mappers_per_round):
_, node, path = parts[part_id]
m = part_id % args.num_mappers_per_round
mapper_results[m, :] = mapper.options(**mapper_opt).remote(
args, part_id, boundaries, path
)
part_id += 1
# Submit merge tasks.
merge_results[round, :] = [
merge_mapper_blocks.options(placement_group=pgs[r]).remote(
args, r, round, *mapper_results[:, r].tolist()
)
for r in range(args.num_reducers)
]
# Delete local references to mapper results.
mapper_results = None
# Submit second-stage reduce tasks.
reducer_results = [
final_merge.options(placement_group=pgs[r]).remote(
args, r, *merge_results[:, r].tolist()
)
for r in range(args.num_reducers)
]
reducer_results = ray.get(reducer_results)
if not args.skip_output:
with open(constants.OUTPUT_MANIFEST_FILE, "w") as fout:
writer = csv.writer(fout)
writer.writerows(reducer_results)
logging.info(ray.internal.internal_api.memory_summary(stats_only=True))
# ------------------------------------------------------------
# Validate Output
# ------------------------------------------------------------
def _run_valsort(args: List[str]):
proc = subprocess.run([constants.VALSORT_PATH] + args, capture_output=True)
if proc.returncode != 0:
logging.critical("\n" + proc.stderr.decode("ascii"))
raise RuntimeError(f"Validation failed: {args}")
@ray.remote
def validate_part(path: Path):
logging_utils.init()
sum_path = path + ".sum"
_run_valsort(["-o", sum_path, path])
logging.info(f"Validated output {path}")
with open(sum_path, "rb") as fin:
return os.path.getsize(path), fin.read()
def validate_output(args: Args):
if args.skip_sorting or args.skip_output:
return
partitions = _load_manifest(args, constants.OUTPUT_MANIFEST_FILE)
results = []
for _, node, path in partitions:
results.append(validate_part.options(**_node_res(node)).remote(path))
logging.info(f"Validating {len(results)} partitions")
results = ray.get(results)
total = sum(s for s, _ in results)
assert total == args.total_data_size, total - args.total_data_size
all_checksum = b"".join(c for _, c in results)
with tempfile.NamedTemporaryFile() as fout:
fout.write(all_checksum)
fout.flush()
_run_valsort(["-s", fout.name])
logging.info("All OK!")
# ------------------------------------------------------------
# Main
# ------------------------------------------------------------
def init(args: Args):
if not args.ray_address:
ray.init(resources={"worker": os.cpu_count()})
else:
ray.init(address=args.ray_address)
logging_utils.init()
logging.info(args)
os.makedirs(constants.WORK_DIR, exist_ok=True)
resources = ray.cluster_resources()
logging.info(resources)
args.num_workers = resources["worker"]
progress_tracker = tracing_utils.create_progress_tracker(args)
return progress_tracker
def main(args: Args):
# Keep the actor handle in scope for the duration of the program.
_progress_tracker = init(args) # noqa F841
if args.generate_input:
generate_input(args)
if args.sort:
sort_main(args)
if args.validate_output:
validate_output(args)
if __name__ == "__main__":
main(get_args())
|
[
"numpy.fromfile",
"numpy.random.bytes",
"ray.cluster_resources",
"ray.experimental.raysort.types.PartInfo",
"os.cpu_count",
"ray.init",
"logging.info",
"os.path.exists",
"os.listdir",
"argparse.ArgumentParser",
"ray.experimental.raysort.logging_utils.init",
"subprocess.run",
"ray.util.remove_placement_group",
"ray.experimental.raysort.tracing_utils.create_progress_tracker",
"numpy.empty",
"ray.experimental.raysort.constants.bytes_to_records",
"tempfile.NamedTemporaryFile",
"ray.remote",
"csv.reader",
"ray.experimental.raysort.types.ByteCount",
"numpy.ceil",
"random.choice",
"os.path.getsize",
"ray.get",
"csv.writer",
"ray.experimental.raysort.tracing_utils.timeit",
"ray.internal.internal_api.memory_summary",
"ray.experimental.raysort.sortlib.get_boundaries",
"ray.util.placement_group",
"os.makedirs",
"os.path.join",
"tempfile.gettempdir",
"ray.experimental.raysort.constants.merge_part_ids"
] |
[((6260, 6287), 'ray.experimental.raysort.tracing_utils.timeit', 'tracing_utils.timeit', (['"""map"""'], {}), "('map')\n", (6280, 6287), False, 'from ray.experimental.raysort import tracing_utils\n'), ((7656, 7703), 'ray.remote', 'ray.remote', ([], {'num_cpus': '(0)', 'resources': "{'worker': 1}"}), "(num_cpus=0, resources={'worker': 1})\n", (7666, 7703), False, 'import ray\n'), ((7705, 7734), 'ray.experimental.raysort.tracing_utils.timeit', 'tracing_utils.timeit', (['"""merge"""'], {}), "('merge')\n", (7725, 7734), False, 'from ray.experimental.raysort import tracing_utils\n'), ((8200, 8247), 'ray.remote', 'ray.remote', ([], {'num_cpus': '(0)', 'resources': "{'worker': 1}"}), "(num_cpus=0, resources={'worker': 1})\n", (8210, 8247), False, 'import ray\n'), ((8249, 8279), 'ray.experimental.raysort.tracing_utils.timeit', 'tracing_utils.timeit', (['"""reduce"""'], {}), "('reduce')\n", (8269, 8279), False, 'from ray.experimental.raysort import tracing_utils\n'), ((9675, 9721), 'ray.experimental.raysort.tracing_utils.timeit', 'tracing_utils.timeit', (['"""sort"""'], {'report_time': '(True)'}), "('sort', report_time=True)\n", (9695, 9721), False, 'from ray.experimental.raysort import tracing_utils\n'), ((746, 771), 'argparse.ArgumentParser', 'argparse.ArgumentParser', ([], {}), '()\n', (769, 771), False, 'import argparse\n'), ((2819, 2869), 'ray.experimental.raysort.types.ByteCount', 'ByteCount', (['(args.total_data_size / args.num_mappers)'], {}), '(args.total_data_size / args.num_mappers)\n', (2828, 2869), False, 'from ray.experimental.raysort.types import BlockInfo, ByteCount, RecordCount, PartId, PartInfo, Path\n'), ((3346, 3365), 'os.path.exists', 'os.path.exists', (['mnt'], {}), '(mnt)\n', (3360, 3365), False, 'import os\n'), ((3784, 3816), 'random.choice', 'random.choice', (['args.mount_points'], {}), '(args.mount_points)\n', (3797, 3816), False, 'import random\n'), ((3878, 3911), 'ray.experimental.raysort.types.PartInfo', 'PartInfo', (['part_id', 'node', 'filepath'], {}), '(part_id, node, filepath)\n', (3886, 3911), False, 'from ray.experimental.raysort.types import BlockInfo, ByteCount, RecordCount, PartId, PartInfo, Path\n'), ((4116, 4151), 'os.makedirs', 'os.makedirs', (['dirpath'], {'exist_ok': '(True)'}), '(dirpath, exist_ok=True)\n', (4127, 4151), False, 'import os\n'), ((4267, 4298), 'os.path.join', 'os.path.join', (['dirpath', 'filename'], {}), '(dirpath, filename)\n', (4279, 4298), False, 'import os\n'), ((4443, 4463), 'ray.experimental.raysort.logging_utils.init', 'logging_utils.init', ([], {}), '()\n', (4461, 4463), False, 'from ray.experimental.raysort import logging_utils\n'), ((4506, 4601), 'subprocess.run', 'subprocess.run', (["[constants.GENSORT_PATH, f'-b{offset}', f'{size}', pinfo.path]"], {'check': '(True)'}), "([constants.GENSORT_PATH, f'-b{offset}', f'{size}', pinfo.\n path], check=True)\n", (4520, 4601), False, 'import subprocess\n'), ((4615, 4655), 'logging.info', 'logging.info', (['f"""Generated input {pinfo}"""'], {}), "(f'Generated input {pinfo}')\n", (4627, 4655), False, 'import logging\n'), ((4757, 4805), 'ray.experimental.raysort.constants.bytes_to_records', 'constants.bytes_to_records', (['args.input_part_size'], {}), '(args.input_part_size)\n', (4783, 4805), False, 'from ray.experimental.raysort import constants\n'), ((5119, 5133), 'ray.get', 'ray.get', (['tasks'], {}), '(tasks)\n', (5126, 5133), False, 'import ray\n'), ((5916, 5949), 'numpy.fromfile', 'np.fromfile', (['path'], {'dtype': 'np.uint8'}), '(path, dtype=np.uint8)\n', (5927, 5949), True, 'import numpy as np\n'), ((6396, 6416), 'ray.experimental.raysort.logging_utils.init', 'logging_utils.init', ([], {}), '()\n', (6414, 6416), False, 'from ray.experimental.raysort import logging_utils\n'), ((7870, 7917), 'ray.experimental.raysort.constants.merge_part_ids', 'constants.merge_part_ids', (['reducer_id', 'mapper_id'], {}), '(reducer_id, mapper_id)\n', (7894, 7917), False, 'from ray.experimental.raysort import constants\n'), ((9872, 9913), 'ray.experimental.raysort.sortlib.get_boundaries', 'sortlib.get_boundaries', (['args.num_reducers'], {}), '(args.num_reducers)\n', (9894, 9913), False, 'from ray.experimental.raysort import sortlib\n'), ((10126, 10186), 'numpy.empty', 'np.empty', (['(args.num_rounds, args.num_reducers)'], {'dtype': 'object'}), '((args.num_rounds, args.num_reducers), dtype=object)\n', (10134, 10186), True, 'import numpy as np\n'), ((12301, 12369), 'subprocess.run', 'subprocess.run', (['([constants.VALSORT_PATH] + args)'], {'capture_output': '(True)'}), '([constants.VALSORT_PATH] + args, capture_output=True)\n', (12315, 12369), False, 'import subprocess\n'), ((12566, 12586), 'ray.experimental.raysort.logging_utils.init', 'logging_utils.init', ([], {}), '()\n', (12584, 12586), False, 'from ray.experimental.raysort import logging_utils\n'), ((12661, 12701), 'logging.info', 'logging.info', (['f"""Validated output {path}"""'], {}), "(f'Validated output {path}')\n", (12673, 12701), False, 'import logging\n'), ((13159, 13175), 'ray.get', 'ray.get', (['results'], {}), '(results)\n', (13166, 13175), False, 'import ray\n'), ((13483, 13506), 'logging.info', 'logging.info', (['"""All OK!"""'], {}), "('All OK!')\n", (13495, 13506), False, 'import logging\n'), ((13811, 13831), 'ray.experimental.raysort.logging_utils.init', 'logging_utils.init', ([], {}), '()\n', (13829, 13831), False, 'from ray.experimental.raysort import logging_utils\n'), ((13836, 13854), 'logging.info', 'logging.info', (['args'], {}), '(args)\n', (13848, 13854), False, 'import logging\n'), ((13859, 13905), 'os.makedirs', 'os.makedirs', (['constants.WORK_DIR'], {'exist_ok': '(True)'}), '(constants.WORK_DIR, exist_ok=True)\n', (13870, 13905), False, 'import os\n'), ((13922, 13945), 'ray.cluster_resources', 'ray.cluster_resources', ([], {}), '()\n', (13943, 13945), False, 'import ray\n'), ((13950, 13973), 'logging.info', 'logging.info', (['resources'], {}), '(resources)\n', (13962, 13973), False, 'import logging\n'), ((14040, 14083), 'ray.experimental.raysort.tracing_utils.create_progress_tracker', 'tracing_utils.create_progress_tracker', (['args'], {}), '(args)\n', (14077, 14083), False, 'from ray.experimental.raysort import tracing_utils\n'), ((3299, 3320), 'tempfile.gettempdir', 'tempfile.gettempdir', ([], {}), '()\n', (3318, 3320), False, 'import tempfile\n'), ((4996, 5044), 'ray.experimental.raysort.constants.bytes_to_records', 'constants.bytes_to_records', (['args.total_data_size'], {}), '(args.total_data_size)\n', (5022, 5044), False, 'from ray.experimental.raysort import constants\n'), ((5210, 5226), 'csv.writer', 'csv.writer', (['fout'], {}), '(fout)\n', (5220, 5226), False, 'import csv\n'), ((5605, 5620), 'csv.reader', 'csv.reader', (['fin'], {}), '(fin)\n', (5615, 5620), False, 'import csv\n'), ((6103, 6125), 'numpy.ceil', 'np.ceil', (['(part.size / N)'], {}), '(part.size / N)\n', (6110, 6125), True, 'import numpy as np\n'), ((8486, 8598), 'numpy.fromfile', 'np.fromfile', (['pinfo.path'], {'dtype': 'np.uint8', 'count': 'args.reducer_input_chunk', 'offset': '(d * args.reducer_input_chunk)'}), '(pinfo.path, dtype=np.uint8, count=args.reducer_input_chunk,\n offset=d * args.reducer_input_chunk)\n', (8497, 8598), True, 'import numpy as np\n'), ((9444, 9485), 'ray.util.placement_group', 'ray.util.placement_group', (["[{'worker': 1}]"], {}), "([{'worker': 1}])\n", (9468, 9485), False, 'import ray\n'), ((11823, 11847), 'ray.get', 'ray.get', (['reducer_results'], {}), '(reducer_results)\n', (11830, 11847), False, 'import ray\n'), ((12044, 12101), 'ray.internal.internal_api.memory_summary', 'ray.internal.internal_api.memory_summary', ([], {'stats_only': '(True)'}), '(stats_only=True)\n', (12084, 12101), False, 'import ray\n'), ((13346, 13375), 'tempfile.NamedTemporaryFile', 'tempfile.NamedTemporaryFile', ([], {}), '()\n', (13373, 13375), False, 'import tempfile\n'), ((13772, 13806), 'ray.init', 'ray.init', ([], {'address': 'args.ray_address'}), '(address=args.ray_address)\n', (13780, 13806), False, 'import ray\n'), ((3382, 3402), 'os.path.join', 'os.path.join', (['mnt', 'd'], {}), '(mnt, d)\n', (3394, 3402), False, 'import os\n'), ((5502, 5525), 'ray.experimental.raysort.types.PartInfo', 'PartInfo', (['i', 'None', 'None'], {}), '(i, None, None)\n', (5510, 5525), False, 'from ray.experimental.raysort.types import BlockInfo, ByteCount, RecordCount, PartId, PartInfo, Path\n'), ((9636, 9671), 'ray.util.remove_placement_group', 'ray.util.remove_placement_group', (['pg'], {}), '(pg)\n', (9667, 9671), False, 'import ray\n'), ((9996, 10010), 'os.cpu_count', 'os.cpu_count', ([], {}), '()\n', (10008, 10010), False, 'import os\n'), ((10706, 10777), 'numpy.empty', 'np.empty', (['(args.num_mappers_per_round, args.num_reducers)'], {'dtype': 'object'}), '((args.num_mappers_per_round, args.num_reducers), dtype=object)\n', (10714, 10777), True, 'import numpy as np\n'), ((11963, 11979), 'csv.writer', 'csv.writer', (['fout'], {}), '(fout)\n', (11973, 11979), False, 'import csv\n'), ((12755, 12776), 'os.path.getsize', 'os.path.getsize', (['path'], {}), '(path)\n', (12770, 12776), False, 'import os\n'), ((3412, 3427), 'os.listdir', 'os.listdir', (['mnt'], {}), '(mnt)\n', (3422, 3427), False, 'import os\n'), ((5834, 5871), 'numpy.random.bytes', 'np.random.bytes', (['args.input_part_size'], {}), '(args.input_part_size)\n', (5849, 5871), True, 'import numpy as np\n'), ((13737, 13751), 'os.cpu_count', 'os.cpu_count', ([], {}), '()\n', (13749, 13751), False, 'import os\n')]
|
__author__ = 'sibirrer'
from lenstronomy.LensModel.Profiles.flexion import Flexion
from lenstronomy.LensModel.lens_model import LensModel
import numpy as np
import numpy.testing as npt
import pytest
class TestExternalShear(object):
"""
tests the Gaussian methods
"""
def setup(self):
self.flex = Flexion()
g1, g2, g3, g4 = 0.01, 0.02, 0.03, 0.04
self.kwargs_lens = {'g1': g1, 'g2': g2, 'g3': g3, 'g4': g4}
def test_function(self):
x = np.array([1])
y = np.array([2])
values = self.flex.function(x, y, **self.kwargs_lens)
npt.assert_almost_equal(values[0], 0.135, decimal=5)
x = np.array([0])
y = np.array([0])
values = self.flex.function(x, y, **self.kwargs_lens)
npt.assert_almost_equal(values[0], 0, decimal=5)
x = np.array([2, 3, 4])
y = np.array([1, 1, 1])
values = self.flex.function(x, y, **self.kwargs_lens)
npt.assert_almost_equal(values[0], 0.09, decimal=5)
npt.assert_almost_equal(values[1], 0.18666666666666668, decimal=5)
def test_derivatives(self):
x = np.array([1])
y = np.array([2])
f_x, f_y = self.flex.derivatives(x, y, **self.kwargs_lens)
npt.assert_almost_equal(f_x[0], 0.105, decimal=5)
npt.assert_almost_equal(f_y[0], 0.15, decimal=5)
x = np.array([1, 3, 4])
y = np.array([2, 1, 1])
values = self.flex.derivatives(x, y, **self.kwargs_lens)
npt.assert_almost_equal(values[0][0], 0.105, decimal=5)
npt.assert_almost_equal(values[1][0], 0.15, decimal=5)
def test_hessian(self):
x = np.array(1)
y = np.array(2)
f_xx, f_xy, f_yx, f_yy = self.flex.hessian(x, y, **self.kwargs_lens)
npt.assert_almost_equal(f_xx, 0.05, decimal=5)
npt.assert_almost_equal(f_yy, 0.11, decimal=5)
npt.assert_almost_equal(f_xy, 0.08, decimal=5)
npt.assert_almost_equal(f_xy, f_yx, decimal=8)
x = np.array([1,3,4])
y = np.array([2,1,1])
values = self.flex.hessian(x, y, **self.kwargs_lens)
npt.assert_almost_equal(values[0][0], 0.05, decimal=5)
npt.assert_almost_equal(values[3][0], 0.11, decimal=5)
npt.assert_almost_equal(values[2][0], 0.08, decimal=5)
npt.assert_almost_equal(values[1][0], 0.08, decimal=5)
def test_flexion(self):
x = np.array(0)
y = np.array(2)
flex = LensModel(['FLEXION'])
f_xxx, f_xxy, f_xyy, f_yyy = flex.flexion(x, y, [self.kwargs_lens])
npt.assert_almost_equal(f_xxx, self.kwargs_lens['g1'], decimal=9)
npt.assert_almost_equal(f_xxy, self.kwargs_lens['g2'], decimal=9)
npt.assert_almost_equal(f_xyy, self.kwargs_lens['g3'], decimal=9)
npt.assert_almost_equal(f_yyy, self.kwargs_lens['g4'], decimal=9)
def test_magnification(self):
ra_0, dec_0 = 1, -1
flex = LensModel(['FLEXION'])
g1, g2, g3, g4 = 0.01, 0.02, 0.03, 0.04
kwargs = {'g1': g1, 'g2': g2, 'g3': g3, 'g4': g4, 'ra_0': ra_0, 'dec_0': dec_0}
mag = flex.magnification(ra_0, dec_0, [kwargs])
npt.assert_almost_equal(mag, 1, decimal=8)
if __name__ == '__main__':
pytest.main()
|
[
"lenstronomy.LensModel.lens_model.LensModel",
"pytest.main",
"numpy.testing.assert_almost_equal",
"numpy.array",
"lenstronomy.LensModel.Profiles.flexion.Flexion"
] |
[((3229, 3242), 'pytest.main', 'pytest.main', ([], {}), '()\n', (3240, 3242), False, 'import pytest\n'), ((325, 334), 'lenstronomy.LensModel.Profiles.flexion.Flexion', 'Flexion', ([], {}), '()\n', (332, 334), False, 'from lenstronomy.LensModel.Profiles.flexion import Flexion\n'), ((494, 507), 'numpy.array', 'np.array', (['[1]'], {}), '([1])\n', (502, 507), True, 'import numpy as np\n'), ((520, 533), 'numpy.array', 'np.array', (['[2]'], {}), '([2])\n', (528, 533), True, 'import numpy as np\n'), ((604, 656), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['values[0]', '(0.135)'], {'decimal': '(5)'}), '(values[0], 0.135, decimal=5)\n', (627, 656), True, 'import numpy.testing as npt\n'), ((669, 682), 'numpy.array', 'np.array', (['[0]'], {}), '([0])\n', (677, 682), True, 'import numpy as np\n'), ((695, 708), 'numpy.array', 'np.array', (['[0]'], {}), '([0])\n', (703, 708), True, 'import numpy as np\n'), ((779, 827), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['values[0]', '(0)'], {'decimal': '(5)'}), '(values[0], 0, decimal=5)\n', (802, 827), True, 'import numpy.testing as npt\n'), ((841, 860), 'numpy.array', 'np.array', (['[2, 3, 4]'], {}), '([2, 3, 4])\n', (849, 860), True, 'import numpy as np\n'), ((873, 892), 'numpy.array', 'np.array', (['[1, 1, 1]'], {}), '([1, 1, 1])\n', (881, 892), True, 'import numpy as np\n'), ((963, 1014), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['values[0]', '(0.09)'], {'decimal': '(5)'}), '(values[0], 0.09, decimal=5)\n', (986, 1014), True, 'import numpy.testing as npt\n'), ((1024, 1090), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['values[1]', '(0.18666666666666668)'], {'decimal': '(5)'}), '(values[1], 0.18666666666666668, decimal=5)\n', (1047, 1090), True, 'import numpy.testing as npt\n'), ((1136, 1149), 'numpy.array', 'np.array', (['[1]'], {}), '([1])\n', (1144, 1149), True, 'import numpy as np\n'), ((1162, 1175), 'numpy.array', 'np.array', (['[2]'], {}), '([2])\n', (1170, 1175), True, 'import numpy as np\n'), ((1251, 1300), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['f_x[0]', '(0.105)'], {'decimal': '(5)'}), '(f_x[0], 0.105, decimal=5)\n', (1274, 1300), True, 'import numpy.testing as npt\n'), ((1309, 1357), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['f_y[0]', '(0.15)'], {'decimal': '(5)'}), '(f_y[0], 0.15, decimal=5)\n', (1332, 1357), True, 'import numpy.testing as npt\n'), ((1371, 1390), 'numpy.array', 'np.array', (['[1, 3, 4]'], {}), '([1, 3, 4])\n', (1379, 1390), True, 'import numpy as np\n'), ((1403, 1422), 'numpy.array', 'np.array', (['[2, 1, 1]'], {}), '([2, 1, 1])\n', (1411, 1422), True, 'import numpy as np\n'), ((1496, 1551), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['values[0][0]', '(0.105)'], {'decimal': '(5)'}), '(values[0][0], 0.105, decimal=5)\n', (1519, 1551), True, 'import numpy.testing as npt\n'), ((1560, 1614), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['values[1][0]', '(0.15)'], {'decimal': '(5)'}), '(values[1][0], 0.15, decimal=5)\n', (1583, 1614), True, 'import numpy.testing as npt\n'), ((1656, 1667), 'numpy.array', 'np.array', (['(1)'], {}), '(1)\n', (1664, 1667), True, 'import numpy as np\n'), ((1680, 1691), 'numpy.array', 'np.array', (['(2)'], {}), '(2)\n', (1688, 1691), True, 'import numpy as np\n'), ((1777, 1823), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['f_xx', '(0.05)'], {'decimal': '(5)'}), '(f_xx, 0.05, decimal=5)\n', (1800, 1823), True, 'import numpy.testing as npt\n'), ((1832, 1878), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['f_yy', '(0.11)'], {'decimal': '(5)'}), '(f_yy, 0.11, decimal=5)\n', (1855, 1878), True, 'import numpy.testing as npt\n'), ((1887, 1933), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['f_xy', '(0.08)'], {'decimal': '(5)'}), '(f_xy, 0.08, decimal=5)\n', (1910, 1933), True, 'import numpy.testing as npt\n'), ((1942, 1988), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['f_xy', 'f_yx'], {'decimal': '(8)'}), '(f_xy, f_yx, decimal=8)\n', (1965, 1988), True, 'import numpy.testing as npt\n'), ((2002, 2021), 'numpy.array', 'np.array', (['[1, 3, 4]'], {}), '([1, 3, 4])\n', (2010, 2021), True, 'import numpy as np\n'), ((2032, 2051), 'numpy.array', 'np.array', (['[2, 1, 1]'], {}), '([2, 1, 1])\n', (2040, 2051), True, 'import numpy as np\n'), ((2119, 2173), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['values[0][0]', '(0.05)'], {'decimal': '(5)'}), '(values[0][0], 0.05, decimal=5)\n', (2142, 2173), True, 'import numpy.testing as npt\n'), ((2182, 2236), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['values[3][0]', '(0.11)'], {'decimal': '(5)'}), '(values[3][0], 0.11, decimal=5)\n', (2205, 2236), True, 'import numpy.testing as npt\n'), ((2245, 2299), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['values[2][0]', '(0.08)'], {'decimal': '(5)'}), '(values[2][0], 0.08, decimal=5)\n', (2268, 2299), True, 'import numpy.testing as npt\n'), ((2308, 2362), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['values[1][0]', '(0.08)'], {'decimal': '(5)'}), '(values[1][0], 0.08, decimal=5)\n', (2331, 2362), True, 'import numpy.testing as npt\n'), ((2404, 2415), 'numpy.array', 'np.array', (['(0)'], {}), '(0)\n', (2412, 2415), True, 'import numpy as np\n'), ((2428, 2439), 'numpy.array', 'np.array', (['(2)'], {}), '(2)\n', (2436, 2439), True, 'import numpy as np\n'), ((2455, 2477), 'lenstronomy.LensModel.lens_model.LensModel', 'LensModel', (["['FLEXION']"], {}), "(['FLEXION'])\n", (2464, 2477), False, 'from lenstronomy.LensModel.lens_model import LensModel\n'), ((2562, 2627), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['f_xxx', "self.kwargs_lens['g1']"], {'decimal': '(9)'}), "(f_xxx, self.kwargs_lens['g1'], decimal=9)\n", (2585, 2627), True, 'import numpy.testing as npt\n'), ((2636, 2701), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['f_xxy', "self.kwargs_lens['g2']"], {'decimal': '(9)'}), "(f_xxy, self.kwargs_lens['g2'], decimal=9)\n", (2659, 2701), True, 'import numpy.testing as npt\n'), ((2710, 2775), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['f_xyy', "self.kwargs_lens['g3']"], {'decimal': '(9)'}), "(f_xyy, self.kwargs_lens['g3'], decimal=9)\n", (2733, 2775), True, 'import numpy.testing as npt\n'), ((2784, 2849), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['f_yyy', "self.kwargs_lens['g4']"], {'decimal': '(9)'}), "(f_yyy, self.kwargs_lens['g4'], decimal=9)\n", (2807, 2849), True, 'import numpy.testing as npt\n'), ((2929, 2951), 'lenstronomy.LensModel.lens_model.LensModel', 'LensModel', (["['FLEXION']"], {}), "(['FLEXION'])\n", (2938, 2951), False, 'from lenstronomy.LensModel.lens_model import LensModel\n'), ((3153, 3195), 'numpy.testing.assert_almost_equal', 'npt.assert_almost_equal', (['mag', '(1)'], {'decimal': '(8)'}), '(mag, 1, decimal=8)\n', (3176, 3195), True, 'import numpy.testing as npt\n')]
|
from unittest.mock import Mock, PropertyMock, MagicMock, patch
import numpy as np
import gym_connect4
from test_fixtures import Connect4Task
import regym
from regym.environments import EnvType
from regym.rl_algorithms import build_MCTS_Agent
from regym.rl_algorithms.agents import Agent, build_Deterministic_Agent, DeterministicAgent
from regym.rl_loops import Trajectory
from regym.rl_algorithms import build_Deterministic_Agent, build_MCTS_Agent
from regym.rl_loops.multiagent_loops.vectorenv_sequential_action_rl_loop import async_run_episode
from regym.rl_loops.multiagent_loops.sequential_action_rl_loop import propagate_experience, propagate_last_experience
def test_sequential_trajectories_feature_agent_predictions_single_env(Connect4Task):
agent_1 = build_Deterministic_Agent(
Connect4Task, {'action': 0}, 'Col-0-DeterministicAgent')
agent_1.requires_opponents_prediction = True # Required!
agent_2 = build_Deterministic_Agent(
Connect4Task, {'action': 1}, 'Col-0-DeterministicAgent')
trajectory = Connect4Task.run_episode([agent_1, agent_2], training=False)
expected_prediction_1 = {'a': 0, 'probs': [[1., 0., 0., 0., 0., 0., 0.]]}
expected_prediction_2 = {'a': 1, 'probs': [[0., 1., 0., 0., 0., 0., 0.]]}
expected_predictions = [expected_prediction_1,
expected_prediction_2]
compare_trajectory_extra_info_against_expected(trajectory, expected_predictions)
def test_sequential_trajectories_feature_agent_predictions_multienv(Connect4Task):
agent_1 = build_Deterministic_Agent(
Connect4Task, {'action': 0}, 'Col-0-DeterministicAgent')
agent_1.requires_opponents_prediction = True # Required!
agent_2 = build_Deterministic_Agent(
Connect4Task, {'action': 1}, 'Col-0-DeterministicAgent')
trajectories = Connect4Task.run_episodes([agent_1, agent_2], training=False,
num_envs=2, num_episodes=2)
# on single agents there's a batch dimension in 'probs', but not
# on multiagent_loops. Does this matter?
expected_prediction_1 = {'a': 0, 'probs': [1., 0., 0., 0., 0., 0., 0.]}
expected_prediction_2 = {'a': 1, 'probs': [0., 1., 0., 0., 0., 0., 0.]}
expected_predictions = [expected_prediction_1, expected_prediction_2]
for trajectory in trajectories:
compare_trajectory_extra_info_against_expected(trajectory, expected_predictions)
def test_agents_in_sequential_environments_handle_experiences_with_extra_info_single_env(Connect4Task):
'''
In this test we want to ensure that when agents process experiences
via `Agent.handle_experience(...)` calls, they obtain the are able
to observe the `predicions` of other agents.
There are 2 cases to consider:
- Handling an experience in the middle of a trajectory
- Handling an experience when the episode just finshed and some agents
need to process the last (terminal) timestep
'''
mock_agent_1 = Mock(spec=DeterministicAgent)
mock_agent_2 = Mock(spec=DeterministicAgent)
agent_vector = [mock_agent_1, mock_agent_2]
mock_agent_1.requires_opponents_prediction = True
mock_agent_1.training = True
mock_agent_2.requires_opponents_prediction = False
mock_agent_2.training = True
prediction_1 = {'a': 0, 'probs': [1., 0., 0., 0., 0., 0., 0.]}
prediction_2 = {'a': 1, 'probs': [0., 1., 0., 0., 0., 0., 0.]}
predictions = [prediction_1, prediction_2]
'''
Creates a trajectory for the game of Connect4 looks like this
Total timesteps 7. P1 (x) actions: 4. P2 (o) actions: 3.
Board: | |
| |
|x |
|x o |
|x o |
|x o . . . . . |
|--------------|
'''
sample_trajectory = Trajectory(
env_type=EnvType.MULTIAGENT_SEQUENTIAL_ACTION, num_agents=2)
o, a, r, succ_o, done = [None, None], None, [0, 0], [None, None], False
sample_trajectory.add_timestep(
o, a, r, succ_o, done, acting_agents=[0],
extra_info={0: predictions[0]})
sample_trajectory.add_timestep(
o, a, r, succ_o, done, acting_agents=[1],
extra_info={1: predictions[1]})
# Update agent 0
propagate_experience(agent_vector, sample_trajectory)
sample_trajectory.add_timestep(
o, a, r, succ_o, done, acting_agents=[0],
extra_info={0: predictions[0]})
# Update agent 1
propagate_experience(agent_vector, sample_trajectory)
sample_trajectory.add_timestep(
o, a, r, succ_o, done, acting_agents=[1],
extra_info={1: predictions[1]})
# Update agent 0
propagate_experience(agent_vector, sample_trajectory)
sample_trajectory.add_timestep(
o, a, r, succ_o, done, acting_agents=[0],
extra_info={0: predictions[0]})
# Update agent 1
propagate_experience(agent_vector, sample_trajectory)
sample_trajectory.add_timestep(
o, a, r, succ_o, done, acting_agents=[1],
extra_info={1: predictions[1]})
# Update agent 0
propagate_experience(agent_vector, sample_trajectory)
done = True
sample_trajectory.add_timestep(
o, a, [1, -1], succ_o, done, acting_agents=[0],
extra_info={0: predictions[0]})
# Update player 1
propagate_experience(agent_vector, sample_trajectory)
# Episode termination
# Update player 0 (After done flag)
propagate_last_experience(agent_vector, sample_trajectory)
def compare_trajectory_extra_info_against_expected(trajectory, expected_predictions):
for timestep in trajectory:
# Only one agent acts at a time in Connect4
a_i = timestep.acting_agents[0]
actual_prediction = timestep.extra_info[a_i]
assert 'a' in actual_prediction
assert 'probs' in actual_prediction
assert actual_prediction['a'] == expected_predictions[a_i]['a']
np.testing.assert_array_equal(
actual_prediction['probs'], expected_predictions[a_i]['probs'])
|
[
"unittest.mock.Mock",
"regym.rl_algorithms.build_Deterministic_Agent",
"regym.rl_loops.Trajectory",
"test_fixtures.Connect4Task.run_episodes",
"regym.rl_loops.multiagent_loops.sequential_action_rl_loop.propagate_last_experience",
"regym.rl_loops.multiagent_loops.sequential_action_rl_loop.propagate_experience",
"numpy.testing.assert_array_equal",
"test_fixtures.Connect4Task.run_episode"
] |
[((771, 857), 'regym.rl_algorithms.build_Deterministic_Agent', 'build_Deterministic_Agent', (['Connect4Task', "{'action': 0}", '"""Col-0-DeterministicAgent"""'], {}), "(Connect4Task, {'action': 0},\n 'Col-0-DeterministicAgent')\n", (796, 857), False, 'from regym.rl_algorithms import build_Deterministic_Agent, build_MCTS_Agent\n'), ((939, 1025), 'regym.rl_algorithms.build_Deterministic_Agent', 'build_Deterministic_Agent', (['Connect4Task', "{'action': 1}", '"""Col-0-DeterministicAgent"""'], {}), "(Connect4Task, {'action': 1},\n 'Col-0-DeterministicAgent')\n", (964, 1025), False, 'from regym.rl_algorithms import build_Deterministic_Agent, build_MCTS_Agent\n'), ((1049, 1109), 'test_fixtures.Connect4Task.run_episode', 'Connect4Task.run_episode', (['[agent_1, agent_2]'], {'training': '(False)'}), '([agent_1, agent_2], training=False)\n', (1073, 1109), False, 'from test_fixtures import Connect4Task\n'), ((1554, 1640), 'regym.rl_algorithms.build_Deterministic_Agent', 'build_Deterministic_Agent', (['Connect4Task', "{'action': 0}", '"""Col-0-DeterministicAgent"""'], {}), "(Connect4Task, {'action': 0},\n 'Col-0-DeterministicAgent')\n", (1579, 1640), False, 'from regym.rl_algorithms import build_Deterministic_Agent, build_MCTS_Agent\n'), ((1722, 1808), 'regym.rl_algorithms.build_Deterministic_Agent', 'build_Deterministic_Agent', (['Connect4Task', "{'action': 1}", '"""Col-0-DeterministicAgent"""'], {}), "(Connect4Task, {'action': 1},\n 'Col-0-DeterministicAgent')\n", (1747, 1808), False, 'from regym.rl_algorithms import build_Deterministic_Agent, build_MCTS_Agent\n'), ((1834, 1927), 'test_fixtures.Connect4Task.run_episodes', 'Connect4Task.run_episodes', (['[agent_1, agent_2]'], {'training': '(False)', 'num_envs': '(2)', 'num_episodes': '(2)'}), '([agent_1, agent_2], training=False, num_envs=2,\n num_episodes=2)\n', (1859, 1927), False, 'from test_fixtures import Connect4Task\n'), ((3002, 3031), 'unittest.mock.Mock', 'Mock', ([], {'spec': 'DeterministicAgent'}), '(spec=DeterministicAgent)\n', (3006, 3031), False, 'from unittest.mock import Mock, PropertyMock, MagicMock, patch\n'), ((3051, 3080), 'unittest.mock.Mock', 'Mock', ([], {'spec': 'DeterministicAgent'}), '(spec=DeterministicAgent)\n', (3055, 3080), False, 'from unittest.mock import Mock, PropertyMock, MagicMock, patch\n'), ((3885, 3956), 'regym.rl_loops.Trajectory', 'Trajectory', ([], {'env_type': 'EnvType.MULTIAGENT_SEQUENTIAL_ACTION', 'num_agents': '(2)'}), '(env_type=EnvType.MULTIAGENT_SEQUENTIAL_ACTION, num_agents=2)\n', (3895, 3956), False, 'from regym.rl_loops import Trajectory\n'), ((4323, 4376), 'regym.rl_loops.multiagent_loops.sequential_action_rl_loop.propagate_experience', 'propagate_experience', (['agent_vector', 'sample_trajectory'], {}), '(agent_vector, sample_trajectory)\n', (4343, 4376), False, 'from regym.rl_loops.multiagent_loops.sequential_action_rl_loop import propagate_experience, propagate_last_experience\n'), ((4530, 4583), 'regym.rl_loops.multiagent_loops.sequential_action_rl_loop.propagate_experience', 'propagate_experience', (['agent_vector', 'sample_trajectory'], {}), '(agent_vector, sample_trajectory)\n', (4550, 4583), False, 'from regym.rl_loops.multiagent_loops.sequential_action_rl_loop import propagate_experience, propagate_last_experience\n'), ((4737, 4790), 'regym.rl_loops.multiagent_loops.sequential_action_rl_loop.propagate_experience', 'propagate_experience', (['agent_vector', 'sample_trajectory'], {}), '(agent_vector, sample_trajectory)\n', (4757, 4790), False, 'from regym.rl_loops.multiagent_loops.sequential_action_rl_loop import propagate_experience, propagate_last_experience\n'), ((4944, 4997), 'regym.rl_loops.multiagent_loops.sequential_action_rl_loop.propagate_experience', 'propagate_experience', (['agent_vector', 'sample_trajectory'], {}), '(agent_vector, sample_trajectory)\n', (4964, 4997), False, 'from regym.rl_loops.multiagent_loops.sequential_action_rl_loop import propagate_experience, propagate_last_experience\n'), ((5151, 5204), 'regym.rl_loops.multiagent_loops.sequential_action_rl_loop.propagate_experience', 'propagate_experience', (['agent_vector', 'sample_trajectory'], {}), '(agent_vector, sample_trajectory)\n', (5171, 5204), False, 'from regym.rl_loops.multiagent_loops.sequential_action_rl_loop import propagate_experience, propagate_last_experience\n'), ((5381, 5434), 'regym.rl_loops.multiagent_loops.sequential_action_rl_loop.propagate_experience', 'propagate_experience', (['agent_vector', 'sample_trajectory'], {}), '(agent_vector, sample_trajectory)\n', (5401, 5434), False, 'from regym.rl_loops.multiagent_loops.sequential_action_rl_loop import propagate_experience, propagate_last_experience\n'), ((5506, 5564), 'regym.rl_loops.multiagent_loops.sequential_action_rl_loop.propagate_last_experience', 'propagate_last_experience', (['agent_vector', 'sample_trajectory'], {}), '(agent_vector, sample_trajectory)\n', (5531, 5564), False, 'from regym.rl_loops.multiagent_loops.sequential_action_rl_loop import propagate_experience, propagate_last_experience\n'), ((5995, 6092), 'numpy.testing.assert_array_equal', 'np.testing.assert_array_equal', (["actual_prediction['probs']", "expected_predictions[a_i]['probs']"], {}), "(actual_prediction['probs'],\n expected_predictions[a_i]['probs'])\n", (6024, 6092), True, 'import numpy as np\n')]
|
import numpy as np
# import os
# current_directory = os.path.dirname(os.path.abspath(__file__)).replace('\\','/')
# from ctypes import *
# bro = cdll.LoadLibrary(current_directory+"/broken.so")
# bro.broken_frame.argtypes = [np.ctypeslib.ndpointer(dtype=np.int16, ndim=1, flags="C_CONTIGUOUS"),
# c_int,
# np.ctypeslib.ndpointer(dtype=np.float, ndim=1), #
# c_int]
# bro.broken_frame.restype = c_int
def detect_broken_frame(wdata, framerate):
'''
To detect broken frame.
Parameters
----------
wdata: wave data Type:[array]
framerate: sample rate. Type:[int]
Returns
----------
bf: broken frame Type:[list]
_______ _______
| | |
| | |
| | |
| amp0 | amp1 |
| | |_______
| | | amp |
|_______|_______|_______|
_____________
\
|
|_______
_______________
| | |
| | |
| | |
| | amp |
_______| | |
| amp0 | amp1 | |
|_______|_______|_______|
_________
/
|
________________|
'''
# num = 0
# bf = np.zeros(5)
# num = bro.broken_frame(wdata,len(wdata),bf,framerate)
# if num == 0:
# bf = []
# return list(bf)
# frame length: 5ms
FRAME_LENGTH = 0.005
AMP_THRESHOLD = 4
up_edge = False
# print framerate
w = int(framerate*FRAME_LENGTH)
amp0 = amp1 = 0
bf = []
last_dis = 0
AMP_ARRAY = []
n = 0
for i in xrange(len(wdata)/w):
tem = np.sum(np.abs(wdata[i*w:(i+1)*w]))
if tem !=0:
amp = np.log10(tem) #amplitude
else:
amp = 0
#Up edge detection
if up_edge is False:
dis = amp1-amp
ldis = amp0-amp
if (dis >= AMP_THRESHOLD) and (ldis>=AMP_THRESHOLD):# and (distance1 > 0):#AMP_THRESHOLD-0.2
bft = round((i*w)/float(framerate),3)
up_edge = True
n = 0
#Falling edge detection
else:
n += 1
dis = amp1-amp0
ldis = amp-amp0
if (dis >= AMP_THRESHOLD) and (ldis>=AMP_THRESHOLD):#AMP_THRESHOLD-0.2 (distance0 > 0) and
# print dis-ldis,i,amp0,amp1,amp
up_edge = False
n = 0
bf.append(bft)
#if detect a falling edge, but it can`t detect a up edge within 5 seconds, we will reset the FLAG
elif n%400 == 0:
n = 0
up_edge = False
#Update amp0 & amp1
amp0 = amp1
amp1 = amp
# #Up edge detection
# if up_edge is False:
# distance0 = amp0-amp
# if (distance0 > AMP_THRESHOLD):# and (distance1 > 0):#AMP_THRESHOLD-0.2
# bft = round((i*w)/float(framerate),3)
# up_edge = True
# #Falling edge detection
# else:
# distance0 = amp-amp0
# distance1 = amp1-amp0
# if (distance1 > AMP_THRESHOLD):#AMP_THRESHOLD-0.2 (distance0 > 0) and
# up_edge = False
# bf.append(bft)
# #if detect a falling edge, but it can`t detect a up edge within 5 seconds, we will reset the FLAG
# elif i%100 == 0:
# up_edge = False
# #Update amp0 & amp1
# amp0,amp1=amp1,amp
# #######################################
# import matplotlib.pyplot as plt
# x = range(len(wdata)/w)
# plt.title("")
# plt.xlabel('Window')
# plt.ylabel('Amplitude (log)')#
# plt.plot(x,AMP_ARRAY)
# plt.show()
# #######################################
# import matplotlib.mlab as mlab
# import matplotlib.pyplot as plt
# num_bins = 90
# # the histogram of the data
# n, bins, patches = plt.hist(AMP_ARRAY, num_bins, normed = True, facecolor='green', alpha=0.5)
# plt.xlabel('Distance')
# plt.ylabel('Probability(100%)')
# plt.title(r'Histogram of amplitude')
# plt.show()
if len(bf) == 0:
return 0
else:
return bf
|
[
"numpy.abs",
"numpy.log10"
] |
[((1532, 1564), 'numpy.abs', 'np.abs', (['wdata[i * w:(i + 1) * w]'], {}), '(wdata[i * w:(i + 1) * w])\n', (1538, 1564), True, 'import numpy as np\n'), ((1584, 1597), 'numpy.log10', 'np.log10', (['tem'], {}), '(tem)\n', (1592, 1597), True, 'import numpy as np\n')]
|
from os import listdir
from os.path import isfile
from PIL import Image
from tqdm import tqdm
import numpy as np
import imgaug.augmenters as iaa
import os
import random
from os.path import join
import matplotlib.pyplot as plt
DATA_DIR = 'DATA DIR'
os.chdir(DATA_DIR)
IMAGE_DIR = join(DATA_DIR, 'dataset\\PascalVOC-OG-flipped\\JPEGImages')
ANN_DIR = join(DATA_DIR, 'dataset\\PascalVOC-OG-flipped\\Annotations')
NEW_IMAGE_DIR = join(DATA_DIR, 'dataset\\PascalVOC-OG-all\\JPEGImages')
NEW_ANN_DIR = join(DATA_DIR, 'dataset\\PascalVOC-OG-all\\Annotations')
NEW_IMAGE_SETS_DIR = join(DATA_DIR, 'dataset\\PascalVOC-OG-all\\ImageSets\\Main')
MAX = 2
with open(join(NEW_IMAGE_SETS_DIR, f"pipe-augmented-degrade.txt"), 'w+') as f:
pass
image_files = [f for f in listdir(IMAGE_DIR) if isfile(join(IMAGE_DIR, f))]
shuffled_image_files = random.sample(image_files, len(image_files))
shuffled_image_files = random.sample(image_files, len(shuffled_image_files))[:MAX]
seq = iaa.Sequential([
iaa.JpegCompression(compression=(99, 99))
])
for image in tqdm(shuffled_image_files):
if len(image) > 0:
# Åpne bildet
im = Image.open(join(IMAGE_DIR, image))
# Gjøre om til array med type uint8, (1920, 1080, 3)
im = np.asarray(im).astype(np.uint8)
# Ekspandere arrayet til å se ut som (1, 1920, 1080, 3), nødvendig siden iaa forventer en 4D matrise
im_expand = np.expand_dims(im, 0)
# Augmentere bildet
augmented_image_array = seq(images=im_expand)
# Fjerne ekstra dimensjonen satt på tidligere på første akse, resultat: (1920, 1080, 3)
augmented_image_array = np.squeeze(augmented_image_array, axis=0)
# Laste inn array som bilde
augmented_image = Image.fromarray(augmented_image_array)
# Laste inn bildet igjen fra matriseformat.
im = Image.fromarray(im)
im.save('im1.jpeg')
augmented_image.save('im2.jpeg')
fig, ax = plt.subplots(nrows=1, ncols=2)
# Plotting
plt.subplot(1, 2, 1)
plt.imshow(im)
plt.subplot(1, 2, 2)
plt.imshow(augmented_image)
plt.show()
|
[
"matplotlib.pyplot.imshow",
"PIL.Image.fromarray",
"os.listdir",
"tqdm.tqdm",
"os.path.join",
"numpy.asarray",
"numpy.squeeze",
"os.chdir",
"numpy.expand_dims",
"imgaug.augmenters.JpegCompression",
"matplotlib.pyplot.subplot",
"matplotlib.pyplot.subplots",
"matplotlib.pyplot.show"
] |
[((265, 283), 'os.chdir', 'os.chdir', (['DATA_DIR'], {}), '(DATA_DIR)\n', (273, 283), False, 'import os\n'), ((301, 360), 'os.path.join', 'join', (['DATA_DIR', '"""dataset\\\\PascalVOC-OG-flipped\\\\JPEGImages"""'], {}), "(DATA_DIR, 'dataset\\\\PascalVOC-OG-flipped\\\\JPEGImages')\n", (305, 360), False, 'from os.path import join\n'), ((372, 432), 'os.path.join', 'join', (['DATA_DIR', '"""dataset\\\\PascalVOC-OG-flipped\\\\Annotations"""'], {}), "(DATA_DIR, 'dataset\\\\PascalVOC-OG-flipped\\\\Annotations')\n", (376, 432), False, 'from os.path import join\n'), ((452, 507), 'os.path.join', 'join', (['DATA_DIR', '"""dataset\\\\PascalVOC-OG-all\\\\JPEGImages"""'], {}), "(DATA_DIR, 'dataset\\\\PascalVOC-OG-all\\\\JPEGImages')\n", (456, 507), False, 'from os.path import join\n'), ((523, 579), 'os.path.join', 'join', (['DATA_DIR', '"""dataset\\\\PascalVOC-OG-all\\\\Annotations"""'], {}), "(DATA_DIR, 'dataset\\\\PascalVOC-OG-all\\\\Annotations')\n", (527, 579), False, 'from os.path import join\n'), ((604, 664), 'os.path.join', 'join', (['DATA_DIR', '"""dataset\\\\PascalVOC-OG-all\\\\ImageSets\\\\Main"""'], {}), "(DATA_DIR, 'dataset\\\\PascalVOC-OG-all\\\\ImageSets\\\\Main')\n", (608, 664), False, 'from os.path import join\n'), ((1095, 1121), 'tqdm.tqdm', 'tqdm', (['shuffled_image_files'], {}), '(shuffled_image_files)\n', (1099, 1121), False, 'from tqdm import tqdm\n'), ((689, 744), 'os.path.join', 'join', (['NEW_IMAGE_SETS_DIR', 'f"""pipe-augmented-degrade.txt"""'], {}), "(NEW_IMAGE_SETS_DIR, f'pipe-augmented-degrade.txt')\n", (693, 744), False, 'from os.path import join\n'), ((797, 815), 'os.listdir', 'listdir', (['IMAGE_DIR'], {}), '(IMAGE_DIR)\n', (804, 815), False, 'from os import listdir\n'), ((1031, 1072), 'imgaug.augmenters.JpegCompression', 'iaa.JpegCompression', ([], {'compression': '(99, 99)'}), '(compression=(99, 99))\n', (1050, 1072), True, 'import imgaug.augmenters as iaa\n'), ((1462, 1483), 'numpy.expand_dims', 'np.expand_dims', (['im', '(0)'], {}), '(im, 0)\n', (1476, 1483), True, 'import numpy as np\n'), ((1702, 1743), 'numpy.squeeze', 'np.squeeze', (['augmented_image_array'], {'axis': '(0)'}), '(augmented_image_array, axis=0)\n', (1712, 1743), True, 'import numpy as np\n'), ((1810, 1848), 'PIL.Image.fromarray', 'Image.fromarray', (['augmented_image_array'], {}), '(augmented_image_array)\n', (1825, 1848), False, 'from PIL import Image\n'), ((1918, 1937), 'PIL.Image.fromarray', 'Image.fromarray', (['im'], {}), '(im)\n', (1933, 1937), False, 'from PIL import Image\n'), ((2028, 2058), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {'nrows': '(1)', 'ncols': '(2)'}), '(nrows=1, ncols=2)\n', (2040, 2058), True, 'import matplotlib.pyplot as plt\n'), ((2090, 2110), 'matplotlib.pyplot.subplot', 'plt.subplot', (['(1)', '(2)', '(1)'], {}), '(1, 2, 1)\n', (2101, 2110), True, 'import matplotlib.pyplot as plt\n'), ((2120, 2134), 'matplotlib.pyplot.imshow', 'plt.imshow', (['im'], {}), '(im)\n', (2130, 2134), True, 'import matplotlib.pyplot as plt\n'), ((2146, 2166), 'matplotlib.pyplot.subplot', 'plt.subplot', (['(1)', '(2)', '(2)'], {}), '(1, 2, 2)\n', (2157, 2166), True, 'import matplotlib.pyplot as plt\n'), ((2176, 2203), 'matplotlib.pyplot.imshow', 'plt.imshow', (['augmented_image'], {}), '(augmented_image)\n', (2186, 2203), True, 'import matplotlib.pyplot as plt\n'), ((2213, 2223), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (2221, 2223), True, 'import matplotlib.pyplot as plt\n'), ((826, 844), 'os.path.join', 'join', (['IMAGE_DIR', 'f'], {}), '(IMAGE_DIR, f)\n', (830, 844), False, 'from os.path import join\n'), ((1195, 1217), 'os.path.join', 'join', (['IMAGE_DIR', 'image'], {}), '(IMAGE_DIR, image)\n', (1199, 1217), False, 'from os.path import join\n'), ((1297, 1311), 'numpy.asarray', 'np.asarray', (['im'], {}), '(im)\n', (1307, 1311), True, 'import numpy as np\n')]
|
from typing import Dict
import numpy as np
def buffer_from_example(example: Dict[str, np.ndarray],
leading_dims) -> Dict[str, np.ndarray]:
buf = {}
for key, value in example.items():
buf[key] = np.zeros(leading_dims + value.shape, dtype=value.dtype)
return buf
def get_leading_dims(dictionary, n_dims=1):
values = iter(dictionary.values())
leading_dims = next(values).shape[:n_dims]
if not all(leading_dims == value.shape[:n_dims] for value in values):
key, shape = [(key, value.shape[:n_dims])
for key, value in dictionary.items()
if leading_dims != value.shape[:n_dims]][0]
raise ValueError((f'Dimensions do not match: {leading_dims} vs. '
f'{shape} (for key `{key}`)'))
return leading_dims
|
[
"numpy.zeros"
] |
[((237, 292), 'numpy.zeros', 'np.zeros', (['(leading_dims + value.shape)'], {'dtype': 'value.dtype'}), '(leading_dims + value.shape, dtype=value.dtype)\n', (245, 292), True, 'import numpy as np\n')]
|
import numpy as np
import cv2
image = cv2.imread('images/unsharp_bird.jpg')
kernel = np.array([
[0, -1, 0],
[-1, 5, -1],
[0, -1, 0]
])
sharpen_iamge = cv2.filter2D(image, -1, kernel)
cv2.imshow("original image", image)
cv2.imshow("sharpen image", sharpen_iamge)
cv2.waitKey(0)
cv2.destroyAllWindows()
|
[
"cv2.filter2D",
"cv2.imshow",
"numpy.array",
"cv2.destroyAllWindows",
"cv2.waitKey",
"cv2.imread"
] |
[((39, 76), 'cv2.imread', 'cv2.imread', (['"""images/unsharp_bird.jpg"""'], {}), "('images/unsharp_bird.jpg')\n", (49, 76), False, 'import cv2\n'), ((87, 134), 'numpy.array', 'np.array', (['[[0, -1, 0], [-1, 5, -1], [0, -1, 0]]'], {}), '([[0, -1, 0], [-1, 5, -1], [0, -1, 0]])\n', (95, 134), True, 'import numpy as np\n'), ((200, 231), 'cv2.filter2D', 'cv2.filter2D', (['image', '(-1)', 'kernel'], {}), '(image, -1, kernel)\n', (212, 231), False, 'import cv2\n'), ((233, 268), 'cv2.imshow', 'cv2.imshow', (['"""original image"""', 'image'], {}), "('original image', image)\n", (243, 268), False, 'import cv2\n'), ((269, 311), 'cv2.imshow', 'cv2.imshow', (['"""sharpen image"""', 'sharpen_iamge'], {}), "('sharpen image', sharpen_iamge)\n", (279, 311), False, 'import cv2\n'), ((312, 326), 'cv2.waitKey', 'cv2.waitKey', (['(0)'], {}), '(0)\n', (323, 326), False, 'import cv2\n'), ((327, 350), 'cv2.destroyAllWindows', 'cv2.destroyAllWindows', ([], {}), '()\n', (348, 350), False, 'import cv2\n')]
|
# -*- coding: utf-8 -*-
import csv
import logging
import math
import multiprocessing
import os
import shutil
import time
from concurrent.futures import ThreadPoolExecutor
from pathlib import Path
from typing import Dict, List, Tuple
from django.utils import timezone
import numpy as np
import pandas as pd
import psutil
import rpy2.robjects as ro
import simplejson as json
from rpy2.robjects import pandas2ri, r as rlang
from rpy2.robjects.packages import importr
from data_refinery_common.logging import get_and_configure_logger
from data_refinery_common.models import ComputedFile, Sample
from data_refinery_common.utils import get_env_variable
from data_refinery_workers.processors import utils
MULTIPROCESSING_MAX_THREAD_COUNT = max(1, math.floor(multiprocessing.cpu_count() / 2) - 1)
RESULTS_BUCKET = get_env_variable("S3_RESULTS_BUCKET_NAME", "refinebio-results-bucket")
S3_BUCKET_NAME = get_env_variable("S3_BUCKET_NAME", "data-refinery")
BODY_HTML = (
Path("data_refinery_workers/processors/smasher_email.min.html").read_text().replace("\n", "")
)
BODY_ERROR_HTML = (
Path("data_refinery_workers/processors/smasher_email_error.min.html")
.read_text()
.replace("\n", "")
)
BYTES_IN_GB = 1024 * 1024 * 1024
QN_CHUNK_SIZE = 10000
logger = get_and_configure_logger(__name__)
### DEBUG ###
logger.setLevel(logging.getLevelName("DEBUG"))
def log_state(message, job_id, start_time=False):
if logger.isEnabledFor(logging.DEBUG):
process = psutil.Process(os.getpid())
ram_in_GB = process.memory_info().rss / BYTES_IN_GB
logger.debug(message, total_cpu=psutil.cpu_percent(), process_ram=ram_in_GB, job_id=job_id)
if start_time:
logger.debug("Duration: %s" % (time.time() - start_time), job_id=job_id)
else:
return time.time()
def prepare_files(job_context: Dict) -> Dict:
"""
Fetches and prepares the files to smash.
"""
start_prepare_files = log_state("start prepare files", job_context["job"].id)
found_files = False
job_context["filtered_samples"] = {}
job_context["input_files"] = {}
# `key` can either be the species name or experiment accession.
for key, samples in job_context["samples"].items():
smashable_files = []
seen_files = set()
for sample in samples:
smashable_file = sample.get_most_recent_smashable_result_file()
if smashable_file is not None and smashable_file not in seen_files:
smashable_files = smashable_files + [(smashable_file, sample)]
seen_files.add(smashable_file)
found_files = True
else:
sample_metadata = sample.to_metadata_dict()
job_context["filtered_samples"][sample.accession_code] = {
**sample_metadata,
"reason": "This sample did not have a processed file associated with it in our database.",
"experiment_accession_code": get_experiment_accession(
sample.accession_code, job_context["dataset"].data
),
}
job_context["input_files"][key] = smashable_files
job_context["num_input_files"] = len(job_context["input_files"])
job_context["group_by_keys"] = list(job_context["input_files"].keys())
if not found_files:
raise utils.ProcessorJobError(
"Couldn't get any files to smash for Smash job!!",
success=False,
dataset_id=job_context["dataset"].id,
num_samples=len(job_context["samples"]),
)
dataset_id = str(job_context["dataset"].pk)
job_context["work_dir"] = "/home/user/data_store/smashed/" + dataset_id + "/"
# Ensure we have a fresh smash directory
shutil.rmtree(job_context["work_dir"], ignore_errors=True)
os.makedirs(job_context["work_dir"])
job_context["output_dir"] = job_context["work_dir"] + "output/"
os.makedirs(job_context["output_dir"])
log_state("end prepare files", job_context["job"].id, start_prepare_files)
return job_context
def _load_and_sanitize_file(computed_file_path) -> pd.DataFrame:
""" Read and sanitize a computed file """
data = pd.read_csv(
computed_file_path,
sep="\t",
header=0,
index_col=0,
dtype={0: str, 1: np.float32},
error_bad_lines=False,
)
# Strip any funky whitespace
data.columns = data.columns.str.strip()
data = data.dropna(axis="columns", how="all")
# Make sure the index type is correct
data.index = data.index.map(str)
# Ensure that we don't have any dangling Brainarray-generated probe symbols.
# BA likes to leave '_at', signifying probe identifiers,
# on their converted, non-probe identifiers. It makes no sense.
# So, we chop them off and don't worry about it.
data.index = data.index.str.replace("_at", "")
# Remove any lingering Affymetrix control probes ("AFFX-")
data = data[~data.index.str.contains("AFFX-")]
# If there are any _versioned_ gene identifiers, remove that
# version information. We're using the latest brainarray for everything anyway.
# Jackie says this is okay.
# She also says that in the future, we may only want to do this
# for cross-technology smashes.
# This regex needs to be able to handle EGIDs in the form:
# ENSGXXXXYYYZZZZ.6
# and
# fgenesh2_kg.7__3016__AT5G35080.1 (via http://plants.ensembl.org/Arabidopsis_lyrata/ \
# Gene/Summary?g=fgenesh2_kg.7__3016__AT5G35080.1;r=7:17949732-17952000;t=fgenesh2_kg. \
# 7__3016__AT5G35080.1;db=core)
data.index = data.index.str.replace(r"(\.[^.]*)$", "")
# Squish duplicated rows together.
# XXX/TODO: Is mean the appropriate method here?
# We can make this an option in future.
# Discussion here: https://github.com/AlexsLemonade/refinebio/issues/186#issuecomment-395516419
data = data.groupby(data.index, sort=False).mean()
return data
def process_frame(work_dir, computed_file, sample_accession_code, aggregate_by) -> pd.DataFrame:
""" Downloads the computed file from S3 and tries to see if it's smashable.
Returns a data frame if the file can be processed or False otherwise. """
try:
# Download the file to a job-specific location so it
# won't disappear while we're using it.
computed_file_path = computed_file.get_synced_file_path(
path="%s%s" % (work_dir, computed_file.filename)
)
# Bail appropriately if this isn't a real file.
if not computed_file_path or not os.path.exists(computed_file_path):
logger.warning(
"Smasher received non-existent file path.",
computed_file_path=computed_file_path,
computed_file_id=computed_file.id,
)
return None
data = _load_and_sanitize_file(computed_file_path)
if len(data.columns) > 2:
# Most of the time, >1 is actually bad, but we also need to support
# two-channel samples. I think ultimately those should be given some kind of
# special consideration.
logger.info(
"Found a frame with more than 2 columns - this shouldn't happen!",
computed_file_path=computed_file_path,
computed_file_id=computed_file.id,
)
return None
# via https://github.com/AlexsLemonade/refinebio/issues/330:
# aggregating by experiment -> return untransformed output from tximport
# aggregating by species -> log2(x + 1) tximport output
if aggregate_by == "SPECIES" and computed_file.has_been_log2scaled():
data = data + 1
data = np.log2(data)
# Ideally done in the NO-OPPER, but sanity check here.
if (not computed_file.has_been_log2scaled()) and (data.max() > 100).any():
logger.info("Detected non-log2 microarray data.", computed_file_id=computed_file.id)
data = np.log2(data)
# Explicitly title this dataframe
try:
data.columns = [sample_accession_code]
except ValueError:
# This sample might have multiple channels, or something else.
# Don't mess with it.
logger.warn(
"Smasher found multi-channel column (probably) - skipping!",
exc_info=1,
computed_file_path=computed_file_path,
)
return None
except Exception:
# Okay, somebody probably forgot to create a SampleComputedFileAssociation
# Don't mess with it.
logger.warn(
"Smasher found very bad column title - skipping!",
exc_info=1,
computed_file_path=computed_file_path,
)
return None
except Exception:
logger.exception("Unable to smash file", file=computed_file_path)
return None
# TEMPORARY for iterating on compendia more quickly.
# finally:
# # Delete before archiving the work dir
# if computed_file_path and os.path.exists(computed_file_path):
# os.remove(computed_file_path)
return data
def load_first_pass_data_if_cached(work_dir: str):
path = os.path.join(work_dir, "first_pass.csv")
try:
with open(path, newline="") as csvfile:
reader = csv.reader(csvfile)
gene_ids = next(reader)
microarray_columns = next(reader)
rnaseq_columns = next(reader)
return {
"gene_ids": gene_ids,
"microarray_columns": microarray_columns,
"rnaseq_columns": rnaseq_columns,
}
# If the file doesn't exist then the gene ids aren't cached. Any
# other exception should be handled and higher in the stack.
except FileNotFoundError:
return None
def cache_first_pass(
job_context: Dict, gene_ids: List[str], microarray_columns: List[str], rnaseq_columns: List[str]
):
try:
path = os.path.join(job_context["work_dir"], "first_pass.csv")
logger.info(
"Caching gene_ids, microarray_columns, and rnaseq_columns to %s",
path,
job_id=job_context["job"].id,
)
with open(path, "w", newline="") as csvfile:
writer = csv.writer(csvfile)
writer.writerow(gene_ids)
writer.writerow(microarray_columns)
writer.writerow(rnaseq_columns)
# Nothing in the above try should raise an exception, but if it
# does don't waste the work we did in the first pass.
except Exception:
logger.exception(
"Error writing gene identifiers to CSV file.", job_id=job_context["job"].id
)
def process_frames_for_key(
key: str, input_files: List[Tuple[ComputedFile, Sample]], job_context: Dict
) -> Dict:
"""Download, read, and chunk processed sample files from s3.
`key` is the species or experiment whose samples are contained in `input_files`.
Will add to job_context the keys 'microarray_matrix' and
'rnaseq_matrix' with pandas dataframes containing all of the
samples' data. Also adds the key 'unsmashable_files' containing a
list of paths that were determined to be unsmashable.
"""
start_gene_ids = log_state(
"Collecting all gene identifiers for key {}".format(key), job_context["job"].id
)
# Build up a list of gene identifiers because these will be the
# rows of our matrices, and we want to preallocate them so we need
# to know them all.
## We may have built this list in a previous job, check to see if it's cached:
cached_data = load_first_pass_data_if_cached(job_context["work_dir"])
first_pass_was_cached = False
if cached_data:
logger.info(
(
"The data from the first pass was cached, so we're using "
"that and skipping the first pass."
),
job_id=job_context["job"].id,
)
first_pass_was_cached = True
all_gene_identifiers = cached_data["gene_ids"]
microarray_columns = cached_data["microarray_columns"]
rnaseq_columns = cached_data["rnaseq_columns"]
else:
gene_identifier_counts = {}
microarray_columns = []
rnaseq_columns = []
for index, (computed_file, sample) in enumerate(input_files):
log_state("1st processing frame {}".format(index), job_context["job"].id)
frame_data = process_frame(
job_context["work_dir"],
computed_file,
sample.accession_code,
job_context["dataset"].aggregate_by,
)
if frame_data is None:
# we were unable to process this sample, so we drop
logger.warning(
"Unable to smash file",
computed_file=computed_file.id,
dataset_id=job_context["dataset"].id,
job_id=job_context["job"].id,
)
sample_metadata = sample.to_metadata_dict()
job_context["filtered_samples"][sample.accession_code] = {
**sample_metadata,
"reason": "The file associated with this sample did not pass the QC checks we apply before aggregating.",
"filename": computed_file.filename,
"experiment_accession_code": get_experiment_accession(
sample.accession_code, job_context["dataset"].data
),
}
continue
# Count how many frames are in each tech so we can preallocate
# the matrices in both directions.
for gene_id in frame_data.index:
if gene_id in gene_identifier_counts:
gene_identifier_counts[gene_id] += 1
else:
gene_identifier_counts[gene_id] = 1
# Each dataframe should only have 1 column, but it's
# returned as a list so use extend.
if sample.technology == "MICROARRAY":
microarray_columns.extend(frame_data.columns)
elif sample.technology == "RNA-SEQ":
rnaseq_columns.extend(frame_data.columns)
# We only want to use gene identifiers which are present
# in >50% of the samples. We're doing this because a large
# number of gene identifiers present in only a modest
# number of experiments have leaked through. We wouldn't
# necessarily want to do this if we'd mapped all the data
# to ENSEMBL identifiers successfully.
total_samples = len(microarray_columns) + len(rnaseq_columns)
all_gene_identifiers = [
gene_id
for gene_id in gene_identifier_counts
if gene_identifier_counts[gene_id] > (total_samples * 0.5)
]
all_gene_identifiers.sort()
del gene_identifier_counts
log_template = (
"Collected {0} gene identifiers for {1} across"
" {2} micrarry samples and {3} RNA-Seq samples."
)
log_state(
log_template.format(
len(all_gene_identifiers), key, len(microarray_columns), len(rnaseq_columns)
),
job_context["job"].id,
start_gene_ids,
)
# Temporarily only cache mouse compendia because it may not succeed.
if not first_pass_was_cached and key == "MUS_MUSCULUS":
cache_first_pass(job_context, all_gene_identifiers, microarray_columns, rnaseq_columns)
start_build_matrix = log_state("Beginning to build the full matrices.", job_context["job"].id)
# Sort the columns so that the matrices are in predictable orders.
microarray_columns.sort()
rnaseq_columns.sort()
# Preallocate the matrices to be the exact size we will need. This
# should prevent any operations from happening while we build it
# up, so the only RAM used will be needed.
job_context["microarray_matrix"] = pd.DataFrame(
data=None, index=all_gene_identifiers, columns=microarray_columns, dtype=np.float32
)
job_context["rnaseq_matrix"] = pd.DataFrame(
data=None, index=all_gene_identifiers, columns=rnaseq_columns, dtype=np.float32
)
for index, (computed_file, sample) in enumerate(input_files):
log_state("2nd processing frame {}".format(index), job_context["job"].id)
frame_data = process_frame(
job_context["work_dir"],
computed_file,
sample.accession_code,
job_context["dataset"].aggregate_by,
)
if frame_data is None:
job_context["unsmashable_files"].append(computed_file.filename)
sample_metadata = sample.to_metadata_dict()
job_context["filtered_samples"][sample.accession_code] = {
**sample_metadata,
"reason": "The file associated with this sample did not contain a vector that fit the expected dimensions of the matrix.",
"filename": computed_file.filename,
"experiment_accession_code": get_experiment_accession(
sample.accession_code, job_context["dataset"].data
),
}
continue
frame_data = frame_data.reindex(all_gene_identifiers)
# The dataframe for each sample will only have one column
# whose header will be the accession code.
column = frame_data.columns[0]
if sample.technology == "MICROARRAY":
job_context["microarray_matrix"][column] = frame_data.values
elif sample.technology == "RNA-SEQ":
job_context["rnaseq_matrix"][column] = frame_data.values
job_context["num_samples"] = 0
if job_context["microarray_matrix"] is not None:
job_context["num_samples"] += len(job_context["microarray_matrix"].columns)
if job_context["rnaseq_matrix"] is not None:
job_context["num_samples"] += len(job_context["rnaseq_matrix"].columns)
log_state(
"Built full matrices for key {}".format(key), job_context["job"].id, start_build_matrix
)
return job_context
# Modified from: http://yaoyao.codes/pandas/2018/01/23/pandas-split-a-dataframe-into-chunks
def _index_marks(num_columns, chunk_size):
return range(chunk_size, math.ceil(num_columns / chunk_size) * chunk_size, chunk_size)
def _split_dataframe_columns(dataframe, chunk_size):
indices = _index_marks(dataframe.shape[1], chunk_size)
return np.split(dataframe, indices, axis=1)
def _quantile_normalize_matrix(target_vector, original_matrix):
preprocessCore = importr("preprocessCore")
as_numeric = rlang("as.numeric")
data_matrix = rlang("data.matrix")
# Convert the smashed frames to an R numeric Matrix
target_vector = as_numeric(target_vector)
# Do so in chunks if the matrix is too large.
if original_matrix.shape[1] <= QN_CHUNK_SIZE:
merged_matrix = data_matrix(original_matrix)
normalized_matrix = preprocessCore.normalize_quantiles_use_target(
x=merged_matrix, target=target_vector, copy=True
)
# And finally convert back to Pandas
ar = np.array(normalized_matrix)
new_merged = pd.DataFrame(ar, columns=original_matrix.columns, index=original_matrix.index)
else:
matrix_chunks = _split_dataframe_columns(original_matrix, QN_CHUNK_SIZE)
for i, chunk in enumerate(matrix_chunks):
R_chunk = data_matrix(chunk)
normalized_chunk = preprocessCore.normalize_quantiles_use_target(
x=R_chunk, target=target_vector, copy=True
)
ar = np.array(normalized_chunk)
start_column = i * QN_CHUNK_SIZE
end_column = (i + 1) * QN_CHUNK_SIZE
original_matrix.iloc[:, start_column:end_column] = ar
new_merged = original_matrix
return new_merged
def _test_qn(merged_matrix):
""" Selects a list of 100 random pairs of columns and performs the KS Test on them.
Returns a list of tuples with the results of the KN test (statistic, pvalue) """
# Verify this QN, related:
# https://github.com/AlexsLemonade/refinebio/issues/599#issuecomment-422132009
data_matrix = rlang("data.matrix")
as_numeric = rlang("as.numeric")
set_seed = rlang("set.seed")
combn = rlang("combn")
ncol = rlang("ncol")
ks_test = rlang("ks.test")
which = rlang("which")
merged_R_matrix = data_matrix(merged_matrix)
set_seed(123)
n = ncol(merged_R_matrix)[0]
m = 2
# Not enough columns to perform KS test - either bad smash or single sample smash.
if n < m:
return None
# This wont work with larger matricies
# https://github.com/AlexsLemonade/refinebio/issues/1860
ncolumns = ncol(merged_R_matrix)
if ncolumns[0] <= 200:
# Convert to NP, Shuffle, Return to R
combos = combn(ncolumns, 2)
ar = np.array(combos)
np.random.shuffle(np.transpose(ar))
else:
indexes = [*range(ncolumns[0])]
np.random.shuffle(indexes)
ar = np.array([*zip(indexes[0:100], indexes[100:200])])
nr, nc = ar.shape
combos = ro.r.matrix(ar, nrow=nr, ncol=nc)
result = []
# adapted from
# https://stackoverflow.com/questions/9661469/r-t-test-over-all-columns
# apply KS test to randomly selected pairs of columns (samples)
for i in range(1, min(ncol(combos)[0], 100)):
value1 = combos.rx(1, i)[0]
value2 = combos.rx(2, i)[0]
test_a = merged_R_matrix.rx(True, value1)
test_b = merged_R_matrix.rx(True, value2)
# RNA-seq has a lot of zeroes in it, which
# breaks the ks_test. Therefore we want to
# filter them out. To do this we drop the
# lowest half of the values. If there's
# still zeroes in there, then that's
# probably too many zeroes so it's okay to
# fail.
median_a = np.median(test_a)
median_b = np.median(test_b)
# `which` returns indices which are
# 1-indexed. Python accesses lists with
# zero-indexes, even if that list is
# actually an R vector. Therefore subtract
# 1 to account for the difference.
test_a = [test_a[i - 1] for i in which(test_a > median_a)]
test_b = [test_b[i - 1] for i in which(test_b > median_b)]
# The python list comprehension gives us a
# python list, but ks_test wants an R
# vector so let's go back.
test_a = as_numeric(test_a)
test_b = as_numeric(test_b)
ks_res = ks_test(test_a, test_b)
statistic = ks_res.rx("statistic")[0][0]
pvalue = ks_res.rx("p.value")[0][0]
result.append((statistic, pvalue))
return result
def quantile_normalize(job_context: Dict, ks_check=True, ks_stat=0.001) -> Dict:
"""
Apply quantile normalization.
"""
# Prepare our QN target file
organism = job_context["organism"]
if not organism.qn_target:
raise utils.ProcessorJobError(
"Could not find QN target for Organism: " + str(organism),
success=False,
organism=organism,
dataset_id=job_context["dataset"].id,
)
qn_target_path = organism.qn_target.computedfile_set.latest().sync_from_s3()
qn_target_frame = pd.read_csv(
qn_target_path, sep="\t", header=None, index_col=None, error_bad_lines=False
)
# Prepare our RPy2 bridge
pandas2ri.activate()
# Remove un-quantiled normalized matrix from job_context
# because we no longer need it.
merged_no_qn = job_context.pop("merged_no_qn")
# Perform the Actual QN
new_merged = _quantile_normalize_matrix(qn_target_frame[0], merged_no_qn)
# And add the quantile normalized matrix to job_context.
job_context["merged_qn"] = new_merged
# For now, don't test the QN for mouse/human. This never fails on
# smasher jobs and is OOM-killing our very large compendia
# jobs. Let's run this manually after we have a compendia job
# actually finish.
if organism.name in ["MUS_MUSCULUS", "HOMO_SAPIENS"]:
return job_context
ks_res = _test_qn(new_merged)
if ks_res:
for (statistic, pvalue) in ks_res:
job_context["ks_statistic"] = statistic
job_context["ks_pvalue"] = pvalue
# We're unsure of how strigent to be about
# the pvalue just yet, so we're extra lax
# rather than failing tons of tests. This may need tuning.
if ks_check and (statistic > ks_stat or pvalue < 0.8):
job_context["ks_warning"] = (
"Failed Kolmogorov Smirnov test! Stat: "
+ str(statistic)
+ ", PVal: "
+ str(pvalue)
)
else:
logger.warning(
"Not enough columns to perform KS test - either bad smash or single sample smash.",
dataset_id=job_context["dataset"].id,
)
return job_context
def compile_metadata(job_context: Dict) -> Dict:
"""Compiles metadata about the job.
Returns a new dict containing the metadata, not the job_context.
"""
metadata = {}
metadata["num_samples"] = job_context["num_samples"]
metadata["num_experiments"] = job_context["experiments"].count()
metadata["quant_sf_only"] = job_context["dataset"].quant_sf_only
if not job_context["dataset"].quant_sf_only:
metadata["aggregate_by"] = job_context["dataset"].aggregate_by
metadata["scale_by"] = job_context["dataset"].scale_by
# https://github.com/AlexsLemonade/refinebio/pull/421#discussion_r203799646
# TODO: do something with these.
# metadata['non_aggregated_files'] = job_context["unsmashable_files"]
metadata["ks_statistic"] = job_context.get("ks_statistic", None)
metadata["ks_pvalue"] = job_context.get("ks_pvalue", None)
metadata["ks_warning"] = job_context.get("ks_warning", None)
metadata["quantile_normalized"] = job_context["dataset"].quantile_normalize
filtered_samples = job_context["filtered_samples"]
samples = {}
for sample in job_context["dataset"].get_samples():
if sample.accession_code in filtered_samples:
# skip the samples that were filtered
continue
samples[sample.accession_code] = sample.to_metadata_dict()
metadata["samples"] = samples
experiments = {}
for experiment in job_context["dataset"].get_experiments():
experiment_metadata = experiment.to_metadata_dict()
# exclude filtered samples from experiment metadata
all_samples = experiment_metadata["sample_accession_codes"]
all_samples = [code for code in all_samples if code not in filtered_samples]
experiment_metadata["sample_accession_codes"] = all_samples
experiments[experiment.accession_code] = experiment_metadata
metadata["experiments"] = experiments
return metadata
def write_non_data_files(job_context: Dict) -> Dict:
"""Writes the files that are not the actual data of the dataset.
This include LICENSE.txt and README.md files and the metadata.
Adds the key `metadata` to job_context and populates it with all
the metadata that needs to be written.
"""
job_context["metadata"] = compile_metadata(job_context)
shutil.copy("README_DATASET.md", job_context["output_dir"] + "README.md")
shutil.copy("LICENSE_DATASET.txt", job_context["output_dir"] + "LICENSE.TXT")
# Write samples metadata to TSV
try:
write_tsv_json(job_context)
# Metadata to JSON
job_context["metadata"]["created_at"] = timezone.now().strftime("%Y-%m-%dT%H:%M:%S")
aggregated_metadata_path = os.path.join(
job_context["output_dir"], "aggregated_metadata.json"
)
with open(aggregated_metadata_path, "w", encoding="utf-8") as metadata_file:
json.dump(job_context["metadata"], metadata_file, indent=4, sort_keys=True)
if job_context["filtered_samples"]:
# generate filtered samples file only if some samples were skipped
filtered_samples_path = os.path.join(
job_context["output_dir"], "filtered_samples_metadata.json"
)
with open(filtered_samples_path, "w", encoding="utf-8") as metadata_file:
json.dump(job_context["filtered_samples"], metadata_file, indent=4, sort_keys=True)
columns = get_tsv_columns(job_context["filtered_samples"])
filtered_samples_tsv_path = os.path.join(
job_context["output_dir"], "filtered_samples_metadata.tsv"
)
with open(filtered_samples_tsv_path, "w", encoding="utf-8") as tsv_file:
dw = csv.DictWriter(tsv_file, columns, delimiter="\t", extrasaction="ignore")
dw.writeheader()
for sample_metadata in job_context["filtered_samples"].values():
dw.writerow(get_tsv_row_data(sample_metadata, job_context["dataset"].data))
except Exception:
raise utils.ProcessorJobError("Failed to write metadata TSV!", success=False)
return job_context
def get_experiment_accession(sample_accession_code, dataset_data):
for experiment_accession, samples in dataset_data.items():
if sample_accession_code in samples:
return experiment_accession
return "" # Should never happen, because the sample is by definition in the dataset
def _add_annotation_column(annotation_columns, column_name):
"""Add annotation column names in place.
Any column_name that starts with "refinebio_" will be skipped.
"""
if not column_name.startswith("refinebio_"):
annotation_columns.add(column_name)
def _add_annotation_value(row_data, col_name, col_value, sample_accession_code):
"""Adds a new `col_name` key whose value is `col_value` to row_data.
If col_name already exists in row_data with different value, print
out a warning message.
"""
# Generate a warning message if annotation field name starts with
# "refinebio_". This should rarely (if ever) happen.
if col_name.startswith("refinebio_"):
logger.warning(
"Annotation value skipped",
annotation_field=col_name,
annotation_value=col_value,
sample_accession_code=sample_accession_code,
)
elif col_name not in row_data:
row_data[col_name] = col_value
# Generate a warning message in case of conflicts of annotation values.
# (Requested by Dr. <NAME>)
elif row_data[col_name] != col_value:
logger.warning(
"Conflict of values found in column %s: %s vs. %s"
% (col_name, row_data[col_name], col_value),
sample_accession_code=sample_accession_code,
)
def get_tsv_row_data(sample_metadata, dataset_data):
"""Returns field values based on input sample_metadata.
Some annotation fields are treated specially because they are more
important. See `get_tsv_columns` function above for details.
"""
sample_accession_code = sample_metadata.get("refinebio_accession_code", "")
row_data = dict()
for meta_key, meta_value in sample_metadata.items():
# If the field is a refinebio-specific field, simply copy it.
if meta_key != "refinebio_annotations":
row_data[meta_key] = meta_value
continue
# Decompose sample_metadata["refinebio_annotations"], which is
# an array of annotations.
for annotation in meta_value:
for annotation_key, annotation_value in annotation.items():
# "characteristic" in ArrayExpress annotation
if (
sample_metadata.get("refinebio_source_database", "") == "ARRAY_EXPRESS"
and annotation_key == "characteristic"
):
for pair_dict in annotation_value:
if "category" in pair_dict and "value" in pair_dict:
col_name, col_value = pair_dict["category"], pair_dict["value"]
_add_annotation_value(
row_data, col_name, col_value, sample_accession_code
)
# "variable" in ArrayExpress annotation
elif (
sample_metadata.get("refinebio_source_database", "") == "ARRAY_EXPRESS"
and annotation_key == "variable"
):
for pair_dict in annotation_value:
if "name" in pair_dict and "value" in pair_dict:
col_name, col_value = pair_dict["name"], pair_dict["value"]
_add_annotation_value(
row_data, col_name, col_value, sample_accession_code
)
# Skip "source" field ArrayExpress sample's annotation
elif (
sample_metadata.get("refinebio_source_database", "") == "ARRAY_EXPRESS"
and annotation_key == "source"
):
continue
# "characteristics_ch1" in GEO annotation
elif (
sample_metadata.get("refinebio_source_database", "") == "GEO"
and annotation_key == "characteristics_ch1"
): # array of strings
for pair_str in annotation_value:
if ":" in pair_str:
col_name, col_value = pair_str.split(":", 1)
col_value = col_value.strip()
_add_annotation_value(
row_data, col_name, col_value, sample_accession_code
)
# If annotation_value includes only a 'name' key, extract its value directly:
elif (
isinstance(annotation_value, dict)
and len(annotation_value) == 1
and "name" in annotation_value
):
_add_annotation_value(
row_data, annotation_key, annotation_value["name"], sample_accession_code
)
# If annotation_value is a single-element array, extract the element directly:
elif isinstance(annotation_value, list) and len(annotation_value) == 1:
_add_annotation_value(
row_data, annotation_key, annotation_value[0], sample_accession_code
)
# Otherwise save all annotation fields in separate columns
else:
_add_annotation_value(
row_data, annotation_key, annotation_value, sample_accession_code
)
row_data["experiment_accession"] = get_experiment_accession(sample_accession_code, dataset_data)
return row_data
def get_tsv_columns(samples_metadata):
"""Returns an array of strings that will be written as a TSV file's
header. The columns are based on fields found in samples_metadata.
Some nested annotation fields are taken out as separate columns
because they are more important than the others.
"""
refinebio_columns = set()
annotation_columns = set()
for sample_metadata in samples_metadata.values():
for meta_key, meta_value in sample_metadata.items():
if meta_key != "refinebio_annotations":
refinebio_columns.add(meta_key)
continue
# Decompose sample_metadata["annotations"], which is an array of annotations!
for annotation in meta_value:
for annotation_key, annotation_value in annotation.items():
# For ArrayExpress samples, take out the fields
# nested in "characteristic" as separate columns.
if (
sample_metadata.get("refinebio_source_database", "") == "ARRAY_EXPRESS"
and annotation_key == "characteristic"
):
for pair_dict in annotation_value:
if "category" in pair_dict and "value" in pair_dict:
_add_annotation_column(annotation_columns, pair_dict["category"])
# For ArrayExpress samples, also take out the fields
# nested in "variable" as separate columns.
elif (
sample_metadata.get("refinebio_source_database", "") == "ARRAY_EXPRESS"
and annotation_key == "variable"
):
for pair_dict in annotation_value:
if "name" in pair_dict and "value" in pair_dict:
_add_annotation_column(annotation_columns, pair_dict["name"])
# For ArrayExpress samples, skip "source" field
elif (
sample_metadata.get("refinebio_source_database", "") == "ARRAY_EXPRESS"
and annotation_key == "source"
):
continue
# For GEO samples, take out the fields nested in
# "characteristics_ch1" as separate columns.
elif (
sample_metadata.get("refinebio_source_database", "") == "GEO"
and annotation_key == "characteristics_ch1"
): # array of strings
for pair_str in annotation_value:
if ":" in pair_str:
tokens = pair_str.split(":", 1)
_add_annotation_column(annotation_columns, tokens[0])
# Saves all other annotation fields in separate columns
else:
_add_annotation_column(annotation_columns, annotation_key)
# Return sorted columns, in which "refinebio_accession_code" and "experiment_accession" are
# always first, followed by the other refinebio columns (in alphabetic order), and
# annotation columns (in alphabetic order) at the end.
refinebio_columns.discard("refinebio_accession_code")
return (
["refinebio_accession_code", "experiment_accession"]
+ sorted(refinebio_columns)
+ sorted(annotation_columns)
)
def write_tsv_json(job_context):
"""Writes tsv files on disk.
If the dataset is aggregated by species, also write species-level
JSON file.
"""
# Avoid pulling this out of job_context repeatedly.
metadata = job_context["metadata"]
# Uniform TSV header per dataset
columns = get_tsv_columns(metadata["samples"])
# Per-Experiment Metadata
if job_context["dataset"].aggregate_by == "EXPERIMENT":
tsv_paths = []
for experiment_title, experiment_data in metadata["experiments"].items():
experiment_dir = job_context["output_dir"] + experiment_title + "/"
experiment_dir = experiment_dir.encode("ascii", "ignore")
os.makedirs(experiment_dir, exist_ok=True)
tsv_path = experiment_dir.decode("utf-8") + "metadata_" + experiment_title + ".tsv"
tsv_path = tsv_path.encode("ascii", "ignore")
tsv_paths.append(tsv_path)
with open(tsv_path, "w", encoding="utf-8") as tsv_file:
dw = csv.DictWriter(tsv_file, columns, delimiter="\t", extrasaction="ignore")
dw.writeheader()
for sample_accession_code, sample_metadata in metadata["samples"].items():
if sample_accession_code in experiment_data["sample_accession_codes"]:
row_data = get_tsv_row_data(sample_metadata, job_context["dataset"].data)
dw.writerow(row_data)
return tsv_paths
# Per-Species Metadata
elif job_context["dataset"].aggregate_by == "SPECIES":
tsv_paths = []
for species in job_context["group_by_keys"]:
species_dir = job_context["output_dir"] + species + "/"
os.makedirs(species_dir, exist_ok=True)
samples_in_species = []
tsv_path = species_dir + "metadata_" + species + ".tsv"
tsv_paths.append(tsv_path)
with open(tsv_path, "w", encoding="utf-8") as tsv_file:
# See http://www.lucainvernizzi.net/blog/2015/08/03/8x-speed-up-for-python-s-csv-dictwriter/
# about extrasaction.
dw = csv.DictWriter(tsv_file, columns, delimiter="\t", extrasaction="ignore")
dw.writeheader()
i = 0
for sample_metadata in metadata["samples"].values():
if sample_metadata.get("refinebio_organism", "") == species:
row_data = get_tsv_row_data(sample_metadata, job_context["dataset"].data)
dw.writerow(row_data)
samples_in_species.append(sample_metadata)
i = i + 1
if i % 1000 == 0:
progress_template = (
"Done with {0} out of {1} lines of metadata " "for species {2}"
)
log_state(
progress_template.format(i, len(metadata["samples"]), species),
job_context["job"].id,
)
# Writes a json file for current species:
if len(samples_in_species):
species_metadata = {"species": species, "samples": samples_in_species}
json_path = species_dir + "metadata_" + species + ".json"
with open(json_path, "w", encoding="utf-8") as json_file:
json.dump(species_metadata, json_file, indent=4, sort_keys=True)
return tsv_paths
# All Metadata
else:
all_dir = job_context["output_dir"] + "ALL/"
os.makedirs(all_dir, exist_ok=True)
tsv_path = all_dir + "metadata_ALL.tsv"
with open(tsv_path, "w", encoding="utf-8") as tsv_file:
dw = csv.DictWriter(tsv_file, columns, delimiter="\t", extrasaction="ignore")
dw.writeheader()
for sample_metadata in metadata["samples"].values():
row_data = get_tsv_row_data(sample_metadata, job_context["dataset"].data)
dw.writerow(row_data)
return [tsv_path]
def download_computed_file(download_tuple: Tuple[ComputedFile, str]):
""" this function downloads the latest computed file. Receives a tuple with
the computed file and the path where it needs to be downloaded
This is used to parallelize downloading quantsf files. """
(latest_computed_file, output_file_path) = download_tuple
try:
latest_computed_file.get_synced_file_path(path=output_file_path)
except:
# Let's not fail if there's an error syncing one of the quant.sf files
logger.exception("Failed to sync computed file", computed_file_id=latest_computed_file.pk)
def sync_quant_files(output_path, samples: List[Sample]):
""" Takes a list of ComputedFiles and copies the ones that are quant files to the provided directory.
Returns the total number of samples that were included """
num_samples = 0
page_size = 100
# split the samples in groups and download each one individually
with ThreadPoolExecutor(max_workers=MULTIPROCESSING_MAX_THREAD_COUNT) as executor:
# for each sample we need it's latest quant.sf file we don't want to query the db
# for all of them, so we do it in groups of 100, and then download all of the computed_files
# in parallel
for sample_page in (
samples[i * page_size : i + page_size] for i in range(0, len(samples), page_size)
):
sample_and_computed_files = []
for sample in sample_page:
latest_computed_file = sample.get_most_recent_quant_sf_file()
if not latest_computed_file:
continue
output_file_path = output_path + sample.accession_code + "_quant.sf"
sample_and_computed_files.append((latest_computed_file, output_file_path))
# download this set of files, this will take a few seconds that should also help the db recover
executor.map(download_computed_file, sample_and_computed_files)
num_samples += len(sample_and_computed_files)
return num_samples
|
[
"csv.DictWriter",
"rpy2.robjects.pandas2ri.activate",
"pandas.read_csv",
"data_refinery_common.logging.get_and_configure_logger",
"multiprocessing.cpu_count",
"numpy.array",
"rpy2.robjects.r",
"os.path.exists",
"pathlib.Path",
"rpy2.robjects.packages.importr",
"django.utils.timezone.now",
"os.getpid",
"pandas.DataFrame",
"csv.reader",
"psutil.cpu_percent",
"csv.writer",
"data_refinery_common.utils.get_env_variable",
"logging.getLevelName",
"simplejson.dump",
"shutil.copy",
"numpy.log2",
"numpy.transpose",
"time.time",
"data_refinery_workers.processors.utils.ProcessorJobError",
"numpy.median",
"math.ceil",
"os.makedirs",
"concurrent.futures.ThreadPoolExecutor",
"os.path.join",
"rpy2.robjects.r.matrix",
"numpy.split",
"shutil.rmtree",
"numpy.random.shuffle"
] |
[((811, 881), 'data_refinery_common.utils.get_env_variable', 'get_env_variable', (['"""S3_RESULTS_BUCKET_NAME"""', '"""refinebio-results-bucket"""'], {}), "('S3_RESULTS_BUCKET_NAME', 'refinebio-results-bucket')\n", (827, 881), False, 'from data_refinery_common.utils import get_env_variable\n'), ((899, 950), 'data_refinery_common.utils.get_env_variable', 'get_env_variable', (['"""S3_BUCKET_NAME"""', '"""data-refinery"""'], {}), "('S3_BUCKET_NAME', 'data-refinery')\n", (915, 950), False, 'from data_refinery_common.utils import get_env_variable\n'), ((1265, 1299), 'data_refinery_common.logging.get_and_configure_logger', 'get_and_configure_logger', (['__name__'], {}), '(__name__)\n', (1289, 1299), False, 'from data_refinery_common.logging import get_and_configure_logger\n'), ((1330, 1359), 'logging.getLevelName', 'logging.getLevelName', (['"""DEBUG"""'], {}), "('DEBUG')\n", (1350, 1359), False, 'import logging\n'), ((3782, 3840), 'shutil.rmtree', 'shutil.rmtree', (["job_context['work_dir']"], {'ignore_errors': '(True)'}), "(job_context['work_dir'], ignore_errors=True)\n", (3795, 3840), False, 'import shutil\n'), ((3845, 3881), 'os.makedirs', 'os.makedirs', (["job_context['work_dir']"], {}), "(job_context['work_dir'])\n", (3856, 3881), False, 'import os\n'), ((3955, 3993), 'os.makedirs', 'os.makedirs', (["job_context['output_dir']"], {}), "(job_context['output_dir'])\n", (3966, 3993), False, 'import os\n'), ((4221, 4347), 'pandas.read_csv', 'pd.read_csv', (['computed_file_path'], {'sep': '"""\t"""', 'header': '(0)', 'index_col': '(0)', 'dtype': '{(0): str, (1): np.float32}', 'error_bad_lines': '(False)'}), "(computed_file_path, sep='\\t', header=0, index_col=0, dtype={(0):\n str, (1): np.float32}, error_bad_lines=False)\n", (4232, 4347), True, 'import pandas as pd\n'), ((9367, 9407), 'os.path.join', 'os.path.join', (['work_dir', '"""first_pass.csv"""'], {}), "(work_dir, 'first_pass.csv')\n", (9379, 9407), False, 'import os\n'), ((16220, 16322), 'pandas.DataFrame', 'pd.DataFrame', ([], {'data': 'None', 'index': 'all_gene_identifiers', 'columns': 'microarray_columns', 'dtype': 'np.float32'}), '(data=None, index=all_gene_identifiers, columns=\n microarray_columns, dtype=np.float32)\n', (16232, 16322), True, 'import pandas as pd\n'), ((16367, 16464), 'pandas.DataFrame', 'pd.DataFrame', ([], {'data': 'None', 'index': 'all_gene_identifiers', 'columns': 'rnaseq_columns', 'dtype': 'np.float32'}), '(data=None, index=all_gene_identifiers, columns=rnaseq_columns,\n dtype=np.float32)\n', (16379, 16464), True, 'import pandas as pd\n'), ((18725, 18761), 'numpy.split', 'np.split', (['dataframe', 'indices'], {'axis': '(1)'}), '(dataframe, indices, axis=1)\n', (18733, 18761), True, 'import numpy as np\n'), ((18849, 18874), 'rpy2.robjects.packages.importr', 'importr', (['"""preprocessCore"""'], {}), "('preprocessCore')\n", (18856, 18874), False, 'from rpy2.robjects.packages import importr\n'), ((18892, 18911), 'rpy2.robjects.r', 'rlang', (['"""as.numeric"""'], {}), "('as.numeric')\n", (18897, 18911), True, 'from rpy2.robjects import pandas2ri, r as rlang\n'), ((18930, 18950), 'rpy2.robjects.r', 'rlang', (['"""data.matrix"""'], {}), "('data.matrix')\n", (18935, 18950), True, 'from rpy2.robjects import pandas2ri, r as rlang\n'), ((20474, 20494), 'rpy2.robjects.r', 'rlang', (['"""data.matrix"""'], {}), "('data.matrix')\n", (20479, 20494), True, 'from rpy2.robjects import pandas2ri, r as rlang\n'), ((20512, 20531), 'rpy2.robjects.r', 'rlang', (['"""as.numeric"""'], {}), "('as.numeric')\n", (20517, 20531), True, 'from rpy2.robjects import pandas2ri, r as rlang\n'), ((20547, 20564), 'rpy2.robjects.r', 'rlang', (['"""set.seed"""'], {}), "('set.seed')\n", (20552, 20564), True, 'from rpy2.robjects import pandas2ri, r as rlang\n'), ((20577, 20591), 'rpy2.robjects.r', 'rlang', (['"""combn"""'], {}), "('combn')\n", (20582, 20591), True, 'from rpy2.robjects import pandas2ri, r as rlang\n'), ((20603, 20616), 'rpy2.robjects.r', 'rlang', (['"""ncol"""'], {}), "('ncol')\n", (20608, 20616), True, 'from rpy2.robjects import pandas2ri, r as rlang\n'), ((20631, 20647), 'rpy2.robjects.r', 'rlang', (['"""ks.test"""'], {}), "('ks.test')\n", (20636, 20647), True, 'from rpy2.robjects import pandas2ri, r as rlang\n'), ((20660, 20674), 'rpy2.robjects.r', 'rlang', (['"""which"""'], {}), "('which')\n", (20665, 20674), True, 'from rpy2.robjects import pandas2ri, r as rlang\n'), ((21421, 21454), 'rpy2.robjects.r.matrix', 'ro.r.matrix', (['ar'], {'nrow': 'nr', 'ncol': 'nc'}), '(ar, nrow=nr, ncol=nc)\n', (21432, 21454), True, 'import rpy2.robjects as ro\n'), ((23583, 23676), 'pandas.read_csv', 'pd.read_csv', (['qn_target_path'], {'sep': '"""\t"""', 'header': 'None', 'index_col': 'None', 'error_bad_lines': '(False)'}), "(qn_target_path, sep='\\t', header=None, index_col=None,\n error_bad_lines=False)\n", (23594, 23676), True, 'import pandas as pd\n'), ((23722, 23742), 'rpy2.robjects.pandas2ri.activate', 'pandas2ri.activate', ([], {}), '()\n', (23740, 23742), False, 'from rpy2.robjects import pandas2ri, r as rlang\n'), ((27651, 27724), 'shutil.copy', 'shutil.copy', (['"""README_DATASET.md"""', "(job_context['output_dir'] + 'README.md')"], {}), "('README_DATASET.md', job_context['output_dir'] + 'README.md')\n", (27662, 27724), False, 'import shutil\n'), ((27729, 27806), 'shutil.copy', 'shutil.copy', (['"""LICENSE_DATASET.txt"""', "(job_context['output_dir'] + 'LICENSE.TXT')"], {}), "('LICENSE_DATASET.txt', job_context['output_dir'] + 'LICENSE.TXT')\n", (27740, 27806), False, 'import shutil\n'), ((10147, 10202), 'os.path.join', 'os.path.join', (["job_context['work_dir']", '"""first_pass.csv"""'], {}), "(job_context['work_dir'], 'first_pass.csv')\n", (10159, 10202), False, 'import os\n'), ((19412, 19439), 'numpy.array', 'np.array', (['normalized_matrix'], {}), '(normalized_matrix)\n', (19420, 19439), True, 'import numpy as np\n'), ((19461, 19539), 'pandas.DataFrame', 'pd.DataFrame', (['ar'], {'columns': 'original_matrix.columns', 'index': 'original_matrix.index'}), '(ar, columns=original_matrix.columns, index=original_matrix.index)\n', (19473, 19539), True, 'import pandas as pd\n'), ((21175, 21191), 'numpy.array', 'np.array', (['combos'], {}), '(combos)\n', (21183, 21191), True, 'import numpy as np\n'), ((21294, 21320), 'numpy.random.shuffle', 'np.random.shuffle', (['indexes'], {}), '(indexes)\n', (21311, 21320), True, 'import numpy as np\n'), ((22190, 22207), 'numpy.median', 'np.median', (['test_a'], {}), '(test_a)\n', (22199, 22207), True, 'import numpy as np\n'), ((22227, 22244), 'numpy.median', 'np.median', (['test_b'], {}), '(test_b)\n', (22236, 22244), True, 'import numpy as np\n'), ((28044, 28111), 'os.path.join', 'os.path.join', (["job_context['output_dir']", '"""aggregated_metadata.json"""'], {}), "(job_context['output_dir'], 'aggregated_metadata.json')\n", (28056, 28111), False, 'import os\n'), ((43948, 44012), 'concurrent.futures.ThreadPoolExecutor', 'ThreadPoolExecutor', ([], {'max_workers': 'MULTIPROCESSING_MAX_THREAD_COUNT'}), '(max_workers=MULTIPROCESSING_MAX_THREAD_COUNT)\n', (43966, 44012), False, 'from concurrent.futures import ThreadPoolExecutor\n'), ((1489, 1500), 'os.getpid', 'os.getpid', ([], {}), '()\n', (1498, 1500), False, 'import os\n'), ((1804, 1815), 'time.time', 'time.time', ([], {}), '()\n', (1813, 1815), False, 'import time\n'), ((7815, 7828), 'numpy.log2', 'np.log2', (['data'], {}), '(data)\n', (7822, 7828), True, 'import numpy as np\n'), ((8092, 8105), 'numpy.log2', 'np.log2', (['data'], {}), '(data)\n', (8099, 8105), True, 'import numpy as np\n'), ((9486, 9505), 'csv.reader', 'csv.reader', (['csvfile'], {}), '(csvfile)\n', (9496, 9505), False, 'import csv\n'), ((10446, 10465), 'csv.writer', 'csv.writer', (['csvfile'], {}), '(csvfile)\n', (10456, 10465), False, 'import csv\n'), ((18538, 18573), 'math.ceil', 'math.ceil', (['(num_columns / chunk_size)'], {}), '(num_columns / chunk_size)\n', (18547, 18573), False, 'import math\n'), ((19890, 19916), 'numpy.array', 'np.array', (['normalized_chunk'], {}), '(normalized_chunk)\n', (19898, 19916), True, 'import numpy as np\n'), ((21218, 21234), 'numpy.transpose', 'np.transpose', (['ar'], {}), '(ar)\n', (21230, 21234), True, 'import numpy as np\n'), ((28231, 28306), 'simplejson.dump', 'json.dump', (["job_context['metadata']", 'metadata_file'], {'indent': '(4)', 'sort_keys': '(True)'}), "(job_context['metadata'], metadata_file, indent=4, sort_keys=True)\n", (28240, 28306), True, 'import simplejson as json\n'), ((28467, 28540), 'os.path.join', 'os.path.join', (["job_context['output_dir']", '"""filtered_samples_metadata.json"""'], {}), "(job_context['output_dir'], 'filtered_samples_metadata.json')\n", (28479, 28540), False, 'import os\n'), ((28869, 28941), 'os.path.join', 'os.path.join', (["job_context['output_dir']", '"""filtered_samples_metadata.tsv"""'], {}), "(job_context['output_dir'], 'filtered_samples_metadata.tsv')\n", (28881, 28941), False, 'import os\n'), ((29397, 29468), 'data_refinery_workers.processors.utils.ProcessorJobError', 'utils.ProcessorJobError', (['"""Failed to write metadata TSV!"""'], {'success': '(False)'}), "('Failed to write metadata TSV!', success=False)\n", (29420, 29468), False, 'from data_refinery_workers.processors import utils\n'), ((39595, 39637), 'os.makedirs', 'os.makedirs', (['experiment_dir'], {'exist_ok': '(True)'}), '(experiment_dir, exist_ok=True)\n', (39606, 39637), False, 'import os\n'), ((42494, 42529), 'os.makedirs', 'os.makedirs', (['all_dir'], {'exist_ok': '(True)'}), '(all_dir, exist_ok=True)\n', (42505, 42529), False, 'import os\n'), ((756, 783), 'multiprocessing.cpu_count', 'multiprocessing.cpu_count', ([], {}), '()\n', (781, 783), False, 'import multiprocessing\n'), ((969, 1032), 'pathlib.Path', 'Path', (['"""data_refinery_workers/processors/smasher_email.min.html"""'], {}), "('data_refinery_workers/processors/smasher_email.min.html')\n", (973, 1032), False, 'from pathlib import Path\n'), ((1089, 1158), 'pathlib.Path', 'Path', (['"""data_refinery_workers/processors/smasher_email_error.min.html"""'], {}), "('data_refinery_workers/processors/smasher_email_error.min.html')\n", (1093, 1158), False, 'from pathlib import Path\n'), ((1602, 1622), 'psutil.cpu_percent', 'psutil.cpu_percent', ([], {}), '()\n', (1620, 1622), False, 'import psutil\n'), ((6650, 6684), 'os.path.exists', 'os.path.exists', (['computed_file_path'], {}), '(computed_file_path)\n', (6664, 6684), False, 'import os\n'), ((27964, 27978), 'django.utils.timezone.now', 'timezone.now', ([], {}), '()\n', (27976, 27978), False, 'from django.utils import timezone\n'), ((28673, 28760), 'simplejson.dump', 'json.dump', (["job_context['filtered_samples']", 'metadata_file'], {'indent': '(4)', 'sort_keys': '(True)'}), "(job_context['filtered_samples'], metadata_file, indent=4,\n sort_keys=True)\n", (28682, 28760), True, 'import simplejson as json\n'), ((29078, 29150), 'csv.DictWriter', 'csv.DictWriter', (['tsv_file', 'columns'], {'delimiter': '"""\t"""', 'extrasaction': '"""ignore"""'}), "(tsv_file, columns, delimiter='\\t', extrasaction='ignore')\n", (29092, 29150), False, 'import csv\n'), ((39920, 39992), 'csv.DictWriter', 'csv.DictWriter', (['tsv_file', 'columns'], {'delimiter': '"""\t"""', 'extrasaction': '"""ignore"""'}), "(tsv_file, columns, delimiter='\\t', extrasaction='ignore')\n", (39934, 39992), False, 'import csv\n'), ((40619, 40658), 'os.makedirs', 'os.makedirs', (['species_dir'], {'exist_ok': '(True)'}), '(species_dir, exist_ok=True)\n', (40630, 40658), False, 'import os\n'), ((42659, 42731), 'csv.DictWriter', 'csv.DictWriter', (['tsv_file', 'columns'], {'delimiter': '"""\t"""', 'extrasaction': '"""ignore"""'}), "(tsv_file, columns, delimiter='\\t', extrasaction='ignore')\n", (42673, 42731), False, 'import csv\n'), ((41038, 41110), 'csv.DictWriter', 'csv.DictWriter', (['tsv_file', 'columns'], {'delimiter': '"""\t"""', 'extrasaction': '"""ignore"""'}), "(tsv_file, columns, delimiter='\\t', extrasaction='ignore')\n", (41052, 41110), False, 'import csv\n'), ((1729, 1740), 'time.time', 'time.time', ([], {}), '()\n', (1738, 1740), False, 'import time\n'), ((42314, 42378), 'simplejson.dump', 'json.dump', (['species_metadata', 'json_file'], {'indent': '(4)', 'sort_keys': '(True)'}), '(species_metadata, json_file, indent=4, sort_keys=True)\n', (42323, 42378), True, 'import simplejson as json\n')]
|
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
# from smac.env.multiagentenv import MultiAgentEnv
# from smac.env.starcraft2.maps import get_map_params
from ..multiagentenv import MultiAgentEnv
from ..starcraft2.maps import get_map_params
import atexit
from operator import attrgetter
from copy import deepcopy
import numpy as np
import enum
import math, time
from absl import logging
from pysc2 import maps
from pysc2 import run_configs
from pysc2.lib import protocol
from s2clientprotocol import common_pb2 as sc_common
from s2clientprotocol import sc2api_pb2 as sc_pb
from s2clientprotocol import raw_pb2 as r_pb
from s2clientprotocol import debug_pb2 as d_pb
races = {
"R": sc_common.Random,
"P": sc_common.Protoss,
"T": sc_common.Terran,
"Z": sc_common.Zerg,
}
difficulties = {
"1": sc_pb.VeryEasy,
"2": sc_pb.Easy,
"3": sc_pb.Medium,
"4": sc_pb.MediumHard,
"5": sc_pb.Hard,
"6": sc_pb.Harder,
"7": sc_pb.VeryHard,
"8": sc_pb.CheatVision,
"9": sc_pb.CheatMoney,
"A": sc_pb.CheatInsane,
}
actions = {
"move": 16, # target: PointOrUnit
"attack": 23, # target: PointOrUnit
"stop": 4, # target: None
"heal": 386, # Unit
}
class Direction(enum.IntEnum):
NORTH = 0
SOUTH = 1
EAST = 2
WEST = 3
class StarCraftWrappedEnv(MultiAgentEnv):
"""The StarCraft II environment for decentralised multi-agent
micromanagement scenarios.
"""
def __init__(
self,
map_name="8m",
step_mul=8,
move_amount=2,
difficulty="7",
game_version=None,
seed=None,
continuing_episode=False,
obs_all_health=True,
obs_own_health=True,
obs_last_action=False,
obs_pathing_grid=False,
obs_terrain_height=False,
obs_instead_of_state=False,
obs_timestep_number=False,
state_last_action=True,
state_timestep_number=False,
reward_sparse=False,
reward_only_positive=True,
reward_death_value=10,
reward_win=200,
reward_defeat=0,
reward_negative_scale=0.5,
reward_scale=True,
reward_scale_rate=20,
replay_dir="",
replay_prefix="",
window_size_x=1920,
window_size_y=1200,
heuristic_ai=False,
debug=False,
is_replay=False
):
"""
Create a StarCraftC2Env environment.
Parameters
----------
map_name : str, optional
The name of the SC2 map to play (default is "8m"). The full list
can be found by running bin/map_list.
step_mul : int, optional
How many game steps per agent step (default is 8). None
indicates to use the default map step_mul.
move_amount : float, optional
How far away units are ordered to move per step (default is 2).
difficulty : str, optional
The difficulty of built-in computer AI bot (default is "7").
game_version : str, optional
StarCraft II game version (default is None). None indicates the
latest version.
seed : int, optional
Random seed used during game initialisation. This allows to
continuing_episode : bool, optional
Whether to consider episodes continuing or finished after time
limit is reached (default is False).
obs_all_health : bool, optional
Agents receive the health of all units (in the sight range) as part
of observations (default is True).
obs_own_health : bool, optional
Agents receive their own health as a part of observations (default
is False). This flag is ignored when obs_all_health == True.
obs_last_action : bool, optional
Agents receive the last actions of all units (in the sight range)
as part of observations (default is False).
obs_pathing_grid : bool, optional
Whether observations include pathing values surrounding the agent
(default is False).
obs_terrain_height : bool, optional
Whether observations include terrain height values surrounding the
agent (default is False).
obs_instead_of_state : bool, optional
Use combination of all agents' observations as the global state
(default is False).
obs_timestep_number : bool, optional
Whether observations include the current timestep of the episode
(default is False).
state_last_action : bool, optional
Include the last actions of all agents as part of the global state
(default is True).
state_timestep_number : bool, optional
Whether the state include the current timestep of the episode
(default is False).
reward_sparse : bool, optional
Receive 1/-1 reward for winning/loosing an episode (default is
False). Whe rest of reward parameters are ignored if True.
reward_only_positive : bool, optional
Reward is always positive (default is True).
reward_death_value : float, optional
The amount of reward received for killing an enemy unit (default
is 10). This is also the negative penalty for having an allied unit
killed if reward_only_positive == False.
reward_win : float, optional
The reward for winning in an episode (default is 200).
reward_defeat : float, optional
The reward for loosing in an episode (default is 0). This value
should be nonpositive.
reward_negative_scale : float, optional
Scaling factor for negative rewards (default is 0.5). This
parameter is ignored when reward_only_positive == True.
reward_scale : bool, optional
Whether or not to scale the reward (default is True).
reward_scale_rate : float, optional
Reward scale rate (default is 20). When reward_scale == True, the
reward received by the agents is divided by (max_reward /
reward_scale_rate), where max_reward is the maximum possible
reward per episode without considering the shield regeneration
of Protoss units.
replay_dir : str, optional
The directory to save replays (default is None). If None, the
replay will be saved in Replays directory where StarCraft II is
installed.
replay_prefix : str, optional
The prefix of the replay to be saved (default is None). If None,
the name of the map will be used.
window_size_x : int, optional
The length of StarCraft II window size (default is 1920).
window_size_y: int, optional
The height of StarCraft II window size (default is 1200).
heuristic_ai: bool, optional
Whether or not to use a non-learning heuristic AI (default False).
debug: bool, optional
Log messages about observations, state, actions and rewards for
debugging purposes (default is False).
"""
# Map arguments
print("inside the new env..")
time.sleep(100)
self.map_name = map_name
map_params = get_map_params(self.map_name)
self.n_agents = map_params["n_agents"]
self.n_enemies = map_params["n_enemies"]
self.episode_limit = map_params["limit"]
self._move_amount = move_amount
self._step_mul = step_mul
self.difficulty = difficulty
# Observations and state
self.obs_own_health = obs_own_health
self.obs_all_health = obs_all_health
self.obs_instead_of_state = obs_instead_of_state
self.obs_last_action = obs_last_action
self.obs_pathing_grid = obs_pathing_grid
self.obs_terrain_height = obs_terrain_height
self.obs_timestep_number = obs_timestep_number
self.state_last_action = state_last_action
self.state_timestep_number = state_timestep_number
if self.obs_all_health:
self.obs_own_health = True
self.n_obs_pathing = 8
self.n_obs_height = 9
# Rewards args
self.reward_sparse = reward_sparse
self.reward_only_positive = reward_only_positive
self.reward_negative_scale = reward_negative_scale
self.reward_death_value = reward_death_value
self.reward_win = reward_win
self.reward_defeat = reward_defeat
self.reward_scale = reward_scale
self.reward_scale_rate = reward_scale_rate
# Other
self.game_version = game_version
self.continuing_episode = continuing_episode
self._seed = seed
self.heuristic_ai = heuristic_ai
self.debug = debug
self.is_replay = is_replay
self.window_size = (window_size_x, window_size_y)
self.replay_dir = replay_dir
self.replay_prefix = replay_prefix
# Actions
self.n_actions_no_attack = 6
self.n_actions_move = 4
self.n_actions = self.n_actions_no_attack + self.n_enemies
# Map info
self._agent_race = map_params["a_race"]
self._bot_race = map_params["b_race"]
self.shield_bits_ally = 1 if self._agent_race == "P" else 0
self.shield_bits_enemy = 1 if self._bot_race == "P" else 0
self.unit_type_bits = map_params["unit_type_bits"]
self.map_type = map_params["map_type"]
self.max_reward = (
self.n_enemies * self.reward_death_value + self.reward_win
)
self.agents = {}
self.enemies = {}
self._episode_count = 0
self._episode_steps = 0
self._total_steps = 0
self._obs = None
self.battles_won = 0
self.battles_game = 0
self.timeouts = 0
self.force_restarts = 0
self.last_stats = None
self.death_tracker_ally = np.zeros(self.n_agents)
self.death_tracker_enemy = np.zeros(self.n_enemies)
self.previous_ally_units = None
self.previous_enemy_units = None
self.last_action = np.zeros((self.n_agents, self.n_actions))
self._min_unit_type = 0
self.marine_id = self.marauder_id = self.medivac_id = 0
self.hydralisk_id = self.zergling_id = self.baneling_id = 0
self.stalker_id = self.colossus_id = self.zealot_id = self.sentry_id = 0
self.void_ray_id = 0
self.max_distance_x = 0
self.max_distance_y = 0
self.map_x = 0
self.map_y = 0
self.terrain_height = None
self.pathing_grid = None
self._run_config = None
self._sc2_proc = None
self._controller = None
# Try to avoid leaking SC2 processes on shutdown
atexit.register(lambda: self.close())
def _launch(self):
"""Launch the StarCraft II game."""
# self._run_config = run_configs.get(version=self.game_version)
self._run_config = run_configs.get()
_map = maps.get(self.map_name)
# Setting up the interface
interface_options = sc_pb.InterfaceOptions(raw=True, score=False)
self._sc2_proc = self._run_config.start(window_size=self.window_size)
self._controller = self._sc2_proc.controller
# Request to create the game
create = sc_pb.RequestCreateGame(
local_map=sc_pb.LocalMap(
map_path=_map.path,
map_data=self._run_config.map_data(_map.path)),
realtime=False,
random_seed=self._seed)
create.player_setup.add(type=sc_pb.Participant)
create.player_setup.add(type=sc_pb.Computer, race=races[self._bot_race],
difficulty=difficulties[self.difficulty])
self._controller.create_game(create)
join = sc_pb.RequestJoinGame(race=races[self._agent_race],
options=interface_options)
self._controller.join_game(join)
game_info = self._controller.game_info()
map_info = game_info.start_raw
map_play_area_min = map_info.playable_area.p0
map_play_area_max = map_info.playable_area.p1
self.max_distance_x = map_play_area_max.x - map_play_area_min.x
self.max_distance_y = map_play_area_max.y - map_play_area_min.y
self.map_x = map_info.map_size.x
self.map_y = map_info.map_size.y
if map_info.pathing_grid.bits_per_pixel == 1:
vals = np.array(list(map_info.pathing_grid.data)).reshape(
self.map_x, int(self.map_y / 8))
self.pathing_grid = np.transpose(np.array([
[(b >> i) & 1 for b in row for i in range(7, -1, -1)]
for row in vals], dtype=np.bool))
else:
self.pathing_grid = np.invert(np.flip(np.transpose(np.array(
list(map_info.pathing_grid.data), dtype=np.bool).reshape(
self.map_x, self.map_y)), axis=1))
self.terrain_height = np.flip(
np.transpose(np.array(list(map_info.terrain_height.data))
.reshape(self.map_x, self.map_y)), 1) / 255
def reset(self):
"""Reset the environment. Required after each full episode.
Returns initial observations and states.
"""
self._episode_steps = 0
if self._episode_count == 0:
# Launch StarCraft II
self._launch()
else:
self._restart()
# Information kept for counting the reward
self.death_tracker_ally = np.zeros(self.n_agents)
self.death_tracker_enemy = np.zeros(self.n_enemies)
self.previous_ally_units = None
self.previous_enemy_units = None
self.win_counted = False
self.defeat_counted = False
self.last_action = np.zeros((self.n_agents, self.n_actions))
if self.heuristic_ai:
self.heuristic_targets = [None] * self.n_agents
try:
self._obs = self._controller.observe()
self.init_units()
except (protocol.ProtocolError, protocol.ConnectionError):
self.full_restart()
if self.debug:
logging.debug("Started Episode {}"
.format(self._episode_count).center(60, "*"))
return self.get_obs(), self.get_state()
def _restart(self):
"""Restart the environment by killing all units on the map.
There is a trigger in the SC2Map file, which restarts the
episode when there are no units left.
"""
try:
self._kill_all_units()
self._controller.step(2)
except (protocol.ProtocolError, protocol.ConnectionError):
self.full_restart()
def full_restart(self):
"""Full restart. Closes the SC2 process and launches a new one. """
self._sc2_proc.close()
self._launch()
self.force_restarts += 1
def step(self, actions):
"""A single environment step. Returns reward, terminated, info."""
if self.is_replay:
positions = []
for agent_id in range(self.n_agents):
unit = self.get_unit_by_id(agent_id)
positions.append([agent_id, unit.pos.x, unit.pos.y, unit.health])
for e_id, e_unit in self.enemies.items():
positions.append([e_id, e_unit.pos.x, e_unit.pos.y, e_unit.health])
# positions.insert(0,self._episode_steps)
print(positions, ",")
actions = [int(a) for a in actions]
self.last_action = np.eye(self.n_actions)[np.array(actions)]
# Collect individual actions
sc_actions = []
if self.debug:
logging.debug("Actions".center(60, "-"))
for a_id, action in enumerate(actions):
if not self.heuristic_ai:
agent_action = self.get_agent_action(a_id, action)
else:
agent_action = self.get_agent_action_heuristic(a_id, action)
if agent_action:
sc_actions.append(agent_action)
# Send action request
req_actions = sc_pb.RequestAction(actions=sc_actions)
try:
self._controller.actions(req_actions)
# Make step in SC2, i.e. apply actions
self._controller.step(self._step_mul)
# Observe here so that we know if the episode is over.
self._obs = self._controller.observe()
except (protocol.ProtocolError, protocol.ConnectionError):
self.full_restart()
return 0, True, {}
self._total_steps += 1
self._episode_steps += 1
# Update units
game_end_code = self.update_units()
terminated = False
reward = self.reward_battle()
info = {"battle_won": False}
if game_end_code is not None:
# Battle is over
terminated = True
self.battles_game += 1
if game_end_code == 1 and not self.win_counted:
self.battles_won += 1
self.win_counted = True
info["battle_won"] = True
if not self.reward_sparse:
reward += self.reward_win
else:
reward = 1
elif game_end_code == -1 and not self.defeat_counted:
self.defeat_counted = True
if not self.reward_sparse:
reward += self.reward_defeat
else:
reward = -1
elif self._episode_steps >= self.episode_limit:
# Episode limit reached
terminated = True
if self.continuing_episode:
info["episode_limit"] = True
self.battles_game += 1
self.timeouts += 1
if self.debug:
logging.debug("Reward = {}".format(reward).center(60, '-'))
if terminated:
self._episode_count += 1
if self.is_replay:
positions = []
for agent_id in range(self.n_agents):
unit = self.get_unit_by_id(agent_id)
positions.append([agent_id, unit.pos.x, unit.pos.y, unit.health])
for e_id, e_unit in self.enemies.items():
positions.append([e_id, e_unit.pos.x, e_unit.pos.y, e_unit.health])
# positions.insert(0,self._episode_steps)
print(positions, ",")
if self.reward_scale:
reward /= self.max_reward / self.reward_scale_rate
print("type of reward returned from within starcraft is: ", type(reward))
return 2 * reward, terminated, info
def get_agent_action(self, a_id, action):
"""Construct the action for agent a_id."""
avail_actions = self.get_avail_agent_actions(a_id)
assert avail_actions[action] == 1, \
"Agent {} cannot perform action {}".format(a_id, action)
unit = self.get_unit_by_id(a_id)
tag = unit.tag
x = unit.pos.x
y = unit.pos.y
if action == 0:
# no-op (valid only when dead)
assert unit.health == 0, "No-op only available for dead agents."
if self.debug:
logging.debug("Agent {}: Dead".format(a_id))
return None
elif action == 1:
# stop
cmd = r_pb.ActionRawUnitCommand(
ability_id=actions["stop"],
unit_tags=[tag],
queue_command=False)
if self.debug:
logging.debug("Agent {}: Stop".format(a_id))
elif action == 2:
# move north
cmd = r_pb.ActionRawUnitCommand(
ability_id=actions["move"],
target_world_space_pos=sc_common.Point2D(
x=x, y=y + self._move_amount),
unit_tags=[tag],
queue_command=False)
if self.debug:
logging.debug("Agent {}: Move North".format(a_id))
elif action == 3:
# move south
cmd = r_pb.ActionRawUnitCommand(
ability_id=actions["move"],
target_world_space_pos=sc_common.Point2D(
x=x, y=y - self._move_amount),
unit_tags=[tag],
queue_command=False)
if self.debug:
logging.debug("Agent {}: Move South".format(a_id))
elif action == 4:
# move east
cmd = r_pb.ActionRawUnitCommand(
ability_id=actions["move"],
target_world_space_pos=sc_common.Point2D(
x=x + self._move_amount, y=y),
unit_tags=[tag],
queue_command=False)
if self.debug:
logging.debug("Agent {}: Move East".format(a_id))
elif action == 5:
# move west
cmd = r_pb.ActionRawUnitCommand(
ability_id=actions["move"],
target_world_space_pos=sc_common.Point2D(
x=x - self._move_amount, y=y),
unit_tags=[tag],
queue_command=False)
if self.debug:
logging.debug("Agent {}: Move West".format(a_id))
else:
# attack/heal units that are in range
target_id = action - self.n_actions_no_attack
if self.map_type in ["MMM", "GMMM"] and unit.unit_type == self.medivac_id:
target_unit = self.agents[target_id]
action_name = "heal"
else:
target_unit = self.enemies[target_id]
action_name = "attack"
action_id = actions[action_name]
target_tag = target_unit.tag
cmd = r_pb.ActionRawUnitCommand(
ability_id=action_id,
target_unit_tag=target_tag,
unit_tags=[tag],
queue_command=False)
if self.debug:
logging.debug("Agent {} {}s unit # {}".format(
a_id, action_name, target_id))
sc_action = sc_pb.Action(action_raw=r_pb.ActionRaw(unit_command=cmd))
return sc_action
def get_agent_action_heuristic(self, a_id, action):
unit = self.get_unit_by_id(a_id)
tag = unit.tag
target = self.heuristic_targets[a_id]
if unit.unit_type == self.medivac_id:
if (target is None or self.agents[target].health == 0 or
self.agents[target].health == self.agents[target].health_max):
min_dist = math.hypot(self.max_distance_x, self.max_distance_y)
min_id = -1
for al_id, al_unit in self.agents.items():
if al_unit.unit_type == self.medivac_id:
continue
if (al_unit.health != 0 and
al_unit.health != al_unit.health_max):
dist = self.distance(unit.pos.x, unit.pos.y,
al_unit.pos.x, al_unit.pos.y)
if dist < min_dist:
min_dist = dist
min_id = al_id
self.heuristic_targets[a_id] = min_id
if min_id == -1:
self.heuristic_targets[a_id] = None
return None
action_id = actions['heal']
target_tag = self.agents[self.heuristic_targets[a_id]].tag
else:
if target is None or self.enemies[target].health == 0:
min_dist = math.hypot(self.max_distance_x, self.max_distance_y)
min_id = -1
for e_id, e_unit in self.enemies.items():
if (unit.unit_type == self.marauder_id and
e_unit.unit_type == self.medivac_id):
continue
if e_unit.health > 0:
dist = self.distance(unit.pos.x, unit.pos.y,
e_unit.pos.x, e_unit.pos.y)
if dist < min_dist:
min_dist = dist
min_id = e_id
self.heuristic_targets[a_id] = min_id
action_id = actions['attack']
target_tag = self.enemies[self.heuristic_targets[a_id]].tag
cmd = r_pb.ActionRawUnitCommand(
ability_id=action_id,
target_unit_tag=target_tag,
unit_tags=[tag],
queue_command=False)
sc_action = sc_pb.Action(action_raw=r_pb.ActionRaw(unit_command=cmd))
return sc_action
def reward_battle(self):
"""Reward function when self.reward_spare==False.
Returns accumulative hit/shield point damage dealt to the enemy
+ reward_death_value per enemy unit killed, and, in case
self.reward_only_positive == False, - (damage dealt to ally units
+ reward_death_value per ally unit killed) * self.reward_negative_scale
"""
if self.reward_sparse:
return 0
reward = 0
delta_deaths = 0
delta_ally = 0
delta_enemy = 0
neg_scale = self.reward_negative_scale
# update deaths
for al_id, al_unit in self.agents.items():
if not self.death_tracker_ally[al_id]:
# did not die so far
prev_health = (
self.previous_ally_units[al_id].health
+ self.previous_ally_units[al_id].shield
)
if al_unit.health == 0:
# just died
self.death_tracker_ally[al_id] = 1
if not self.reward_only_positive:
delta_deaths -= self.reward_death_value * neg_scale
delta_ally += prev_health * neg_scale
else:
# still alive
delta_ally += neg_scale * (
prev_health - al_unit.health - al_unit.shield
)
for e_id, e_unit in self.enemies.items():
if not self.death_tracker_enemy[e_id]:
prev_health = (
self.previous_enemy_units[e_id].health
+ self.previous_enemy_units[e_id].shield
)
if e_unit.health == 0:
self.death_tracker_enemy[e_id] = 1
delta_deaths += self.reward_death_value
delta_enemy += prev_health
else:
delta_enemy += prev_health - e_unit.health - e_unit.shield
if self.reward_only_positive:
reward = abs(delta_enemy + delta_deaths) # shield regeneration
else:
reward = delta_enemy + delta_deaths - delta_ally
return reward
def get_total_actions(self):
"""Returns the total number of actions an agent could ever take."""
return self.n_actions
@staticmethod
def distance(x1, y1, x2, y2):
"""Distance between two points."""
return math.hypot(x2 - x1, y2 - y1)
def unit_shoot_range(self, agent_id):
"""Returns the shooting range for an agent."""
return 6
def unit_sight_range(self, agent_id):
"""Returns the sight range for an agent."""
return 9
def unit_max_cooldown(self, unit):
"""Returns the maximal cooldown for a unit."""
switcher = {
self.marine_id: 15,
self.marauder_id: 25,
self.medivac_id: 200, # max energy
self.stalker_id: 35,
self.void_ray_id: 35,
self.sentry_id: 22,
self.zealot_id: 22,
self.colossus_id: 24,
self.hydralisk_id: 10,
self.zergling_id: 11,
self.baneling_id: 1
}
return switcher.get(unit.unit_type, 15)
def save_replay(self):
"""Save a replay."""
prefix = self.replay_prefix or self.map_name
replay_dir = self.replay_dir or ""
replay_path = self._run_config.save_replay(
self._controller.save_replay(), replay_dir=replay_dir, prefix=prefix)
logging.info("Replay saved at: %s" % replay_path)
def unit_max_shield(self, unit):
"""Returns maximal shield for a given unit."""
if unit.unit_type == 74 or unit.unit_type == self.stalker_id:
return 80 # Protoss's Stalker
if unit.unit_type == 73 or unit.unit_type == self.zealot_id:
return 50 # Protoss's Zealot
if unit.unit_type == 4 or unit.unit_type == self.colossus_id:
return 150 # Protoss's Colossus
if unit.unit_type == 77 or unit.unit_type == self.sentry_id:
return 40 # Protoss's Sentry
if unit.unit_type == self.void_ray_id:
return 100 # Protoss's Void Ray
def can_move(self, unit, direction):
"""Whether a unit can move in a given direction."""
m = self._move_amount / 2
if direction == Direction.NORTH:
x, y = int(unit.pos.x), int(unit.pos.y + m)
elif direction == Direction.SOUTH:
x, y = int(unit.pos.x), int(unit.pos.y - m)
elif direction == Direction.EAST:
x, y = int(unit.pos.x + m), int(unit.pos.y)
else:
x, y = int(unit.pos.x - m), int(unit.pos.y)
if self.check_bounds(x, y) and self.pathing_grid[x, y]:
return True
return False
def get_surrounding_points(self, unit, include_self=False):
"""Returns the surrounding points of the unit in 8 directions."""
x = int(unit.pos.x)
y = int(unit.pos.y)
ma = self._move_amount
points = [
(x, y + 2 * ma),
(x, y - 2 * ma),
(x + 2 * ma, y),
(x - 2 * ma, y),
(x + ma, y + ma),
(x - ma, y - ma),
(x + ma, y - ma),
(x - ma, y + ma),
]
if include_self:
points.append((x, y))
return points
def check_bounds(self, x, y):
"""Whether a point is within the map bounds."""
return (0 <= x < self.map_x and 0 <= y < self.map_y)
def get_surrounding_pathing(self, unit):
"""Returns pathing values of the grid surrounding the given unit."""
points = self.get_surrounding_points(unit, include_self=False)
vals = [
self.pathing_grid[x, y] if self.check_bounds(x, y) else 1
for x, y in points
]
return vals
def get_surrounding_height(self, unit):
"""Returns height values of the grid surrounding the given unit."""
points = self.get_surrounding_points(unit, include_self=True)
vals = [
self.terrain_height[x, y] if self.check_bounds(x, y) else 1
for x, y in points
]
return vals
def get_own_feature_size(self):
nf_own = self.unit_type_bits
if self.obs_own_health:
nf_own += 1 + self.shield_bits_ally
return nf_own
def get_units_type_id(self):
self.reset()
type_ids = []
for agent_i in range(self.n_agents):
agent = self.get_unit_by_id(agent_i)
type_ids.append(self.get_unit_type_id(agent, True))
print('>>>', type_ids)
return type_ids
def get_obs_agent(self, agent_id):
"""Returns observation for agent_id.
NOTE: Agents should have access only to their local observations
during decentralised execution.
"""
unit = self.get_unit_by_id(agent_id)
nf_al = 4 + self.unit_type_bits
nf_en = 4 + self.unit_type_bits
if self.obs_all_health:
nf_al += 1 + self.shield_bits_ally
nf_en += 1 + self.shield_bits_enemy
if self.obs_last_action:
nf_al += self.n_actions
nf_own = self.unit_type_bits
if self.obs_own_health:
nf_own += 1 + self.shield_bits_ally
move_feats_len = self.n_actions_move
if self.obs_pathing_grid:
move_feats_len += self.n_obs_pathing
if self.obs_terrain_height:
move_feats_len += self.n_obs_height
move_feats = np.zeros(move_feats_len, dtype=np.float32)
enemy_feats = np.zeros((self.n_enemies, nf_en), dtype=np.float32)
ally_feats = np.zeros((self.n_agents - 1, nf_al), dtype=np.float32)
own_feats = np.zeros(nf_own, dtype=np.float32)
if unit.health > 0: # otherwise dead, return all zeros
x = unit.pos.x
y = unit.pos.y
sight_range = self.unit_sight_range(agent_id)
# Movement features
avail_actions = self.get_avail_agent_actions(agent_id)
for m in range(self.n_actions_move):
move_feats[m] = avail_actions[m + 2]
ind = self.n_actions_move
if self.obs_pathing_grid:
move_feats[
ind: ind + self.n_obs_pathing
] = self.get_surrounding_pathing(unit)
ind += self.n_obs_pathing
if self.obs_terrain_height:
move_feats[ind:] = self.get_surrounding_height(unit)
# Enemy features
for e_id, e_unit in self.enemies.items():
e_x = e_unit.pos.x
e_y = e_unit.pos.y
dist = self.distance(x, y, e_x, e_y)
if (
dist < sight_range and e_unit.health > 0
): # visible and alive
# Sight range > shoot range
enemy_feats[e_id, 0] = avail_actions[
self.n_actions_no_attack + e_id
] # available
enemy_feats[e_id, 1] = dist / sight_range # distance
enemy_feats[e_id, 2] = (
e_x - x
) / sight_range # relative X
enemy_feats[e_id, 3] = (
e_y - y
) / sight_range # relative Y
ind = 4
if self.obs_all_health:
enemy_feats[e_id, ind] = (
e_unit.health / e_unit.health_max
) # health
ind += 1
if self.shield_bits_enemy > 0:
max_shield = self.unit_max_shield(e_unit)
enemy_feats[e_id, ind] = (
e_unit.shield / max_shield
) # shield
ind += 1
if self.unit_type_bits > 0:
type_id = self.get_unit_type_id(e_unit, False)
enemy_feats[e_id, ind + type_id] = 1 # unit type
# Ally features
al_ids = [
al_id for al_id in range(self.n_agents) if al_id != agent_id
]
for i, al_id in enumerate(al_ids):
al_unit = self.get_unit_by_id(al_id)
al_x = al_unit.pos.x
al_y = al_unit.pos.y
dist = self.distance(x, y, al_x, al_y)
if (
dist < sight_range and al_unit.health > 0
): # visible and alive
ally_feats[i, 0] = 1 # visible
ally_feats[i, 1] = dist / sight_range # distance
ally_feats[i, 2] = (al_x - x) / sight_range # relative X
ally_feats[i, 3] = (al_y - y) / sight_range # relative Y
ind = 4
if self.obs_all_health:
ally_feats[i, ind] = (
al_unit.health / al_unit.health_max
) # health
ind += 1
if self.shield_bits_ally > 0:
max_shield = self.unit_max_shield(al_unit)
ally_feats[i, ind] = (
al_unit.shield / max_shield
) # shield
ind += 1
if self.unit_type_bits > 0:
type_id = self.get_unit_type_id(al_unit, True)
ally_feats[i, ind + type_id] = 1
ind += self.unit_type_bits
if self.obs_last_action:
ally_feats[i, ind:] = self.last_action[al_id]
# Own features
ind = 0
if self.obs_own_health:
own_feats[ind] = unit.health / unit.health_max
ind += 1
if self.shield_bits_ally > 0:
max_shield = self.unit_max_shield(unit)
own_feats[ind] = unit.shield / max_shield
ind += 1
if self.unit_type_bits > 0:
type_id = self.get_unit_type_id(unit, True)
own_feats[ind + type_id] = 1
agent_obs = np.concatenate(
(
move_feats.flatten(),
enemy_feats.flatten(),
ally_feats.flatten(),
own_feats.flatten(),
)
)
if self.obs_timestep_number:
agent_obs = np.append(agent_obs,
self._episode_steps / self.episode_limit)
if self.debug:
logging.debug("Obs Agent: {}".format(agent_id).center(60, "-"))
logging.debug("Avail. actions {}".format(
self.get_avail_agent_actions(agent_id)))
logging.debug("Move feats {}".format(move_feats))
logging.debug("Enemy feats {}".format(enemy_feats))
logging.debug("Ally feats {}".format(ally_feats))
logging.debug("Own feats {}".format(own_feats))
return agent_obs
def get_obs(self):
"""Returns all agent observations in a list.
NOTE: Agents should have access only to their local observations
during decentralised execution.
"""
agents_obs = [self.get_obs_agent(i) for i in range(self.n_agents)]
return agents_obs
def get_state(self):
"""Returns the global state.
NOTE: This functon should not be used during decentralised execution.
"""
if self.obs_instead_of_state:
obs_concat = np.concatenate(self.get_obs(), axis=0).astype(
np.float32
)
return obs_concat
nf_al = 4 + self.shield_bits_ally + self.unit_type_bits
nf_en = 3 + self.shield_bits_enemy + self.unit_type_bits
ally_state = np.zeros((self.n_agents, nf_al))
enemy_state = np.zeros((self.n_enemies, nf_en))
center_x = self.map_x / 2
center_y = self.map_y / 2
for al_id, al_unit in self.agents.items():
if al_unit.health > 0:
x = al_unit.pos.x
y = al_unit.pos.y
max_cd = self.unit_max_cooldown(al_unit)
ally_state[al_id, 0] = (
al_unit.health / al_unit.health_max
) # health
if (
self.map_type in ["MMM", "GMMM"]
and al_unit.unit_type == self.medivac_id
):
ally_state[al_id, 1] = al_unit.energy / max_cd # energy
else:
ally_state[al_id, 1] = (
al_unit.weapon_cooldown / max_cd
) # cooldown
ally_state[al_id, 2] = (
x - center_x
) / self.max_distance_x # relative X
ally_state[al_id, 3] = (
y - center_y
) / self.max_distance_y # relative Y
ind = 4
if self.shield_bits_ally > 0:
max_shield = self.unit_max_shield(al_unit)
ally_state[al_id, ind] = (
al_unit.shield / max_shield
) # shield
ind += 1
if self.unit_type_bits > 0:
type_id = self.get_unit_type_id(al_unit, True)
ally_state[al_id, ind + type_id] = 1
for e_id, e_unit in self.enemies.items():
if e_unit.health > 0:
x = e_unit.pos.x
y = e_unit.pos.y
enemy_state[e_id, 0] = (
e_unit.health / e_unit.health_max
) # health
enemy_state[e_id, 1] = (
x - center_x
) / self.max_distance_x # relative X
enemy_state[e_id, 2] = (
y - center_y
) / self.max_distance_y # relative Y
ind = 3
if self.shield_bits_enemy > 0:
max_shield = self.unit_max_shield(e_unit)
enemy_state[e_id, ind] = (
e_unit.shield / max_shield
) # shield
ind += 1
if self.unit_type_bits > 0:
type_id = self.get_unit_type_id(e_unit, False)
enemy_state[e_id, ind + type_id] = 1
state = np.append(ally_state.flatten(), enemy_state.flatten())
if self.state_last_action:
state = np.append(state, self.last_action.flatten())
if self.state_timestep_number:
state = np.append(state,
self._episode_steps / self.episode_limit)
state = state.astype(dtype=np.float32)
if self.debug:
logging.debug("STATE".center(60, "-"))
logging.debug("Ally state {}".format(ally_state))
logging.debug("Enemy state {}".format(enemy_state))
if self.state_last_action:
logging.debug("Last actions {}".format(self.last_action))
return state
def get_obs_size(self):
"""Returns the size of the observation."""
nf_al = 4 + self.unit_type_bits
nf_en = 4 + self.unit_type_bits
if self.obs_all_health:
nf_al += 1 + self.shield_bits_ally
nf_en += 1 + self.shield_bits_enemy
own_feats = self.unit_type_bits
if self.obs_own_health:
own_feats += 1 + self.shield_bits_ally
if self.obs_timestep_number:
own_feats += 1
if self.obs_last_action:
nf_al += self.n_actions
move_feats = self.n_actions_move
if self.obs_pathing_grid:
move_feats += self.n_obs_pathing
if self.obs_terrain_height:
move_feats += self.n_obs_height
enemy_feats = self.n_enemies * nf_en
ally_feats = (self.n_agents - 1) * nf_al
return move_feats + enemy_feats + ally_feats + own_feats
def get_state_size(self):
"""Returns the size of the global state."""
if self.obs_instead_of_state:
return self.get_obs_size() * self.n_agents
nf_al = 4 + self.shield_bits_ally + self.unit_type_bits
nf_en = 3 + self.shield_bits_enemy + self.unit_type_bits
enemy_state = self.n_enemies * nf_en
ally_state = self.n_agents * nf_al
size = enemy_state + ally_state
if self.state_last_action:
size += self.n_agents * self.n_actions
if self.state_timestep_number:
size += 1
return size
def get_unit_type_id(self, unit, ally):
"""Returns the ID of unit type in the given scenario."""
if ally: # use new SC2 unit types
type_id = unit.unit_type - self._min_unit_type
else: # use default SC2 unit types
if self.map_type == "stalkers_and_zealots":
# id(Stalker) = 74, id(Zealot) = 73
type_id = unit.unit_type - 73
if self.map_type == "bane_vs_sz":
# id(Stalker) = 74, id(Zealot) = 73
type_id = unit.unit_type - 73
if self.map_type == "stalkers_and_zealots_vs_zb":
# id(Stalker) = 74, id() =
if unit.unit_type == 9:
type_id = 0
else:
type_id = 1
elif self.map_type == "colossi_stalkers_zealots":
# id(Stalker) = 74, id(Zealot) = 73, id(Colossus) = 4
if unit.unit_type == 4:
type_id = 0
elif unit.unit_type == 74:
type_id = 1
else:
type_id = 2
elif self.map_type == "stalkers_and_sentries":
# id(Stalker) = 74, id(Sentry) = 77
if unit.unit_type == 77:
type_id = 1
elif unit.unit_type == 74:
type_id = 0
elif self.map_type == "zv_mb":
# id(Battlecrusier) = 57, id(Marine) = 48
if unit.unit_type == 57:
type_id = 1
elif unit.unit_type == 48:
type_id = 0
elif self.map_type == "bane":
if unit.unit_type == 9:
type_id = 0
else:
type_id = 1
elif self.map_type == "MMM":
if unit.unit_type == 51:
type_id = 0
elif unit.unit_type == 48:
type_id = 1
else:
type_id = 2
elif self.map_type == "GMMM":
if unit.unit_type == 51:
type_id = 0
elif unit.unit_type == 48:
type_id = 1
elif unit.unit_type == 54:
type_id = 2
else:
type_id = 3
return type_id
def get_avail_agent_actions(self, agent_id):
"""Returns the available actions for agent_id."""
unit = self.get_unit_by_id(agent_id)
if unit.health > 0:
# cannot choose no-op when alive
avail_actions = [0] * self.n_actions
# stop should be allowed
avail_actions[1] = 1
# see if we can move
if self.can_move(unit, Direction.NORTH):
avail_actions[2] = 1
if self.can_move(unit, Direction.SOUTH):
avail_actions[3] = 1
if self.can_move(unit, Direction.EAST):
avail_actions[4] = 1
if self.can_move(unit, Direction.WEST):
avail_actions[5] = 1
# Can attack only alive units that are alive in the shooting range
shoot_range = self.unit_shoot_range(agent_id)
target_items = self.enemies.items()
if self.map_type in ["MMM", "GMMM"] and unit.unit_type == self.medivac_id:
# Medivacs cannot heal themselves or other flying units
target_items = [
(t_id, t_unit)
for (t_id, t_unit) in self.agents.items()
if t_unit.unit_type != self.medivac_id
]
for t_id, t_unit in target_items:
if t_unit.health > 0:
dist = self.distance(
unit.pos.x, unit.pos.y, t_unit.pos.x, t_unit.pos.y
)
if dist <= shoot_range:
avail_actions[t_id + self.n_actions_no_attack] = 1
return avail_actions
else:
# only no-op allowed
return [1] + [0] * (self.n_actions - 1)
def get_avail_actions(self):
"""Returns the available actions of all agents in a list."""
avail_actions = []
for agent_id in range(self.n_agents):
avail_agent = self.get_avail_agent_actions(agent_id)
avail_actions.append(avail_agent)
return avail_actions
def close(self):
"""Close StarCraft II."""
if self._sc2_proc:
self._sc2_proc.close()
def seed(self):
"""Returns the random seed used by the environment."""
return self._seed
def render(self):
"""Not implemented."""
pass
def _kill_all_units(self):
"""Kill all units on the map."""
units_alive = [
unit.tag for unit in self.agents.values() if unit.health > 0
] + [unit.tag for unit in self.enemies.values() if unit.health > 0]
debug_command = [
d_pb.DebugCommand(kill_unit=d_pb.DebugKillUnit(tag=units_alive))
]
self._controller.debug(debug_command)
def init_units(self):
"""Initialise the units."""
while True:
# Sometimes not all units have yet been created by SC2
self.agents = {}
self.enemies = {}
ally_units = [
unit
for unit in self._obs.observation.raw_data.units
if unit.owner == 1
]
ally_units_sorted = sorted(
ally_units,
key=attrgetter("unit_type", "pos.x", "pos.y"),
reverse=False,
)
for i in range(len(ally_units_sorted)):
self.agents[i] = ally_units_sorted[i]
if self.debug:
logging.debug(
"Unit {} is {}, x = {}, y = {}".format(
len(self.agents),
self.agents[i].unit_type,
self.agents[i].pos.x,
self.agents[i].pos.y,
)
)
for unit in self._obs.observation.raw_data.units:
if unit.owner == 2:
self.enemies[len(self.enemies)] = unit
if self._episode_count == 0:
self.max_reward += unit.health_max + unit.shield_max
if self._episode_count == 0:
min_unit_type = min(
unit.unit_type for unit in self.agents.values()
)
self._init_ally_unit_types(min_unit_type)
all_agents_created = (len(self.agents) == self.n_agents)
all_enemies_created = (len(self.enemies) == self.n_enemies)
if all_agents_created and all_enemies_created: # all good
return
try:
self._controller.step(1)
self._obs = self._controller.observe()
except (protocol.ProtocolError, protocol.ConnectionError):
self.full_restart()
self.reset()
def update_units(self):
"""Update units after an environment step.
This function assumes that self._obs is up-to-date.
"""
n_ally_alive = 0
n_enemy_alive = 0
# Store previous state
self.previous_ally_units = deepcopy(self.agents)
self.previous_enemy_units = deepcopy(self.enemies)
for al_id, al_unit in self.agents.items():
updated = False
for unit in self._obs.observation.raw_data.units:
if al_unit.tag == unit.tag:
self.agents[al_id] = unit
updated = True
n_ally_alive += 1
break
if not updated: # dead
al_unit.health = 0
for e_id, e_unit in self.enemies.items():
updated = False
for unit in self._obs.observation.raw_data.units:
if e_unit.tag == unit.tag:
self.enemies[e_id] = unit
updated = True
n_enemy_alive += 1
break
if not updated: # dead
e_unit.health = 0
if (n_ally_alive == 0 and n_enemy_alive > 0
or self.only_medivac_left(ally=True)):
return -1 # lost
if (n_ally_alive > 0 and n_enemy_alive == 0
or self.only_medivac_left(ally=False)):
return 1 # won
if n_ally_alive == 0 and n_enemy_alive == 0:
return 0
return None
def _init_ally_unit_types(self, min_unit_type):
"""Initialise ally unit types. Should be called once from the
init_units function.
"""
self._min_unit_type = min_unit_type
if self.map_type == "marines":
self.marine_id = min_unit_type
elif self.map_type == "stalkers_and_zealots":
self.stalker_id = min_unit_type
self.zealot_id = min_unit_type + 1
elif self.map_type == "stalkers_and_zealots_vs_zb":
self.stalker_id = min_unit_type
self.zealot_id = min_unit_type + 1
elif self.map_type == "stalkers_and_sentries":
self.stalker_id = min_unit_type + 1
self.sentry_id = min_unit_type
elif self.map_type == "colossi_stalkers_zealots":
self.colossus_id = min_unit_type
self.stalker_id = min_unit_type + 1
self.zealot_id = min_unit_type + 2
elif self.map_type == "zv_mb":
self.void_ray_id = min_unit_type
self.zealot_id = min_unit_type + 1
elif self.map_type == "MMM":
self.marauder_id = min_unit_type
self.marine_id = min_unit_type + 1
self.medivac_id = min_unit_type + 2
elif self.map_type == 'GMMM':
self.marauder_id = min_unit_type
self.marine_id = min_unit_type + 1
self.medivac_id = min_unit_type + 2
self.ghost_id = min_unit_type + 3
elif self.map_type == "zealots":
self.zealot_id = min_unit_type
elif self.map_type == "hydralisks":
self.hydralisk_id = min_unit_type
elif self.map_type == "stalkers":
self.stalker_id = min_unit_type
elif self.map_type == "colossus":
self.colossus_id = min_unit_type
elif self.map_type == "bane":
self.baneling_id = min_unit_type
self.zergling_id = min_unit_type + 1
elif self.map_type == "bane_vs_sz":
self.baneling_id = min_unit_type
self.zergling_id = min_unit_type + 1
def only_medivac_left(self, ally):
"""Check if only Medivac units are left."""
if self.map_type not in ["MMM", "GMMM"]:
return False
if ally:
units_alive = [
a
for a in self.agents.values()
if (a.health > 0 and a.unit_type != self.medivac_id)
]
if len(units_alive) == 0:
return True
return False
else:
units_alive = [
a
for a in self.enemies.values()
if (a.health > 0 and a.unit_type != self.medivac_id)
]
if len(units_alive) == 1 and units_alive[0].unit_type == 54:
return True
return False
def get_unit_by_id(self, a_id):
"""Get unit by ID."""
return self.agents[a_id]
def get_stats(self):
stats = {
"battles_won": self.battles_won,
"battles_game": self.battles_game,
"battles_draw": self.timeouts,
"win_rate": self.battles_won / self.battles_game,
"timeouts": self.timeouts,
"restarts": self.force_restarts,
}
return stats
|
[
"operator.attrgetter",
"numpy.eye",
"s2clientprotocol.raw_pb2.ActionRawUnitCommand",
"copy.deepcopy",
"s2clientprotocol.sc2api_pb2.InterfaceOptions",
"s2clientprotocol.sc2api_pb2.RequestJoinGame",
"absl.logging.info",
"time.sleep",
"s2clientprotocol.raw_pb2.ActionRaw",
"numpy.append",
"numpy.array",
"numpy.zeros",
"pysc2.maps.get",
"s2clientprotocol.debug_pb2.DebugKillUnit",
"pysc2.run_configs.get",
"s2clientprotocol.sc2api_pb2.RequestAction",
"math.hypot",
"s2clientprotocol.common_pb2.Point2D"
] |
[((7454, 7469), 'time.sleep', 'time.sleep', (['(100)'], {}), '(100)\n', (7464, 7469), False, 'import math, time\n'), ((10200, 10223), 'numpy.zeros', 'np.zeros', (['self.n_agents'], {}), '(self.n_agents)\n', (10208, 10223), True, 'import numpy as np\n'), ((10259, 10283), 'numpy.zeros', 'np.zeros', (['self.n_enemies'], {}), '(self.n_enemies)\n', (10267, 10283), True, 'import numpy as np\n'), ((10392, 10433), 'numpy.zeros', 'np.zeros', (['(self.n_agents, self.n_actions)'], {}), '((self.n_agents, self.n_actions))\n', (10400, 10433), True, 'import numpy as np\n'), ((11251, 11268), 'pysc2.run_configs.get', 'run_configs.get', ([], {}), '()\n', (11266, 11268), False, 'from pysc2 import run_configs\n'), ((11284, 11307), 'pysc2.maps.get', 'maps.get', (['self.map_name'], {}), '(self.map_name)\n', (11292, 11307), False, 'from pysc2 import maps\n'), ((11372, 11417), 's2clientprotocol.sc2api_pb2.InterfaceOptions', 'sc_pb.InterfaceOptions', ([], {'raw': '(True)', 'score': '(False)'}), '(raw=True, score=False)\n', (11394, 11417), True, 'from s2clientprotocol import sc2api_pb2 as sc_pb\n'), ((12103, 12181), 's2clientprotocol.sc2api_pb2.RequestJoinGame', 'sc_pb.RequestJoinGame', ([], {'race': 'races[self._agent_race]', 'options': 'interface_options'}), '(race=races[self._agent_race], options=interface_options)\n', (12124, 12181), True, 'from s2clientprotocol import sc2api_pb2 as sc_pb\n'), ((13834, 13857), 'numpy.zeros', 'np.zeros', (['self.n_agents'], {}), '(self.n_agents)\n', (13842, 13857), True, 'import numpy as np\n'), ((13893, 13917), 'numpy.zeros', 'np.zeros', (['self.n_enemies'], {}), '(self.n_enemies)\n', (13901, 13917), True, 'import numpy as np\n'), ((14096, 14137), 'numpy.zeros', 'np.zeros', (['(self.n_agents, self.n_actions)'], {}), '((self.n_agents, self.n_actions))\n', (14104, 14137), True, 'import numpy as np\n'), ((16410, 16449), 's2clientprotocol.sc2api_pb2.RequestAction', 'sc_pb.RequestAction', ([], {'actions': 'sc_actions'}), '(actions=sc_actions)\n', (16429, 16449), True, 'from s2clientprotocol import sc2api_pb2 as sc_pb\n'), ((24693, 24810), 's2clientprotocol.raw_pb2.ActionRawUnitCommand', 'r_pb.ActionRawUnitCommand', ([], {'ability_id': 'action_id', 'target_unit_tag': 'target_tag', 'unit_tags': '[tag]', 'queue_command': '(False)'}), '(ability_id=action_id, target_unit_tag=target_tag,\n unit_tags=[tag], queue_command=False)\n', (24718, 24810), True, 'from s2clientprotocol import raw_pb2 as r_pb\n'), ((27446, 27474), 'math.hypot', 'math.hypot', (['(x2 - x1)', '(y2 - y1)'], {}), '(x2 - x1, y2 - y1)\n', (27456, 27474), False, 'import math, time\n'), ((28551, 28600), 'absl.logging.info', 'logging.info', (["('Replay saved at: %s' % replay_path)"], {}), "('Replay saved at: %s' % replay_path)\n", (28563, 28600), False, 'from absl import logging\n'), ((32616, 32658), 'numpy.zeros', 'np.zeros', (['move_feats_len'], {'dtype': 'np.float32'}), '(move_feats_len, dtype=np.float32)\n', (32624, 32658), True, 'import numpy as np\n'), ((32681, 32732), 'numpy.zeros', 'np.zeros', (['(self.n_enemies, nf_en)'], {'dtype': 'np.float32'}), '((self.n_enemies, nf_en), dtype=np.float32)\n', (32689, 32732), True, 'import numpy as np\n'), ((32754, 32808), 'numpy.zeros', 'np.zeros', (['(self.n_agents - 1, nf_al)'], {'dtype': 'np.float32'}), '((self.n_agents - 1, nf_al), dtype=np.float32)\n', (32762, 32808), True, 'import numpy as np\n'), ((32829, 32863), 'numpy.zeros', 'np.zeros', (['nf_own'], {'dtype': 'np.float32'}), '(nf_own, dtype=np.float32)\n', (32837, 32863), True, 'import numpy as np\n'), ((39210, 39242), 'numpy.zeros', 'np.zeros', (['(self.n_agents, nf_al)'], {}), '((self.n_agents, nf_al))\n', (39218, 39242), True, 'import numpy as np\n'), ((39265, 39298), 'numpy.zeros', 'np.zeros', (['(self.n_enemies, nf_en)'], {}), '((self.n_enemies, nf_en))\n', (39273, 39298), True, 'import numpy as np\n'), ((51741, 51762), 'copy.deepcopy', 'deepcopy', (['self.agents'], {}), '(self.agents)\n', (51749, 51762), False, 'from copy import deepcopy\n'), ((51799, 51821), 'copy.deepcopy', 'deepcopy', (['self.enemies'], {}), '(self.enemies)\n', (51807, 51821), False, 'from copy import deepcopy\n'), ((15851, 15873), 'numpy.eye', 'np.eye', (['self.n_actions'], {}), '(self.n_actions)\n', (15857, 15873), True, 'import numpy as np\n'), ((15874, 15891), 'numpy.array', 'np.array', (['actions'], {}), '(actions)\n', (15882, 15891), True, 'import numpy as np\n'), ((37839, 37901), 'numpy.append', 'np.append', (['agent_obs', '(self._episode_steps / self.episode_limit)'], {}), '(agent_obs, self._episode_steps / self.episode_limit)\n', (37848, 37901), True, 'import numpy as np\n'), ((42266, 42324), 'numpy.append', 'np.append', (['state', '(self._episode_steps / self.episode_limit)'], {}), '(state, self._episode_steps / self.episode_limit)\n', (42275, 42324), True, 'import numpy as np\n'), ((19662, 19757), 's2clientprotocol.raw_pb2.ActionRawUnitCommand', 'r_pb.ActionRawUnitCommand', ([], {'ability_id': "actions['stop']", 'unit_tags': '[tag]', 'queue_command': '(False)'}), "(ability_id=actions['stop'], unit_tags=[tag],\n queue_command=False)\n", (19687, 19757), True, 'from s2clientprotocol import raw_pb2 as r_pb\n'), ((22425, 22457), 's2clientprotocol.raw_pb2.ActionRaw', 'r_pb.ActionRaw', ([], {'unit_command': 'cmd'}), '(unit_command=cmd)\n', (22439, 22457), True, 'from s2clientprotocol import raw_pb2 as r_pb\n'), ((22877, 22929), 'math.hypot', 'math.hypot', (['self.max_distance_x', 'self.max_distance_y'], {}), '(self.max_distance_x, self.max_distance_y)\n', (22887, 22929), False, 'import math, time\n'), ((23895, 23947), 'math.hypot', 'math.hypot', (['self.max_distance_x', 'self.max_distance_y'], {}), '(self.max_distance_x, self.max_distance_y)\n', (23905, 23947), False, 'import math, time\n'), ((24901, 24933), 's2clientprotocol.raw_pb2.ActionRaw', 'r_pb.ActionRaw', ([], {'unit_command': 'cmd'}), '(unit_command=cmd)\n', (24915, 24933), True, 'from s2clientprotocol import raw_pb2 as r_pb\n'), ((49351, 49386), 's2clientprotocol.debug_pb2.DebugKillUnit', 'd_pb.DebugKillUnit', ([], {'tag': 'units_alive'}), '(tag=units_alive)\n', (49369, 49386), True, 'from s2clientprotocol import debug_pb2 as d_pb\n'), ((49904, 49945), 'operator.attrgetter', 'attrgetter', (['"""unit_type"""', '"""pos.x"""', '"""pos.y"""'], {}), "('unit_type', 'pos.x', 'pos.y')\n", (49914, 49945), False, 'from operator import attrgetter\n'), ((20071, 20118), 's2clientprotocol.common_pb2.Point2D', 'sc_common.Point2D', ([], {'x': 'x', 'y': '(y + self._move_amount)'}), '(x=x, y=y + self._move_amount)\n', (20088, 20118), True, 'from s2clientprotocol import common_pb2 as sc_common\n'), ((20485, 20532), 's2clientprotocol.common_pb2.Point2D', 'sc_common.Point2D', ([], {'x': 'x', 'y': '(y - self._move_amount)'}), '(x=x, y=y - self._move_amount)\n', (20502, 20532), True, 'from s2clientprotocol import common_pb2 as sc_common\n'), ((22059, 22176), 's2clientprotocol.raw_pb2.ActionRawUnitCommand', 'r_pb.ActionRawUnitCommand', ([], {'ability_id': 'action_id', 'target_unit_tag': 'target_tag', 'unit_tags': '[tag]', 'queue_command': '(False)'}), '(ability_id=action_id, target_unit_tag=target_tag,\n unit_tags=[tag], queue_command=False)\n', (22084, 22176), True, 'from s2clientprotocol import raw_pb2 as r_pb\n'), ((20898, 20945), 's2clientprotocol.common_pb2.Point2D', 'sc_common.Point2D', ([], {'x': '(x + self._move_amount)', 'y': 'y'}), '(x=x + self._move_amount, y=y)\n', (20915, 20945), True, 'from s2clientprotocol import common_pb2 as sc_common\n'), ((21310, 21357), 's2clientprotocol.common_pb2.Point2D', 'sc_common.Point2D', ([], {'x': '(x - self._move_amount)', 'y': 'y'}), '(x=x - self._move_amount, y=y)\n', (21327, 21357), True, 'from s2clientprotocol import common_pb2 as sc_common\n')]
|
import numbers
from typing import Any, Dict, List, Optional, Sequence, Tuple, Union
import numpy as np
import torch
from PIL import Image, ImageOps, ImageEnhance
from typing_extensions import Literal
try:
import accimage
except ImportError:
accimage = None
@torch.jit.unused
def _is_pil_image(img: Any) -> bool:
if accimage is not None:
return isinstance(img, (Image.Image, accimage.Image))
else:
return isinstance(img, Image.Image)
@torch.jit.unused
def get_image_size(img: Any) -> List[int]:
if _is_pil_image(img):
return list(img.size)
raise TypeError(f"Unexpected type {type(img)}")
@torch.jit.unused
def get_image_num_channels(img: Any) -> int:
if _is_pil_image(img):
return 1 if img.mode == "L" else 3
raise TypeError(f"Unexpected type {type(img)}")
@torch.jit.unused
def hflip(img: Image.Image) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
return img.transpose(Image.FLIP_LEFT_RIGHT)
@torch.jit.unused
def vflip(img: Image.Image) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
return img.transpose(Image.FLIP_TOP_BOTTOM)
@torch.jit.unused
def adjust_brightness(img: Image.Image, brightness_factor: float) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
enhancer = ImageEnhance.Brightness(img)
img = enhancer.enhance(brightness_factor)
return img
@torch.jit.unused
def adjust_contrast(img: Image.Image, contrast_factor: float) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
enhancer = ImageEnhance.Contrast(img)
img = enhancer.enhance(contrast_factor)
return img
@torch.jit.unused
def adjust_saturation(img: Image.Image, saturation_factor: float) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
enhancer = ImageEnhance.Color(img)
img = enhancer.enhance(saturation_factor)
return img
@torch.jit.unused
def adjust_hue(img: Image.Image, hue_factor: float) -> Image.Image:
if not (-0.5 <= hue_factor <= 0.5):
raise ValueError(f"hue_factor ({hue_factor}) is not in [-0.5, 0.5].")
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
input_mode = img.mode
if input_mode in {"L", "1", "I", "F"}:
return img
h, s, v = img.convert("HSV").split()
np_h = np.array(h, dtype=np.uint8)
# uint8 addition take cares of rotation across boundaries
with np.errstate(over="ignore"):
np_h += np.uint8(hue_factor * 255)
h = Image.fromarray(np_h, "L")
img = Image.merge("HSV", (h, s, v)).convert(input_mode)
return img
@torch.jit.unused
def adjust_gamma(
img: Image.Image,
gamma: float,
gain: float = 1.0,
) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
if gamma < 0:
raise ValueError("Gamma should be a non-negative real number")
input_mode = img.mode
img = img.convert("RGB")
gamma_map = [int((255 + 1 - 1e-3) * gain * pow(ele / 255.0, gamma)) for ele in range(256)] * 3
img = img.point(gamma_map) # use PIL's point-function to accelerate this part
img = img.convert(input_mode)
return img
@torch.jit.unused
def pad(
img: Image.Image,
padding: Union[int, List[int], Tuple[int, ...]],
fill: Optional[Union[float, List[float], Tuple[float, ...]]] = 0,
padding_mode: Literal["constant", "edge", "reflect", "symmetric"] = "constant",
) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
if not isinstance(padding, (numbers.Number, tuple, list)):
raise TypeError("Got inappropriate padding arg")
if not isinstance(fill, (numbers.Number, str, tuple)):
raise TypeError("Got inappropriate fill arg")
if not isinstance(padding_mode, str):
raise TypeError("Got inappropriate padding_mode arg")
if isinstance(padding, list):
padding = tuple(padding)
if isinstance(padding, tuple) and len(padding) not in [1, 2, 4]:
raise ValueError(f"Padding must be an int or a 1, 2, or 4 element tuple, not a {len(padding)} element tuple")
if isinstance(padding, tuple) and len(padding) == 1:
# Compatibility with `functional_tensor.pad`
padding = padding[0]
if padding_mode not in ["constant", "edge", "reflect", "symmetric"]:
raise ValueError("Padding mode should be either constant, edge, reflect or symmetric")
if padding_mode == "constant":
opts = _parse_fill(fill, img, name="fill")
if img.mode == "P":
palette = img.getpalette()
image = ImageOps.expand(img, border=padding, **opts)
image.putpalette(palette)
return image
return ImageOps.expand(img, border=padding, **opts)
else:
if isinstance(padding, int):
pad_left = pad_right = pad_top = pad_bottom = padding
if isinstance(padding, tuple) and len(padding) == 2:
pad_left = pad_right = padding[0]
pad_top = pad_bottom = padding[1]
if isinstance(padding, tuple) and len(padding) == 4:
pad_left = padding[0]
pad_top = padding[1]
pad_right = padding[2]
pad_bottom = padding[3]
p = [pad_left, pad_top, pad_right, pad_bottom]
cropping = -np.minimum(p, 0)
if cropping.any():
crop_left, crop_top, crop_right, crop_bottom = cropping
img = img.crop((crop_left, crop_top, img.width - crop_right, img.height - crop_bottom))
pad_left, pad_top, pad_right, pad_bottom = np.maximum(p, 0)
if img.mode == "P":
palette = img.getpalette()
img = np.asarray(img)
img = np.pad(img, ((pad_top, pad_bottom), (pad_left, pad_right)), mode=padding_mode)
img = Image.fromarray(img)
img.putpalette(palette)
return img
img = np.asarray(img)
# RGB image
if len(img.shape) == 3:
img = np.pad(img, ((pad_top, pad_bottom), (pad_left, pad_right), (0, 0)), padding_mode)
# Grayscale image
if len(img.shape) == 2:
img = np.pad(img, ((pad_top, pad_bottom), (pad_left, pad_right)), padding_mode)
return Image.fromarray(img)
@torch.jit.unused
def crop(
img: Image.Image,
top: int,
left: int,
height: int,
width: int,
) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
return img.crop((left, top, left + width, top + height))
@torch.jit.unused
def resize(
img: Image.Image,
size: Union[Sequence[int], int],
interpolation: int = Image.BILINEAR,
max_size: Optional[int] = None,
) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
if not (isinstance(size, int) or (isinstance(size, Sequence) and len(size) in (1, 2))):
raise TypeError(f"Got inappropriate size arg: {size}")
if isinstance(size, Sequence) and len(size) == 1:
size = size[0]
if isinstance(size, int):
w, h = img.size
short, long = (w, h) if w <= h else (h, w)
if short == size:
return img
new_short, new_long = size, int(size * long / short)
if max_size is not None:
if max_size <= size:
raise ValueError(
f"max_size = {max_size} must be strictly greater than the requested "
f"size for the smaller edge size = {size}"
)
if new_long > max_size:
new_short, new_long = int(max_size * new_short / new_long), max_size
new_w, new_h = (new_short, new_long) if w <= h else (new_long, new_short)
return img.resize((new_w, new_h), interpolation)
else:
if max_size is not None:
raise ValueError(
"max_size should only be passed if size specifies the length of the smaller edge, "
"i.e. size should be an int or a sequence of length 1 in torchscript mode."
)
return img.resize(size[::-1], interpolation)
@torch.jit.unused
def _parse_fill(
fill: Optional[Union[float, List[float], Tuple[float, ...]]],
img: Image.Image,
name: str = "fillcolor",
) -> Dict[str, Optional[Union[float, List[float], Tuple[float, ...]]]]:
# Process fill color for affine transforms
num_bands = len(img.getbands())
if fill is None:
fill = 0
if isinstance(fill, (int, float)) and num_bands > 1:
fill = tuple([fill] * num_bands)
if isinstance(fill, (list, tuple)):
if len(fill) != num_bands:
msg = "The number of elements in 'fill' does not match the number of bands of the image ({} != {})"
raise ValueError(msg.format(len(fill), num_bands))
fill = tuple(fill)
return {name: fill}
@torch.jit.unused
def affine(
img: Image.Image,
matrix: List[float],
interpolation: int = Image.NEAREST,
fill: Optional[Union[float, List[float], Tuple[float, ...]]] = 0,
) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
output_size = img.size
opts = _parse_fill(fill, img)
return img.transform(output_size, Image.AFFINE, matrix, interpolation, **opts)
@torch.jit.unused
def rotate(
img: Image.Image,
angle: float,
interpolation: int = Image.NEAREST,
expand: bool = False,
center: Optional[Tuple[int, int]] = None,
fill: Optional[Union[float, List[float], Tuple[float, ...]]] = 0,
) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
opts = _parse_fill(fill, img)
return img.rotate(angle, interpolation, expand, center, **opts)
@torch.jit.unused
def perspective(
img: Image.Image,
perspective_coeffs: float,
interpolation: int = Image.BICUBIC,
fill: Optional[Union[float, List[float], Tuple[float, ...]]] = 0,
) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
opts = _parse_fill(fill, img)
return img.transform(img.size, Image.PERSPECTIVE, perspective_coeffs, interpolation, **opts)
@torch.jit.unused
def to_grayscale(img: Image.Image, num_output_channels: int) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
if num_output_channels == 1:
img = img.convert("L")
elif num_output_channels == 3:
img = img.convert("L")
np_img = np.array(img, dtype=np.uint8)
np_img = np.dstack([np_img, np_img, np_img])
img = Image.fromarray(np_img, "RGB")
else:
raise ValueError("num_output_channels should be either 1 or 3")
return img
@torch.jit.unused
def invert(img: Image.Image) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
return ImageOps.invert(img)
@torch.jit.unused
def posterize(img: Image.Image, bits: int) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
return ImageOps.posterize(img, bits)
@torch.jit.unused
def solarize(img: Image.Image, threshold: int) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
return ImageOps.solarize(img, threshold)
@torch.jit.unused
def adjust_sharpness(img: Image.Image, sharpness_factor: float) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
enhancer = ImageEnhance.Sharpness(img)
img = enhancer.enhance(sharpness_factor)
return img
@torch.jit.unused
def autocontrast(img: Image.Image) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
return ImageOps.autocontrast(img)
@torch.jit.unused
def equalize(img: Image.Image) -> Image.Image:
if not _is_pil_image(img):
raise TypeError(f"img should be PIL Image. Got {type(img)}")
return ImageOps.equalize(img)
|
[
"numpy.uint8",
"PIL.ImageEnhance.Contrast",
"numpy.array",
"PIL.ImageOps.posterize",
"PIL.ImageOps.autocontrast",
"PIL.ImageOps.expand",
"numpy.asarray",
"PIL.ImageEnhance.Sharpness",
"PIL.ImageEnhance.Color",
"PIL.ImageOps.invert",
"numpy.maximum",
"PIL.ImageOps.equalize",
"PIL.ImageOps.solarize",
"PIL.ImageEnhance.Brightness",
"numpy.dstack",
"PIL.Image.fromarray",
"numpy.minimum",
"numpy.errstate",
"numpy.pad",
"PIL.Image.merge"
] |
[((1472, 1500), 'PIL.ImageEnhance.Brightness', 'ImageEnhance.Brightness', (['img'], {}), '(img)\n', (1495, 1500), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((1776, 1802), 'PIL.ImageEnhance.Contrast', 'ImageEnhance.Contrast', (['img'], {}), '(img)\n', (1797, 1802), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((2080, 2103), 'PIL.ImageEnhance.Color', 'ImageEnhance.Color', (['img'], {}), '(img)\n', (2098, 2103), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((2615, 2642), 'numpy.array', 'np.array', (['h'], {'dtype': 'np.uint8'}), '(h, dtype=np.uint8)\n', (2623, 2642), True, 'import numpy as np\n'), ((2793, 2819), 'PIL.Image.fromarray', 'Image.fromarray', (['np_h', '"""L"""'], {}), "(np_h, 'L')\n", (2808, 2819), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((11381, 11401), 'PIL.ImageOps.invert', 'ImageOps.invert', (['img'], {}), '(img)\n', (11396, 11401), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((11592, 11621), 'PIL.ImageOps.posterize', 'ImageOps.posterize', (['img', 'bits'], {}), '(img, bits)\n', (11610, 11621), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((11816, 11849), 'PIL.ImageOps.solarize', 'ImageOps.solarize', (['img', 'threshold'], {}), '(img, threshold)\n', (11833, 11849), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((12066, 12093), 'PIL.ImageEnhance.Sharpness', 'ImageEnhance.Sharpness', (['img'], {}), '(img)\n', (12088, 12093), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((12336, 12362), 'PIL.ImageOps.autocontrast', 'ImageOps.autocontrast', (['img'], {}), '(img)\n', (12357, 12362), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((12541, 12563), 'PIL.ImageOps.equalize', 'ImageOps.equalize', (['img'], {}), '(img)\n', (12558, 12563), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((2714, 2740), 'numpy.errstate', 'np.errstate', ([], {'over': '"""ignore"""'}), "(over='ignore')\n", (2725, 2740), True, 'import numpy as np\n'), ((2758, 2784), 'numpy.uint8', 'np.uint8', (['(hue_factor * 255)'], {}), '(hue_factor * 255)\n', (2766, 2784), True, 'import numpy as np\n'), ((5072, 5116), 'PIL.ImageOps.expand', 'ImageOps.expand', (['img'], {'border': 'padding'}), '(img, border=padding, **opts)\n', (5087, 5116), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((5923, 5939), 'numpy.maximum', 'np.maximum', (['p', '(0)'], {}), '(p, 0)\n', (5933, 5939), True, 'import numpy as np\n'), ((6252, 6267), 'numpy.asarray', 'np.asarray', (['img'], {}), '(img)\n', (6262, 6267), True, 'import numpy as np\n'), ((6586, 6606), 'PIL.Image.fromarray', 'Image.fromarray', (['img'], {}), '(img)\n', (6601, 6606), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((2831, 2860), 'PIL.Image.merge', 'Image.merge', (['"""HSV"""', '(h, s, v)'], {}), "('HSV', (h, s, v))\n", (2842, 2860), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((4948, 4992), 'PIL.ImageOps.expand', 'ImageOps.expand', (['img'], {'border': 'padding'}), '(img, border=padding, **opts)\n', (4963, 4992), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((5658, 5674), 'numpy.minimum', 'np.minimum', (['p', '(0)'], {}), '(p, 0)\n', (5668, 5674), True, 'import numpy as np\n'), ((6026, 6041), 'numpy.asarray', 'np.asarray', (['img'], {}), '(img)\n', (6036, 6041), True, 'import numpy as np\n'), ((6060, 6138), 'numpy.pad', 'np.pad', (['img', '((pad_top, pad_bottom), (pad_left, pad_right))'], {'mode': 'padding_mode'}), '(img, ((pad_top, pad_bottom), (pad_left, pad_right)), mode=padding_mode)\n', (6066, 6138), True, 'import numpy as np\n'), ((6157, 6177), 'PIL.Image.fromarray', 'Image.fromarray', (['img'], {}), '(img)\n', (6172, 6177), False, 'from PIL import Image, ImageOps, ImageEnhance\n'), ((6338, 6423), 'numpy.pad', 'np.pad', (['img', '((pad_top, pad_bottom), (pad_left, pad_right), (0, 0))', 'padding_mode'], {}), '(img, ((pad_top, pad_bottom), (pad_left, pad_right), (0, 0)),\n padding_mode)\n', (6344, 6423), True, 'import numpy as np\n'), ((6496, 6569), 'numpy.pad', 'np.pad', (['img', '((pad_top, pad_bottom), (pad_left, pad_right))', 'padding_mode'], {}), '(img, ((pad_top, pad_bottom), (pad_left, pad_right)), padding_mode)\n', (6502, 6569), True, 'import numpy as np\n'), ((10979, 11008), 'numpy.array', 'np.array', (['img'], {'dtype': 'np.uint8'}), '(img, dtype=np.uint8)\n', (10987, 11008), True, 'import numpy as np\n'), ((11026, 11061), 'numpy.dstack', 'np.dstack', (['[np_img, np_img, np_img]'], {}), '([np_img, np_img, np_img])\n', (11035, 11061), True, 'import numpy as np\n'), ((11076, 11106), 'PIL.Image.fromarray', 'Image.fromarray', (['np_img', '"""RGB"""'], {}), "(np_img, 'RGB')\n", (11091, 11106), False, 'from PIL import Image, ImageOps, ImageEnhance\n')]
|
import numpy as np
from pyray.shapes.twod.paraboloid import *
from pyray.shapes.twod.functional import *
from pyray.rotation import *
from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt
from matplotlib import cm
from matplotlib.ticker import LinearLocator, FormatStrFormatter
import matplotlib as mpl
import os
basedir = '.\\Images\\RotatingCube\\'
if os.name == 'posix':
basedir = 'Images/RotatingCube/'
def draw_cubic():
fn = lambda x,y: x**3+y**3
for i in range(20):
im = Image.new("RGB", (2048, 2048), "black")
draw = ImageDraw.Draw(im, 'RGBA')
r = general_rotation(np.array([1,0,0]),np.pi/120*i)
#drawFunctionalXYGridInCircle(draw, r, fn=fn, scale=10.0)
im.save(basedir + 'im' + str(i) + '.png')
def three_d_grid():
fig = plt.figure()
ax = fig.gca(projection='3d')
# Make data.
X = np.arange(-5, 5, 0.25)
Y = np.arange(-5, 5, 0.25)
X, Y = np.meshgrid(X, Y)
R = (X**3 + Y**3)
Z = R
# Plot the surface.
surf = ax.plot_surface(X, Y, Z, cmap=cm.coolwarm,
linewidth=0, antialiased=False)
# Customize the z axis.
#ax.set_zlim(-1.01, 1.01)
#ax.zaxis.set_major_locator(LinearLocator(10))
#ax.zaxis.set_major_formatter(FormatStrFormatter('%.02f'))
# Add a color bar which maps values to colors.
fig.colorbar(surf, shrink=0.5, aspect=5)
plt.show()
mpl.rcParams['legend.fontsize'] = 10
fig = plt.figure()
ax = fig.gca(projection='3d')
theta = np.linspace(0, 2 * np.pi, 100)
for r in np.arange(0.1,1.0,0.1):
#r = 1.0
x = r * np.sin(theta)
y = r * np.cos(theta)
z = x**3+y**3
ax.plot(x, y, z, label='parametric curve')
#ax.legend()
plt.show()
def paraboloid_w_grad(im_ind=0, scale=200, shift=np.array([1000,1000,0]), opacity=60,
basepath='.\\'):
r1 = np.eye(4)
rot = general_rotation(np.array([0,0,1]), np.pi/20.0 * (8 + im_ind/3.0))
j=4
r = rotation(3, 2 * np.pi* j /30.0)
rr = general_rotation(np.array([0,1,0]), np.pi/20.0 * (im_ind/7.0))
r = np.dot(r,rr)
r = np.dot(r, rot)
r1[:3,:3] = r
im = Image.new("RGB", (2048, 2048), "black")
draw = ImageDraw.Draw(im, 'RGBA')
render_scene_4d_axis(draw, r1, 4, scale, shift)
# This is what draws the pink paraboloid.
for z in np.arange(0.001, 3.5, 0.02):
point1 = np.array([np.sqrt(z),0,z])
generalized_arc(draw, r, center=np.array([0,0,z]), vec=np.array([0,0,1]),
point=point1, radius=np.sqrt(z), prcnt=1.0,
rgba=(255,20,147,50))
xax1=np.array([-100.0,0,0.0]);xax1=np.dot(r,xax1)*scale+shift
xax2=np.array([100.0,0,0.0]);xax2=np.dot(r,xax2)*scale+shift
draw.line((xax1[0], xax1[1], xax2[0], xax2[1]), fill=(255,255,0), width=4)
xax1=np.array([0.0,-100,0.0]);xax1=np.dot(r,xax1)*scale+shift
xax2=np.array([0.0,100,0.0]);xax2=np.dot(r,xax2)*scale+shift
draw.line((xax1[0], xax1[1], xax2[0], xax2[1]), fill=(255,255,0), width=4)
#gradients(draw,r)
pt = shift
draw.ellipse((pt[0]-10, pt[1]-10, pt[0]+10, pt[1]+10), fill = (0,255,0))
draw_paraboloid_plane(draw,r,3.3)
draw_paraboloid_plane(draw,r,2.0,extent=1.4)
draw_paraboloid_plane(draw,r,1.0,extent=1.0)
im.save(basepath + 'im' + str(im_ind) + '.png')
def gradients(draw,r):
#for z in [0.3,1.3,2.3,3.3]:
for z in [3.3,2.0,1.0]:
x = np.sqrt(z)
for x in np.arange(-x,x,x/2):
y = np.sqrt(z-x*x)
arrowV1(draw,r,np.array([y,x,z]), np.array([1.5*y,1.5*x,z]), (204,102,255))
if z>3.0:
arrowV1(draw,r,np.array([-y,x,z]), np.array([-1.5*y,1.5*x,z]), (204,102,255))
def draw_paraboloid_plane(draw,r,z=3.3,scale=200,shift=np.array([1000,1000,0]),extent=2):
pt1=np.array([extent,extent,z]);pt1=np.dot(r,pt1)*scale+shift
pt2=np.array([extent,-extent,z]);pt2=np.dot(r,pt2)*scale+shift
pt3=np.array([-extent,-extent,z]);pt3=np.dot(r,pt3)*scale+shift
pt4=np.array([-extent,extent,z]);pt4=np.dot(r,pt4)*scale+shift
draw.polygon([(pt1[0], pt1[1]), (pt2[0], pt2[1]), (pt3[0], pt3[1]), (pt4[0], pt4[1])],\
(0,102,255,50))
point1 = np.array([np.sqrt(z),0,z])
generalized_arc(draw, r, center=np.array([0,0,z]), vec=np.array([0,0,1]),
point=point1, radius=np.sqrt(z), prcnt=1.0,scale=scale,
rgba=(255,20,10,100),width=10)
def plane_w_arrows(im_ind=0, scale=200,\
shift=np.array([824,824,0]),\
basepath='.\\'):
r1 = np.eye(4)
rot = general_rotation(np.array([0,0,1]), np.pi/20.0*(8 + im_ind/3.0))
j=4
r = rotation(3, 2*np.pi*j/30.0)
rr = general_rotation(np.array([0,1,0]), np.pi/20.0*(im_ind/7.0))
r = np.dot(r,rr)
r = np.dot(r, rot)
r1[:3,:3] = r
im = Image.new("RGB", (1648, 1648), "black")
draw = ImageDraw.Draw(im, 'RGBA')
pt1 = 3*np.array([1.0,-1.0,0]); pt2 = 3*np.array([1.0,1.0,0])
z = 1.2**2+1
pt3 = 3*np.array([-1.0,1.0,0]); pt4 = 3*np.array([-1.0,-1.0,0])
pt1 = np.dot(r,pt1)*scale+shift; pt2 = np.dot(r,pt2)*scale+shift
pt3 = np.dot(r,pt3)*scale+shift; pt4 = np.dot(r,pt4)*scale+shift
draw.polygon([(pt1[0], pt1[1]), (pt2[0], pt2[1]), (pt3[0], pt3[1]), (pt4[0], pt4[1])],\
(0,102,255,50))
draw_arrows(draw,r,rgba=(255,250,47),shift=shift)
draw_arrows(draw,r,rot_angl=np.pi/2.0, rgba=(73,200,250),shift=shift)
draw_arrows(draw,r,rot_angl=np.pi/2.0+np.pi/3, rgba=(255,20,147),shift=shift)
arrowV1(draw,r,np.array([0,0,0]), np.array([0,0,2.5]), shift=shift,rgb=(20,200,25))
arrowV1(draw,r,np.array([0,0,0]), np.array([0,0,-2.5]), shift=shift,rgb=(255,20,25))
im.save(basepath + 'im' + str(im_ind) + '.png')
def draw_arrows(draw,r,rot_angl=np.pi/6.0,rgba=(255,20,147),shift=np.array([1000,1000,0])):
base = np.array([0,0,1.5])
for theta in np.arange(0,np.pi*2,2*np.pi/3):
a = np.array([np.cos(theta),np.sin(theta),0])
rr = general_rotation(a, rot_angl)
arrow1 = np.dot(rr,base)
arrowV1(draw,r,np.array([0,0,0]), arrow1, rgb=rgba,shift=shift)
rgba = rgba+(150,)
generalized_arc(draw, r, center=np.array([0,0,1.5*np.cos(rot_angl)]),
vec=np.array([0,0,1]),
point=1.5*np.array([0,np.sin(rot_angl),np.cos(rot_angl)]),
radius=100, prcnt=1.0,
rgba=rgba,shift=shift)
#####################
## Paraboloid with Lagrange visualized.
im = Image.new("RGB", (2048, 2048), (1, 1, 1))
draw = ImageDraw.Draw(im, 'RGBA')
scale=5.0; ind=0; sep = 24; i = 2.0; base_coeff = 0.02; start_line = -12.0
shift = np.array([1000.0, 1000.0, 0.0])
r1 = np.eye(4); j=24
r = rotation(3, np.pi/30*j)
r1[:3,:3] = r
render_scene_4d_axis(draw, r1, 4)
fn = lambda x, y : paraboloid(x, y, coeff=i*base_coeff, intercept=i)
drawFunctionalXYGrid(draw, r, scale=scale, fn=fn,
extent=60, rgba2=(255,20,147,80),
saperatingPlane=np.array([-1,-1,sep]))
three_d_parabola(draw, r, r2)
im.save(basedir + 'im' + str(0) + '.png')
|
[
"numpy.eye",
"numpy.sqrt",
"numpy.array",
"matplotlib.pyplot.figure",
"numpy.linspace",
"numpy.dot",
"numpy.cos",
"numpy.sin",
"numpy.meshgrid",
"numpy.arange",
"matplotlib.pyplot.show"
] |
[((1463, 1475), 'matplotlib.pyplot.figure', 'plt.figure', ([], {}), '()\n', (1473, 1475), True, 'import matplotlib.pyplot as plt\n'), ((1514, 1544), 'numpy.linspace', 'np.linspace', (['(0)', '(2 * np.pi)', '(100)'], {}), '(0, 2 * np.pi, 100)\n', (1525, 1544), True, 'import numpy as np\n'), ((1555, 1579), 'numpy.arange', 'np.arange', (['(0.1)', '(1.0)', '(0.1)'], {}), '(0.1, 1.0, 0.1)\n', (1564, 1579), True, 'import numpy as np\n'), ((1727, 1737), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (1735, 1737), True, 'import matplotlib.pyplot as plt\n'), ((6730, 6761), 'numpy.array', 'np.array', (['[1000.0, 1000.0, 0.0]'], {}), '([1000.0, 1000.0, 0.0])\n', (6738, 6761), True, 'import numpy as np\n'), ((6768, 6777), 'numpy.eye', 'np.eye', (['(4)'], {}), '(4)\n', (6774, 6777), True, 'import numpy as np\n'), ((809, 821), 'matplotlib.pyplot.figure', 'plt.figure', ([], {}), '()\n', (819, 821), True, 'import matplotlib.pyplot as plt\n'), ((882, 904), 'numpy.arange', 'np.arange', (['(-5)', '(5)', '(0.25)'], {}), '(-5, 5, 0.25)\n', (891, 904), True, 'import numpy as np\n'), ((913, 935), 'numpy.arange', 'np.arange', (['(-5)', '(5)', '(0.25)'], {}), '(-5, 5, 0.25)\n', (922, 935), True, 'import numpy as np\n'), ((947, 964), 'numpy.meshgrid', 'np.meshgrid', (['X', 'Y'], {}), '(X, Y)\n', (958, 964), True, 'import numpy as np\n'), ((1406, 1416), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (1414, 1416), True, 'import matplotlib.pyplot as plt\n'), ((1789, 1814), 'numpy.array', 'np.array', (['[1000, 1000, 0]'], {}), '([1000, 1000, 0])\n', (1797, 1814), True, 'import numpy as np\n'), ((1872, 1881), 'numpy.eye', 'np.eye', (['(4)'], {}), '(4)\n', (1878, 1881), True, 'import numpy as np\n'), ((2087, 2100), 'numpy.dot', 'np.dot', (['r', 'rr'], {}), '(r, rr)\n', (2093, 2100), True, 'import numpy as np\n'), ((2108, 2122), 'numpy.dot', 'np.dot', (['r', 'rot'], {}), '(r, rot)\n', (2114, 2122), True, 'import numpy as np\n'), ((2339, 2366), 'numpy.arange', 'np.arange', (['(0.001)', '(3.5)', '(0.02)'], {}), '(0.001, 3.5, 0.02)\n', (2348, 2366), True, 'import numpy as np\n'), ((2619, 2645), 'numpy.array', 'np.array', (['[-100.0, 0, 0.0]'], {}), '([-100.0, 0, 0.0])\n', (2627, 2645), True, 'import numpy as np\n'), ((2685, 2710), 'numpy.array', 'np.array', (['[100.0, 0, 0.0]'], {}), '([100.0, 0, 0.0])\n', (2693, 2710), True, 'import numpy as np\n'), ((2829, 2855), 'numpy.array', 'np.array', (['[0.0, -100, 0.0]'], {}), '([0.0, -100, 0.0])\n', (2837, 2855), True, 'import numpy as np\n'), ((2895, 2920), 'numpy.array', 'np.array', (['[0.0, 100, 0.0]'], {}), '([0.0, 100, 0.0])\n', (2903, 2920), True, 'import numpy as np\n'), ((3772, 3797), 'numpy.array', 'np.array', (['[1000, 1000, 0]'], {}), '([1000, 1000, 0])\n', (3780, 3797), True, 'import numpy as np\n'), ((3815, 3844), 'numpy.array', 'np.array', (['[extent, extent, z]'], {}), '([extent, extent, z])\n', (3823, 3844), True, 'import numpy as np\n'), ((3881, 3911), 'numpy.array', 'np.array', (['[extent, -extent, z]'], {}), '([extent, -extent, z])\n', (3889, 3911), True, 'import numpy as np\n'), ((3948, 3979), 'numpy.array', 'np.array', (['[-extent, -extent, z]'], {}), '([-extent, -extent, z])\n', (3956, 3979), True, 'import numpy as np\n'), ((4016, 4046), 'numpy.array', 'np.array', (['[-extent, extent, z]'], {}), '([-extent, extent, z])\n', (4024, 4046), True, 'import numpy as np\n'), ((4526, 4549), 'numpy.array', 'np.array', (['[824, 824, 0]'], {}), '([824, 824, 0])\n', (4534, 4549), True, 'import numpy as np\n'), ((4596, 4605), 'numpy.eye', 'np.eye', (['(4)'], {}), '(4)\n', (4602, 4605), True, 'import numpy as np\n'), ((4803, 4816), 'numpy.dot', 'np.dot', (['r', 'rr'], {}), '(r, rr)\n', (4809, 4816), True, 'import numpy as np\n'), ((4824, 4838), 'numpy.dot', 'np.dot', (['r', 'rot'], {}), '(r, rot)\n', (4830, 4838), True, 'import numpy as np\n'), ((5870, 5895), 'numpy.array', 'np.array', (['[1000, 1000, 0]'], {}), '([1000, 1000, 0])\n', (5878, 5895), True, 'import numpy as np\n'), ((5907, 5928), 'numpy.array', 'np.array', (['[0, 0, 1.5]'], {}), '([0, 0, 1.5])\n', (5915, 5928), True, 'import numpy as np\n'), ((5944, 5982), 'numpy.arange', 'np.arange', (['(0)', '(np.pi * 2)', '(2 * np.pi / 3)'], {}), '(0, np.pi * 2, 2 * np.pi / 3)\n', (5953, 5982), True, 'import numpy as np\n'), ((1604, 1617), 'numpy.sin', 'np.sin', (['theta'], {}), '(theta)\n', (1610, 1617), True, 'import numpy as np\n'), ((1630, 1643), 'numpy.cos', 'np.cos', (['theta'], {}), '(theta)\n', (1636, 1643), True, 'import numpy as np\n'), ((1909, 1928), 'numpy.array', 'np.array', (['[0, 0, 1]'], {}), '([0, 0, 1])\n', (1917, 1928), True, 'import numpy as np\n'), ((2033, 2052), 'numpy.array', 'np.array', (['[0, 1, 0]'], {}), '([0, 1, 0])\n', (2041, 2052), True, 'import numpy as np\n'), ((3431, 3441), 'numpy.sqrt', 'np.sqrt', (['z'], {}), '(z)\n', (3438, 3441), True, 'import numpy as np\n'), ((3459, 3482), 'numpy.arange', 'np.arange', (['(-x)', 'x', '(x / 2)'], {}), '(-x, x, x / 2)\n', (3468, 3482), True, 'import numpy as np\n'), ((4633, 4652), 'numpy.array', 'np.array', (['[0, 0, 1]'], {}), '([0, 0, 1])\n', (4641, 4652), True, 'import numpy as np\n'), ((4751, 4770), 'numpy.array', 'np.array', (['[0, 1, 0]'], {}), '([0, 1, 0])\n', (4759, 4770), True, 'import numpy as np\n'), ((4956, 4980), 'numpy.array', 'np.array', (['[1.0, -1.0, 0]'], {}), '([1.0, -1.0, 0])\n', (4964, 4980), True, 'import numpy as np\n'), ((4988, 5011), 'numpy.array', 'np.array', (['[1.0, 1.0, 0]'], {}), '([1.0, 1.0, 0])\n', (4996, 5011), True, 'import numpy as np\n'), ((5039, 5063), 'numpy.array', 'np.array', (['[-1.0, 1.0, 0]'], {}), '([-1.0, 1.0, 0])\n', (5047, 5063), True, 'import numpy as np\n'), ((5071, 5096), 'numpy.array', 'np.array', (['[-1.0, -1.0, 0]'], {}), '([-1.0, -1.0, 0])\n', (5079, 5096), True, 'import numpy as np\n'), ((5592, 5611), 'numpy.array', 'np.array', (['[0, 0, 0]'], {}), '([0, 0, 0])\n', (5600, 5611), True, 'import numpy as np\n'), ((5611, 5632), 'numpy.array', 'np.array', (['[0, 0, 2.5]'], {}), '([0, 0, 2.5])\n', (5619, 5632), True, 'import numpy as np\n'), ((5680, 5699), 'numpy.array', 'np.array', (['[0, 0, 0]'], {}), '([0, 0, 0])\n', (5688, 5699), True, 'import numpy as np\n'), ((5699, 5721), 'numpy.array', 'np.array', (['[0, 0, -2.5]'], {}), '([0, 0, -2.5])\n', (5707, 5721), True, 'import numpy as np\n'), ((6090, 6106), 'numpy.dot', 'np.dot', (['rr', 'base'], {}), '(rr, base)\n', (6096, 6106), True, 'import numpy as np\n'), ((7065, 7088), 'numpy.array', 'np.array', (['[-1, -1, sep]'], {}), '([-1, -1, sep])\n', (7073, 7088), True, 'import numpy as np\n'), ((630, 649), 'numpy.array', 'np.array', (['[1, 0, 0]'], {}), '([1, 0, 0])\n', (638, 649), True, 'import numpy as np\n'), ((2649, 2664), 'numpy.dot', 'np.dot', (['r', 'xax1'], {}), '(r, xax1)\n', (2655, 2664), True, 'import numpy as np\n'), ((2714, 2729), 'numpy.dot', 'np.dot', (['r', 'xax2'], {}), '(r, xax2)\n', (2720, 2729), True, 'import numpy as np\n'), ((2859, 2874), 'numpy.dot', 'np.dot', (['r', 'xax1'], {}), '(r, xax1)\n', (2865, 2874), True, 'import numpy as np\n'), ((2924, 2939), 'numpy.dot', 'np.dot', (['r', 'xax2'], {}), '(r, xax2)\n', (2930, 2939), True, 'import numpy as np\n'), ((3496, 3514), 'numpy.sqrt', 'np.sqrt', (['(z - x * x)'], {}), '(z - x * x)\n', (3503, 3514), True, 'import numpy as np\n'), ((3847, 3861), 'numpy.dot', 'np.dot', (['r', 'pt1'], {}), '(r, pt1)\n', (3853, 3861), True, 'import numpy as np\n'), ((3914, 3928), 'numpy.dot', 'np.dot', (['r', 'pt2'], {}), '(r, pt2)\n', (3920, 3928), True, 'import numpy as np\n'), ((3982, 3996), 'numpy.dot', 'np.dot', (['r', 'pt3'], {}), '(r, pt3)\n', (3988, 3996), True, 'import numpy as np\n'), ((4049, 4063), 'numpy.dot', 'np.dot', (['r', 'pt4'], {}), '(r, pt4)\n', (4055, 4063), True, 'import numpy as np\n'), ((4226, 4236), 'numpy.sqrt', 'np.sqrt', (['z'], {}), '(z)\n', (4233, 4236), True, 'import numpy as np\n'), ((4279, 4298), 'numpy.array', 'np.array', (['[0, 0, z]'], {}), '([0, 0, z])\n', (4287, 4298), True, 'import numpy as np\n'), ((4302, 4321), 'numpy.array', 'np.array', (['[0, 0, 1]'], {}), '([0, 0, 1])\n', (4310, 4321), True, 'import numpy as np\n'), ((4367, 4377), 'numpy.sqrt', 'np.sqrt', (['z'], {}), '(z)\n', (4374, 4377), True, 'import numpy as np\n'), ((5105, 5119), 'numpy.dot', 'np.dot', (['r', 'pt1'], {}), '(r, pt1)\n', (5111, 5119), True, 'import numpy as np\n'), ((5138, 5152), 'numpy.dot', 'np.dot', (['r', 'pt2'], {}), '(r, pt2)\n', (5144, 5152), True, 'import numpy as np\n'), ((5174, 5188), 'numpy.dot', 'np.dot', (['r', 'pt3'], {}), '(r, pt3)\n', (5180, 5188), True, 'import numpy as np\n'), ((5207, 5221), 'numpy.dot', 'np.dot', (['r', 'pt4'], {}), '(r, pt4)\n', (5213, 5221), True, 'import numpy as np\n'), ((6129, 6148), 'numpy.array', 'np.array', (['[0, 0, 0]'], {}), '([0, 0, 0])\n', (6137, 6148), True, 'import numpy as np\n'), ((6303, 6322), 'numpy.array', 'np.array', (['[0, 0, 1]'], {}), '([0, 0, 1])\n', (6311, 6322), True, 'import numpy as np\n'), ((2395, 2405), 'numpy.sqrt', 'np.sqrt', (['z'], {}), '(z)\n', (2402, 2405), True, 'import numpy as np\n'), ((2452, 2471), 'numpy.array', 'np.array', (['[0, 0, z]'], {}), '([0, 0, z])\n', (2460, 2471), True, 'import numpy as np\n'), ((2475, 2494), 'numpy.array', 'np.array', (['[0, 0, 1]'], {}), '([0, 0, 1])\n', (2483, 2494), True, 'import numpy as np\n'), ((2540, 2550), 'numpy.sqrt', 'np.sqrt', (['z'], {}), '(z)\n', (2547, 2550), True, 'import numpy as np\n'), ((3538, 3557), 'numpy.array', 'np.array', (['[y, x, z]'], {}), '([y, x, z])\n', (3546, 3557), True, 'import numpy as np\n'), ((3557, 3588), 'numpy.array', 'np.array', (['[1.5 * y, 1.5 * x, z]'], {}), '([1.5 * y, 1.5 * x, z])\n', (3565, 3588), True, 'import numpy as np\n'), ((5998, 6011), 'numpy.cos', 'np.cos', (['theta'], {}), '(theta)\n', (6004, 6011), True, 'import numpy as np\n'), ((6012, 6025), 'numpy.sin', 'np.sin', (['theta'], {}), '(theta)\n', (6018, 6025), True, 'import numpy as np\n'), ((3652, 3672), 'numpy.array', 'np.array', (['[-y, x, z]'], {}), '([-y, x, z])\n', (3660, 3672), True, 'import numpy as np\n'), ((3672, 3704), 'numpy.array', 'np.array', (['[-1.5 * y, 1.5 * x, z]'], {}), '([-1.5 * y, 1.5 * x, z])\n', (3680, 3704), True, 'import numpy as np\n'), ((6255, 6271), 'numpy.cos', 'np.cos', (['rot_angl'], {}), '(rot_angl)\n', (6261, 6271), True, 'import numpy as np\n'), ((6368, 6384), 'numpy.sin', 'np.sin', (['rot_angl'], {}), '(rot_angl)\n', (6374, 6384), True, 'import numpy as np\n'), ((6385, 6401), 'numpy.cos', 'np.cos', (['rot_angl'], {}), '(rot_angl)\n', (6391, 6401), True, 'import numpy as np\n')]
|
import os
import numpy as np
from vmaf import plt
from vmaf.core.cross_validation import ModelCrossValidation
from vmaf.core.feature_assembler import FeatureAssembler
from vmaf.core.quality_runner import VmafQualityRunner
from vmaf.core.result_store import FileSystemResultStore
from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension
from vmaf.config import VmafConfig, DisplayConfig
from vmaf.core.asset import Asset
from vmaf.core.train_test_model import TrainTestModel, RegressorMixin, ClassifierMixin
from vmaf.core.local_explainer import LocalExplainer
__copyright__ = "Copyright 2016-2020, Netflix, Inc."
__license__ = "BSD+Patent"
def read_dataset(dataset, **kwargs):
groundtruth_key = kwargs['groundtruth_key'] if 'groundtruth_key' in kwargs else None
skip_asset_with_none_groundtruth = kwargs['skip_asset_with_none_groundtruth'] \
if 'skip_asset_with_none_groundtruth' in kwargs else False
content_ids = kwargs['content_ids'] if 'content_ids' in kwargs else None
asset_ids = kwargs['asset_ids'] if 'asset_ids' in kwargs else None
workdir_root = kwargs['workdir_root'] if 'workdir_root' in kwargs else VmafConfig.workdir_path()
# asserts, can add more to the list...
assert hasattr(dataset, 'dataset_name')
assert hasattr(dataset, 'ref_videos')
assert hasattr(dataset, 'dis_videos')
assert hasattr(dataset, 'yuv_fmt') or all(['yuv_fmt' in ref_video for ref_video in dataset.ref_videos])
data_set_name = dataset.dataset_name
ref_videos = dataset.ref_videos
dis_videos = dataset.dis_videos
width = dataset.width if hasattr(dataset, 'width') else None
height = dataset.height if hasattr(dataset, 'height') else None
yuv_fmt = dataset.yuv_fmt if hasattr(dataset, 'yuv_fmt') else None
quality_width = dataset.quality_width if hasattr(dataset, 'quality_width') else None
quality_height = dataset.quality_height if hasattr(dataset, 'quality_height') else None
resampling_type = dataset.resampling_type if hasattr(dataset, 'resampling_type') else None
crop_cmd = dataset.crop_cmd if hasattr(dataset, 'crop_cmd') else None
pad_cmd = dataset.pad_cmd if hasattr(dataset, 'pad_cmd') else None
workfile_yuv_type = dataset.workfile_yuv_type if hasattr(dataset, 'workfile_yuv_type') else None
duration_sec = dataset.duration_sec if hasattr(dataset, 'duration_sec') else None
fps = dataset.fps if hasattr(dataset, 'fps') else None
start_frame = dataset.start_frame if hasattr(dataset, 'start_frame') else None
end_frame = dataset.end_frame if hasattr(dataset, 'end_frame') else None
ref_dict = {} # dictionary of content_id -> path for ref videos
for ref_video in ref_videos:
ref_dict[ref_video['content_id']] = ref_video
assets = []
for dis_video in dis_videos:
if content_ids is not None and dis_video['content_id'] not in content_ids:
continue
if asset_ids is not None and dis_video['asset_id'] not in asset_ids:
continue
if groundtruth_key is not None:
groundtruth = dis_video[groundtruth_key]
else:
if 'dmos' in dis_video:
groundtruth = dis_video['dmos']
elif 'mos' in dis_video:
groundtruth = dis_video['mos']
elif 'groundtruth' in dis_video:
groundtruth = dis_video['groundtruth']
else:
groundtruth = None
if 'os' in dis_video:
raw_groundtruth = dis_video['os']
else:
raw_groundtruth = None
if 'groundtruth_std' in dis_video:
groundtruth_std = dis_video['groundtruth_std']
else:
groundtruth_std = None
if 'rebuf_indices' in dis_video:
rebuf_indices = dis_video['rebuf_indices']
else:
rebuf_indices = None
ref_video = ref_dict[dis_video['content_id']]
ref_path = ref_video['path']
ref_yuv_fmt_ = yuv_fmt if yuv_fmt is not None else ref_dict[dis_video['content_id']]['yuv_fmt']
dis_yuv_fmt_ = dis_video['yuv_fmt'] if 'yuv_fmt' in dis_video else ref_yuv_fmt_
if width is not None:
width_ = width
elif 'width' in ref_video and 'width' not in dis_video:
width_ = ref_video['width']
elif 'width' in dis_video and 'width' not in ref_video:
width_ = dis_video['width']
elif 'width' in ref_video and 'width' in dis_video:
assert ref_video['width'] == dis_video['width']
width_ = ref_video['width']
else:
width_ = None
if height is not None:
height_ = height
elif 'height' in ref_video and 'height' not in dis_video:
height_ = ref_video['height']
elif 'height' in dis_video and 'height' not in ref_video:
height_ = dis_video['height']
elif 'height' in ref_video and 'height' in dis_video:
assert ref_video['height'] == dis_video['height']
height_ = ref_video['height']
else:
height_ = None
if quality_width is not None:
quality_width_ = quality_width
elif 'quality_width' in dis_video:
quality_width_ = dis_video['quality_width']
else:
quality_width_ = None
if quality_height is not None:
quality_height_ = quality_height
elif 'quality_height' in dis_video:
quality_height_ = dis_video['quality_height']
else:
quality_height_ = None
if resampling_type is not None:
resampling_type_ = resampling_type
elif 'resampling_type' in dis_video:
resampling_type_ = dis_video['resampling_type']
else:
resampling_type_ = None
if crop_cmd is not None:
ref_crop_cmd_ = crop_cmd
dis_crop_cmd_ = crop_cmd
else:
if 'crop_cmd' in ref_video:
ref_crop_cmd_ = ref_video['crop_cmd']
else:
ref_crop_cmd_ = None
if 'crop_cmd' in dis_video:
dis_crop_cmd_ = dis_video['crop_cmd']
else:
dis_crop_cmd_ = None
if pad_cmd is not None:
ref_pad_cmd_ = pad_cmd
dis_pad_cmd_ = pad_cmd
else:
if 'pad_cmd' in ref_video:
ref_pad_cmd_ = ref_video['pad_cmd']
else:
ref_pad_cmd_ = None
if 'pad_cmd' in dis_video:
dis_pad_cmd_ = dis_video['pad_cmd']
else:
dis_pad_cmd_ = None
if duration_sec is not None:
duration_sec_ = duration_sec
elif 'duration_sec' in dis_video:
duration_sec_ = dis_video['duration_sec']
else:
duration_sec_ = None
if fps is not None:
fps_ = fps
elif 'fps' in dis_video:
fps_ = dis_video['fps']
else:
fps_ = None
if start_frame is not None:
start_frame_ = start_frame
elif 'start_frame' in dis_video:
start_frame_ = dis_video['start_frame']
else:
start_frame_ = None
if end_frame is not None:
end_frame_ = end_frame
elif 'end_frame' in dis_video:
end_frame_ = dis_video['end_frame']
else:
end_frame_ = None
asset_dict = {'ref_yuv_type': ref_yuv_fmt_, 'dis_yuv_type': dis_yuv_fmt_}
if width_ is not None:
if asset_dict['ref_yuv_type'] != 'notyuv':
asset_dict['ref_width'] = width_
if asset_dict['dis_yuv_type'] != 'notyuv':
asset_dict['dis_width'] = width_
if height_ is not None:
if asset_dict['ref_yuv_type'] != 'notyuv':
asset_dict['ref_height'] = height_
if asset_dict['dis_yuv_type'] != 'notyuv':
asset_dict['dis_height'] = height_
if groundtruth is not None:
asset_dict['groundtruth'] = groundtruth
if raw_groundtruth is not None:
asset_dict['raw_groundtruth'] = raw_groundtruth
if groundtruth_std is not None:
asset_dict['groundtruth_std'] = groundtruth_std
if quality_width_ is not None:
asset_dict['quality_width'] = quality_width_
if quality_height_ is not None:
asset_dict['quality_height'] = quality_height_
if resampling_type_ is not None:
asset_dict['resampling_type'] = resampling_type_
if ref_crop_cmd_ is not None:
asset_dict['ref_crop_cmd'] = ref_crop_cmd_
if dis_crop_cmd_ is not None:
asset_dict['dis_crop_cmd'] = dis_crop_cmd_
if ref_pad_cmd_ is not None:
asset_dict['ref_pad_cmd'] = ref_pad_cmd_
if dis_pad_cmd_ is not None:
asset_dict['dis_pad_cmd'] = dis_pad_cmd_
if duration_sec_ is not None:
asset_dict['duration_sec'] = duration_sec_
if workfile_yuv_type is not None:
asset_dict['workfile_yuv_type'] = workfile_yuv_type
if rebuf_indices is not None:
asset_dict['rebuf_indices'] = rebuf_indices
if fps_ is not None:
asset_dict['fps'] = fps_
if start_frame_ is not None:
asset_dict['start_frame'] = start_frame_
if end_frame_ is not None:
asset_dict['end_frame'] = end_frame_
if groundtruth is None and skip_asset_with_none_groundtruth:
pass
else:
asset = Asset(dataset=data_set_name,
content_id=dis_video['content_id'],
asset_id=dis_video['asset_id'],
workdir_root=workdir_root,
ref_path=ref_path,
dis_path=dis_video['path'],
asset_dict=asset_dict,
)
assets.append(asset)
return assets
def run_test_on_dataset(test_dataset, runner_class, ax,
result_store, model_filepath,
parallelize=True, fifo_mode=True,
aggregate_method=np.mean,
type='regressor',
**kwargs):
test_assets = read_dataset(test_dataset, **kwargs)
test_raw_assets = None
try:
for test_asset in test_assets:
assert test_asset.groundtruth is not None
except AssertionError:
# no groundtruth, try do subjective modeling
from sureal.dataset_reader import RawDatasetReader
from sureal.subjective_model import DmosModel
subj_model_class = kwargs['subj_model_class'] if 'subj_model_class' in kwargs and kwargs['subj_model_class'] is not None else DmosModel
dataset_reader_class = kwargs['dataset_reader_class'] if 'dataset_reader_class' in kwargs else RawDatasetReader
subjective_model = subj_model_class(dataset_reader_class(test_dataset))
subjective_model.run_modeling(**kwargs)
test_dataset_aggregate = subjective_model.to_aggregated_dataset(**kwargs)
test_raw_assets = test_assets
test_assets = read_dataset(test_dataset_aggregate, **kwargs)
if model_filepath is not None:
optional_dict = {'model_filepath': model_filepath}
if 'model_720_filepath' in kwargs and kwargs['model_720_filepath'] is not None:
optional_dict['720model_filepath'] = kwargs['model_720_filepath']
if 'model_480_filepath' in kwargs and kwargs['model_480_filepath'] is not None:
optional_dict['480model_filepath'] = kwargs['model_480_filepath']
else:
optional_dict = None
if 'enable_transform_score' in kwargs and kwargs['enable_transform_score'] is not None:
if not optional_dict:
optional_dict = {}
optional_dict['enable_transform_score'] = kwargs['enable_transform_score']
if 'disable_clip_score' in kwargs and kwargs['disable_clip_score'] is not None:
if not optional_dict:
optional_dict = {}
optional_dict['disable_clip_score'] = kwargs['disable_clip_score']
if 'subsample' in kwargs and kwargs['subsample'] is not None:
if not optional_dict:
optional_dict = {}
optional_dict['subsample'] = kwargs['subsample']
# run
runner = runner_class(
test_assets,
None, fifo_mode=fifo_mode,
delete_workdir=True,
result_store=result_store,
optional_dict=optional_dict,
optional_dict2=None,
)
runner.run(parallelize=parallelize)
results = runner.results
for result in results:
result.set_score_aggregate_method(aggregate_method)
try:
model_type = runner.get_train_test_model_class()
except:
if type == 'regressor':
model_type = RegressorMixin
elif type == 'classifier':
model_type = ClassifierMixin
else:
assert False
split_test_indices_for_perf_ci = kwargs['split_test_indices_for_perf_ci'] \
if 'split_test_indices_for_perf_ci' in kwargs else False
# plot
groundtruths = list(map(lambda asset: asset.groundtruth, test_assets))
predictions = list(map(lambda result: result[runner_class.get_score_key()], results))
raw_grountruths = None if test_raw_assets is None else \
list(map(lambda asset: asset.raw_groundtruth, test_raw_assets))
groundtruths_std = None if test_assets is None else \
list(map(lambda asset: asset.groundtruth_std, test_assets))
try:
predictions_bagging = list(map(lambda result: result[runner_class.get_bagging_score_key()], results))
predictions_stddev = list(map(lambda result: result[runner_class.get_stddev_score_key()], results))
predictions_ci95_low = list(map(lambda result: result[runner_class.get_ci95_low_score_key()], results))
predictions_ci95_high = list(map(lambda result: result[runner_class.get_ci95_high_score_key()], results))
predictions_all_models = list(map(lambda result: result[runner_class.get_all_models_score_key()], results))
# need to revert the list of lists, so that the outer list has the predictions for each model separately
predictions_all_models = np.array(predictions_all_models).T.tolist()
num_models = np.shape(predictions_all_models)[0]
stats = model_type.get_stats(groundtruths, predictions,
ys_label_raw=raw_grountruths,
ys_label_pred_bagging=predictions_bagging,
ys_label_pred_stddev=predictions_stddev,
ys_label_pred_ci95_low=predictions_ci95_low,
ys_label_pred_ci95_high=predictions_ci95_high,
ys_label_pred_all_models=predictions_all_models,
ys_label_stddev=groundtruths_std,
split_test_indices_for_perf_ci=split_test_indices_for_perf_ci)
except Exception as e:
print('Stats calculation failed, using default stats calculation. Error cause: ')
print(e)
stats = model_type.get_stats(groundtruths, predictions,
ys_label_raw=raw_grountruths,
ys_label_stddev=groundtruths_std,
split_test_indices_for_perf_ci=split_test_indices_for_perf_ci)
num_models = 1
print('Stats on testing data: {}'.format(model_type.format_stats_for_print(stats)))
# printing stats if multiple models are present
if 'SRCC_across_model_distribution' in stats \
and 'PCC_across_model_distribution' in stats \
and 'RMSE_across_model_distribution' in stats:
print('Stats on testing data (across multiple models, using all test indices): {}'.format(
model_type.format_across_model_stats_for_print(model_type.extract_across_model_stats(stats))))
if split_test_indices_for_perf_ci:
print('Stats on testing data (single model, multiple test sets): {}'
.format(model_type.format_stats_across_test_splits_for_print(model_type.extract_across_test_splits_stats(stats))))
if ax is not None:
content_ids = list(map(lambda asset: asset.content_id, test_assets))
if 'point_label' in kwargs:
if kwargs['point_label'] == 'asset_id':
point_labels = list(map(lambda asset: asset.asset_id, test_assets))
elif kwargs['point_label'] == 'dis_path':
point_labels = list(map(lambda asset: get_file_name_without_extension(asset.dis_path), test_assets))
else:
raise AssertionError("Unknown point_label {}".format(kwargs['point_label']))
else:
point_labels = None
model_type.plot_scatter(ax, stats, content_ids=content_ids, point_labels=point_labels, **kwargs)
ax.set_xlabel('True Score')
ax.set_ylabel("Predicted Score")
ax.grid()
ax.set_title("{runner}{num_models}\n{stats}".format(
dataset=test_assets[0].dataset,
runner=runner_class.TYPE,
stats=model_type.format_stats_for_plot(stats),
num_models=", {} models".format(num_models) if num_models > 1 else "",
))
return test_assets, results
def print_matplotlib_warning():
print("Warning: cannot import matplotlib, no picture displayed. " \
"If you are on Mac OS and have installed matplotlib, you " \
"possibly need to run: \nsudo pip uninstall python-dateutil \n" \
"sudo pip install python-dateutil==2.2 \n" \
"Refer to: http://stackoverflow.com/questions/27630114/matplotlib-issue-on-os-x-importerror-cannot-import-name-thread")
def train_test_vmaf_on_dataset(train_dataset, test_dataset,
feature_param, model_param,
train_ax, test_ax, result_store,
parallelize=True, logger=None, fifo_mode=True,
output_model_filepath=None,
aggregate_method=np.mean,
**kwargs):
train_assets = read_dataset(train_dataset, **kwargs)
train_raw_assets = None
try:
for train_asset in train_assets:
assert train_asset.groundtruth is not None
except AssertionError:
# no groundtruth, try do subjective modeling
from sureal.dataset_reader import RawDatasetReader
from sureal.subjective_model import DmosModel
subj_model_class = kwargs['subj_model_class'] if 'subj_model_class' in kwargs and kwargs['subj_model_class'] is not None else DmosModel
dataset_reader_class = kwargs['dataset_reader_class'] if 'dataset_reader_class' in kwargs else RawDatasetReader
subjective_model = subj_model_class(dataset_reader_class(train_dataset))
subjective_model.run_modeling(**kwargs)
train_dataset_aggregate = subjective_model.to_aggregated_dataset(**kwargs)
train_raw_assets = train_assets
train_assets = read_dataset(train_dataset_aggregate, **kwargs)
train_fassembler = FeatureAssembler(
feature_dict=feature_param.feature_dict,
feature_option_dict=None,
assets=train_assets,
logger=logger,
fifo_mode=fifo_mode,
delete_workdir=True,
result_store=result_store,
optional_dict=None,
optional_dict2=None,
parallelize=parallelize,
)
train_fassembler.run()
train_features = train_fassembler.results
for result in train_features:
result.set_score_aggregate_method(aggregate_method)
model_type = model_param.model_type
model_param_dict = model_param.model_param_dict
model_class = TrainTestModel.find_subclass(model_type)
train_xys = model_class.get_xys_from_results(train_features)
train_xs = model_class.get_xs_from_results(train_features)
train_ys = model_class.get_ys_from_results(train_features)
model = model_class(model_param_dict, logger)
model.train(train_xys, **kwargs)
# append additional information to model before saving, so that
# VmafQualityRunner can read and process
model.append_info('feature_dict', feature_param.feature_dict)
if 'score_clip' in model_param_dict:
VmafQualityRunner.set_clip_score(model, model_param_dict['score_clip'])
if 'score_transform' in model_param_dict:
VmafQualityRunner.set_transform_score(model, model_param_dict['score_transform'])
train_ys_pred = VmafQualityRunner.predict_with_model(model, train_xs, **kwargs)['ys_pred']
raw_groundtruths = None if train_raw_assets is None else \
list(map(lambda asset: asset.raw_groundtruth, train_raw_assets))
train_stats = model.get_stats(train_ys['label'], train_ys_pred, ys_label_raw=raw_groundtruths)
log = 'Stats on training data: {}'.format(model.format_stats_for_print(train_stats))
if logger:
logger.info(log)
else:
print(log)
# save model
if output_model_filepath is not None:
model.to_file(output_model_filepath)
if train_ax is not None:
train_content_ids = list(map(lambda asset: asset.content_id, train_assets))
model_class.plot_scatter(train_ax, train_stats, content_ids=train_content_ids)
train_ax.set_xlabel('True Score')
train_ax.set_ylabel("Predicted Score")
train_ax.grid()
train_ax.set_title("Dataset: {dataset}, Model: {model}\n{stats}".format(
dataset=train_dataset.dataset_name,
model=model.model_id,
stats=model_class.format_stats_for_plot(train_stats)
))
# === test model on test dataset ===
if test_dataset is None:
test_assets = None
test_stats = None
test_fassembler = None
else:
test_assets = read_dataset(test_dataset, **kwargs)
test_raw_assets = None
try:
for test_asset in test_assets:
assert test_asset.groundtruth is not None
except AssertionError:
# no groundtruth, try do subjective modeling
from sureal.dataset_reader import RawDatasetReader
from sureal.subjective_model import DmosModel
subj_model_class = kwargs['subj_model_class'] if 'subj_model_class' in kwargs and kwargs['subj_model_class'] is not None else DmosModel
dataset_reader_class = kwargs['dataset_reader_class'] if 'dataset_reader_class' in kwargs else RawDatasetReader
subjective_model = subj_model_class(dataset_reader_class(test_dataset))
subjective_model.run_modeling(**kwargs)
test_dataset_aggregate = subjective_model.to_aggregated_dataset(**kwargs)
test_raw_assets = test_assets
test_assets = read_dataset(test_dataset_aggregate, **kwargs)
test_fassembler = FeatureAssembler(
feature_dict=feature_param.feature_dict,
feature_option_dict=None,
assets=test_assets,
logger=logger,
fifo_mode=fifo_mode,
delete_workdir=True,
result_store=result_store,
optional_dict=None,
optional_dict2=None,
parallelize=True,
)
test_fassembler.run()
test_features = test_fassembler.results
for result in test_features:
result.set_score_aggregate_method(aggregate_method)
test_xs = model_class.get_xs_from_results(test_features)
test_ys = model_class.get_ys_from_results(test_features)
test_ys_pred = VmafQualityRunner.predict_with_model(model, test_xs, **kwargs)['ys_pred']
raw_groundtruths = None if test_raw_assets is None else \
list(map(lambda asset: asset.raw_groundtruth, test_raw_assets))
test_stats = model.get_stats(test_ys['label'], test_ys_pred, ys_label_raw=raw_groundtruths)
log = 'Stats on testing data: {}'.format(model_class.format_stats_for_print(test_stats))
if logger:
logger.info(log)
else:
print(log)
if test_ax is not None:
test_content_ids = list(map(lambda asset: asset.content_id, test_assets))
model_class.plot_scatter(test_ax, test_stats, content_ids=test_content_ids)
test_ax.set_xlabel('True Score')
test_ax.set_ylabel("Predicted Score")
test_ax.grid()
test_ax.set_title("Dataset: {dataset}, Model: {model}\n{stats}".format(
dataset=test_dataset.dataset_name,
model=model.model_id,
stats=model_class.format_stats_for_plot(test_stats)
))
return train_fassembler, train_assets, train_stats, test_fassembler, test_assets, test_stats, model
def construct_kfold_list(assets, contentid_groups):
# construct cross validation kfold input list
content_ids = list(map(lambda asset: asset.content_id, assets))
kfold = []
for curr_content_group in contentid_groups:
curr_indices = indices(content_ids, lambda x: x in curr_content_group)
kfold.append(curr_indices)
return kfold
def cv_on_dataset(dataset, feature_param, model_param, ax, result_store,
contentid_groups, logger=None, aggregate_method=np.mean):
assets = read_dataset(dataset)
kfold = construct_kfold_list(assets, contentid_groups)
fassembler = FeatureAssembler(
feature_dict=feature_param.feature_dict,
feature_option_dict=None,
assets=assets,
logger=logger,
delete_workdir=True,
result_store=result_store,
optional_dict=None,
optional_dict2=None,
parallelize=True, fifo_mode=True,
# parallelize=False, fifo_mode=False, # VQM
)
fassembler.run()
results = fassembler.results
for result in results:
result.set_score_aggregate_method(aggregate_method)
model_class = TrainTestModel.find_subclass(model_param.model_type)
# run nested kfold cv for each combintation
cv_output = ModelCrossValidation.run_kfold_cross_validation(
model_class,
model_param.model_param_dict,
results,
kfold,
logger=logger,
)
print('Feature parameters: {}'.format(feature_param.feature_dict))
print('Model type: {}'.format(model_param.model_type))
print('Model parameters: {}'.format(model_param.model_param_dict))
print('Stats: {}'.format(model_class.format_stats_for_print(cv_output['aggr_stats'])))
if ax is not None:
model_class.plot_scatter(ax, cv_output['aggr_stats'], cv_output['contentids'])
ax.set_xlabel('True Score')
ax.set_ylabel("Predicted Score")
ax.grid()
ax.set_title("Dataset: {dataset}, Model: {model},\n{stats}".format(
dataset=dataset.dataset_name,
model=model_param.model_type,
stats=model_class.format_stats_for_plot(cv_output['aggr_stats'])
))
return assets, cv_output
def run_remove_results_for_dataset(result_store, dataset, executor_class):
assets = read_dataset(dataset)
executor = executor_class(assets=assets, logger=None, result_store=result_store)
executor.remove_results()
def run_vmaf_cv(train_dataset_filepath,
test_dataset_filepath,
param_filepath,
output_model_filepath=None,
**kwargs):
result_store_dir = kwargs['result_store_dir'] if 'result_store_dir' in kwargs else VmafConfig.file_result_store_path()
logger = get_stdout_logger()
result_store = FileSystemResultStore(result_store_dir)
train_dataset = import_python_file(train_dataset_filepath)
test_dataset = import_python_file(test_dataset_filepath) if test_dataset_filepath is not None else None
param = import_python_file(param_filepath)
# === plot scatter ===
nrows = 1
ncols = 2
fig, axs = plt.subplots(figsize=(5*ncols, 5*nrows), nrows=nrows, ncols=ncols)
train_test_vmaf_on_dataset(train_dataset, test_dataset, param, param, axs[0], axs[1],
result_store, parallelize=True, logger=None,
output_model_filepath=output_model_filepath,
**kwargs)
if 'xlim' in kwargs:
axs[0].set_xlim(kwargs['xlim'])
axs[1].set_xlim(kwargs['xlim'])
if 'ylim' in kwargs:
axs[0].set_ylim(kwargs['ylim'])
axs[1].set_ylim(kwargs['ylim'])
bbox = {'facecolor':'white', 'alpha':1, 'pad':20}
axs[0].annotate('Training Set', xy=(0.1, 0.85), xycoords='axes fraction', bbox=bbox)
axs[1].annotate('Testing Set', xy=(0.1, 0.85), xycoords='axes fraction', bbox=bbox)
plt.tight_layout()
# === clean up ===
close_logger(logger)
def run_vmaf_kfold_cv(dataset_filepath,
contentid_groups,
param_filepath,
aggregate_method,
result_store_dir=VmafConfig.file_result_store_path(),
):
logger = get_stdout_logger()
result_store = FileSystemResultStore(result_store_dir)
dataset = import_python_file(dataset_filepath)
param = import_python_file(param_filepath)
fig, ax = plt.subplots(figsize=(5, 5), nrows=1, ncols=1)
cv_on_dataset(dataset, param, param, ax, result_store, contentid_groups,
logger, aggregate_method)
ax.set_xlim([0, 120])
ax.set_ylim([0, 120])
plt.tight_layout()
# === clean up ===
close_logger(logger)
def explain_model_on_dataset(model, test_assets_selected_indexs,
test_dataset_filepath,
result_store_dir=VmafConfig.file_result_store_path()):
def print_assets(test_assets):
print('\n'.join(map(
lambda tasset: "Asset {i}: {name}".format(
i=tasset[0], name=get_file_name_without_extension(tasset[1].dis_path)),
enumerate(test_assets)
)))
test_dataset = import_python_file(test_dataset_filepath)
test_assets = read_dataset(test_dataset)
print_assets(test_assets)
print("Assets selected for local explanation: {}".format(
test_assets_selected_indexs))
result_store = FileSystemResultStore(result_store_dir)
test_assets = [test_assets[i] for i in test_assets_selected_indexs]
test_fassembler = FeatureAssembler(
feature_dict=model.model_dict['feature_dict'],
feature_option_dict=None,
assets=test_assets,
logger=None,
fifo_mode=True,
delete_workdir=True,
result_store=result_store,
optional_dict=None,
optional_dict2=None,
parallelize=True,
)
test_fassembler.run()
test_feature_results = test_fassembler.results
test_xs = model.get_xs_from_results(test_feature_results)
test_ys = model.get_ys_from_results(test_feature_results)
test_ys_pred = model.predict(test_xs)['ys_label_pred']
explainer = LocalExplainer(neighbor_samples=1000)
test_exps = explainer.explain(model, test_xs)
explainer.print_explanations(test_exps, assets=test_assets, ys=test_ys, ys_pred=test_ys_pred)
explainer.plot_explanations(test_exps, assets=test_assets, ys=test_ys, ys_pred=test_ys_pred)
DisplayConfig.show()
def generate_dataset_from_raw(raw_dataset_filepath, output_dataset_filepath, **kwargs):
if raw_dataset_filepath:
from sureal.subjective_model import DmosModel
subj_model_class = kwargs['subj_model_class'] if 'subj_model_class' in kwargs else DmosModel
content_ids = kwargs['content_ids'] if 'content_ids' in kwargs else None
asset_ids = kwargs['asset_ids'] if 'asset_ids' in kwargs else None
subjective_model = subj_model_class.from_dataset_file(raw_dataset_filepath,
content_ids=content_ids,
asset_ids=asset_ids)
subjective_model.run_modeling(**kwargs)
subjective_model.to_aggregated_dataset_file(output_dataset_filepath, **kwargs)
def run_vmaf_cv_from_raw(train_dataset_raw_filepath, test_dataset_raw_filepath,
param_filepath, output_model_filepath, **kwargs):
if 'train_quality_wh' in kwargs and kwargs['train_quality_wh'] is not None:
train_quality_width, train_quality_height = kwargs['train_quality_wh']
else:
train_quality_width = None
train_quality_height = None
if 'test_quality_wh' in kwargs and kwargs['test_quality_wh'] is not None:
test_quality_width, test_quality_height = kwargs['test_quality_wh']
else:
test_quality_width = None
test_quality_height = None
if 'train_transform_final' in kwargs and kwargs['train_transform_final'] is not None:
train_transform_final = kwargs['train_transform_final']
else:
train_transform_final = None
if 'test_transform_final' in kwargs and kwargs['test_transform_final'] is not None:
test_transform_final = kwargs['test_transform_final']
else:
test_transform_final = None
workspace_path = kwargs['workspace_path'] if 'workspace_path' in kwargs else VmafConfig.workspace_path()
train_output_dataset_filepath = os.path.join(workspace_path, 'dataset', 'train_dataset.py')
generate_dataset_from_raw(raw_dataset_filepath=train_dataset_raw_filepath,
output_dataset_filepath=train_output_dataset_filepath,
quality_width=train_quality_width,
quality_height=train_quality_height,
transform_final=train_transform_final,
**kwargs)
test_output_dataset_filepath = os.path.join(workspace_path, 'dataset', 'test_dataset.py') \
if test_dataset_raw_filepath is not None else None
generate_dataset_from_raw(raw_dataset_filepath=test_dataset_raw_filepath,
output_dataset_filepath=test_output_dataset_filepath,
quality_width=test_quality_width,
quality_height=test_quality_height,
transform_final=test_transform_final,
**kwargs)
run_vmaf_cv(
train_dataset_filepath=train_output_dataset_filepath,
test_dataset_filepath=test_output_dataset_filepath,
param_filepath=param_filepath,
output_model_filepath=output_model_filepath,
**kwargs
)
|
[
"vmaf.core.train_test_model.TrainTestModel.find_subclass",
"vmaf.core.quality_runner.VmafQualityRunner.predict_with_model",
"numpy.array",
"vmaf.core.local_explainer.LocalExplainer",
"vmaf.config.VmafConfig.file_result_store_path",
"vmaf.config.VmafConfig.workspace_path",
"vmaf.plt.subplots",
"vmaf.core.feature_assembler.FeatureAssembler",
"vmaf.core.cross_validation.ModelCrossValidation.run_kfold_cross_validation",
"vmaf.config.DisplayConfig.show",
"vmaf.tools.misc.get_stdout_logger",
"numpy.shape",
"vmaf.tools.misc.import_python_file",
"vmaf.core.quality_runner.VmafQualityRunner.set_clip_score",
"vmaf.core.asset.Asset",
"vmaf.plt.tight_layout",
"vmaf.tools.misc.close_logger",
"vmaf.core.result_store.FileSystemResultStore",
"os.path.join",
"vmaf.config.VmafConfig.workdir_path",
"vmaf.core.quality_runner.VmafQualityRunner.set_transform_score",
"vmaf.tools.misc.get_file_name_without_extension",
"vmaf.tools.misc.indices"
] |
[((19415, 19682), 'vmaf.core.feature_assembler.FeatureAssembler', 'FeatureAssembler', ([], {'feature_dict': 'feature_param.feature_dict', 'feature_option_dict': 'None', 'assets': 'train_assets', 'logger': 'logger', 'fifo_mode': 'fifo_mode', 'delete_workdir': '(True)', 'result_store': 'result_store', 'optional_dict': 'None', 'optional_dict2': 'None', 'parallelize': 'parallelize'}), '(feature_dict=feature_param.feature_dict,\n feature_option_dict=None, assets=train_assets, logger=logger, fifo_mode\n =fifo_mode, delete_workdir=True, result_store=result_store,\n optional_dict=None, optional_dict2=None, parallelize=parallelize)\n', (19431, 19682), False, 'from vmaf.core.feature_assembler import FeatureAssembler\n'), ((20037, 20077), 'vmaf.core.train_test_model.TrainTestModel.find_subclass', 'TrainTestModel.find_subclass', (['model_type'], {}), '(model_type)\n', (20065, 20077), False, 'from vmaf.core.train_test_model import TrainTestModel, RegressorMixin, ClassifierMixin\n'), ((25696, 25946), 'vmaf.core.feature_assembler.FeatureAssembler', 'FeatureAssembler', ([], {'feature_dict': 'feature_param.feature_dict', 'feature_option_dict': 'None', 'assets': 'assets', 'logger': 'logger', 'delete_workdir': '(True)', 'result_store': 'result_store', 'optional_dict': 'None', 'optional_dict2': 'None', 'parallelize': '(True)', 'fifo_mode': '(True)'}), '(feature_dict=feature_param.feature_dict,\n feature_option_dict=None, assets=assets, logger=logger, delete_workdir=\n True, result_store=result_store, optional_dict=None, optional_dict2=\n None, parallelize=True, fifo_mode=True)\n', (25712, 25946), False, 'from vmaf.core.feature_assembler import FeatureAssembler\n'), ((26225, 26277), 'vmaf.core.train_test_model.TrainTestModel.find_subclass', 'TrainTestModel.find_subclass', (['model_param.model_type'], {}), '(model_param.model_type)\n', (26253, 26277), False, 'from vmaf.core.train_test_model import TrainTestModel, RegressorMixin, ClassifierMixin\n'), ((26342, 26468), 'vmaf.core.cross_validation.ModelCrossValidation.run_kfold_cross_validation', 'ModelCrossValidation.run_kfold_cross_validation', (['model_class', 'model_param.model_param_dict', 'results', 'kfold'], {'logger': 'logger'}), '(model_class, model_param.\n model_param_dict, results, kfold, logger=logger)\n', (26389, 26468), False, 'from vmaf.core.cross_validation import ModelCrossValidation\n'), ((27837, 27856), 'vmaf.tools.misc.get_stdout_logger', 'get_stdout_logger', ([], {}), '()\n', (27854, 27856), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n'), ((27876, 27915), 'vmaf.core.result_store.FileSystemResultStore', 'FileSystemResultStore', (['result_store_dir'], {}), '(result_store_dir)\n', (27897, 27915), False, 'from vmaf.core.result_store import FileSystemResultStore\n'), ((27937, 27979), 'vmaf.tools.misc.import_python_file', 'import_python_file', (['train_dataset_filepath'], {}), '(train_dataset_filepath)\n', (27955, 27979), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n'), ((28101, 28135), 'vmaf.tools.misc.import_python_file', 'import_python_file', (['param_filepath'], {}), '(param_filepath)\n', (28119, 28135), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n'), ((28208, 28278), 'vmaf.plt.subplots', 'plt.subplots', ([], {'figsize': '(5 * ncols, 5 * nrows)', 'nrows': 'nrows', 'ncols': 'ncols'}), '(figsize=(5 * ncols, 5 * nrows), nrows=nrows, ncols=ncols)\n', (28220, 28278), False, 'from vmaf import plt\n'), ((29008, 29026), 'vmaf.plt.tight_layout', 'plt.tight_layout', ([], {}), '()\n', (29024, 29026), False, 'from vmaf import plt\n'), ((29055, 29075), 'vmaf.tools.misc.close_logger', 'close_logger', (['logger'], {}), '(logger)\n', (29067, 29075), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n'), ((29275, 29310), 'vmaf.config.VmafConfig.file_result_store_path', 'VmafConfig.file_result_store_path', ([], {}), '()\n', (29308, 29310), False, 'from vmaf.config import VmafConfig, DisplayConfig\n'), ((29351, 29370), 'vmaf.tools.misc.get_stdout_logger', 'get_stdout_logger', ([], {}), '()\n', (29368, 29370), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n'), ((29390, 29429), 'vmaf.core.result_store.FileSystemResultStore', 'FileSystemResultStore', (['result_store_dir'], {}), '(result_store_dir)\n', (29411, 29429), False, 'from vmaf.core.result_store import FileSystemResultStore\n'), ((29444, 29480), 'vmaf.tools.misc.import_python_file', 'import_python_file', (['dataset_filepath'], {}), '(dataset_filepath)\n', (29462, 29480), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n'), ((29493, 29527), 'vmaf.tools.misc.import_python_file', 'import_python_file', (['param_filepath'], {}), '(param_filepath)\n', (29511, 29527), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n'), ((29543, 29589), 'vmaf.plt.subplots', 'plt.subplots', ([], {'figsize': '(5, 5)', 'nrows': '(1)', 'ncols': '(1)'}), '(figsize=(5, 5), nrows=1, ncols=1)\n', (29555, 29589), False, 'from vmaf import plt\n'), ((29769, 29787), 'vmaf.plt.tight_layout', 'plt.tight_layout', ([], {}), '()\n', (29785, 29787), False, 'from vmaf import plt\n'), ((29816, 29836), 'vmaf.tools.misc.close_logger', 'close_logger', (['logger'], {}), '(logger)\n', (29828, 29836), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n'), ((30002, 30037), 'vmaf.config.VmafConfig.file_result_store_path', 'VmafConfig.file_result_store_path', ([], {}), '()\n', (30035, 30037), False, 'from vmaf.config import VmafConfig, DisplayConfig\n'), ((30315, 30356), 'vmaf.tools.misc.import_python_file', 'import_python_file', (['test_dataset_filepath'], {}), '(test_dataset_filepath)\n', (30333, 30356), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n'), ((30551, 30590), 'vmaf.core.result_store.FileSystemResultStore', 'FileSystemResultStore', (['result_store_dir'], {}), '(result_store_dir)\n', (30572, 30590), False, 'from vmaf.core.result_store import FileSystemResultStore\n'), ((30685, 30944), 'vmaf.core.feature_assembler.FeatureAssembler', 'FeatureAssembler', ([], {'feature_dict': "model.model_dict['feature_dict']", 'feature_option_dict': 'None', 'assets': 'test_assets', 'logger': 'None', 'fifo_mode': '(True)', 'delete_workdir': '(True)', 'result_store': 'result_store', 'optional_dict': 'None', 'optional_dict2': 'None', 'parallelize': '(True)'}), "(feature_dict=model.model_dict['feature_dict'],\n feature_option_dict=None, assets=test_assets, logger=None, fifo_mode=\n True, delete_workdir=True, result_store=result_store, optional_dict=\n None, optional_dict2=None, parallelize=True)\n", (30701, 30944), False, 'from vmaf.core.feature_assembler import FeatureAssembler\n'), ((31294, 31331), 'vmaf.core.local_explainer.LocalExplainer', 'LocalExplainer', ([], {'neighbor_samples': '(1000)'}), '(neighbor_samples=1000)\n', (31308, 31331), False, 'from vmaf.core.local_explainer import LocalExplainer\n'), ((31582, 31602), 'vmaf.config.DisplayConfig.show', 'DisplayConfig.show', ([], {}), '()\n', (31600, 31602), False, 'from vmaf.config import VmafConfig, DisplayConfig\n'), ((33594, 33653), 'os.path.join', 'os.path.join', (['workspace_path', '"""dataset"""', '"""train_dataset.py"""'], {}), "(workspace_path, 'dataset', 'train_dataset.py')\n", (33606, 33653), False, 'import os\n'), ((1210, 1235), 'vmaf.config.VmafConfig.workdir_path', 'VmafConfig.workdir_path', ([], {}), '()\n', (1233, 1235), False, 'from vmaf.config import VmafConfig, DisplayConfig\n'), ((20588, 20659), 'vmaf.core.quality_runner.VmafQualityRunner.set_clip_score', 'VmafQualityRunner.set_clip_score', (['model', "model_param_dict['score_clip']"], {}), "(model, model_param_dict['score_clip'])\n", (20620, 20659), False, 'from vmaf.core.quality_runner import VmafQualityRunner\n'), ((20714, 20800), 'vmaf.core.quality_runner.VmafQualityRunner.set_transform_score', 'VmafQualityRunner.set_transform_score', (['model', "model_param_dict['score_transform']"], {}), "(model, model_param_dict[\n 'score_transform'])\n", (20751, 20800), False, 'from vmaf.core.quality_runner import VmafQualityRunner\n'), ((20817, 20880), 'vmaf.core.quality_runner.VmafQualityRunner.predict_with_model', 'VmafQualityRunner.predict_with_model', (['model', 'train_xs'], {}), '(model, train_xs, **kwargs)\n', (20853, 20880), False, 'from vmaf.core.quality_runner import VmafQualityRunner\n'), ((23162, 23421), 'vmaf.core.feature_assembler.FeatureAssembler', 'FeatureAssembler', ([], {'feature_dict': 'feature_param.feature_dict', 'feature_option_dict': 'None', 'assets': 'test_assets', 'logger': 'logger', 'fifo_mode': 'fifo_mode', 'delete_workdir': '(True)', 'result_store': 'result_store', 'optional_dict': 'None', 'optional_dict2': 'None', 'parallelize': '(True)'}), '(feature_dict=feature_param.feature_dict,\n feature_option_dict=None, assets=test_assets, logger=logger, fifo_mode=\n fifo_mode, delete_workdir=True, result_store=result_store,\n optional_dict=None, optional_dict2=None, parallelize=True)\n', (23178, 23421), False, 'from vmaf.core.feature_assembler import FeatureAssembler\n'), ((25324, 25379), 'vmaf.tools.misc.indices', 'indices', (['content_ids', '(lambda x: x in curr_content_group)'], {}), '(content_ids, lambda x: x in curr_content_group)\n', (25331, 25379), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n'), ((27787, 27822), 'vmaf.config.VmafConfig.file_result_store_path', 'VmafConfig.file_result_store_path', ([], {}), '()\n', (27820, 27822), False, 'from vmaf.config import VmafConfig, DisplayConfig\n'), ((27999, 28040), 'vmaf.tools.misc.import_python_file', 'import_python_file', (['test_dataset_filepath'], {}), '(test_dataset_filepath)\n', (28017, 28040), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n'), ((33529, 33556), 'vmaf.config.VmafConfig.workspace_path', 'VmafConfig.workspace_path', ([], {}), '()\n', (33554, 33556), False, 'from vmaf.config import VmafConfig, DisplayConfig\n'), ((34050, 34108), 'os.path.join', 'os.path.join', (['workspace_path', '"""dataset"""', '"""test_dataset.py"""'], {}), "(workspace_path, 'dataset', 'test_dataset.py')\n", (34062, 34108), False, 'import os\n'), ((9667, 9869), 'vmaf.core.asset.Asset', 'Asset', ([], {'dataset': 'data_set_name', 'content_id': "dis_video['content_id']", 'asset_id': "dis_video['asset_id']", 'workdir_root': 'workdir_root', 'ref_path': 'ref_path', 'dis_path': "dis_video['path']", 'asset_dict': 'asset_dict'}), "(dataset=data_set_name, content_id=dis_video['content_id'], asset_id=\n dis_video['asset_id'], workdir_root=workdir_root, ref_path=ref_path,\n dis_path=dis_video['path'], asset_dict=asset_dict)\n", (9672, 9869), False, 'from vmaf.core.asset import Asset\n'), ((14458, 14490), 'numpy.shape', 'np.shape', (['predictions_all_models'], {}), '(predictions_all_models)\n', (14466, 14490), True, 'import numpy as np\n'), ((23875, 23937), 'vmaf.core.quality_runner.VmafQualityRunner.predict_with_model', 'VmafQualityRunner.predict_with_model', (['model', 'test_xs'], {}), '(model, test_xs, **kwargs)\n', (23911, 23937), False, 'from vmaf.core.quality_runner import VmafQualityRunner\n'), ((14393, 14425), 'numpy.array', 'np.array', (['predictions_all_models'], {}), '(predictions_all_models)\n', (14401, 14425), True, 'import numpy as np\n'), ((16811, 16858), 'vmaf.tools.misc.get_file_name_without_extension', 'get_file_name_without_extension', (['asset.dis_path'], {}), '(asset.dis_path)\n', (16842, 16858), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n'), ((30194, 30245), 'vmaf.tools.misc.get_file_name_without_extension', 'get_file_name_without_extension', (['tasset[1].dis_path'], {}), '(tasset[1].dis_path)\n', (30225, 30245), False, 'from vmaf.tools.misc import indices, get_stdout_logger, import_python_file, close_logger, get_file_name_without_extension\n')]
|
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from __future__ import print_function
import os
import unittest
import paddle.fluid as fluid
import paddle.fluid.core as core
from paddle.fluid.dygraph.nn import Embedding, Linear
import paddle.fluid.framework as framework
from paddle.fluid.optimizer import Adam
from paddle.fluid.dygraph.base import to_variable
from paddle.fluid.dygraph.learning_rate_scheduler import LearningRateDecay
from test_imperative_base import new_program_scope
import numpy as np
import six
class SimpleLSTMRNN(fluid.Layer):
def __init__(self,
hidden_size,
num_steps,
num_layers=2,
init_scale=0.1,
dropout=None):
super(SimpleLSTMRNN, self).__init__()
self._hidden_size = hidden_size
self._num_layers = num_layers
self._init_scale = init_scale
self._dropout = dropout
self._input = None
self._num_steps = num_steps
self.cell_array = []
self.hidden_array = []
self.weight_1_arr = []
self.weight_2_arr = []
self.bias_arr = []
self.mask_array = []
for i in range(self._num_layers):
weight_1 = self.create_parameter(
attr=fluid.ParamAttr(
initializer=fluid.initializer.UniformInitializer(
low=-self._init_scale, high=self._init_scale)),
shape=[self._hidden_size * 2, self._hidden_size * 4],
dtype="float32",
default_initializer=fluid.initializer.UniformInitializer(
low=-self._init_scale, high=self._init_scale))
self.weight_1_arr.append(self.add_parameter('w_%d' % i, weight_1))
bias_1 = self.create_parameter(
attr=fluid.ParamAttr(
initializer=fluid.initializer.UniformInitializer(
low=-self._init_scale, high=self._init_scale)),
shape=[self._hidden_size * 4],
dtype="float32",
default_initializer=fluid.initializer.Constant(0.0))
self.bias_arr.append(self.add_parameter('b_%d' % i, bias_1))
def forward(self, input_embedding, init_hidden=None, init_cell=None):
self.cell_array = []
self.hidden_array = []
for i in range(self._num_layers):
pre_hidden = fluid.layers.slice(
init_hidden, axes=[0], starts=[i], ends=[i + 1])
pre_cell = fluid.layers.slice(
init_cell, axes=[0], starts=[i], ends=[i + 1])
pre_hidden = fluid.layers.reshape(
pre_hidden, shape=[-1, self._hidden_size])
pre_cell = fluid.layers.reshape(
pre_cell, shape=[-1, self._hidden_size])
self.hidden_array.append(pre_hidden)
self.cell_array.append(pre_cell)
res = []
for index in range(self._num_steps):
self._input = fluid.layers.slice(
input_embedding, axes=[1], starts=[index], ends=[index + 1])
self._input = fluid.layers.reshape(
self._input, shape=[-1, self._hidden_size])
for k in range(self._num_layers):
pre_hidden = self.hidden_array[k]
pre_cell = self.cell_array[k]
weight_1 = self.weight_1_arr[k]
bias = self.bias_arr[k]
nn = fluid.layers.concat([self._input, pre_hidden], 1)
gate_input = fluid.layers.matmul(x=nn, y=weight_1)
gate_input = fluid.layers.elementwise_add(gate_input, bias)
i, j, f, o = fluid.layers.split(
gate_input, num_or_sections=4, dim=-1)
c = pre_cell * fluid.layers.sigmoid(f) + fluid.layers.sigmoid(
i) * fluid.layers.tanh(j)
m = fluid.layers.tanh(c) * fluid.layers.sigmoid(o)
self.hidden_array[k] = m
self.cell_array[k] = c
self._input = m
if self._dropout is not None and self._dropout > 0.0:
self._input = fluid.layers.dropout(
self._input,
dropout_prob=self._dropout,
dropout_implementation='upscale_in_train')
res.append(
fluid.layers.reshape(
self._input, shape=[1, -1, self._hidden_size]))
real_res = fluid.layers.concat(res, 0)
real_res = fluid.layers.transpose(x=real_res, perm=[1, 0, 2])
last_hidden = fluid.layers.concat(self.hidden_array, 1)
last_hidden = fluid.layers.reshape(
last_hidden, shape=[-1, self._num_layers, self._hidden_size])
last_hidden = fluid.layers.transpose(x=last_hidden, perm=[1, 0, 2])
last_cell = fluid.layers.concat(self.cell_array, 1)
last_cell = fluid.layers.reshape(
last_cell, shape=[-1, self._num_layers, self._hidden_size])
last_cell = fluid.layers.transpose(x=last_cell, perm=[1, 0, 2])
return real_res, last_hidden, last_cell
class PtbModel(fluid.Layer):
def __init__(self,
hidden_size,
vocab_size,
num_layers=2,
num_steps=20,
init_scale=0.1,
dropout=None):
super(PtbModel, self).__init__()
self.hidden_size = hidden_size
self.vocab_size = vocab_size
self.init_scale = init_scale
self.num_layers = num_layers
self.num_steps = num_steps
self.dropout = dropout
self.simple_lstm_rnn = SimpleLSTMRNN(
hidden_size,
num_steps,
num_layers=num_layers,
init_scale=init_scale,
dropout=dropout)
self.embedding = Embedding(
size=[vocab_size, hidden_size],
dtype='float32',
is_sparse=False,
param_attr=fluid.ParamAttr(
name='embedding_para',
initializer=fluid.initializer.UniformInitializer(
low=-init_scale, high=init_scale)))
self.softmax_weight = self.create_parameter(
attr=fluid.ParamAttr(),
shape=[self.hidden_size, self.vocab_size],
dtype="float32",
default_initializer=fluid.initializer.UniformInitializer(
low=-self.init_scale, high=self.init_scale))
self.softmax_bias = self.create_parameter(
attr=fluid.ParamAttr(),
shape=[self.vocab_size],
dtype="float32",
default_initializer=fluid.initializer.UniformInitializer(
low=-self.init_scale, high=self.init_scale))
def forward(self, input, label, init_hidden, init_cell):
init_h = fluid.layers.reshape(
init_hidden, shape=[self.num_layers, -1, self.hidden_size])
init_c = fluid.layers.reshape(
init_cell, shape=[self.num_layers, -1, self.hidden_size])
x_emb = self.embedding(input)
x_emb = fluid.layers.reshape(
x_emb, shape=[-1, self.num_steps, self.hidden_size])
if self.dropout is not None and self.dropout > 0.0:
x_emb = fluid.layers.dropout(
x_emb,
dropout_prob=self.drop_out,
dropout_implementation='upscale_in_train')
rnn_out, last_hidden, last_cell = self.simple_lstm_rnn(x_emb, init_h,
init_c)
rnn_out = fluid.layers.reshape(
rnn_out, shape=[-1, self.num_steps, self.hidden_size])
projection = fluid.layers.matmul(rnn_out, self.softmax_weight)
projection = fluid.layers.elementwise_add(projection, self.softmax_bias)
projection = fluid.layers.reshape(
projection, shape=[-1, self.vocab_size])
loss = fluid.layers.softmax_with_cross_entropy(
logits=projection, label=label, soft_label=False)
loss = fluid.layers.reshape(loss, shape=[-1, self.num_steps])
loss = fluid.layers.reduce_mean(loss, dim=[0])
loss = fluid.layers.reduce_sum(loss)
return loss, last_hidden, last_cell
class TestDygraphPtbRnn(unittest.TestCase):
def setUp(self):
seed = 90
hidden_size = 10
vocab_size = 1000
num_layers = 1
num_steps = 3
init_scale = 0.1
batch_size = 4
batch_num = 200
with fluid.dygraph.guard():
fluid.default_startup_program().random_seed = seed
fluid.default_main_program().random_seed = seed
# TODO: marsyang1993 Change seed to
ptb_model = PtbModel(
hidden_size=hidden_size,
vocab_size=vocab_size,
num_layers=num_layers,
num_steps=num_steps,
init_scale=init_scale)
bd = []
lr_arr = [1.0]
# this a fake lr decay strategy
for i in range(1, 10):
bd.append(100 * i)
new_lr = 1.0
lr_arr.append(new_lr)
place = fluid.CPUPlace() if not core.is_compiled_with_cuda(
) else fluid.CUDAPlace(0)
adam = Adam(
learning_rate=fluid.layers.piecewise_decay(
boundaries=bd, values=lr_arr),
parameter_list=ptb_model.parameters())
dy_param_updated = dict()
dy_param_init = dict()
dy_loss = None
last_hidden = None
last_cell = None
for i in range(batch_num):
x_data = np.arange(12).reshape(4, 3).astype('int64')
y_data = np.arange(1, 13).reshape(4, 3).astype('int64')
y_data = y_data.reshape((-1, 1))
init_hidden_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
init_cell_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
x = to_variable(x_data)
y = to_variable(y_data)
init_hidden = to_variable(init_hidden_data)
init_cell = to_variable(init_cell_data)
dy_loss, last_hidden, last_cell = ptb_model(x, y, init_hidden,
init_cell)
if i == 0:
for param in ptb_model.parameters():
dy_param_init[param.name] = param.numpy()
dy_loss.backward()
adam.minimize(dy_loss)
ptb_model.clear_gradients()
if i == batch_num - 1:
for param in ptb_model.parameters():
dy_param_updated[param.name] = param.numpy()
# check optimizer
self.opti_dict = adam.state_dict()
self.base_opti = {}
for k, v in self.opti_dict.items():
self.base_opti[v.name] = v.numpy()
self.assertTrue(np.sum(np.abs(v.numpy())) != 0)
fluid.save_dygraph(self.opti_dict, "./test_dy")
self.state_dict = ptb_model.state_dict()
self.model_base = {}
for k, v in self.state_dict.items():
np_t = v.numpy()
self.model_base[k] = np_t
fluid.save_dygraph(self.state_dict, "./test_dy")
def testLoadAndSetVarBase(self):
seed = 90
hidden_size = 10
vocab_size = 1000
num_layers = 1
num_steps = 3
init_scale = 0.1
batch_size = 4
batch_num = 200
with fluid.dygraph.guard():
fluid.default_startup_program().random_seed = seed
fluid.default_main_program().random_seed = seed
# TODO: marsyang1993 Change seed to
ptb_model = PtbModel(
hidden_size=hidden_size,
vocab_size=vocab_size,
num_layers=num_layers,
num_steps=num_steps,
init_scale=init_scale)
bd = []
lr_arr = [1.0]
# this a fake lr decay strategy
for i in range(1, 10):
bd.append(100 * i)
new_lr = 1.0
lr_arr.append(new_lr)
place = fluid.CPUPlace() if not core.is_compiled_with_cuda(
) else fluid.CUDAPlace(0)
adam = Adam(
learning_rate=fluid.layers.piecewise_decay(
boundaries=bd, values=lr_arr),
parameter_list=ptb_model.parameters())
dy_param_updated = dict()
dy_param_init = dict()
dy_loss = None
last_hidden = None
last_cell = None
for i in range(batch_num):
x_data = np.arange(12).reshape(4, 3).astype('int64')
y_data = np.arange(1, 13).reshape(4, 3).astype('int64')
y_data = y_data.reshape((-1, 1))
init_hidden_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
init_cell_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
x = to_variable(x_data)
y = to_variable(y_data)
init_hidden = to_variable(init_hidden_data)
init_cell = to_variable(init_cell_data)
dy_loss, last_hidden, last_cell = ptb_model(x, y, init_hidden,
init_cell)
if i == 0:
for param in ptb_model.parameters():
dy_param_init[param.name] = param.numpy()
dy_loss.backward()
adam.minimize(dy_loss)
ptb_model.clear_gradients()
if i == batch_num - 1:
for param in ptb_model.parameters():
dy_param_updated[param.name] = param.numpy()
# check optimizer
opti_dict = adam.state_dict()
# set to zero
for k, v in opti_dict.items():
np_t = v.numpy()
var = v.value().get_tensor()
var.set(np.zeros_like(np_t), place)
self.assertTrue(np.sum(np.abs(v.numpy())) == 0)
if isinstance(adam._learning_rate, LearningRateDecay):
adam._learning_rate.step_num = 0
para_state_dict, opti_state_dict = fluid.load_dygraph("./test_dy")
adam.set_dict(opti_state_dict)
opti_dict = adam.state_dict()
for k, v in opti_dict.items():
self.assertTrue(
np.array_equal(v.numpy(), self.base_opti[v.name]))
# check parameter
state_dict = ptb_model.state_dict()
for k, v in state_dict.items():
np_t = v.numpy()
var = v.value().get_tensor()
var.set(np.zeros_like(np_t), place)
ptb_model.set_dict(para_state_dict)
state_dict = ptb_model.state_dict()
for k, v in state_dict.items():
new_t = v.numpy()
base_t = self.model_base[k]
self.assertTrue(np.array_equal(new_t, base_t))
def testSetVariable(self):
seed = 90
hidden_size = 10
vocab_size = 1000
num_layers = 1
num_steps = 3
init_scale = 0.1
batch_size = 4
batch_num = 200
with fluid.dygraph.guard():
fluid.default_startup_program().random_seed = seed
fluid.default_main_program().random_seed = seed
# TODO: marsyang1993 Change seed to
ptb_model = PtbModel(
hidden_size=hidden_size,
vocab_size=vocab_size,
num_layers=num_layers,
num_steps=num_steps,
init_scale=init_scale)
bd = []
lr_arr = [1.0]
# this a fake lr decay strategy
for i in range(1, 10):
bd.append(100 * i)
new_lr = 1.0
lr_arr.append(new_lr)
place = fluid.CPUPlace() if not core.is_compiled_with_cuda(
) else fluid.CUDAPlace(0)
adam = Adam(
learning_rate=fluid.layers.piecewise_decay(
boundaries=bd, values=lr_arr),
parameter_list=ptb_model.parameters())
dy_param_updated = dict()
dy_param_init = dict()
dy_loss = None
last_hidden = None
last_cell = None
for i in range(batch_num):
x_data = np.arange(12).reshape(4, 3).astype('int64')
y_data = np.arange(1, 13).reshape(4, 3).astype('int64')
y_data = y_data.reshape((-1, 1))
init_hidden_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
init_cell_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
x = to_variable(x_data)
y = to_variable(y_data)
init_hidden = to_variable(init_hidden_data)
init_cell = to_variable(init_cell_data)
dy_loss, last_hidden, last_cell = ptb_model(x, y, init_hidden,
init_cell)
if i == 0:
for param in ptb_model.parameters():
dy_param_init[param.name] = param.numpy()
dy_loss.backward()
adam.minimize(dy_loss)
ptb_model.clear_gradients()
if i == batch_num - 1:
for param in ptb_model.parameters():
dy_param_updated[param.name] = param.numpy()
# check optimizer
opti_dict = adam.state_dict()
# set to zero
for k, v in opti_dict.items():
np_t = v.numpy()
var = v.value().get_tensor()
var.set(np.zeros_like(np_t), place)
self.assertTrue(np.sum(np.abs(v.numpy())) == 0)
if isinstance(adam._learning_rate, LearningRateDecay):
adam._learning_rate.step_num = 0
adam.set_dict(self.opti_dict)
opti_dict = adam.state_dict()
for k, v in opti_dict.items():
self.assertTrue(
np.array_equal(v.numpy(), self.base_opti[v.name]))
# check parameter
state_dict = ptb_model.state_dict()
for k, v in state_dict.items():
np_t = v.numpy()
var = v.value().get_tensor()
var.set(np.zeros_like(np_t), place)
ptb_model.set_dict(self.state_dict)
state_dict = ptb_model.state_dict()
for k, v in state_dict.items():
new_t = v.numpy()
base_t = self.model_base[k]
self.assertTrue(np.array_equal(new_t, base_t))
def testSetNumpy(self):
seed = 90
hidden_size = 10
vocab_size = 1000
num_layers = 1
num_steps = 3
init_scale = 0.1
batch_size = 4
batch_num = 200
with fluid.dygraph.guard():
fluid.default_startup_program().random_seed = seed
fluid.default_main_program().random_seed = seed
# TODO: marsyang1993 Change seed to
ptb_model = PtbModel(
hidden_size=hidden_size,
vocab_size=vocab_size,
num_layers=num_layers,
num_steps=num_steps,
init_scale=init_scale)
bd = []
lr_arr = [1.0]
# this a fake lr decay strategy
for i in range(1, 10):
bd.append(100 * i)
new_lr = 1.0
lr_arr.append(new_lr)
place = fluid.CPUPlace() if not core.is_compiled_with_cuda(
) else fluid.CUDAPlace(0)
adam = Adam(
learning_rate=fluid.layers.piecewise_decay(
boundaries=bd, values=lr_arr),
parameter_list=ptb_model.parameters())
dy_param_updated = dict()
dy_param_init = dict()
dy_loss = None
last_hidden = None
last_cell = None
for i in range(batch_num):
x_data = np.arange(12).reshape(4, 3).astype('int64')
y_data = np.arange(1, 13).reshape(4, 3).astype('int64')
y_data = y_data.reshape((-1, 1))
init_hidden_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
init_cell_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
x = to_variable(x_data)
y = to_variable(y_data)
init_hidden = to_variable(init_hidden_data)
init_cell = to_variable(init_cell_data)
dy_loss, last_hidden, last_cell = ptb_model(x, y, init_hidden,
init_cell)
if i == 0:
for param in ptb_model.parameters():
dy_param_init[param.name] = param.numpy()
dy_loss.backward()
adam.minimize(dy_loss)
ptb_model.clear_gradients()
if i == batch_num - 1:
for param in ptb_model.parameters():
dy_param_updated[param.name] = param.numpy()
# check optimizer
opti_dict = adam.state_dict()
np_opti_dict = {}
# set to zero
for k, v in opti_dict.items():
np_t = v.numpy()
np_opti_dict[v.name] = np_t
var = v.value().get_tensor()
var.set(np.zeros_like(np_t), place)
self.assertTrue(np.sum(np.abs(v.numpy())) == 0)
if isinstance(adam._learning_rate, LearningRateDecay):
adam._learning_rate.step_num = 0
adam.set_dict(np_opti_dict)
opti_dict = adam.state_dict()
for k, v in opti_dict.items():
self.assertTrue(
np.array_equal(v.numpy(), self.base_opti[v.name]))
# check parameter
state_dict = ptb_model.state_dict()
np_state_dict = {}
for k, v in state_dict.items():
np_t = v.numpy()
np_state_dict[k] = np_t
var = v.value().get_tensor()
var.set(np.zeros_like(np_t), place)
ptb_model.set_dict(np_state_dict)
state_dict = ptb_model.state_dict()
for k, v in state_dict.items():
new_t = v.numpy()
base_t = self.model_base[k]
self.assertTrue(np.array_equal(new_t, base_t))
def testSetVariableBeforeTrain(self):
seed = 90
hidden_size = 10
vocab_size = 1000
num_layers = 1
num_steps = 3
init_scale = 0.1
batch_size = 4
batch_num = 200
with fluid.dygraph.guard():
fluid.default_startup_program().random_seed = seed
fluid.default_main_program().random_seed = seed
# TODO: marsyang1993 Change seed to
ptb_model = PtbModel(
hidden_size=hidden_size,
vocab_size=vocab_size,
num_layers=num_layers,
num_steps=num_steps,
init_scale=init_scale)
place = fluid.CPUPlace() if not core.is_compiled_with_cuda(
) else fluid.CUDAPlace(0)
adam = Adam(
learning_rate=0.0,
beta1=0.8,
beta2=0.6,
parameter_list=ptb_model.parameters())
dy_param_updated = dict()
dy_param_init = dict()
dy_loss = None
last_hidden = None
last_cell = None
adam.set_dict(self.opti_dict)
ptb_model.set_dict(self.state_dict)
for i in range(1):
x_data = np.arange(12).reshape(4, 3).astype('int64')
y_data = np.arange(1, 13).reshape(4, 3).astype('int64')
y_data = y_data.reshape((-1, 1))
init_hidden_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
init_cell_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
x = to_variable(x_data)
y = to_variable(y_data)
init_hidden = to_variable(init_hidden_data)
init_cell = to_variable(init_cell_data)
dy_loss, last_hidden, last_cell = ptb_model(x, y, init_hidden,
init_cell)
dy_loss.backward()
adam.minimize(dy_loss)
ptb_model.clear_gradients()
opti_dict = adam.state_dict()
for k, v in opti_dict.items():
if k == "global_step":
self.assertTrue(
np.array_equal(v.numpy(), self.base_opti[v.name] + 1))
if k.find("beta1_pow_acc_0") > 0:
self.assertTrue(
np.array_equal(v.numpy(), self.base_opti[v.name] *
adam._beta1))
if k.find("beta2_pow_acc_0") > 0:
self.assertTrue(
np.array_equal(v.numpy(), self.base_opti[v.name] *
adam._beta2))
state_dict = ptb_model.state_dict()
for k, v in state_dict.items():
new_t = v.numpy()
base_t = self.model_base[k]
self.assertTrue(np.array_equal(new_t, base_t))
def testLoadAndSetVarBaseBeforeTrain(self):
seed = 90
hidden_size = 10
vocab_size = 1000
num_layers = 1
num_steps = 3
init_scale = 0.1
batch_size = 4
batch_num = 200
with fluid.dygraph.guard():
fluid.default_startup_program().random_seed = seed
fluid.default_main_program().random_seed = seed
# TODO: marsyang1993 Change seed to
ptb_model = PtbModel(
hidden_size=hidden_size,
vocab_size=vocab_size,
num_layers=num_layers,
num_steps=num_steps,
init_scale=init_scale)
bd = []
lr_arr = [0.0]
# this a fake lr decay strategy
for i in range(1, 10):
bd.append(100 * i)
# set lr to zero not update parameter
new_lr = 0.0
lr_arr.append(new_lr)
place = fluid.CPUPlace() if not core.is_compiled_with_cuda(
) else fluid.CUDAPlace(0)
adam = Adam(
learning_rate=0.0,
beta1=0.8,
beta2=0.6,
parameter_list=ptb_model.parameters())
dy_param_updated = dict()
dy_param_init = dict()
dy_loss = None
last_hidden = None
last_cell = None
state_dict, opti_dict = fluid.load_dygraph("./test_dy")
adam.set_dict(opti_dict)
ptb_model.set_dict(state_dict)
for i in range(1):
x_data = np.arange(12).reshape(4, 3).astype('int64')
y_data = np.arange(1, 13).reshape(4, 3).astype('int64')
y_data = y_data.reshape((-1, 1))
init_hidden_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
init_cell_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
x = to_variable(x_data)
y = to_variable(y_data)
init_hidden = to_variable(init_hidden_data)
init_cell = to_variable(init_cell_data)
dy_loss, last_hidden, last_cell = ptb_model(x, y, init_hidden,
init_cell)
dy_loss.backward()
adam.minimize(dy_loss)
ptb_model.clear_gradients()
opti_dict = adam.state_dict()
for k, v in opti_dict.items():
if k == "global_step":
self.assertTrue(
np.array_equal(v.numpy(), self.base_opti[v.name] + 1))
if k.find("beta1_pow_acc_0") > 0:
self.assertTrue(
np.array_equal(v.numpy(), self.base_opti[v.name] *
adam._beta1))
if k.find("beta2_pow_acc_0") > 0:
self.assertTrue(
np.array_equal(v.numpy(), self.base_opti[v.name] *
adam._beta2))
# check parameter
state_dict = ptb_model.state_dict()
for k, v in state_dict.items():
new_t = v.numpy()
base_t = self.model_base[k]
self.assertTrue(np.array_equal(new_t, base_t))
def testSetNumpyBeforeTrain(self):
seed = 90
hidden_size = 10
vocab_size = 1000
num_layers = 1
num_steps = 3
init_scale = 0.1
batch_size = 4
batch_num = 200
with fluid.dygraph.guard():
fluid.default_startup_program().random_seed = seed
fluid.default_main_program().random_seed = seed
# TODO: marsyang1993 Change seed to
ptb_model = PtbModel(
hidden_size=hidden_size,
vocab_size=vocab_size,
num_layers=num_layers,
num_steps=num_steps,
init_scale=init_scale)
bd = []
lr_arr = [0.0]
# this a fake lr decay strategy
for i in range(1, 10):
bd.append(100 * i)
# set lr to 0.0, not update parameter
new_lr = 0.0
lr_arr.append(new_lr)
place = fluid.CPUPlace() if not core.is_compiled_with_cuda(
) else fluid.CUDAPlace(0)
adam = Adam(
learning_rate=fluid.layers.piecewise_decay(
boundaries=bd, values=lr_arr),
beta1=0.8,
beta2=0.6,
parameter_list=ptb_model.parameters())
dy_param_updated = dict()
dy_param_init = dict()
dy_loss = None
last_hidden = None
last_cell = None
np_opti_dict = {}
np_state_dict = {}
for k, v in self.opti_dict.items():
np_opti_dict[v.name] = v.numpy()
for k, v in self.state_dict.items():
np_state_dict[k] = v.numpy()
adam.set_dict(np_opti_dict)
ptb_model.set_dict(np_state_dict)
for i in range(1):
x_data = np.arange(12).reshape(4, 3).astype('int64')
y_data = np.arange(1, 13).reshape(4, 3).astype('int64')
y_data = y_data.reshape((-1, 1))
init_hidden_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
init_cell_data = np.zeros(
(num_layers, batch_size, hidden_size), dtype='float32')
x = to_variable(x_data)
y = to_variable(y_data)
init_hidden = to_variable(init_hidden_data)
init_cell = to_variable(init_cell_data)
dy_loss, last_hidden, last_cell = ptb_model(x, y, init_hidden,
init_cell)
dy_loss.backward()
adam.minimize(dy_loss)
ptb_model.clear_gradients()
opti_dict = adam.state_dict()
for k, v in opti_dict.items():
if k == "global_step":
self.assertTrue(
np.array_equal(v.numpy(), self.base_opti[v.name] + 1))
if k.find("beta1_pow_acc_0") > 0:
self.assertTrue(
np.array_equal(v.numpy(), self.base_opti[v.name] *
adam._beta1))
if k.find("beta2_pow_acc_0") > 0:
self.assertTrue(
np.array_equal(v.numpy(), self.base_opti[v.name] *
adam._beta2))
# check parameter
state_dict = ptb_model.state_dict()
for k, v in state_dict.items():
new_t = v.numpy()
base_t = self.model_base[k]
self.assertTrue(np.array_equal(new_t, base_t))
def testOnlyLoadParams(self):
with fluid.dygraph.guard():
emb = fluid.dygraph.Embedding([10, 10])
state_dict = emb.state_dict()
fluid.save_dygraph(state_dict, os.path.join('saved_dy', 'emb_dy'))
para_state_dict, opti_state_dict = fluid.load_dygraph(
os.path.join('saved_dy', 'emb_dy'))
self.assertTrue(opti_state_dict == None)
if __name__ == '__main__':
unittest.main()
|
[
"paddle.fluid.dygraph.guard",
"paddle.fluid.dygraph.Embedding",
"paddle.fluid.layers.split",
"paddle.fluid.layers.piecewise_decay",
"unittest.main",
"paddle.fluid.layers.transpose",
"paddle.fluid.layers.matmul",
"numpy.arange",
"paddle.fluid.layers.reshape",
"paddle.fluid.save_dygraph",
"paddle.fluid.layers.reduce_mean",
"paddle.fluid.layers.tanh",
"paddle.fluid.default_startup_program",
"paddle.fluid.layers.reduce_sum",
"paddle.fluid.default_main_program",
"paddle.fluid.core.is_compiled_with_cuda",
"paddle.fluid.layers.elementwise_add",
"paddle.fluid.ParamAttr",
"paddle.fluid.load_dygraph",
"paddle.fluid.dygraph.base.to_variable",
"paddle.fluid.layers.dropout",
"paddle.fluid.CPUPlace",
"paddle.fluid.layers.sigmoid",
"paddle.fluid.layers.softmax_with_cross_entropy",
"paddle.fluid.layers.concat",
"paddle.fluid.layers.slice",
"os.path.join",
"paddle.fluid.initializer.Constant",
"paddle.fluid.initializer.UniformInitializer",
"numpy.zeros",
"numpy.array_equal",
"paddle.fluid.CUDAPlace",
"numpy.zeros_like"
] |
[((34298, 34313), 'unittest.main', 'unittest.main', ([], {}), '()\n', (34311, 34313), False, 'import unittest\n'), ((5049, 5076), 'paddle.fluid.layers.concat', 'fluid.layers.concat', (['res', '(0)'], {}), '(res, 0)\n', (5068, 5076), True, 'import paddle.fluid as fluid\n'), ((5096, 5146), 'paddle.fluid.layers.transpose', 'fluid.layers.transpose', ([], {'x': 'real_res', 'perm': '[1, 0, 2]'}), '(x=real_res, perm=[1, 0, 2])\n', (5118, 5146), True, 'import paddle.fluid as fluid\n'), ((5169, 5210), 'paddle.fluid.layers.concat', 'fluid.layers.concat', (['self.hidden_array', '(1)'], {}), '(self.hidden_array, 1)\n', (5188, 5210), True, 'import paddle.fluid as fluid\n'), ((5233, 5320), 'paddle.fluid.layers.reshape', 'fluid.layers.reshape', (['last_hidden'], {'shape': '[-1, self._num_layers, self._hidden_size]'}), '(last_hidden, shape=[-1, self._num_layers, self.\n _hidden_size])\n', (5253, 5320), True, 'import paddle.fluid as fluid\n'), ((5351, 5404), 'paddle.fluid.layers.transpose', 'fluid.layers.transpose', ([], {'x': 'last_hidden', 'perm': '[1, 0, 2]'}), '(x=last_hidden, perm=[1, 0, 2])\n', (5373, 5404), True, 'import paddle.fluid as fluid\n'), ((5425, 5464), 'paddle.fluid.layers.concat', 'fluid.layers.concat', (['self.cell_array', '(1)'], {}), '(self.cell_array, 1)\n', (5444, 5464), True, 'import paddle.fluid as fluid\n'), ((5485, 5570), 'paddle.fluid.layers.reshape', 'fluid.layers.reshape', (['last_cell'], {'shape': '[-1, self._num_layers, self._hidden_size]'}), '(last_cell, shape=[-1, self._num_layers, self._hidden_size]\n )\n', (5505, 5570), True, 'import paddle.fluid as fluid\n'), ((5599, 5650), 'paddle.fluid.layers.transpose', 'fluid.layers.transpose', ([], {'x': 'last_cell', 'perm': '[1, 0, 2]'}), '(x=last_cell, perm=[1, 0, 2])\n', (5621, 5650), True, 'import paddle.fluid as fluid\n'), ((7396, 7481), 'paddle.fluid.layers.reshape', 'fluid.layers.reshape', (['init_hidden'], {'shape': '[self.num_layers, -1, self.hidden_size]'}), '(init_hidden, shape=[self.num_layers, -1, self.hidden_size]\n )\n', (7416, 7481), True, 'import paddle.fluid as fluid\n'), ((7508, 7586), 'paddle.fluid.layers.reshape', 'fluid.layers.reshape', (['init_cell'], {'shape': '[self.num_layers, -1, self.hidden_size]'}), '(init_cell, shape=[self.num_layers, -1, self.hidden_size])\n', (7528, 7586), True, 'import paddle.fluid as fluid\n'), ((7655, 7728), 'paddle.fluid.layers.reshape', 'fluid.layers.reshape', (['x_emb'], {'shape': '[-1, self.num_steps, self.hidden_size]'}), '(x_emb, shape=[-1, self.num_steps, self.hidden_size])\n', (7675, 7728), True, 'import paddle.fluid as fluid\n'), ((8137, 8212), 'paddle.fluid.layers.reshape', 'fluid.layers.reshape', (['rnn_out'], {'shape': '[-1, self.num_steps, self.hidden_size]'}), '(rnn_out, shape=[-1, self.num_steps, self.hidden_size])\n', (8157, 8212), True, 'import paddle.fluid as fluid\n'), ((8248, 8297), 'paddle.fluid.layers.matmul', 'fluid.layers.matmul', (['rnn_out', 'self.softmax_weight'], {}), '(rnn_out, self.softmax_weight)\n', (8267, 8297), True, 'import paddle.fluid as fluid\n'), ((8319, 8378), 'paddle.fluid.layers.elementwise_add', 'fluid.layers.elementwise_add', (['projection', 'self.softmax_bias'], {}), '(projection, self.softmax_bias)\n', (8347, 8378), True, 'import paddle.fluid as fluid\n'), ((8400, 8461), 'paddle.fluid.layers.reshape', 'fluid.layers.reshape', (['projection'], {'shape': '[-1, self.vocab_size]'}), '(projection, shape=[-1, self.vocab_size])\n', (8420, 8461), True, 'import paddle.fluid as fluid\n'), ((8490, 8583), 'paddle.fluid.layers.softmax_with_cross_entropy', 'fluid.layers.softmax_with_cross_entropy', ([], {'logits': 'projection', 'label': 'label', 'soft_label': '(False)'}), '(logits=projection, label=label,\n soft_label=False)\n', (8529, 8583), True, 'import paddle.fluid as fluid\n'), ((8608, 8662), 'paddle.fluid.layers.reshape', 'fluid.layers.reshape', (['loss'], {'shape': '[-1, self.num_steps]'}), '(loss, shape=[-1, self.num_steps])\n', (8628, 8662), True, 'import paddle.fluid as fluid\n'), ((8678, 8717), 'paddle.fluid.layers.reduce_mean', 'fluid.layers.reduce_mean', (['loss'], {'dim': '[0]'}), '(loss, dim=[0])\n', (8702, 8717), True, 'import paddle.fluid as fluid\n'), ((8733, 8762), 'paddle.fluid.layers.reduce_sum', 'fluid.layers.reduce_sum', (['loss'], {}), '(loss)\n', (8756, 8762), True, 'import paddle.fluid as fluid\n'), ((2972, 3039), 'paddle.fluid.layers.slice', 'fluid.layers.slice', (['init_hidden'], {'axes': '[0]', 'starts': '[i]', 'ends': '[i + 1]'}), '(init_hidden, axes=[0], starts=[i], ends=[i + 1])\n', (2990, 3039), True, 'import paddle.fluid as fluid\n'), ((3080, 3145), 'paddle.fluid.layers.slice', 'fluid.layers.slice', (['init_cell'], {'axes': '[0]', 'starts': '[i]', 'ends': '[i + 1]'}), '(init_cell, axes=[0], starts=[i], ends=[i + 1])\n', (3098, 3145), True, 'import paddle.fluid as fluid\n'), ((3188, 3251), 'paddle.fluid.layers.reshape', 'fluid.layers.reshape', (['pre_hidden'], {'shape': '[-1, self._hidden_size]'}), '(pre_hidden, shape=[-1, self._hidden_size])\n', (3208, 3251), True, 'import paddle.fluid as fluid\n'), ((3292, 3353), 'paddle.fluid.layers.reshape', 'fluid.layers.reshape', (['pre_cell'], {'shape': '[-1, self._hidden_size]'}), '(pre_cell, shape=[-1, self._hidden_size])\n', (3312, 3353), True, 'import paddle.fluid as fluid\n'), ((3554, 3633), 'paddle.fluid.layers.slice', 'fluid.layers.slice', (['input_embedding'], {'axes': '[1]', 'starts': '[index]', 'ends': '[index + 1]'}), '(input_embedding, axes=[1], starts=[index], ends=[index + 1])\n', (3572, 3633), True, 'import paddle.fluid as fluid\n'), ((3677, 3741), 'paddle.fluid.layers.reshape', 'fluid.layers.reshape', (['self._input'], {'shape': '[-1, self._hidden_size]'}), '(self._input, shape=[-1, self._hidden_size])\n', (3697, 3741), True, 'import paddle.fluid as fluid\n'), ((7822, 7924), 'paddle.fluid.layers.dropout', 'fluid.layers.dropout', (['x_emb'], {'dropout_prob': 'self.drop_out', 'dropout_implementation': '"""upscale_in_train"""'}), "(x_emb, dropout_prob=self.drop_out,\n dropout_implementation='upscale_in_train')\n", (7842, 7924), True, 'import paddle.fluid as fluid\n'), ((9075, 9096), 'paddle.fluid.dygraph.guard', 'fluid.dygraph.guard', ([], {}), '()\n', (9094, 9096), True, 'import paddle.fluid as fluid\n'), ((11724, 11771), 'paddle.fluid.save_dygraph', 'fluid.save_dygraph', (['self.opti_dict', '"""./test_dy"""'], {}), "(self.opti_dict, './test_dy')\n", (11742, 11771), True, 'import paddle.fluid as fluid\n'), ((11997, 12045), 'paddle.fluid.save_dygraph', 'fluid.save_dygraph', (['self.state_dict', '"""./test_dy"""'], {}), "(self.state_dict, './test_dy')\n", (12015, 12045), True, 'import paddle.fluid as fluid\n'), ((12284, 12305), 'paddle.fluid.dygraph.guard', 'fluid.dygraph.guard', ([], {}), '()\n', (12303, 12305), True, 'import paddle.fluid as fluid\n'), ((15149, 15180), 'paddle.fluid.load_dygraph', 'fluid.load_dygraph', (['"""./test_dy"""'], {}), "('./test_dy')\n", (15167, 15180), True, 'import paddle.fluid as fluid\n'), ((16186, 16207), 'paddle.fluid.dygraph.guard', 'fluid.dygraph.guard', ([], {}), '()\n', (16205, 16207), True, 'import paddle.fluid as fluid\n'), ((20005, 20026), 'paddle.fluid.dygraph.guard', 'fluid.dygraph.guard', ([], {}), '()\n', (20024, 20026), True, 'import paddle.fluid as fluid\n'), ((23979, 24000), 'paddle.fluid.dygraph.guard', 'fluid.dygraph.guard', ([], {}), '()\n', (23998, 24000), True, 'import paddle.fluid as fluid\n'), ((27017, 27038), 'paddle.fluid.dygraph.guard', 'fluid.dygraph.guard', ([], {}), '()\n', (27036, 27038), True, 'import paddle.fluid as fluid\n'), ((28200, 28231), 'paddle.fluid.load_dygraph', 'fluid.load_dygraph', (['"""./test_dy"""'], {}), "('./test_dy')\n", (28218, 28231), True, 'import paddle.fluid as fluid\n'), ((30418, 30439), 'paddle.fluid.dygraph.guard', 'fluid.dygraph.guard', ([], {}), '()\n', (30437, 30439), True, 'import paddle.fluid as fluid\n'), ((33895, 33916), 'paddle.fluid.dygraph.guard', 'fluid.dygraph.guard', ([], {}), '()\n', (33914, 33916), True, 'import paddle.fluid as fluid\n'), ((33936, 33969), 'paddle.fluid.dygraph.Embedding', 'fluid.dygraph.Embedding', (['[10, 10]'], {}), '([10, 10])\n', (33959, 33969), True, 'import paddle.fluid as fluid\n'), ((4011, 4060), 'paddle.fluid.layers.concat', 'fluid.layers.concat', (['[self._input, pre_hidden]', '(1)'], {}), '([self._input, pre_hidden], 1)\n', (4030, 4060), True, 'import paddle.fluid as fluid\n'), ((4090, 4127), 'paddle.fluid.layers.matmul', 'fluid.layers.matmul', ([], {'x': 'nn', 'y': 'weight_1'}), '(x=nn, y=weight_1)\n', (4109, 4127), True, 'import paddle.fluid as fluid\n'), ((4158, 4204), 'paddle.fluid.layers.elementwise_add', 'fluid.layers.elementwise_add', (['gate_input', 'bias'], {}), '(gate_input, bias)\n', (4186, 4204), True, 'import paddle.fluid as fluid\n'), ((4234, 4291), 'paddle.fluid.layers.split', 'fluid.layers.split', (['gate_input'], {'num_or_sections': '(4)', 'dim': '(-1)'}), '(gate_input, num_or_sections=4, dim=-1)\n', (4252, 4291), True, 'import paddle.fluid as fluid\n'), ((4940, 5007), 'paddle.fluid.layers.reshape', 'fluid.layers.reshape', (['self._input'], {'shape': '[1, -1, self._hidden_size]'}), '(self._input, shape=[1, -1, self._hidden_size])\n', (4960, 5007), True, 'import paddle.fluid as fluid\n'), ((6799, 6816), 'paddle.fluid.ParamAttr', 'fluid.ParamAttr', ([], {}), '()\n', (6814, 6816), True, 'import paddle.fluid as fluid\n'), ((6934, 7019), 'paddle.fluid.initializer.UniformInitializer', 'fluid.initializer.UniformInitializer', ([], {'low': '(-self.init_scale)', 'high': 'self.init_scale'}), '(low=-self.init_scale, high=self.init_scale\n )\n', (6970, 7019), True, 'import paddle.fluid as fluid\n'), ((7101, 7118), 'paddle.fluid.ParamAttr', 'fluid.ParamAttr', ([], {}), '()\n', (7116, 7118), True, 'import paddle.fluid as fluid\n'), ((7218, 7303), 'paddle.fluid.initializer.UniformInitializer', 'fluid.initializer.UniformInitializer', ([], {'low': '(-self.init_scale)', 'high': 'self.init_scale'}), '(low=-self.init_scale, high=self.init_scale\n )\n', (7254, 7303), True, 'import paddle.fluid as fluid\n'), ((9110, 9141), 'paddle.fluid.default_startup_program', 'fluid.default_startup_program', ([], {}), '()\n', (9139, 9141), True, 'import paddle.fluid as fluid\n'), ((9173, 9201), 'paddle.fluid.default_main_program', 'fluid.default_main_program', ([], {}), '()\n', (9199, 9201), True, 'import paddle.fluid as fluid\n'), ((9748, 9764), 'paddle.fluid.CPUPlace', 'fluid.CPUPlace', ([], {}), '()\n', (9762, 9764), True, 'import paddle.fluid as fluid\n'), ((9819, 9837), 'paddle.fluid.CUDAPlace', 'fluid.CUDAPlace', (['(0)'], {}), '(0)\n', (9834, 9837), True, 'import paddle.fluid as fluid\n'), ((10454, 10518), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (10462, 10518), True, 'import numpy as np\n'), ((10573, 10637), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (10581, 10637), True, 'import numpy as np\n'), ((10679, 10698), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['x_data'], {}), '(x_data)\n', (10690, 10698), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((10719, 10738), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['y_data'], {}), '(y_data)\n', (10730, 10738), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((10769, 10798), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_hidden_data'], {}), '(init_hidden_data)\n', (10780, 10798), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((10827, 10854), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_cell_data'], {}), '(init_cell_data)\n', (10838, 10854), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((12319, 12350), 'paddle.fluid.default_startup_program', 'fluid.default_startup_program', ([], {}), '()\n', (12348, 12350), True, 'import paddle.fluid as fluid\n'), ((12382, 12410), 'paddle.fluid.default_main_program', 'fluid.default_main_program', ([], {}), '()\n', (12408, 12410), True, 'import paddle.fluid as fluid\n'), ((12957, 12973), 'paddle.fluid.CPUPlace', 'fluid.CPUPlace', ([], {}), '()\n', (12971, 12973), True, 'import paddle.fluid as fluid\n'), ((13028, 13046), 'paddle.fluid.CUDAPlace', 'fluid.CUDAPlace', (['(0)'], {}), '(0)\n', (13043, 13046), True, 'import paddle.fluid as fluid\n'), ((13663, 13727), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (13671, 13727), True, 'import numpy as np\n'), ((13782, 13846), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (13790, 13846), True, 'import numpy as np\n'), ((13888, 13907), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['x_data'], {}), '(x_data)\n', (13899, 13907), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((13928, 13947), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['y_data'], {}), '(y_data)\n', (13939, 13947), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((13978, 14007), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_hidden_data'], {}), '(init_hidden_data)\n', (13989, 14007), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((14036, 14063), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_cell_data'], {}), '(init_cell_data)\n', (14047, 14063), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((16221, 16252), 'paddle.fluid.default_startup_program', 'fluid.default_startup_program', ([], {}), '()\n', (16250, 16252), True, 'import paddle.fluid as fluid\n'), ((16284, 16312), 'paddle.fluid.default_main_program', 'fluid.default_main_program', ([], {}), '()\n', (16310, 16312), True, 'import paddle.fluid as fluid\n'), ((16859, 16875), 'paddle.fluid.CPUPlace', 'fluid.CPUPlace', ([], {}), '()\n', (16873, 16875), True, 'import paddle.fluid as fluid\n'), ((16930, 16948), 'paddle.fluid.CUDAPlace', 'fluid.CUDAPlace', (['(0)'], {}), '(0)\n', (16945, 16948), True, 'import paddle.fluid as fluid\n'), ((17565, 17629), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (17573, 17629), True, 'import numpy as np\n'), ((17684, 17748), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (17692, 17748), True, 'import numpy as np\n'), ((17790, 17809), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['x_data'], {}), '(x_data)\n', (17801, 17809), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((17830, 17849), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['y_data'], {}), '(y_data)\n', (17841, 17849), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((17880, 17909), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_hidden_data'], {}), '(init_hidden_data)\n', (17891, 17909), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((17938, 17965), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_cell_data'], {}), '(init_cell_data)\n', (17949, 17965), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((20040, 20071), 'paddle.fluid.default_startup_program', 'fluid.default_startup_program', ([], {}), '()\n', (20069, 20071), True, 'import paddle.fluid as fluid\n'), ((20103, 20131), 'paddle.fluid.default_main_program', 'fluid.default_main_program', ([], {}), '()\n', (20129, 20131), True, 'import paddle.fluid as fluid\n'), ((20678, 20694), 'paddle.fluid.CPUPlace', 'fluid.CPUPlace', ([], {}), '()\n', (20692, 20694), True, 'import paddle.fluid as fluid\n'), ((20749, 20767), 'paddle.fluid.CUDAPlace', 'fluid.CUDAPlace', (['(0)'], {}), '(0)\n', (20764, 20767), True, 'import paddle.fluid as fluid\n'), ((21384, 21448), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (21392, 21448), True, 'import numpy as np\n'), ((21503, 21567), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (21511, 21567), True, 'import numpy as np\n'), ((21609, 21628), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['x_data'], {}), '(x_data)\n', (21620, 21628), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((21649, 21668), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['y_data'], {}), '(y_data)\n', (21660, 21668), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((21699, 21728), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_hidden_data'], {}), '(init_hidden_data)\n', (21710, 21728), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((21757, 21784), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_cell_data'], {}), '(init_cell_data)\n', (21768, 21784), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((24014, 24045), 'paddle.fluid.default_startup_program', 'fluid.default_startup_program', ([], {}), '()\n', (24043, 24045), True, 'import paddle.fluid as fluid\n'), ((24077, 24105), 'paddle.fluid.default_main_program', 'fluid.default_main_program', ([], {}), '()\n', (24103, 24105), True, 'import paddle.fluid as fluid\n'), ((24423, 24439), 'paddle.fluid.CPUPlace', 'fluid.CPUPlace', ([], {}), '()\n', (24437, 24439), True, 'import paddle.fluid as fluid\n'), ((24494, 24512), 'paddle.fluid.CUDAPlace', 'fluid.CUDAPlace', (['(0)'], {}), '(0)\n', (24509, 24512), True, 'import paddle.fluid as fluid\n'), ((25190, 25254), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (25198, 25254), True, 'import numpy as np\n'), ((25309, 25373), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (25317, 25373), True, 'import numpy as np\n'), ((25415, 25434), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['x_data'], {}), '(x_data)\n', (25426, 25434), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((25455, 25474), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['y_data'], {}), '(y_data)\n', (25466, 25474), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((25505, 25534), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_hidden_data'], {}), '(init_hidden_data)\n', (25516, 25534), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((25563, 25590), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_cell_data'], {}), '(init_cell_data)\n', (25574, 25590), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((27052, 27083), 'paddle.fluid.default_startup_program', 'fluid.default_startup_program', ([], {}), '()\n', (27081, 27083), True, 'import paddle.fluid as fluid\n'), ((27115, 27143), 'paddle.fluid.default_main_program', 'fluid.default_main_program', ([], {}), '()\n', (27141, 27143), True, 'import paddle.fluid as fluid\n'), ((27744, 27760), 'paddle.fluid.CPUPlace', 'fluid.CPUPlace', ([], {}), '()\n', (27758, 27760), True, 'import paddle.fluid as fluid\n'), ((27815, 27833), 'paddle.fluid.CUDAPlace', 'fluid.CUDAPlace', (['(0)'], {}), '(0)\n', (27830, 27833), True, 'import paddle.fluid as fluid\n'), ((28569, 28633), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (28577, 28633), True, 'import numpy as np\n'), ((28688, 28752), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (28696, 28752), True, 'import numpy as np\n'), ((28794, 28813), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['x_data'], {}), '(x_data)\n', (28805, 28813), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((28834, 28853), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['y_data'], {}), '(y_data)\n', (28845, 28853), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((28884, 28913), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_hidden_data'], {}), '(init_hidden_data)\n', (28895, 28913), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((28942, 28969), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_cell_data'], {}), '(init_cell_data)\n', (28953, 28969), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((30453, 30484), 'paddle.fluid.default_startup_program', 'fluid.default_startup_program', ([], {}), '()\n', (30482, 30484), True, 'import paddle.fluid as fluid\n'), ((30516, 30544), 'paddle.fluid.default_main_program', 'fluid.default_main_program', ([], {}), '()\n', (30542, 30544), True, 'import paddle.fluid as fluid\n'), ((31145, 31161), 'paddle.fluid.CPUPlace', 'fluid.CPUPlace', ([], {}), '()\n', (31159, 31161), True, 'import paddle.fluid as fluid\n'), ((31216, 31234), 'paddle.fluid.CUDAPlace', 'fluid.CUDAPlace', (['(0)'], {}), '(0)\n', (31231, 31234), True, 'import paddle.fluid as fluid\n'), ((32238, 32302), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (32246, 32302), True, 'import numpy as np\n'), ((32357, 32421), 'numpy.zeros', 'np.zeros', (['(num_layers, batch_size, hidden_size)'], {'dtype': '"""float32"""'}), "((num_layers, batch_size, hidden_size), dtype='float32')\n", (32365, 32421), True, 'import numpy as np\n'), ((32463, 32482), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['x_data'], {}), '(x_data)\n', (32474, 32482), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((32503, 32522), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['y_data'], {}), '(y_data)\n', (32514, 32522), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((32553, 32582), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_hidden_data'], {}), '(init_hidden_data)\n', (32564, 32582), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((32611, 32638), 'paddle.fluid.dygraph.base.to_variable', 'to_variable', (['init_cell_data'], {}), '(init_cell_data)\n', (32622, 32638), False, 'from paddle.fluid.dygraph.base import to_variable\n'), ((34055, 34089), 'os.path.join', 'os.path.join', (['"""saved_dy"""', '"""emb_dy"""'], {}), "('saved_dy', 'emb_dy')\n", (34067, 34089), False, 'import os\n'), ((34175, 34209), 'os.path.join', 'os.path.join', (['"""saved_dy"""', '"""emb_dy"""'], {}), "('saved_dy', 'emb_dy')\n", (34187, 34209), False, 'import os\n'), ((2139, 2226), 'paddle.fluid.initializer.UniformInitializer', 'fluid.initializer.UniformInitializer', ([], {'low': '(-self._init_scale)', 'high': 'self._init_scale'}), '(low=-self._init_scale, high=self.\n _init_scale)\n', (2175, 2226), True, 'import paddle.fluid as fluid\n'), ((2663, 2694), 'paddle.fluid.initializer.Constant', 'fluid.initializer.Constant', (['(0.0)'], {}), '(0.0)\n', (2689, 2694), True, 'import paddle.fluid as fluid\n'), ((4458, 4478), 'paddle.fluid.layers.tanh', 'fluid.layers.tanh', (['c'], {}), '(c)\n', (4475, 4478), True, 'import paddle.fluid as fluid\n'), ((4481, 4504), 'paddle.fluid.layers.sigmoid', 'fluid.layers.sigmoid', (['o'], {}), '(o)\n', (4501, 4504), True, 'import paddle.fluid as fluid\n'), ((4722, 4830), 'paddle.fluid.layers.dropout', 'fluid.layers.dropout', (['self._input'], {'dropout_prob': 'self._dropout', 'dropout_implementation': '"""upscale_in_train"""'}), "(self._input, dropout_prob=self._dropout,\n dropout_implementation='upscale_in_train')\n", (4742, 4830), True, 'import paddle.fluid as fluid\n'), ((9772, 9800), 'paddle.fluid.core.is_compiled_with_cuda', 'core.is_compiled_with_cuda', ([], {}), '()\n', (9798, 9800), True, 'import paddle.fluid.core as core\n'), ((9893, 9951), 'paddle.fluid.layers.piecewise_decay', 'fluid.layers.piecewise_decay', ([], {'boundaries': 'bd', 'values': 'lr_arr'}), '(boundaries=bd, values=lr_arr)\n', (9921, 9951), True, 'import paddle.fluid as fluid\n'), ((12981, 13009), 'paddle.fluid.core.is_compiled_with_cuda', 'core.is_compiled_with_cuda', ([], {}), '()\n', (13007, 13009), True, 'import paddle.fluid.core as core\n'), ((13102, 13160), 'paddle.fluid.layers.piecewise_decay', 'fluid.layers.piecewise_decay', ([], {'boundaries': 'bd', 'values': 'lr_arr'}), '(boundaries=bd, values=lr_arr)\n', (13130, 13160), True, 'import paddle.fluid as fluid\n'), ((14891, 14910), 'numpy.zeros_like', 'np.zeros_like', (['np_t'], {}), '(np_t)\n', (14904, 14910), True, 'import numpy as np\n'), ((15640, 15659), 'numpy.zeros_like', 'np.zeros_like', (['np_t'], {}), '(np_t)\n', (15653, 15659), True, 'import numpy as np\n'), ((15923, 15952), 'numpy.array_equal', 'np.array_equal', (['new_t', 'base_t'], {}), '(new_t, base_t)\n', (15937, 15952), True, 'import numpy as np\n'), ((16883, 16911), 'paddle.fluid.core.is_compiled_with_cuda', 'core.is_compiled_with_cuda', ([], {}), '()\n', (16909, 16911), True, 'import paddle.fluid.core as core\n'), ((17004, 17062), 'paddle.fluid.layers.piecewise_decay', 'fluid.layers.piecewise_decay', ([], {'boundaries': 'bd', 'values': 'lr_arr'}), '(boundaries=bd, values=lr_arr)\n', (17032, 17062), True, 'import paddle.fluid as fluid\n'), ((18793, 18812), 'numpy.zeros_like', 'np.zeros_like', (['np_t'], {}), '(np_t)\n', (18806, 18812), True, 'import numpy as np\n'), ((19462, 19481), 'numpy.zeros_like', 'np.zeros_like', (['np_t'], {}), '(np_t)\n', (19475, 19481), True, 'import numpy as np\n'), ((19745, 19774), 'numpy.array_equal', 'np.array_equal', (['new_t', 'base_t'], {}), '(new_t, base_t)\n', (19759, 19774), True, 'import numpy as np\n'), ((20702, 20730), 'paddle.fluid.core.is_compiled_with_cuda', 'core.is_compiled_with_cuda', ([], {}), '()\n', (20728, 20730), True, 'import paddle.fluid.core as core\n'), ((20823, 20881), 'paddle.fluid.layers.piecewise_decay', 'fluid.layers.piecewise_decay', ([], {'boundaries': 'bd', 'values': 'lr_arr'}), '(boundaries=bd, values=lr_arr)\n', (20851, 20881), True, 'import paddle.fluid as fluid\n'), ((22686, 22705), 'numpy.zeros_like', 'np.zeros_like', (['np_t'], {}), '(np_t)\n', (22699, 22705), True, 'import numpy as np\n'), ((23424, 23443), 'numpy.zeros_like', 'np.zeros_like', (['np_t'], {}), '(np_t)\n', (23437, 23443), True, 'import numpy as np\n'), ((23705, 23734), 'numpy.array_equal', 'np.array_equal', (['new_t', 'base_t'], {}), '(new_t, base_t)\n', (23719, 23734), True, 'import numpy as np\n'), ((24447, 24475), 'paddle.fluid.core.is_compiled_with_cuda', 'core.is_compiled_with_cuda', ([], {}), '()\n', (24473, 24475), True, 'import paddle.fluid.core as core\n'), ((26737, 26766), 'numpy.array_equal', 'np.array_equal', (['new_t', 'base_t'], {}), '(new_t, base_t)\n', (26751, 26766), True, 'import numpy as np\n'), ((27768, 27796), 'paddle.fluid.core.is_compiled_with_cuda', 'core.is_compiled_with_cuda', ([], {}), '()\n', (27794, 27796), True, 'import paddle.fluid.core as core\n'), ((30147, 30176), 'numpy.array_equal', 'np.array_equal', (['new_t', 'base_t'], {}), '(new_t, base_t)\n', (30161, 30176), True, 'import numpy as np\n'), ((31169, 31197), 'paddle.fluid.core.is_compiled_with_cuda', 'core.is_compiled_with_cuda', ([], {}), '()\n', (31195, 31197), True, 'import paddle.fluid.core as core\n'), ((31290, 31348), 'paddle.fluid.layers.piecewise_decay', 'fluid.layers.piecewise_decay', ([], {'boundaries': 'bd', 'values': 'lr_arr'}), '(boundaries=bd, values=lr_arr)\n', (31318, 31348), True, 'import paddle.fluid as fluid\n'), ((33816, 33845), 'numpy.array_equal', 'np.array_equal', (['new_t', 'base_t'], {}), '(new_t, base_t)\n', (33830, 33845), True, 'import numpy as np\n'), ((4344, 4367), 'paddle.fluid.layers.sigmoid', 'fluid.layers.sigmoid', (['f'], {}), '(f)\n', (4364, 4367), True, 'import paddle.fluid as fluid\n'), ((4370, 4393), 'paddle.fluid.layers.sigmoid', 'fluid.layers.sigmoid', (['i'], {}), '(i)\n', (4390, 4393), True, 'import paddle.fluid as fluid\n'), ((4417, 4437), 'paddle.fluid.layers.tanh', 'fluid.layers.tanh', (['j'], {}), '(j)\n', (4434, 4437), True, 'import paddle.fluid as fluid\n'), ((6634, 6704), 'paddle.fluid.initializer.UniformInitializer', 'fluid.initializer.UniformInitializer', ([], {'low': '(-init_scale)', 'high': 'init_scale'}), '(low=-init_scale, high=init_scale)\n', (6670, 6704), True, 'import paddle.fluid as fluid\n'), ((1890, 1977), 'paddle.fluid.initializer.UniformInitializer', 'fluid.initializer.UniformInitializer', ([], {'low': '(-self._init_scale)', 'high': 'self._init_scale'}), '(low=-self._init_scale, high=self.\n _init_scale)\n', (1926, 1977), True, 'import paddle.fluid as fluid\n'), ((2437, 2524), 'paddle.fluid.initializer.UniformInitializer', 'fluid.initializer.UniformInitializer', ([], {'low': '(-self._init_scale)', 'high': 'self._init_scale'}), '(low=-self._init_scale, high=self.\n _init_scale)\n', (2473, 2524), True, 'import paddle.fluid as fluid\n'), ((10254, 10267), 'numpy.arange', 'np.arange', (['(12)'], {}), '(12)\n', (10263, 10267), True, 'import numpy as np\n'), ((10323, 10339), 'numpy.arange', 'np.arange', (['(1)', '(13)'], {}), '(1, 13)\n', (10332, 10339), True, 'import numpy as np\n'), ((13463, 13476), 'numpy.arange', 'np.arange', (['(12)'], {}), '(12)\n', (13472, 13476), True, 'import numpy as np\n'), ((13532, 13548), 'numpy.arange', 'np.arange', (['(1)', '(13)'], {}), '(1, 13)\n', (13541, 13548), True, 'import numpy as np\n'), ((17365, 17378), 'numpy.arange', 'np.arange', (['(12)'], {}), '(12)\n', (17374, 17378), True, 'import numpy as np\n'), ((17434, 17450), 'numpy.arange', 'np.arange', (['(1)', '(13)'], {}), '(1, 13)\n', (17443, 17450), True, 'import numpy as np\n'), ((21184, 21197), 'numpy.arange', 'np.arange', (['(12)'], {}), '(12)\n', (21193, 21197), True, 'import numpy as np\n'), ((21253, 21269), 'numpy.arange', 'np.arange', (['(1)', '(13)'], {}), '(1, 13)\n', (21262, 21269), True, 'import numpy as np\n'), ((24990, 25003), 'numpy.arange', 'np.arange', (['(12)'], {}), '(12)\n', (24999, 25003), True, 'import numpy as np\n'), ((25059, 25075), 'numpy.arange', 'np.arange', (['(1)', '(13)'], {}), '(1, 13)\n', (25068, 25075), True, 'import numpy as np\n'), ((28369, 28382), 'numpy.arange', 'np.arange', (['(12)'], {}), '(12)\n', (28378, 28382), True, 'import numpy as np\n'), ((28438, 28454), 'numpy.arange', 'np.arange', (['(1)', '(13)'], {}), '(1, 13)\n', (28447, 28454), True, 'import numpy as np\n'), ((32038, 32051), 'numpy.arange', 'np.arange', (['(12)'], {}), '(12)\n', (32047, 32051), True, 'import numpy as np\n'), ((32107, 32123), 'numpy.arange', 'np.arange', (['(1)', '(13)'], {}), '(1, 13)\n', (32116, 32123), True, 'import numpy as np\n')]
|
# Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import image_util
from paddle.utils.image_util import *
import random
from PIL import Image
import numpy as np
import xml.etree.ElementTree
import os
class Settings(object):
def __init__(self, data_dir, label_file, resize_h, resize_w, mean_value,
apply_distort, apply_expand):
self._data_dir = data_dir
self._label_list = []
label_fpath = os.path.join(data_dir, label_file)
for line in open(label_fpath):
self._label_list.append(line.strip())
self._apply_distort = apply_distort
self._apply_expand = apply_expand
self._resize_height = resize_h
self._resize_width = resize_w
self._img_mean = np.array(mean_value)[:, np.newaxis, np.newaxis].astype(
'float32')
self._expand_prob = 0.5
self._expand_max_ratio = 4
self._hue_prob = 0.5
self._hue_delta = 18
self._contrast_prob = 0.5
self._contrast_delta = 0.5
self._saturation_prob = 0.5
self._saturation_delta = 0.5
self._brightness_prob = 0.5
self._brightness_delta = 0.125
@property
def apply_distort(self):
return self._apply_expand
@property
def apply_distort(self):
return self._apply_distort
@property
def data_dir(self):
return self._data_dir
@property
def label_list(self):
return self._label_list
@property
def resize_h(self):
return self._resize_height
@property
def resize_w(self):
return self._resize_width
@property
def img_mean(self):
return self._img_mean
def _reader_creator(settings, file_list, mode, shuffle):
def reader():
with open(file_list) as flist:
lines = [line.strip() for line in flist]
if shuffle:
random.shuffle(lines)
for line in lines:
if mode == 'train' or mode == 'test':
img_path, label_path = line.split()
img_path = os.path.join(settings.data_dir, img_path)
label_path = os.path.join(settings.data_dir, label_path)
elif mode == 'infer':
img_path = os.path.join(settings.data_dir, line)
img = Image.open(img_path)
img_width, img_height = img.size
# layout: label | xmin | ymin | xmax | ymax | difficult
if mode == 'train' or mode == 'test':
bbox_labels = []
root = xml.etree.ElementTree.parse(label_path).getroot()
for object in root.findall('object'):
bbox_sample = []
# start from 1
bbox_sample.append(
float(
settings.label_list.index(
object.find('name').text)))
bbox = object.find('bndbox')
difficult = float(object.find('difficult').text)
bbox_sample.append(
float(bbox.find('xmin').text) / img_width)
bbox_sample.append(
float(bbox.find('ymin').text) / img_height)
bbox_sample.append(
float(bbox.find('xmax').text) / img_width)
bbox_sample.append(
float(bbox.find('ymax').text) / img_height)
bbox_sample.append(difficult)
bbox_labels.append(bbox_sample)
sample_labels = bbox_labels
if mode == 'train':
if settings._apply_distort:
img = image_util.distort_image(img, settings)
if settings._apply_expand:
img, bbox_labels = image_util.expand_image(
img, bbox_labels, img_width, img_height,
settings)
batch_sampler = []
# hard-code here
batch_sampler.append(
image_util.sampler(1, 1, 1.0, 1.0, 1.0, 1.0, 0.0,
0.0))
batch_sampler.append(
image_util.sampler(1, 50, 0.3, 1.0, 0.5, 2.0, 0.1,
0.0))
batch_sampler.append(
image_util.sampler(1, 50, 0.3, 1.0, 0.5, 2.0, 0.3,
0.0))
batch_sampler.append(
image_util.sampler(1, 50, 0.3, 1.0, 0.5, 2.0, 0.5,
0.0))
batch_sampler.append(
image_util.sampler(1, 50, 0.3, 1.0, 0.5, 2.0, 0.7,
0.0))
batch_sampler.append(
image_util.sampler(1, 50, 0.3, 1.0, 0.5, 2.0, 0.9,
0.0))
batch_sampler.append(
image_util.sampler(1, 50, 0.3, 1.0, 0.5, 2.0, 0.0,
1.0))
""" random crop """
sampled_bbox = image_util.generate_batch_samples(
batch_sampler, bbox_labels, img_width, img_height)
img = np.array(img)
if len(sampled_bbox) > 0:
idx = int(random.uniform(0, len(sampled_bbox)))
img, sample_labels = image_util.crop_image(
img, bbox_labels, sampled_bbox[idx], img_width,
img_height)
img = Image.fromarray(img)
img = img.resize((settings.resize_w, settings.resize_h),
Image.ANTIALIAS)
img = np.array(img)
if mode == 'train':
mirror = int(random.uniform(0, 2))
if mirror == 1:
img = img[:, ::-1, :]
for i in xrange(len(sample_labels)):
tmp = sample_labels[i][1]
sample_labels[i][1] = 1 - sample_labels[i][3]
sample_labels[i][3] = 1 - tmp
if len(img.shape) == 3:
img = np.swapaxes(img, 1, 2)
img = np.swapaxes(img, 1, 0)
img = img[[2, 1, 0], :, :]
img = img.astype('float32')
img -= settings.img_mean
img = img.flatten()
img = img * 0.007843
sample_labels = np.array(sample_labels)
if mode == 'train' or mode == 'test':
if mode == 'train' and len(sample_labels) == 0: continue
yield img.astype(
'float32'
), sample_labels[:, 1:5], sample_labels[:, 0].astype(
'int32'), sample_labels[:, -1].astype('int32')
elif mode == 'infer':
yield img.astype('float32')
return reader
def train(settings, file_list, shuffle=True):
return _reader_creator(settings, file_list, 'train', shuffle)
def test(settings, file_list):
return _reader_creator(settings, file_list, 'test', False)
def infer(settings, file_list):
return _reader_creator(settings, file_list, 'infer', False)
|
[
"image_util.expand_image",
"PIL.Image.fromarray",
"PIL.Image.open",
"random.uniform",
"random.shuffle",
"image_util.sampler",
"image_util.generate_batch_samples",
"image_util.distort_image",
"os.path.join",
"numpy.swapaxes",
"numpy.array",
"image_util.crop_image"
] |
[((996, 1030), 'os.path.join', 'os.path.join', (['data_dir', 'label_file'], {}), '(data_dir, label_file)\n', (1008, 1030), False, 'import os\n'), ((2454, 2475), 'random.shuffle', 'random.shuffle', (['lines'], {}), '(lines)\n', (2468, 2475), False, 'import random\n'), ((2897, 2917), 'PIL.Image.open', 'Image.open', (['img_path'], {}), '(img_path)\n', (2907, 2917), False, 'from PIL import Image\n'), ((6789, 6802), 'numpy.array', 'np.array', (['img'], {}), '(img)\n', (6797, 6802), True, 'import numpy as np\n'), ((7598, 7621), 'numpy.array', 'np.array', (['sample_labels'], {}), '(sample_labels)\n', (7606, 7621), True, 'import numpy as np\n'), ((1309, 1329), 'numpy.array', 'np.array', (['mean_value'], {}), '(mean_value)\n', (1317, 1329), True, 'import numpy as np\n'), ((2648, 2689), 'os.path.join', 'os.path.join', (['settings.data_dir', 'img_path'], {}), '(settings.data_dir, img_path)\n', (2660, 2689), False, 'import os\n'), ((2723, 2766), 'os.path.join', 'os.path.join', (['settings.data_dir', 'label_path'], {}), '(settings.data_dir, label_path)\n', (2735, 2766), False, 'import os\n'), ((7291, 7313), 'numpy.swapaxes', 'np.swapaxes', (['img', '(1)', '(2)'], {}), '(img, 1, 2)\n', (7302, 7313), True, 'import numpy as np\n'), ((7340, 7362), 'numpy.swapaxes', 'np.swapaxes', (['img', '(1)', '(0)'], {}), '(img, 1, 0)\n', (7351, 7362), True, 'import numpy as np\n'), ((2836, 2873), 'os.path.join', 'os.path.join', (['settings.data_dir', 'line'], {}), '(settings.data_dir, line)\n', (2848, 2873), False, 'import os\n'), ((6111, 6199), 'image_util.generate_batch_samples', 'image_util.generate_batch_samples', (['batch_sampler', 'bbox_labels', 'img_width', 'img_height'], {}), '(batch_sampler, bbox_labels, img_width,\n img_height)\n', (6144, 6199), False, 'import image_util\n'), ((6256, 6269), 'numpy.array', 'np.array', (['img'], {}), '(img)\n', (6264, 6269), True, 'import numpy as np\n'), ((6623, 6643), 'PIL.Image.fromarray', 'Image.fromarray', (['img'], {}), '(img)\n', (6638, 6643), False, 'from PIL import Image\n'), ((6873, 6893), 'random.uniform', 'random.uniform', (['(0)', '(2)'], {}), '(0, 2)\n', (6887, 6893), False, 'import random\n'), ((4421, 4460), 'image_util.distort_image', 'image_util.distort_image', (['img', 'settings'], {}), '(img, settings)\n', (4445, 4460), False, 'import image_util\n'), ((4559, 4633), 'image_util.expand_image', 'image_util.expand_image', (['img', 'bbox_labels', 'img_width', 'img_height', 'settings'], {}), '(img, bbox_labels, img_width, img_height, settings)\n', (4582, 4633), False, 'import image_util\n'), ((4857, 4911), 'image_util.sampler', 'image_util.sampler', (['(1)', '(1)', '(1.0)', '(1.0)', '(1.0)', '(1.0)', '(0.0)', '(0.0)'], {}), '(1, 1, 1.0, 1.0, 1.0, 1.0, 0.0, 0.0)\n', (4875, 4911), False, 'import image_util\n'), ((5034, 5089), 'image_util.sampler', 'image_util.sampler', (['(1)', '(50)', '(0.3)', '(1.0)', '(0.5)', '(2.0)', '(0.1)', '(0.0)'], {}), '(1, 50, 0.3, 1.0, 0.5, 2.0, 0.1, 0.0)\n', (5052, 5089), False, 'import image_util\n'), ((5212, 5267), 'image_util.sampler', 'image_util.sampler', (['(1)', '(50)', '(0.3)', '(1.0)', '(0.5)', '(2.0)', '(0.3)', '(0.0)'], {}), '(1, 50, 0.3, 1.0, 0.5, 2.0, 0.3, 0.0)\n', (5230, 5267), False, 'import image_util\n'), ((5390, 5445), 'image_util.sampler', 'image_util.sampler', (['(1)', '(50)', '(0.3)', '(1.0)', '(0.5)', '(2.0)', '(0.5)', '(0.0)'], {}), '(1, 50, 0.3, 1.0, 0.5, 2.0, 0.5, 0.0)\n', (5408, 5445), False, 'import image_util\n'), ((5568, 5623), 'image_util.sampler', 'image_util.sampler', (['(1)', '(50)', '(0.3)', '(1.0)', '(0.5)', '(2.0)', '(0.7)', '(0.0)'], {}), '(1, 50, 0.3, 1.0, 0.5, 2.0, 0.7, 0.0)\n', (5586, 5623), False, 'import image_util\n'), ((5746, 5801), 'image_util.sampler', 'image_util.sampler', (['(1)', '(50)', '(0.3)', '(1.0)', '(0.5)', '(2.0)', '(0.9)', '(0.0)'], {}), '(1, 50, 0.3, 1.0, 0.5, 2.0, 0.9, 0.0)\n', (5764, 5801), False, 'import image_util\n'), ((5924, 5979), 'image_util.sampler', 'image_util.sampler', (['(1)', '(50)', '(0.3)', '(1.0)', '(0.5)', '(2.0)', '(0.0)', '(1.0)'], {}), '(1, 50, 0.3, 1.0, 0.5, 2.0, 0.0, 1.0)\n', (5942, 5979), False, 'import image_util\n'), ((6445, 6530), 'image_util.crop_image', 'image_util.crop_image', (['img', 'bbox_labels', 'sampled_bbox[idx]', 'img_width', 'img_height'], {}), '(img, bbox_labels, sampled_bbox[idx], img_width,\n img_height)\n', (6466, 6530), False, 'import image_util\n')]
|
import torch
import unittest
import numpy as np
from torch.autograd import Variable
from losses.svm import SmoothTop1SVM, SmoothTopkSVM, MaxTop1SVM, MaxTopkSVM
from losses.functional import Topk_Smooth_SVM
from tests.utils import assert_all_close, V
from tests.py_ref import svm_topk_smooth_py_1, svm_topk_smooth_py_2,\
smooth_svm_py, max_svm_py, svm_topk_max_py
from torch.autograd.gradcheck import gradcheck
class TestMaxSVM(unittest.TestCase):
def setUp(self):
torch.manual_seed(1234)
np.random.seed(1234)
self.n_samples = 20
self.n_classes = 7
self.alpha = 1.
self.x = torch.randn(self.n_samples, self.n_classes)
self.y = torch.from_numpy(np.random.randint(0, self.n_classes,
size=self.n_samples))
self.k = 3
def testMaxSVM(self):
max_svm_th = MaxTop1SVM(self.n_classes, alpha=self.alpha)
res_th = max_svm_th(V(self.x), V(self.y))
res_py = max_svm_py(V(self.x), V(self.y), alpha=self.alpha)
assert_all_close(res_th, res_py)
def testMaxSVMtopk(self):
max_svm_th = MaxTopkSVM(self.n_classes, k=self.k)
res_th = max_svm_th(V(self.x), V(self.y))
res_py = svm_topk_max_py(V(self.x), V(self.y), k=self.k)
assert_all_close(res_th, res_py)
class TestSmoothSVM(unittest.TestCase):
def setUp(self):
torch.manual_seed(1234)
np.random.seed(1234)
self.n_samples = 20
self.n_classes = 7
self.tau = float(2.)
self.x = torch.randn(self.n_samples, self.n_classes)
self.y = torch.from_numpy(np.random.randint(0, self.n_classes,
size=self.n_samples))
def testSmoothSVM(self):
smooth_svm_th = SmoothTop1SVM(self.n_classes, tau=self.tau)
res_th = smooth_svm_th(V(self.x), V(self.y))
res_py = smooth_svm_py(V(self.x), V(self.y), self.tau)
assert_all_close(res_th, res_py)
class TestSmoothSVMTopk(unittest.TestCase):
def setUp(self):
torch.manual_seed(1234)
np.random.seed(1234)
self.n_samples = 2
self.n_classes = 7
self.k = 5
self.tau = float(2.)
self.x = torch.randn(self.n_samples, self.n_classes)
self.y = torch.from_numpy(np.random.randint(0, self.n_classes,
size=self.n_samples))
self.labels = torch.from_numpy(np.arange(self.n_classes))
def testSmoothSVMpy(self):
res_py_1 = svm_topk_smooth_py_1(V(self.x), V(self.y), self.tau, self.k)
res_py_2 = svm_topk_smooth_py_2(V(self.x), V(self.y), self.tau, self.k)
assert_all_close(res_py_1, res_py_2)
def testSmoothSVMth_functional(self):
F = Topk_Smooth_SVM(self.labels, self.k, self.tau)
res_th = F(V(self.x), V(self.y))
res_py = svm_topk_smooth_py_1(V(self.x), V(self.y), self.tau, self.k)
assert_all_close(res_th, res_py)
def testSmoothSVMth_loss(self):
svm_topk_smooth_th = SmoothTopkSVM(self.n_classes, tau=self.tau,
k=self.k)
res_th = svm_topk_smooth_th(V(self.x), V(self.y))
res_py = svm_topk_smooth_py_1(V(self.x),
V(self.y),
self.tau, self.k).mean()
assert_all_close(res_th, res_py)
def testSmoothSVMth_loss_scales(self):
svm_topk_smooth_th = SmoothTopkSVM(self.n_classes, tau=self.tau, k=self.k)
for scale in (1e-4, 1e-3, 1e-2, 1e-1, 1e0, 1e1, 1e2, 1e3):
x = self.x * scale
res_th = svm_topk_smooth_th(V(x), V(self.y))
res_py = svm_topk_smooth_py_1(V(x), V(self.y), self.tau, self.k).mean()
assert_all_close(res_th, res_py)
def testGradSmoothSVMth_loss(self):
svm_topk_smooth_th = SmoothTopkSVM(self.n_classes, tau=self.tau, k=self.k)
for scale in (1e-4, 1e-3, 1e-2, 1e-1, 1e0, 1e1, 1e2, 1e3, 1e4):
x = self.x * scale
x = Variable(x, requires_grad=True)
assert gradcheck(lambda x: svm_topk_smooth_th(x, V(self.y)),
(x,), atol=1e-2, rtol=1e-3, eps=max(1e-4 * scale, 1e-2)), \
"failed with scale {}".format(scale)
|
[
"torch.manual_seed",
"numpy.arange",
"losses.svm.MaxTopkSVM",
"tests.utils.assert_all_close",
"losses.functional.Topk_Smooth_SVM",
"numpy.random.randint",
"tests.utils.V",
"numpy.random.seed",
"losses.svm.MaxTop1SVM",
"torch.autograd.Variable",
"torch.randn",
"losses.svm.SmoothTop1SVM",
"losses.svm.SmoothTopkSVM"
] |
[((486, 509), 'torch.manual_seed', 'torch.manual_seed', (['(1234)'], {}), '(1234)\n', (503, 509), False, 'import torch\n'), ((518, 538), 'numpy.random.seed', 'np.random.seed', (['(1234)'], {}), '(1234)\n', (532, 538), True, 'import numpy as np\n'), ((636, 679), 'torch.randn', 'torch.randn', (['self.n_samples', 'self.n_classes'], {}), '(self.n_samples, self.n_classes)\n', (647, 679), False, 'import torch\n'), ((893, 937), 'losses.svm.MaxTop1SVM', 'MaxTop1SVM', (['self.n_classes'], {'alpha': 'self.alpha'}), '(self.n_classes, alpha=self.alpha)\n', (903, 937), False, 'from losses.svm import SmoothTop1SVM, SmoothTopkSVM, MaxTop1SVM, MaxTopkSVM\n'), ((1065, 1097), 'tests.utils.assert_all_close', 'assert_all_close', (['res_th', 'res_py'], {}), '(res_th, res_py)\n', (1081, 1097), False, 'from tests.utils import assert_all_close, V\n'), ((1151, 1187), 'losses.svm.MaxTopkSVM', 'MaxTopkSVM', (['self.n_classes'], {'k': 'self.k'}), '(self.n_classes, k=self.k)\n', (1161, 1187), False, 'from losses.svm import SmoothTop1SVM, SmoothTopkSVM, MaxTop1SVM, MaxTopkSVM\n'), ((1312, 1344), 'tests.utils.assert_all_close', 'assert_all_close', (['res_th', 'res_py'], {}), '(res_th, res_py)\n', (1328, 1344), False, 'from tests.utils import assert_all_close, V\n'), ((1418, 1441), 'torch.manual_seed', 'torch.manual_seed', (['(1234)'], {}), '(1234)\n', (1435, 1441), False, 'import torch\n'), ((1450, 1470), 'numpy.random.seed', 'np.random.seed', (['(1234)'], {}), '(1234)\n', (1464, 1470), True, 'import numpy as np\n'), ((1573, 1616), 'torch.randn', 'torch.randn', (['self.n_samples', 'self.n_classes'], {}), '(self.n_samples, self.n_classes)\n', (1584, 1616), False, 'import torch\n'), ((1817, 1860), 'losses.svm.SmoothTop1SVM', 'SmoothTop1SVM', (['self.n_classes'], {'tau': 'self.tau'}), '(self.n_classes, tau=self.tau)\n', (1830, 1860), False, 'from losses.svm import SmoothTop1SVM, SmoothTopkSVM, MaxTop1SVM, MaxTopkSVM\n'), ((1986, 2018), 'tests.utils.assert_all_close', 'assert_all_close', (['res_th', 'res_py'], {}), '(res_th, res_py)\n', (2002, 2018), False, 'from tests.utils import assert_all_close, V\n'), ((2096, 2119), 'torch.manual_seed', 'torch.manual_seed', (['(1234)'], {}), '(1234)\n', (2113, 2119), False, 'import torch\n'), ((2128, 2148), 'numpy.random.seed', 'np.random.seed', (['(1234)'], {}), '(1234)\n', (2142, 2148), True, 'import numpy as np\n'), ((2269, 2312), 'torch.randn', 'torch.randn', (['self.n_samples', 'self.n_classes'], {}), '(self.n_samples, self.n_classes)\n', (2280, 2312), False, 'import torch\n'), ((2726, 2762), 'tests.utils.assert_all_close', 'assert_all_close', (['res_py_1', 'res_py_2'], {}), '(res_py_1, res_py_2)\n', (2742, 2762), False, 'from tests.utils import assert_all_close, V\n'), ((2819, 2865), 'losses.functional.Topk_Smooth_SVM', 'Topk_Smooth_SVM', (['self.labels', 'self.k', 'self.tau'], {}), '(self.labels, self.k, self.tau)\n', (2834, 2865), False, 'from losses.functional import Topk_Smooth_SVM\n'), ((2994, 3026), 'tests.utils.assert_all_close', 'assert_all_close', (['res_th', 'res_py'], {}), '(res_th, res_py)\n', (3010, 3026), False, 'from tests.utils import assert_all_close, V\n'), ((3094, 3147), 'losses.svm.SmoothTopkSVM', 'SmoothTopkSVM', (['self.n_classes'], {'tau': 'self.tau', 'k': 'self.k'}), '(self.n_classes, tau=self.tau, k=self.k)\n', (3107, 3147), False, 'from losses.svm import SmoothTop1SVM, SmoothTopkSVM, MaxTop1SVM, MaxTopkSVM\n'), ((3419, 3451), 'tests.utils.assert_all_close', 'assert_all_close', (['res_th', 'res_py'], {}), '(res_th, res_py)\n', (3435, 3451), False, 'from tests.utils import assert_all_close, V\n'), ((3526, 3579), 'losses.svm.SmoothTopkSVM', 'SmoothTopkSVM', (['self.n_classes'], {'tau': 'self.tau', 'k': 'self.k'}), '(self.n_classes, tau=self.tau, k=self.k)\n', (3539, 3579), False, 'from losses.svm import SmoothTop1SVM, SmoothTopkSVM, MaxTop1SVM, MaxTopkSVM\n'), ((3935, 3988), 'losses.svm.SmoothTopkSVM', 'SmoothTopkSVM', (['self.n_classes'], {'tau': 'self.tau', 'k': 'self.k'}), '(self.n_classes, tau=self.tau, k=self.k)\n', (3948, 3988), False, 'from losses.svm import SmoothTop1SVM, SmoothTopkSVM, MaxTop1SVM, MaxTopkSVM\n'), ((714, 771), 'numpy.random.randint', 'np.random.randint', (['(0)', 'self.n_classes'], {'size': 'self.n_samples'}), '(0, self.n_classes, size=self.n_samples)\n', (731, 771), True, 'import numpy as np\n'), ((966, 975), 'tests.utils.V', 'V', (['self.x'], {}), '(self.x)\n', (967, 975), False, 'from tests.utils import assert_all_close, V\n'), ((977, 986), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (978, 986), False, 'from tests.utils import assert_all_close, V\n'), ((1016, 1025), 'tests.utils.V', 'V', (['self.x'], {}), '(self.x)\n', (1017, 1025), False, 'from tests.utils import assert_all_close, V\n'), ((1027, 1036), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (1028, 1036), False, 'from tests.utils import assert_all_close, V\n'), ((1216, 1225), 'tests.utils.V', 'V', (['self.x'], {}), '(self.x)\n', (1217, 1225), False, 'from tests.utils import assert_all_close, V\n'), ((1227, 1236), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (1228, 1236), False, 'from tests.utils import assert_all_close, V\n'), ((1271, 1280), 'tests.utils.V', 'V', (['self.x'], {}), '(self.x)\n', (1272, 1280), False, 'from tests.utils import assert_all_close, V\n'), ((1282, 1291), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (1283, 1291), False, 'from tests.utils import assert_all_close, V\n'), ((1651, 1708), 'numpy.random.randint', 'np.random.randint', (['(0)', 'self.n_classes'], {'size': 'self.n_samples'}), '(0, self.n_classes, size=self.n_samples)\n', (1668, 1708), True, 'import numpy as np\n'), ((1892, 1901), 'tests.utils.V', 'V', (['self.x'], {}), '(self.x)\n', (1893, 1901), False, 'from tests.utils import assert_all_close, V\n'), ((1903, 1912), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (1904, 1912), False, 'from tests.utils import assert_all_close, V\n'), ((1945, 1954), 'tests.utils.V', 'V', (['self.x'], {}), '(self.x)\n', (1946, 1954), False, 'from tests.utils import assert_all_close, V\n'), ((1956, 1965), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (1957, 1965), False, 'from tests.utils import assert_all_close, V\n'), ((2347, 2404), 'numpy.random.randint', 'np.random.randint', (['(0)', 'self.n_classes'], {'size': 'self.n_samples'}), '(0, self.n_classes, size=self.n_samples)\n', (2364, 2404), True, 'import numpy as np\n'), ((2497, 2522), 'numpy.arange', 'np.arange', (['self.n_classes'], {}), '(self.n_classes)\n', (2506, 2522), True, 'import numpy as np\n'), ((2597, 2606), 'tests.utils.V', 'V', (['self.x'], {}), '(self.x)\n', (2598, 2606), False, 'from tests.utils import assert_all_close, V\n'), ((2608, 2617), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (2609, 2617), False, 'from tests.utils import assert_all_close, V\n'), ((2677, 2686), 'tests.utils.V', 'V', (['self.x'], {}), '(self.x)\n', (2678, 2686), False, 'from tests.utils import assert_all_close, V\n'), ((2688, 2697), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (2689, 2697), False, 'from tests.utils import assert_all_close, V\n'), ((2885, 2894), 'tests.utils.V', 'V', (['self.x'], {}), '(self.x)\n', (2886, 2894), False, 'from tests.utils import assert_all_close, V\n'), ((2896, 2905), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (2897, 2905), False, 'from tests.utils import assert_all_close, V\n'), ((2945, 2954), 'tests.utils.V', 'V', (['self.x'], {}), '(self.x)\n', (2946, 2954), False, 'from tests.utils import assert_all_close, V\n'), ((2956, 2965), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (2957, 2965), False, 'from tests.utils import assert_all_close, V\n'), ((3227, 3236), 'tests.utils.V', 'V', (['self.x'], {}), '(self.x)\n', (3228, 3236), False, 'from tests.utils import assert_all_close, V\n'), ((3238, 3247), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (3239, 3247), False, 'from tests.utils import assert_all_close, V\n'), ((3831, 3863), 'tests.utils.assert_all_close', 'assert_all_close', (['res_th', 'res_py'], {}), '(res_th, res_py)\n', (3847, 3863), False, 'from tests.utils import assert_all_close, V\n'), ((4108, 4139), 'torch.autograd.Variable', 'Variable', (['x'], {'requires_grad': '(True)'}), '(x, requires_grad=True)\n', (4116, 4139), False, 'from torch.autograd import Variable\n'), ((3718, 3722), 'tests.utils.V', 'V', (['x'], {}), '(x)\n', (3719, 3722), False, 'from tests.utils import assert_all_close, V\n'), ((3724, 3733), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (3725, 3733), False, 'from tests.utils import assert_all_close, V\n'), ((3287, 3296), 'tests.utils.V', 'V', (['self.x'], {}), '(self.x)\n', (3288, 3296), False, 'from tests.utils import assert_all_close, V\n'), ((3336, 3345), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (3337, 3345), False, 'from tests.utils import assert_all_close, V\n'), ((3777, 3781), 'tests.utils.V', 'V', (['x'], {}), '(x)\n', (3778, 3781), False, 'from tests.utils import assert_all_close, V\n'), ((3783, 3792), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (3784, 3792), False, 'from tests.utils import assert_all_close, V\n'), ((4201, 4210), 'tests.utils.V', 'V', (['self.y'], {}), '(self.y)\n', (4202, 4210), False, 'from tests.utils import assert_all_close, V\n')]
|
# pylint: disable=g-bad-file-header
# Copyright 2019 DeepMind Technologies Limited. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""Tests for bsuite.experiments.mnist."""
# Import all required packages
from absl.testing import absltest
from bsuite.experiments.mnist import mnist
from dm_env import test_utils
import numpy as np
class CatchInterfaceTest(test_utils.EnvironmentTestMixin, absltest.TestCase):
def make_object_under_test(self):
return mnist.MNISTBandit(seed=101)
def make_action_sequence(self):
num_actions = self.environment.action_spec().num_values
rng = np.random.RandomState(42)
for _ in range(100):
yield rng.randint(num_actions)
if __name__ == '__main__':
absltest.main()
|
[
"bsuite.experiments.mnist.mnist.MNISTBandit",
"absl.testing.absltest.main",
"numpy.random.RandomState"
] |
[((1312, 1327), 'absl.testing.absltest.main', 'absltest.main', ([], {}), '()\n', (1325, 1327), False, 'from absl.testing import absltest\n'), ((1060, 1087), 'bsuite.experiments.mnist.mnist.MNISTBandit', 'mnist.MNISTBandit', ([], {'seed': '(101)'}), '(seed=101)\n', (1077, 1087), False, 'from bsuite.experiments.mnist import mnist\n'), ((1193, 1218), 'numpy.random.RandomState', 'np.random.RandomState', (['(42)'], {}), '(42)\n', (1214, 1218), True, 'import numpy as np\n')]
|
# Copyright 2018 The TensorFlow Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
"""extenders tests."""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import tensorflow as tf
import numpy as np
from tensorflow.python.data.ops import dataset_ops
from tensorflow.python.feature_column import feature_column_lib as fc
from tensorflow.python.framework import constant_op
from tensorflow.python.keras import metrics as metrics_module
from tensorflow.python.platform import test
from tensorflow_estimator.python.estimator import extenders
from tensorflow_estimator.python.estimator import run_config
from tensorflow_estimator.python.estimator.canned import linear
def get_input_fn(x, y):
def input_fn():
dataset = tf.compat.v1.data.Dataset.from_tensor_slices({'x': x, 'y': y})
iterator = tf.compat.v1.data.make_one_shot_iterator(dataset)
features = iterator.get_next()
labels = features.pop('y')
return features, labels
return input_fn
class AddMetricsTest(tf.test.TestCase):
def test_should_add_metrics(self):
def _test_metric_fn(metric_fn):
input_fn = get_input_fn(
x=np.arange(4)[:, None, None], y=np.ones(4)[:, None])
config = run_config.RunConfig(log_step_count_steps=1)
estimator = linear.LinearClassifierV2([tf.feature_column.numeric_column('x')],
config=config)
estimator = extenders.add_metrics(estimator, metric_fn)
estimator.train(input_fn=input_fn)
metrics = estimator.evaluate(input_fn=input_fn)
self.assertIn('mean_x', metrics)
self.assertEqual(1.5, metrics['mean_x'])
# assert that it keeps original estimators metrics
self.assertIn('auc', metrics)
def metric_fn(features):
metric = metrics_module.Mean()
metric.update_state(features['x'])
return {'mean_x': metric}
_test_metric_fn(metric_fn)
def test_should_error_out_for_not_recognized_args(self):
estimator = linear.LinearClassifierV2([tf.feature_column.numeric_column('x')])
def metric_fn(features, not_recognized):
_, _ = features, not_recognized
return {}
with self.assertRaisesRegexp(ValueError, 'not_recognized'):
estimator = extenders.add_metrics(estimator, metric_fn)
def test_all_supported_args(self):
input_fn = get_input_fn(x=[[[0.]]], y=[[[1]]])
estimator = linear.LinearClassifierV2([tf.feature_column.numeric_column('x')])
def metric_fn(features, predictions, labels, config):
self.assertIn('x', features)
self.assertIsNotNone(labels)
self.assertIn('logistic', predictions)
self.assertTrue(isinstance(config, run_config.RunConfig))
return {}
estimator = extenders.add_metrics(estimator, metric_fn)
estimator.train(input_fn=input_fn)
estimator.evaluate(input_fn=input_fn)
def test_all_supported_args_in_different_order(self):
input_fn = get_input_fn(x=[[[0.]]], y=[[[1]]])
estimator = linear.LinearClassifierV2([tf.feature_column.numeric_column('x')])
def metric_fn(labels, config, features, predictions):
self.assertIn('x', features)
self.assertIsNotNone(labels)
self.assertIn('logistic', predictions)
self.assertTrue(isinstance(config, run_config.RunConfig))
return {}
estimator = extenders.add_metrics(estimator, metric_fn)
estimator.train(input_fn=input_fn)
estimator.evaluate(input_fn=input_fn)
def test_all_args_are_optional(self):
def _test_metric_fn(metric_fn):
input_fn = get_input_fn(x=[[[0.]]], y=[[[1]]])
estimator = linear.LinearClassifierV2([tf.feature_column.numeric_column('x')])
estimator = extenders.add_metrics(estimator, metric_fn)
estimator.train(input_fn=input_fn)
metrics = estimator.evaluate(input_fn=input_fn)
self.assertEqual(2., metrics['two'])
def metric_fn():
metric = metrics_module.Mean()
metric.update_state(tf.constant([2.]))
return {'two': metric}
_test_metric_fn(metric_fn)
def test_overrides_existing_metrics(self):
def _test_metric_fn(metric_fn):
input_fn = get_input_fn(x=[[[0.]]], y=[[[1]]])
estimator = linear.LinearClassifierV2([tf.feature_column.numeric_column('x')])
estimator.train(input_fn=input_fn)
metrics = estimator.evaluate(input_fn=input_fn)
self.assertNotEqual(2., metrics['auc'])
estimator = extenders.add_metrics(estimator, metric_fn)
metrics = estimator.evaluate(input_fn=input_fn)
self.assertEqual(2., metrics['auc'])
def metric_fn():
metric = metrics_module.Mean()
metric.update_state(tf.constant([2.]))
return {'auc': metric}
_test_metric_fn(metric_fn)
if __name__ == '__main__':
tf.test.main()
|
[
"tensorflow_estimator.python.estimator.run_config.RunConfig",
"tensorflow.python.keras.metrics.Mean",
"numpy.ones",
"tensorflow.compat.v1.data.make_one_shot_iterator",
"tensorflow.test.main",
"tensorflow.feature_column.numeric_column",
"tensorflow.compat.v1.data.Dataset.from_tensor_slices",
"tensorflow.constant",
"tensorflow_estimator.python.estimator.extenders.add_metrics",
"numpy.arange"
] |
[((5382, 5396), 'tensorflow.test.main', 'tf.test.main', ([], {}), '()\n', (5394, 5396), True, 'import tensorflow as tf\n'), ((1391, 1453), 'tensorflow.compat.v1.data.Dataset.from_tensor_slices', 'tf.compat.v1.data.Dataset.from_tensor_slices', (["{'x': x, 'y': y}"], {}), "({'x': x, 'y': y})\n", (1435, 1453), True, 'import tensorflow as tf\n'), ((1469, 1518), 'tensorflow.compat.v1.data.make_one_shot_iterator', 'tf.compat.v1.data.make_one_shot_iterator', (['dataset'], {}), '(dataset)\n', (1509, 1518), True, 'import tensorflow as tf\n'), ((3370, 3413), 'tensorflow_estimator.python.estimator.extenders.add_metrics', 'extenders.add_metrics', (['estimator', 'metric_fn'], {}), '(estimator, metric_fn)\n', (3391, 3413), False, 'from tensorflow_estimator.python.estimator import extenders\n'), ((3958, 4001), 'tensorflow_estimator.python.estimator.extenders.add_metrics', 'extenders.add_metrics', (['estimator', 'metric_fn'], {}), '(estimator, metric_fn)\n', (3979, 4001), False, 'from tensorflow_estimator.python.estimator import extenders\n'), ((1858, 1902), 'tensorflow_estimator.python.estimator.run_config.RunConfig', 'run_config.RunConfig', ([], {'log_step_count_steps': '(1)'}), '(log_step_count_steps=1)\n', (1878, 1902), False, 'from tensorflow_estimator.python.estimator import run_config\n'), ((2066, 2109), 'tensorflow_estimator.python.estimator.extenders.add_metrics', 'extenders.add_metrics', (['estimator', 'metric_fn'], {}), '(estimator, metric_fn)\n', (2087, 2109), False, 'from tensorflow_estimator.python.estimator import extenders\n'), ((2430, 2451), 'tensorflow.python.keras.metrics.Mean', 'metrics_module.Mean', ([], {}), '()\n', (2449, 2451), True, 'from tensorflow.python.keras import metrics as metrics_module\n'), ((2883, 2926), 'tensorflow_estimator.python.estimator.extenders.add_metrics', 'extenders.add_metrics', (['estimator', 'metric_fn'], {}), '(estimator, metric_fn)\n', (2904, 2926), False, 'from tensorflow_estimator.python.estimator import extenders\n'), ((4317, 4360), 'tensorflow_estimator.python.estimator.extenders.add_metrics', 'extenders.add_metrics', (['estimator', 'metric_fn'], {}), '(estimator, metric_fn)\n', (4338, 4360), False, 'from tensorflow_estimator.python.estimator import extenders\n'), ((4537, 4558), 'tensorflow.python.keras.metrics.Mean', 'metrics_module.Mean', ([], {}), '()\n', (4556, 4558), True, 'from tensorflow.python.keras import metrics as metrics_module\n'), ((5045, 5088), 'tensorflow_estimator.python.estimator.extenders.add_metrics', 'extenders.add_metrics', (['estimator', 'metric_fn'], {}), '(estimator, metric_fn)\n', (5066, 5088), False, 'from tensorflow_estimator.python.estimator import extenders\n'), ((5223, 5244), 'tensorflow.python.keras.metrics.Mean', 'metrics_module.Mean', ([], {}), '()\n', (5242, 5244), True, 'from tensorflow.python.keras import metrics as metrics_module\n'), ((2660, 2697), 'tensorflow.feature_column.numeric_column', 'tf.feature_column.numeric_column', (['"""x"""'], {}), "('x')\n", (2692, 2697), True, 'import tensorflow as tf\n'), ((3059, 3096), 'tensorflow.feature_column.numeric_column', 'tf.feature_column.numeric_column', (['"""x"""'], {}), "('x')\n", (3091, 3096), True, 'import tensorflow as tf\n'), ((3647, 3684), 'tensorflow.feature_column.numeric_column', 'tf.feature_column.numeric_column', (['"""x"""'], {}), "('x')\n", (3679, 3684), True, 'import tensorflow as tf\n'), ((4585, 4603), 'tensorflow.constant', 'tf.constant', (['[2.0]'], {}), '([2.0])\n', (4596, 4603), True, 'import tensorflow as tf\n'), ((5271, 5289), 'tensorflow.constant', 'tf.constant', (['[2.0]'], {}), '([2.0])\n', (5282, 5289), True, 'import tensorflow as tf\n'), ((1948, 1985), 'tensorflow.feature_column.numeric_column', 'tf.feature_column.numeric_column', (['"""x"""'], {}), "('x')\n", (1980, 1985), True, 'import tensorflow as tf\n'), ((4259, 4296), 'tensorflow.feature_column.numeric_column', 'tf.feature_column.numeric_column', (['"""x"""'], {}), "('x')\n", (4291, 4296), True, 'import tensorflow as tf\n'), ((4845, 4882), 'tensorflow.feature_column.numeric_column', 'tf.feature_column.numeric_column', (['"""x"""'], {}), "('x')\n", (4877, 4882), True, 'import tensorflow as tf\n'), ((1791, 1803), 'numpy.arange', 'np.arange', (['(4)'], {}), '(4)\n', (1800, 1803), True, 'import numpy as np\n'), ((1822, 1832), 'numpy.ones', 'np.ones', (['(4)'], {}), '(4)\n', (1829, 1832), True, 'import numpy as np\n')]
|
import numpy as np
from scipy.sparse import csc_matrix, save_npz, hstack
import time
import argparse
import gzip
from pysam import VariantFile, TabixFile
import json
import os
import itertools
parser = argparse.ArgumentParser(description='Pull genotypes.')
parser.add_argument('vcf_file', type=str, help='VCF file to pull from.')
parser.add_argument('assembly', type=str, help='Human genome reference used.')
parser.add_argument('out_directory', type=str, help='Output directory.')
parser.add_argument('chrom', type=str, help='Chromosome of interest.')
parser.add_argument('--batch_size', type=int, default=-1, help='Restrict number of positions per file to batch_size.')
parser.add_argument('--batch_num', type=int, default=0, help='To be used along with batch_size to restrict positions per file. Will include positions >= batch_num*batch_size and <= (batch_num+1)*batch_size')
parser.add_argument('--maxsize', type=int, default=500000000, help='Amount of memory per block.')
parser.add_argument('--additional_vcf_files', type=str, nargs='+', help='Additional VCF files to pull data from.')
parser.add_argument('--id_mapper_file', type=str, default=None, help='File that maps old ids to new ones.')
parser.add_argument('--id_mapper_sep', type=str, default='\t', help='Separater to parse id_mapper_file.')
parser.add_argument('--old_id_index', type=int, default=0, help='Index of old_id in id_mapper_file.')
parser.add_argument('--new_id_index', type=int, default=1, help='Index of new_id in id_mapper_file.')
args = parser.parse_args()
t0 = time.time()
chrom_int = 23 if args.chrom == 'X' else 24 if args.chrom == 'Y' else 25 if args.chrom == 'MT' else int(args.chrom)
gen_mapping = {'./.': -1, '0/0': 0, '0|0': 0, '0/1': 1, '0|1': 1, '1/0': 1, '1|0': 1, '1/1': 2, '1|1': 2}
def process_header(vcf):
sample_ids = [x.replace('.', '_') for x in vcf.header.samples]
if args.id_mapper_file is not None:
old_id_to_new_id = dict()
with open(args.id_mapper_file, 'r') as f:
for line in f:
pieces = line.strip().split(args.id_mapper_sep)
if len(pieces)>args.old_id_index and len(pieces)>args.new_id_index:
old_id_to_new_id[pieces[args.old_id_index]] = pieces[args.new_id_index]
sample_ids = [old_id_to_new_id[x] for x in sample_ids]
sample_file = '%s/samples.json' % args.out_directory
if os.path.isfile(sample_file):
with open(sample_file, 'r') as f:
stored_sample_ids = json.load(f)
assert sample_ids == stored_sample_ids
else:
with open(sample_file, 'w+') as f:
json.dump(sample_ids, f)
return sample_ids, vcf.header.contigs
def process_body(records, sample_ids):
data, indices, indptr, index = np.zeros((args.maxsize,), dtype=np.int8), np.zeros((args.maxsize,), dtype=int), [0], 0
chrom_coord = []
with gzip.open('%s/chr.%s.%d.gen.variants.txt.gz' % (args.out_directory, args.chrom, args.batch_num), 'wt') as variant_f:
for line in records:
pieces = line.strip().split('\t')
fmt = pieces[8].strip().split(':')
# Write variant to file
variant_f.write('\t'.join(pieces[:9]) + '\n')
# pull chrom_coord information
pos, _, ref, alt = pieces[1:5]
is_biallelic_snp = 1 if len(ref) == 1 and len(alt) == 1 and ref != '.' and alt != '.' else 0
is_pass = pieces[6] == 'PASS'
chrom_coord.append((chrom_int, int(pos), is_biallelic_snp, is_pass))
# pull genotypes
gen_index = fmt.index('GT')
for i, piece in enumerate(pieces[9:]):
segment = piece.split(':', maxsplit=gen_index+1)
gt = gen_mapping.get(segment[gen_index], -1) # For now we mark multi-base loci as unknown
if gt != 0:
indices[index] = i
data[index] = gt
index += 1
indptr.append(index)
gen = csc_matrix((data[:index], indices[:index], indptr), shape=(len(sample_ids), len(indptr)-1), dtype=np.int8)
# Save to file
save_npz('%s/chr.%s.%d.gen' % (args.out_directory, args.chrom, args.batch_num), gen)
np.save('%s/chr.%s.%d.gen.coordinates' % (args.out_directory, args.chrom, args.batch_num), np.asarray(np.asarray(chrom_coord, dtype=int), dtype=int))
print('Completed in ', time.time()-t0, 'sec')
with open('%s/info.json' % args.out_directory, 'w+') as f:
json.dump({'assembly': args.assembly, 'batch_size': args.batch_size, 'vcf_directory': '/'.join(args.vcf_file.split('/')[:-1])}, f)
vcf = VariantFile(args.vcf_file)
sample_ids, contigs = process_header(vcf)
if args.additional_vcf_files is not None:
for vcf_file in args.additional_vcf_files:
if os.path.isfile(vcf_file):
new_vcf = VariantFile(vcf_file)
new_sample_ids, _ = process_header(new_vcf)
assert sample_ids == new_sample_ids
else:
print(vcf_file, 'does not exist')
contig = None
if args.chrom in contigs:
contig = contigs[args.chrom]
elif 'chr%s' % args.chrom in contigs:
contig = contigs['chr%s' % args.chrom]
else:
raise Exception('Trouble finding contig', args.chrom, 'in', contigs)
print('Chrom length', contig.length)
vcf_files = [args.vcf_file]
if args.additional_vcf_files is not None:
vcf_files.extend(args.additional_vcf_files)
if np.all([os.path.isfile(vcf_file + '.tbi') for vcf_file in vcf_files]):
vcfs = [TabixFile(vcf_file, parser=None) for vcf_file in vcf_files]
if args.batch_size != -1:
start_pos, end_pos = args.batch_num*args.batch_size, (args.batch_num+1)*args.batch_size
print('Interval', start_pos, end_pos)
if start_pos < contig.length:
process_body(itertools.chain(*[vcf.fetch(reference=contig.name, start=start_pos, end=end_pos) for vcf in vcfs]), sample_ids)
else:
print('Interval (%d-%d) is longer than chromosome (length=%d).' % (start_pos, end_pos, contig.length))
else:
process_body(itertools.chain(*[vcf.fetch(reference=contig.name) for vcf in vcfs]), sample_ids)
else:
print('Error, .tbi files are missing.')
|
[
"pysam.VariantFile",
"argparse.ArgumentParser",
"gzip.open",
"numpy.asarray",
"os.path.isfile",
"numpy.zeros",
"pysam.TabixFile",
"json.load",
"scipy.sparse.save_npz",
"time.time",
"json.dump"
] |
[((204, 258), 'argparse.ArgumentParser', 'argparse.ArgumentParser', ([], {'description': '"""Pull genotypes."""'}), "(description='Pull genotypes.')\n", (227, 258), False, 'import argparse\n'), ((1546, 1557), 'time.time', 'time.time', ([], {}), '()\n', (1555, 1557), False, 'import time\n'), ((4632, 4658), 'pysam.VariantFile', 'VariantFile', (['args.vcf_file'], {}), '(args.vcf_file)\n', (4643, 4658), False, 'from pysam import VariantFile, TabixFile\n'), ((2395, 2422), 'os.path.isfile', 'os.path.isfile', (['sample_file'], {}), '(sample_file)\n', (2409, 2422), False, 'import os\n'), ((4141, 4230), 'scipy.sparse.save_npz', 'save_npz', (["('%s/chr.%s.%d.gen' % (args.out_directory, args.chrom, args.batch_num))", 'gen'], {}), "('%s/chr.%s.%d.gen' % (args.out_directory, args.chrom, args.\n batch_num), gen)\n", (4149, 4230), False, 'from scipy.sparse import csc_matrix, save_npz, hstack\n'), ((2771, 2811), 'numpy.zeros', 'np.zeros', (['(args.maxsize,)'], {'dtype': 'np.int8'}), '((args.maxsize,), dtype=np.int8)\n', (2779, 2811), True, 'import numpy as np\n'), ((2813, 2849), 'numpy.zeros', 'np.zeros', (['(args.maxsize,)'], {'dtype': 'int'}), '((args.maxsize,), dtype=int)\n', (2821, 2849), True, 'import numpy as np\n'), ((2889, 2996), 'gzip.open', 'gzip.open', (["('%s/chr.%s.%d.gen.variants.txt.gz' % (args.out_directory, args.chrom, args\n .batch_num))", '"""wt"""'], {}), "('%s/chr.%s.%d.gen.variants.txt.gz' % (args.out_directory, args.\n chrom, args.batch_num), 'wt')\n", (2898, 2996), False, 'import gzip\n'), ((4802, 4826), 'os.path.isfile', 'os.path.isfile', (['vcf_file'], {}), '(vcf_file)\n', (4816, 4826), False, 'import os\n'), ((5439, 5472), 'os.path.isfile', 'os.path.isfile', (["(vcf_file + '.tbi')"], {}), "(vcf_file + '.tbi')\n", (5453, 5472), False, 'import os\n'), ((5514, 5546), 'pysam.TabixFile', 'TabixFile', (['vcf_file'], {'parser': 'None'}), '(vcf_file, parser=None)\n', (5523, 5546), False, 'from pysam import VariantFile, TabixFile\n'), ((2498, 2510), 'json.load', 'json.load', (['f'], {}), '(f)\n', (2507, 2510), False, 'import json\n'), ((2627, 2651), 'json.dump', 'json.dump', (['sample_ids', 'f'], {}), '(sample_ids, f)\n', (2636, 2651), False, 'import json\n'), ((4332, 4366), 'numpy.asarray', 'np.asarray', (['chrom_coord'], {'dtype': 'int'}), '(chrom_coord, dtype=int)\n', (4342, 4366), True, 'import numpy as np\n'), ((4407, 4418), 'time.time', 'time.time', ([], {}), '()\n', (4416, 4418), False, 'import time\n'), ((4850, 4871), 'pysam.VariantFile', 'VariantFile', (['vcf_file'], {}), '(vcf_file)\n', (4861, 4871), False, 'from pysam import VariantFile, TabixFile\n')]
|
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from pathlib import Path
import numpy as np
import pandas as pd
import librosa
import csv
from paddle import fluid
from parakeet import g2p
from parakeet import audio
from parakeet.data.sampler import *
from parakeet.data.datacargo import DataCargo
from parakeet.data.batch import TextIDBatcher, SpecBatcher
from parakeet.data.dataset import DatasetMixin, TransformDataset, CacheDataset, SliceDataset
from parakeet.models.transformer_tts.utils import *
class LJSpeechLoader:
def __init__(self,
config,
args,
nranks,
rank,
is_vocoder=False,
shuffle=True):
place = fluid.CUDAPlace(rank) if args.use_gpu else fluid.CPUPlace()
LJSPEECH_ROOT = Path(args.data_path)
metadata = LJSpeechMetaData(LJSPEECH_ROOT)
transformer = LJSpeech(config)
dataset = TransformDataset(metadata, transformer)
dataset = CacheDataset(dataset)
sampler = DistributedSampler(
len(dataset), nranks, rank, shuffle=shuffle)
assert args.batch_size % nranks == 0
each_bs = args.batch_size // nranks
if is_vocoder:
dataloader = DataCargo(
dataset,
sampler=sampler,
batch_size=each_bs,
shuffle=shuffle,
batch_fn=batch_examples_vocoder,
drop_last=True)
else:
dataloader = DataCargo(
dataset,
sampler=sampler,
batch_size=each_bs,
shuffle=shuffle,
batch_fn=batch_examples,
drop_last=True)
self.reader = fluid.io.DataLoader.from_generator(
capacity=32,
iterable=True,
use_double_buffer=True,
return_list=True)
self.reader.set_batch_generator(dataloader, place)
class LJSpeechMetaData(DatasetMixin):
def __init__(self, root):
self.root = Path(root)
self._wav_dir = self.root.joinpath("wavs")
csv_path = self.root.joinpath("metadata.csv")
self._table = pd.read_csv(
csv_path,
sep="|",
header=None,
quoting=csv.QUOTE_NONE,
names=["fname", "raw_text", "normalized_text"])
def get_example(self, i):
fname, raw_text, normalized_text = self._table.iloc[i]
fname = str(self._wav_dir.joinpath(fname + ".wav"))
return fname, raw_text, normalized_text
def __len__(self):
return len(self._table)
class LJSpeech(object):
def __init__(self, config):
super(LJSpeech, self).__init__()
self.config = config
self._ljspeech_processor = audio.AudioProcessor(
sample_rate=config['audio']['sr'],
num_mels=config['audio']['num_mels'],
min_level_db=config['audio']['min_level_db'],
ref_level_db=config['audio']['ref_level_db'],
n_fft=config['audio']['n_fft'],
win_length=config['audio']['win_length'],
hop_length=config['audio']['hop_length'],
power=config['audio']['power'],
preemphasis=config['audio']['preemphasis'],
signal_norm=True,
symmetric_norm=False,
max_norm=1.,
mel_fmin=0,
mel_fmax=None,
clip_norm=True,
griffin_lim_iters=60,
do_trim_silence=False,
sound_norm=False)
def __call__(self, metadatum):
"""All the code for generating an Example from a metadatum. If you want a
different preprocessing pipeline, you can override this method.
This method may require several processor, each of which has a lot of options.
In this case, you'd better pass a composed transform and pass it to the init
method.
"""
fname, raw_text, normalized_text = metadatum
# load -> trim -> preemphasis -> stft -> magnitude -> mel_scale -> logscale -> normalize
wav = self._ljspeech_processor.load_wav(str(fname))
mag = self._ljspeech_processor.spectrogram(wav).astype(np.float32)
mel = self._ljspeech_processor.melspectrogram(wav).astype(np.float32)
phonemes = np.array(
g2p.en.text_to_sequence(normalized_text), dtype=np.int64)
return (mag, mel, phonemes
) # maybe we need to implement it as a map in the future
def batch_examples(batch):
texts = []
mels = []
mel_inputs = []
mel_lens = []
text_lens = []
pos_texts = []
pos_mels = []
for data in batch:
_, mel, text = data
mel_inputs.append(
np.concatenate(
[np.zeros([mel.shape[0], 1], np.float32), mel[:, :-1]],
axis=-1))
mel_lens.append(mel.shape[1])
text_lens.append(len(text))
pos_texts.append(np.arange(1, len(text) + 1))
pos_mels.append(np.arange(1, mel.shape[1] + 1))
mels.append(mel)
texts.append(text)
# Sort by text_len in descending order
texts = [
i
for i, _ in sorted(
zip(texts, text_lens), key=lambda x: x[1], reverse=True)
]
mels = [
i
for i, _ in sorted(
zip(mels, text_lens), key=lambda x: x[1], reverse=True)
]
mel_inputs = [
i
for i, _ in sorted(
zip(mel_inputs, text_lens), key=lambda x: x[1], reverse=True)
]
mel_lens = [
i
for i, _ in sorted(
zip(mel_lens, text_lens), key=lambda x: x[1], reverse=True)
]
pos_texts = [
i
for i, _ in sorted(
zip(pos_texts, text_lens), key=lambda x: x[1], reverse=True)
]
pos_mels = [
i
for i, _ in sorted(
zip(pos_mels, text_lens), key=lambda x: x[1], reverse=True)
]
text_lens = sorted(text_lens, reverse=True)
# Pad sequence with largest len of the batch
texts = TextIDBatcher(pad_id=0)(texts) #(B, T)
pos_texts = TextIDBatcher(pad_id=0)(pos_texts) #(B,T)
pos_mels = TextIDBatcher(pad_id=0)(pos_mels) #(B,T)
mels = np.transpose(
SpecBatcher(pad_value=0.)(mels), axes=(0, 2, 1)) #(B,T,num_mels)
mel_inputs = np.transpose(
SpecBatcher(pad_value=0.)(mel_inputs), axes=(0, 2, 1)) #(B,T,num_mels)
enc_slf_mask = get_attn_key_pad_mask(pos_texts).astype(np.float32)
enc_query_mask = get_non_pad_mask(pos_texts).astype(np.float32)
dec_slf_mask = get_dec_attn_key_pad_mask(pos_mels,
mel_inputs).astype(np.float32)
enc_dec_mask = get_attn_key_pad_mask(enc_query_mask[:, :, 0]).astype(
np.float32)
dec_query_slf_mask = get_non_pad_mask(pos_mels).astype(np.float32)
dec_query_mask = get_non_pad_mask(pos_mels).astype(np.float32)
return (texts, mels, mel_inputs, pos_texts, pos_mels, np.array(text_lens),
np.array(mel_lens), enc_slf_mask, enc_query_mask, dec_slf_mask,
enc_dec_mask, dec_query_slf_mask, dec_query_mask)
def batch_examples_vocoder(batch):
mels = []
mags = []
for data in batch:
mag, mel, _ = data
mels.append(mel)
mags.append(mag)
mels = np.transpose(SpecBatcher(pad_value=0.)(mels), axes=(0, 2, 1))
mags = np.transpose(SpecBatcher(pad_value=0.)(mags), axes=(0, 2, 1))
return (mels, mags)
|
[
"parakeet.data.datacargo.DataCargo",
"parakeet.data.dataset.TransformDataset",
"parakeet.audio.AudioProcessor",
"pandas.read_csv",
"pathlib.Path",
"paddle.fluid.io.DataLoader.from_generator",
"parakeet.g2p.en.text_to_sequence",
"paddle.fluid.CPUPlace",
"numpy.array",
"parakeet.data.batch.TextIDBatcher",
"parakeet.data.dataset.CacheDataset",
"numpy.zeros",
"paddle.fluid.CUDAPlace",
"parakeet.data.batch.SpecBatcher",
"numpy.arange"
] |
[((1375, 1395), 'pathlib.Path', 'Path', (['args.data_path'], {}), '(args.data_path)\n', (1379, 1395), False, 'from pathlib import Path\n'), ((1504, 1543), 'parakeet.data.dataset.TransformDataset', 'TransformDataset', (['metadata', 'transformer'], {}), '(metadata, transformer)\n', (1520, 1543), False, 'from parakeet.data.dataset import DatasetMixin, TransformDataset, CacheDataset, SliceDataset\n'), ((1562, 1583), 'parakeet.data.dataset.CacheDataset', 'CacheDataset', (['dataset'], {}), '(dataset)\n', (1574, 1583), False, 'from parakeet.data.dataset import DatasetMixin, TransformDataset, CacheDataset, SliceDataset\n'), ((2309, 2417), 'paddle.fluid.io.DataLoader.from_generator', 'fluid.io.DataLoader.from_generator', ([], {'capacity': '(32)', 'iterable': '(True)', 'use_double_buffer': '(True)', 'return_list': '(True)'}), '(capacity=32, iterable=True,\n use_double_buffer=True, return_list=True)\n', (2343, 2417), False, 'from paddle import fluid\n'), ((2612, 2622), 'pathlib.Path', 'Path', (['root'], {}), '(root)\n', (2616, 2622), False, 'from pathlib import Path\n'), ((2750, 2870), 'pandas.read_csv', 'pd.read_csv', (['csv_path'], {'sep': '"""|"""', 'header': 'None', 'quoting': 'csv.QUOTE_NONE', 'names': "['fname', 'raw_text', 'normalized_text']"}), "(csv_path, sep='|', header=None, quoting=csv.QUOTE_NONE, names=[\n 'fname', 'raw_text', 'normalized_text'])\n", (2761, 2870), True, 'import pandas as pd\n'), ((3348, 3917), 'parakeet.audio.AudioProcessor', 'audio.AudioProcessor', ([], {'sample_rate': "config['audio']['sr']", 'num_mels': "config['audio']['num_mels']", 'min_level_db': "config['audio']['min_level_db']", 'ref_level_db': "config['audio']['ref_level_db']", 'n_fft': "config['audio']['n_fft']", 'win_length': "config['audio']['win_length']", 'hop_length': "config['audio']['hop_length']", 'power': "config['audio']['power']", 'preemphasis': "config['audio']['preemphasis']", 'signal_norm': '(True)', 'symmetric_norm': '(False)', 'max_norm': '(1.0)', 'mel_fmin': '(0)', 'mel_fmax': 'None', 'clip_norm': '(True)', 'griffin_lim_iters': '(60)', 'do_trim_silence': '(False)', 'sound_norm': '(False)'}), "(sample_rate=config['audio']['sr'], num_mels=config[\n 'audio']['num_mels'], min_level_db=config['audio']['min_level_db'],\n ref_level_db=config['audio']['ref_level_db'], n_fft=config['audio'][\n 'n_fft'], win_length=config['audio']['win_length'], hop_length=config[\n 'audio']['hop_length'], power=config['audio']['power'], preemphasis=\n config['audio']['preemphasis'], signal_norm=True, symmetric_norm=False,\n max_norm=1.0, mel_fmin=0, mel_fmax=None, clip_norm=True,\n griffin_lim_iters=60, do_trim_silence=False, sound_norm=False)\n", (3368, 3917), False, 'from parakeet import audio\n'), ((6602, 6625), 'parakeet.data.batch.TextIDBatcher', 'TextIDBatcher', ([], {'pad_id': '(0)'}), '(pad_id=0)\n', (6615, 6625), False, 'from parakeet.data.batch import TextIDBatcher, SpecBatcher\n'), ((6658, 6681), 'parakeet.data.batch.TextIDBatcher', 'TextIDBatcher', ([], {'pad_id': '(0)'}), '(pad_id=0)\n', (6671, 6681), False, 'from parakeet.data.batch import TextIDBatcher, SpecBatcher\n'), ((6716, 6739), 'parakeet.data.batch.TextIDBatcher', 'TextIDBatcher', ([], {'pad_id': '(0)'}), '(pad_id=0)\n', (6729, 6739), False, 'from parakeet.data.batch import TextIDBatcher, SpecBatcher\n'), ((7530, 7549), 'numpy.array', 'np.array', (['text_lens'], {}), '(text_lens)\n', (7538, 7549), True, 'import numpy as np\n'), ((7563, 7581), 'numpy.array', 'np.array', (['mel_lens'], {}), '(mel_lens)\n', (7571, 7581), True, 'import numpy as np\n'), ((1290, 1311), 'paddle.fluid.CUDAPlace', 'fluid.CUDAPlace', (['rank'], {}), '(rank)\n', (1305, 1311), False, 'from paddle import fluid\n'), ((1333, 1349), 'paddle.fluid.CPUPlace', 'fluid.CPUPlace', ([], {}), '()\n', (1347, 1349), False, 'from paddle import fluid\n'), ((1818, 1943), 'parakeet.data.datacargo.DataCargo', 'DataCargo', (['dataset'], {'sampler': 'sampler', 'batch_size': 'each_bs', 'shuffle': 'shuffle', 'batch_fn': 'batch_examples_vocoder', 'drop_last': '(True)'}), '(dataset, sampler=sampler, batch_size=each_bs, shuffle=shuffle,\n batch_fn=batch_examples_vocoder, drop_last=True)\n', (1827, 1943), False, 'from parakeet.data.datacargo import DataCargo\n'), ((2076, 2193), 'parakeet.data.datacargo.DataCargo', 'DataCargo', (['dataset'], {'sampler': 'sampler', 'batch_size': 'each_bs', 'shuffle': 'shuffle', 'batch_fn': 'batch_examples', 'drop_last': '(True)'}), '(dataset, sampler=sampler, batch_size=each_bs, shuffle=shuffle,\n batch_fn=batch_examples, drop_last=True)\n', (2085, 2193), False, 'from parakeet.data.datacargo import DataCargo\n'), ((4899, 4939), 'parakeet.g2p.en.text_to_sequence', 'g2p.en.text_to_sequence', (['normalized_text'], {}), '(normalized_text)\n', (4922, 4939), False, 'from parakeet import g2p\n'), ((5574, 5604), 'numpy.arange', 'np.arange', (['(1)', '(mel.shape[1] + 1)'], {}), '(1, mel.shape[1] + 1)\n', (5583, 5604), True, 'import numpy as np\n'), ((6791, 6817), 'parakeet.data.batch.SpecBatcher', 'SpecBatcher', ([], {'pad_value': '(0.0)'}), '(pad_value=0.0)\n', (6802, 6817), False, 'from parakeet.data.batch import TextIDBatcher, SpecBatcher\n'), ((6896, 6922), 'parakeet.data.batch.SpecBatcher', 'SpecBatcher', ([], {'pad_value': '(0.0)'}), '(pad_value=0.0)\n', (6907, 6922), False, 'from parakeet.data.batch import TextIDBatcher, SpecBatcher\n'), ((7879, 7905), 'parakeet.data.batch.SpecBatcher', 'SpecBatcher', ([], {'pad_value': '(0.0)'}), '(pad_value=0.0)\n', (7890, 7905), False, 'from parakeet.data.batch import TextIDBatcher, SpecBatcher\n'), ((7952, 7978), 'parakeet.data.batch.SpecBatcher', 'SpecBatcher', ([], {'pad_value': '(0.0)'}), '(pad_value=0.0)\n', (7963, 7978), False, 'from parakeet.data.batch import TextIDBatcher, SpecBatcher\n'), ((5341, 5380), 'numpy.zeros', 'np.zeros', (['[mel.shape[0], 1]', 'np.float32'], {}), '([mel.shape[0], 1], np.float32)\n', (5349, 5380), True, 'import numpy as np\n')]
|
import numpy as np
import geocoder
def process_df(df):
"""
df: pd.DataFrame
"""
df.dropna(subset=['lat', 'lon'], axis=0, inplace=True)
df.reset_index(drop=True, inplace=True)
# Add new column to hold the years
df["year"] = [int(x.split("-")[0]) for x in df['date']]
# Convert coordinates to decimal degrees ISO 6709 format i.e:(14.76º,
# -23.2234º)
lat_flip = np.logical_and(df["lat-dir"] == "S", df["lat"] >= 0)
df.loc[lat_flip, "lat"] *= -1
lon_flip = np.logical_and(df["lon-dir"] == "W", df["lon"] >= 0)
df.loc[lon_flip, "lon"] *= -1
legend = []
print('Starting to pull data from Google Geolocation API')
for i in range(len(df['impact-e'])):
print(i+1, "of {}".format(len(df)+1))
g = geocoder.google([df['lat'][i], df['lon'][i]], method='reverse')
city = '{}'.format(g.city) if g.city else "N/A"
country = '{}'.format(g.country) if g.country else "N/A"
if city is not None and country is not None:
location = "Location: {},{}<br>".format(city, country)
if df['impact-e'][i] < 10:
legend.append('{}<10 kt<br>{}'.format(location, str(df['date'][i])))
else:
legend.append('{}{} kt<br>{}'.format(location, df['impact-e'][i], str(df['date'][i])))
df['legend'] = legend
return df
|
[
"geocoder.google",
"numpy.logical_and"
] |
[((409, 461), 'numpy.logical_and', 'np.logical_and', (["(df['lat-dir'] == 'S')", "(df['lat'] >= 0)"], {}), "(df['lat-dir'] == 'S', df['lat'] >= 0)\n", (423, 461), True, 'import numpy as np\n'), ((511, 563), 'numpy.logical_and', 'np.logical_and', (["(df['lon-dir'] == 'W')", "(df['lon'] >= 0)"], {}), "(df['lon-dir'] == 'W', df['lon'] >= 0)\n", (525, 563), True, 'import numpy as np\n'), ((777, 840), 'geocoder.google', 'geocoder.google', (["[df['lat'][i], df['lon'][i]]"], {'method': '"""reverse"""'}), "([df['lat'][i], df['lon'][i]], method='reverse')\n", (792, 840), False, 'import geocoder\n')]
|
"""
Offset Mirror Classes.
This module contains all the classes relating to the offset mirrors used in the
FEE and XRT. Each offset mirror contains a stepper motor and piezo motor to
control the pitch, and two pairs of motors to control the horizontal and
vertical gantries.
"""
import logging
import numpy as np
from ophyd import Component as Cpt
from ophyd import Device, EpicsSignal, EpicsSignalRO
from ophyd import FormattedComponent as FCpt
from ophyd import PVPositioner
from .device import GroupDevice
from .doc_stubs import basic_positioner_init
from .epics_motor import BeckhoffAxisNoOffset
from .inout import InOutRecordPositioner
from .interface import BaseInterface, FltMvInterface
from .pmps import TwinCATStatePMPS
from .signal import PytmcSignal
from .utils import get_status_value
logger = logging.getLogger(__name__)
class OMMotor(FltMvInterface, PVPositioner):
"""Base class for each motor in the LCLS offset mirror system."""
__doc__ += basic_positioner_init
# position
readback = Cpt(EpicsSignalRO, ':RBV', auto_monitor=True, kind='hinted')
setpoint = Cpt(EpicsSignal, ':VAL', auto_monitor=True, limits=True,
kind='normal')
done = Cpt(EpicsSignalRO, ':DMOV', auto_monitor=True, kind='omitted')
motor_egu = Cpt(EpicsSignal, ':RBV.EGU', kind='omitted')
# status
interlock = Cpt(EpicsSignalRO, ':INTERLOCK', kind='omitted')
enabled = Cpt(EpicsSignalRO, ':ENABLED', kind='omitted')
# limit switches
low_limit_switch = Cpt(EpicsSignalRO, ":LLS", kind='omitted')
high_limit_switch = Cpt(EpicsSignalRO, ":HLS", kind='omitted')
@property
def egu(self):
"""
Returns the Engineering Units of the readback PV, as reported by EPICS.
"""
return self.motor_egu.get()
def check_value(self, position):
"""
Checks that the value is both valid and within the motor's soft limits.
Parameters
----------
position : float
Position to check for validity.
Raises
------
ValueError
If position is `None`, `~numpy.NaN` or `~numpy.Inf`.
LimitError
If the position is outside the soft limits.
"""
# Check that we do not have a NaN or an Inf as those will
# will make the PLC very unhappy ...
if position is None or np.isnan(position) or np.isinf(position):
raise ValueError("Invalid value inputted: '{0}'".format(position))
# Use the built-in PVPositioner check_value
super().check_value(position)
class Pitch(OMMotor):
"""
HOMS Pitch Mechanism.
The axis is actually a piezo actuator and a stepper motor in series, and
this is reflected in the PV naming.
"""
__doc__ += basic_positioner_init
piezo_volts = FCpt(EpicsSignalRO, "{self._piezo}:VRBV", kind='normal')
stop_signal = FCpt(EpicsSignal, "{self._piezo}:STOP", kind='omitted')
# TODO: Limits will be added soon, but not present yet
def __init__(self, prefix, **kwargs):
# Predict the prefix of all piezo pvs
self._piezo = prefix.replace('MIRR', 'PIEZO')
super().__init__(prefix, **kwargs)
class Gantry(OMMotor):
"""
Gantry Axis.
The horizontal and vertical motion of the OffsetMirror are controlled by
two coupled stepper motors. Instructions are sent to both by simply
requesting a move on the primary so they are represented here as a single
motor with additional diagnostics and interlock.
Parameters
----------
prefix : str
Base prefix for both stepper motors e.g. 'XRT:M1H'. Do not include the
'P' or 'S' to indicate primary or secondary steppers.
gantry_prefix : str, optional
Prefix for the shared gantry diagnostics if it is different than the
stepper motor prefix.
"""
# Readbacks for gantry information
gantry_difference = FCpt(EpicsSignalRO, '{self.gantry_prefix}:GDIF',
kind='normal')
decoupled = FCpt(EpicsSignalRO, '{self.gantry_prefix}:DECOUPLE',
kind='config')
# Readbacks for the secondary motor
follower_readback = FCpt(EpicsSignalRO, '{self.follow_prefix}:RBV',
kind='normal')
follower_low_limit_switch = FCpt(EpicsSignalRO, '{self.follow_prefix}:LLS',
kind='omitted')
follower_high_limit_switch = FCpt(EpicsSignalRO,
'{self.follow_prefix}:HLS',
kind='omitted')
def __init__(self, prefix, *, gantry_prefix=None, **kwargs):
self.gantry_prefix = gantry_prefix or 'GANTRY:' + prefix
self.follow_prefix = prefix + ':S'
super().__init__(prefix + ':P', **kwargs)
def check_value(self, pos):
"""
Add additional check for the gantry coupling.
This is not a safety measure, but instead just here largely
for bookkeeping and to give the operator further feedback on why the
requested move is not completed.
"""
# Check that the gantry is not decoupled
if self.decoupled.get():
raise PermissionError("The gantry is not currently coupled")
# Allow OMMotor to check the value
super().check_value(pos)
class OffsetMirror(BaseInterface, GroupDevice):
"""
X-ray Offset Mirror class.
This is for each individual mirror system used in the FEE
and XRT. Controls for the pitch, and primary gantry x- and y-motors are
included.
When controlling the pitch motor, if the piezo is set to 'PID' mode, then
the pitch mechanism is setup to first move the stepper as close to the
desired position, then the piezo will kick in to constantly try and correct
any positional changes.
Parameters
----------
prefix : str
The EPICS base PV of the pitch motor.
prefix_xy : str
The EPICS base PV of the gantry x and y gantry motors.
xgantry_prefix : str
The name of the horizontal gantry if not identical to the prefix.
name : str
The name of the offset mirror.
"""
# Pitch Motor
pitch = FCpt(Pitch, "MIRR:{self.prefix}", kind='hinted')
# Gantry motors
xgantry = FCpt(Gantry, "{self._prefix_xy}:X",
gantry_prefix="{self._xgantry}",
add_prefix=['suffix', 'gantry_prefix'],
kind='normal')
ygantry = FCpt(Gantry, "{self._prefix_xy}:Y",
gantry_prefix='GANTRY:{self.prefix}:Y',
add_prefix=['suffix', 'gantry_prefix'],
kind='config')
# Transmission for Lightpath Interface
transmission = 1.0
# QIcon for UX
_icon = 'fa.minus-square'
# Subscription types
SUB_STATE = 'state'
tab_whitelist = ['pitch', 'xgantry', 'ygantry']
def __init__(self, prefix, *, prefix_xy=None,
xgantry_prefix=None, **kwargs):
# Handle prefix mangling
self._prefix_xy = prefix_xy or prefix
self._xgantry = xgantry_prefix or 'GANTRY:' + prefix + ':X'
super().__init__(prefix, **kwargs)
@property
def inserted(self):
"""Returns `True`. Treats OffsetMirror as always inserted."""
return True
@property
def removed(self):
"""Returns :keyword:`False`. Treats OffsetMirror as always inserted."""
return False
def format_status_info(self, status_info):
"""
Override status info handler to render the `OffsetMirror`.
Display `OffsetMirror` status info in the ipython terminal.
Parameters
----------
status_info: dict
Nested dictionary. Each level has keys name, kind, and is_device.
If is_device is True, subdevice dictionaries may follow. Otherwise,
the only other key in the dictionary will be value.
Returns
-------
status: str
Formatted string with all relevant status information.
"""
# happi metadata
try:
md = self.root.md
except AttributeError:
name = f'{self.prefix}'
else:
beamline = get_status_value(md, 'beamline')
functional_group = get_status_value(md, 'functional_group')
if functional_group is not None:
name = f'{self.prefix} ({beamline} {functional_group})'
else:
name = f'{self.prefix} ({beamline})'
p_position = get_status_value(status_info, 'pitch', 'position')
p_setpoint = get_status_value(status_info, 'pitch',
'setpoint', 'value')
p_units = get_status_value(status_info, 'pitch', 'setpoint',
'units')
return f"""\
{name}
------
pitch: ({self.pitch.prefix})
------
position: {p_position}
setpoint: {p_setpoint} [{p_units}]
"""
class PointingMirror(InOutRecordPositioner, OffsetMirror):
"""
Retractable `OffsetMirror`.
Both XRT M1H and XRT M2H can be completely removed from the beam depending
on the beam destination. In this case, the X gantry can be controlled via
the standard PCDS states record. This class has all the functionality of
`OffsetMirror` with the addition of the records that control the
overall state.
Parameters
----------
in_lines : list, optional
List of beamlines that are delivered beam when the mirror is in.
out_lines : list, optional
List of beamlines thate are delivered beam when the mirror is out.
"""
# Reverse state order as PointingMirror is non-standard
states_list = ['OUT', 'IN']
# Moving PointingMirror moves the x gantry
stage_group = [OffsetMirror.xgantry]
def __init__(self, prefix, *, out_lines=None, in_lines=None, **kwargs):
# Branching pattern
self.in_lines = in_lines or list()
self.out_lines = out_lines or list()
super().__init__(prefix, **kwargs)
@property
def destination(self):
"""Current list of destinations the mirror currently supports."""
# Inserted
if self.inserted and not self.removed:
return self.in_lines
# Removed
elif self.removed and not self.inserted:
return self.out_lines
# Unknown
else:
return []
@property
def branches(self):
"""Return all possible beamlines for mirror destinations."""
return self.in_lines + self.out_lines
def check_value(self, pos):
"""Check that our gantry is coupled before state moves."""
# Check the X gantry
if self.xgantry.decoupled.get():
raise PermissionError("Can not move the horizontal gantry is "
"uncoupled")
# Allow StatePositioner to check the state
return super().check_value(pos)
class XOffsetMirror(BaseInterface, GroupDevice):
"""
X-ray Offset Mirror.
1st and 2nd gen Axilon designs with LCLS-II Beckhoff motion architecture.
Parameters
----------
prefix : str
Base PV for the mirror.
name : str
Alias for the device.
"""
# UI representation
_icon = 'fa.minus-square'
# Motor components: can read/write positions
y_up = Cpt(BeckhoffAxisNoOffset, ':MMS:YUP', kind='hinted',
doc='Yupstream master axis [um]')
x_up = Cpt(BeckhoffAxisNoOffset, ':MMS:XUP', kind='hinted',
doc='Xupstream master [um]')
pitch = Cpt(BeckhoffAxisNoOffset, ':MMS:PITCH', kind='hinted',
doc='Pitch stepper and piezo axes [urad]')
bender = Cpt(BeckhoffAxisNoOffset, ':MMS:BENDER', kind='normal',
doc='Bender motor [um]')
y_dwn = Cpt(BeckhoffAxisNoOffset, ':MMS:YDWN', kind='config',
doc='Ydwnstream slave axis [um]')
x_dwn = Cpt(BeckhoffAxisNoOffset, ':MMS:XDWN', kind='config',
doc='Xdwnstream slave axis [um]')
# Gantry components
gantry_x = Cpt(PytmcSignal, ':GANTRY_X', io='i', kind='normal',
doc='X gantry difference [um]')
gantry_y = Cpt(PytmcSignal, ':GANTRY_Y', io='i', kind='normal',
doc='Y gantry difference [um]')
couple_y = Cpt(PytmcSignal, ':COUPLE_Y', io='o', kind='config',
doc='Couple Y motors [bool]')
couple_x = Cpt(PytmcSignal, ':COUPLE_X', io='o', kind='config',
doc='Couple X motors [bool]')
decouple_y = Cpt(PytmcSignal, ':DECOUPLE_Y', io='o', kind='config',
doc='Decouple Y motors [bool]')
decouple_x = Cpt(PytmcSignal, ':DECOUPLE_X', io='o', kind='config',
doc='Decouple X motors [bool]')
couple_status_y = Cpt(PytmcSignal, ':ALREADY_COUPLED_Y', io='i',
kind='normal')
couple_status_x = Cpt(PytmcSignal, ':ALREADY_COUPLED_X', io='i',
kind='normal')
# RMS Cpts:
y_enc_rms = Cpt(PytmcSignal, ':ENC:Y:RMS', io='i', kind='normal',
doc='Yup encoder RMS deviation [um]')
x_enc_rms = Cpt(PytmcSignal, ':ENC:X:RMS', io='i', kind='normal',
doc='Xup encoder RMS deviation [um]')
pitch_enc_rms = Cpt(PytmcSignal, ':ENC:PITCH:RMS', io='i', kind='normal',
doc='Pitch encoder RMS deviation [urad]')
bender_enc_rms = Cpt(PytmcSignal, ':ENC:BENDER:RMS', io='i',
kind='normal',
doc='Bender encoder RMS deviation [um]')
# Lightpath config: implement inserted, removed, transmission, subscribe
# For now, keep it simple. Some mirrors need more than this, but it is
# sufficient for MR1L0 and MR2L0 for today.
inserted = True
removed = False
transmission = 1
SUB_STATE = 'state'
def format_status_info(self, status_info):
"""
Override status info handler to render the Hard X-ray Offset Mirror.
Display homs status info in the ipython terminal.
Parameters
----------
status_info: dict
Nested dictionary. Each level has keys name, kind, and is_device.
If is_device is True, subdevice dictionaries may follow. Otherwise,
the only other key in the dictionary will be value.
Returns
-------
status: str
Formatted string with all relevant status information.
"""
# happi metadata
try:
md = self.root.md
except AttributeError:
name = f'{self.prefix}'
else:
beamline = get_status_value(md, 'beamline')
functional_group = get_status_value(md, 'functional_group')
if functional_group is not None:
name = f'{self.prefix} ({beamline} {functional_group})'
else:
name = f'{self.prefix} ({beamline})'
x_position = get_status_value(status_info, 'x_up', 'position')
x_user_setpoint = get_status_value(status_info, 'x_up',
'user_setpoint', 'value')
x_units = get_status_value(status_info, 'x_up', 'user_setpoint',
'units')
x_description = get_status_value(status_info, 'x_up', 'description',
'value')
p_position = get_status_value(status_info, 'pitch', 'position')
p_user_setpoint = get_status_value(status_info, 'pitch',
'user_setpoint', 'value')
p_units = get_status_value(status_info, 'pitch', 'user_setpoint',
'units')
p_description = get_status_value(status_info, 'pitch', 'description',
'value')
p_enc_rms = get_status_value(status_info, 'pitch_enc_rms', 'value')
return f"""\
{name}
------
x_up: ({self.x_up.prefix})
------
position: {x_position}
user_setpoint: {x_user_setpoint} [{x_units}]
description: {x_description}
------
pitch: ({self.pitch.prefix})
------
position: {p_position}
user_setpoint: {p_user_setpoint} [{p_units}]
description: {p_description}
pitch_enc_rms: {p_enc_rms}
"""
class XOffsetMirrorBend(XOffsetMirror):
"""
X-ray Offset Mirror with 2 bender acutators.
1st and 2nd gen Axilon designs with LCLS-II Beckhoff motion architecture.
Parameters
----------
prefix : str
Base PV for the mirror.
name : str
Alias for the device.
"""
# UI representation
_icon = 'fa.minus-square'
# Do a dumb thing and kill inherited single bender
bender = None
bender_enc_rms = None
# Motor components: can read/write positions
bender_us = Cpt(BeckhoffAxisNoOffset, ':MMS:US', kind='hinted')
bender_ds = Cpt(BeckhoffAxisNoOffset, ':MMS:DS', kind='hinted')
# RMS Cpts:
bender_us_enc_rms = Cpt(PytmcSignal, ':ENC:US:RMS', io='i',
kind='normal')
bender_ds_enc_rms = Cpt(PytmcSignal, ':ENC:DS:RMS', io='i',
kind='normal')
# Bender RTD Cpts:
us_rtd = Cpt(EpicsSignalRO, ':RTD:US:1_RBV', kind='normal')
ds_rtd = Cpt(EpicsSignalRO, ':RTD:DS:1_RBV', kind='normal')
# Maintain backward compatibility
XOffsetMirror2 = XOffsetMirrorBend
class XOffsetMirrorSwitch(XOffsetMirror):
"""
X-ray Offset Mirror with Yleft/Yright
1st and 2nd gen Axilon designs with LCLS-II Beckhoff motion architecture.
Parameters
----------
prefix : str
Base PV for the mirror.
name : str
Alias for the device.
"""
# UI representation
_icon = 'fa.minus-square'
# Do a dumb thing and kill inherited/unused components
y_up = None
y_dwn = None
bender = None
bender_enc_rms = None
# Motor components: can read/write positions
y_left = Cpt(BeckhoffAxisNoOffset, ':MMS:YLEFT', kind='hinted',
doc='Yleft master axis [um]')
y_right = Cpt(BeckhoffAxisNoOffset, ':MMS:YRIGHT', kind='config',
doc='Yright slave axis [um]')
class KBOMirror(BaseInterface, GroupDevice):
"""
Kirkpatrick-Baez Mirror with Bender Axes.
1st gen Toyama designs with LCLS-II Beckhoff motion architecture.
Parameters
----------
prefix : str
Base PV for the mirror.
name : str
Alias for the device.
"""
# UI representation
_icon = 'fa.minus-square'
# Motor components: can read/write positions
x = Cpt(BeckhoffAxisNoOffset, ':MMS:X', kind='hinted')
y = Cpt(BeckhoffAxisNoOffset, ':MMS:Y', kind='hinted')
pitch = Cpt(BeckhoffAxisNoOffset, ':MMS:PITCH', kind='hinted')
bender_us = Cpt(BeckhoffAxisNoOffset, ':MMS:BEND:US', kind='hinted')
bender_ds = Cpt(BeckhoffAxisNoOffset, ':MMS:BEND:DS', kind='hinted')
# RMS Cpts:
x_enc_rms = Cpt(PytmcSignal, ':ENC:X:RMS', io='i', kind='normal')
y_enc_rms = Cpt(PytmcSignal, ':ENC:Y:RMS', io='i', kind='normal')
pitch_enc_rms = Cpt(PytmcSignal, ':ENC:PITCH:RMS', io='i', kind='normal')
bender_us_enc_rms = Cpt(PytmcSignal, ':ENC:BEND:US:RMS', io='i',
kind='normal')
bender_ds_enc_rms = Cpt(PytmcSignal, ':ENC:BEND:DS:RMS', io='i',
kind='normal')
# Bender RTD Cpts:
us_rtd = Cpt(EpicsSignalRO, ':RTD:BEND:US:1_RBV', kind='normal')
ds_rtd = Cpt(EpicsSignalRO, ':RTD:BEND:DS:1_RBV', kind='normal')
# Lightpath config: implement inserted, removed, transmission, subscribe
inserted = True
removed = False
transmission = 1
SUB_STATE = 'state'
def format_status_info(self, status_info):
"""
Override status info handler to render the `KBOMirror`.
Display `KBOMirror` status info in the ipython terminal.
Parameters
----------
status_info: dict
Nested dictionary. Each level has keys name, kind, and is_device.
If is_device is True, subdevice dictionaries may follow. Otherwise,
the only other key in the dictionary will be value.
Returns
-------
status: str
Formatted string with all relevant status information.
"""
# happi metadata
try:
md = self.root.md
except AttributeError:
name = f'{self.prefix}'
else:
beamline = get_status_value(md, 'beamline')
functional_group = get_status_value(md, 'functional_group')
if functional_group is not None:
name = f'{self.prefix} ({beamline} {functional_group})'
else:
name = f'{self.prefix} ({beamline})'
x_position = get_status_value(status_info, 'x', 'position')
x_user_setpoint = get_status_value(status_info, 'x',
'user_setpoint', 'value')
x_units = get_status_value(status_info, 'x', 'user_setpoint',
'units')
x_description = get_status_value(status_info, 'x', 'description',
'value')
p_position = get_status_value(status_info, 'pitch', 'position')
p_user_setpoint = get_status_value(status_info, 'pitch',
'user_setpoint', 'value')
p_units = get_status_value(status_info, 'pitch', 'user_setpoint',
'units')
p_description = get_status_value(status_info, 'pitch', 'description',
'value')
p_enc_rms = get_status_value(status_info, 'pitch_enc_rms', 'value')
b_us_position = get_status_value(status_info, 'bender_us', 'position')
b_us_setpoint = get_status_value(status_info, 'bender_us',
'user_setpoint', 'value')
b_us_units = get_status_value(status_info, 'bender_us',
'user_setpoint', 'units')
b_us_description = get_status_value(status_info, 'bender_us',
'description', 'value')
b_us_enc_rms = get_status_value(status_info, 'bender_us_enc_rms',
'value')
b_ds_position = get_status_value(status_info, 'bender_ds', 'position')
b_ds_setpoint = get_status_value(status_info, 'bender_ds',
'user_setpoint', 'value')
b_ds_units = get_status_value(status_info, 'bender_ds',
'user_setpoint', 'units')
b_ds_description = get_status_value(status_info, 'bender_ds',
'description', 'value')
b_ds_enc_rms = get_status_value(status_info, 'bender_ds_enc_rms',
'value')
return f"""\
{name}
------
x_up: ({self.x.prefix})
------
position: {x_position}
user_setpoint: {x_user_setpoint} [{x_units}]
description: {x_description}
------
pitch: ({self.pitch.prefix})
------
position: {p_position}
user_setpoint: {p_user_setpoint} [{p_units}]
description: {p_description}
pitch_enc_rms: {p_enc_rms}
---------
bender_us ({self.bender_us.prefix})
---------
position {b_us_position}
user_setpoint: {b_us_setpoint} [{b_us_units}]
description: {b_us_description}
bender_us_enc_rms: {b_us_enc_rms}
---------
bender_ds ({self.bender_ds.prefix})
---------
position: {b_ds_position}
user_setpoint: {b_ds_setpoint} [{b_ds_units}]
description: {b_ds_description}
bender_ds_enc_rms: {b_ds_enc_rms}
"""
class FFMirror(BaseInterface, GroupDevice):
"""
Fixed Focus Kirkpatrick-Baez Mirror.
1st gen Toyama designs with LCLS-II Beckhoff motion architecture.
Parameters
----------
prefix : str
Base PV for the mirror.
name : str
Alias for the device.
"""
# UI representation
_icon = 'fa.minus-square'
# Motor components: can read/write positions
x = Cpt(BeckhoffAxisNoOffset, ':MMS:X', kind='hinted')
y = Cpt(BeckhoffAxisNoOffset, ':MMS:Y', kind='hinted')
pitch = Cpt(BeckhoffAxisNoOffset, ':MMS:PITCH', kind='hinted')
# RMS Cpts:
x_enc_rms = Cpt(PytmcSignal, ':ENC:X:RMS', io='i', kind='normal')
y_enc_rms = Cpt(PytmcSignal, ':ENC:Y:RMS', io='i', kind='normal')
pitch_enc_rms = Cpt(PytmcSignal, ':ENC:PITCH:RMS', io='i', kind='normal')
# Lightpath config: implement inserted, removed, transmission, subscribe
inserted = True
removed = False
transmission = 1
SUB_STATE = 'state'
def format_status_info(self, status_info):
"""
Override status info handler to render the `FFMirror`.
Display `FFMirror` status info in the ipython terminal.
Parameters
----------
status_info: dict
Nested dictionary. Each level has keys name, kind, and is_device.
If is_device is True, subdevice dictionaries may follow. Otherwise,
the only other key in the dictionary will be value.
Returns
-------
status: str
Formatted string with all relevant status information.
"""
# happi metadata
try:
md = self.root.md
except AttributeError:
name = f'{self.prefix}'
else:
beamline = get_status_value(md, 'beamline')
functional_group = get_status_value(md, 'functional_group')
if functional_group is not None:
name = f'{self.prefix} ({beamline} {functional_group})'
else:
name = f'{self.prefix} ({beamline})'
x_position = get_status_value(status_info, 'x', 'position')
x_user_setpoint = get_status_value(status_info, 'x',
'user_setpoint', 'value')
x_units = get_status_value(status_info, 'x', 'user_setpoint',
'units')
x_description = get_status_value(status_info, 'x', 'description',
'value')
p_position = get_status_value(status_info, 'pitch', 'position')
p_user_setpoint = get_status_value(status_info, 'pitch',
'user_setpoint', 'value')
p_units = get_status_value(status_info, 'pitch', 'user_setpoint',
'units')
p_description = get_status_value(status_info, 'pitch', 'description',
'value')
p_enc_rms = get_status_value(status_info, 'pitch_enc_rms', 'value')
return f"""\
{name}
------
x_up: ({self.x.prefix})
------
position: {x_position}
user_setpoint: {x_user_setpoint} [{x_units}]
description: {x_description}
------
pitch: ({self.pitch.prefix})
------
position: {p_position}
user_setpoint: {p_user_setpoint} [{p_units}]
description: {p_description}
pitch_enc_rms: {p_enc_rms}
"""
class TwinCATMirrorStripe(TwinCATStatePMPS):
"""
Subclass of TwinCATStatePMPS for the mirror coatings.
Unless most TwinCATStatePMPS, we have:
- Only in_states
- No in_states block the beam
We also clear the states_list and set _in_if_not_out to True
to automatically pick up the coatings from each mirror enum.
"""
states_list = []
in_states = []
out_states = []
_in_if_not_out = True
@property
def transmission(self):
"""The mirror coating never blocks the beam."""
return 1
class CoatingState(Device):
"""
Extra parent class to put "coating" as the first device in order.
This makes it appear at the top of the screen in typhos.
"""
coating = Cpt(TwinCATMirrorStripe, ':COATING:STATE', kind='hinted',
doc='Control of the coating states via saved positions.')
class XOffsetMirrorState(XOffsetMirror, CoatingState):
"""
X-ray Offset Mirror with Yleft/Yright
1st and 2nd gen Axilon designs with LCLS-II Beckhoff motion architecture.
With Coating State selection implemented
Parameters
----------
prefix : str
Base PV for the mirror.
name : str
Alias for the device.
"""
# UI representation
_icon = 'fa.minus-square'
|
[
"logging.getLogger",
"ophyd.Component",
"numpy.isnan",
"numpy.isinf",
"ophyd.FormattedComponent"
] |
[((810, 837), 'logging.getLogger', 'logging.getLogger', (['__name__'], {}), '(__name__)\n', (827, 837), False, 'import logging\n'), ((1023, 1083), 'ophyd.Component', 'Cpt', (['EpicsSignalRO', '""":RBV"""'], {'auto_monitor': '(True)', 'kind': '"""hinted"""'}), "(EpicsSignalRO, ':RBV', auto_monitor=True, kind='hinted')\n", (1026, 1083), True, 'from ophyd import Component as Cpt\n'), ((1099, 1170), 'ophyd.Component', 'Cpt', (['EpicsSignal', '""":VAL"""'], {'auto_monitor': '(True)', 'limits': '(True)', 'kind': '"""normal"""'}), "(EpicsSignal, ':VAL', auto_monitor=True, limits=True, kind='normal')\n", (1102, 1170), True, 'from ophyd import Component as Cpt\n'), ((1201, 1263), 'ophyd.Component', 'Cpt', (['EpicsSignalRO', '""":DMOV"""'], {'auto_monitor': '(True)', 'kind': '"""omitted"""'}), "(EpicsSignalRO, ':DMOV', auto_monitor=True, kind='omitted')\n", (1204, 1263), True, 'from ophyd import Component as Cpt\n'), ((1280, 1324), 'ophyd.Component', 'Cpt', (['EpicsSignal', '""":RBV.EGU"""'], {'kind': '"""omitted"""'}), "(EpicsSignal, ':RBV.EGU', kind='omitted')\n", (1283, 1324), True, 'from ophyd import Component as Cpt\n'), ((1355, 1403), 'ophyd.Component', 'Cpt', (['EpicsSignalRO', '""":INTERLOCK"""'], {'kind': '"""omitted"""'}), "(EpicsSignalRO, ':INTERLOCK', kind='omitted')\n", (1358, 1403), True, 'from ophyd import Component as Cpt\n'), ((1418, 1464), 'ophyd.Component', 'Cpt', (['EpicsSignalRO', '""":ENABLED"""'], {'kind': '"""omitted"""'}), "(EpicsSignalRO, ':ENABLED', kind='omitted')\n", (1421, 1464), True, 'from ophyd import Component as Cpt\n'), ((1509, 1551), 'ophyd.Component', 'Cpt', (['EpicsSignalRO', '""":LLS"""'], {'kind': '"""omitted"""'}), "(EpicsSignalRO, ':LLS', kind='omitted')\n", (1512, 1551), True, 'from ophyd import Component as Cpt\n'), ((1576, 1618), 'ophyd.Component', 'Cpt', (['EpicsSignalRO', '""":HLS"""'], {'kind': '"""omitted"""'}), "(EpicsSignalRO, ':HLS', kind='omitted')\n", (1579, 1618), True, 'from ophyd import Component as Cpt\n'), ((2829, 2885), 'ophyd.FormattedComponent', 'FCpt', (['EpicsSignalRO', '"""{self._piezo}:VRBV"""'], {'kind': '"""normal"""'}), "(EpicsSignalRO, '{self._piezo}:VRBV', kind='normal')\n", (2833, 2885), True, 'from ophyd import FormattedComponent as FCpt\n'), ((2904, 2959), 'ophyd.FormattedComponent', 'FCpt', (['EpicsSignal', '"""{self._piezo}:STOP"""'], {'kind': '"""omitted"""'}), "(EpicsSignal, '{self._piezo}:STOP', kind='omitted')\n", (2908, 2959), True, 'from ophyd import FormattedComponent as FCpt\n'), ((3939, 4002), 'ophyd.FormattedComponent', 'FCpt', (['EpicsSignalRO', '"""{self.gantry_prefix}:GDIF"""'], {'kind': '"""normal"""'}), "(EpicsSignalRO, '{self.gantry_prefix}:GDIF', kind='normal')\n", (3943, 4002), True, 'from ophyd import FormattedComponent as FCpt\n'), ((4048, 4115), 'ophyd.FormattedComponent', 'FCpt', (['EpicsSignalRO', '"""{self.gantry_prefix}:DECOUPLE"""'], {'kind': '"""config"""'}), "(EpicsSignalRO, '{self.gantry_prefix}:DECOUPLE', kind='config')\n", (4052, 4115), True, 'from ophyd import FormattedComponent as FCpt\n'), ((4201, 4263), 'ophyd.FormattedComponent', 'FCpt', (['EpicsSignalRO', '"""{self.follow_prefix}:RBV"""'], {'kind': '"""normal"""'}), "(EpicsSignalRO, '{self.follow_prefix}:RBV', kind='normal')\n", (4205, 4263), True, 'from ophyd import FormattedComponent as FCpt\n'), ((4325, 4388), 'ophyd.FormattedComponent', 'FCpt', (['EpicsSignalRO', '"""{self.follow_prefix}:LLS"""'], {'kind': '"""omitted"""'}), "(EpicsSignalRO, '{self.follow_prefix}:LLS', kind='omitted')\n", (4329, 4388), True, 'from ophyd import FormattedComponent as FCpt\n'), ((4459, 4522), 'ophyd.FormattedComponent', 'FCpt', (['EpicsSignalRO', '"""{self.follow_prefix}:HLS"""'], {'kind': '"""omitted"""'}), "(EpicsSignalRO, '{self.follow_prefix}:HLS', kind='omitted')\n", (4463, 4522), True, 'from ophyd import FormattedComponent as FCpt\n'), ((6229, 6277), 'ophyd.FormattedComponent', 'FCpt', (['Pitch', '"""MIRR:{self.prefix}"""'], {'kind': '"""hinted"""'}), "(Pitch, 'MIRR:{self.prefix}', kind='hinted')\n", (6233, 6277), True, 'from ophyd import FormattedComponent as FCpt\n'), ((6312, 6439), 'ophyd.FormattedComponent', 'FCpt', (['Gantry', '"""{self._prefix_xy}:X"""'], {'gantry_prefix': '"""{self._xgantry}"""', 'add_prefix': "['suffix', 'gantry_prefix']", 'kind': '"""normal"""'}), "(Gantry, '{self._prefix_xy}:X', gantry_prefix='{self._xgantry}',\n add_prefix=['suffix', 'gantry_prefix'], kind='normal')\n", (6316, 6439), True, 'from ophyd import FormattedComponent as FCpt\n'), ((6507, 6641), 'ophyd.FormattedComponent', 'FCpt', (['Gantry', '"""{self._prefix_xy}:Y"""'], {'gantry_prefix': '"""GANTRY:{self.prefix}:Y"""', 'add_prefix': "['suffix', 'gantry_prefix']", 'kind': '"""config"""'}), "(Gantry, '{self._prefix_xy}:Y', gantry_prefix='GANTRY:{self.prefix}:Y',\n add_prefix=['suffix', 'gantry_prefix'], kind='config')\n", (6511, 6641), True, 'from ophyd import FormattedComponent as FCpt\n'), ((11405, 11496), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:YUP"""'], {'kind': '"""hinted"""', 'doc': '"""Yupstream master axis [um]"""'}), "(BeckhoffAxisNoOffset, ':MMS:YUP', kind='hinted', doc=\n 'Yupstream master axis [um]')\n", (11408, 11496), True, 'from ophyd import Component as Cpt\n'), ((11518, 11604), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:XUP"""'], {'kind': '"""hinted"""', 'doc': '"""Xupstream master [um]"""'}), "(BeckhoffAxisNoOffset, ':MMS:XUP', kind='hinted', doc=\n 'Xupstream master [um]')\n", (11521, 11604), True, 'from ophyd import Component as Cpt\n'), ((11627, 11729), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:PITCH"""'], {'kind': '"""hinted"""', 'doc': '"""Pitch stepper and piezo axes [urad]"""'}), "(BeckhoffAxisNoOffset, ':MMS:PITCH', kind='hinted', doc=\n 'Pitch stepper and piezo axes [urad]')\n", (11630, 11729), True, 'from ophyd import Component as Cpt\n'), ((11754, 11839), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:BENDER"""'], {'kind': '"""normal"""', 'doc': '"""Bender motor [um]"""'}), "(BeckhoffAxisNoOffset, ':MMS:BENDER', kind='normal', doc='Bender motor [um]'\n )\n", (11757, 11839), True, 'from ophyd import Component as Cpt\n'), ((11864, 11956), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:YDWN"""'], {'kind': '"""config"""', 'doc': '"""Ydwnstream slave axis [um]"""'}), "(BeckhoffAxisNoOffset, ':MMS:YDWN', kind='config', doc=\n 'Ydwnstream slave axis [um]')\n", (11867, 11956), True, 'from ophyd import Component as Cpt\n'), ((11980, 12072), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:XDWN"""'], {'kind': '"""config"""', 'doc': '"""Xdwnstream slave axis [um]"""'}), "(BeckhoffAxisNoOffset, ':MMS:XDWN', kind='config', doc=\n 'Xdwnstream slave axis [um]')\n", (11983, 12072), True, 'from ophyd import Component as Cpt\n'), ((12124, 12213), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":GANTRY_X"""'], {'io': '"""i"""', 'kind': '"""normal"""', 'doc': '"""X gantry difference [um]"""'}), "(PytmcSignal, ':GANTRY_X', io='i', kind='normal', doc=\n 'X gantry difference [um]')\n", (12127, 12213), True, 'from ophyd import Component as Cpt\n'), ((12243, 12332), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":GANTRY_Y"""'], {'io': '"""i"""', 'kind': '"""normal"""', 'doc': '"""Y gantry difference [um]"""'}), "(PytmcSignal, ':GANTRY_Y', io='i', kind='normal', doc=\n 'Y gantry difference [um]')\n", (12246, 12332), True, 'from ophyd import Component as Cpt\n'), ((12362, 12449), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":COUPLE_Y"""'], {'io': '"""o"""', 'kind': '"""config"""', 'doc': '"""Couple Y motors [bool]"""'}), "(PytmcSignal, ':COUPLE_Y', io='o', kind='config', doc=\n 'Couple Y motors [bool]')\n", (12365, 12449), True, 'from ophyd import Component as Cpt\n'), ((12479, 12566), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":COUPLE_X"""'], {'io': '"""o"""', 'kind': '"""config"""', 'doc': '"""Couple X motors [bool]"""'}), "(PytmcSignal, ':COUPLE_X', io='o', kind='config', doc=\n 'Couple X motors [bool]')\n", (12482, 12566), True, 'from ophyd import Component as Cpt\n'), ((12598, 12689), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":DECOUPLE_Y"""'], {'io': '"""o"""', 'kind': '"""config"""', 'doc': '"""Decouple Y motors [bool]"""'}), "(PytmcSignal, ':DECOUPLE_Y', io='o', kind='config', doc=\n 'Decouple Y motors [bool]')\n", (12601, 12689), True, 'from ophyd import Component as Cpt\n'), ((12723, 12814), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":DECOUPLE_X"""'], {'io': '"""o"""', 'kind': '"""config"""', 'doc': '"""Decouple X motors [bool]"""'}), "(PytmcSignal, ':DECOUPLE_X', io='o', kind='config', doc=\n 'Decouple X motors [bool]')\n", (12726, 12814), True, 'from ophyd import Component as Cpt\n'), ((12853, 12914), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ALREADY_COUPLED_Y"""'], {'io': '"""i"""', 'kind': '"""normal"""'}), "(PytmcSignal, ':ALREADY_COUPLED_Y', io='i', kind='normal')\n", (12856, 12914), True, 'from ophyd import Component as Cpt\n'), ((12963, 13024), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ALREADY_COUPLED_X"""'], {'io': '"""i"""', 'kind': '"""normal"""'}), "(PytmcSignal, ':ALREADY_COUPLED_X', io='i', kind='normal')\n", (12966, 13024), True, 'from ophyd import Component as Cpt\n'), ((13083, 13179), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:Y:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""', 'doc': '"""Yup encoder RMS deviation [um]"""'}), "(PytmcSignal, ':ENC:Y:RMS', io='i', kind='normal', doc=\n 'Yup encoder RMS deviation [um]')\n", (13086, 13179), True, 'from ophyd import Component as Cpt\n'), ((13211, 13307), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:X:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""', 'doc': '"""Xup encoder RMS deviation [um]"""'}), "(PytmcSignal, ':ENC:X:RMS', io='i', kind='normal', doc=\n 'Xup encoder RMS deviation [um]')\n", (13214, 13307), True, 'from ophyd import Component as Cpt\n'), ((13343, 13447), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:PITCH:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""', 'doc': '"""Pitch encoder RMS deviation [urad]"""'}), "(PytmcSignal, ':ENC:PITCH:RMS', io='i', kind='normal', doc=\n 'Pitch encoder RMS deviation [urad]')\n", (13346, 13447), True, 'from ophyd import Component as Cpt\n'), ((13488, 13592), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:BENDER:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""', 'doc': '"""Bender encoder RMS deviation [um]"""'}), "(PytmcSignal, ':ENC:BENDER:RMS', io='i', kind='normal', doc=\n 'Bender encoder RMS deviation [um]')\n", (13491, 13592), True, 'from ophyd import Component as Cpt\n'), ((16880, 16931), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:US"""'], {'kind': '"""hinted"""'}), "(BeckhoffAxisNoOffset, ':MMS:US', kind='hinted')\n", (16883, 16931), True, 'from ophyd import Component as Cpt\n'), ((16948, 16999), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:DS"""'], {'kind': '"""hinted"""'}), "(BeckhoffAxisNoOffset, ':MMS:DS', kind='hinted')\n", (16951, 16999), True, 'from ophyd import Component as Cpt\n'), ((17041, 17095), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:US:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""'}), "(PytmcSignal, ':ENC:US:RMS', io='i', kind='normal')\n", (17044, 17095), True, 'from ophyd import Component as Cpt\n'), ((17148, 17202), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:DS:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""'}), "(PytmcSignal, ':ENC:DS:RMS', io='i', kind='normal')\n", (17151, 17202), True, 'from ophyd import Component as Cpt\n'), ((17268, 17318), 'ophyd.Component', 'Cpt', (['EpicsSignalRO', '""":RTD:US:1_RBV"""'], {'kind': '"""normal"""'}), "(EpicsSignalRO, ':RTD:US:1_RBV', kind='normal')\n", (17271, 17318), True, 'from ophyd import Component as Cpt\n'), ((17332, 17382), 'ophyd.Component', 'Cpt', (['EpicsSignalRO', '""":RTD:DS:1_RBV"""'], {'kind': '"""normal"""'}), "(EpicsSignalRO, ':RTD:DS:1_RBV', kind='normal')\n", (17335, 17382), True, 'from ophyd import Component as Cpt\n'), ((18015, 18104), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:YLEFT"""'], {'kind': '"""hinted"""', 'doc': '"""Yleft master axis [um]"""'}), "(BeckhoffAxisNoOffset, ':MMS:YLEFT', kind='hinted', doc=\n 'Yleft master axis [um]')\n", (18018, 18104), True, 'from ophyd import Component as Cpt\n'), ((18131, 18221), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:YRIGHT"""'], {'kind': '"""config"""', 'doc': '"""Yright slave axis [um]"""'}), "(BeckhoffAxisNoOffset, ':MMS:YRIGHT', kind='config', doc=\n 'Yright slave axis [um]')\n", (18134, 18221), True, 'from ophyd import Component as Cpt\n'), ((18653, 18703), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:X"""'], {'kind': '"""hinted"""'}), "(BeckhoffAxisNoOffset, ':MMS:X', kind='hinted')\n", (18656, 18703), True, 'from ophyd import Component as Cpt\n'), ((18712, 18762), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:Y"""'], {'kind': '"""hinted"""'}), "(BeckhoffAxisNoOffset, ':MMS:Y', kind='hinted')\n", (18715, 18762), True, 'from ophyd import Component as Cpt\n'), ((18775, 18829), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:PITCH"""'], {'kind': '"""hinted"""'}), "(BeckhoffAxisNoOffset, ':MMS:PITCH', kind='hinted')\n", (18778, 18829), True, 'from ophyd import Component as Cpt\n'), ((18846, 18902), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:BEND:US"""'], {'kind': '"""hinted"""'}), "(BeckhoffAxisNoOffset, ':MMS:BEND:US', kind='hinted')\n", (18849, 18902), True, 'from ophyd import Component as Cpt\n'), ((18919, 18975), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:BEND:DS"""'], {'kind': '"""hinted"""'}), "(BeckhoffAxisNoOffset, ':MMS:BEND:DS', kind='hinted')\n", (18922, 18975), True, 'from ophyd import Component as Cpt\n'), ((19009, 19062), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:X:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""'}), "(PytmcSignal, ':ENC:X:RMS', io='i', kind='normal')\n", (19012, 19062), True, 'from ophyd import Component as Cpt\n'), ((19079, 19132), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:Y:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""'}), "(PytmcSignal, ':ENC:Y:RMS', io='i', kind='normal')\n", (19082, 19132), True, 'from ophyd import Component as Cpt\n'), ((19153, 19210), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:PITCH:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""'}), "(PytmcSignal, ':ENC:PITCH:RMS', io='i', kind='normal')\n", (19156, 19210), True, 'from ophyd import Component as Cpt\n'), ((19235, 19294), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:BEND:US:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""'}), "(PytmcSignal, ':ENC:BEND:US:RMS', io='i', kind='normal')\n", (19238, 19294), True, 'from ophyd import Component as Cpt\n'), ((19347, 19406), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:BEND:DS:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""'}), "(PytmcSignal, ':ENC:BEND:DS:RMS', io='i', kind='normal')\n", (19350, 19406), True, 'from ophyd import Component as Cpt\n'), ((19472, 19527), 'ophyd.Component', 'Cpt', (['EpicsSignalRO', '""":RTD:BEND:US:1_RBV"""'], {'kind': '"""normal"""'}), "(EpicsSignalRO, ':RTD:BEND:US:1_RBV', kind='normal')\n", (19475, 19527), True, 'from ophyd import Component as Cpt\n'), ((19541, 19596), 'ophyd.Component', 'Cpt', (['EpicsSignalRO', '""":RTD:BEND:DS:1_RBV"""'], {'kind': '"""normal"""'}), "(EpicsSignalRO, ':RTD:BEND:DS:1_RBV', kind='normal')\n", (19544, 19596), True, 'from ophyd import Component as Cpt\n'), ((24196, 24246), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:X"""'], {'kind': '"""hinted"""'}), "(BeckhoffAxisNoOffset, ':MMS:X', kind='hinted')\n", (24199, 24246), True, 'from ophyd import Component as Cpt\n'), ((24255, 24305), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:Y"""'], {'kind': '"""hinted"""'}), "(BeckhoffAxisNoOffset, ':MMS:Y', kind='hinted')\n", (24258, 24305), True, 'from ophyd import Component as Cpt\n'), ((24318, 24372), 'ophyd.Component', 'Cpt', (['BeckhoffAxisNoOffset', '""":MMS:PITCH"""'], {'kind': '"""hinted"""'}), "(BeckhoffAxisNoOffset, ':MMS:PITCH', kind='hinted')\n", (24321, 24372), True, 'from ophyd import Component as Cpt\n'), ((24406, 24459), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:X:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""'}), "(PytmcSignal, ':ENC:X:RMS', io='i', kind='normal')\n", (24409, 24459), True, 'from ophyd import Component as Cpt\n'), ((24476, 24529), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:Y:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""'}), "(PytmcSignal, ':ENC:Y:RMS', io='i', kind='normal')\n", (24479, 24529), True, 'from ophyd import Component as Cpt\n'), ((24550, 24607), 'ophyd.Component', 'Cpt', (['PytmcSignal', '""":ENC:PITCH:RMS"""'], {'io': '"""i"""', 'kind': '"""normal"""'}), "(PytmcSignal, ':ENC:PITCH:RMS', io='i', kind='normal')\n", (24553, 24607), True, 'from ophyd import Component as Cpt\n'), ((27917, 28037), 'ophyd.Component', 'Cpt', (['TwinCATMirrorStripe', '""":COATING:STATE"""'], {'kind': '"""hinted"""', 'doc': '"""Control of the coating states via saved positions."""'}), "(TwinCATMirrorStripe, ':COATING:STATE', kind='hinted', doc=\n 'Control of the coating states via saved positions.')\n", (27920, 28037), True, 'from ophyd import Component as Cpt\n'), ((2377, 2395), 'numpy.isnan', 'np.isnan', (['position'], {}), '(position)\n', (2385, 2395), True, 'import numpy as np\n'), ((2399, 2417), 'numpy.isinf', 'np.isinf', (['position'], {}), '(position)\n', (2407, 2417), True, 'import numpy as np\n')]
|
# Copyright (c) 2020 <NAME>
from baselines.common import Dataset, explained_variance, fmt_row, zipsame
from baselines import logger
import baselines.common.tf_util as U
import tensorflow as tf, numpy as np
import time
from baselines.common.mpi_adam import MpiAdam
from baselines.common.mpi_moments import mpi_moments
from mpi4py import MPI
from collections import deque
import pdb
import os
import shutil
from scipy import spatial
import gym
def traj_segment_generator(pi, env, horizon, stochastic, num_options,saves,results,rewbuffer,dc):
# sample state action pairs, i.e. sample rollouts on the real system
max_action = env.action_space.high
t = 0
glob_count = 0
glob_count_thresh = -1
ac = env.action_space.sample() # not used, just so we have the datatype
new = True # marks if we're on first timestep of an episode
ob = env.reset()
ob_env_shape = np.shape(ob)
ac_env_shape = np.shape(ac)
ac = pi.reset_last_act().eval()
ob = np.concatenate((ob,ac))
cur_ep_ret = 0 # return in current episode
cur_ep_len = 0 # len of current episode
ep_rets = [] # returns of completed episodes in this segment
ep_lens = [] # lengths of ...
# Initialize history arrays
obs = np.array([ob for _ in range(horizon)])
rews = np.zeros(horizon, 'float32')
realrews = np.zeros(horizon, 'float32')
vpreds = np.zeros(horizon, 'float32')
news = np.zeros(horizon, 'int32')
opts = np.zeros(horizon, 'int32')
acs = np.array([ac for _ in range(horizon)])
prevacs = acs.copy()
option = pi.get_option(ob)
if (glob_count<glob_count_thresh):
option = 1
optpol_p=[]
term_p=[]
value_val=[]
opt_duration = [[] for _ in range(num_options)]
logstds = [[] for _ in range(num_options)]
curr_opt_duration = 0.
while True:
# in here collect the state action pairs:
prevac = ac
# remember u[k-1]
ob[ob_env_shape[0]:] = ac
# evaluate policy and recieve action
ac, vpred, feats,logstd = pi.act(stochastic, ob, option)
logstds[option].append(logstd)
# Slight weirdness here because we need value function at time T
# before returning segment [0, T-1] so we get the correct
# terminal value
if t > 0 and t % horizon == 0:
yield {"ob" : obs, "rew" : rews, "realrew": realrews, "vpred" : vpreds, "new" : news,
"ac" : acs, "opts" : opts, "prevac" : prevacs, "nextvpred": vpred * (1 - new),
"ep_rets" : ep_rets, "ep_lens" : ep_lens, 'term_p': term_p, 'value_val': value_val,
"opt_dur": opt_duration, "optpol_p":optpol_p, "logstds": logstds}
# Be careful!!! if you change the downstream algorithm to aggregate
# several of these batches, then be sure to do a deepcopy
ep_rets = []
ep_lens = []
term_p = []
value_val=[]
opt_duration = [[] for _ in range(num_options)]
logstds = [[] for _ in range(num_options)]
curr_opt_duration = 0.
glob_count += 1
i = t % horizon
obs[i] = ob
vpreds[i] = vpred
news[i] = new
opts[i] = option
acs[i] = ac
prevacs[i] = prevac
# Careful: Without this "copy" operation the variable ac is actually modified...
# Apply the action to the environment
ob[:ob_env_shape[0]], rew, new, _ = env.step(max_action*np.copy(ac))
# IMPORTANT: here there is no triggering decision
rew = rew*1.0
rew = rew/10 if num_options > 1 else rew # To stabilize learning.
rews[i] = rew
realrews[i] = rew
curr_opt_duration += 1
### Book-keeping
t_p = []
v_val = []
for oopt in range(num_options):
v_val.append(pi.get_vpred([ob],[oopt])[0][0])
t_p.append(pi.get_tpred([ob],[oopt])[0][0])
term_p.append(t_p)
optpol_p.append(pi._get_op([ob])[0][0])
value_val.append(v_val)
term = pi.get_term([ob],[option])[0][0]
# in case of termination, decide which option to execute next:
if term:
opt_duration[option].append(curr_opt_duration)
curr_opt_duration = 0.
option = pi.get_option(ob)
if (glob_count<glob_count_thresh):
option = 1
cur_ep_ret += rew*10 if num_options > 1 else rew
cur_ep_len += 1
if new:
# if new rollout starts -> reset last action and start anew
ep_rets.append(cur_ep_ret)
ep_lens.append(cur_ep_len)
cur_ep_ret = 0
cur_ep_len = 0
ob[:ob_env_shape[0]] = env.reset()
ob[ob_env_shape[0]:] = pi.reset_last_act().eval()
ac = pi.reset_last_act().eval()
option = pi.get_option(ob)
if (glob_count<glob_count_thresh):
option = 1
t += 1
def add_vtarg_and_adv(seg, gamma, lam):
"""
Compute target value using TD(lambda) estimator, and advantage with GAE(lambda)
"""
new = np.append(seg["new"], 0) # last element is only used for last vtarg, but we already zeroed it if last new = 1
vpred = np.append(seg["vpred"], seg["nextvpred"])
T = len(seg["rew"])
seg["adv"] = gaelam = np.empty(T, 'float32')
rew = seg["rew"]
lastgaelam = 0
for t in reversed(range(T)):
nonterminal = 1-new[t+1]
delta = rew[t] + gamma * vpred[t+1] * nonterminal - vpred[t]
gaelam[t] = lastgaelam = delta + gamma * lam * nonterminal * lastgaelam
seg["tdlamret"] = seg["adv"] + seg["vpred"]
def learn(env, policy_func, *,
timesteps_per_batch, # timesteps per actor per update
clip_param, entcoeff, # clipping parameter epsilon, entropy coeff
optim_epochs, optim_stepsize, optim_batchsize,# optimization hypers
gamma, lam, # advantage estimation
max_timesteps=0, max_episodes=0, max_iters=0, max_seconds=0, # time constraint
callback=None, # you can do anything in the callback, since it takes locals(), globals()
adam_epsilon=1e-5,
schedule='constant', # annealing for stepsize parameters (epsilon and adam)
num_options=1,
app='',
saves=False,
wsaves=False,
epoch=-1,
seed=1,
dc=0
):
optim_batchsize_ideal = optim_batchsize
np.random.seed(seed)
tf.set_random_seed(seed)
env.seed(seed)
### Book-keeping
gamename = env.spec.id[:-3].lower()
gamename += 'seed' + str(seed)
gamename += app
# This variable: "version name, defines the name of the training"
version_name = 'NORM-ACT-LOWER-LR-len-400-wNoise-update1-ppo-ESCH-1-own-impl-both-equal'
dirname = '{}_{}_{}opts_saves/'.format(version_name,gamename,num_options)
print (dirname)
# retrieve everything using relative paths. Create a train_results folder where the repo has been cloned
dirname_rel = os.path.dirname(__file__)
splitted = dirname_rel.split("/")
dirname_rel = ("/".join(dirname_rel.split("/")[:len(splitted)-3])+"/")
dirname = dirname_rel + "train_results/" + dirname
# if saving -> create the necessary directories
if wsaves:
first=True
if not os.path.exists(dirname):
os.makedirs(dirname)
first = False
# copy also the original files into the folder where the training results are stored
files = ['pposgd_simple.py','mlp_policy.py','run_mujoco.py']
first = True
for i in range(len(files)):
src = os.path.join(dirname_rel,'baselines/baselines/ppo1/') + files[i]
print (src)
#dest = os.path.join('/home/nfunk/results_NEW/ppo1/') + dirname
dest = dirname + "src_code/"
if (first):
os.makedirs(dest)
first = False
print (dest)
shutil.copy2(src,dest)
# brute force copy normal env file at end of copying process:
src = os.path.join(dirname_rel,'nfunk/envs_nf/pendulum_nf.py')
shutil.copy2(src,dest)
shutil.copy2(src,dest)
os.makedirs(dest+"assets/")
src = os.path.join(dirname_rel,'nfunk/envs_nf/assets/clockwise.png')
shutil.copy2(src,dest+"assets/")
###
# Setup losses and stuff
# ----------------------------------------
ob_space = env.observation_space
ac_space = env.action_space
max_action = env.action_space.high
# add the dimension in the observation space!
ob_space.shape =((ob_space.shape[0] + ac_space.shape[0]),)
print (ob_space.shape)
print (ac_space.shape)
pi = policy_func("pi", ob_space, ac_space) # Construct network for new policy
oldpi = policy_func("oldpi", ob_space, ac_space) # Network for old policy
atarg = tf.placeholder(dtype=tf.float32, shape=[None]) # Target advantage function
ret = tf.placeholder(dtype=tf.float32, shape=[None]) # Empirical return
pol_ov_op_ent = tf.placeholder(dtype=tf.float32, shape=None) # Entropy coefficient for policy over options
lrmult = tf.placeholder(name='lrmult', dtype=tf.float32, shape=[]) # learning rate multiplier, updated with schedule
clip_param = clip_param * lrmult # Annealed cliping parameter epislon for PPO
# setup observation, option and terminal advantace
ob = U.get_placeholder_cached(name="ob")
option = U.get_placeholder_cached(name="option")
term_adv = U.get_placeholder(name='term_adv', dtype=tf.float32, shape=[None])
# create variable for action
ac = pi.pdtype.sample_placeholder([None])
kloldnew = oldpi.pd.kl(pi.pd)
ent = pi.pd.entropy()
meankl = U.mean(kloldnew)
meanent = U.mean(ent)
pol_entpen = (-entcoeff) * meanent
# propability of choosing action under new policy vs old policy (PPO)
ratio = tf.exp(pi.pd.logp(ac) - oldpi.pd.logp(ac))
# advantage of choosing the action
atarg_clip = atarg
# surrogate 1:
surr1 = ratio * atarg_clip #atarg # surrogate from conservative policy iteration
# surrogate 2:
surr2 = U.clip(ratio, 1.0 - clip_param, 1.0 + clip_param) * atarg_clip
# PPO's pessimistic surrogate (L^CLIP)
pol_surr = - U.mean(tf.minimum(surr1, surr2))
# Loss on the Q-function
vf_loss = U.mean(tf.square(pi.vpred - ret))
# calculate the total loss
total_loss = pol_surr + vf_loss
losses = [pol_surr, pol_entpen, vf_loss, meankl, meanent]
loss_names = ["pol_surr", "pol_entpen", "vf_loss", "kl", "ent"]
# calculate logarithm of propability of policy over options
log_pi = tf.log(tf.clip_by_value(pi.op_pi, 1e-5, 1.0))
# calculate logarithm of propability of policy over options old parameter
old_log_pi = tf.log(tf.clip_by_value(oldpi.op_pi, 1e-5, 1.0))
# calculate entropy of policy over options
entropy = -tf.reduce_sum(pi.op_pi * log_pi, reduction_indices=1)
# calculate the ppo update for the policy over options:
ratio_pol_ov_op = tf.exp(tf.transpose(log_pi)[option[0]] - tf.transpose(old_log_pi)[option[0]]) # pnew / pold
term_adv_clip = term_adv
surr1_pol_ov_op = ratio_pol_ov_op * term_adv_clip # surrogate from conservative policy iteration
surr2_pol_ov_op = U.clip(ratio_pol_ov_op, 1.0 - clip_param, 1.0 + clip_param) * term_adv_clip #
pol_surr_pol_ov_op = - U.mean(tf.minimum(surr1_pol_ov_op, surr2_pol_ov_op)) # PPO's pessimistic surrogate (L^CLIP)
op_loss = pol_surr_pol_ov_op - pol_ov_op_ent*tf.reduce_sum(entropy)
# add loss of policy over options to total loss
total_loss += op_loss
var_list = pi.get_trainable_variables()
term_list = var_list[6:8]
# define function that we will later do gradien descent on
lossandgrad = U.function([ob, ac, atarg, ret, lrmult,option, term_adv,pol_ov_op_ent], losses + [U.flatgrad(total_loss, var_list)])
# define adam optimizer
adam = MpiAdam(var_list, epsilon=adam_epsilon)
# define function that will assign the current parameters to the old policy
assign_old_eq_new = U.function([],[], updates=[tf.assign(oldv, newv)
for (oldv, newv) in zipsame(oldpi.get_variables(), pi.get_variables())])
compute_losses = U.function([ob, ac, atarg, ret, lrmult, option], losses)
U.initialize()
adam.sync()
# NOW: all the stuff for training was defined, from here on we start with the execution:
# initialize "savers" which will store the results
saver = tf.train.Saver(max_to_keep=10000)
saver_best = tf.train.Saver(max_to_keep=1)
### Define the names of the .csv files that are going to be stored
results=[]
if saves:
results = open(dirname + version_name + '_' + gamename +'_'+str(num_options)+'opts_'+'_results.csv','w')
results_best_model = open(dirname + version_name + '_' + gamename +'_'+str(num_options)+'opts_'+'_bestmodel.csv','w')
out = 'epoch,avg_reward'
for opt in range(num_options): out += ',option {} dur'.format(opt)
for opt in range(num_options): out += ',option {} std'.format(opt)
for opt in range(num_options): out += ',option {} term'.format(opt)
for opt in range(num_options): out += ',option {} adv'.format(opt)
out+='\n'
results.write(out)
# results.write('epoch,avg_reward,option 1 dur, option 2 dur, option 1 term, option 2 term\n')
results.flush()
# speciality: if running the training with epoch argument -> a model is loaded
if epoch >= 0:
dirname = '{}_{}opts_saves/'.format(gamename,num_options)
print("Loading weights from iteration: " + str(epoch))
filename = dirname + '{}_epoch_{}.ckpt'.format(gamename,epoch)
saver.restore(U.get_session(),filename)
###
# start training
episodes_so_far = 0
timesteps_so_far = 0
global iters_so_far
iters_so_far = 0
des_pol_op_ent = 0.1 # define policy over options entropy scheduling
max_val = -100000 # define max_val, this will be updated to always store the best model
tstart = time.time()
lenbuffer = deque(maxlen=100) # rolling buffer for episode lengths
rewbuffer = deque(maxlen=100) # rolling buffer for episode rewards
assert sum([max_iters>0, max_timesteps>0, max_episodes>0, max_seconds>0])==1, "Only one time constraint permitted"
# Prepare for rollouts
# ----------------------------------------
seg_gen = traj_segment_generator(pi, env, timesteps_per_batch, stochastic=True, num_options=num_options,saves=saves,results=results,rewbuffer=rewbuffer,dc=dc)
datas = [0 for _ in range(num_options)]
while True:
if callback: callback(locals(), globals())
if max_timesteps and timesteps_so_far >= max_timesteps:
break
elif max_episodes and episodes_so_far >= max_episodes:
break
elif max_iters and iters_so_far >= max_iters:
break
elif max_seconds and time.time() - tstart >= max_seconds:
break
if schedule == 'constant':
cur_lrmult = 1.0
elif schedule == 'linear':
cur_lrmult = max(1.0 - float(timesteps_so_far) / max_timesteps, 0)
else:
raise NotImplementedError
logger.log("********** Iteration %i ************"%iters_so_far)
# Sample (s,a)-Transitions
seg = seg_gen.__next__()
# Calculate A(s,a,o) using GAE
add_vtarg_and_adv(seg, gamma, lam)
# calculate information for logging
opt_d = []
for i in range(num_options):
dur = np.mean(seg['opt_dur'][i]) if len(seg['opt_dur'][i]) > 0 else 0.
opt_d.append(dur)
std = []
for i in range(num_options):
logstd = np.mean(seg['logstds'][i]) if len(seg['logstds'][i]) > 0 else 0.
std.append(np.exp(logstd))
print("mean opt dur:", opt_d)
print("mean op pol:", np.mean(np.array(seg['optpol_p']),axis=0))
print("mean term p:", np.mean(np.array(seg['term_p']),axis=0))
print("mean value val:", np.mean(np.array(seg['value_val']),axis=0))
ob, ac, opts, atarg, tdlamret = seg["ob"], seg["ac"], seg["opts"], seg["adv"], seg["tdlamret"]
vpredbefore = seg["vpred"] # predicted value function before udpate
atarg = (atarg - atarg.mean()) / atarg.std() # standardized advantage function estimate
if hasattr(pi, "ob_rms"): pi.ob_rms.update(ob) # update running mean/std for policy
if hasattr(pi, "ob_rms_only"): pi.ob_rms_only.update(ob[:,:-ac_space.shape[0]]) # update running mean/std for policy
assign_old_eq_new() # set old parameter values to new parameter values
# if iterations modulo 1000 -> adapt entropy scheduling coefficient
if (iters_so_far+1)%1000 == 0:
des_pol_op_ent = des_pol_op_ent/10
# every 50 epochs save the best model
if iters_so_far % 50 == 0 and wsaves:
print("weights are saved...")
filename = dirname + '{}_epoch_{}.ckpt'.format(gamename,iters_so_far)
save_path = saver.save(U.get_session(),filename)
# adaptively save best model -> if current reward is highest, save the model
if (np.mean(rewbuffer)>max_val) and wsaves:
max_val = np.mean(rewbuffer)
results_best_model.write('epoch: '+str(iters_so_far) + 'rew: ' + str(np.mean(rewbuffer)) + '\n')
results_best_model.flush()
filename = dirname + 'best.ckpt'.format(gamename,iters_so_far)
save_path = saver_best.save(U.get_session(),filename)
# minimum batch size:
min_batch=160
t_advs = [[] for _ in range(num_options)]
# select all the samples concering one of the options
# Note: so far the update is that we first use all samples from option 0 to update, then we use all samples from option 1 to update
for opt in range(num_options):
indices = np.where(opts==opt)[0]
print("batch size:",indices.size)
opt_d[opt] = indices.size
if not indices.size:
t_advs[opt].append(0.)
continue
### This part is only necessasry when we use options. We proceed to these verifications in order not to discard any collected trajectories.
if datas[opt] != 0:
if (indices.size < min_batch and datas[opt].n > min_batch):
datas[opt] = Dataset(dict(ob=ob[indices], ac=ac[indices], atarg=atarg[indices], vtarg=tdlamret[indices]), shuffle=not pi.recurrent)
t_advs[opt].append(0.)
continue
elif indices.size + datas[opt].n < min_batch:
# pdb.set_trace()
oldmap = datas[opt].data_map
cat_ob = np.concatenate((oldmap['ob'],ob[indices]))
cat_ac = np.concatenate((oldmap['ac'],ac[indices]))
cat_atarg = np.concatenate((oldmap['atarg'],atarg[indices]))
cat_vtarg = np.concatenate((oldmap['vtarg'],tdlamret[indices]))
datas[opt] = Dataset(dict(ob=cat_ob, ac=cat_ac, atarg=cat_atarg, vtarg=cat_vtarg), shuffle=not pi.recurrent)
t_advs[opt].append(0.)
continue
elif (indices.size + datas[opt].n > min_batch and datas[opt].n < min_batch) or (indices.size > min_batch and datas[opt].n < min_batch):
oldmap = datas[opt].data_map
cat_ob = np.concatenate((oldmap['ob'],ob[indices]))
cat_ac = np.concatenate((oldmap['ac'],ac[indices]))
cat_atarg = np.concatenate((oldmap['atarg'],atarg[indices]))
cat_vtarg = np.concatenate((oldmap['vtarg'],tdlamret[indices]))
datas[opt] = d = Dataset(dict(ob=cat_ob, ac=cat_ac, atarg=cat_atarg, vtarg=cat_vtarg), shuffle=not pi.recurrent)
if (indices.size > min_batch and datas[opt].n > min_batch):
datas[opt] = d = Dataset(dict(ob=ob[indices], ac=ac[indices], atarg=atarg[indices], vtarg=tdlamret[indices]), shuffle=not pi.recurrent)
elif datas[opt] == 0:
datas[opt] = d = Dataset(dict(ob=ob[indices], ac=ac[indices], atarg=atarg[indices], vtarg=tdlamret[indices]), shuffle=not pi.recurrent)
###
# define the batchsize of the optimizer:
optim_batchsize = optim_batchsize or ob.shape[0]
print("optim epochs:", optim_epochs)
logger.log("Optimizing...")
# Here we do a bunch of optimization epochs over the data
for _ in range(optim_epochs):
losses = [] # list of tuples, each of which gives the loss for a minibatch
for batch in d.iterate_once(optim_batchsize):
# Calculate advantage for using specific option here
tadv,nodc_adv = pi.get_opt_adv(batch["ob"],[opt])
tadv = tadv if num_options > 1 else np.zeros_like(tadv)
t_advs[opt].append(nodc_adv)
# calculate the gradient
*newlosses, grads = lossandgrad(batch["ob"], batch["ac"], batch["atarg"], batch["vtarg"], cur_lrmult, [opt], tadv,des_pol_op_ent)
# perform gradient update
adam.update(grads, optim_stepsize * cur_lrmult)
losses.append(newlosses)
# do logging:
lrlocal = (seg["ep_lens"], seg["ep_rets"]) # local values
listoflrpairs = MPI.COMM_WORLD.allgather(lrlocal) # list of tuples
lens, rews = map(flatten_lists, zip(*listoflrpairs))
lenbuffer.extend(lens)
rewbuffer.extend(rews)
logger.record_tabular("EpLenMean", np.mean(lenbuffer))
logger.record_tabular("EpRewMean", np.mean(rewbuffer))
logger.record_tabular("EpThisIter", len(lens))
episodes_so_far += len(lens)
timesteps_so_far += sum(lens)
iters_so_far += 1
logger.record_tabular("EpisodesSoFar", episodes_so_far)
logger.record_tabular("TimestepsSoFar", timesteps_so_far)
logger.record_tabular("TimeElapsed", time.time() - tstart)
if MPI.COMM_WORLD.Get_rank()==0:
logger.dump_tabular()
### Book keeping
if saves:
out = "{},{}"
for _ in range(num_options): out+=",{},{},{},{}"
out+="\n"
info = [iters_so_far, np.mean(rewbuffer)]
for i in range(num_options): info.append(opt_d[i])
for i in range(num_options): info.append(std[i])
for i in range(num_options): info.append(np.mean(np.array(seg['term_p']),axis=0)[i])
for i in range(num_options):
info.append(np.mean(t_advs[i]))
results.write(out.format(*info))
results.flush()
###
def flatten_lists(listoflists):
return [el for list_ in listoflists for el in list_]
|
[
"baselines.common.tf_util.get_session",
"baselines.common.mpi_adam.MpiAdam",
"tensorflow.transpose",
"tensorflow.reduce_sum",
"mpi4py.MPI.COMM_WORLD.allgather",
"numpy.array",
"baselines.logger.log",
"tensorflow.set_random_seed",
"baselines.common.tf_util.get_placeholder_cached",
"baselines.common.tf_util.clip",
"baselines.common.tf_util.get_placeholder",
"os.path.exists",
"numpy.mean",
"collections.deque",
"shutil.copy2",
"numpy.where",
"tensorflow.placeholder",
"numpy.exp",
"tensorflow.assign",
"baselines.common.tf_util.mean",
"numpy.empty",
"numpy.random.seed",
"numpy.concatenate",
"tensorflow.square",
"tensorflow.clip_by_value",
"baselines.common.tf_util.flatgrad",
"os.path.dirname",
"baselines.common.tf_util.initialize",
"numpy.shape",
"time.time",
"tensorflow.minimum",
"numpy.copy",
"baselines.logger.record_tabular",
"os.makedirs",
"tensorflow.train.Saver",
"os.path.join",
"baselines.logger.dump_tabular",
"numpy.append",
"numpy.zeros",
"numpy.zeros_like",
"mpi4py.MPI.COMM_WORLD.Get_rank",
"baselines.common.tf_util.function"
] |
[((895, 907), 'numpy.shape', 'np.shape', (['ob'], {}), '(ob)\n', (903, 907), True, 'import tensorflow as tf, numpy as np\n'), ((927, 939), 'numpy.shape', 'np.shape', (['ac'], {}), '(ac)\n', (935, 939), True, 'import tensorflow as tf, numpy as np\n'), ((987, 1011), 'numpy.concatenate', 'np.concatenate', (['(ob, ac)'], {}), '((ob, ac))\n', (1001, 1011), True, 'import tensorflow as tf, numpy as np\n'), ((1295, 1323), 'numpy.zeros', 'np.zeros', (['horizon', '"""float32"""'], {}), "(horizon, 'float32')\n", (1303, 1323), True, 'import tensorflow as tf, numpy as np\n'), ((1339, 1367), 'numpy.zeros', 'np.zeros', (['horizon', '"""float32"""'], {}), "(horizon, 'float32')\n", (1347, 1367), True, 'import tensorflow as tf, numpy as np\n'), ((1381, 1409), 'numpy.zeros', 'np.zeros', (['horizon', '"""float32"""'], {}), "(horizon, 'float32')\n", (1389, 1409), True, 'import tensorflow as tf, numpy as np\n'), ((1421, 1447), 'numpy.zeros', 'np.zeros', (['horizon', '"""int32"""'], {}), "(horizon, 'int32')\n", (1429, 1447), True, 'import tensorflow as tf, numpy as np\n'), ((1459, 1485), 'numpy.zeros', 'np.zeros', (['horizon', '"""int32"""'], {}), "(horizon, 'int32')\n", (1467, 1485), True, 'import tensorflow as tf, numpy as np\n'), ((5192, 5216), 'numpy.append', 'np.append', (["seg['new']", '(0)'], {}), "(seg['new'], 0)\n", (5201, 5216), True, 'import tensorflow as tf, numpy as np\n'), ((5314, 5355), 'numpy.append', 'np.append', (["seg['vpred']", "seg['nextvpred']"], {}), "(seg['vpred'], seg['nextvpred'])\n", (5323, 5355), True, 'import tensorflow as tf, numpy as np\n'), ((5406, 5428), 'numpy.empty', 'np.empty', (['T', '"""float32"""'], {}), "(T, 'float32')\n", (5414, 5428), True, 'import tensorflow as tf, numpy as np\n'), ((6507, 6527), 'numpy.random.seed', 'np.random.seed', (['seed'], {}), '(seed)\n', (6521, 6527), True, 'import tensorflow as tf, numpy as np\n'), ((6532, 6556), 'tensorflow.set_random_seed', 'tf.set_random_seed', (['seed'], {}), '(seed)\n', (6550, 6556), True, 'import tensorflow as tf, numpy as np\n'), ((7084, 7109), 'os.path.dirname', 'os.path.dirname', (['__file__'], {}), '(__file__)\n', (7099, 7109), False, 'import os\n'), ((8949, 8995), 'tensorflow.placeholder', 'tf.placeholder', ([], {'dtype': 'tf.float32', 'shape': '[None]'}), '(dtype=tf.float32, shape=[None])\n', (8963, 8995), True, 'import tensorflow as tf, numpy as np\n'), ((9035, 9081), 'tensorflow.placeholder', 'tf.placeholder', ([], {'dtype': 'tf.float32', 'shape': '[None]'}), '(dtype=tf.float32, shape=[None])\n', (9049, 9081), True, 'import tensorflow as tf, numpy as np\n'), ((9121, 9165), 'tensorflow.placeholder', 'tf.placeholder', ([], {'dtype': 'tf.float32', 'shape': 'None'}), '(dtype=tf.float32, shape=None)\n', (9135, 9165), True, 'import tensorflow as tf, numpy as np\n'), ((9227, 9284), 'tensorflow.placeholder', 'tf.placeholder', ([], {'name': '"""lrmult"""', 'dtype': 'tf.float32', 'shape': '[]'}), "(name='lrmult', dtype=tf.float32, shape=[])\n", (9241, 9284), True, 'import tensorflow as tf, numpy as np\n'), ((9483, 9518), 'baselines.common.tf_util.get_placeholder_cached', 'U.get_placeholder_cached', ([], {'name': '"""ob"""'}), "(name='ob')\n", (9507, 9518), True, 'import baselines.common.tf_util as U\n'), ((9532, 9571), 'baselines.common.tf_util.get_placeholder_cached', 'U.get_placeholder_cached', ([], {'name': '"""option"""'}), "(name='option')\n", (9556, 9571), True, 'import baselines.common.tf_util as U\n'), ((9587, 9653), 'baselines.common.tf_util.get_placeholder', 'U.get_placeholder', ([], {'name': '"""term_adv"""', 'dtype': 'tf.float32', 'shape': '[None]'}), "(name='term_adv', dtype=tf.float32, shape=[None])\n", (9604, 9653), True, 'import baselines.common.tf_util as U\n'), ((9808, 9824), 'baselines.common.tf_util.mean', 'U.mean', (['kloldnew'], {}), '(kloldnew)\n', (9814, 9824), True, 'import baselines.common.tf_util as U\n'), ((9839, 9850), 'baselines.common.tf_util.mean', 'U.mean', (['ent'], {}), '(ent)\n', (9845, 9850), True, 'import baselines.common.tf_util as U\n'), ((12038, 12077), 'baselines.common.mpi_adam.MpiAdam', 'MpiAdam', (['var_list'], {'epsilon': 'adam_epsilon'}), '(var_list, epsilon=adam_epsilon)\n', (12045, 12077), False, 'from baselines.common.mpi_adam import MpiAdam\n'), ((12334, 12390), 'baselines.common.tf_util.function', 'U.function', (['[ob, ac, atarg, ret, lrmult, option]', 'losses'], {}), '([ob, ac, atarg, ret, lrmult, option], losses)\n', (12344, 12390), True, 'import baselines.common.tf_util as U\n'), ((12397, 12411), 'baselines.common.tf_util.initialize', 'U.initialize', ([], {}), '()\n', (12409, 12411), True, 'import baselines.common.tf_util as U\n'), ((12591, 12624), 'tensorflow.train.Saver', 'tf.train.Saver', ([], {'max_to_keep': '(10000)'}), '(max_to_keep=10000)\n', (12605, 12624), True, 'import tensorflow as tf, numpy as np\n'), ((12642, 12671), 'tensorflow.train.Saver', 'tf.train.Saver', ([], {'max_to_keep': '(1)'}), '(max_to_keep=1)\n', (12656, 12671), True, 'import tensorflow as tf, numpy as np\n'), ((14199, 14210), 'time.time', 'time.time', ([], {}), '()\n', (14208, 14210), False, 'import time\n'), ((14227, 14244), 'collections.deque', 'deque', ([], {'maxlen': '(100)'}), '(maxlen=100)\n', (14232, 14244), False, 'from collections import deque\n'), ((14298, 14315), 'collections.deque', 'deque', ([], {'maxlen': '(100)'}), '(maxlen=100)\n', (14303, 14315), False, 'from collections import deque\n'), ((8141, 8198), 'os.path.join', 'os.path.join', (['dirname_rel', '"""nfunk/envs_nf/pendulum_nf.py"""'], {}), "(dirname_rel, 'nfunk/envs_nf/pendulum_nf.py')\n", (8153, 8198), False, 'import os\n'), ((8206, 8229), 'shutil.copy2', 'shutil.copy2', (['src', 'dest'], {}), '(src, dest)\n', (8218, 8229), False, 'import shutil\n'), ((8237, 8260), 'shutil.copy2', 'shutil.copy2', (['src', 'dest'], {}), '(src, dest)\n', (8249, 8260), False, 'import shutil\n'), ((8268, 8297), 'os.makedirs', 'os.makedirs', (["(dest + 'assets/')"], {}), "(dest + 'assets/')\n", (8279, 8297), False, 'import os\n'), ((8310, 8373), 'os.path.join', 'os.path.join', (['dirname_rel', '"""nfunk/envs_nf/assets/clockwise.png"""'], {}), "(dirname_rel, 'nfunk/envs_nf/assets/clockwise.png')\n", (8322, 8373), False, 'import os\n'), ((8381, 8416), 'shutil.copy2', 'shutil.copy2', (['src', "(dest + 'assets/')"], {}), "(src, dest + 'assets/')\n", (8393, 8416), False, 'import shutil\n'), ((10218, 10267), 'baselines.common.tf_util.clip', 'U.clip', (['ratio', '(1.0 - clip_param)', '(1.0 + clip_param)'], {}), '(ratio, 1.0 - clip_param, 1.0 + clip_param)\n', (10224, 10267), True, 'import baselines.common.tf_util as U\n'), ((10427, 10452), 'tensorflow.square', 'tf.square', (['(pi.vpred - ret)'], {}), '(pi.vpred - ret)\n', (10436, 10452), True, 'import tensorflow as tf, numpy as np\n'), ((10736, 10774), 'tensorflow.clip_by_value', 'tf.clip_by_value', (['pi.op_pi', '(1e-05)', '(1.0)'], {}), '(pi.op_pi, 1e-05, 1.0)\n', (10752, 10774), True, 'import tensorflow as tf, numpy as np\n'), ((10877, 10918), 'tensorflow.clip_by_value', 'tf.clip_by_value', (['oldpi.op_pi', '(1e-05)', '(1.0)'], {}), '(oldpi.op_pi, 1e-05, 1.0)\n', (10893, 10918), True, 'import tensorflow as tf, numpy as np\n'), ((10981, 11034), 'tensorflow.reduce_sum', 'tf.reduce_sum', (['(pi.op_pi * log_pi)'], {'reduction_indices': '(1)'}), '(pi.op_pi * log_pi, reduction_indices=1)\n', (10994, 11034), True, 'import tensorflow as tf, numpy as np\n'), ((11363, 11422), 'baselines.common.tf_util.clip', 'U.clip', (['ratio_pol_ov_op', '(1.0 - clip_param)', '(1.0 + clip_param)'], {}), '(ratio_pol_ov_op, 1.0 - clip_param, 1.0 + clip_param)\n', (11369, 11422), True, 'import baselines.common.tf_util as U\n'), ((15384, 15449), 'baselines.logger.log', 'logger.log', (["('********** Iteration %i ************' % iters_so_far)"], {}), "('********** Iteration %i ************' % iters_so_far)\n", (15394, 15449), False, 'from baselines import logger\n'), ((21787, 21820), 'mpi4py.MPI.COMM_WORLD.allgather', 'MPI.COMM_WORLD.allgather', (['lrlocal'], {}), '(lrlocal)\n', (21811, 21820), False, 'from mpi4py import MPI\n'), ((22251, 22306), 'baselines.logger.record_tabular', 'logger.record_tabular', (['"""EpisodesSoFar"""', 'episodes_so_far'], {}), "('EpisodesSoFar', episodes_so_far)\n", (22272, 22306), False, 'from baselines import logger\n'), ((22315, 22372), 'baselines.logger.record_tabular', 'logger.record_tabular', (['"""TimestepsSoFar"""', 'timesteps_so_far'], {}), "('TimestepsSoFar', timesteps_so_far)\n", (22336, 22372), False, 'from baselines import logger\n'), ((7380, 7403), 'os.path.exists', 'os.path.exists', (['dirname'], {}), '(dirname)\n', (7394, 7403), False, 'import os\n'), ((7417, 7437), 'os.makedirs', 'os.makedirs', (['dirname'], {}), '(dirname)\n', (7428, 7437), False, 'import os\n'), ((8034, 8057), 'shutil.copy2', 'shutil.copy2', (['src', 'dest'], {}), '(src, dest)\n', (8046, 8057), False, 'import shutil\n'), ((10349, 10373), 'tensorflow.minimum', 'tf.minimum', (['surr1', 'surr2'], {}), '(surr1, surr2)\n', (10359, 10373), True, 'import tensorflow as tf, numpy as np\n'), ((11475, 11519), 'tensorflow.minimum', 'tf.minimum', (['surr1_pol_ov_op', 'surr2_pol_ov_op'], {}), '(surr1_pol_ov_op, surr2_pol_ov_op)\n', (11485, 11519), True, 'import tensorflow as tf, numpy as np\n'), ((11614, 11636), 'tensorflow.reduce_sum', 'tf.reduce_sum', (['entropy'], {}), '(entropy)\n', (11627, 11636), True, 'import tensorflow as tf, numpy as np\n'), ((13857, 13872), 'baselines.common.tf_util.get_session', 'U.get_session', ([], {}), '()\n', (13870, 13872), True, 'import baselines.common.tf_util as U\n'), ((17457, 17475), 'numpy.mean', 'np.mean', (['rewbuffer'], {}), '(rewbuffer)\n', (17464, 17475), True, 'import tensorflow as tf, numpy as np\n'), ((20752, 20779), 'baselines.logger.log', 'logger.log', (['"""Optimizing..."""'], {}), "('Optimizing...')\n", (20762, 20779), False, 'from baselines import logger\n'), ((22004, 22022), 'numpy.mean', 'np.mean', (['lenbuffer'], {}), '(lenbuffer)\n', (22011, 22022), True, 'import tensorflow as tf, numpy as np\n'), ((22067, 22085), 'numpy.mean', 'np.mean', (['rewbuffer'], {}), '(rewbuffer)\n', (22074, 22085), True, 'import tensorflow as tf, numpy as np\n'), ((22451, 22476), 'mpi4py.MPI.COMM_WORLD.Get_rank', 'MPI.COMM_WORLD.Get_rank', ([], {}), '()\n', (22474, 22476), False, 'from mpi4py import MPI\n'), ((22493, 22514), 'baselines.logger.dump_tabular', 'logger.dump_tabular', ([], {}), '()\n', (22512, 22514), False, 'from baselines import logger\n'), ((3522, 3533), 'numpy.copy', 'np.copy', (['ac'], {}), '(ac)\n', (3529, 3533), True, 'import tensorflow as tf, numpy as np\n'), ((7703, 7757), 'os.path.join', 'os.path.join', (['dirname_rel', '"""baselines/baselines/ppo1/"""'], {}), "(dirname_rel, 'baselines/baselines/ppo1/')\n", (7715, 7757), False, 'import os\n'), ((7949, 7966), 'os.makedirs', 'os.makedirs', (['dest'], {}), '(dest)\n', (7960, 7966), False, 'import os\n'), ((11125, 11145), 'tensorflow.transpose', 'tf.transpose', (['log_pi'], {}), '(log_pi)\n', (11137, 11145), True, 'import tensorflow as tf, numpy as np\n'), ((11159, 11183), 'tensorflow.transpose', 'tf.transpose', (['old_log_pi'], {}), '(old_log_pi)\n', (11171, 11183), True, 'import tensorflow as tf, numpy as np\n'), ((11959, 11991), 'baselines.common.tf_util.flatgrad', 'U.flatgrad', (['total_loss', 'var_list'], {}), '(total_loss, var_list)\n', (11969, 11991), True, 'import baselines.common.tf_util as U\n'), ((12210, 12231), 'tensorflow.assign', 'tf.assign', (['oldv', 'newv'], {}), '(oldv, newv)\n', (12219, 12231), True, 'import tensorflow as tf, numpy as np\n'), ((15719, 15745), 'numpy.mean', 'np.mean', (["seg['opt_dur'][i]"], {}), "(seg['opt_dur'][i])\n", (15726, 15745), True, 'import tensorflow as tf, numpy as np\n'), ((15890, 15916), 'numpy.mean', 'np.mean', (["seg['logstds'][i]"], {}), "(seg['logstds'][i])\n", (15897, 15916), True, 'import tensorflow as tf, numpy as np\n'), ((15978, 15992), 'numpy.exp', 'np.exp', (['logstd'], {}), '(logstd)\n', (15984, 15992), True, 'import tensorflow as tf, numpy as np\n'), ((16083, 16108), 'numpy.array', 'np.array', (["seg['optpol_p']"], {}), "(seg['optpol_p'])\n", (16091, 16108), True, 'import tensorflow as tf, numpy as np\n'), ((16165, 16188), 'numpy.array', 'np.array', (["seg['term_p']"], {}), "(seg['term_p'])\n", (16173, 16188), True, 'import tensorflow as tf, numpy as np\n'), ((16239, 16265), 'numpy.array', 'np.array', (["seg['value_val']"], {}), "(seg['value_val'])\n", (16247, 16265), True, 'import tensorflow as tf, numpy as np\n'), ((17271, 17286), 'baselines.common.tf_util.get_session', 'U.get_session', ([], {}), '()\n', (17284, 17286), True, 'import baselines.common.tf_util as U\n'), ((17395, 17413), 'numpy.mean', 'np.mean', (['rewbuffer'], {}), '(rewbuffer)\n', (17402, 17413), True, 'import tensorflow as tf, numpy as np\n'), ((17739, 17754), 'baselines.common.tf_util.get_session', 'U.get_session', ([], {}), '()\n', (17752, 17754), True, 'import baselines.common.tf_util as U\n'), ((18143, 18164), 'numpy.where', 'np.where', (['(opts == opt)'], {}), '(opts == opt)\n', (18151, 18164), True, 'import tensorflow as tf, numpy as np\n'), ((22418, 22429), 'time.time', 'time.time', ([], {}), '()\n', (22427, 22429), False, 'import time\n'), ((22716, 22734), 'numpy.mean', 'np.mean', (['rewbuffer'], {}), '(rewbuffer)\n', (22723, 22734), True, 'import tensorflow as tf, numpy as np\n'), ((23027, 23045), 'numpy.mean', 'np.mean', (['t_advs[i]'], {}), '(t_advs[i])\n', (23034, 23045), True, 'import tensorflow as tf, numpy as np\n'), ((19013, 19056), 'numpy.concatenate', 'np.concatenate', (["(oldmap['ob'], ob[indices])"], {}), "((oldmap['ob'], ob[indices]))\n", (19027, 19056), True, 'import tensorflow as tf, numpy as np\n'), ((19085, 19128), 'numpy.concatenate', 'np.concatenate', (["(oldmap['ac'], ac[indices])"], {}), "((oldmap['ac'], ac[indices]))\n", (19099, 19128), True, 'import tensorflow as tf, numpy as np\n'), ((19160, 19209), 'numpy.concatenate', 'np.concatenate', (["(oldmap['atarg'], atarg[indices])"], {}), "((oldmap['atarg'], atarg[indices]))\n", (19174, 19209), True, 'import tensorflow as tf, numpy as np\n'), ((19241, 19293), 'numpy.concatenate', 'np.concatenate', (["(oldmap['vtarg'], tdlamret[indices])"], {}), "((oldmap['vtarg'], tdlamret[indices]))\n", (19255, 19293), True, 'import tensorflow as tf, numpy as np\n'), ((21247, 21266), 'numpy.zeros_like', 'np.zeros_like', (['tadv'], {}), '(tadv)\n', (21260, 21266), True, 'import tensorflow as tf, numpy as np\n'), ((17557, 17575), 'numpy.mean', 'np.mean', (['rewbuffer'], {}), '(rewbuffer)\n', (17564, 17575), True, 'import tensorflow as tf, numpy as np\n'), ((19726, 19769), 'numpy.concatenate', 'np.concatenate', (["(oldmap['ob'], ob[indices])"], {}), "((oldmap['ob'], ob[indices]))\n", (19740, 19769), True, 'import tensorflow as tf, numpy as np\n'), ((19798, 19841), 'numpy.concatenate', 'np.concatenate', (["(oldmap['ac'], ac[indices])"], {}), "((oldmap['ac'], ac[indices]))\n", (19812, 19841), True, 'import tensorflow as tf, numpy as np\n'), ((19873, 19922), 'numpy.concatenate', 'np.concatenate', (["(oldmap['atarg'], atarg[indices])"], {}), "((oldmap['atarg'], atarg[indices]))\n", (19887, 19922), True, 'import tensorflow as tf, numpy as np\n'), ((19954, 20006), 'numpy.concatenate', 'np.concatenate', (["(oldmap['vtarg'], tdlamret[indices])"], {}), "((oldmap['vtarg'], tdlamret[indices]))\n", (19968, 20006), True, 'import tensorflow as tf, numpy as np\n'), ((22921, 22944), 'numpy.array', 'np.array', (["seg['term_p']"], {}), "(seg['term_p'])\n", (22929, 22944), True, 'import tensorflow as tf, numpy as np\n'), ((15088, 15099), 'time.time', 'time.time', ([], {}), '()\n', (15097, 15099), False, 'import time\n')]
|
import time
from unittest.case import SkipTest
from ddtrace.context import Context
from ddtrace.constants import ANALYTICS_SAMPLE_RATE_KEY
from ddtrace.span import Span
from ddtrace.ext import errors
def test_ids():
s = Span(tracer=None, name='span.test')
assert s.trace_id
assert s.span_id
assert not s.parent_id
s2 = Span(tracer=None, name='t', trace_id=1, span_id=2, parent_id=1)
assert s2.trace_id == 1
assert s2.span_id == 2
assert s2.parent_id == 1
def test_tags():
s = Span(tracer=None, name='test.span')
s.set_tag('a', 'a')
s.set_tag('b', 1)
s.set_tag('c', '1')
d = s.to_dict()
expected = {
'a': 'a',
'b': '1',
'c': '1',
}
assert d['meta'] == expected
def test_set_valid_metrics():
s = Span(tracer=None, name='test.span')
s.set_metric('a', 0)
s.set_metric('b', -12)
s.set_metric('c', 12.134)
s.set_metric('d', 1231543543265475686787869123)
s.set_metric('e', '12.34')
d = s.to_dict()
expected = {
'a': 0,
'b': -12,
'c': 12.134,
'd': 1231543543265475686787869123,
'e': 12.34,
}
assert d['metrics'] == expected
def test_set_invalid_metric():
s = Span(tracer=None, name='test.span')
invalid_metrics = [
None,
{},
[],
s,
'quarante-douze',
float('nan'),
float('inf'),
1j
]
for i, m in enumerate(invalid_metrics):
k = str(i)
s.set_metric(k, m)
assert s.get_metric(k) is None
def test_set_numpy_metric():
try:
import numpy as np
except ImportError:
raise SkipTest('numpy not installed')
s = Span(tracer=None, name='test.span')
s.set_metric('a', np.int64(1))
assert s.get_metric('a') == 1
assert type(s.get_metric('a')) == float
def test_tags_not_string():
# ensure we can cast as strings
class Foo(object):
def __repr__(self):
1 / 0
s = Span(tracer=None, name='test.span')
s.set_tag('a', Foo())
def test_finish():
# ensure finish will record a span
dt = DummyTracer()
ctx = Context()
s = Span(dt, 'test.span', context=ctx)
ctx.add_span(s)
assert s.duration is None
sleep = 0.05
with s as s1:
assert s is s1
time.sleep(sleep)
assert s.duration >= sleep, '%s < %s' % (s.duration, sleep)
assert 1 == dt.spans_recorded
def test_finish_no_tracer():
# ensure finish works with no tracer without raising exceptions
s = Span(tracer=None, name='test.span')
s.finish()
def test_finish_called_multiple_times():
# we should only record a span the first time finish is called on it
dt = DummyTracer()
ctx = Context()
s = Span(dt, 'bar', context=ctx)
ctx.add_span(s)
s.finish()
s.finish()
assert dt.spans_recorded == 1
def test_finish_set_span_duration():
# If set the duration on a span, the span should be recorded with this
# duration
s = Span(tracer=None, name='test.span')
s.duration = 1337.0
s.finish()
assert s.duration == 1337.0
def test_traceback_with_error():
s = Span(None, 'test.span')
try:
1 / 0
except ZeroDivisionError:
s.set_traceback()
else:
assert 0, 'should have failed'
assert s.error
assert 'by zero' in s.get_tag(errors.ERROR_MSG)
assert 'ZeroDivisionError' in s.get_tag(errors.ERROR_TYPE)
def test_traceback_without_error():
s = Span(None, 'test.span')
s.set_traceback()
assert not s.error
assert not s.get_tag(errors.ERROR_MSG)
assert not s.get_tag(errors.ERROR_TYPE)
assert 'in test_traceback_without_error' in s.get_tag(errors.ERROR_STACK)
def test_ctx_mgr():
dt = DummyTracer()
s = Span(dt, 'bar')
assert not s.duration
assert not s.error
e = Exception('boo')
try:
with s:
time.sleep(0.01)
raise e
except Exception as out:
assert out == e
assert s.duration > 0, s.duration
assert s.error
assert s.get_tag(errors.ERROR_MSG) == 'boo'
assert 'Exception' in s.get_tag(errors.ERROR_TYPE)
assert s.get_tag(errors.ERROR_STACK)
else:
assert 0, 'should have failed'
def test_span_to_dict():
s = Span(tracer=None, name='test.span', service='s', resource='r')
s.span_type = 'foo'
s.set_tag('a', '1')
s.set_meta('b', '2')
s.finish()
d = s.to_dict()
assert d
assert d['span_id'] == s.span_id
assert d['trace_id'] == s.trace_id
assert d['parent_id'] == s.parent_id
assert d['meta'] == {'a': '1', 'b': '2'}
assert d['type'] == 'foo'
assert d['error'] == 0
assert type(d['error']) == int
def test_span_to_dict_sub():
parent = Span(tracer=None, name='test.span', service='s', resource='r')
s = Span(tracer=None, name='test.span', service='s', resource='r')
s._parent = parent
s.span_type = 'foo'
s.set_tag('a', '1')
s.set_meta('b', '2')
s.finish()
d = s.to_dict()
assert d
assert d['span_id'] == s.span_id
assert d['trace_id'] == s.trace_id
assert d['parent_id'] == s.parent_id
assert d['meta'] == {'a': '1', 'b': '2'}
assert d['type'] == 'foo'
assert d['error'] == 0
assert type(d['error']) == int
def test_span_boolean_err():
s = Span(tracer=None, name='foo.bar', service='s', resource='r')
s.error = True
s.finish()
d = s.to_dict()
assert d
assert d['error'] == 1
assert type(d['error']) == int
def test_numeric_tags_none():
s = Span(tracer=None, name='test.span')
s.set_tag(ANALYTICS_SAMPLE_RATE_KEY, None)
d = s.to_dict()
assert d
assert 'metrics' not in d
def test_numeric_tags_true():
s = Span(tracer=None, name='test.span')
s.set_tag(ANALYTICS_SAMPLE_RATE_KEY, True)
d = s.to_dict()
assert d
expected = {
ANALYTICS_SAMPLE_RATE_KEY: 1.0
}
assert d['metrics'] == expected
def test_numeric_tags_value():
s = Span(tracer=None, name='test.span')
s.set_tag(ANALYTICS_SAMPLE_RATE_KEY, 0.5)
d = s.to_dict()
assert d
expected = {
ANALYTICS_SAMPLE_RATE_KEY: 0.5
}
assert d['metrics'] == expected
def test_numeric_tags_bad_value():
s = Span(tracer=None, name='test.span')
s.set_tag(ANALYTICS_SAMPLE_RATE_KEY, 'Hello')
d = s.to_dict()
assert d
assert 'metrics' not in d
class DummyTracer(object):
def __init__(self):
self.debug_logging = False
self.last_span = None
self.spans_recorded = 0
def record(self, span):
self.last_span = span
self.spans_recorded += 1
|
[
"ddtrace.span.Span",
"numpy.int64",
"unittest.case.SkipTest",
"time.sleep",
"ddtrace.context.Context"
] |
[((228, 263), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""span.test"""'}), "(tracer=None, name='span.test')\n", (232, 263), False, 'from ddtrace.span import Span\n'), ((344, 407), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""t"""', 'trace_id': '(1)', 'span_id': '(2)', 'parent_id': '(1)'}), "(tracer=None, name='t', trace_id=1, span_id=2, parent_id=1)\n", (348, 407), False, 'from ddtrace.span import Span\n'), ((519, 554), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""'}), "(tracer=None, name='test.span')\n", (523, 554), False, 'from ddtrace.span import Span\n'), ((795, 830), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""'}), "(tracer=None, name='test.span')\n", (799, 830), False, 'from ddtrace.span import Span\n'), ((1234, 1269), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""'}), "(tracer=None, name='test.span')\n", (1238, 1269), False, 'from ddtrace.span import Span\n'), ((1706, 1741), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""'}), "(tracer=None, name='test.span')\n", (1710, 1741), False, 'from ddtrace.span import Span\n'), ((1999, 2034), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""'}), "(tracer=None, name='test.span')\n", (2003, 2034), False, 'from ddtrace.span import Span\n'), ((2154, 2163), 'ddtrace.context.Context', 'Context', ([], {}), '()\n', (2161, 2163), False, 'from ddtrace.context import Context\n'), ((2172, 2206), 'ddtrace.span.Span', 'Span', (['dt', '"""test.span"""'], {'context': 'ctx'}), "(dt, 'test.span', context=ctx)\n", (2176, 2206), False, 'from ddtrace.span import Span\n'), ((2547, 2582), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""'}), "(tracer=None, name='test.span')\n", (2551, 2582), False, 'from ddtrace.span import Span\n'), ((2747, 2756), 'ddtrace.context.Context', 'Context', ([], {}), '()\n', (2754, 2756), False, 'from ddtrace.context import Context\n'), ((2765, 2793), 'ddtrace.span.Span', 'Span', (['dt', '"""bar"""'], {'context': 'ctx'}), "(dt, 'bar', context=ctx)\n", (2769, 2793), False, 'from ddtrace.span import Span\n'), ((3015, 3050), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""'}), "(tracer=None, name='test.span')\n", (3019, 3050), False, 'from ddtrace.span import Span\n'), ((3165, 3188), 'ddtrace.span.Span', 'Span', (['None', '"""test.span"""'], {}), "(None, 'test.span')\n", (3169, 3188), False, 'from ddtrace.span import Span\n'), ((3498, 3521), 'ddtrace.span.Span', 'Span', (['None', '"""test.span"""'], {}), "(None, 'test.span')\n", (3502, 3521), False, 'from ddtrace.span import Span\n'), ((3785, 3800), 'ddtrace.span.Span', 'Span', (['dt', '"""bar"""'], {}), "(dt, 'bar')\n", (3789, 3800), False, 'from ddtrace.span import Span\n'), ((4309, 4371), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""', 'service': '"""s"""', 'resource': '"""r"""'}), "(tracer=None, name='test.span', service='s', resource='r')\n", (4313, 4371), False, 'from ddtrace.span import Span\n'), ((4792, 4854), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""', 'service': '"""s"""', 'resource': '"""r"""'}), "(tracer=None, name='test.span', service='s', resource='r')\n", (4796, 4854), False, 'from ddtrace.span import Span\n'), ((4863, 4925), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""', 'service': '"""s"""', 'resource': '"""r"""'}), "(tracer=None, name='test.span', service='s', resource='r')\n", (4867, 4925), False, 'from ddtrace.span import Span\n'), ((5364, 5424), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""foo.bar"""', 'service': '"""s"""', 'resource': '"""r"""'}), "(tracer=None, name='foo.bar', service='s', resource='r')\n", (5368, 5424), False, 'from ddtrace.span import Span\n'), ((5595, 5630), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""'}), "(tracer=None, name='test.span')\n", (5599, 5630), False, 'from ddtrace.span import Span\n'), ((5781, 5816), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""'}), "(tracer=None, name='test.span')\n", (5785, 5816), False, 'from ddtrace.span import Span\n'), ((6036, 6071), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""'}), "(tracer=None, name='test.span')\n", (6040, 6071), False, 'from ddtrace.span import Span\n'), ((6294, 6329), 'ddtrace.span.Span', 'Span', ([], {'tracer': 'None', 'name': '"""test.span"""'}), "(tracer=None, name='test.span')\n", (6298, 6329), False, 'from ddtrace.span import Span\n'), ((1764, 1775), 'numpy.int64', 'np.int64', (['(1)'], {}), '(1)\n', (1772, 1775), True, 'import numpy as np\n'), ((2324, 2341), 'time.sleep', 'time.sleep', (['sleep'], {}), '(sleep)\n', (2334, 2341), False, 'import time\n'), ((1666, 1697), 'unittest.case.SkipTest', 'SkipTest', (['"""numpy not installed"""'], {}), "('numpy not installed')\n", (1674, 1697), False, 'from unittest.case import SkipTest\n'), ((3913, 3929), 'time.sleep', 'time.sleep', (['(0.01)'], {}), '(0.01)\n', (3923, 3929), False, 'import time\n')]
|
from gzip import (
compress,
GzipFile
)
import numpy as np
from .record import Record
UNK = '<unk>'
PAD = '<pad>'
class Vocab(Record):
__attributes__ = ['words', 'counts']
def __init__(self, words, counts):
self.words = words
self.counts = counts
self.word_ids = {
word: id
for id, word in enumerate(self.words)
}
self.unk_id = self.word_ids.get(UNK)
self.pad_id = self.word_ids.get(PAD)
def __getitem__(self, word):
return self.word_ids[word]
def __contains__(self, word):
return word in self.word_ids
def get(self, word, default=None):
if word in self:
return self[word]
return default
def count(self, word):
return self.counts[self.word_ids[word]]
def top(self, count=None):
return sorted(
self.words,
key=self.count,
reverse=True
)[:count]
def sampled(self, words):
words = list(words)
counts = [
self.counts[self.word_ids[_]]
for _ in words
]
return Vocab(words, counts)
def __repr__(self):
return '{name}(words=[...], counts=[...])'.format(
name=self.__class__.__name__
)
def _repr_pretty_(self, printer, cycle):
printer.text(repr(self))
@classmethod
def from_glove(cls, words, counts):
# for some reason glove vocab may have words with broken
# unicode
words = [_.decode('utf8', errors='ignore') for _ in words]
# emb has unk in the end
for word in (UNK, PAD):
words.append(word)
counts.append(0)
return cls(words, counts)
@property
def as_glove(self):
for word, count in zip(self.words, self.counts):
if word in (UNK, PAD):
continue
word = word.encode('utf8')
yield word, count
@property
def as_bytes(self):
meta = [len(self.counts)]
meta = np.array(meta).astype(np.uint32).tobytes()
words = '\n'.join(self.words)
words = words.encode('utf8')
counts = np.array(self.counts, dtype=np.uint32).tobytes()
return compress(meta + counts + words)
@classmethod
def from_file(cls, file):
file = GzipFile(mode='rb', fileobj=file)
buffer = file.read(4)
size, = np.frombuffer(buffer, np.uint32)
buffer = file.read(4 * size)
counts = np.frombuffer(buffer, np.uint32).tolist()
text = file.read().decode('utf8')
words = text.splitlines()
return cls(words, counts)
|
[
"gzip.GzipFile",
"numpy.array",
"numpy.frombuffer",
"gzip.compress"
] |
[((2259, 2290), 'gzip.compress', 'compress', (['(meta + counts + words)'], {}), '(meta + counts + words)\n', (2267, 2290), False, 'from gzip import compress, GzipFile\n'), ((2354, 2387), 'gzip.GzipFile', 'GzipFile', ([], {'mode': '"""rb"""', 'fileobj': 'file'}), "(mode='rb', fileobj=file)\n", (2362, 2387), False, 'from gzip import compress, GzipFile\n'), ((2435, 2467), 'numpy.frombuffer', 'np.frombuffer', (['buffer', 'np.uint32'], {}), '(buffer, np.uint32)\n', (2448, 2467), True, 'import numpy as np\n'), ((2195, 2233), 'numpy.array', 'np.array', (['self.counts'], {'dtype': 'np.uint32'}), '(self.counts, dtype=np.uint32)\n', (2203, 2233), True, 'import numpy as np\n'), ((2523, 2555), 'numpy.frombuffer', 'np.frombuffer', (['buffer', 'np.uint32'], {}), '(buffer, np.uint32)\n', (2536, 2555), True, 'import numpy as np\n'), ((2058, 2072), 'numpy.array', 'np.array', (['meta'], {}), '(meta)\n', (2066, 2072), True, 'import numpy as np\n')]
|
# -*- coding: utf-8 -*-
from EXOSIMS.Prototypes.OpticalSystem import OpticalSystem
import astropy.units as u
import numpy as np
import scipy.stats as st
import scipy.optimize as opt
class Nemati(OpticalSystem):
"""Nemati Optical System class
This class contains all variables and methods necessary to perform
Optical System Module calculations in exoplanet mission simulation using
the model from Nemati 2014.
Args:
\*\*specs:
user specified values
"""
def __init__(self, **specs):
OpticalSystem.__init__(self, **specs)
def calc_intTime(self, TL, sInds, fZ, fEZ, dMag, WA, mode):
"""Finds integration times of target systems for a specific observing
mode (imaging or characterization), based on Nemati 2014 (SPIE).
Args:
TL (TargetList module):
TargetList class object
sInds (integer ndarray):
Integer indices of the stars of interest
fZ (astropy Quantity array):
Surface brightness of local zodiacal light in units of 1/arcsec2
fEZ (astropy Quantity array):
Surface brightness of exo-zodiacal light in units of 1/arcsec2
dMag (float ndarray):
Differences in magnitude between planets and their host star
WA (astropy Quantity array):
Working angles of the planets of interest in units of arcsec
mode (dict):
Selected observing mode
Returns:
intTime (astropy Quantity array):
Integration times in units of day
"""
# electron counts
C_p, C_b, C_sp = self.Cp_Cb_Csp(TL, sInds, fZ, fEZ, dMag, WA, mode)
# get SNR threshold
SNR = mode['SNR']
# calculate integration time based on Nemati 2014
with np.errstate(divide='ignore', invalid='ignore'):
intTime = np.true_divide(SNR**2*C_b, (C_p**2 - (SNR*C_sp)**2))
# infinite and NAN are set to zero
intTime[np.isinf(intTime) | np.isnan(intTime)] = 0.*u.d
# negative values are set to zero
intTime[intTime < 0] = 0.*u.d
return intTime.to('day')
def calc_dMag_per_intTime(self, intTimes, TL, sInds, fZ, fEZ, WA, mode, C_b=None, C_sp=None):
"""Finds achievable dMag for one integration time per star in the input
list at one working angle.
Args:
intTimes (astropy Quantity array):
Integration times
TL (TargetList module):
TargetList class object
sInds (integer ndarray):
Integer indices of the stars of interest
fZ (astropy Quantity array):
Surface brightness of local zodiacal light for each star in sInds
in units of 1/arcsec2
fEZ (astropy Quantity array):
Surface brightness of exo-zodiacal light for each star in sInds
in units of 1/arcsec2
WA (astropy Quantity array):
Working angle for each star in sInds in units of arcsec
mode (dict):
Selected observing mode
C_b (astropy Quantity array):
Background noise electron count rate in units of 1/s (optional)
C_sp (astropy Quantity array):
Residual speckle spatial structure (systematic error) in units of 1/s
(optional)
Returns:
dMag (ndarray):
Achievable dMag for given integration time and working angle
"""
# cast sInds, WA, fZ, fEZ, and intTimes to arrays
sInds = np.array(sInds, ndmin=1, copy=False)
WA = np.array(WA.value, ndmin=1)*WA.unit
fZ = np.array(fZ.value, ndmin=1)*fZ.unit
fEZ = np.array(fEZ.value, ndmin=1)*fEZ.unit
intTimes = np.array(intTimes.value, ndmin=1)*intTimes.unit
assert len(intTimes) == len(sInds), "intTimes and sInds must be same length"
assert len(fEZ) == len(sInds), "fEZ must be an array of length len(sInds)"
assert len(fZ) == len(sInds), "fZ must be an array of length len(sInds)"
assert len(WA) == len(sInds), "WA must be an array of length len(sInds)"
# get scienceInstrument and starlightSuppressionSystem
inst = mode['inst']
syst = mode['syst']
# get mode wavelength
lam = mode['lam']
# get mode bandwidth (including any IFS spectral resolving power)
deltaLam = lam/inst['Rs'] if 'spec' in inst['name'].lower() else mode['deltaLam']
# get star magnitude
mV = TL.starMag(sInds, lam)
# get signal to noise ratio
SNR = mode['SNR']
# spectral flux density = F0 * A * Dlam * QE * T (attenuation due to optics)
attenuation = inst['optics']*syst['optics']
C_F0 = self.F0(lam)*self.pupilArea*deltaLam*inst['QE'](lam)*attenuation
# get core_thruput
core_thruput = syst['core_thruput'](lam, WA)
# calculate planet delta magnitude
dMagLim = np.zeros(len(sInds)) + 25
if (C_b is None) or (C_sp is None):
_, C_b, C_sp = self.Cp_Cb_Csp(TL, sInds, fZ, fEZ, dMagLim, WA, mode)
dMag = -2.5*np.log10((SNR*np.sqrt(C_b/intTimes + C_sp**2)/(C_F0*10.0**(-0.4*mV)*core_thruput*inst['PCeff'])).decompose().value)
return dMag
def ddMag_dt(self, intTimes, TL, sInds, fZ, fEZ, WA, mode, C_b=None, C_sp=None):
"""Finds derivative of achievable dMag with respect to integration time
Args:
intTimes (astropy Quantity array):
Integration times
TL (TargetList module):
TargetList class object
sInds (integer ndarray):
Integer indices of the stars of interest
fZ (astropy Quantity array):
Surface brightness of local zodiacal light for each star in sInds
in units of 1/arcsec2
fEZ (astropy Quantity array):
Surface brightness of exo-zodiacal light for each star in sInds
in units of 1/arcsec2
WA (astropy Quantity array):
Working angle for each star in sInds in units of arcsec
mode (dict):
Selected observing mode
C_b (astropy Quantity array):
Background noise electron count rate in units of 1/s (optional)
C_sp (astropy Quantity array):
Residual speckle spatial structure (systematic error) in units of 1/s
(optional)
Returns:
ddMagdt (ndarray):
Derivative of achievable dMag with respect to integration time
"""
# cast sInds, WA, fZ, fEZ, and intTimes to arrays
sInds = np.array(sInds, ndmin=1, copy=False)
WA = np.array(WA.value, ndmin=1)*WA.unit
fZ = np.array(fZ.value, ndmin=1)*fZ.unit
fEZ = np.array(fEZ.value, ndmin=1)*fEZ.unit
intTimes = np.array(intTimes.value, ndmin=1)*intTimes.unit
assert len(intTimes) == len(sInds), "intTimes and sInds must be same length"
assert len(fEZ) == len(sInds), "fEZ must be an array of length len(sInds)"
assert len(fZ) == len(sInds), "fZ must be an array of length len(sInds)"
assert len(WA) == len(sInds), "WA must be an array of length len(sInds)"
dMagLim = np.zeros(len(sInds)) + 25
if (C_b is None) or (C_sp is None):
_, C_b, C_sp = self.Cp_Cb_Csp(TL, sInds, fZ, fEZ, dMagLim, WA, mode)
ddMagdt = 2.5/(2.0*np.log(10.0))*(C_b/(C_b*intTimes + (C_sp*intTimes)**2)).to('1/s').value
return ddMagdt/u.s
|
[
"numpy.sqrt",
"numpy.log",
"EXOSIMS.Prototypes.OpticalSystem.OpticalSystem.__init__",
"numpy.array",
"numpy.errstate",
"numpy.isnan",
"numpy.true_divide",
"numpy.isinf"
] |
[((590, 627), 'EXOSIMS.Prototypes.OpticalSystem.OpticalSystem.__init__', 'OpticalSystem.__init__', (['self'], {}), '(self, **specs)\n', (612, 627), False, 'from EXOSIMS.Prototypes.OpticalSystem import OpticalSystem\n'), ((3905, 3941), 'numpy.array', 'np.array', (['sInds'], {'ndmin': '(1)', 'copy': '(False)'}), '(sInds, ndmin=1, copy=False)\n', (3913, 3941), True, 'import numpy as np\n'), ((7218, 7254), 'numpy.array', 'np.array', (['sInds'], {'ndmin': '(1)', 'copy': '(False)'}), '(sInds, ndmin=1, copy=False)\n', (7226, 7254), True, 'import numpy as np\n'), ((1997, 2043), 'numpy.errstate', 'np.errstate', ([], {'divide': '"""ignore"""', 'invalid': '"""ignore"""'}), "(divide='ignore', invalid='ignore')\n", (2008, 2043), True, 'import numpy as np\n'), ((2068, 2128), 'numpy.true_divide', 'np.true_divide', (['(SNR ** 2 * C_b)', '(C_p ** 2 - (SNR * C_sp) ** 2)'], {}), '(SNR ** 2 * C_b, C_p ** 2 - (SNR * C_sp) ** 2)\n', (2082, 2128), True, 'import numpy as np\n'), ((3956, 3983), 'numpy.array', 'np.array', (['WA.value'], {'ndmin': '(1)'}), '(WA.value, ndmin=1)\n', (3964, 3983), True, 'import numpy as np\n'), ((4006, 4033), 'numpy.array', 'np.array', (['fZ.value'], {'ndmin': '(1)'}), '(fZ.value, ndmin=1)\n', (4014, 4033), True, 'import numpy as np\n'), ((4057, 4085), 'numpy.array', 'np.array', (['fEZ.value'], {'ndmin': '(1)'}), '(fEZ.value, ndmin=1)\n', (4065, 4085), True, 'import numpy as np\n'), ((4115, 4148), 'numpy.array', 'np.array', (['intTimes.value'], {'ndmin': '(1)'}), '(intTimes.value, ndmin=1)\n', (4123, 4148), True, 'import numpy as np\n'), ((7269, 7296), 'numpy.array', 'np.array', (['WA.value'], {'ndmin': '(1)'}), '(WA.value, ndmin=1)\n', (7277, 7296), True, 'import numpy as np\n'), ((7319, 7346), 'numpy.array', 'np.array', (['fZ.value'], {'ndmin': '(1)'}), '(fZ.value, ndmin=1)\n', (7327, 7346), True, 'import numpy as np\n'), ((7370, 7398), 'numpy.array', 'np.array', (['fEZ.value'], {'ndmin': '(1)'}), '(fEZ.value, ndmin=1)\n', (7378, 7398), True, 'import numpy as np\n'), ((7428, 7461), 'numpy.array', 'np.array', (['intTimes.value'], {'ndmin': '(1)'}), '(intTimes.value, ndmin=1)\n', (7436, 7461), True, 'import numpy as np\n'), ((2182, 2199), 'numpy.isinf', 'np.isinf', (['intTime'], {}), '(intTime)\n', (2190, 2199), True, 'import numpy as np\n'), ((2202, 2219), 'numpy.isnan', 'np.isnan', (['intTime'], {}), '(intTime)\n', (2210, 2219), True, 'import numpy as np\n'), ((8020, 8032), 'numpy.log', 'np.log', (['(10.0)'], {}), '(10.0)\n', (8026, 8032), True, 'import numpy as np\n'), ((5589, 5624), 'numpy.sqrt', 'np.sqrt', (['(C_b / intTimes + C_sp ** 2)'], {}), '(C_b / intTimes + C_sp ** 2)\n', (5596, 5624), True, 'import numpy as np\n')]
|
#!/usr/bin/env python
import numpy as np
from olympus.surfaces import AbstractSurface
class AckleyPath(AbstractSurface):
def __init__(self, param_dim=2, noise=None):
"""Ackley path function.
Args:
param_dim (int): Number of input dimensions. Default is 2.
noise (Noise): Noise object that injects noise into the evaluations of the surface. Default is None.
"""
AbstractSurface.__init__(**locals())
@property
def minima(self):
# minimum at the centre
params = [0.5] * self.param_dim
value = self._run(params)
return [{'params': params, 'value': value}]
@property
def maxima(self):
return None
def _run(self, params):
params = np.array(params)
params = 64 * np.array(params) - 32 # rescale onto [-32, 32]
a = 20.
b = 0.2
c = 2 * np.pi
n = float(len(params))
params = np.array(params)
result = - a * np.exp(- b * np.sqrt(np.sum(params ** 2) / n)) - np.exp(np.sum(np.cos(c * params)) / n) + a + np.exp(1.)
if self.noise is None:
return result
else:
return self.noise(result)
|
[
"numpy.exp",
"numpy.array",
"numpy.sum",
"numpy.cos"
] |
[((761, 777), 'numpy.array', 'np.array', (['params'], {}), '(params)\n', (769, 777), True, 'import numpy as np\n'), ((950, 966), 'numpy.array', 'np.array', (['params'], {}), '(params)\n', (958, 966), True, 'import numpy as np\n'), ((1084, 1095), 'numpy.exp', 'np.exp', (['(1.0)'], {}), '(1.0)\n', (1090, 1095), True, 'import numpy as np\n'), ((800, 816), 'numpy.array', 'np.array', (['params'], {}), '(params)\n', (808, 816), True, 'import numpy as np\n'), ((1053, 1071), 'numpy.cos', 'np.cos', (['(c * params)'], {}), '(c * params)\n', (1059, 1071), True, 'import numpy as np\n'), ((1011, 1030), 'numpy.sum', 'np.sum', (['(params ** 2)'], {}), '(params ** 2)\n', (1017, 1030), True, 'import numpy as np\n')]
|
import random
import numpy as np
import cv2
from utils.transforms.transforms import CustomTransform
class RandomFlip(CustomTransform):
def __init__(self, prob_x=0, prob_y=0):
"""
Arguments:
----------
prob_x: range [0, 1], probability to use horizontal flip, setting to 0 means disabling flip
prob_y: range [0, 1], probability to use vertical flip
"""
self.prob_x = prob_x
self.prob_y = prob_y
def __call__(self, sample):
img = sample.get('img').copy()
segLabel = sample.get('segLabel', None)
if segLabel is not None:
segLabel = segLabel.copy()
flip_x = np.random.choice([False, True], p=(1 - self.prob_x, self.prob_x))
flip_y = np.random.choice([False, True], p=(1 - self.prob_y, self.prob_y))
if flip_x:
img = np.ascontiguousarray(np.flip(img, axis=1))
if segLabel is not None:
segLabel = np.ascontiguousarray(np.flip(segLabel, axis=1))
if flip_y:
img = np.ascontiguousarray(np.flip(img, axis=0))
if segLabel is not None:
segLabel = np.ascontiguousarray(np.flip(segLabel, axis=0))
_sample = sample.copy()
_sample['img'] = img
_sample['segLabel'] = segLabel
return _sample
class Darkness(CustomTransform):
def __init__(self, coeff):
assert coeff >= 1., "Darkness coefficient must be greater than 1"
self.coeff = coeff
def __call__(self, sample):
img = sample.get('img')
coeff = np.random.uniform(1., self.coeff)
img = (img.astype('float32') / coeff).astype('uint8')
_sample = sample.copy()
_sample['img'] = img
return _sample
|
[
"numpy.random.choice",
"numpy.flip",
"numpy.random.uniform"
] |
[((675, 740), 'numpy.random.choice', 'np.random.choice', (['[False, True]'], {'p': '(1 - self.prob_x, self.prob_x)'}), '([False, True], p=(1 - self.prob_x, self.prob_x))\n', (691, 740), True, 'import numpy as np\n'), ((758, 823), 'numpy.random.choice', 'np.random.choice', (['[False, True]'], {'p': '(1 - self.prob_y, self.prob_y)'}), '([False, True], p=(1 - self.prob_y, self.prob_y))\n', (774, 823), True, 'import numpy as np\n'), ((1581, 1615), 'numpy.random.uniform', 'np.random.uniform', (['(1.0)', 'self.coeff'], {}), '(1.0, self.coeff)\n', (1598, 1615), True, 'import numpy as np\n'), ((882, 902), 'numpy.flip', 'np.flip', (['img'], {'axis': '(1)'}), '(img, axis=1)\n', (889, 902), True, 'import numpy as np\n'), ((1075, 1095), 'numpy.flip', 'np.flip', (['img'], {'axis': '(0)'}), '(img, axis=0)\n', (1082, 1095), True, 'import numpy as np\n'), ((989, 1014), 'numpy.flip', 'np.flip', (['segLabel'], {'axis': '(1)'}), '(segLabel, axis=1)\n', (996, 1014), True, 'import numpy as np\n'), ((1182, 1207), 'numpy.flip', 'np.flip', (['segLabel'], {'axis': '(0)'}), '(segLabel, axis=0)\n', (1189, 1207), True, 'import numpy as np\n')]
|
'''Utility functions and classes for handling image datasets.'''
import cPickle
import cv2
import os.path as osp
import numpy as np
import tensorflow as tf
from config_tfvgg import cfg
FLAGS = tf.app.flags.FLAGS
def get_facebox_dims(img_shape,face_bbox,target_size,crop_size,spec,crop_ind):
face_bbox = np.zeros_like(face_bbox)
center = np.floor(np.array([face_bbox[2] + face_bbox[0], face_bbox[3] + face_bbox[1]]) / 2)
dims = np.array([face_bbox[2] - face_bbox[0], face_bbox[3] - face_bbox[1]]) * scale
face_bbox[0] = max(0, (center[0] - dims[0] / 2).astype(np.int32))
face_bbox[2] = min(img_shape[1], (center[0] + dims[0] / 2).astype(np.int32))
face_bbox[1] = max(0, (center[1] - dims[1] / 2).astype(np.int32))
face_bbox[3] = min(img_shape[0], (center[1] + dims[1] / 2).astype(np.int32))
(scale, isotropic, crop, mean) = (spec.scale_size, spec.isotropic, spec.crop_size)
img_shape = np.array((face_bbox[3]-face_bbox[1]+1,face_bbox[2]-face_bbox[0]+1))
min_length = np.min(img_shape)
new_shape = np.ceil((target_size * 1.0 / min_length) * img_shape)
offset = ((new_shape - crop) / 2).astype(np.int32)
return new_shape,offset
def process_image_reg(img_path,spec, flip=False, crop_ind=0, face_bbox=None,face_box_scale=1):
'''Crops, scales, and normalizes the given image.
scale : The image wil be first scaled to this size.
If isotropic is true, the smaller side is rescaled to this,
preserving the aspect ratio.
crop : After scaling, depending on crop_ind a crop of the image is given.
crope_ind: 0 center, 1 SW, 2 SE, 3 NE, 4 NW crop
flip: Whether to flip the image
mean : Subtracted from the image
'''
(scale, isotropic, crop, mean) = (spec.scale_size, spec.isotropic, spec.crop_size, spec.mean)
img = cv2.imread(img_path)
if face_bbox is not None:
face_bbox = np.array(face_bbox)
center =np.floor(np.array([face_bbox[2]+face_bbox[0],face_bbox[3]+face_bbox[1]])/2)
dims =np.array([face_bbox[2]-face_bbox[0],face_bbox[3]-face_bbox[1]]) * face_box_scale
face_bbox[0] = max(0,(center[0] - dims[0] / 2).astype(np.int32))
face_bbox[2] = min(img.shape[1],(center[0] + dims[0] / 2).astype(np.int32))
face_bbox[1] = max(0,(center[1] - dims[1] / 2).astype(np.int32))
face_bbox[3] = min(img.shape[0],(center[1] + dims[1] / 2).astype(np.int32))
img = img[face_bbox[1]:face_bbox[3],face_bbox[0]:face_bbox[2],:]
# Rescale
if flip:
img = img[:,::-1,:]
if isotropic:
img_shape = np.array(img.shape[:2])
min_length = np.min(img_shape)
new_shape = np.ceil((scale *1.0/ min_length) * img_shape)
else:
new_shape = np.array([scale, scale])
img = cv2.resize(img, tuple(new_shape.astype(np.int32).tolist()[::-1]))
offset = [0,0]
if crop_ind == 1:
offset[0] = new_shape[0]-crop
offset = new_shape-crop
elif crop_ind == 2:
offset = new_shape-crop
elif crop_ind == 3:
offset[1] = new_shape[1]-crop
elif crop_ind == 4:
offset = [0,0]
else:
offset = ((new_shape - crop) / 2).astype(np.int32)
img = img[offset[0]:offset[0]+crop,offset[1]:offset[1]+crop,:]
# Mean subtraction
return img.astype(np.float) - mean
def process_image(img, scale, isotropic, crop, mean, flip=False, crop_ind=0, face_bbox=None):
'''Crops, scales, and normalizes the given image.
scale : The image wil be first scaled to this size.
If isotropic is true, the smaller side is rescaled to this,
preserving the aspect ratio.
crop : After scaling, depending on crop_ind a crop of the image is given.
crope_ind: 0 center, 1 SW, 2 SE, 3 NE, 4 NW crop
flip: Whether to flip the image
mean : Subtracted from the image
'''
if face_bbox is not None:
img = tf.slice(img, begin=tf.pack([face_bbox[0], face_bbox[1], 0]), size=tf.pack([face_bbox[2]-face_bbox[0], face_bbox[3]-face_bbox[1], -1]))
# Rescale
if flip:
img = tf.reverse(img,[False,True,False])
if isotropic:
img_shape = tf.to_float(tf.shape(img)[:2])
min_length = tf.minimum(img_shape[0], img_shape[1])
new_shape = tf.to_int32((scale / min_length) * img_shape)
else:
new_shape = tf.pack([scale, scale])
img = tf.image.resize_images(img, new_shape)
offset = [0,0]
if crop_ind == 1:
offset[0] = new_shape[0]-crop
offset = new_shape-crop
elif crop_ind == 2:
offset = new_shape-crop
elif crop_ind == 3:
offset[1] = new_shape[1]-crop
elif crop_ind == 4:
offset = [0,0]
else:
offset = (new_shape - crop) / 2
img = tf.slice(img, begin=tf.pack([offset[0], offset[1], 0]), size=tf.pack([crop, crop, -1]))
# Mean subtraction
return tf.to_float(img) - mean
class ImageProducer(object):
'''
Loads and processes batches of images in parallel.
'''
def __init__(self, image_paths, data_spec, num_concurrent=4, batch_size=None, labels=None):
# The data specifications describe how to process the image
self.data_spec = data_spec
# A list of full image paths
self.image_paths = image_paths
# An optional list of labels corresponding to each image path
self.labels = labels
# A boolean flag per image indicating whether its a JPEG or PNG
self.extension_mask = self.create_extension_mask(self.image_paths)
# Create the loading and processing operations
self.setup(batch_size=batch_size, num_concurrent=num_concurrent)
def start(self, session, coordinator, num_concurrent=4):
'''Start the processing worker threads.'''
# Queue all paths
session.run(self.enqueue_paths_op)
# Close the path queue
session.run(self.close_path_queue_op)
# Start the queue runner and return the created threads
return self.queue_runner.create_threads(session, coord=coordinator, start=True)
def get(self, session):
'''
Get a single batch of images along with their indices. If a set of labels were provided,
the corresponding labels are returned instead of the indices.
'''
(indices, images) = session.run(self.dequeue_op)
if self.labels is not None:
labels = [self.labels[idx] for idx in indices]
return (labels, images)
return (indices, images)
def batches(self, session):
'''Yield a batch until no more images are left.'''
for _ in xrange(self.num_batches):
yield self.get(session=session)
def load_image(self, image_path, is_jpeg):
# Read the file
file_data = tf.read_file(image_path)
# Decode the image data
img = tf.cond(
is_jpeg,
lambda: tf.image.decode_jpeg(file_data, channels=self.data_spec.channels),
lambda: tf.image.decode_png(file_data, channels=self.data_spec.channels))
if self.data_spec.expects_bgr:
# Convert from RGB channel ordering to BGR
# This matches, for instance, how OpenCV orders the channels.
img = tf.reverse(img, [False, False, True])
return img
def process(self,crop_flip):
# Dequeue a single image path
idx, is_jpeg, image_path = self.path_bbox_queue.dequeue()
# Load the image
# Process the image
img_list = []
idx_list = []
for (c,f) in crop_flip:
img = self.load_image(image_path, is_jpeg)
processed_img = process_image(img=img,
scale=self.data_spec.scale_size,
isotropic=self.data_spec.isotropic,
crop=self.data_spec.crop_size,
mean=self.data_spec.mean,
flip=f,
crop_ind=c)
img_list.append(processed_img)
idx_list.append(idx)
# Return the processed image, along with its index
processed_idx_list = tf.pack(idx_list)
processed_img_list = tf.pack(img_list)
return (processed_idx_list, processed_img_list)
@staticmethod
def create_extension_mask(paths):
def is_jpeg(path):
extension = osp.splitext(path)[-1].lower()
if extension in ('.jpg', '.jpeg'):
return True
if extension != '.png':
raise ValueError('Unsupported image format: {}'.format(extension))
return False
return [is_jpeg(p) for p in paths]
def __len__(self):
return len(self.image_paths)
def setup(self, batch_size, num_concurrent):
pass
class VGGFaceProducer(ImageProducer):
def __init__(self, image_paths, data_spec ,num_concurrent=4,bbox_fp=None,labels=None):
round_rect = lambda x: [int(p) for p in x]
try:
v = self.face_bboxes
except AttributeError:
self.face_bboxes = None
if bbox_fp is not None:
face_bboxes=cPickle.load(open(bbox_fp,'r'))
self.face_bboxes = [round_rect(face_bboxes[p][0]) for p in image_paths]
# Initialize base
super(VGGFaceProducer, self).__init__(image_paths=image_paths,
data_spec=data_spec,num_concurrent=num_concurrent,labels=labels)
def setup(self, batch_size, num_concurrent):
# Validate the batch size
num_images = len(self.image_paths)
batch_size = min(num_images, batch_size or self.data_spec.batch_size)
if num_images % batch_size != 0:
raise ValueError(
'The total number of images ({}) must be divisible by the batch size ({}).'.format(
num_images, batch_size))
self.num_batches = num_images / batch_size
# Create a queue that will contain image paths (and their indices and extension indicator)
if self.face_bboxes is None:
self.path_bbox_queue = tf.FIFOQueue(capacity=num_images,
dtypes=[tf.int32, tf.bool, tf.string],
name='path_queue')
indices = tf.range(num_images)
self.enqueue_paths_op = self.path_bbox_queue.enqueue_many([indices, self.extension_mask,
self.image_paths])
else:
self.path_bbox_queue = tf.FIFOQueue(capacity=num_images,
dtypes=[tf.int32, tf.bool, tf.string, tf.int32],
name='path_queue')
indices = tf.range(num_images)
self.enqueue_paths_op = self.path_bbox_queue.enqueue_many([indices, self.extension_mask,
self.image_paths,self.face_bboxes])
# Close the path queue (no more additions)
self.close_path_queue_op = self.path_bbox_queue.close()
# Create an operation that dequeues a single path and returns a processed image
crop_flip = [[0,False]]
if cfg.CROP:
for i in range(1,5):
crop_flip.append([i,False])
if cfg.FLIP:
for i in range(len(crop_flip)):
crop_flip.append((crop_flip[i][0],True))
(processed_idx_list,processed_img_list) = self.process(crop_flip)
# Create a queue that will contain the processed images (and their indices)
image_shape = (self.data_spec.crop_size, self.data_spec.crop_size, self.data_spec.channels)
processed_queue = tf.FIFOQueue(capacity=int(np.ceil(len(crop_flip)*num_images / float(num_concurrent))),
dtypes=[tf.int32, tf.float32],
shapes=[(), image_shape],
name='processed_queue')
# Enqueue the processed image and path
enqueue_processed_op = processed_queue.enqueue_many([processed_idx_list,processed_img_list])
# Create a dequeue op that fetches a batch of processed images off the queue
[self.ind_deq,self.img_deq] = processed_queue.dequeue_many(batch_size)
self.dequeue_op = [self.ind_deq,self.img_deq]
# Create a queue runner to perform the processing operations in parallel
num_concurrent = min(num_concurrent, num_images)
self.queue_runner = tf.train.QueueRunner(processed_queue,
[enqueue_processed_op] * num_concurrent)
self.num_imgs = len(crop_flip)*num_images
self.num_feats_per_image = len(crop_flip)
def process(self,crop_flip):
# Dequeue a single image path
if self.face_bboxes is None:
idx, is_jpeg, image_path = self.path_bbox_queue.dequeue()
face_bbox = None
else:
idx, is_jpeg, image_path, face_bbox = self.path_bbox_queue.dequeue()
# Load the image
# Process the image
img_list = []
idx_list = []
for (c,f) in crop_flip:
img = self.load_image(image_path, is_jpeg)
processed_img = process_image(img=img,
scale=self.data_spec.scale_size,
isotropic=self.data_spec.isotropic,
crop=self.data_spec.crop_size,
mean=self.data_spec.mean,
flip=f,
crop_ind=c,
face_bbox=face_bbox)
img_list.append(processed_img)
idx_list.append(idx)
# Return the processed image, along with its index
processed_idx_list = tf.pack(idx_list)
processed_img_list = tf.pack(img_list)
return (processed_idx_list, processed_img_list)
class LFWProducer(VGGFaceProducer):
def __init__(self, val_path, data_path, data_spec,bbox_fp=None,num_concurrent=4,labels=None):
round_rect = lambda x: [int(p) for p in x]
image_paths = [osp.join(data_path, p) for p in val_path]
self.face_bboxes=None
if bbox_fp is not None:
face_bboxes=cPickle.load(open(bbox_fp,'r'))
self.face_bboxes = [round_rect(face_bboxes[p][0]) for p in val_path]
super(LFWProducer, self).__init__(image_paths=image_paths,
data_spec=data_spec,num_concurrent=num_concurrent,labels=labels)
|
[
"tensorflow.image.resize_images",
"tensorflow.shape",
"tensorflow.read_file",
"numpy.array",
"tensorflow.FIFOQueue",
"numpy.min",
"numpy.ceil",
"tensorflow.reverse",
"os.path.splitext",
"tensorflow.to_int32",
"tensorflow.range",
"cv2.imread",
"tensorflow.minimum",
"tensorflow.image.decode_png",
"tensorflow.to_float",
"os.path.join",
"tensorflow.pack",
"tensorflow.train.QueueRunner",
"numpy.zeros_like",
"tensorflow.image.decode_jpeg"
] |
[((309, 333), 'numpy.zeros_like', 'np.zeros_like', (['face_bbox'], {}), '(face_bbox)\n', (322, 333), True, 'import numpy as np\n'), ((925, 1001), 'numpy.array', 'np.array', (['(face_bbox[3] - face_bbox[1] + 1, face_bbox[2] - face_bbox[0] + 1)'], {}), '((face_bbox[3] - face_bbox[1] + 1, face_bbox[2] - face_bbox[0] + 1))\n', (933, 1001), True, 'import numpy as np\n'), ((1010, 1027), 'numpy.min', 'np.min', (['img_shape'], {}), '(img_shape)\n', (1016, 1027), True, 'import numpy as np\n'), ((1044, 1095), 'numpy.ceil', 'np.ceil', (['(target_size * 1.0 / min_length * img_shape)'], {}), '(target_size * 1.0 / min_length * img_shape)\n', (1051, 1095), True, 'import numpy as np\n'), ((1826, 1846), 'cv2.imread', 'cv2.imread', (['img_path'], {}), '(img_path)\n', (1836, 1846), False, 'import cv2\n'), ((4373, 4411), 'tensorflow.image.resize_images', 'tf.image.resize_images', (['img', 'new_shape'], {}), '(img, new_shape)\n', (4395, 4411), True, 'import tensorflow as tf\n'), ((441, 509), 'numpy.array', 'np.array', (['[face_bbox[2] - face_bbox[0], face_bbox[3] - face_bbox[1]]'], {}), '([face_bbox[2] - face_bbox[0], face_bbox[3] - face_bbox[1]])\n', (449, 509), True, 'import numpy as np\n'), ((1898, 1917), 'numpy.array', 'np.array', (['face_bbox'], {}), '(face_bbox)\n', (1906, 1917), True, 'import numpy as np\n'), ((2588, 2611), 'numpy.array', 'np.array', (['img.shape[:2]'], {}), '(img.shape[:2])\n', (2596, 2611), True, 'import numpy as np\n'), ((2633, 2650), 'numpy.min', 'np.min', (['img_shape'], {}), '(img_shape)\n', (2639, 2650), True, 'import numpy as np\n'), ((2671, 2716), 'numpy.ceil', 'np.ceil', (['(scale * 1.0 / min_length * img_shape)'], {}), '(scale * 1.0 / min_length * img_shape)\n', (2678, 2716), True, 'import numpy as np\n'), ((2747, 2771), 'numpy.array', 'np.array', (['[scale, scale]'], {}), '([scale, scale])\n', (2755, 2771), True, 'import numpy as np\n'), ((4078, 4115), 'tensorflow.reverse', 'tf.reverse', (['img', '[False, True, False]'], {}), '(img, [False, True, False])\n', (4088, 4115), True, 'import tensorflow as tf\n'), ((4204, 4242), 'tensorflow.minimum', 'tf.minimum', (['img_shape[0]', 'img_shape[1]'], {}), '(img_shape[0], img_shape[1])\n', (4214, 4242), True, 'import tensorflow as tf\n'), ((4263, 4306), 'tensorflow.to_int32', 'tf.to_int32', (['(scale / min_length * img_shape)'], {}), '(scale / min_length * img_shape)\n', (4274, 4306), True, 'import tensorflow as tf\n'), ((4339, 4362), 'tensorflow.pack', 'tf.pack', (['[scale, scale]'], {}), '([scale, scale])\n', (4346, 4362), True, 'import tensorflow as tf\n'), ((4871, 4887), 'tensorflow.to_float', 'tf.to_float', (['img'], {}), '(img)\n', (4882, 4887), True, 'import tensorflow as tf\n'), ((6770, 6794), 'tensorflow.read_file', 'tf.read_file', (['image_path'], {}), '(image_path)\n', (6782, 6794), True, 'import tensorflow as tf\n'), ((8224, 8241), 'tensorflow.pack', 'tf.pack', (['idx_list'], {}), '(idx_list)\n', (8231, 8241), True, 'import tensorflow as tf\n'), ((8271, 8288), 'tensorflow.pack', 'tf.pack', (['img_list'], {}), '(img_list)\n', (8278, 8288), True, 'import tensorflow as tf\n'), ((12703, 12781), 'tensorflow.train.QueueRunner', 'tf.train.QueueRunner', (['processed_queue', '([enqueue_processed_op] * num_concurrent)'], {}), '(processed_queue, [enqueue_processed_op] * num_concurrent)\n', (12723, 12781), True, 'import tensorflow as tf\n'), ((14098, 14115), 'tensorflow.pack', 'tf.pack', (['idx_list'], {}), '(idx_list)\n', (14105, 14115), True, 'import tensorflow as tf\n'), ((14145, 14162), 'tensorflow.pack', 'tf.pack', (['img_list'], {}), '(img_list)\n', (14152, 14162), True, 'import tensorflow as tf\n'), ((356, 424), 'numpy.array', 'np.array', (['[face_bbox[2] + face_bbox[0], face_bbox[3] + face_bbox[1]]'], {}), '([face_bbox[2] + face_bbox[0], face_bbox[3] + face_bbox[1]])\n', (364, 424), True, 'import numpy as np\n'), ((2024, 2092), 'numpy.array', 'np.array', (['[face_bbox[2] - face_bbox[0], face_bbox[3] - face_bbox[1]]'], {}), '([face_bbox[2] - face_bbox[0], face_bbox[3] - face_bbox[1]])\n', (2032, 2092), True, 'import numpy as np\n'), ((4769, 4803), 'tensorflow.pack', 'tf.pack', (['[offset[0], offset[1], 0]'], {}), '([offset[0], offset[1], 0])\n', (4776, 4803), True, 'import tensorflow as tf\n'), ((4810, 4835), 'tensorflow.pack', 'tf.pack', (['[crop, crop, -1]'], {}), '([crop, crop, -1])\n', (4817, 4835), True, 'import tensorflow as tf\n'), ((7230, 7267), 'tensorflow.reverse', 'tf.reverse', (['img', '[False, False, True]'], {}), '(img, [False, False, True])\n', (7240, 7267), True, 'import tensorflow as tf\n'), ((10206, 10301), 'tensorflow.FIFOQueue', 'tf.FIFOQueue', ([], {'capacity': 'num_images', 'dtypes': '[tf.int32, tf.bool, tf.string]', 'name': '"""path_queue"""'}), "(capacity=num_images, dtypes=[tf.int32, tf.bool, tf.string],\n name='path_queue')\n", (10218, 10301), True, 'import tensorflow as tf\n'), ((10408, 10428), 'tensorflow.range', 'tf.range', (['num_images'], {}), '(num_images)\n', (10416, 10428), True, 'import tensorflow as tf\n'), ((10665, 10771), 'tensorflow.FIFOQueue', 'tf.FIFOQueue', ([], {'capacity': 'num_images', 'dtypes': '[tf.int32, tf.bool, tf.string, tf.int32]', 'name': '"""path_queue"""'}), "(capacity=num_images, dtypes=[tf.int32, tf.bool, tf.string, tf.\n int32], name='path_queue')\n", (10677, 10771), True, 'import tensorflow as tf\n'), ((10885, 10905), 'tensorflow.range', 'tf.range', (['num_images'], {}), '(num_images)\n', (10893, 10905), True, 'import tensorflow as tf\n'), ((14428, 14450), 'os.path.join', 'osp.join', (['data_path', 'p'], {}), '(data_path, p)\n', (14436, 14450), True, 'import os.path as osp\n'), ((1943, 2011), 'numpy.array', 'np.array', (['[face_bbox[2] + face_bbox[0], face_bbox[3] + face_bbox[1]]'], {}), '([face_bbox[2] + face_bbox[0], face_bbox[3] + face_bbox[1]])\n', (1951, 2011), True, 'import numpy as np\n'), ((3920, 3960), 'tensorflow.pack', 'tf.pack', (['[face_bbox[0], face_bbox[1], 0]'], {}), '([face_bbox[0], face_bbox[1], 0])\n', (3927, 3960), True, 'import tensorflow as tf\n'), ((3967, 4038), 'tensorflow.pack', 'tf.pack', (['[face_bbox[2] - face_bbox[0], face_bbox[3] - face_bbox[1], -1]'], {}), '([face_bbox[2] - face_bbox[0], face_bbox[3] - face_bbox[1], -1])\n', (3974, 4038), True, 'import tensorflow as tf\n'), ((4164, 4177), 'tensorflow.shape', 'tf.shape', (['img'], {}), '(img)\n', (4172, 4177), True, 'import tensorflow as tf\n'), ((6891, 6956), 'tensorflow.image.decode_jpeg', 'tf.image.decode_jpeg', (['file_data'], {'channels': 'self.data_spec.channels'}), '(file_data, channels=self.data_spec.channels)\n', (6911, 6956), True, 'import tensorflow as tf\n'), ((6978, 7042), 'tensorflow.image.decode_png', 'tf.image.decode_png', (['file_data'], {'channels': 'self.data_spec.channels'}), '(file_data, channels=self.data_spec.channels)\n', (6997, 7042), True, 'import tensorflow as tf\n'), ((8454, 8472), 'os.path.splitext', 'osp.splitext', (['path'], {}), '(path)\n', (8466, 8472), True, 'import os.path as osp\n')]
|
import spartan
from spartan import core, expr, util, blob_ctx
import numpy as np
from .qr import qr
def svd(A, k=None):
"""
Stochastic SVD.
Parameters
----------
A : spartan matrix
Array to compute the SVD on, of shape (M, N)
k : int, optional
Number of singular values and vectors to compute.
The operations include matrix multiplication and QR decomposition.
We parallelize both of them.
Returns
--------
U : Spartan array of shape (M, k)
S : numpy array of shape (k,)
V : numpy array of shape (k, k)
"""
if k is None: k = A.shape[1]
Omega = expr.randn(A.shape[1], k)
Y = expr.dot(A, Omega)
Q, R = qr(Y)
B = expr.dot(expr.transpose(Q), A)
BTB = expr.dot(B, expr.transpose(B)).optimized().glom()
S, U_ = np.linalg.eig(BTB)
S = np.sqrt(S)
# Sort by eigen values from large to small
si = np.argsort(S)[::-1]
S = S[si]
U_ = U_[:, si]
U = expr.dot(Q, U_).optimized().evaluate()
V = np.dot(np.dot(expr.transpose(B).optimized().glom(), U_), np.diag(np.ones(S.shape[0]) / S))
return U, S, V.T
|
[
"spartan.expr.randn",
"spartan.expr.dot",
"numpy.sqrt",
"numpy.linalg.eig",
"numpy.ones",
"spartan.expr.transpose",
"numpy.argsort"
] |
[((594, 619), 'spartan.expr.randn', 'expr.randn', (['A.shape[1]', 'k'], {}), '(A.shape[1], k)\n', (604, 619), False, 'from spartan import core, expr, util, blob_ctx\n'), ((627, 645), 'spartan.expr.dot', 'expr.dot', (['A', 'Omega'], {}), '(A, Omega)\n', (635, 645), False, 'from spartan import core, expr, util, blob_ctx\n'), ((769, 787), 'numpy.linalg.eig', 'np.linalg.eig', (['BTB'], {}), '(BTB)\n', (782, 787), True, 'import numpy as np\n'), ((794, 804), 'numpy.sqrt', 'np.sqrt', (['S'], {}), '(S)\n', (801, 804), True, 'import numpy as np\n'), ((678, 695), 'spartan.expr.transpose', 'expr.transpose', (['Q'], {}), '(Q)\n', (692, 695), False, 'from spartan import core, expr, util, blob_ctx\n'), ((858, 871), 'numpy.argsort', 'np.argsort', (['S'], {}), '(S)\n', (868, 871), True, 'import numpy as np\n'), ((1024, 1043), 'numpy.ones', 'np.ones', (['S.shape[0]'], {}), '(S.shape[0])\n', (1031, 1043), True, 'import numpy as np\n'), ((914, 929), 'spartan.expr.dot', 'expr.dot', (['Q', 'U_'], {}), '(Q, U_)\n', (922, 929), False, 'from spartan import core, expr, util, blob_ctx\n'), ((720, 737), 'spartan.expr.transpose', 'expr.transpose', (['B'], {}), '(B)\n', (734, 737), False, 'from spartan import core, expr, util, blob_ctx\n'), ((973, 990), 'spartan.expr.transpose', 'expr.transpose', (['B'], {}), '(B)\n', (987, 990), False, 'from spartan import core, expr, util, blob_ctx\n')]
|
from collections import namedtuple
import tensorflow as tf
import numpy as np
from rl.agents.a2c.agent import A2CAgent
TestArgType = namedtuple('ArgType', ['name'])
arg_type = TestArgType('arg')
A = np.array
class A2CAgentTest(tf.test.TestCase):
def test_compute_policy_log_probs(self):
from rl.agents.a2c.agent import compute_policy_log_probs
available_actions = A([[1, 0, 1],
[1, 0, 0],
[1, 1, 1]], dtype=np.float32)
fn_pi = A([[0.2, 0.0, 0.8],
[1.0, 0.0, 0.0],
[0.2, 0.7, 0.1]], dtype=np.float32)
fn_ids = A([2, 0, 1], dtype=np.int32)
arg_pi = {arg_type: A([[0.8, 0.2],
[0.0, 1.0],
[0.5, 0.5]], dtype=np.float32)}
arg_ids = {arg_type: A([0, 1, -1], dtype=np.int32)}
log_probs = compute_policy_log_probs(
available_actions, (fn_pi, arg_pi), (fn_ids, arg_ids)
)
expected_log_probs = np.log([0.8, 1.0, 0.7]) + A([np.log(0.8), np.log(1.0), 0])
with self.test_session() as sess:
log_probs_out = sess.run(log_probs)
self.assertAllClose(log_probs_out, expected_log_probs)
def test_compute_policy_entropy(self):
from rl.agents.a2c.agent import compute_policy_entropy
available_actions = A([[1, 0, 1],
[1, 0, 0],
[1, 1, 1]], dtype=np.float32)
fn_pi = A([[0.2, 0.0, 0.8],
[1.0, 0.0, 0.0],
[0.2, 0.7, 0.1]], dtype=np.float32)
fn_ids = A([2, 0, 1], dtype=np.int32)
arg_pi = {arg_type: A([[0.8, 0.2],
[0.0, 1.0],
[0.5, 0.5]], dtype=np.float32)}
arg_ids = {arg_type: A([0, 1, -1], dtype=np.int32)}
entropy = compute_policy_entropy(
available_actions, (fn_pi, arg_pi), (fn_ids, arg_ids)
)
expected_entropy = (0.50040245 + 0.80181855) / 3.0 + (0.50040245) / 2
with self.test_session() as sess:
entropy_out = sess.run(entropy)
self.assertAllClose(entropy_out, expected_entropy)
if __name__ == '__main__':
tf.test.main()
|
[
"collections.namedtuple",
"rl.agents.a2c.agent.compute_policy_entropy",
"numpy.log",
"tensorflow.test.main",
"rl.agents.a2c.agent.compute_policy_log_probs"
] |
[((137, 168), 'collections.namedtuple', 'namedtuple', (['"""ArgType"""', "['name']"], {}), "('ArgType', ['name'])\n", (147, 168), False, 'from collections import namedtuple\n'), ((2115, 2129), 'tensorflow.test.main', 'tf.test.main', ([], {}), '()\n', (2127, 2129), True, 'import tensorflow as tf\n'), ((862, 941), 'rl.agents.a2c.agent.compute_policy_log_probs', 'compute_policy_log_probs', (['available_actions', '(fn_pi, arg_pi)', '(fn_ids, arg_ids)'], {}), '(available_actions, (fn_pi, arg_pi), (fn_ids, arg_ids))\n', (886, 941), False, 'from rl.agents.a2c.agent import compute_policy_log_probs\n'), ((1785, 1862), 'rl.agents.a2c.agent.compute_policy_entropy', 'compute_policy_entropy', (['available_actions', '(fn_pi, arg_pi)', '(fn_ids, arg_ids)'], {}), '(available_actions, (fn_pi, arg_pi), (fn_ids, arg_ids))\n', (1807, 1862), False, 'from rl.agents.a2c.agent import compute_policy_entropy\n'), ((980, 1003), 'numpy.log', 'np.log', (['[0.8, 1.0, 0.7]'], {}), '([0.8, 1.0, 0.7])\n', (986, 1003), True, 'import numpy as np\n'), ((1009, 1020), 'numpy.log', 'np.log', (['(0.8)'], {}), '(0.8)\n', (1015, 1020), True, 'import numpy as np\n'), ((1022, 1033), 'numpy.log', 'np.log', (['(1.0)'], {}), '(1.0)\n', (1028, 1033), True, 'import numpy as np\n')]
|
# -*- coding: utf-8 -*-
import pandas as pd
import numpy as np
def rescale(data, to=[0, 1]):
"""Rescale data.
Rescale a numeric variable to a new range.
Parameters
----------
data : list, array or Series
Raw data.
to : list
New range of values of the data after rescaling.
Returns
----------
list, array or Series
The rescaled values.
Examples
----------
>>> import neurokit2 as nk
>>>
>>> nk.rescale(data=[3, 1, 2, 4, 6], to=[0, 1])
"""
# Return appropriate type
if isinstance(data, list):
data = list(_rescale(np.array(data), to=to))
else:
data = _rescale(data, to=to)
return data
def _rescale(data, to=[0, 1]):
y = (to[1] - to[0]) / (np.nanmax(data) - np.nanmin(data)) * (data - np.nanmin(data)) + to[0]
return y
|
[
"numpy.nanmin",
"numpy.array",
"numpy.nanmax"
] |
[((623, 637), 'numpy.array', 'np.array', (['data'], {}), '(data)\n', (631, 637), True, 'import numpy as np\n'), ((821, 836), 'numpy.nanmin', 'np.nanmin', (['data'], {}), '(data)\n', (830, 836), True, 'import numpy as np\n'), ((776, 791), 'numpy.nanmax', 'np.nanmax', (['data'], {}), '(data)\n', (785, 791), True, 'import numpy as np\n'), ((794, 809), 'numpy.nanmin', 'np.nanmin', (['data'], {}), '(data)\n', (803, 809), True, 'import numpy as np\n')]
|
#!/usr/bin/env python
# coding: utf-8
# In[1]:
#Libraries
import cv2
import numpy as np
import pyautogui
import keyboard
# In[2]:
#Color to detect BGR
l = [17, 15, 100] #lower
u = [80, 76, 220] #upper
# In[3]:
#region coordinates
k_left, k_top, k_right, k_bottom = 640, 30, 440, 130
h_left, h_top, h_right, h_bottom = 440, 130, 240, 330
s_left, s_top, s_right, s_bottom = 840, 130, 640, 330
f_left, f_top, f_right, f_bottom = 640, 330, 440, 430
# In[4]:
#Key Pressed
current_key_pressed = set()
# In[5]:
#Accelerate
def up():
#print("W")
pyautogui.keyDown('up')
current_key_pressed.add('w')
# In[6]:
#Steering Right
def right():
#print("D")
pyautogui.keyDown('right')
current_key_pressed.add('d')
# In[7]:
#Steering Left
def left():
#print("A")
pyautogui.keyDown('left')
current_key_pressed.add('a')
# In[8]:
#Brakes
def down():
#print("S")
pyautogui.keyDown('down')
current_key_pressed.add('s')
# In[9]:
#Find contours
def findContours(image):
img = cv2.cvtColor(image,cv2.COLOR_BGR2GRAY)
threshold = cv2.threshold(img, 15, 255, cv2.THRESH_BINARY)[1]
(_, cnts, _) = cv2.findContours(threshold.copy(),cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)
return len(cnts)
# In[10]:
#Main function
if __name__=='__main__':
aWeight=0.5
cam=cv2.VideoCapture(0)
cam.set(3,1280)
cam.set(4,720)
cam.set(cv2.CAP_PROP_FPS,60)
while True:
buttonPressed = False
buttonPressed_leftright = False
status, frame = cam.read()
clone = frame.copy()
clone = cv2.flip(clone,1)
clone = cv2.resize(clone,(1280,720))
reg_up = clone[k_top:k_bottom, k_right:k_left]
reg_left = clone[h_top:h_bottom, h_right:h_left]
reg_right = clone[s_top:s_bottom, s_right:s_left]
reg_down = clone[f_top:f_bottom, f_right:f_left]
reg_up = cv2.GaussianBlur(reg_up, (7,7), 0)
reg_right = cv2.GaussianBlur(reg_right, (7,7), 0)
reg_left = cv2.GaussianBlur(reg_left, (7,7), 0)
reg_down = cv2.GaussianBlur(reg_down, (7,7), 0)
hsv = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV)
mask = cv2.inRange(hsv, l, u)
mask = cv2.erode(mask, None, iterations=2)
mask = cv2.dilate(mask, None, iterations=2)
l = np.array(lower, dtype="uint8")
u = np.array(upper, dtype="uint8")
mask_up = cv2.inRange(reg_up, l, u)
mask_right = cv2.inRange(reg_right, l, u)
mask_left = cv2.inRange(reg_left, l, u)
mask_down = cv2.inRange(reg_down, l, u)
out_up = cv2.bitwise_and(reg_up, reg_up, mask=mask_up)
out_right = cv2.bitwise_and(reg_right, reg_right, mask=mask_right)
out_left = cv2.bitwise_and(reg_left, reg_left, mask=mask_left)
out_down = cv2.bitwise_and(reg_down, reg_down, mask=mask_down)
cnts_up = findContours(out_up)
cnts_right = findContours(out_right)
cnts_left = findContours(out_left)
cnts_down = findContours(out_down)
if (cnts_up > 0):
up()
buttonPressed = True
elif (cnts_right > 0):
right()
buttonPressed = True
buttonPressed_leftright = True
elif (cnts_left > 0):
left()
buttonPressed = True
buttonPressed_leftright = True
elif (cnts_down > 0):
down()
buttonPressed = True
image_up = cv2.rectangle(clone, (k_left, k_top), (k_right, k_bottom), (255,0,255,0.5), 2)
image_left = cv2.rectangle(clone, (h_left, h_top), (h_right, h_bottom), (255,0,0,0.5), 2)
image_right = cv2.rectangle(clone, (s_left, s_top), (s_right, s_bottom), (0,0,255,0.5), 2)
image_down = cv2.rectangle(clone, (f_left, f_top), (f_right, f_bottom), (0,255,255,0.5), 2)
cv2.putText(image_up, "W", (k_left-170,k_top+110), cv2.FONT_HERSHEY_SIMPLEX, 0.9, (36,255,12), 2)
cv2.putText(image_left, "A", (h_left-170,h_top+200), cv2.FONT_HERSHEY_SIMPLEX, 0.9, (36,255,12), 2)
cv2.putText(image_right, "D", (s_left-170,s_top+200), cv2.FONT_HERSHEY_SIMPLEX, 0.9, (36,255,12), 2)
cv2.putText(image_down, "S", (f_left-170,f_top+110), cv2.FONT_HERSHEY_SIMPLEX, 0.9, (36,255,12), 2)
cv2.namedWindow("video",cv2.WINDOW_AUTOSIZE)
cv2.imshow("video", clone)
if not buttonPressed and len(current_key_pressed) != 0:
for key in current_key_pressed:
pyautogui.keyUp(key)
current_key_pressed = set()
if not buttonPressed_leftright and (('a' in current_key_pressed) or ('d' in current_key_pressed)):
if 'a' in current_key_pressed:
pyautogui.keyUp('left')
current_key_pressed.remove('a')
elif 'd' in current_key_pressed:
pyautogui.keyUp('right')
current_key_pressed.remove('d')
if cv2.waitKey(1) & 0Xff == ord('q'):
break
cam.release()
cv2.destroyAllWindows()
|
[
"cv2.rectangle",
"cv2.imshow",
"numpy.array",
"cv2.destroyAllWindows",
"cv2.threshold",
"cv2.erode",
"cv2.waitKey",
"pyautogui.keyDown",
"cv2.putText",
"cv2.cvtColor",
"cv2.resize",
"cv2.GaussianBlur",
"cv2.namedWindow",
"cv2.flip",
"cv2.inRange",
"cv2.bitwise_and",
"cv2.VideoCapture",
"pyautogui.keyUp",
"cv2.dilate"
] |
[((566, 589), 'pyautogui.keyDown', 'pyautogui.keyDown', (['"""up"""'], {}), "('up')\n", (583, 589), False, 'import pyautogui\n'), ((685, 711), 'pyautogui.keyDown', 'pyautogui.keyDown', (['"""right"""'], {}), "('right')\n", (702, 711), False, 'import pyautogui\n'), ((805, 830), 'pyautogui.keyDown', 'pyautogui.keyDown', (['"""left"""'], {}), "('left')\n", (822, 830), False, 'import pyautogui\n'), ((917, 942), 'pyautogui.keyDown', 'pyautogui.keyDown', (['"""down"""'], {}), "('down')\n", (934, 942), False, 'import pyautogui\n'), ((1039, 1078), 'cv2.cvtColor', 'cv2.cvtColor', (['image', 'cv2.COLOR_BGR2GRAY'], {}), '(image, cv2.COLOR_BGR2GRAY)\n', (1051, 1078), False, 'import cv2\n'), ((1340, 1359), 'cv2.VideoCapture', 'cv2.VideoCapture', (['(0)'], {}), '(0)\n', (1356, 1359), False, 'import cv2\n'), ((5079, 5102), 'cv2.destroyAllWindows', 'cv2.destroyAllWindows', ([], {}), '()\n', (5100, 5102), False, 'import cv2\n'), ((1094, 1140), 'cv2.threshold', 'cv2.threshold', (['img', '(15)', '(255)', 'cv2.THRESH_BINARY'], {}), '(img, 15, 255, cv2.THRESH_BINARY)\n', (1107, 1140), False, 'import cv2\n'), ((1610, 1628), 'cv2.flip', 'cv2.flip', (['clone', '(1)'], {}), '(clone, 1)\n', (1618, 1628), False, 'import cv2\n'), ((1644, 1674), 'cv2.resize', 'cv2.resize', (['clone', '(1280, 720)'], {}), '(clone, (1280, 720))\n', (1654, 1674), False, 'import cv2\n'), ((1919, 1954), 'cv2.GaussianBlur', 'cv2.GaussianBlur', (['reg_up', '(7, 7)', '(0)'], {}), '(reg_up, (7, 7), 0)\n', (1935, 1954), False, 'import cv2\n'), ((1974, 2012), 'cv2.GaussianBlur', 'cv2.GaussianBlur', (['reg_right', '(7, 7)', '(0)'], {}), '(reg_right, (7, 7), 0)\n', (1990, 2012), False, 'import cv2\n'), ((2031, 2068), 'cv2.GaussianBlur', 'cv2.GaussianBlur', (['reg_left', '(7, 7)', '(0)'], {}), '(reg_left, (7, 7), 0)\n', (2047, 2068), False, 'import cv2\n'), ((2087, 2124), 'cv2.GaussianBlur', 'cv2.GaussianBlur', (['reg_down', '(7, 7)', '(0)'], {}), '(reg_down, (7, 7), 0)\n', (2103, 2124), False, 'import cv2\n'), ((2147, 2185), 'cv2.cvtColor', 'cv2.cvtColor', (['frame', 'cv2.COLOR_BGR2HSV'], {}), '(frame, cv2.COLOR_BGR2HSV)\n', (2159, 2185), False, 'import cv2\n'), ((2210, 2232), 'cv2.inRange', 'cv2.inRange', (['hsv', 'l', 'u'], {}), '(hsv, l, u)\n', (2221, 2232), False, 'import cv2\n'), ((2248, 2283), 'cv2.erode', 'cv2.erode', (['mask', 'None'], {'iterations': '(2)'}), '(mask, None, iterations=2)\n', (2257, 2283), False, 'import cv2\n'), ((2299, 2335), 'cv2.dilate', 'cv2.dilate', (['mask', 'None'], {'iterations': '(2)'}), '(mask, None, iterations=2)\n', (2309, 2335), False, 'import cv2\n'), ((2349, 2379), 'numpy.array', 'np.array', (['lower'], {'dtype': '"""uint8"""'}), "(lower, dtype='uint8')\n", (2357, 2379), True, 'import numpy as np\n'), ((2392, 2422), 'numpy.array', 'np.array', (['upper'], {'dtype': '"""uint8"""'}), "(upper, dtype='uint8')\n", (2400, 2422), True, 'import numpy as np\n'), ((2442, 2467), 'cv2.inRange', 'cv2.inRange', (['reg_up', 'l', 'u'], {}), '(reg_up, l, u)\n', (2453, 2467), False, 'import cv2\n'), ((2489, 2517), 'cv2.inRange', 'cv2.inRange', (['reg_right', 'l', 'u'], {}), '(reg_right, l, u)\n', (2500, 2517), False, 'import cv2\n'), ((2538, 2565), 'cv2.inRange', 'cv2.inRange', (['reg_left', 'l', 'u'], {}), '(reg_left, l, u)\n', (2549, 2565), False, 'import cv2\n'), ((2586, 2613), 'cv2.inRange', 'cv2.inRange', (['reg_down', 'l', 'u'], {}), '(reg_down, l, u)\n', (2597, 2613), False, 'import cv2\n'), ((2632, 2677), 'cv2.bitwise_and', 'cv2.bitwise_and', (['reg_up', 'reg_up'], {'mask': 'mask_up'}), '(reg_up, reg_up, mask=mask_up)\n', (2647, 2677), False, 'import cv2\n'), ((2698, 2752), 'cv2.bitwise_and', 'cv2.bitwise_and', (['reg_right', 'reg_right'], {'mask': 'mask_right'}), '(reg_right, reg_right, mask=mask_right)\n', (2713, 2752), False, 'import cv2\n'), ((2772, 2823), 'cv2.bitwise_and', 'cv2.bitwise_and', (['reg_left', 'reg_left'], {'mask': 'mask_left'}), '(reg_left, reg_left, mask=mask_left)\n', (2787, 2823), False, 'import cv2\n'), ((2843, 2894), 'cv2.bitwise_and', 'cv2.bitwise_and', (['reg_down', 'reg_down'], {'mask': 'mask_down'}), '(reg_down, reg_down, mask=mask_down)\n', (2858, 2894), False, 'import cv2\n'), ((3514, 3600), 'cv2.rectangle', 'cv2.rectangle', (['clone', '(k_left, k_top)', '(k_right, k_bottom)', '(255, 0, 255, 0.5)', '(2)'], {}), '(clone, (k_left, k_top), (k_right, k_bottom), (255, 0, 255, \n 0.5), 2)\n', (3527, 3600), False, 'import cv2\n'), ((3614, 3693), 'cv2.rectangle', 'cv2.rectangle', (['clone', '(h_left, h_top)', '(h_right, h_bottom)', '(255, 0, 0, 0.5)', '(2)'], {}), '(clone, (h_left, h_top), (h_right, h_bottom), (255, 0, 0, 0.5), 2)\n', (3627, 3693), False, 'import cv2\n'), ((3713, 3792), 'cv2.rectangle', 'cv2.rectangle', (['clone', '(s_left, s_top)', '(s_right, s_bottom)', '(0, 0, 255, 0.5)', '(2)'], {}), '(clone, (s_left, s_top), (s_right, s_bottom), (0, 0, 255, 0.5), 2)\n', (3726, 3792), False, 'import cv2\n'), ((3811, 3897), 'cv2.rectangle', 'cv2.rectangle', (['clone', '(f_left, f_top)', '(f_right, f_bottom)', '(0, 255, 255, 0.5)', '(2)'], {}), '(clone, (f_left, f_top), (f_right, f_bottom), (0, 255, 255, \n 0.5), 2)\n', (3824, 3897), False, 'import cv2\n'), ((3905, 4014), 'cv2.putText', 'cv2.putText', (['image_up', '"""W"""', '(k_left - 170, k_top + 110)', 'cv2.FONT_HERSHEY_SIMPLEX', '(0.9)', '(36, 255, 12)', '(2)'], {}), "(image_up, 'W', (k_left - 170, k_top + 110), cv2.\n FONT_HERSHEY_SIMPLEX, 0.9, (36, 255, 12), 2)\n", (3916, 4014), False, 'import cv2\n'), ((4011, 4122), 'cv2.putText', 'cv2.putText', (['image_left', '"""A"""', '(h_left - 170, h_top + 200)', 'cv2.FONT_HERSHEY_SIMPLEX', '(0.9)', '(36, 255, 12)', '(2)'], {}), "(image_left, 'A', (h_left - 170, h_top + 200), cv2.\n FONT_HERSHEY_SIMPLEX, 0.9, (36, 255, 12), 2)\n", (4022, 4122), False, 'import cv2\n'), ((4119, 4231), 'cv2.putText', 'cv2.putText', (['image_right', '"""D"""', '(s_left - 170, s_top + 200)', 'cv2.FONT_HERSHEY_SIMPLEX', '(0.9)', '(36, 255, 12)', '(2)'], {}), "(image_right, 'D', (s_left - 170, s_top + 200), cv2.\n FONT_HERSHEY_SIMPLEX, 0.9, (36, 255, 12), 2)\n", (4130, 4231), False, 'import cv2\n'), ((4228, 4339), 'cv2.putText', 'cv2.putText', (['image_down', '"""S"""', '(f_left - 170, f_top + 110)', 'cv2.FONT_HERSHEY_SIMPLEX', '(0.9)', '(36, 255, 12)', '(2)'], {}), "(image_down, 'S', (f_left - 170, f_top + 110), cv2.\n FONT_HERSHEY_SIMPLEX, 0.9, (36, 255, 12), 2)\n", (4239, 4339), False, 'import cv2\n'), ((4337, 4382), 'cv2.namedWindow', 'cv2.namedWindow', (['"""video"""', 'cv2.WINDOW_AUTOSIZE'], {}), "('video', cv2.WINDOW_AUTOSIZE)\n", (4352, 4382), False, 'import cv2\n'), ((4390, 4416), 'cv2.imshow', 'cv2.imshow', (['"""video"""', 'clone'], {}), "('video', clone)\n", (4400, 4416), False, 'import cv2\n'), ((4542, 4562), 'pyautogui.keyUp', 'pyautogui.keyUp', (['key'], {}), '(key)\n', (4557, 4562), False, 'import pyautogui\n'), ((4783, 4806), 'pyautogui.keyUp', 'pyautogui.keyUp', (['"""left"""'], {}), "('left')\n", (4798, 4806), False, 'import pyautogui\n'), ((5003, 5017), 'cv2.waitKey', 'cv2.waitKey', (['(1)'], {}), '(1)\n', (5014, 5017), False, 'import cv2\n'), ((4918, 4942), 'pyautogui.keyUp', 'pyautogui.keyUp', (['"""right"""'], {}), "('right')\n", (4933, 4942), False, 'import pyautogui\n')]
|
# Copyright 2018 The TensorFlow Probability Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""Tests for the JointDistributionAutoBatched."""
import collections
import os
# Dependency imports
from absl.testing import parameterized
import numpy as np
import tensorflow.compat.v1 as tf1
import tensorflow.compat.v2 as tf
import tensorflow_probability as tfp
from tensorflow_probability.python.internal import test_util
tfb = tfp.bijectors
tfd = tfp.distributions
JAX_MODE = False
Root = tfd.JointDistributionCoroutineAutoBatched.Root
@test_util.test_all_tf_execution_regimes
class JointDistributionAutoBatchedTest(test_util.TestCase):
@parameterized.named_parameters(
{'testcase_name': 'coroutine',
'jd_class': tfd.JointDistributionCoroutineAutoBatched},
{'testcase_name': 'sequential',
'jd_class': tfd.JointDistributionSequentialAutoBatched},
{'testcase_name': 'named',
'jd_class': tfd.JointDistributionNamedAutoBatched})
def test_batch_and_event_shape_with_plate(self, jd_class):
models = {}
def coroutine_model():
g = yield tfd.LogNormal(0., 1.)
df = yield tfd.Exponential(1.)
loc = yield tfd.Sample(tfd.Normal(0, g), 20)
yield tfd.StudentT(tf.expand_dims(df, -1), loc, 1)
models[tfd.JointDistributionCoroutineAutoBatched] = coroutine_model
models[tfd.JointDistributionSequentialAutoBatched] = [
tfd.LogNormal(0., 1.),
tfd.Exponential(1.),
lambda _, g: tfd.Sample(tfd.Normal(0, g), 20),
lambda loc, df: tfd.StudentT(tf.expand_dims(df, -1), loc, 1)
]
models[tfd.JointDistributionNamedAutoBatched] = collections.OrderedDict((
('g', tfd.LogNormal(0., 1.)),
('df', tfd.Exponential(1.)),
('loc', lambda g: tfd.Sample(tfd.Normal(0, g), 20)),
('x', lambda loc, df: tfd.StudentT(tf.expand_dims(df, -1), loc, 1))))
joint = jd_class(models[jd_class], validate_args=True)
# Properties `event_shape` and `batch_shape` should be defined
# even before any sampling calls have occurred.
self.assertAllEqual(joint._model_flatten(joint.event_shape),
[[], [], [20], [20]])
self.assertAllEqual(joint.batch_shape, [])
is_scalar = joint._model_flatten(joint.is_scalar_event())
self.assertAllEqual(is_scalar[0], True)
self.assertAllEqual(is_scalar[1], True)
self.assertAllEqual(is_scalar[2], False)
self.assertAllEqual(is_scalar[3], False)
event_shape = joint._model_flatten(joint.event_shape_tensor())
self.assertAllEqual(event_shape[0], [])
self.assertAllEqual(event_shape[1], [])
self.assertAllEqual(event_shape[2], [20])
self.assertAllEqual(event_shape[3], [20])
self.assertEqual(joint.is_scalar_batch(), True)
batch_shape = joint.batch_shape_tensor()
self.assertAllEqual(batch_shape, [])
@parameterized.named_parameters(
*(dict( # pylint: disable=g-complex-comprehension
testcase_name=jd_type + '_' + sampler_type,
jd_class=getattr(tfd, 'JointDistribution' + jd_type + 'AutoBatched'),
sampler_type=sampler_type)
for jd_type in ('Coroutine', 'Sequential', 'Named')
for sampler_type in ('stateful', 'stateless')))
def test_model_with_nontrivial_batch_shape(self, jd_class, sampler_type):
models = {}
def coroutine_model():
g = yield tfd.LogNormal(0., [1., 2.])
df = yield tfd.Exponential([1., 2.])
loc = yield tfd.Sample(tfd.Normal(0, g), 20)
yield tfd.StudentT(tf.expand_dims(df, -1), loc, 1)
models[tfd.JointDistributionCoroutineAutoBatched] = coroutine_model
models[tfd.JointDistributionSequentialAutoBatched] = [
tfd.LogNormal(0., [1., 2.]),
tfd.Exponential([1., 2.]),
lambda _, g: tfd.Sample(tfd.Normal(0, g), 20),
lambda loc, df: tfd.StudentT(tf.expand_dims(df, -1), loc, 1)
]
models[tfd.JointDistributionNamedAutoBatched] = collections.OrderedDict((
('g', tfd.LogNormal(0., [1., 2.])),
('df', tfd.Exponential([1., 2.])),
('loc', lambda g: tfd.Sample(tfd.Normal(0, g), 20)),
('x', lambda loc, df: tfd.StudentT(tf.expand_dims(df, -1), loc, 1))))
joint = jd_class(models[jd_class], batch_ndims=1, validate_args=True)
self.assertAllEqual(joint._model_flatten(joint.event_shape),
[[], [], [20], [20]])
self.assertAllEqual(joint.batch_shape, [2])
is_scalar = joint._model_flatten(joint.is_scalar_event())
self.assertAllEqual(is_scalar[0], True)
self.assertAllEqual(is_scalar[1], True)
self.assertAllEqual(is_scalar[2], False)
self.assertAllEqual(is_scalar[3], False)
self.assertAllEqual(joint.is_scalar_batch(), False)
batch_shape = self.evaluate(joint.batch_shape_tensor())
self.assertAllEqual(batch_shape, [2])
x = joint.sample([5], seed=test_util.test_seed(sampler_type=sampler_type))
lp = self.evaluate(joint.log_prob(x))
self.assertAllEqual(lp.shape, [5, 2])
def test_model_with_dynamic_batch_ndims(self):
if tf.executing_eagerly():
self.skipTest('Dynamic shape.')
def coroutine_model():
g = yield tfd.LogNormal(0., [1., 2.])
df = yield tfd.Exponential([1., 2.])
loc = yield tfd.Sample(tfd.Normal(0, g), 20)
yield tfd.StudentT(tf.expand_dims(df, -1), loc, 1)
joint = tfd.JointDistributionCoroutineAutoBatched(
coroutine_model,
batch_ndims=tf1.placeholder_with_default(1, shape=[]),
validate_args=True)
batch_shape_tensor = self.evaluate(joint.batch_shape_tensor())
self.assertAllEqual(batch_shape_tensor, [2])
event_shape_tensor = self.evaluate(joint.event_shape_tensor())
self.assertAllEqual(event_shape_tensor[0], [])
self.assertAllEqual(event_shape_tensor[1], [])
self.assertAllEqual(event_shape_tensor[2], [20])
self.assertAllEqual(event_shape_tensor[3], [20])
self.assertAllEqual(joint.batch_shape, tf.TensorShape(None))
self.assertAllEqual(joint._model_flatten(joint.event_shape),
[tf.TensorShape(None)] * 4)
x = joint.sample([5], seed=test_util.test_seed(sampler_type='stateless'))
lp = self.evaluate(joint.log_prob(x))
self.assertAllEqual(lp.shape, [5, 2])
@parameterized.named_parameters(
{'testcase_name': 'coroutine',
'base_jd_class': tfd.JointDistributionCoroutine,
'jda_class': tfd.JointDistributionCoroutineAutoBatched},
{'testcase_name': 'sequential',
'base_jd_class': tfd.JointDistributionSequential,
'jda_class': tfd.JointDistributionSequentialAutoBatched},
{'testcase_name': 'named',
'base_jd_class': tfd.JointDistributionNamed,
'jda_class': tfd.JointDistributionNamedAutoBatched})
def test_broadcast_ragged_batch_shape(self, base_jd_class, jda_class):
base_jd_models = {}
# Writing a JDC with ragged batch shape will broadcast the first
# distribution over the second.
# (though note, this model breaks `log_prob` with nontrivial sample shape).
def coroutine():
x = yield Root(tfd.Normal(0., scale=1.))
yield tfd.Normal(x[..., tf.newaxis], [1., 2., 3., 4., 5.])
base_jd_models[tfd.JointDistributionCoroutine] = coroutine
base_jd_models[tfd.JointDistributionSequential] = [
tfd.Normal(0., scale=1.),
lambda x: tfd.Normal(x[..., tf.newaxis], [1., 2., 3., 4., 5.])
]
base_jd_models[tfd.JointDistributionNamed] = {
'x': tfd.Normal(0., scale=1.),
'y': lambda x: tfd.Normal(x[..., tf.newaxis], [1., 2., 3., 4., 5.])
}
# But we can get equivalent behavior in a JDCA by expanding dims so that
# the batch dimensions line up.
jd_auto_models = {}
def coroutine_auto():
x = yield tfd.Normal(0., scale=[1.])
yield tfd.Normal(x, [1., 2., 3., 4., 5.])
jd_auto_models[tfd.JointDistributionCoroutineAutoBatched] = coroutine_auto
jd_auto_models[tfd.JointDistributionSequentialAutoBatched] = [
tfd.Normal(0., scale=[1.]),
lambda x: tfd.Normal(x, [1., 2., 3., 4., 5.])
]
jd_auto_models[tfd.JointDistributionNamedAutoBatched] = (
collections.OrderedDict((
('x', tfd.Normal(0., scale=[1.])),
('y', lambda x: tfd.Normal(x, [1., 2., 3., 4., 5.])))))
# Writing a JD with ragged batch shape will broadcast the first
# distribution over the second.
# (though note, this model breaks `log_prob` with nontrivial sample shape).
jd_broadcasting = base_jd_class(base_jd_models[base_jd_class])
# This model's broadcasting behavior is a footgun (it can break inference
# routines and cause silently incorrect optimization); it should be
# disallowed by `validate_args`.
with self.assertRaisesRegexp(
Exception,
('Component batch shapes are inconsistent|'
'Broadcasting probably indicates an error in model specification')):
jda_invalid = jda_class(jd_auto_models[jda_class],
batch_ndims=1, validate_args=True)
_ = self.evaluate(jda_invalid.log_prob(
jda_invalid.sample(seed=test_util.test_seed())))
# But, if the user wants to run with no guardrails, one can eke out
# performance wins when evaluating a shared value over multiple models.
jda_broadcasting = jda_class(jd_auto_models[jda_class], batch_ndims=1)
self.assertAllEqual(
jda_broadcasting._model_flatten(jda_broadcasting.event_shape),
[[], []])
self.assertAllEqual(jda_broadcasting.batch_shape, [5])
joint_sample = jda_broadcasting.sample(seed=test_util.test_seed())
x_sample, y_sample = self.evaluate(
list(joint_sample.values()) if hasattr(joint_sample, 'values')
else joint_sample)
# The model samples only a single value for x, shared across the batch.
self.assertAllEqual(x_sample.shape, [1])
self.assertAllEqual(y_sample.shape, [5])
lp_jd_broadcast = self.evaluate(jd_broadcasting.log_prob(
jd_broadcasting._model_unflatten([x_sample[..., 0], y_sample])))
lp_jda_broadcast = self.evaluate(jda_broadcasting.log_prob(
jda_broadcasting._model_unflatten([x_sample, y_sample])))
self.assertAllEqual(lp_jda_broadcast.shape, [5])
self.assertAllEqual(lp_jd_broadcast, lp_jda_broadcast)
# Try drawing multiple samples and computing log-prob.
joint_sample = self.evaluate(jda_broadcasting.sample(
[2, 3], seed=test_util.test_seed()))
lp_jda_broadcast = self.evaluate(jda_broadcasting.log_prob(joint_sample))
self.assertAllEqual(lp_jda_broadcast.shape, [2, 3, 5])
@parameterized.named_parameters(
{'testcase_name': 'coroutine',
'jd_class': tfd.JointDistributionCoroutineAutoBatched},
{'testcase_name': 'sequential',
'jd_class': tfd.JointDistributionSequentialAutoBatched},
{'testcase_name': 'named',
'jd_class': tfd.JointDistributionNamedAutoBatched})
def test_log_prob_and_prob_with_plate(self, jd_class):
models = {}
def coroutine_model():
a = yield tfd.Bernoulli(probs=0.5, dtype=tf.float32)
b = yield tfd.Sample(tfd.Bernoulli(probs=0.25 + 0.5*a,
dtype=tf.float32), 2)
yield tfd.Normal(loc=a, scale=1. + b)
models[tfd.JointDistributionCoroutineAutoBatched] = coroutine_model
models[tfd.JointDistributionSequentialAutoBatched] = [
tfd.Bernoulli(probs=0.5, dtype=tf.float32),
lambda a: tfd.Sample(tfd.Bernoulli( # pylint: disable=g-long-lambda
probs=0.25 + 0.5*a, dtype=tf.float32), 2),
lambda b, a: tfd.Normal(loc=a, scale=1. + b)
]
models[tfd.JointDistributionNamedAutoBatched] = collections.OrderedDict((
('a', tfd.Bernoulli(probs=0.5, dtype=tf.float32)),
('b', lambda a: tfd.Sample(tfd.Bernoulli( # pylint: disable=g-long-lambda
probs=0.25 + 0.5*a, dtype=tf.float32), 2)),
('c', lambda b, a: tfd.Normal(loc=a, scale=1. + b))))
joint = jd_class(models[jd_class], validate_args=True)
z = self.evaluate(joint.sample(seed=test_util.test_seed()))
a, b, c = z.values() if hasattr(z, 'values') else z
log_prob = self.evaluate(joint.log_prob(z))
prob = self.evaluate(joint.prob(z))
expected_log_prob = self.evaluate(
np.log(0.5) +
tf.reduce_sum(tf.math.log(b * (0.25 + 0.5 * a) +
(1 - b) * (0.75 - 0.5 * a))) +
tf.reduce_sum(-0.5 * ((c - a) / (1. + b))**2 -
0.5 * np.log(2. * np.pi) -
tf.math.log((1. + b))))
self.assertAllClose(log_prob, expected_log_prob)
self.assertAllClose(prob, np.exp(expected_log_prob))
@parameterized.named_parameters(
{'testcase_name': 'coroutine',
'jd_class': tfd.JointDistributionCoroutineAutoBatched},
{'testcase_name': 'sequential',
'jd_class': tfd.JointDistributionSequentialAutoBatched},
{'testcase_name': 'named',
'jd_class': tfd.JointDistributionNamedAutoBatched})
def test_log_prob_multiple_samples(self, jd_class):
models = {}
def coroutine_model():
a = yield tfd.Bernoulli(probs=0.5, dtype=tf.float32)
b = yield tfd.Bernoulli(probs=0.25 + 0.5*a,
dtype=tf.float32)
yield tfd.Normal(loc=a, scale=1. + b)
models[tfd.JointDistributionCoroutineAutoBatched] = coroutine_model
models[tfd.JointDistributionSequentialAutoBatched] = [
tfd.Bernoulli(probs=0.5, dtype=tf.float32),
lambda a: tfd.Bernoulli(probs=0.25 + 0.5*a, dtype=tf.float32),
lambda b, a: tfd.Normal(loc=a, scale=1. + b)
]
models[tfd.JointDistributionNamedAutoBatched] = collections.OrderedDict((
('a', tfd.Bernoulli(probs=0.5, dtype=tf.float32)),
('b', lambda a: tfd.Bernoulli(probs=0.25 + 0.5*a, dtype=tf.float32)),
('c', lambda b, a: tfd.Normal(loc=a, scale=1. + b))))
joint = jd_class(models[jd_class], validate_args=True)
z = joint.sample(4, seed=test_util.test_seed())
log_prob = joint.log_prob(z)
a, b, c = z.values() if hasattr(z, 'values') else z # pylint: disable=unbalanced-tuple-unpacking
expected_log_prob = (
np.log(0.5) +
tf.math.log(b * (0.25 + 0.5 * a) +
(1 - b) * (0.75 -0.5 * a)) +
-0.5 * ((c - a) / (1. + b)) ** 2 -
0.5 * np.log(2. * np.pi) -
tf.math.log((1. + b)))
self.assertAllClose(*self.evaluate([log_prob, expected_log_prob]))
@parameterized.named_parameters(
{'testcase_name': 'coroutine',
'jd_class': tfd.JointDistributionCoroutineAutoBatched},
{'testcase_name': 'sequential',
'jd_class': tfd.JointDistributionSequentialAutoBatched},
{'testcase_name': 'named',
'jd_class': tfd.JointDistributionNamedAutoBatched})
def test_sample_and_log_prob(self, jd_class):
# Define a bijector to detect if/when `inverse` is called.
inverted_values = []
class InverseTracingExp(tfb.Exp):
def _inverse(self, y):
inverted_values.append(y)
return tf.math.log(y)
models = {}
def coroutine_model():
g = yield InverseTracingExp()(tfd.Normal(0., 1.), name='g')
df = yield tfd.Exponential(1., name='df')
loc = yield tfd.Sample(tfd.Normal(0, g), 20, name='loc')
yield tfd.StudentT(df, loc, 1, name='x')
models[tfd.JointDistributionCoroutineAutoBatched] = coroutine_model
models[tfd.JointDistributionSequentialAutoBatched] = [
InverseTracingExp()(tfd.Normal(0., 1.), name='g'),
tfd.Exponential(1., name='df'),
lambda _, g: tfd.Sample(tfd.Normal(0, g), 20, name='loc'),
lambda loc, df: tfd.StudentT(df, loc, 1, name='x')
]
models[tfd.JointDistributionNamedAutoBatched] = collections.OrderedDict((
('g', InverseTracingExp()(tfd.Normal(0., 1.))),
('df', tfd.Exponential(1.)),
('loc', lambda g: tfd.Sample(tfd.Normal(0, g), 20)),
('x', lambda loc, df: tfd.StudentT(df, loc, 1))))
joint = jd_class(models[jd_class], validate_args=True)
seed = test_util.test_seed(sampler_type='stateless')
for sample_shape in ([], [5]):
inverted_values.clear()
x1, lp1 = self.evaluate(
joint.experimental_sample_and_log_prob(
sample_shape,
seed=seed,
df=2.7)) # Check that kwargs are supported.
x2 = self.evaluate(
joint.sample(sample_shape, seed=seed, df=2.7))
self.assertAllCloseNested(x1, x2)
self.assertLen(inverted_values, 0)
lp2 = joint.log_prob(x1)
self.assertLen(inverted_values, 1)
self.assertAllClose(lp1, lp2)
@test_util.jax_disable_test_missing_functionality('b/157594634')
def test_sample_distributions(self):
def coroutine_model():
g = yield tfd.Normal(0., 1., name='g')
df = yield tfd.Exponential(1., name='df')
loc = yield tfd.Normal(tf.zeros([20]), g, name='loc')
yield tfd.StudentT(df, loc, 1, name='x')
joint = tfd.JointDistributionCoroutineAutoBatched(coroutine_model)
ds, xs = joint.sample_distributions([4, 3], seed=test_util.test_seed())
for d, x in zip(ds, xs):
self.assertGreaterEqual(len(d.batch_shape), 2)
lp = d.log_prob(x)
self.assertAllEqual(lp.shape[:2], [4, 3])
@test_util.jax_disable_test_missing_functionality('b/201586404')
def test_sample_distributions_not_composite_tensor_raises_error(self):
def coroutine_model():
yield tfd.TransformedDistribution(tfd.Normal(0., 1.),
tfb.Exp(),
name='td')
joint = tfd.JointDistributionCoroutineAutoBatched(coroutine_model)
# Sampling with trivial sample shape avoids the vmap codepath.
ds, _ = joint.sample_distributions([], seed=test_util.test_seed())
self.assertIsInstance(ds[0], tfd.TransformedDistribution)
with self.assertRaisesRegex(
TypeError, r'Some component distribution\(s\) cannot be returned'):
joint.sample_distributions([4, 3], seed=test_util.test_seed())
def test_sample_with_batch_value(self):
@tfd.JointDistributionCoroutineAutoBatched
def dist():
a = yield tfd.Sample(tfd.Normal(0, 1.), 2)
b = yield tfd.Sample(tfd.Normal(0, 1.), 3)
# The following line fails if not autovectorized.
yield tfd.Normal(a[tf.newaxis, ...] * b[..., tf.newaxis], 1.)
x = self.evaluate(dist.sample(123, seed=test_util.test_seed()))
x2 = self.evaluate(dist.sample(value=x, seed=test_util.test_seed()))
self.assertAllCloseNested(x, x2)
# Also test a dict-type value (JDNamed).
dist = tfd.JointDistributionNamedAutoBatched({
'a': tfd.Sample(tfd.Normal(0, 1.), 2),
'b': tfd.Sample(tfd.Normal(0, 1.), 3),
'c': lambda a, b: tfd.Normal( # pylint: disable=g-long-lambda
a[tf.newaxis, ...] * b[..., tf.newaxis], 1.)})
x = self.evaluate(dist.sample(123, seed=test_util.test_seed()))
x2 = self.evaluate(dist.sample(value=x, seed=test_util.test_seed()))
self.assertAllCloseNested(x, x2)
def test_sample_with_value_as_kwarg(self):
@tfd.JointDistributionCoroutineAutoBatched
def dist():
a = yield tfd.Sample(tfd.Normal(0, 1.), 2, name='a')
b = yield tfd.Sample(tfd.Normal(0, 1.), 3, name='b')
# The following line fails if not autovectorized.
yield tfd.Normal(a[tf.newaxis, ...] * b[..., tf.newaxis], 1., name='c')
x = self.evaluate(dist.sample(4, seed=test_util.test_seed()))
x2 = self.evaluate(dist.sample(seed=test_util.test_seed(), a=x.a))
self.assertAllClose(x.a, x2.a)
self.assertAllEqual(x2.b.shape, [4, 3])
self.assertAllEqual(x2.c.shape, [4, 3, 2])
@parameterized.named_parameters(
dict(testcase_name='stateful', sampler_type='stateful'),
dict(testcase_name='stateless', sampler_type='stateless'))
def test_sample_with_partially_specified_value(self, sampler_type):
num_features = 5
def dist():
scale_variance = yield tfd.InverseGamma(0.5, 0.5)
scale_noncentered = yield tfd.Sample(tfd.HalfNormal(1.), num_features)
scale = scale_noncentered * scale_variance[..., None]**0.5
weights_noncentered = yield tfd.Sample(tfd.Normal(0., 1.), num_features)
yield tfd.Deterministic(weights_noncentered * scale)
joint = tfd.JointDistributionCoroutineAutoBatched(dist, validate_args=True)
value_partial_batch_dim = 4
value_ = (3.,
None,
None,
np.ones([value_partial_batch_dim, num_features]))
value = [None if v is None else tf.cast(v, tf.float32) for v in value_]
# The sample should keep the specified values.
xs = self.evaluate(
joint.sample(
value=value, seed=test_util.test_seed(sampler_type=sampler_type)))
self.assertAllEqual(xs[0], tf.fill([value_partial_batch_dim], value[0]))
self.assertAllEqual(xs[1].shape, [value_partial_batch_dim, num_features])
self.assertAllEqual(xs[2].shape, [value_partial_batch_dim, num_features])
self.assertAllEqual(xs[3], value[3])
# With sample shape.
sample_shape = [6, 2]
samples = joint.sample(sample_shape, value=value,
seed=test_util.test_seed(sampler_type=sampler_type))
xs = self.evaluate(samples)
expect_shp = sample_shape + [value_partial_batch_dim, num_features]
self.assertAllEqual(
xs[0], tf.fill(sample_shape + [value_partial_batch_dim], value[0]))
self.assertAllEqual(xs[1].shape, expect_shp)
self.assertAllEqual(xs[2].shape, expect_shp)
self.assertAllEqual(xs[3], value[3] * tf.ones(expect_shp))
sample_shape_dynamic = tf1.placeholder_with_default(
sample_shape, shape=None)
samples = joint.sample(sample_shape_dynamic, value=value,
seed=test_util.test_seed(sampler_type=sampler_type))
xs = self.evaluate(samples)
self.assertAllEqual(
xs[0], tf.fill(sample_shape + [value_partial_batch_dim], value[0]))
self.assertAllEqual(xs[1].shape, expect_shp)
self.assertAllEqual(xs[2].shape, expect_shp)
self.assertAllEqual(xs[3], value[3] * tf.ones(expect_shp))
@parameterized.named_parameters(
dict(testcase_name='stateful', sampler_type='stateful'),
dict(testcase_name='stateless', sampler_type='stateless'))
def test_sample_with_prefix_of_values(self, sampler_type):
num_rows = 4
num_columns = 5
def dist():
a = yield tfd.Sample(tfd.Normal(0., 1.), num_rows, name='a')
b = yield tfd.Sample(tfd.Normal(0., 1.), num_columns, name='b')
yield tfd.Normal(a[..., None] * b[None, ...], 1., name='c')
tuple_joint = tfd.JointDistributionCoroutineAutoBatched(
dist, validate_args=True)
namedtuple_joint = tfd.JointDistributionCoroutineAutoBatched(
dist,
sample_dtype=collections.namedtuple(
'ModelSpec', ['a', 'b', 'c'])(
a=tf.float32, b=tf.float32, c=tf.float32),
validate_args=True)
value_partial_batch_dim = 3
v0 = 3. * np.ones([value_partial_batch_dim, num_rows]).astype(np.float32)
# Tuple (or namedtuple) value contains only the first variable.
tuple_value = (v0,)
namedtuple_value = collections.namedtuple('ValueSpec', ['a'])(a=v0)
for joint in (tuple_joint, namedtuple_joint):
for value in (tuple_value, namedtuple_value):
xs = self.evaluate(
joint.sample(value=value,
seed=test_util.test_seed(sampler_type=sampler_type)))
self.assertAllEqual(xs[0], v0)
self.assertAllEqual(xs[1].shape,
[value_partial_batch_dim, num_columns])
self.assertAllEqual(xs[2].shape,
[value_partial_batch_dim, num_rows, num_columns])
def test_unit_sample_shape_avoids_vectorization(self):
xs = [] # Collect (possibly symbolic) Tensors sampled inside the model.
@tfd.JointDistributionCoroutineAutoBatched
def dist():
x = yield tfd.Normal(0., 1., name='x')
xs.append(x)
# Try sampling with a variety of unit sample shapes.
self.assertEqual(
[1],
dist.sample(
1, seed=test_util.test_seed(sampler_type='seedless')).x.shape)
self.assertEqual(
[1],
dist.sample([1],
seed=test_util.test_seed(sampler_type='seedless')).x.shape)
self.assertEqual(
[1, 1],
dist.sample([1, 1],
seed=test_util.test_seed(sampler_type='seedless')).x.shape)
# Check that the model only ever saw the trivial sample shape.
for x in xs:
self.assertEqual(x.shape, [])
def test_unit_sample_shape(self):
@tfd.JointDistributionCoroutineAutoBatched
def dist():
x = yield tfd.Normal(loc=tf.zeros([3]), scale=1., name='x')
yield tfd.Bernoulli(logits=tf.einsum('n->', x), name='y')
for sample_shape in [(), 1, [1], [1, 1], [2]]:
self.assertAllEqual(
dist.log_prob(
dist.sample(sample_shape,
seed=test_util.test_seed())).shape,
np.reshape(sample_shape, [-1]))
def test_sample_dtype_structures_output(self):
num_features = 4
def dist():
scale_variance = yield Root(tfd.InverseGamma(0.5, 0.5))
scale_noncentered = yield Root(
tfd.Sample(tfd.HalfNormal(1.), num_features))
scale = scale_noncentered * scale_variance[..., None]**0.5
weights_noncentered = yield Root(
tfd.Sample(tfd.Normal(0., 1.), num_features))
yield tfd.Deterministic(weights_noncentered * scale)
# Currently sample_dtype is only used for `tf.nest.pack_structure_as`. In
# the future we may use it for error checking and/or casting.
sample_dtype = collections.namedtuple('Model', [
'scale_variance',
'scale_noncentered',
'weights_noncentered',
'weights',
])(*([None]*4))
joint = tfd.JointDistributionCoroutineAutoBatched(
dist, sample_dtype=sample_dtype, validate_args=True)
self.assertAllEqual(sorted(sample_dtype._fields),
sorted(joint.sample(
seed=test_util.test_seed())._fields))
ds, xs = joint.sample_distributions(seed=test_util.test_seed())
tf.nest.assert_same_structure(sample_dtype, ds)
tf.nest.assert_same_structure(sample_dtype, xs)
self.assertEqual([3, 4], joint.log_prob(joint.sample(
[3, 4], seed=test_util.test_seed())).shape)
def test_repr_with_custom_sample_dtype(self):
sd = collections.namedtuple('Model', ['s', 'w'])(None, None)
def dist():
s = yield tfd.Sample(tfd.InverseGamma(2, 2), 100)
yield tfd.Normal(0, s)
m = tfd.JointDistributionCoroutineAutoBatched(dist, sample_dtype=sd)
self.assertEqual(
('<tfp.distributions.JointDistributionCoroutineAutoBatched'
' \'JointDistributionCoroutineAutoBatched\''
' batch_shape=[]'
' event_shape=Model(s=[100], w=[100])'
' dtype=Model(s=float32, w=float32)>'),
repr(m))
@parameterized.named_parameters(
{'testcase_name': 'coroutine',
'jd_class': tfd.JointDistributionCoroutineAutoBatched},
{'testcase_name': 'sequential',
'jd_class': tfd.JointDistributionSequentialAutoBatched},
{'testcase_name': 'named',
'jd_class': tfd.JointDistributionNamedAutoBatched})
@test_util.jax_disable_variable_test
def test_latent_dirichlet_allocation(self, jd_class): # pylint: disable=g-doc-args
"""Tests Latent Dirichlet Allocation joint model.
The LDA generative process can be written as:
```none
N[i] ~ Poisson(xi)
theta[i] ~ Dirichlet(alpha)
Z[i] ~ Multinomial(N[i], theta[i])
for k in 1...K:
X[i,k] ~ Multinomial(Z[i, k], beta[j])
```
Typically `xi` is specified and `alpha`, `beta` are fit using type-II
maximum likelihood estimators.
Reference: http://www.jmlr.org/papers/volume3/blei03a/blei03a.pdf
"""
seed = test_util.test_seed_stream()
# Hyperparameters.
num_topics = 3
num_words = 10
avg_doc_length = 5
u = tfd.Uniform(low=-1., high=1.)
alpha = tfp.util.TransformedVariable(
u.sample([num_topics], seed=seed()),
tfb.Softplus(), name='alpha')
beta = tf.Variable(u.sample([num_topics, num_words],
seed=seed()), name='beta')
# Note near 1:1 with mathematical specification. The main distinction is the
# use of Independent--this lets us easily aggregate multinomials across
# topics (and in any "shape" of documents).
def lda_coroutine_model():
n = yield Root(tfd.Poisson(rate=avg_doc_length))
theta = yield Root(tfd.Dirichlet(concentration=alpha))
z = yield tfd.Multinomial(total_count=n, probs=theta)
yield tfd.Multinomial(total_count=z, logits=beta)
if jd_class is tfd.JointDistributionCoroutineAutoBatched:
model = lda_coroutine_model
elif jd_class is tfd.JointDistributionSequentialAutoBatched:
model = [
tfd.Poisson(rate=avg_doc_length), # n
tfd.Dirichlet(concentration=alpha), # theta
lambda theta, n: tfd.Multinomial(total_count=n, probs=theta), # z
lambda z: tfd.Multinomial(total_count=z, logits=beta)
]
elif jd_class is tfd.JointDistributionNamedAutoBatched:
model = collections.OrderedDict((
('n', tfd.Poisson(rate=avg_doc_length)),
('theta', tfd.Dirichlet(concentration=alpha)),
('z', lambda theta, n: tfd.Multinomial(total_count=n, probs=theta)),
('X', lambda z: tfd.Multinomial(total_count=z, logits=beta))))
# TODO(b/159842104): Enable autovectorization for Multinomial sampling.
lda = jd_class(model, validate_args=True, use_vectorized_map=False)
# Now, let's sample some "documents" and compute the log-prob of each.
docs_shape = [2, 4] # That is, 8 docs in the shape of [2, 4].
sample = lda.sample(docs_shape, seed=seed())
log_probs = lda.log_prob(sample)
self.assertEqual(docs_shape, log_probs.shape)
# Verify we correctly track trainable variables.
self.assertLen(lda.trainable_variables, 2)
self.assertIs(alpha.pretransformed_input, lda.trainable_variables[0])
self.assertIs(beta, lda.trainable_variables[1])
# Ensure we can compute gradients.
with tf.GradientTape() as tape:
# Note: The samples are not taped, hence implicitly "stop_gradient."
negloglik = -lda.log_prob(sample)
grads = tape.gradient(negloglik, lda.trainable_variables)
self.assertLen(grads, 2)
self.assertAllEqual((alpha.pretransformed_input.shape, beta.shape),
(grads[0].shape, grads[1].shape))
self.assertAllNotNone(grads)
@parameterized.named_parameters(
{'testcase_name': 'coroutine',
'jd_class': tfd.JointDistributionCoroutineAutoBatched},
{'testcase_name': 'sequential',
'jd_class': tfd.JointDistributionSequentialAutoBatched},
{'testcase_name': 'named',
'jd_class': tfd.JointDistributionNamedAutoBatched})
def test_default_event_space_bijector(self, jd_class):
models = {}
def coroutine_model():
high = yield tfd.LogNormal(0., [1.])
yield tfd.Uniform(low=[[-1., -2.]], high=high[..., tf.newaxis])
yield tfd.Deterministic([[0., 1., 2.]])
models[tfd.JointDistributionCoroutineAutoBatched] = coroutine_model
models[tfd.JointDistributionSequentialAutoBatched] = [
tfd.LogNormal(0., [1.]),
lambda high: tfd.Uniform(low=[[-1., -2.]], high=high[..., tf.newaxis]),
tfd.Deterministic([[0., 1., 2.]])
]
models[tfd.JointDistributionNamedAutoBatched] = collections.OrderedDict((
('high', tfd.LogNormal(0., [1.])),
('x', lambda high: tfd.Uniform(low=[[-1., -2.]], # pylint: disable=g-long-lambda
high=high[..., tf.newaxis])),
('y', tfd.Deterministic([[0., 1., 2.]]))))
joint = jd_class(models[jd_class], batch_ndims=1, validate_args=True)
self.assertAllEqual(joint.batch_shape, [1])
self.assertAllEqualNested(tf.nest.flatten(joint.event_shape),
[[], [2], [3]])
joint_bijector = joint.experimental_default_event_space_bijector()
y = self.evaluate(joint.sample([2, 3], seed=test_util.test_seed()))
x = joint_bijector.inverse(y)
self.assertAllCloseNested(y, joint_bijector.forward(x))
fldj = joint_bijector.forward_log_det_jacobian(
x, event_ndims=tf.nest.pack_sequence_as(joint.dtype, [0, 1, 2]))
ildj = joint_bijector.inverse_log_det_jacobian(
y, event_ndims=tf.nest.pack_sequence_as(joint.dtype, [0, 1, 1]))
self.assertAllEqual(fldj.shape, joint.log_prob(y).shape)
self.assertAllClose(fldj, -ildj)
# Passing inputs *without* batch shape should return sane outputs.
y = self.evaluate(joint.sample([], seed=test_util.test_seed()))
# Strip the sample to represent just a single event.
unbatched_y = tf.nest.map_structure(lambda t: t[0, ...], y)
self.assertAllEqualNested(tf.nest.map_structure(tf.shape, unbatched_y),
joint.event_shape_tensor())
ildj = joint_bijector.inverse_log_det_jacobian(
unbatched_y,
event_ndims=tf.nest.pack_sequence_as(joint.dtype, [0, 1, 1]))
self.assertAllEqual(ildj.shape, joint.log_prob(unbatched_y).shape)
@parameterized.named_parameters(
{'testcase_name': 'coroutine',
'jd_class': tfd.JointDistributionCoroutineAutoBatched},
{'testcase_name': 'sequential',
'jd_class': tfd.JointDistributionSequentialAutoBatched},
{'testcase_name': 'named',
'jd_class': tfd.JointDistributionNamedAutoBatched})
def test_default_event_space_bijector_constant_jacobian(self, jd_class):
models = {}
def coroutine_model():
yield tfd.Normal(0., [1., 2.], name='x')
models[tfd.JointDistributionCoroutineAutoBatched] = coroutine_model
models[tfd.JointDistributionSequentialAutoBatched] = [
tfd.Normal(0., [1., 2.], name='x')
]
models[tfd.JointDistributionNamedAutoBatched] = {
'x': tfd.Normal(0., [1., 2.], name='x')}
joint = jd_class(models[jd_class], batch_ndims=1, validate_args=True)
self.assertAllEqual(joint.batch_shape, [2])
joint_bijector = joint.experimental_default_event_space_bijector()
y = self.evaluate(joint.sample([3], seed=test_util.test_seed()))
x = joint_bijector.inverse(y)
self.assertAllCloseNested(y, joint_bijector.forward(x))
fldj = joint_bijector.forward_log_det_jacobian(x)
ildj = joint_bijector.inverse_log_det_jacobian(y)
self.assertAllEqual(fldj.shape, joint.log_prob(y).shape)
self.assertAllClose(fldj, -ildj)
def test_nested_joint_distributions(self):
batch_shape = [2, 3]
def inner_fn():
xy = yield tfd.JointDistributionNamedAutoBatched(
{'x': tfd.Normal(loc=tf.zeros(batch_shape),
scale=tf.ones(batch_shape),
name='x'),
'y': lambda x: tfd.Poisson(log_rate=x, name='y')},
batch_ndims=2,
name='xy')
_ = yield tfd.Normal(loc=0., scale=xy['y'], name='z')
joint = tfd.JointDistributionSequentialAutoBatched([
tfd.JointDistributionCoroutineAutoBatched(inner_fn,
batch_ndims=1,
name='a')])
z = joint.sample(seed=test_util.test_seed())
# Batch and event shape.
self.assertAllEqual(joint.batch_shape, [])
self.assertAllEqualNested(
tf.nest.map_structure(lambda x: tf.TensorShape(x.shape), z),
joint.event_shape)
# Sample shape.
z2 = self.evaluate(
joint.sample(5, seed=test_util.test_seed()))
lp2 = joint.log_prob(z2)
self.assertAllEqual(lp2.shape, [5])
z3 = joint.sample(value=z2, seed=test_util.test_seed())
self.assertAllCloseNested(z2, z3)
@parameterized.named_parameters(*[
dict(testcase_name='_{}{}'.format(jd_class.__name__, # pylint: disable=g-complex-comprehension
'_jit' if jit else ''),
jd_class=jd_class, jit=jit)
for jd_class in (tfd.JointDistributionCoroutineAutoBatched,
tfd.JointDistributionSequentialAutoBatched,
tfd.JointDistributionNamedAutoBatched)
for jit in (False, True)
])
def test_kahan_precision(self, jd_class, jit):
maybe_jit = lambda f: f
if jit:
self.skip_if_no_xla()
if not JAX_MODE and not tf.test.is_gpu_available():
self.skipTest('b/179303849')
maybe_jit = tf.function(jit_compile=True)
def make_models(dtype):
models = {}
def mk_20k_poisson(log_rate):
return tfd.Poisson(log_rate=tf.broadcast_to(log_rate[..., tf.newaxis],
log_rate.shape + (20_000,)))
def coroutine_model():
log_rate = yield tfd.Normal(0., dtype(.2), name='log_rate')
yield mk_20k_poisson(log_rate).copy(name='x')
models[tfd.JointDistributionCoroutineAutoBatched] = coroutine_model
models[tfd.JointDistributionSequentialAutoBatched] = [
tfd.Normal(0., dtype(.2)), mk_20k_poisson
]
models[tfd.JointDistributionNamedAutoBatched] = collections.OrderedDict((
('log_rate', tfd.Normal(0., dtype(.2))), ('x', mk_20k_poisson)))
return models
joint = jd_class(make_models(np.float32)[jd_class], validate_args=True,
experimental_use_kahan_sum=True)
joint64 = jd_class(make_models(np.float64)[jd_class], validate_args=True)
stream = test_util.test_seed_stream()
nsamp = 7
xs = self.evaluate(
joint.sample(log_rate=tf.zeros([nsamp]), seed=stream()))
if isinstance(xs, dict):
xs['log_rate'] = tfd.Normal(0, .2).sample(nsamp, seed=stream())
else:
xs = (tfd.Normal(0, .2).sample(nsamp, seed=stream()), xs[1])
xs64 = tf.nest.map_structure(lambda x: tf.cast(x, tf.float64), xs)
lp = maybe_jit(joint.copy(validate_args=not jit).log_prob)(xs)
lp64 = joint64.log_prob(xs64)
lp, lp64 = self.evaluate((tf.cast(lp, tf.float64), lp64))
# Without Kahan, example max-abs-diff: ~0.06
self.assertAllClose(lp64, lp, rtol=0., atol=.01)
def test_kahan_broadcasting_check(self):
def model():
_ = yield tfd.Normal(0., 1.) # Batch shape ()
_ = yield tfd.Normal([0., 1., 2.], 1.) # Batch shape [3]
dist = tfd.JointDistributionCoroutineAutoBatched(
model, validate_args=True, experimental_use_kahan_sum=True,
batch_ndims=1)
sample = self.evaluate(dist.sample(seed=test_util.test_seed(
sampler_type='stateless')))
with self.assertRaises(ValueError):
self.evaluate(dist.log_prob(sample))
if __name__ == '__main__':
# TODO(b/173158845): XLA:CPU reassociates away the Kahan correction term.
os.environ['XLA_FLAGS'] = '--xla_cpu_enable_fast_math=false'
test_util.main()
|
[
"tensorflow.compat.v2.nest.map_structure",
"numpy.log",
"tensorflow.compat.v2.einsum",
"tensorflow_probability.python.internal.test_util.jax_disable_test_missing_functionality",
"tensorflow.compat.v2.cast",
"tensorflow.compat.v2.nest.pack_sequence_as",
"tensorflow.compat.v2.nest.assert_same_structure",
"numpy.reshape",
"tensorflow.compat.v2.executing_eagerly",
"tensorflow.compat.v2.function",
"tensorflow.compat.v2.math.log",
"tensorflow.compat.v2.expand_dims",
"tensorflow.compat.v2.TensorShape",
"numpy.exp",
"tensorflow.compat.v2.ones",
"tensorflow.compat.v2.nest.flatten",
"tensorflow.compat.v1.placeholder_with_default",
"collections.namedtuple",
"numpy.ones",
"tensorflow.compat.v2.zeros",
"tensorflow_probability.python.internal.test_util.main",
"tensorflow.compat.v2.fill",
"tensorflow.compat.v2.broadcast_to",
"absl.testing.parameterized.named_parameters",
"tensorflow_probability.python.internal.test_util.test_seed_stream",
"tensorflow_probability.python.internal.test_util.test_seed",
"tensorflow.compat.v2.GradientTape",
"tensorflow.compat.v2.test.is_gpu_available"
] |
[((1233, 1535), 'absl.testing.parameterized.named_parameters', 'parameterized.named_parameters', (["{'testcase_name': 'coroutine', 'jd_class': tfd.\n JointDistributionCoroutineAutoBatched}", "{'testcase_name': 'sequential', 'jd_class': tfd.\n JointDistributionSequentialAutoBatched}", "{'testcase_name': 'named', 'jd_class': tfd.JointDistributionNamedAutoBatched}"], {}), "({'testcase_name': 'coroutine', 'jd_class':\n tfd.JointDistributionCoroutineAutoBatched}, {'testcase_name':\n 'sequential', 'jd_class': tfd.JointDistributionSequentialAutoBatched},\n {'testcase_name': 'named', 'jd_class': tfd.\n JointDistributionNamedAutoBatched})\n", (1263, 1535), False, 'from absl.testing import parameterized\n'), ((6819, 7278), 'absl.testing.parameterized.named_parameters', 'parameterized.named_parameters', (["{'testcase_name': 'coroutine', 'base_jd_class': tfd.\n JointDistributionCoroutine, 'jda_class': tfd.\n JointDistributionCoroutineAutoBatched}", "{'testcase_name': 'sequential', 'base_jd_class': tfd.\n JointDistributionSequential, 'jda_class': tfd.\n JointDistributionSequentialAutoBatched}", "{'testcase_name': 'named', 'base_jd_class': tfd.JointDistributionNamed,\n 'jda_class': tfd.JointDistributionNamedAutoBatched}"], {}), "({'testcase_name': 'coroutine',\n 'base_jd_class': tfd.JointDistributionCoroutine, 'jda_class': tfd.\n JointDistributionCoroutineAutoBatched}, {'testcase_name': 'sequential',\n 'base_jd_class': tfd.JointDistributionSequential, 'jda_class': tfd.\n JointDistributionSequentialAutoBatched}, {'testcase_name': 'named',\n 'base_jd_class': tfd.JointDistributionNamed, 'jda_class': tfd.\n JointDistributionNamedAutoBatched})\n", (6849, 7278), False, 'from absl.testing import parameterized\n'), ((11146, 11448), 'absl.testing.parameterized.named_parameters', 'parameterized.named_parameters', (["{'testcase_name': 'coroutine', 'jd_class': tfd.\n JointDistributionCoroutineAutoBatched}", "{'testcase_name': 'sequential', 'jd_class': tfd.\n JointDistributionSequentialAutoBatched}", "{'testcase_name': 'named', 'jd_class': tfd.JointDistributionNamedAutoBatched}"], {}), "({'testcase_name': 'coroutine', 'jd_class':\n tfd.JointDistributionCoroutineAutoBatched}, {'testcase_name':\n 'sequential', 'jd_class': tfd.JointDistributionSequentialAutoBatched},\n {'testcase_name': 'named', 'jd_class': tfd.\n JointDistributionNamedAutoBatched})\n", (11176, 11448), False, 'from absl.testing import parameterized\n'), ((13230, 13532), 'absl.testing.parameterized.named_parameters', 'parameterized.named_parameters', (["{'testcase_name': 'coroutine', 'jd_class': tfd.\n JointDistributionCoroutineAutoBatched}", "{'testcase_name': 'sequential', 'jd_class': tfd.\n JointDistributionSequentialAutoBatched}", "{'testcase_name': 'named', 'jd_class': tfd.JointDistributionNamedAutoBatched}"], {}), "({'testcase_name': 'coroutine', 'jd_class':\n tfd.JointDistributionCoroutineAutoBatched}, {'testcase_name':\n 'sequential', 'jd_class': tfd.JointDistributionSequentialAutoBatched},\n {'testcase_name': 'named', 'jd_class': tfd.\n JointDistributionNamedAutoBatched})\n", (13260, 13532), False, 'from absl.testing import parameterized\n'), ((15023, 15325), 'absl.testing.parameterized.named_parameters', 'parameterized.named_parameters', (["{'testcase_name': 'coroutine', 'jd_class': tfd.\n JointDistributionCoroutineAutoBatched}", "{'testcase_name': 'sequential', 'jd_class': tfd.\n JointDistributionSequentialAutoBatched}", "{'testcase_name': 'named', 'jd_class': tfd.JointDistributionNamedAutoBatched}"], {}), "({'testcase_name': 'coroutine', 'jd_class':\n tfd.JointDistributionCoroutineAutoBatched}, {'testcase_name':\n 'sequential', 'jd_class': tfd.JointDistributionSequentialAutoBatched},\n {'testcase_name': 'named', 'jd_class': tfd.\n JointDistributionNamedAutoBatched})\n", (15053, 15325), False, 'from absl.testing import parameterized\n'), ((17195, 17258), 'tensorflow_probability.python.internal.test_util.jax_disable_test_missing_functionality', 'test_util.jax_disable_test_missing_functionality', (['"""b/157594634"""'], {}), "('b/157594634')\n", (17243, 17258), False, 'from tensorflow_probability.python.internal import test_util\n'), ((17832, 17895), 'tensorflow_probability.python.internal.test_util.jax_disable_test_missing_functionality', 'test_util.jax_disable_test_missing_functionality', (['"""b/201586404"""'], {}), "('b/201586404')\n", (17880, 17895), False, 'from tensorflow_probability.python.internal import test_util\n'), ((27580, 27882), 'absl.testing.parameterized.named_parameters', 'parameterized.named_parameters', (["{'testcase_name': 'coroutine', 'jd_class': tfd.\n JointDistributionCoroutineAutoBatched}", "{'testcase_name': 'sequential', 'jd_class': tfd.\n JointDistributionSequentialAutoBatched}", "{'testcase_name': 'named', 'jd_class': tfd.JointDistributionNamedAutoBatched}"], {}), "({'testcase_name': 'coroutine', 'jd_class':\n tfd.JointDistributionCoroutineAutoBatched}, {'testcase_name':\n 'sequential', 'jd_class': tfd.JointDistributionSequentialAutoBatched},\n {'testcase_name': 'named', 'jd_class': tfd.\n JointDistributionNamedAutoBatched})\n", (27610, 27882), False, 'from absl.testing import parameterized\n'), ((31272, 31574), 'absl.testing.parameterized.named_parameters', 'parameterized.named_parameters', (["{'testcase_name': 'coroutine', 'jd_class': tfd.\n JointDistributionCoroutineAutoBatched}", "{'testcase_name': 'sequential', 'jd_class': tfd.\n JointDistributionSequentialAutoBatched}", "{'testcase_name': 'named', 'jd_class': tfd.JointDistributionNamedAutoBatched}"], {}), "({'testcase_name': 'coroutine', 'jd_class':\n tfd.JointDistributionCoroutineAutoBatched}, {'testcase_name':\n 'sequential', 'jd_class': tfd.JointDistributionSequentialAutoBatched},\n {'testcase_name': 'named', 'jd_class': tfd.\n JointDistributionNamedAutoBatched})\n", (31302, 31574), False, 'from absl.testing import parameterized\n'), ((33918, 34220), 'absl.testing.parameterized.named_parameters', 'parameterized.named_parameters', (["{'testcase_name': 'coroutine', 'jd_class': tfd.\n JointDistributionCoroutineAutoBatched}", "{'testcase_name': 'sequential', 'jd_class': tfd.\n JointDistributionSequentialAutoBatched}", "{'testcase_name': 'named', 'jd_class': tfd.JointDistributionNamedAutoBatched}"], {}), "({'testcase_name': 'coroutine', 'jd_class':\n tfd.JointDistributionCoroutineAutoBatched}, {'testcase_name':\n 'sequential', 'jd_class': tfd.JointDistributionSequentialAutoBatched},\n {'testcase_name': 'named', 'jd_class': tfd.\n JointDistributionNamedAutoBatched})\n", (33948, 34220), False, 'from absl.testing import parameterized\n'), ((39530, 39546), 'tensorflow_probability.python.internal.test_util.main', 'test_util.main', ([], {}), '()\n', (39544, 39546), False, 'from tensorflow_probability.python.internal import test_util\n'), ((5620, 5642), 'tensorflow.compat.v2.executing_eagerly', 'tf.executing_eagerly', ([], {}), '()\n', (5640, 5642), True, 'import tensorflow.compat.v2 as tf\n'), ((16614, 16659), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {'sampler_type': '"""stateless"""'}), "(sampler_type='stateless')\n", (16633, 16659), False, 'from tensorflow_probability.python.internal import test_util\n'), ((22191, 22245), 'tensorflow.compat.v1.placeholder_with_default', 'tf1.placeholder_with_default', (['sample_shape'], {'shape': 'None'}), '(sample_shape, shape=None)\n', (22219, 22245), True, 'import tensorflow.compat.v1 as tf1\n'), ((26791, 26838), 'tensorflow.compat.v2.nest.assert_same_structure', 'tf.nest.assert_same_structure', (['sample_dtype', 'ds'], {}), '(sample_dtype, ds)\n', (26820, 26838), True, 'import tensorflow.compat.v2 as tf\n'), ((26843, 26890), 'tensorflow.compat.v2.nest.assert_same_structure', 'tf.nest.assert_same_structure', (['sample_dtype', 'xs'], {}), '(sample_dtype, xs)\n', (26872, 26890), True, 'import tensorflow.compat.v2 as tf\n'), ((28516, 28544), 'tensorflow_probability.python.internal.test_util.test_seed_stream', 'test_util.test_seed_stream', ([], {}), '()\n', (28542, 28544), False, 'from tensorflow_probability.python.internal import test_util\n'), ((33520, 33565), 'tensorflow.compat.v2.nest.map_structure', 'tf.nest.map_structure', (['(lambda t: t[0, ...])', 'y'], {}), '(lambda t: t[0, ...], y)\n', (33541, 33565), True, 'import tensorflow.compat.v2 as tf\n'), ((38209, 38237), 'tensorflow_probability.python.internal.test_util.test_seed_stream', 'test_util.test_seed_stream', ([], {}), '()\n', (38235, 38237), False, 'from tensorflow_probability.python.internal import test_util\n'), ((6513, 6533), 'tensorflow.compat.v2.TensorShape', 'tf.TensorShape', (['None'], {}), '(None)\n', (6527, 6533), True, 'import tensorflow.compat.v2 as tf\n'), ((13199, 13224), 'numpy.exp', 'np.exp', (['expected_log_prob'], {}), '(expected_log_prob)\n', (13205, 13224), True, 'import numpy as np\n'), ((14924, 14944), 'tensorflow.compat.v2.math.log', 'tf.math.log', (['(1.0 + b)'], {}), '(1.0 + b)\n', (14935, 14944), True, 'import tensorflow.compat.v2 as tf\n'), ((21034, 21082), 'numpy.ones', 'np.ones', (['[value_partial_batch_dim, num_features]'], {}), '([value_partial_batch_dim, num_features])\n', (21041, 21082), True, 'import numpy as np\n'), ((21368, 21412), 'tensorflow.compat.v2.fill', 'tf.fill', (['[value_partial_batch_dim]', 'value[0]'], {}), '([value_partial_batch_dim], value[0])\n', (21375, 21412), True, 'import tensorflow.compat.v2 as tf\n'), ((21941, 22000), 'tensorflow.compat.v2.fill', 'tf.fill', (['(sample_shape + [value_partial_batch_dim])', 'value[0]'], {}), '(sample_shape + [value_partial_batch_dim], value[0])\n', (21948, 22000), True, 'import tensorflow.compat.v2 as tf\n'), ((22469, 22528), 'tensorflow.compat.v2.fill', 'tf.fill', (['(sample_shape + [value_partial_batch_dim])', 'value[0]'], {}), '(sample_shape + [value_partial_batch_dim], value[0])\n', (22476, 22528), True, 'import tensorflow.compat.v2 as tf\n'), ((23750, 23792), 'collections.namedtuple', 'collections.namedtuple', (['"""ValueSpec"""', "['a']"], {}), "('ValueSpec', ['a'])\n", (23772, 23792), False, 'import collections\n'), ((26279, 26389), 'collections.namedtuple', 'collections.namedtuple', (['"""Model"""', "['scale_variance', 'scale_noncentered', 'weights_noncentered', 'weights']"], {}), "('Model', ['scale_variance', 'scale_noncentered',\n 'weights_noncentered', 'weights'])\n", (26301, 26389), False, 'import collections\n'), ((27059, 27102), 'collections.namedtuple', 'collections.namedtuple', (['"""Model"""', "['s', 'w']"], {}), "('Model', ['s', 'w'])\n", (27081, 27102), False, 'import collections\n'), ((30871, 30888), 'tensorflow.compat.v2.GradientTape', 'tf.GradientTape', ([], {}), '()\n', (30886, 30888), True, 'import tensorflow.compat.v2 as tf\n'), ((32636, 32670), 'tensorflow.compat.v2.nest.flatten', 'tf.nest.flatten', (['joint.event_shape'], {}), '(joint.event_shape)\n', (32651, 32670), True, 'import tensorflow.compat.v2 as tf\n'), ((33596, 33640), 'tensorflow.compat.v2.nest.map_structure', 'tf.nest.map_structure', (['tf.shape', 'unbatched_y'], {}), '(tf.shape, unbatched_y)\n', (33617, 33640), True, 'import tensorflow.compat.v2 as tf\n'), ((37191, 37220), 'tensorflow.compat.v2.function', 'tf.function', ([], {'jit_compile': '(True)'}), '(jit_compile=True)\n', (37202, 37220), True, 'import tensorflow.compat.v2 as tf\n'), ((5431, 5477), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {'sampler_type': 'sampler_type'}), '(sampler_type=sampler_type)\n', (5450, 5477), False, 'from tensorflow_probability.python.internal import test_util\n'), ((6006, 6047), 'tensorflow.compat.v1.placeholder_with_default', 'tf1.placeholder_with_default', (['(1)'], {'shape': '[]'}), '(1, shape=[])\n', (6034, 6047), True, 'import tensorflow.compat.v1 as tf1\n'), ((6684, 6729), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {'sampler_type': '"""stateless"""'}), "(sampler_type='stateless')\n", (6703, 6729), False, 'from tensorflow_probability.python.internal import test_util\n'), ((10137, 10158), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (10156, 10158), False, 'from tensorflow_probability.python.internal import test_util\n'), ((14537, 14558), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (14556, 14558), False, 'from tensorflow_probability.python.internal import test_util\n'), ((15604, 15618), 'tensorflow.compat.v2.math.log', 'tf.math.log', (['y'], {}), '(y)\n', (15615, 15618), True, 'import tensorflow.compat.v2 as tf\n'), ((17650, 17671), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (17669, 17671), False, 'from tensorflow_probability.python.internal import test_util\n'), ((18345, 18366), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (18364, 18366), False, 'from tensorflow_probability.python.internal import test_util\n'), ((21120, 21142), 'tensorflow.compat.v2.cast', 'tf.cast', (['v', 'tf.float32'], {}), '(v, tf.float32)\n', (21127, 21142), True, 'import tensorflow.compat.v2 as tf\n'), ((21749, 21795), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {'sampler_type': 'sampler_type'}), '(sampler_type=sampler_type)\n', (21768, 21795), False, 'from tensorflow_probability.python.internal import test_util\n'), ((22142, 22161), 'tensorflow.compat.v2.ones', 'tf.ones', (['expect_shp'], {}), '(expect_shp)\n', (22149, 22161), True, 'import tensorflow.compat.v2 as tf\n'), ((22349, 22395), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {'sampler_type': 'sampler_type'}), '(sampler_type=sampler_type)\n', (22368, 22395), False, 'from tensorflow_probability.python.internal import test_util\n'), ((22670, 22689), 'tensorflow.compat.v2.ones', 'tf.ones', (['expect_shp'], {}), '(expect_shp)\n', (22677, 22689), True, 'import tensorflow.compat.v2 as tf\n'), ((25618, 25648), 'numpy.reshape', 'np.reshape', (['sample_shape', '[-1]'], {}), '(sample_shape, [-1])\n', (25628, 25648), True, 'import numpy as np\n'), ((26764, 26785), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (26783, 26785), False, 'from tensorflow_probability.python.internal import test_util\n'), ((33032, 33080), 'tensorflow.compat.v2.nest.pack_sequence_as', 'tf.nest.pack_sequence_as', (['joint.dtype', '[0, 1, 2]'], {}), '(joint.dtype, [0, 1, 2])\n', (33056, 33080), True, 'import tensorflow.compat.v2 as tf\n'), ((33157, 33205), 'tensorflow.compat.v2.nest.pack_sequence_as', 'tf.nest.pack_sequence_as', (['joint.dtype', '[0, 1, 1]'], {}), '(joint.dtype, [0, 1, 1])\n', (33181, 33205), True, 'import tensorflow.compat.v2 as tf\n'), ((33793, 33841), 'tensorflow.compat.v2.nest.pack_sequence_as', 'tf.nest.pack_sequence_as', (['joint.dtype', '[0, 1, 1]'], {}), '(joint.dtype, [0, 1, 1])\n', (33817, 33841), True, 'import tensorflow.compat.v2 as tf\n'), ((35994, 36015), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (36013, 36015), False, 'from tensorflow_probability.python.internal import test_util\n'), ((36426, 36447), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (36445, 36447), False, 'from tensorflow_probability.python.internal import test_util\n'), ((38560, 38582), 'tensorflow.compat.v2.cast', 'tf.cast', (['x', 'tf.float64'], {}), '(x, tf.float64)\n', (38567, 38582), True, 'import tensorflow.compat.v2 as tf\n'), ((38719, 38742), 'tensorflow.compat.v2.cast', 'tf.cast', (['lp', 'tf.float64'], {}), '(lp, tf.float64)\n', (38726, 38742), True, 'import tensorflow.compat.v2 as tf\n'), ((1816, 1838), 'tensorflow.compat.v2.expand_dims', 'tf.expand_dims', (['df', '(-1)'], {}), '(df, -1)\n', (1830, 1838), True, 'import tensorflow.compat.v2 as tf\n'), ((2132, 2154), 'tensorflow.compat.v2.expand_dims', 'tf.expand_dims', (['df', '(-1)'], {}), '(df, -1)\n', (2146, 2154), True, 'import tensorflow.compat.v2 as tf\n'), ((4092, 4114), 'tensorflow.compat.v2.expand_dims', 'tf.expand_dims', (['df', '(-1)'], {}), '(df, -1)\n', (4106, 4114), True, 'import tensorflow.compat.v2 as tf\n'), ((4420, 4442), 'tensorflow.compat.v2.expand_dims', 'tf.expand_dims', (['df', '(-1)'], {}), '(df, -1)\n', (4434, 4442), True, 'import tensorflow.compat.v2 as tf\n'), ((5873, 5895), 'tensorflow.compat.v2.expand_dims', 'tf.expand_dims', (['df', '(-1)'], {}), '(df, -1)\n', (5887, 5895), True, 'import tensorflow.compat.v2 as tf\n'), ((6625, 6645), 'tensorflow.compat.v2.TensorShape', 'tf.TensorShape', (['None'], {}), '(None)\n', (6639, 6645), True, 'import tensorflow.compat.v2 as tf\n'), ((10981, 11002), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (11000, 11002), False, 'from tensorflow_probability.python.internal import test_util\n'), ((12613, 12634), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (12632, 12634), False, 'from tensorflow_probability.python.internal import test_util\n'), ((12830, 12841), 'numpy.log', 'np.log', (['(0.5)'], {}), '(0.5)\n', (12836, 12841), True, 'import numpy as np\n'), ((14895, 14914), 'numpy.log', 'np.log', (['(2.0 * np.pi)'], {}), '(2.0 * np.pi)\n', (14901, 14914), True, 'import numpy as np\n'), ((17447, 17461), 'tensorflow.compat.v2.zeros', 'tf.zeros', (['[20]'], {}), '([20])\n', (17455, 17461), True, 'import tensorflow.compat.v2 as tf\n'), ((18586, 18607), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (18605, 18607), False, 'from tensorflow_probability.python.internal import test_util\n'), ((18981, 19002), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (19000, 19002), False, 'from tensorflow_probability.python.internal import test_util\n'), ((19054, 19075), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (19073, 19075), False, 'from tensorflow_probability.python.internal import test_util\n'), ((19480, 19501), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (19499, 19501), False, 'from tensorflow_probability.python.internal import test_util\n'), ((19553, 19574), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (19572, 19574), False, 'from tensorflow_probability.python.internal import test_util\n'), ((20018, 20039), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (20037, 20039), False, 'from tensorflow_probability.python.internal import test_util\n'), ((20082, 20103), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (20101, 20103), False, 'from tensorflow_probability.python.internal import test_util\n'), ((21288, 21334), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {'sampler_type': 'sampler_type'}), '(sampler_type=sampler_type)\n', (21307, 21334), False, 'from tensorflow_probability.python.internal import test_util\n'), ((23369, 23421), 'collections.namedtuple', 'collections.namedtuple', (['"""ModelSpec"""', "['a', 'b', 'c']"], {}), "('ModelSpec', ['a', 'b', 'c'])\n", (23391, 23421), False, 'import collections\n'), ((23570, 23614), 'numpy.ones', 'np.ones', (['[value_partial_batch_dim, num_rows]'], {}), '([value_partial_batch_dim, num_rows])\n', (23577, 23614), True, 'import numpy as np\n'), ((32838, 32859), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (32857, 32859), False, 'from tensorflow_probability.python.internal import test_util\n'), ((33421, 33442), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (33440, 33442), False, 'from tensorflow_probability.python.internal import test_util\n'), ((34935, 34956), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (34954, 34956), False, 'from tensorflow_probability.python.internal import test_util\n'), ((36165, 36188), 'tensorflow.compat.v2.TensorShape', 'tf.TensorShape', (['x.shape'], {}), '(x.shape)\n', (36179, 36188), True, 'import tensorflow.compat.v2 as tf\n'), ((36295, 36316), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (36314, 36316), False, 'from tensorflow_probability.python.internal import test_util\n'), ((37108, 37134), 'tensorflow.compat.v2.test.is_gpu_available', 'tf.test.is_gpu_available', ([], {}), '()\n', (37132, 37134), True, 'import tensorflow.compat.v2 as tf\n'), ((38306, 38323), 'tensorflow.compat.v2.zeros', 'tf.zeros', (['[nsamp]'], {}), '([nsamp])\n', (38314, 38323), True, 'import tensorflow.compat.v2 as tf\n'), ((39220, 39265), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {'sampler_type': '"""stateless"""'}), "(sampler_type='stateless')\n", (39239, 39265), False, 'from tensorflow_probability.python.internal import test_util\n'), ((12866, 12928), 'tensorflow.compat.v2.math.log', 'tf.math.log', (['(b * (0.25 + 0.5 * a) + (1 - b) * (0.75 - 0.5 * a))'], {}), '(b * (0.25 + 0.5 * a) + (1 - b) * (0.75 - 0.5 * a))\n', (12877, 12928), True, 'import tensorflow.compat.v2 as tf\n'), ((13092, 13112), 'tensorflow.compat.v2.math.log', 'tf.math.log', (['(1.0 + b)'], {}), '(1.0 + b)\n', (13103, 13112), True, 'import tensorflow.compat.v2 as tf\n'), ((14732, 14743), 'numpy.log', 'np.log', (['(0.5)'], {}), '(0.5)\n', (14738, 14743), True, 'import numpy as np\n'), ((14754, 14816), 'tensorflow.compat.v2.math.log', 'tf.math.log', (['(b * (0.25 + 0.5 * a) + (1 - b) * (0.75 - 0.5 * a))'], {}), '(b * (0.25 + 0.5 * a) + (1 - b) * (0.75 - 0.5 * a))\n', (14765, 14816), True, 'import tensorflow.compat.v2 as tf\n'), ((25303, 25316), 'tensorflow.compat.v2.zeros', 'tf.zeros', (['[3]'], {}), '([3])\n', (25311, 25316), True, 'import tensorflow.compat.v2 as tf\n'), ((25371, 25390), 'tensorflow.compat.v2.einsum', 'tf.einsum', (['"""n->"""', 'x'], {}), "('n->', x)\n", (25380, 25390), True, 'import tensorflow.compat.v2 as tf\n'), ((37340, 37409), 'tensorflow.compat.v2.broadcast_to', 'tf.broadcast_to', (['log_rate[..., tf.newaxis]', '(log_rate.shape + (20000,))'], {}), '(log_rate[..., tf.newaxis], log_rate.shape + (20000,))\n', (37355, 37409), True, 'import tensorflow.compat.v2 as tf\n'), ((2428, 2450), 'tensorflow.compat.v2.expand_dims', 'tf.expand_dims', (['df', '(-1)'], {}), '(df, -1)\n', (2442, 2450), True, 'import tensorflow.compat.v2 as tf\n'), ((4728, 4750), 'tensorflow.compat.v2.expand_dims', 'tf.expand_dims', (['df', '(-1)'], {}), '(df, -1)\n', (4742, 4750), True, 'import tensorflow.compat.v2 as tf\n'), ((9665, 9686), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (9684, 9686), False, 'from tensorflow_probability.python.internal import test_util\n'), ((23998, 24044), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {'sampler_type': 'sampler_type'}), '(sampler_type=sampler_type)\n', (24017, 24044), False, 'from tensorflow_probability.python.internal import test_util\n'), ((24710, 24754), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {'sampler_type': '"""seedless"""'}), "(sampler_type='seedless')\n", (24729, 24754), False, 'from tensorflow_probability.python.internal import test_util\n'), ((24850, 24894), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {'sampler_type': '"""seedless"""'}), "(sampler_type='seedless')\n", (24869, 24894), False, 'from tensorflow_probability.python.internal import test_util\n'), ((24996, 25040), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {'sampler_type': '"""seedless"""'}), "(sampler_type='seedless')\n", (25015, 25040), False, 'from tensorflow_probability.python.internal import test_util\n'), ((26686, 26707), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (26705, 26707), False, 'from tensorflow_probability.python.internal import test_util\n'), ((26970, 26991), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (26989, 26991), False, 'from tensorflow_probability.python.internal import test_util\n'), ((13049, 13068), 'numpy.log', 'np.log', (['(2.0 * np.pi)'], {}), '(2.0 * np.pi)\n', (13055, 13068), True, 'import numpy as np\n'), ((25577, 25598), 'tensorflow_probability.python.internal.test_util.test_seed', 'test_util.test_seed', ([], {}), '()\n', (25596, 25598), False, 'from tensorflow_probability.python.internal import test_util\n'), ((35439, 35460), 'tensorflow.compat.v2.zeros', 'tf.zeros', (['batch_shape'], {}), '(batch_shape)\n', (35447, 35460), True, 'import tensorflow.compat.v2 as tf\n'), ((35495, 35515), 'tensorflow.compat.v2.ones', 'tf.ones', (['batch_shape'], {}), '(batch_shape)\n', (35502, 35515), True, 'import tensorflow.compat.v2 as tf\n')]
|
import numpy as np
from sklearn.decomposition import PCA
import pandas as pd
import matplotlib.pyplot as plt
import random
import seaborn as sns
from sklearn.cluster import KMeans
from sklearn.metrics import confusion_matrix
from sklearn.metrics.cluster import adjusted_rand_score
from sklearn.datasets import fetch_openml
from sklearn.model_selection import train_test_split
import pandas.util.testing as tm
from keras.datasets import mnist
import tensorflow_datasets as tfds
import tensorflow as tf
from google.colab import files
import sys
import itertools as it
#@title ElasticNetSubspaceClustering
import warnings
import progressbar
import spams
import time
from scipy import sparse
from sklearn import cluster
from sklearn.base import BaseEstimator, ClusterMixin
from sklearn.decomposition import sparse_encode
from sklearn.linear_model import orthogonal_mp
from sklearn.neighbors import kneighbors_graph
from sklearn.preprocessing import normalize
from sklearn.utils import check_random_state, check_array, check_symmetric
class SelfRepresentation(BaseEstimator, ClusterMixin):
def __init__(self, n_clusters=8, affinity='symmetrize', random_state=None, n_init=20, n_jobs=1):
self.n_clusters = n_clusters
self.affinity = affinity
self.random_state = random_state
self.n_init = n_init
self.n_jobs = n_jobs
def fit(self, X, y=None):
X = check_array(X, accept_sparse=['csr', 'csc', 'coo'], dtype=np.float64)
time_base = time.time()
self._self_representation(X)
self.timer_self_representation_ = time.time() - time_base
self._representation_to_affinity()
self._spectral_clustering()
self.timer_time_ = time.time() - time_base
return self
def fit_self_representation(self, X, y=None):
X = check_array(X, accept_sparse=['csr', 'csc', 'coo'], dtype=np.float64)
time_base = time.time()
self._self_representation(X)
self.timer_self_representation_ = time.time() - time_base
return self
def _representation_to_affinity(self):
normalized_representation_matrix_ = normalize(self.representation_matrix_, 'l2')
if self.affinity == 'symmetrize':
self.affinity_matrix_ = 0.5 * (np.absolute(normalized_representation_matrix_) + np.absolute(normalized_representation_matrix_.T))
elif self.affinity == 'nearest_neighbors':
neighbors_graph = kneighbors_graph(normalized_representation_matrix_, 3,
mode='connectivity', include_self=False)
self.affinity_matrix_ = 0.5 * (neighbors_graph + neighbors_graph.T)
def _spectral_clustering(self):
affinity_matrix_ = check_symmetric(self.affinity_matrix_)
random_state = check_random_state(self.random_state)
laplacian = sparse.csgraph.laplacian(affinity_matrix_, normed=True)
_, vec = sparse.linalg.eigsh(sparse.identity(laplacian.shape[0]) - laplacian,
k=self.n_clusters, sigma=None, which='LA')
embedding = normalize(vec)
_, self.labels_, _ = cluster.k_means(embedding, self.n_clusters,
random_state=random_state, n_init=self.n_init)
def active_support_elastic_net(X, y, alpha, tau=1.0, algorithm='spams', support_init='knn',
support_size=100, maxiter=40):
n_samples = X.shape[0]
if n_samples <= support_size: # skip active support search for small scale data
supp = np.arange(n_samples, dtype=int) # this results in the following iteration to converge in 1 iteration
else:
if support_init == 'L2':
L2sol = np.linalg.solve(np.identity(y.shape[1]) * alpha + np.dot(X.T, X), y.T)
c0 = np.dot(X, L2sol)[:, 0]
supp = np.argpartition(-np.abs(c0), support_size)[0:support_size]
elif support_init == 'knn':
supp = np.argpartition(-np.abs(np.dot(y, X.T)[0]), support_size)[0:support_size]
curr_obj = float("inf")
for _ in range(maxiter):
Xs = X[supp, :]
if algorithm == 'spams':
cs = spams.lasso(np.asfortranarray(y.T), D=np.asfortranarray(Xs.T),
lambda1=tau*alpha, lambda2=(1.0-tau)*alpha)
cs = np.asarray(cs.todense()).T
else:
cs = sparse_encode(y, Xs, algorithm=algorithm, alpha=alpha)
delta = (y - np.dot(cs, Xs)) / alpha
obj = tau * np.sum(np.abs(cs[0])) + (1.0 - tau)/2.0 * np.sum(np.power(cs[0], 2.0)) + alpha/2.0 * np.sum(np.power(delta, 2.0))
if curr_obj - obj < 1.0e-10 * curr_obj:
break
curr_obj = obj
coherence = np.abs(np.dot(delta, X.T))[0]
coherence[supp] = 0
addedsupp = np.nonzero(coherence > tau + 1.0e-10)[0]
if addedsupp.size == 0: # converged
break
# Find the set of nonzero entries of cs.
activesupp = supp[np.abs(cs[0]) > 1.0e-10]
if activesupp.size > 0.8 * support_size: # this suggests that support_size is too small and needs to be increased
support_size = min([round(max([activesupp.size, support_size]) * 1.1), n_samples])
if addedsupp.size + activesupp.size > support_size:
ord = np.argpartition(-coherence[addedsupp], support_size - activesupp.size)[0:support_size - activesupp.size]
addedsupp = addedsupp[ord]
supp = np.concatenate([activesupp, addedsupp])
c = np.zeros(n_samples)
c[supp] = cs
return c
def elastic_net_subspace_clustering(X, gamma=50.0, gamma_nz=True, tau=1.0, algorithm='lasso_lars',
active_support=True, active_support_params=None, n_nonzero=50):
if algorithm in ('lasso_lars', 'lasso_cd') and tau < 1.0 - 1.0e-10:
warnings.warn('algorithm {} cannot handle tau smaller than 1. Using tau = 1'.format(algorithm))
tau = 1.0
if active_support == True and active_support_params == None:
active_support_params = {}
n_samples = X.shape[0]
rows = np.zeros(n_samples * n_nonzero)
cols = np.zeros(n_samples * n_nonzero)
vals = np.zeros(n_samples * n_nonzero)
curr_pos = 0
for i in progressbar.progressbar(range(n_samples)):
y = X[i, :].copy().reshape(1, -1)
X[i, :] = 0
if algorithm in ('lasso_lars', 'lasso_cd', 'spams'):
if gamma_nz == True:
coh = np.delete(np.absolute(np.dot(X, y.T)), i)
alpha0 = np.amax(coh) / tau # value for which the solution is zero
alpha = alpha0 / gamma
else:
alpha = 1.0 / gamma
if active_support == True:
c = active_support_elastic_net(X, y, alpha, tau, algorithm, **active_support_params)
else:
if algorithm == 'spams':
c = spams.lasso(np.asfortranarray(y.T), D=np.asfortranarray(X.T),
lambda1=tau * alpha, lambda2=(1.0-tau) * alpha)
c = np.asarray(c.todense()).T[0]
else:
c = sparse_encode(y, X, algorithm=algorithm, alpha=alpha)[0]
else:
warnings.warn("algorithm {} not found".format(algorithm))
index = np.flatnonzero(c)
if index.size > n_nonzero:
# warnings.warn("The number of nonzero entries in sparse subspace clustering exceeds n_nonzero")
index = index[np.argsort(-np.absolute(c[index]))[0:n_nonzero]]
rows[curr_pos:curr_pos + len(index)] = i
cols[curr_pos:curr_pos + len(index)] = index
vals[curr_pos:curr_pos + len(index)] = c[index]
curr_pos += len(index)
X[i, :] = y
# affinity = sparse.csr_matrix((vals, (rows, cols)), shape=(n_samples, n_samples)) + sparse.csr_matrix((vals, (cols, rows)), shape=(n_samples, n_samples))
return sparse.csr_matrix((vals, (rows, cols)), shape=(n_samples, n_samples))
class ElasticNetSubspaceClustering(SelfRepresentation):
def __init__(self, n_clusters=8, affinity='symmetrize', random_state=None, n_init=20, n_jobs=1, gamma=50.0, gamma_nz=True, tau=1.0,
algorithm='lasso_lars', active_support=True, active_support_params=None, n_nonzero=50):
self.gamma = gamma
self.gamma_nz = gamma_nz
self.tau = tau
self.algorithm = algorithm
self.active_support = active_support
self.active_support_params = active_support_params
self.n_nonzero = n_nonzero
SelfRepresentation.__init__(self, n_clusters, affinity, random_state, n_init, n_jobs)
def _self_representation(self, X):
self.representation_matrix_ = elastic_net_subspace_clustering(X, self.gamma, self.gamma_nz,
self.tau, self.algorithm,
self.active_support, self.active_support_params,
self.n_nonzero)
def sparse_subspace_clustering_orthogonal_matching_pursuit(X, n_nonzero=10, thr=1.0e-6):
n_samples = X.shape[0]
rows = np.zeros(n_samples * n_nonzero, dtype = int)
cols = np.zeros(n_samples * n_nonzero, dtype = int)
vals = np.zeros(n_samples * n_nonzero)
curr_pos = 0
for i in progressbar.progressbar(range(n_samples)):
# for i in range(n_samples):
residual = X[i, :].copy() # initialize residual
supp = np.empty(shape=(0), dtype = int) # initialize support
residual_norm_thr = np.linalg.norm(X[i, :]) * thr
for t in range(n_nonzero): # for each iteration of OMP
# compute coherence between residuals and X
coherence = abs( np.matmul(residual, X.T) )
coherence[i] = 0.0
# update support
supp = np.append(supp, np.argmax(coherence))
# compute coefficients
c = np.linalg.lstsq( X[supp, :].T, X[i, :].T, rcond=None)[0]
# compute residual
residual = X[i, :] - np.matmul(c.T, X[supp, :])
# check termination
if np.sum(residual **2) < residual_norm_thr:
break
rows[curr_pos:curr_pos + len(supp)] = i
cols[curr_pos:curr_pos + len(supp)] = supp
vals[curr_pos:curr_pos + len(supp)] = c
curr_pos += len(supp)
# affinity = sparse.csr_matrix((vals, (rows, cols)), shape=(n_samples, n_samples)) + sparse.csr_matrix((vals, (cols, rows)), shape=(n_samples, n_samples))
return sparse.csr_matrix((vals, (rows, cols)), shape=(n_samples, n_samples))
class SparseSubspaceClusteringOMP(SelfRepresentation):
def __init__(self, n_clusters=8, affinity='symmetrize', random_state=None, n_init=10, n_jobs=1, n_nonzero=10, thr=1.0e-6):
self.n_nonzero = n_nonzero
self.thr = thr
SelfRepresentation.__init__(self, n_clusters, affinity, random_state, n_init, n_jobs)
def _self_representation(self, X):
self.representation_matrix_ = sparse_subspace_clustering_orthogonal_matching_pursuit(X, self.n_nonzero, self.thr)
def least_squares_subspace_clustering(X, gamma=10.0, exclude_self=False):
n_samples, n_features = X.shape
if exclude_self == False:
if n_samples < n_features:
gram = np.matmul(X, X.T)
return np.linalg.solve(gram + np.eye(n_sample) / gamma, gram).T
else:
tmp = np.linalg.solve(np.matmul(X.T, X) + np.eye(n_features) / gamma, X.T)
return np.matmul(X, tmp).T
else:
if n_samples < n_features:
D = np.linalg.solve(np.matmul(X, X.T) + np.eye(n_sample) / gamma, np.eye(n_sample))
# see Theorem 6 in https://arxiv.org/pdf/1404.6736.pdf
else:
tmp = np.linalg.solve(np.matmul(X.T, X) + np.eye(n_features) / gamma, X.T)
D = eye(n_samples) - np.matmul(X, tmp)
D = D / D.diagonal()[None,:]
np.fill_diagonal(D, 0.0)
return -1.0 * D.T
class LeastSquaresSubspaceClustering(SelfRepresentation):
def __init__(self, n_clusters=8, affinity='symmetrize', random_state=None, n_init=None, n_jobs=1, gamma=10.0, exclude_self=False):
self.gamma = gamma
self.exclude_self = exclude_self
SelfRepresentation.__init__(self, n_clusters, affinity, random_state, n_init, n_jobs)
def _self_representation(self, X):
self.representation_matrix_ = least_squares_subspace_clustering(X, self.gamma, self.exclude_self)
if 'google.colab' in sys.modules:
uploaded = files.upload()
#subtract the mean from every class
def preprocess_substract_mean(X, y):
labels = np.unique(y)
X_processed= X.copy()
for l in labels:
mean = np.average(X_processed[y == l], 0)
X_processed[y == l] = X_processed[y == l]- mean
return X_processed
def q_a(X,y):
#Run PCA on the dataset and plot the projection on the first 2 principal components, with each class marked in a different color/symbol
X_train_processed = preprocess_substract_mean(X, y)
pca = PCA(2) # project from 64 to 2 dimensions
projected = pca.fit_transform(X_train_processed)
#print(X_train_processed)
#print(projected.shape)
plt.scatter(projected[:, 0], projected[:, 1],
c=y_train, edgecolor='none', alpha=0.5,
cmap=plt.cm.get_cmap('tab10', 10))
plt.xlabel('component 1')
plt.ylabel('component 2')
# plt.colorbar();
plt.show()
q_a(X_train,y_train)
def angle_calucalte(p1, p2):
p1_u = p1 / np.linalg.norm(p1)
p2_u = p2 / np.linalg.norm(p2)
return (np.arccos(np.clip(np.dot(p1_u, p2_u), -1.0, 1.0)))
def q_b(X,y):
# Sample at least 5000 pairs of points from the same class and 5000 pairs of points from different classes,
labels = np.unique(y)
n=5000
cos_theta_in_all = np.empty( shape=(0, 0) )
cos_theta_out_all = np.empty( shape=(0, 0) )
num_labels = len(labels)
rand_indx1 = random.choices(range(len(X)), k=int(n))
rand_indx2 = list(pd.Series(rand_indx1).apply(lambda x: random.choices(Y.index[Y ==Y[x]])))
rand_indx2 = [j[0] for j in rand_indx2]
rand_indx3 = list(pd.Series(rand_indx1).apply(lambda x: random.choices(Y.index[Y !=Y[x]])))
rand_indx3 = [j[0] for j in rand_indx3]
points_in_1 = X_train.iloc[rand_indx1,:]
points_in_2 = X_train.iloc[rand_indx2,:]
points_out_1 = X_train.iloc[rand_indx3,:]
#compute the angle between every pair of points
theta_in_all = [angle_calucalte(points_in_1.iloc[i,:],points_in_2.iloc[i,:]) for i in range(len(points_in_1))]
theta_out_all = [angle_calucalte(points_in_1.iloc[i,:],points_out_1.iloc[i,:]) for i in range(len(points_in_1))]
# Plot the distribution of between-cluster angles and within cluster angles.
sns.distplot(theta_in_all,hist=True)
sns.distplot(theta_out_all,hist=True)
plt.legend(labels=['theta in', 'theta out'])
plt.show()
q_b(X_train,y_train)
l=5
pca = PCA()
pca.fit_transform(X_train)
#print(pca.explained_variance_ratio_.round(3))
np.cumsum(pca.explained_variance_ratio_).round(3)
def q_c(X,y):
# Perform PCA for each class separately, and plot for each class the proportion of variance explained vs the number of components ordered from the first PC until the last.
# What number of components would you take for further analysis?
labels = np.unique(y)
#fig1, ax1 = plt.subplots()
fig2, ax2 = plt.subplots()
for l in labels:
pca = PCA()
pca.fit_transform(X[y==l])
exp_var_ratio = pca.explained_variance_ratio_
#ax1.plot(exp_var_ratio,label=f'class {l}')
ax2.plot(np.cumsum(pca.explained_variance_ratio_),label=f'class {l}')
#ax1.set_title("Explained Variance per class")
ax2.set_title("Cumulated Explained Variance per class")
#ax1.legend()
ax2.legend()
#fig1.show()
fig2.show()
# Repeat but now with PCA for the entire dataset
#fig3, ax3 = plt.subplots()
fig4, ax4 = plt.subplots()
pca = PCA()
projected = pca.fit_transform(X)
exp_var_ratio = pca.explained_variance_ratio_
#x3.plot(exp_var_ratio)
ax4.plot(np.cumsum(exp_var_ratio))
#ax3.set_title("Explained Variance Global")
ax4.set_title("Cumulated Explained Variance Global")
#fig3.show()
fig4.show()
q_c(X_train,Y_train)
#What number of components would you take for further analysis?
pca = PCA(0.9)
pca.fit_transform(X_train)
print(f"The number of components necessary to explain 90% of the data is : {pca.n_components_}")
def performance_measure2(k,cluster1,cluster2):
data = {'cluster1': cluster1,'cluster2': cluster2}
clusters = pd.DataFrame(data, index=range(len(cluster1)))
all_per = list(it.permutations(range(k)))
accuracy_rate_all_per = np.zeros(len(all_per))
for l, per in enumerate(all_per) :
c = [i for i in range(k)]
dic = dict(zip(c,per))
clusters['premut_cluster'] = clusters['cluster2'].transform(lambda x: dic[x] if x in dic else None)
m = clusters.groupby(['cluster1','premut_cluster']).size().unstack(fill_value=0)
accuracy_rate_all_per[l]=np.trace(m)
cost_cluster = (accuracy_rate_all_per.max())/len(cluster1)
return (cost_cluster)
def performance_measure3(cluster1,cluster2):
data = {'cluster1': cluster1,'cluster2': cluster2}
clusters = pd.DataFrame(data, index=range(len(cluster1)))
m = -1*np.array(clusters.groupby(['cluster1','cluster2']).size().unstack(fill_value=0))
indx, per = linear_sum_assignment(m)
cost_cluster = -m[indx,per].sum()/len(clusters)
return (cost_cluster)
num_components = 85
pca =PCA(num_components)
pca_X =pca.fit_transform(X_train)
kmeans_after_PCA = KMeans(n_clusters=10).fit(pca_X)
kmeans_after_PCA.labels_
a,b
def q_d(X,y):
#Run the following algorithms on your dataset:
#For each algorithm, compute and report the clustering accuracy from eq. (6). Explain your results.
labels = np.unique(y)
K=10
#i. K-means with K = 10
kmeans = KMeans(n_clusters=K).fit(X)
kmeans_acc = performance_measure2(K,Y_train,kmeans.labels_)
#ii. PCA with the number of components chosen based on (c.), followed by K-means with K = 10 on the projection to the top components.
num_components = PCA(0.9).n_components_
pca =PCA(num_components)
pca_X =pca.fit_transform(X)
kmeans_after_PCA = KMeans(n_clusters=K).fit(pca_X)
kmeans_after_PCA.labels_
kmeans_pca_acc = performance_measure2(K,Y_train,kmeans_after_PCA.labels_)
#iii. A subspace clustering algorithm of your choice (ENsc), where you can set the number of clusters to the correct one, 10.
model_ensc = ElasticNetSubspaceClustering(n_clusters=K, algorithm='spams', gamma=500)
ensc_acc = performance_measure2(K,Y_train,model_ensc.fit(X).lables_)
print(f'kmeans acc is: {kmeans_acc} , pca followed by kmeans acc is : {kmeans_pca_acc}, ensc acc is {ensc_acc}')
q_d(X_train,y_train)
def main():
#X_train, y_train = load_mnist('data/fashion', kind='train')
#X_test, y_test = load_mnist('data/fashion', kind='t10k')
train_data = pd.read_csv('fashion-mnist_train.csv')
X_train = train_data.drop('label', axis=1)
y_train = train_data['label']
#X_train =X_train.astype(np.uint)
#y_train =y_train.astype(np.uint)
#X_test = X_test.astype(np.uint)
#y_test = y_test.astype(np.uint)
q_a(X_train, y_train)
q_b(X_train, y_train)
q_c(X_train, y_train)
q_d(X_train, y_train)
if __name__ == '__main__':
main()
|
[
"numpy.trace",
"pandas.read_csv",
"matplotlib.pyplot.ylabel",
"sklearn.neighbors.kneighbors_graph",
"random.choices",
"numpy.linalg.norm",
"sklearn.decomposition.sparse_encode",
"numpy.arange",
"sklearn.cluster.k_means",
"seaborn.distplot",
"sklearn.decomposition.PCA",
"matplotlib.pyplot.xlabel",
"numpy.flatnonzero",
"numpy.dot",
"numpy.empty",
"numpy.matmul",
"numpy.concatenate",
"numpy.linalg.lstsq",
"scipy.sparse.csr_matrix",
"numpy.identity",
"numpy.abs",
"numpy.eye",
"sklearn.utils.check_random_state",
"numpy.average",
"numpy.argmax",
"numpy.fill_diagonal",
"numpy.asfortranarray",
"numpy.nonzero",
"matplotlib.pyplot.cm.get_cmap",
"scipy.sparse.identity",
"time.time",
"matplotlib.pyplot.legend",
"matplotlib.pyplot.show",
"sklearn.cluster.KMeans",
"sklearn.utils.check_symmetric",
"pandas.Series",
"numpy.unique",
"numpy.argpartition",
"scipy.sparse.csgraph.laplacian",
"numpy.power",
"numpy.absolute",
"google.colab.files.upload",
"numpy.sum",
"numpy.zeros",
"sklearn.utils.check_array",
"numpy.cumsum",
"sklearn.preprocessing.normalize",
"numpy.amax",
"matplotlib.pyplot.subplots"
] |
[((15196, 15201), 'sklearn.decomposition.PCA', 'PCA', ([], {}), '()\n', (15199, 15201), False, 'from sklearn.decomposition import PCA\n'), ((16649, 16657), 'sklearn.decomposition.PCA', 'PCA', (['(0.9)'], {}), '(0.9)\n', (16652, 16657), False, 'from sklearn.decomposition import PCA\n'), ((17838, 17857), 'sklearn.decomposition.PCA', 'PCA', (['num_components'], {}), '(num_components)\n', (17841, 17857), False, 'from sklearn.decomposition import PCA\n'), ((5622, 5641), 'numpy.zeros', 'np.zeros', (['n_samples'], {}), '(n_samples)\n', (5630, 5641), True, 'import numpy as np\n'), ((6214, 6245), 'numpy.zeros', 'np.zeros', (['(n_samples * n_nonzero)'], {}), '(n_samples * n_nonzero)\n', (6222, 6245), True, 'import numpy as np\n'), ((6257, 6288), 'numpy.zeros', 'np.zeros', (['(n_samples * n_nonzero)'], {}), '(n_samples * n_nonzero)\n', (6265, 6288), True, 'import numpy as np\n'), ((6300, 6331), 'numpy.zeros', 'np.zeros', (['(n_samples * n_nonzero)'], {}), '(n_samples * n_nonzero)\n', (6308, 6331), True, 'import numpy as np\n'), ((8066, 8135), 'scipy.sparse.csr_matrix', 'sparse.csr_matrix', (['(vals, (rows, cols))'], {'shape': '(n_samples, n_samples)'}), '((vals, (rows, cols)), shape=(n_samples, n_samples))\n', (8083, 8135), False, 'from scipy import sparse\n'), ((9356, 9398), 'numpy.zeros', 'np.zeros', (['(n_samples * n_nonzero)'], {'dtype': 'int'}), '(n_samples * n_nonzero, dtype=int)\n', (9364, 9398), True, 'import numpy as np\n'), ((9412, 9454), 'numpy.zeros', 'np.zeros', (['(n_samples * n_nonzero)'], {'dtype': 'int'}), '(n_samples * n_nonzero, dtype=int)\n', (9420, 9454), True, 'import numpy as np\n'), ((9468, 9499), 'numpy.zeros', 'np.zeros', (['(n_samples * n_nonzero)'], {}), '(n_samples * n_nonzero)\n', (9476, 9499), True, 'import numpy as np\n'), ((10749, 10818), 'scipy.sparse.csr_matrix', 'sparse.csr_matrix', (['(vals, (rows, cols))'], {'shape': '(n_samples, n_samples)'}), '((vals, (rows, cols)), shape=(n_samples, n_samples))\n', (10766, 10818), False, 'from scipy import sparse\n'), ((12777, 12791), 'google.colab.files.upload', 'files.upload', ([], {}), '()\n', (12789, 12791), False, 'from google.colab import files\n'), ((12879, 12891), 'numpy.unique', 'np.unique', (['y'], {}), '(y)\n', (12888, 12891), True, 'import numpy as np\n'), ((13286, 13292), 'sklearn.decomposition.PCA', 'PCA', (['(2)'], {}), '(2)\n', (13289, 13292), False, 'from sklearn.decomposition import PCA\n'), ((13600, 13625), 'matplotlib.pyplot.xlabel', 'plt.xlabel', (['"""component 1"""'], {}), "('component 1')\n", (13610, 13625), True, 'import matplotlib.pyplot as plt\n'), ((13630, 13655), 'matplotlib.pyplot.ylabel', 'plt.ylabel', (['"""component 2"""'], {}), "('component 2')\n", (13640, 13655), True, 'import matplotlib.pyplot as plt\n'), ((13682, 13692), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (13690, 13692), True, 'import matplotlib.pyplot as plt\n'), ((14023, 14035), 'numpy.unique', 'np.unique', (['y'], {}), '(y)\n', (14032, 14035), True, 'import numpy as np\n'), ((14070, 14092), 'numpy.empty', 'np.empty', ([], {'shape': '(0, 0)'}), '(shape=(0, 0))\n', (14078, 14092), True, 'import numpy as np\n'), ((14119, 14141), 'numpy.empty', 'np.empty', ([], {'shape': '(0, 0)'}), '(shape=(0, 0))\n', (14127, 14141), True, 'import numpy as np\n'), ((15016, 15053), 'seaborn.distplot', 'sns.distplot', (['theta_in_all'], {'hist': '(True)'}), '(theta_in_all, hist=True)\n', (15028, 15053), True, 'import seaborn as sns\n'), ((15057, 15095), 'seaborn.distplot', 'sns.distplot', (['theta_out_all'], {'hist': '(True)'}), '(theta_out_all, hist=True)\n', (15069, 15095), True, 'import seaborn as sns\n'), ((15099, 15143), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'labels': "['theta in', 'theta out']"}), "(labels=['theta in', 'theta out'])\n", (15109, 15143), True, 'import matplotlib.pyplot as plt\n'), ((15148, 15158), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (15156, 15158), True, 'import matplotlib.pyplot as plt\n'), ((15606, 15618), 'numpy.unique', 'np.unique', (['y'], {}), '(y)\n', (15615, 15618), True, 'import numpy as np\n'), ((15667, 15681), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {}), '()\n', (15679, 15681), True, 'import matplotlib.pyplot as plt\n'), ((16224, 16238), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {}), '()\n', (16236, 16238), True, 'import matplotlib.pyplot as plt\n'), ((16249, 16254), 'sklearn.decomposition.PCA', 'PCA', ([], {}), '()\n', (16252, 16254), False, 'from sklearn.decomposition import PCA\n'), ((18164, 18176), 'numpy.unique', 'np.unique', (['y'], {}), '(y)\n', (18173, 18176), True, 'import numpy as np\n'), ((18522, 18541), 'sklearn.decomposition.PCA', 'PCA', (['num_components'], {}), '(num_components)\n', (18525, 18541), False, 'from sklearn.decomposition import PCA\n'), ((19330, 19368), 'pandas.read_csv', 'pd.read_csv', (['"""fashion-mnist_train.csv"""'], {}), "('fashion-mnist_train.csv')\n", (19341, 19368), True, 'import pandas as pd\n'), ((1406, 1475), 'sklearn.utils.check_array', 'check_array', (['X'], {'accept_sparse': "['csr', 'csc', 'coo']", 'dtype': 'np.float64'}), "(X, accept_sparse=['csr', 'csc', 'coo'], dtype=np.float64)\n", (1417, 1475), False, 'from sklearn.utils import check_random_state, check_array, check_symmetric\n'), ((1496, 1507), 'time.time', 'time.time', ([], {}), '()\n', (1505, 1507), False, 'import time\n'), ((1844, 1913), 'sklearn.utils.check_array', 'check_array', (['X'], {'accept_sparse': "['csr', 'csc', 'coo']", 'dtype': 'np.float64'}), "(X, accept_sparse=['csr', 'csc', 'coo'], dtype=np.float64)\n", (1855, 1913), False, 'from sklearn.utils import check_random_state, check_array, check_symmetric\n'), ((1934, 1945), 'time.time', 'time.time', ([], {}), '()\n', (1943, 1945), False, 'import time\n'), ((2175, 2219), 'sklearn.preprocessing.normalize', 'normalize', (['self.representation_matrix_', '"""l2"""'], {}), "(self.representation_matrix_, 'l2')\n", (2184, 2219), False, 'from sklearn.preprocessing import normalize\n'), ((2767, 2805), 'sklearn.utils.check_symmetric', 'check_symmetric', (['self.affinity_matrix_'], {}), '(self.affinity_matrix_)\n', (2782, 2805), False, 'from sklearn.utils import check_random_state, check_array, check_symmetric\n'), ((2829, 2866), 'sklearn.utils.check_random_state', 'check_random_state', (['self.random_state'], {}), '(self.random_state)\n', (2847, 2866), False, 'from sklearn.utils import check_random_state, check_array, check_symmetric\n'), ((2896, 2951), 'scipy.sparse.csgraph.laplacian', 'sparse.csgraph.laplacian', (['affinity_matrix_'], {'normed': '(True)'}), '(affinity_matrix_, normed=True)\n', (2920, 2951), False, 'from scipy import sparse\n'), ((3139, 3153), 'sklearn.preprocessing.normalize', 'normalize', (['vec'], {}), '(vec)\n', (3148, 3153), False, 'from sklearn.preprocessing import normalize\n'), ((3183, 3277), 'sklearn.cluster.k_means', 'cluster.k_means', (['embedding', 'self.n_clusters'], {'random_state': 'random_state', 'n_init': 'self.n_init'}), '(embedding, self.n_clusters, random_state=random_state,\n n_init=self.n_init)\n', (3198, 3277), False, 'from sklearn import cluster\n'), ((3605, 3636), 'numpy.arange', 'np.arange', (['n_samples'], {'dtype': 'int'}), '(n_samples, dtype=int)\n', (3614, 3636), True, 'import numpy as np\n'), ((5569, 5608), 'numpy.concatenate', 'np.concatenate', (['[activesupp, addedsupp]'], {}), '([activesupp, addedsupp])\n', (5583, 5608), True, 'import numpy as np\n'), ((7447, 7464), 'numpy.flatnonzero', 'np.flatnonzero', (['c'], {}), '(c)\n', (7461, 7464), True, 'import numpy as np\n'), ((9679, 9707), 'numpy.empty', 'np.empty', ([], {'shape': '(0)', 'dtype': 'int'}), '(shape=0, dtype=int)\n', (9687, 9707), True, 'import numpy as np\n'), ((12166, 12190), 'numpy.fill_diagonal', 'np.fill_diagonal', (['D', '(0.0)'], {}), '(D, 0.0)\n', (12182, 12190), True, 'import numpy as np\n'), ((12954, 12988), 'numpy.average', 'np.average', (['X_processed[y == l]', '(0)'], {}), '(X_processed[y == l], 0)\n', (12964, 12988), True, 'import numpy as np\n'), ((13765, 13783), 'numpy.linalg.norm', 'np.linalg.norm', (['p1'], {}), '(p1)\n', (13779, 13783), True, 'import numpy as np\n'), ((13800, 13818), 'numpy.linalg.norm', 'np.linalg.norm', (['p2'], {}), '(p2)\n', (13814, 13818), True, 'import numpy as np\n'), ((15278, 15318), 'numpy.cumsum', 'np.cumsum', (['pca.explained_variance_ratio_'], {}), '(pca.explained_variance_ratio_)\n', (15287, 15318), True, 'import numpy as np\n'), ((15717, 15722), 'sklearn.decomposition.PCA', 'PCA', ([], {}), '()\n', (15720, 15722), False, 'from sklearn.decomposition import PCA\n'), ((16385, 16409), 'numpy.cumsum', 'np.cumsum', (['exp_var_ratio'], {}), '(exp_var_ratio)\n', (16394, 16409), True, 'import numpy as np\n'), ((17350, 17361), 'numpy.trace', 'np.trace', (['m'], {}), '(m)\n', (17358, 17361), True, 'import numpy as np\n'), ((17911, 17932), 'sklearn.cluster.KMeans', 'KMeans', ([], {'n_clusters': '(10)'}), '(n_clusters=10)\n', (17917, 17932), False, 'from sklearn.cluster import KMeans\n'), ((18490, 18498), 'sklearn.decomposition.PCA', 'PCA', (['(0.9)'], {}), '(0.9)\n', (18493, 18498), False, 'from sklearn.decomposition import PCA\n'), ((1596, 1607), 'time.time', 'time.time', ([], {}), '()\n', (1605, 1607), False, 'import time\n'), ((1735, 1746), 'time.time', 'time.time', ([], {}), '()\n', (1744, 1746), False, 'import time\n'), ((2034, 2045), 'time.time', 'time.time', ([], {}), '()\n', (2043, 2045), False, 'import time\n'), ((4436, 4490), 'sklearn.decomposition.sparse_encode', 'sparse_encode', (['y', 'Xs'], {'algorithm': 'algorithm', 'alpha': 'alpha'}), '(y, Xs, algorithm=algorithm, alpha=alpha)\n', (4449, 4490), False, 'from sklearn.decomposition import sparse_encode\n'), ((4871, 4906), 'numpy.nonzero', 'np.nonzero', (['(coherence > tau + 1e-10)'], {}), '(coherence > tau + 1e-10)\n', (4881, 4906), True, 'import numpy as np\n'), ((9762, 9785), 'numpy.linalg.norm', 'np.linalg.norm', (['X[i, :]'], {}), '(X[i, :])\n', (9776, 9785), True, 'import numpy as np\n'), ((11525, 11542), 'numpy.matmul', 'np.matmul', (['X', 'X.T'], {}), '(X, X.T)\n', (11534, 11542), True, 'import numpy as np\n'), ((13566, 13594), 'matplotlib.pyplot.cm.get_cmap', 'plt.cm.get_cmap', (['"""tab10"""', '(10)'], {}), "('tab10', 10)\n", (13581, 13594), True, 'import matplotlib.pyplot as plt\n'), ((13849, 13867), 'numpy.dot', 'np.dot', (['p1_u', 'p2_u'], {}), '(p1_u, p2_u)\n', (13855, 13867), True, 'import numpy as np\n'), ((15882, 15922), 'numpy.cumsum', 'np.cumsum', (['pca.explained_variance_ratio_'], {}), '(pca.explained_variance_ratio_)\n', (15891, 15922), True, 'import numpy as np\n'), ((18228, 18248), 'sklearn.cluster.KMeans', 'KMeans', ([], {'n_clusters': 'K'}), '(n_clusters=K)\n', (18234, 18248), False, 'from sklearn.cluster import KMeans\n'), ((18597, 18617), 'sklearn.cluster.KMeans', 'KMeans', ([], {'n_clusters': 'K'}), '(n_clusters=K)\n', (18603, 18617), False, 'from sklearn.cluster import KMeans\n'), ((2485, 2584), 'sklearn.neighbors.kneighbors_graph', 'kneighbors_graph', (['normalized_representation_matrix_', '(3)'], {'mode': '"""connectivity"""', 'include_self': '(False)'}), "(normalized_representation_matrix_, 3, mode='connectivity',\n include_self=False)\n", (2501, 2584), False, 'from sklearn.neighbors import kneighbors_graph\n'), ((2989, 3024), 'scipy.sparse.identity', 'sparse.identity', (['laplacian.shape[0]'], {}), '(laplacian.shape[0])\n', (3004, 3024), False, 'from scipy import sparse\n'), ((3862, 3878), 'numpy.dot', 'np.dot', (['X', 'L2sol'], {}), '(X, L2sol)\n', (3868, 3878), True, 'import numpy as np\n'), ((4236, 4258), 'numpy.asfortranarray', 'np.asfortranarray', (['y.T'], {}), '(y.T)\n', (4253, 4258), True, 'import numpy as np\n'), ((4519, 4533), 'numpy.dot', 'np.dot', (['cs', 'Xs'], {}), '(cs, Xs)\n', (4525, 4533), True, 'import numpy as np\n'), ((4800, 4818), 'numpy.dot', 'np.dot', (['delta', 'X.T'], {}), '(delta, X.T)\n', (4806, 4818), True, 'import numpy as np\n'), ((5060, 5073), 'numpy.abs', 'np.abs', (['cs[0]'], {}), '(cs[0])\n', (5066, 5073), True, 'import numpy as np\n'), ((5401, 5471), 'numpy.argpartition', 'np.argpartition', (['(-coherence[addedsupp])', '(support_size - activesupp.size)'], {}), '(-coherence[addedsupp], support_size - activesupp.size)\n', (5416, 5471), True, 'import numpy as np\n'), ((9948, 9972), 'numpy.matmul', 'np.matmul', (['residual', 'X.T'], {}), '(residual, X.T)\n', (9957, 9972), True, 'import numpy as np\n'), ((10070, 10090), 'numpy.argmax', 'np.argmax', (['coherence'], {}), '(coherence)\n', (10079, 10090), True, 'import numpy as np\n'), ((10143, 10195), 'numpy.linalg.lstsq', 'np.linalg.lstsq', (['X[supp, :].T', 'X[i, :].T'], {'rcond': 'None'}), '(X[supp, :].T, X[i, :].T, rcond=None)\n', (10158, 10195), True, 'import numpy as np\n'), ((10264, 10290), 'numpy.matmul', 'np.matmul', (['c.T', 'X[supp, :]'], {}), '(c.T, X[supp, :])\n', (10273, 10290), True, 'import numpy as np\n'), ((10338, 10359), 'numpy.sum', 'np.sum', (['(residual ** 2)'], {}), '(residual ** 2)\n', (10344, 10359), True, 'import numpy as np\n'), ((11739, 11756), 'numpy.matmul', 'np.matmul', (['X', 'tmp'], {}), '(X, tmp)\n', (11748, 11756), True, 'import numpy as np\n'), ((11882, 11898), 'numpy.eye', 'np.eye', (['n_sample'], {}), '(n_sample)\n', (11888, 11898), True, 'import numpy as np\n'), ((12103, 12120), 'numpy.matmul', 'np.matmul', (['X', 'tmp'], {}), '(X, tmp)\n', (12112, 12120), True, 'import numpy as np\n'), ((14252, 14273), 'pandas.Series', 'pd.Series', (['rand_indx1'], {}), '(rand_indx1)\n', (14261, 14273), True, 'import pandas as pd\n'), ((14290, 14324), 'random.choices', 'random.choices', (['Y.index[Y == Y[x]]'], {}), '(Y.index[Y == Y[x]])\n', (14304, 14324), False, 'import random\n'), ((14392, 14413), 'pandas.Series', 'pd.Series', (['rand_indx1'], {}), '(rand_indx1)\n', (14401, 14413), True, 'import pandas as pd\n'), ((14430, 14464), 'random.choices', 'random.choices', (['Y.index[Y != Y[x]]'], {}), '(Y.index[Y != Y[x]])\n', (14444, 14464), False, 'import random\n'), ((2305, 2351), 'numpy.absolute', 'np.absolute', (['normalized_representation_matrix_'], {}), '(normalized_representation_matrix_)\n', (2316, 2351), True, 'import numpy as np\n'), ((2354, 2402), 'numpy.absolute', 'np.absolute', (['normalized_representation_matrix_.T'], {}), '(normalized_representation_matrix_.T)\n', (2365, 2402), True, 'import numpy as np\n'), ((3824, 3838), 'numpy.dot', 'np.dot', (['X.T', 'X'], {}), '(X.T, X)\n', (3830, 3838), True, 'import numpy as np\n'), ((4262, 4285), 'numpy.asfortranarray', 'np.asfortranarray', (['Xs.T'], {}), '(Xs.T)\n', (4279, 4285), True, 'import numpy as np\n'), ((4658, 4678), 'numpy.power', 'np.power', (['delta', '(2.0)'], {}), '(delta, 2.0)\n', (4666, 4678), True, 'import numpy as np\n'), ((6661, 6673), 'numpy.amax', 'np.amax', (['coh'], {}), '(coh)\n', (6668, 6673), True, 'import numpy as np\n'), ((11667, 11684), 'numpy.matmul', 'np.matmul', (['X.T', 'X'], {}), '(X.T, X)\n', (11676, 11684), True, 'import numpy as np\n'), ((11836, 11853), 'numpy.matmul', 'np.matmul', (['X', 'X.T'], {}), '(X, X.T)\n', (11845, 11853), True, 'import numpy as np\n'), ((12017, 12034), 'numpy.matmul', 'np.matmul', (['X.T', 'X'], {}), '(X.T, X)\n', (12026, 12034), True, 'import numpy as np\n'), ((3790, 3813), 'numpy.identity', 'np.identity', (['y.shape[1]'], {}), '(y.shape[1])\n', (3801, 3813), True, 'import numpy as np\n'), ((3921, 3931), 'numpy.abs', 'np.abs', (['c0'], {}), '(c0)\n', (3927, 3931), True, 'import numpy as np\n'), ((4573, 4586), 'numpy.abs', 'np.abs', (['cs[0]'], {}), '(cs[0])\n', (4579, 4586), True, 'import numpy as np\n'), ((4615, 4635), 'numpy.power', 'np.power', (['cs[0]', '(2.0)'], {}), '(cs[0], 2.0)\n', (4623, 4635), True, 'import numpy as np\n'), ((6616, 6630), 'numpy.dot', 'np.dot', (['X', 'y.T'], {}), '(X, y.T)\n', (6622, 6630), True, 'import numpy as np\n'), ((7049, 7071), 'numpy.asfortranarray', 'np.asfortranarray', (['y.T'], {}), '(y.T)\n', (7066, 7071), True, 'import numpy as np\n'), ((7283, 7336), 'sklearn.decomposition.sparse_encode', 'sparse_encode', (['y', 'X'], {'algorithm': 'algorithm', 'alpha': 'alpha'}), '(y, X, algorithm=algorithm, alpha=alpha)\n', (7296, 7336), False, 'from sklearn.decomposition import sparse_encode\n'), ((11687, 11705), 'numpy.eye', 'np.eye', (['n_features'], {}), '(n_features)\n', (11693, 11705), True, 'import numpy as np\n'), ((11856, 11872), 'numpy.eye', 'np.eye', (['n_sample'], {}), '(n_sample)\n', (11862, 11872), True, 'import numpy as np\n'), ((12037, 12055), 'numpy.eye', 'np.eye', (['n_features'], {}), '(n_features)\n', (12043, 12055), True, 'import numpy as np\n'), ((7075, 7097), 'numpy.asfortranarray', 'np.asfortranarray', (['X.T'], {}), '(X.T)\n', (7092, 7097), True, 'import numpy as np\n'), ((7642, 7663), 'numpy.absolute', 'np.absolute', (['c[index]'], {}), '(c[index])\n', (7653, 7663), True, 'import numpy as np\n'), ((11585, 11601), 'numpy.eye', 'np.eye', (['n_sample'], {}), '(n_sample)\n', (11591, 11601), True, 'import numpy as np\n'), ((4042, 4056), 'numpy.dot', 'np.dot', (['y', 'X.T'], {}), '(y, X.T)\n', (4048, 4056), True, 'import numpy as np\n')]
|
#!/usr/bin/env python3
# class file uppergeodesic.py
# started as a script to visualize what happens to hyperbolic plane
# if different isometries act on it
import geodesic as gd
import numpy as np
import numpy.linalg as lina
import matplotlib.pyplot as plt
# upper half space as the basic model
class UpperGeodesic(gd.Geodesic):
"""UpperGeodesic line in upper half space
takes endpoints on boundary as arguments
stores x and y data as points in x and y
string "inf" is point at infinity, i.e. y=inf
"""
xmin = 0
xmax = 0
ymin = 0 # just for consistency, shouldn't change
ymax = 0
inf = "inf"
def __init__(self, a, b, color="b", label=''):
"""initialize UpperGeodesic by endpoints
a, b - x values of the endpoints or "inf" if infinity
res is resolution
"""
super().__init__(a, b, color, label)
# adjust the boundaries of hyperbolic space
if self.start != UpperGeodesic.inf:
if self.start < UpperGeodesic.xmin:
UpperGeodesic.xmin = self.start
if self.end > UpperGeodesic.xmax:
UpperGeodesic.xmax = self.end
UpperGeodesic.ymax = (UpperGeodesic.xmax - UpperGeodesic.xmin)/2
@classmethod
def vertical(cls, real):
return cls(cls.inf, real)
@classmethod
def from_midpoint_and_radius(cls, m, r):
"""
m is only the real part of the circle thing
"""
return cls(m-r, m+r)
def sort_se(self):
"""sort start and end"""
if self.end == self.inf:
# just want to assume that the first value is inf if any
self.end = self.start
self.start = self.inf
if self.start != self.inf and self.end < self.start:
# swap a and self.end such that a < self.end
c = self.start
self.start = self.end
self.end = c
def get_data(self):
if self.start == UpperGeodesic.inf:
# vertical line
xs = [self.end, self.end]
ys = [self.ymin, self.ymax]
else:
# calculate semicircle
t = np.linspace(0, np.pi, self.res)
r = (self.end - self.start)/2
xs = r*(1 + np.cos(t)) + self.start
ys = r*np.sin(t)
return(xs, ys)
## the next two functions create new geodesics from existing ones
def new_geod(self, a, b, c, d):
"""return new geodesic by received by moebius trafo
apply the matrix
| a b |
| c d |
on the geodesic self and return the resulting geodesic
"""
start = self.apply_moebius(a, b, c, d, self.start)
end = self.apply_moebius(a, b, c, d, self.end)
return(UpperGeodesic(start, end))
def new_from_matrix(self, M):
return self.new_geod(M[0,0], M[0,1], M[1,0], M[1,1])
## apply transformations to ONE geodesic
def apply_matrix(self, M):
self.start = self.apply_moebius(M[0,0], M[0,1], M[1, 0], M[1,1],
self.start)
self.end = self.apply_moebius(M[0,0], M[0,1], M[1, 0], M[1,1],
self.end)
self.sort_se()
def translate_one_geod(self, dx):
if self.start != UpperGeodesic.inf:
self.start += dx
if self.end != UpperGeodesic.inf:
self.end += dx
def translate_one_at_zero(self, dx):
"""inverts at unit sphere, translates and inverts again"""
a = self.inversion_on_unit_circle(self.start)
b = self.inversion_on_unit_circle(self.end)
if a != UpperGeodesic.inf:
a += dx
if b != UpperGeodesic.inf:
b += dx
self.start = self.inversion_on_unit_circle(a)
self.end = self.inversion_on_unit_circle(b)
self.sort_se()
def rotate_one_geod(self, phi):
"""rotates the geodesic on upper half space
conjugate to a rotation around origin in the disc model
"""
if self.start == UpperGeodesic.inf:
alpha = -np.pi/2
else:
alpha = self.from_upper_to_disc(self.start)
beta = self.from_upper_to_disc(self.end)
alpha += phi
beta += phi
self.start = self.from_disc_to_upper(alpha)
self.end = self.from_disc_to_upper(beta)
self.sort_se()
def hyperbolic_translate_one(self, dmult=1.001):
"""translates one geodesic along UpperGeodesic(-1,1)"""
diag = (dmult + 1.0/dmult)/2.0
off = (1.0/dmult - dmult)/2.0
matrix = np.matrix([[diag, off], [off, diag]])
self.apply_matrix(matrix)
# tesselate hyperbolic space
@classmethod
def tesselate(self, depth=10):
"""Tesselates according to SL(2,Z)"""
g0 = UpperGeodesic(-1,1, "r")
g1 = UpperGeodesic(-0.5,self.inf, "r")
g2 = UpperGeodesic(0.5,self.inf, "r")
first = [g0,g1,g2]
for k in range(1, depth):
for g in first:
g.new_geod(1, k, 0, 1)
g.new_geod(1, -k, 0, 1)
kmax = len(UpperGeodesic.all_geods)
for geod in UpperGeodesic.all_geods[:kmax]:
temp = [geod.new_geod(0, -1, 1, 0)]
for k in range(1, 2*depth):
temp.append(geod.new_geod(1, 0, k, 1))
temp.append(geod.new_geod(1, 0, -k, 1))
for k in range(1, depth//2):
for t in temp:
t.new_geod(1, k, 0, 1)
t.new_geod(1, -k, 0, 1)
UpperGeodesic.xmin= -3
UpperGeodesic.xmax= 3
UpperGeodesic.ymax= 3
## plot commands
@classmethod
def set_plot_limits(cls):
highest = max(abs(i)
for i in [cls.ymin, cls.ymax, cls.xmax, cls.xmin])
cls.ax.axis([-highest, highest, 0, highest])
@classmethod
def plot_all(cls):
if UpperGeodesic.ymax <= UpperGeodesic.ymin:
UpperGeodesic.ymax = UpperGeodesic.ymin + 1 # else nothing to plot
super().plot_all()
|
[
"numpy.sin",
"numpy.matrix",
"numpy.linspace",
"numpy.cos"
] |
[((4613, 4650), 'numpy.matrix', 'np.matrix', (['[[diag, off], [off, diag]]'], {}), '([[diag, off], [off, diag]])\n', (4622, 4650), True, 'import numpy as np\n'), ((2161, 2192), 'numpy.linspace', 'np.linspace', (['(0)', 'np.pi', 'self.res'], {}), '(0, np.pi, self.res)\n', (2172, 2192), True, 'import numpy as np\n'), ((2302, 2311), 'numpy.sin', 'np.sin', (['t'], {}), '(t)\n', (2308, 2311), True, 'import numpy as np\n'), ((2259, 2268), 'numpy.cos', 'np.cos', (['t'], {}), '(t)\n', (2265, 2268), True, 'import numpy as np\n')]
|
"""
Created on Thu Sept 24 2020-
@author: <NAME>
GitHub username: esgomezm
"""
import tensorflow as tf
from tensorflow.keras import backend as K
from tensorflow.keras.losses import binary_crossentropy
import numpy as np
from tensorflow.keras import losses
# --------------------------------
# ## Unet with tf 2.0.0
# https://www.kaggle.com/advaitsave/tensorflow-2-nuclei-segmentation-unet
# ## binary weighted loss example
# https://www.kaggle.com/lyakaap/weighing-boundary-pixels-loss-script-by-keras2
# https://stackoverflow.com/questions/48555820/keras-binary-segmentation-add-weight-to-loss-function/48577360
# https://stackoverflow.com/questions/55213599/u-net-with-pixel-wise-weighted-cross-entropy-input-dimension-errors
# https://lars76.github.io/neural-networks/object-detection/losses-for-segmentation/
# https://stackoverflow.com/questions/46858016/keras-custom-loss-function-to-pass-arguments-other-than-y-true-and-y-pred
# --------------------------------
# weight: weighted tensor(same s☺hape as mask image)
def weighted_bce(y_true, y_pred, weight):
# avoiding overflow
epsilon = 1e-7
y_pred = K.clip(y_pred, epsilon, 1. - epsilon)
logit_y_pred = K.log(y_pred / (1. - y_pred))
# https://www.tensorflow.org/api_docs/python/tf/nn/weighted_cross_entropy_with_logits
loss = (1. - y_true) * logit_y_pred + (1. + (weight - 1.) * y_true) * \
(K.log(1. + K.exp(-K.abs(logit_y_pred))) + K.maximum(-logit_y_pred, 0.))
return K.sum(loss) / K.sum(weight)
def weighted_dice(y_true, y_pred, weight):
smooth = 1.
w, m1, m2 = weight * weight, y_true, y_pred
intersection = (m1 * m2)
score = (2. * K.sum(w * intersection) + smooth) / (K.sum(w * m1) + K.sum(w * m2) + smooth)
loss = 1. - K.sum(score)
return loss
def weighted_bce_dice_loss(y_true, y_pred):
y_true = K.cast(y_true, 'float32')
y_pred = K.cast(y_pred, 'float32')
# if we want to get same size of output, kernel size must be odd number
averaged_mask = K.pool2d(
y_true, pool_size=(11, 11), strides=(1, 1), padding='same', pool_mode='avg')
border = K.cast(K.greater(averaged_mask, 0.005), 'float32') * K.cast(K.less(averaged_mask, 0.995), 'float32')
weight = K.ones_like(averaged_mask)
w0 = K.sum(weight)
weight += border * 2
w1 = K.sum(weight)
weight *= (w0 / w1)
loss = weighted_bce(y_true, y_pred, weight) + weighted_dice(y_true, y_pred, weight)
return loss
def bce_loss(X):
# y_true, y_pred, weight = X
y_true, y_pred = X
loss = binary_crossentropy(y_true, y_pred)
loss = tf.expand_dims(loss, 3)
# loss = multiply([loss, weight])
return loss
def identity_loss(y_true, y_pred):
# return K.mean(y_pred, axis=-1)
return y_pred
def jaccard_multiple_output(y_true, y_pred, from_logits = True):
"""Define Jaccard index for multiple labels.
Args:
y_true (tensor): ground truth masks.
y_pred (tensor): predicted masks.
Return:
jac (tensor): Jaccard index value
"""
if from_logits:
# run activation to evaluate the jaccard index
y_pred_ = tf.sigmoid(y_pred)
y_pred_ = y_pred_ > 0.5
y_pred_ = tf.cast(y_pred_, dtype=tf.int8)
y_true_ = tf.cast(y_true, dtype=tf.int8)
TP = tf.math.count_nonzero(y_pred_ * y_true_)
FP = tf.math.count_nonzero(y_pred_ * (1 - y_true_))
FN = tf.math.count_nonzero((1 - y_pred_) * y_true_)
jac = tf.cond(tf.greater((TP + FP + FN), 0), lambda: TP / (TP + FP + FN),
lambda: tf.cast(0.000, dtype='float64'))
return jac
def jaccard_sparse(y_true, y_pred, skip_background=True):
"""Define Jaccard index (multi-class).
Args:
y_true (tensor): ground truth masks.
y_pred (tensor): predicted masks.
skip_background (bool, optional): skip background label.
Return:
jac (tensor): Jaccard index value
"""
# number of classes (last dimension of predictions)
num_classes = tf.shape(y_pred)[-1]
# one_hot representation of predicted segmentation
y_pred_ = tf.cast(y_pred, dtype=tf.int32)
y_pred_ = tf.one_hot(tf.math.argmax(y_pred_, axis=-1), num_classes, axis=-1)
# one_hot representation of ground truth segmentation
y_true_ = tf.cast(y_true[...,0], dtype=tf.int32)
y_true_ = tf.one_hot(y_true_, num_classes, axis=-1)
if skip_background:
y_pred_ = y_pred_[...,1:]
y_true_ = y_true_[...,1:]
TP = tf.math.count_nonzero(y_pred_ * y_true_)
FP = tf.math.count_nonzero(y_pred_ * (y_true_ - 1))
FN = tf.math.count_nonzero((y_pred_ - 1) * y_true_)
jac = tf.cond(tf.greater((TP + FP + FN), 0), lambda: TP / (TP + FP + FN),
lambda: tf.cast(0.000, dtype='float64'))
return jac
def jaccard_cce(y_true, y_pred, skip_background=True):
"""Define Jaccard index for multiple labels.
Args:
y_true (tensor): ground truth masks.
y_pred (tensor): predicted masks.
skip_background (bool, optional): skip 0-label from calculation
Return:
jac (tensor): Jaccard index value
"""
# We read the number of classes from the last dimension of the true labels
num_classes = tf.shape(y_true)[-1]
# one_hot representation of predicted segmentation after argmax
y_pred_ = tf.cast(y_pred, dtype=tf.float32)
y_pred_ = tf.one_hot(tf.math.argmax(y_pred_, axis=-1), num_classes, axis=-1)
# y_true is already one-hot encoded
y_true_ = tf.cast(y_true, dtype=tf.float32)
# skip background pixels from the Jaccard index calculation
if skip_background:
y_true_ = y_true_[...,1:]
y_pred_ = y_pred_[...,1:]
TP = tf.math.count_nonzero(y_pred_ * y_true_)
FP = tf.math.count_nonzero(y_pred_ * (y_true_ - 1))
FN = tf.math.count_nonzero((y_pred_ - 1) * y_true_)
jac = tf.cond(tf.greater((TP + FP + FN), 0), lambda: TP / (TP + FP + FN),
lambda: tf.cast(0.000, dtype='float64'))
return jac
## Code taken from DeepSTORM at ZeroCostDL4Mic. Please cite when using it
# Define a matlab like gaussian 2D filter
def matlab_style_gauss2D(shape=(7,7),sigma=1):
"""
2D gaussian filter - should give the same result as:
MATLAB's fspecial('gaussian',[shape],[sigma])
"""
m,n = [(ss-1.)/2. for ss in shape]
y,x = np.ogrid[-m:m+1,-n:n+1]
h = np.exp( -(x*x + y*y) / (2.*sigma*sigma) )
h.astype(dtype=K.floatx())
h[ h < np.finfo(h.dtype).eps*h.max() ] = 0
sumh = h.sum()
if sumh != 0:
h /= sumh
h = h*2.0
h = h.astype('float32')
return h
# Expand the filter dimensions
## We changed the kernel size from 7 to 10.
# psf_heatmap = matlab_style_gauss2D(shape=(14, 14), sigma=2)
# gfilter = tf.reshape(psf_heatmap, [14, 14, 1, 1])
# Combined MSE + L1 loss
def L1L2loss(input_shape, gfilter, strides=(1, 1)):
"""
Args:
input_shape: (512,512,1)
Returns:
"""
def bump_mse(heatmap_true, spikes_pred):
# generate the heatmap corresponding to the predicted spikes
if len(strides) == 2:
heatmap_pred = K.conv2d(spikes_pred, gfilter, strides=strides, padding='same')
elif len(strides) == 3:
heatmap_pred = K.conv3d(spikes_pred, gfilter, strides=strides, padding='same')
# heatmaps MSE
loss_heatmaps = losses.mean_squared_error(heatmap_true,heatmap_pred)
# l1 on the predicted spikes
loss_spikes = losses.mean_absolute_error(spikes_pred,tf.zeros(input_shape))
return loss_heatmaps + loss_spikes
return bump_mse
|
[
"tensorflow.keras.backend.log",
"tensorflow.shape",
"tensorflow.keras.backend.floatx",
"tensorflow.keras.backend.greater",
"tensorflow.keras.backend.ones_like",
"tensorflow.keras.losses.binary_crossentropy",
"tensorflow.cast",
"tensorflow.keras.backend.conv2d",
"tensorflow.keras.backend.conv3d",
"tensorflow.keras.backend.maximum",
"numpy.exp",
"tensorflow.math.count_nonzero",
"tensorflow.keras.backend.cast",
"tensorflow.greater",
"tensorflow.zeros",
"tensorflow.one_hot",
"tensorflow.math.argmax",
"tensorflow.sigmoid",
"tensorflow.keras.backend.less",
"tensorflow.keras.losses.mean_squared_error",
"tensorflow.expand_dims",
"numpy.finfo",
"tensorflow.keras.backend.sum",
"tensorflow.keras.backend.pool2d",
"tensorflow.keras.backend.clip",
"tensorflow.keras.backend.abs"
] |
[((1129, 1167), 'tensorflow.keras.backend.clip', 'K.clip', (['y_pred', 'epsilon', '(1.0 - epsilon)'], {}), '(y_pred, epsilon, 1.0 - epsilon)\n', (1135, 1167), True, 'from tensorflow.keras import backend as K\n'), ((1186, 1216), 'tensorflow.keras.backend.log', 'K.log', (['(y_pred / (1.0 - y_pred))'], {}), '(y_pred / (1.0 - y_pred))\n', (1191, 1216), True, 'from tensorflow.keras import backend as K\n'), ((1843, 1868), 'tensorflow.keras.backend.cast', 'K.cast', (['y_true', '"""float32"""'], {}), "(y_true, 'float32')\n", (1849, 1868), True, 'from tensorflow.keras import backend as K\n'), ((1882, 1907), 'tensorflow.keras.backend.cast', 'K.cast', (['y_pred', '"""float32"""'], {}), "(y_pred, 'float32')\n", (1888, 1907), True, 'from tensorflow.keras import backend as K\n'), ((2004, 2093), 'tensorflow.keras.backend.pool2d', 'K.pool2d', (['y_true'], {'pool_size': '(11, 11)', 'strides': '(1, 1)', 'padding': '"""same"""', 'pool_mode': '"""avg"""'}), "(y_true, pool_size=(11, 11), strides=(1, 1), padding='same',\n pool_mode='avg')\n", (2012, 2093), True, 'from tensorflow.keras import backend as K\n'), ((2226, 2252), 'tensorflow.keras.backend.ones_like', 'K.ones_like', (['averaged_mask'], {}), '(averaged_mask)\n', (2237, 2252), True, 'from tensorflow.keras import backend as K\n'), ((2262, 2275), 'tensorflow.keras.backend.sum', 'K.sum', (['weight'], {}), '(weight)\n', (2267, 2275), True, 'from tensorflow.keras import backend as K\n'), ((2310, 2323), 'tensorflow.keras.backend.sum', 'K.sum', (['weight'], {}), '(weight)\n', (2315, 2323), True, 'from tensorflow.keras import backend as K\n'), ((2538, 2573), 'tensorflow.keras.losses.binary_crossentropy', 'binary_crossentropy', (['y_true', 'y_pred'], {}), '(y_true, y_pred)\n', (2557, 2573), False, 'from tensorflow.keras.losses import binary_crossentropy\n'), ((2585, 2608), 'tensorflow.expand_dims', 'tf.expand_dims', (['loss', '(3)'], {}), '(loss, 3)\n', (2599, 2608), True, 'import tensorflow as tf\n'), ((3206, 3237), 'tensorflow.cast', 'tf.cast', (['y_pred_'], {'dtype': 'tf.int8'}), '(y_pred_, dtype=tf.int8)\n', (3213, 3237), True, 'import tensorflow as tf\n'), ((3252, 3282), 'tensorflow.cast', 'tf.cast', (['y_true'], {'dtype': 'tf.int8'}), '(y_true, dtype=tf.int8)\n', (3259, 3282), True, 'import tensorflow as tf\n'), ((3293, 3333), 'tensorflow.math.count_nonzero', 'tf.math.count_nonzero', (['(y_pred_ * y_true_)'], {}), '(y_pred_ * y_true_)\n', (3314, 3333), True, 'import tensorflow as tf\n'), ((3343, 3389), 'tensorflow.math.count_nonzero', 'tf.math.count_nonzero', (['(y_pred_ * (1 - y_true_))'], {}), '(y_pred_ * (1 - y_true_))\n', (3364, 3389), True, 'import tensorflow as tf\n'), ((3399, 3445), 'tensorflow.math.count_nonzero', 'tf.math.count_nonzero', (['((1 - y_pred_) * y_true_)'], {}), '((1 - y_pred_) * y_true_)\n', (3420, 3445), True, 'import tensorflow as tf\n'), ((4117, 4148), 'tensorflow.cast', 'tf.cast', (['y_pred'], {'dtype': 'tf.int32'}), '(y_pred, dtype=tf.int32)\n', (4124, 4148), True, 'import tensorflow as tf\n'), ((4312, 4351), 'tensorflow.cast', 'tf.cast', (['y_true[..., 0]'], {'dtype': 'tf.int32'}), '(y_true[..., 0], dtype=tf.int32)\n', (4319, 4351), True, 'import tensorflow as tf\n'), ((4365, 4406), 'tensorflow.one_hot', 'tf.one_hot', (['y_true_', 'num_classes'], {'axis': '(-1)'}), '(y_true_, num_classes, axis=-1)\n', (4375, 4406), True, 'import tensorflow as tf\n'), ((4514, 4554), 'tensorflow.math.count_nonzero', 'tf.math.count_nonzero', (['(y_pred_ * y_true_)'], {}), '(y_pred_ * y_true_)\n', (4535, 4554), True, 'import tensorflow as tf\n'), ((4564, 4610), 'tensorflow.math.count_nonzero', 'tf.math.count_nonzero', (['(y_pred_ * (y_true_ - 1))'], {}), '(y_pred_ * (y_true_ - 1))\n', (4585, 4610), True, 'import tensorflow as tf\n'), ((4620, 4666), 'tensorflow.math.count_nonzero', 'tf.math.count_nonzero', (['((y_pred_ - 1) * y_true_)'], {}), '((y_pred_ - 1) * y_true_)\n', (4641, 4666), True, 'import tensorflow as tf\n'), ((5380, 5413), 'tensorflow.cast', 'tf.cast', (['y_pred'], {'dtype': 'tf.float32'}), '(y_pred, dtype=tf.float32)\n', (5387, 5413), True, 'import tensorflow as tf\n'), ((5554, 5587), 'tensorflow.cast', 'tf.cast', (['y_true'], {'dtype': 'tf.float32'}), '(y_true, dtype=tf.float32)\n', (5561, 5587), True, 'import tensorflow as tf\n'), ((5750, 5790), 'tensorflow.math.count_nonzero', 'tf.math.count_nonzero', (['(y_pred_ * y_true_)'], {}), '(y_pred_ * y_true_)\n', (5771, 5790), True, 'import tensorflow as tf\n'), ((5800, 5846), 'tensorflow.math.count_nonzero', 'tf.math.count_nonzero', (['(y_pred_ * (y_true_ - 1))'], {}), '(y_pred_ * (y_true_ - 1))\n', (5821, 5846), True, 'import tensorflow as tf\n'), ((5856, 5902), 'tensorflow.math.count_nonzero', 'tf.math.count_nonzero', (['((y_pred_ - 1) * y_true_)'], {}), '((y_pred_ - 1) * y_true_)\n', (5877, 5902), True, 'import tensorflow as tf\n'), ((6426, 6474), 'numpy.exp', 'np.exp', (['(-(x * x + y * y) / (2.0 * sigma * sigma))'], {}), '(-(x * x + y * y) / (2.0 * sigma * sigma))\n', (6432, 6474), True, 'import numpy as np\n'), ((1478, 1489), 'tensorflow.keras.backend.sum', 'K.sum', (['loss'], {}), '(loss)\n', (1483, 1489), True, 'from tensorflow.keras import backend as K\n'), ((1492, 1505), 'tensorflow.keras.backend.sum', 'K.sum', (['weight'], {}), '(weight)\n', (1497, 1505), True, 'from tensorflow.keras import backend as K\n'), ((1755, 1767), 'tensorflow.keras.backend.sum', 'K.sum', (['score'], {}), '(score)\n', (1760, 1767), True, 'from tensorflow.keras import backend as K\n'), ((3145, 3163), 'tensorflow.sigmoid', 'tf.sigmoid', (['y_pred'], {}), '(y_pred)\n', (3155, 3163), True, 'import tensorflow as tf\n'), ((3465, 3492), 'tensorflow.greater', 'tf.greater', (['(TP + FP + FN)', '(0)'], {}), '(TP + FP + FN, 0)\n', (3475, 3492), True, 'import tensorflow as tf\n'), ((4022, 4038), 'tensorflow.shape', 'tf.shape', (['y_pred'], {}), '(y_pred)\n', (4030, 4038), True, 'import tensorflow as tf\n'), ((4174, 4206), 'tensorflow.math.argmax', 'tf.math.argmax', (['y_pred_'], {'axis': '(-1)'}), '(y_pred_, axis=-1)\n', (4188, 4206), True, 'import tensorflow as tf\n'), ((4686, 4713), 'tensorflow.greater', 'tf.greater', (['(TP + FP + FN)', '(0)'], {}), '(TP + FP + FN, 0)\n', (4696, 4713), True, 'import tensorflow as tf\n'), ((5277, 5293), 'tensorflow.shape', 'tf.shape', (['y_true'], {}), '(y_true)\n', (5285, 5293), True, 'import tensorflow as tf\n'), ((5439, 5471), 'tensorflow.math.argmax', 'tf.math.argmax', (['y_pred_'], {'axis': '(-1)'}), '(y_pred_, axis=-1)\n', (5453, 5471), True, 'import tensorflow as tf\n'), ((5922, 5949), 'tensorflow.greater', 'tf.greater', (['(TP + FP + FN)', '(0)'], {}), '(TP + FP + FN, 0)\n', (5932, 5949), True, 'import tensorflow as tf\n'), ((7406, 7459), 'tensorflow.keras.losses.mean_squared_error', 'losses.mean_squared_error', (['heatmap_true', 'heatmap_pred'], {}), '(heatmap_true, heatmap_pred)\n', (7431, 7459), False, 'from tensorflow.keras import losses\n'), ((2119, 2150), 'tensorflow.keras.backend.greater', 'K.greater', (['averaged_mask', '(0.005)'], {}), '(averaged_mask, 0.005)\n', (2128, 2150), True, 'from tensorflow.keras import backend as K\n'), ((2172, 2200), 'tensorflow.keras.backend.less', 'K.less', (['averaged_mask', '(0.995)'], {}), '(averaged_mask, 0.995)\n', (2178, 2200), True, 'from tensorflow.keras import backend as K\n'), ((3551, 3580), 'tensorflow.cast', 'tf.cast', (['(0.0)'], {'dtype': '"""float64"""'}), "(0.0, dtype='float64')\n", (3558, 3580), True, 'import tensorflow as tf\n'), ((4772, 4801), 'tensorflow.cast', 'tf.cast', (['(0.0)'], {'dtype': '"""float64"""'}), "(0.0, dtype='float64')\n", (4779, 4801), True, 'import tensorflow as tf\n'), ((6008, 6037), 'tensorflow.cast', 'tf.cast', (['(0.0)'], {'dtype': '"""float64"""'}), "(0.0, dtype='float64')\n", (6015, 6037), True, 'import tensorflow as tf\n'), ((6487, 6497), 'tensorflow.keras.backend.floatx', 'K.floatx', ([], {}), '()\n', (6495, 6497), True, 'from tensorflow.keras import backend as K\n'), ((7170, 7233), 'tensorflow.keras.backend.conv2d', 'K.conv2d', (['spikes_pred', 'gfilter'], {'strides': 'strides', 'padding': '"""same"""'}), "(spikes_pred, gfilter, strides=strides, padding='same')\n", (7178, 7233), True, 'from tensorflow.keras import backend as K\n'), ((7558, 7579), 'tensorflow.zeros', 'tf.zeros', (['input_shape'], {}), '(input_shape)\n', (7566, 7579), True, 'import tensorflow as tf\n'), ((1437, 1466), 'tensorflow.keras.backend.maximum', 'K.maximum', (['(-logit_y_pred)', '(0.0)'], {}), '(-logit_y_pred, 0.0)\n', (1446, 1466), True, 'from tensorflow.keras import backend as K\n'), ((1662, 1685), 'tensorflow.keras.backend.sum', 'K.sum', (['(w * intersection)'], {}), '(w * intersection)\n', (1667, 1685), True, 'from tensorflow.keras import backend as K\n'), ((1699, 1712), 'tensorflow.keras.backend.sum', 'K.sum', (['(w * m1)'], {}), '(w * m1)\n', (1704, 1712), True, 'from tensorflow.keras import backend as K\n'), ((1715, 1728), 'tensorflow.keras.backend.sum', 'K.sum', (['(w * m2)'], {}), '(w * m2)\n', (1720, 1728), True, 'from tensorflow.keras import backend as K\n'), ((7294, 7357), 'tensorflow.keras.backend.conv3d', 'K.conv3d', (['spikes_pred', 'gfilter'], {'strides': 'strides', 'padding': '"""same"""'}), "(spikes_pred, gfilter, strides=strides, padding='same')\n", (7302, 7357), True, 'from tensorflow.keras import backend as K\n'), ((6510, 6527), 'numpy.finfo', 'np.finfo', (['h.dtype'], {}), '(h.dtype)\n', (6518, 6527), True, 'import numpy as np\n'), ((1413, 1432), 'tensorflow.keras.backend.abs', 'K.abs', (['logit_y_pred'], {}), '(logit_y_pred)\n', (1418, 1432), True, 'from tensorflow.keras import backend as K\n')]
|
import json
import bz2
import gzip
import _pickle as cPickle
import gym
import numpy as np
import quaternion
import skimage.morphology
import habitat
from envs.utils.fmm_planner import FMMPlanner
from constants import coco_categories
import envs.utils.pose as pu
class ObjectGoal_Env(habitat.RLEnv):
"""The Object Goal Navigation environment class. The class is responsible
for loading the dataset, generating episodes, and computing evaluation
metrics.
"""
def __init__(self, args, rank, config_env, dataset):
self.args = args
self.rank = rank
super().__init__(config_env, dataset)
# Loading dataset info file
self.split = config_env.DATASET.SPLIT
self.episodes_dir = config_env.DATASET.EPISODES_DIR.format(
split=self.split)
if args.custom_eps:
with open("{}/train_episode_data.json".format(args.custom_eps), 'r') as f:
episodes_all = json.load(f)
self.episodes_all = {}
for ep in episodes_all:
if ep["scene"] in self.episodes_all:
self.episodes_all[ep["scene"]].append(ep)
else:
self.episodes_all[ep["scene"]] = [ep]
dataset_info_file = self.episodes_dir + \
"{split}_info.pbz2".format(split=self.split)
with bz2.BZ2File(dataset_info_file, 'rb') as f:
self.dataset_info = cPickle.load(f)
# Specifying action and observation space
self.action_space = gym.spaces.Discrete(3)
self.observation_space = gym.spaces.Box(0, 255,
(3, args.frame_height,
args.frame_width),
dtype='uint8')
# Initializations
self.episode_no = 0
# Scene info
self.last_scene_path = None
self.scene_path = None
self.scene_name = None
# Episode Dataset info
self.eps_data = None
self.eps_data_idx = None
self.gen_ep_idx = 1
self.gt_planner = None
self.object_boundary = None
self.goal_idx = None
self.goal_name = None
self.map_obj_origin = None
self.starting_loc = None
self.starting_distance = None
if args.eval and args.shuffle:
self.shuffled_indices = np.arange(args.num_eval_episodes)
np.random.shuffle(self.shuffled_indices)
# Episode tracking info
self.curr_distance = None
self.prev_distance = None
self.timestep = None
self.stopped = None
self.path_length = None
self.last_sim_location = None
self.trajectory_states = []
self.info = {}
self.info['distance_to_goal'] = None
self.info['spl'] = None
self.info['success'] = None
def load_new_episode(self):
"""The function loads a fixed episode from the episode dataset. This
function is used for evaluating a trained model on the val split.
"""
args = self.args
self.scene_path = self.habitat_env.sim.config.SCENE
scene_name = self.scene_path.split("/")[-1].split(".")[0]
if self.scene_path != self.last_scene_path:
if not args.testset:
episodes_file = self.episodes_dir + \
"content/{}_episodes.json.gz".format(scene_name)
print("Loading episodes from: {}".format(episodes_file))
with gzip.open(episodes_file, 'r') as f:
self.eps_data = json.loads(
f.read().decode('utf-8'))["episodes"]
else:
episodes_file = self.episodes_dir + \
"content/{}_test_episodes.json".format(scene_name)
print("Loading episodes from: {}".format(episodes_file))
with open(episodes_file, 'r') as f:
self.eps_data = json.load(f)
self.eps_data_idx = 0
self.last_scene_path = self.scene_path
# Load episode info
if self.args.shuffle:
episode = self.eps_data[self.shuffled_indices[self.eps_data_idx]]
else:
episode = self.eps_data[self.eps_data_idx]
self.info["episode_data"] = episode
self.eps_data_idx += 1
self.eps_data_idx = self.eps_data_idx % len(self.eps_data)
pos = episode["start_position"]
rot = quaternion.from_float_array(episode["start_rotation"])
goal_name = episode["object_category"]
goal_idx = episode["object_id"]
floor_idx = episode["floor_id"]
# Load scene info
scene_info = self.dataset_info[scene_name]
sem_map = scene_info[floor_idx]['sem_map']
map_obj_origin = scene_info[floor_idx]['origin']
# Setup ground truth planner
object_boundary = args.success_dist
map_resolution = args.map_resolution
selem = skimage.morphology.disk(2)
traversible = skimage.morphology.binary_dilation(
sem_map[0], selem) != True
traversible = 1 - traversible
planner = FMMPlanner(traversible)
selem = skimage.morphology.disk(
int(object_boundary * 100. / map_resolution))
goal_map = skimage.morphology.binary_dilation(
sem_map[goal_idx + 1], selem) != True
goal_map = 1 - goal_map
planner.set_multi_goal(goal_map)
# Get starting loc in GT map coordinates
x = -pos[2]
y = -pos[0]
min_x, min_y = map_obj_origin / 100.0
map_loc = int((-y - min_y) * 20.), int((-x - min_x) * 20.)
self.gt_planner = planner
self.starting_loc = map_loc
self.object_boundary = object_boundary
self.goal_idx = goal_idx
self.goal_name = goal_name
self.map_obj_origin = map_obj_origin
self.starting_distance = self.gt_planner.fmm_dist[self.starting_loc]\
/ 20.0 + self.object_boundary
self.info["episode_data"]["shortest_dist"] = self.starting_distance
self.prev_distance = self.starting_distance
self._env.sim.set_agent_state(pos, rot)
self.info["sim_pos"] = pos
self.info["sim_rot"] = rot
self.info["scene"] = scene_name
self.info["floor_idx"] = floor_idx
# The following two should match approximately
#print(self.starting_loc)
#print(self.sim_continuous_to_sim_map(self.get_sim_location()))
self.info['gt_pos'] = self.sim_continuous_to_sim_map(self.get_sim_location())
obs = self._env.sim.get_observations_at(pos, rot)
return obs
def load_incomplete_episode(self):
args = self.args
self.scene_path = self.habitat_env.sim.config.SCENE
scene_name = self.scene_path.split("/")[-1].split(".")[0]
if self.scene_path != self.last_scene_path:
print("Loading episodes from: {}".format(scene_name))
self.eps_data_idx = 0
self.last_scene_path = self.scene_path
episode = self.episodes_all[scene_name][self.eps_data_idx]
self.info["episode_data"] = episode
self.eps_data_idx += 1
self.eps_data_idx = self.eps_data_idx % len(self.episodes_all[scene_name])
pos = episode["sim_pos"]
rot = quaternion.from_rotation_vector(episode["sim_rot"])
goal_name = episode["goal_name"]
goal_idx = episode["goal_cat_id"]
floor_idx = episode["floor_idx"]
# Load scene info
scene_info = self.dataset_info[scene_name]
sem_map = scene_info[floor_idx]['sem_map']
map_obj_origin = scene_info[floor_idx]['origin']
# Setup ground truth planner
object_boundary = args.success_dist
map_resolution = args.map_resolution
selem = skimage.morphology.disk(2)
traversible = skimage.morphology.binary_dilation(
sem_map[0], selem) != True
traversible = 1 - traversible
planner = FMMPlanner(traversible)
selem = skimage.morphology.disk(
int(object_boundary * 100. / map_resolution))
goal_map = skimage.morphology.binary_dilation(
sem_map[goal_idx + 1], selem) != True
goal_map = 1 - goal_map
planner.set_multi_goal(goal_map)
# Get starting loc in GT map coordinates
x = -pos[2]
y = -pos[0]
min_x, min_y = map_obj_origin / 100.0
map_loc = int((-y - min_y) * 20.), int((-x - min_x) * 20.)
self.gt_planner = planner
self.starting_loc = map_loc
self.object_boundary = object_boundary
self.goal_idx = goal_idx
self.goal_name = goal_name
self.map_obj_origin = map_obj_origin
self.starting_distance = self.gt_planner.fmm_dist[self.starting_loc]\
/ 20.0 + self.object_boundary
self.info["episode_data"]["shortest_dist"] = self.starting_distance
self.prev_distance = self.starting_distance
self._env.sim.set_agent_state(pos, rot)
self.info["sim_pos"] = pos
self.info["sim_rot"] = rot
# The following two should match approximately
#print(self.starting_loc)
#print(self.sim_continuous_to_sim_map(self.get_sim_location()))
self.info['gt_pos'] = self.sim_continuous_to_sim_map(self.get_sim_location())
obs = self._env.sim.get_observations_at(pos, rot)
return obs
def generate_new_episode(self):
"""The function generates a random valid episode. This function is used
for training a model on the train split.
"""
args = self.args
self.scene_path = self.habitat_env.sim.config.SCENE
scene_name = self.scene_path.split("/")[-1].split(".")[0]
scene_info = self.dataset_info[scene_name]
map_resolution = args.map_resolution
floor_idx = np.random.randint(len(scene_info.keys()))
floor_height = scene_info[floor_idx]['floor_height']
sem_map = scene_info[floor_idx]['sem_map']
map_obj_origin = scene_info[floor_idx]['origin']
cat_counts = sem_map.sum(2).sum(1)
possible_cats = list(np.arange(6))
for i in range(6):
if cat_counts[i + 1] == 0:
possible_cats.remove(i)
object_boundary = args.success_dist
loc_found = False
while not loc_found:
if len(possible_cats) == 0:
print("No valid objects for {}".format(floor_height))
eps = eps - 1
continue
goal_idx = np.random.choice(possible_cats)
for key, value in coco_categories.items():
if value == goal_idx:
goal_name = key
selem = skimage.morphology.disk(2)
traversible = skimage.morphology.binary_dilation(
sem_map[0], selem) != True
traversible = 1 - traversible
planner = FMMPlanner(traversible)
selem = skimage.morphology.disk(
int(object_boundary * 100. / map_resolution))
goal_map = skimage.morphology.binary_dilation(
sem_map[goal_idx + 1], selem) != True
goal_map = 1 - goal_map
planner.set_multi_goal(goal_map)
m1 = sem_map[0] > 0
m2 = planner.fmm_dist > (args.min_d - object_boundary) * 20.0
m3 = planner.fmm_dist < (args.max_d - object_boundary) * 20.0
possible_starting_locs = np.logical_and(m1, m2)
possible_starting_locs = np.logical_and(
possible_starting_locs, m3) * 1.
if possible_starting_locs.sum() != 0:
loc_found = True
else:
print("Invalid object: {} / {} / {}".format(
scene_name, floor_height, goal_name))
possible_cats.remove(goal_idx)
scene_info[floor_idx]["sem_map"][goal_idx + 1, :, :] = 0.
self.dataset_info[scene_name][floor_idx][
"sem_map"][goal_idx + 1, :, :] = 0.
loc_found = False
while not loc_found:
pos = self._env.sim.sample_navigable_point()
x = -pos[2]
y = -pos[0]
min_x, min_y = map_obj_origin / 100.0
map_loc = int((-y - min_y) * 20.), int((-x - min_x) * 20.)
if abs(pos[1] - floor_height) < args.floor_thr / 100.0 and \
possible_starting_locs[map_loc[0], map_loc[1]] == 1:
loc_found = True
agent_state = self._env.sim.get_agent_state(0)
rotation = agent_state.rotation
rvec = quaternion.as_rotation_vector(rotation)
rvec[1] = np.random.rand() * 2 * np.pi
rot = quaternion.from_rotation_vector(rvec)
self.gt_planner = planner
self.starting_loc = map_loc
self.object_boundary = object_boundary
self.goal_idx = goal_idx
self.goal_name = goal_name
self.map_obj_origin = map_obj_origin
self.starting_distance = self.gt_planner.fmm_dist[self.starting_loc] \
/ 20.0 + self.object_boundary
self.prev_distance = self.starting_distance
self._env.sim.set_agent_state(pos, rot)
self.info["sim_pos"] = pos
self.info["sim_rot"] = quaternion.as_float_array(rot)
self.info["episode_id"] = self.gen_ep_idx
self.gen_ep_idx += 1
self.info["scene"] = scene_name
self.info["floor_idx"] = floor_idx
self.info["goal_name"] = goal_name
# The following two should match approximately
# print(starting_loc)
# print(self.sim_continuous_to_sim_map(self.get_sim_location()))
self.info['gt_pos'] = self.sim_continuous_to_sim_map(self.get_sim_location())
obs = self._env.sim.get_observations_at(pos, rot)
return obs
def sim_map_to_sim_continuous(self, coords):
"""Converts ground-truth 2D Map coordinates to absolute Habitat
simulator position and rotation.
"""
agent_state = self._env.sim.get_agent_state(0)
y, x = coords
min_x, min_y = self.map_obj_origin / 100.0
cont_x = x / 20. + min_x
cont_y = y / 20. + min_y
agent_state.position[0] = cont_y
agent_state.position[2] = cont_x
rotation = agent_state.rotation
rvec = quaternion.as_rotation_vector(rotation)
if self.args.train_single_eps:
rvec[1] = 0.0
else:
rvec[1] = np.random.rand() * 2 * np.pi
rot = quaternion.from_rotation_vector(rvec)
return agent_state.position, rot
def sim_continuous_to_sim_map(self, sim_loc):
"""Converts absolute Habitat simulator pose to ground-truth 2D Map
coordinates.
"""
x, y, o = sim_loc
min_x, min_y = self.map_obj_origin / 100.0
x, y = int((-x - min_x) * 20.), int((-y - min_y) * 20.)
o = np.rad2deg(o) + 180.0
return y, x, o
def reset(self):
"""Resets the environment to a new episode.
Returns:
obs (ndarray): RGBD observations (4 x H x W)
info (dict): contains timestep, pose, goal category and
evaluation metric info
"""
args = self.args
new_scene = self.episode_no % args.num_train_episodes == 0
self.episode_no += 1
# Initializations
self.timestep = 0
self.stopped = False
self.path_length = 1e-5
self.trajectory_states = []
if new_scene:
obs = super().reset()
self.scene_name = self.habitat_env.sim.config.SCENE
print("Changing scene: {}/{}".format(self.rank, self.scene_name))
self.scene_path = self.habitat_env.sim.config.SCENE
if args.gen_episode:
obs = self.generate_new_episode()
elif args.custom_eps:
obs = self.load_incomplete_episode()
elif self.split == "val":
obs = self.load_new_episode()
else:
obs = self.generate_new_episode()
rgb = obs['rgb'].astype(np.uint8)
depth = obs['depth']
state = np.concatenate((rgb, depth), axis=2).transpose(2, 0, 1)
self.last_sim_location = self.get_sim_location()
# Set info
self.info['time'] = self.timestep
self.info['sensor_pose'] = [0., 0., 0.]
self.info['goal_cat_id'] = self.goal_idx
self.info['goal_name'] = self.goal_name
return state, self.info
def step(self, action):
"""Function to take an action in the environment.
Args:
action (dict):
dict with following keys:
'action' (int): 0: stop, 1: forward, 2: left, 3: right
Returns:
obs (ndarray): RGBD observations (4 x H x W)
reward (float): amount of reward returned after previous action
done (bool): whether the episode has ended
info (dict): contains timestep, pose, goal category and
evaluation metric info
"""
action = action["action"]
if action == 0:
self.stopped = True
# Not sending stop to simulator, resetting manually
action = 3
obs, rew, done, _ = super().step(action)
# Get pose change
dx, dy, do = self.get_pose_change()
self.info['sensor_pose'] = [dx, dy, do]
self.path_length += pu.get_l2_distance(0, dx, 0, dy)
spl, success, dist = 0., 0., 0.
if done:
spl, success, dist = self.get_metrics()
self.info['distance_to_goal'] = dist
self.info['spl'] = spl
self.info['success'] = success
rgb = obs['rgb'].astype(np.uint8)
depth = obs['depth']
state = np.concatenate((rgb, depth), axis=2).transpose(2, 0, 1)
self.timestep += 1
self.info['time'] = self.timestep
return state, rew, done, self.info
def get_reward_range(self):
"""This function is not used, Habitat-RLEnv requires this function"""
return (0., 1.0)
def get_reward(self, observations):
curr_loc = self.sim_continuous_to_sim_map(self.get_sim_location())
self.curr_distance = self.gt_planner.fmm_dist[curr_loc[0],
curr_loc[1]] / 20.0
reward = (self.prev_distance - self.curr_distance) * \
self.args.reward_coeff
self.prev_distance = self.curr_distance
return reward
def get_metrics(self):
"""This function computes evaluation metrics for the Object Goal task
Returns:
spl (float): Success weighted by Path Length
(See https://arxiv.org/pdf/1807.06757.pdf)
success (int): 0: Failure, 1: Successful
dist (float): Distance to Success (DTS), distance of the agent
from the success threshold boundary in meters.
(See https://arxiv.org/pdf/2007.00643.pdf)
"""
curr_loc = self.sim_continuous_to_sim_map(self.get_sim_location())
dist = self.gt_planner.fmm_dist[curr_loc[0], curr_loc[1]] / 20.0
if dist == 0.0:
success = 1
else:
success = 0
spl = min(success * self.starting_distance / self.path_length, 1)
return spl, success, dist
def get_done(self, observations):
if self.info['time'] >= self.args.max_episode_length - 1:
done = True
elif self.stopped:
done = True
else:
done = False
return done
def get_info(self, observations):
"""This function is not used, Habitat-RLEnv requires this function"""
info = {}
return info
def get_spaces(self):
"""Returns observation and action spaces for the ObjectGoal task."""
return self.observation_space, self.action_space
def get_sim_location(self):
"""Returns x, y, o pose of the agent in the Habitat simulator."""
agent_state = super().habitat_env.sim.get_agent_state(0)
x = -agent_state.position[2]
y = -agent_state.position[0]
axis = quaternion.as_euler_angles(agent_state.rotation)[0]
if (axis % (2 * np.pi)) < 0.1 or (axis %
(2 * np.pi)) > 2 * np.pi - 0.1:
o = quaternion.as_euler_angles(agent_state.rotation)[1]
else:
o = 2 * np.pi - quaternion.as_euler_angles(agent_state.rotation)[1]
if o > np.pi:
o -= 2 * np.pi
return x, y, o
def get_pose_change(self):
"""Returns dx, dy, do pose change of the agent relative to the last
timestep."""
curr_sim_pose = self.get_sim_location()
dx, dy, do = pu.get_rel_pose_change(
curr_sim_pose, self.last_sim_location)
self.last_sim_location = curr_sim_pose
return dx, dy, do
|
[
"quaternion.as_rotation_vector",
"quaternion.as_float_array",
"envs.utils.fmm_planner.FMMPlanner",
"numpy.random.rand",
"gzip.open",
"envs.utils.pose.get_l2_distance",
"numpy.arange",
"_pickle.load",
"bz2.BZ2File",
"numpy.concatenate",
"constants.coco_categories.items",
"numpy.rad2deg",
"quaternion.from_float_array",
"numpy.random.choice",
"gym.spaces.Discrete",
"quaternion.from_rotation_vector",
"numpy.logical_and",
"quaternion.as_euler_angles",
"gym.spaces.Box",
"json.load",
"envs.utils.pose.get_rel_pose_change",
"numpy.random.shuffle"
] |
[((1554, 1576), 'gym.spaces.Discrete', 'gym.spaces.Discrete', (['(3)'], {}), '(3)\n', (1573, 1576), False, 'import gym\n'), ((1611, 1690), 'gym.spaces.Box', 'gym.spaces.Box', (['(0)', '(255)', '(3, args.frame_height, args.frame_width)'], {'dtype': '"""uint8"""'}), "(0, 255, (3, args.frame_height, args.frame_width), dtype='uint8')\n", (1625, 1690), False, 'import gym\n'), ((4531, 4585), 'quaternion.from_float_array', 'quaternion.from_float_array', (["episode['start_rotation']"], {}), "(episode['start_rotation'])\n", (4558, 4585), False, 'import quaternion\n'), ((5223, 5246), 'envs.utils.fmm_planner.FMMPlanner', 'FMMPlanner', (['traversible'], {}), '(traversible)\n', (5233, 5246), False, 'from envs.utils.fmm_planner import FMMPlanner\n'), ((7404, 7455), 'quaternion.from_rotation_vector', 'quaternion.from_rotation_vector', (["episode['sim_rot']"], {}), "(episode['sim_rot'])\n", (7435, 7455), False, 'import quaternion\n'), ((8090, 8113), 'envs.utils.fmm_planner.FMMPlanner', 'FMMPlanner', (['traversible'], {}), '(traversible)\n', (8100, 8113), False, 'from envs.utils.fmm_planner import FMMPlanner\n'), ((12741, 12780), 'quaternion.as_rotation_vector', 'quaternion.as_rotation_vector', (['rotation'], {}), '(rotation)\n', (12770, 12780), False, 'import quaternion\n'), ((12842, 12879), 'quaternion.from_rotation_vector', 'quaternion.from_rotation_vector', (['rvec'], {}), '(rvec)\n', (12873, 12879), False, 'import quaternion\n'), ((13400, 13430), 'quaternion.as_float_array', 'quaternion.as_float_array', (['rot'], {}), '(rot)\n', (13425, 13430), False, 'import quaternion\n'), ((14469, 14508), 'quaternion.as_rotation_vector', 'quaternion.as_rotation_vector', (['rotation'], {}), '(rotation)\n', (14498, 14508), False, 'import quaternion\n'), ((14654, 14691), 'quaternion.from_rotation_vector', 'quaternion.from_rotation_vector', (['rvec'], {}), '(rvec)\n', (14685, 14691), False, 'import quaternion\n'), ((17587, 17619), 'envs.utils.pose.get_l2_distance', 'pu.get_l2_distance', (['(0)', 'dx', '(0)', 'dy'], {}), '(0, dx, 0, dy)\n', (17605, 17619), True, 'import envs.utils.pose as pu\n'), ((20971, 21032), 'envs.utils.pose.get_rel_pose_change', 'pu.get_rel_pose_change', (['curr_sim_pose', 'self.last_sim_location'], {}), '(curr_sim_pose, self.last_sim_location)\n', (20993, 21032), True, 'import envs.utils.pose as pu\n'), ((1384, 1420), 'bz2.BZ2File', 'bz2.BZ2File', (['dataset_info_file', '"""rb"""'], {}), "(dataset_info_file, 'rb')\n", (1395, 1420), False, 'import bz2\n'), ((1459, 1474), '_pickle.load', 'cPickle.load', (['f'], {}), '(f)\n', (1471, 1474), True, 'import _pickle as cPickle\n'), ((2440, 2473), 'numpy.arange', 'np.arange', (['args.num_eval_episodes'], {}), '(args.num_eval_episodes)\n', (2449, 2473), True, 'import numpy as np\n'), ((2486, 2526), 'numpy.random.shuffle', 'np.random.shuffle', (['self.shuffled_indices'], {}), '(self.shuffled_indices)\n', (2503, 2526), True, 'import numpy as np\n'), ((10252, 10264), 'numpy.arange', 'np.arange', (['(6)'], {}), '(6)\n', (10261, 10264), True, 'import numpy as np\n'), ((10663, 10694), 'numpy.random.choice', 'np.random.choice', (['possible_cats'], {}), '(possible_cats)\n', (10679, 10694), True, 'import numpy as np\n'), ((10726, 10749), 'constants.coco_categories.items', 'coco_categories.items', ([], {}), '()\n', (10747, 10749), False, 'from constants import coco_categories\n'), ((11043, 11066), 'envs.utils.fmm_planner.FMMPlanner', 'FMMPlanner', (['traversible'], {}), '(traversible)\n', (11053, 11066), False, 'from envs.utils.fmm_planner import FMMPlanner\n'), ((11589, 11611), 'numpy.logical_and', 'np.logical_and', (['m1', 'm2'], {}), '(m1, m2)\n', (11603, 11611), True, 'import numpy as np\n'), ((15047, 15060), 'numpy.rad2deg', 'np.rad2deg', (['o'], {}), '(o)\n', (15057, 15060), True, 'import numpy as np\n'), ((20364, 20412), 'quaternion.as_euler_angles', 'quaternion.as_euler_angles', (['agent_state.rotation'], {}), '(agent_state.rotation)\n', (20390, 20412), False, 'import quaternion\n'), ((960, 972), 'json.load', 'json.load', (['f'], {}), '(f)\n', (969, 972), False, 'import json\n'), ((11649, 11691), 'numpy.logical_and', 'np.logical_and', (['possible_starting_locs', 'm3'], {}), '(possible_starting_locs, m3)\n', (11663, 11691), True, 'import numpy as np\n'), ((12799, 12815), 'numpy.random.rand', 'np.random.rand', ([], {}), '()\n', (12813, 12815), True, 'import numpy as np\n'), ((16280, 16316), 'numpy.concatenate', 'np.concatenate', (['(rgb, depth)'], {'axis': '(2)'}), '((rgb, depth), axis=2)\n', (16294, 16316), True, 'import numpy as np\n'), ((17945, 17981), 'numpy.concatenate', 'np.concatenate', (['(rgb, depth)'], {'axis': '(2)'}), '((rgb, depth), axis=2)\n', (17959, 17981), True, 'import numpy as np\n'), ((20555, 20603), 'quaternion.as_euler_angles', 'quaternion.as_euler_angles', (['agent_state.rotation'], {}), '(agent_state.rotation)\n', (20581, 20603), False, 'import quaternion\n'), ((3579, 3608), 'gzip.open', 'gzip.open', (['episodes_file', '"""r"""'], {}), "(episodes_file, 'r')\n", (3588, 3608), False, 'import gzip\n'), ((4030, 4042), 'json.load', 'json.load', (['f'], {}), '(f)\n', (4039, 4042), False, 'import json\n'), ((14611, 14627), 'numpy.random.rand', 'np.random.rand', ([], {}), '()\n', (14625, 14627), True, 'import numpy as np\n'), ((20649, 20697), 'quaternion.as_euler_angles', 'quaternion.as_euler_angles', (['agent_state.rotation'], {}), '(agent_state.rotation)\n', (20675, 20697), False, 'import quaternion\n')]
|
import math
import re
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
from go_utils.cleanup import ( # isort: skip
rename_latlon_cols,
replace_column_prefix,
round_cols,
standardize_null_vals,
)
from go_utils.plot import ( # isort: skip
completeness_histogram,
plot_freq_bar,
plot_int_distribution,
)
__doc__ = r"""
## Mosquito Specific Cleanup Procedures
### Converting Larvae Data to Integers
Larvae Data is stored as a string in the raw GLOBE Observer dataset. To facillitate analysis, [this method](#larvae_to_num) converts this data to numerical data.
It needs to account for 4 types of data:
1. Regular Data: Converts it to a number
2. Extraneously large data ($\geq 100$ as its hard to count more than that amount accurately): To maintain the information from that entry, the `LarvaeCountMagnitude` flag is used to indicate the real value
3. Ranges (e.g. "25-50"): Chooses the lower bound and set the `LarvaeCountIsRangeFlag` to true.
4. Null Values: Sets null values to $-9999$
It generates the following flags:
- `LarvaeCountMagnitude`: The integer flag contains the order of magnitude (0-4) by which the larvae count exceeds the maximum Larvae Count of 100. This is calculated by $1 + \lfloor \log{\frac{num}{100}} \rfloor$. As a result:
- `0`: Corresponds to a Larvae Count $\leq 100$
- `1`: Corresponds to a Larvae Count between $100$ and $999$
- `2`: Corresponds to a Larvae Count between $1000$ and $9999$
- `3`: Corresponds to a Larvae Count between $10,000$ and $99,999$
- `4`: Corresponds to a Larvae Count $\geq 100,000$
- `LarvaeCountIsRange`: Either a $1$ which indicates the entry was a range (e.g. 25-50) or $0$ which indicates the entry wasn't a range.
Additionally, there were extremely large values that Python was unable to process (`1e+27`) and so there was an initial preprocessing step to set those numbers to 100000 (which corresponds to the maximum magnitude flag).
"""
def cleanup_column_prefix(df, inplace=False):
"""Method for shortening raw mosquito habitat mapper column names.
Parameters
----------
df : pd.DataFrame
The DataFrame containing raw mosquito habitat mapper data.
inplace : bool, default=False
Whether to return a new DataFrame. If True then no DataFrame copy is not returned and the operation is performed in place.
Returns
-------
pd.DataFrame or None
A DataFrame with the cleaned up column prefixes. If `inplace=True` it returns None.
"""
return replace_column_prefix(df, "mosquitohabitatmapper", "mhm", inplace=inplace)
def _entry_to_num(entry):
try:
if entry == "more than 100":
return 101, 1, 1
if pd.isna(entry):
return -9999, 0, 0
elif float(entry) > 100:
return 101, min(math.floor(math.log10(float(entry) / 100)) + 1, 4), 0
return float(entry), 0, 0
except ValueError:
return float(re.sub(r"-.*", "", entry)), 0, 1
def larvae_to_num(
mhm_df,
larvae_count_col="mhm_LarvaeCount",
magnitude="mhm_LarvaeCountMagnitude",
range_flag="mhm_LarvaeCountIsRangeFlag",
inplace=False,
):
"""Converts the Larvae Count of the Mosquito Habitat Mapper Dataset from being stored as a string to integers.
See [here](#converting-larvae-data-to-integers) for more information.
Parameters
----------
mhm_df : pd.DataFrame
A DataFrame of Mosquito Habitat Mapper data that needs the larvae counts to be set to numbers
larvae_count_col : str, default="mhm_LarvaeCount"
The name of the column storing the larvae count. **Note**: The columns will be output in the format: `prefix_ColumnName` where `prefix` is all the characters that preceed the words `LarvaeCount` in the specified name.
magnitude: str, default="mhm_LarvaeCountMagnitude"
The name of the column which will store the generated LarvaeCountMagnitude output
range_flag : str, default="mhm_LarvaeCountIsRangeFlag"
The name of the column which will store the generated LarvaeCountIsRange flag
inplace : bool, default=False
Whether to return a new DataFrame. If True then no DataFrame copy is not returned and the operation is performed in place.
Returns
-------
pd.DataFrame
A DataFrame with the larvae count as integers. If `inplace=True` it returns None.
"""
if not inplace:
mhm_df = mhm_df.copy()
# Preprocessing step to remove extremely erroneous values
for i in mhm_df.index:
count = mhm_df[larvae_count_col][i]
if not pd.isna(count) and type(count) is str and "e+" in count:
mhm_df.at[i, larvae_count_col] = "100000"
larvae_conversion = np.vectorize(_entry_to_num)
(
mhm_df[larvae_count_col],
mhm_df[magnitude],
mhm_df[range_flag],
) = larvae_conversion(mhm_df[larvae_count_col].to_numpy())
if not inplace:
return mhm_df
def has_genus_flag(df, genus_col="mhm_Genus", bit_col="mhm_HasGenus", inplace=False):
"""
Creates a bit flag: `mhm_HasGenus` where 1 denotes a recorded Genus and 0 denotes the contrary.
Parameters
----------
df : pd.DataFrame
A mosquito habitat mapper DataFrame
genus_col : str, default="mhm_Genus"
The name of the column in the mosquito habitat mapper DataFrame that contains the genus records.
bit_col : str, default="mhm_HasGenus"
The name of the column which will store the generated HasGenus flag
inplace : bool, default=False
Whether to return a new DataFrame. If True then no DataFrame copy is not returned and the operation is performed in place.
Returns
-------
pd.DataFrame
A DataFrame with the HasGenus flag. If `inplace=True` it returns None.
"""
if not inplace:
df = df.copy()
df[bit_col] = (~pd.isna(df[genus_col].to_numpy())).astype(int)
if not inplace:
return df
def infectious_genus_flag(
df, genus_col="mhm_Genus", bit_col="mhm_IsGenusOfInterest", inplace=False
):
"""
Creates a bit flag: `mhm_IsGenusOfInterest` where 1 denotes a Genus of a infectious mosquito and 0 denotes the contrary.
Parameters
----------
df : pd.DataFrame
A mosquito habitat mapper DataFrame
genus_col : str, default="mhm_Genus"
The name of the column in the mosquito habitat mapper DataFrame that contains the genus records.
bit_col : str, default="mhm_HasGenus"
The name of the column which will store the generated IsGenusOfInterest flag
inplace : bool, default=False
Whether to return a new DataFrame. If True then no DataFrame copy is not returned and the operation is performed in place.
Returns
-------
pd.DataFrame
A DataFrame with the IsGenusOfInterest flag. If `inplace=True` it returns None.
"""
if not inplace:
df = df.copy()
infectious_genus_flag = np.vectorize(
lambda genus: genus in ["Aedes", "Anopheles", "Culex"]
)
df[bit_col] = infectious_genus_flag(df[genus_col].to_numpy()).astype(int)
if not inplace:
return df
def is_container_flag(
df,
watersource_col="mhm_WaterSourceType",
bit_col="mhm_IsWaterSourceContainer",
inplace=False,
):
"""
Creates a bit flag: `mhm_IsWaterSourceContainer` where 1 denotes if a watersource is a container (e.g. ovitrap, pots, tires, etc.) and 0 denotes the contrary.
Parameters
----------
df : pd.DataFrame
A mosquito habitat mapper DataFrame
watersource_col : str, default="mhm_WaterSourceType"
The name of the column in the mosquito habitat mapper DataFrame that contains the watersource type records.
bit_col : str, default="mhm_IsWaterSourceContainer"
The name of the column which will store the generated IsWaterSourceContainer flag
inplace : bool, default=False
Whether to return a new DataFrame. If True then no DataFrame copy is not returned and the operation is performed in place.
Returns
-------
pd.DataFrame
A DataFrame with the IsContainer flag. If `inplace=True` it returns None.
"""
if not inplace:
df = df.copy()
mark_containers = np.vectorize(lambda container: "container" in container)
df[bit_col] = mark_containers(df[watersource_col].to_numpy()).astype(int)
if not inplace:
return df
def has_watersource_flag(
df, watersource_col="mhm_WaterSource", bit_col="mhm_HasWaterSource", inplace=False
):
"""
Creates a bit flag: `mhm_HasWaterSource` where 1 denotes if there is a watersource and 0 denotes the contrary.
Parameters
----------
df : pd.DataFrame
A mosquito habitat mapper DataFrame
watersource_col : str, default="mhm_WaterSource"
The name of the column in the mosquito habitat mapper DataFrame that contains the watersource records.
bit_col : str, default="mhm_IsWaterSourceContainer"
The name of the column which will store the generated HasWaterSource flag
inplace : bool, default=False
Whether to return a new DataFrame. If True then no DataFrame copy is not returned and the operation is performed in place.
Returns
-------
pd.DataFrame
A DataFrame with the HasWaterSource flag. If `inplace=True` it returns None.
"""
if not inplace:
df = df.copy()
has_watersource = np.vectorize(lambda watersource: int(not pd.isna(watersource)))
df[bit_col] = has_watersource(df[watersource_col].to_numpy())
if not inplace:
return df
def photo_bit_flags(
df,
watersource_photos="mhm_WaterSourcePhotoUrls",
larvae_photos="mhm_LarvaFullBodyPhotoUrls",
abdomen_photos="mhm_AbdomenCloseupPhotoUrls",
photo_count="mhm_PhotoCount",
rejected_count="mhm_RejectedCount",
pending_count="mhm_PendingCount",
photo_bit_binary="mhm_PhotoBitBinary",
photo_bit_decimal="mhm_PhotoBitDecimal",
inplace=False,
):
"""
Creates the following flags:
- `PhotoCount`: The number of valid photos per record.
- `RejectedCount`: The number of photos that were rejected per record.
- `PendingCount`: The number of photos that are pending approval per record.
- `PhotoBitBinary`: A string that represents the presence of a photo in the order of watersource, larvae, and abdomen. For example, if the entry is `110`, that indicates that there is a water source photo and a larvae photo, but no abdomen photo.
- `PhotoBitDecimal`: The numerical representation of the mhm_PhotoBitBinary string.
Parameters
----------
df : pd.DataFrame
A mosquito habitat mapper DataFrame
watersource_photos : str, default="mhm_WaterSourcePhotoUrls"
The name of the column in the mosquito habitat mapper DataFrame that contains the watersource photo url records.
larvae_photos : str, default="mhm_LarvaFullBodyPhotoUrls"
The name of the column in the mosquito habitat mapper DataFrame that contains the larvae photo url records.
abdomen_photos : str, default="mhm_AbdomenCloseupPhotoUrls"
The name of the column in the mosquito habitat mapper DataFrame that contains the abdomen photo url records.
photo_count : str, default="mhm_PhotoCount"
The name of the column that will store the PhotoCount flag.
rejected_count : str, default="mhm_RejectedCount"
The name of the column that will store the RejectedCount flag.
pending_count : str, default="mhm_PendingCount"
The name of the column that will store the PendingCount flag.
photo_bit_binary : str, default="mhm_PhotoBitBinary"
The name of the column that will store the PhotoBitBinary flag.
photo_bit_decimal : str, default="mhm_PhotoBitDecimal"
The name of the column that will store the PhotoBitDecimal flag.
inplace : bool, default=False
Whether to return a new DataFrame. If True then no DataFrame copy is not returned and the operation is performed in place.
Returns
-------
pd.DataFrame
A DataFrame with the photo flags. If `inplace=True` it returns None.
"""
def pic_data(*args):
pic_count = 0
rejected_count = 0
pending_count = 0
valid_photo_bit_mask = ""
# bit_power = len(args) - 1
# For url string -- if we see ANY http, add 1
# also count all valid photos, rejected photos,
# If there are NO http then add 0, to empty photo field
for url_string in args:
if not pd.isna(url_string):
if "http" not in url_string:
valid_photo_bit_mask += "0"
else:
valid_photo_bit_mask += "1"
pic_count += url_string.count("http")
pending_count += url_string.count("pending")
rejected_count += url_string.count("rejected")
else:
valid_photo_bit_mask += "0"
return (
pic_count,
rejected_count,
pending_count,
valid_photo_bit_mask,
int(valid_photo_bit_mask, 2),
)
if not inplace:
df = df.copy()
get_photo_data = np.vectorize(pic_data)
(
df[photo_count],
df[rejected_count],
df[pending_count],
df[photo_bit_binary],
df[photo_bit_decimal],
) = get_photo_data(
df[watersource_photos].to_numpy(),
df[larvae_photos].to_numpy(),
df[abdomen_photos].to_numpy(),
)
if not inplace:
return df
def completion_score_flag(
df,
photo_bit_binary="mhm_PhotoBitBinary",
has_genus="mhm_HasGenus",
sub_completeness="mhm_SubCompletenessScore",
completeness="mhm_CumulativeCompletenessScore",
inplace=False,
):
"""
Adds the following completness score flags:
- `SubCompletenessScore`: The percentage of the watersource photos, larvae photos, abdomen photos, and genus columns that are filled out.
- `CumulativeCompletenessScore`: The percentage of non null values out of all the columns.
Parameters
----------
df : pd.DataFrame
A mosquito habitat mapper DataFrame with the [`PhotoBitDecimal`](#photo_bit_flags) and [`HasGenus`](#has_genus_flags) flags.
photo_bit_binary: str, default="mhm_PhotoBitBinary"
The name of the column in the mosquito habitat mapper DataFrame that contains the PhotoBitBinary flag.
sub_completeness : str, default="mhm_HasGenus"
The name of the column in the mosquito habitat mapper DataFrame that will contain the generated SubCompletenessScore flag.
completeness : str, default="mhm_SubCompletenessScore"
The name of the column in the mosquito habitat mapper DataFrame that will contain the generated CumulativeCompletenessScore flag.
inplace : bool, default=False
Whether to return a new DataFrame. If True then no DataFrame copy is not returned and the operation is performed in place.
Returns
-------
pd.DataFrame
A DataFrame with completion score flags. If `inplace=True` it returns None.
"""
def sum_bit_mask(bit_mask="0"):
total = 0.0
for char in bit_mask:
total += int(char)
return total
if not inplace:
df = df.copy()
scores = {}
scores["sub_score"] = []
# Cummulative Completion Score
scores["cumulative_score"] = round(df.count(axis=1) / len(df.columns), 2)
# Sub-Score
for index in df.index:
bit_mask = df[photo_bit_binary][index]
sub_score = df[has_genus][index] + sum_bit_mask(bit_mask=bit_mask)
sub_score /= 4.0
scores["sub_score"].append(sub_score)
df[sub_completeness], df[completeness] = (
scores["sub_score"],
scores["cumulative_score"],
)
if not inplace:
return df
def apply_cleanup(mhm_df):
"""Applies a full cleanup procedure to the mosquito habitat mapper data. Only returns a copy.
It follows the following steps:
- Removes Homogenous Columns
- Renames Latitude and Longitudes
- Cleans the Column Naming
- Converts Larvae Count to Numbers
- Rounds Columns
- Standardizes Null Values
Parameters
----------
mhm_df : pd.DataFrame
A DataFrame containing **raw** Mosquito Habitat Mapper Data from the API.
Returns
-------
pd.DataFrame
A DataFrame containing the cleaned up Mosquito Habitat Mapper Data
"""
mhm_df = mhm_df.copy()
rename_latlon_cols(mhm_df, inplace=True)
cleanup_column_prefix(mhm_df, inplace=True)
larvae_to_num(mhm_df, inplace=True)
round_cols(mhm_df, inplace=True)
standardize_null_vals(mhm_df, inplace=True)
return mhm_df
def add_flags(mhm_df):
"""Adds the following flags to the Mosquito Habitat Mapper Data:
- Has Genus
- Is Infectious Genus/Genus of Interest
- Is Container
- Has WaterSource
- Photo Bit Flags
- Completion Score Flag
This returns a copy of the original DataFrame with the flags added onto it.
Parameters
----------
mhm_df : pd.DataFrame
A DataFrame containing cleaned up Mosquito Habitat Mapper Data ideally from the method.
Returns
-------
pd.DataFrame
A DataFrame containing the flagged Mosquito Habitat Mapper Data
"""
mhm_df = mhm_df.copy()
has_genus_flag(mhm_df, inplace=True)
infectious_genus_flag(mhm_df, inplace=True)
is_container_flag(mhm_df, inplace=True)
has_watersource_flag(mhm_df, inplace=True)
photo_bit_flags(mhm_df, inplace=True)
completion_score_flag(mhm_df, inplace=True)
return mhm_df
def plot_valid_entries(df, bit_col, entry_type):
"""
Plots the number of entries with photos and the number of entries without photos
Parameters
----------
df : pd.DataFrame
The DataFrame containing Mosquito Habitat Mapper Data with the PhotoBitDecimal Flag.
"""
plt.figure()
num_valid = len(df[df[bit_col] > 0])
plt.title(f"Entries with {entry_type} vs No {entry_type}")
plt.ylabel("Number of Entries")
plt.bar(entry_type, num_valid, color="#e34a33")
plt.bar(f"No {entry_type}", len(df) - num_valid, color="#fdcc8a")
def photo_subjects(mhm_df):
"""
Plots the amount of photos for each photo area (Larvae, Abdomen, Watersource)
Parameters
----------
mhm_df : pd.DataFrame
The DataFrame containing Mosquito Habitat Mapper Data with the PhotoBitDecimal Flag.
"""
total_dict = {"Larvae Photos": 0, "Abdomen Photos": 0, "Watersource Photos": 0}
for number in mhm_df["mhm_PhotoBitDecimal"]:
total_dict["Watersource Photos"] += number & 4
total_dict["Larvae Photos"] += number & 2
total_dict["Abdomen Photos"] += number & 1
for key in total_dict.keys():
if total_dict[key] != 0:
total_dict[key] = math.log10(total_dict[key])
else:
total_dict[key] = 0
plt.figure(figsize=(10, 5))
plt.title("Mosquito Habitat Mapper - Photo Subject Frequencies (Log Scale)")
plt.xlabel("Photo Type")
plt.ylabel("Frequency (Log Scale)")
plt.bar(total_dict.keys(), total_dict.values(), color="lightblue")
def diagnostic_plots(mhm_df):
"""
Generates (but doesn't display) diagnostic plots to gain insight into the current data.
Plots:
- Larvae Count Distribution (where a negative entry denotes null data)
- Photo Subject Distribution
- Number of valid photos vs no photos
- Completeness Score Distribution
- Subcompleteness Score Distribution
Parameters
----------
mhm_df : pd.DataFrame
The DataFrame containing Flagged and Cleaned Mosquito Habitat Mapper Data.
"""
plot_int_distribution(mhm_df, "mhm_LarvaeCount", "Larvae Count")
photo_subjects(mhm_df)
plot_freq_bar(mhm_df, "Mosquito Habitat Mapper", "mhm_Genus", "Genus Types")
plot_valid_entries(mhm_df, "mhm_HasGenus", "Genus Classifications")
plot_valid_entries(mhm_df, "mhm_PhotoBitDecimal", "Valid Photos")
completeness_histogram(
mhm_df,
"Mosquito Habitat Mapper",
"mhm_CumulativeCompletenessScore",
"Cumulative Completeness",
)
completeness_histogram(
mhm_df,
"Mosquito Habitat Mapper",
"mhm_SubCompletenessScore",
"Sub Completeness",
)
def qa_filter(
mhm_df,
has_genus=False,
min_larvae_count=-9999,
has_photos=False,
is_container=False,
):
"""
Can filter a cleaned and flagged mosquito habitat mapper DataFrame based on the following criteria:
- `Has Genus`: If the entry has an identified genus
- `Min Larvae Count` : Minimum larvae count needed for an entry
- `Has Photos` : If the entry contains valid photo entries
- `Is Container` : If the entry's watersource was a container
Returns a copy of the DataFrame
Parameters
----------
has_genus : bool, default=False
If True, only entries with an identified genus will be returned.
min_larvae_count : int, default=-9999
Only entries with a larvae count greater than or equal to this parameter will be included.
has_photos : bool, default=False
If True, only entries with recorded photos will be returned
is_container : bool, default=False
If True, only entries with containers will be returned
Returns
-------
pd.DataFrame
A DataFrame of the applied filters.
"""
mhm_df = mhm_df[mhm_df["mhm_LarvaeCount"] >= min_larvae_count]
if has_genus:
mhm_df = mhm_df[mhm_df["mhm_HasGenus"] == 1]
if has_photos:
mhm_df = mhm_df[mhm_df["mhm_PhotoBitDecimal"] > 0]
if is_container:
mhm_df = mhm_df[mhm_df["mhm_IsWaterSourceContainer"] == 1]
return mhm_df
|
[
"go_utils.plot.completeness_histogram",
"go_utils.plot.plot_int_distribution",
"matplotlib.pyplot.ylabel",
"go_utils.cleanup.round_cols",
"matplotlib.pyplot.xlabel",
"go_utils.cleanup.rename_latlon_cols",
"matplotlib.pyplot.figure",
"matplotlib.pyplot.bar",
"re.sub",
"go_utils.cleanup.standardize_null_vals",
"pandas.isna",
"matplotlib.pyplot.title",
"math.log10",
"go_utils.cleanup.replace_column_prefix",
"numpy.vectorize",
"go_utils.plot.plot_freq_bar"
] |
[((2550, 2624), 'go_utils.cleanup.replace_column_prefix', 'replace_column_prefix', (['df', '"""mosquitohabitatmapper"""', '"""mhm"""'], {'inplace': 'inplace'}), "(df, 'mosquitohabitatmapper', 'mhm', inplace=inplace)\n", (2571, 2624), False, 'from go_utils.cleanup import rename_latlon_cols, replace_column_prefix, round_cols, standardize_null_vals\n'), ((4754, 4781), 'numpy.vectorize', 'np.vectorize', (['_entry_to_num'], {}), '(_entry_to_num)\n', (4766, 4781), True, 'import numpy as np\n'), ((6970, 7038), 'numpy.vectorize', 'np.vectorize', (["(lambda genus: genus in ['Aedes', 'Anopheles', 'Culex'])"], {}), "(lambda genus: genus in ['Aedes', 'Anopheles', 'Culex'])\n", (6982, 7038), True, 'import numpy as np\n'), ((8261, 8317), 'numpy.vectorize', 'np.vectorize', (["(lambda container: 'container' in container)"], {}), "(lambda container: 'container' in container)\n", (8273, 8317), True, 'import numpy as np\n'), ((13235, 13257), 'numpy.vectorize', 'np.vectorize', (['pic_data'], {}), '(pic_data)\n', (13247, 13257), True, 'import numpy as np\n'), ((16539, 16579), 'go_utils.cleanup.rename_latlon_cols', 'rename_latlon_cols', (['mhm_df'], {'inplace': '(True)'}), '(mhm_df, inplace=True)\n', (16557, 16579), False, 'from go_utils.cleanup import rename_latlon_cols, replace_column_prefix, round_cols, standardize_null_vals\n'), ((16672, 16704), 'go_utils.cleanup.round_cols', 'round_cols', (['mhm_df'], {'inplace': '(True)'}), '(mhm_df, inplace=True)\n', (16682, 16704), False, 'from go_utils.cleanup import rename_latlon_cols, replace_column_prefix, round_cols, standardize_null_vals\n'), ((16709, 16752), 'go_utils.cleanup.standardize_null_vals', 'standardize_null_vals', (['mhm_df'], {'inplace': '(True)'}), '(mhm_df, inplace=True)\n', (16730, 16752), False, 'from go_utils.cleanup import rename_latlon_cols, replace_column_prefix, round_cols, standardize_null_vals\n'), ((17989, 18001), 'matplotlib.pyplot.figure', 'plt.figure', ([], {}), '()\n', (17999, 18001), True, 'import matplotlib.pyplot as plt\n'), ((18047, 18105), 'matplotlib.pyplot.title', 'plt.title', (['f"""Entries with {entry_type} vs No {entry_type}"""'], {}), "(f'Entries with {entry_type} vs No {entry_type}')\n", (18056, 18105), True, 'import matplotlib.pyplot as plt\n'), ((18110, 18141), 'matplotlib.pyplot.ylabel', 'plt.ylabel', (['"""Number of Entries"""'], {}), "('Number of Entries')\n", (18120, 18141), True, 'import matplotlib.pyplot as plt\n'), ((18146, 18193), 'matplotlib.pyplot.bar', 'plt.bar', (['entry_type', 'num_valid'], {'color': '"""#e34a33"""'}), "(entry_type, num_valid, color='#e34a33')\n", (18153, 18193), True, 'import matplotlib.pyplot as plt\n'), ((19009, 19036), 'matplotlib.pyplot.figure', 'plt.figure', ([], {'figsize': '(10, 5)'}), '(figsize=(10, 5))\n', (19019, 19036), True, 'import matplotlib.pyplot as plt\n'), ((19041, 19117), 'matplotlib.pyplot.title', 'plt.title', (['"""Mosquito Habitat Mapper - Photo Subject Frequencies (Log Scale)"""'], {}), "('Mosquito Habitat Mapper - Photo Subject Frequencies (Log Scale)')\n", (19050, 19117), True, 'import matplotlib.pyplot as plt\n'), ((19122, 19146), 'matplotlib.pyplot.xlabel', 'plt.xlabel', (['"""Photo Type"""'], {}), "('Photo Type')\n", (19132, 19146), True, 'import matplotlib.pyplot as plt\n'), ((19151, 19186), 'matplotlib.pyplot.ylabel', 'plt.ylabel', (['"""Frequency (Log Scale)"""'], {}), "('Frequency (Log Scale)')\n", (19161, 19186), True, 'import matplotlib.pyplot as plt\n'), ((19783, 19847), 'go_utils.plot.plot_int_distribution', 'plot_int_distribution', (['mhm_df', '"""mhm_LarvaeCount"""', '"""Larvae Count"""'], {}), "(mhm_df, 'mhm_LarvaeCount', 'Larvae Count')\n", (19804, 19847), False, 'from go_utils.plot import completeness_histogram, plot_freq_bar, plot_int_distribution\n'), ((19879, 19955), 'go_utils.plot.plot_freq_bar', 'plot_freq_bar', (['mhm_df', '"""Mosquito Habitat Mapper"""', '"""mhm_Genus"""', '"""Genus Types"""'], {}), "(mhm_df, 'Mosquito Habitat Mapper', 'mhm_Genus', 'Genus Types')\n", (19892, 19955), False, 'from go_utils.plot import completeness_histogram, plot_freq_bar, plot_int_distribution\n'), ((20102, 20225), 'go_utils.plot.completeness_histogram', 'completeness_histogram', (['mhm_df', '"""Mosquito Habitat Mapper"""', '"""mhm_CumulativeCompletenessScore"""', '"""Cumulative Completeness"""'], {}), "(mhm_df, 'Mosquito Habitat Mapper',\n 'mhm_CumulativeCompletenessScore', 'Cumulative Completeness')\n", (20124, 20225), False, 'from go_utils.plot import completeness_histogram, plot_freq_bar, plot_int_distribution\n'), ((20265, 20374), 'go_utils.plot.completeness_histogram', 'completeness_histogram', (['mhm_df', '"""Mosquito Habitat Mapper"""', '"""mhm_SubCompletenessScore"""', '"""Sub Completeness"""'], {}), "(mhm_df, 'Mosquito Habitat Mapper',\n 'mhm_SubCompletenessScore', 'Sub Completeness')\n", (20287, 20374), False, 'from go_utils.plot import completeness_histogram, plot_freq_bar, plot_int_distribution\n'), ((2739, 2753), 'pandas.isna', 'pd.isna', (['entry'], {}), '(entry)\n', (2746, 2753), True, 'import pandas as pd\n'), ((18931, 18958), 'math.log10', 'math.log10', (['total_dict[key]'], {}), '(total_dict[key])\n', (18941, 18958), False, 'import math\n'), ((4618, 4632), 'pandas.isna', 'pd.isna', (['count'], {}), '(count)\n', (4625, 4632), True, 'import pandas as pd\n'), ((12562, 12581), 'pandas.isna', 'pd.isna', (['url_string'], {}), '(url_string)\n', (12569, 12581), True, 'import pandas as pd\n'), ((2979, 3003), 're.sub', 're.sub', (['"""-.*"""', '""""""', 'entry'], {}), "('-.*', '', entry)\n", (2985, 3003), False, 'import re\n'), ((9482, 9502), 'pandas.isna', 'pd.isna', (['watersource'], {}), '(watersource)\n', (9489, 9502), True, 'import pandas as pd\n')]
|
# -*- coding: utf-8 -*-
# This code is part of Qiskit.
#
# (C) Copyright IBM 2020.
#
# This code is licensed under the Apache License, Version 2.0. You may
# obtain a copy of this license in the LICENSE.txt file in the root directory
# of this source tree or at http://www.apache.org/licenses/LICENSE-2.0.
#
# Any modifications or derivative works of this code must retain this
# copyright notice, and modified files need to carry a notice indicating
# that they have been altered from the originals.
"""FCIDump dumper."""
from typing import List, Optional
from io import TextIOWrapper
import itertools
import numpy as np
def dump(outpath: str, norb: int, nelec: int, hijs: List[float], hijkls: List[float], einact: float,
ms2: int = 0, orbsym: Optional[List[int]] = None, isym: int = 1
) -> None:
# pylint: disable=wrong-spelling-in-docstring
"""Generates a FCIDump output.
Args:
outpath: Path to the output file.
norb: The number of orbitals.
nelec: The number of electrons.
hijs: The pair of alpha and beta 1-electron integrals. The latter may be None.
hijkls: The triplet of alpha/alpha, beta/alpha and beta/beta 2-electron integrals. The
latter two may be None.
einact: The inactive energy.
ms2: 2*S, where S is the spin quantum number.
orbsym: A list of spatial symmetries of the orbitals.
isym: The spatial symmetry of the wave function.
"""
hij, hij_b = hijs
hijkl, hijkl_ba, hijkl_bb = hijkls
# assert that either all beta variables are None or all of them are not
assert all([h is None for h in [hij_b, hijkl_ba, hijkl_bb]]) \
or all([h is not None for h in [hij_b, hijkl_ba, hijkl_bb]])
assert norb == hij.shape[0] == hijkl.shape[0]
mos = range(norb)
with open(outpath, 'w') as outfile:
# print header
outfile.write('&FCI NORB={:4d},NELEC={:4d},MS2={:4d}\n'.format(norb, nelec, ms2))
if orbsym is None:
outfile.write(' ORBSYM=' + '1,'*norb + '\n')
else:
assert len(orbsym) == norb
outfile.write(' ORBSYM=' + ','.join(orbsym) + '\n')
outfile.write(' ISYM={:d},\n/&END\n'.format(isym))
# append 2e integrals
_dump_2e_ints(hijkl, mos, outfile)
if hijkl_ba is not None:
_dump_2e_ints(hijkl_ba.transpose(), mos, outfile, beta=1)
if hijkl_bb is not None:
_dump_2e_ints(hijkl_bb, mos, outfile, beta=2)
# append 1e integrals
_dump_1e_ints(hij, mos, outfile)
if hij_b is not None:
_dump_1e_ints(hij_b, mos, outfile, beta=True)
# TODO append MO energies (last three indices are 0)
# append inactive energy
_write_to_outfile(outfile, einact, (0, 0, 0, 0))
def _dump_1e_ints(hij: List[float], mos: List[int], outfile: TextIOWrapper,
beta: bool = False) -> None:
idx_offset = 1 if not beta else 1+len(mos)
hij_elements = set()
for i, j in itertools.product(mos, repeat=2):
if i == j:
_write_to_outfile(outfile, hij[i][j], (i+idx_offset, j+idx_offset, 0, 0))
continue
if (j, i) in hij_elements and np.isclose(hij[i][j], hij[j][i]):
continue
_write_to_outfile(outfile, hij[i][j], (i+idx_offset, j+idx_offset, 0, 0))
hij_elements.add((i, j))
def _dump_2e_ints(hijkl: List[float], mos: List[int], outfile: TextIOWrapper,
beta: int = 0) -> None:
idx_offsets = [1, 1]
for b in range(beta):
idx_offsets[1-b] += len(mos)
hijkl_elements = set()
# pylint: disable=invalid-name
for elem in itertools.product(mos, repeat=4):
if np.isclose(hijkl[elem], 0.0, atol=1e-14):
continue
if len(set(elem)) == 1:
_write_to_outfile(outfile, hijkl[elem], (*[e+idx_offsets[0] for e in elem[:2]],
*[e+idx_offsets[1] for e in elem[2:]]))
continue
if beta != 1 and elem[::-1] in hijkl_elements and \
np.isclose(hijkl[elem], hijkl[elem[::-1]]):
continue
bra_perms = set(itertools.permutations(elem[:2]))
ket_perms = set(itertools.permutations(elem[2:]))
if beta == 1:
permutations = itertools.product(bra_perms, ket_perms)
else:
permutations = itertools.chain(
itertools.product(bra_perms, ket_perms),
itertools.product(ket_perms, bra_perms)
)
for perm in {e1 + e2 for e1, e2 in permutations}:
if perm in hijkl_elements and np.isclose(hijkl[elem], hijkl[perm]):
break
else:
_write_to_outfile(outfile, hijkl[elem], (*[e+idx_offsets[0] for e in elem[:2]],
*[e+idx_offsets[1] for e in elem[2:]]))
hijkl_elements.add(elem)
def _write_to_outfile(outfile: str, value: float, indices: List[int]):
outfile.write('{:23.16E}{:4d}{:4d}{:4d}{:4d}\n'.format(value, *indices))
|
[
"itertools.permutations",
"itertools.product",
"numpy.isclose"
] |
[((3021, 3053), 'itertools.product', 'itertools.product', (['mos'], {'repeat': '(2)'}), '(mos, repeat=2)\n', (3038, 3053), False, 'import itertools\n'), ((3677, 3709), 'itertools.product', 'itertools.product', (['mos'], {'repeat': '(4)'}), '(mos, repeat=4)\n', (3694, 3709), False, 'import itertools\n'), ((3722, 3762), 'numpy.isclose', 'np.isclose', (['hijkl[elem]', '(0.0)'], {'atol': '(1e-14)'}), '(hijkl[elem], 0.0, atol=1e-14)\n', (3732, 3762), True, 'import numpy as np\n'), ((3219, 3251), 'numpy.isclose', 'np.isclose', (['hij[i][j]', 'hij[j][i]'], {}), '(hij[i][j], hij[j][i])\n', (3229, 3251), True, 'import numpy as np\n'), ((4099, 4141), 'numpy.isclose', 'np.isclose', (['hijkl[elem]', 'hijkl[elem[::-1]]'], {}), '(hijkl[elem], hijkl[elem[::-1]])\n', (4109, 4141), True, 'import numpy as np\n'), ((4188, 4220), 'itertools.permutations', 'itertools.permutations', (['elem[:2]'], {}), '(elem[:2])\n', (4210, 4220), False, 'import itertools\n'), ((4246, 4278), 'itertools.permutations', 'itertools.permutations', (['elem[2:]'], {}), '(elem[2:])\n', (4268, 4278), False, 'import itertools\n'), ((4329, 4368), 'itertools.product', 'itertools.product', (['bra_perms', 'ket_perms'], {}), '(bra_perms, ket_perms)\n', (4346, 4368), False, 'import itertools\n'), ((4443, 4482), 'itertools.product', 'itertools.product', (['bra_perms', 'ket_perms'], {}), '(bra_perms, ket_perms)\n', (4460, 4482), False, 'import itertools\n'), ((4500, 4539), 'itertools.product', 'itertools.product', (['ket_perms', 'bra_perms'], {}), '(ket_perms, bra_perms)\n', (4517, 4539), False, 'import itertools\n'), ((4654, 4690), 'numpy.isclose', 'np.isclose', (['hijkl[elem]', 'hijkl[perm]'], {}), '(hijkl[elem], hijkl[perm])\n', (4664, 4690), True, 'import numpy as np\n')]
|
from flask import Flask, flash, request, redirect, url_for, render_template
from werkzeug.utils import secure_filename
import os
from keras.models import load_model
from keras.applications.inception_resnet_v2 import InceptionResNetV2
import tensorflow as tf
from skimage.io import imsave
from skimage.transform import resize
import numpy as np
from keras.preprocessing.image import ImageDataGenerator, array_to_img, img_to_array, load_img
from skimage.color import rgb2lab, lab2rgb, rgb2gray, gray2rgb
from keras.applications.inception_resnet_v2 import preprocess_input
from PIL import Image,ImageChops
import logging
global graph
graph = tf.get_default_graph()
app = Flask(__name__)
app.secret_key = "hello"
ALLOWED_EXTENSIONS = set(['png', 'jpg', 'jpeg'])
model = load_model('trained-model.h5')
UPLOAD_FOLDER = '/home/nubaf/Git-Projects/colorization/files'
app.config['UPLOAD_FOLDER'] = UPLOAD_FOLDER
files = [f for f in os.listdir('.') if os.path.isfile(f)]
checkInception = False
for f in files:
if f == "inception.h5":
checkInception = True
inception = load_model('inception.h5', compile=False)
break
if not checkInception:
inception = InceptionResNetV2(weights='imagenet', include_top=True)
inception.save('inception.h5')
inception.graph = graph
def create_inception_embedding(grayscaled_rgb):
grayscaled_rgb_resized = []
for i in grayscaled_rgb:
i = resize(i, (299, 299, 3), mode='constant')
grayscaled_rgb_resized.append(i)
grayscaled_rgb_resized = np.array(grayscaled_rgb_resized)
grayscaled_rgb_resized = preprocess_input(grayscaled_rgb_resized)
with graph.as_default():
embed = inception.predict(grayscaled_rgb_resized)
return embed
def allowed_file(filename):
return '.' in filename and \
filename.rsplit('.', 1)[1].lower() in ALLOWED_EXTENSIONS
@app.route('/', methods=['GET', 'POST'])
def upload_file():
if request.method == 'POST':
try:
url = request.form['url']
if 'examples' in url:
color_file = process(url)
return render_template('index.html', res='static/examples/girl.jpg')
# check if the post request has the file part
except:
logging.exception('')
if 'file' not in request.files:
flash('No file part')
return redirect(request.url)
file = request.files['file']
# if user does not select file, browser also
# submit an empty part without filename
if file.filename == '':
flash('No selected file')
return redirect(request.url)
if file and allowed_file(file.filename):
filename = secure_filename(file.filename)
file.save(os.path.join(app.config['UPLOAD_FOLDER'], filename))
color_file = process(file.filename)
return render_template('index.html', og=color_file[0], res=color_file[1])
return render_template('index.html')
def process(img):
if 'examples' in img:
im = Image.open(img)
name = img.split('.')[0].split('/')[-1]
else:
im = Image.open('files/' + img)
name = img.split('.')[0]
old_size = im.size # old_size[0] is in (width, height) format
ratio = float(256)/max(old_size)
new_size = tuple([int(x*ratio) for x in old_size])
im = im.resize(new_size, Image.ANTIALIAS)
new_im = Image.new("RGB", (256, 256))
new_im.paste(im, ((256-new_size[0])//2,(256-new_size[1])//2))
new_im.save('static/processed_png/' + name + ".png","PNG")
a = np.array(img_to_array(load_img('static/processed_png/' + name +'.png')))
a = a.reshape(1,256,256,3)
#gray_me = gray2rgb(rgb2gray(1.0/255*a))
color_me_embed = create_inception_embedding(a)
a = rgb2lab(1.0/255*a)[:,:,:,0]
a = a.reshape(a.shape+(1,))
with graph.as_default():
output = model.predict([a, color_me_embed])
output = output * 128
for i in range(len(output)):
cur = np.zeros((256, 256, 3))
cur[:,:,0] = a[i][:,:,0]
cur[:,:,1:] = output[i]
imsave(f'static/colored_img/{name}.png',(lab2rgb(cur)))
trim(Image.open(f'static/processed_png/{name}.png')).save(f'static/processed_png/{name}.png')
trim(Image.open(f'static/colored_img/{name}.png')).save(f'static/colored_img/{name}.png')
return (f'static/processed_png/{name}.png',f'static/colored_img/{name}.png')
def trim(im):
bg = Image.new(im.mode, im.size, im.getpixel((0,0)))
diff = ImageChops.difference(im, bg)
diff = ImageChops.add(diff, diff, 2.0, -100)
bbox = diff.getbbox()
if bbox:
return im.crop(bbox)
if __name__ == "__main__":
app.run(debug=True)
|
[
"flask.render_template",
"flask.Flask",
"PIL.Image.new",
"logging.exception",
"numpy.array",
"werkzeug.utils.secure_filename",
"os.listdir",
"skimage.color.rgb2lab",
"flask.flash",
"keras.applications.inception_resnet_v2.preprocess_input",
"skimage.color.lab2rgb",
"keras.applications.inception_resnet_v2.InceptionResNetV2",
"PIL.ImageChops.add",
"tensorflow.get_default_graph",
"PIL.ImageChops.difference",
"os.path.isfile",
"flask.redirect",
"skimage.transform.resize",
"keras.preprocessing.image.load_img",
"PIL.Image.open",
"keras.models.load_model",
"os.path.join",
"numpy.zeros"
] |
[((642, 664), 'tensorflow.get_default_graph', 'tf.get_default_graph', ([], {}), '()\n', (662, 664), True, 'import tensorflow as tf\n'), ((671, 686), 'flask.Flask', 'Flask', (['__name__'], {}), '(__name__)\n', (676, 686), False, 'from flask import Flask, flash, request, redirect, url_for, render_template\n'), ((769, 799), 'keras.models.load_model', 'load_model', (['"""trained-model.h5"""'], {}), "('trained-model.h5')\n", (779, 799), False, 'from keras.models import load_model\n'), ((1176, 1231), 'keras.applications.inception_resnet_v2.InceptionResNetV2', 'InceptionResNetV2', ([], {'weights': '"""imagenet"""', 'include_top': '(True)'}), "(weights='imagenet', include_top=True)\n", (1193, 1231), False, 'from keras.applications.inception_resnet_v2 import InceptionResNetV2\n'), ((1526, 1558), 'numpy.array', 'np.array', (['grayscaled_rgb_resized'], {}), '(grayscaled_rgb_resized)\n', (1534, 1558), True, 'import numpy as np\n'), ((1588, 1628), 'keras.applications.inception_resnet_v2.preprocess_input', 'preprocess_input', (['grayscaled_rgb_resized'], {}), '(grayscaled_rgb_resized)\n', (1604, 1628), False, 'from keras.applications.inception_resnet_v2 import preprocess_input\n'), ((2969, 2998), 'flask.render_template', 'render_template', (['"""index.html"""'], {}), "('index.html')\n", (2984, 2998), False, 'from flask import Flask, flash, request, redirect, url_for, render_template\n'), ((3422, 3450), 'PIL.Image.new', 'Image.new', (['"""RGB"""', '(256, 256)'], {}), "('RGB', (256, 256))\n", (3431, 3450), False, 'from PIL import Image, ImageChops\n'), ((4569, 4598), 'PIL.ImageChops.difference', 'ImageChops.difference', (['im', 'bg'], {}), '(im, bg)\n', (4590, 4598), False, 'from PIL import Image, ImageChops\n'), ((4610, 4647), 'PIL.ImageChops.add', 'ImageChops.add', (['diff', 'diff', '(2.0)', '(-100)'], {}), '(diff, diff, 2.0, -100)\n', (4624, 4647), False, 'from PIL import Image, ImageChops\n'), ((926, 941), 'os.listdir', 'os.listdir', (['"""."""'], {}), "('.')\n", (936, 941), False, 'import os\n'), ((945, 962), 'os.path.isfile', 'os.path.isfile', (['f'], {}), '(f)\n', (959, 962), False, 'import os\n'), ((1081, 1122), 'keras.models.load_model', 'load_model', (['"""inception.h5"""'], {'compile': '(False)'}), "('inception.h5', compile=False)\n", (1091, 1122), False, 'from keras.models import load_model\n'), ((1414, 1455), 'skimage.transform.resize', 'resize', (['i', '(299, 299, 3)'], {'mode': '"""constant"""'}), "(i, (299, 299, 3), mode='constant')\n", (1420, 1455), False, 'from skimage.transform import resize\n'), ((3057, 3072), 'PIL.Image.open', 'Image.open', (['img'], {}), '(img)\n', (3067, 3072), False, 'from PIL import Image, ImageChops\n'), ((3144, 3170), 'PIL.Image.open', 'Image.open', (["('files/' + img)"], {}), "('files/' + img)\n", (3154, 3170), False, 'from PIL import Image, ImageChops\n'), ((3796, 3818), 'skimage.color.rgb2lab', 'rgb2lab', (['(1.0 / 255 * a)'], {}), '(1.0 / 255 * a)\n', (3803, 3818), False, 'from skimage.color import rgb2lab, lab2rgb, rgb2gray, gray2rgb\n'), ((2332, 2353), 'flask.flash', 'flash', (['"""No file part"""'], {}), "('No file part')\n", (2337, 2353), False, 'from flask import Flask, flash, request, redirect, url_for, render_template\n'), ((2373, 2394), 'flask.redirect', 'redirect', (['request.url'], {}), '(request.url)\n', (2381, 2394), False, 'from flask import Flask, flash, request, redirect, url_for, render_template\n'), ((2577, 2602), 'flask.flash', 'flash', (['"""No selected file"""'], {}), "('No selected file')\n", (2582, 2602), False, 'from flask import Flask, flash, request, redirect, url_for, render_template\n'), ((2622, 2643), 'flask.redirect', 'redirect', (['request.url'], {}), '(request.url)\n', (2630, 2643), False, 'from flask import Flask, flash, request, redirect, url_for, render_template\n'), ((2716, 2746), 'werkzeug.utils.secure_filename', 'secure_filename', (['file.filename'], {}), '(file.filename)\n', (2731, 2746), False, 'from werkzeug.utils import secure_filename\n'), ((2889, 2955), 'flask.render_template', 'render_template', (['"""index.html"""'], {'og': 'color_file[0]', 'res': 'color_file[1]'}), "('index.html', og=color_file[0], res=color_file[1])\n", (2904, 2955), False, 'from flask import Flask, flash, request, redirect, url_for, render_template\n'), ((3610, 3659), 'keras.preprocessing.image.load_img', 'load_img', (["('static/processed_png/' + name + '.png')"], {}), "('static/processed_png/' + name + '.png')\n", (3618, 3659), False, 'from keras.preprocessing.image import ImageDataGenerator, array_to_img, img_to_array, load_img\n'), ((4022, 4045), 'numpy.zeros', 'np.zeros', (['(256, 256, 3)'], {}), '((256, 256, 3))\n', (4030, 4045), True, 'import numpy as np\n'), ((2114, 2175), 'flask.render_template', 'render_template', (['"""index.html"""'], {'res': '"""static/examples/girl.jpg"""'}), "('index.html', res='static/examples/girl.jpg')\n", (2129, 2175), False, 'from flask import Flask, flash, request, redirect, url_for, render_template\n'), ((2258, 2279), 'logging.exception', 'logging.exception', (['""""""'], {}), "('')\n", (2275, 2279), False, 'import logging\n'), ((2769, 2820), 'os.path.join', 'os.path.join', (["app.config['UPLOAD_FOLDER']", 'filename'], {}), "(app.config['UPLOAD_FOLDER'], filename)\n", (2781, 2820), False, 'import os\n'), ((4172, 4184), 'skimage.color.lab2rgb', 'lab2rgb', (['cur'], {}), '(cur)\n', (4179, 4184), False, 'from skimage.color import rgb2lab, lab2rgb, rgb2gray, gray2rgb\n'), ((4204, 4250), 'PIL.Image.open', 'Image.open', (['f"""static/processed_png/{name}.png"""'], {}), "(f'static/processed_png/{name}.png')\n", (4214, 4250), False, 'from PIL import Image, ImageChops\n'), ((4310, 4354), 'PIL.Image.open', 'Image.open', (['f"""static/colored_img/{name}.png"""'], {}), "(f'static/colored_img/{name}.png')\n", (4320, 4354), False, 'from PIL import Image, ImageChops\n')]
|
from collections import ChainMap
from collections.abc import Mapping, Iterable
from itertools import groupby
from operator import itemgetter
import numpy as np
from probability import RowKey
from probability import TableColumns
# from probability.core_1 import RowKey
# from probability.core_1 import TableColumns
def to_dict(groupby_index, value_index):
def make_dict(sorted_items):
# It groups the sorted item based on
# the element as groupby_index
# and then sum the values at value_index
return {
k: sum([item[value_index] for item in g2])
for k, g2 in groupby(sorted_items, key=itemgetter(groupby_index))
}
return make_dict
class Table(dict):
def __init__(self, rows, names=None, _internal_=False, _children_names_=None):
if _internal_:
# rows are dictionary for internal calls
key_values = rows
try:
self._row_sample_ = next(iter(rows))
except StopIteration:
# Rows are empty
super().__init__(key_values)
self._row_sample_ = None
self.names = names
if _children_names_ is None:
self.children_names = []
self.columns = TableColumns(
names=names, children_names=[], table=self
)
else:
self.children_names = _children_names_
self.columns = TableColumns(
names=names, children_names=_children_names_, table=self
)
return
else:
if isinstance(rows, Mapping):
key_values = [(RowKey(k), value) for k, value in rows.items()]
elif isinstance(rows, Iterable):
key_values = [(RowKey(k), value) for k, value in rows]
else:
raise ValueError("Table expect rows as Mapping/Iterable")
self._row_sample_ = key_values[0][0]
if names is None:
names = [f"X{i+1}" for i, _ in enumerate(self._row_sample_)]
if len(names) != len(self._row_sample_):
raise ValueError("The length of column names and columns are not the same.")
super().__init__(key_values)
self.names = names
value_sample = super().__getitem__(self._row_sample_)
if isinstance(value_sample, Table):
self.columns = TableColumns(
names=names, children_names=value_sample.names, table=self
)
self.children_names = value_sample.names
else:
if _children_names_ is None:
self.children_names = []
self.columns = TableColumns(names=names, children_names=[], table=self)
else:
self.children_names = _children_names_
self.columns = TableColumns(
names=names, children_names=_children_names_, table=self
)
def __missing__(self, key):
return None
def __getitem__(self, args):
"""Override the dict by converting the
comma separated arguments to RowKey
"""
# This is faster than isinstance
# We are sure there is not any inheritance
# to deal with
if type(args) is RowKey:
return super().__getitem__(args)
if self.columns.size == 1:
key = self.columns.to_key(args)
else:
key = self.columns.to_key(*args)
return super().__getitem__(key)
def _check_keys_consistencies_(self):
# We suppose each column is positioned
# in a fix place of the n-tuple.
# Therefore, the levels of the column can be
# found by iterating over each tuple's item
# Convert each features line to tuple
first_row_types = [type(item) for item in self._row_sample_]
for row in self.keys():
# compair length
if len(row) != self.columns.size:
raise ValueError("The length of the 'factors' are not consistence.")
# compair row's elements type
comparisions = [
isinstance(element, type_1)
for element, type_1 in zip(row, first_row_types)
]
if not all(comparisions):
raise ValueError("The types of the 'factors' are not consistence.")
def to_2d_array(self):
"""Convert the distribution ( or the self._counter's
key:value) to a 2D numpy array where the array
rows are [[(RV_1, RV_2, ..., RV_n, count)],[...
Returns:
numpy ndarray:
A 2D numpy array that the its last column
is the counts.
"""
return np.array([k + (v,) for k, v in self.items()], dtype=np.object)
def _product_(self, right):
"""Multiply two Tables.
Args:
right ([type]): [description]
Raises:
ValueError: [description]
Returns:
[type]: [description]
"""
if not isinstance(right, Table):
raise ValueError("The 'right' argument must be a Table.")
# Find common variables
# reorder commons based on their order in left_common_indices
commons = [
name for name in self.names if name in (set(self.names) & set(right.names))
]
# When there is no common variable, it is just a simple product
if len(commons) == 0:
names = np.r_[self.names, right.names]
return (
{
k1 + k2: v1 * v2
for k1, v1 in self.items()
for k2, v2 in right.items()
},
names,
)
# In the case that there is one or more common variables,
# the operation is similar to SQL inner join
# So, create a lookup for the left table, by using the
# common variables as key.
left_common_indices = [
i for i, name in enumerate(self.names) if name in commons
]
# the order in right must be the same as the left
# so we reorder the indices base on its left order
right_common_indices = [
i
for name in commons
for i, name2 in enumerate(right.names)
if name == name2
]
right_complement_indices = [
i for i, name in enumerate(right.names) if name not in commons
]
# Methods to split the keys
def l_comm(key):
return tuple([key[i] for i in left_common_indices])
def r_comm(key):
return tuple([key[i] for i in right_common_indices])
def r_comp(key):
return tuple([key[i] for i in right_complement_indices])
# left and right tables lookup
# left : (key:value) == (common_key: (left_key, left_value))
left_lookup = {}
for k, value in self.items():
comm = l_comm(k)
if comm in left_lookup:
left_lookup[comm] += [(k, value)]
else:
left_lookup[comm] = [(k, value)]
# right : (key:value) == (common_key: (right_compliment_key, right_value))
right_lookup = {}
for k, value in right.items():
comm = r_comm(k)
if comm in right_lookup:
right_lookup[comm] += [(r_comp(k), value)]
else:
right_lookup[comm] = [(r_comp(k), value)]
# The inner join happens over keys of two dictionaries (left_lookup and
# right_lookup).
prodcut_dict = {}
for comm, l_values in left_lookup.items():
if comm not in right_lookup:
continue
for left_key, left_value in l_values:
for right_comp, right_value in right_lookup[comm]:
# prodcut_dict values must be multiplied.
# prodcut_dict keys are the combination: (left, right_compliment).
prodcut_dict[left_key + right_comp] = left_value * right_value
# names are the combination of [left_names, right_compelements_names]
combined_names = np.r_[
self.names,
[name for name in right.names if name not in commons],
]
return (prodcut_dict, combined_names)
def marginal(self, *args, normalise=True):
"""Marginal of (group by) the Table over a set of columns.
P(X, Y, Z) -> P(X, Y) or P(X, Z) or P(Y, Z)
Args:
args (list):
List of column names to marginalised.
Raises:
ValueError:
Raises when one of the column names is
not defined.
Or raises when requested for all column names.
Returns:
Table: (rows, names).
"""
# check the validity of operation based on column names
if len(args) == self.columns.size:
raise ValueError("Cannot marginalize on all column names.")
# split columns to indices and comp_indices
columns_info = self.columns.split_columns(*args)
#
# Convert the key:values to 2D numpy array
# the array rows are (row, value)
arr = self.to_2d_array()
# filter the compliment columns
filtered_arr = np.c_[arr[:, columns_info.complimnet_indices], arr[:, -1]]
# split the 2d array's rows to a tuple of
# compliment columns (row[comp_indices])
# and count row[-1]
arr_gen = ((RowKey(row[:-1]), row[-1]) for row in filtered_arr)
# Before calling the groupby, we have to sort the generator
# by the tuple of compliment columns (index zero in itemgetter)
sorted_arr = sorted(arr_gen, key=itemgetter(0))
# since the values in each 'group' are
# (compliment columns, value)
# here we group by 'compliment columns' and apply
# the sum on the value. Then the dictionary of
# compliment columns:op_of_values
# is an acceptable argument for Table
grouped_arr = {
k: sum([item[1] for item in g])
for k, g in groupby(sorted_arr, key=itemgetter(0))
}
table = Table(grouped_arr, columns_info.complimnet_names, _internal_=True)
if normalise:
table.normalise()
return table
def condition_on(self, *args, normalise=True):
"""Creates the conditional based on
the provided names of columns.
P(X, Y) -> P(X | Y) or P(Y | X)
Args:
args (list):
List of names of provided random
variables.
Raises:
ValueError:
If the provided RV names do not exist
in the distribution.
Returns:
MultiTable
"""
if self.columns.size == 1:
raise ValueError("This is a single column Table and cannot condition on.")
if len(args) == self.columns.size:
raise ValueError("Cannot condition on all columns.")
# split columns to indices and comp_indices
columns_info = self.columns.split_columns(*args)
# Convert the key:value to 2D numpy array
# the array rows are (rows, value)
arr = self.to_2d_array()
# divide the 2d array's rows to a tuple of columns,
# (row[indices]), compliment columns (row[comp_indices])
# and values row[-1]
arr_gen = (
(
RowKey(row[columns_info.indices]),
RowKey(row[columns_info.complimnet_indices]),
row[-1],
)
for row in arr
)
# Before calling the groupby, we have to sort the generator
# by the tuple of columns (index zero in itemgetter)
# And since later we will call the group by on group,
# for each key we do the inner sort too (index one in itemgetter)
sorted_arr = sorted(arr_gen, key=itemgetter(0, 1))
# This method convert a group to a dictionary
def make_dict(group):
# since the values in 'group' argument are
# (columns, compliment columns, value)
# here we group by 'compliment columns' and sum
# the values.
return {
k: sum([item[2] for item in g2])
for k, g2 in groupby(group, key=itemgetter(1))
}
# For each group (belongs a unique values), we create
# a dictionary in a dictionary comprehension
grouped_arr = {
k: make_dict(g) for k, g in groupby(sorted_arr, key=itemgetter(0))
}
# The above dictionary is dictionary of dictionaries
# # the first set of names is for parent dictionary
# and the second set is for children
table = MultiTable(
{
key: Table(values, columns_info.complimnet_names, _internal_=True)
for key, values in grouped_arr.items()
},
columns_info.indices_names,
)
if normalise:
table.normalise()
return table
def reduce(self, **kwargs):
"""Reduce the Table by one or more columns.
P(X, Y) -> P(X = x, Y) or P(X, Y = y)
Args:
kwargs (dict):
A dictionary that its 'key' is the name
of the column and its 'value'
is the value that must be reduced by.
Raises:
ValueError:
If the provided names do not exist in the Table.
Returns:
[Table]: A reduce Table.
"""
# split columns to indices and comp_indices
columns = list(kwargs.keys())
if len(columns) == self.columns.size:
raise ValueError("Cannot reduce on all column names.")
columns_info = self.columns.split_columns(*columns)
values = np.array([value for _, value in kwargs.items()], dtype=np.object)
#
# Convert the key:values to 2D numpy array
# the array rows are (keys, value)
arr_counter = self.to_2d_array()
# filter the 2d array rows by provided values of the reduce
# conditioned_arr is a boolean one, and filtering happens
# in the second line
conditioned_arr = np.all(arr_counter[:, columns_info.indices] == values, axis=1)
sliced_arr = arr_counter[conditioned_arr, :]
# filter the 2d array columns (the compliment columns)
# plus the value column (which is the last column)
sliced_arr = sliced_arr[:, columns_info.complimnet_indices + [-1]]
# divide the 2d array's rows to a tuple of columns
# and value
# So, we make a generator that divide the rows to the tuple of
# columns (tuple(row[:-1]) and value (row[-1])
arr_gen = ((RowKey(row[:-1]), row[-1]) for row in sliced_arr)
# Before calling the groupby, we have to sort the generator
# by the tuple of column (index zero in itemgetter)
sorted_slice_arr = sorted(arr_gen, key=itemgetter(0))
# group by the filtered columns (compliment
# columns) and sum the value per key
# Note that the 'itemgetter' read the first index which
# is the tuple of compliment columns
return Table(
{
k: sum([item[1] for item in g])
for k, g in groupby(sorted_slice_arr, key=itemgetter(0))
},
columns_info.complimnet_names,
_internal_=True,
)
def get(self, *args, **kwargs):
key = self.columns.to_key(*args, **kwargs)
return super().__getitem__(key)
def to_table(self, sort=False, value_title=""):
arr = self.to_2d_array().astype("U")
arr_len = np.apply_along_axis(lambda row: [len(item) for item in row], 0, arr)
max_levels_len = np.max(arr_len[:, :-1], axis=0)
max_freq_len = max(np.max(arr_len[:, -1]), len(value_title))
def padding(max_len):
def str_padding(value):
return "".join([" "] * (max_len - len(str(value))))
return str_padding
r_padding = padding(max_freq_len)
if sort: # sort by values
items = reversed(sorted(self.items(), key=lambda item: item[1]))
else: # sort by keys
items = sorted(self.items())
rows = ""
header = ""
horizontal_line = ""
for i, name in enumerate(self.names):
header += f"|{name}{padding(max_levels_len[i])(name)}"
horizontal_line += "|" + "".join(["-"] * max_levels_len[i])
header += "|" + "".join([" "] * max_freq_len) + "|"
horizontal_line += "|" + "".join(["-"] * max_freq_len) + "|"
for k, value in items:
key_str = ""
for i, k_part in enumerate(k):
key_str += f"|{padding(max_levels_len[i])(k_part)}{k_part}"
freq_padding = r_padding(value)
rows += f"{key_str}|{value}{freq_padding}|\n"
return f"{header}\n{horizontal_line}\n{rows}"
def add(self, that):
"""Combines two FrequencyTable and return
a new one. All the frequencies are sum together.
This is not a mathematical sum.
"""
#############################################
# check the validity of operation based on column names
if not isinstance(that, Table):
raise ValueError("Table can only adds to Table.")
if self.columns.size != that.columns.size:
raise ValueError("Two adding Table do not have the same columns.")
if len(self.children_names) != len(that.children_names):
raise ValueError("Two adding Table do not have the same children columns.")
for i, name in enumerate(self.names):
if name != that.names[i]:
raise ValueError(
"Two adding Table do not have the same columns "
"(order must be the same too)."
)
for i, name in enumerate(self.children_names):
if name != that.children_names[i]:
raise ValueError(
"Two adding Table do not have the same children columns "
"(order must be the same too)."
)
#############################################
# Algorithm
#
def add_internal(this, that, names):
if that is not None:
for key in that.keys():
if key in this:
this[key] += that[key]
else:
this[key] = that[key]
return Table(this, names=names, _internal_=True)
############################################
# MultiTable handeling
if self.columns.is_multitable():
return Table(
{
k: add_internal(table.copy(), that[k], self.children_names)
for k, table in self.items()
},
self.names,
_internal_=True,
)
return add_internal(self.copy(), that, self.names)
def total(self):
if self.columns.is_multitable():
return {k: table.total() for k, table in self.items()}
return sum(self.values())
def normalise(self):
if self.columns.is_multitable():
for k, total in self.total().items():
if total == 0:
continue
table = self[k]
for k2 in table:
table[k2] /= total
else:
total = self.total()
if total != 0:
for k in self.keys():
self[k] /= total
def __mul__(self, right):
"""Multiplies a table with this one.
P(X, Y) * k -> P(X, Y)
P(X) * P(Y, Z) -> P(X, Y, Z)
Args:
right ([type]): [description]
Raises:
ValueError: [description]
Returns:
[type]: [description]
"""
if not isinstance(right, Table):
raise ValueError("The 'right' argument must be a 'Table'.")
(rows, names) = self._product_(right)
return Table(rows, names, _internal_=True)
def __rmul__(self, left):
"""Multiplies a table with this one.
k * P(X, Y) -> P(X, Y)
P(X) * P(Y, Z) -> P(X, Y, Z)
Args:
right ([type]): [description]
Raises:
ValueError: [description]
Returns:
[type]: [description]
"""
if not isinstance(left, Table):
raise ValueError("The 'right' argument must be a 'Table'.")
(rows, names) = left._product_(self)
return Table(rows, names, _internal_=True)
def __add__(self, right):
return self.add(right)
def prod_right(table, key2, value2):
# Product a table with kay and value
if value2 is None:
return {}
return {key1 + key2: value1 * value2 for key1, value1 in table.items()}
def prod_left(table, key2, value2):
# Product a table with kay and value
if value2 is None:
return {}
return {key2 + key1: value1 * value2 for key1, value1 in table.items()}
def multi_table_to_table_product(left, right, all_ordered_names):
"""Multiply two tables.
P(X, Y | Z) * P(Z) -> P(X, Y, Z)
P(X, Y | Z, W) * P(Z) -> P(X, Y, Z | W)
"""
# Case P(X, Y | Z) * P(Z) -> P(X, Y, Z)
if list(left.names) == list(right.names):
return Table(
ChainMap(
*[
prod_right(table, key2=k, value2=right[k])
for k, table in left.items()
]
),
left.columns.children_names + left.names,
_internal_=True,
)
# Case P(X, Y | Z, W) * P(Z) -> P(X, Y, Z | W)
for name in right.names:
if not left.columns:
raise ValueError(
f"Column name '{name}'in right table is not defined on "
"conditioned columns of the left Table (name mismatched)."
)
# e.g. P(X, Y | Z, W) * P(Z) : indices of [W]
indices = [i for i, name in enumerate(left.names) if name not in right.names]
# e.g. P(X, Y | Z, W) * P(Z) : indices of [Z]
compliment_indices = [i for i in range(left.columns.size) if i not in indices]
# e.g. P(X, Y | Z, W) * P(Z) : [W]
reduced_names = [left.names[i] for i in indices]
children_names = [
names for names in all_ordered_names if names not in reduced_names
]
def reduced_key(key):
# Method to split the keys
return {left.names[i]: key[i] for i in indices}
def compliment_key(key):
# Method to make a split key
return RowKey(*[key[i] for i in compliment_indices])
# Case: P(X, Y | Z, W) * P(Z) -> P(X, Y, Z | W)
if right.columns.size == len(indices):
return MultiTable(
ChainMap(
*[
prod_right(table, key2=k, value2=right[k])
for k, table in left.items()
]
),
reduced_names,
_children_names_=children_names,
)
return MultiTable(
{
compliment_key(k): table * right.reduce(**reduced_key(k))
for k, table in left.items()
},
reduced_names,
_children_names_=children_names,
)
def table_to_multi_table_product(left, right, all_ordered_names):
"""Multiply two tables.
P(Z) * P(X, Y | Z) -> P(Z, X, Y)
P(Z) * P(X, Y | Z, W) -> P(Z, X, Y | W)
"""
# Case P(Z) * P(X, Y | Z) -> P(Z, X, Y)
if list(left.names) == list(right.names):
return Table(
ChainMap(
*[
prod_left(table, key2=k, value2=left[k])
for k, table in right.items()
]
),
right.names + right.columns.children_names,
_internal_=True,
)
# Case P(Z) * P(X, Y | Z, W) -> P(Z, X, Y | W)
for name in left.names:
if not right.columns:
raise ValueError(
f"Column name '{name}'in left table is not defined on "
"conditioned columns of the right Table (name mismatched)."
)
# e.g. P(Z) * P(X, Y | Z, W) : indices of [W]
indices = [i for i, name in enumerate(right.names) if name not in left.names]
# e.g. P(Z) * P(X, Y | Z, W) : indices of [Z]
compliment_indices = [i for i in range(right.columns.size) if i not in indices]
# e.g. P(Z) * P(X, Y | Z, W) : [W]
reduced_names = [right.names[i] for i in indices]
children_names = [
names for names in all_ordered_names if names not in reduced_names
]
def reduced_key(key):
# Method to split the keys
return {right.names[i]: key[i] for i in indices}
def compliment_key(key):
# Method to make a split key
return RowKey(*[key[i] for i in compliment_indices])
# Case: P(Z) * P(X, Y | Z, W) -> P(Z, X, Y | W)
if left.columns.size == len(indices):
return MultiTable(
ChainMap(
*[
prod_left(table, key2=k, value2=left[k])
for k, table in right.items()
]
),
reduced_names,
_children_names_=children_names,
)
return MultiTable(
{
compliment_key(k): table * left.reduce(**reduced_key(k))
for k, table in right.items()
},
reduced_names,
_children_names_=children_names,
)
def multi_table_to_multi_table_product(table_main, table_side, all_ordered_names):
indices = [
i for i, name in enumerate(table_main.names) if name not in table_side.names
]
compliment_indices = [i for i in range(table_main.columns.size) if i not in indices]
reduced_names = [table_main.names[i] for i in compliment_indices]
children_names = [
names for names in all_ordered_names if names not in reduced_names
]
def reduced_key(key):
# Method to split the keys
return {table_main.names[i]: key[i] for i in indices}
def compliment_key(key):
# Method to split the keys
return RowKey(*[key[i] for i in compliment_indices])
if len(table_side.columns.children_names) == len(indices):
def prod2(key1, table1):
table_side_table2 = table_side[key1]
if table_side_table2 is None:
return {}
return {
compliment_key(key1): table1 * table2
for key2, table2 in table_side_table2
}
return MultiTable(
ChainMap(*[prod2(key1, table1) for key1, table1 in table_main.items()]),
reduced_names,
_children_names_=children_names,
)
return MultiTable(
{
compliment_key(key1): table1 * table2
for key1, table1 in table_main.items()
for key2, table2 in table_side.reduce(**reduced_key(key1))
},
reduced_names,
_children_names_=children_names,
)
def multi_table_product(left, right):
"""Multiply two tables.
P(X, Y | Z) * P(Z) -> P(X, Y , Z)
P(X, Y | Z, W) * P(Z) -> P(X, Y , Z | W)
P(X, Y | Z, U) * P(Z | U) -> P(X, Y , Z | U)
P(X, Y | Z, U, W) * P(Z | U, W) -> P(X, Y , Z | U, W)
in the case of two conditionals, the longer one defines
the order of variables
e.g.
P(X, Y | Z, U, W) * P(Z | W, U) -> P(X, Y , Z | U, W)
P(Z | W, U) * P(X, Y | Z, U, W) -> P(X, Y , Z | U, W)
Args:
left ([type]): [description]
right ([type]): [description]
Raises:
ValueError: [description]
Returns:
[type]: [description]
"""
# Cases:
# P(X, Y | Z) * P(Z) -> P(X, Y, Z)
# P(X, Y | Z, W) * P(Z) -> P(X, Y, Z | W)
if not isinstance(right, MultiTable):
if sorted(right.names) != sorted(left.names):
raise ValueError("The right names is" " not equal to conditionals of left.")
all_ordered_names = left.columns.children_names + right.columns.names
return multi_table_to_table_product(left, right, all_ordered_names)
# Cases:
# P(Z) * P(X, Y | Z) -> P(Z, X, Y)
# P(Z) * P(X, Y | Z, W) -> P(Z, X, Y | W)
if not isinstance(left, MultiTable):
if sorted(right.names) != sorted(left.names):
raise ValueError("The left names is" " not equal to conditionals of right.")
all_ordered_names = left.names + right.columns.children_names
return table_to_multi_table_product(left, right, all_ordered_names)
# Cases:
# P(X, Y | Z, U) * P(Z | U) -> P(X, Y, Z | U)
# P(X, Y | Z, U, W) * P(Z | U, W) -> P(X, Y, Z | U, W)
# P(X, Y, Z| U, W) * P(U | W) -> P(X, Y, Z, U | W
# P(X, Y, Z| U, V, W) * P(U, V | W) -> P(X, Y, Z, U, V | W)
def in_the_other(first, second):
for name in first:
if name not in second:
return False
return True
common_conditions = [name for name in left.names if name in right.names]
right_compliment_conditions = [
name for name in right.names if name not in common_conditions
]
left_compliment_conditions = [
name for name in left.names if name not in common_conditions
]
# To check the crossed cases
# e.g. P(X | Y) * P(Y | X)
# after removing common names on conditionals,
# one of them must remains conditionless
# e.g.
# 1) P(X, Y | Z, U) * P(Z | U)
# removes commons: P(X, Y | Z) * P(Z)
# 2) P(Z | U, W) * P(X, Y | Z, U, W)
# removes commons: P(Z) * P(X, Y | Z)
# 3) P(X | Y) * P(Y | X)
# remove commons fails
if len(right_compliment_conditions) == 0:
if not in_the_other(right.columns.children_names, left.names):
raise ValueError(
"Columns in right is not defined in conditional names of left."
)
all_ordered_names = left.columns.children_names + right.columns.children_names
return multi_table_to_multi_table_product(left, right, all_ordered_names)
elif len(left_compliment_conditions) == 0:
if not in_the_other(left.columns.children_names, right.names):
raise ValueError(
"Columns in left is not defined in conditional names of right."
)
all_ordered_names = left.columns.children_names + right.columns.children_names
return multi_table_to_multi_table_product(right, left, all_ordered_names)
else:
raise ValueError("Columns and conditional names mismatch.")
class MultiTable(Table):
def __init__(self, rows, names=None, _children_names_=None):
super().__init__(
rows, names, _internal_=True, _children_names_=_children_names_
)
def marginal(self, *args, normalise=True):
"""[summary]
P(X, Y | Z) -> P(X | Z) or P(Y | Z)
Args:
normalise (bool, optional): [description]. Defaults to True.
Raises:
ValueError: [description]
Returns:
MultiTable: [description]
"""
for name in args:
if name in self.names:
raise ValueError(f"Cannot marginalize on conditioned columns:'{name}'.")
table = Table(
{
k: table.marginal(*args, normalise=normalise)
for k, table in self.items()
},
self.names,
_internal_=True,
)
if normalise:
table.normalise()
return table
def condition_on(self, *args, normalise=True):
"""Creates the conditional based on
the provided names of columns.
P(X, Y | Z) -> P(X | Y, Z) or P(Y | X, Z)
Args:
args (list):
List of names of provided random
variables.
Raises:
ValueError:
If the provided RV names do not exist
in the distribution.
Returns:
(row, names)
"""
for name in args:
if name in self.names:
raise ValueError(f"Cannot condition on conditioned columns:'{name}'.")
conditioned_children = (
(k, table.condition_on(*args, normalise=normalise))
for k, table in self.items()
)
return MultiTable(
{
key2 + key1: table
for key1, key2_table in conditioned_children
for key2, table in key2_table.items()
},
# It results in: P(X, Y | Z) -> P(X | Y, Z)
# inversing the order turns it P(X, Y | Z) -> P(X | Z, Y)
# Maybe more controls is needed here
list(args) + self.names,
)
def reduce(self, **kwargs):
"""Reduce the Table by one or more columns.
P(X, Y | Z) -> P(X = x, Y | Z) or P(X, Y = y | Z)
Args:
kwargs (dict):
A dictionary that its 'key' is the name
of the column and its 'value'
is the value that must be reduced by.
Raises:
ValueError:
If the provided names do not exist in the Table.
Returns:
[Table]: A reduce Table.
"""
return MultiTable(
{k: table.reduce(**kwargs) for k, table in self.items()},
self.names,
)
def __mul__(self, right):
if not isinstance(right, Table):
raise ValueError("The 'right' argument must be a 'Table'.")
return multi_table_product(self, right)
def __rmul__(self, left):
if not isinstance(left, Table):
raise ValueError("The 'left' argument must be a 'Table'.")
return multi_table_product(left, self)
|
[
"probability.RowKey",
"numpy.max",
"operator.itemgetter",
"numpy.all",
"probability.TableColumns"
] |
[((14467, 14529), 'numpy.all', 'np.all', (['(arr_counter[:, columns_info.indices] == values)'], {'axis': '(1)'}), '(arr_counter[:, columns_info.indices] == values, axis=1)\n', (14473, 14529), True, 'import numpy as np\n'), ((16044, 16075), 'numpy.max', 'np.max', (['arr_len[:, :-1]'], {'axis': '(0)'}), '(arr_len[:, :-1], axis=0)\n', (16050, 16075), True, 'import numpy as np\n'), ((23022, 23067), 'probability.RowKey', 'RowKey', (['*[key[i] for i in compliment_indices]'], {}), '(*[key[i] for i in compliment_indices])\n', (23028, 23067), False, 'from probability import RowKey\n'), ((25228, 25273), 'probability.RowKey', 'RowKey', (['*[key[i] for i in compliment_indices]'], {}), '(*[key[i] for i in compliment_indices])\n', (25234, 25273), False, 'from probability import RowKey\n'), ((26548, 26593), 'probability.RowKey', 'RowKey', (['*[key[i] for i in compliment_indices]'], {}), '(*[key[i] for i in compliment_indices])\n', (26554, 26593), False, 'from probability import RowKey\n'), ((2490, 2562), 'probability.TableColumns', 'TableColumns', ([], {'names': 'names', 'children_names': 'value_sample.names', 'table': 'self'}), '(names=names, children_names=value_sample.names, table=self)\n', (2502, 2562), False, 'from probability import TableColumns\n'), ((16104, 16126), 'numpy.max', 'np.max', (['arr_len[:, -1]'], {}), '(arr_len[:, -1])\n', (16110, 16126), True, 'import numpy as np\n'), ((2773, 2829), 'probability.TableColumns', 'TableColumns', ([], {'names': 'names', 'children_names': '[]', 'table': 'self'}), '(names=names, children_names=[], table=self)\n', (2785, 2829), False, 'from probability import TableColumns\n'), ((2934, 3004), 'probability.TableColumns', 'TableColumns', ([], {'names': 'names', 'children_names': '_children_names_', 'table': 'self'}), '(names=names, children_names=_children_names_, table=self)\n', (2946, 3004), False, 'from probability import TableColumns\n'), ((9663, 9679), 'probability.RowKey', 'RowKey', (['row[:-1]'], {}), '(row[:-1])\n', (9669, 9679), False, 'from probability import RowKey\n'), ((9896, 9909), 'operator.itemgetter', 'itemgetter', (['(0)'], {}), '(0)\n', (9906, 9909), False, 'from operator import itemgetter\n'), ((11646, 11679), 'probability.RowKey', 'RowKey', (['row[columns_info.indices]'], {}), '(row[columns_info.indices])\n', (11652, 11679), False, 'from probability import RowKey\n'), ((11697, 11741), 'probability.RowKey', 'RowKey', (['row[columns_info.complimnet_indices]'], {}), '(row[columns_info.complimnet_indices])\n', (11703, 11741), False, 'from probability import RowKey\n'), ((12125, 12141), 'operator.itemgetter', 'itemgetter', (['(0)', '(1)'], {}), '(0, 1)\n', (12135, 12141), False, 'from operator import itemgetter\n'), ((15005, 15021), 'probability.RowKey', 'RowKey', (['row[:-1]'], {}), '(row[:-1])\n', (15011, 15021), False, 'from probability import RowKey\n'), ((15230, 15243), 'operator.itemgetter', 'itemgetter', (['(0)'], {}), '(0)\n', (15240, 15243), False, 'from operator import itemgetter\n'), ((647, 672), 'operator.itemgetter', 'itemgetter', (['groupby_index'], {}), '(groupby_index)\n', (657, 672), False, 'from operator import itemgetter\n'), ((1300, 1356), 'probability.TableColumns', 'TableColumns', ([], {'names': 'names', 'children_names': '[]', 'table': 'self'}), '(names=names, children_names=[], table=self)\n', (1312, 1356), False, 'from probability import TableColumns\n'), ((1519, 1589), 'probability.TableColumns', 'TableColumns', ([], {'names': 'names', 'children_names': '_children_names_', 'table': 'self'}), '(names=names, children_names=_children_names_, table=self)\n', (1531, 1589), False, 'from probability import TableColumns\n'), ((1746, 1755), 'probability.RowKey', 'RowKey', (['k'], {}), '(k)\n', (1752, 1755), False, 'from probability import RowKey\n'), ((10313, 10326), 'operator.itemgetter', 'itemgetter', (['(0)'], {}), '(0)\n', (10323, 10326), False, 'from operator import itemgetter\n'), ((12771, 12784), 'operator.itemgetter', 'itemgetter', (['(0)'], {}), '(0)\n', (12781, 12784), False, 'from operator import itemgetter\n'), ((1870, 1879), 'probability.RowKey', 'RowKey', (['k'], {}), '(k)\n', (1876, 1879), False, 'from probability import RowKey\n'), ((12538, 12551), 'operator.itemgetter', 'itemgetter', (['(1)'], {}), '(1)\n', (12548, 12551), False, 'from operator import itemgetter\n'), ((15593, 15606), 'operator.itemgetter', 'itemgetter', (['(0)'], {}), '(0)\n', (15603, 15606), False, 'from operator import itemgetter\n')]
|
# Copyright (C) 2020-2021 Intel Corporation
# SPDX-License-Identifier: Apache-2.0
from copy import deepcopy
from functools import partial
import numpy as np
import scipy
from addict import Dict
from ....algorithms.quantization import utils as eu
from ....engines.ac_engine import ACEngine
from ....graph.model_utils import get_nodes_by_type
from ....graph.node_utils import get_all_node_outputs
from ....graph.utils import find_operation_matches
from ....samplers.creator import create_sampler
SPECIAL_METRICS = ['cmc', 'reid_map', 'pairwise_accuracy_subsets', 'pairwise_accuracy', 'normalized_embedding_accuracy',
'face_recognition_tafa_pair_metric', 'localization_recall',
'coco_orig_keypoints_precision', 'coco_orig_segm_precision', 'coco_orig_keypoints_precision']
METRICS_CONFIGS = {'sigmoid_recom_loss': {'metrics': 'log_loss',
'postprocessing': 'sigmoid_normalize_recommendation'},
'coco_precision': {'metrics': 'coco_precision'},
'coco_segm_precision': {'metrics': 'coco_segm_precision'}}
METRIC2PROXY_METRIC = {
'hit_ratio':
{
'persample': 'sigmoid_recom_loss',
'ranking': 'sigmoid_recom_loss'
},
'ndcg':
{
'persample': 'sigmoid_recom_loss',
'ranking': 'sigmoid_recom_loss'
},
'coco_orig_precision':
{
'persample': 'coco_precision'
},
'coco_orig_keypoints_precision':
{
'persample': 'coco_precision'
},
'coco_orig_segm_precision':
{
'persample': 'coco_segm_precision'
}
}
def create_metric_config(engine, algo_config: Dict, force_logit_comparison=False,
logit_distance_type='cosine') -> Dict:
def create_metric_params(metric_name):
engine_metrics_attributes = engine.get_metrics_attributes()
if metric_name not in engine_metrics_attributes:
RuntimeError('Couldn\'t create metric parameters. '
'Metric {} not registered in the engine.'.format(metric_name))
params = Dict()
params.name = metric_name
params.type = engine_metrics_attributes[metric_name]['type']
params.is_special = (params.type in SPECIAL_METRICS) or force_logit_comparison
if engine_metrics_attributes[metric_name]['direction'] == 'higher-better':
params.comparator = (lambda a: a)
elif engine_metrics_attributes[metric_name]['direction'] == 'higher-worse':
params.comparator = (lambda a: -a)
else:
raise ValueError('Unexpected {} metric direction value.'.format(metric_name))
params.sort_fn = partial(sort_by_logit_distance, distance=logit_distance_type) \
if params.is_special else partial(sort_by_metric_difference, comp_fn=params.comparator)
return params
def metric_to_proxy_map(metrics):
"""Determines which metrics need proxy metrics and creates metrics to proxy metrics map.
:param metrics: optimizable metrics names
:returns a dictionary of metrics to proxy metrics mapping {metric_name: 'persample': proxy_name,
'ranking': proxy_name}
a list of proxy metrics names to register
"""
def update_proxy_list(proxy_metric_name):
"""Updates a list of proxy metrics names to register.
:return a proxy metric name in accordance with the engine naming
"""
proxy_config = METRICS_CONFIGS.get(proxy_metric_name, {})
metric_config = proxy_config.get('metrics')
postprocessing_config = proxy_config.get('postprocessing')
if metric_config or postprocessing_config:
to_register.add(proxy_metric_name)
return metric_name_from_config(metric_config)
match_names_config = Dict({metric_name: {} for metric_name in metrics})
to_register = set()
for metric_name, metric_type in metrics:
if metric_type in METRIC2PROXY_METRIC:
persample_metric_name = METRIC2PROXY_METRIC[metric_type].get('persample')
persample_proxy_metric_name = update_proxy_list(persample_metric_name)
if persample_proxy_metric_name:
match_names_config[metric_name].persample = persample_proxy_metric_name
ranking_metric_name = METRIC2PROXY_METRIC[metric_type].get('ranking')
ranking_proxy_metric_name = update_proxy_list(ranking_metric_name)
if ranking_proxy_metric_name:
match_names_config[metric_name].ranking = ranking_proxy_metric_name
return match_names_config, list(to_register)
metrics_attributes = engine.get_metrics_attributes()
# configure which metrics to optimize
if algo_config.metrics:
metrics_names = []
for metric in algo_config.metrics:
metric_type = metric.type if metric.type else metric.name
metrics_names.append((metric.name, metric_type))
else:
metrics_names = [(metric_name, metric_attr.get('type', metric_name)) for metric_name, metric_attr
in metrics_attributes.items()]
# register proxy metrics
metrics_to_proxy_map, metrics_to_register = metric_to_proxy_map(metrics_names)
register_metrics(engine, metrics_to_register)
metrics_config = Dict()
for metric, _ in metrics_names:
persample_name = metrics_to_proxy_map[metric].get('persample', metric)
ranking_name = metrics_to_proxy_map[metric].get('ranking', metric)
metrics_config[metric].persample = create_metric_params(persample_name)
metrics_config[metric].ranking = create_metric_params(ranking_name)
metrics_config[metric].update(create_metric_params(metric))
return metrics_config
def metric_name_from_config(metric_config):
if isinstance(metric_config, str):
return metric_config
if isinstance(metric_config, dict):
return metric_config.get('name', metric_config['type'])
return None
def register_metrics(engine, metrics_names: list):
"""Registers metrics and postprocessing in the engine.
:param engine: an engine in which metrics will be registered
:param metrics_names: a list of metrics names
"""
registered_metrics = engine.get_metrics_attributes()
for metric in metrics_names:
if metric not in METRICS_CONFIGS:
raise ValueError('Cannot register metric. Unsupported name {}.'.format(metric))
proxy_config = METRICS_CONFIGS.get(metric, {})
if 'metrics' in proxy_config:
metric_config = proxy_config['metrics']
if metric_name_from_config(metric_config) not in registered_metrics:
register_metric(engine, metric_config)
if 'postprocessing' in proxy_config:
postprocessing_config = proxy_config['postprocessing']
register_postprocessing(engine, postprocessing_config)
def sort_by_logit_distance(u, v, reverse=False, distance='cosine'):
if len(u) != len(v):
raise RuntimeError('Cannot compare samples. '
'Lists of per-sample metric results should be the same length.')
kd_distance = lambda u, v: scipy.stats.entropy(scipy.special.softmax(u),
scipy.special.softmax(v))
mse_distance = lambda u, v: np.mean((u - v) ** 2)
distance_function = {
'cosine': scipy.spatial.distance.cosine,
'kd': kd_distance,
'mse': mse_distance,
}
distance_between_samples = np.array([distance_function[distance](ui.flatten(), vi.flatten())
for ui, vi in zip(u, v)])
sorted_arr = np.argsort(distance_between_samples)
if reverse:
sorted_arr = np.flip(sorted_arr)
return sorted_arr
def sort_by_metric_difference(u, v, comp_fn=lambda a: a, reverse=False):
if len(u) != len(v):
raise RuntimeError('Cannot compare samples. '
'Lists of per-sample metric results should be the same length.')
u = np.asarray(u)
v = np.asarray(v)
sorted_arr = np.argsort(comp_fn(u - v))
if reverse:
sorted_arr = np.flip(sorted_arr)
return sorted_arr
def register_metric(engine, metric_config):
if isinstance(engine, ACEngine):
engine.add_metric(metric_config)
else:
raise NotImplementedError('{} engine cannot register new metrics.'
.format(type(engine).__name__))
def register_postprocessing(engine, postprocessing_config):
if isinstance(engine, ACEngine):
engine.add_postprocessing(postprocessing_config)
else:
raise NotImplementedError('{} engine cannot register new postprocessing.'
.format(type(engine).__name__))
def is_preset_performance(config: Dict):
if config.weights.mode == 'symmetric' and config.activations.mode == 'symmetric':
return True
if config.weights.mode == 'asymmetric' or config.activations.mode == 'asymmetric':
return False
if config.preset == 'performance':
return True
return False
def get_mixed_preset_config(config: Dict):
config = deepcopy(config)
config.update(preset='mixed')
if config.activations.mode:
config.activations.mode = 'asymmetric'
if config.weights.mode:
config.weights.mode = 'symmetric'
return config
def get_num_of_quantized_ops(model, quantizable_operations):
quantized_ops = set()
nodes_to_see = []
for fq_node in get_nodes_by_type(model, ['FakeQuantize']):
nodes_to_see.extend(get_all_node_outputs(fq_node))
while nodes_to_see:
child = nodes_to_see.pop()
if find_operation_matches(quantizable_operations, child):
quantized_ops.add(child)
continue
nodes_to_see.extend(get_all_node_outputs(child))
return len(quantized_ops)
def evaluate_model(
model, engine,
dataset_size,
subset_indices=None,
print_progress=True,
metrics_config=None,
per_sample_subset_indices=None,
output_node_name=None,
stats_layout=None,
):
"""Evaluates the model and processes metrics values
:param model: model to evaluate
:param subset_indices: image indices to evaluate on. If None evaluate on whole dataset
:param per_sample_subset_indices: image indices for which to return per-sample metrics.
If None for all predicted images
:param print_progress: Whether to print inference progress
:returns a dictionary of predicted metrics {metric_name: value}
a dictionary of per-sample metrics values {metric_name: [values]}
"""
engine.set_model(model)
eu.select_evaluation_dataset(engine)
if not subset_indices:
subset_indices = range(dataset_size)
index_sampler = create_sampler(engine, samples=subset_indices)
(metrics_per_sample, metrics), raw_output = engine.predict(stats_layout=stats_layout,
sampler=index_sampler,
metric_per_sample=True,
print_progress=print_progress)
raw_output = process_raw_output(raw_output, output_node_name)
metrics_per_sample = process_per_sample_metrics(metrics_per_sample,
metrics_config,
per_sample_subset_indices,
raw_output=raw_output)
metrics = dict((name, value) for name, value in metrics.items() if name in metrics_config)
eu.reset_dataset_to_default(engine)
return metrics, metrics_per_sample
def process_raw_output(output, output_node_name):
if not output:
return []
return output[output_node_name]['output_logits']
def process_per_sample_metrics(metrics_per_sample, metrics_config,
indices=None, raw_output=None):
"""Creates a dictionary of per-sample metrics values {metric_name: [values]}
:param metrics_per_sample: list of per-sample metrics
:param indices: indices of samples to be considered. All if None
:param raw_output: raw output from the model
:return processed dictionary
"""
metrics_to_keep = {config.persample.name: config.persample
for config in metrics_config.values()}
if not metrics_to_keep:
return {}
processed_metrics_per_sample = dict((name, []) for name in metrics_to_keep)
for metric_name, metric_params in metrics_to_keep.items():
if metric_params.is_special:
processed_metrics_per_sample[metric_name] = raw_output
for value in metrics_per_sample:
if value['metric_name'] in metrics_to_keep:
if metrics_to_keep[value['metric_name']].is_special:
continue
if value['result'] is not None:
result_value = np.nanmean(value['result'])
else:
result_value = None
processed_metrics_per_sample[value['metric_name']].append(result_value)
# check that all metrics have equal number of samples
if not len({len(value) for value in processed_metrics_per_sample.values()}) == 1:
raise RuntimeError('Inconsistent number of per-sample metric values')
if indices:
for name, values in processed_metrics_per_sample.items():
processed_metrics_per_sample[name] = [values[i] for i in indices]
return processed_metrics_per_sample
|
[
"addict.Dict",
"numpy.mean",
"numpy.flip",
"numpy.asarray",
"numpy.argsort",
"numpy.nanmean",
"functools.partial",
"copy.deepcopy",
"scipy.special.softmax"
] |
[((5558, 5564), 'addict.Dict', 'Dict', ([], {}), '()\n', (5562, 5564), False, 'from addict import Dict\n'), ((7929, 7965), 'numpy.argsort', 'np.argsort', (['distance_between_samples'], {}), '(distance_between_samples)\n', (7939, 7965), True, 'import numpy as np\n'), ((8299, 8312), 'numpy.asarray', 'np.asarray', (['u'], {}), '(u)\n', (8309, 8312), True, 'import numpy as np\n'), ((8321, 8334), 'numpy.asarray', 'np.asarray', (['v'], {}), '(v)\n', (8331, 8334), True, 'import numpy as np\n'), ((9438, 9454), 'copy.deepcopy', 'deepcopy', (['config'], {}), '(config)\n', (9446, 9454), False, 'from copy import deepcopy\n'), ((2174, 2180), 'addict.Dict', 'Dict', ([], {}), '()\n', (2178, 2180), False, 'from addict import Dict\n'), ((4016, 4066), 'addict.Dict', 'Dict', (['{metric_name: {} for metric_name in metrics}'], {}), '({metric_name: {} for metric_name in metrics})\n', (4020, 4066), False, 'from addict import Dict\n'), ((7587, 7608), 'numpy.mean', 'np.mean', (['((u - v) ** 2)'], {}), '((u - v) ** 2)\n', (7594, 7608), True, 'import numpy as np\n'), ((8003, 8022), 'numpy.flip', 'np.flip', (['sorted_arr'], {}), '(sorted_arr)\n', (8010, 8022), True, 'import numpy as np\n'), ((8416, 8435), 'numpy.flip', 'np.flip', (['sorted_arr'], {}), '(sorted_arr)\n', (8423, 8435), True, 'import numpy as np\n'), ((2763, 2824), 'functools.partial', 'partial', (['sort_by_logit_distance'], {'distance': 'logit_distance_type'}), '(sort_by_logit_distance, distance=logit_distance_type)\n', (2770, 2824), False, 'from functools import partial\n'), ((2865, 2926), 'functools.partial', 'partial', (['sort_by_metric_difference'], {'comp_fn': 'params.comparator'}), '(sort_by_metric_difference, comp_fn=params.comparator)\n', (2872, 2926), False, 'from functools import partial\n'), ((7452, 7476), 'scipy.special.softmax', 'scipy.special.softmax', (['u'], {}), '(u)\n', (7473, 7476), False, 'import scipy\n'), ((7529, 7553), 'scipy.special.softmax', 'scipy.special.softmax', (['v'], {}), '(v)\n', (7550, 7553), False, 'import scipy\n'), ((13422, 13449), 'numpy.nanmean', 'np.nanmean', (["value['result']"], {}), "(value['result'])\n", (13432, 13449), True, 'import numpy as np\n')]
|
from collections import defaultdict
import time
from joblib import Parallel, delayed
from multiprocessing import cpu_count
from math import ceil
import torch
from torch import nn
import torch.multiprocessing as mp
import torch.distributed as dist
from torch.nn.parallel import DistributedDataParallel as DDP
from torch.utils.data import TensorDataset, DataLoader, SequentialSampler
from torch.utils.data.distributed import DistributedSampler
from nltk.corpus import stopwords
from transformers import AdamW, get_linear_schedule_with_warmup
# =================================================================================
#from transformers.models.camembert.tokenization_camembert import CamembertTokenizer
from transformers import BertTokenizer
# =================================================================================
import numpy as np
import os
import shutil
import sys
from tqdm import tqdm
from LOTClass.bert.model import LOTClassModel
import warnings
warnings.filterwarnings("ignore")
class LOTClassTrainer(object):
def __init__(self, args):
self.args = args
self.verbose = args.verbose
self.max_len = args.max_len
self.dataset_dir = args.dataset_dir
self.dist_port = args.dist_port
self.num_cpus = min(10, cpu_count() - 1) if cpu_count() > 1 else 1
self.world_size = args.gpus
self.train_batch_size = args.train_batch_size
self.eval_batch_size = args.eval_batch_size
self.accum_steps = args.accum_steps
eff_batch_size = self.train_batch_size * self.world_size * self.accum_steps
assert abs(eff_batch_size - 128) < 10, f"Make sure the effective training batch size is around 128, current: {eff_batch_size}"
print(f"Effective training batch size: {eff_batch_size}")
self.pretrained_lm = args.pretrained_lm
self.tokenizer = BertTokenizer.from_pretrained(self.pretrained_lm, do_lower_case=True)
#self.tokenizer = CamembertTokenizer.from_pretrained(self.pretrained_lm, force_download=True)
self.vocab = self.tokenizer.get_vocab()
self.vocab_size = len(self.vocab)
self.mask_id = self.vocab[self.tokenizer.mask_token]
self.inv_vocab = {k:v for v, k in self.vocab.items()}
self.read_label_names(args.dataset_dir, args.label_names_file)
self.num_class = len(self.label_name_dict)
self.model = LOTClassModel.from_pretrained(self.pretrained_lm,
output_attentions=False,
output_hidden_states=False,
num_labels=self.num_class)
self.read_data(args.dataset_dir, args.train_file, args.test_file, args.test_label_file)
self.with_test_label = True if args.test_label_file is not None else False
self.temp_dir = f'tmp_{self.dist_port}'
self.mcp_loss = nn.CrossEntropyLoss()
self.st_loss = nn.KLDivLoss(reduction='batchmean')
self.update_interval = args.update_interval
self.early_stop = args.early_stop
# set up distributed training
def set_up_dist(self, rank):
dist.init_process_group(
backend='nccl',
init_method=f'tcp://localhost:{self.dist_port}',
world_size=self.world_size,
rank=rank
)
# create local model
model = self.model.to(rank)
model = DDP(model, device_ids=[rank], find_unused_parameters=True)
return model
# get document truncation statistics with the defined max length
def corpus_trunc_stats(self, docs):
doc_len = []
for doc in docs:
input_ids = self.tokenizer.encode(doc, add_special_tokens=True)
doc_len.append(len(input_ids))
print(f"Document max length: {np.max(doc_len)}, avg length: {np.mean(doc_len)}, std length: {np.std(doc_len)}")
trunc_frac = np.sum(np.array(doc_len) > self.max_len) / len(doc_len)
print(f"Truncated fraction of all documents: {trunc_frac}")
# convert a list of strings to token ids
def encode(self, docs):
encoded_dict = self.tokenizer.batch_encode_plus(docs, add_special_tokens=True, max_length=self.max_len, padding='max_length',
return_attention_mask=True, truncation=True, return_tensors='pt')
input_ids = encoded_dict['input_ids']
if self.verbose:
print(f"input_ids size (from encode): {input_ids.size()}")
print(f"input_ids (from encode): {input_ids}")
attention_masks = encoded_dict['attention_mask']
return input_ids, attention_masks
# convert list of token ids to list of strings
def decode(self, ids):
strings = self.tokenizer.batch_decode(ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)
return strings
# convert dataset into tensors
def create_dataset(self, dataset_dir, text_file, label_file, loader_name, find_label_name=False, label_name_loader_name=None):
loader_file = os.path.join(dataset_dir, loader_name)
if os.path.exists(loader_file):
print(f"Loading encoded texts from {loader_file}")
data = torch.load(loader_file)
else:
print(f"Reading texts from {os.path.join(dataset_dir, text_file)}")
corpus = open(os.path.join(dataset_dir, text_file), encoding="utf-8")
docs = [doc.strip() for doc in corpus.readlines()]
print(f"Converting texts into tensors.")
chunk_size = ceil(len(docs) / self.num_cpus)
chunks = [docs[x:x+chunk_size] for x in range(0, len(docs), chunk_size)]
results = Parallel(n_jobs=self.num_cpus)(delayed(self.encode)(docs=chunk) for chunk in chunks)
input_ids = torch.cat([result[0] for result in results])
print(f"Concatenated input_ids size: {input_ids.size()}")
attention_masks = torch.cat([result[1] for result in results])
print(f"Saving encoded texts into {loader_file}")
if label_file is not None:
print(f"Reading labels from {os.path.join(dataset_dir, label_file)}")
truth = open(os.path.join(dataset_dir, label_file))
labels = [int(label.strip()) for label in truth.readlines()]
labels = torch.tensor(labels)
data = {"input_ids": input_ids, "attention_masks": attention_masks, "labels": labels}
else:
data = {"input_ids": input_ids, "attention_masks": attention_masks}
torch.save(data, loader_file)
if find_label_name:
loader_file = os.path.join(dataset_dir, label_name_loader_name)
if os.path.exists(loader_file):
print(f"Loading texts with label names from {loader_file}")
label_name_data = torch.load(loader_file)
else:
print(f"Reading texts from {os.path.join(dataset_dir, text_file)}")
corpus = open(os.path.join(dataset_dir, text_file), encoding="utf-8")
docs = [doc.strip() for doc in corpus.readlines()]
print("Locating label names in the corpus.")
chunk_size = ceil(len(docs) / self.num_cpus)
chunks = [docs[x:x+chunk_size] for x in range(0, len(docs), chunk_size)]
results = Parallel(n_jobs=self.num_cpus)(delayed(self.label_name_occurrence)(docs=chunk) for chunk in chunks)
input_ids_with_label_name = torch.cat([result[0] for result in results])
attention_masks_with_label_name = torch.cat([result[1] for result in results])
label_name_idx = torch.cat([result[2] for result in results])
print(f"Concatenated input_ids_with_label_name size: {input_ids_with_label_name.size()}")
assert len(input_ids_with_label_name) > 0, "No label names appear in corpus!"
label_name_data = {"input_ids": input_ids_with_label_name, "attention_masks": attention_masks_with_label_name, "labels": label_name_idx}
loader_file = os.path.join(dataset_dir, label_name_loader_name)
print(f"Saving texts with label names into {loader_file}")
torch.save(label_name_data, loader_file)
return data, label_name_data
else:
return data
# find label name indices and replace out-of-vocab label names with [MASK]
def label_name_in_doc(self, doc):
doc = self.tokenizer.tokenize(doc)
if self.verbose:
print(doc)
label_idx = -1 * torch.ones(self.max_len, dtype=torch.long)
new_doc = []
wordpcs = []
idx = 1 # index starts at 1 due to [CLS] token
for i, wordpc in enumerate(doc):
wordpcs.append(wordpc[2:] if wordpc.startswith("##") else wordpc)
if self.verbose:
print(wordpcs)
if idx >= self.max_len - 1: # last index will be [SEP] token
break
if i == len(doc) - 1 or not doc[i+1].startswith("##"):
word = ''.join(wordpcs)
if word in self.label2class:
label_idx[idx] = self.label2class[word]
# replace label names that are not in tokenizer's vocabulary with the [MASK] token
if word not in self.vocab:
wordpcs = [self.tokenizer.mask_token]
new_word = ''.join(wordpcs)
if new_word != self.tokenizer.unk_token:
idx += len(wordpcs)
new_doc.append(new_word)
wordpcs = []
if (label_idx >= 0).any():
return ' '.join(new_doc), label_idx
else:
return None
# find label name occurrences in the corpus
def label_name_occurrence(self, docs):
text_with_label = []
label_name_idx = []
for doc in docs:
result = self.label_name_in_doc(doc)
if result is not None:
text_with_label.append(result[0])
label_name_idx.append(result[1].unsqueeze(0))
if len(text_with_label) > 0:
encoded_dict = self.tokenizer.batch_encode_plus(text_with_label, add_special_tokens=True, max_length=self.max_len,
padding='max_length', return_attention_mask=True, truncation=True, return_tensors='pt')
input_ids_with_label_name = encoded_dict['input_ids']
attention_masks_with_label_name = encoded_dict['attention_mask']
label_name_idx = torch.cat(label_name_idx, dim=0)
else:
input_ids_with_label_name = torch.ones(0, self.max_len, dtype=torch.long)
attention_masks_with_label_name = torch.ones(0, self.max_len, dtype=torch.long)
label_name_idx = torch.ones(0, self.max_len, dtype=torch.long)
return input_ids_with_label_name, attention_masks_with_label_name, label_name_idx
# read text corpus and labels from files
def read_data(self, dataset_dir, train_file, test_file, test_label_file):
self.train_data, self.label_name_data = self.create_dataset(dataset_dir, train_file, None, "train.pt",
find_label_name=True, label_name_loader_name="label_name_data.pt")
if test_file is not None:
self.test_data = self.create_dataset(dataset_dir, test_file, test_label_file, "test.pt")
# read label names from file
def read_label_names(self, dataset_dir, label_name_file):
label_name_file = open(os.path.join(dataset_dir, label_name_file))
label_names = label_name_file.readlines()
self.label_name_dict = {i: [word.lower() for word in category_words.strip().split()] for i, category_words in enumerate(label_names)}
print(f"Label names used for each class are: {self.label_name_dict}")
self.label2class = {}
self.all_label_name_ids = [self.mask_id]
self.all_label_names = [self.tokenizer.mask_token]
for class_idx in self.label_name_dict:
for word in self.label_name_dict[class_idx]:
assert word not in self.label2class, f"\"{word}\" used as the label name by multiple classes!"
self.label2class[word] = class_idx
if word in self.vocab:
self.all_label_name_ids.append(self.vocab[word])
self.all_label_names.append(word)
# create dataset loader
def make_dataloader(self, rank, data_dict, batch_size):
if self.verbose:
print(f"data_dict['input_ids']: {data_dict['input_ids']}")
if "labels" in data_dict:
dataset = TensorDataset(data_dict["input_ids"], data_dict["attention_masks"], data_dict["labels"])
else:
dataset = TensorDataset(data_dict["input_ids"], data_dict["attention_masks"])
sampler = DistributedSampler(dataset, num_replicas=self.world_size, rank=rank)
dataset_loader = DataLoader(dataset, sampler=sampler, batch_size=batch_size, shuffle=False)
return dataset_loader
# filter out stop words and words in multiple categories
def filter_keywords(self, category_vocab_size=100):
all_words = defaultdict(list)
sorted_dicts = {}
for i, cat_dict in self.category_words_freq.items():
sorted_dict = {k:v for k, v in sorted(cat_dict.items(), key=lambda item: item[1], reverse=True)[:category_vocab_size]}
sorted_dicts[i] = sorted_dict
for word_id in sorted_dict:
all_words[word_id].append(i)
repeat_words = []
for word_id in all_words:
if len(all_words[word_id]) > 1:
repeat_words.append(word_id)
self.category_vocab = {}
for i, sorted_dict in sorted_dicts.items():
self.category_vocab[i] = np.array(list(sorted_dict.keys()))
stopwords_vocab = stopwords.words('english')
for i, word_list in self.category_vocab.items():
delete_idx = []
for j, word_id in enumerate(word_list):
word = self.inv_vocab[word_id]
if word in self.label_name_dict[i]:
continue
if not word.isalpha() or len(word) == 1 or word in stopwords_vocab or word_id in repeat_words:
delete_idx.append(j)
self.category_vocab[i] = np.delete(self.category_vocab[i], delete_idx)
def print_predictions(self, word_list):
if not self.verbose: return
print(40*'=')
print(self.decode(word_list))
print(40*'=')
# construct category vocabulary (distributed function)
def category_vocabulary_dist(self, rank, top_pred_num=50, loader_name="category_vocab.pt"):
if self.world_size > 1:
model = self.set_up_dist(rank)
else:
self.model.to(rank)
model = self.model
model.eval()
label_name_dataset_loader = self.make_dataloader(rank, self.label_name_data, self.eval_batch_size)
category_words_freq = {i: defaultdict(float) for i in range(self.num_class)}
wrap_label_name_dataset_loader = tqdm(label_name_dataset_loader) if rank == 0 else label_name_dataset_loader
try:
for batch in wrap_label_name_dataset_loader:
with torch.no_grad():
input_ids = batch[0].to(rank)
input_mask = batch[1].to(rank)
label_pos = batch[2].to(rank)
match_idx = label_pos >= 0
if self.verbose:
print(match_idx)
for input_id in input_ids:
self.print_predictions(input_id)
for attention_mask in input_mask:
print(attention_mask)
predictions = model(input_ids,
pred_mode="mlm",
token_type_ids=None,
attention_mask=input_mask)
if self.verbose:
print(predictions.size())
_, sorted_res = torch.topk(predictions[match_idx], top_pred_num, dim=-1)
label_idx = label_pos[match_idx]
for i, word_list in enumerate(sorted_res):
self.print_predictions(word_list)
for j, word_id in enumerate(word_list):
category_words_freq[label_idx[i].item()][word_id.item()] += 1
if self.verbose:
print(category_words_freq)
save_file = os.path.join(self.temp_dir, f"{rank}_"+loader_name)
torch.save(category_words_freq, save_file)
except RuntimeError as err:
self.cuda_mem_error(err, "eval", rank)
# construct category vocabulary
def category_vocabulary(self, top_pred_num=50, category_vocab_size=100, loader_name="category_vocab.pt"):
loader_file = os.path.join(self.dataset_dir, loader_name)
if os.path.exists(loader_file):
print(f"Loading category vocabulary from {loader_file}")
self.category_vocab = torch.load(loader_file)
else:
print("Contructing category vocabulary.")
if not os.path.exists(self.temp_dir):
os.makedirs(self.temp_dir)
if self.verbose:
print(f"Args: ({top_pred_num, loader_name}); World size: {self.world_size}")
#mp.spawn(self.category_vocabulary_dist, nprocs=self.world_size, args=(top_pred_num, loader_name))
self.category_vocabulary_dist(0, top_pred_num, loader_name)
gather_res = []
for f in os.listdir(self.temp_dir):
if f[-3:] == '.pt':
gather_res.append(torch.load(os.path.join(self.temp_dir, f)))
assert len(gather_res) == self.world_size, "Number of saved files not equal to number of processes!"
self.category_words_freq = {i: defaultdict(float) for i in range(self.num_class)}
for i in range(self.num_class):
for category_words_freq in gather_res:
for word_id, freq in category_words_freq[i].items():
self.category_words_freq[i][word_id] += freq
self.filter_keywords(category_vocab_size)
torch.save(self.category_vocab, loader_file)
if os.path.exists(self.temp_dir):
shutil.rmtree(self.temp_dir)
for i, category_vocab in self.category_vocab.items():
print(f"Class {i} category vocabulary: {[self.inv_vocab[w] for w in category_vocab]}\n")
# prepare self supervision for masked category prediction (distributed function)
def prepare_mcp_dist(self, rank, top_pred_num=50, match_threshold=20, loader_name="mcp_train.pt"):
if self.world_size > 1:
model = self.set_up_dist(rank)
else:
model = self.model
model.eval()
train_dataset_loader = self.make_dataloader(rank, self.train_data, self.eval_batch_size)
all_input_ids = []
all_mask_label = []
all_input_mask = []
category_doc_num = defaultdict(int)
wrap_train_dataset_loader = tqdm(train_dataset_loader) if rank == 0 else train_dataset_loader
try:
for batch in wrap_train_dataset_loader:
with torch.no_grad():
input_ids = batch[0].to(rank)
input_mask = batch[1].to(rank)
predictions = model(input_ids,
pred_mode="mlm",
token_type_ids=None,
attention_mask=input_mask)
_, sorted_res = torch.topk(predictions, top_pred_num, dim=-1)
for i, category_vocab in self.category_vocab.items():
match_idx = torch.zeros_like(sorted_res).bool()
for word_id in category_vocab:
match_idx = (sorted_res == word_id) | match_idx
match_count = torch.sum(match_idx.int(), dim=-1)
valid_idx = (match_count > match_threshold) & (input_mask > 0)
valid_doc = torch.sum(valid_idx, dim=-1) > 0
if valid_doc.any():
mask_label = -1 * torch.ones_like(input_ids)
mask_label[valid_idx] = i
all_input_ids.append(input_ids[valid_doc].cpu())
all_mask_label.append(mask_label[valid_doc].cpu())
all_input_mask.append(input_mask[valid_doc].cpu())
category_doc_num[i] += valid_doc.int().sum().item()
all_input_ids = torch.cat(all_input_ids, dim=0)
all_mask_label = torch.cat(all_mask_label, dim=0)
all_input_mask = torch.cat(all_input_mask, dim=0)
save_dict = {
"all_input_ids": all_input_ids,
"all_mask_label": all_mask_label,
"all_input_mask": all_input_mask,
"category_doc_num": category_doc_num,
}
save_file = os.path.join(self.temp_dir, f"{rank}_"+loader_name)
torch.save(save_dict, save_file)
except RuntimeError as err:
self.cuda_mem_error(err, "eval", rank)
# prepare self supervision for masked category prediction
def prepare_mcp(self, top_pred_num=50, match_threshold=20, loader_name="mcp_train.pt"):
loader_file = os.path.join(self.dataset_dir, loader_name)
if os.path.exists(loader_file):
print(f"Loading masked category prediction data from {loader_file}")
self.mcp_data = torch.load(loader_file)
else:
loader_file = os.path.join(self.dataset_dir, loader_name)
print("Preparing self supervision for masked category prediction.")
if not os.path.exists(self.temp_dir):
os.makedirs(self.temp_dir)
#mp.spawn(self.prepare_mcp_dist, nprocs=self.world_size, args=(top_pred_num, match_threshold, loader_name))
self.prepare_mcp_dist(0, top_pred_num, match_threshold, loader_name)
gather_res = []
for f in os.listdir(self.temp_dir):
if f[-3:] == '.pt':
gather_res.append(torch.load(os.path.join(self.temp_dir, f)))
assert len(gather_res) == self.world_size, "Number of saved files not equal to number of processes!"
all_input_ids = torch.cat([res["all_input_ids"] for res in gather_res], dim=0)
all_mask_label = torch.cat([res["all_mask_label"] for res in gather_res], dim=0)
all_input_mask = torch.cat([res["all_input_mask"] for res in gather_res], dim=0)
category_doc_num = {i: 0 for i in range(self.num_class)}
for i in category_doc_num:
for res in gather_res:
if i in res["category_doc_num"]:
category_doc_num[i] += res["category_doc_num"][i]
print(f"Number of documents with category indicative terms found for each category is: {category_doc_num}")
self.mcp_data = {"input_ids": all_input_ids, "attention_masks": all_input_mask, "labels": all_mask_label}
torch.save(self.mcp_data, loader_file)
if os.path.exists(self.temp_dir):
shutil.rmtree(self.temp_dir)
for i in category_doc_num:
assert category_doc_num[i] > 10, f"Too few ({category_doc_num[i]}) documents with category indicative terms found for category {i}; " \
"try to add more unlabeled documents to the training corpus (recommend) or reduce `--match_threshold` (not recommend)"
print(f"There are totally {len(self.mcp_data['input_ids'])} documents with category indicative terms.")
# masked category prediction (distributed function)
def mcp_dist(self, rank, epochs=5, loader_name="mcp_model.pt"):
if self.world_size > 1:
model = self.set_up_dist(rank)
else:
model = self.model
mcp_dataset_loader = self.make_dataloader(rank, self.mcp_data, self.train_batch_size)
total_steps = len(mcp_dataset_loader) * epochs / self.accum_steps
optimizer = AdamW(filter(lambda p: p.requires_grad, model.parameters()), lr=2e-5, eps=1e-8)
scheduler = get_linear_schedule_with_warmup(optimizer, num_warmup_steps=0.1*total_steps, num_training_steps=total_steps)
try:
for i in range(epochs):
model.train()
total_train_loss = 0
if rank == 0:
print(f"Epoch {i+1}:")
wrap_mcp_dataset_loader = tqdm(mcp_dataset_loader) if rank == 0 else mcp_dataset_loader
model.zero_grad()
for j, batch in enumerate(wrap_mcp_dataset_loader):
input_ids = batch[0].to(rank)
input_mask = batch[1].to(rank)
labels = batch[2].to(rank)
mask_pos = labels >= 0
labels = labels[mask_pos]
# mask out category indicative words
input_ids[mask_pos] = self.mask_id
logits = model(input_ids,
pred_mode="classification",
token_type_ids=None,
attention_mask=input_mask)
logits = logits[mask_pos]
loss = self.mcp_loss(logits.view(-1, self.num_class), labels.view(-1)) / self.accum_steps
total_train_loss += loss.item()
loss.backward()
if (j+1) % self.accum_steps == 0:
# Clip the norm of the gradients to 1.0.
nn.utils.clip_grad_norm_(model.parameters(), 1.0)
optimizer.step()
scheduler.step()
model.zero_grad()
avg_train_loss = torch.tensor([total_train_loss / len(mcp_dataset_loader) * self.accum_steps]).to(rank)
gather_list = [torch.ones_like(avg_train_loss) for _ in range(self.world_size)]
if self.world_size > 1:
dist.all_gather(gather_list, avg_train_loss)
avg_train_loss = torch.tensor(gather_list)
if rank == 0:
print(f"Average training loss: {avg_train_loss.mean().item()}")
if rank == 0:
loader_file = os.path.join(self.dataset_dir, loader_name)
torch.save(self.model.state_dict(), loader_file)
except RuntimeError as err:
self.cuda_mem_error(err, "train", rank)
# masked category prediction
def mcp(self, top_pred_num=50, match_threshold=20, epochs=5, loader_name="mcp_model.pt"):
loader_file = os.path.join(self.dataset_dir, loader_name)
if os.path.exists(loader_file):
print(f"\nLoading model trained via masked category prediction from {loader_file}")
else:
self.prepare_mcp(top_pred_num, match_threshold)
print(f"\nTraining model via masked category prediction.")
#mp.spawn(self.mcp_dist, nprocs=self.world_size, args=(epochs, loader_name))
self.mcp_dist(0, epochs, loader_name)
self.model.load_state_dict(torch.load(loader_file))
# prepare self training data and target distribution
def prepare_self_train_data(self, rank, model, idx):
target_num = min(self.world_size * self.train_batch_size * self.update_interval * self.accum_steps, len(self.train_data["input_ids"]))
if idx + target_num >= len(self.train_data["input_ids"]):
select_idx = torch.cat((torch.arange(idx, len(self.train_data["input_ids"])),
torch.arange(idx + target_num - len(self.train_data["input_ids"]))))
else:
select_idx = torch.arange(idx, idx + target_num)
assert len(select_idx) == target_num
idx = (idx + len(select_idx)) % len(self.train_data["input_ids"])
select_dataset = {"input_ids": self.train_data["input_ids"][select_idx],
"attention_masks": self.train_data["attention_masks"][select_idx]}
dataset_loader = self.make_dataloader(rank, select_dataset, self.eval_batch_size)
input_ids, input_mask, preds = self.inference(model, dataset_loader, rank, return_type="data")
gather_input_ids = [torch.ones_like(input_ids) for _ in range(self.world_size)]
gather_input_mask = [torch.ones_like(input_mask) for _ in range(self.world_size)]
gather_preds = [torch.ones_like(preds) for _ in range(self.world_size)]
dist.all_gather(gather_input_ids, input_ids)
dist.all_gather(gather_input_mask, input_mask)
dist.all_gather(gather_preds, preds)
input_ids = torch.cat(gather_input_ids, dim=0).cpu()
input_mask = torch.cat(gather_input_mask, dim=0).cpu()
all_preds = torch.cat(gather_preds, dim=0).cpu()
weight = all_preds**2 / torch.sum(all_preds, dim=0)
target_dist = (weight.t() / torch.sum(weight, dim=1)).t()
all_target_pred = target_dist.argmax(dim=-1)
agree = (all_preds.argmax(dim=-1) == all_target_pred).int().sum().item() / len(all_target_pred)
self_train_dict = {"input_ids": input_ids, "attention_masks": input_mask, "labels": target_dist}
return self_train_dict, idx, agree
# train a model on batches of data with target labels
def self_train_batches(self, rank, model, self_train_loader, optimizer, scheduler, test_dataset_loader):
model.train()
total_train_loss = 0
wrap_train_dataset_loader = tqdm(self_train_loader) if rank == 0 else self_train_loader
model.zero_grad()
try:
for j, batch in enumerate(wrap_train_dataset_loader):
input_ids = batch[0].to(rank)
input_mask = batch[1].to(rank)
target_dist = batch[2].to(rank)
logits = model(input_ids,
pred_mode="classification",
token_type_ids=None,
attention_mask=input_mask)
logits = logits[:, 0, :]
preds = nn.LogSoftmax(dim=-1)(logits)
loss = self.st_loss(preds.view(-1, self.num_class), target_dist.view(-1, self.num_class)) / self.accum_steps
total_train_loss += loss.item()
loss.backward()
if (j+1) % self.accum_steps == 0:
# Clip the norm of the gradients to 1.0.
nn.utils.clip_grad_norm_(model.parameters(), 1.0)
optimizer.step()
scheduler.step()
model.zero_grad()
if self.with_test_label:
acc = self.inference(model, test_dataset_loader, rank, return_type="acc")
gather_acc = [torch.ones_like(acc) for _ in range(self.world_size)]
dist.all_gather(gather_acc, acc)
acc = torch.tensor(gather_acc).mean().item()
avg_train_loss = torch.tensor([total_train_loss / len(wrap_train_dataset_loader) * self.accum_steps]).to(rank)
gather_list = [torch.ones_like(avg_train_loss) for _ in range(self.world_size)]
dist.all_gather(gather_list, avg_train_loss)
avg_train_loss = torch.tensor(gather_list)
if rank == 0:
print(f"lr: {optimizer.param_groups[0]['lr']:.4g}")
print(f"Average training loss: {avg_train_loss.mean().item()}")
if self.with_test_label:
print(f"Test acc: {acc}")
except RuntimeError as err:
self.cuda_mem_error(err, "train", rank)
# self training (distributed function)
def self_train_dist(self, rank, epochs, loader_name="final_model.pt"):
model = self.set_up_dist(rank)
test_dataset_loader = self.make_dataloader(rank, self.test_data, self.eval_batch_size) if self.with_test_label else None
total_steps = int(len(self.train_data["input_ids"]) * epochs / (self.world_size * self.train_batch_size * self.accum_steps))
optimizer = AdamW(filter(lambda p: p.requires_grad, model.parameters()), lr=1e-6, eps=1e-8)
scheduler = get_linear_schedule_with_warmup(optimizer, num_warmup_steps=0.1*total_steps, num_training_steps=total_steps)
idx = 0
if self.early_stop:
agree_count = 0
for i in tqdm(range(int(total_steps / self.update_interval))):
self_train_dict, idx, agree = self.prepare_self_train_data(rank, model, idx)
# early stop if current prediction agrees with target distribution for 3 consecutive updates
if self.early_stop:
if 1 - agree < 1e-3:
agree_count += 1
else:
agree_count = 0
if agree_count >= 3:
break
self_train_dataset_loader = self.make_dataloader(rank, self_train_dict, self.train_batch_size)
self.self_train_batches(rank, model, self_train_dataset_loader, optimizer, scheduler, test_dataset_loader)
if rank == 0:
loader_file = os.path.join(self.dataset_dir, loader_name)
print(f"Saving final model to {loader_file}")
torch.save(model.module.state_dict(), loader_file)
# self training
def self_train(self, epochs, loader_name="final_model.pt"):
loader_file = os.path.join(self.dataset_dir, loader_name)
if os.path.exists(loader_file):
print(f"\nFinal model {loader_file} found, skip self-training")
else:
rand_idx = torch.randperm(len(self.train_data["input_ids"]))
self.train_data = {"input_ids": self.train_data["input_ids"][rand_idx],
"attention_masks": self.train_data["attention_masks"][rand_idx]}
print(f"\nStart self-training.")
if self.world_size > 1:
mp.spawn(self.self_train_dist, nprocs=self.world_size, args=(epochs, loader_name))
else:
self.self_train_dist(0, epochs, loader_name)
# use a model to do inference on a dataloader
def inference(self, model, dataset_loader, rank, return_type):
if return_type == "data":
all_input_ids = []
all_input_mask = []
all_preds = []
elif return_type == "acc":
pred_labels = []
truth_labels = []
elif return_type == "pred":
pred_labels = []
model.eval()
try:
for batch in dataset_loader:
with torch.no_grad():
input_ids = batch[0].to(rank)
input_mask = batch[1].to(rank)
logits = model(input_ids,
pred_mode="classification",
token_type_ids=None,
attention_mask=input_mask)
logits = logits[:, 0, :]
if return_type == "data":
all_input_ids.append(input_ids)
all_input_mask.append(input_mask)
all_preds.append(nn.Softmax(dim=-1)(logits))
elif return_type == "acc":
labels = batch[2]
pred_labels.append(torch.argmax(logits, dim=-1).cpu())
truth_labels.append(labels)
elif return_type == "pred":
pred_labels.append(torch.argmax(logits, dim=-1).cpu())
if return_type == "data":
all_input_ids = torch.cat(all_input_ids, dim=0)
all_input_mask = torch.cat(all_input_mask, dim=0)
all_preds = torch.cat(all_preds, dim=0)
return all_input_ids, all_input_mask, all_preds
elif return_type == "acc":
pred_labels = torch.cat(pred_labels, dim=0)
truth_labels = torch.cat(truth_labels, dim=0)
samples = len(truth_labels)
acc = (pred_labels == truth_labels).float().sum() / samples
return acc.to(rank)
elif return_type == "pred":
pred_labels = torch.cat(pred_labels, dim=0)
return pred_labels
except RuntimeError as err:
self.cuda_mem_error(err, "eval", rank)
# use trained model to make predictions on the test set
def write_results(self, loader_name="final_model.pt", out_file="out.txt"):
loader_file = os.path.join(self.dataset_dir, loader_name)
assert os.path.exists(loader_file)
print(f"\nLoading final model from {loader_file}")
self.model.load_state_dict(torch.load(loader_file))
self.model.to(0)
test_set = TensorDataset(self.test_data["input_ids"], self.test_data["attention_masks"])
test_dataset_loader = DataLoader(test_set, sampler=SequentialSampler(test_set), batch_size=self.eval_batch_size)
pred_labels = self.inference(self.model, test_dataset_loader, 0, return_type="pred")
out_file = os.path.join(self.dataset_dir, out_file)
print(f"Writing prediction results to {out_file}")
f_out = open(out_file, 'w')
for label in pred_labels:
f_out.write(str(label.item()) + '\n')
# print error message based on CUDA memory error
def cuda_mem_error(self, err, mode, rank):
if rank == 0:
print(err)
if "CUDA out of memory" in str(err):
if mode == "eval":
print(f"Your GPUs can't hold the current batch size for evaluation, try to reduce `--eval_batch_size`, current: {self.eval_batch_size}")
else:
print(f"Your GPUs can't hold the current batch size for training, try to reduce `--train_batch_size`, current: {self.train_batch_size}")
sys.exit(1)
|
[
"torch.nn.CrossEntropyLoss",
"multiprocessing.cpu_count",
"numpy.array",
"torch.utils.data.distributed.DistributedSampler",
"torch.sum",
"sys.exit",
"joblib.delayed",
"torch.arange",
"os.path.exists",
"numpy.mean",
"os.listdir",
"nltk.corpus.stopwords.words",
"numpy.delete",
"numpy.max",
"LOTClass.bert.model.LOTClassModel.from_pretrained",
"torch.zeros_like",
"torch.distributed.all_gather",
"torch.nn.parallel.DistributedDataParallel",
"torch.argmax",
"torch.ones_like",
"torch.topk",
"torch.nn.KLDivLoss",
"torch.utils.data.SequentialSampler",
"torch.utils.data.TensorDataset",
"shutil.rmtree",
"torch.save",
"numpy.std",
"warnings.filterwarnings",
"torch.cat",
"os.makedirs",
"torch.multiprocessing.spawn",
"torch.nn.Softmax",
"transformers.get_linear_schedule_with_warmup",
"torch.load",
"tqdm.tqdm",
"transformers.BertTokenizer.from_pretrained",
"os.path.join",
"joblib.Parallel",
"torch.tensor",
"collections.defaultdict",
"torch.nn.LogSoftmax",
"torch.utils.data.DataLoader",
"torch.no_grad",
"torch.distributed.init_process_group",
"torch.ones"
] |
[((970, 1003), 'warnings.filterwarnings', 'warnings.filterwarnings', (['"""ignore"""'], {}), "('ignore')\n", (993, 1003), False, 'import warnings\n'), ((1868, 1937), 'transformers.BertTokenizer.from_pretrained', 'BertTokenizer.from_pretrained', (['self.pretrained_lm'], {'do_lower_case': '(True)'}), '(self.pretrained_lm, do_lower_case=True)\n', (1897, 1937), False, 'from transformers import BertTokenizer\n'), ((2396, 2529), 'LOTClass.bert.model.LOTClassModel.from_pretrained', 'LOTClassModel.from_pretrained', (['self.pretrained_lm'], {'output_attentions': '(False)', 'output_hidden_states': '(False)', 'num_labels': 'self.num_class'}), '(self.pretrained_lm, output_attentions=False,\n output_hidden_states=False, num_labels=self.num_class)\n', (2425, 2529), False, 'from LOTClass.bert.model import LOTClassModel\n'), ((2930, 2951), 'torch.nn.CrossEntropyLoss', 'nn.CrossEntropyLoss', ([], {}), '()\n', (2949, 2951), False, 'from torch import nn\n'), ((2975, 3010), 'torch.nn.KLDivLoss', 'nn.KLDivLoss', ([], {'reduction': '"""batchmean"""'}), "(reduction='batchmean')\n", (2987, 3010), False, 'from torch import nn\n'), ((3181, 3313), 'torch.distributed.init_process_group', 'dist.init_process_group', ([], {'backend': '"""nccl"""', 'init_method': 'f"""tcp://localhost:{self.dist_port}"""', 'world_size': 'self.world_size', 'rank': 'rank'}), "(backend='nccl', init_method=\n f'tcp://localhost:{self.dist_port}', world_size=self.world_size, rank=rank)\n", (3204, 3313), True, 'import torch.distributed as dist\n'), ((3448, 3506), 'torch.nn.parallel.DistributedDataParallel', 'DDP', (['model'], {'device_ids': '[rank]', 'find_unused_parameters': '(True)'}), '(model, device_ids=[rank], find_unused_parameters=True)\n', (3451, 3506), True, 'from torch.nn.parallel import DistributedDataParallel as DDP\n'), ((5102, 5140), 'os.path.join', 'os.path.join', (['dataset_dir', 'loader_name'], {}), '(dataset_dir, loader_name)\n', (5114, 5140), False, 'import os\n'), ((5152, 5179), 'os.path.exists', 'os.path.exists', (['loader_file'], {}), '(loader_file)\n', (5166, 5179), False, 'import os\n'), ((13068, 13136), 'torch.utils.data.distributed.DistributedSampler', 'DistributedSampler', (['dataset'], {'num_replicas': 'self.world_size', 'rank': 'rank'}), '(dataset, num_replicas=self.world_size, rank=rank)\n', (13086, 13136), False, 'from torch.utils.data.distributed import DistributedSampler\n'), ((13162, 13236), 'torch.utils.data.DataLoader', 'DataLoader', (['dataset'], {'sampler': 'sampler', 'batch_size': 'batch_size', 'shuffle': '(False)'}), '(dataset, sampler=sampler, batch_size=batch_size, shuffle=False)\n', (13172, 13236), False, 'from torch.utils.data import TensorDataset, DataLoader, SequentialSampler\n'), ((13405, 13422), 'collections.defaultdict', 'defaultdict', (['list'], {}), '(list)\n', (13416, 13422), False, 'from collections import defaultdict\n'), ((14100, 14126), 'nltk.corpus.stopwords.words', 'stopwords.words', (['"""english"""'], {}), "('english')\n", (14115, 14126), False, 'from nltk.corpus import stopwords\n'), ((17233, 17276), 'os.path.join', 'os.path.join', (['self.dataset_dir', 'loader_name'], {}), '(self.dataset_dir, loader_name)\n', (17245, 17276), False, 'import os\n'), ((17288, 17315), 'os.path.exists', 'os.path.exists', (['loader_file'], {}), '(loader_file)\n', (17302, 17315), False, 'import os\n'), ((19454, 19470), 'collections.defaultdict', 'defaultdict', (['int'], {}), '(int)\n', (19465, 19470), False, 'from collections import defaultdict\n'), ((21898, 21941), 'os.path.join', 'os.path.join', (['self.dataset_dir', 'loader_name'], {}), '(self.dataset_dir, loader_name)\n', (21910, 21941), False, 'import os\n'), ((21953, 21980), 'os.path.exists', 'os.path.exists', (['loader_file'], {}), '(loader_file)\n', (21967, 21980), False, 'import os\n'), ((24789, 24903), 'transformers.get_linear_schedule_with_warmup', 'get_linear_schedule_with_warmup', (['optimizer'], {'num_warmup_steps': '(0.1 * total_steps)', 'num_training_steps': 'total_steps'}), '(optimizer, num_warmup_steps=0.1 *\n total_steps, num_training_steps=total_steps)\n', (24820, 24903), False, 'from transformers import AdamW, get_linear_schedule_with_warmup\n'), ((27329, 27372), 'os.path.join', 'os.path.join', (['self.dataset_dir', 'loader_name'], {}), '(self.dataset_dir, loader_name)\n', (27341, 27372), False, 'import os\n'), ((27384, 27411), 'os.path.exists', 'os.path.exists', (['loader_file'], {}), '(loader_file)\n', (27398, 27411), False, 'import os\n'), ((29199, 29243), 'torch.distributed.all_gather', 'dist.all_gather', (['gather_input_ids', 'input_ids'], {}), '(gather_input_ids, input_ids)\n', (29214, 29243), True, 'import torch.distributed as dist\n'), ((29252, 29298), 'torch.distributed.all_gather', 'dist.all_gather', (['gather_input_mask', 'input_mask'], {}), '(gather_input_mask, input_mask)\n', (29267, 29298), True, 'import torch.distributed as dist\n'), ((29307, 29343), 'torch.distributed.all_gather', 'dist.all_gather', (['gather_preds', 'preds'], {}), '(gather_preds, preds)\n', (29322, 29343), True, 'import torch.distributed as dist\n'), ((32858, 32972), 'transformers.get_linear_schedule_with_warmup', 'get_linear_schedule_with_warmup', (['optimizer'], {'num_warmup_steps': '(0.1 * total_steps)', 'num_training_steps': 'total_steps'}), '(optimizer, num_warmup_steps=0.1 *\n total_steps, num_training_steps=total_steps)\n', (32889, 32972), False, 'from transformers import AdamW, get_linear_schedule_with_warmup\n'), ((34077, 34120), 'os.path.join', 'os.path.join', (['self.dataset_dir', 'loader_name'], {}), '(self.dataset_dir, loader_name)\n', (34089, 34120), False, 'import os\n'), ((34132, 34159), 'os.path.exists', 'os.path.exists', (['loader_file'], {}), '(loader_file)\n', (34146, 34159), False, 'import os\n'), ((37219, 37262), 'os.path.join', 'os.path.join', (['self.dataset_dir', 'loader_name'], {}), '(self.dataset_dir, loader_name)\n', (37231, 37262), False, 'import os\n'), ((37278, 37305), 'os.path.exists', 'os.path.exists', (['loader_file'], {}), '(loader_file)\n', (37292, 37305), False, 'import os\n'), ((37469, 37546), 'torch.utils.data.TensorDataset', 'TensorDataset', (["self.test_data['input_ids']", "self.test_data['attention_masks']"], {}), "(self.test_data['input_ids'], self.test_data['attention_masks'])\n", (37482, 37546), False, 'from torch.utils.data import TensorDataset, DataLoader, SequentialSampler\n'), ((37780, 37820), 'os.path.join', 'os.path.join', (['self.dataset_dir', 'out_file'], {}), '(self.dataset_dir, out_file)\n', (37792, 37820), False, 'import os\n'), ((38574, 38585), 'sys.exit', 'sys.exit', (['(1)'], {}), '(1)\n', (38582, 38585), False, 'import sys\n'), ((5263, 5286), 'torch.load', 'torch.load', (['loader_file'], {}), '(loader_file)\n', (5273, 5286), False, 'import torch\n'), ((5852, 5896), 'torch.cat', 'torch.cat', (['[result[0] for result in results]'], {}), '([result[0] for result in results])\n', (5861, 5896), False, 'import torch\n'), ((5997, 6041), 'torch.cat', 'torch.cat', (['[result[1] for result in results]'], {}), '([result[1] for result in results])\n', (6006, 6041), False, 'import torch\n'), ((6636, 6665), 'torch.save', 'torch.save', (['data', 'loader_file'], {}), '(data, loader_file)\n', (6646, 6665), False, 'import torch\n'), ((6720, 6769), 'os.path.join', 'os.path.join', (['dataset_dir', 'label_name_loader_name'], {}), '(dataset_dir, label_name_loader_name)\n', (6732, 6769), False, 'import os\n'), ((6785, 6812), 'os.path.exists', 'os.path.exists', (['loader_file'], {}), '(loader_file)\n', (6799, 6812), False, 'import os\n'), ((8684, 8726), 'torch.ones', 'torch.ones', (['self.max_len'], {'dtype': 'torch.long'}), '(self.max_len, dtype=torch.long)\n', (8694, 8726), False, 'import torch\n'), ((10713, 10745), 'torch.cat', 'torch.cat', (['label_name_idx'], {'dim': '(0)'}), '(label_name_idx, dim=0)\n', (10722, 10745), False, 'import torch\n'), ((10800, 10845), 'torch.ones', 'torch.ones', (['(0)', 'self.max_len'], {'dtype': 'torch.long'}), '(0, self.max_len, dtype=torch.long)\n', (10810, 10845), False, 'import torch\n'), ((10892, 10937), 'torch.ones', 'torch.ones', (['(0)', 'self.max_len'], {'dtype': 'torch.long'}), '(0, self.max_len, dtype=torch.long)\n', (10902, 10937), False, 'import torch\n'), ((10967, 11012), 'torch.ones', 'torch.ones', (['(0)', 'self.max_len'], {'dtype': 'torch.long'}), '(0, self.max_len, dtype=torch.long)\n', (10977, 11012), False, 'import torch\n'), ((11736, 11778), 'os.path.join', 'os.path.join', (['dataset_dir', 'label_name_file'], {}), '(dataset_dir, label_name_file)\n', (11748, 11778), False, 'import os\n'), ((12857, 12949), 'torch.utils.data.TensorDataset', 'TensorDataset', (["data_dict['input_ids']", "data_dict['attention_masks']", "data_dict['labels']"], {}), "(data_dict['input_ids'], data_dict['attention_masks'],\n data_dict['labels'])\n", (12870, 12949), False, 'from torch.utils.data import TensorDataset, DataLoader, SequentialSampler\n'), ((12982, 13049), 'torch.utils.data.TensorDataset', 'TensorDataset', (["data_dict['input_ids']", "data_dict['attention_masks']"], {}), "(data_dict['input_ids'], data_dict['attention_masks'])\n", (12995, 13049), False, 'from torch.utils.data import TensorDataset, DataLoader, SequentialSampler\n'), ((14581, 14626), 'numpy.delete', 'np.delete', (['self.category_vocab[i]', 'delete_idx'], {}), '(self.category_vocab[i], delete_idx)\n', (14590, 14626), True, 'import numpy as np\n'), ((15257, 15275), 'collections.defaultdict', 'defaultdict', (['float'], {}), '(float)\n', (15268, 15275), False, 'from collections import defaultdict\n'), ((15349, 15380), 'tqdm.tqdm', 'tqdm', (['label_name_dataset_loader'], {}), '(label_name_dataset_loader)\n', (15353, 15380), False, 'from tqdm import tqdm\n'), ((16870, 16923), 'os.path.join', 'os.path.join', (['self.temp_dir', "(f'{rank}_' + loader_name)"], {}), "(self.temp_dir, f'{rank}_' + loader_name)\n", (16882, 16923), False, 'import os\n'), ((16934, 16976), 'torch.save', 'torch.save', (['category_words_freq', 'save_file'], {}), '(category_words_freq, save_file)\n', (16944, 16976), False, 'import torch\n'), ((17420, 17443), 'torch.load', 'torch.load', (['loader_file'], {}), '(loader_file)\n', (17430, 17443), False, 'import torch\n'), ((17959, 17984), 'os.listdir', 'os.listdir', (['self.temp_dir'], {}), '(self.temp_dir)\n', (17969, 17984), False, 'import os\n'), ((18618, 18662), 'torch.save', 'torch.save', (['self.category_vocab', 'loader_file'], {}), '(self.category_vocab, loader_file)\n', (18628, 18662), False, 'import torch\n'), ((18678, 18707), 'os.path.exists', 'os.path.exists', (['self.temp_dir'], {}), '(self.temp_dir)\n', (18692, 18707), False, 'import os\n'), ((19507, 19533), 'tqdm.tqdm', 'tqdm', (['train_dataset_loader'], {}), '(train_dataset_loader)\n', (19511, 19533), False, 'from tqdm import tqdm\n'), ((21115, 21146), 'torch.cat', 'torch.cat', (['all_input_ids'], {'dim': '(0)'}), '(all_input_ids, dim=0)\n', (21124, 21146), False, 'import torch\n'), ((21176, 21208), 'torch.cat', 'torch.cat', (['all_mask_label'], {'dim': '(0)'}), '(all_mask_label, dim=0)\n', (21185, 21208), False, 'import torch\n'), ((21238, 21270), 'torch.cat', 'torch.cat', (['all_input_mask'], {'dim': '(0)'}), '(all_input_mask, dim=0)\n', (21247, 21270), False, 'import torch\n'), ((21537, 21590), 'os.path.join', 'os.path.join', (['self.temp_dir', "(f'{rank}_' + loader_name)"], {}), "(self.temp_dir, f'{rank}_' + loader_name)\n", (21549, 21590), False, 'import os\n'), ((21601, 21633), 'torch.save', 'torch.save', (['save_dict', 'save_file'], {}), '(save_dict, save_file)\n', (21611, 21633), False, 'import torch\n'), ((22091, 22114), 'torch.load', 'torch.load', (['loader_file'], {}), '(loader_file)\n', (22101, 22114), False, 'import torch\n'), ((22155, 22198), 'os.path.join', 'os.path.join', (['self.dataset_dir', 'loader_name'], {}), '(self.dataset_dir, loader_name)\n', (22167, 22198), False, 'import os\n'), ((22622, 22647), 'os.listdir', 'os.listdir', (['self.temp_dir'], {}), '(self.temp_dir)\n', (22632, 22647), False, 'import os\n'), ((22908, 22970), 'torch.cat', 'torch.cat', (["[res['all_input_ids'] for res in gather_res]"], {'dim': '(0)'}), "([res['all_input_ids'] for res in gather_res], dim=0)\n", (22917, 22970), False, 'import torch\n'), ((23000, 23063), 'torch.cat', 'torch.cat', (["[res['all_mask_label'] for res in gather_res]"], {'dim': '(0)'}), "([res['all_mask_label'] for res in gather_res], dim=0)\n", (23009, 23063), False, 'import torch\n'), ((23093, 23156), 'torch.cat', 'torch.cat', (["[res['all_input_mask'] for res in gather_res]"], {'dim': '(0)'}), "([res['all_input_mask'] for res in gather_res], dim=0)\n", (23102, 23156), False, 'import torch\n'), ((23681, 23719), 'torch.save', 'torch.save', (['self.mcp_data', 'loader_file'], {}), '(self.mcp_data, loader_file)\n', (23691, 23719), False, 'import torch\n'), ((23735, 23764), 'os.path.exists', 'os.path.exists', (['self.temp_dir'], {}), '(self.temp_dir)\n', (23749, 23764), False, 'import os\n'), ((27828, 27851), 'torch.load', 'torch.load', (['loader_file'], {}), '(loader_file)\n', (27838, 27851), False, 'import torch\n'), ((28411, 28446), 'torch.arange', 'torch.arange', (['idx', '(idx + target_num)'], {}), '(idx, idx + target_num)\n', (28423, 28446), False, 'import torch\n'), ((28961, 28987), 'torch.ones_like', 'torch.ones_like', (['input_ids'], {}), '(input_ids)\n', (28976, 28987), False, 'import torch\n'), ((29050, 29077), 'torch.ones_like', 'torch.ones_like', (['input_mask'], {}), '(input_mask)\n', (29065, 29077), False, 'import torch\n'), ((29135, 29157), 'torch.ones_like', 'torch.ones_like', (['preds'], {}), '(preds)\n', (29150, 29157), False, 'import torch\n'), ((29557, 29584), 'torch.sum', 'torch.sum', (['all_preds'], {'dim': '(0)'}), '(all_preds, dim=0)\n', (29566, 29584), False, 'import torch\n'), ((30211, 30234), 'tqdm.tqdm', 'tqdm', (['self_train_loader'], {}), '(self_train_loader)\n', (30215, 30234), False, 'from tqdm import tqdm\n'), ((31869, 31913), 'torch.distributed.all_gather', 'dist.all_gather', (['gather_list', 'avg_train_loss'], {}), '(gather_list, avg_train_loss)\n', (31884, 31913), True, 'import torch.distributed as dist\n'), ((31943, 31968), 'torch.tensor', 'torch.tensor', (['gather_list'], {}), '(gather_list)\n', (31955, 31968), False, 'import torch\n'), ((33805, 33848), 'os.path.join', 'os.path.join', (['self.dataset_dir', 'loader_name'], {}), '(self.dataset_dir, loader_name)\n', (33817, 33848), False, 'import os\n'), ((37400, 37423), 'torch.load', 'torch.load', (['loader_file'], {}), '(loader_file)\n', (37410, 37423), False, 'import torch\n'), ((1301, 1312), 'multiprocessing.cpu_count', 'cpu_count', ([], {}), '()\n', (1310, 1312), False, 'from multiprocessing import cpu_count\n'), ((5407, 5443), 'os.path.join', 'os.path.join', (['dataset_dir', 'text_file'], {}), '(dataset_dir, text_file)\n', (5419, 5443), False, 'import os\n'), ((5743, 5773), 'joblib.Parallel', 'Parallel', ([], {'n_jobs': 'self.num_cpus'}), '(n_jobs=self.num_cpus)\n', (5751, 5773), False, 'from joblib import Parallel, delayed\n'), ((6399, 6419), 'torch.tensor', 'torch.tensor', (['labels'], {}), '(labels)\n', (6411, 6419), False, 'import torch\n'), ((6924, 6947), 'torch.load', 'torch.load', (['loader_file'], {}), '(loader_file)\n', (6934, 6947), False, 'import torch\n'), ((7584, 7628), 'torch.cat', 'torch.cat', (['[result[0] for result in results]'], {}), '([result[0] for result in results])\n', (7593, 7628), False, 'import torch\n'), ((7679, 7723), 'torch.cat', 'torch.cat', (['[result[1] for result in results]'], {}), '([result[1] for result in results])\n', (7688, 7723), False, 'import torch\n'), ((7757, 7801), 'torch.cat', 'torch.cat', (['[result[2] for result in results]'], {}), '([result[2] for result in results])\n', (7766, 7801), False, 'import torch\n'), ((8185, 8234), 'os.path.join', 'os.path.join', (['dataset_dir', 'label_name_loader_name'], {}), '(dataset_dir, label_name_loader_name)\n', (8197, 8234), False, 'import os\n'), ((8326, 8366), 'torch.save', 'torch.save', (['label_name_data', 'loader_file'], {}), '(label_name_data, loader_file)\n', (8336, 8366), False, 'import torch\n'), ((17531, 17560), 'os.path.exists', 'os.path.exists', (['self.temp_dir'], {}), '(self.temp_dir)\n', (17545, 17560), False, 'import os\n'), ((17578, 17604), 'os.makedirs', 'os.makedirs', (['self.temp_dir'], {}), '(self.temp_dir)\n', (17589, 17604), False, 'import os\n'), ((18260, 18278), 'collections.defaultdict', 'defaultdict', (['float'], {}), '(float)\n', (18271, 18278), False, 'from collections import defaultdict\n'), ((18725, 18753), 'shutil.rmtree', 'shutil.rmtree', (['self.temp_dir'], {}), '(self.temp_dir)\n', (18738, 18753), False, 'import shutil\n'), ((22298, 22327), 'os.path.exists', 'os.path.exists', (['self.temp_dir'], {}), '(self.temp_dir)\n', (22312, 22327), False, 'import os\n'), ((22345, 22371), 'os.makedirs', 'os.makedirs', (['self.temp_dir'], {}), '(self.temp_dir)\n', (22356, 22371), False, 'import os\n'), ((23782, 23810), 'shutil.rmtree', 'shutil.rmtree', (['self.temp_dir'], {}), '(self.temp_dir)\n', (23795, 23810), False, 'import shutil\n'), ((26786, 26811), 'torch.tensor', 'torch.tensor', (['gather_list'], {}), '(gather_list)\n', (26798, 26811), False, 'import torch\n'), ((26982, 27025), 'os.path.join', 'os.path.join', (['self.dataset_dir', 'loader_name'], {}), '(self.dataset_dir, loader_name)\n', (26994, 27025), False, 'import os\n'), ((29364, 29398), 'torch.cat', 'torch.cat', (['gather_input_ids'], {'dim': '(0)'}), '(gather_input_ids, dim=0)\n', (29373, 29398), False, 'import torch\n'), ((29426, 29461), 'torch.cat', 'torch.cat', (['gather_input_mask'], {'dim': '(0)'}), '(gather_input_mask, dim=0)\n', (29435, 29461), False, 'import torch\n'), ((29488, 29518), 'torch.cat', 'torch.cat', (['gather_preds'], {'dim': '(0)'}), '(gather_preds, dim=0)\n', (29497, 29518), False, 'import torch\n'), ((31548, 31580), 'torch.distributed.all_gather', 'dist.all_gather', (['gather_acc', 'acc'], {}), '(gather_acc, acc)\n', (31563, 31580), True, 'import torch.distributed as dist\n'), ((31792, 31823), 'torch.ones_like', 'torch.ones_like', (['avg_train_loss'], {}), '(avg_train_loss)\n', (31807, 31823), False, 'import torch\n'), ((34601, 34687), 'torch.multiprocessing.spawn', 'mp.spawn', (['self.self_train_dist'], {'nprocs': 'self.world_size', 'args': '(epochs, loader_name)'}), '(self.self_train_dist, nprocs=self.world_size, args=(epochs,\n loader_name))\n', (34609, 34687), True, 'import torch.multiprocessing as mp\n'), ((36296, 36327), 'torch.cat', 'torch.cat', (['all_input_ids'], {'dim': '(0)'}), '(all_input_ids, dim=0)\n', (36305, 36327), False, 'import torch\n'), ((36361, 36393), 'torch.cat', 'torch.cat', (['all_input_mask'], {'dim': '(0)'}), '(all_input_mask, dim=0)\n', (36370, 36393), False, 'import torch\n'), ((36422, 36449), 'torch.cat', 'torch.cat', (['all_preds'], {'dim': '(0)'}), '(all_preds, dim=0)\n', (36431, 36449), False, 'import torch\n'), ((37606, 37633), 'torch.utils.data.SequentialSampler', 'SequentialSampler', (['test_set'], {}), '(test_set)\n', (37623, 37633), False, 'from torch.utils.data import TensorDataset, DataLoader, SequentialSampler\n'), ((1281, 1292), 'multiprocessing.cpu_count', 'cpu_count', ([], {}), '()\n', (1290, 1292), False, 'from multiprocessing import cpu_count\n'), ((3841, 3856), 'numpy.max', 'np.max', (['doc_len'], {}), '(doc_len)\n', (3847, 3856), True, 'import numpy as np\n'), ((3872, 3888), 'numpy.mean', 'np.mean', (['doc_len'], {}), '(doc_len)\n', (3879, 3888), True, 'import numpy as np\n'), ((3904, 3919), 'numpy.std', 'np.std', (['doc_len'], {}), '(doc_len)\n', (3910, 3919), True, 'import numpy as np\n'), ((3951, 3968), 'numpy.array', 'np.array', (['doc_len'], {}), '(doc_len)\n', (3959, 3968), True, 'import numpy as np\n'), ((6258, 6295), 'os.path.join', 'os.path.join', (['dataset_dir', 'label_file'], {}), '(dataset_dir, label_file)\n', (6270, 6295), False, 'import os\n'), ((7080, 7116), 'os.path.join', 'os.path.join', (['dataset_dir', 'text_file'], {}), '(dataset_dir, text_file)\n', (7092, 7116), False, 'import os\n'), ((7440, 7470), 'joblib.Parallel', 'Parallel', ([], {'n_jobs': 'self.num_cpus'}), '(n_jobs=self.num_cpus)\n', (7448, 7470), False, 'from joblib import Parallel, delayed\n'), ((15516, 15531), 'torch.no_grad', 'torch.no_grad', ([], {}), '()\n', (15529, 15531), False, 'import torch\n'), ((16373, 16429), 'torch.topk', 'torch.topk', (['predictions[match_idx]', 'top_pred_num'], {'dim': '(-1)'}), '(predictions[match_idx], top_pred_num, dim=-1)\n', (16383, 16429), False, 'import torch\n'), ((19659, 19674), 'torch.no_grad', 'torch.no_grad', ([], {}), '()\n', (19672, 19674), False, 'import torch\n'), ((20049, 20094), 'torch.topk', 'torch.topk', (['predictions', 'top_pred_num'], {'dim': '(-1)'}), '(predictions, top_pred_num, dim=-1)\n', (20059, 20094), False, 'import torch\n'), ((25129, 25153), 'tqdm.tqdm', 'tqdm', (['mcp_dataset_loader'], {}), '(mcp_dataset_loader)\n', (25133, 25153), False, 'from tqdm import tqdm\n'), ((26583, 26614), 'torch.ones_like', 'torch.ones_like', (['avg_train_loss'], {}), '(avg_train_loss)\n', (26598, 26614), False, 'import torch\n'), ((26708, 26752), 'torch.distributed.all_gather', 'dist.all_gather', (['gather_list', 'avg_train_loss'], {}), '(gather_list, avg_train_loss)\n', (26723, 26752), True, 'import torch.distributed as dist\n'), ((29621, 29645), 'torch.sum', 'torch.sum', (['weight'], {'dim': '(1)'}), '(weight, dim=1)\n', (29630, 29645), False, 'import torch\n'), ((30793, 30814), 'torch.nn.LogSoftmax', 'nn.LogSoftmax', ([], {'dim': '(-1)'}), '(dim=-1)\n', (30806, 30814), False, 'from torch import nn\n'), ((31478, 31498), 'torch.ones_like', 'torch.ones_like', (['acc'], {}), '(acc)\n', (31493, 31498), False, 'import torch\n'), ((35260, 35275), 'torch.no_grad', 'torch.no_grad', ([], {}), '()\n', (35273, 35275), False, 'import torch\n'), ((36583, 36612), 'torch.cat', 'torch.cat', (['pred_labels'], {'dim': '(0)'}), '(pred_labels, dim=0)\n', (36592, 36612), False, 'import torch\n'), ((36644, 36674), 'torch.cat', 'torch.cat', (['truth_labels'], {'dim': '(0)'}), '(truth_labels, dim=0)\n', (36653, 36674), False, 'import torch\n'), ((5341, 5377), 'os.path.join', 'os.path.join', (['dataset_dir', 'text_file'], {}), '(dataset_dir, text_file)\n', (5353, 5377), False, 'import os\n'), ((5774, 5794), 'joblib.delayed', 'delayed', (['self.encode'], {}), '(self.encode)\n', (5781, 5794), False, 'from joblib import Parallel, delayed\n'), ((36901, 36930), 'torch.cat', 'torch.cat', (['pred_labels'], {'dim': '(0)'}), '(pred_labels, dim=0)\n', (36910, 36930), False, 'import torch\n'), ((6188, 6225), 'os.path.join', 'os.path.join', (['dataset_dir', 'label_file'], {}), '(dataset_dir, label_file)\n', (6200, 6225), False, 'import os\n'), ((7010, 7046), 'os.path.join', 'os.path.join', (['dataset_dir', 'text_file'], {}), '(dataset_dir, text_file)\n', (7022, 7046), False, 'import os\n'), ((7471, 7506), 'joblib.delayed', 'delayed', (['self.label_name_occurrence'], {}), '(self.label_name_occurrence)\n', (7478, 7506), False, 'from joblib import Parallel, delayed\n'), ((18071, 18101), 'os.path.join', 'os.path.join', (['self.temp_dir', 'f'], {}), '(self.temp_dir, f)\n', (18083, 18101), False, 'import os\n'), ((20568, 20596), 'torch.sum', 'torch.sum', (['valid_idx'], {'dim': '(-1)'}), '(valid_idx, dim=-1)\n', (20577, 20596), False, 'import torch\n'), ((22734, 22764), 'os.path.join', 'os.path.join', (['self.temp_dir', 'f'], {}), '(self.temp_dir, f)\n', (22746, 22764), False, 'import os\n'), ((20205, 20233), 'torch.zeros_like', 'torch.zeros_like', (['sorted_res'], {}), '(sorted_res)\n', (20221, 20233), False, 'import torch\n'), ((20691, 20717), 'torch.ones_like', 'torch.ones_like', (['input_ids'], {}), '(input_ids)\n', (20706, 20717), False, 'import torch\n'), ((31603, 31627), 'torch.tensor', 'torch.tensor', (['gather_acc'], {}), '(gather_acc)\n', (31615, 31627), False, 'import torch\n'), ((35851, 35869), 'torch.nn.Softmax', 'nn.Softmax', ([], {'dim': '(-1)'}), '(dim=-1)\n', (35861, 35869), False, 'from torch import nn\n'), ((36011, 36039), 'torch.argmax', 'torch.argmax', (['logits'], {'dim': '(-1)'}), '(logits, dim=-1)\n', (36023, 36039), False, 'import torch\n'), ((36190, 36218), 'torch.argmax', 'torch.argmax', (['logits'], {'dim': '(-1)'}), '(logits, dim=-1)\n', (36202, 36218), False, 'import torch\n')]
|
#!/usr/bin/env python
import os
os.environ["TF_CPP_MIN_LOG_LEVEL"] = '2'
import sys
import csv
import numpy as np
import pandas as pd
import random
from time import time, strftime, gmtime, sleep
from optparse import OptionParser
from pylsl import StreamInlet, resolve_byprop
from sklearn.linear_model import LinearRegression
import subprocess
currentpath = os.path.dirname(os.path.realpath(sys.argv[0]))
# dejitter timestamps
dejitter = False
# how long to wait for the Muse device to connect
muse_connect_timout = 5
parser = OptionParser()
parser.add_option("-d", "--duration",
dest="duration", type='int', default=10,
help="duration of the recording in seconds.")
parser.add_option("-p", "--path",
dest="path", type='str',
help="Directory for the recording file.")
parser.add_option("-s", "--sample",
dest="sample", type='str',
help="Record sample for specific term (1/2/3)")
(options, args) = parser.parse_args()
record_sample = False
record_sample_term = None
if options.sample:
record_sample = True
print("NOTICE: Creating sample dataset for term: " + str(options.sample))
record_sample_term = options.sample
else:
print("NOTICE: Creating training dataset with random terms")
if not options.path:
print("ERROR: please use -p to specify the datapath to the recorded csv files!")
sys.exit(1)
fname = (options.path + "/data_%s.csv" % strftime("%Y-%m-%d-%H.%M.%S", gmtime()))
# search for the last word id in eventually already existing datafiles
last_data_file = None
for file in sorted(os.listdir(options.path)):
if file.endswith(".csv"):
if not "eeg_data" in file:
print(os.path.join(options.path, file))
last_data_file = os.path.join(options.path, file)
if last_data_file:
print("Found existing datafiles! Getting currentWord from last datafiles: " + last_data_file)
line = subprocess.check_output(['tail', '-1', last_data_file])
line = str(line)
linesplit = line.split(",")
#print(linesplit[1])
print("Starting currentWord from " + str(linesplit[1]))
currentWord = int(linesplit[1])
currentWord = currentWord + 1
else:
print("Did not found any existing datafiles! Starting currentWord from 1")
currentWord = 1
print("-- currentWord: " + str(currentWord))
eeg_stream = False
print("looking for an EEG stream...")
streams = resolve_byprop('type', 'EEG', timeout=2)
if len(streams) == 0:
print("No EEG stream running yet. Trying to start the Muse EEG stream ...")
eeg_stream = subprocess.Popen([ currentpath + "/bci-stream"])
sleep(muse_connect_timout)
streams = resolve_byprop('type', 'EEG', timeout=2)
if len(streams) == 0:
raise(RuntimeError, "Cant find EEG stream")
else:
print("Success: found Muse EEG stream")
print("Start aquiring data")
inlet = StreamInlet(streams[0], max_chunklen=12)
eeg_time_correction = inlet.time_correction()
inlet_marker = False
#print("looking for a Markers stream...")
#marker_streams = resolve_byprop('type', 'Markers', timeout=2)
#if marker_streams:
# inlet_marker = StreamInlet(marker_streams[0])
# marker_time_correction = inlet_marker.time_correction()
#else:
# inlet_marker = False
# print("Cant find Markers stream")
info = inlet.info()
description = info.desc()
freq = info.nominal_srate()
Nchan = info.channel_count()
ch = description.child('channels').first_child()
ch_names = [ch.child_value('label')]
for i in range(1, Nchan):
ch = ch.next_sibling()
ch_names.append(ch.child_value('label'))
# Word Capturing
#currentWord = 1
currentTerm = "1"
t_word = time() + 1 * 2
words = []
terms = []
termBank = ["1", "2", "3"]
subdisplay = False
res = []
timestamps = []
markers = []
t_init = time()
print('Start recording at time t=%.3f' % t_init)
print(currentTerm)
while (time() - t_init) < options.duration:
if time() >= t_word:
if subdisplay:
subdisplay.kill()
# Check for new word
if time() >= t_word:
# sample or training data recording ?
if record_sample:
currentTerm = record_sample_term
else:
currentTerm = random.choice(termBank)
print(str(currentWord) +": " +currentTerm)
subdisplay = subprocess.Popen([ "/usr/bin/display", currentpath + "/images/" + currentTerm + ".png"])
currentWord += 1
t_word = time() + 1 * 2
try:
data, timestamp = inlet.pull_chunk(timeout=1.0, max_samples=12)
if timestamp:
res.append(data)
timestamps.extend(timestamp)
words.extend([currentWord] * len(timestamp))
terms.extend([currentTerm] * len(timestamp))
if inlet_marker:
marker, timestamp = inlet_marker.pull_sample(timeout=0.0)
if timestamp:
markers.append([marker, timestamp])
except KeyboardInterrupt:
break
if subdisplay:
subdisplay.kill()
res = np.concatenate(res, axis=0)
timestamps = np.array(timestamps)
if dejitter:
y = timestamps
X = np.atleast_2d(np.arange(0, len(y))).T
lr = LinearRegression()
lr.fit(X, y)
timestamps = lr.predict(X)
res = np.c_[timestamps, words, terms, res]
data = pd.DataFrame(data=res, columns=['timestamps'] + ['words'] + ['terms'] + ch_names)
data['Marker'] = 0
# process markers:
for marker in markers:
# find index of margers
ix = np.argmin(np.abs(marker[1] - timestamps))
val = timestamps[ix]
data.loc[ix, 'Marker'] = marker[0][0]
data.to_csv(fname, float_format='%.3f', index=False)
print('Wrote datafile: ' + fname)
if eeg_stream:
print("Found running EEG stream. Stopping it")
eeg_stream.kill()
print("Success")
|
[
"subprocess.check_output",
"numpy.abs",
"os.listdir",
"random.choice",
"pylsl.StreamInlet",
"subprocess.Popen",
"os.path.join",
"optparse.OptionParser",
"pylsl.resolve_byprop",
"time.sleep",
"os.path.realpath",
"numpy.array",
"time.gmtime",
"numpy.concatenate",
"sys.exit",
"pandas.DataFrame",
"time.time",
"sklearn.linear_model.LinearRegression"
] |
[((531, 545), 'optparse.OptionParser', 'OptionParser', ([], {}), '()\n', (543, 545), False, 'from optparse import OptionParser\n'), ((2457, 2497), 'pylsl.resolve_byprop', 'resolve_byprop', (['"""type"""', '"""EEG"""'], {'timeout': '(2)'}), "('type', 'EEG', timeout=2)\n", (2471, 2497), False, 'from pylsl import StreamInlet, resolve_byprop\n'), ((2916, 2956), 'pylsl.StreamInlet', 'StreamInlet', (['streams[0]'], {'max_chunklen': '(12)'}), '(streams[0], max_chunklen=12)\n', (2927, 2956), False, 'from pylsl import StreamInlet, resolve_byprop\n'), ((3822, 3828), 'time.time', 'time', ([], {}), '()\n', (3826, 3828), False, 'from time import time, strftime, gmtime, sleep\n'), ((5016, 5043), 'numpy.concatenate', 'np.concatenate', (['res'], {'axis': '(0)'}), '(res, axis=0)\n', (5030, 5043), True, 'import numpy as np\n'), ((5057, 5077), 'numpy.array', 'np.array', (['timestamps'], {}), '(timestamps)\n', (5065, 5077), True, 'import numpy as np\n'), ((5284, 5369), 'pandas.DataFrame', 'pd.DataFrame', ([], {'data': 'res', 'columns': "(['timestamps'] + ['words'] + ['terms'] + ch_names)"}), "(data=res, columns=['timestamps'] + ['words'] + ['terms'] +\n ch_names)\n", (5296, 5369), True, 'import pandas as pd\n'), ((375, 404), 'os.path.realpath', 'os.path.realpath', (['sys.argv[0]'], {}), '(sys.argv[0])\n', (391, 404), False, 'import os\n'), ((1424, 1435), 'sys.exit', 'sys.exit', (['(1)'], {}), '(1)\n', (1432, 1435), False, 'import sys\n'), ((1632, 1656), 'os.listdir', 'os.listdir', (['options.path'], {}), '(options.path)\n', (1642, 1656), False, 'import os\n'), ((1972, 2027), 'subprocess.check_output', 'subprocess.check_output', (["['tail', '-1', last_data_file]"], {}), "(['tail', '-1', last_data_file])\n", (1995, 2027), False, 'import subprocess\n'), ((2618, 2665), 'subprocess.Popen', 'subprocess.Popen', (["[currentpath + '/bci-stream']"], {}), "([currentpath + '/bci-stream'])\n", (2634, 2665), False, 'import subprocess\n'), ((2671, 2697), 'time.sleep', 'sleep', (['muse_connect_timout'], {}), '(muse_connect_timout)\n', (2676, 2697), False, 'from time import time, strftime, gmtime, sleep\n'), ((2712, 2752), 'pylsl.resolve_byprop', 'resolve_byprop', (['"""type"""', '"""EEG"""'], {'timeout': '(2)'}), "('type', 'EEG', timeout=2)\n", (2726, 2752), False, 'from pylsl import StreamInlet, resolve_byprop\n'), ((3691, 3697), 'time.time', 'time', ([], {}), '()\n', (3695, 3697), False, 'from time import time, strftime, gmtime, sleep\n'), ((5166, 5184), 'sklearn.linear_model.LinearRegression', 'LinearRegression', ([], {}), '()\n', (5182, 5184), False, 'from sklearn.linear_model import LinearRegression\n'), ((3904, 3910), 'time.time', 'time', ([], {}), '()\n', (3908, 3910), False, 'from time import time, strftime, gmtime, sleep\n'), ((3948, 3954), 'time.time', 'time', ([], {}), '()\n', (3952, 3954), False, 'from time import time, strftime, gmtime, sleep\n'), ((4048, 4054), 'time.time', 'time', ([], {}), '()\n', (4052, 4054), False, 'from time import time, strftime, gmtime, sleep\n'), ((4320, 4411), 'subprocess.Popen', 'subprocess.Popen', (["['/usr/bin/display', currentpath + '/images/' + currentTerm + '.png']"], {}), "(['/usr/bin/display', currentpath + '/images/' +\n currentTerm + '.png'])\n", (4336, 4411), False, 'import subprocess\n'), ((5475, 5505), 'numpy.abs', 'np.abs', (['(marker[1] - timestamps)'], {}), '(marker[1] - timestamps)\n', (5481, 5505), True, 'import numpy as np\n'), ((1508, 1516), 'time.gmtime', 'gmtime', ([], {}), '()\n', (1514, 1516), False, 'from time import time, strftime, gmtime, sleep\n'), ((1805, 1837), 'os.path.join', 'os.path.join', (['options.path', 'file'], {}), '(options.path, file)\n', (1817, 1837), False, 'import os\n'), ((4223, 4246), 'random.choice', 'random.choice', (['termBank'], {}), '(termBank)\n', (4236, 4246), False, 'import random\n'), ((4452, 4458), 'time.time', 'time', ([], {}), '()\n', (4456, 4458), False, 'from time import time, strftime, gmtime, sleep\n'), ((1742, 1774), 'os.path.join', 'os.path.join', (['options.path', 'file'], {}), '(options.path, file)\n', (1754, 1774), False, 'import os\n')]
|
# Hungarian algorithm (Kuhn-Munkres) for solving the linear sum assignment
# problem. Taken from scikit-learn. Based on original code by <NAME>,
# adapted to NumPy by <NAME>.
# Further improvements by <NAME>, <NAME> and <NAME>.
#
# Copyright (c) 2008 <NAME> <<EMAIL>>, <NAME>
# Author: <NAME>, <NAME>
# License: 3-clause BSD
import numpy as np
def linear_sum_assignment(cost_matrix):
"""Solve the linear sum assignment problem.
The linear sum assignment problem is also known as minimum weight matching
in bipartite graphs. A problem instance is described by a matrix C, where
each C[i,j] is the cost of matching vertex i of the first partite set
(a "worker") and vertex j of the second set (a "job"). The goal is to find
a complete assignment of workers to jobs of minimal cost.
Formally, let X be a boolean matrix where :math:`X[i,j] = 1` iff row i is
assigned to column j. Then the optimal assignment has cost
.. math::
\min \sum_i \sum_j C_{i,j} X_{i,j}
s.t. each row is assignment to at most one column, and each column to at
most one row.
This function can also solve a generalization of the classic assignment
problem where the cost matrix is rectangular. If it has more rows than
columns, then not every row needs to be assigned to a column, and vice
versa.
The method used is the Hungarian algorithm, also known as the Munkres or
Kuhn-Munkres algorithm.
Parameters
----------
cost_matrix : array
The cost matrix of the bipartite graph.
Returns
-------
row_ind, col_ind : array
An array of row indices and one of corresponding column indices giving
the optimal assignment. The cost of the assignment can be computed
as ``cost_matrix[row_ind, col_ind].sum()``. The row indices will be
sorted; in the case of a square cost matrix they will be equal to
``numpy.arange(cost_matrix.shape[0])``.
Notes
-----
.. versionadded:: 0.17.0
Examples
--------
>>> cost = np.array([[4, 1, 3], [2, 0, 5], [3, 2, 2]])
>>> from scipy.optimize import linear_sum_assignment
>>> row_ind, col_ind = linear_sum_assignment(cost)
>>> col_ind
array([1, 0, 2])
>>> cost[row_ind, col_ind].sum()
5
References
----------
1. http://csclab.murraystate.edu/bob.pilgrim/445/munkres.html
2. <NAME>. The Hungarian Method for the assignment problem.
*Naval Research Logistics Quarterly*, 2:83-97, 1955.
3. <NAME>. Variants of the Hungarian method for assignment
problems. *Naval Research Logistics Quarterly*, 3: 253-258, 1956.
4. <NAME>. Algorithms for the Assignment and Transportation Problems.
*J. SIAM*, 5(1):32-38, March, 1957.
5. https://en.wikipedia.org/wiki/Hungarian_algorithm
"""
cost_matrix = np.asarray(cost_matrix)
if len(cost_matrix.shape) != 2:
raise ValueError("expected a matrix (2-d array), got a %r array"
% (cost_matrix.shape,))
# The algorithm expects more columns than rows in the cost matrix.
if cost_matrix.shape[1] < cost_matrix.shape[0]:
cost_matrix = cost_matrix.T
transposed = True
else:
transposed = False
state = _Hungary(cost_matrix)
# No need to bother with assignments if one of the dimensions
# of the cost matrix is zero-length.
step = None if 0 in cost_matrix.shape else _step1
while step is not None:
step = step(state)
if transposed:
marked = state.marked.T
else:
marked = state.marked
return np.where(marked == 1)
class _Hungary(object):
"""State of the Hungarian algorithm.
Parameters
----------
cost_matrix : 2D matrix
The cost matrix. Must have shape[1] >= shape[0].
"""
def __init__(self, cost_matrix):
self.C = cost_matrix.copy()
n, m = self.C.shape
self.row_uncovered = np.ones(n, dtype=bool)
self.col_uncovered = np.ones(m, dtype=bool)
self.Z0_r = 0
self.Z0_c = 0
self.path = np.zeros((n + m, 2), dtype=int)
self.marked = np.zeros((n, m), dtype=int)
def _clear_covers(self):
"""Clear all covered matrix cells"""
self.row_uncovered[:] = True
self.col_uncovered[:] = True
# Individual steps of the algorithm follow, as a state machine: they return
# the next step to be taken (function to be called), if any.
def _step1(state):
"""Steps 1 and 2 in the Wikipedia page."""
# Step 1: For each row of the matrix, find the smallest element and
# subtract it from every element in its row.
state.C -= state.C.min(axis=1)[:, np.newaxis]
# Step 2: Find a zero (Z) in the resulting matrix. If there is no
# starred zero in its row or column, star Z. Repeat for each element
# in the matrix.
for i, j in zip(*np.where(state.C == 0)):
if state.col_uncovered[j] and state.row_uncovered[i]:
state.marked[i, j] = 1
state.col_uncovered[j] = False
state.row_uncovered[i] = False
state._clear_covers()
return _step3
def _step3(state):
"""
Cover each column containing a starred zero. If n columns are covered,
the starred zeros describe a complete set of unique assignments.
In this case, Go to DONE, otherwise, Go to Step 4.
"""
marked = (state.marked == 1)
state.col_uncovered[np.any(marked, axis=0)] = False
if marked.sum() < state.C.shape[0]:
return _step4
def _step4(state):
"""
Find a noncovered zero and prime it. If there is no starred zero
in the row containing this primed zero, Go to Step 5. Otherwise,
cover this row and uncover the column containing the starred
zero. Continue in this manner until there are no uncovered zeros
left. Save the smallest uncovered value and Go to Step 6.
"""
# We convert to int as numpy operations are faster on int
C = (state.C == 0).astype(int)
covered_C = C * state.row_uncovered[:, np.newaxis]
covered_C *= np.asarray(state.col_uncovered, dtype=int)
n = state.C.shape[0]
m = state.C.shape[1]
while True:
# Find an uncovered zero
row, col = np.unravel_index(np.argmax(covered_C), (n, m))
if covered_C[row, col] == 0:
return _step6
else:
state.marked[row, col] = 2
# Find the first starred element in the row
star_col = np.argmax(state.marked[row] == 1)
if state.marked[row, star_col] != 1:
# Could not find one
state.Z0_r = row
state.Z0_c = col
return _step5
else:
col = star_col
state.row_uncovered[row] = False
state.col_uncovered[col] = True
covered_C[:, col] = C[:, col] * (
np.asarray(state.row_uncovered, dtype=int))
covered_C[row] = 0
def _step5(state):
"""
Construct a series of alternating primed and starred zeros as follows.
Let Z0 represent the uncovered primed zero found in Step 4.
Let Z1 denote the starred zero in the column of Z0 (if any).
Let Z2 denote the primed zero in the row of Z1 (there will always be one).
Continue until the series terminates at a primed zero that has no starred
zero in its column. Unstar each starred zero of the series, star each
primed zero of the series, erase all primes and uncover every line in the
matrix. Return to Step 3
"""
count = 0
path = state.path
path[count, 0] = state.Z0_r
path[count, 1] = state.Z0_c
while True:
# Find the first starred element in the col defined by
# the path.
row = np.argmax(state.marked[:, path[count, 1]] == 1)
if state.marked[row, path[count, 1]] != 1:
# Could not find one
break
else:
count += 1
path[count, 0] = row
path[count, 1] = path[count - 1, 1]
# Find the first prime element in the row defined by the
# first path step
col = np.argmax(state.marked[path[count, 0]] == 2)
if state.marked[row, col] != 2:
col = -1
count += 1
path[count, 0] = path[count - 1, 0]
path[count, 1] = col
# Convert paths
for i in range(count + 1):
if state.marked[path[i, 0], path[i, 1]] == 1:
state.marked[path[i, 0], path[i, 1]] = 0
else:
state.marked[path[i, 0], path[i, 1]] = 1
state._clear_covers()
# Erase all prime markings
state.marked[state.marked == 2] = 0
return _step3
def _step6(state):
"""
Add the value found in Step 4 to every element of each covered row,
and subtract it from every element of each uncovered column.
Return to Step 4 without altering any stars, primes, or covered lines.
"""
# the smallest uncovered value in the matrix
if np.any(state.row_uncovered) and np.any(state.col_uncovered):
minval = np.min(state.C[state.row_uncovered], axis=0)
minval = np.min(minval[state.col_uncovered])
state.C[~state.row_uncovered] += minval
state.C[:, state.col_uncovered] -= minval
return _step4
|
[
"numpy.ones",
"numpy.where",
"numpy.asarray",
"numpy.argmax",
"numpy.any",
"numpy.zeros",
"numpy.min"
] |
[((2927, 2950), 'numpy.asarray', 'np.asarray', (['cost_matrix'], {}), '(cost_matrix)\n', (2937, 2950), True, 'import numpy as np\n'), ((3713, 3734), 'numpy.where', 'np.where', (['(marked == 1)'], {}), '(marked == 1)\n', (3721, 3734), True, 'import numpy as np\n'), ((6247, 6289), 'numpy.asarray', 'np.asarray', (['state.col_uncovered'], {'dtype': 'int'}), '(state.col_uncovered, dtype=int)\n', (6257, 6289), True, 'import numpy as np\n'), ((4074, 4096), 'numpy.ones', 'np.ones', (['n'], {'dtype': 'bool'}), '(n, dtype=bool)\n', (4081, 4096), True, 'import numpy as np\n'), ((4127, 4149), 'numpy.ones', 'np.ones', (['m'], {'dtype': 'bool'}), '(m, dtype=bool)\n', (4134, 4149), True, 'import numpy as np\n'), ((4217, 4248), 'numpy.zeros', 'np.zeros', (['(n + m, 2)'], {'dtype': 'int'}), '((n + m, 2), dtype=int)\n', (4225, 4248), True, 'import numpy as np\n'), ((4272, 4299), 'numpy.zeros', 'np.zeros', (['(n, m)'], {'dtype': 'int'}), '((n, m), dtype=int)\n', (4280, 4299), True, 'import numpy as np\n'), ((5595, 5617), 'numpy.any', 'np.any', (['marked'], {'axis': '(0)'}), '(marked, axis=0)\n', (5601, 5617), True, 'import numpy as np\n'), ((8001, 8048), 'numpy.argmax', 'np.argmax', (['(state.marked[:, path[count, 1]] == 1)'], {}), '(state.marked[:, path[count, 1]] == 1)\n', (8010, 8048), True, 'import numpy as np\n'), ((8386, 8430), 'numpy.argmax', 'np.argmax', (['(state.marked[path[count, 0]] == 2)'], {}), '(state.marked[path[count, 0]] == 2)\n', (8395, 8430), True, 'import numpy as np\n'), ((9258, 9285), 'numpy.any', 'np.any', (['state.row_uncovered'], {}), '(state.row_uncovered)\n', (9264, 9285), True, 'import numpy as np\n'), ((9290, 9317), 'numpy.any', 'np.any', (['state.col_uncovered'], {}), '(state.col_uncovered)\n', (9296, 9317), True, 'import numpy as np\n'), ((9337, 9381), 'numpy.min', 'np.min', (['state.C[state.row_uncovered]'], {'axis': '(0)'}), '(state.C[state.row_uncovered], axis=0)\n', (9343, 9381), True, 'import numpy as np\n'), ((9400, 9435), 'numpy.min', 'np.min', (['minval[state.col_uncovered]'], {}), '(minval[state.col_uncovered])\n', (9406, 9435), True, 'import numpy as np\n'), ((5032, 5054), 'numpy.where', 'np.where', (['(state.C == 0)'], {}), '(state.C == 0)\n', (5040, 5054), True, 'import numpy as np\n'), ((6432, 6452), 'numpy.argmax', 'np.argmax', (['covered_C'], {}), '(covered_C)\n', (6441, 6452), True, 'import numpy as np\n'), ((6663, 6696), 'numpy.argmax', 'np.argmax', (['(state.marked[row] == 1)'], {}), '(state.marked[row] == 1)\n', (6672, 6696), True, 'import numpy as np\n'), ((7106, 7148), 'numpy.asarray', 'np.asarray', (['state.row_uncovered'], {'dtype': 'int'}), '(state.row_uncovered, dtype=int)\n', (7116, 7148), True, 'import numpy as np\n')]
|
from abc import ABC, abstractmethod
import numpy as np
import pandas as pd
from pvrpm.core.enums import ConfigKeys as ck
from pvrpm.core.case import SamCase
from pvrpm.core.utils import sample, get_higher_components
from pvrpm.core.modules.monitor import IndepMonitor
class Failure(ABC):
"""
This abstract class defines how a failure should be set up
"""
def __init__(
self,
level: str,
comp_level_df: pd.DataFrame,
case: SamCase,
indep_monitoring: IndepMonitor = None,
):
"""
Initalizes a failure instance
Args:
level (str): The component level this failure is apart of
comp_level_df (:obj:`pd.DataFrame`): The component level dataframe containing the simulation data
case (:obj:`SamCase`): The SAM case for this simulation
indep_monitoring (:obj:`IndepMonitoring`, Optional): For updating static monitoring during simulation
"""
super().__init__()
self.level = level
self.df = comp_level_df
self.case = case
self.fails_per_day = {}
self.indep_monitoring = indep_monitoring
self.last_failure_day = 0
self.mean = None
self.initialize_components()
@abstractmethod
def initialize_components(self):
"""
Initalizes failure data for all components to be tracked during simulation for this failure type
Note:
Updates the underlying dataframes in place
"""
pass
@abstractmethod
def reinitialize_components(self, df: pd.DataFrame) -> pd.DataFrame:
"""
Reinitalize components in a dataframe similiar to the inital initalization. Used for when repairs or other things may occur
Args:
df (:obj:`pd.DataFrame`): The dataframe containing the components to reinitalize
Returns:
:obj:`pd.DataFrame`: The reinitalized components
"""
pass
@abstractmethod
def update(self, day: int):
"""
Perform a failure update for one day in the simulation:
Changes state of a component to failed, incrementing failures and checking warranty only for failed components of each failure type
Args:
day (int): Current day in the simulation
Note:
Updates the underlying dataframes in place
"""
pass
class TotalFailure(Failure):
"""
Describes how a total failure of a component should operate
"""
def initialize_components(self):
component_info = self.case.config[self.level]
df = self.df
failure_modes = list(component_info.get(ck.FAILURE, {}).keys())
self.mean = {} # init mean for each failure mode
possible_failure_times = np.zeros((component_info[ck.NUM_COMPONENT], len(failure_modes)))
for i, mode in enumerate(failure_modes):
self.mean[mode] = 0
# initalize failure mode by type
df[f"failure_by_type_{mode}"] = 0
fail = component_info[ck.FAILURE][mode]
if fail.get(ck.FRAC, None) or fail.get(ck.DECAY_FRAC, None):
frac = fail[ck.FRAC] if ck.FRAC in fail else fail[ck.DECAY_FRAC]
# choose a percentage of components to be defective
sample_ = np.random.random_sample(size=component_info[ck.NUM_COMPONENT])
defective = sample_ < frac
sample_ = sample(fail[ck.DIST], fail[ck.PARAM], component_info[ck.NUM_COMPONENT])
# only give a possible failure time if the module is defective, otherwise it is set to numpy max float value (which won't be used)
possible_failure_times[:, i] = np.where(list(defective), sample_, np.finfo(np.float32).max)
else:
# setup failure times for each component
possible_failure_times[:, i] = sample(fail[ck.DIST], fail[ck.PARAM], component_info[ck.NUM_COMPONENT])
# initalize failures per day for this failure mode
self.fails_per_day[mode] = np.zeros(self.case.config[ck.LIFETIME_YRS] * 365)
failure_ind = np.argmin(possible_failure_times, axis=1)
df["time_to_failure"] = np.amin(possible_failure_times, axis=1)
df["failure_type"] = [failure_modes[i] for i in failure_ind]
def reinitialize_components(self, df: pd.DataFrame) -> pd.DataFrame:
component_info = self.case.config[self.level]
failure_modes = list(component_info.get(ck.FAILURE, {}).keys())
fraction_failures = []
num_repaired = len(df)
possible_failure_times = np.zeros((num_repaired, len(failure_modes)))
for i, mode in enumerate(failure_modes):
fail = component_info[ck.FAILURE][mode]
if fail.get(ck.FRAC, None) or fail.get(ck.DECAY_FRAC, None):
frac = 0
if fail.get(ck.FRAC, None):
fraction_failures.append(mode)
frac = fail[ck.FRAC]
else:
frac = fail[ck.DECAY_FRAC]
# choose a percentage of modules to be defective
sample_ = np.random.random_sample(size=num_repaired)
defective = sample_ < frac
sample_ = sample(fail[ck.DIST], fail[ck.PARAM], num_repaired)
# only give a possible failure time if the module is defective, otherwise it is set to numpy max float value (which won't be used)
possible_failure_times[:, i] = np.where(
list(defective),
sample_,
np.finfo(np.float32).max,
)
else:
# setup failure times for each component
possible_failure_times[:, i] = sample(fail[ck.DIST], fail[ck.PARAM], num_repaired)
failure_ind = np.argmin(possible_failure_times, axis=1)
df["time_to_failure"] = np.amin(possible_failure_times, axis=1)
df["failure_type"] = [failure_modes[i] for i in failure_ind]
# now, need to make sure that our fractional failures percentages are met for all components in this level
# TODO: need to speed this up somehow
if fraction_failures:
# removes the diminishing effect where at the beginning of the simulation frac modules are a defective failure, then frac of frac is defective, etc.
# possible failure times will also include whatever the current failure time is for the component, if its less then a defective one it doesn't change
possible_failure_times = np.zeros((len(self.df), len(fraction_failures) + 1))
possible_failure_times.fill(np.finfo(np.float32).max)
# NOTE: i think i should just instead of doing the whole df, find the fraction, then sample that fraction from the components and just update those using the same method below
for i, mode in enumerate(fraction_failures):
counts = (self.df["failure_type"].astype(str) == mode).sum()
frac = counts / len(self.df)
fail = component_info[ck.FAILURE][mode]
if frac >= fail[ck.FRAC]:
continue
sample_ = np.random.random_sample(size=len(self.df))
# we just want the difference in fractions to bump it up to the failure fraction
defective = sample_ < (fail[ck.FRAC] - frac)
sample_ = sample(fail[ck.DIST], fail[ck.PARAM], len(self.df))
# only give a possible failure time if the module is defective, otherwise it is set to numpy max float value (which won't be used)
possible_failure_times[:, i] = np.where(
list(defective),
sample_,
np.finfo(np.float32).max,
)
possible_failure_times[:, -1] = self.df["time_to_failure"]
failure_ind = np.argmin(possible_failure_times, axis=1)
types = []
for comp, i in enumerate(failure_ind):
if i != len(fraction_failures):
types.append(fraction_failures[i])
else:
types.append(self.df["failure_type"].iloc[comp])
self.df["time_to_failure"] = np.amin(possible_failure_times, axis=1)
self.df["failure_type"] = np.array(types).astype(str)
return df
def update(self, day: int):
df = self.df
# decrement time to failures for operational modules
# TODO: change this to state > 0 once partial failures implemented
df["time_to_failure"] -= 1
failure_modes = list(self.case.config[self.level][ck.FAILURE].keys())
# TODO: change this to state > 0 once partial failures implemented
mask = (df["state"] == 1) & (df["time_to_failure"] < 1)
failed_comps = df.loc[mask].copy()
if len(failed_comps) > 0:
self.last_failure_day = day
failed_comps["time_to_failure"] = 0
failed_comps["cumulative_failures"] += 1
for fail in failure_modes:
fail_mask = failed_comps["failure_type"].astype(str) == fail
failed_comps.loc[fail_mask, f"failure_by_type_{fail}"] += 1
self.fails_per_day[fail][day] += len(failed_comps.loc[fail_mask])
warranty_mask = failed_comps["time_left_on_warranty"] <= 0
failed_comps.loc[warranty_mask, "cumulative_oow_failures"] += 1
failed_comps["state"] = 0
# update time to detection times for component levels with only independent monitoring
# which will have None for monitor times
try:
if failed_comps["monitor_times"].isnull().any():
# monitor and time to detection will be the time to next indep monitoring
indep_monitors = list(self.case.config[self.level][ck.INDEP_MONITOR].keys())
# next indep monitoring is the min of the possible indep monitors for this component level
failed_comps["monitor_times"] = np.amin(self.indep_monitoring.indep_monitoring[indep_monitors])
# in order to calculate the time to detection for component levels only monitoring by an
# independment monitoring with a threshold (no interval), need to instead
# set the nans that will be there to the day in the simulation when these components failed
# so it can be calculated later
failed_comps["monitor_times"] = failed_comps["monitor_times"].fillna(day)
failed_comps["time_to_detection"] = None # failed_comps["monitor_times"].copy()
# fails if no monitoring defined, faster then just doing a check if the column exists or whatever
except KeyError:
pass
df.loc[mask] = failed_comps
else:
# check to see when last failure was for fraction failure, and update components with new failures
# if its been longer then the mean time of the distribution
# this is so if repairs arent occuring due to poor monitoring, failures are still occuring
failure_modes = list(self.case.config[self.level].get(ck.FAILURE, {}).keys())
fraction_failures = []
for mode in failure_modes:
fail = self.case.config[self.level][ck.FAILURE][mode]
if fail.get(ck.FRAC, None):
# extract mean, since some distributions might not have mean defined in params
if self.mean[mode] == 0:
self.mean[mode] = sample(fail[ck.DIST], fail[ck.PARAM], 10000).mean()
if day > (self.mean[mode] + self.last_failure_day):
fraction_failures.append(mode)
self.last_failure_day = day
for mode in fraction_failures:
# fail new fraction of components
# possible failure times will also include whatever the current failure time is for the component, if its less then a defective one it doesn't change
possible_failure_times = np.zeros((len(self.df), len(fraction_failures) + 1))
possible_failure_times.fill(np.finfo(np.float32).max)
# NOTE: i think i should just instead of doing the whole df, find the fraction, then sample that fraction from the components and just update those using the same method below
for i, mode in enumerate(fraction_failures):
fail = self.case.config[self.level][ck.FAILURE][mode]
sample_ = np.random.random_sample(size=len(self.df))
defective = sample_ < fail[ck.FRAC]
sample_ = sample(fail[ck.DIST], fail[ck.PARAM], len(self.df))
# only give a possible failure time if the module is defective, otherwise it is set to numpy max float value (which won't be used)
possible_failure_times[:, i] = np.where(
list(defective),
sample_,
np.finfo(np.float32).max,
)
possible_failure_times[:, -1] = self.df["time_to_failure"]
failure_ind = np.argmin(possible_failure_times, axis=1)
types = []
for comp, i in enumerate(failure_ind):
if i != len(fraction_failures):
types.append(fraction_failures[i])
else:
types.append(self.df["failure_type"].iloc[comp])
self.df["time_to_failure"] = np.amin(possible_failure_times, axis=1)
self.df["failure_type"] = np.array(types).astype(str)
class PartialFailure(Failure):
"""
Specifies a decrease in the state of a component via a failure
Unlike total failures, every defined partial failure will have its own object, instead of manaing all of them at once
"""
def __init__(
self,
level: str,
comp_level_df: pd.DataFrame,
case: SamCase,
mode: str,
indep_monitoring: IndepMonitor = None,
):
"""
Initalizes a partial failure instance
Args:
level (str): The component level this failure is apart of
comp_level_df (:obj:`pd.DataFrame`): The component level dataframe containing the simulation data
case (:obj:`SamCase`): The SAM case for this simulation
mode (str): The name of the partial failure mode
indep_monitoring (:obj:`IndepMonitoring`, Optional): For updating static monitoring during simulation
"""
self.mode = mode
super().__init__(level, comp_level_df, case, indep_monitoring=indep_monitoring)
def initialize_components(self):
component_info = self.case.config[self.level]
df = self.df
mode = self.mode
failure_times = None
# initalize failure mode by type
df[f"failure_by_type_{mode}"] = 0
fail = component_info[ck.PARTIAL_FAIL][mode]
if fail.get(ck.FRAC, None) or fail.get(ck.DECAY_FRAC, None):
frac = fail[ck.FRAC] if ck.FRAC in fail else fail[ck.DECAY_FRAC]
# choose a percentage of components to be defective
sample_ = np.random.random_sample(size=component_info[ck.NUM_COMPONENT])
defective = sample_ < frac
sample_ = sample(fail[ck.DIST], fail[ck.PARAM], component_info[ck.NUM_COMPONENT])
# only give a possible failure time if the module is defective, otherwise it is set to numpy max float value (which won't be used)
failure_times = np.where(list(defective), sample_, np.nan)
else:
# setup failure times for each component
failure_times = sample(fail[ck.DIST], fail[ck.PARAM], component_info[ck.NUM_COMPONENT])
# initalize failures per day for this failure mode
self.fails_per_day = {self.mode: np.zeros(self.case.config[ck.LIFETIME_YRS] * 365)}
df[f"time_to_failure_{mode}"] = failure_times
def reinitialize_components(self, df: pd.DataFrame) -> pd.DataFrame:
component_info = self.case.config[self.level]
num_repaired = len(df)
fraction_failure = False
failure_times = None
mode = self.mode
fail = component_info[ck.PARTIAL_FAIL][mode]
if fail.get(ck.FRAC, None) or fail.get(ck.DECAY_FRAC, None):
if fail.get(ck.FRAC, None):
fraction_failure = True
frac = fail[ck.FRAC]
else:
frac = fail[ck.DECAY_FRAC]
# choose a percentage of modules to be defective
sample_ = np.random.random_sample(size=num_repaired)
defective = sample_ < frac
sample_ = sample(fail[ck.DIST], fail[ck.PARAM], num_repaired)
# only give a possible failure time if the module is defective, otherwise it is set to nan, partial failure is not applied
failure_times = np.where(list(defective), sample_, np.nan)
else:
# setup failure times for each component
failure_times = sample(fail[ck.DIST], fail[ck.PARAM], num_repaired)
df[f"time_to_failure_{mode}"] = failure_times
# now, need to make sure that our fractional failure percentage is met for all components in this level
# TODO: need to speed this up somehow
if fraction_failure:
# removes the diminishing effect where at the beginning of the simulation frac modules are a defective failure, then frac of frac is defective, etc.
# NOTE: i think i should just instead of doing the whole df, find the fraction, then sample that fraction from the components and just update those using the same method below
# number currently with failure mode is going to be the number non nan time_to_failures
counts = self.df[f"time_to_failure_{mode}"].isna()
update_df = self.df.loc[counts].copy()
frac = (~counts).sum() / len(self.df)
if frac >= fail[ck.FRAC]:
return df
sample_ = np.random.random_sample(size=len(update_df))
# we just want the difference in fractions to bump it up to the failure fraction
defective = sample_ < (fail[ck.FRAC] - frac)
sample_ = sample(fail[ck.DIST], fail[ck.PARAM], len(update_df))
# only give a possible failure time if the module is defective, otherwise it is set to numpy max float value (which won't be used)
failure_times = np.where(
list(defective),
sample_,
np.nan,
)
update_df[f"time_to_failure_{mode}"] = failure_times
self.df.loc[counts] = update_df
return df
def update(self, day: int):
df = self.df
# decrement time to failures
df[f"time_to_failure_{self.mode}"] -= 1
mask = (df["state"] == 1) & (df[f"time_to_failure_{self.mode}"] < 1)
failed_comps = df.loc[mask].copy()
if len(failed_comps) > 0:
self.last_failure_day = day
failed_comps["cumulative_failures"] += 1
failed_comps[f"failure_by_type_{self.mode}"] += 1
self.fails_per_day[self.mode][day] += len(failed_comps)
warranty_mask = failed_comps["time_left_on_warranty"] <= 0
failed_comps.loc[warranty_mask, "cumulative_oow_failures"] += 1
failed_comps["state"] = 0
# update time to detection times for component levels with only static monitoring
# which will have None for monitor times
try:
if failed_comps["monitor_times"].isnull().any():
# monitor and time to detection will be the time to next static monitoring
indep_monitors = list(self.case.config[self.level][ck.INDEP_MONITOR].keys())
# next static monitoring is the min of the possible static monitors for this component level
failed_comps["monitor_times"] = np.amin(self.indep_monitoring.indep_monitoring[indep_monitors])
# in order to calculate the time to detection for component levels only monitoring by an
# independment monitoring with a threshold (no interval), need to instead
# set the nans that will be there to the day in the simulation when these components failed
# so it can be calculated later
failed_comps["monitor_times"] = failed_comps["monitor_times"].fillna(day)
failed_comps["time_to_detection"] = None # failed_comps["monitor_times"].copy()
# fails if no monitoring defined, faster then just doing a check if the column exists or whatever
except KeyError:
pass
df.loc[mask] = failed_comps
else:
# check to see when last failure was for fraction failure, and update components with new failures
# if its been longer then the mean time of the distribution
# this is so if repairs arent occuring due to poor monitoring, failures are still occuring
fail = self.case.config[self.level][ck.PARTIAL_FAIL][self.mode]
if fail.get(ck.FRAC, None):
# extract mean, since some distributions might not have mean defined in params
if not self.mean:
self.mean = sample(fail[ck.DIST], fail[ck.PARAM], 10000).mean()
if day > (self.mean + self.last_failure_day):
# fail new fraction of components
counts = self.df[f"time_to_failure_{self.mode}"].isna()
update_df = self.df.loc[counts].copy()
sample_ = np.random.random_sample(size=len(update_df))
# we just want the difference in fractions to bump it up to the failure fraction
defective = sample_ < fail[ck.FRAC]
sample_ = sample(fail[ck.DIST], fail[ck.PARAM], len(update_df))
# only give a possible failure time if the module is defective, otherwise it is set to numpy max float value (which won't be used)
failure_times = np.where(
list(defective),
sample_,
np.nan,
)
update_df[f"time_to_failure_{self.mode}"] = failure_times
self.df.loc[counts] = update_df
self.last_failure_day = day
|
[
"numpy.random.random_sample",
"numpy.amin",
"pvrpm.core.utils.sample",
"numpy.array",
"numpy.zeros",
"numpy.finfo",
"numpy.argmin"
] |
[((4175, 4216), 'numpy.argmin', 'np.argmin', (['possible_failure_times'], {'axis': '(1)'}), '(possible_failure_times, axis=1)\n', (4184, 4216), True, 'import numpy as np\n'), ((4249, 4288), 'numpy.amin', 'np.amin', (['possible_failure_times'], {'axis': '(1)'}), '(possible_failure_times, axis=1)\n', (4256, 4288), True, 'import numpy as np\n'), ((5894, 5935), 'numpy.argmin', 'np.argmin', (['possible_failure_times'], {'axis': '(1)'}), '(possible_failure_times, axis=1)\n', (5903, 5935), True, 'import numpy as np\n'), ((5968, 6007), 'numpy.amin', 'np.amin', (['possible_failure_times'], {'axis': '(1)'}), '(possible_failure_times, axis=1)\n', (5975, 6007), True, 'import numpy as np\n'), ((4102, 4151), 'numpy.zeros', 'np.zeros', (['(self.case.config[ck.LIFETIME_YRS] * 365)'], {}), '(self.case.config[ck.LIFETIME_YRS] * 365)\n', (4110, 4151), True, 'import numpy as np\n'), ((7981, 8022), 'numpy.argmin', 'np.argmin', (['possible_failure_times'], {'axis': '(1)'}), '(possible_failure_times, axis=1)\n', (7990, 8022), True, 'import numpy as np\n'), ((8333, 8372), 'numpy.amin', 'np.amin', (['possible_failure_times'], {'axis': '(1)'}), '(possible_failure_times, axis=1)\n', (8340, 8372), True, 'import numpy as np\n'), ((15496, 15558), 'numpy.random.random_sample', 'np.random.random_sample', ([], {'size': 'component_info[ck.NUM_COMPONENT]'}), '(size=component_info[ck.NUM_COMPONENT])\n', (15519, 15558), True, 'import numpy as np\n'), ((15621, 15692), 'pvrpm.core.utils.sample', 'sample', (['fail[ck.DIST]', 'fail[ck.PARAM]', 'component_info[ck.NUM_COMPONENT]'], {}), '(fail[ck.DIST], fail[ck.PARAM], component_info[ck.NUM_COMPONENT])\n', (15627, 15692), False, 'from pvrpm.core.utils import sample, get_higher_components\n'), ((16004, 16075), 'pvrpm.core.utils.sample', 'sample', (['fail[ck.DIST]', 'fail[ck.PARAM]', 'component_info[ck.NUM_COMPONENT]'], {}), '(fail[ck.DIST], fail[ck.PARAM], component_info[ck.NUM_COMPONENT])\n', (16010, 16075), False, 'from pvrpm.core.utils import sample, get_higher_components\n'), ((16177, 16226), 'numpy.zeros', 'np.zeros', (['(self.case.config[ck.LIFETIME_YRS] * 365)'], {}), '(self.case.config[ck.LIFETIME_YRS] * 365)\n', (16185, 16226), True, 'import numpy as np\n'), ((16914, 16956), 'numpy.random.random_sample', 'np.random.random_sample', ([], {'size': 'num_repaired'}), '(size=num_repaired)\n', (16937, 16956), True, 'import numpy as np\n'), ((17019, 17070), 'pvrpm.core.utils.sample', 'sample', (['fail[ck.DIST]', 'fail[ck.PARAM]', 'num_repaired'], {}), '(fail[ck.DIST], fail[ck.PARAM], num_repaired)\n', (17025, 17070), False, 'from pvrpm.core.utils import sample, get_higher_components\n'), ((17373, 17424), 'pvrpm.core.utils.sample', 'sample', (['fail[ck.DIST]', 'fail[ck.PARAM]', 'num_repaired'], {}), '(fail[ck.DIST], fail[ck.PARAM], num_repaired)\n', (17379, 17424), False, 'from pvrpm.core.utils import sample, get_higher_components\n'), ((3343, 3405), 'numpy.random.random_sample', 'np.random.random_sample', ([], {'size': 'component_info[ck.NUM_COMPONENT]'}), '(size=component_info[ck.NUM_COMPONENT])\n', (3366, 3405), True, 'import numpy as np\n'), ((3476, 3547), 'pvrpm.core.utils.sample', 'sample', (['fail[ck.DIST]', 'fail[ck.PARAM]', 'component_info[ck.NUM_COMPONENT]'], {}), '(fail[ck.DIST], fail[ck.PARAM], component_info[ck.NUM_COMPONENT])\n', (3482, 3547), False, 'from pvrpm.core.utils import sample, get_higher_components\n'), ((3927, 3998), 'pvrpm.core.utils.sample', 'sample', (['fail[ck.DIST]', 'fail[ck.PARAM]', 'component_info[ck.NUM_COMPONENT]'], {}), '(fail[ck.DIST], fail[ck.PARAM], component_info[ck.NUM_COMPONENT])\n', (3933, 3998), False, 'from pvrpm.core.utils import sample, get_higher_components\n'), ((5196, 5238), 'numpy.random.random_sample', 'np.random.random_sample', ([], {'size': 'num_repaired'}), '(size=num_repaired)\n', (5219, 5238), True, 'import numpy as np\n'), ((5309, 5360), 'pvrpm.core.utils.sample', 'sample', (['fail[ck.DIST]', 'fail[ck.PARAM]', 'num_repaired'], {}), '(fail[ck.DIST], fail[ck.PARAM], num_repaired)\n', (5315, 5360), False, 'from pvrpm.core.utils import sample, get_higher_components\n'), ((5819, 5870), 'pvrpm.core.utils.sample', 'sample', (['fail[ck.DIST]', 'fail[ck.PARAM]', 'num_repaired'], {}), '(fail[ck.DIST], fail[ck.PARAM], num_repaired)\n', (5825, 5870), False, 'from pvrpm.core.utils import sample, get_higher_components\n'), ((13424, 13465), 'numpy.argmin', 'np.argmin', (['possible_failure_times'], {'axis': '(1)'}), '(possible_failure_times, axis=1)\n', (13433, 13465), True, 'import numpy as np\n'), ((13804, 13843), 'numpy.amin', 'np.amin', (['possible_failure_times'], {'axis': '(1)'}), '(possible_failure_times, axis=1)\n', (13811, 13843), True, 'import numpy as np\n'), ((6722, 6742), 'numpy.finfo', 'np.finfo', (['np.float32'], {}), '(np.float32)\n', (6730, 6742), True, 'import numpy as np\n'), ((8411, 8426), 'numpy.array', 'np.array', (['types'], {}), '(types)\n', (8419, 8426), True, 'import numpy as np\n'), ((10172, 10235), 'numpy.amin', 'np.amin', (['self.indep_monitoring.indep_monitoring[indep_monitors]'], {}), '(self.indep_monitoring.indep_monitoring[indep_monitors])\n', (10179, 10235), True, 'import numpy as np\n'), ((20340, 20403), 'numpy.amin', 'np.amin', (['self.indep_monitoring.indep_monitoring[indep_monitors]'], {}), '(self.indep_monitoring.indep_monitoring[indep_monitors])\n', (20347, 20403), True, 'import numpy as np\n'), ((3778, 3798), 'numpy.finfo', 'np.finfo', (['np.float32'], {}), '(np.float32)\n', (3786, 3798), True, 'import numpy as np\n'), ((5652, 5672), 'numpy.finfo', 'np.finfo', (['np.float32'], {}), '(np.float32)\n', (5660, 5672), True, 'import numpy as np\n'), ((7839, 7859), 'numpy.finfo', 'np.finfo', (['np.float32'], {}), '(np.float32)\n', (7847, 7859), True, 'import numpy as np\n'), ((12393, 12413), 'numpy.finfo', 'np.finfo', (['np.float32'], {}), '(np.float32)\n', (12401, 12413), True, 'import numpy as np\n'), ((13886, 13901), 'numpy.array', 'np.array', (['types'], {}), '(types)\n', (13894, 13901), True, 'import numpy as np\n'), ((13270, 13290), 'numpy.finfo', 'np.finfo', (['np.float32'], {}), '(np.float32)\n', (13278, 13290), True, 'import numpy as np\n'), ((21744, 21788), 'pvrpm.core.utils.sample', 'sample', (['fail[ck.DIST]', 'fail[ck.PARAM]', '(10000)'], {}), '(fail[ck.DIST], fail[ck.PARAM], 10000)\n', (21750, 21788), False, 'from pvrpm.core.utils import sample, get_higher_components\n'), ((11763, 11807), 'pvrpm.core.utils.sample', 'sample', (['fail[ck.DIST]', 'fail[ck.PARAM]', '(10000)'], {}), '(fail[ck.DIST], fail[ck.PARAM], 10000)\n', (11769, 11807), False, 'from pvrpm.core.utils import sample, get_higher_components\n')]
|
'''
load lottery tickets and evaluation
support datasets: cifar10, Fashionmnist, cifar100
'''
import os
import time
import random
import shutil
import argparse
import numpy as np
from copy import deepcopy
import matplotlib.pyplot as plt
import torch
import torch.optim
import torch.nn as nn
import torch.utils.data
import torch.nn.functional as F
import torchvision.models as models
import torch.backends.cudnn as cudnn
import torchvision.transforms as transforms
import torchvision.datasets as datasets
from torch.utils.data.sampler import SubsetRandomSampler
from advertorch.utils import NormalizeByChannelMeanStd
from utils import *
from pruning_utils_2 import *
from pruning_utils_unprune import *
parser = argparse.ArgumentParser(description='PyTorch Evaluation Tickets')
##################################### general setting #################################################
parser.add_argument('--data', type=str, default='../../data', help='location of the data corpus')
parser.add_argument('--dataset', type=str, default='cifar10', help='dataset')
parser.add_argument('--arch', type=str, default='res18', help='model architecture')
parser.add_argument('--seed', default=None, type=int, help='random seed')
parser.add_argument('--save_dir', help='The directory used to save the trained models', default=None, type=str)
parser.add_argument('--gpu', type=int, default=0, help='gpu device id')
parser.add_argument('--save_model', action="store_true", help="whether saving model")
##################################### training setting #################################################
parser.add_argument('--optim', type=str, default='sgd', help='optimizer')
parser.add_argument('--batch_size', type=int, default=128, help='batch size')
parser.add_argument('--lr', default=0.1, type=float, help='initial learning rate')
parser.add_argument('--momentum', default=0.9, type=float, help='momentum')
parser.add_argument('--weight_decay', default=1e-4, type=float, help='weight decay')
parser.add_argument('--epochs', default=182, type=int, help='number of total epochs to run')
parser.add_argument('--warmup', default=0, type=int, help='warm up epochs')
parser.add_argument('--print_freq', default=50, type=int, help='print frequency')
parser.add_argument('--decreasing_lr', default='91,136', help='decreasing strategy')
##################################### Pruning setting #################################################
parser.add_argument('--pretrained', default=None, type=str, help='pretrained weight for pt')
parser.add_argument('--mask_dir', default=None, type=str, help='mask direction for ticket')
parser.add_argument('--conv1', action="store_true", help="whether pruning&rewind conv1")
parser.add_argument('--fc', action="store_true", help="whether rewind fc")
parser.add_argument('--type', type=str, default=None, choices=['ewp', 'random_path', 'betweenness', 'hessian_abs', 'taylor1_abs','intgrads','identity', 'omp'])
parser.add_argument('--add-back', action="store_true", help="add back weights")
parser.add_argument('--prune-type', type=str, choices=["lt", 'pt', 'st', 'mt', 'trained', 'transfer'])
parser.add_argument('--num-paths', default=50000, type=int)
parser.add_argument('--evaluate', action="store_true")
parser.add_argument('--evaluate-p', type=float, default=0.00)
parser.add_argument('--evaluate-random', action="store_true")
parser.add_argument('--evaluate-full', action="store_true")
parser.add_argument('--checkpoint', type=str)
best_sa = 0
def main():
global args, best_sa
args = parser.parse_args()
print(args)
print('*'*50)
print('conv1 included for prune and rewind: {}'.format(args.conv1))
print('fc included for rewind: {}'.format(args.fc))
print('*'*50)
torch.cuda.set_device(int(args.gpu))
os.makedirs(args.save_dir, exist_ok=True)
if args.seed:
setup_seed(args.seed)
# prepare dataset
model, train_loader, val_loader, test_loader = setup_model_dataset(args)
criterion = nn.CrossEntropyLoss()
if args.evaluate:
state_dict = torch.load(args.checkpoint, map_location="cpu")['state_dict']
if not args.evaluate_full:
current_mask = extract_mask(state_dict)
print(current_mask.keys())
prune_model_custom(model, current_mask, conv1=False)
check_sparsity(model, conv1=False)
try:
model.load_state_dict(state_dict)
except:
state_dict['normalize.mean'] = model.state_dict()['normalize.mean']
state_dict['normalize.std'] = model.state_dict()['normalize.std']
model.load_state_dict(state_dict)
model.cuda()
validate(val_loader, model, criterion)
if args.evaluate_p > 0:
pruning_model(model, args.evaluate_p, random=args.evaluate_random)
check_sparsity(model, conv1=False)
tacc = validate(val_loader, model, criterion)
# evaluate on test set
test_tacc = validate(test_loader, model, criterion)
print(tacc)
print(test_tacc)
return
#loading tickets
model.cuda()
load_ticket(model, args)
decreasing_lr = list(map(int, args.decreasing_lr.split(',')))
optimizer = torch.optim.SGD(model.parameters(), args.lr,
momentum=args.momentum,
weight_decay=args.weight_decay)
scheduler = torch.optim.lr_scheduler.MultiStepLR(optimizer, milestones=decreasing_lr, gamma=0.1)
all_result = {}
all_result['train'] = []
all_result['test_ta'] = []
all_result['ta'] = []
start_epoch = 0
remain_weight = check_sparsity(model, conv1=args.conv1)
for epoch in range(start_epoch, args.epochs):
print(optimizer.state_dict()['param_groups'][0]['lr'])
acc = train(train_loader, model, criterion, optimizer, epoch)
# evaluate on validation set
tacc = validate(val_loader, model, criterion)
# evaluate on test set
test_tacc = validate(test_loader, model, criterion)
scheduler.step()
all_result['train'].append(acc)
all_result['ta'].append(tacc)
all_result['test_ta'].append(test_tacc)
all_result['remain_weight'] = remain_weight
# remember best prec@1 and save checkpoint
is_best_sa = tacc > best_sa
best_sa = max(tacc, best_sa)
if args.save_model:
save_checkpoint({
'result': all_result,
'epoch': epoch + 1,
'state_dict': model.state_dict(),
'best_sa': best_sa,
'optimizer': optimizer.state_dict(),
'scheduler': scheduler.state_dict()
}, is_SA_best=is_best_sa, save_path=args.save_dir)
else:
save_checkpoint({
'result': all_result
}, is_SA_best=False, save_path=args.save_dir)
plt.plot(all_result['train'], label='train_acc')
plt.plot(all_result['ta'], label='val_acc')
plt.plot(all_result['test_ta'], label='test_acc')
plt.legend()
plt.savefig(os.path.join(args.save_dir, 'net_train.png'))
plt.close()
check_sparsity(model, conv1=args.conv1)
print('* best SA={}'.format(all_result['test_ta'][np.argmax(np.array(all_result['ta']))]))
def train(train_loader, model, criterion, optimizer, epoch):
losses = AverageMeter()
top1 = AverageMeter()
# switch to train mode
model.train()
start = time.time()
for i, (image, target) in enumerate(train_loader):
if epoch < args.warmup:
warmup_lr(epoch, i+1, optimizer, one_epoch_step=len(train_loader))
image = image.cuda()
target = target.cuda()
# compute output
output_clean = model(image)
loss = criterion(output_clean, target)
optimizer.zero_grad()
loss.backward()
optimizer.step()
output = output_clean.float()
loss = loss.float()
# measure accuracy and record loss
prec1 = accuracy(output.data, target)[0]
losses.update(loss.item(), image.size(0))
top1.update(prec1.item(), image.size(0))
if i % args.print_freq == 0:
end = time.time()
print('Epoch: [{0}][{1}/{2}]\t'
'Loss {loss.val:.4f} ({loss.avg:.4f})\t'
'Accuracy {top1.val:.3f} ({top1.avg:.3f})\t'
'Time {3:.2f}'.format(
epoch, i, len(train_loader), end-start, loss=losses, top1=top1))
start = time.time()
print('train_accuracy {top1.avg:.3f}'.format(top1=top1))
return top1.avg
def validate(val_loader, model, criterion):
"""
Run evaluation
"""
losses = AverageMeter()
top1 = AverageMeter()
# switch to evaluate mode
model.eval()
for i, (image, target) in enumerate(val_loader):
image = image.cuda()
target = target.cuda()
# compute output
with torch.no_grad():
output = model(image)
loss = criterion(output, target)
output = output.float()
loss = loss.float()
# measure accuracy and record loss
prec1 = accuracy(output.data, target)[0]
losses.update(loss.item(), image.size(0))
top1.update(prec1.item(), image.size(0))
if i % args.print_freq == 0:
print('Test: [{0}/{1}]\t'
'Loss {loss.val:.4f} ({loss.avg:.4f})\t'
'Accuracy {top1.val:.3f} ({top1.avg:.3f})'.format(
i, len(val_loader), loss=losses, top1=top1))
print('valid_accuracy {top1.avg:.3f}'
.format(top1=top1))
return top1.avg
def save_checkpoint(state, is_SA_best, save_path, filename='checkpoint.pth.tar'):
filepath = os.path.join(save_path, filename)
torch.save(state, filepath)
if is_SA_best:
shutil.copyfile(filepath, os.path.join(save_path, 'model_SA_best.pth.tar'))
def load_ticket(model, args):
# weight
if args.pretrained:
initalization = torch.load(args.pretrained, map_location = torch.device('cuda:'+str(args.gpu)))
if 'init_weight' in initalization.keys():
print('loading from init_weight')
initalization = initalization['init_weight']
elif 'state_dict' in initalization.keys():
print('loading from state_dict')
initalization = initalization['state_dict']
loading_weight = extract_main_weight(initalization, fc=True, conv1=True)
new_initialization = model.state_dict()
if not 'normalize.std' in loading_weight:
loading_weight['normalize.std'] = new_initialization['normalize.std']
loading_weight['normalize.mean'] = new_initialization['normalize.mean']
if not (args.prune_type == 'lt' or args.prune_type == 'trained'):
keys = list(loading_weight.keys())
for key in keys:
if key.startswith('fc') or key.startswith('conv1'):
del loading_weight[key]
loading_weight['fc.weight'] = new_initialization['fc.weight']
loading_weight['fc.bias'] = new_initialization['fc.bias']
loading_weight['conv1.weight'] = new_initialization['conv1.weight']
print('*number of loading weight={}'.format(len(loading_weight.keys())))
print('*number of model weight={}'.format(len(model.state_dict().keys())))
model.load_state_dict(loading_weight)
# mask
if args.mask_dir:
print('loading mask')
current_mask_weight = torch.load(args.mask_dir, map_location = torch.device('cuda:'+str(args.gpu)))
if 'state_dict' in current_mask_weight.keys():
current_mask_weight = current_mask_weight['state_dict']
current_mask = extract_mask(current_mask_weight)
#check_sparsity(model, conv1=args.conv1)
if args.arch == 'res18':
downsample = 100
else:
downsample = 1000
custom_prune(model, current_mask, args.type, args.num_paths, args, args.add_back)
#prune_random_betweeness(model, current_mask, int(args.num_paths), downsample=downsample, conv1=args.conv1)
check_sparsity(model, conv1=args.conv1)
def warmup_lr(epoch, step, optimizer, one_epoch_step):
overall_steps = args.warmup*one_epoch_step
current_steps = epoch*one_epoch_step + step
lr = args.lr * current_steps/overall_steps
lr = min(lr, args.lr)
for p in optimizer.param_groups:
p['lr']=lr
class AverageMeter(object):
"""Computes and stores the average and current value"""
def __init__(self):
self.reset()
def reset(self):
self.val = 0
self.avg = 0
self.sum = 0
self.count = 0
def update(self, val, n=1):
self.val = val
self.sum += val * n
self.count += n
self.avg = self.sum / self.count
def accuracy(output, target, topk=(1,)):
"""Computes the precision@k for the specified values of k"""
maxk = max(topk)
batch_size = target.size(0)
_, pred = output.topk(maxk, 1, True, True)
pred = pred.t()
correct = pred.eq(target.view(1, -1).expand_as(pred))
res = []
for k in topk:
correct_k = correct[:k].view(-1).float().sum(0)
res.append(correct_k.mul_(100.0 / batch_size))
return res
def setup_seed(seed):
torch.manual_seed(seed)
torch.cuda.manual_seed_all(seed)
np.random.seed(seed)
random.seed(seed)
torch.backends.cudnn.deterministic = True
if __name__ == '__main__':
main()
|
[
"torch.cuda.manual_seed_all",
"torch.manual_seed",
"torch.optim.lr_scheduler.MultiStepLR",
"os.makedirs",
"torch.nn.CrossEntropyLoss",
"argparse.ArgumentParser",
"torch.load",
"matplotlib.pyplot.plot",
"os.path.join",
"random.seed",
"matplotlib.pyplot.close",
"numpy.array",
"numpy.random.seed",
"torch.save",
"torch.no_grad",
"time.time",
"matplotlib.pyplot.legend"
] |
[((719, 784), 'argparse.ArgumentParser', 'argparse.ArgumentParser', ([], {'description': '"""PyTorch Evaluation Tickets"""'}), "(description='PyTorch Evaluation Tickets')\n", (742, 784), False, 'import argparse\n'), ((3787, 3828), 'os.makedirs', 'os.makedirs', (['args.save_dir'], {'exist_ok': '(True)'}), '(args.save_dir, exist_ok=True)\n', (3798, 3828), False, 'import os\n'), ((3999, 4020), 'torch.nn.CrossEntropyLoss', 'nn.CrossEntropyLoss', ([], {}), '()\n', (4018, 4020), True, 'import torch.nn as nn\n'), ((5437, 5525), 'torch.optim.lr_scheduler.MultiStepLR', 'torch.optim.lr_scheduler.MultiStepLR', (['optimizer'], {'milestones': 'decreasing_lr', 'gamma': '(0.1)'}), '(optimizer, milestones=decreasing_lr,\n gamma=0.1)\n', (5473, 5525), False, 'import torch\n'), ((7533, 7544), 'time.time', 'time.time', ([], {}), '()\n', (7542, 7544), False, 'import time\n'), ((9838, 9871), 'os.path.join', 'os.path.join', (['save_path', 'filename'], {}), '(save_path, filename)\n', (9850, 9871), False, 'import os\n'), ((9876, 9903), 'torch.save', 'torch.save', (['state', 'filepath'], {}), '(state, filepath)\n', (9886, 9903), False, 'import torch\n'), ((13467, 13490), 'torch.manual_seed', 'torch.manual_seed', (['seed'], {}), '(seed)\n', (13484, 13490), False, 'import torch\n'), ((13496, 13528), 'torch.cuda.manual_seed_all', 'torch.cuda.manual_seed_all', (['seed'], {}), '(seed)\n', (13522, 13528), False, 'import torch\n'), ((13534, 13554), 'numpy.random.seed', 'np.random.seed', (['seed'], {}), '(seed)\n', (13548, 13554), True, 'import numpy as np\n'), ((13560, 13577), 'random.seed', 'random.seed', (['seed'], {}), '(seed)\n', (13571, 13577), False, 'import random\n'), ((6947, 6995), 'matplotlib.pyplot.plot', 'plt.plot', (["all_result['train']"], {'label': '"""train_acc"""'}), "(all_result['train'], label='train_acc')\n", (6955, 6995), True, 'import matplotlib.pyplot as plt\n'), ((7004, 7047), 'matplotlib.pyplot.plot', 'plt.plot', (["all_result['ta']"], {'label': '"""val_acc"""'}), "(all_result['ta'], label='val_acc')\n", (7012, 7047), True, 'import matplotlib.pyplot as plt\n'), ((7056, 7105), 'matplotlib.pyplot.plot', 'plt.plot', (["all_result['test_ta']"], {'label': '"""test_acc"""'}), "(all_result['test_ta'], label='test_acc')\n", (7064, 7105), True, 'import matplotlib.pyplot as plt\n'), ((7114, 7126), 'matplotlib.pyplot.legend', 'plt.legend', ([], {}), '()\n', (7124, 7126), True, 'import matplotlib.pyplot as plt\n'), ((7201, 7212), 'matplotlib.pyplot.close', 'plt.close', ([], {}), '()\n', (7210, 7212), True, 'import matplotlib.pyplot as plt\n'), ((4065, 4112), 'torch.load', 'torch.load', (['args.checkpoint'], {'map_location': '"""cpu"""'}), "(args.checkpoint, map_location='cpu')\n", (4075, 4112), False, 'import torch\n'), ((7147, 7191), 'os.path.join', 'os.path.join', (['args.save_dir', '"""net_train.png"""'], {}), "(args.save_dir, 'net_train.png')\n", (7159, 7191), False, 'import os\n'), ((8277, 8288), 'time.time', 'time.time', ([], {}), '()\n', (8286, 8288), False, 'import time\n'), ((8595, 8606), 'time.time', 'time.time', ([], {}), '()\n', (8604, 8606), False, 'import time\n'), ((9034, 9049), 'torch.no_grad', 'torch.no_grad', ([], {}), '()\n', (9047, 9049), False, 'import torch\n'), ((9957, 10005), 'os.path.join', 'os.path.join', (['save_path', '"""model_SA_best.pth.tar"""'], {}), "(save_path, 'model_SA_best.pth.tar')\n", (9969, 10005), False, 'import os\n'), ((7322, 7348), 'numpy.array', 'np.array', (["all_result['ta']"], {}), "(all_result['ta'])\n", (7330, 7348), True, 'import numpy as np\n')]
|
import time
from absl import app, flags, logging
from absl.flags import FLAGS
import cv2
import numpy as np
import tensorflow as tf
from yolov3_tf2.models import (
YoloV3, YoloV3Tiny
)
from yolov3_tf2.dataset import transform_images, load_tfrecord_dataset
from yolov3_tf2.utils import draw_outputs
flags.DEFINE_string('classes', './data/vocmine.names', 'path to classes file')
flags.DEFINE_string('weights', './checkpoints/yolov3_train_9.tf',
'path to weights file')
flags.DEFINE_boolean('tiny', False, 'yolov3 or yolov3-tiny')
flags.DEFINE_integer('size', 416, 'resize images to')
flags.DEFINE_string('image', './data/girl.png', 'path to input image')
flags.DEFINE_string('tfrecord', None, 'tfrecord instead of image')
flags.DEFINE_string('output', './output.jpg', 'path to output image')
flags.DEFINE_integer('num_classes', 80, 'number of classes in the model')
def main(_argv):
physical_devices = tf.config.experimental.list_physical_devices('GPU')
for physical_device in physical_devices:
tf.config.experimental.set_memory_growth(physical_device, True)
if FLAGS.tiny:
yolo = YoloV3Tiny(classes=FLAGS.num_classes)
else:
yolo = YoloV3(classes=FLAGS.num_classes)
yolo.load_weights(FLAGS.weights).expect_partial()
logging.info('weights loaded')
class_names = [c.strip() for c in open(FLAGS.classes).readlines()]
logging.info('classes loaded')
if FLAGS.tfrecord:
dataset = load_tfrecord_dataset(
FLAGS.tfrecord, FLAGS.classes, FLAGS.size)
dataset = dataset.shuffle(512)
img_raw, _label = next(iter(dataset.take(1)))
else:
img_raw = tf.image.decode_image(
open(FLAGS.image, 'rb').read(), channels=3)
img = tf.expand_dims(img_raw, 0)
img = transform_images(img, FLAGS.size)
t1 = time.time()
boxes, scores, classes, nums = yolo(img)
t2 = time.time()
logging.info('time: {}'.format(t2 - t1))
logging.info('detections:')
for i in range(nums[0]):
logging.info('\t{}, {}, {}'.format(class_names[int(classes[0][i])],
np.array(scores[0][i]),
np.array(boxes[0][i])))
img = cv2.cvtColor(img_raw.numpy(), cv2.COLOR_RGB2BGR)
img = draw_outputs(img, (boxes, scores, classes, nums), class_names)
cv2.imwrite(FLAGS.output, img)
logging.info('output saved to: {}'.format(FLAGS.output))
if __name__ == '__main__':
try:
app.run(main)
except SystemExit:
pass
|
[
"cv2.imwrite",
"yolov3_tf2.dataset.transform_images",
"tensorflow.config.experimental.set_memory_growth",
"absl.flags.DEFINE_integer",
"absl.logging.info",
"absl.flags.DEFINE_boolean",
"absl.app.run",
"numpy.array",
"yolov3_tf2.dataset.load_tfrecord_dataset",
"yolov3_tf2.utils.draw_outputs",
"time.time",
"tensorflow.expand_dims",
"yolov3_tf2.models.YoloV3",
"absl.flags.DEFINE_string",
"yolov3_tf2.models.YoloV3Tiny",
"tensorflow.config.experimental.list_physical_devices"
] |
[((303, 381), 'absl.flags.DEFINE_string', 'flags.DEFINE_string', (['"""classes"""', '"""./data/vocmine.names"""', '"""path to classes file"""'], {}), "('classes', './data/vocmine.names', 'path to classes file')\n", (322, 381), False, 'from absl import app, flags, logging\n'), ((382, 475), 'absl.flags.DEFINE_string', 'flags.DEFINE_string', (['"""weights"""', '"""./checkpoints/yolov3_train_9.tf"""', '"""path to weights file"""'], {}), "('weights', './checkpoints/yolov3_train_9.tf',\n 'path to weights file')\n", (401, 475), False, 'from absl import app, flags, logging\n'), ((492, 552), 'absl.flags.DEFINE_boolean', 'flags.DEFINE_boolean', (['"""tiny"""', '(False)', '"""yolov3 or yolov3-tiny"""'], {}), "('tiny', False, 'yolov3 or yolov3-tiny')\n", (512, 552), False, 'from absl import app, flags, logging\n'), ((553, 606), 'absl.flags.DEFINE_integer', 'flags.DEFINE_integer', (['"""size"""', '(416)', '"""resize images to"""'], {}), "('size', 416, 'resize images to')\n", (573, 606), False, 'from absl import app, flags, logging\n'), ((607, 677), 'absl.flags.DEFINE_string', 'flags.DEFINE_string', (['"""image"""', '"""./data/girl.png"""', '"""path to input image"""'], {}), "('image', './data/girl.png', 'path to input image')\n", (626, 677), False, 'from absl import app, flags, logging\n'), ((678, 744), 'absl.flags.DEFINE_string', 'flags.DEFINE_string', (['"""tfrecord"""', 'None', '"""tfrecord instead of image"""'], {}), "('tfrecord', None, 'tfrecord instead of image')\n", (697, 744), False, 'from absl import app, flags, logging\n'), ((745, 814), 'absl.flags.DEFINE_string', 'flags.DEFINE_string', (['"""output"""', '"""./output.jpg"""', '"""path to output image"""'], {}), "('output', './output.jpg', 'path to output image')\n", (764, 814), False, 'from absl import app, flags, logging\n'), ((815, 888), 'absl.flags.DEFINE_integer', 'flags.DEFINE_integer', (['"""num_classes"""', '(80)', '"""number of classes in the model"""'], {}), "('num_classes', 80, 'number of classes in the model')\n", (835, 888), False, 'from absl import app, flags, logging\n'), ((931, 982), 'tensorflow.config.experimental.list_physical_devices', 'tf.config.experimental.list_physical_devices', (['"""GPU"""'], {}), "('GPU')\n", (975, 982), True, 'import tensorflow as tf\n'), ((1291, 1321), 'absl.logging.info', 'logging.info', (['"""weights loaded"""'], {}), "('weights loaded')\n", (1303, 1321), False, 'from absl import app, flags, logging\n'), ((1398, 1428), 'absl.logging.info', 'logging.info', (['"""classes loaded"""'], {}), "('classes loaded')\n", (1410, 1428), False, 'from absl import app, flags, logging\n'), ((1760, 1786), 'tensorflow.expand_dims', 'tf.expand_dims', (['img_raw', '(0)'], {}), '(img_raw, 0)\n', (1774, 1786), True, 'import tensorflow as tf\n'), ((1797, 1830), 'yolov3_tf2.dataset.transform_images', 'transform_images', (['img', 'FLAGS.size'], {}), '(img, FLAGS.size)\n', (1813, 1830), False, 'from yolov3_tf2.dataset import transform_images, load_tfrecord_dataset\n'), ((1841, 1852), 'time.time', 'time.time', ([], {}), '()\n', (1850, 1852), False, 'import time\n'), ((1907, 1918), 'time.time', 'time.time', ([], {}), '()\n', (1916, 1918), False, 'import time\n'), ((1969, 1996), 'absl.logging.info', 'logging.info', (['"""detections:"""'], {}), "('detections:')\n", (1981, 1996), False, 'from absl import app, flags, logging\n'), ((2306, 2368), 'yolov3_tf2.utils.draw_outputs', 'draw_outputs', (['img', '(boxes, scores, classes, nums)', 'class_names'], {}), '(img, (boxes, scores, classes, nums), class_names)\n', (2318, 2368), False, 'from yolov3_tf2.utils import draw_outputs\n'), ((2373, 2403), 'cv2.imwrite', 'cv2.imwrite', (['FLAGS.output', 'img'], {}), '(FLAGS.output, img)\n', (2384, 2403), False, 'import cv2\n'), ((1036, 1099), 'tensorflow.config.experimental.set_memory_growth', 'tf.config.experimental.set_memory_growth', (['physical_device', '(True)'], {}), '(physical_device, True)\n', (1076, 1099), True, 'import tensorflow as tf\n'), ((1135, 1172), 'yolov3_tf2.models.YoloV3Tiny', 'YoloV3Tiny', ([], {'classes': 'FLAGS.num_classes'}), '(classes=FLAGS.num_classes)\n', (1145, 1172), False, 'from yolov3_tf2.models import YoloV3, YoloV3Tiny\n'), ((1198, 1231), 'yolov3_tf2.models.YoloV3', 'YoloV3', ([], {'classes': 'FLAGS.num_classes'}), '(classes=FLAGS.num_classes)\n', (1204, 1231), False, 'from yolov3_tf2.models import YoloV3, YoloV3Tiny\n'), ((1471, 1535), 'yolov3_tf2.dataset.load_tfrecord_dataset', 'load_tfrecord_dataset', (['FLAGS.tfrecord', 'FLAGS.classes', 'FLAGS.size'], {}), '(FLAGS.tfrecord, FLAGS.classes, FLAGS.size)\n', (1492, 1535), False, 'from yolov3_tf2.dataset import transform_images, load_tfrecord_dataset\n'), ((2511, 2524), 'absl.app.run', 'app.run', (['main'], {}), '(main)\n', (2518, 2524), False, 'from absl import app, flags, logging\n'), ((2145, 2167), 'numpy.array', 'np.array', (['scores[0][i]'], {}), '(scores[0][i])\n', (2153, 2167), True, 'import numpy as np\n'), ((2212, 2233), 'numpy.array', 'np.array', (['boxes[0][i]'], {}), '(boxes[0][i])\n', (2220, 2233), True, 'import numpy as np\n')]
|
import matplotlib.pyplot as plt
from time import time
import numpy as np
from .plotter_utils import figure_ratio, xarray_set_axes_labels, retrieve_or_create_fig_ax
# Change the bands (RGB) here if you want other false color combinations
def rgb(dataset, at_index=0, x_coord='longitude', y_coord='latitude',
bands=['red', 'green', 'blue'], paint_on_mask = [],
min_possible=0, max_possible=10000, use_data_min=False,
use_data_max=False, min_inten=0.15, max_inten=1.0,
width=10, fig=None, ax=None, imshow_kwargs=None):
"""
Creates a figure showing an area, using three specified bands as the rgb componenets.
Parameters
----------
dataset: xarray.Dataset
A Dataset containing at least latitude and longitude coordinates and optionally time.
The coordinate order should be time, latitude, and finally longitude.
Must contain the data variables specified in the `bands` parameter.
at_index: int
The time index to show.
x_coord, y_coord, time_coord: str
Names of DataArrays in `dataset_in` to use as x, y, and time coordinates.
bands: list-like
A list-like containing 3 names of data variables in `dataset` to use as the red, green, and blue
bands, respectively.
min_possible, max_possible: int
The minimum and maximum valid values for the selected bands according to
the platform used to retrieve the data in `dataset`.
For example, for Landsat these are generally 0 and 10000, respectively.
use_data_min: bool
Whether to use `min_possible` or the minimum among all selected bands
as the band value which has a minimal intensity.
use_data_max: bool
Whether to use `max_possible` or the maximum among all selected bands
as the band value which has a maximal intensity.
min_inten, max_inten: float
The min and max intensities for any band. These can be in range [0,1].
These can be used to brighten or darken the image.
width: int
The width of the figure in inches.
fig: matplotlib.figure.Figure
The figure to use for the plot.
If only `fig` is supplied, the Axes object used will be the first.
ax: matplotlib.axes.Axes
The axes to use for the plot.
imshow_kwargs: dict
The dictionary of keyword arguments passed to `ax.imshow()`.
You can pass a colormap here with the key 'cmap'.
Returns
-------
fig, ax: matplotlib.figure.Figure, matplotlib.axes.Axes
The figure and axes used for the plot.
"""
imshow_kwargs = {} if imshow_kwargs is None else imshow_kwargs
### < Dataset to RGB Format, needs float values between 0-1
rgb = np.stack([dataset[bands[0]],
dataset[bands[1]],
dataset[bands[2]]], axis = -1)
# Interpolate values to be in the range [0,1] for creating the image.
min_rgb = np.nanmin(rgb) if use_data_min else min_possible
max_rgb = np.nanmax(rgb) if use_data_max else max_possible
rgb = np.interp(rgb, (min_rgb, max_rgb), [min_inten,max_inten])
rgb = rgb.astype(float)
### >
### < takes a T/F mask, apply a color to T areas
for mask, color in paint_on_mask:
rgb[mask] = np.array(color)/ 255.0
### >
fig, ax = retrieve_or_create_fig_ax(fig, ax, figsize=figure_ratio(rgb.shape[:2], fixed_width = width))
xarray_set_axes_labels(dataset, ax, x_coord, y_coord)
if 'time' in dataset.dims:
ax.imshow(rgb[at_index], **imshow_kwargs)
else:
ax.imshow(rgb, **imshow_kwargs)
return fig, ax
|
[
"numpy.stack",
"numpy.array",
"numpy.nanmax",
"numpy.interp",
"numpy.nanmin"
] |
[((2732, 2808), 'numpy.stack', 'np.stack', (['[dataset[bands[0]], dataset[bands[1]], dataset[bands[2]]]'], {'axis': '(-1)'}), '([dataset[bands[0]], dataset[bands[1]], dataset[bands[2]]], axis=-1)\n', (2740, 2808), True, 'import numpy as np\n'), ((3061, 3119), 'numpy.interp', 'np.interp', (['rgb', '(min_rgb, max_rgb)', '[min_inten, max_inten]'], {}), '(rgb, (min_rgb, max_rgb), [min_inten, max_inten])\n', (3070, 3119), True, 'import numpy as np\n'), ((2939, 2953), 'numpy.nanmin', 'np.nanmin', (['rgb'], {}), '(rgb)\n', (2948, 2953), True, 'import numpy as np\n'), ((3002, 3016), 'numpy.nanmax', 'np.nanmax', (['rgb'], {}), '(rgb)\n', (3011, 3016), True, 'import numpy as np\n'), ((3284, 3299), 'numpy.array', 'np.array', (['color'], {}), '(color)\n', (3292, 3299), True, 'import numpy as np\n')]
|
import logging
import os
import numpy as np
import xml.etree.ElementTree as ET
from PIL import Image
from paths import DATASETS_ROOT
log = logging.getLogger()
VOC_CATS = ['__background__', 'aeroplane', 'bicycle', 'bird', 'boat', 'bottle',
'bus', 'car', 'cat', 'chair', 'cow', 'diningtable', 'dog', 'horse',
'motorbike', 'person', 'pottedplant', 'sheep', 'sofa', 'train',
'tvmonitor']
class VOCLoader():
def __init__(self, year, split, segmentation=False, augmented_seg=False):
assert year in ['07', '12']
self.dataset = 'voc'
self.year = year
self.root = os.path.join(DATASETS_ROOT, 'VOCdevkit/VOC20%s/' % year)
self.split = split
assert split in ['train', 'val', 'trainval', 'test']
cats = VOC_CATS
self.cats_to_ids = dict(map(reversed, enumerate(cats)))
self.ids_to_cats = dict(enumerate(cats))
self.num_classes = len(cats)
self.categories = cats[1:]
self.segmentation = segmentation
self.augmented_seg = augmented_seg
assert not self.segmentation or self.segmentation and self.year == '12'
if self.augmented_seg:
filelist = 'ImageSets/SegmentationAug/%s.txt'
elif self.segmentation:
filelist = 'ImageSets/Segmentation/%s.txt'
else:
filelist = 'ImageSets/Main/%s.txt'
with open(os.path.join(self.root, filelist % self.split), 'r') as f:
self.filenames = f.read().split('\n')[:-1]
log.info("Created a loader VOC%s %s with %i images" % (year, split, len(self.filenames)))
def load_image(self, name):
im = Image.open('%sJPEGImages/%s.jpg' % (self.root, name)).convert('RGB')
im = np.array(im) / 255.0
im = im.astype(np.float32)
return im
def get_filenames(self):
return self.filenames
def read_annotations(self, name):
bboxes = []
cats = []
tree = ET.parse('%sAnnotations/%s.xml' % (self.root, name))
root = tree.getroot()
width = int(root.find('size/width').text)
height = int(root.find('size/height').text)
difficulty = []
for obj in root.findall('object'):
cat = self.cats_to_ids[obj.find('name').text]
difficult = (int(obj.find('difficult').text) != 0)
difficulty.append(difficult)
cats.append(cat)
bbox_tag = obj.find('bndbox')
x = int(bbox_tag.find('xmin').text)
y = int(bbox_tag.find('ymin').text)
w = int(bbox_tag.find('xmax').text)-x
h = int(bbox_tag.find('ymax').text)-y
bboxes.append((x, y, w, h))
gt_cats = np.array(cats)
gt_bboxes = np.array(bboxes).reshape((len(bboxes), 4))
difficulty = np.array(difficulty)
seg_gt = self.read_segmentations(name, height, width)
output = gt_bboxes, seg_gt, gt_cats, width, height, difficulty
return output
def read_segmentations(self, name, height, width):
if self.segmentation:
try:
seg_folder = self.root + 'SegmentationClass/'
seg_file = seg_folder + name + '.png'
seg_map = Image.open(seg_file)
except:
assert self.augmented_seg
seg_folder = self.root + 'SegmentationClassAug/'
seg_file = seg_folder + name + '.png'
seg_map = Image.open(seg_file)
segmentation = np.array(seg_map, dtype=np.uint8)
else:
# if there is no segmentation for a particular image we fill the mask
# with zeros to keep the same amount of tensors but don't learn from it
segmentation = np.zeros([height, width], dtype=np.uint8) + 255
return segmentation
|
[
"logging.getLogger",
"PIL.Image.open",
"xml.etree.ElementTree.parse",
"os.path.join",
"numpy.array",
"numpy.zeros"
] |
[((143, 162), 'logging.getLogger', 'logging.getLogger', ([], {}), '()\n', (160, 162), False, 'import logging\n'), ((634, 690), 'os.path.join', 'os.path.join', (['DATASETS_ROOT', "('VOCdevkit/VOC20%s/' % year)"], {}), "(DATASETS_ROOT, 'VOCdevkit/VOC20%s/' % year)\n", (646, 690), False, 'import os\n'), ((1978, 2030), 'xml.etree.ElementTree.parse', 'ET.parse', (["('%sAnnotations/%s.xml' % (self.root, name))"], {}), "('%sAnnotations/%s.xml' % (self.root, name))\n", (1986, 2030), True, 'import xml.etree.ElementTree as ET\n'), ((2718, 2732), 'numpy.array', 'np.array', (['cats'], {}), '(cats)\n', (2726, 2732), True, 'import numpy as np\n'), ((2817, 2837), 'numpy.array', 'np.array', (['difficulty'], {}), '(difficulty)\n', (2825, 2837), True, 'import numpy as np\n'), ((1751, 1763), 'numpy.array', 'np.array', (['im'], {}), '(im)\n', (1759, 1763), True, 'import numpy as np\n'), ((3516, 3549), 'numpy.array', 'np.array', (['seg_map'], {'dtype': 'np.uint8'}), '(seg_map, dtype=np.uint8)\n', (3524, 3549), True, 'import numpy as np\n'), ((1411, 1457), 'os.path.join', 'os.path.join', (['self.root', '(filelist % self.split)'], {}), '(self.root, filelist % self.split)\n', (1423, 1457), False, 'import os\n'), ((1669, 1722), 'PIL.Image.open', 'Image.open', (["('%sJPEGImages/%s.jpg' % (self.root, name))"], {}), "('%sJPEGImages/%s.jpg' % (self.root, name))\n", (1679, 1722), False, 'from PIL import Image\n'), ((2753, 2769), 'numpy.array', 'np.array', (['bboxes'], {}), '(bboxes)\n', (2761, 2769), True, 'import numpy as np\n'), ((3240, 3260), 'PIL.Image.open', 'Image.open', (['seg_file'], {}), '(seg_file)\n', (3250, 3260), False, 'from PIL import Image\n'), ((3757, 3798), 'numpy.zeros', 'np.zeros', (['[height, width]'], {'dtype': 'np.uint8'}), '([height, width], dtype=np.uint8)\n', (3765, 3798), True, 'import numpy as np\n'), ((3468, 3488), 'PIL.Image.open', 'Image.open', (['seg_file'], {}), '(seg_file)\n', (3478, 3488), False, 'from PIL import Image\n')]
|
import numpy as np
from skmultiflow.drift_detection import ADWIN
def demo():
""" _test_adwin
In this demo, an ADWIN object evaluates a sequence of numbers corresponding to 2 distributions.
The ADWIN object indicates the indices where change is detected.
The first half of the data is a sequence of randomly generated 0's and 1's.
The second half of the data is a normal distribution of integers from 0 to 7.
"""
adwin = ADWIN()
size = 2000
change_start = 999
np.random.seed(1)
data_stream = np.random.randint(2, size=size)
data_stream[change_start:] = np.random.randint(8, size=size-change_start)
for i in range(size):
adwin.add_element(data_stream[i])
if adwin.detected_change():
print('Change has been detected in data: ' + str(data_stream[i]) + ' - of index: ' + str(i))
if __name__ == '__main__':
demo()
|
[
"skmultiflow.drift_detection.ADWIN",
"numpy.random.randint",
"numpy.random.seed"
] |
[((463, 470), 'skmultiflow.drift_detection.ADWIN', 'ADWIN', ([], {}), '()\n', (468, 470), False, 'from skmultiflow.drift_detection import ADWIN\n'), ((514, 531), 'numpy.random.seed', 'np.random.seed', (['(1)'], {}), '(1)\n', (528, 531), True, 'import numpy as np\n'), ((550, 581), 'numpy.random.randint', 'np.random.randint', (['(2)'], {'size': 'size'}), '(2, size=size)\n', (567, 581), True, 'import numpy as np\n'), ((615, 661), 'numpy.random.randint', 'np.random.randint', (['(8)'], {'size': '(size - change_start)'}), '(8, size=size - change_start)\n', (632, 661), True, 'import numpy as np\n')]
|
# ==========================================================================
#
# Copyright NumFOCUS
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0.txt
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
# ==========================================================================*/
import re
from typing import Optional, Union, Dict, Any, List, Tuple, Sequence, TYPE_CHECKING
from sys import stderr as system_error_stream
import numpy as np
try:
from numpy.typing import ArrayLike
except ImportError:
from numpy import ndarray as ArrayLike
import warnings
from sys import stderr as system_error_stream
import os
import builtins
fileiotype = Union[str, bytes, os.PathLike]
import itk.support.types as itkt
from .helpers import wasm_type_from_image_type, image_type_from_wasm_type
from .helpers import wasm_type_from_mesh_type, mesh_type_from_wasm_type, python_to_js
from .helpers import wasm_type_from_pointset_type, pointset_type_from_wasm_type
if TYPE_CHECKING:
try:
import xarray as xr
except ImportError:
pass
try:
import vtk
except ImportError:
pass
__all__ = [
"output",
"image",
"set_nthreads",
"get_nthreads",
"echo",
"size",
"physical_size",
"spacing",
"origin",
"index",
"region",
"GetArrayFromImage",
"array_from_image",
"GetArrayViewFromImage",
"array_view_from_image",
"GetImageFromArray",
"image_from_array",
"GetImageViewFromArray",
"image_view_from_array",
"array_from_vector_container",
"array_view_from_vector_container",
"vector_container_from_array",
"GetArrayFromVnlVector",
"array_from_vnl_vector",
"GetVnlVectorFromArray",
"vnl_vector_from_array",
"GetArrayViewFromVnlVector",
"array_view_from_vnl_vector",
"GetVnlMatrixFromArray",
"vnl_matrix_from_array",
"GetArrayFromVnlMatrix",
"array_from_vnl_matrix",
"GetArrayViewFromVnlMatrix",
"array_view_from_vnl_matrix",
"GetArrayFromMatrix",
"array_from_matrix",
"GetMatrixFromArray",
"matrix_from_array",
"xarray_from_image",
"image_from_xarray",
"vtk_image_from_image",
"image_from_vtk_image",
"dict_from_image",
"image_from_dict",
"image_intensity_min_max",
"imwrite",
"imread",
"meshwrite",
"meshread",
"mesh_from_dict",
"dict_from_mesh",
"pointset_from_dict",
"dict_from_pointset",
"transformwrite",
"transformread",
"search",
"set_inputs",
"templated_class",
"pipeline",
"auto_pipeline",
"down_cast",
"template",
"class_",
"ctype",
"python_type",
"range",
"TemplateTypeError",
]
def output(input):
"""
If input object has attribute "GetOutput()" then return an itk image,
otherwise this function simply returns the input value
"""
if hasattr(input, "GetOutput"):
return input.GetOutput()
return input
def image(input):
warnings.warn(
"WrapITK warning: itk.image() is deprecated. " "Use itk.output() instead."
)
return output(input)
def set_nthreads(number_of_threads: int) -> None:
"""
Support convenient set of the number of threads.
Use example (in python):
import itk
itk.set_nthreads(4) ## use 4 threads
"""
assert number_of_threads > 0, (
"Please set a positive number of threads instead of %d" % number_of_threads
)
import itk
threader = itk.MultiThreaderBase.New()
threader.SetGlobalDefaultNumberOfThreads(number_of_threads)
def get_nthreads() -> int:
"""
Get the number of threads
"""
import itk
threader = itk.MultiThreaderBase.New()
return threader.GetGlobalDefaultNumberOfThreads()
def echo(obj, f=system_error_stream) -> None:
"""Print an object to stream
If the object has a method Print(), this method is used.
repr(obj) is used otherwise
"""
print(f, obj)
def size(image_or_filter: "itkt.ImageOrImageSource") -> Sequence[int]:
"""Return the size of an image, or of the output image of a filter
This method take care of updating the needed information
"""
# we don't need the entire output, only its size
import itk
image_or_filter.UpdateOutputInformation()
img = itk.output(image_or_filter)
return img.GetLargestPossibleRegion().GetSize()
def physical_size(image_or_filter: "itkt.ImageOrImageSource") -> Sequence[float]:
"""Return the physical size of an image, or of the output image of a filter
This method take care of updating the needed information
"""
# required because range is overloaded in this module
from builtins import range
spacing_ = spacing(image_or_filter)
size_ = size(image_or_filter)
result = []
for i in range(0, spacing_.Size()):
result.append(spacing_.GetElement(i) * size_.GetElement(i))
return result
def spacing(image_or_filter: "itkt.ImageOrImageSource") -> Sequence[float]:
"""Return the spacing of an image, or of the output image of a filter
This method take care of updating the needed information
"""
import itk
# we don't need the entire output, only its size
image_or_filter.UpdateOutputInformation()
img = itk.output(image_or_filter)
return img.GetSpacing()
def origin(image_or_filter: "itkt.ImageOrImageSource") -> Sequence[float]:
"""Return the origin of an image, or of the output image of a filter
This method take care of updating the needed information
"""
import itk
# we don't need the entire output, only its size
image_or_filter.UpdateOutputInformation()
img = itk.output(image_or_filter)
return img.GetOrigin()
def index(image_or_filter: "itkt.ImageOrImageSource") -> Sequence[int]:
"""Return the index of an image, or of the output image of a filter
This method take care of updating the needed information
"""
import itk
# we don't need the entire output, only its size
image_or_filter.UpdateOutputInformation()
img = itk.output(image_or_filter)
return img.GetLargestPossibleRegion().GetIndex()
def region(image_or_filter: "itkt.ImageOrImageSource") -> "itkt.ImageRegion":
"""Return the region of an image, or of the output image of a filter
This method take care of updating the needed information
"""
import itk
# we don't need the entire output, only its size
image_or_filter.UpdateOutputInformation()
img = itk.output(image_or_filter)
return img.GetLargestPossibleRegion()
def _get_itk_pixelid(numpy_array_type):
"""Returns a ITK PixelID given a numpy array."""
import itk
def _long_type():
if os.name == "nt":
return itk.ULL
else:
return itk.UL
# This is a Mapping from numpy array types to itk pixel types.
_np_itk = {
np.uint8: itk.UC,
np.uint16: itk.US,
np.uint32: itk.UI,
np.uint64: _long_type(),
np.int8: itk.SC,
np.int16: itk.SS,
np.int32: itk.SI,
np.int64: itk.SL,
np.float32: itk.F,
np.float64: itk.D,
np.complex64: itk.complex[itk.F],
np.complex128: itk.complex[itk.D],
}
try:
return _np_itk[numpy_array_type.dtype.type]
except KeyError as e:
for key in _np_itk:
if np.issubdtype(numpy_array_type.dtype.type, key):
return _np_itk[key]
raise e
def _GetArrayFromImage(
image_or_filter, function_name: str, keep_axes: bool, update: bool, ttype
) -> np.ndarray:
"""Get an Array with the content of the image buffer"""
# Finds the image type
import itk
img = itk.output(image_or_filter)
if ttype is not None:
if isinstance(ttype, (tuple, list)):
if len(ttype) != 1:
raise RuntimeError("Expected 1 component in ttype tuple.")
ImageType = ttype[0]
else:
ImageType = ttype
else:
ImageType = img.__class__
keys = [k for k in itk.PyBuffer.keys() if k[0] == ImageType]
if len(keys) == 0:
raise RuntimeError("No suitable template parameter can be found.")
# Create a numpy array of the type of the input image
templatedFunction = getattr(itk.PyBuffer[keys[0]], function_name)
return templatedFunction(img, keep_axes, update)
def GetArrayFromImage(
image_or_filter: "itkt.ImageOrImageSource",
keep_axes: bool = False,
update: bool = True,
ttype=None,
) -> np.ndarray:
"""Get an array with the content of the image buffer"""
return _GetArrayFromImage(
image_or_filter, "GetArrayFromImage", keep_axes, update, ttype
)
array_from_image = GetArrayFromImage
def GetArrayViewFromImage(
image_or_filter: "itkt.ImageOrImageSource",
keep_axes: bool = False,
update: bool = True,
ttype=None,
) -> np.ndarray:
"""Get an array view with the content of the image buffer"""
return _GetArrayFromImage(
image_or_filter, "GetArrayViewFromImage", keep_axes, update, ttype
)
array_view_from_image = GetArrayViewFromImage
def _GetImageFromArray(arr: ArrayLike, function_name: str, is_vector: bool, ttype):
"""Get an ITK image from a Python array."""
import itk
# Verify inputs
if not isinstance(arr, np.ndarray):
arr = np.asarray(arr)
if ttype is not None:
if is_vector:
raise RuntimeError("Cannot specify both `is_vector` and `ttype`.")
if isinstance(ttype, (tuple, list)):
if len(ttype) != 1:
raise RuntimeError("Expected 1 component in ttype tuple.")
ImageType = ttype[0]
else:
ImageType = ttype
if type(itk.template(ImageType)) != tuple or len(itk.template(ImageType)) < 2:
raise RuntimeError("Cannot determine pixel type from supplied ttype.")
is_vector = (
type(itk.template(ImageType)[1][0]) != itk.support.types.itkCType
or itk.template(ImageType)[0] == itk.VectorImage
)
else:
PixelType = _get_itk_pixelid(arr)
Dimension = arr.ndim
if is_vector:
Dimension = arr.ndim - 1
if arr.flags["C_CONTIGUOUS"]:
VectorDimension = arr.shape[-1]
else:
VectorDimension = arr.shape[0]
if PixelType == itk.UC:
if VectorDimension == 3:
ImageType = itk.Image[itk.RGBPixel[itk.UC], Dimension]
elif VectorDimension == 4:
ImageType = itk.Image[itk.RGBAPixel[itk.UC], Dimension]
else:
ImageType = itk.VectorImage[PixelType, Dimension]
else:
ImageType = itk.VectorImage[PixelType, Dimension]
else:
ImageType = itk.Image[PixelType, Dimension]
keys = [k for k in itk.PyBuffer.keys() if k[0] == ImageType]
if len(keys) == 0:
raise RuntimeError(
"""No suitable template parameter can be found.
Please specify an output type via the 'ttype' keyword parameter."""
)
templatedFunction = getattr(itk.PyBuffer[keys[0]], function_name)
return templatedFunction(arr, is_vector)
def GetImageFromArray(
arr: ArrayLike, is_vector: bool = False, ttype=None
) -> "itkt.ImageBase":
"""Get an ITK image from a Python array."""
return _GetImageFromArray(arr, "GetImageFromArray", is_vector, ttype)
image_from_array = GetImageFromArray
def GetImageViewFromArray(
arr: ArrayLike, is_vector: bool = False, ttype=None
) -> "itkt.ImageBase":
"""Get an ITK image view from a Python array."""
return _GetImageFromArray(arr, "GetImageViewFromArray", is_vector, ttype)
image_view_from_array = GetImageViewFromArray
def array_from_vector_container(
container: "itkt.VectorContainer", ttype=None
) -> np.ndarray:
"""Get an Array with the content of the vector container"""
import itk
container_template = itk.template(container)
IndexType = container_template[1][0]
# Find container data type
if ttype is not None:
if isinstance(ttype, (tuple, list)):
if len(ttype) != 1:
raise RuntimeError("Expected 1 component in ttype tuple.")
DataType = ttype[0]
else:
DataType = ttype
else:
DataType = container_template[1][1]
keys = [k for k in itk.PyVectorContainer.keys() if k == (IndexType, DataType)]
if len(keys) == 0:
raise RuntimeError("No suitable template parameter can be found.")
# Create numpy array of the type of the input container
return itk.PyVectorContainer[keys[0]].array_from_vector_container(container)
def array_view_from_vector_container(
container: "itkt.VectorContainer", ttype=None
) -> np.ndarray:
"""Get an Array view with the content of the vector container"""
import itk
container_template = itk.template(container)
IndexType = container_template[1][0]
# Find container type
if ttype is not None:
if isinstance(ttype, (tuple, list)):
if len(ttype) != 1:
raise RuntimeError("Expected 1 component in ttype tuple.")
DataType = ttype[0]
else:
DataType = ttype
else:
DataType = container_template[1][1]
keys = [k for k in itk.PyVectorContainer.keys() if k == (IndexType, DataType)]
if len(keys) == 0:
raise RuntimeError("No suitable template parameter can be found.")
# Create numpy array of the type of the input container
return itk.PyVectorContainer[keys[0]].array_view_from_vector_container(container)
def vector_container_from_array(arr: ArrayLike, ttype=None) -> "itkt.VectorContainer":
"""Get a vector container from a Python array"""
import itk
# Verify inputs
if not isinstance(arr, np.ndarray):
arr = np.asarray(arr)
# Return VectorContainer with 64-bit index type
if os.name == "nt":
IndexType = itk.ULL
else:
IndexType = itk.UL
# Find container type
if ttype is not None:
if isinstance(ttype, (tuple, list)):
if len(ttype) != 1:
raise RuntimeError("Expected 1 component in ttype tuple.")
DataType = ttype[0]
else:
DataType = ttype
else:
DataType = _get_itk_pixelid(arr)
keys = [k for k in itk.PyVectorContainer.keys() if k == (IndexType, DataType)]
if len(keys) == 0:
raise RuntimeError("No suitable template parameter can be found.")
# Create numpy array of the type of the input container
return itk.PyVectorContainer[keys[0]].vector_container_from_array(arr)
def _GetArrayFromVnlObject(vnl_object, function_name: str, ttype) -> np.ndarray:
"""Get an array with the content of vnl_object"""
# Finds the vnl object type
import itk
if ttype is not None:
if isinstance(ttype, (tuple, list)):
if len(ttype) != 1:
raise RuntimeError("Expected 1 component in ttype tuple.")
PixelType = ttype[0]
else:
PixelType = ttype
else:
PixelType = itk.template(vnl_object)[1][0]
keys = [k for k in itk.PyVnl.keys() if k[0] == PixelType]
if len(keys) == 0:
raise RuntimeError("No suitable template parameter can be found.")
# Create a numpy array of the type of the vnl object
templatedFunction = getattr(itk.PyVnl[keys[0]], function_name)
return templatedFunction(vnl_object)
def GetArrayFromVnlVector(vnl_vector, ttype=None) -> np.ndarray:
"""Get an array with the content of vnl_vector"""
return _GetArrayFromVnlObject(vnl_vector, "GetArrayFromVnlVector", ttype)
array_from_vnl_vector = GetArrayFromVnlVector
def GetArrayViewFromVnlVector(vnl_vector, ttype=None) -> np.ndarray:
"""Get an array view of vnl_vector"""
return _GetArrayFromVnlObject(vnl_vector, "GetArrayViewFromVnlVector", ttype)
array_view_from_vnl_vector = GetArrayFromVnlVector
def GetArrayFromVnlMatrix(vnl_matrix, ttype=None) -> np.ndarray:
"""Get an array with the content of vnl_matrix"""
return _GetArrayFromVnlObject(vnl_matrix, "GetArrayFromVnlMatrix", ttype)
array_from_vnl_matrix = GetArrayFromVnlMatrix
def GetArrayViewFromVnlMatrix(vnl_matrix, ttype=None) -> np.ndarray:
"""Get an array view of vnl_matrix"""
return _GetArrayFromVnlObject(vnl_matrix, "GetArrayViewFromVnlMatrix", ttype)
array_view_from_vnl_matrix = GetArrayViewFromVnlMatrix
def _GetVnlObjectFromArray(arr: ArrayLike, function_name: str, ttype):
"""Get a vnl object from a Python array."""
import itk
# Verify inputs
if not isinstance(arr, np.ndarray):
arr = np.asarray(arr)
if ttype is not None:
if isinstance(ttype, (tuple, list)):
if len(ttype) != 1:
raise RuntimeError("Expected 1 component in ttype tuple.")
PixelType = ttype[0]
else:
PixelType = ttype
else:
PixelType = _get_itk_pixelid(arr)
keys = [k for k in itk.PyVnl.keys() if k[0] == PixelType]
if len(keys) == 0:
raise RuntimeError("No suitable template parameter can be found.")
templatedFunction = getattr(itk.PyVnl[keys[0]], function_name)
return templatedFunction(arr)
def GetVnlVectorFromArray(arr: ArrayLike, ttype=None):
"""Get a vnl vector from a Python array."""
return _GetVnlObjectFromArray(arr, "GetVnlVectorFromArray", ttype)
vnl_vector_from_array = GetVnlVectorFromArray
def GetVnlMatrixFromArray(arr: ArrayLike, ttype=None):
"""Get a vnl matrix from a Python array."""
return _GetVnlObjectFromArray(arr, "GetVnlMatrixFromArray", ttype)
vnl_matrix_from_array = GetVnlMatrixFromArray
def GetArrayFromMatrix(itk_matrix) -> np.ndarray:
return GetArrayFromVnlMatrix(itk_matrix.GetVnlMatrix().as_matrix())
array_from_matrix = GetArrayFromMatrix
def GetMatrixFromArray(arr: ArrayLike) -> "itkt.Matrix":
import itk
# Verify inputs
if not isinstance(arr, np.ndarray):
arr = np.asarray(arr)
vnl_matrix = GetVnlMatrixFromArray(arr)
dims = arr.shape
PixelType = _get_itk_pixelid(arr)
m = itk.Matrix[PixelType, dims[0], dims[1]](vnl_matrix)
return m
matrix_from_array = GetMatrixFromArray
def xarray_from_image(l_image: "itkt.ImageOrImageSource") -> "xr.DataArray":
"""Convert an itk.Image to an xarray.DataArray.
Origin and spacing metadata is preserved in the xarray's coords. The
Direction is set in the `direction` attribute.
Dims are labeled as `x`, `y`, `z`, `t`, and `c`.
This interface is and behavior is experimental and is subject to possible
future changes."""
import xarray as xr
import itk
import numpy as np
array_view = itk.array_view_from_image(l_image)
l_spacing = itk.spacing(l_image)
l_origin = itk.origin(l_image)
l_size = itk.size(l_image)
direction = np.flip(itk.array_from_matrix(l_image.GetDirection()))
image_dimension = l_image.GetImageDimension()
image_dims: Tuple[str, str, str] = ("x", "y", "z", "t")
coords = {}
for l_index, dim in enumerate(image_dims[:image_dimension]):
coords[dim] = np.linspace(
l_origin[l_index],
l_origin[l_index] + (l_size[l_index] - 1) * l_spacing[l_index],
l_size[l_index],
dtype=np.float64,
)
dims = list(reversed(image_dims[:image_dimension]))
components = l_image.GetNumberOfComponentsPerPixel()
if components > 1:
dims.append("c")
coords["c"] = np.arange(components, dtype=np.uint32)
direction = np.flip(itk.array_from_matrix(l_image.GetDirection()))
attrs = {"direction": direction}
metadata = dict(l_image)
ignore_keys = {"direction", "origin", "spacing"}
for key in metadata:
if not key in ignore_keys:
attrs[key] = metadata[key]
data_array = xr.DataArray(array_view, dims=dims, coords=coords, attrs=attrs)
return data_array
def image_from_xarray(data_array: "xr.DataArray") -> "itkt.ImageBase":
"""Convert an xarray.DataArray to an itk.Image.
Metadata encoded with xarray_from_image is applied to the itk.Image.
This interface is and behavior is experimental and is subject to possible
future changes."""
import numpy as np
import itk
if not {"t", "z", "y", "x", "c"}.issuperset(data_array.dims):
raise ValueError('Unsupported dims, supported dims: "t", "z", "y", "x", "c".')
image_dims = list({"t", "z", "y", "x"}.intersection(set(data_array.dims)))
image_dims.sort(reverse=True)
image_dimension = len(image_dims)
ordered_dims = ("t", "z", "y", "x")[-image_dimension:]
is_vector = "c" in data_array.dims
if is_vector:
ordered_dims = ordered_dims + ("c",)
values = data_array.values
if ordered_dims != data_array.dims:
dest = list(builtins.range(len(ordered_dims)))
source = dest.copy()
for ii in builtins.range(len(ordered_dims)):
source[ii] = data_array.dims.index(ordered_dims[ii])
values = np.moveaxis(values, source, dest).copy()
itk_image = itk.image_view_from_array(values, is_vector=is_vector)
l_origin = [0.0] * image_dimension
l_spacing = [1.0] * image_dimension
for l_index, dim in enumerate(image_dims):
coords = data_array.coords[dim]
if coords.shape[0] > 1:
l_origin[l_index] = float(coords[0])
l_spacing[l_index] = float(coords[1]) - float(coords[0])
l_spacing.reverse()
itk_image.SetSpacing(l_spacing)
l_origin.reverse()
itk_image.SetOrigin(l_origin)
if "direction" in data_array.attrs:
direction = data_array.attrs["direction"]
itk_image.SetDirection(np.flip(direction))
ignore_keys = {"direction", "origin", "spacing"}
for key in data_array.attrs:
if not key in ignore_keys:
itk_image[key] = data_array.attrs[key]
return itk_image
def vtk_image_from_image(l_image: "itkt.ImageOrImageSource") -> "vtk.vtkImageData":
"""Convert an itk.Image to a vtk.vtkImageData."""
import itk
import vtk
from vtk.util.numpy_support import numpy_to_vtk
array = itk.array_view_from_image(l_image)
vtk_image = vtk.vtkImageData()
data_array = numpy_to_vtk(array.reshape(-1))
data_array.SetNumberOfComponents(l_image.GetNumberOfComponentsPerPixel())
data_array.SetName("Scalars")
# Always set Scalars for (future?) multi-component volume rendering
vtk_image.GetPointData().SetScalars(data_array)
dim = l_image.GetImageDimension()
l_spacing = [1.0] * 3
l_spacing[:dim] = l_image.GetSpacing()
vtk_image.SetSpacing(l_spacing)
l_origin = [0.0] * 3
l_origin[:dim] = l_image.GetOrigin()
vtk_image.SetOrigin(l_origin)
dims = [1] * 3
dims[:dim] = itk.size(l_image)
vtk_image.SetDimensions(dims)
# Copy direction matrix for VTK>=9
import vtk
if vtk.vtkVersion.GetVTKMajorVersion() >= 9:
l_direction = l_image.GetDirection()
direction = itk.array_from_matrix(l_direction).flatten().tolist()
if len(direction) == 4:
# Change 2d matrix to 3d
direction = [
direction[0],
direction[1],
0.0,
direction[2],
direction[3],
0.0,
0.0,
0.0,
1.0,
]
vtk_image.SetDirectionMatrix(direction)
if l_image.GetImageDimension() == 3:
PixelType = itk.template(l_image)[1][0]
if PixelType == itk.Vector:
vtk_image.GetPointData().SetVectors(data_array)
elif PixelType == itk.CovariantVector:
vtk_image.GetPointData().SetVectors(data_array)
elif PixelType == itk.SymmetricSecondRankTensor:
vtk_image.GetPointData().SetTensors(data_array)
elif PixelType == itk.DiffusionTensor3D:
vtk_image.GetPointData().SetTensors(data_array)
return vtk_image
def image_from_vtk_image(vtk_image: "vtk.vtkImageData") -> "itkt.ImageBase":
"""Convert a vtk.vtkImageData to an itk.Image."""
import itk
from vtk.util.numpy_support import vtk_to_numpy
point_data = vtk_image.GetPointData()
array = vtk_to_numpy(point_data.GetScalars())
array = array.reshape(-1)
is_vector = point_data.GetScalars().GetNumberOfComponents() != 1
dims = list(vtk_image.GetDimensions())
if is_vector and dims[-1] == 1:
# 2D
dims = dims[:2]
dims.reverse()
dims.append(point_data.GetScalars().GetNumberOfComponents())
else:
dims.reverse()
array.shape = tuple(dims)
l_image = itk.image_view_from_array(array, is_vector)
dim = l_image.GetImageDimension()
l_spacing = [1.0] * dim
l_spacing[:dim] = vtk_image.GetSpacing()[:dim]
l_image.SetSpacing(l_spacing)
l_origin = [0.0] * dim
l_origin[:dim] = vtk_image.GetOrigin()[:dim]
l_image.SetOrigin(l_origin)
# Direction support with VTK 9
import vtk
if vtk.vtkVersion.GetVTKMajorVersion() >= 9:
direction = vtk_image.GetDirectionMatrix()
if dim == 3:
direction_array = np.identity(3)
for y in (0, 1, 2):
for x in (0, 1, 2):
direction_array[x, y] = direction.GetElement(x, y)
elif dim == 2:
direction_array = np.identity(2)
for y in (0, 1):
for x in (0, 1):
direction_array[x, y] = direction.GetElement(x, y)
l_direction = itk.matrix_from_array(direction_array)
l_image.SetDirection(l_direction)
return l_image
def dict_from_image(image: "itkt.Image") -> Dict:
"""Serialize a Python itk.Image object to a pickable Python dictionary."""
import itk
pixel_arr = itk.array_view_from_image(image)
imageType = wasm_type_from_image_type(image)
return dict(
imageType=imageType,
origin=tuple(image.GetOrigin()),
spacing=tuple(image.GetSpacing()),
size=tuple(image.GetBufferedRegion().GetSize()),
direction=np.asarray(image.GetDirection()),
data=pixel_arr,
)
def image_from_dict(image_dict: Dict) -> "itkt.Image":
"""Deserialize an dictionary representing an itk.Image object."""
import itk
ImageType = image_type_from_wasm_type(image_dict["imageType"])
image = itk.PyBuffer[ImageType].GetImageViewFromArray(image_dict["data"])
image.SetOrigin(image_dict["origin"])
image.SetSpacing(image_dict["spacing"])
image.SetDirection(image_dict["direction"])
return image
def mesh_from_dict(mesh_dict: Dict) -> "itkt.Mesh":
"""Deserialize an dictionary representing an itk.Mesh object."""
import itk
MeshType = mesh_type_from_wasm_type(mesh_dict["meshType"])
mesh = MeshType.New()
mesh.SetObjectName(mesh_dict["name"])
points = mesh_dict["points"]
points = itk.vector_container_from_array(points)
mesh.SetPoints(points)
point_data = mesh_dict["pointData"]
point_data = itk.vector_container_from_array(point_data)
mesh.SetPointData(point_data)
cells = mesh_dict["cells"]
cells = itk.vector_container_from_array(cells)
mesh.SetCellsArray(cells)
cell_data = mesh_dict["cellData"]
cell_data = itk.vector_container_from_array(cell_data)
mesh.SetCellData(cell_data)
return mesh
def dict_from_mesh(mesh: "itkt.Mesh") -> Dict:
"""Serialize a Python itk.Mesh object to a pickable Python dictionary."""
import itk
mesh_template = itk.template(mesh)
pixel_type, mangle, pixel_type_components = wasm_type_from_mesh_type(mesh)
number_of_points = mesh.GetNumberOfPoints()
number_of_cells = mesh.GetNumberOfCells()
if number_of_cells == 0:
cells_array = np.array([], np.uint)
else:
cells_array = itk.array_view_from_vector_container(mesh.GetCellsArray())
if number_of_points == 0:
points_array = np.array([], np.float32)
else:
points_array = itk.array_view_from_vector_container(mesh.GetPoints()).flatten()
point_data = mesh.GetPointData()
if point_data.Size() == 0:
point_data_numpy = np.array([], mangle)
else:
point_data_numpy = itk.array_view_from_vector_container(point_data)
cell_data = mesh.GetCellData()
if cell_data.Size() == 0:
cell_data_numpy = np.array([], mangle)
else:
cell_data_numpy = itk.array_view_from_vector_container(cell_data)
if os.name == "nt":
cell_component_type = python_to_js(itk.ULL)
else:
cell_component_type = python_to_js(itk.UL)
point_component_type = python_to_js(itk.F)
# Currently use the same data type for point and cell data
mesh_type = dict()
mesh_type["dimension"] = mesh_template[1][1]
mesh_type["pointComponentType"] = point_component_type
mesh_type["pointPixelComponentType"] = mangle
mesh_type["pointPixelType"] = pixel_type
mesh_type["pointPixelComponents"] = pixel_type_components
mesh_type["cellComponentType"] = cell_component_type
mesh_type["cellPixelComponentType"] = mangle
mesh_type["cellPixelType"] = pixel_type
mesh_type["cellPixelComponents"] = pixel_type_components
cell_buffer_size = cells_array.size
return dict(
meshType=mesh_type,
name=mesh.GetObjectName(),
dimension=mesh_template[1][1],
numberOfPoints=number_of_points,
points=points_array,
numberOfPointPixels=point_data.Size(),
pointData=point_data_numpy,
numberOfCells=number_of_cells,
cells=cells_array,
numberOfCellPixels=cell_data.Size(),
cellData=cell_data_numpy,
cellBufferSize=cell_buffer_size,
)
def pointset_from_dict(pointset_dict: Dict) -> "itkt.PointSet":
"""Deserialize an dictionary representing an itk.PointSet object."""
import itk
MeshType = pointset_type_from_wasm_type(pointset_dict["pointSetType"])
mesh = MeshType.New()
mesh.SetObjectName(pointset_dict["name"])
points = pointset_dict["points"]
points = itk.vector_container_from_array(points)
mesh.SetPoints(points)
point_data = pointset_dict["pointData"]
point_data = itk.vector_container_from_array(point_data)
mesh.SetPointData(point_data)
return mesh
def dict_from_pointset(pointset: "itkt.PointSet") -> Dict:
"""Serialize a Python itk.PointSet object to a pickable Python dictionary."""
import itk
pointset_template = itk.template(pointset)
pixel_type, mangle, pixel_type_components = wasm_type_from_pointset_type(pointset)
number_of_points = pointset.GetNumberOfPoints()
if number_of_points == 0:
points_array = np.array([], np.float32)
else:
points_array = itk.array_view_from_vector_container(pointset.GetPoints()).flatten()
point_data = pointset.GetPointData()
if point_data.Size() == 0:
point_data_numpy = np.array([], mangle)
else:
point_data_numpy = itk.array_view_from_vector_container(point_data)
if os.name == "nt":
cell_component_type = python_to_js(itk.ULL)
else:
cell_component_type = python_to_js(itk.UL)
point_component_type = python_to_js(itk.F)
# Currently use the same data type for point and cell data
pointset_type = dict()
pointset_type["dimension"] = pointset_template[1][1]
pointset_type["pointComponentType"] = point_component_type
pointset_type["pointPixelComponentType"] = mangle
pointset_type["pointPixelType"] = pixel_type
pointset_type["pointPixelComponents"] = pixel_type_components
return dict(
pointSetType=pointset_type,
name=pointset.GetObjectName(),
dimension=pointset_template[1][1],
numberOfPoints=number_of_points,
points=points_array,
numberOfPointPixels=point_data.Size(),
pointData=point_data_numpy,
)
def image_intensity_min_max(image_or_filter: "itkt.ImageOrImageSource"):
"""Return the minimum and maximum of values in a image of in the output image of a filter
The minimum and maximum values are returned in a tuple: (min, max)
image_intensity_min_max() take care of updating the pipeline
"""
import itk
img = itk.output(image_or_filter)
img.UpdateOutputInformation()
img.Update()
# don't put that calculator in the automatic pipeline
tmp_auto_pipeline = auto_pipeline.current
auto_pipeline.current = None
comp = itk.MinimumMaximumImageCalculator[img].New(Image=img)
auto_pipeline.current = tmp_auto_pipeline
comp.Compute()
return comp.GetMinimum(), comp.GetMaximum()
# range is a python function, and should not be overridden
# the current use of the function name "range" is for backward
# compatibility, but should be considered for removal in the future
def range(image_or_filter):
return image_intensity_min_max(image_or_filter)
def imwrite(
image_or_filter: "itkt.ImageOrImageSource",
filename: fileiotype,
compression: bool = False,
imageio: Optional["itkt.ImageIOBase"] = None,
) -> None:
"""Write a image or the output image of a filter to a file.
Parameters
----------
image_or_filter :
Image or filter that produces an image to write to the file.
filename :
Target output file path.
compression :
Use compression when writing if the format supports it.
imageio :
Use the provided itk.ImageIOBase derived instance to write the file.
The writer is instantiated with the image type of the image in
parameter (or, again, with the output image of the filter in parameter).
"""
import itk
img = itk.output(image_or_filter)
img.UpdateOutputInformation()
# don't put that writer in the automatic pipeline
tmp_auto_pipeline = auto_pipeline.current
auto_pipeline.current = None
writer = itk.ImageFileWriter[type(img)].New(
Input=img, FileName=f"{filename}", UseCompression=compression
)
auto_pipeline.current = tmp_auto_pipeline
if imageio:
writer.SetImageIO(imageio)
writer.Update()
def imread(
filename: fileiotype,
pixel_type: Optional["itkt.PixelTypes"] = None,
fallback_only: bool = False,
imageio: Optional["itkt.ImageIOBase"] = None,
) -> "itkt.ImageBase":
"""Read an image from a file or series of files and return an itk.Image.
Parameters
----------
filename :
File path for a single file, a list of files for an image series, or a
directory for a DICOM image series.
pixel_type :
Image pixel type to cast to when loading.
fallback_only :
If true, first try to automatically deduce the image pixel type, and
only use the given `pixel_type` if automatic deduction fails.
imageio :
Use the provided itk.ImageIOBase derived instance to read the file.
Returns
-------
image :
The resulting itk.Image.
The reader is instantiated with the image type of the image file if
`pixel_type` is not provided (default). The dimension of the image is
automatically deduced from the dimension stored on disk.
If the filename provided is a directory then the directory is assumed to
be for a DICOM series volume. If there is exactly one DICOM series
volume in that directory, the reader will use an itk.ImageSeriesReader
object to read the the DICOM filenames within that directory.
If the given filename is a list or a tuple of file names, the reader
will use an itk.ImageSeriesReader object to read the files.
If `fallback_only` is set to `True`, `imread()` will first try to
automatically deduce the image pixel_type, and only use the given
`pixel_type` if automatic deduction fails. Failures typically happen if
the pixel type is not supported (e.g. it is not currently wrapped).
"""
import itk
from itk.support.extras import TemplateTypeError
if fallback_only:
if pixel_type is None:
raise Exception(
"pixel_type must be set when using the fallback_only option"
)
try:
return imread(filename)
except (KeyError, TemplateTypeError):
pass
if type(filename) not in [list, tuple]:
import os
if os.path.isdir(filename):
# read DICOM series of 1 image in a folder, refer to: https://github.com/RSIP-Vision/medio
names_generator = itk.GDCMSeriesFileNames.New()
names_generator.SetUseSeriesDetails(True)
names_generator.AddSeriesRestriction("0008|0021") # Series Date
names_generator.SetDirectory(f"{filename}")
series_uid = names_generator.GetSeriesUIDs()
if len(series_uid) == 0:
raise FileNotFoundError(f"no DICOMs in: {filename}.")
if len(series_uid) > 1:
raise OSError(
f"the directory: {filename} contains more than one DICOM series."
)
series_identifier = series_uid[0]
filename = names_generator.GetFileNames(series_identifier)
if type(filename) in [list, tuple]:
template_reader_type = itk.ImageSeriesReader
io_filename = f"{filename[0]}"
increase_dimension = True
kwargs = {"FileNames": [f"{f}" for f in filename]}
else:
template_reader_type = itk.ImageFileReader
io_filename = f"{filename}"
increase_dimension = False
kwargs = {"FileName": f"{filename}"}
if imageio:
kwargs["ImageIO"] = imageio
if pixel_type:
image_IO = itk.ImageIOFactory.CreateImageIO(
io_filename, itk.CommonEnums.IOFileMode_ReadMode
)
if not image_IO:
raise RuntimeError("No ImageIO is registered to handle the given file.")
image_IO.SetFileName(io_filename)
image_IO.ReadImageInformation()
dimension = image_IO.GetNumberOfDimensions()
# Increase dimension if last dimension is not of size one.
if increase_dimension and image_IO.GetDimensions(dimension - 1) != 1:
dimension += 1
is_vlv = False
try:
is_vlv = itk.template(pixel_type)[0] is itk.VariableLengthVector
except KeyError:
pass
if is_vlv:
ImageType = itk.VectorImage[itk.template(pixel_type)[1][0], dimension]
else:
ImageType = itk.Image[pixel_type, dimension]
reader = template_reader_type[ImageType].New(**kwargs)
else:
reader = template_reader_type.New(**kwargs)
reader.Update()
return reader.GetOutput()
def meshwrite(
mesh: "itkt.Mesh", filename: fileiotype, compression: bool = False
) -> None:
"""Write a mesh to a file.
The writer is instantiated according to the type of the input mesh.
"""
import itk
mesh.UpdateOutputInformation()
# don't put that writer in the automatic pipeline
tmp_auto_pipeline = auto_pipeline.current
auto_pipeline.current = None
writer = itk.MeshFileWriter[type(mesh)].New(
Input=mesh, FileName=f"{filename}", UseCompression=compression
)
auto_pipeline.current = tmp_auto_pipeline
writer.Update()
def meshread(
filename: fileiotype,
pixel_type: Optional["itkt.PixelTypes"] = None,
fallback_only: bool = False,
) -> "itkt.Mesh":
"""Read a mesh from a file and return an itk.Mesh.
The reader is instantiated with the mesh type of the mesh file if
`pixel_type` is not provided (default). The dimension of the mesh is
automatically found.
If `fallback_only` is set to `True`, `meshread()` will first try to
automatically deduce the image pixel_type, and only use the given
`pixel_type` if automatic deduction fails. Failures typically
happen if the pixel type is not supported (e.g. it is not currently
wrapped).
"""
import itk
if fallback_only:
if pixel_type is None:
raise Exception(
"pixel_type must be set when using the fallback_only option"
)
try:
return meshread(filename)
except (KeyError, itk.TemplateTypeError):
pass
TemplateReaderType = itk.MeshFileReader
io_filename = f"{filename}"
increase_dimension = False
kwargs = {"FileName": f"{filename}"}
if pixel_type:
meshIO = itk.MeshIOFactory.CreateMeshIO(
io_filename, itk.CommonEnums.IOFileMode_ReadMode
)
if not meshIO:
raise RuntimeError("No MeshIO is registered to handle the given file.")
meshIO.SetFileName(io_filename)
meshIO.ReadMeshInformation()
dimension = meshIO.GetPointDimension()
# Increase dimension if last dimension is not of size one.
if increase_dimension and meshIO.GetDimensions(dimension - 1) != 1:
dimension += 1
MeshType = itk.Mesh[pixel_type, dimension]
reader = TemplateReaderType[MeshType].New(**kwargs)
else:
reader = TemplateReaderType.New(**kwargs)
reader.Update()
return reader.GetOutput()
def transformread(filename: fileiotype) -> List["itkt.TransformBase"]:
"""Read an itk Transform file.
Parameters
----------
filename:
Path to the transform file (typically a .h5 file).
Returns
-------
A Python list containing the transforms in the file.
"""
import itk
reader = itk.TransformFileReaderTemplate[itk.D].New()
reader.SetFileName(f"{filename}")
reader.Update()
transforms = []
transform_list = reader.GetModifiableTransformList()
while not transform_list.empty():
transform = transform_list.pop()
transforms.append(itk.down_cast(transform))
transforms.reverse()
return transforms
def transformwrite(
transforms: List["itkt.TransformBase"],
filename: fileiotype,
compression: bool = False,
) -> None:
"""Write an itk Transform file.
Parameters
----------
transforms: list of itk.TransformBaseTemplate[itk.D]
Python list of the transforms to write.
filename:
Path to the transform file (typically a .h5 file).
compression:
Use compression, if the file format supports it.
"""
import itk
writer = itk.TransformFileWriterTemplate[itk.D].New()
writer.SetFileName(f"{filename}")
writer.SetUseCompression(compression)
for transform in transforms:
writer.AddTransform(transform)
writer.Update()
def search(s: str, case_sensitive: bool = False) -> List[str]: # , fuzzy=True):
"""Search for a class name in the itk module."""
s = s.replace(" ", "")
if not case_sensitive:
s = s.lower()
import itk
names = sorted(dir(itk))
# exact match first
if case_sensitive:
res = [n for n in names if s == n]
else:
res = [n for n in names if s == n.lower()]
# then exact match inside the name
if case_sensitive:
res += [n for n in names if s in n and s != n]
else:
res += [n for n in names if s in n.lower() and s != n.lower()]
# if fuzzy:
# try:
# everything now requires editdist
# import editdist
# if case_sensitive:
# res.sort(key=lambda x: editdist.distance(x, s))
# else:
# res.sort(key=lambda x: (editdist.distance(x.lower(), s), x))
# except:
# pass
return res
def _snake_to_camel(keyword: str):
# Helpers for set_inputs snake case to CamelCase keyword argument conversion
_snake_underscore_re = re.compile("(_)([a-z0-9A-Z])")
def _underscore_upper(match_obj):
return match_obj.group(2).upper()
camel = keyword[0].upper()
if _snake_underscore_re.search(keyword[1:]):
return camel + _snake_underscore_re.sub(_underscore_upper, keyword[1:])
return camel + keyword[1:]
def set_inputs(
new_itk_object,
inargs: Optional[Sequence[Any]] = None,
inkargs: Optional[Dict[str, Any]] = None,
):
"""Set the inputs of the given objects, according to the non named or the
named parameters in args and kargs
This function tries to assign all the non named parameters in the input of
the new_itk_object
- the first non named parameter in the first input, etc.
The named parameters are used by calling the method with the same name
prefixed by 'Set'.
set_inputs( obj, kargs={'Threshold': 10} ) calls obj.SetThreshold(10)
This is the function use in the enhanced New() method to manage the inputs.
It can be used to produce a similar behavior:
def SetInputs(self, *args, **kargs):
import itk
itk.set_inputs(self, *args, **kargs)
"""
# Fix bug with Mutable Default Arguments
# https://docs.python-guide.org/writing/gotchas/
args: List[Any] = inargs if inargs else []
kargs: Dict[str, Any] = inkargs if inkargs else {}
# try to get the images from the filters in args
args = [output(arg) for arg in args]
# args without name are filter used to set input image
#
# count SetInput calls to call SetInput, SetInput2, SetInput3, ...
# useful with filter which take 2 input (or more) like SubtractImageFiler
# Ex: subtract image2.png to image1.png and save the result in result.png
# r1 = itk.ImageFileReader.US2.New(FileName='image1.png')
# r2 = itk.ImageFileReader.US2.New(FileName='image2.png')
# s = itk.SubtractImageFilter.US2US2US2.New(r1, r2)
# itk.ImageFileWriter.US2.New(s, FileName='result.png').Update()
setInputNb: int = -1
try:
for setInputNb, arg in enumerate(args):
methodName = "SetInput%i" % (setInputNb + 1)
if methodName in dir(new_itk_object):
# first try to use methods called SetInput1, SetInput2, ...
# those method should have more chances to work in case of
# multiple input types
getattr(new_itk_object, methodName)(arg)
else:
# no method called SetInput?
# try with the standard SetInput(nb, input)
new_itk_object.SetInput(setInputNb, arg)
except TypeError as e:
# the exception have (at least) to possible reasons:
# + the filter don't take the input number as first argument
# + arg is an object of wrong type
#
# if it's not the first input, re-raise the exception
if setInputNb != 0:
raise e
# it's the first input, try to use the SetInput() method without input
# number
new_itk_object.SetInput(args[0])
# but raise an exception if there is more than 1 argument
if len(args) > 1:
raise TypeError("Object accepts only 1 input.")
except AttributeError:
# There is no SetInput() method, try SetImage
# but before, check the number of inputs
if len(args) > 1:
raise TypeError("Object accepts only 1 input.")
methodList = ["SetImage", "SetInputImage"]
methodName = None
for m in methodList:
if m in dir(new_itk_object):
methodName = m
if methodName:
getattr(new_itk_object, methodName)(args[0])
else:
raise AttributeError("No method found to set the input.")
# named args : name is the function name, value is argument(s)
for attribName, value in kargs.items():
# use Set as prefix. It allow to use a shorter and more intuitive
# call (Ex: itk.ImageFileReader.UC2.New(FileName='image.png')) than
# with the full name
# (Ex: itk.ImageFileReader.UC2.New(SetFileName='image.png'))
if attribName not in ["auto_progress", "template_parameters"]:
if attribName.islower():
attribName = _snake_to_camel(attribName)
attrib = getattr(new_itk_object, "Set" + attribName)
# Do not use try-except mechanism as this leads to
# segfaults. Instead limit the number of types that are
# tested. The list of tested type could maybe be replaced by
# a test that would check for iterables.
import itk
if type(value) in [list, tuple]:
try:
output_value = [itk.output(x) for x in value]
attrib(*output_value)
except Exception:
attrib(itk.output(value))
else:
attrib(itk.output(value))
class templated_class:
"""This class is used to mimic the behavior of the templated C++ classes.
It is used this way:
class CustomClass:
# class definition here
CustomClass = templated_class(CustomClass)
customObject = CustomClass[template, parameters].New()
The template parameters are passed to the custom class constructor as a
named parameter 'template_parameters' in a tuple.
The custom class may implement a static method
check_template_parameters(parameters) which should raise an exception if
the template parameters provided are not suitable to instantiate the custom
class.
"""
def __init__(self, cls) -> None:
"""cls is the custom class"""
self.__cls__ = cls
self.__templates__ = {}
def New(self, *args, **kargs):
"""Use the parameters to infer the types of the template parameters."""
# extract the types from the arguments to instantiate the class
import itk
types = tuple(class_(o) for o in args)
return self[types].New(*args, **kargs)
def __getitem__(self, template_parameters):
"""Return a pair class-template parameters ready to be instantiated.
The template parameters may be validated if the custom class provide
the static method check_template_parameters(parameters).
"""
if not isinstance(template_parameters, tuple):
template_parameters = (template_parameters,)
return templated_class.__templated_class_and_parameters__(
self, template_parameters
)
def check_template_parameters(self, template_parameters) -> None:
"""Check the template parameters passed in parameter."""
# this method is there mainly to make possible to reuse it in the
# custom class constructor after having used templated_class().
# Without that, the following example doesn't work:
#
# class CustomClass:
# def __init__(self, *args, **kargs):
# template_parameters = kargs["template_parameters"]
# CustomClass.check_template_parameters(template_parameters)
# other init stuff
# def check_template_parameters(template_parameters):
# check, really
# pass
# CustomClass = templated_class(CustomClass)
#
self.__cls__.check_template_parameters(template_parameters)
def add_template(self, name: str, params):
if not isinstance(params, list) and not isinstance(params, tuple):
params = (params,)
params = tuple(params)
val = self[params]
self.__templates__[params] = val
setattr(self, name, val)
def add_image_templates(self, *args) -> None:
import itk
if not args:
return
combinations = [[t] for t in args[0]]
for types in args[1:]:
temp = []
for t in types:
for c in combinations:
temp.append(c + [t])
combinations = temp
for d in itk.DIMS:
for c in combinations:
parameters = []
name = ""
for t in c:
parameters.append(itk.Image[t, d])
name += "I" + t.short_name + str(d)
self.add_template(name, tuple(parameters))
class __templated_class_and_parameters__:
"""Inner class used to store the pair class-template parameters ready
to instantiate.
"""
def __init__(self, l_templated_class, l_template_parameters) -> None:
self.__templated_class__ = l_templated_class
self.__template_parameters__ = l_template_parameters
if "check_template_parameters" in dir(l_templated_class.__cls__):
l_templated_class.__cls__.check_template_parameters(
l_template_parameters
)
def New(self, *args, **kargs):
"""A New() method to mimic the ITK default behavior, even if the
class doesn't provide any New() method.
"""
kargs["template_parameters"] = self.__template_parameters__
if "New" in dir(self.__templated_class__.__cls__):
obj = self.__templated_class__.__cls__.New(*args, **kargs)
else:
obj = self.__templated_class__.__cls__(*args, **kargs)
setattr(obj, "__template_parameters__", self.__template_parameters__)
setattr(obj, "__templated_class__", self.__templated_class__)
return obj
def __call__(self, *args, **kargs):
return self.New(*args, **kargs)
def keys(self):
return self.__templates__.keys()
def values(self):
return list(self.__templates__.values())
def items(self):
return list(self.__templates__.items())
# everything after this comment is for dict interface
# and is a copy/paste from DictMixin
# only methods to edit dictionary are not there
def __iter__(self) -> str:
yield from self.keys()
def has_key(self, key: str):
return key in self.__templates__
def __contains__(self, key: str):
return key in self
def get(self, key: str, default: Optional[str] = None) -> Optional[str]:
return self.get(key, default)
def __len__(self):
return len(self.keys())
class pipeline:
"""A convenient class to store the reference to the filters of a pipeline
With this class, a method can create a pipeline of several filters and
return it without losing the references to the filters in this pipeline.
The pipeline object act almost like a filter (it has a GetOutput() method)
and thus can be simply integrated in another pipeline.
"""
def __init__(self, *args, **kargs) -> None:
self.clear()
self.input = None
self.filters: List[Any] = []
set_inputs(self, args, kargs)
def connect(self, l_filter) -> None:
"""Connect a new l_filter to the pipeline
The output of the first l_filter will be used as the input of this
one and the l_filter passed as parameter will be added to the list
"""
if self.GetOutput() is not None:
set_inputs(l_filter, [self.GetOutput()])
self.append(l_filter)
def append(self, l_filter) -> None:
"""Add a new l_filter to the pipeline
The new l_filter will not be connected. The user must connect it.
"""
self.filters.append(l_filter)
def clear(self) -> None:
"""Clear the filter list"""
self.filters = []
def GetOutput(self, l_index: int = 0):
"""Return the output of the pipeline
If another output is needed, use
pipeline.filters[-1].GetAnotherOutput() instead of this method,
subclass pipeline to implement another GetOutput() method, or use
expose()
"""
if len(self.filters) == 0:
return self.GetInput()
else:
l_filter = self.filters[-1]
if hasattr(l_filter, "__getitem__"):
return l_filter[l_index]
try:
return l_filter.GetOutput(l_index)
except Exception:
if l_index == 0:
return l_filter.GetOutput()
else:
raise ValueError("Index can only be 0 on that object")
def GetNumberOfOutputs(self) -> int:
"""Return the number of outputs"""
if len(self.filters) == 0:
return 1
else:
return self.filters[-1].GetNumberOfOutputs()
def SetInput(self, l_input) -> None:
"""Set the l_input of the pipeline"""
if len(self.filters) != 0:
set_inputs(self.filters[0], [l_input])
self.l_input = l_input
def GetInput(self):
"""Get the input of the pipeline"""
return self.input
def Update(self):
"""Update the pipeline"""
if len(self.filters) > 0:
return self.filters[-1].Update()
def UpdateLargestPossibleRegion(self):
"""Update the pipeline"""
if len(self.filters) > 0:
return self.filters[-1].UpdateLargestPossibleRegion()
def UpdateOutputInformation(self) -> None:
if "UpdateOutputInformation" in dir(self.filters[-1]):
self.filters[-1].UpdateOutputInformation()
else:
self.Update()
def __len__(self):
return self.GetNumberOfOutputs()
def __getitem__(self, item):
return self.GetOutput(item)
def __call__(self, *args, **kargs):
set_inputs(self, args, kargs)
self.UpdateLargestPossibleRegion()
return self
def expose(self, name: str, new_name: Optional[str] = None, position: int = -1):
"""Expose an attribute from a filter of the mini-pipeline.
Once called, the pipeline instance has a new Set/Get set of methods to
access directly the corresponding method of one of the filter of the
pipeline.
Ex: p.expose( "Radius" )
p.SetRadius( 5 )
p.GetRadius( 5 )
By default, the attribute usable on the pipeline instance has the same
name than the one of the filter, but it can be changed by providing a
value to new_name.
The last filter of the pipeline is used by default, but another one may
be used by giving its position.
Ex: p.expose("Radius", "SmoothingNeighborhood", 2)
p.GetSmoothingNeighborhood()
"""
if new_name is None:
new_name = name
src = self.filters[position]
ok: bool = False
set_name: str = "Set" + name
if set_name in dir(src):
setattr(self, "Set" + new_name, getattr(src, set_name))
ok = True
get_name = "Get" + name
if get_name in dir(src):
setattr(self, "Get" + new_name, getattr(src, get_name))
ok = True
if not ok:
raise RuntimeError(f"No attribute {name} at position {position}.")
class auto_pipeline(pipeline):
current = None
def __init__(self, *args, **kargs) -> None:
pipeline.__init__(self, *args, **kargs)
self.Start()
def Start(self) -> None:
auto_pipeline.current = self
@staticmethod
def Stop() -> None:
auto_pipeline.current = None
def down_cast(obj: "itkt.LightObject"):
"""Down cast an itk.LightObject (or a object of a subclass) to its most
specialized type.
"""
import itk
from itk.support.template_class import itkTemplate
class_name: str = obj.GetNameOfClass()
t = getattr(itk, class_name)
if isinstance(t, itkTemplate):
for c in t.values():
try:
return c.cast(obj)
except Exception:
# fail silently for now
pass
raise RuntimeError(f"Can't downcast to a specialization of {class_name}")
else:
return t.cast(obj)
def attribute_list(inputobject, name: str):
"""Returns a list of the specified attributes for the objects in the image.
i: the input LabelImage
name: the attribute name
"""
import itk
img = itk.output(inputobject)
relabel = itk.StatisticsRelabelLabelMapFilter[img].New(
img, Attribute=name, ReverseOrdering=True, InPlace=False
)
relabel.UpdateLargestPossibleRegion()
r = relabel.GetOutput()
l_list: List[Any] = []
# required because range is overloaded in this module
import sys
from builtins import range
for i in range(1, r.GetNumberOfLabelObjects() + 1):
l_list.append(r.GetLabelObject(i).__getattribute__("Get" + name)())
return l_list
def attributes_list(inputObject, names: List[str]):
"""Returns a list of the specified attributes for the objects in the image.
i: the input LabelImage
name: the attribute name
"""
import itk
img = itk.output(inputObject)
relabel = itk.StatisticsRelabelLabelMapFilter[img].New(
img, Attribute=names[0], ReverseOrdering=True, InPlace=False
)
relabel.UpdateLargestPossibleRegion()
r = relabel.GetOutput()
l_list: List[Any] = []
# required because range is overloaded in this module
from builtins import range
for i in range(1, r.GetNumberOfLabelObjects() + 1):
attrs = []
for name in names:
attrs.append(r.GetLabelObject(i).__getattribute__("Get" + name)())
l_list.append(tuple(attrs))
return l_list
def attribute_dict(inputobject, name: str):
"""Returns a dict with the attribute values in keys and a list of the
corresponding objects in value
i: the input LabelImage
name: the name of the attribute
"""
import itk
img = itk.output(inputobject)
relabel = itk.StatisticsRelabelLabelMapFilter[img].New(
img, Attribute=name, ReverseOrdering=True, InPlace=False
)
relabel.UpdateLargestPossibleRegion()
r = relabel.GetOutput()
d = {}
# required because range is overloaded in this module
from builtins import range
for i in range(1, r.GetNumberOfLabelObjects() + 1):
lo = r.GetLabelObject(i)
v = lo.__getattribute__("Get" + name)()
l_list = d.get(v, [])
l_list.append(lo)
d[v] = l_list
return d
def number_of_objects(image_or_filter) -> int:
"""Returns the number of objets in the image.
img: the input LabelImage
"""
import itk
image_or_filter.UpdateLargestPossibleRegion()
img = itk.output(image_or_filter)
return img.GetNumberOfLabelObjects()
def ipython_kw_matches(text: str):
"""Match named ITK object's named parameters"""
import IPython
import itk
import re
import inspect
from itk.support import template_class
regexp = re.compile(
r"""
'.*?' | # single quoted strings or
".*?" | # double quoted strings or
\w+ | # identifier
\S # other characters
""",
re.VERBOSE | re.DOTALL,
)
ip = IPython.get_ipython()
if "." in text: # a parameter cannot be dotted
return []
# 1. Find the nearest identifier that comes before an unclosed
# parenthesis e.g. for "foo (1+bar(x), pa", the candidate is "foo".
if ip.Completer.readline:
text_until_cursor = ip.Completer.readline.get_line_buffer()[
: ip.Completer.readline.get_endidx()
]
else:
# IPython >= 5.0.0, which is based on the Python Prompt Toolkit
text_until_cursor = ip.Completer.text_until_cursor
tokens = regexp.findall(text_until_cursor)
tokens.reverse()
iter_tokens = iter(tokens)
open_par = 0
for token in iter_tokens:
if token == ")":
open_par -= 1
elif token == "(":
open_par += 1
if open_par > 0:
# found the last unclosed parenthesis
break
else:
return []
# 2. Concatenate dotted names ("foo.bar" for "foo.bar(x, pa" )
ids = []
is_id = re.compile(r"\w+$").match
while True:
try:
ids.append(iter_tokens.next())
if not is_id(ids[-1]):
ids.pop()
break
if not iter_tokens.next() == ".":
break
except StopIteration:
break
# lookup the candidate callable matches either using global_matches
# or attr_matches for dotted names
if len(ids) == 1:
callable_matches = ip.Completer.global_matches(ids[0])
else:
callable_matches = ip.Completer.attr_matches(".".join(ids[::-1]))
arg_matches = []
for callable_match in callable_matches:
# drop the .New at this end, so we can search in the class members
if callable_match.endswith(".New"):
callable_match = callable_match[:-4]
elif not re.findall("([A-Z])", callable_match): # True if snake case
# Split at the last '.' occurrence
split_name_parts = callable_match.split(".")
namespace = split_name_parts[:-1]
function_name = split_name_parts[-1]
# Find corresponding object name
object_name = _snake_to_camel(function_name)
# Check that this object actually exists
try:
object_callable_match = ".".join(namespace + [object_name])
eval(object_callable_match, ip.Completer.namespace)
# Reconstruct full object name
callable_match = object_callable_match
except AttributeError:
# callable_match is not a snake case function with a
# corresponding object.
pass
try:
l_object = eval(callable_match, ip.Completer.namespace)
if isinstance(l_object, template_class.itkTemplate):
# this is a template - lets grab the first entry to search for
# the methods
l_object = l_object.values()[0]
named_args = []
is_in: bool = isinstance(l_object, itk.LightObject)
if is_in or (
inspect.isclass(l_object) and issubclass(l_object, itk.LightObject)
):
named_args = [n[3:] for n in dir(l_object) if n.startswith("Set")]
except Exception as e:
print(e)
continue
for namedArg in named_args:
if namedArg.startswith(text):
arg_matches.append(f"{namedArg}=")
return arg_matches
def template(cl):
"""Return the template of a class (or of the class of an object) and
its parameters
template() returns a tuple with 2 elements:
- the first one is the itkTemplate object
- the second is a tuple containing the template parameters
"""
from itk.support.template_class import itkTemplateBase
return itkTemplateBase.__template_instantiations_object_to_name__[class_(cl)]
def ctype(s: str) -> "itkt.itkCType":
"""Return the c type corresponding to the string passed in parameter
The string can contain some extra spaces.
see also itkCType
"""
from itk.support.types import itkCType
ret = itkCType.GetCType(" ".join(s.split()))
if ret is None:
raise KeyError(f"Unrecognized C type '{s}'")
return ret
def class_(obj):
"""Return a class from an object
Often in itk, the __class__ is not what the user is expecting.
class_() should do a better job
"""
import inspect
if inspect.isclass(obj):
# obj is already a class !
return obj
else:
return obj.__class__
def python_type(object_ref) -> str:
"""Returns the Python type name of an object
The Python name corresponding to the given instantiated object is printed.
This includes both the Python name and the parameters of the object. A user
can copy and paste the printed value to instantiate a new object of the
same type."""
from itk.support.template_class import itkTemplate
from itk.support.types import itkCType
def in_itk(name):
import itk
# Remove "itk::" and "std::" from template name.
# Only happens for ITK objects.
shortname: str = name.split("::")[-1]
shortname = shortname.split("itk")[-1]
namespace = itk
# A type cannot be part of ITK if its name was not modified above. This
# check avoids having an input of type `list` and return `itk.list` that
# also exists.
likely_itk: bool = shortname != name or name[:3] == "vnl"
if likely_itk and hasattr(namespace, shortname):
return namespace.__name__ + "." + shortname # Prepend name with 'itk.'
else:
return name
def recursive(l_obj, level: int):
try:
type_name, param_list = template(l_obj)
name = in_itk(type_name.__name__)
parameters = []
for t in param_list:
parameters.append(recursive(t, level + 1))
return name + "[" + ",".join(parameters) + "]"
except KeyError:
if isinstance(l_obj, itkCType): # Handles CTypes differently
return "itk." + l_obj.short_name
elif hasattr(l_obj, "__name__"):
# This should be where most ITK types end up.
return in_itk(l_obj.__name__)
elif (
not isinstance(l_obj, type)
and type(l_obj) != itkTemplate
and level != 0
):
# l_obj should actually be considered a value, not a type,
# or it is already an itkTemplate type.
# A value can be an integer that is a template parameter.
# This does not happen at the first level of the recursion
# as it is not possible that this object would be a template
# parameter. Checking the level `0` allows e.g. to find the
# type of an object that is a `list` or an `int`.
return str(l_obj)
else:
return in_itk(type(l_obj).__name__)
return recursive(object_ref, 0)
class TemplateTypeError(TypeError):
def __init__(self, template_type, input_type):
def tuple_to_string_type(t):
if type(t) == tuple:
return ", ".join(python_type(x) for x in t)
else:
python_type(t)
import itk
# Special case for ITK readers: Add extra information.
extra_eg: str = ""
if template_type in [
itk.ImageFileReader,
itk.ImageSeriesReader,
itk.MeshFileReader,
]:
extra_eg = """
or
e.g.: image = itk.imread(my_input_filename, itk.F)
"""
python_template_type = python_type(template_type)
python_input_type = tuple_to_string_type(input_type)
type_list = "\n".join([python_type(x[0]) for x in template_type.keys()])
eg_type = ", ".join([python_type(x) for x in list(template_type.keys())[0]])
msg: str = """{template_type} is not wrapped for input type `{input_type}`.
To limit the size of the package, only a limited number of
types are available in ITK Python. To print the supported
types, run the following command in your python environment:
{template_type}.GetTypes()
Possible solutions:
* If you are an application user:
** Convert your input image into a supported format (see below).
** Contact developer to report the issue.
* If you are an application developer, force input images to be
loaded in a supported pixel type.
e.g.: instance = {template_type}[{eg_type}].New(my_input){extra_eg}
* (Advanced) If you are an application developer, build ITK Python yourself and
turned to `ON` the corresponding CMake option to wrap the pixel type or image
dimension you need. When configuring ITK with CMake, you can set
`ITK_WRAP_${{type}}` (replace ${{type}} with appropriate pixel type such as
`double`). If you need to support images with 4 or 5 dimensions, you can add
these dimensions to the list of dimensions in the CMake variable
`ITK_WRAP_IMAGE_DIMS`.
Supported input types:
{type_list}
""".format(
template_type=python_template_type,
input_type=python_input_type,
type_list=type_list,
eg_type=eg_type,
extra_eg=extra_eg,
)
TypeError.__init__(self, msg)
# install progress callback and custom completer if we are in ipython
# interpreter
try:
import itkConfig
import IPython
if IPython.get_ipython():
IPython.get_ipython().Completer.matchers.insert(0, ipython_kw_matches)
# some cleanup
del itkConfig, IPython
except (ImportError, AttributeError):
# fail silently
pass
|
[
"itk.PyBuffer.keys",
"re.compile",
"itk.down_cast",
"itk.output",
"numpy.array",
"numpy.moveaxis",
"numpy.arange",
"itk.array_from_matrix",
"numpy.flip",
"itk.origin",
"numpy.asarray",
"itk.ImageIOFactory.CreateImageIO",
"itk.MeshIOFactory.CreateMeshIO",
"numpy.issubdtype",
"numpy.linspace",
"os.path.isdir",
"itk.PyVnl.keys",
"warnings.warn",
"itk.PyVectorContainer.keys",
"itk.MultiThreaderBase.New",
"IPython.get_ipython",
"numpy.identity",
"vtk.vtkVersion.GetVTKMajorVersion",
"itk.template",
"itk.GDCMSeriesFileNames.New",
"vtk.vtkImageData",
"re.findall",
"itk.size",
"itk.spacing",
"itk.array_view_from_vector_container",
"itk.vector_container_from_array",
"itk.array_view_from_image",
"itk.matrix_from_array",
"xarray.DataArray",
"inspect.isclass",
"itk.image_view_from_array"
] |
[((3438, 3529), 'warnings.warn', 'warnings.warn', (['"""WrapITK warning: itk.image() is deprecated. Use itk.output() instead."""'], {}), "(\n 'WrapITK warning: itk.image() is deprecated. Use itk.output() instead.')\n", (3451, 3529), False, 'import warnings\n'), ((3940, 3967), 'itk.MultiThreaderBase.New', 'itk.MultiThreaderBase.New', ([], {}), '()\n', (3965, 3967), False, 'import itk\n'), ((4138, 4165), 'itk.MultiThreaderBase.New', 'itk.MultiThreaderBase.New', ([], {}), '()\n', (4163, 4165), False, 'import itk\n'), ((4761, 4788), 'itk.output', 'itk.output', (['image_or_filter'], {}), '(image_or_filter)\n', (4771, 4788), False, 'import itk\n'), ((5728, 5755), 'itk.output', 'itk.output', (['image_or_filter'], {}), '(image_or_filter)\n', (5738, 5755), False, 'import itk\n'), ((6129, 6156), 'itk.output', 'itk.output', (['image_or_filter'], {}), '(image_or_filter)\n', (6139, 6156), False, 'import itk\n'), ((6525, 6552), 'itk.output', 'itk.output', (['image_or_filter'], {}), '(image_or_filter)\n', (6535, 6552), False, 'import itk\n'), ((6954, 6981), 'itk.output', 'itk.output', (['image_or_filter'], {}), '(image_or_filter)\n', (6964, 6981), False, 'import itk\n'), ((8166, 8193), 'itk.output', 'itk.output', (['image_or_filter'], {}), '(image_or_filter)\n', (8176, 8193), False, 'import itk\n'), ((12472, 12495), 'itk.template', 'itk.template', (['container'], {}), '(container)\n', (12484, 12495), False, 'import itk\n'), ((13415, 13438), 'itk.template', 'itk.template', (['container'], {}), '(container)\n', (13427, 13438), False, 'import itk\n'), ((19282, 19316), 'itk.array_view_from_image', 'itk.array_view_from_image', (['l_image'], {}), '(l_image)\n', (19307, 19316), False, 'import itk\n'), ((19333, 19353), 'itk.spacing', 'itk.spacing', (['l_image'], {}), '(l_image)\n', (19344, 19353), False, 'import itk\n'), ((19369, 19388), 'itk.origin', 'itk.origin', (['l_image'], {}), '(l_image)\n', (19379, 19388), False, 'import itk\n'), ((19402, 19419), 'itk.size', 'itk.size', (['l_image'], {}), '(l_image)\n', (19410, 19419), False, 'import itk\n'), ((20424, 20487), 'xarray.DataArray', 'xr.DataArray', (['array_view'], {'dims': 'dims', 'coords': 'coords', 'attrs': 'attrs'}), '(array_view, dims=dims, coords=coords, attrs=attrs)\n', (20436, 20487), True, 'import xarray as xr\n'), ((21664, 21718), 'itk.image_view_from_array', 'itk.image_view_from_array', (['values'], {'is_vector': 'is_vector'}), '(values, is_vector=is_vector)\n', (21689, 21718), False, 'import itk\n'), ((22723, 22757), 'itk.array_view_from_image', 'itk.array_view_from_image', (['l_image'], {}), '(l_image)\n', (22748, 22757), False, 'import itk\n'), ((22775, 22793), 'vtk.vtkImageData', 'vtk.vtkImageData', ([], {}), '()\n', (22791, 22793), False, 'import vtk\n'), ((23358, 23375), 'itk.size', 'itk.size', (['l_image'], {}), '(l_image)\n', (23366, 23375), False, 'import itk\n'), ((25231, 25274), 'itk.image_view_from_array', 'itk.image_view_from_array', (['array', 'is_vector'], {}), '(array, is_vector)\n', (25256, 25274), False, 'import itk\n'), ((26377, 26409), 'itk.array_view_from_image', 'itk.array_view_from_image', (['image'], {}), '(image)\n', (26402, 26409), False, 'import itk\n'), ((27485, 27524), 'itk.vector_container_from_array', 'itk.vector_container_from_array', (['points'], {}), '(points)\n', (27516, 27524), False, 'import itk\n'), ((27610, 27653), 'itk.vector_container_from_array', 'itk.vector_container_from_array', (['point_data'], {}), '(point_data)\n', (27641, 27653), False, 'import itk\n'), ((27732, 27770), 'itk.vector_container_from_array', 'itk.vector_container_from_array', (['cells'], {}), '(cells)\n', (27763, 27770), False, 'import itk\n'), ((27856, 27898), 'itk.vector_container_from_array', 'itk.vector_container_from_array', (['cell_data'], {}), '(cell_data)\n', (27887, 27898), False, 'import itk\n'), ((28111, 28129), 'itk.template', 'itk.template', (['mesh'], {}), '(mesh)\n', (28123, 28129), False, 'import itk\n'), ((30654, 30693), 'itk.vector_container_from_array', 'itk.vector_container_from_array', (['points'], {}), '(points)\n', (30685, 30693), False, 'import itk\n'), ((30783, 30826), 'itk.vector_container_from_array', 'itk.vector_container_from_array', (['point_data'], {}), '(point_data)\n', (30814, 30826), False, 'import itk\n'), ((31060, 31082), 'itk.template', 'itk.template', (['pointset'], {}), '(pointset)\n', (31072, 31082), False, 'import itk\n'), ((32811, 32838), 'itk.output', 'itk.output', (['image_or_filter'], {}), '(image_or_filter)\n', (32821, 32838), False, 'import itk\n'), ((34248, 34275), 'itk.output', 'itk.output', (['image_or_filter'], {}), '(image_or_filter)\n', (34258, 34275), False, 'import itk\n'), ((44250, 44280), 're.compile', 're.compile', (['"""(_)([a-z0-9A-Z])"""'], {}), "('(_)([a-z0-9A-Z])')\n", (44260, 44280), False, 'import re\n'), ((60532, 60555), 'itk.output', 'itk.output', (['inputobject'], {}), '(inputobject)\n', (60542, 60555), False, 'import itk\n'), ((61265, 61288), 'itk.output', 'itk.output', (['inputObject'], {}), '(inputObject)\n', (61275, 61288), False, 'import itk\n'), ((62100, 62123), 'itk.output', 'itk.output', (['inputobject'], {}), '(inputobject)\n', (62110, 62123), False, 'import itk\n'), ((62868, 62895), 'itk.output', 'itk.output', (['image_or_filter'], {}), '(image_or_filter)\n', (62878, 62895), False, 'import itk\n'), ((63150, 63424), 're.compile', 're.compile', (['"""\n \'.*?\' | # single quoted strings or\n ".*?" | # double quoted strings or\n \\\\w+ | # identifier\n \\\\S # other characters\n """', '(re.VERBOSE | re.DOTALL)'], {}), '(\n """\n \'.*?\' | # single quoted strings or\n ".*?" | # double quoted strings or\n \\\\w+ | # identifier\n \\\\S # other characters\n """\n , re.VERBOSE | re.DOTALL)\n', (63160, 63424), False, 'import re\n'), ((63446, 63467), 'IPython.get_ipython', 'IPython.get_ipython', ([], {}), '()\n', (63465, 63467), False, 'import IPython\n'), ((67944, 67964), 'inspect.isclass', 'inspect.isclass', (['obj'], {}), '(obj)\n', (67959, 67964), False, 'import inspect\n'), ((73021, 73042), 'IPython.get_ipython', 'IPython.get_ipython', ([], {}), '()\n', (73040, 73042), False, 'import IPython\n'), ((9817, 9832), 'numpy.asarray', 'np.asarray', (['arr'], {}), '(arr)\n', (9827, 9832), True, 'import numpy as np\n'), ((14373, 14388), 'numpy.asarray', 'np.asarray', (['arr'], {}), '(arr)\n', (14383, 14388), True, 'import numpy as np\n'), ((17209, 17224), 'numpy.asarray', 'np.asarray', (['arr'], {}), '(arr)\n', (17219, 17224), True, 'import numpy as np\n'), ((18556, 18571), 'numpy.asarray', 'np.asarray', (['arr'], {}), '(arr)\n', (18566, 18571), True, 'import numpy as np\n'), ((19705, 19838), 'numpy.linspace', 'np.linspace', (['l_origin[l_index]', '(l_origin[l_index] + (l_size[l_index] - 1) * l_spacing[l_index])', 'l_size[l_index]'], {'dtype': 'np.float64'}), '(l_origin[l_index], l_origin[l_index] + (l_size[l_index] - 1) *\n l_spacing[l_index], l_size[l_index], dtype=np.float64)\n', (19716, 19838), True, 'import numpy as np\n'), ((20078, 20116), 'numpy.arange', 'np.arange', (['components'], {'dtype': 'np.uint32'}), '(components, dtype=np.uint32)\n', (20087, 20116), True, 'import numpy as np\n'), ((23472, 23507), 'vtk.vtkVersion.GetVTKMajorVersion', 'vtk.vtkVersion.GetVTKMajorVersion', ([], {}), '()\n', (23505, 23507), False, 'import vtk\n'), ((25593, 25628), 'vtk.vtkVersion.GetVTKMajorVersion', 'vtk.vtkVersion.GetVTKMajorVersion', ([], {}), '()\n', (25626, 25628), False, 'import vtk\n'), ((26114, 26152), 'itk.matrix_from_array', 'itk.matrix_from_array', (['direction_array'], {}), '(direction_array)\n', (26135, 26152), False, 'import itk\n'), ((28356, 28377), 'numpy.array', 'np.array', (['[]', 'np.uint'], {}), '([], np.uint)\n', (28364, 28377), True, 'import numpy as np\n'), ((28523, 28547), 'numpy.array', 'np.array', (['[]', 'np.float32'], {}), '([], np.float32)\n', (28531, 28547), True, 'import numpy as np\n'), ((28742, 28762), 'numpy.array', 'np.array', (['[]', 'mangle'], {}), '([], mangle)\n', (28750, 28762), True, 'import numpy as np\n'), ((28800, 28848), 'itk.array_view_from_vector_container', 'itk.array_view_from_vector_container', (['point_data'], {}), '(point_data)\n', (28836, 28848), False, 'import itk\n'), ((28941, 28961), 'numpy.array', 'np.array', (['[]', 'mangle'], {}), '([], mangle)\n', (28949, 28961), True, 'import numpy as np\n'), ((28998, 29045), 'itk.array_view_from_vector_container', 'itk.array_view_from_vector_container', (['cell_data'], {}), '(cell_data)\n', (29034, 29045), False, 'import itk\n'), ((31277, 31301), 'numpy.array', 'np.array', (['[]', 'np.float32'], {}), '([], np.float32)\n', (31285, 31301), True, 'import numpy as np\n'), ((31504, 31524), 'numpy.array', 'np.array', (['[]', 'mangle'], {}), '([], mangle)\n', (31512, 31524), True, 'import numpy as np\n'), ((31562, 31610), 'itk.array_view_from_vector_container', 'itk.array_view_from_vector_container', (['point_data'], {}), '(point_data)\n', (31598, 31610), False, 'import itk\n'), ((36891, 36914), 'os.path.isdir', 'os.path.isdir', (['filename'], {}), '(filename)\n', (36904, 36914), False, 'import os\n'), ((38210, 38297), 'itk.ImageIOFactory.CreateImageIO', 'itk.ImageIOFactory.CreateImageIO', (['io_filename', 'itk.CommonEnums.IOFileMode_ReadMode'], {}), '(io_filename, itk.CommonEnums.\n IOFileMode_ReadMode)\n', (38242, 38297), False, 'import itk\n'), ((40985, 41070), 'itk.MeshIOFactory.CreateMeshIO', 'itk.MeshIOFactory.CreateMeshIO', (['io_filename', 'itk.CommonEnums.IOFileMode_ReadMode'], {}), '(io_filename, itk.CommonEnums.IOFileMode_ReadMode\n )\n', (41015, 41070), False, 'import itk\n'), ((64452, 64471), 're.compile', 're.compile', (['"""\\\\w+$"""'], {}), "('\\\\w+$')\n", (64462, 64471), False, 'import re\n'), ((8516, 8535), 'itk.PyBuffer.keys', 'itk.PyBuffer.keys', ([], {}), '()\n', (8533, 8535), False, 'import itk\n'), ((11366, 11385), 'itk.PyBuffer.keys', 'itk.PyBuffer.keys', ([], {}), '()\n', (11383, 11385), False, 'import itk\n'), ((12899, 12927), 'itk.PyVectorContainer.keys', 'itk.PyVectorContainer.keys', ([], {}), '()\n', (12925, 12927), False, 'import itk\n'), ((13837, 13865), 'itk.PyVectorContainer.keys', 'itk.PyVectorContainer.keys', ([], {}), '()\n', (13863, 13865), False, 'import itk\n'), ((14885, 14913), 'itk.PyVectorContainer.keys', 'itk.PyVectorContainer.keys', ([], {}), '()\n', (14911, 14913), False, 'import itk\n'), ((15702, 15718), 'itk.PyVnl.keys', 'itk.PyVnl.keys', ([], {}), '()\n', (15716, 15718), False, 'import itk\n'), ((17556, 17572), 'itk.PyVnl.keys', 'itk.PyVnl.keys', ([], {}), '()\n', (17570, 17572), False, 'import itk\n'), ((22274, 22292), 'numpy.flip', 'np.flip', (['direction'], {}), '(direction)\n', (22281, 22292), True, 'import numpy as np\n'), ((25737, 25751), 'numpy.identity', 'np.identity', (['(3)'], {}), '(3)\n', (25748, 25751), True, 'import numpy as np\n'), ((37049, 37078), 'itk.GDCMSeriesFileNames.New', 'itk.GDCMSeriesFileNames.New', ([], {}), '()\n', (37076, 37078), False, 'import itk\n'), ((42329, 42353), 'itk.down_cast', 'itk.down_cast', (['transform'], {}), '(transform)\n', (42342, 42353), False, 'import itk\n'), ((7827, 7874), 'numpy.issubdtype', 'np.issubdtype', (['numpy_array_type.dtype.type', 'key'], {}), '(numpy_array_type.dtype.type, key)\n', (7840, 7874), True, 'import numpy as np\n'), ((15648, 15672), 'itk.template', 'itk.template', (['vnl_object'], {}), '(vnl_object)\n', (15660, 15672), False, 'import itk\n'), ((21607, 21640), 'numpy.moveaxis', 'np.moveaxis', (['values', 'source', 'dest'], {}), '(values, source, dest)\n', (21618, 21640), True, 'import numpy as np\n'), ((24076, 24097), 'itk.template', 'itk.template', (['l_image'], {}), '(l_image)\n', (24088, 24097), False, 'import itk\n'), ((25944, 25958), 'numpy.identity', 'np.identity', (['(2)'], {}), '(2)\n', (25955, 25958), True, 'import numpy as np\n'), ((65279, 65316), 're.findall', 're.findall', (['"""([A-Z])"""', 'callable_match'], {}), "('([A-Z])', callable_match)\n", (65289, 65316), False, 'import re\n'), ((10206, 10229), 'itk.template', 'itk.template', (['ImageType'], {}), '(ImageType)\n', (10218, 10229), False, 'import itk\n'), ((10247, 10270), 'itk.template', 'itk.template', (['ImageType'], {}), '(ImageType)\n', (10259, 10270), False, 'import itk\n'), ((10475, 10498), 'itk.template', 'itk.template', (['ImageType'], {}), '(ImageType)\n', (10487, 10498), False, 'import itk\n'), ((38789, 38813), 'itk.template', 'itk.template', (['pixel_type'], {}), '(pixel_type)\n', (38801, 38813), False, 'import itk\n'), ((49172, 49189), 'itk.output', 'itk.output', (['value'], {}), '(value)\n', (49182, 49189), False, 'import itk\n'), ((66559, 66584), 'inspect.isclass', 'inspect.isclass', (['l_object'], {}), '(l_object)\n', (66574, 66584), False, 'import inspect\n'), ((23579, 23613), 'itk.array_from_matrix', 'itk.array_from_matrix', (['l_direction'], {}), '(l_direction)\n', (23600, 23613), False, 'import itk\n'), ((48979, 48992), 'itk.output', 'itk.output', (['x'], {}), '(x)\n', (48989, 48992), False, 'import itk\n'), ((73052, 73073), 'IPython.get_ipython', 'IPython.get_ipython', ([], {}), '()\n', (73071, 73073), False, 'import IPython\n'), ((10399, 10422), 'itk.template', 'itk.template', (['ImageType'], {}), '(ImageType)\n', (10411, 10422), False, 'import itk\n'), ((38946, 38970), 'itk.template', 'itk.template', (['pixel_type'], {}), '(pixel_type)\n', (38958, 38970), False, 'import itk\n'), ((49112, 49129), 'itk.output', 'itk.output', (['value'], {}), '(value)\n', (49122, 49129), False, 'import itk\n')]
|
import pandas as pd
from scipy.stats import t
import numpy as np
import requests
def make_dataframe(r):
rows = []
for item in r['data']:
rows.append([item['lat'], item['lon'], item['aqi'], item['station']['name']])
df = pd.DataFrame(rows, columns=['lat', 'lon', 'aqi', 'name'])
df['aqi'] = pd.to_numeric(df.aqi, errors='coerce')
return df
def one_samp_t_test(df, diff):
return diff / (
df['aqi'].var() / df.count()) ** (1 / 2)
def get_request_data(url):
return requests.get(url).json()
class Air_Quality_Analytics():
def __init__(self):
self.base_url = "https://api.waqi.info/feed/"
self.city_str = ""
self.url = self.base_url + self.city_str + "/?token=<PASSWORD>"
def get_local_air_quality_comparison(self, city_str, tolerance=2.0):
self.city_str = city_str
token = "<PASSWORD>"
req_data = get_request_data(self.base_url + self.city_str + "/?token=" + token)
lat, lng = req_data['data']['city']['geo']
latlngbx = str(lat) + "," + str(lng) + "," + str(lat + tolerance) + "," + str(lng + tolerance)
r = requests.get("https://api.waqi.info/" + f"/map/bounds/?latlng={latlngbx}&token={token}").json()
if len(r['data']) > 0:
local_df = make_dataframe(r)
air_quality_comp = {
'deviation': 'Not found',
'probability': 'Not found'
}
deviation = local_df[local_df['name'].str.contains(city_str)]['aqi'].mean() - local_df['aqi'].mean()
if not np.isnan(deviation):
air_quality_comp['deviation'] = deviation
probability = one_samp_t_test(local_df[local_df['name'].str.contains(city_str)], deviation)
probability = t.sf(np.abs(probability), local_df.count() - 1)[0]
if not np.isnan(probability):
air_quality_comp['probability'] = probability
return air_quality_comp
def get_air_quality_index(self, city_str):
self.city_str = city_str
try:
return get_request_data(self.base_url + self.city_str + "/?token=<PASSWORD>")[
'data']["aqi"]
except:
pass
# AQA = Air_Quality_Analytics()
# print(AQA.get_local_air_quality_comparison('Los Angeles'))
|
[
"numpy.abs",
"requests.get",
"pandas.to_numeric",
"numpy.isnan",
"pandas.DataFrame"
] |
[((242, 299), 'pandas.DataFrame', 'pd.DataFrame', (['rows'], {'columns': "['lat', 'lon', 'aqi', 'name']"}), "(rows, columns=['lat', 'lon', 'aqi', 'name'])\n", (254, 299), True, 'import pandas as pd\n'), ((316, 354), 'pandas.to_numeric', 'pd.to_numeric', (['df.aqi'], {'errors': '"""coerce"""'}), "(df.aqi, errors='coerce')\n", (329, 354), True, 'import pandas as pd\n'), ((511, 528), 'requests.get', 'requests.get', (['url'], {}), '(url)\n', (523, 528), False, 'import requests\n'), ((1138, 1230), 'requests.get', 'requests.get', (["('https://api.waqi.info/' + f'/map/bounds/?latlng={latlngbx}&token={token}')"], {}), "('https://api.waqi.info/' +\n f'/map/bounds/?latlng={latlngbx}&token={token}')\n", (1150, 1230), False, 'import requests\n'), ((1572, 1591), 'numpy.isnan', 'np.isnan', (['deviation'], {}), '(deviation)\n', (1580, 1591), True, 'import numpy as np\n'), ((1853, 1874), 'numpy.isnan', 'np.isnan', (['probability'], {}), '(probability)\n', (1861, 1874), True, 'import numpy as np\n'), ((1787, 1806), 'numpy.abs', 'np.abs', (['probability'], {}), '(probability)\n', (1793, 1806), True, 'import numpy as np\n')]
|
# Copyright 2017 The TensorFlow Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
"""Tests for object_detection.core.preprocessor."""
import numpy as np
import six
import tensorflow as tf
from object_detection.tensorflow_detect.core import standard_fields as fields, \
preprocessor, preprocessor_cache
if six.PY2:
import mock # pylint: disable=g-import-not-at-top
else:
from unittest import mock # pylint: disable=g-import-not-at-top
class PreprocessorTest(tf.test.TestCase):
def createColorfulTestImage(self):
ch255 = tf.fill([1, 100, 200, 1], tf.constant(255, dtype=tf.uint8))
ch128 = tf.fill([1, 100, 200, 1], tf.constant(128, dtype=tf.uint8))
ch0 = tf.fill([1, 100, 200, 1], tf.constant(0, dtype=tf.uint8))
imr = tf.concat([ch255, ch0, ch0], 3)
img = tf.concat([ch255, ch255, ch0], 3)
imb = tf.concat([ch255, ch0, ch255], 3)
imw = tf.concat([ch128, ch128, ch128], 3)
imu = tf.concat([imr, img], 2)
imd = tf.concat([imb, imw], 2)
im = tf.concat([imu, imd], 1)
return im
def createTestImages(self):
images_r = tf.constant([[[128, 128, 128, 128], [0, 0, 128, 128],
[0, 128, 128, 128], [192, 192, 128, 128]]],
dtype=tf.uint8)
images_r = tf.expand_dims(images_r, 3)
images_g = tf.constant([[[0, 0, 128, 128], [0, 0, 128, 128],
[0, 128, 192, 192], [192, 192, 128, 192]]],
dtype=tf.uint8)
images_g = tf.expand_dims(images_g, 3)
images_b = tf.constant([[[128, 128, 192, 0], [0, 0, 128, 192],
[0, 128, 128, 0], [192, 192, 192, 128]]],
dtype=tf.uint8)
images_b = tf.expand_dims(images_b, 3)
images = tf.concat([images_r, images_g, images_b], 3)
return images
def createEmptyTestBoxes(self):
boxes = tf.constant([[]], dtype=tf.float32)
return boxes
def createTestBoxes(self):
boxes = tf.constant(
[[0.0, 0.25, 0.75, 1.0], [0.25, 0.5, 0.75, 1.0]], dtype=tf.float32)
return boxes
def createTestLabelScores(self):
return tf.constant([1.0, 0.5], dtype=tf.float32)
def createTestLabelScoresWithMissingScore(self):
return tf.constant([0.5, np.nan], dtype=tf.float32)
def createTestMasks(self):
mask = np.array([
[[255.0, 0.0, 0.0],
[255.0, 0.0, 0.0],
[255.0, 0.0, 0.0]],
[[255.0, 255.0, 0.0],
[255.0, 255.0, 0.0],
[255.0, 255.0, 0.0]]])
return tf.constant(mask, dtype=tf.float32)
def createTestKeypoints(self):
keypoints = np.array([
[[0.1, 0.1], [0.2, 0.2], [0.3, 0.3]],
[[0.4, 0.4], [0.5, 0.5], [0.6, 0.6]],
])
return tf.constant(keypoints, dtype=tf.float32)
def createTestKeypointsInsideCrop(self):
keypoints = np.array([
[[0.4, 0.4], [0.5, 0.5], [0.6, 0.6]],
[[0.4, 0.4], [0.5, 0.5], [0.6, 0.6]],
])
return tf.constant(keypoints, dtype=tf.float32)
def createTestKeypointsOutsideCrop(self):
keypoints = np.array([
[[0.1, 0.1], [0.2, 0.2], [0.3, 0.3]],
[[0.1, 0.1], [0.2, 0.2], [0.3, 0.3]],
])
return tf.constant(keypoints, dtype=tf.float32)
def createKeypointFlipPermutation(self):
return np.array([0, 2, 1], dtype=np.int32)
def createTestLabels(self):
labels = tf.constant([1, 2], dtype=tf.int32)
return labels
def createTestBoxesOutOfImage(self):
boxes = tf.constant(
[[-0.1, 0.25, 0.75, 1], [0.25, 0.5, 0.75, 1.1]], dtype=tf.float32)
return boxes
def createTestMultiClassScores(self):
return tf.constant([[1.0, 0.0], [0.5, 0.5]], dtype=tf.float32)
def expectedImagesAfterNormalization(self):
images_r = tf.constant([[[0, 0, 0, 0], [-1, -1, 0, 0],
[-1, 0, 0, 0], [0.5, 0.5, 0, 0]]],
dtype=tf.float32)
images_r = tf.expand_dims(images_r, 3)
images_g = tf.constant([[[-1, -1, 0, 0], [-1, -1, 0, 0],
[-1, 0, 0.5, 0.5], [0.5, 0.5, 0, 0.5]]],
dtype=tf.float32)
images_g = tf.expand_dims(images_g, 3)
images_b = tf.constant([[[0, 0, 0.5, -1], [-1, -1, 0, 0.5],
[-1, 0, 0, -1], [0.5, 0.5, 0.5, 0]]],
dtype=tf.float32)
images_b = tf.expand_dims(images_b, 3)
images = tf.concat([images_r, images_g, images_b], 3)
return images
def expectedMaxImageAfterColorScale(self):
images_r = tf.constant([[[0.1, 0.1, 0.1, 0.1], [-0.9, -0.9, 0.1, 0.1],
[-0.9, 0.1, 0.1, 0.1], [0.6, 0.6, 0.1, 0.1]]],
dtype=tf.float32)
images_r = tf.expand_dims(images_r, 3)
images_g = tf.constant([[[-0.9, -0.9, 0.1, 0.1], [-0.9, -0.9, 0.1, 0.1],
[-0.9, 0.1, 0.6, 0.6], [0.6, 0.6, 0.1, 0.6]]],
dtype=tf.float32)
images_g = tf.expand_dims(images_g, 3)
images_b = tf.constant([[[0.1, 0.1, 0.6, -0.9], [-0.9, -0.9, 0.1, 0.6],
[-0.9, 0.1, 0.1, -0.9], [0.6, 0.6, 0.6, 0.1]]],
dtype=tf.float32)
images_b = tf.expand_dims(images_b, 3)
images = tf.concat([images_r, images_g, images_b], 3)
return images
def expectedMinImageAfterColorScale(self):
images_r = tf.constant([[[-0.1, -0.1, -0.1, -0.1], [-1, -1, -0.1, -0.1],
[-1, -0.1, -0.1, -0.1], [0.4, 0.4, -0.1, -0.1]]],
dtype=tf.float32)
images_r = tf.expand_dims(images_r, 3)
images_g = tf.constant([[[-1, -1, -0.1, -0.1], [-1, -1, -0.1, -0.1],
[-1, -0.1, 0.4, 0.4], [0.4, 0.4, -0.1, 0.4]]],
dtype=tf.float32)
images_g = tf.expand_dims(images_g, 3)
images_b = tf.constant([[[-0.1, -0.1, 0.4, -1], [-1, -1, -0.1, 0.4],
[-1, -0.1, -0.1, -1], [0.4, 0.4, 0.4, -0.1]]],
dtype=tf.float32)
images_b = tf.expand_dims(images_b, 3)
images = tf.concat([images_r, images_g, images_b], 3)
return images
def expectedImagesAfterLeftRightFlip(self):
images_r = tf.constant([[[0, 0, 0, 0], [0, 0, -1, -1],
[0, 0, 0, -1], [0, 0, 0.5, 0.5]]],
dtype=tf.float32)
images_r = tf.expand_dims(images_r, 3)
images_g = tf.constant([[[0, 0, -1, -1], [0, 0, -1, -1],
[0.5, 0.5, 0, -1], [0.5, 0, 0.5, 0.5]]],
dtype=tf.float32)
images_g = tf.expand_dims(images_g, 3)
images_b = tf.constant([[[-1, 0.5, 0, 0], [0.5, 0, -1, -1],
[-1, 0, 0, -1], [0, 0.5, 0.5, 0.5]]],
dtype=tf.float32)
images_b = tf.expand_dims(images_b, 3)
images = tf.concat([images_r, images_g, images_b], 3)
return images
def expectedImagesAfterUpDownFlip(self):
images_r = tf.constant([[[0.5, 0.5, 0, 0], [-1, 0, 0, 0],
[-1, -1, 0, 0], [0, 0, 0, 0]]],
dtype=tf.float32)
images_r = tf.expand_dims(images_r, 3)
images_g = tf.constant([[[0.5, 0.5, 0, 0.5], [-1, 0, 0.5, 0.5],
[-1, -1, 0, 0], [-1, -1, 0, 0]]],
dtype=tf.float32)
images_g = tf.expand_dims(images_g, 3)
images_b = tf.constant([[[0.5, 0.5, 0.5, 0], [-1, 0, 0, -1],
[-1, -1, 0, 0.5], [0, 0, 0.5, -1]]],
dtype=tf.float32)
images_b = tf.expand_dims(images_b, 3)
images = tf.concat([images_r, images_g, images_b], 3)
return images
def expectedImagesAfterRot90(self):
images_r = tf.constant([[[0, 0, 0, 0], [0, 0, 0, 0],
[0, -1, 0, 0.5], [0, -1, -1, 0.5]]],
dtype=tf.float32)
images_r = tf.expand_dims(images_r, 3)
images_g = tf.constant([[[0, 0, 0.5, 0.5], [0, 0, 0.5, 0],
[-1, -1, 0, 0.5], [-1, -1, -1, 0.5]]],
dtype=tf.float32)
images_g = tf.expand_dims(images_g, 3)
images_b = tf.constant([[[-1, 0.5, -1, 0], [0.5, 0, 0, 0.5],
[0, -1, 0, 0.5], [0, -1, -1, 0.5]]],
dtype=tf.float32)
images_b = tf.expand_dims(images_b, 3)
images = tf.concat([images_r, images_g, images_b], 3)
return images
def expectedBoxesAfterLeftRightFlip(self):
boxes = tf.constant([[0.0, 0.0, 0.75, 0.75], [0.25, 0.0, 0.75, 0.5]],
dtype=tf.float32)
return boxes
def expectedBoxesAfterUpDownFlip(self):
boxes = tf.constant([[0.25, 0.25, 1.0, 1.0], [0.25, 0.5, 0.75, 1.0]],
dtype=tf.float32)
return boxes
def expectedBoxesAfterRot90(self):
boxes = tf.constant(
[[0.0, 0.0, 0.75, 0.75], [0.0, 0.25, 0.5, 0.75]], dtype=tf.float32)
return boxes
def expectedMasksAfterLeftRightFlip(self):
mask = np.array([
[[0.0, 0.0, 255.0],
[0.0, 0.0, 255.0],
[0.0, 0.0, 255.0]],
[[0.0, 255.0, 255.0],
[0.0, 255.0, 255.0],
[0.0, 255.0, 255.0]]])
return tf.constant(mask, dtype=tf.float32)
def expectedMasksAfterUpDownFlip(self):
mask = np.array([
[[255.0, 0.0, 0.0],
[255.0, 0.0, 0.0],
[255.0, 0.0, 0.0]],
[[255.0, 255.0, 0.0],
[255.0, 255.0, 0.0],
[255.0, 255.0, 0.0]]])
return tf.constant(mask, dtype=tf.float32)
def expectedMasksAfterRot90(self):
mask = np.array([
[[0.0, 0.0, 0.0],
[0.0, 0.0, 0.0],
[255.0, 255.0, 255.0]],
[[0.0, 0.0, 0.0],
[255.0, 255.0, 255.0],
[255.0, 255.0, 255.0]]])
return tf.constant(mask, dtype=tf.float32)
def expectedLabelScoresAfterThresholding(self):
return tf.constant([1.0], dtype=tf.float32)
def expectedBoxesAfterThresholding(self):
return tf.constant([[0.0, 0.25, 0.75, 1.0]], dtype=tf.float32)
def expectedLabelsAfterThresholding(self):
return tf.constant([1], dtype=tf.float32)
def expectedMultiClassScoresAfterThresholding(self):
return tf.constant([[1.0, 0.0]], dtype=tf.float32)
def expectedMasksAfterThresholding(self):
mask = np.array([
[[255.0, 0.0, 0.0],
[255.0, 0.0, 0.0],
[255.0, 0.0, 0.0]]])
return tf.constant(mask, dtype=tf.float32)
def expectedKeypointsAfterThresholding(self):
keypoints = np.array([
[[0.1, 0.1], [0.2, 0.2], [0.3, 0.3]]
])
return tf.constant(keypoints, dtype=tf.float32)
def expectedLabelScoresAfterThresholdingWithMissingScore(self):
return tf.constant([np.nan], dtype=tf.float32)
def expectedBoxesAfterThresholdingWithMissingScore(self):
return tf.constant([[0.25, 0.5, 0.75, 1]], dtype=tf.float32)
def expectedLabelsAfterThresholdingWithMissingScore(self):
return tf.constant([2], dtype=tf.float32)
def testRgbToGrayscale(self):
images = self.createTestImages()
grayscale_images = preprocessor._rgb_to_grayscale(images)
expected_images = tf.image.rgb_to_grayscale(images)
with self.test_session() as sess:
(grayscale_images, expected_images) = sess.run(
[grayscale_images, expected_images])
self.assertAllEqual(expected_images, grayscale_images)
def testNormalizeImage(self):
preprocess_options = [(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 256,
'target_minval': -1,
'target_maxval': 1
})]
images = self.createTestImages()
tensor_dict = {fields.InputDataFields.image: images}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocess_options)
images = tensor_dict[fields.InputDataFields.image]
images_expected = self.expectedImagesAfterNormalization()
with self.test_session() as sess:
(images_, images_expected_) = sess.run(
[images, images_expected])
images_shape_ = images_.shape
images_expected_shape_ = images_expected_.shape
expected_shape = [1, 4, 4, 3]
self.assertAllEqual(images_expected_shape_, images_shape_)
self.assertAllEqual(images_shape_, expected_shape)
self.assertAllClose(images_, images_expected_)
def testRetainBoxesAboveThreshold(self):
boxes = self.createTestBoxes()
labels = self.createTestLabels()
label_scores = self.createTestLabelScores()
(retained_boxes, retained_labels,
retained_label_scores) = preprocessor.retain_boxes_above_threshold(
boxes, labels, label_scores, threshold=0.6)
with self.test_session() as sess:
(retained_boxes_, retained_labels_, retained_label_scores_,
expected_retained_boxes_, expected_retained_labels_,
expected_retained_label_scores_) = sess.run([
retained_boxes, retained_labels, retained_label_scores,
self.expectedBoxesAfterThresholding(),
self.expectedLabelsAfterThresholding(),
self.expectedLabelScoresAfterThresholding()])
self.assertAllClose(
retained_boxes_, expected_retained_boxes_)
self.assertAllClose(
retained_labels_, expected_retained_labels_)
self.assertAllClose(
retained_label_scores_, expected_retained_label_scores_)
def testRetainBoxesAboveThresholdWithMultiClassScores(self):
boxes = self.createTestBoxes()
labels = self.createTestLabels()
label_scores = self.createTestLabelScores()
multiclass_scores = self.createTestMultiClassScores()
(_, _, _,
retained_multiclass_scores) = preprocessor.retain_boxes_above_threshold(
boxes,
labels,
label_scores,
multiclass_scores=multiclass_scores,
threshold=0.6)
with self.test_session() as sess:
(retained_multiclass_scores_,
expected_retained_multiclass_scores_) = sess.run([
retained_multiclass_scores,
self.expectedMultiClassScoresAfterThresholding()
])
self.assertAllClose(retained_multiclass_scores_,
expected_retained_multiclass_scores_)
def testRetainBoxesAboveThresholdWithMasks(self):
boxes = self.createTestBoxes()
labels = self.createTestLabels()
label_scores = self.createTestLabelScores()
masks = self.createTestMasks()
_, _, _, retained_masks = preprocessor.retain_boxes_above_threshold(
boxes, labels, label_scores, masks, threshold=0.6)
with self.test_session() as sess:
retained_masks_, expected_retained_masks_ = sess.run([
retained_masks, self.expectedMasksAfterThresholding()])
self.assertAllClose(
retained_masks_, expected_retained_masks_)
def testRetainBoxesAboveThresholdWithKeypoints(self):
boxes = self.createTestBoxes()
labels = self.createTestLabels()
label_scores = self.createTestLabelScores()
keypoints = self.createTestKeypoints()
(_, _, _, retained_keypoints) = preprocessor.retain_boxes_above_threshold(
boxes, labels, label_scores, keypoints=keypoints, threshold=0.6)
with self.test_session() as sess:
(retained_keypoints_,
expected_retained_keypoints_) = sess.run([
retained_keypoints,
self.expectedKeypointsAfterThresholding()])
self.assertAllClose(
retained_keypoints_, expected_retained_keypoints_)
def testRetainBoxesAboveThresholdWithMissingScore(self):
boxes = self.createTestBoxes()
labels = self.createTestLabels()
label_scores = self.createTestLabelScoresWithMissingScore()
(retained_boxes, retained_labels,
retained_label_scores) = preprocessor.retain_boxes_above_threshold(
boxes, labels, label_scores, threshold=0.6)
with self.test_session() as sess:
(retained_boxes_, retained_labels_, retained_label_scores_,
expected_retained_boxes_, expected_retained_labels_,
expected_retained_label_scores_) = sess.run([
retained_boxes, retained_labels, retained_label_scores,
self.expectedBoxesAfterThresholdingWithMissingScore(),
self.expectedLabelsAfterThresholdingWithMissingScore(),
self.expectedLabelScoresAfterThresholdingWithMissingScore()])
self.assertAllClose(
retained_boxes_, expected_retained_boxes_)
self.assertAllClose(
retained_labels_, expected_retained_labels_)
self.assertAllClose(
retained_label_scores_, expected_retained_label_scores_)
def testFlipBoxesLeftRight(self):
boxes = self.createTestBoxes()
flipped_boxes = preprocessor._flip_boxes_left_right(boxes)
expected_boxes = self.expectedBoxesAfterLeftRightFlip()
with self.test_session() as sess:
flipped_boxes, expected_boxes = sess.run([flipped_boxes, expected_boxes])
self.assertAllEqual(flipped_boxes.flatten(), expected_boxes.flatten())
def testFlipBoxesUpDown(self):
boxes = self.createTestBoxes()
flipped_boxes = preprocessor._flip_boxes_up_down(boxes)
expected_boxes = self.expectedBoxesAfterUpDownFlip()
with self.test_session() as sess:
flipped_boxes, expected_boxes = sess.run([flipped_boxes, expected_boxes])
self.assertAllEqual(flipped_boxes.flatten(), expected_boxes.flatten())
def testRot90Boxes(self):
boxes = self.createTestBoxes()
rotated_boxes = preprocessor._rot90_boxes(boxes)
expected_boxes = self.expectedBoxesAfterRot90()
with self.test_session() as sess:
rotated_boxes, expected_boxes = sess.run([rotated_boxes, expected_boxes])
self.assertAllEqual(rotated_boxes.flatten(), expected_boxes.flatten())
def testFlipMasksLeftRight(self):
test_mask = self.createTestMasks()
flipped_mask = preprocessor._flip_masks_left_right(test_mask)
expected_mask = self.expectedMasksAfterLeftRightFlip()
with self.test_session() as sess:
flipped_mask, expected_mask = sess.run([flipped_mask, expected_mask])
self.assertAllEqual(flipped_mask.flatten(), expected_mask.flatten())
def testFlipMasksUpDown(self):
test_mask = self.createTestMasks()
flipped_mask = preprocessor._flip_masks_up_down(test_mask)
expected_mask = self.expectedMasksAfterUpDownFlip()
with self.test_session() as sess:
flipped_mask, expected_mask = sess.run([flipped_mask, expected_mask])
self.assertAllEqual(flipped_mask.flatten(), expected_mask.flatten())
def testRot90Masks(self):
test_mask = self.createTestMasks()
rotated_mask = preprocessor._rot90_masks(test_mask)
expected_mask = self.expectedMasksAfterRot90()
with self.test_session() as sess:
rotated_mask, expected_mask = sess.run([rotated_mask, expected_mask])
self.assertAllEqual(rotated_mask.flatten(), expected_mask.flatten())
def _testPreprocessorCache(self,
preprocess_options,
test_boxes=False,
test_masks=False,
test_keypoints=False,
num_runs=4):
cache = preprocessor_cache.PreprocessorCache()
images = self.createTestImages()
boxes = self.createTestBoxes()
classes = self.createTestLabels()
masks = self.createTestMasks()
keypoints = self.createTestKeypoints()
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_instance_masks=test_masks, include_keypoints=test_keypoints)
out = []
for i in range(num_runs):
tensor_dict = {
fields.InputDataFields.image: images,
}
num_outputs = 1
if test_boxes:
tensor_dict[fields.InputDataFields.groundtruth_boxes] = boxes
tensor_dict[fields.InputDataFields.groundtruth_classes] = classes
num_outputs += 1
if test_masks:
tensor_dict[fields.InputDataFields.groundtruth_instance_masks] = masks
num_outputs += 1
if test_keypoints:
tensor_dict[fields.InputDataFields.groundtruth_keypoints] = keypoints
num_outputs += 1
out.append(preprocessor.preprocess(
tensor_dict, preprocess_options, preprocessor_arg_map, cache))
with self.test_session() as sess:
to_run = []
for i in range(num_runs):
to_run.append(out[i][fields.InputDataFields.image])
if test_boxes:
to_run.append(out[i][fields.InputDataFields.groundtruth_boxes])
if test_masks:
to_run.append(
out[i][fields.InputDataFields.groundtruth_instance_masks])
if test_keypoints:
to_run.append(out[i][fields.InputDataFields.groundtruth_keypoints])
out_array = sess.run(to_run)
for i in range(num_outputs, len(out_array)):
self.assertAllClose(out_array[i], out_array[i - num_outputs])
def testRandomHorizontalFlip(self):
preprocess_options = [(preprocessor.random_horizontal_flip, {})]
images = self.expectedImagesAfterNormalization()
boxes = self.createTestBoxes()
tensor_dict = {fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes}
images_expected1 = self.expectedImagesAfterLeftRightFlip()
boxes_expected1 = self.expectedBoxesAfterLeftRightFlip()
images_expected2 = images
boxes_expected2 = boxes
tensor_dict = preprocessor.preprocess(tensor_dict, preprocess_options)
images = tensor_dict[fields.InputDataFields.image]
boxes = tensor_dict[fields.InputDataFields.groundtruth_boxes]
boxes_diff1 = tf.squared_difference(boxes, boxes_expected1)
boxes_diff2 = tf.squared_difference(boxes, boxes_expected2)
boxes_diff = tf.multiply(boxes_diff1, boxes_diff2)
boxes_diff_expected = tf.zeros_like(boxes_diff)
images_diff1 = tf.squared_difference(images, images_expected1)
images_diff2 = tf.squared_difference(images, images_expected2)
images_diff = tf.multiply(images_diff1, images_diff2)
images_diff_expected = tf.zeros_like(images_diff)
with self.test_session() as sess:
(images_diff_, images_diff_expected_, boxes_diff_,
boxes_diff_expected_) = sess.run([images_diff, images_diff_expected,
boxes_diff, boxes_diff_expected])
self.assertAllClose(boxes_diff_, boxes_diff_expected_)
self.assertAllClose(images_diff_, images_diff_expected_)
def testRandomHorizontalFlipWithEmptyBoxes(self):
preprocess_options = [(preprocessor.random_horizontal_flip, {})]
images = self.expectedImagesAfterNormalization()
boxes = self.createEmptyTestBoxes()
tensor_dict = {fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes}
images_expected1 = self.expectedImagesAfterLeftRightFlip()
boxes_expected = self.createEmptyTestBoxes()
images_expected2 = images
tensor_dict = preprocessor.preprocess(tensor_dict, preprocess_options)
images = tensor_dict[fields.InputDataFields.image]
boxes = tensor_dict[fields.InputDataFields.groundtruth_boxes]
images_diff1 = tf.squared_difference(images, images_expected1)
images_diff2 = tf.squared_difference(images, images_expected2)
images_diff = tf.multiply(images_diff1, images_diff2)
images_diff_expected = tf.zeros_like(images_diff)
with self.test_session() as sess:
(images_diff_, images_diff_expected_, boxes_,
boxes_expected_) = sess.run([images_diff, images_diff_expected, boxes,
boxes_expected])
self.assertAllClose(boxes_, boxes_expected_)
self.assertAllClose(images_diff_, images_diff_expected_)
def testRandomHorizontalFlipWithCache(self):
keypoint_flip_permutation = self.createKeypointFlipPermutation()
preprocess_options = [
(preprocessor.random_horizontal_flip,
{'keypoint_flip_permutation': keypoint_flip_permutation})]
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=True,
test_keypoints=True)
def testRunRandomHorizontalFlipWithMaskAndKeypoints(self):
preprocess_options = [(preprocessor.random_horizontal_flip, {})]
image_height = 3
image_width = 3
images = tf.random_uniform([1, image_height, image_width, 3])
boxes = self.createTestBoxes()
masks = self.createTestMasks()
keypoints = self.createTestKeypoints()
keypoint_flip_permutation = self.createKeypointFlipPermutation()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_instance_masks: masks,
fields.InputDataFields.groundtruth_keypoints: keypoints
}
preprocess_options = [
(preprocessor.random_horizontal_flip,
{'keypoint_flip_permutation': keypoint_flip_permutation})]
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_instance_masks=True, include_keypoints=True)
tensor_dict = preprocessor.preprocess(
tensor_dict, preprocess_options, func_arg_map=preprocessor_arg_map)
boxes = tensor_dict[fields.InputDataFields.groundtruth_boxes]
masks = tensor_dict[fields.InputDataFields.groundtruth_instance_masks]
keypoints = tensor_dict[fields.InputDataFields.groundtruth_keypoints]
with self.test_session() as sess:
boxes, masks, keypoints = sess.run([boxes, masks, keypoints])
self.assertTrue(boxes is not None)
self.assertTrue(masks is not None)
self.assertTrue(keypoints is not None)
def testRandomVerticalFlip(self):
preprocess_options = [(preprocessor.random_vertical_flip, {})]
images = self.expectedImagesAfterNormalization()
boxes = self.createTestBoxes()
tensor_dict = {fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes}
images_expected1 = self.expectedImagesAfterUpDownFlip()
boxes_expected1 = self.expectedBoxesAfterUpDownFlip()
images_expected2 = images
boxes_expected2 = boxes
tensor_dict = preprocessor.preprocess(tensor_dict, preprocess_options)
images = tensor_dict[fields.InputDataFields.image]
boxes = tensor_dict[fields.InputDataFields.groundtruth_boxes]
boxes_diff1 = tf.squared_difference(boxes, boxes_expected1)
boxes_diff2 = tf.squared_difference(boxes, boxes_expected2)
boxes_diff = tf.multiply(boxes_diff1, boxes_diff2)
boxes_diff_expected = tf.zeros_like(boxes_diff)
images_diff1 = tf.squared_difference(images, images_expected1)
images_diff2 = tf.squared_difference(images, images_expected2)
images_diff = tf.multiply(images_diff1, images_diff2)
images_diff_expected = tf.zeros_like(images_diff)
with self.test_session() as sess:
(images_diff_, images_diff_expected_, boxes_diff_,
boxes_diff_expected_) = sess.run([images_diff, images_diff_expected,
boxes_diff, boxes_diff_expected])
self.assertAllClose(boxes_diff_, boxes_diff_expected_)
self.assertAllClose(images_diff_, images_diff_expected_)
def testRandomVerticalFlipWithEmptyBoxes(self):
preprocess_options = [(preprocessor.random_vertical_flip, {})]
images = self.expectedImagesAfterNormalization()
boxes = self.createEmptyTestBoxes()
tensor_dict = {fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes}
images_expected1 = self.expectedImagesAfterUpDownFlip()
boxes_expected = self.createEmptyTestBoxes()
images_expected2 = images
tensor_dict = preprocessor.preprocess(tensor_dict, preprocess_options)
images = tensor_dict[fields.InputDataFields.image]
boxes = tensor_dict[fields.InputDataFields.groundtruth_boxes]
images_diff1 = tf.squared_difference(images, images_expected1)
images_diff2 = tf.squared_difference(images, images_expected2)
images_diff = tf.multiply(images_diff1, images_diff2)
images_diff_expected = tf.zeros_like(images_diff)
with self.test_session() as sess:
(images_diff_, images_diff_expected_, boxes_,
boxes_expected_) = sess.run([images_diff, images_diff_expected, boxes,
boxes_expected])
self.assertAllClose(boxes_, boxes_expected_)
self.assertAllClose(images_diff_, images_diff_expected_)
def testRandomVerticalFlipWithCache(self):
keypoint_flip_permutation = self.createKeypointFlipPermutation()
preprocess_options = [
(preprocessor.random_vertical_flip,
{'keypoint_flip_permutation': keypoint_flip_permutation})]
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=True,
test_keypoints=True)
def testRunRandomVerticalFlipWithMaskAndKeypoints(self):
preprocess_options = [(preprocessor.random_vertical_flip, {})]
image_height = 3
image_width = 3
images = tf.random_uniform([1, image_height, image_width, 3])
boxes = self.createTestBoxes()
masks = self.createTestMasks()
keypoints = self.createTestKeypoints()
keypoint_flip_permutation = self.createKeypointFlipPermutation()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_instance_masks: masks,
fields.InputDataFields.groundtruth_keypoints: keypoints
}
preprocess_options = [
(preprocessor.random_vertical_flip,
{'keypoint_flip_permutation': keypoint_flip_permutation})]
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_instance_masks=True, include_keypoints=True)
tensor_dict = preprocessor.preprocess(
tensor_dict, preprocess_options, func_arg_map=preprocessor_arg_map)
boxes = tensor_dict[fields.InputDataFields.groundtruth_boxes]
masks = tensor_dict[fields.InputDataFields.groundtruth_instance_masks]
keypoints = tensor_dict[fields.InputDataFields.groundtruth_keypoints]
with self.test_session() as sess:
boxes, masks, keypoints = sess.run([boxes, masks, keypoints])
self.assertTrue(boxes is not None)
self.assertTrue(masks is not None)
self.assertTrue(keypoints is not None)
def testRandomRotation90(self):
preprocess_options = [(preprocessor.random_rotation90, {})]
images = self.expectedImagesAfterNormalization()
boxes = self.createTestBoxes()
tensor_dict = {fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes}
images_expected1 = self.expectedImagesAfterRot90()
boxes_expected1 = self.expectedBoxesAfterRot90()
images_expected2 = images
boxes_expected2 = boxes
tensor_dict = preprocessor.preprocess(tensor_dict, preprocess_options)
images = tensor_dict[fields.InputDataFields.image]
boxes = tensor_dict[fields.InputDataFields.groundtruth_boxes]
boxes_diff1 = tf.squared_difference(boxes, boxes_expected1)
boxes_diff2 = tf.squared_difference(boxes, boxes_expected2)
boxes_diff = tf.multiply(boxes_diff1, boxes_diff2)
boxes_diff_expected = tf.zeros_like(boxes_diff)
images_diff1 = tf.squared_difference(images, images_expected1)
images_diff2 = tf.squared_difference(images, images_expected2)
images_diff = tf.multiply(images_diff1, images_diff2)
images_diff_expected = tf.zeros_like(images_diff)
with self.test_session() as sess:
(images_diff_, images_diff_expected_, boxes_diff_,
boxes_diff_expected_) = sess.run([images_diff, images_diff_expected,
boxes_diff, boxes_diff_expected])
self.assertAllClose(boxes_diff_, boxes_diff_expected_)
self.assertAllClose(images_diff_, images_diff_expected_)
def testRandomRotation90WithEmptyBoxes(self):
preprocess_options = [(preprocessor.random_rotation90, {})]
images = self.expectedImagesAfterNormalization()
boxes = self.createEmptyTestBoxes()
tensor_dict = {fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes}
images_expected1 = self.expectedImagesAfterRot90()
boxes_expected = self.createEmptyTestBoxes()
images_expected2 = images
tensor_dict = preprocessor.preprocess(tensor_dict, preprocess_options)
images = tensor_dict[fields.InputDataFields.image]
boxes = tensor_dict[fields.InputDataFields.groundtruth_boxes]
images_diff1 = tf.squared_difference(images, images_expected1)
images_diff2 = tf.squared_difference(images, images_expected2)
images_diff = tf.multiply(images_diff1, images_diff2)
images_diff_expected = tf.zeros_like(images_diff)
with self.test_session() as sess:
(images_diff_, images_diff_expected_, boxes_,
boxes_expected_) = sess.run([images_diff, images_diff_expected, boxes,
boxes_expected])
self.assertAllClose(boxes_, boxes_expected_)
self.assertAllClose(images_diff_, images_diff_expected_)
def testRandomRotation90WithCache(self):
preprocess_options = [(preprocessor.random_rotation90, {})]
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=True,
test_keypoints=True)
def testRunRandomRotation90WithMaskAndKeypoints(self):
preprocess_options = [(preprocessor.random_rotation90, {})]
image_height = 3
image_width = 3
images = tf.random_uniform([1, image_height, image_width, 3])
boxes = self.createTestBoxes()
masks = self.createTestMasks()
keypoints = self.createTestKeypoints()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_instance_masks: masks,
fields.InputDataFields.groundtruth_keypoints: keypoints
}
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_instance_masks=True, include_keypoints=True)
tensor_dict = preprocessor.preprocess(
tensor_dict, preprocess_options, func_arg_map=preprocessor_arg_map)
boxes = tensor_dict[fields.InputDataFields.groundtruth_boxes]
masks = tensor_dict[fields.InputDataFields.groundtruth_instance_masks]
keypoints = tensor_dict[fields.InputDataFields.groundtruth_keypoints]
with self.test_session() as sess:
boxes, masks, keypoints = sess.run([boxes, masks, keypoints])
self.assertTrue(boxes is not None)
self.assertTrue(masks is not None)
self.assertTrue(keypoints is not None)
def testRandomPixelValueScale(self):
preprocessing_options = []
preprocessing_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocessing_options.append((preprocessor.random_pixel_value_scale, {}))
images = self.createTestImages()
tensor_dict = {fields.InputDataFields.image: images}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocessing_options)
images_min = tf.to_float(images) * 0.9 / 255.0
images_max = tf.to_float(images) * 1.1 / 255.0
images = tensor_dict[fields.InputDataFields.image]
values_greater = tf.greater_equal(images, images_min)
values_less = tf.less_equal(images, images_max)
values_true = tf.fill([1, 4, 4, 3], True)
with self.test_session() as sess:
(values_greater_, values_less_, values_true_) = sess.run(
[values_greater, values_less, values_true])
self.assertAllClose(values_greater_, values_true_)
self.assertAllClose(values_less_, values_true_)
def testRandomPixelValueScaleWithCache(self):
preprocess_options = []
preprocess_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocess_options.append((preprocessor.random_pixel_value_scale, {}))
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=False,
test_keypoints=False)
def testRandomImageScale(self):
preprocess_options = [(preprocessor.random_image_scale, {})]
images_original = self.createTestImages()
tensor_dict = {fields.InputDataFields.image: images_original}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocess_options)
images_scaled = tensor_dict[fields.InputDataFields.image]
images_original_shape = tf.shape(images_original)
images_scaled_shape = tf.shape(images_scaled)
with self.test_session() as sess:
(images_original_shape_, images_scaled_shape_) = sess.run(
[images_original_shape, images_scaled_shape])
self.assertTrue(
images_original_shape_[1] * 0.5 <= images_scaled_shape_[1])
self.assertTrue(
images_original_shape_[1] * 2.0 >= images_scaled_shape_[1])
self.assertTrue(
images_original_shape_[2] * 0.5 <= images_scaled_shape_[2])
self.assertTrue(
images_original_shape_[2] * 2.0 >= images_scaled_shape_[2])
def testRandomImageScaleWithCache(self):
preprocess_options = [(preprocessor.random_image_scale, {})]
self._testPreprocessorCache(preprocess_options,
test_boxes=False,
test_masks=False,
test_keypoints=False)
def testRandomRGBtoGray(self):
preprocess_options = [(preprocessor.random_rgb_to_gray, {})]
images_original = self.createTestImages()
tensor_dict = {fields.InputDataFields.image: images_original}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocess_options)
images_gray = tensor_dict[fields.InputDataFields.image]
images_gray_r, images_gray_g, images_gray_b = tf.split(
value=images_gray, num_or_size_splits=3, axis=3)
images_r, images_g, images_b = tf.split(
value=images_original, num_or_size_splits=3, axis=3)
images_r_diff1 = tf.squared_difference(tf.to_float(images_r),
tf.to_float(images_gray_r))
images_r_diff2 = tf.squared_difference(tf.to_float(images_gray_r),
tf.to_float(images_gray_g))
images_r_diff = tf.multiply(images_r_diff1, images_r_diff2)
images_g_diff1 = tf.squared_difference(tf.to_float(images_g),
tf.to_float(images_gray_g))
images_g_diff2 = tf.squared_difference(tf.to_float(images_gray_g),
tf.to_float(images_gray_b))
images_g_diff = tf.multiply(images_g_diff1, images_g_diff2)
images_b_diff1 = tf.squared_difference(tf.to_float(images_b),
tf.to_float(images_gray_b))
images_b_diff2 = tf.squared_difference(tf.to_float(images_gray_b),
tf.to_float(images_gray_r))
images_b_diff = tf.multiply(images_b_diff1, images_b_diff2)
image_zero1 = tf.constant(0, dtype=tf.float32, shape=[1, 4, 4, 1])
with self.test_session() as sess:
(images_r_diff_, images_g_diff_, images_b_diff_, image_zero1_) = sess.run(
[images_r_diff, images_g_diff, images_b_diff, image_zero1])
self.assertAllClose(images_r_diff_, image_zero1_)
self.assertAllClose(images_g_diff_, image_zero1_)
self.assertAllClose(images_b_diff_, image_zero1_)
def testRandomRGBtoGrayWithCache(self):
preprocess_options = [(
preprocessor.random_rgb_to_gray, {'probability': 0.5})]
self._testPreprocessorCache(preprocess_options,
test_boxes=False,
test_masks=False,
test_keypoints=False)
def testRandomAdjustBrightness(self):
preprocessing_options = []
preprocessing_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocessing_options.append((preprocessor.random_adjust_brightness, {}))
images_original = self.createTestImages()
tensor_dict = {fields.InputDataFields.image: images_original}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocessing_options)
images_bright = tensor_dict[fields.InputDataFields.image]
image_original_shape = tf.shape(images_original)
image_bright_shape = tf.shape(images_bright)
with self.test_session() as sess:
(image_original_shape_, image_bright_shape_) = sess.run(
[image_original_shape, image_bright_shape])
self.assertAllEqual(image_original_shape_, image_bright_shape_)
def testRandomAdjustBrightnessWithCache(self):
preprocess_options = []
preprocess_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocess_options.append((preprocessor.random_adjust_brightness, {}))
self._testPreprocessorCache(preprocess_options,
test_boxes=False,
test_masks=False,
test_keypoints=False)
def testRandomAdjustContrast(self):
preprocessing_options = []
preprocessing_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocessing_options.append((preprocessor.random_adjust_contrast, {}))
images_original = self.createTestImages()
tensor_dict = {fields.InputDataFields.image: images_original}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocessing_options)
images_contrast = tensor_dict[fields.InputDataFields.image]
image_original_shape = tf.shape(images_original)
image_contrast_shape = tf.shape(images_contrast)
with self.test_session() as sess:
(image_original_shape_, image_contrast_shape_) = sess.run(
[image_original_shape, image_contrast_shape])
self.assertAllEqual(image_original_shape_, image_contrast_shape_)
def testRandomAdjustContrastWithCache(self):
preprocess_options = []
preprocess_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocess_options.append((preprocessor.random_adjust_contrast, {}))
self._testPreprocessorCache(preprocess_options,
test_boxes=False,
test_masks=False,
test_keypoints=False)
def testRandomAdjustHue(self):
preprocessing_options = []
preprocessing_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocessing_options.append((preprocessor.random_adjust_hue, {}))
images_original = self.createTestImages()
tensor_dict = {fields.InputDataFields.image: images_original}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocessing_options)
images_hue = tensor_dict[fields.InputDataFields.image]
image_original_shape = tf.shape(images_original)
image_hue_shape = tf.shape(images_hue)
with self.test_session() as sess:
(image_original_shape_, image_hue_shape_) = sess.run(
[image_original_shape, image_hue_shape])
self.assertAllEqual(image_original_shape_, image_hue_shape_)
def testRandomAdjustHueWithCache(self):
preprocess_options = []
preprocess_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocess_options.append((preprocessor.random_adjust_hue, {}))
self._testPreprocessorCache(preprocess_options,
test_boxes=False,
test_masks=False,
test_keypoints=False)
def testRandomDistortColor(self):
preprocessing_options = []
preprocessing_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocessing_options.append((preprocessor.random_distort_color, {}))
images_original = self.createTestImages()
images_original_shape = tf.shape(images_original)
tensor_dict = {fields.InputDataFields.image: images_original}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocessing_options)
images_distorted_color = tensor_dict[fields.InputDataFields.image]
images_distorted_color_shape = tf.shape(images_distorted_color)
with self.test_session() as sess:
(images_original_shape_, images_distorted_color_shape_) = sess.run(
[images_original_shape, images_distorted_color_shape])
self.assertAllEqual(images_original_shape_, images_distorted_color_shape_)
def testRandomDistortColorWithCache(self):
preprocess_options = []
preprocess_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocess_options.append((preprocessor.random_distort_color, {}))
self._testPreprocessorCache(preprocess_options,
test_boxes=False,
test_masks=False,
test_keypoints=False)
def testRandomJitterBoxes(self):
preprocessing_options = []
preprocessing_options.append((preprocessor.random_jitter_boxes, {}))
boxes = self.createTestBoxes()
boxes_shape = tf.shape(boxes)
tensor_dict = {fields.InputDataFields.groundtruth_boxes: boxes}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocessing_options)
distorted_boxes = tensor_dict[fields.InputDataFields.groundtruth_boxes]
distorted_boxes_shape = tf.shape(distorted_boxes)
with self.test_session() as sess:
(boxes_shape_, distorted_boxes_shape_) = sess.run(
[boxes_shape, distorted_boxes_shape])
self.assertAllEqual(boxes_shape_, distorted_boxes_shape_)
def testRandomCropImage(self):
preprocessing_options = []
preprocessing_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocessing_options.append((preprocessor.random_crop_image, {}))
images = self.createTestImages()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
}
distorted_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
distorted_images = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
boxes_rank = tf.rank(boxes)
distorted_boxes_rank = tf.rank(distorted_boxes)
images_rank = tf.rank(images)
distorted_images_rank = tf.rank(distorted_images)
self.assertEqual(3, distorted_images.get_shape()[3])
with self.test_session() as sess:
(boxes_rank_, distorted_boxes_rank_, images_rank_,
distorted_images_rank_) = sess.run([
boxes_rank, distorted_boxes_rank, images_rank, distorted_images_rank
])
self.assertAllEqual(boxes_rank_, distorted_boxes_rank_)
self.assertAllEqual(images_rank_, distorted_images_rank_)
def testRandomCropImageWithCache(self):
preprocess_options = [(preprocessor.random_rgb_to_gray,
{'probability': 0.5}),
(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1,
}),
(preprocessor.random_crop_image, {})]
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=False,
test_keypoints=False)
def testRandomCropImageGrayscale(self):
preprocessing_options = [(preprocessor.rgb_to_gray, {}),
(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1,
}),
(preprocessor.random_crop_image, {})]
images = self.createTestImages()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
}
distorted_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options)
distorted_images = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
boxes_rank = tf.rank(boxes)
distorted_boxes_rank = tf.rank(distorted_boxes)
images_rank = tf.rank(images)
distorted_images_rank = tf.rank(distorted_images)
self.assertEqual(1, distorted_images.get_shape()[3])
with self.test_session() as sess:
session_results = sess.run([
boxes_rank, distorted_boxes_rank, images_rank, distorted_images_rank
])
(boxes_rank_, distorted_boxes_rank_, images_rank_,
distorted_images_rank_) = session_results
self.assertAllEqual(boxes_rank_, distorted_boxes_rank_)
self.assertAllEqual(images_rank_, distorted_images_rank_)
def testRandomCropImageWithBoxOutOfImage(self):
preprocessing_options = []
preprocessing_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocessing_options.append((preprocessor.random_crop_image, {}))
images = self.createTestImages()
boxes = self.createTestBoxesOutOfImage()
labels = self.createTestLabels()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
}
distorted_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
distorted_images = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
boxes_rank = tf.rank(boxes)
distorted_boxes_rank = tf.rank(distorted_boxes)
images_rank = tf.rank(images)
distorted_images_rank = tf.rank(distorted_images)
with self.test_session() as sess:
(boxes_rank_, distorted_boxes_rank_, images_rank_,
distorted_images_rank_) = sess.run(
[boxes_rank, distorted_boxes_rank, images_rank,
distorted_images_rank])
self.assertAllEqual(boxes_rank_, distorted_boxes_rank_)
self.assertAllEqual(images_rank_, distorted_images_rank_)
def testRandomCropImageWithRandomCoefOne(self):
preprocessing_options = [(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
})]
images = self.createTestImages()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
label_scores = self.createTestLabelScores()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.groundtruth_label_scores: label_scores
}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocessing_options)
images = tensor_dict[fields.InputDataFields.image]
preprocessing_options = [(preprocessor.random_crop_image, {
'random_coef': 1.0
})]
distorted_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
distorted_images = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
distorted_labels = distorted_tensor_dict[
fields.InputDataFields.groundtruth_classes]
distorted_label_scores = distorted_tensor_dict[
fields.InputDataFields.groundtruth_label_scores]
boxes_shape = tf.shape(boxes)
distorted_boxes_shape = tf.shape(distorted_boxes)
images_shape = tf.shape(images)
distorted_images_shape = tf.shape(distorted_images)
with self.test_session() as sess:
(boxes_shape_, distorted_boxes_shape_, images_shape_,
distorted_images_shape_, images_, distorted_images_,
boxes_, distorted_boxes_, labels_, distorted_labels_,
label_scores_, distorted_label_scores_) = sess.run(
[boxes_shape, distorted_boxes_shape, images_shape,
distorted_images_shape, images, distorted_images,
boxes, distorted_boxes, labels, distorted_labels,
label_scores, distorted_label_scores])
self.assertAllEqual(boxes_shape_, distorted_boxes_shape_)
self.assertAllEqual(images_shape_, distorted_images_shape_)
self.assertAllClose(images_, distorted_images_)
self.assertAllClose(boxes_, distorted_boxes_)
self.assertAllEqual(labels_, distorted_labels_)
self.assertAllEqual(label_scores_, distorted_label_scores_)
def testRandomCropWithMockSampleDistortedBoundingBox(self):
preprocessing_options = [(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
})]
images = self.createColorfulTestImage()
boxes = tf.constant([[0.1, 0.1, 0.8, 0.3],
[0.2, 0.4, 0.75, 0.75],
[0.3, 0.1, 0.4, 0.7]], dtype=tf.float32)
labels = tf.constant([1, 7, 11], dtype=tf.int32)
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocessing_options)
images = tensor_dict[fields.InputDataFields.image]
preprocessing_options = [(preprocessor.random_crop_image, {})]
with mock.patch.object(
tf.image,
'sample_distorted_bounding_box') as mock_sample_distorted_bounding_box:
mock_sample_distorted_bounding_box.return_value = (tf.constant(
[6, 143, 0], dtype=tf.int32), tf.constant(
[190, 237, -1], dtype=tf.int32), tf.constant(
[[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32))
distorted_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
distorted_labels = distorted_tensor_dict[
fields.InputDataFields.groundtruth_classes]
expected_boxes = tf.constant([[0.178947, 0.07173, 0.75789469, 0.66244733],
[0.28421, 0.0, 0.38947365, 0.57805908]],
dtype=tf.float32)
expected_labels = tf.constant([7, 11], dtype=tf.int32)
with self.test_session() as sess:
(distorted_boxes_, distorted_labels_,
expected_boxes_, expected_labels_) = sess.run(
[distorted_boxes, distorted_labels,
expected_boxes, expected_labels])
self.assertAllClose(distorted_boxes_, expected_boxes_)
self.assertAllEqual(distorted_labels_, expected_labels_)
def testRandomCropImageWithMultiClassScores(self):
preprocessing_options = []
preprocessing_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocessing_options.append((preprocessor.random_crop_image, {}))
images = self.createTestImages()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
multiclass_scores = self.createTestMultiClassScores()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.multiclass_scores: multiclass_scores
}
distorted_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
distorted_images = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
distorted_multiclass_scores = distorted_tensor_dict[
fields.InputDataFields.multiclass_scores]
boxes_rank = tf.rank(boxes)
distorted_boxes_rank = tf.rank(distorted_boxes)
images_rank = tf.rank(images)
distorted_images_rank = tf.rank(distorted_images)
multiclass_scores_rank = tf.rank(multiclass_scores)
distorted_multiclass_scores_rank = tf.rank(distorted_multiclass_scores)
with self.test_session() as sess:
(boxes_rank_, distorted_boxes_, distorted_boxes_rank_, images_rank_,
distorted_images_rank_, multiclass_scores_rank_,
distorted_multiclass_scores_rank_,
distorted_multiclass_scores_) = sess.run([
boxes_rank, distorted_boxes, distorted_boxes_rank, images_rank,
distorted_images_rank, multiclass_scores_rank,
distorted_multiclass_scores_rank, distorted_multiclass_scores
])
self.assertAllEqual(boxes_rank_, distorted_boxes_rank_)
self.assertAllEqual(images_rank_, distorted_images_rank_)
self.assertAllEqual(multiclass_scores_rank_,
distorted_multiclass_scores_rank_)
self.assertAllEqual(distorted_boxes_.shape[0],
distorted_multiclass_scores_.shape[0])
def testStrictRandomCropImageWithLabelScores(self):
image = self.createColorfulTestImage()[0]
boxes = self.createTestBoxes()
labels = self.createTestLabels()
label_scores = self.createTestLabelScores()
with mock.patch.object(
tf.image,
'sample_distorted_bounding_box'
) as mock_sample_distorted_bounding_box:
mock_sample_distorted_bounding_box.return_value = (
tf.constant([6, 143, 0], dtype=tf.int32),
tf.constant([190, 237, -1], dtype=tf.int32),
tf.constant([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32))
new_image, new_boxes, new_labels, new_label_scores = (
preprocessor._strict_random_crop_image(
image, boxes, labels, label_scores))
with self.test_session() as sess:
new_image, new_boxes, new_labels, new_label_scores = (
sess.run(
[new_image, new_boxes, new_labels, new_label_scores])
)
expected_boxes = np.array(
[[0.0, 0.0, 0.75789469, 1.0],
[0.23157893, 0.24050637, 0.75789469, 1.0]], dtype=np.float32)
self.assertAllEqual(new_image.shape, [190, 237, 3])
self.assertAllEqual(new_label_scores, [1.0, 0.5])
self.assertAllClose(
new_boxes.flatten(), expected_boxes.flatten())
def testStrictRandomCropImageWithMasks(self):
image = self.createColorfulTestImage()[0]
boxes = self.createTestBoxes()
labels = self.createTestLabels()
masks = tf.random_uniform([2, 200, 400], dtype=tf.float32)
with mock.patch.object(
tf.image,
'sample_distorted_bounding_box'
) as mock_sample_distorted_bounding_box:
mock_sample_distorted_bounding_box.return_value = (
tf.constant([6, 143, 0], dtype=tf.int32),
tf.constant([190, 237, -1], dtype=tf.int32),
tf.constant([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32))
new_image, new_boxes, new_labels, new_masks = (
preprocessor._strict_random_crop_image(
image, boxes, labels, masks=masks))
with self.test_session() as sess:
new_image, new_boxes, new_labels, new_masks = sess.run(
[new_image, new_boxes, new_labels, new_masks])
expected_boxes = np.array(
[[0.0, 0.0, 0.75789469, 1.0],
[0.23157893, 0.24050637, 0.75789469, 1.0]], dtype=np.float32)
self.assertAllEqual(new_image.shape, [190, 237, 3])
self.assertAllEqual(new_masks.shape, [2, 190, 237])
self.assertAllClose(
new_boxes.flatten(), expected_boxes.flatten())
def testStrictRandomCropImageWithKeypoints(self):
image = self.createColorfulTestImage()[0]
boxes = self.createTestBoxes()
labels = self.createTestLabels()
keypoints = self.createTestKeypoints()
with mock.patch.object(
tf.image,
'sample_distorted_bounding_box'
) as mock_sample_distorted_bounding_box:
mock_sample_distorted_bounding_box.return_value = (
tf.constant([6, 143, 0], dtype=tf.int32),
tf.constant([190, 237, -1], dtype=tf.int32),
tf.constant([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32))
new_image, new_boxes, new_labels, new_keypoints = (
preprocessor._strict_random_crop_image(
image, boxes, labels, keypoints=keypoints))
with self.test_session() as sess:
new_image, new_boxes, new_labels, new_keypoints = sess.run(
[new_image, new_boxes, new_labels, new_keypoints])
expected_boxes = np.array([
[0.0, 0.0, 0.75789469, 1.0],
[0.23157893, 0.24050637, 0.75789469, 1.0],], dtype=np.float32)
expected_keypoints = np.array([
[[np.nan, np.nan],
[np.nan, np.nan],
[np.nan, np.nan]],
[[0.38947368, 0.07173],
[0.49473682, 0.24050637],
[0.60000002, 0.40928277]]
], dtype=np.float32)
self.assertAllEqual(new_image.shape, [190, 237, 3])
self.assertAllClose(
new_boxes.flatten(), expected_boxes.flatten())
self.assertAllClose(
new_keypoints.flatten(), expected_keypoints.flatten())
def testRunRandomCropImageWithMasks(self):
image = self.createColorfulTestImage()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
masks = tf.random_uniform([2, 200, 400], dtype=tf.float32)
tensor_dict = {
fields.InputDataFields.image: image,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.groundtruth_instance_masks: masks,
}
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_instance_masks=True)
preprocessing_options = [(preprocessor.random_crop_image, {})]
with mock.patch.object(
tf.image,
'sample_distorted_bounding_box'
) as mock_sample_distorted_bounding_box:
mock_sample_distorted_bounding_box.return_value = (
tf.constant([6, 143, 0], dtype=tf.int32),
tf.constant([190, 237, -1], dtype=tf.int32),
tf.constant([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32))
distorted_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
distorted_image = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
distorted_labels = distorted_tensor_dict[
fields.InputDataFields.groundtruth_classes]
distorted_masks = distorted_tensor_dict[
fields.InputDataFields.groundtruth_instance_masks]
with self.test_session() as sess:
(distorted_image_, distorted_boxes_, distorted_labels_,
distorted_masks_) = sess.run(
[distorted_image, distorted_boxes, distorted_labels,
distorted_masks])
expected_boxes = np.array([
[0.0, 0.0, 0.75789469, 1.0],
[0.23157893, 0.24050637, 0.75789469, 1.0],
], dtype=np.float32)
self.assertAllEqual(distorted_image_.shape, [1, 190, 237, 3])
self.assertAllEqual(distorted_masks_.shape, [2, 190, 237])
self.assertAllEqual(distorted_labels_, [1, 2])
self.assertAllClose(
distorted_boxes_.flatten(), expected_boxes.flatten())
def testRunRandomCropImageWithKeypointsInsideCrop(self):
image = self.createColorfulTestImage()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
keypoints = self.createTestKeypointsInsideCrop()
tensor_dict = {
fields.InputDataFields.image: image,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.groundtruth_keypoints: keypoints
}
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_keypoints=True)
preprocessing_options = [(preprocessor.random_crop_image, {})]
with mock.patch.object(
tf.image,
'sample_distorted_bounding_box'
) as mock_sample_distorted_bounding_box:
mock_sample_distorted_bounding_box.return_value = (
tf.constant([6, 143, 0], dtype=tf.int32),
tf.constant([190, 237, -1], dtype=tf.int32),
tf.constant([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32))
distorted_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
distorted_image = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
distorted_labels = distorted_tensor_dict[
fields.InputDataFields.groundtruth_classes]
distorted_keypoints = distorted_tensor_dict[
fields.InputDataFields.groundtruth_keypoints]
with self.test_session() as sess:
(distorted_image_, distorted_boxes_, distorted_labels_,
distorted_keypoints_) = sess.run(
[distorted_image, distorted_boxes, distorted_labels,
distorted_keypoints])
expected_boxes = np.array([
[0.0, 0.0, 0.75789469, 1.0],
[0.23157893, 0.24050637, 0.75789469, 1.0],
], dtype=np.float32)
expected_keypoints = np.array([
[[0.38947368, 0.07173],
[0.49473682, 0.24050637],
[0.60000002, 0.40928277]],
[[0.38947368, 0.07173],
[0.49473682, 0.24050637],
[0.60000002, 0.40928277]]
])
self.assertAllEqual(distorted_image_.shape, [1, 190, 237, 3])
self.assertAllEqual(distorted_labels_, [1, 2])
self.assertAllClose(
distorted_boxes_.flatten(), expected_boxes.flatten())
self.assertAllClose(
distorted_keypoints_.flatten(), expected_keypoints.flatten())
def testRunRandomCropImageWithKeypointsOutsideCrop(self):
image = self.createColorfulTestImage()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
keypoints = self.createTestKeypointsOutsideCrop()
tensor_dict = {
fields.InputDataFields.image: image,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.groundtruth_keypoints: keypoints
}
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_keypoints=True)
preprocessing_options = [(preprocessor.random_crop_image, {})]
with mock.patch.object(
tf.image,
'sample_distorted_bounding_box'
) as mock_sample_distorted_bounding_box:
mock_sample_distorted_bounding_box.return_value = (
tf.constant([6, 143, 0], dtype=tf.int32),
tf.constant([190, 237, -1], dtype=tf.int32),
tf.constant([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32))
distorted_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
distorted_image = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
distorted_labels = distorted_tensor_dict[
fields.InputDataFields.groundtruth_classes]
distorted_keypoints = distorted_tensor_dict[
fields.InputDataFields.groundtruth_keypoints]
with self.test_session() as sess:
(distorted_image_, distorted_boxes_, distorted_labels_,
distorted_keypoints_) = sess.run(
[distorted_image, distorted_boxes, distorted_labels,
distorted_keypoints])
expected_boxes = np.array([
[0.0, 0.0, 0.75789469, 1.0],
[0.23157893, 0.24050637, 0.75789469, 1.0],
], dtype=np.float32)
expected_keypoints = np.array([
[[np.nan, np.nan],
[np.nan, np.nan],
[np.nan, np.nan]],
[[np.nan, np.nan],
[np.nan, np.nan],
[np.nan, np.nan]],
])
self.assertAllEqual(distorted_image_.shape, [1, 190, 237, 3])
self.assertAllEqual(distorted_labels_, [1, 2])
self.assertAllClose(
distorted_boxes_.flatten(), expected_boxes.flatten())
self.assertAllClose(
distorted_keypoints_.flatten(), expected_keypoints.flatten())
def testRunRetainBoxesAboveThreshold(self):
boxes = self.createTestBoxes()
labels = self.createTestLabels()
label_scores = self.createTestLabelScores()
tensor_dict = {
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.groundtruth_label_scores: label_scores
}
preprocessing_options = [
(preprocessor.retain_boxes_above_threshold, {'threshold': 0.6})
]
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_label_scores=True)
retained_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
retained_boxes = retained_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
retained_labels = retained_tensor_dict[
fields.InputDataFields.groundtruth_classes]
retained_label_scores = retained_tensor_dict[
fields.InputDataFields.groundtruth_label_scores]
with self.test_session() as sess:
(retained_boxes_, retained_labels_,
retained_label_scores_, expected_retained_boxes_,
expected_retained_labels_, expected_retained_label_scores_) = sess.run(
[retained_boxes, retained_labels, retained_label_scores,
self.expectedBoxesAfterThresholding(),
self.expectedLabelsAfterThresholding(),
self.expectedLabelScoresAfterThresholding()])
self.assertAllClose(retained_boxes_, expected_retained_boxes_)
self.assertAllClose(retained_labels_, expected_retained_labels_)
self.assertAllClose(
retained_label_scores_, expected_retained_label_scores_)
def testRunRetainBoxesAboveThresholdWithMasks(self):
boxes = self.createTestBoxes()
labels = self.createTestLabels()
label_scores = self.createTestLabelScores()
masks = self.createTestMasks()
tensor_dict = {
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.groundtruth_label_scores: label_scores,
fields.InputDataFields.groundtruth_instance_masks: masks
}
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_label_scores=True,
include_instance_masks=True)
preprocessing_options = [
(preprocessor.retain_boxes_above_threshold, {'threshold': 0.6})
]
retained_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
retained_masks = retained_tensor_dict[
fields.InputDataFields.groundtruth_instance_masks]
with self.test_session() as sess:
(retained_masks_, expected_masks_) = sess.run(
[retained_masks,
self.expectedMasksAfterThresholding()])
self.assertAllClose(retained_masks_, expected_masks_)
def testRunRetainBoxesAboveThresholdWithKeypoints(self):
boxes = self.createTestBoxes()
labels = self.createTestLabels()
label_scores = self.createTestLabelScores()
keypoints = self.createTestKeypoints()
tensor_dict = {
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.groundtruth_label_scores: label_scores,
fields.InputDataFields.groundtruth_keypoints: keypoints
}
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_label_scores=True,
include_keypoints=True)
preprocessing_options = [
(preprocessor.retain_boxes_above_threshold, {'threshold': 0.6})
]
retained_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
retained_keypoints = retained_tensor_dict[
fields.InputDataFields.groundtruth_keypoints]
with self.test_session() as sess:
(retained_keypoints_, expected_keypoints_) = sess.run(
[retained_keypoints,
self.expectedKeypointsAfterThresholding()])
self.assertAllClose(retained_keypoints_, expected_keypoints_)
def testRandomCropToAspectRatioWithCache(self):
preprocess_options = [(preprocessor.random_crop_to_aspect_ratio, {})]
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=False,
test_keypoints=False)
def testRunRandomCropToAspectRatioWithMasks(self):
image = self.createColorfulTestImage()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
masks = tf.random_uniform([2, 200, 400], dtype=tf.float32)
tensor_dict = {
fields.InputDataFields.image: image,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.groundtruth_instance_masks: masks
}
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_instance_masks=True)
preprocessing_options = [(preprocessor.random_crop_to_aspect_ratio, {})]
with mock.patch.object(preprocessor,
'_random_integer') as mock_random_integer:
mock_random_integer.return_value = tf.constant(0, dtype=tf.int32)
distorted_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
distorted_image = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
distorted_labels = distorted_tensor_dict[
fields.InputDataFields.groundtruth_classes]
distorted_masks = distorted_tensor_dict[
fields.InputDataFields.groundtruth_instance_masks]
with self.test_session() as sess:
(distorted_image_, distorted_boxes_, distorted_labels_,
distorted_masks_) = sess.run([
distorted_image, distorted_boxes, distorted_labels, distorted_masks
])
expected_boxes = np.array([0.0, 0.5, 0.75, 1.0], dtype=np.float32)
self.assertAllEqual(distorted_image_.shape, [1, 200, 200, 3])
self.assertAllEqual(distorted_labels_, [1])
self.assertAllClose(distorted_boxes_.flatten(),
expected_boxes.flatten())
self.assertAllEqual(distorted_masks_.shape, [1, 200, 200])
def testRunRandomCropToAspectRatioWithKeypoints(self):
image = self.createColorfulTestImage()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
keypoints = self.createTestKeypoints()
tensor_dict = {
fields.InputDataFields.image: image,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.groundtruth_keypoints: keypoints
}
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_keypoints=True)
preprocessing_options = [(preprocessor.random_crop_to_aspect_ratio, {})]
with mock.patch.object(preprocessor,
'_random_integer') as mock_random_integer:
mock_random_integer.return_value = tf.constant(0, dtype=tf.int32)
distorted_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
distorted_image = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
distorted_labels = distorted_tensor_dict[
fields.InputDataFields.groundtruth_classes]
distorted_keypoints = distorted_tensor_dict[
fields.InputDataFields.groundtruth_keypoints]
with self.test_session() as sess:
(distorted_image_, distorted_boxes_, distorted_labels_,
distorted_keypoints_) = sess.run([
distorted_image, distorted_boxes, distorted_labels,
distorted_keypoints
])
expected_boxes = np.array([0.0, 0.5, 0.75, 1.0], dtype=np.float32)
expected_keypoints = np.array(
[[0.1, 0.2], [0.2, 0.4], [0.3, 0.6]], dtype=np.float32)
self.assertAllEqual(distorted_image_.shape, [1, 200, 200, 3])
self.assertAllEqual(distorted_labels_, [1])
self.assertAllClose(distorted_boxes_.flatten(),
expected_boxes.flatten())
self.assertAllClose(distorted_keypoints_.flatten(),
expected_keypoints.flatten())
def testRandomPadToAspectRatioWithCache(self):
preprocess_options = [(preprocessor.random_pad_to_aspect_ratio, {})]
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=True,
test_keypoints=True)
def testRunRandomPadToAspectRatioWithMinMaxPaddedSizeRatios(self):
image = self.createColorfulTestImage()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
tensor_dict = {
fields.InputDataFields.image: image,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels
}
preprocessor_arg_map = preprocessor.get_default_func_arg_map()
preprocessing_options = [(preprocessor.random_pad_to_aspect_ratio,
{'min_padded_size_ratio': (4.0, 4.0),
'max_padded_size_ratio': (4.0, 4.0)})]
distorted_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
distorted_image = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
distorted_labels = distorted_tensor_dict[
fields.InputDataFields.groundtruth_classes]
with self.test_session() as sess:
distorted_image_, distorted_boxes_, distorted_labels_ = sess.run([
distorted_image, distorted_boxes, distorted_labels])
expected_boxes = np.array(
[[0.0, 0.125, 0.1875, 0.5], [0.0625, 0.25, 0.1875, 0.5]],
dtype=np.float32)
self.assertAllEqual(distorted_image_.shape, [1, 800, 800, 3])
self.assertAllEqual(distorted_labels_, [1, 2])
self.assertAllClose(distorted_boxes_.flatten(),
expected_boxes.flatten())
def testRunRandomPadToAspectRatioWithMasks(self):
image = self.createColorfulTestImage()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
masks = tf.random_uniform([2, 200, 400], dtype=tf.float32)
tensor_dict = {
fields.InputDataFields.image: image,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.groundtruth_instance_masks: masks
}
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_instance_masks=True)
preprocessing_options = [(preprocessor.random_pad_to_aspect_ratio, {})]
distorted_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
distorted_image = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
distorted_labels = distorted_tensor_dict[
fields.InputDataFields.groundtruth_classes]
distorted_masks = distorted_tensor_dict[
fields.InputDataFields.groundtruth_instance_masks]
with self.test_session() as sess:
(distorted_image_, distorted_boxes_, distorted_labels_,
distorted_masks_) = sess.run([
distorted_image, distorted_boxes, distorted_labels, distorted_masks
])
expected_boxes = np.array(
[[0.0, 0.25, 0.375, 1.0], [0.125, 0.5, 0.375, 1.0]], dtype=np.float32)
self.assertAllEqual(distorted_image_.shape, [1, 400, 400, 3])
self.assertAllEqual(distorted_labels_, [1, 2])
self.assertAllClose(distorted_boxes_.flatten(),
expected_boxes.flatten())
self.assertAllEqual(distorted_masks_.shape, [2, 400, 400])
def testRunRandomPadToAspectRatioWithKeypoints(self):
image = self.createColorfulTestImage()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
keypoints = self.createTestKeypoints()
tensor_dict = {
fields.InputDataFields.image: image,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.groundtruth_keypoints: keypoints
}
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_keypoints=True)
preprocessing_options = [(preprocessor.random_pad_to_aspect_ratio, {})]
distorted_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
distorted_image = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
distorted_labels = distorted_tensor_dict[
fields.InputDataFields.groundtruth_classes]
distorted_keypoints = distorted_tensor_dict[
fields.InputDataFields.groundtruth_keypoints]
with self.test_session() as sess:
(distorted_image_, distorted_boxes_, distorted_labels_,
distorted_keypoints_) = sess.run([
distorted_image, distorted_boxes, distorted_labels,
distorted_keypoints
])
expected_boxes = np.array(
[[0.0, 0.25, 0.375, 1.0], [0.125, 0.5, 0.375, 1.0]], dtype=np.float32)
expected_keypoints = np.array([
[[0.05, 0.1], [0.1, 0.2], [0.15, 0.3]],
[[0.2, 0.4], [0.25, 0.5], [0.3, 0.6]],
], dtype=np.float32)
self.assertAllEqual(distorted_image_.shape, [1, 400, 400, 3])
self.assertAllEqual(distorted_labels_, [1, 2])
self.assertAllClose(distorted_boxes_.flatten(),
expected_boxes.flatten())
self.assertAllClose(distorted_keypoints_.flatten(),
expected_keypoints.flatten())
def testRandomPadImageWithCache(self):
preprocess_options = [(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1,}), (preprocessor.random_pad_image, {})]
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=True,
test_keypoints=True)
def testRandomPadImage(self):
preprocessing_options = [(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
})]
images = self.createTestImages()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocessing_options)
images = tensor_dict[fields.InputDataFields.image]
preprocessing_options = [(preprocessor.random_pad_image, {})]
padded_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
padded_images = padded_tensor_dict[fields.InputDataFields.image]
padded_boxes = padded_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
boxes_shape = tf.shape(boxes)
padded_boxes_shape = tf.shape(padded_boxes)
images_shape = tf.shape(images)
padded_images_shape = tf.shape(padded_images)
with self.test_session() as sess:
(boxes_shape_, padded_boxes_shape_, images_shape_,
padded_images_shape_, boxes_, padded_boxes_) = sess.run(
[boxes_shape, padded_boxes_shape, images_shape,
padded_images_shape, boxes, padded_boxes])
self.assertAllEqual(boxes_shape_, padded_boxes_shape_)
self.assertTrue((images_shape_[1] >= padded_images_shape_[1] * 0.5).all)
self.assertTrue((images_shape_[2] >= padded_images_shape_[2] * 0.5).all)
self.assertTrue((images_shape_[1] <= padded_images_shape_[1]).all)
self.assertTrue((images_shape_[2] <= padded_images_shape_[2]).all)
self.assertTrue(np.all((boxes_[:, 2] - boxes_[:, 0]) >= (
padded_boxes_[:, 2] - padded_boxes_[:, 0])))
self.assertTrue(np.all((boxes_[:, 3] - boxes_[:, 1]) >= (
padded_boxes_[:, 3] - padded_boxes_[:, 1])))
def testRandomCropPadImageWithCache(self):
preprocess_options = [(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1,}), (preprocessor.random_crop_pad_image, {})]
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=True,
test_keypoints=True)
def testRandomCropPadImageWithRandomCoefOne(self):
preprocessing_options = [(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
})]
images = self.createTestImages()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
}
tensor_dict = preprocessor.preprocess(tensor_dict, preprocessing_options)
images = tensor_dict[fields.InputDataFields.image]
preprocessing_options = [(preprocessor.random_crop_pad_image, {
'random_coef': 1.0
})]
padded_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
padded_images = padded_tensor_dict[fields.InputDataFields.image]
padded_boxes = padded_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
boxes_shape = tf.shape(boxes)
padded_boxes_shape = tf.shape(padded_boxes)
images_shape = tf.shape(images)
padded_images_shape = tf.shape(padded_images)
with self.test_session() as sess:
(boxes_shape_, padded_boxes_shape_, images_shape_,
padded_images_shape_, boxes_, padded_boxes_) = sess.run(
[boxes_shape, padded_boxes_shape, images_shape,
padded_images_shape, boxes, padded_boxes])
self.assertAllEqual(boxes_shape_, padded_boxes_shape_)
self.assertTrue((images_shape_[1] >= padded_images_shape_[1] * 0.5).all)
self.assertTrue((images_shape_[2] >= padded_images_shape_[2] * 0.5).all)
self.assertTrue((images_shape_[1] <= padded_images_shape_[1]).all)
self.assertTrue((images_shape_[2] <= padded_images_shape_[2]).all)
self.assertTrue(np.all((boxes_[:, 2] - boxes_[:, 0]) >= (
padded_boxes_[:, 2] - padded_boxes_[:, 0])))
self.assertTrue(np.all((boxes_[:, 3] - boxes_[:, 1]) >= (
padded_boxes_[:, 3] - padded_boxes_[:, 1])))
def testRandomCropToAspectRatio(self):
images = self.createTestImages()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
}
tensor_dict = preprocessor.preprocess(tensor_dict, [])
images = tensor_dict[fields.InputDataFields.image]
preprocessing_options = [(preprocessor.random_crop_to_aspect_ratio, {
'aspect_ratio': 2.0
})]
cropped_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
cropped_images = cropped_tensor_dict[fields.InputDataFields.image]
cropped_boxes = cropped_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
boxes_shape = tf.shape(boxes)
cropped_boxes_shape = tf.shape(cropped_boxes)
images_shape = tf.shape(images)
cropped_images_shape = tf.shape(cropped_images)
with self.test_session() as sess:
(boxes_shape_, cropped_boxes_shape_, images_shape_,
cropped_images_shape_) = sess.run([
boxes_shape, cropped_boxes_shape, images_shape, cropped_images_shape
])
self.assertAllEqual(boxes_shape_, cropped_boxes_shape_)
self.assertEqual(images_shape_[1], cropped_images_shape_[1] * 2)
self.assertEqual(images_shape_[2], cropped_images_shape_[2])
def testRandomPadToAspectRatio(self):
images = self.createTestImages()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
}
tensor_dict = preprocessor.preprocess(tensor_dict, [])
images = tensor_dict[fields.InputDataFields.image]
preprocessing_options = [(preprocessor.random_pad_to_aspect_ratio, {
'aspect_ratio': 2.0
})]
padded_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
padded_images = padded_tensor_dict[fields.InputDataFields.image]
padded_boxes = padded_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
boxes_shape = tf.shape(boxes)
padded_boxes_shape = tf.shape(padded_boxes)
images_shape = tf.shape(images)
padded_images_shape = tf.shape(padded_images)
with self.test_session() as sess:
(boxes_shape_, padded_boxes_shape_, images_shape_,
padded_images_shape_) = sess.run([
boxes_shape, padded_boxes_shape, images_shape, padded_images_shape
])
self.assertAllEqual(boxes_shape_, padded_boxes_shape_)
self.assertEqual(images_shape_[1], padded_images_shape_[1])
self.assertEqual(2 * images_shape_[2], padded_images_shape_[2])
def testRandomBlackPatchesWithCache(self):
preprocess_options = []
preprocess_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocess_options.append((preprocessor.random_black_patches, {
'size_to_image_ratio': 0.5
}))
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=True,
test_keypoints=True)
def testRandomBlackPatches(self):
preprocessing_options = []
preprocessing_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocessing_options.append((preprocessor.random_black_patches, {
'size_to_image_ratio': 0.5
}))
images = self.createTestImages()
tensor_dict = {fields.InputDataFields.image: images}
blacked_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
blacked_images = blacked_tensor_dict[fields.InputDataFields.image]
images_shape = tf.shape(images)
blacked_images_shape = tf.shape(blacked_images)
with self.test_session() as sess:
(images_shape_, blacked_images_shape_) = sess.run(
[images_shape, blacked_images_shape])
self.assertAllEqual(images_shape_, blacked_images_shape_)
def testRandomResizeMethodWithCache(self):
preprocess_options = []
preprocess_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocess_options.append((preprocessor.random_resize_method, {
'target_size': (75, 150)
}))
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=True,
test_keypoints=True)
def testRandomResizeMethod(self):
preprocessing_options = []
preprocessing_options.append((preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}))
preprocessing_options.append((preprocessor.random_resize_method, {
'target_size': (75, 150)
}))
images = self.createTestImages()
tensor_dict = {fields.InputDataFields.image: images}
resized_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
resized_images = resized_tensor_dict[fields.InputDataFields.image]
resized_images_shape = tf.shape(resized_images)
expected_images_shape = tf.constant([1, 75, 150, 3], dtype=tf.int32)
with self.test_session() as sess:
(expected_images_shape_, resized_images_shape_) = sess.run(
[expected_images_shape, resized_images_shape])
self.assertAllEqual(expected_images_shape_,
resized_images_shape_)
def testResizeImageWithMasks(self):
"""Tests image resizing, checking output sizes."""
in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
in_masks_shape_list = [[15, 60, 40], [10, 15, 30]]
height = 50
width = 100
expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
expected_masks_shape_list = [[15, 50, 100], [10, 50, 100]]
for (in_image_shape, expected_image_shape, in_masks_shape,
expected_mask_shape) in zip(in_image_shape_list,
expected_image_shape_list,
in_masks_shape_list,
expected_masks_shape_list):
in_image = tf.random_uniform(in_image_shape)
in_masks = tf.random_uniform(in_masks_shape)
out_image, out_masks, _ = preprocessor.resize_image(
in_image, in_masks, new_height=height, new_width=width)
out_image_shape = tf.shape(out_image)
out_masks_shape = tf.shape(out_masks)
with self.test_session() as sess:
out_image_shape, out_masks_shape = sess.run(
[out_image_shape, out_masks_shape])
self.assertAllEqual(out_image_shape, expected_image_shape)
self.assertAllEqual(out_masks_shape, expected_mask_shape)
def testResizeImageWithMasksTensorInputHeightAndWidth(self):
"""Tests image resizing, checking output sizes."""
in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
in_masks_shape_list = [[15, 60, 40], [10, 15, 30]]
height = tf.constant(50, dtype=tf.int32)
width = tf.constant(100, dtype=tf.int32)
expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
expected_masks_shape_list = [[15, 50, 100], [10, 50, 100]]
for (in_image_shape, expected_image_shape, in_masks_shape,
expected_mask_shape) in zip(in_image_shape_list,
expected_image_shape_list,
in_masks_shape_list,
expected_masks_shape_list):
in_image = tf.random_uniform(in_image_shape)
in_masks = tf.random_uniform(in_masks_shape)
out_image, out_masks, _ = preprocessor.resize_image(
in_image, in_masks, new_height=height, new_width=width)
out_image_shape = tf.shape(out_image)
out_masks_shape = tf.shape(out_masks)
with self.test_session() as sess:
out_image_shape, out_masks_shape = sess.run(
[out_image_shape, out_masks_shape])
self.assertAllEqual(out_image_shape, expected_image_shape)
self.assertAllEqual(out_masks_shape, expected_mask_shape)
def testResizeImageWithNoInstanceMask(self):
"""Tests image resizing, checking output sizes."""
in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
in_masks_shape_list = [[0, 60, 40], [0, 15, 30]]
height = 50
width = 100
expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
expected_masks_shape_list = [[0, 50, 100], [0, 50, 100]]
for (in_image_shape, expected_image_shape, in_masks_shape,
expected_mask_shape) in zip(in_image_shape_list,
expected_image_shape_list,
in_masks_shape_list,
expected_masks_shape_list):
in_image = tf.random_uniform(in_image_shape)
in_masks = tf.random_uniform(in_masks_shape)
out_image, out_masks, _ = preprocessor.resize_image(
in_image, in_masks, new_height=height, new_width=width)
out_image_shape = tf.shape(out_image)
out_masks_shape = tf.shape(out_masks)
with self.test_session() as sess:
out_image_shape, out_masks_shape = sess.run(
[out_image_shape, out_masks_shape])
self.assertAllEqual(out_image_shape, expected_image_shape)
self.assertAllEqual(out_masks_shape, expected_mask_shape)
def testResizeToRangePreservesStaticSpatialShape(self):
"""Tests image resizing, checking output sizes."""
in_shape_list = [[60, 40, 3], [15, 30, 3], [15, 50, 3]]
min_dim = 50
max_dim = 100
expected_shape_list = [[75, 50, 3], [50, 100, 3], [30, 100, 3]]
for in_shape, expected_shape in zip(in_shape_list, expected_shape_list):
in_image = tf.random_uniform(in_shape)
out_image, _ = preprocessor.resize_to_range(
in_image, min_dimension=min_dim, max_dimension=max_dim)
self.assertAllEqual(out_image.get_shape().as_list(), expected_shape)
def testResizeToRangeWithDynamicSpatialShape(self):
"""Tests image resizing, checking output sizes."""
in_shape_list = [[60, 40, 3], [15, 30, 3], [15, 50, 3]]
min_dim = 50
max_dim = 100
expected_shape_list = [[75, 50, 3], [50, 100, 3], [30, 100, 3]]
for in_shape, expected_shape in zip(in_shape_list, expected_shape_list):
in_image = tf.placeholder(tf.float32, shape=(None, None, 3))
out_image, _ = preprocessor.resize_to_range(
in_image, min_dimension=min_dim, max_dimension=max_dim)
out_image_shape = tf.shape(out_image)
with self.test_session() as sess:
out_image_shape = sess.run(out_image_shape,
feed_dict={in_image:
np.random.randn(*in_shape)})
self.assertAllEqual(out_image_shape, expected_shape)
def testResizeToRangeWithPadToMaxDimensionReturnsCorrectShapes(self):
in_shape_list = [[60, 40, 3], [15, 30, 3], [15, 50, 3]]
min_dim = 50
max_dim = 100
expected_shape_list = [[100, 100, 3], [100, 100, 3], [100, 100, 3]]
for in_shape, expected_shape in zip(in_shape_list, expected_shape_list):
in_image = tf.placeholder(tf.float32, shape=(None, None, 3))
out_image, _ = preprocessor.resize_to_range(
in_image,
min_dimension=min_dim,
max_dimension=max_dim,
pad_to_max_dimension=True)
self.assertAllEqual(out_image.shape.as_list(), expected_shape)
out_image_shape = tf.shape(out_image)
with self.test_session() as sess:
out_image_shape = sess.run(
out_image_shape, feed_dict={in_image: np.random.randn(*in_shape)})
self.assertAllEqual(out_image_shape, expected_shape)
def testResizeToRangeWithPadToMaxDimensionReturnsCorrectTensor(self):
in_image_np = np.array([[[0, 1, 2]]], np.float32)
ex_image_np = np.array(
[[[0, 1, 2], [123.68, 116.779, 103.939]],
[[123.68, 116.779, 103.939], [123.68, 116.779, 103.939]]], np.float32)
min_dim = 1
max_dim = 2
in_image = tf.placeholder(tf.float32, shape=(None, None, 3))
out_image, _ = preprocessor.resize_to_range(
in_image,
min_dimension=min_dim,
max_dimension=max_dim,
pad_to_max_dimension=True,
per_channel_pad_value=(123.68, 116.779, 103.939))
with self.test_session() as sess:
out_image_np = sess.run(out_image, feed_dict={in_image: in_image_np})
self.assertAllClose(ex_image_np, out_image_np)
def testResizeToRangeWithMasksPreservesStaticSpatialShape(self):
"""Tests image resizing, checking output sizes."""
in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
in_masks_shape_list = [[15, 60, 40], [10, 15, 30]]
min_dim = 50
max_dim = 100
expected_image_shape_list = [[75, 50, 3], [50, 100, 3]]
expected_masks_shape_list = [[15, 75, 50], [10, 50, 100]]
for (in_image_shape, expected_image_shape, in_masks_shape,
expected_mask_shape) in zip(in_image_shape_list,
expected_image_shape_list,
in_masks_shape_list,
expected_masks_shape_list):
in_image = tf.random_uniform(in_image_shape)
in_masks = tf.random_uniform(in_masks_shape)
out_image, out_masks, _ = preprocessor.resize_to_range(
in_image, in_masks, min_dimension=min_dim, max_dimension=max_dim)
self.assertAllEqual(out_masks.get_shape().as_list(), expected_mask_shape)
self.assertAllEqual(out_image.get_shape().as_list(), expected_image_shape)
def testResizeToRangeWithMasksAndPadToMaxDimension(self):
"""Tests image resizing, checking output sizes."""
in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
in_masks_shape_list = [[15, 60, 40], [10, 15, 30]]
min_dim = 50
max_dim = 100
expected_image_shape_list = [[100, 100, 3], [100, 100, 3]]
expected_masks_shape_list = [[15, 100, 100], [10, 100, 100]]
for (in_image_shape,
expected_image_shape, in_masks_shape, expected_mask_shape) in zip(
in_image_shape_list, expected_image_shape_list,
in_masks_shape_list, expected_masks_shape_list):
in_image = tf.placeholder(tf.float32, shape=(None, None, 3))
in_masks = tf.placeholder(tf.float32, shape=(None, None, None))
out_image, out_masks, _ = preprocessor.resize_to_range(
in_image,
in_masks,
min_dimension=min_dim,
max_dimension=max_dim,
pad_to_max_dimension=True)
out_image_shape = tf.shape(out_image)
out_masks_shape = tf.shape(out_masks)
with self.test_session() as sess:
out_image_shape, out_masks_shape = sess.run(
[out_image_shape, out_masks_shape],
feed_dict={
in_image: np.random.randn(*in_image_shape),
in_masks: np.random.randn(*in_masks_shape)
})
self.assertAllEqual(out_image_shape, expected_image_shape)
self.assertAllEqual(out_masks_shape, expected_mask_shape)
def testResizeToRangeWithMasksAndDynamicSpatialShape(self):
"""Tests image resizing, checking output sizes."""
in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
in_masks_shape_list = [[15, 60, 40], [10, 15, 30]]
min_dim = 50
max_dim = 100
expected_image_shape_list = [[75, 50, 3], [50, 100, 3]]
expected_masks_shape_list = [[15, 75, 50], [10, 50, 100]]
for (in_image_shape, expected_image_shape, in_masks_shape,
expected_mask_shape) in zip(in_image_shape_list,
expected_image_shape_list,
in_masks_shape_list,
expected_masks_shape_list):
in_image = tf.placeholder(tf.float32, shape=(None, None, 3))
in_masks = tf.placeholder(tf.float32, shape=(None, None, None))
in_masks = tf.random_uniform(in_masks_shape)
out_image, out_masks, _ = preprocessor.resize_to_range(
in_image, in_masks, min_dimension=min_dim, max_dimension=max_dim)
out_image_shape = tf.shape(out_image)
out_masks_shape = tf.shape(out_masks)
with self.test_session() as sess:
out_image_shape, out_masks_shape = sess.run(
[out_image_shape, out_masks_shape],
feed_dict={
in_image: np.random.randn(*in_image_shape),
in_masks: np.random.randn(*in_masks_shape)
})
self.assertAllEqual(out_image_shape, expected_image_shape)
self.assertAllEqual(out_masks_shape, expected_mask_shape)
def testResizeToRangeWithInstanceMasksTensorOfSizeZero(self):
"""Tests image resizing, checking output sizes."""
in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
in_masks_shape_list = [[0, 60, 40], [0, 15, 30]]
min_dim = 50
max_dim = 100
expected_image_shape_list = [[75, 50, 3], [50, 100, 3]]
expected_masks_shape_list = [[0, 75, 50], [0, 50, 100]]
for (in_image_shape, expected_image_shape, in_masks_shape,
expected_mask_shape) in zip(in_image_shape_list,
expected_image_shape_list,
in_masks_shape_list,
expected_masks_shape_list):
in_image = tf.random_uniform(in_image_shape)
in_masks = tf.random_uniform(in_masks_shape)
out_image, out_masks, _ = preprocessor.resize_to_range(
in_image, in_masks, min_dimension=min_dim, max_dimension=max_dim)
out_image_shape = tf.shape(out_image)
out_masks_shape = tf.shape(out_masks)
with self.test_session() as sess:
out_image_shape, out_masks_shape = sess.run(
[out_image_shape, out_masks_shape])
self.assertAllEqual(out_image_shape, expected_image_shape)
self.assertAllEqual(out_masks_shape, expected_mask_shape)
def testResizeToRange4DImageTensor(self):
image = tf.random_uniform([1, 200, 300, 3])
with self.assertRaises(ValueError):
preprocessor.resize_to_range(image, 500, 600)
def testResizeToRangeSameMinMax(self):
"""Tests image resizing, checking output sizes."""
in_shape_list = [[312, 312, 3], [299, 299, 3]]
min_dim = 320
max_dim = 320
expected_shape_list = [[320, 320, 3], [320, 320, 3]]
for in_shape, expected_shape in zip(in_shape_list, expected_shape_list):
in_image = tf.random_uniform(in_shape)
out_image, _ = preprocessor.resize_to_range(
in_image, min_dimension=min_dim, max_dimension=max_dim)
out_image_shape = tf.shape(out_image)
with self.test_session() as sess:
out_image_shape = sess.run(out_image_shape)
self.assertAllEqual(out_image_shape, expected_shape)
def testResizeToMinDimensionTensorShapes(self):
in_image_shape_list = [[60, 55, 3], [15, 30, 3]]
in_masks_shape_list = [[15, 60, 55], [10, 15, 30]]
min_dim = 50
expected_image_shape_list = [[60, 55, 3], [50, 100, 3]]
expected_masks_shape_list = [[15, 60, 55], [10, 50, 100]]
for (in_image_shape, expected_image_shape, in_masks_shape,
expected_mask_shape) in zip(in_image_shape_list,
expected_image_shape_list,
in_masks_shape_list,
expected_masks_shape_list):
in_image = tf.placeholder(tf.float32, shape=(None, None, 3))
in_masks = tf.placeholder(tf.float32, shape=(None, None, None))
in_masks = tf.random_uniform(in_masks_shape)
out_image, out_masks, _ = preprocessor.resize_to_min_dimension(
in_image, in_masks, min_dimension=min_dim)
out_image_shape = tf.shape(out_image)
out_masks_shape = tf.shape(out_masks)
with self.test_session() as sess:
out_image_shape, out_masks_shape = sess.run(
[out_image_shape, out_masks_shape],
feed_dict={
in_image: np.random.randn(*in_image_shape),
in_masks: np.random.randn(*in_masks_shape)
})
self.assertAllEqual(out_image_shape, expected_image_shape)
self.assertAllEqual(out_masks_shape, expected_mask_shape)
def testResizeToMinDimensionWithInstanceMasksTensorOfSizeZero(self):
"""Tests image resizing, checking output sizes."""
in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
in_masks_shape_list = [[0, 60, 40], [0, 15, 30]]
min_dim = 50
expected_image_shape_list = [[75, 50, 3], [50, 100, 3]]
expected_masks_shape_list = [[0, 75, 50], [0, 50, 100]]
for (in_image_shape, expected_image_shape, in_masks_shape,
expected_mask_shape) in zip(in_image_shape_list,
expected_image_shape_list,
in_masks_shape_list,
expected_masks_shape_list):
in_image = tf.random_uniform(in_image_shape)
in_masks = tf.random_uniform(in_masks_shape)
out_image, out_masks, _ = preprocessor.resize_to_min_dimension(
in_image, in_masks, min_dimension=min_dim)
out_image_shape = tf.shape(out_image)
out_masks_shape = tf.shape(out_masks)
with self.test_session() as sess:
out_image_shape, out_masks_shape = sess.run(
[out_image_shape, out_masks_shape])
self.assertAllEqual(out_image_shape, expected_image_shape)
self.assertAllEqual(out_masks_shape, expected_mask_shape)
def testResizeToMinDimensionRaisesErrorOn4DImage(self):
image = tf.random_uniform([1, 200, 300, 3])
with self.assertRaises(ValueError):
preprocessor.resize_to_min_dimension(image, 500)
def testScaleBoxesToPixelCoordinates(self):
"""Tests box scaling, checking scaled values."""
in_shape = [60, 40, 3]
in_boxes = [[0.1, 0.2, 0.4, 0.6],
[0.5, 0.3, 0.9, 0.7]]
expected_boxes = [[6., 8., 24., 24.],
[30., 12., 54., 28.]]
in_image = tf.random_uniform(in_shape)
in_boxes = tf.constant(in_boxes)
_, out_boxes = preprocessor.scale_boxes_to_pixel_coordinates(
in_image, boxes=in_boxes)
with self.test_session() as sess:
out_boxes = sess.run(out_boxes)
self.assertAllClose(out_boxes, expected_boxes)
def testScaleBoxesToPixelCoordinatesWithKeypoints(self):
"""Tests box and keypoint scaling, checking scaled values."""
in_shape = [60, 40, 3]
in_boxes = self.createTestBoxes()
in_keypoints = self.createTestKeypoints()
expected_boxes = [[0., 10., 45., 40.],
[15., 20., 45., 40.]]
expected_keypoints = [
[[6., 4.], [12., 8.], [18., 12.]],
[[24., 16.], [30., 20.], [36., 24.]],
]
in_image = tf.random_uniform(in_shape)
_, out_boxes, out_keypoints = preprocessor.scale_boxes_to_pixel_coordinates(
in_image, boxes=in_boxes, keypoints=in_keypoints)
with self.test_session() as sess:
out_boxes_, out_keypoints_ = sess.run([out_boxes, out_keypoints])
self.assertAllClose(out_boxes_, expected_boxes)
self.assertAllClose(out_keypoints_, expected_keypoints)
def testSubtractChannelMean(self):
"""Tests whether channel means have been subtracted."""
with self.test_session():
image = tf.zeros((240, 320, 3))
means = [1, 2, 3]
actual = preprocessor.subtract_channel_mean(image, means=means)
actual = actual.eval()
self.assertTrue((actual[:, :, 0] == -1).all())
self.assertTrue((actual[:, :, 1] == -2).all())
self.assertTrue((actual[:, :, 2] == -3).all())
def testOneHotEncoding(self):
"""Tests one hot encoding of multiclass labels."""
with self.test_session():
labels = tf.constant([1, 4, 2], dtype=tf.int32)
one_hot = preprocessor.one_hot_encoding(labels, num_classes=5)
one_hot = one_hot.eval()
self.assertAllEqual([0, 1, 1, 0, 1], one_hot)
def testSSDRandomCropWithCache(self):
preprocess_options = [
(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}),
(preprocessor.ssd_random_crop, {})]
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=False,
test_keypoints=False)
def testSSDRandomCrop(self):
preprocessing_options = [
(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}),
(preprocessor.ssd_random_crop, {})]
images = self.createTestImages()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
}
distorted_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
distorted_images = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
images_rank = tf.rank(images)
distorted_images_rank = tf.rank(distorted_images)
boxes_rank = tf.rank(boxes)
distorted_boxes_rank = tf.rank(distorted_boxes)
with self.test_session() as sess:
(boxes_rank_, distorted_boxes_rank_, images_rank_,
distorted_images_rank_) = sess.run(
[boxes_rank, distorted_boxes_rank, images_rank,
distorted_images_rank])
self.assertAllEqual(boxes_rank_, distorted_boxes_rank_)
self.assertAllEqual(images_rank_, distorted_images_rank_)
def testSSDRandomCropWithMultiClassScores(self):
preprocessing_options = [(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}), (preprocessor.ssd_random_crop, {})]
images = self.createTestImages()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
multiclass_scores = self.createTestMultiClassScores()
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
fields.InputDataFields.multiclass_scores: multiclass_scores,
}
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_multiclass_scores=True)
distorted_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
distorted_images = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
distorted_multiclass_scores = distorted_tensor_dict[
fields.InputDataFields.multiclass_scores]
images_rank = tf.rank(images)
distorted_images_rank = tf.rank(distorted_images)
boxes_rank = tf.rank(boxes)
distorted_boxes_rank = tf.rank(distorted_boxes)
multiclass_scores_rank = tf.rank(multiclass_scores)
distorted_multiclass_scores_rank = tf.rank(distorted_multiclass_scores)
with self.test_session() as sess:
(boxes_rank_, distorted_boxes_, distorted_boxes_rank_, images_rank_,
distorted_images_rank_, multiclass_scores_rank_,
distorted_multiclass_scores_,
distorted_multiclass_scores_rank_) = sess.run([
boxes_rank, distorted_boxes, distorted_boxes_rank, images_rank,
distorted_images_rank, multiclass_scores_rank,
distorted_multiclass_scores, distorted_multiclass_scores_rank
])
self.assertAllEqual(boxes_rank_, distorted_boxes_rank_)
self.assertAllEqual(images_rank_, distorted_images_rank_)
self.assertAllEqual(multiclass_scores_rank_,
distorted_multiclass_scores_rank_)
self.assertAllEqual(distorted_boxes_.shape[0],
distorted_multiclass_scores_.shape[0])
def testSSDRandomCropPad(self):
images = self.createTestImages()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
preprocessing_options = [
(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}),
(preprocessor.ssd_random_crop_pad, {})]
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
}
distorted_tensor_dict = preprocessor.preprocess(tensor_dict,
preprocessing_options)
distorted_images = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
images_rank = tf.rank(images)
distorted_images_rank = tf.rank(distorted_images)
boxes_rank = tf.rank(boxes)
distorted_boxes_rank = tf.rank(distorted_boxes)
with self.test_session() as sess:
(boxes_rank_, distorted_boxes_rank_, images_rank_,
distorted_images_rank_) = sess.run([
boxes_rank, distorted_boxes_rank, images_rank, distorted_images_rank
])
self.assertAllEqual(boxes_rank_, distorted_boxes_rank_)
self.assertAllEqual(images_rank_, distorted_images_rank_)
def testSSDRandomCropFixedAspectRatioWithCache(self):
preprocess_options = [
(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}),
(preprocessor.ssd_random_crop_fixed_aspect_ratio, {})]
self._testPreprocessorCache(preprocess_options,
test_boxes=True,
test_masks=False,
test_keypoints=False)
def _testSSDRandomCropFixedAspectRatio(self,
include_label_scores,
include_multiclass_scores,
include_instance_masks,
include_keypoints):
images = self.createTestImages()
boxes = self.createTestBoxes()
labels = self.createTestLabels()
preprocessing_options = [(preprocessor.normalize_image, {
'original_minval': 0,
'original_maxval': 255,
'target_minval': 0,
'target_maxval': 1
}), (preprocessor.ssd_random_crop_fixed_aspect_ratio, {})]
tensor_dict = {
fields.InputDataFields.image: images,
fields.InputDataFields.groundtruth_boxes: boxes,
fields.InputDataFields.groundtruth_classes: labels,
}
if include_label_scores:
label_scores = self.createTestLabelScores()
tensor_dict[fields.InputDataFields.groundtruth_label_scores] = (
label_scores)
if include_multiclass_scores:
multiclass_scores = self.createTestMultiClassScores()
tensor_dict[fields.InputDataFields.multiclass_scores] = (
multiclass_scores)
if include_instance_masks:
masks = self.createTestMasks()
tensor_dict[fields.InputDataFields.groundtruth_instance_masks] = masks
if include_keypoints:
keypoints = self.createTestKeypoints()
tensor_dict[fields.InputDataFields.groundtruth_keypoints] = keypoints
preprocessor_arg_map = preprocessor.get_default_func_arg_map(
include_label_scores=include_label_scores,
include_multiclass_scores=include_multiclass_scores,
include_instance_masks=include_instance_masks,
include_keypoints=include_keypoints)
distorted_tensor_dict = preprocessor.preprocess(
tensor_dict, preprocessing_options, func_arg_map=preprocessor_arg_map)
distorted_images = distorted_tensor_dict[fields.InputDataFields.image]
distorted_boxes = distorted_tensor_dict[
fields.InputDataFields.groundtruth_boxes]
images_rank = tf.rank(images)
distorted_images_rank = tf.rank(distorted_images)
boxes_rank = tf.rank(boxes)
distorted_boxes_rank = tf.rank(distorted_boxes)
with self.test_session() as sess:
(boxes_rank_, distorted_boxes_rank_, images_rank_,
distorted_images_rank_) = sess.run(
[boxes_rank, distorted_boxes_rank, images_rank,
distorted_images_rank])
self.assertAllEqual(boxes_rank_, distorted_boxes_rank_)
self.assertAllEqual(images_rank_, distorted_images_rank_)
def testSSDRandomCropFixedAspectRatio(self):
self._testSSDRandomCropFixedAspectRatio(include_label_scores=False,
include_multiclass_scores=False,
include_instance_masks=False,
include_keypoints=False)
def testSSDRandomCropFixedAspectRatioWithMultiClassScores(self):
self._testSSDRandomCropFixedAspectRatio(include_label_scores=False,
include_multiclass_scores=True,
include_instance_masks=False,
include_keypoints=False)
def testSSDRandomCropFixedAspectRatioWithMasksAndKeypoints(self):
self._testSSDRandomCropFixedAspectRatio(include_label_scores=False,
include_multiclass_scores=False,
include_instance_masks=True,
include_keypoints=True)
def testSSDRandomCropFixedAspectRatioWithLabelScoresMasksAndKeypoints(self):
self._testSSDRandomCropFixedAspectRatio(include_label_scores=True,
include_multiclass_scores=False,
include_instance_masks=True,
include_keypoints=True)
def testConvertClassLogitsToSoftmax(self):
multiclass_scores = tf.constant(
[[1.0, 0.0], [0.5, 0.5], [1000, 1]], dtype=tf.float32)
temperature = 2.0
converted_multiclass_scores = (
preprocessor.convert_class_logits_to_softmax(
multiclass_scores=multiclass_scores, temperature=temperature))
expected_converted_multiclass_scores = [[[0.62245935, 0.37754068],
[0.5, 0.5], [1, 0]]]
with self.test_session() as sess:
(converted_multiclass_scores_) = sess.run([converted_multiclass_scores])
self.assertAllClose(converted_multiclass_scores_,
expected_converted_multiclass_scores)
if __name__ == '__main__':
tf.test.main()
|
[
"tensorflow.shape",
"tensorflow.split",
"tensorflow.multiply",
"numpy.array",
"object_detection.tensorflow_detect.core.preprocessor.resize_image",
"object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map",
"object_detection.tensorflow_detect.core.preprocessor.scale_boxes_to_pixel_coordinates",
"object_detection.tensorflow_detect.core.preprocessor.subtract_channel_mean",
"object_detection.tensorflow_detect.core.preprocessor._rot90_masks",
"tensorflow.squared_difference",
"tensorflow.placeholder",
"tensorflow.rank",
"tensorflow.concat",
"object_detection.tensorflow_detect.core.preprocessor._strict_random_crop_image",
"tensorflow.zeros_like",
"object_detection.tensorflow_detect.core.preprocessor._flip_masks_up_down",
"object_detection.tensorflow_detect.core.preprocessor._flip_boxes_left_right",
"object_detection.tensorflow_detect.core.preprocessor.convert_class_logits_to_softmax",
"tensorflow.zeros",
"tensorflow.image.rgb_to_grayscale",
"object_detection.tensorflow_detect.core.preprocessor.resize_to_min_dimension",
"object_detection.tensorflow_detect.core.preprocessor.one_hot_encoding",
"object_detection.tensorflow_detect.core.preprocessor._rot90_boxes",
"object_detection.tensorflow_detect.core.preprocessor.resize_to_range",
"tensorflow.expand_dims",
"object_detection.tensorflow_detect.core.preprocessor_cache.PreprocessorCache",
"numpy.random.randn",
"object_detection.tensorflow_detect.core.preprocessor._flip_boxes_up_down",
"object_detection.tensorflow_detect.core.preprocessor.retain_boxes_above_threshold",
"tensorflow.fill",
"tensorflow.to_float",
"object_detection.tensorflow_detect.core.preprocessor._rgb_to_grayscale",
"tensorflow.test.main",
"object_detection.tensorflow_detect.core.preprocessor._flip_masks_left_right",
"tensorflow.random_uniform",
"tensorflow.constant",
"tensorflow.less_equal",
"unittest.mock.patch.object",
"object_detection.tensorflow_detect.core.preprocessor.preprocess",
"numpy.all",
"tensorflow.greater_equal"
] |
[((125987, 126001), 'tensorflow.test.main', 'tf.test.main', ([], {}), '()\n', (125999, 126001), True, 'import tensorflow as tf\n'), ((1360, 1391), 'tensorflow.concat', 'tf.concat', (['[ch255, ch0, ch0]', '(3)'], {}), '([ch255, ch0, ch0], 3)\n', (1369, 1391), True, 'import tensorflow as tf\n'), ((1402, 1435), 'tensorflow.concat', 'tf.concat', (['[ch255, ch255, ch0]', '(3)'], {}), '([ch255, ch255, ch0], 3)\n', (1411, 1435), True, 'import tensorflow as tf\n'), ((1446, 1479), 'tensorflow.concat', 'tf.concat', (['[ch255, ch0, ch255]', '(3)'], {}), '([ch255, ch0, ch255], 3)\n', (1455, 1479), True, 'import tensorflow as tf\n'), ((1490, 1525), 'tensorflow.concat', 'tf.concat', (['[ch128, ch128, ch128]', '(3)'], {}), '([ch128, ch128, ch128], 3)\n', (1499, 1525), True, 'import tensorflow as tf\n'), ((1536, 1560), 'tensorflow.concat', 'tf.concat', (['[imr, img]', '(2)'], {}), '([imr, img], 2)\n', (1545, 1560), True, 'import tensorflow as tf\n'), ((1571, 1595), 'tensorflow.concat', 'tf.concat', (['[imb, imw]', '(2)'], {}), '([imb, imw], 2)\n', (1580, 1595), True, 'import tensorflow as tf\n'), ((1605, 1629), 'tensorflow.concat', 'tf.concat', (['[imu, imd]', '(1)'], {}), '([imu, imd], 1)\n', (1614, 1629), True, 'import tensorflow as tf\n'), ((1690, 1808), 'tensorflow.constant', 'tf.constant', (['[[[128, 128, 128, 128], [0, 0, 128, 128], [0, 128, 128, 128], [192, 192, \n 128, 128]]]'], {'dtype': 'tf.uint8'}), '([[[128, 128, 128, 128], [0, 0, 128, 128], [0, 128, 128, 128], [\n 192, 192, 128, 128]]], dtype=tf.uint8)\n', (1701, 1808), True, 'import tensorflow as tf\n'), ((1875, 1902), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_r', '(3)'], {}), '(images_r, 3)\n', (1889, 1902), True, 'import tensorflow as tf\n'), ((1918, 2031), 'tensorflow.constant', 'tf.constant', (['[[[0, 0, 128, 128], [0, 0, 128, 128], [0, 128, 192, 192], [192, 192, 128, 192]]\n ]'], {'dtype': 'tf.uint8'}), '([[[0, 0, 128, 128], [0, 0, 128, 128], [0, 128, 192, 192], [192,\n 192, 128, 192]]], dtype=tf.uint8)\n', (1929, 2031), True, 'import tensorflow as tf\n'), ((2099, 2126), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_g', '(3)'], {}), '(images_g, 3)\n', (2113, 2126), True, 'import tensorflow as tf\n'), ((2142, 2255), 'tensorflow.constant', 'tf.constant', (['[[[128, 128, 192, 0], [0, 0, 128, 192], [0, 128, 128, 0], [192, 192, 192, 128]]\n ]'], {'dtype': 'tf.uint8'}), '([[[128, 128, 192, 0], [0, 0, 128, 192], [0, 128, 128, 0], [192,\n 192, 192, 128]]], dtype=tf.uint8)\n', (2153, 2255), True, 'import tensorflow as tf\n'), ((2323, 2350), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_b', '(3)'], {}), '(images_b, 3)\n', (2337, 2350), True, 'import tensorflow as tf\n'), ((2364, 2408), 'tensorflow.concat', 'tf.concat', (['[images_r, images_g, images_b]', '(3)'], {}), '([images_r, images_g, images_b], 3)\n', (2373, 2408), True, 'import tensorflow as tf\n'), ((2474, 2509), 'tensorflow.constant', 'tf.constant', (['[[]]'], {'dtype': 'tf.float32'}), '([[]], dtype=tf.float32)\n', (2485, 2509), True, 'import tensorflow as tf\n'), ((2569, 2648), 'tensorflow.constant', 'tf.constant', (['[[0.0, 0.25, 0.75, 1.0], [0.25, 0.5, 0.75, 1.0]]'], {'dtype': 'tf.float32'}), '([[0.0, 0.25, 0.75, 1.0], [0.25, 0.5, 0.75, 1.0]], dtype=tf.float32)\n', (2580, 2648), True, 'import tensorflow as tf\n'), ((2722, 2763), 'tensorflow.constant', 'tf.constant', (['[1.0, 0.5]'], {'dtype': 'tf.float32'}), '([1.0, 0.5], dtype=tf.float32)\n', (2733, 2763), True, 'import tensorflow as tf\n'), ((2827, 2871), 'tensorflow.constant', 'tf.constant', (['[0.5, np.nan]'], {'dtype': 'tf.float32'}), '([0.5, np.nan], dtype=tf.float32)\n', (2838, 2871), True, 'import tensorflow as tf\n'), ((2913, 3052), 'numpy.array', 'np.array', (['[[[255.0, 0.0, 0.0], [255.0, 0.0, 0.0], [255.0, 0.0, 0.0]], [[255.0, 255.0,\n 0.0], [255.0, 255.0, 0.0], [255.0, 255.0, 0.0]]]'], {}), '([[[255.0, 0.0, 0.0], [255.0, 0.0, 0.0], [255.0, 0.0, 0.0]], [[\n 255.0, 255.0, 0.0], [255.0, 255.0, 0.0], [255.0, 255.0, 0.0]]])\n', (2921, 3052), True, 'import numpy as np\n'), ((3112, 3147), 'tensorflow.constant', 'tf.constant', (['mask'], {'dtype': 'tf.float32'}), '(mask, dtype=tf.float32)\n', (3123, 3147), True, 'import tensorflow as tf\n'), ((3198, 3289), 'numpy.array', 'np.array', (['[[[0.1, 0.1], [0.2, 0.2], [0.3, 0.3]], [[0.4, 0.4], [0.5, 0.5], [0.6, 0.6]]]'], {}), '([[[0.1, 0.1], [0.2, 0.2], [0.3, 0.3]], [[0.4, 0.4], [0.5, 0.5], [\n 0.6, 0.6]]])\n', (3206, 3289), True, 'import numpy as np\n'), ((3319, 3359), 'tensorflow.constant', 'tf.constant', (['keypoints'], {'dtype': 'tf.float32'}), '(keypoints, dtype=tf.float32)\n', (3330, 3359), True, 'import tensorflow as tf\n'), ((3420, 3511), 'numpy.array', 'np.array', (['[[[0.4, 0.4], [0.5, 0.5], [0.6, 0.6]], [[0.4, 0.4], [0.5, 0.5], [0.6, 0.6]]]'], {}), '([[[0.4, 0.4], [0.5, 0.5], [0.6, 0.6]], [[0.4, 0.4], [0.5, 0.5], [\n 0.6, 0.6]]])\n', (3428, 3511), True, 'import numpy as np\n'), ((3541, 3581), 'tensorflow.constant', 'tf.constant', (['keypoints'], {'dtype': 'tf.float32'}), '(keypoints, dtype=tf.float32)\n', (3552, 3581), True, 'import tensorflow as tf\n'), ((3643, 3734), 'numpy.array', 'np.array', (['[[[0.1, 0.1], [0.2, 0.2], [0.3, 0.3]], [[0.1, 0.1], [0.2, 0.2], [0.3, 0.3]]]'], {}), '([[[0.1, 0.1], [0.2, 0.2], [0.3, 0.3]], [[0.1, 0.1], [0.2, 0.2], [\n 0.3, 0.3]]])\n', (3651, 3734), True, 'import numpy as np\n'), ((3764, 3804), 'tensorflow.constant', 'tf.constant', (['keypoints'], {'dtype': 'tf.float32'}), '(keypoints, dtype=tf.float32)\n', (3775, 3804), True, 'import tensorflow as tf\n'), ((3860, 3895), 'numpy.array', 'np.array', (['[0, 2, 1]'], {'dtype': 'np.int32'}), '([0, 2, 1], dtype=np.int32)\n', (3868, 3895), True, 'import numpy as np\n'), ((3940, 3975), 'tensorflow.constant', 'tf.constant', (['[1, 2]'], {'dtype': 'tf.int32'}), '([1, 2], dtype=tf.int32)\n', (3951, 3975), True, 'import tensorflow as tf\n'), ((4046, 4124), 'tensorflow.constant', 'tf.constant', (['[[-0.1, 0.25, 0.75, 1], [0.25, 0.5, 0.75, 1.1]]'], {'dtype': 'tf.float32'}), '([[-0.1, 0.25, 0.75, 1], [0.25, 0.5, 0.75, 1.1]], dtype=tf.float32)\n', (4057, 4124), True, 'import tensorflow as tf\n'), ((4203, 4258), 'tensorflow.constant', 'tf.constant', (['[[1.0, 0.0], [0.5, 0.5]]'], {'dtype': 'tf.float32'}), '([[1.0, 0.0], [0.5, 0.5]], dtype=tf.float32)\n', (4214, 4258), True, 'import tensorflow as tf\n'), ((4321, 4422), 'tensorflow.constant', 'tf.constant', (['[[[0, 0, 0, 0], [-1, -1, 0, 0], [-1, 0, 0, 0], [0.5, 0.5, 0, 0]]]'], {'dtype': 'tf.float32'}), '([[[0, 0, 0, 0], [-1, -1, 0, 0], [-1, 0, 0, 0], [0.5, 0.5, 0, 0]\n ]], dtype=tf.float32)\n', (4332, 4422), True, 'import tensorflow as tf\n'), ((4489, 4516), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_r', '(3)'], {}), '(images_r, 3)\n', (4503, 4516), True, 'import tensorflow as tf\n'), ((4532, 4640), 'tensorflow.constant', 'tf.constant', (['[[[-1, -1, 0, 0], [-1, -1, 0, 0], [-1, 0, 0.5, 0.5], [0.5, 0.5, 0, 0.5]]]'], {'dtype': 'tf.float32'}), '([[[-1, -1, 0, 0], [-1, -1, 0, 0], [-1, 0, 0.5, 0.5], [0.5, 0.5,\n 0, 0.5]]], dtype=tf.float32)\n', (4543, 4640), True, 'import tensorflow as tf\n'), ((4708, 4735), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_g', '(3)'], {}), '(images_g, 3)\n', (4722, 4735), True, 'import tensorflow as tf\n'), ((4751, 4859), 'tensorflow.constant', 'tf.constant', (['[[[0, 0, 0.5, -1], [-1, -1, 0, 0.5], [-1, 0, 0, -1], [0.5, 0.5, 0.5, 0]]]'], {'dtype': 'tf.float32'}), '([[[0, 0, 0.5, -1], [-1, -1, 0, 0.5], [-1, 0, 0, -1], [0.5, 0.5,\n 0.5, 0]]], dtype=tf.float32)\n', (4762, 4859), True, 'import tensorflow as tf\n'), ((4927, 4954), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_b', '(3)'], {}), '(images_b, 3)\n', (4941, 4954), True, 'import tensorflow as tf\n'), ((4968, 5012), 'tensorflow.concat', 'tf.concat', (['[images_r, images_g, images_b]', '(3)'], {}), '([images_r, images_g, images_b], 3)\n', (4977, 5012), True, 'import tensorflow as tf\n'), ((5092, 5220), 'tensorflow.constant', 'tf.constant', (['[[[0.1, 0.1, 0.1, 0.1], [-0.9, -0.9, 0.1, 0.1], [-0.9, 0.1, 0.1, 0.1], [0.6,\n 0.6, 0.1, 0.1]]]'], {'dtype': 'tf.float32'}), '([[[0.1, 0.1, 0.1, 0.1], [-0.9, -0.9, 0.1, 0.1], [-0.9, 0.1, 0.1,\n 0.1], [0.6, 0.6, 0.1, 0.1]]], dtype=tf.float32)\n', (5103, 5220), True, 'import tensorflow as tf\n'), ((5288, 5315), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_r', '(3)'], {}), '(images_r, 3)\n', (5302, 5315), True, 'import tensorflow as tf\n'), ((5331, 5462), 'tensorflow.constant', 'tf.constant', (['[[[-0.9, -0.9, 0.1, 0.1], [-0.9, -0.9, 0.1, 0.1], [-0.9, 0.1, 0.6, 0.6], [\n 0.6, 0.6, 0.1, 0.6]]]'], {'dtype': 'tf.float32'}), '([[[-0.9, -0.9, 0.1, 0.1], [-0.9, -0.9, 0.1, 0.1], [-0.9, 0.1, \n 0.6, 0.6], [0.6, 0.6, 0.1, 0.6]]], dtype=tf.float32)\n', (5342, 5462), True, 'import tensorflow as tf\n'), ((5529, 5556), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_g', '(3)'], {}), '(images_g, 3)\n', (5543, 5556), True, 'import tensorflow as tf\n'), ((5572, 5703), 'tensorflow.constant', 'tf.constant', (['[[[0.1, 0.1, 0.6, -0.9], [-0.9, -0.9, 0.1, 0.6], [-0.9, 0.1, 0.1, -0.9], [\n 0.6, 0.6, 0.6, 0.1]]]'], {'dtype': 'tf.float32'}), '([[[0.1, 0.1, 0.6, -0.9], [-0.9, -0.9, 0.1, 0.6], [-0.9, 0.1, \n 0.1, -0.9], [0.6, 0.6, 0.6, 0.1]]], dtype=tf.float32)\n', (5583, 5703), True, 'import tensorflow as tf\n'), ((5770, 5797), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_b', '(3)'], {}), '(images_b, 3)\n', (5784, 5797), True, 'import tensorflow as tf\n'), ((5811, 5855), 'tensorflow.concat', 'tf.concat', (['[images_r, images_g, images_b]', '(3)'], {}), '([images_r, images_g, images_b], 3)\n', (5820, 5855), True, 'import tensorflow as tf\n'), ((5935, 6069), 'tensorflow.constant', 'tf.constant', (['[[[-0.1, -0.1, -0.1, -0.1], [-1, -1, -0.1, -0.1], [-1, -0.1, -0.1, -0.1], [\n 0.4, 0.4, -0.1, -0.1]]]'], {'dtype': 'tf.float32'}), '([[[-0.1, -0.1, -0.1, -0.1], [-1, -1, -0.1, -0.1], [-1, -0.1, -\n 0.1, -0.1], [0.4, 0.4, -0.1, -0.1]]], dtype=tf.float32)\n', (5946, 6069), True, 'import tensorflow as tf\n'), ((6136, 6163), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_r', '(3)'], {}), '(images_r, 3)\n', (6150, 6163), True, 'import tensorflow as tf\n'), ((6179, 6306), 'tensorflow.constant', 'tf.constant', (['[[[-1, -1, -0.1, -0.1], [-1, -1, -0.1, -0.1], [-1, -0.1, 0.4, 0.4], [0.4, \n 0.4, -0.1, 0.4]]]'], {'dtype': 'tf.float32'}), '([[[-1, -1, -0.1, -0.1], [-1, -1, -0.1, -0.1], [-1, -0.1, 0.4, \n 0.4], [0.4, 0.4, -0.1, 0.4]]], dtype=tf.float32)\n', (6190, 6306), True, 'import tensorflow as tf\n'), ((6373, 6400), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_g', '(3)'], {}), '(images_g, 3)\n', (6387, 6400), True, 'import tensorflow as tf\n'), ((6416, 6543), 'tensorflow.constant', 'tf.constant', (['[[[-0.1, -0.1, 0.4, -1], [-1, -1, -0.1, 0.4], [-1, -0.1, -0.1, -1], [0.4, \n 0.4, 0.4, -0.1]]]'], {'dtype': 'tf.float32'}), '([[[-0.1, -0.1, 0.4, -1], [-1, -1, -0.1, 0.4], [-1, -0.1, -0.1, \n -1], [0.4, 0.4, 0.4, -0.1]]], dtype=tf.float32)\n', (6427, 6543), True, 'import tensorflow as tf\n'), ((6610, 6637), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_b', '(3)'], {}), '(images_b, 3)\n', (6624, 6637), True, 'import tensorflow as tf\n'), ((6651, 6695), 'tensorflow.concat', 'tf.concat', (['[images_r, images_g, images_b]', '(3)'], {}), '([images_r, images_g, images_b], 3)\n', (6660, 6695), True, 'import tensorflow as tf\n'), ((6776, 6877), 'tensorflow.constant', 'tf.constant', (['[[[0, 0, 0, 0], [0, 0, -1, -1], [0, 0, 0, -1], [0, 0, 0.5, 0.5]]]'], {'dtype': 'tf.float32'}), '([[[0, 0, 0, 0], [0, 0, -1, -1], [0, 0, 0, -1], [0, 0, 0.5, 0.5]\n ]], dtype=tf.float32)\n', (6787, 6877), True, 'import tensorflow as tf\n'), ((6944, 6971), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_r', '(3)'], {}), '(images_r, 3)\n', (6958, 6971), True, 'import tensorflow as tf\n'), ((6987, 7096), 'tensorflow.constant', 'tf.constant', (['[[[0, 0, -1, -1], [0, 0, -1, -1], [0.5, 0.5, 0, -1], [0.5, 0, 0.5, 0.5]]]'], {'dtype': 'tf.float32'}), '([[[0, 0, -1, -1], [0, 0, -1, -1], [0.5, 0.5, 0, -1], [0.5, 0, \n 0.5, 0.5]]], dtype=tf.float32)\n', (6998, 7096), True, 'import tensorflow as tf\n'), ((7163, 7190), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_g', '(3)'], {}), '(images_g, 3)\n', (7177, 7190), True, 'import tensorflow as tf\n'), ((7206, 7315), 'tensorflow.constant', 'tf.constant', (['[[[-1, 0.5, 0, 0], [0.5, 0, -1, -1], [-1, 0, 0, -1], [0, 0.5, 0.5, 0.5]]]'], {'dtype': 'tf.float32'}), '([[[-1, 0.5, 0, 0], [0.5, 0, -1, -1], [-1, 0, 0, -1], [0, 0.5, \n 0.5, 0.5]]], dtype=tf.float32)\n', (7217, 7315), True, 'import tensorflow as tf\n'), ((7382, 7409), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_b', '(3)'], {}), '(images_b, 3)\n', (7396, 7409), True, 'import tensorflow as tf\n'), ((7423, 7467), 'tensorflow.concat', 'tf.concat', (['[images_r, images_g, images_b]', '(3)'], {}), '([images_r, images_g, images_b], 3)\n', (7432, 7467), True, 'import tensorflow as tf\n'), ((7545, 7646), 'tensorflow.constant', 'tf.constant', (['[[[0.5, 0.5, 0, 0], [-1, 0, 0, 0], [-1, -1, 0, 0], [0, 0, 0, 0]]]'], {'dtype': 'tf.float32'}), '([[[0.5, 0.5, 0, 0], [-1, 0, 0, 0], [-1, -1, 0, 0], [0, 0, 0, 0]\n ]], dtype=tf.float32)\n', (7556, 7646), True, 'import tensorflow as tf\n'), ((7713, 7740), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_r', '(3)'], {}), '(images_r, 3)\n', (7727, 7740), True, 'import tensorflow as tf\n'), ((7756, 7865), 'tensorflow.constant', 'tf.constant', (['[[[0.5, 0.5, 0, 0.5], [-1, 0, 0.5, 0.5], [-1, -1, 0, 0], [-1, -1, 0, 0]]]'], {'dtype': 'tf.float32'}), '([[[0.5, 0.5, 0, 0.5], [-1, 0, 0.5, 0.5], [-1, -1, 0, 0], [-1, -\n 1, 0, 0]]], dtype=tf.float32)\n', (7767, 7865), True, 'import tensorflow as tf\n'), ((7932, 7959), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_g', '(3)'], {}), '(images_g, 3)\n', (7946, 7959), True, 'import tensorflow as tf\n'), ((7975, 8084), 'tensorflow.constant', 'tf.constant', (['[[[0.5, 0.5, 0.5, 0], [-1, 0, 0, -1], [-1, -1, 0, 0.5], [0, 0, 0.5, -1]]]'], {'dtype': 'tf.float32'}), '([[[0.5, 0.5, 0.5, 0], [-1, 0, 0, -1], [-1, -1, 0, 0.5], [0, 0, \n 0.5, -1]]], dtype=tf.float32)\n', (7986, 8084), True, 'import tensorflow as tf\n'), ((8151, 8178), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_b', '(3)'], {}), '(images_b, 3)\n', (8165, 8178), True, 'import tensorflow as tf\n'), ((8192, 8236), 'tensorflow.concat', 'tf.concat', (['[images_r, images_g, images_b]', '(3)'], {}), '([images_r, images_g, images_b], 3)\n', (8201, 8236), True, 'import tensorflow as tf\n'), ((8309, 8410), 'tensorflow.constant', 'tf.constant', (['[[[0, 0, 0, 0], [0, 0, 0, 0], [0, -1, 0, 0.5], [0, -1, -1, 0.5]]]'], {'dtype': 'tf.float32'}), '([[[0, 0, 0, 0], [0, 0, 0, 0], [0, -1, 0, 0.5], [0, -1, -1, 0.5]\n ]], dtype=tf.float32)\n', (8320, 8410), True, 'import tensorflow as tf\n'), ((8477, 8504), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_r', '(3)'], {}), '(images_r, 3)\n', (8491, 8504), True, 'import tensorflow as tf\n'), ((8520, 8629), 'tensorflow.constant', 'tf.constant', (['[[[0, 0, 0.5, 0.5], [0, 0, 0.5, 0], [-1, -1, 0, 0.5], [-1, -1, -1, 0.5]]]'], {'dtype': 'tf.float32'}), '([[[0, 0, 0.5, 0.5], [0, 0, 0.5, 0], [-1, -1, 0, 0.5], [-1, -1, \n -1, 0.5]]], dtype=tf.float32)\n', (8531, 8629), True, 'import tensorflow as tf\n'), ((8696, 8723), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_g', '(3)'], {}), '(images_g, 3)\n', (8710, 8723), True, 'import tensorflow as tf\n'), ((8739, 8848), 'tensorflow.constant', 'tf.constant', (['[[[-1, 0.5, -1, 0], [0.5, 0, 0, 0.5], [0, -1, 0, 0.5], [0, -1, -1, 0.5]]]'], {'dtype': 'tf.float32'}), '([[[-1, 0.5, -1, 0], [0.5, 0, 0, 0.5], [0, -1, 0, 0.5], [0, -1, \n -1, 0.5]]], dtype=tf.float32)\n', (8750, 8848), True, 'import tensorflow as tf\n'), ((8915, 8942), 'tensorflow.expand_dims', 'tf.expand_dims', (['images_b', '(3)'], {}), '(images_b, 3)\n', (8929, 8942), True, 'import tensorflow as tf\n'), ((8956, 9000), 'tensorflow.concat', 'tf.concat', (['[images_r, images_g, images_b]', '(3)'], {}), '([images_r, images_g, images_b], 3)\n', (8965, 9000), True, 'import tensorflow as tf\n'), ((9077, 9156), 'tensorflow.constant', 'tf.constant', (['[[0.0, 0.0, 0.75, 0.75], [0.25, 0.0, 0.75, 0.5]]'], {'dtype': 'tf.float32'}), '([[0.0, 0.0, 0.75, 0.75], [0.25, 0.0, 0.75, 0.5]], dtype=tf.float32)\n', (9088, 9156), True, 'import tensorflow as tf\n'), ((9253, 9332), 'tensorflow.constant', 'tf.constant', (['[[0.25, 0.25, 1.0, 1.0], [0.25, 0.5, 0.75, 1.0]]'], {'dtype': 'tf.float32'}), '([[0.25, 0.25, 1.0, 1.0], [0.25, 0.5, 0.75, 1.0]], dtype=tf.float32)\n', (9264, 9332), True, 'import tensorflow as tf\n'), ((9424, 9503), 'tensorflow.constant', 'tf.constant', (['[[0.0, 0.0, 0.75, 0.75], [0.0, 0.25, 0.5, 0.75]]'], {'dtype': 'tf.float32'}), '([[0.0, 0.0, 0.75, 0.75], [0.0, 0.25, 0.5, 0.75]], dtype=tf.float32)\n', (9435, 9503), True, 'import tensorflow as tf\n'), ((9587, 9725), 'numpy.array', 'np.array', (['[[[0.0, 0.0, 255.0], [0.0, 0.0, 255.0], [0.0, 0.0, 255.0]], [[0.0, 255.0, \n 255.0], [0.0, 255.0, 255.0], [0.0, 255.0, 255.0]]]'], {}), '([[[0.0, 0.0, 255.0], [0.0, 0.0, 255.0], [0.0, 0.0, 255.0]], [[0.0,\n 255.0, 255.0], [0.0, 255.0, 255.0], [0.0, 255.0, 255.0]]])\n', (9595, 9725), True, 'import numpy as np\n'), ((9786, 9821), 'tensorflow.constant', 'tf.constant', (['mask'], {'dtype': 'tf.float32'}), '(mask, dtype=tf.float32)\n', (9797, 9821), True, 'import tensorflow as tf\n'), ((9876, 10015), 'numpy.array', 'np.array', (['[[[255.0, 0.0, 0.0], [255.0, 0.0, 0.0], [255.0, 0.0, 0.0]], [[255.0, 255.0,\n 0.0], [255.0, 255.0, 0.0], [255.0, 255.0, 0.0]]]'], {}), '([[[255.0, 0.0, 0.0], [255.0, 0.0, 0.0], [255.0, 0.0, 0.0]], [[\n 255.0, 255.0, 0.0], [255.0, 255.0, 0.0], [255.0, 255.0, 0.0]]])\n', (9884, 10015), True, 'import numpy as np\n'), ((10075, 10110), 'tensorflow.constant', 'tf.constant', (['mask'], {'dtype': 'tf.float32'}), '(mask, dtype=tf.float32)\n', (10086, 10110), True, 'import tensorflow as tf\n'), ((10160, 10298), 'numpy.array', 'np.array', (['[[[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [255.0, 255.0, 255.0]], [[0.0, 0.0, 0.0\n ], [255.0, 255.0, 255.0], [255.0, 255.0, 255.0]]]'], {}), '([[[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [255.0, 255.0, 255.0]], [[0.0,\n 0.0, 0.0], [255.0, 255.0, 255.0], [255.0, 255.0, 255.0]]])\n', (10168, 10298), True, 'import numpy as np\n'), ((10359, 10394), 'tensorflow.constant', 'tf.constant', (['mask'], {'dtype': 'tf.float32'}), '(mask, dtype=tf.float32)\n', (10370, 10394), True, 'import tensorflow as tf\n'), ((10457, 10493), 'tensorflow.constant', 'tf.constant', (['[1.0]'], {'dtype': 'tf.float32'}), '([1.0], dtype=tf.float32)\n', (10468, 10493), True, 'import tensorflow as tf\n'), ((10550, 10605), 'tensorflow.constant', 'tf.constant', (['[[0.0, 0.25, 0.75, 1.0]]'], {'dtype': 'tf.float32'}), '([[0.0, 0.25, 0.75, 1.0]], dtype=tf.float32)\n', (10561, 10605), True, 'import tensorflow as tf\n'), ((10663, 10697), 'tensorflow.constant', 'tf.constant', (['[1]'], {'dtype': 'tf.float32'}), '([1], dtype=tf.float32)\n', (10674, 10697), True, 'import tensorflow as tf\n'), ((10765, 10808), 'tensorflow.constant', 'tf.constant', (['[[1.0, 0.0]]'], {'dtype': 'tf.float32'}), '([[1.0, 0.0]], dtype=tf.float32)\n', (10776, 10808), True, 'import tensorflow as tf\n'), ((10865, 10934), 'numpy.array', 'np.array', (['[[[255.0, 0.0, 0.0], [255.0, 0.0, 0.0], [255.0, 0.0, 0.0]]]'], {}), '([[[255.0, 0.0, 0.0], [255.0, 0.0, 0.0], [255.0, 0.0, 0.0]]])\n', (10873, 10934), True, 'import numpy as np\n'), ((10973, 11008), 'tensorflow.constant', 'tf.constant', (['mask'], {'dtype': 'tf.float32'}), '(mask, dtype=tf.float32)\n', (10984, 11008), True, 'import tensorflow as tf\n'), ((11074, 11122), 'numpy.array', 'np.array', (['[[[0.1, 0.1], [0.2, 0.2], [0.3, 0.3]]]'], {}), '([[[0.1, 0.1], [0.2, 0.2], [0.3, 0.3]]])\n', (11082, 11122), True, 'import numpy as np\n'), ((11148, 11188), 'tensorflow.constant', 'tf.constant', (['keypoints'], {'dtype': 'tf.float32'}), '(keypoints, dtype=tf.float32)\n', (11159, 11188), True, 'import tensorflow as tf\n'), ((11267, 11306), 'tensorflow.constant', 'tf.constant', (['[np.nan]'], {'dtype': 'tf.float32'}), '([np.nan], dtype=tf.float32)\n', (11278, 11306), True, 'import tensorflow as tf\n'), ((11379, 11432), 'tensorflow.constant', 'tf.constant', (['[[0.25, 0.5, 0.75, 1]]'], {'dtype': 'tf.float32'}), '([[0.25, 0.5, 0.75, 1]], dtype=tf.float32)\n', (11390, 11432), True, 'import tensorflow as tf\n'), ((11506, 11540), 'tensorflow.constant', 'tf.constant', (['[2]'], {'dtype': 'tf.float32'}), '([2], dtype=tf.float32)\n', (11517, 11540), True, 'import tensorflow as tf\n'), ((11634, 11672), 'object_detection.tensorflow_detect.core.preprocessor._rgb_to_grayscale', 'preprocessor._rgb_to_grayscale', (['images'], {}), '(images)\n', (11664, 11672), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((11695, 11728), 'tensorflow.image.rgb_to_grayscale', 'tf.image.rgb_to_grayscale', (['images'], {}), '(images)\n', (11720, 11728), True, 'import tensorflow as tf\n'), ((12259, 12315), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options'], {}), '(tensor_dict, preprocess_options)\n', (12282, 12315), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((13088, 13177), 'object_detection.tensorflow_detect.core.preprocessor.retain_boxes_above_threshold', 'preprocessor.retain_boxes_above_threshold', (['boxes', 'labels', 'label_scores'], {'threshold': '(0.6)'}), '(boxes, labels, label_scores,\n threshold=0.6)\n', (13129, 13177), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((14173, 14299), 'object_detection.tensorflow_detect.core.preprocessor.retain_boxes_above_threshold', 'preprocessor.retain_boxes_above_threshold', (['boxes', 'labels', 'label_scores'], {'multiclass_scores': 'multiclass_scores', 'threshold': '(0.6)'}), '(boxes, labels, label_scores,\n multiclass_scores=multiclass_scores, threshold=0.6)\n', (14214, 14299), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((14941, 15037), 'object_detection.tensorflow_detect.core.preprocessor.retain_boxes_above_threshold', 'preprocessor.retain_boxes_above_threshold', (['boxes', 'labels', 'label_scores', 'masks'], {'threshold': '(0.6)'}), '(boxes, labels, label_scores,\n masks, threshold=0.6)\n', (14982, 15037), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((15545, 15655), 'object_detection.tensorflow_detect.core.preprocessor.retain_boxes_above_threshold', 'preprocessor.retain_boxes_above_threshold', (['boxes', 'labels', 'label_scores'], {'keypoints': 'keypoints', 'threshold': '(0.6)'}), '(boxes, labels, label_scores,\n keypoints=keypoints, threshold=0.6)\n', (15586, 15655), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((16216, 16305), 'object_detection.tensorflow_detect.core.preprocessor.retain_boxes_above_threshold', 'preprocessor.retain_boxes_above_threshold', (['boxes', 'labels', 'label_scores'], {'threshold': '(0.6)'}), '(boxes, labels, label_scores,\n threshold=0.6)\n', (16257, 16305), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((17150, 17192), 'object_detection.tensorflow_detect.core.preprocessor._flip_boxes_left_right', 'preprocessor._flip_boxes_left_right', (['boxes'], {}), '(boxes)\n', (17185, 17192), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((17537, 17576), 'object_detection.tensorflow_detect.core.preprocessor._flip_boxes_up_down', 'preprocessor._flip_boxes_up_down', (['boxes'], {}), '(boxes)\n', (17569, 17576), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((17913, 17945), 'object_detection.tensorflow_detect.core.preprocessor._rot90_boxes', 'preprocessor._rot90_boxes', (['boxes'], {}), '(boxes)\n', (17938, 17945), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((18288, 18334), 'object_detection.tensorflow_detect.core.preprocessor._flip_masks_left_right', 'preprocessor._flip_masks_left_right', (['test_mask'], {}), '(test_mask)\n', (18323, 18334), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((18675, 18718), 'object_detection.tensorflow_detect.core.preprocessor._flip_masks_up_down', 'preprocessor._flip_masks_up_down', (['test_mask'], {}), '(test_mask)\n', (18707, 18718), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((19051, 19087), 'object_detection.tensorflow_detect.core.preprocessor._rot90_masks', 'preprocessor._rot90_masks', (['test_mask'], {}), '(test_mask)\n', (19076, 19087), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((19612, 19650), 'object_detection.tensorflow_detect.core.preprocessor_cache.PreprocessorCache', 'preprocessor_cache.PreprocessorCache', ([], {}), '()\n', (19648, 19650), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((19866, 19976), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_instance_masks': 'test_masks', 'include_keypoints': 'test_keypoints'}), '(include_instance_masks=test_masks,\n include_keypoints=test_keypoints)\n', (19903, 19976), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((21833, 21889), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options'], {}), '(tensor_dict, preprocess_options)\n', (21856, 21889), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((22030, 22075), 'tensorflow.squared_difference', 'tf.squared_difference', (['boxes', 'boxes_expected1'], {}), '(boxes, boxes_expected1)\n', (22051, 22075), True, 'import tensorflow as tf\n'), ((22094, 22139), 'tensorflow.squared_difference', 'tf.squared_difference', (['boxes', 'boxes_expected2'], {}), '(boxes, boxes_expected2)\n', (22115, 22139), True, 'import tensorflow as tf\n'), ((22157, 22194), 'tensorflow.multiply', 'tf.multiply', (['boxes_diff1', 'boxes_diff2'], {}), '(boxes_diff1, boxes_diff2)\n', (22168, 22194), True, 'import tensorflow as tf\n'), ((22221, 22246), 'tensorflow.zeros_like', 'tf.zeros_like', (['boxes_diff'], {}), '(boxes_diff)\n', (22234, 22246), True, 'import tensorflow as tf\n'), ((22267, 22314), 'tensorflow.squared_difference', 'tf.squared_difference', (['images', 'images_expected1'], {}), '(images, images_expected1)\n', (22288, 22314), True, 'import tensorflow as tf\n'), ((22334, 22381), 'tensorflow.squared_difference', 'tf.squared_difference', (['images', 'images_expected2'], {}), '(images, images_expected2)\n', (22355, 22381), True, 'import tensorflow as tf\n'), ((22400, 22439), 'tensorflow.multiply', 'tf.multiply', (['images_diff1', 'images_diff2'], {}), '(images_diff1, images_diff2)\n', (22411, 22439), True, 'import tensorflow as tf\n'), ((22467, 22493), 'tensorflow.zeros_like', 'tf.zeros_like', (['images_diff'], {}), '(images_diff)\n', (22480, 22493), True, 'import tensorflow as tf\n'), ((23365, 23421), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options'], {}), '(tensor_dict, preprocess_options)\n', (23388, 23421), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((23563, 23610), 'tensorflow.squared_difference', 'tf.squared_difference', (['images', 'images_expected1'], {}), '(images, images_expected1)\n', (23584, 23610), True, 'import tensorflow as tf\n'), ((23630, 23677), 'tensorflow.squared_difference', 'tf.squared_difference', (['images', 'images_expected2'], {}), '(images, images_expected2)\n', (23651, 23677), True, 'import tensorflow as tf\n'), ((23696, 23735), 'tensorflow.multiply', 'tf.multiply', (['images_diff1', 'images_diff2'], {}), '(images_diff1, images_diff2)\n', (23707, 23735), True, 'import tensorflow as tf\n'), ((23763, 23789), 'tensorflow.zeros_like', 'tf.zeros_like', (['images_diff'], {}), '(images_diff)\n', (23776, 23789), True, 'import tensorflow as tf\n'), ((24772, 24824), 'tensorflow.random_uniform', 'tf.random_uniform', (['[1, image_height, image_width, 3]'], {}), '([1, image_height, image_width, 3])\n', (24789, 24824), True, 'import tensorflow as tf\n'), ((25434, 25528), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_instance_masks': '(True)', 'include_keypoints': '(True)'}), '(include_instance_masks=True,\n include_keypoints=True)\n', (25471, 25528), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((25552, 25648), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocess_options, func_arg_map=\n preprocessor_arg_map)\n', (25575, 25648), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((26612, 26668), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options'], {}), '(tensor_dict, preprocess_options)\n', (26635, 26668), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((26809, 26854), 'tensorflow.squared_difference', 'tf.squared_difference', (['boxes', 'boxes_expected1'], {}), '(boxes, boxes_expected1)\n', (26830, 26854), True, 'import tensorflow as tf\n'), ((26873, 26918), 'tensorflow.squared_difference', 'tf.squared_difference', (['boxes', 'boxes_expected2'], {}), '(boxes, boxes_expected2)\n', (26894, 26918), True, 'import tensorflow as tf\n'), ((26936, 26973), 'tensorflow.multiply', 'tf.multiply', (['boxes_diff1', 'boxes_diff2'], {}), '(boxes_diff1, boxes_diff2)\n', (26947, 26973), True, 'import tensorflow as tf\n'), ((27000, 27025), 'tensorflow.zeros_like', 'tf.zeros_like', (['boxes_diff'], {}), '(boxes_diff)\n', (27013, 27025), True, 'import tensorflow as tf\n'), ((27046, 27093), 'tensorflow.squared_difference', 'tf.squared_difference', (['images', 'images_expected1'], {}), '(images, images_expected1)\n', (27067, 27093), True, 'import tensorflow as tf\n'), ((27113, 27160), 'tensorflow.squared_difference', 'tf.squared_difference', (['images', 'images_expected2'], {}), '(images, images_expected2)\n', (27134, 27160), True, 'import tensorflow as tf\n'), ((27179, 27218), 'tensorflow.multiply', 'tf.multiply', (['images_diff1', 'images_diff2'], {}), '(images_diff1, images_diff2)\n', (27190, 27218), True, 'import tensorflow as tf\n'), ((27246, 27272), 'tensorflow.zeros_like', 'tf.zeros_like', (['images_diff'], {}), '(images_diff)\n', (27259, 27272), True, 'import tensorflow as tf\n'), ((28137, 28193), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options'], {}), '(tensor_dict, preprocess_options)\n', (28160, 28193), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((28335, 28382), 'tensorflow.squared_difference', 'tf.squared_difference', (['images', 'images_expected1'], {}), '(images, images_expected1)\n', (28356, 28382), True, 'import tensorflow as tf\n'), ((28402, 28449), 'tensorflow.squared_difference', 'tf.squared_difference', (['images', 'images_expected2'], {}), '(images, images_expected2)\n', (28423, 28449), True, 'import tensorflow as tf\n'), ((28468, 28507), 'tensorflow.multiply', 'tf.multiply', (['images_diff1', 'images_diff2'], {}), '(images_diff1, images_diff2)\n', (28479, 28507), True, 'import tensorflow as tf\n'), ((28535, 28561), 'tensorflow.zeros_like', 'tf.zeros_like', (['images_diff'], {}), '(images_diff)\n', (28548, 28561), True, 'import tensorflow as tf\n'), ((29536, 29588), 'tensorflow.random_uniform', 'tf.random_uniform', (['[1, image_height, image_width, 3]'], {}), '([1, image_height, image_width, 3])\n', (29553, 29588), True, 'import tensorflow as tf\n'), ((30196, 30290), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_instance_masks': '(True)', 'include_keypoints': '(True)'}), '(include_instance_masks=True,\n include_keypoints=True)\n', (30233, 30290), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((30314, 30410), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocess_options, func_arg_map=\n preprocessor_arg_map)\n', (30337, 30410), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((31359, 31415), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options'], {}), '(tensor_dict, preprocess_options)\n', (31382, 31415), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((31556, 31601), 'tensorflow.squared_difference', 'tf.squared_difference', (['boxes', 'boxes_expected1'], {}), '(boxes, boxes_expected1)\n', (31577, 31601), True, 'import tensorflow as tf\n'), ((31620, 31665), 'tensorflow.squared_difference', 'tf.squared_difference', (['boxes', 'boxes_expected2'], {}), '(boxes, boxes_expected2)\n', (31641, 31665), True, 'import tensorflow as tf\n'), ((31683, 31720), 'tensorflow.multiply', 'tf.multiply', (['boxes_diff1', 'boxes_diff2'], {}), '(boxes_diff1, boxes_diff2)\n', (31694, 31720), True, 'import tensorflow as tf\n'), ((31747, 31772), 'tensorflow.zeros_like', 'tf.zeros_like', (['boxes_diff'], {}), '(boxes_diff)\n', (31760, 31772), True, 'import tensorflow as tf\n'), ((31793, 31840), 'tensorflow.squared_difference', 'tf.squared_difference', (['images', 'images_expected1'], {}), '(images, images_expected1)\n', (31814, 31840), True, 'import tensorflow as tf\n'), ((31860, 31907), 'tensorflow.squared_difference', 'tf.squared_difference', (['images', 'images_expected2'], {}), '(images, images_expected2)\n', (31881, 31907), True, 'import tensorflow as tf\n'), ((31926, 31965), 'tensorflow.multiply', 'tf.multiply', (['images_diff1', 'images_diff2'], {}), '(images_diff1, images_diff2)\n', (31937, 31965), True, 'import tensorflow as tf\n'), ((31993, 32019), 'tensorflow.zeros_like', 'tf.zeros_like', (['images_diff'], {}), '(images_diff)\n', (32006, 32019), True, 'import tensorflow as tf\n'), ((32874, 32930), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options'], {}), '(tensor_dict, preprocess_options)\n', (32897, 32930), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((33072, 33119), 'tensorflow.squared_difference', 'tf.squared_difference', (['images', 'images_expected1'], {}), '(images, images_expected1)\n', (33093, 33119), True, 'import tensorflow as tf\n'), ((33139, 33186), 'tensorflow.squared_difference', 'tf.squared_difference', (['images', 'images_expected2'], {}), '(images, images_expected2)\n', (33160, 33186), True, 'import tensorflow as tf\n'), ((33205, 33244), 'tensorflow.multiply', 'tf.multiply', (['images_diff1', 'images_diff2'], {}), '(images_diff1, images_diff2)\n', (33216, 33244), True, 'import tensorflow as tf\n'), ((33272, 33298), 'tensorflow.zeros_like', 'tf.zeros_like', (['images_diff'], {}), '(images_diff)\n', (33285, 33298), True, 'import tensorflow as tf\n'), ((34122, 34174), 'tensorflow.random_uniform', 'tf.random_uniform', (['[1, image_height, image_width, 3]'], {}), '([1, image_height, image_width, 3])\n', (34139, 34174), True, 'import tensorflow as tf\n'), ((34574, 34668), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_instance_masks': '(True)', 'include_keypoints': '(True)'}), '(include_instance_masks=True,\n include_keypoints=True)\n', (34611, 34668), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((34692, 34788), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocess_options, func_arg_map=\n preprocessor_arg_map)\n', (34715, 34788), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((35693, 35752), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (35716, 35752), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((35931, 35967), 'tensorflow.greater_equal', 'tf.greater_equal', (['images', 'images_min'], {}), '(images, images_min)\n', (35947, 35967), True, 'import tensorflow as tf\n'), ((35986, 36019), 'tensorflow.less_equal', 'tf.less_equal', (['images', 'images_max'], {}), '(images, images_max)\n', (35999, 36019), True, 'import tensorflow as tf\n'), ((36038, 36065), 'tensorflow.fill', 'tf.fill', (['[1, 4, 4, 3]', '(True)'], {}), '([1, 4, 4, 3], True)\n', (36045, 36065), True, 'import tensorflow as tf\n'), ((37108, 37164), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options'], {}), '(tensor_dict, preprocess_options)\n', (37131, 37164), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((37255, 37280), 'tensorflow.shape', 'tf.shape', (['images_original'], {}), '(images_original)\n', (37263, 37280), True, 'import tensorflow as tf\n'), ((37307, 37330), 'tensorflow.shape', 'tf.shape', (['images_scaled'], {}), '(images_scaled)\n', (37315, 37330), True, 'import tensorflow as tf\n'), ((38406, 38462), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options'], {}), '(tensor_dict, preprocess_options)\n', (38429, 38462), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((38573, 38630), 'tensorflow.split', 'tf.split', ([], {'value': 'images_gray', 'num_or_size_splits': '(3)', 'axis': '(3)'}), '(value=images_gray, num_or_size_splits=3, axis=3)\n', (38581, 38630), True, 'import tensorflow as tf\n'), ((38675, 38736), 'tensorflow.split', 'tf.split', ([], {'value': 'images_original', 'num_or_size_splits': '(3)', 'axis': '(3)'}), '(value=images_original, num_or_size_splits=3, axis=3)\n', (38683, 38736), True, 'import tensorflow as tf\n'), ((39045, 39088), 'tensorflow.multiply', 'tf.multiply', (['images_r_diff1', 'images_r_diff2'], {}), '(images_r_diff1, images_r_diff2)\n', (39056, 39088), True, 'import tensorflow as tf\n'), ((39388, 39431), 'tensorflow.multiply', 'tf.multiply', (['images_g_diff1', 'images_g_diff2'], {}), '(images_g_diff1, images_g_diff2)\n', (39399, 39431), True, 'import tensorflow as tf\n'), ((39731, 39774), 'tensorflow.multiply', 'tf.multiply', (['images_b_diff1', 'images_b_diff2'], {}), '(images_b_diff1, images_b_diff2)\n', (39742, 39774), True, 'import tensorflow as tf\n'), ((39793, 39845), 'tensorflow.constant', 'tf.constant', (['(0)'], {'dtype': 'tf.float32', 'shape': '[1, 4, 4, 1]'}), '(0, dtype=tf.float32, shape=[1, 4, 4, 1])\n', (39804, 39845), True, 'import tensorflow as tf\n'), ((41015, 41074), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (41038, 41074), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((41164, 41189), 'tensorflow.shape', 'tf.shape', (['images_original'], {}), '(images_original)\n', (41172, 41189), True, 'import tensorflow as tf\n'), ((41215, 41238), 'tensorflow.shape', 'tf.shape', (['images_bright'], {}), '(images_bright)\n', (41223, 41238), True, 'import tensorflow as tf\n'), ((42478, 42537), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (42501, 42537), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((42629, 42654), 'tensorflow.shape', 'tf.shape', (['images_original'], {}), '(images_original)\n', (42637, 42654), True, 'import tensorflow as tf\n'), ((42682, 42707), 'tensorflow.shape', 'tf.shape', (['images_contrast'], {}), '(images_contrast)\n', (42690, 42707), True, 'import tensorflow as tf\n'), ((43939, 43998), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (43962, 43998), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((44085, 44110), 'tensorflow.shape', 'tf.shape', (['images_original'], {}), '(images_original)\n', (44093, 44110), True, 'import tensorflow as tf\n'), ((44133, 44153), 'tensorflow.shape', 'tf.shape', (['images_hue'], {}), '(images_hue)\n', (44141, 44153), True, 'import tensorflow as tf\n'), ((45310, 45335), 'tensorflow.shape', 'tf.shape', (['images_original'], {}), '(images_original)\n', (45318, 45335), True, 'import tensorflow as tf\n'), ((45420, 45479), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (45443, 45479), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((45586, 45618), 'tensorflow.shape', 'tf.shape', (['images_distorted_color'], {}), '(images_distorted_color)\n', (45594, 45618), True, 'import tensorflow as tf\n'), ((46609, 46624), 'tensorflow.shape', 'tf.shape', (['boxes'], {}), '(boxes)\n', (46617, 46624), True, 'import tensorflow as tf\n'), ((46711, 46770), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (46734, 46770), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((46875, 46900), 'tensorflow.shape', 'tf.shape', (['distorted_boxes'], {}), '(distorted_boxes)\n', (46883, 46900), True, 'import tensorflow as tf\n'), ((47762, 47821), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (47785, 47821), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((48061, 48075), 'tensorflow.rank', 'tf.rank', (['boxes'], {}), '(boxes)\n', (48068, 48075), True, 'import tensorflow as tf\n'), ((48103, 48127), 'tensorflow.rank', 'tf.rank', (['distorted_boxes'], {}), '(distorted_boxes)\n', (48110, 48127), True, 'import tensorflow as tf\n'), ((48146, 48161), 'tensorflow.rank', 'tf.rank', (['images'], {}), '(images)\n', (48153, 48161), True, 'import tensorflow as tf\n'), ((48190, 48215), 'tensorflow.rank', 'tf.rank', (['distorted_images'], {}), '(distorted_images)\n', (48197, 48215), True, 'import tensorflow as tf\n'), ((50156, 50215), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (50179, 50215), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((50412, 50426), 'tensorflow.rank', 'tf.rank', (['boxes'], {}), '(boxes)\n', (50419, 50426), True, 'import tensorflow as tf\n'), ((50454, 50478), 'tensorflow.rank', 'tf.rank', (['distorted_boxes'], {}), '(distorted_boxes)\n', (50461, 50478), True, 'import tensorflow as tf\n'), ((50497, 50512), 'tensorflow.rank', 'tf.rank', (['images'], {}), '(images)\n', (50504, 50512), True, 'import tensorflow as tf\n'), ((50541, 50566), 'tensorflow.rank', 'tf.rank', (['distorted_images'], {}), '(distorted_images)\n', (50548, 50566), True, 'import tensorflow as tf\n'), ((51702, 51761), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (51725, 51761), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((52001, 52015), 'tensorflow.rank', 'tf.rank', (['boxes'], {}), '(boxes)\n', (52008, 52015), True, 'import tensorflow as tf\n'), ((52043, 52067), 'tensorflow.rank', 'tf.rank', (['distorted_boxes'], {}), '(distorted_boxes)\n', (52050, 52067), True, 'import tensorflow as tf\n'), ((52086, 52101), 'tensorflow.rank', 'tf.rank', (['images'], {}), '(images)\n', (52093, 52101), True, 'import tensorflow as tf\n'), ((52130, 52155), 'tensorflow.rank', 'tf.rank', (['distorted_images'], {}), '(distorted_images)\n', (52137, 52155), True, 'import tensorflow as tf\n'), ((53189, 53248), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (53212, 53248), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((53432, 53491), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (53455, 53491), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((53940, 53955), 'tensorflow.shape', 'tf.shape', (['boxes'], {}), '(boxes)\n', (53948, 53955), True, 'import tensorflow as tf\n'), ((53984, 54009), 'tensorflow.shape', 'tf.shape', (['distorted_boxes'], {}), '(distorted_boxes)\n', (53992, 54009), True, 'import tensorflow as tf\n'), ((54029, 54045), 'tensorflow.shape', 'tf.shape', (['images'], {}), '(images)\n', (54037, 54045), True, 'import tensorflow as tf\n'), ((54075, 54101), 'tensorflow.shape', 'tf.shape', (['distorted_images'], {}), '(distorted_images)\n', (54083, 54101), True, 'import tensorflow as tf\n'), ((55281, 55385), 'tensorflow.constant', 'tf.constant', (['[[0.1, 0.1, 0.8, 0.3], [0.2, 0.4, 0.75, 0.75], [0.3, 0.1, 0.4, 0.7]]'], {'dtype': 'tf.float32'}), '([[0.1, 0.1, 0.8, 0.3], [0.2, 0.4, 0.75, 0.75], [0.3, 0.1, 0.4, \n 0.7]], dtype=tf.float32)\n', (55292, 55385), True, 'import tensorflow as tf\n'), ((55444, 55483), 'tensorflow.constant', 'tf.constant', (['[1, 7, 11]'], {'dtype': 'tf.int32'}), '([1, 7, 11], dtype=tf.int32)\n', (55455, 55483), True, 'import tensorflow as tf\n'), ((55692, 55751), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (55715, 55751), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((58040, 58099), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (58063, 58099), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((58446, 58460), 'tensorflow.rank', 'tf.rank', (['boxes'], {}), '(boxes)\n', (58453, 58460), True, 'import tensorflow as tf\n'), ((58488, 58512), 'tensorflow.rank', 'tf.rank', (['distorted_boxes'], {}), '(distorted_boxes)\n', (58495, 58512), True, 'import tensorflow as tf\n'), ((58531, 58546), 'tensorflow.rank', 'tf.rank', (['images'], {}), '(images)\n', (58538, 58546), True, 'import tensorflow as tf\n'), ((58575, 58600), 'tensorflow.rank', 'tf.rank', (['distorted_images'], {}), '(distorted_images)\n', (58582, 58600), True, 'import tensorflow as tf\n'), ((58630, 58656), 'tensorflow.rank', 'tf.rank', (['multiclass_scores'], {}), '(multiclass_scores)\n', (58637, 58656), True, 'import tensorflow as tf\n'), ((58696, 58732), 'tensorflow.rank', 'tf.rank', (['distorted_multiclass_scores'], {}), '(distorted_multiclass_scores)\n', (58703, 58732), True, 'import tensorflow as tf\n'), ((61062, 61112), 'tensorflow.random_uniform', 'tf.random_uniform', (['[2, 200, 400]'], {'dtype': 'tf.float32'}), '([2, 200, 400], dtype=tf.float32)\n', (61079, 61112), True, 'import tensorflow as tf\n'), ((63926, 63976), 'tensorflow.random_uniform', 'tf.random_uniform', (['[2, 200, 400]'], {'dtype': 'tf.float32'}), '([2, 200, 400], dtype=tf.float32)\n', (63943, 63976), True, 'import tensorflow as tf\n'), ((64260, 64326), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_instance_masks': '(True)'}), '(include_instance_masks=True)\n', (64297, 64326), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((66494, 66555), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_keypoints': '(True)'}), '(include_keypoints=True)\n', (66531, 66555), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((69048, 69109), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_keypoints': '(True)'}), '(include_keypoints=True)\n', (69085, 69109), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((71567, 71631), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_label_scores': '(True)'}), '(include_label_scores=True)\n', (71604, 71631), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((71668, 71767), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (71691, 71767), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((73268, 73365), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_label_scores': '(True)', 'include_instance_masks': '(True)'}), '(include_label_scores=True,\n include_instance_masks=True)\n', (73305, 73365), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((73516, 73615), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (73539, 73615), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((74482, 74574), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_label_scores': '(True)', 'include_keypoints': '(True)'}), '(include_label_scores=True,\n include_keypoints=True)\n', (74519, 74574), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((74725, 74824), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (74748, 74824), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((75695, 75745), 'tensorflow.random_uniform', 'tf.random_uniform', (['[2, 200, 400]'], {'dtype': 'tf.float32'}), '([2, 200, 400], dtype=tf.float32)\n', (75712, 75745), True, 'import tensorflow as tf\n'), ((76028, 76094), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_instance_masks': '(True)'}), '(include_instance_masks=True)\n', (76065, 76094), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((77996, 78057), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_keypoints': '(True)'}), '(include_keypoints=True)\n', (78033, 78057), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((80367, 80406), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {}), '()\n', (80404, 80406), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((80645, 80744), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (80668, 80744), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((81727, 81777), 'tensorflow.random_uniform', 'tf.random_uniform', (['[2, 200, 400]'], {'dtype': 'tf.float32'}), '([2, 200, 400], dtype=tf.float32)\n', (81744, 81777), True, 'import tensorflow as tf\n'), ((82060, 82126), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_instance_masks': '(True)'}), '(include_instance_masks=True)\n', (82097, 82126), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((82242, 82341), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (82265, 82341), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((83847, 83908), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_keypoints': '(True)'}), '(include_keypoints=True)\n', (83884, 83908), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((84024, 84123), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (84047, 84123), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((86363, 86422), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (86386, 86422), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((86570, 86629), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (86593, 86629), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((86856, 86871), 'tensorflow.shape', 'tf.shape', (['boxes'], {}), '(boxes)\n', (86864, 86871), True, 'import tensorflow as tf\n'), ((86897, 86919), 'tensorflow.shape', 'tf.shape', (['padded_boxes'], {}), '(padded_boxes)\n', (86905, 86919), True, 'import tensorflow as tf\n'), ((86939, 86955), 'tensorflow.shape', 'tf.shape', (['images'], {}), '(images)\n', (86947, 86955), True, 'import tensorflow as tf\n'), ((86982, 87005), 'tensorflow.shape', 'tf.shape', (['padded_images'], {}), '(padded_images)\n', (86990, 87005), True, 'import tensorflow as tf\n'), ((88912, 88971), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (88935, 88971), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((89156, 89215), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (89179, 89215), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((89442, 89457), 'tensorflow.shape', 'tf.shape', (['boxes'], {}), '(boxes)\n', (89450, 89457), True, 'import tensorflow as tf\n'), ((89483, 89505), 'tensorflow.shape', 'tf.shape', (['padded_boxes'], {}), '(padded_boxes)\n', (89491, 89505), True, 'import tensorflow as tf\n'), ((89525, 89541), 'tensorflow.shape', 'tf.shape', (['images'], {}), '(images)\n', (89533, 89541), True, 'import tensorflow as tf\n'), ((89568, 89591), 'tensorflow.shape', 'tf.shape', (['padded_images'], {}), '(padded_images)\n', (89576, 89591), True, 'import tensorflow as tf\n'), ((90827, 90867), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', '[]'], {}), '(tensor_dict, [])\n', (90850, 90867), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((91060, 91119), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (91083, 91119), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((91351, 91366), 'tensorflow.shape', 'tf.shape', (['boxes'], {}), '(boxes)\n', (91359, 91366), True, 'import tensorflow as tf\n'), ((91393, 91416), 'tensorflow.shape', 'tf.shape', (['cropped_boxes'], {}), '(cropped_boxes)\n', (91401, 91416), True, 'import tensorflow as tf\n'), ((91436, 91452), 'tensorflow.shape', 'tf.shape', (['images'], {}), '(images)\n', (91444, 91452), True, 'import tensorflow as tf\n'), ((91480, 91504), 'tensorflow.shape', 'tf.shape', (['cropped_images'], {}), '(cropped_images)\n', (91488, 91504), True, 'import tensorflow as tf\n'), ((92292, 92332), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', '[]'], {}), '(tensor_dict, [])\n', (92315, 92332), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((92523, 92582), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (92546, 92582), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((92809, 92824), 'tensorflow.shape', 'tf.shape', (['boxes'], {}), '(boxes)\n', (92817, 92824), True, 'import tensorflow as tf\n'), ((92850, 92872), 'tensorflow.shape', 'tf.shape', (['padded_boxes'], {}), '(padded_boxes)\n', (92858, 92872), True, 'import tensorflow as tf\n'), ((92892, 92908), 'tensorflow.shape', 'tf.shape', (['images'], {}), '(images)\n', (92900, 92908), True, 'import tensorflow as tf\n'), ((92935, 92958), 'tensorflow.shape', 'tf.shape', (['padded_images'], {}), '(padded_images)\n', (92943, 92958), True, 'import tensorflow as tf\n'), ((94451, 94510), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (94474, 94510), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((94651, 94667), 'tensorflow.shape', 'tf.shape', (['images'], {}), '(images)\n', (94659, 94667), True, 'import tensorflow as tf\n'), ((94695, 94719), 'tensorflow.shape', 'tf.shape', (['blacked_images'], {}), '(blacked_images)\n', (94703, 94719), True, 'import tensorflow as tf\n'), ((95993, 96052), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (96016, 96052), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((96201, 96225), 'tensorflow.shape', 'tf.shape', (['resized_images'], {}), '(resized_images)\n', (96209, 96225), True, 'import tensorflow as tf\n'), ((96254, 96298), 'tensorflow.constant', 'tf.constant', (['[1, 75, 150, 3]'], {'dtype': 'tf.int32'}), '([1, 75, 150, 3], dtype=tf.int32)\n', (96265, 96298), True, 'import tensorflow as tf\n'), ((98057, 98088), 'tensorflow.constant', 'tf.constant', (['(50)'], {'dtype': 'tf.int32'}), '(50, dtype=tf.int32)\n', (98068, 98088), True, 'import tensorflow as tf\n'), ((98101, 98133), 'tensorflow.constant', 'tf.constant', (['(100)'], {'dtype': 'tf.int32'}), '(100, dtype=tf.int32)\n', (98112, 98133), True, 'import tensorflow as tf\n'), ((102853, 102888), 'numpy.array', 'np.array', (['[[[0, 1, 2]]]', 'np.float32'], {}), '([[[0, 1, 2]]], np.float32)\n', (102861, 102888), True, 'import numpy as np\n'), ((102907, 103033), 'numpy.array', 'np.array', (['[[[0, 1, 2], [123.68, 116.779, 103.939]], [[123.68, 116.779, 103.939], [\n 123.68, 116.779, 103.939]]]', 'np.float32'], {}), '([[[0, 1, 2], [123.68, 116.779, 103.939]], [[123.68, 116.779, \n 103.939], [123.68, 116.779, 103.939]]], np.float32)\n', (102915, 103033), True, 'import numpy as np\n'), ((103095, 103144), 'tensorflow.placeholder', 'tf.placeholder', (['tf.float32'], {'shape': '(None, None, 3)'}), '(tf.float32, shape=(None, None, 3))\n', (103109, 103144), True, 'import tensorflow as tf\n'), ((103164, 103335), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_range', 'preprocessor.resize_to_range', (['in_image'], {'min_dimension': 'min_dim', 'max_dimension': 'max_dim', 'pad_to_max_dimension': '(True)', 'per_channel_pad_value': '(123.68, 116.779, 103.939)'}), '(in_image, min_dimension=min_dim, max_dimension\n =max_dim, pad_to_max_dimension=True, per_channel_pad_value=(123.68, \n 116.779, 103.939))\n', (103192, 103335), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((108997, 109032), 'tensorflow.random_uniform', 'tf.random_uniform', (['[1, 200, 300, 3]'], {}), '([1, 200, 300, 3])\n', (109014, 109032), True, 'import tensorflow as tf\n'), ((112581, 112616), 'tensorflow.random_uniform', 'tf.random_uniform', (['[1, 200, 300, 3]'], {}), '([1, 200, 300, 3])\n', (112598, 112616), True, 'import tensorflow as tf\n'), ((113018, 113045), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_shape'], {}), '(in_shape)\n', (113035, 113045), True, 'import tensorflow as tf\n'), ((113061, 113082), 'tensorflow.constant', 'tf.constant', (['in_boxes'], {}), '(in_boxes)\n', (113072, 113082), True, 'import tensorflow as tf\n'), ((113102, 113173), 'object_detection.tensorflow_detect.core.preprocessor.scale_boxes_to_pixel_coordinates', 'preprocessor.scale_boxes_to_pixel_coordinates', (['in_image'], {'boxes': 'in_boxes'}), '(in_image, boxes=in_boxes)\n', (113147, 113173), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((113775, 113802), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_shape'], {}), '(in_shape)\n', (113792, 113802), True, 'import tensorflow as tf\n'), ((113837, 113936), 'object_detection.tensorflow_detect.core.preprocessor.scale_boxes_to_pixel_coordinates', 'preprocessor.scale_boxes_to_pixel_coordinates', (['in_image'], {'boxes': 'in_boxes', 'keypoints': 'in_keypoints'}), '(in_image, boxes=in_boxes,\n keypoints=in_keypoints)\n', (113882, 113936), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((116063, 116122), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (116086, 116122), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((116364, 116379), 'tensorflow.rank', 'tf.rank', (['images'], {}), '(images)\n', (116371, 116379), True, 'import tensorflow as tf\n'), ((116408, 116433), 'tensorflow.rank', 'tf.rank', (['distorted_images'], {}), '(distorted_images)\n', (116415, 116433), True, 'import tensorflow as tf\n'), ((116451, 116465), 'tensorflow.rank', 'tf.rank', (['boxes'], {}), '(boxes)\n', (116458, 116465), True, 'import tensorflow as tf\n'), ((116493, 116517), 'tensorflow.rank', 'tf.rank', (['distorted_boxes'], {}), '(distorted_boxes)\n', (116500, 116517), True, 'import tensorflow as tf\n'), ((117606, 117675), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_multiclass_scores': '(True)'}), '(include_multiclass_scores=True)\n', (117643, 117675), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((117713, 117812), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (117736, 117812), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((118113, 118128), 'tensorflow.rank', 'tf.rank', (['images'], {}), '(images)\n', (118120, 118128), True, 'import tensorflow as tf\n'), ((118157, 118182), 'tensorflow.rank', 'tf.rank', (['distorted_images'], {}), '(distorted_images)\n', (118164, 118182), True, 'import tensorflow as tf\n'), ((118200, 118214), 'tensorflow.rank', 'tf.rank', (['boxes'], {}), '(boxes)\n', (118207, 118214), True, 'import tensorflow as tf\n'), ((118242, 118266), 'tensorflow.rank', 'tf.rank', (['distorted_boxes'], {}), '(distorted_boxes)\n', (118249, 118266), True, 'import tensorflow as tf\n'), ((118296, 118322), 'tensorflow.rank', 'tf.rank', (['multiclass_scores'], {}), '(multiclass_scores)\n', (118303, 118322), True, 'import tensorflow as tf\n'), ((118362, 118398), 'tensorflow.rank', 'tf.rank', (['distorted_multiclass_scores'], {}), '(distorted_multiclass_scores)\n', (118369, 118398), True, 'import tensorflow as tf\n'), ((119858, 119917), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (119881, 119917), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((120159, 120174), 'tensorflow.rank', 'tf.rank', (['images'], {}), '(images)\n', (120166, 120174), True, 'import tensorflow as tf\n'), ((120203, 120228), 'tensorflow.rank', 'tf.rank', (['distorted_images'], {}), '(distorted_images)\n', (120210, 120228), True, 'import tensorflow as tf\n'), ((120246, 120260), 'tensorflow.rank', 'tf.rank', (['boxes'], {}), '(boxes)\n', (120253, 120260), True, 'import tensorflow as tf\n'), ((120288, 120312), 'tensorflow.rank', 'tf.rank', (['distorted_boxes'], {}), '(distorted_boxes)\n', (120295, 120312), True, 'import tensorflow as tf\n'), ((122733, 122965), 'object_detection.tensorflow_detect.core.preprocessor.get_default_func_arg_map', 'preprocessor.get_default_func_arg_map', ([], {'include_label_scores': 'include_label_scores', 'include_multiclass_scores': 'include_multiclass_scores', 'include_instance_masks': 'include_instance_masks', 'include_keypoints': 'include_keypoints'}), '(include_label_scores=\n include_label_scores, include_multiclass_scores=\n include_multiclass_scores, include_instance_masks=\n include_instance_masks, include_keypoints=include_keypoints)\n', (122770, 122965), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((123012, 123111), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (123035, 123111), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((123304, 123319), 'tensorflow.rank', 'tf.rank', (['images'], {}), '(images)\n', (123311, 123319), True, 'import tensorflow as tf\n'), ((123348, 123373), 'tensorflow.rank', 'tf.rank', (['distorted_images'], {}), '(distorted_images)\n', (123355, 123373), True, 'import tensorflow as tf\n'), ((123391, 123405), 'tensorflow.rank', 'tf.rank', (['boxes'], {}), '(boxes)\n', (123398, 123405), True, 'import tensorflow as tf\n'), ((123433, 123457), 'tensorflow.rank', 'tf.rank', (['distorted_boxes'], {}), '(distorted_boxes)\n', (123440, 123457), True, 'import tensorflow as tf\n'), ((125315, 125381), 'tensorflow.constant', 'tf.constant', (['[[1.0, 0.0], [0.5, 0.5], [1000, 1]]'], {'dtype': 'tf.float32'}), '([[1.0, 0.0], [0.5, 0.5], [1000, 1]], dtype=tf.float32)\n', (125326, 125381), True, 'import tensorflow as tf\n'), ((125458, 125569), 'object_detection.tensorflow_detect.core.preprocessor.convert_class_logits_to_softmax', 'preprocessor.convert_class_logits_to_softmax', ([], {'multiclass_scores': 'multiclass_scores', 'temperature': 'temperature'}), '(multiclass_scores=\n multiclass_scores, temperature=temperature)\n', (125502, 125569), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((1176, 1208), 'tensorflow.constant', 'tf.constant', (['(255)'], {'dtype': 'tf.uint8'}), '(255, dtype=tf.uint8)\n', (1187, 1208), True, 'import tensorflow as tf\n'), ((1248, 1280), 'tensorflow.constant', 'tf.constant', (['(128)'], {'dtype': 'tf.uint8'}), '(128, dtype=tf.uint8)\n', (1259, 1280), True, 'import tensorflow as tf\n'), ((1318, 1348), 'tensorflow.constant', 'tf.constant', (['(0)'], {'dtype': 'tf.uint8'}), '(0, dtype=tf.uint8)\n', (1329, 1348), True, 'import tensorflow as tf\n'), ((38789, 38810), 'tensorflow.to_float', 'tf.to_float', (['images_r'], {}), '(images_r)\n', (38800, 38810), True, 'import tensorflow as tf\n'), ((38855, 38881), 'tensorflow.to_float', 'tf.to_float', (['images_gray_r'], {}), '(images_gray_r)\n', (38866, 38881), True, 'import tensorflow as tf\n'), ((38926, 38952), 'tensorflow.to_float', 'tf.to_float', (['images_gray_r'], {}), '(images_gray_r)\n', (38937, 38952), True, 'import tensorflow as tf\n'), ((38997, 39023), 'tensorflow.to_float', 'tf.to_float', (['images_gray_g'], {}), '(images_gray_g)\n', (39008, 39023), True, 'import tensorflow as tf\n'), ((39132, 39153), 'tensorflow.to_float', 'tf.to_float', (['images_g'], {}), '(images_g)\n', (39143, 39153), True, 'import tensorflow as tf\n'), ((39198, 39224), 'tensorflow.to_float', 'tf.to_float', (['images_gray_g'], {}), '(images_gray_g)\n', (39209, 39224), True, 'import tensorflow as tf\n'), ((39269, 39295), 'tensorflow.to_float', 'tf.to_float', (['images_gray_g'], {}), '(images_gray_g)\n', (39280, 39295), True, 'import tensorflow as tf\n'), ((39340, 39366), 'tensorflow.to_float', 'tf.to_float', (['images_gray_b'], {}), '(images_gray_b)\n', (39351, 39366), True, 'import tensorflow as tf\n'), ((39475, 39496), 'tensorflow.to_float', 'tf.to_float', (['images_b'], {}), '(images_b)\n', (39486, 39496), True, 'import tensorflow as tf\n'), ((39541, 39567), 'tensorflow.to_float', 'tf.to_float', (['images_gray_b'], {}), '(images_gray_b)\n', (39552, 39567), True, 'import tensorflow as tf\n'), ((39612, 39638), 'tensorflow.to_float', 'tf.to_float', (['images_gray_b'], {}), '(images_gray_b)\n', (39623, 39638), True, 'import tensorflow as tf\n'), ((39683, 39709), 'tensorflow.to_float', 'tf.to_float', (['images_gray_r'], {}), '(images_gray_r)\n', (39694, 39709), True, 'import tensorflow as tf\n'), ((55884, 55944), 'unittest.mock.patch.object', 'mock.patch.object', (['tf.image', '"""sample_distorted_bounding_box"""'], {}), "(tf.image, 'sample_distorted_bounding_box')\n", (55901, 55944), False, 'from unittest import mock\n'), ((56284, 56343), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {}), '(tensor_dict, preprocessing_options)\n', (56307, 56343), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((56623, 56744), 'tensorflow.constant', 'tf.constant', (['[[0.178947, 0.07173, 0.75789469, 0.66244733], [0.28421, 0.0, 0.38947365, \n 0.57805908]]'], {'dtype': 'tf.float32'}), '([[0.178947, 0.07173, 0.75789469, 0.66244733], [0.28421, 0.0, \n 0.38947365, 0.57805908]], dtype=tf.float32)\n', (56634, 56744), True, 'import tensorflow as tf\n'), ((56835, 56871), 'tensorflow.constant', 'tf.constant', (['[7, 11]'], {'dtype': 'tf.int32'}), '([7, 11], dtype=tf.int32)\n', (56846, 56871), True, 'import tensorflow as tf\n'), ((59797, 59857), 'unittest.mock.patch.object', 'mock.patch.object', (['tf.image', '"""sample_distorted_bounding_box"""'], {}), "(tf.image, 'sample_distorted_bounding_box')\n", (59814, 59857), False, 'from unittest import mock\n'), ((60228, 60302), 'object_detection.tensorflow_detect.core.preprocessor._strict_random_crop_image', 'preprocessor._strict_random_crop_image', (['image', 'boxes', 'labels', 'label_scores'], {}), '(image, boxes, labels, label_scores)\n', (60266, 60302), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((61122, 61182), 'unittest.mock.patch.object', 'mock.patch.object', (['tf.image', '"""sample_distorted_bounding_box"""'], {}), "(tf.image, 'sample_distorted_bounding_box')\n", (61139, 61182), False, 'from unittest import mock\n'), ((61546, 61619), 'object_detection.tensorflow_detect.core.preprocessor._strict_random_crop_image', 'preprocessor._strict_random_crop_image', (['image', 'boxes', 'labels'], {'masks': 'masks'}), '(image, boxes, labels, masks=masks)\n', (61584, 61619), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((62382, 62442), 'unittest.mock.patch.object', 'mock.patch.object', (['tf.image', '"""sample_distorted_bounding_box"""'], {}), "(tf.image, 'sample_distorted_bounding_box')\n", (62399, 62442), False, 'from unittest import mock\n'), ((62810, 62896), 'object_detection.tensorflow_detect.core.preprocessor._strict_random_crop_image', 'preprocessor._strict_random_crop_image', (['image', 'boxes', 'labels'], {'keypoints': 'keypoints'}), '(image, boxes, labels, keypoints=\n keypoints)\n', (62848, 62896), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((64414, 64474), 'unittest.mock.patch.object', 'mock.patch.object', (['tf.image', '"""sample_distorted_bounding_box"""'], {}), "(tf.image, 'sample_distorted_bounding_box')\n", (64431, 64474), False, 'from unittest import mock\n'), ((64804, 64903), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (64827, 64903), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((66643, 66703), 'unittest.mock.patch.object', 'mock.patch.object', (['tf.image', '"""sample_distorted_bounding_box"""'], {}), "(tf.image, 'sample_distorted_bounding_box')\n", (66660, 66703), False, 'from unittest import mock\n'), ((67033, 67132), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (67056, 67132), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((69197, 69257), 'unittest.mock.patch.object', 'mock.patch.object', (['tf.image', '"""sample_distorted_bounding_box"""'], {}), "(tf.image, 'sample_distorted_bounding_box')\n", (69214, 69257), False, 'from unittest import mock\n'), ((69587, 69686), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (69610, 69686), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((76192, 76242), 'unittest.mock.patch.object', 'mock.patch.object', (['preprocessor', '"""_random_integer"""'], {}), "(preprocessor, '_random_integer')\n", (76209, 76242), False, 'from unittest import mock\n'), ((76335, 76365), 'tensorflow.constant', 'tf.constant', (['(0)'], {'dtype': 'tf.int32'}), '(0, dtype=tf.int32)\n', (76346, 76365), True, 'import tensorflow as tf\n'), ((76396, 76495), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (76419, 76495), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((78155, 78205), 'unittest.mock.patch.object', 'mock.patch.object', (['preprocessor', '"""_random_integer"""'], {}), "(preprocessor, '_random_integer')\n", (78172, 78205), False, 'from unittest import mock\n'), ((78298, 78328), 'tensorflow.constant', 'tf.constant', (['(0)'], {'dtype': 'tf.int32'}), '(0, dtype=tf.int32)\n', (78309, 78328), True, 'import tensorflow as tf\n'), ((78359, 78458), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocessing_options'], {'func_arg_map': 'preprocessor_arg_map'}), '(tensor_dict, preprocessing_options, func_arg_map=\n preprocessor_arg_map)\n', (78382, 78458), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((81214, 81303), 'numpy.array', 'np.array', (['[[0.0, 0.125, 0.1875, 0.5], [0.0625, 0.25, 0.1875, 0.5]]'], {'dtype': 'np.float32'}), '([[0.0, 0.125, 0.1875, 0.5], [0.0625, 0.25, 0.1875, 0.5]], dtype=np\n .float32)\n', (81222, 81303), True, 'import numpy as np\n'), ((82968, 83047), 'numpy.array', 'np.array', (['[[0.0, 0.25, 0.375, 1.0], [0.125, 0.5, 0.375, 1.0]]'], {'dtype': 'np.float32'}), '([[0.0, 0.25, 0.375, 1.0], [0.125, 0.5, 0.375, 1.0]], dtype=np.float32)\n', (82976, 83047), True, 'import numpy as np\n'), ((84768, 84847), 'numpy.array', 'np.array', (['[[0.0, 0.25, 0.375, 1.0], [0.125, 0.5, 0.375, 1.0]]'], {'dtype': 'np.float32'}), '([[0.0, 0.25, 0.375, 1.0], [0.125, 0.5, 0.375, 1.0]], dtype=np.float32)\n', (84776, 84847), True, 'import numpy as np\n'), ((84886, 84997), 'numpy.array', 'np.array', (['[[[0.05, 0.1], [0.1, 0.2], [0.15, 0.3]], [[0.2, 0.4], [0.25, 0.5], [0.3, 0.6]]]'], {'dtype': 'np.float32'}), '([[[0.05, 0.1], [0.1, 0.2], [0.15, 0.3]], [[0.2, 0.4], [0.25, 0.5],\n [0.3, 0.6]]], dtype=np.float32)\n', (84894, 84997), True, 'import numpy as np\n'), ((97244, 97277), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_image_shape'], {}), '(in_image_shape)\n', (97261, 97277), True, 'import tensorflow as tf\n'), ((97295, 97328), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_masks_shape'], {}), '(in_masks_shape)\n', (97312, 97328), True, 'import tensorflow as tf\n'), ((97361, 97447), 'object_detection.tensorflow_detect.core.preprocessor.resize_image', 'preprocessor.resize_image', (['in_image', 'in_masks'], {'new_height': 'height', 'new_width': 'width'}), '(in_image, in_masks, new_height=height, new_width=\n width)\n', (97386, 97447), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((97478, 97497), 'tensorflow.shape', 'tf.shape', (['out_image'], {}), '(out_image)\n', (97486, 97497), True, 'import tensorflow as tf\n'), ((97522, 97541), 'tensorflow.shape', 'tf.shape', (['out_masks'], {}), '(out_masks)\n', (97530, 97541), True, 'import tensorflow as tf\n'), ((98584, 98617), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_image_shape'], {}), '(in_image_shape)\n', (98601, 98617), True, 'import tensorflow as tf\n'), ((98635, 98668), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_masks_shape'], {}), '(in_masks_shape)\n', (98652, 98668), True, 'import tensorflow as tf\n'), ((98701, 98787), 'object_detection.tensorflow_detect.core.preprocessor.resize_image', 'preprocessor.resize_image', (['in_image', 'in_masks'], {'new_height': 'height', 'new_width': 'width'}), '(in_image, in_masks, new_height=height, new_width=\n width)\n', (98726, 98787), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((98818, 98837), 'tensorflow.shape', 'tf.shape', (['out_image'], {}), '(out_image)\n', (98826, 98837), True, 'import tensorflow as tf\n'), ((98862, 98881), 'tensorflow.shape', 'tf.shape', (['out_masks'], {}), '(out_masks)\n', (98870, 98881), True, 'import tensorflow as tf\n'), ((99846, 99879), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_image_shape'], {}), '(in_image_shape)\n', (99863, 99879), True, 'import tensorflow as tf\n'), ((99897, 99930), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_masks_shape'], {}), '(in_masks_shape)\n', (99914, 99930), True, 'import tensorflow as tf\n'), ((99963, 100049), 'object_detection.tensorflow_detect.core.preprocessor.resize_image', 'preprocessor.resize_image', (['in_image', 'in_masks'], {'new_height': 'height', 'new_width': 'width'}), '(in_image, in_masks, new_height=height, new_width=\n width)\n', (99988, 100049), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((100080, 100099), 'tensorflow.shape', 'tf.shape', (['out_image'], {}), '(out_image)\n', (100088, 100099), True, 'import tensorflow as tf\n'), ((100124, 100143), 'tensorflow.shape', 'tf.shape', (['out_masks'], {}), '(out_masks)\n', (100132, 100143), True, 'import tensorflow as tf\n'), ((100791, 100818), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_shape'], {}), '(in_shape)\n', (100808, 100818), True, 'import tensorflow as tf\n'), ((100840, 100929), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_range', 'preprocessor.resize_to_range', (['in_image'], {'min_dimension': 'min_dim', 'max_dimension': 'max_dim'}), '(in_image, min_dimension=min_dim, max_dimension\n =max_dim)\n', (100868, 100929), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((101379, 101428), 'tensorflow.placeholder', 'tf.placeholder', (['tf.float32'], {'shape': '(None, None, 3)'}), '(tf.float32, shape=(None, None, 3))\n', (101393, 101428), True, 'import tensorflow as tf\n'), ((101450, 101539), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_range', 'preprocessor.resize_to_range', (['in_image'], {'min_dimension': 'min_dim', 'max_dimension': 'max_dim'}), '(in_image, min_dimension=min_dim, max_dimension\n =max_dim)\n', (101478, 101539), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((101570, 101589), 'tensorflow.shape', 'tf.shape', (['out_image'], {}), '(out_image)\n', (101578, 101589), True, 'import tensorflow as tf\n'), ((102209, 102258), 'tensorflow.placeholder', 'tf.placeholder', (['tf.float32'], {'shape': '(None, None, 3)'}), '(tf.float32, shape=(None, None, 3))\n', (102223, 102258), True, 'import tensorflow as tf\n'), ((102280, 102396), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_range', 'preprocessor.resize_to_range', (['in_image'], {'min_dimension': 'min_dim', 'max_dimension': 'max_dim', 'pad_to_max_dimension': '(True)'}), '(in_image, min_dimension=min_dim, max_dimension\n =max_dim, pad_to_max_dimension=True)\n', (102308, 102396), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((102526, 102545), 'tensorflow.shape', 'tf.shape', (['out_image'], {}), '(out_image)\n', (102534, 102545), True, 'import tensorflow as tf\n'), ((104249, 104282), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_image_shape'], {}), '(in_image_shape)\n', (104266, 104282), True, 'import tensorflow as tf\n'), ((104300, 104333), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_masks_shape'], {}), '(in_masks_shape)\n', (104317, 104333), True, 'import tensorflow as tf\n'), ((104366, 104464), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_range', 'preprocessor.resize_to_range', (['in_image', 'in_masks'], {'min_dimension': 'min_dim', 'max_dimension': 'max_dim'}), '(in_image, in_masks, min_dimension=min_dim,\n max_dimension=max_dim)\n', (104394, 104464), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((105262, 105311), 'tensorflow.placeholder', 'tf.placeholder', (['tf.float32'], {'shape': '(None, None, 3)'}), '(tf.float32, shape=(None, None, 3))\n', (105276, 105311), True, 'import tensorflow as tf\n'), ((105329, 105381), 'tensorflow.placeholder', 'tf.placeholder', (['tf.float32'], {'shape': '(None, None, None)'}), '(tf.float32, shape=(None, None, None))\n', (105343, 105381), True, 'import tensorflow as tf\n'), ((105414, 105539), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_range', 'preprocessor.resize_to_range', (['in_image', 'in_masks'], {'min_dimension': 'min_dim', 'max_dimension': 'max_dim', 'pad_to_max_dimension': '(True)'}), '(in_image, in_masks, min_dimension=min_dim,\n max_dimension=max_dim, pad_to_max_dimension=True)\n', (105442, 105539), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((105611, 105630), 'tensorflow.shape', 'tf.shape', (['out_image'], {}), '(out_image)\n', (105619, 105630), True, 'import tensorflow as tf\n'), ((105655, 105674), 'tensorflow.shape', 'tf.shape', (['out_masks'], {}), '(out_masks)\n', (105663, 105674), True, 'import tensorflow as tf\n'), ((106817, 106866), 'tensorflow.placeholder', 'tf.placeholder', (['tf.float32'], {'shape': '(None, None, 3)'}), '(tf.float32, shape=(None, None, 3))\n', (106831, 106866), True, 'import tensorflow as tf\n'), ((106884, 106936), 'tensorflow.placeholder', 'tf.placeholder', (['tf.float32'], {'shape': '(None, None, None)'}), '(tf.float32, shape=(None, None, None))\n', (106898, 106936), True, 'import tensorflow as tf\n'), ((106954, 106987), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_masks_shape'], {}), '(in_masks_shape)\n', (106971, 106987), True, 'import tensorflow as tf\n'), ((107020, 107118), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_range', 'preprocessor.resize_to_range', (['in_image', 'in_masks'], {'min_dimension': 'min_dim', 'max_dimension': 'max_dim'}), '(in_image, in_masks, min_dimension=min_dim,\n max_dimension=max_dim)\n', (107048, 107118), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((107150, 107169), 'tensorflow.shape', 'tf.shape', (['out_image'], {}), '(out_image)\n', (107158, 107169), True, 'import tensorflow as tf\n'), ((107194, 107213), 'tensorflow.shape', 'tf.shape', (['out_masks'], {}), '(out_masks)\n', (107202, 107213), True, 'import tensorflow as tf\n'), ((108354, 108387), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_image_shape'], {}), '(in_image_shape)\n', (108371, 108387), True, 'import tensorflow as tf\n'), ((108405, 108438), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_masks_shape'], {}), '(in_masks_shape)\n', (108422, 108438), True, 'import tensorflow as tf\n'), ((108471, 108569), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_range', 'preprocessor.resize_to_range', (['in_image', 'in_masks'], {'min_dimension': 'min_dim', 'max_dimension': 'max_dim'}), '(in_image, in_masks, min_dimension=min_dim,\n max_dimension=max_dim)\n', (108499, 108569), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((108601, 108620), 'tensorflow.shape', 'tf.shape', (['out_image'], {}), '(out_image)\n', (108609, 108620), True, 'import tensorflow as tf\n'), ((108645, 108664), 'tensorflow.shape', 'tf.shape', (['out_masks'], {}), '(out_masks)\n', (108653, 108664), True, 'import tensorflow as tf\n'), ((109079, 109124), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_range', 'preprocessor.resize_to_range', (['image', '(500)', '(600)'], {}), '(image, 500, 600)\n', (109107, 109124), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((109461, 109488), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_shape'], {}), '(in_shape)\n', (109478, 109488), True, 'import tensorflow as tf\n'), ((109510, 109599), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_range', 'preprocessor.resize_to_range', (['in_image'], {'min_dimension': 'min_dim', 'max_dimension': 'max_dim'}), '(in_image, min_dimension=min_dim, max_dimension\n =max_dim)\n', (109538, 109599), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((109630, 109649), 'tensorflow.shape', 'tf.shape', (['out_image'], {}), '(out_image)\n', (109638, 109649), True, 'import tensorflow as tf\n'), ((110428, 110477), 'tensorflow.placeholder', 'tf.placeholder', (['tf.float32'], {'shape': '(None, None, 3)'}), '(tf.float32, shape=(None, None, 3))\n', (110442, 110477), True, 'import tensorflow as tf\n'), ((110495, 110547), 'tensorflow.placeholder', 'tf.placeholder', (['tf.float32'], {'shape': '(None, None, None)'}), '(tf.float32, shape=(None, None, None))\n', (110509, 110547), True, 'import tensorflow as tf\n'), ((110565, 110598), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_masks_shape'], {}), '(in_masks_shape)\n', (110582, 110598), True, 'import tensorflow as tf\n'), ((110631, 110710), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_min_dimension', 'preprocessor.resize_to_min_dimension', (['in_image', 'in_masks'], {'min_dimension': 'min_dim'}), '(in_image, in_masks, min_dimension=min_dim)\n', (110667, 110710), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((110746, 110765), 'tensorflow.shape', 'tf.shape', (['out_image'], {}), '(out_image)\n', (110754, 110765), True, 'import tensorflow as tf\n'), ((110790, 110809), 'tensorflow.shape', 'tf.shape', (['out_masks'], {}), '(out_masks)\n', (110798, 110809), True, 'import tensorflow as tf\n'), ((111939, 111972), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_image_shape'], {}), '(in_image_shape)\n', (111956, 111972), True, 'import tensorflow as tf\n'), ((111990, 112023), 'tensorflow.random_uniform', 'tf.random_uniform', (['in_masks_shape'], {}), '(in_masks_shape)\n', (112007, 112023), True, 'import tensorflow as tf\n'), ((112056, 112135), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_min_dimension', 'preprocessor.resize_to_min_dimension', (['in_image', 'in_masks'], {'min_dimension': 'min_dim'}), '(in_image, in_masks, min_dimension=min_dim)\n', (112092, 112135), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((112171, 112190), 'tensorflow.shape', 'tf.shape', (['out_image'], {}), '(out_image)\n', (112179, 112190), True, 'import tensorflow as tf\n'), ((112215, 112234), 'tensorflow.shape', 'tf.shape', (['out_masks'], {}), '(out_masks)\n', (112223, 112234), True, 'import tensorflow as tf\n'), ((112663, 112711), 'object_detection.tensorflow_detect.core.preprocessor.resize_to_min_dimension', 'preprocessor.resize_to_min_dimension', (['image', '(500)'], {}), '(image, 500)\n', (112699, 112711), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((114310, 114333), 'tensorflow.zeros', 'tf.zeros', (['(240, 320, 3)'], {}), '((240, 320, 3))\n', (114318, 114333), True, 'import tensorflow as tf\n'), ((114373, 114427), 'object_detection.tensorflow_detect.core.preprocessor.subtract_channel_mean', 'preprocessor.subtract_channel_mean', (['image'], {'means': 'means'}), '(image, means=means)\n', (114407, 114427), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((114750, 114788), 'tensorflow.constant', 'tf.constant', (['[1, 4, 2]'], {'dtype': 'tf.int32'}), '([1, 4, 2], dtype=tf.int32)\n', (114761, 114788), True, 'import tensorflow as tf\n'), ((114805, 114857), 'object_detection.tensorflow_detect.core.preprocessor.one_hot_encoding', 'preprocessor.one_hot_encoding', (['labels'], {'num_classes': '(5)'}), '(labels, num_classes=5)\n', (114834, 114857), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((20585, 20674), 'object_detection.tensorflow_detect.core.preprocessor.preprocess', 'preprocessor.preprocess', (['tensor_dict', 'preprocess_options', 'preprocessor_arg_map', 'cache'], {}), '(tensor_dict, preprocess_options,\n preprocessor_arg_map, cache)\n', (20608, 20674), False, 'from object_detection.tensorflow_detect.core import standard_fields as fields, preprocessor, preprocessor_cache\n'), ((35770, 35789), 'tensorflow.to_float', 'tf.to_float', (['images'], {}), '(images)\n', (35781, 35789), True, 'import tensorflow as tf\n'), ((35821, 35840), 'tensorflow.to_float', 'tf.to_float', (['images'], {}), '(images)\n', (35832, 35840), True, 'import tensorflow as tf\n'), ((56058, 56098), 'tensorflow.constant', 'tf.constant', (['[6, 143, 0]'], {'dtype': 'tf.int32'}), '([6, 143, 0], dtype=tf.int32)\n', (56069, 56098), True, 'import tensorflow as tf\n'), ((56111, 56154), 'tensorflow.constant', 'tf.constant', (['[190, 237, -1]'], {'dtype': 'tf.int32'}), '([190, 237, -1], dtype=tf.int32)\n', (56122, 56154), True, 'import tensorflow as tf\n'), ((56171, 56232), 'tensorflow.constant', 'tf.constant', (['[[[0.03, 0.3575, 0.98, 0.95]]]'], {'dtype': 'tf.float32'}), '([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32)\n', (56182, 56232), True, 'import tensorflow as tf\n'), ((59987, 60027), 'tensorflow.constant', 'tf.constant', (['[6, 143, 0]'], {'dtype': 'tf.int32'}), '([6, 143, 0], dtype=tf.int32)\n', (59998, 60027), True, 'import tensorflow as tf\n'), ((60039, 60082), 'tensorflow.constant', 'tf.constant', (['[190, 237, -1]'], {'dtype': 'tf.int32'}), '([190, 237, -1], dtype=tf.int32)\n', (60050, 60082), True, 'import tensorflow as tf\n'), ((60094, 60155), 'tensorflow.constant', 'tf.constant', (['[[[0.03, 0.3575, 0.98, 0.95]]]'], {'dtype': 'tf.float32'}), '([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32)\n', (60105, 60155), True, 'import tensorflow as tf\n'), ((60550, 60654), 'numpy.array', 'np.array', (['[[0.0, 0.0, 0.75789469, 1.0], [0.23157893, 0.24050637, 0.75789469, 1.0]]'], {'dtype': 'np.float32'}), '([[0.0, 0.0, 0.75789469, 1.0], [0.23157893, 0.24050637, 0.75789469,\n 1.0]], dtype=np.float32)\n', (60558, 60654), True, 'import numpy as np\n'), ((61312, 61352), 'tensorflow.constant', 'tf.constant', (['[6, 143, 0]'], {'dtype': 'tf.int32'}), '([6, 143, 0], dtype=tf.int32)\n', (61323, 61352), True, 'import tensorflow as tf\n'), ((61364, 61407), 'tensorflow.constant', 'tf.constant', (['[190, 237, -1]'], {'dtype': 'tf.int32'}), '([190, 237, -1], dtype=tf.int32)\n', (61375, 61407), True, 'import tensorflow as tf\n'), ((61419, 61480), 'tensorflow.constant', 'tf.constant', (['[[[0.03, 0.3575, 0.98, 0.95]]]'], {'dtype': 'tf.float32'}), '([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32)\n', (61430, 61480), True, 'import tensorflow as tf\n'), ((61824, 61928), 'numpy.array', 'np.array', (['[[0.0, 0.0, 0.75789469, 1.0], [0.23157893, 0.24050637, 0.75789469, 1.0]]'], {'dtype': 'np.float32'}), '([[0.0, 0.0, 0.75789469, 1.0], [0.23157893, 0.24050637, 0.75789469,\n 1.0]], dtype=np.float32)\n', (61832, 61928), True, 'import numpy as np\n'), ((62572, 62612), 'tensorflow.constant', 'tf.constant', (['[6, 143, 0]'], {'dtype': 'tf.int32'}), '([6, 143, 0], dtype=tf.int32)\n', (62583, 62612), True, 'import tensorflow as tf\n'), ((62624, 62667), 'tensorflow.constant', 'tf.constant', (['[190, 237, -1]'], {'dtype': 'tf.int32'}), '([190, 237, -1], dtype=tf.int32)\n', (62635, 62667), True, 'import tensorflow as tf\n'), ((62679, 62740), 'tensorflow.constant', 'tf.constant', (['[[[0.03, 0.3575, 0.98, 0.95]]]'], {'dtype': 'tf.float32'}), '([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32)\n', (62690, 62740), True, 'import tensorflow as tf\n'), ((63105, 63209), 'numpy.array', 'np.array', (['[[0.0, 0.0, 0.75789469, 1.0], [0.23157893, 0.24050637, 0.75789469, 1.0]]'], {'dtype': 'np.float32'}), '([[0.0, 0.0, 0.75789469, 1.0], [0.23157893, 0.24050637, 0.75789469,\n 1.0]], dtype=np.float32)\n', (63113, 63209), True, 'import numpy as np\n'), ((63261, 63432), 'numpy.array', 'np.array', (['[[[np.nan, np.nan], [np.nan, np.nan], [np.nan, np.nan]], [[0.38947368, \n 0.07173], [0.49473682, 0.24050637], [0.60000002, 0.40928277]]]'], {'dtype': 'np.float32'}), '([[[np.nan, np.nan], [np.nan, np.nan], [np.nan, np.nan]], [[\n 0.38947368, 0.07173], [0.49473682, 0.24050637], [0.60000002, 0.40928277\n ]]], dtype=np.float32)\n', (63269, 63432), True, 'import numpy as np\n'), ((64604, 64644), 'tensorflow.constant', 'tf.constant', (['[6, 143, 0]'], {'dtype': 'tf.int32'}), '([6, 143, 0], dtype=tf.int32)\n', (64615, 64644), True, 'import tensorflow as tf\n'), ((64656, 64699), 'tensorflow.constant', 'tf.constant', (['[190, 237, -1]'], {'dtype': 'tf.int32'}), '([190, 237, -1], dtype=tf.int32)\n', (64667, 64699), True, 'import tensorflow as tf\n'), ((64711, 64772), 'tensorflow.constant', 'tf.constant', (['[[[0.03, 0.3575, 0.98, 0.95]]]'], {'dtype': 'tf.float32'}), '([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32)\n', (64722, 64772), True, 'import tensorflow as tf\n'), ((65562, 65666), 'numpy.array', 'np.array', (['[[0.0, 0.0, 0.75789469, 1.0], [0.23157893, 0.24050637, 0.75789469, 1.0]]'], {'dtype': 'np.float32'}), '([[0.0, 0.0, 0.75789469, 1.0], [0.23157893, 0.24050637, 0.75789469,\n 1.0]], dtype=np.float32)\n', (65570, 65666), True, 'import numpy as np\n'), ((66833, 66873), 'tensorflow.constant', 'tf.constant', (['[6, 143, 0]'], {'dtype': 'tf.int32'}), '([6, 143, 0], dtype=tf.int32)\n', (66844, 66873), True, 'import tensorflow as tf\n'), ((66885, 66928), 'tensorflow.constant', 'tf.constant', (['[190, 237, -1]'], {'dtype': 'tf.int32'}), '([190, 237, -1], dtype=tf.int32)\n', (66896, 66928), True, 'import tensorflow as tf\n'), ((66940, 67001), 'tensorflow.constant', 'tf.constant', (['[[[0.03, 0.3575, 0.98, 0.95]]]'], {'dtype': 'tf.float32'}), '([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32)\n', (66951, 67001), True, 'import tensorflow as tf\n'), ((67798, 67902), 'numpy.array', 'np.array', (['[[0.0, 0.0, 0.75789469, 1.0], [0.23157893, 0.24050637, 0.75789469, 1.0]]'], {'dtype': 'np.float32'}), '([[0.0, 0.0, 0.75789469, 1.0], [0.23157893, 0.24050637, 0.75789469,\n 1.0]], dtype=np.float32)\n', (67806, 67902), True, 'import numpy as np\n'), ((67963, 68137), 'numpy.array', 'np.array', (['[[[0.38947368, 0.07173], [0.49473682, 0.24050637], [0.60000002, 0.40928277]\n ], [[0.38947368, 0.07173], [0.49473682, 0.24050637], [0.60000002, \n 0.40928277]]]'], {}), '([[[0.38947368, 0.07173], [0.49473682, 0.24050637], [0.60000002, \n 0.40928277]], [[0.38947368, 0.07173], [0.49473682, 0.24050637], [\n 0.60000002, 0.40928277]]])\n', (67971, 68137), True, 'import numpy as np\n'), ((69387, 69427), 'tensorflow.constant', 'tf.constant', (['[6, 143, 0]'], {'dtype': 'tf.int32'}), '([6, 143, 0], dtype=tf.int32)\n', (69398, 69427), True, 'import tensorflow as tf\n'), ((69439, 69482), 'tensorflow.constant', 'tf.constant', (['[190, 237, -1]'], {'dtype': 'tf.int32'}), '([190, 237, -1], dtype=tf.int32)\n', (69450, 69482), True, 'import tensorflow as tf\n'), ((69494, 69555), 'tensorflow.constant', 'tf.constant', (['[[[0.03, 0.3575, 0.98, 0.95]]]'], {'dtype': 'tf.float32'}), '([[[0.03, 0.3575, 0.98, 0.95]]], dtype=tf.float32)\n', (69505, 69555), True, 'import tensorflow as tf\n'), ((70352, 70456), 'numpy.array', 'np.array', (['[[0.0, 0.0, 0.75789469, 1.0], [0.23157893, 0.24050637, 0.75789469, 1.0]]'], {'dtype': 'np.float32'}), '([[0.0, 0.0, 0.75789469, 1.0], [0.23157893, 0.24050637, 0.75789469,\n 1.0]], dtype=np.float32)\n', (70360, 70456), True, 'import numpy as np\n'), ((70517, 70643), 'numpy.array', 'np.array', (['[[[np.nan, np.nan], [np.nan, np.nan], [np.nan, np.nan]], [[np.nan, np.nan],\n [np.nan, np.nan], [np.nan, np.nan]]]'], {}), '([[[np.nan, np.nan], [np.nan, np.nan], [np.nan, np.nan]], [[np.nan,\n np.nan], [np.nan, np.nan], [np.nan, np.nan]]])\n', (70525, 70643), True, 'import numpy as np\n'), ((77150, 77199), 'numpy.array', 'np.array', (['[0.0, 0.5, 0.75, 1.0]'], {'dtype': 'np.float32'}), '([0.0, 0.5, 0.75, 1.0], dtype=np.float32)\n', (77158, 77199), True, 'import numpy as np\n'), ((79133, 79182), 'numpy.array', 'np.array', (['[0.0, 0.5, 0.75, 1.0]'], {'dtype': 'np.float32'}), '([0.0, 0.5, 0.75, 1.0], dtype=np.float32)\n', (79141, 79182), True, 'import numpy as np\n'), ((79212, 79276), 'numpy.array', 'np.array', (['[[0.1, 0.2], [0.2, 0.4], [0.3, 0.6]]'], {'dtype': 'np.float32'}), '([[0.1, 0.2], [0.2, 0.4], [0.3, 0.6]], dtype=np.float32)\n', (79220, 79276), True, 'import numpy as np\n'), ((87667, 87752), 'numpy.all', 'np.all', (['(boxes_[:, 2] - boxes_[:, 0] >= padded_boxes_[:, 2] - padded_boxes_[:, 0])'], {}), '(boxes_[:, 2] - boxes_[:, 0] >= padded_boxes_[:, 2] - padded_boxes_[:, 0]\n )\n', (87673, 87752), True, 'import numpy as np\n'), ((87786, 87871), 'numpy.all', 'np.all', (['(boxes_[:, 3] - boxes_[:, 1] >= padded_boxes_[:, 3] - padded_boxes_[:, 1])'], {}), '(boxes_[:, 3] - boxes_[:, 1] >= padded_boxes_[:, 3] - padded_boxes_[:, 1]\n )\n', (87792, 87871), True, 'import numpy as np\n'), ((90253, 90338), 'numpy.all', 'np.all', (['(boxes_[:, 2] - boxes_[:, 0] >= padded_boxes_[:, 2] - padded_boxes_[:, 0])'], {}), '(boxes_[:, 2] - boxes_[:, 0] >= padded_boxes_[:, 2] - padded_boxes_[:, 0]\n )\n', (90259, 90338), True, 'import numpy as np\n'), ((90372, 90457), 'numpy.all', 'np.all', (['(boxes_[:, 3] - boxes_[:, 1] >= padded_boxes_[:, 3] - padded_boxes_[:, 1])'], {}), '(boxes_[:, 3] - boxes_[:, 1] >= padded_boxes_[:, 3] - padded_boxes_[:, 1]\n )\n', (90378, 90457), True, 'import numpy as np\n'), ((101784, 101810), 'numpy.random.randn', 'np.random.randn', (['*in_shape'], {}), '(*in_shape)\n', (101799, 101810), True, 'import numpy as np\n'), ((102672, 102698), 'numpy.random.randn', 'np.random.randn', (['*in_shape'], {}), '(*in_shape)\n', (102687, 102698), True, 'import numpy as np\n'), ((105867, 105899), 'numpy.random.randn', 'np.random.randn', (['*in_image_shape'], {}), '(*in_image_shape)\n', (105882, 105899), True, 'import numpy as np\n'), ((105927, 105959), 'numpy.random.randn', 'np.random.randn', (['*in_masks_shape'], {}), '(*in_masks_shape)\n', (105942, 105959), True, 'import numpy as np\n'), ((107406, 107438), 'numpy.random.randn', 'np.random.randn', (['*in_image_shape'], {}), '(*in_image_shape)\n', (107421, 107438), True, 'import numpy as np\n'), ((107466, 107498), 'numpy.random.randn', 'np.random.randn', (['*in_masks_shape'], {}), '(*in_masks_shape)\n', (107481, 107498), True, 'import numpy as np\n'), ((111002, 111034), 'numpy.random.randn', 'np.random.randn', (['*in_image_shape'], {}), '(*in_image_shape)\n', (111017, 111034), True, 'import numpy as np\n'), ((111062, 111094), 'numpy.random.randn', 'np.random.randn', (['*in_masks_shape'], {}), '(*in_masks_shape)\n', (111077, 111094), True, 'import numpy as np\n')]
|
# -*- coding: utf-8 -*-
import logging
import utool as ut
import numpy as np
import vtool as vt
import pandas as pd
from wbia.algo.graph.nx_utils import ensure_multi_index
from wbia.algo.graph.state import POSTV, NEGTV, INCMP
print, rrr, profile = ut.inject2(__name__)
logger = logging.getLogger('wbia')
class Groundtruth(object):
def is_comparable(infr, aid_pairs, allow_guess=True):
"""
Guesses by default when real comparable information is not available.
"""
if infr.ibs is not None:
return infr.wbia_is_comparable(aid_pairs, allow_guess)
is_comp = list(
infr.gen_edge_values(
'gt_comparable', edges=aid_pairs, default=True, on_missing='default'
)
)
return np.array(is_comp)
def is_photobomb(infr, aid_pairs):
if infr.ibs is not None:
return infr.wbia_is_photobomb(aid_pairs)
return np.array([False] * len(aid_pairs))
def is_same(infr, aid_pairs):
if infr.ibs is not None:
return infr.wbia_is_same(aid_pairs)
node_dict = ut.nx_node_dict(infr.graph)
nid1 = [node_dict[n1]['orig_name_label'] for n1, n2 in aid_pairs]
nid2 = [node_dict[n2]['orig_name_label'] for n1, n2 in aid_pairs]
return np.equal(nid1, nid2)
def apply_edge_truth(infr, edges=None):
if edges is None:
edges = list(infr.edges())
edge_truth_df = infr.match_state_df(edges)
edge_truth = edge_truth_df.idxmax(axis=1).to_dict()
infr.set_edge_attrs('truth', edge_truth)
infr.edge_truth.update(edge_truth)
def match_state_df(infr, index):
""" Returns groundtruth state based on wbia controller """
index = ensure_multi_index(index, ('aid1', 'aid2'))
aid_pairs = np.asarray(index.tolist())
aid_pairs = vt.ensure_shape(aid_pairs, (None, 2))
is_same = infr.is_same(aid_pairs)
is_comp = infr.is_comparable(aid_pairs)
match_state_df = pd.DataFrame.from_dict(
dict(
[
(NEGTV, ~is_same & is_comp),
(POSTV, is_same & is_comp),
(INCMP, ~is_comp),
]
)
)
match_state_df.index = index
return match_state_df
def match_state_gt(infr, edge):
if edge in infr.edge_truth:
truth = infr.edge_truth[edge]
elif hasattr(infr, 'dummy_verif'):
truth = infr.dummy_verif._get_truth(edge)
else:
aid_pairs = np.asarray([edge])
is_same = infr.is_same(aid_pairs)[0]
is_comp = infr.is_comparable(aid_pairs)[0]
match_state = pd.Series(
dict(
[
(NEGTV, ~is_same & is_comp),
(POSTV, is_same & is_comp),
(INCMP, ~is_comp),
]
)
)
truth = match_state.idxmax()
return truth
def edge_attr_df(infr, key, edges=None, default=ut.NoParam):
""" constructs DataFrame using current predictions """
edge_states = infr.gen_edge_attrs(key, edges=edges, default=default)
edge_states = list(edge_states)
if isinstance(edges, pd.MultiIndex):
index = edges
else:
if edges is None:
edges_ = ut.take_column(edge_states, 0)
else:
edges_ = ut.lmap(tuple, ut.aslist(edges))
index = pd.MultiIndex.from_tuples(edges_, names=('aid1', 'aid2'))
records = ut.itake_column(edge_states, 1)
edge_df = pd.Series.from_array(records)
edge_df.name = key
edge_df.index = index
return edge_df
|
[
"logging.getLogger",
"utool.inject2",
"utool.itake_column",
"wbia.algo.graph.nx_utils.ensure_multi_index",
"numpy.asarray",
"numpy.equal",
"numpy.array",
"utool.take_column",
"vtool.ensure_shape",
"utool.nx_node_dict",
"pandas.MultiIndex.from_tuples",
"utool.aslist",
"pandas.Series.from_array"
] |
[((249, 269), 'utool.inject2', 'ut.inject2', (['__name__'], {}), '(__name__)\n', (259, 269), True, 'import utool as ut\n'), ((279, 304), 'logging.getLogger', 'logging.getLogger', (['"""wbia"""'], {}), "('wbia')\n", (296, 304), False, 'import logging\n'), ((776, 793), 'numpy.array', 'np.array', (['is_comp'], {}), '(is_comp)\n', (784, 793), True, 'import numpy as np\n'), ((1106, 1133), 'utool.nx_node_dict', 'ut.nx_node_dict', (['infr.graph'], {}), '(infr.graph)\n', (1121, 1133), True, 'import utool as ut\n'), ((1297, 1317), 'numpy.equal', 'np.equal', (['nid1', 'nid2'], {}), '(nid1, nid2)\n', (1305, 1317), True, 'import numpy as np\n'), ((1752, 1795), 'wbia.algo.graph.nx_utils.ensure_multi_index', 'ensure_multi_index', (['index', "('aid1', 'aid2')"], {}), "(index, ('aid1', 'aid2'))\n", (1770, 1795), False, 'from wbia.algo.graph.nx_utils import ensure_multi_index\n'), ((1863, 1900), 'vtool.ensure_shape', 'vt.ensure_shape', (['aid_pairs', '(None, 2)'], {}), '(aid_pairs, (None, 2))\n', (1878, 1900), True, 'import vtool as vt\n'), ((3628, 3659), 'utool.itake_column', 'ut.itake_column', (['edge_states', '(1)'], {}), '(edge_states, 1)\n', (3643, 3659), True, 'import utool as ut\n'), ((3678, 3707), 'pandas.Series.from_array', 'pd.Series.from_array', (['records'], {}), '(records)\n', (3698, 3707), True, 'import pandas as pd\n'), ((3552, 3609), 'pandas.MultiIndex.from_tuples', 'pd.MultiIndex.from_tuples', (['edges_'], {'names': "('aid1', 'aid2')"}), "(edges_, names=('aid1', 'aid2'))\n", (3577, 3609), True, 'import pandas as pd\n'), ((2571, 2589), 'numpy.asarray', 'np.asarray', (['[edge]'], {}), '([edge])\n', (2581, 2589), True, 'import numpy as np\n'), ((3425, 3455), 'utool.take_column', 'ut.take_column', (['edge_states', '(0)'], {}), '(edge_states, 0)\n', (3439, 3455), True, 'import utool as ut\n'), ((3514, 3530), 'utool.aslist', 'ut.aslist', (['edges'], {}), '(edges)\n', (3523, 3530), True, 'import utool as ut\n')]
|
from collections import OrderedDict
import numpy as np
from pypospack.qoi import Qoi
class ThermalExpansion(Qoi):
"""
Args:
temperature_min (float,int): beginning of the temperature range in Kelvin
temperature_max (float,int): end of the temperature range in Kelvin
temperature_step (float,int): increments of the temperature range in Kelvin
time_total (int): total simulation time in fs
time_step (int): simulation time step in fs
"""
def __init__(self,qoi_name,structures,
temperature_min=0,
temperature_max=2700,
temperature_step=100,
time_total=10,
time_step=0.001,
supercell=[5,5,5]):
_qoi_name = qoi_name
_qoi_type = 'lmps_thermal_expansion'
_structures = OrderedDict()
_structures['ideal'] = structures['ideal']
Qoi.__init__(self,
qoi_name=_qoi_name,
qoi_type=_qoi_type,
structures=_structures)
self.temperature_min = temperature_min
self.temperature_max = temperature_max
self.temperature_step = temperature_step
self.time_total=time_total
self.time_step=time_step
self.supercell = supercell
def determine_tasks(self):
T = self.temperature_min
while T <= self.temperature_max:
if T == 0:
_ideal_structure_name = self.structures['ideal']
_ideal_task_type = 'lmps_min_all'
_ideal_task_name = '{}.{}'.format(
_ideal_structure_name,
_ideal_task_type
)
_bulk_structure_name = None
self.add_task(
task_type=_ideal_task_type,
task_name=_ideal_task_name,
task_structure=_ideal_structure_name,
bulk_structure_name=_bulk_structure_name,
)
else:
_ideal_structure_name = self.structures['ideal']
_ideal_task_type = 'lmps_npt'.format(T)
_ideal_task_name = '{}.{}_{}'.format(
_ideal_structure_name,
_ideal_task_type,
T
)
_bulk_structure_name = None
self.add_task(
task_type=_ideal_task_type,
task_name=_ideal_task_name,
task_structure=_ideal_structure_name,
bulk_structure_name=_bulk_structure_name,
task_options={
'temperature':T,
'time_total':self.time_total,
'time_step':self.time_step,
'supercell':self.supercell}
)
T = T + self.temperature_step
def calculate_thermal_expansion_coefficient(self,temperatures,lattice_constants):
assert isinstance(temperatures,list)
assert isinstance(lattice_constants,list)
T = list(temperatures)
a0 = list(lattice_constants)
a0_at_0K = a0[0]
for i,v in enumerate(a0):
a0[i] = v/a0_at_0K-1
T = np.array(temperatures)
a0 = np.array(a0)
print(T)
print(a0)
T = T[:,np.newaxis] # T needs to be a column vector
# model is y = a*x
alpha_L,_,_,_ = np.linalg.lstsq(T,a0)
print('alpha_L:{}'.format(alpha_L[0]))
return alpha_L[0]
def calculate_qois(self,task_results):
_prefix = '{}.{}'.format(self.structures['ideal'],self.qoi_type)
s = self.structures['ideal']
T = self.temperature_min
lattice_constants = OrderedDict()
while T <= self.temperature_max:
if T == 0:
lattice_constants[T] = np.sqrt(
task_results["{}.lmps_min_all.a11".format(s)]**2 \
+ task_results['{}.lmps_min_all.a12'.format(s)]**2 \
+ task_results["{}.lmps_min_all.a13".format(s)]**2
)
else:
try:
lattice_constants[T] = task_results["{}.lmps_npt_{}.a1".format(s,T)]
except KeyError as e:
for k,v in task_results.items():
print(k,v)
raise
T = T + self.temperature_step
self.qois = OrderedDict()
# add lattice constants at different temperatures
for k,v in lattice_constants.items():
self.qois['{}.a0_{}'.format(_prefix,T)] = v
_temperatures = [k for k,v in lattice_constants.items()]
_lattice_constants = [v for k,v in lattice_constants.items()]
# add thermal expansion coefficient
self.qois['{}.thermal_expansion_coefficient'.format(_prefix)] = \
self.calculate_thermal_expansion_coefficient(
temperatures=_temperatures,
lattice_constants=_lattice_constants
)
|
[
"numpy.array",
"collections.OrderedDict",
"pypospack.qoi.Qoi.__init__",
"numpy.linalg.lstsq"
] |
[((819, 832), 'collections.OrderedDict', 'OrderedDict', ([], {}), '()\n', (830, 832), False, 'from collections import OrderedDict\n'), ((893, 980), 'pypospack.qoi.Qoi.__init__', 'Qoi.__init__', (['self'], {'qoi_name': '_qoi_name', 'qoi_type': '_qoi_type', 'structures': '_structures'}), '(self, qoi_name=_qoi_name, qoi_type=_qoi_type, structures=\n _structures)\n', (905, 980), False, 'from pypospack.qoi import Qoi\n'), ((3368, 3390), 'numpy.array', 'np.array', (['temperatures'], {}), '(temperatures)\n', (3376, 3390), True, 'import numpy as np\n'), ((3404, 3416), 'numpy.array', 'np.array', (['a0'], {}), '(a0)\n', (3412, 3416), True, 'import numpy as np\n'), ((3576, 3598), 'numpy.linalg.lstsq', 'np.linalg.lstsq', (['T', 'a0'], {}), '(T, a0)\n', (3591, 3598), True, 'import numpy as np\n'), ((3900, 3913), 'collections.OrderedDict', 'OrderedDict', ([], {}), '()\n', (3911, 3913), False, 'from collections import OrderedDict\n'), ((4604, 4617), 'collections.OrderedDict', 'OrderedDict', ([], {}), '()\n', (4615, 4617), False, 'from collections import OrderedDict\n')]
|
import sys
import math
import random
from collections import namedtuple
import time
from pyrf.util import (compute_usable_bins, adjust_usable_fstart_fstop,
trim_to_usable_fstart_fstop, find_saturation)
import numpy as np
from twisted.internet import defer
from pyrf.numpy_util import compute_fft
import struct
MAXIMUM_SPP = 32768
class correction_vector_acquire(object):
data_buffer = ""
v_type = "SIGNAL"
dut = None
complete_buffer = False
d = None
offset = 0
size = 0
transfer_size = 16*1024
def get_vector_loop(self, data):
self.data_buffer = b"".join([self.data_buffer, data])
self.offset += len(data)
if self.offset >= self.size:
# we have gotten all out data, return this object
if self.d is not None:
self.d.callback(self)
else:
# more data, grab another set of data
data1 = self.dut.correction_data(self.v_type, self.offset,
self.transfer_size)
# and add this function to the call back
data1.addCallback(self.get_vector_loop)
def get_vector_data(self, size):
# We got out size
if size is None:
# size is return None threw our created deffered in get_vector
if self.d is not None:
self.d.callback(None)
return
self.size = int(size)
if self.size == 0:
if self.d is not None:
self.d.callback(None)
return
if self.size < self.transfer_size:
self.transfer_size = self.size
# Grab our first set of data (deffered)
data = self.dut.correction_data(self.v_type, self.offset,
self.transfer_size)
# add the self.get_vector_loop call back
data.addCallback(self.get_vector_loop)
# what happens to error back here?
def error_b(self, failure):
if self.d is not None:
self.d.callback(None)
return None
def get_vector(self, v_type=None):
#
self.v_type = v_type
self.offset = 0
self.data_buffer = ""
# Create a defered
d = defer.Deferred()
self.d = d
# get our size (deffered)
size = self.dut.correction_size(self.v_type)
size.addCallback(self.get_vector_data)
size.addErrback(self.error_b)
# return our deferred
return d
class correction_vector(object):
correction_vectors = None
frequency_index = None
digest = None
def __init__(self):
self.frequency_index = []
self.dy = np.dtype(np.int32)
self.dy = self.dy.newbyteorder('>')
self.correction_vectors = {}
def _binary_search(self, freq):
# Simple binary search, modified to work the object's datastructure
lo = 0
hi = len(self.frequency_index)
while lo < hi:
mid = (lo + hi) // 2
if self.frequency_index[mid][0] * 1e3 < freq:
lo = mid + 1
else:
hi = mid
return lo
def _interp(self, in_array, number_of_points):
# array index of our orignal from 0 to size of vector - 1
x = np.arange(0.0, self.vector_size, 1.0)
# our new index
z = np.linspace(0.0, self.vector_size - 1, number_of_points)
# interpolate to get our new vector array
out_array = np.interp(z, x, in_array)
return out_array
def get_correction_vector(self, freq, number_of_points):
# binary search, retunrs our index
index = self._binary_search(freq)
# get the case where we go off the end
if index == len(self.frequency_index):
index = index - 1
# get our vector
vector = self.correction_vectors[self.frequency_index[index][1]]
# convert from micro db to db
vector = vector / 1000000.0
# interpolate our vector to the wanted size
resampled_vector = self._interp(vector, number_of_points)
return resampled_vector
def buffer_to_vector(self, buffer_in):
if buffer_in is None:
raise ValueError
if len(buffer_in) < 8 + 40:
raise ValueError
# Get the first 8 bytes
offset = 0
size = 8
input_buffer = buffer_in[offset:offset + size]
version, freq_num, vector_num, self.vector_size = struct.unpack("!HHHH", input_buffer)
offset = size
# Ignore the next 40 bytes, as not used know
offset += 40
# grab our frequency list
size = 6 * freq_num
input_buffer = buffer_in[offset:offset + size]
offset += size
if len(input_buffer) < size:
raise ValueError
# loop over our buffer, adding a frequency pair to the array
for i in range(freq_num):
freq, index = struct.unpack("!LH", input_buffer[i*6:i*6+6])
self.frequency_index.append([freq, index])
# grab our correction vectors
for i in range(vector_num):
# Grab out index
size = 2
input_buffer = buffer_in[offset:offset + size]
index = struct.unpack(">H", input_buffer)[0]
offset += size
# get our correction vector
size = 4 * self.vector_size
input_buffer = buffer_in[offset:offset + size]
micro_db = np.frombuffer(input_buffer, dtype=self.dy,
count=self.vector_size)
self.correction_vectors[index] = micro_db
offset += size
class SweepDeviceError(Exception):
"""
Exception for the sweep device to state an error() has occured
"""
pass
class SweepSettings(object):
"""
An object used to keep track of the sweep settings
"""
def __init__(self):
# start frequency of the results we will eventually return
self.bandstart = 0.0
# stop frequency of the results we will eventually return
self.bandstop = 0.0
# sweep entry's start frequency
self.fstart = 0.0
# sweep entry's stop frequency
self.fstop = 0.0
# sweep entry frequency step
self.fstep = 0.0
# sweep entry's RFE mode
self.rfe_mode = None
# determine if a second entry is required
self.dd_mode = False
# determines if a non dd entry is needed
self.beyond_dd = True
# entry attenuation
self.attenuation = 0
# entry ppb
self.ppb = 1
# sweep entry's spp
self.spp = 0.0
# sweep capture iterations
self.iterations = 0
# expected spectral points
self.spectral_points = 0
# determines if a sweep entry is required at the end
self.make_end_entry = False
# determine the frequency of the end entry
self.end_entry_freq = 0.0
# how many steps are in this sweep
self.step_count = 0
# what's the actual RBW of what we're capturing
self.rbw = 0
def __str__(self):
return "SweepSettings[ bandstart = %d, bandstop = %d, fstart = %d, fstop = %d, fstep = %d, step_count = %d, rfe_mode = %s, dd_mode = %s, beyond_dd = %s, attenuation = %s, ppb = %d, spp = %d, iterations = %d, spectral_points = %d, make_end_entry = %s, end_entry_freq = %d, rbw = %f ]" % (self.bandstart, self.bandstop, self.fstart, self.fstop, self.fstep, self.step_count, self.rfe_mode, self.dd_mode, self.beyond_dd, self.attenuation, self.ppb, self.spp, self.iterations, self.spectral_points, self.make_end_entry, self.end_entry_freq, self.rbw)
class SweepPlanner(object):
"""
An object that plans a sweep based on given paramaters.
:param dev_prop: the sweep device properties
:type dev_prop: dict
"""
def __init__(self, dev_prop):
self.dev_properties = dev_prop
self._prev_settings = SweepSettings()
def plan_sweep(self, fstart, fstop, rbw, mode, dev_settings = {}):
"""
Plan the sweep given the inputs
"""
# initialize the sweep settings variable
sweep_settings = SweepSettings()
# assign the sweep mode and start/stop
sweep_settings.rfe_mode = mode
sweep_settings.bandstart = fstart
sweep_settings.bandstop = fstop
if 'attenuator' in dev_settings:
sweep_settings.attenuation = dev_settings['attenuator']
# grab the usable bw of the current mode
usable_bw = self.dev_properties.USABLE_BW[mode]
# calculate the required SPP to get the RBW desired
sweep_settings.spp = self.dev_properties.FULL_BW[mode] / rbw
# find closest multiple of 32 because hardware
sweep_settings.spp = int(32 * round(float(sweep_settings.spp) / 32))
# double the points for SH/SHN mode
if mode in ['SH', 'SHN']:
sweep_settings.spp = sweep_settings.spp * 2
# if we're using zif mode, but we have a DD entry, we have half the SPP avaible, since DD is I-only and ZIF is IQ
if (mode == 'ZIF') and sweep_settings.dd_mode:
maxspp = self.dev_properties.MAX_SPP / 2
else:
maxspp = self.dev_properties.MAX_SPP
# adjust SPP if it's too big
sweep_settings.spp = min(maxspp, sweep_settings.spp)
# figure out our actual RBW (account for real vs complex data)
sweep_settings.rbw = self.dev_properties.FULL_BW[mode] / sweep_settings.spp
if not (mode == 'ZIF'):
sweep_settings.rbw = sweep_settings.rbw * 2
# make sure our result is atleast 1 RBW big
if (sweep_settings.bandstop - sweep_settings.bandstart) < sweep_settings.rbw:
fstop = sweep_settings.bandstart + sweep_settings.rbw
sweep_settings.bandstop = fstop
# change fstart and stop by a bit to account for floating point errors
# TODO: make this take into account tuning resolution
fstart -= sweep_settings.rbw * 4
fstop += sweep_settings.rbw * 4
# calculate fstart frequency
if fstart < self.dev_properties.MIN_TUNABLE[mode]:
sweep_settings.dd_mode = True
sweep_settings.fstart = self.dev_properties.MIN_TUNABLE[mode] + (usable_bw / 2)
sweep_settings.step_count += 1
# make sure we don't accidentally make an fstart that's beyond our tuning range
elif (fstart + (usable_bw / 2)) > self.dev_properties.MAX_TUNABLE[mode]:
sweep_settings.dd_mode = False
sweep_settings.fstart = self.dev_properties.MAX_TUNABLE[mode] - (usable_bw / 2)
else:
sweep_settings.dd_mode = False
sweep_settings.fstart = fstart + (usable_bw / 2)
# check if non-dd mode is required
if fstop <= self.dev_properties.MIN_TUNABLE[mode]:
sweep_settings.beyond_dd = False
else:
sweep_settings.beyond_dd = True
sweep_settings.step_count += 1
# assign the sweep entry's step frequency reducing by a couple rbw to account for floating point errors
# TODO: make this take into account tuning resolution
sweep_settings.fstep = usable_bw - (sweep_settings.rbw * 4)
# calculate the fstop of the sweep entry from fstart and how many usable_bw's we need
fspan = fstop - sweep_settings.fstart - sweep_settings.rbw
required_steps = round(fspan / sweep_settings.fstep)
sweep_settings.fstop = sweep_settings.fstart + (required_steps * sweep_settings.fstep)
sweep_settings.step_count += required_steps
# make sure fstop is lower than max tunable
# - it can sometimes be higher if an fstart is chosen, such that our
# fstep causes our fstop to go beyond fmax to cover all the band required
sweep_settings.make_end_entry = False
sweep_settings.end_entry_freq = 0
if sweep_settings.fstop > self.dev_properties.MAX_TUNABLE[mode]:
# go back one step
sweep_settings.fstop -= sweep_settings.fstep
# add an entry for fmax
sweep_settings.make_end_entry = True
sweep_settings.end_entry_freq = self.dev_properties.MAX_TUNABLE[mode] - (usable_bw / 2)
# calculate the expected number of spectral bins required for the SweepEntry
sweep_settings.spectral_points = int(round((sweep_settings.bandstop - sweep_settings.bandstart) / sweep_settings.rbw))
# return the sweep_settings
return sweep_settings
class SweepDevice(object):
"""
Virtual device that generates power spectrum from a given frequency range
by sweeping the frequencies with a real device and piecing together the FFT results.
:param real_device: the RF device that will be used for capturing data,
typically a :class:`pyrf.devices.thinkrf.WSA` instance.
:param async_callback: a callback to use for async operation (not used if
*real_device* is using a blocking :class:`PlainSocketConnector`)
"""
# keep track of the mode
rfe_mode = None
# keep track of the fstart/fstop and rbw
fstart = None
fstop = None
rbw = None
# keep track of non-standard device settings
device_settings = None
# keep track of whether DD mode is needed
dd_mode = False
# keep track of the sweep settings
_sweep_settings = None
# keep track of the packet count
packet_count = 0
# determine if a new entry is required
_new_entry = True
# array to place spectral data
spectral_data = []
capture_count = 0
sp_corr_obj = None
nf_corr_obj = None
_flattening_enabled = True
def __init__(self, real_device, async_callback=None):
# init log string
self.logstr = ''
self.logtype = 'NONE'
# initialize the real device
self.real_device = real_device
# request read permission from device
self.real_device.request_read_perm()
# keep track of the device properties
self.dev_properties = self.real_device.properties
# initialize the geolocation callback
self._geo_callback_func = None
self._geo_callback_data = None
# initialize the sweep planner
self._sweep_planner = SweepPlanner(self.dev_properties)
# make sure user passes async callback if the device has async connector
if real_device.async_connector():
if not async_callback:
raise SweepDeviceError(
"async_callback required for async operation")
# disable receiving data until we are expecting it
real_device.set_async_callback(None)
# Function to be called when async data is done capturing
def _save_correction_vector(data_buffer):
if data_buffer is None:
return None
try:
if data_buffer.v_type == "SIGNAL":
self.sp_corr_obj = correction_vector()
self.sp_corr_obj.buffer_to_vector(data_buffer.data_buffer)
elif data_buffer.v_type == "NOISE":
self.nf_corr_obj = correction_vector()
self.nf_corr_obj.buffer_to_vector(data_buffer.data_buffer)
except AttributeError:
if data_buffer.v_type == "SIGNAL":
self.sp_corr_obj = None
elif data_buffer.v_type == "NOISE":
self.nf_corr_obj = None
# function to catch the errback of the async code. Used to handle
# the case when we can get the correction vectors.
def _catch_timeout(failure):
failure.trap(IOError)
return None
vector_obj = correction_vector_acquire()
vector_obj.dut = real_device
vector_obj1 = correction_vector_acquire()
vector_obj1.dut = real_device
d1 = vector_obj.get_vector("NOISE")
d1.addCallback(_save_correction_vector)
d1.addErrback(_catch_timeout)
d2 = vector_obj1.get_vector("SIGNAL")
d2.addCallback(_save_correction_vector)
d2.addErrback(_catch_timeout)
else:
# make sure user doesnt pass async callback if the connector uses blocking sockets
if async_callback:
raise SweepDeviceError(
"async_callback not applicable for sync operation")
def _get_correction(dut, v_type=None):
if v_type.upper() == "SIGNAL" or v_type.upper() == "NOISE":
v_type = v_type.upper()
else:
raise ValueError
max_buf_size = 16*1024
offset = 0
bin_data = ""
try:
signal_size = dut.correction_size(v_type)
except (IOError, OSError): # this will handle socket.error's
raise ValueError
# We have nothing to transfer
if signal_size == 0:
return None
# check to see if tere is more data than can be transfer in one
# go
if signal_size > max_buf_size:
# if so transfer our max buffer size
transfer_size = max_buf_size
else:
# if not grab only what we need
transfer_size = signal_size
# While we still have data remaining
while offset < signal_size:
# get the data
data_buffer = dut.correction_data(v_type, offset,
transfer_size)
# figure out how many bytes were transfered
transfered = len(data_buffer)
# append the data to the buffer of what we have allready
# got
bin_data = b"".join([bin_data, data_buffer])
# increase the offset
offset = offset + transfered
return bin_data
self.sp_corr_obj = correction_vector()
try:
self.sp_corr_obj.buffer_to_vector(_get_correction(self.real_device, "SIGNAL"))
except ValueError:
self.sp_corr_obj = None
self.nf_corr_obj = correction_vector()
try:
self.nf_corr_obj.buffer_to_vector(_get_correction(self.real_device, "NOISE"))
except ValueError:
self.nf_corr_obj = None
self.async_callback = async_callback
self.continuous = False
# init the sweep id
self._next_sweep_id = 0
# init last finished (technically, it hasn't finished, but for our purposes, it has)
self._last_finished = True
# Private function
def log(self, firstmsg, *msgs):
if self.logtype == 'LOG':
self.logstr += firstmsg.__str__()
for msg in msgs:
self.logstr += ", "
self.logstr += msg.__str__()
self.logstr += "\n"
elif self.logtype == 'PRINT':
sys.stdout.write(firstmsg.__str__())
for msg in msgs:
sys.stdout.write(", ")
sys.stdout.write(msg.__str__())
sys.stdout.write("\n")
def enable_flattening(self, enable=None):
"""
:param enable: enable or disable spectral flattening
:type enable: bool or None
"""
if enable is None:
return self._flattening_enabled
else:
self._flattening_enabled = enable
def set_geolocation_callback(self, func, data = None):
"""
set a callback that will get called whenever the geolocation information
of the device is updated.
The callback function should accept two parameters. The first parameter
will be the callback data that was passed in this function
set_geolocation_callback(func, data, geolocation_dictionary).
The geolocation_dictionary will have the following properties:
- oui
- seconds
- altitude
- longitude
- speedoverground
- secondsfractional
- track
- latitude
- magneticvariation
- heading
See the programmer's guide for usage on each of these properties.
:param func: the function to be called
:param data: the data to be passed to the function
:returns: None
"""
self._geo_callback_func = func
self._geo_callback_data = data
def capture_power_spectrum(self,
fstart,
fstop,
rbw,
device_settings=None,
mode='SH',
continuous=False):
"""
Initiate a data capture from the *real_device* by setting up a sweep list
and starting a single sweep, and then return power spectral density data
along with the **actual** sweep start and stop frequencies set (which
might not be exactly the same as the requested *fstart* and *fstop*).
.. note:: This function does not pipeline, and if the last sweep isn't received before starting a new one, it will generate a failure.
:param int fstart: sweep starting frequency in Hz
:param int fstop: sweep ending frequency in Hz
:param float rbw: the resolution bandwidth (RBW) in Hz of the data to be captured (output RBW may be smaller than requested)
:param device_settings: attenuation and other device settings
:type device_settings: dict
:param str mode: sweep mode, 'ZIF', 'SH', or 'SHN'
:param bool continuous: set sweep to be continuously or not (once only)
:returns: fstart, fstop, power_data
"""
self.log("- capture_power_spectrum", fstart, fstop, rbw, device_settings, mode, continuous)
if continuous and not self.async_callback:
raise SweepDeviceError(
"continuous mode only applies to async operation")
# see if the last sweep has finished
if not self._last_finished:
raise SweepDeviceError(
"previous sweep must have finished before starting a new one")
self._last_finished = False
# increment the sweep id
if self._next_sweep_id < 0x00000000ffffffff:
self._next_sweep_id += 1
else:
self._next_sweep_id = 0
# keep track if this is a continuous sweep
self.continuous = continuous
# plan the sweep
self._sweep_planner = SweepPlanner(self.dev_properties)
self._sweep_settings = self._sweep_planner.plan_sweep(fstart, fstop, rbw, mode, device_settings)
self.log("self._sweep_settings = %s" % self._sweep_settings)
# remember our last sweep for optimization purposes
self._last_sweep = (fstart, fstop, rbw, mode, device_settings, continuous)
# configure the device with the sweep_settings
self.real_device.sweep_clear()
self.real_device.sweep_add(self._sweep_settings)
# configure the iteration
self.real_device.sweep_iterations(1)
# capture the sweep data
return self._perform_full_sweep()
def _perform_full_sweep(self):
# perform the sweep using async socket
if self.async_callback:
# set the async callback
self.real_device.set_async_callback(self._vrt_receive)
# start the sweep sequence
self._start_sweep()
return
# perform sweep using blocking sockets
self._start_sweep()
result = None
while result is None:
result = self._vrt_receive(self.real_device.read())
return result
def _start_sweep(self):
self._vrt_context = {}
# initialize the array we'll use to hold results
self.spectral_data = np.zeros(self._sweep_settings.spectral_points)
# keep track of packets recieved
self.packet_count = 0
self.real_device.sweep_start(self._next_sweep_id)
def _vrt_receive(self, packet):
# context packet just update our context dictionary
if packet.is_context_packet():
# look for any geolocation info
geo = { }
for field in [ 'latitude', 'longitude', 'altitude', 'speedoverground', 'heading', 'track', 'magneticvariation' ]:
if field in packet.fields:
geo[field] = packet.fields[field]
if geo and self._geo_callback_func:
# execute callback
func = self._geo_callback_func
func(self._geo_callback_data, geo)
self._vrt_context.update(packet.fields)
self.log(packet)
return
# check to see if we recieved our sweep ID
if not ('sweepid' in self._vrt_context):
return
# make sure we are receiving packets for the right sweep
if not (self._vrt_context['sweepid'] == self._next_sweep_id):
raise SweepDeviceError("data packets received before start of sweep received! cur = %d, next = %d" % (self._vrt_context['sweepid'], self._next_sweep_id))
# increment the packet count
self.packet_count += 1
self.log("#%d of %d - %s" % (self.packet_count, self._sweep_settings.step_count, packet))
# retrieve the frequency and usable BW of the packet
packet_freq = self._vrt_context['rffreq']
usable_bw = self.dev_properties.USABLE_BW[self._sweep_settings.rfe_mode]
# compute the fft
pow_data = compute_fft(self.real_device, packet, self._vrt_context)
# calc rbw for this packet
rbw = float(self.dev_properties.FULL_BW[self._sweep_settings.rfe_mode]) / len(pow_data)
self.log("rbw = %f, %f" % (rbw, self._sweep_settings.rbw))
if self._flattening_enabled:
# Check if we are above 50 MHz and in SH mode
if packet_freq >= 50e6 and self._sweep_settings.rfe_mode == "SH":
number_of_points = len(pow_data)
# check if we have correction vectors (Noise)
if self.nf_corr_obj is not None:
# if so grab them
nf_cal = \
self.nf_corr_obj.get_correction_vector(packet_freq,
number_of_points)
else:
# if no set it to 0
nf_cal = np.zeros(number_of_points)
# check if we have corrrection vectors (Spectrum)
if self.sp_corr_obj is not None:
# if so grab them
sp_cal = \
self.sp_corr_obj.get_correction_vector(packet_freq,
number_of_points)
else:
# if not set it to 0
sp_cal = np.zeros(number_of_points)
# if the data is spectraly inverted, invert the vectors
if packet.spec_inv:
nf_cal = np.flipud(nf_cal)
sp_cal = np.flipud(sp_cal)
# calculate the correction threshold
correction_thresh = (-135.0 + ((10.0 * packet_freq / 1e6)
/ 27000.0) + 10.0
* np.log10(rbw)
+ self._sweep_settings.attenuation)
# creat the spectrum. per bin, if the ampltitude is above
# correction threshold do pow_data - sp_cal else do pow_data -
# nf_cal
pow_data = np.where(pow_data < correction_thresh,
pow_data - nf_cal, pow_data - sp_cal)
# check if DD mode was used in this sweep
if self.packet_count == 1 and self._sweep_settings.dd_mode:
# copy the data into the result array
self._copy_data(0, self.dev_properties.FULL_BW['DD'], pow_data, self._sweep_settings.bandstart, self._sweep_settings.bandstop, self.spectral_data);
if self._sweep_settings.beyond_dd:
return
else:
return self._emit_data()
# determine the usable bins in this config
self.log("===> compute_usable_bins()", self._sweep_settings.rfe_mode, self._sweep_settings.spp, 1, 0)
usable_bins = compute_usable_bins(self.dev_properties,
self._sweep_settings.rfe_mode,
self._sweep_settings.spp,
1,
0)
self.log("<--- usable_bins", usable_bins)
# adjust the usable range based on spectral inversion
self.log("===> adjust_usable_fstart_fstop()", "self.dev_properties", self._sweep_settings.rfe_mode, len(pow_data) * 2, 1, packet_freq, packet.spec_inv, usable_bins)
usable_bins, packet_start, packet_stop = adjust_usable_fstart_fstop(self.dev_properties,
self._sweep_settings.rfe_mode,
len(pow_data) * 2,
1,
packet_freq,
packet.spec_inv,
usable_bins)
self.log("<--- adjust_usable_fstart_fstop", packet_start, packet_stop, usable_bins)
#
# WARNING: the start and stop returned from this function are HIGHLY sketchy
#
# calculate packet frequency range
#packet_start = packet_freq - (self.dev_properties.FULL_BW[self._sweep_settings.rfe_mode] / 2)
#packet_stop = packet_freq + (self.dev_properties.FULL_BW[self._sweep_settings.rfe_mode] / 2)
#print "packet start/stop", packet_start, packet_stop
#trim the FFT data, note decimation is 1, fshift is 0
self.log("===> trim_to_usable_fstart_fstop()", "pow_data", usable_bins, packet_start, packet_stop)
trimmed_spectrum, edge_data, usable_start, usable_stop = trim_to_usable_fstart_fstop(pow_data,
usable_bins,
packet_start,
packet_stop)
self.log("<--- trim_to_usable_fstart_fstop", usable_start, usable_stop, "trimmed_spectrum", edge_data)
# copy the data
self._copy_data(usable_start, usable_stop, trimmed_spectrum, self._sweep_settings.bandstart, self._sweep_settings.bandstop, self.spectral_data);
# if there's no more packets, emit result
if self.packet_count == self._sweep_settings.step_count:
return self._emit_data()
# all done
return
def _emit_data(self):
# note that we finished this sweep
self._last_finished = True
# if async callback is available, emit the data
if self.async_callback:
self.async_callback(self._sweep_settings.bandstart, self._sweep_settings.bandstop, self.spectral_data)
return
# return the values if using blocking sockets
else:
return (self._sweep_settings.bandstart, self._sweep_settings.bandstop, self.spectral_data)
def _copy_data(self, src_fstart, src_fstop, src_psd, dst_fstart, dst_fstop, dst_psd):
self.log("_copy_data(%d, %d, src_psd, %d, %d, dst_psd)" % (src_fstart, src_fstop, dst_fstart, dst_fstop))
# calc src len and dst len
srclen = len(src_psd)
dstlen = len(dst_psd)
self.log("len -- src = %d, dst = %d" % (srclen, dstlen))
# calc src and dest rbw
srcrbw = float(src_fstop - src_fstart) / srclen
dstrbw = float(dst_fstop - dst_fstart) / dstlen
self.log("rbw = %f, %f, %f" % (srcrbw, dstrbw, self._sweep_settings.rbw))
# check if packet start is before sweep start. shouldn't happen, but check anyway
self.log("boundary(start) = %f / %f" % (src_fstart, dst_fstart))
if src_fstart < dst_fstart:
self.log("foo")
src_start_bin = int(float(dst_fstart - src_fstart) / srcrbw)
else:
self.log("bar")
src_start_bin = 0
# check if packet stop is after sweep stop. this means we don't need the whole packet
self.log("boundary(stop) = %f / %f" % (src_fstop, dst_fstop))
if src_fstop > dst_fstop:
self.log("foo")
src_stop_bin = srclen - int(float(src_fstop - dst_fstop) / srcrbw)
else:
self.log("bar")
src_stop_bin = srclen
# how many values are we copying?
tocopy = src_stop_bin - src_start_bin
# calculate dest start index
if src_fstart < dst_fstart:
dst_start_bin = 0
else:
dst_start_bin = int(round(float(src_fstart - dst_fstart) / dstrbw))
# calculate dest stop index
dst_stop_bin = dst_start_bin + tocopy
if dst_stop_bin > dstlen:
dst_stop_bin = dstlen
# adjust tocopy
tocopy = dst_stop_bin - dst_start_bin
# adjust src stop bin because we adjusted tocopy
src_stop_bin = src_start_bin + tocopy
# copy the data, if there's data that needs copying
if ((dst_stop_bin - dst_start_bin) > 0) and ((src_stop_bin - src_start_bin) > 0):
self.log("dst_psd[%d:%d] = src_psd[%d:%d]" % (dst_start_bin, dst_stop_bin, src_start_bin, src_stop_bin))
dst_psd[dst_start_bin:dst_stop_bin] = src_psd[src_start_bin:src_stop_bin]
|
[
"numpy.log10",
"numpy.arange",
"pyrf.util.compute_usable_bins",
"numpy.where",
"numpy.flipud",
"pyrf.util.trim_to_usable_fstart_fstop",
"pyrf.numpy_util.compute_fft",
"numpy.linspace",
"struct.unpack",
"numpy.zeros",
"numpy.interp",
"numpy.frombuffer",
"numpy.dtype",
"twisted.internet.defer.Deferred",
"sys.stdout.write"
] |
[((2239, 2255), 'twisted.internet.defer.Deferred', 'defer.Deferred', ([], {}), '()\n', (2253, 2255), False, 'from twisted.internet import defer\n'), ((2682, 2700), 'numpy.dtype', 'np.dtype', (['np.int32'], {}), '(np.int32)\n', (2690, 2700), True, 'import numpy as np\n'), ((3283, 3320), 'numpy.arange', 'np.arange', (['(0.0)', 'self.vector_size', '(1.0)'], {}), '(0.0, self.vector_size, 1.0)\n', (3292, 3320), True, 'import numpy as np\n'), ((3357, 3413), 'numpy.linspace', 'np.linspace', (['(0.0)', '(self.vector_size - 1)', 'number_of_points'], {}), '(0.0, self.vector_size - 1, number_of_points)\n', (3368, 3413), True, 'import numpy as np\n'), ((3484, 3509), 'numpy.interp', 'np.interp', (['z', 'x', 'in_array'], {}), '(z, x, in_array)\n', (3493, 3509), True, 'import numpy as np\n'), ((4480, 4516), 'struct.unpack', 'struct.unpack', (['"""!HHHH"""', 'input_buffer'], {}), "('!HHHH', input_buffer)\n", (4493, 4516), False, 'import struct\n'), ((24410, 24456), 'numpy.zeros', 'np.zeros', (['self._sweep_settings.spectral_points'], {}), '(self._sweep_settings.spectral_points)\n', (24418, 24456), True, 'import numpy as np\n'), ((26125, 26181), 'pyrf.numpy_util.compute_fft', 'compute_fft', (['self.real_device', 'packet', 'self._vrt_context'], {}), '(self.real_device, packet, self._vrt_context)\n', (26136, 26181), False, 'from pyrf.numpy_util import compute_fft\n'), ((29019, 29126), 'pyrf.util.compute_usable_bins', 'compute_usable_bins', (['self.dev_properties', 'self._sweep_settings.rfe_mode', 'self._sweep_settings.spp', '(1)', '(0)'], {}), '(self.dev_properties, self._sweep_settings.rfe_mode,\n self._sweep_settings.spp, 1, 0)\n', (29038, 29126), False, 'from pyrf.util import compute_usable_bins, adjust_usable_fstart_fstop, trim_to_usable_fstart_fstop, find_saturation\n'), ((30885, 30962), 'pyrf.util.trim_to_usable_fstart_fstop', 'trim_to_usable_fstart_fstop', (['pow_data', 'usable_bins', 'packet_start', 'packet_stop'], {}), '(pow_data, usable_bins, packet_start, packet_stop)\n', (30912, 30962), False, 'from pyrf.util import compute_usable_bins, adjust_usable_fstart_fstop, trim_to_usable_fstart_fstop, find_saturation\n'), ((4951, 5002), 'struct.unpack', 'struct.unpack', (['"""!LH"""', 'input_buffer[i * 6:i * 6 + 6]'], {}), "('!LH', input_buffer[i * 6:i * 6 + 6])\n", (4964, 5002), False, 'import struct\n'), ((5484, 5550), 'numpy.frombuffer', 'np.frombuffer', (['input_buffer'], {'dtype': 'self.dy', 'count': 'self.vector_size'}), '(input_buffer, dtype=self.dy, count=self.vector_size)\n', (5497, 5550), True, 'import numpy as np\n'), ((5257, 5290), 'struct.unpack', 'struct.unpack', (['""">H"""', 'input_buffer'], {}), "('>H', input_buffer)\n", (5270, 5290), False, 'import struct\n'), ((19649, 19671), 'sys.stdout.write', 'sys.stdout.write', (['"""\n"""'], {}), "('\\n')\n", (19665, 19671), False, 'import sys\n'), ((28262, 28338), 'numpy.where', 'np.where', (['(pow_data < correction_thresh)', '(pow_data - nf_cal)', '(pow_data - sp_cal)'], {}), '(pow_data < correction_thresh, pow_data - nf_cal, pow_data - sp_cal)\n', (28270, 28338), True, 'import numpy as np\n'), ((19566, 19588), 'sys.stdout.write', 'sys.stdout.write', (['""", """'], {}), "(', ')\n", (19582, 19588), False, 'import sys\n'), ((27039, 27065), 'numpy.zeros', 'np.zeros', (['number_of_points'], {}), '(number_of_points)\n', (27047, 27065), True, 'import numpy as np\n'), ((27508, 27534), 'numpy.zeros', 'np.zeros', (['number_of_points'], {}), '(number_of_points)\n', (27516, 27534), True, 'import numpy as np\n'), ((27673, 27690), 'numpy.flipud', 'np.flipud', (['nf_cal'], {}), '(nf_cal)\n', (27682, 27690), True, 'import numpy as np\n'), ((27720, 27737), 'numpy.flipud', 'np.flipud', (['sp_cal'], {}), '(sp_cal)\n', (27729, 27737), True, 'import numpy as np\n'), ((27970, 27983), 'numpy.log10', 'np.log10', (['rbw'], {}), '(rbw)\n', (27978, 27983), True, 'import numpy as np\n')]
|
# Import libraries and set random seeds for reproducibility
random_seed = 1237
import random
random.seed( random_seed )
import numpy as np
np.random.seed( random_seed )
import tensorflow as tf
tf.set_random_seed( random_seed )
# Import model and instance loader
import model
from instance_loader import InstanceLoader
import os, sys, itertools, util
from scipy import stats
METRICS_LIST = [ "ABS", "REL", "KENDALLTAUB", "KENDALLTAUB_P", "PEARSON", "PEARSON_P" ]
def get_metrics_from_batch( predictions_list, labels_list ):
"""
Gets all the metrics from a batch for a specific centrality
"""
bABS, bREL, bKENDALLTAUB, bKENDALLTAUB_P, bPEARSON, bPEARSON_P = (0 for _ in METRICS_LIST)
for predictions, labels in zip( predictions_list, labels_list ):
ABS, REL, KENDALLTAUB, KENDALLTAUB_P, PEARSON, PEARSON_P = get_metrics( predictions, labels )
bABS, bREL, bKENDALLTAUB, bKENDALLTAUB_P, bPEARSON, bPEARSON_P = map(
lambda pair: pair[0] + pair[1],
zip(
(bABS, bREL, bKENDALLTAUB, bKENDALLTAUB_P, bPEARSON, bPEARSON_P),
( ABS, REL, KENDALLTAUB, KENDALLTAUB_P, PEARSON, PEARSON_P)
)
)
#end for
b = len( labels_list )
bABS, bREL, bKENDALLTAUB, bKENDALLTAUB_P, bPEARSON, bPEARSON_P = map(
lambda x: x / b,
(bABS, bREL, bKENDALLTAUB, bKENDALLTAUB_P, bPEARSON, bPEARSON_P)
)
return bABS, bREL, bKENDALLTAUB, bKENDALLTAUB_P, bPEARSON, bPEARSON_P
#end get_metrics_batch
def get_metrics( predictions, labels ):
"""
Gets all the metrics for a specific centrality
"""
ABS = [ abs( p - l ) for p, l in zip( predictions, labels ) ]
ABS = sum(ABS) / len(ABS)
REL = [ abs( p - l ) / l for p, l in zip( predictions, labels ) if l != 0 ]
REL = sum(REL) / len(REL)
KENDALLTAUB, KENDALLTAUB_P = stats.kendalltau(predictions,labels)
PEARSON, PEARSON_P = stats.pearsonr(predictions,labels)
return ABS, REL, KENDALLTAUB, KENDALLTAUB_P, PEARSON, PEARSON_P
#end
def build_metrics_dict( centralities, header = False, header_prepend = "" ):
"""
Builds the dictionary used to log the values or the dictionary containing the headers
"""
metrics_dict = dict()
for metric in METRICS_LIST:
metrics_dict[metric] = header_prepend + metric if header else 0
for centrality in centralities:
centrality_metric = "{c}_{m}".format(c=centrality,m=metric)
metrics_dict[centrality_metric] = header_prepend + centrality_metric if header else 0
#end for
#end for
metrics_dict["loss"] = header_prepend + "loss" if header else 0
for centrality in centralities:
centrality_cost = "{c}_cost".format(c=centrality)
metrics_dict[centrality_cost] = header_prepend + centrality_cost if header else 0
#end for
return metrics_dict
#end build_metrics_dict
def log_metrics_dict( metrics_dict, centralities, log_file ):
"""
Log a dictionary to a file.
Note that it assumes one will write at least some value before the values being logged in the file and it also doesn't end a line.
"""
print(
"\t{val}".format(
val = metrics_dict["loss"]
),
end = "",
file = log_file
)
for centrality in centralities:
print(
"\t{val}".format(
val = metrics_dict["{c}_cost".format(c=centrality)]
),
end = "",
file = log_file
)
#end for
for metric in METRICS_LIST:
print(
"\t{val}".format(
val = metrics_dict[metric]
),
end = "",
file = log_file
)
#end for
for centrality in centralities:
for metric in METRICS_LIST:
print(
"\t{val}".format(
val = metrics_dict["{c}_{m}".format(c=centrality,m=metric)]
),
end = "",
file = log_file
)
#end for
#end for
#end log_metrics_dict
def train(
session,
model_dict,
time_steps,
centralities,
epochs_to_run,
train_instance_loader,
batch_size,
test_batch_size,
batches_per_epoch,
test_instance_loader,
epoch_logging_file,
batch_logging_file,
model_checkpoint_filename,
log_to_stdout = False
):
"""
Runs the training procedure, logs the metrics and saves the model's weights to a checkpoint after every epoch.
"""
log_epoch( "epoch_id", build_metrics_dict( centralities, header = True, header_prepend = "train_" ), build_metrics_dict( centralities, header = True, header_prepend = "test_" ), centralities, epoch_logging_file )
log_batch( "epoch_id", "batch_id", build_metrics_dict( centralities, header = True ), centralities, batch_logging_file )
print( "Starting training for {} epochs".format( epochs_to_run ) )
for epoch_id in range( epochs_to_run ):
if log_to_stdout:
print( "Epoch\t{}".format( epoch_id ), end = "", file = sys.stdout )
log_metrics_dict( build_metrics_dict( centralities, header = True ), centralities, sys.stdout )
print( "", flush = True, file = sys.stdout )
#end if
run_epoch(
epoch_id,
session,
model_dict,
time_steps,
centralities,
train_instance_loader,
batch_size,
test_batch_size if epoch_id != epochs_to_run - 1 else 1,
batches_per_epoch,
test_instance_loader,
epoch_logging_file,
batch_logging_file,
log_to_stdout = log_to_stdout
)
print( "SAVING MODEL WEIGHTS TO {}".format( model_checkpoint_filename ) )
util.save_weights( session, model_checkpoint_filename )
#end for
#end train
def log_epoch(
epoch_id,
epoch_train_metrics_dict,
epoch_test_metrics_dict,
centralities,
epoch_logging_file
):
# Log the training part of the epoch
print( epoch_id, end = "", file = epoch_logging_file )
log_metrics_dict( epoch_train_metrics_dict, centralities, epoch_logging_file )
# Log the testing part of the epoch and flush the line
log_metrics_dict( epoch_test_metrics_dict, centralities, epoch_logging_file )
print( "", flush = True, file = epoch_logging_file )
#end log_epoch
def run_epoch(
epoch_id,
session,
model_dict,
time_steps,
centralities,
train_instance_loader,
batch_size,
test_batch_size,
batches_per_epoch,
test_instance_loader,
epoch_logging_file,
batch_logging_file,
log_to_stdout = False
):
"""
Runs and logs a training/testing epoch
"""
# Build the metrics dictionary for logging
epoch_train_metrics_dict = build_metrics_dict( centralities )
# Reset instance loader
train_instance_loader.reset()
for batch_id, batch in itertools.islice( enumerate( train_instance_loader.get_batches( batch_size ) ), batches_per_epoch ):
# Run and log every training batch, accumulating the metrics
batch_metrics_dict = run_batch(
epoch_id,
batch_id,
session,
model_dict,
time_steps,
centralities,
batch,
batch_logging_file,
train = True,
log_to_stdout = log_to_stdout
)
for metric in epoch_train_metrics_dict:
epoch_train_metrics_dict[metric] += batch_metrics_dict[metric]
#end for
#end for
# Normalize the metrics by the number of batches
for metric in epoch_train_metrics_dict:
epoch_train_metrics_dict[metric] /= batches_per_epoch
#end for
# Test
# Build the metrics dictionary for logging
epoch_test_metrics_dict = build_metrics_dict( centralities )
# Reset instance loader
test_instance_loader.reset()
# Counter for the number of instances
number_of_test_batches = 0
for cbat, batch in enumerate( test_instance_loader.get_batches( test_batch_size ) ):
# Open a null file so that we don't log every test instance being ran as a separate batch
with open(os.devnull, 'w') as nullfile:
# Run and log every test instance, accumulating the metrics
batch_metrics_dict = run_batch(
epoch_id,
"test",
session,
model_dict,
time_steps,
centralities,
batch,
nullfile,
train = False,
log_to_stdout = log_to_stdout
)
#end with
for metric in epoch_test_metrics_dict:
epoch_test_metrics_dict[metric] += batch_metrics_dict[metric]
#end for
number_of_test_batches += 1
#end for
# Normalize the metrics by the number of test instances
for metric in epoch_test_metrics_dict:
epoch_test_metrics_dict[metric] /= number_of_test_batches
#end for
log_epoch( epoch_id, epoch_train_metrics_dict, epoch_test_metrics_dict, centralities, epoch_logging_file )
if log_to_stdout:
print( "EPOCH\t", end = "" )
log_epoch( "summary", epoch_train_metrics_dict, epoch_test_metrics_dict, centralities, sys.stdout )
#end if
#end run_epoch
def log_batch(
epoch_id,
batch_id,
batch_metrics_dict,
centralities,
batch_logging_file
):
print(
"{eid}\t{bid}".format(
eid = epoch_id,
bid = batch_id
),
end = "",
file = batch_logging_file
)
log_metrics_dict( batch_metrics_dict, centralities, batch_logging_file )
print( "", flush = True, file = batch_logging_file )
#end
def run_batch(
epoch_id,
batch_id,
session,
model_dict,
time_steps,
centralities,
batch,
batch_logging_file,
train = False,
log_to_stdout = False
):
"""
Runs and logs a batch
"""
# Build metrics dictionary for logging
batch_metrics_dict = build_metrics_dict( centralities )
# Transform sparse batch labels to dense
labels = {
centrality: util.flatten( batch["{c}".format(c=centrality)] )
for centrality in centralities
}
# Build the feed_dict
feed_dict = {
model_dict["{c}_labels".format(c=centrality)]: labels[centrality]
for centrality in centralities
}
feed_dict[ model_dict["gnn"].matrix_placeholders["M"] ] = util.sparse_to_dense( batch["matrix"] )
feed_dict[ model_dict["gnn"].time_steps ] = time_steps
feed_dict[ model_dict["nodes_n"] ] = batch["problem_n"]
# Train if required
if train:
returned_values = session.run(
model_dict["train_step"],
feed_dict = feed_dict
)
#end if
# Get logits for batch
returned_predictions = session.run(
[
model_dict["{c}_predictions".format( c = centrality ) ]
for centrality in centralities
],
feed_dict = feed_dict
)
# Get losses for batch
returned_losses = session.run(
[
model_dict["loss"]
] + [
model_dict["{c}_cost".format( c = centrality ) ]
for centrality in centralities
],
feed_dict = feed_dict
)
# Update the overall loss
batch_metrics_dict["loss"] = returned_losses[0]
# Update each centrality's value
for centrality, predictions, cost in zip( centralities, returned_predictions, returned_losses[1:] ):
metric_values = get_metrics_from_batch(
model.separate_batch(
predictions,
batch["problem_n"]
),
model.separate_batch(
labels[centrality],
batch["problem_n"]
)
)
# Update loss for the centrality
batch_metrics_dict["{c}_cost".format(c=centrality)] = cost
# Update every other metric for the centrality
for metric, value in zip( METRICS_LIST, metric_values ):
batch_metrics_dict["{c}_{m}".format(c=centrality,m=metric)] = value
#end for
#end for
# For every metric, comput the average over the centralities
for metric in METRICS_LIST:
for centrality in centralities:
batch_metrics_dict[metric] += batch_metrics_dict["{c}_{m}".format(c=centrality,m=metric)]
#end for
batch_metrics_dict[metric] /= len( centralities )
#end for
# Log the batch
log_batch( epoch_id, batch_id, batch_metrics_dict, centralities, batch_logging_file )
if log_to_stdout:
log_batch( "batch", batch_id, batch_metrics_dict, centralities, sys.stdout )
#end if
return batch_metrics_dict
#end run_batch
if __name__ == "__main__":
embedding_size = 32
epochs_to_run = 32
batches_per_epoch = 32
batch_size = 32
test_batch_size = 32
time_steps = 32
centralities = sorted( sys.argv[1:] )
if len( centralities ) <= 0:
raise ValueError( "No centrality passed" )
#end if
for centrality in centralities:
if centrality not in [ "betweenness","closeness","degree","eigenvector" ]:
raise ValueError( "Centrality {c} not one of the accepted ones.".format( c=centrality ) )
#end if
#end for
fname = "centrality-" + "-".join( centralities )
# Build model
print( "Building model ..." )
GNN = model.model_builder( embedding_size, centralities )
# Load instances with a predefined seed and separate random generator for reproducibility
training_instance_loader_random_generator = random.Random( random_seed )
test_instance_loader_random_generator = random.Random( random_seed )
training_instance_loader = InstanceLoader("./instances", rng = training_instance_loader_random_generator )
test_instance_loader = InstanceLoader("./test-instances", rng = test_instance_loader_random_generator )
epoch_logging_file = open( "{fname}.epoch.log".format( fname = fname ), "w" )
batch_logging_file = open( "{fname}.batch.log".format( fname = fname ), "w" )
model_checkpoint_filename = fname
# Disallow GPU use
config = tf.ConfigProto(
# device_count = {"GPU": 0 },
# inter_op_parallelism_threads = 1,
# intra_op_parallelism_threads = 1,
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=1/5.2)
)
with tf.Session(config=config) as sess:
# Initialize global variables
print( "Initializing global variables ... " )
sess.run( tf.global_variables_initializer() )
train(
sess,
GNN,
time_steps,
centralities,
epochs_to_run,
training_instance_loader,
batch_size,
test_batch_size,
batches_per_epoch,
test_instance_loader,
epoch_logging_file,
batch_logging_file,
model_checkpoint_filename,
log_to_stdout = True
)
#end Session
|
[
"instance_loader.InstanceLoader",
"util.sparse_to_dense",
"random.Random",
"model.separate_batch",
"tensorflow.Session",
"random.seed",
"util.save_weights",
"tensorflow.global_variables_initializer",
"model.model_builder",
"numpy.random.seed",
"scipy.stats.pearsonr",
"tensorflow.set_random_seed",
"tensorflow.GPUOptions",
"scipy.stats.kendalltau"
] |
[((93, 117), 'random.seed', 'random.seed', (['random_seed'], {}), '(random_seed)\n', (104, 117), False, 'import random\n'), ((139, 166), 'numpy.random.seed', 'np.random.seed', (['random_seed'], {}), '(random_seed)\n', (153, 166), True, 'import numpy as np\n'), ((193, 224), 'tensorflow.set_random_seed', 'tf.set_random_seed', (['random_seed'], {}), '(random_seed)\n', (211, 224), True, 'import tensorflow as tf\n'), ((1769, 1806), 'scipy.stats.kendalltau', 'stats.kendalltau', (['predictions', 'labels'], {}), '(predictions, labels)\n', (1785, 1806), False, 'from scipy import stats\n'), ((1829, 1864), 'scipy.stats.pearsonr', 'stats.pearsonr', (['predictions', 'labels'], {}), '(predictions, labels)\n', (1843, 1864), False, 'from scipy import stats\n'), ((9790, 9827), 'util.sparse_to_dense', 'util.sparse_to_dense', (["batch['matrix']"], {}), "(batch['matrix'])\n", (9810, 9827), False, 'import os, sys, itertools, util\n'), ((12476, 12525), 'model.model_builder', 'model.model_builder', (['embedding_size', 'centralities'], {}), '(embedding_size, centralities)\n', (12495, 12525), False, 'import model\n'), ((12666, 12692), 'random.Random', 'random.Random', (['random_seed'], {}), '(random_seed)\n', (12679, 12692), False, 'import random\n'), ((12737, 12763), 'random.Random', 'random.Random', (['random_seed'], {}), '(random_seed)\n', (12750, 12763), False, 'import random\n'), ((12796, 12872), 'instance_loader.InstanceLoader', 'InstanceLoader', (['"""./instances"""'], {'rng': 'training_instance_loader_random_generator'}), "('./instances', rng=training_instance_loader_random_generator)\n", (12810, 12872), False, 'from instance_loader import InstanceLoader\n'), ((12901, 12978), 'instance_loader.InstanceLoader', 'InstanceLoader', (['"""./test-instances"""'], {'rng': 'test_instance_loader_random_generator'}), "('./test-instances', rng=test_instance_loader_random_generator)\n", (12915, 12978), False, 'from instance_loader import InstanceLoader\n'), ((5512, 5565), 'util.save_weights', 'util.save_weights', (['session', 'model_checkpoint_filename'], {}), '(session, model_checkpoint_filename)\n', (5529, 5565), False, 'import os, sys, itertools, util\n'), ((13422, 13447), 'tensorflow.Session', 'tf.Session', ([], {'config': 'config'}), '(config=config)\n', (13432, 13447), True, 'import tensorflow as tf\n'), ((10789, 10842), 'model.separate_batch', 'model.separate_batch', (['predictions', "batch['problem_n']"], {}), "(predictions, batch['problem_n'])\n", (10809, 10842), False, 'import model\n'), ((10874, 10934), 'model.separate_batch', 'model.separate_batch', (['labels[centrality]', "batch['problem_n']"], {}), "(labels[centrality], batch['problem_n'])\n", (10894, 10934), False, 'import model\n'), ((13358, 13412), 'tensorflow.GPUOptions', 'tf.GPUOptions', ([], {'per_process_gpu_memory_fraction': '(1 / 5.2)'}), '(per_process_gpu_memory_fraction=1 / 5.2)\n', (13371, 13412), True, 'import tensorflow as tf\n'), ((13555, 13588), 'tensorflow.global_variables_initializer', 'tf.global_variables_initializer', ([], {}), '()\n', (13586, 13588), True, 'import tensorflow as tf\n')]
|
#!/usr/bin/env python
import rospy
import math
import numpy as np
from sensor_msgs.msg import LaserScan
#######################################
# Laser Scan:
# Header: Seq, Stamp, frame_id
# Angle_min, Angle_max, Angle_Increment, Time_Increment
# Scan time, range_min, range_max, ranges, intensities
#######################################
class Noise_class:
def __init__(self):
#rospy.on_shutdown(self.save_csv)
self.laser_sub = rospy.Subscriber('/base_scan', LaserScan, self.laser_callback)
self.scan_pub = rospy.Publisher('/gaus_err_laser_scan', LaserScan, queue_size= 1)
def laser_callback(self, msg):
filtered_values = LaserScan()
distance = np.array(msg.ranges)
filtered_values.header = msg.header
filtered_values.angle_increment = msg.angle_increment
filtered_values.time_increment = msg.time_increment
filtered_values.scan_time = msg.scan_time
filtered_values.range_min = msg.range_min
filtered_values.range_max = msg.range_max
filtered_values.intensities = msg.intensities
angle = filtered_values.angle_increment
min_angle = msg.angle_min
max_angle = msg.angle_max
laser_noise_variance = rospy.get_param('laser_noise_variance')
if laser_noise_variance <= 0:
laser_noise_variance = 0.1
filtered_values_ranges = np.zeros(len(distance))
noise_values_ranges = np.random.normal(loc = 0, scale=laser_noise_variance, size=len(distance))
for i in range(len(distance)):
filtered_values_ranges[i] = noise_values_ranges[i]+distance[i]
filtered_values.ranges = filtered_values_ranges
filtered_values.angle_min = min_angle
filtered_values.angle_max = max_angle
self.scan_pub.publish(filtered_values)
if __name__ == '__main__':
rospy.init_node('noiser', anonymous=True)
noisy = Noise_class()
try:
rospy.spin()
except KeyboardInterrupt:
print("Shutting down")
|
[
"sensor_msgs.msg.LaserScan",
"rospy.Subscriber",
"rospy.init_node",
"rospy.get_param",
"numpy.array",
"rospy.spin",
"rospy.Publisher"
] |
[((1698, 1739), 'rospy.init_node', 'rospy.init_node', (['"""noiser"""'], {'anonymous': '(True)'}), "('noiser', anonymous=True)\n", (1713, 1739), False, 'import rospy\n'), ((441, 503), 'rospy.Subscriber', 'rospy.Subscriber', (['"""/base_scan"""', 'LaserScan', 'self.laser_callback'], {}), "('/base_scan', LaserScan, self.laser_callback)\n", (457, 503), False, 'import rospy\n'), ((522, 586), 'rospy.Publisher', 'rospy.Publisher', (['"""/gaus_err_laser_scan"""', 'LaserScan'], {'queue_size': '(1)'}), "('/gaus_err_laser_scan', LaserScan, queue_size=1)\n", (537, 586), False, 'import rospy\n'), ((642, 653), 'sensor_msgs.msg.LaserScan', 'LaserScan', ([], {}), '()\n', (651, 653), False, 'from sensor_msgs.msg import LaserScan\n'), ((667, 687), 'numpy.array', 'np.array', (['msg.ranges'], {}), '(msg.ranges)\n', (675, 687), True, 'import numpy as np\n'), ((1142, 1181), 'rospy.get_param', 'rospy.get_param', (['"""laser_noise_variance"""'], {}), "('laser_noise_variance')\n", (1157, 1181), False, 'import rospy\n'), ((1771, 1783), 'rospy.spin', 'rospy.spin', ([], {}), '()\n', (1781, 1783), False, 'import rospy\n')]
|
#!/usr/bin/env python3
import sys
import numpy as np
from example import AmiciExample
class ExampleCalvetti(AmiciExample):
def __init__(self):
AmiciExample.__init__( self )
self.numX = 6
self.numP = 0
self.numK = 6
self.modelOptions['theta'] = []
self.modelOptions['kappa'] = [0.29, 0.74, 0.44, 0.08, 0.27, 0.18]
self.modelOptions['ts'] = np.linspace(0, 20, 201)
self.modelOptions['pscale'] = 0
self.solverOptions['atol'] = 1e-6
self.solverOptions['rtol'] = 1e-4
self.solverOptions['sens_ind'] = []
self.solverOptions['sensi'] = 0
self.solverOptions['sensi_meth'] = 1
def writeNoSensi(filename):
ex = ExampleCalvetti()
ex.writeToFile(filename, '/model_calvetti/nosensi/')
def main():
if len(sys.argv) < 2:
print("Error: Must provide output file as first and only argument.")
sys.exit(1)
filename = sys.argv[1]
writeNoSensi(filename)
if __name__ == "__main__":
main()
|
[
"numpy.linspace",
"example.AmiciExample.__init__",
"sys.exit"
] |
[((158, 185), 'example.AmiciExample.__init__', 'AmiciExample.__init__', (['self'], {}), '(self)\n', (179, 185), False, 'from example import AmiciExample\n'), ((404, 427), 'numpy.linspace', 'np.linspace', (['(0)', '(20)', '(201)'], {}), '(0, 20, 201)\n', (415, 427), True, 'import numpy as np\n'), ((922, 933), 'sys.exit', 'sys.exit', (['(1)'], {}), '(1)\n', (930, 933), False, 'import sys\n')]
|
from PIL import Image
from tflite_runtime.interpreter import Interpreter
from tflite_runtime.interpreter import load_delegate
from video import create_capture
import numpy as np
import cv2 as cv
import io
import picamera
import simpleaudio as sa
# tf model upload
def load_labels(path):
with open(path, 'r') as f:
return {i: line.strip() for i, line in enumerate(f.readlines())}
def set_input_tensor(interpreter, image):
tensor_index = interpreter.get_input_details()[0]['index']
input_tensor = interpreter.tensor(tensor_index)()[0]
input_tensor[:, :] = image
# check whether user wears helmet
def classify_image(interpreter, image, top_k=1):
set_input_tensor(interpreter, image)
interpreter.invoke()
output_details = interpreter.get_output_details()[0]
output = np.squeeze(interpreter.get_tensor(output_details['index']))
# If the model is quantized (uint8 data), then dequantize the results
if output_details['dtype'] == np.uint8:
scale, zero_point = output_details['quantization']
output = scale * (output - zero_point)
ordered = np.argpartition(-output, top_k)
# if 0.90 above then regard user is wearing a helmet
if (top_k==1) and (output[1] > 0.9):
res = 1
else:
res = 0
return res
# for detect human face
def detect(img, cascade):
rects = cascade.detectMultiScale(img, scaleFactor=1.3, minNeighbors=4, minSize=(30, 30), flags=cv.CASCADE_SCALE_IMAGE)
if len(rects) == 0:
return []
rects[:,2:] += rects[:,:2]
return rects
def main():
import sys, getopt
checknum = 0
while True:
try:
# face recognizing code
print('face camera ')
args, video_src = getopt.getopt(sys.argv[1:2], '', ['cascade=', 'nested-cascade='])
try:
video_src = video_src[0]
except:
video_src = 0
args = dict(args)
cascade_fn = args.get('--cascade', "data/haarcascades/haarcascade_frontalface_alt.xml")
nested_fn = args.get('--nested-cascade', "data/haarcascades/haarcascade_eye.xml")
cascade = cv.CascadeClassifier(cv.samples.findFile(cascade_fn))
nested = cv.CascadeClassifier(cv.samples.findFile(nested_fn))
cam = create_capture(video_src, fallback='synth:bg={}:noise=0.05'.format(cv.samples.findFile('samples/data/lena.jpg')))
while True:
ret, img = cam.read()
gray = cv.cvtColor(img, cv.COLOR_BGR2GRAY)
gray = cv.equalizeHist(gray)
rects = detect(gray, cascade)
vis = img.copy()
if len(rects):
if not nested.empty():
print('into nested') # 사람이 들어왔을 때
for x1, y1, x2, y2 in rects:
roi = gray[y1:y2, x1:x2]
vis_roi = vis[y1:y2, x1:x2]
print('findrects')
subrects = detect(roi.copy(), nested)
if subrects!='[]':
faceok = 'faceok.wav'
fa = sa.WaveObject.from_wave_file(faceok)
face = fa.play()
face.wait_done()
print('detect!!')
break
cam.release() # face recognition camera off
print("helmet camera")
# helmet detectecting code
filename = 'helmet.wav'
wave_obj = sa.WaveObject.from_wave_file(filename)
helmetok = 'helmetok.wav'
wave = sa.WaveObject.from_wave_file(helmetok)
labels = "labels.txt"
model = "model_edgetpu.tflite"
interpreter = Interpreter(model, experimental_delegates=[load_delegate('libedgetpu.so.1.0')])
interpreter.allocate_tensors()
_, height, width, _ = interpreter.get_input_details()[0]['shape']
# helmet detect camera on
with picamera.PiCamera(resolution=(640, 480), framerate=30) as camera:
camera.start_preview()
try:
stream = io.BytesIO()
for _ in camera.capture_continuous(stream, format='jpeg', use_video_port=True):
stream.seek(0)
image = Image.open(stream).convert('RGB').resize((width, height),Image.ANTIALIAS)
results = classify_image(interpreter, image)
print("result:")
print(results)
stream.seek(0)
stream.truncate()
# 헬멧 착용여부 판단
if results==0:
play_obj = wave_obj.play()
play_obj.wait_done()
checknum += 1
if checknum==3:
checknum = 0
break;
else:
helm = wave.play()
helm.wait_done()
print('GoodBoy')
break
finally:
camera.stop_preview()
except KeyboardInterrupt:
break
if __name__ == '__main__':
main()
cv.destroyAllWindows()
|
[
"simpleaudio.WaveObject.from_wave_file",
"getopt.getopt",
"PIL.Image.open",
"numpy.argpartition",
"cv2.samples.findFile",
"io.BytesIO",
"picamera.PiCamera",
"cv2.equalizeHist",
"tflite_runtime.interpreter.load_delegate",
"cv2.destroyAllWindows",
"cv2.cvtColor"
] |
[((1193, 1224), 'numpy.argpartition', 'np.argpartition', (['(-output)', 'top_k'], {}), '(-output, top_k)\n', (1208, 1224), True, 'import numpy as np\n'), ((5829, 5851), 'cv2.destroyAllWindows', 'cv.destroyAllWindows', ([], {}), '()\n', (5849, 5851), True, 'import cv2 as cv\n'), ((1887, 1952), 'getopt.getopt', 'getopt.getopt', (['sys.argv[1:2]', '""""""', "['cascade=', 'nested-cascade=']"], {}), "(sys.argv[1:2], '', ['cascade=', 'nested-cascade='])\n", (1900, 1952), False, 'import sys, getopt\n'), ((3844, 3882), 'simpleaudio.WaveObject.from_wave_file', 'sa.WaveObject.from_wave_file', (['filename'], {}), '(filename)\n', (3872, 3882), True, 'import simpleaudio as sa\n'), ((3942, 3980), 'simpleaudio.WaveObject.from_wave_file', 'sa.WaveObject.from_wave_file', (['helmetok'], {}), '(helmetok)\n', (3970, 3980), True, 'import simpleaudio as sa\n'), ((2355, 2386), 'cv2.samples.findFile', 'cv.samples.findFile', (['cascade_fn'], {}), '(cascade_fn)\n', (2374, 2386), True, 'import cv2 as cv\n'), ((2431, 2461), 'cv2.samples.findFile', 'cv.samples.findFile', (['nested_fn'], {}), '(nested_fn)\n', (2450, 2461), True, 'import cv2 as cv\n'), ((2704, 2739), 'cv2.cvtColor', 'cv.cvtColor', (['img', 'cv.COLOR_BGR2GRAY'], {}), '(img, cv.COLOR_BGR2GRAY)\n', (2715, 2739), True, 'import cv2 as cv\n'), ((2764, 2785), 'cv2.equalizeHist', 'cv.equalizeHist', (['gray'], {}), '(gray)\n', (2779, 2785), True, 'import cv2 as cv\n'), ((4355, 4409), 'picamera.PiCamera', 'picamera.PiCamera', ([], {'resolution': '(640, 480)', 'framerate': '(30)'}), '(resolution=(640, 480), framerate=30)\n', (4372, 4409), False, 'import picamera\n'), ((4513, 4525), 'io.BytesIO', 'io.BytesIO', ([], {}), '()\n', (4523, 4525), False, 'import io\n'), ((2549, 2593), 'cv2.samples.findFile', 'cv.samples.findFile', (['"""samples/data/lena.jpg"""'], {}), "('samples/data/lena.jpg')\n", (2568, 2593), True, 'import cv2 as cv\n'), ((4134, 4168), 'tflite_runtime.interpreter.load_delegate', 'load_delegate', (['"""libedgetpu.so.1.0"""'], {}), "('libedgetpu.so.1.0')\n", (4147, 4168), False, 'from tflite_runtime.interpreter import load_delegate\n'), ((3419, 3455), 'simpleaudio.WaveObject.from_wave_file', 'sa.WaveObject.from_wave_file', (['faceok'], {}), '(faceok)\n', (3447, 3455), True, 'import simpleaudio as sa\n'), ((4700, 4718), 'PIL.Image.open', 'Image.open', (['stream'], {}), '(stream)\n', (4710, 4718), False, 'from PIL import Image\n')]
|
"""
tellotracker:
Allows manual operation of the drone and demo tracking mode.
Requires mplayer to record/save video.
Controls:
- tab to lift off
- WASD to move the drone
- space/shift to ascend/escent slowly
- Q/E to yaw slowly
- arrow keys to ascend, descend, or yaw quickly
- backspace to land, or P to palm-land
- enter to take a picture
- R to start recording video, R again to stop recording
(video and photos will be saved to a timestamped file in ~/Pictures/)
- Z to toggle camera zoom state
(zoomed-in widescreen or high FOV 4:3)
- T to toggle tracking
@author <NAME>, <NAME> and <NAME>
@copyright 2018 see license file for details
"""
import time
import datetime
import os
import tellopy
import numpy
import av
import cv2
from pynput import keyboard
from tracker import Tracker
#posenet
import os
import numpy as np
import sys
from tensorflow.lite.python.interpreter import Interpreter
from PIL import Image
import math
import threading
import traceback
frame = None
run_recv_thread = True
def sigmoid(x):
return 1 / (1 + math.exp(-x))
def argmax2d(inp_3d):
"""
Get the x,y positions of the heatmap of each part's argmax()
"""
heatmapPositions = np.zeros(shape=(17,2))
heatmapConf = np.zeros(shape=(17,1))
for i in range(17):
argmax_i = np.unravel_index(inp_3d[:,:,i].argmax(), inp_3d[:,:,i].shape)
max_i = inp_3d[:,:,i].max()
heatmapPositions[i,:] = argmax_i
heatmapConf[i,:] = max_i
return heatmapPositions,heatmapConf
def get_offsetVector(heatmapPositions=None,offsets=None):
allArrays = np.zeros(shape=(17,2))
for idx,el in enumerate(heatmapPositions):
# print(el)
allArrays[idx,0] = offsets[int(el[0]),int(el[1]),idx]
allArrays[idx,1] = offsets[int(el[0]),int(el[1]),17+idx]
return allArrays
MODEL_NAME = "pose_TFLite_model"
GRAPH_NAME = 'detect.tflite'
LABELMAP_NAME = 'labelmap.txt'
resW, resH = '952x720'.split('x')
imW, imH = int(resW), int(resH)
use_TPU = False
min_thresh = 0.7
# Get path to current working directory
CWD_PATH = os.getcwd()
# Path to .tflite file, which contains the model that is used for object detection
PATH_TO_CKPT = os.path.join(CWD_PATH,MODEL_NAME,GRAPH_NAME)
# Path to label map file
PATH_TO_LABELS = os.path.join(CWD_PATH,MODEL_NAME,LABELMAP_NAME)
# Load the label map
with open(PATH_TO_LABELS, 'r') as f:
labels = [line.strip() for line in f.readlines()]
# Have to do a weird fix for label map if using the COCO "starter model" from
# https://www.tensorflow.org/lite/models/object_detection/overview
# First label is '???', which has to be removed.
if labels[0] == '???':
del(labels[0])
# Load the Tensorflow Lite model.
# If using Edge TPU, use special load_delegate argument
if use_TPU:
interpreter = Interpreter(model_path=PATH_TO_CKPT,
experimental_delegates=[load_delegate('libedgetpu.so.1.0')])
print(PATH_TO_CKPT)
else:
interpreter = Interpreter(model_path=PATH_TO_CKPT)
interpreter.allocate_tensors()
# Get model details
input_details = interpreter.get_input_details()
output_details = interpreter.get_output_details()
height = input_details[0]['shape'][1]
width = input_details[0]['shape'][2]
floating_model = (input_details[0]['dtype'] == np.float32)
input_mean = width/2
input_std = width/2
# Initialize frame rate calculation
frame_rate_calc = 1
freq = cv2.getTickFrequency()
#posenet
def main():
""" Create a tello controller and show the video feed."""
tellotrack = TelloCV()
# for packet in tellotrack.container.demux((tellotrack.vid_stream,)):
# for frame in packet.decode():
# start = time.time()
# image = tellotrack.process_frame(frame)
# print("image_time",time.time()-start)
# cv2.imshow('tello', image)
# _ = cv2.waitKey(1) & 0xFF
#posenet
try:
threading.Thread(target=tellotrack.recv_thread).start()
while True:
if frame is None:
time.sleep(0.01)
else:
# print("frame FOUNDD")
image = tellotrack.process_frame(frame)
cv2.imshow('Original', image)
# cv2.imshow('Canny', cv2.Canny(image, 100, 200))
cv2.waitKey(1)
# long delay
# time.sleep(0.5)
image = None
except Exception as ex:
exc_type, exc_value, exc_traceback = sys.exc_info()
traceback.print_exception(exc_type, exc_value, exc_traceback)
print(ex)
finally:
run_recv_thread = False
cv2.destroyAllWindows()
#posenet
class TelloCV(object):
"""
TelloTracker builds keyboard controls on top of TelloPy as well
as generating images from the video stream and enabling opencv support
"""
def __init__(self):
self.prev_flight_data = None
self.record = False
self.tracking = False
self.keydown = False
self.date_fmt = '%Y-%m-%d_%H%M%S'
self.speed = 30
self.drone = tellopy.Tello()
self.init_drone() #posenet
self.init_controls()
# container for processing the packets into frames
self.container = av.open(self.drone.get_video_stream())
self.vid_stream = self.container.streams.video[0]
self.out_file = None
self.out_stream = None
self.out_name = None
self.start_time = time.time()
# tracking a color
green_lower = (30, 50, 50)
green_upper = (80, 255, 255)
#red_lower = (0, 50, 50)
# red_upper = (20, 255, 255)
# blue_lower = (110, 50, 50)
# upper_blue = (130, 255, 255)
self.track_cmd = ""
# self.tracker = Tracker(self.vid_stream.height,
# self.vid_stream.width,
# green_lower, green_upper) #posenet
self.tracker = Tracker(720,
960,
green_lower, green_upper) #posenet
#posenet
def recv_thread(self):
global frame
global run_recv_thread
print('start recv_thread()')
# drone = tellopy.Tello()
try:
# self.drone.connect()
# self.drone.wait_for_connection(60.0)
# #posenet
# self.drone.start_video()
# self.drone.subscribe(self.drone.EVENT_FLIGHT_DATA,
# self.flight_data_handler)
# self.drone.subscribe(self.drone.EVENT_FILE_RECEIVED,
# self.handle_flight_received)
#posenet
# container = av.open(self.drone.get_video_stream())
frame_count = 0
while run_recv_thread:
for f in self.container.decode(video=0):
frame_count = frame_count + 1
# skip first 300 frames
if frame_count < 300:
continue
frame = f
time.sleep(0.01)
except Exception as ex:
exc_type, exc_value, exc_traceback = sys.exc_info()
traceback.print_exception(exc_type, exc_value, exc_traceback)
print(ex)
finally:
self.drone.quit()
#posenet
def init_drone(self):
"""Connect, uneable streaming and subscribe to events"""
# self.drone.log.set_level(2)
self.drone.connect()
self.drone.wait_for_connection(60.0) #posenet
self.drone.start_video()
self.drone.subscribe(self.drone.EVENT_FLIGHT_DATA,
self.flight_data_handler)
self.drone.subscribe(self.drone.EVENT_FILE_RECEIVED,
self.handle_flight_received)
def on_press(self, keyname):
"""handler for keyboard listener"""
if self.keydown:
return
try:
self.keydown = True
keyname = str(keyname).strip('\'')
print('+' + keyname)
if keyname == 'Key.esc':
self.drone.quit()
exit(0)
if keyname in self.controls:
key_handler = self.controls[keyname]
if isinstance(key_handler, str):
getattr(self.drone, key_handler)(self.speed)
else:
key_handler(self.speed)
except AttributeError:
print('special key {0} pressed'.format(keyname))
def on_release(self, keyname):
"""Reset on key up from keyboard listener"""
self.keydown = False
keyname = str(keyname).strip('\'')
print('-' + keyname)
if keyname in self.controls:
key_handler = self.controls[keyname]
if isinstance(key_handler, str):
getattr(self.drone, key_handler)(0)
else:
key_handler(0)
def init_controls(self):
"""Define keys and add listener"""
self.controls = {
'w': lambda speed: self.drone.forward(speed),#'forward',
's': 'backward',
'a': 'left',
'd': 'right',
'Key.space': 'up',
'Key.shift': 'down',
'Key.shift_r': 'down',
'q': 'counter_clockwise',
'e': 'clockwise',
'i': lambda speed: self.drone.flip_forward(),
'k': lambda speed: self.drone.flip_back(),
'j': lambda speed: self.drone.flip_left(),
'l': lambda speed: self.drone.flip_right(),
# arrow keys for fast turns and altitude adjustments
'Key.left': lambda speed: self.drone.counter_clockwise(speed),
'Key.right': lambda speed: self.drone.clockwise(speed),
'Key.up': lambda speed: self.drone.up(speed),
'Key.down': lambda speed: self.drone.down(speed),
'Key.tab': lambda speed: self.drone.takeoff(),
'Key.backspace': lambda speed: self.drone.land(),
'p': lambda speed: self.palm_land(speed),
't': lambda speed: self.toggle_tracking(speed),
'r': lambda speed: self.toggle_recording(speed),
'z': lambda speed: self.toggle_zoom(speed),
'Key.enter': lambda speed: self.take_picture(speed)
}
self.key_listener = keyboard.Listener(on_press=self.on_press,
on_release=self.on_release)
self.key_listener.start()
# self.key_listener.join()
def process_frame(self, frame):
"""convert frame to cv2 image and show"""
# Start timer (for calculating frame rate)
t1 = cv2.getTickCount()
image = cv2.cvtColor(numpy.array(
frame.to_image()), cv2.COLOR_RGB2BGR)
image = self.write_hud(image)
if self.record:
self.record_vid(frame)
# xoff, yoff = self.tracker.track(image)
xoff, yoff = 0,0
xLeftWrist, yLeftWrist =0,0
xNose, yNose =0,0
# print("CV xoff{}, yoff {}".format(xoff, yoff))
#posenet
frame_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
frame_resized = cv2.resize(frame_rgb, (width, height))
input_data = np.expand_dims(frame_resized, axis=0)
# Normalize pixel values if using a floating model (i.e. if model is non-quantized)
if floating_model:
input_data = (np.float32(input_data) - input_mean) / input_std
# Perform the actual detection by running the model with the image as input
interpreter.set_tensor(input_details[0]['index'],input_data)
interpreter.invoke()
heatmapscores = interpreter.get_tensor(output_details[0]['index'])[0] # Bounding box coordinates of detected objects
offsets = interpreter.get_tensor(output_details[1]['index'])[0] # Class index of detected objects
# define vectorized sigmoid
sigmoid_v = np.vectorize(sigmoid)
# 1 sigmoid
sigmoheatmapscores = sigmoid_v(heatmapscores)
# 2 argmax2d
heatmapPositions,heatmapConfidence = argmax2d(sigmoheatmapscores)
# 3 offsetVectors
offsetVectors = get_offsetVector(heatmapPositions,offsets)
# 4 keypointPositions
outputStride = 32
keypointPositions = heatmapPositions * outputStride + offsetVectors
# 5 draw keypoints
for idx,el in enumerate(heatmapConfidence):
if heatmapConfidence[idx][0] >= min_thresh:
x = round((keypointPositions[idx][1]/width)*imW)
y = round((keypointPositions[idx][0]/height)*imH)
if 'right' in labels[idx]:
cv2.circle(image,(int(x),int(y)), 5, (0,255,0), -1)
elif 'left' in labels[idx]:
cv2.circle(image,(int(x),int(y)), 5, (0,0,255), -1)
elif 'nose' in labels[idx]:
xNose, yNose = int(x),int(y)
xoff, yoff = (x-int(960/2)),(int(720/2)-y)
# print("NOSE xoff{}, yoff {}".format(xoff, yoff))
cv2.circle(image,(int(x),int(y)), 5, (255,0,0), -1)
if 'leftWri' in labels[idx]:
xLeftWrist, yLeftWrist = int(x),int(y)
#posenet
def draw_arrows(frame):
"""Show the direction vector output in the cv2 window"""
#cv2.putText(frame,"Color:", (0, 35), cv2.FONT_HERSHEY_SIMPLEX, 1, 255, thickness=2)
cv2.arrowedLine(frame, (int(960/2), int(720/2)),
(int(960/2 + xoff), int(720/2 - yoff)),
(0, 0, 255), 1)
return frame
# image = self.tracker.draw_arrows(image)
image = draw_arrows(image)
# Calculate framerate
t2 = cv2.getTickCount()
time1 = (t2-t1)/freq
frame_rate_calc= 1/time1
# Draw framerate in corner of frame
cv2.putText(image,
'FPS: {0:.2f}'.format(frame_rate_calc),
(imW-200,30),
cv2.FONT_HERSHEY_SIMPLEX,
1,
(255,255,0),
1,
cv2.LINE_AA)
distance = 150
cmd = ""
# print(yoff)
# print("WRIST {}>>>> NOSE {}???? ".format(yLeftWrist,yNose),yLeftWrist>yNose)
if self.tracking:
# if yLeftWrist>yNose:
# print("RECORDING",yLeftWrist)
# cmd = "r"
# lambda speed: self.toggle_recording(speed)
if xoff < -distance and xoff>-960/2:
cmd = "counter_clockwise"
elif xoff > distance and xoff<960/2:
cmd = "clockwise"
elif yoff < -distance and yoff>-720/2:
cmd = "down"
elif yoff > distance and yoff<720/2:
print("UPPPPPPPPPPPPPPP",yoff)
cmd = "up"
else:
if self.track_cmd is not "":
getattr(self.drone, self.track_cmd)(0)
self.track_cmd = ""
if cmd is not self.track_cmd:
if cmd is not "":
print("track command:", cmd)
getattr(self.drone, cmd)(self.speed)
self.track_cmd = cmd
return image
def write_hud(self, frame):
"""Draw drone info, tracking and record on frame"""
stats = self.prev_flight_data.split('|')
stats.append("Tracking:" + str(self.tracking))
if self.drone.zoom:
stats.append("VID")
else:
stats.append("PIC")
if self.record:
diff = int(time.time() - self.start_time)
mins, secs = divmod(diff, 60)
stats.append("REC {:02d}:{:02d}".format(mins, secs))
for idx, stat in enumerate(stats):
text = stat.lstrip()
cv2.putText(frame, text, (0, 30 + (idx * 30)),
cv2.FONT_HERSHEY_SIMPLEX,
1.0, (255, 0, 0), lineType=30)
return frame
def toggle_recording(self, speed):
"""Handle recording keypress, creates output stream and file"""
if speed == 0:
return
self.record = not self.record
if self.record:
datename = [os.getenv('HOME'), datetime.datetime.now().strftime(self.date_fmt)]
self.out_name = '{}/Pictures/tello-{}.mp4'.format(*datename)
print("Outputting video to:", self.out_name)
self.out_file = av.open(self.out_name, 'w')
self.start_time = time.time()
self.out_stream = self.out_file.add_stream(
'mpeg4', self.vid_stream.rate)
self.out_stream.pix_fmt = 'yuv420p'
self.out_stream.width = self.vid_stream.width
self.out_stream.height = self.vid_stream.height
if not self.record:
print("Video saved to ", self.out_name)
self.out_file.close()
self.out_stream = None
def record_vid(self, frame):
"""
convert frames to packets and write to file
"""
new_frame = av.VideoFrame(
width=frame.width, height=frame.height, format=frame.format.name)
for i in range(len(frame.planes)):
new_frame.planes[i].update(frame.planes[i])
pkt = None
try:
pkt = self.out_stream.encode(new_frame)
except IOError as err:
print("encoding failed: {0}".format(err))
if pkt is not None:
try:
self.out_file.mux(pkt)
except IOError:
print('mux failed: ' + str(pkt))
def take_picture(self, speed):
"""Tell drone to take picture, image sent to file handler"""
if speed == 0:
return
self.drone.take_picture()
def palm_land(self, speed):
"""Tell drone to land"""
if speed == 0:
return
self.drone.palm_land()
def toggle_tracking(self, speed):
""" Handle tracking keypress"""
if speed == 0: # handle key up event
return
self.tracking = not self.tracking
print("tracking:", self.tracking)
return
def toggle_zoom(self, speed):
"""
In "video" mode the self.drone sends 1280x720 frames.
In "photo" mode it sends 2592x1936 (952x720) frames.
The video will always be centered in the window.
In photo mode, if we keep the window at 1280x720 that gives us ~160px on
each side for status information, which is ample.
Video mode is harder because then we need to abandon the 16:9 display size
if we want to put the HUD next to the video.
"""
if speed == 0:
return
self.drone.set_video_mode(not self.drone.zoom)
def flight_data_handler(self, event, sender, data):
"""Listener to flight data from the drone."""
text = str(data)
if self.prev_flight_data != text:
self.prev_flight_data = text
def handle_flight_received(self, event, sender, data):
"""Create a file in ~/Pictures/ to receive image from the drone"""
path = '%s/Pictures/tello-%s.jpeg' % (
os.getenv('HOME'),
datetime.datetime.now().strftime(self.date_fmt))
with open(path, 'wb') as out_file:
out_file.write(data)
print('Saved photo to %s' % path)
if __name__ == '__main__':
main()
|
[
"time.sleep",
"cv2.imshow",
"sys.exc_info",
"av.open",
"cv2.destroyAllWindows",
"math.exp",
"traceback.print_exception",
"tracker.Tracker",
"cv2.waitKey",
"cv2.getTickFrequency",
"numpy.float32",
"cv2.putText",
"cv2.cvtColor",
"threading.Thread",
"cv2.resize",
"time.time",
"numpy.vectorize",
"pynput.keyboard.Listener",
"av.VideoFrame",
"os.getenv",
"os.path.join",
"os.getcwd",
"datetime.datetime.now",
"numpy.zeros",
"cv2.getTickCount",
"numpy.expand_dims",
"tellopy.Tello",
"tensorflow.lite.python.interpreter.Interpreter"
] |
[((2062, 2073), 'os.getcwd', 'os.getcwd', ([], {}), '()\n', (2071, 2073), False, 'import os\n'), ((2173, 2219), 'os.path.join', 'os.path.join', (['CWD_PATH', 'MODEL_NAME', 'GRAPH_NAME'], {}), '(CWD_PATH, MODEL_NAME, GRAPH_NAME)\n', (2185, 2219), False, 'import os\n'), ((2261, 2310), 'os.path.join', 'os.path.join', (['CWD_PATH', 'MODEL_NAME', 'LABELMAP_NAME'], {}), '(CWD_PATH, MODEL_NAME, LABELMAP_NAME)\n', (2273, 2310), False, 'import os\n'), ((3385, 3407), 'cv2.getTickFrequency', 'cv2.getTickFrequency', ([], {}), '()\n', (3405, 3407), False, 'import cv2\n'), ((1183, 1206), 'numpy.zeros', 'np.zeros', ([], {'shape': '(17, 2)'}), '(shape=(17, 2))\n', (1191, 1206), True, 'import numpy as np\n'), ((1224, 1247), 'numpy.zeros', 'np.zeros', ([], {'shape': '(17, 1)'}), '(shape=(17, 1))\n', (1232, 1247), True, 'import numpy as np\n'), ((1578, 1601), 'numpy.zeros', 'np.zeros', ([], {'shape': '(17, 2)'}), '(shape=(17, 2))\n', (1586, 1601), True, 'import numpy as np\n'), ((2956, 2992), 'tensorflow.lite.python.interpreter.Interpreter', 'Interpreter', ([], {'model_path': 'PATH_TO_CKPT'}), '(model_path=PATH_TO_CKPT)\n', (2967, 2992), False, 'from tensorflow.lite.python.interpreter import Interpreter\n'), ((4607, 4630), 'cv2.destroyAllWindows', 'cv2.destroyAllWindows', ([], {}), '()\n', (4628, 4630), False, 'import cv2\n'), ((5063, 5078), 'tellopy.Tello', 'tellopy.Tello', ([], {}), '()\n', (5076, 5078), False, 'import tellopy\n'), ((5440, 5451), 'time.time', 'time.time', ([], {}), '()\n', (5449, 5451), False, 'import time\n'), ((5938, 5981), 'tracker.Tracker', 'Tracker', (['(720)', '(960)', 'green_lower', 'green_upper'], {}), '(720, 960, green_lower, green_upper)\n', (5945, 5981), False, 'from tracker import Tracker\n'), ((10356, 10425), 'pynput.keyboard.Listener', 'keyboard.Listener', ([], {'on_press': 'self.on_press', 'on_release': 'self.on_release'}), '(on_press=self.on_press, on_release=self.on_release)\n', (10373, 10425), False, 'from pynput import keyboard\n'), ((10701, 10719), 'cv2.getTickCount', 'cv2.getTickCount', ([], {}), '()\n', (10717, 10719), False, 'import cv2\n'), ((11141, 11179), 'cv2.cvtColor', 'cv2.cvtColor', (['image', 'cv2.COLOR_BGR2RGB'], {}), '(image, cv2.COLOR_BGR2RGB)\n', (11153, 11179), False, 'import cv2\n'), ((11204, 11242), 'cv2.resize', 'cv2.resize', (['frame_rgb', '(width, height)'], {}), '(frame_rgb, (width, height))\n', (11214, 11242), False, 'import cv2\n'), ((11264, 11301), 'numpy.expand_dims', 'np.expand_dims', (['frame_resized'], {'axis': '(0)'}), '(frame_resized, axis=0)\n', (11278, 11301), True, 'import numpy as np\n'), ((11993, 12014), 'numpy.vectorize', 'np.vectorize', (['sigmoid'], {}), '(sigmoid)\n', (12005, 12014), True, 'import numpy as np\n'), ((13843, 13861), 'cv2.getTickCount', 'cv2.getTickCount', ([], {}), '()\n', (13859, 13861), False, 'import cv2\n'), ((17168, 17247), 'av.VideoFrame', 'av.VideoFrame', ([], {'width': 'frame.width', 'height': 'frame.height', 'format': 'frame.format.name'}), '(width=frame.width, height=frame.height, format=frame.format.name)\n', (17181, 17247), False, 'import av\n'), ((1043, 1055), 'math.exp', 'math.exp', (['(-x)'], {}), '(-x)\n', (1051, 1055), False, 'import math\n'), ((4451, 4465), 'sys.exc_info', 'sys.exc_info', ([], {}), '()\n', (4463, 4465), False, 'import sys\n'), ((4474, 4535), 'traceback.print_exception', 'traceback.print_exception', (['exc_type', 'exc_value', 'exc_traceback'], {}), '(exc_type, exc_value, exc_traceback)\n', (4499, 4535), False, 'import traceback\n'), ((15909, 16014), 'cv2.putText', 'cv2.putText', (['frame', 'text', '(0, 30 + idx * 30)', 'cv2.FONT_HERSHEY_SIMPLEX', '(1.0)', '(255, 0, 0)'], {'lineType': '(30)'}), '(frame, text, (0, 30 + idx * 30), cv2.FONT_HERSHEY_SIMPLEX, 1.0,\n (255, 0, 0), lineType=30)\n', (15920, 16014), False, 'import cv2\n'), ((16549, 16576), 'av.open', 'av.open', (['self.out_name', '"""w"""'], {}), "(self.out_name, 'w')\n", (16556, 16576), False, 'import av\n'), ((16607, 16618), 'time.time', 'time.time', ([], {}), '()\n', (16616, 16618), False, 'import time\n'), ((3888, 3935), 'threading.Thread', 'threading.Thread', ([], {'target': 'tellotrack.recv_thread'}), '(target=tellotrack.recv_thread)\n', (3904, 3935), False, 'import threading\n'), ((4011, 4027), 'time.sleep', 'time.sleep', (['(0.01)'], {}), '(0.01)\n', (4021, 4027), False, 'import time\n'), ((4158, 4187), 'cv2.imshow', 'cv2.imshow', (['"""Original"""', 'image'], {}), "('Original', image)\n", (4168, 4187), False, 'import cv2\n'), ((4270, 4284), 'cv2.waitKey', 'cv2.waitKey', (['(1)'], {}), '(1)\n', (4281, 4284), False, 'import cv2\n'), ((7056, 7072), 'time.sleep', 'time.sleep', (['(0.01)'], {}), '(0.01)\n', (7066, 7072), False, 'import time\n'), ((7154, 7168), 'sys.exc_info', 'sys.exc_info', ([], {}), '()\n', (7166, 7168), False, 'import sys\n'), ((7181, 7242), 'traceback.print_exception', 'traceback.print_exception', (['exc_type', 'exc_value', 'exc_traceback'], {}), '(exc_type, exc_value, exc_traceback)\n', (7206, 7242), False, 'import traceback\n'), ((16323, 16340), 'os.getenv', 'os.getenv', (['"""HOME"""'], {}), "('HOME')\n", (16332, 16340), False, 'import os\n'), ((19277, 19294), 'os.getenv', 'os.getenv', (['"""HOME"""'], {}), "('HOME')\n", (19286, 19294), False, 'import os\n'), ((11447, 11469), 'numpy.float32', 'np.float32', (['input_data'], {}), '(input_data)\n', (11457, 11469), True, 'import numpy as np\n'), ((15682, 15693), 'time.time', 'time.time', ([], {}), '()\n', (15691, 15693), False, 'import time\n'), ((16342, 16365), 'datetime.datetime.now', 'datetime.datetime.now', ([], {}), '()\n', (16363, 16365), False, 'import datetime\n'), ((19308, 19331), 'datetime.datetime.now', 'datetime.datetime.now', ([], {}), '()\n', (19329, 19331), False, 'import datetime\n')]
|
# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from auto_scan_test import PassAutoScanTest, IgnoreReasons
from program_config import TensorConfig, ProgramConfig, OpConfig
import numpy as np
import paddle.inference as paddle_infer
from functools import partial
from typing import Optional, List, Callable, Dict, Any, Set
import unittest
import hypothesis
from hypothesis import given, settings, seed, example, assume, reproduce_failure
import hypothesis.strategies as st
class TestFcFusePass(PassAutoScanTest):
"""
x_var y_var(persistable)
\ /
mul bias_var(persistable)
|
mul_out_var bias_var(persistable)
\ /
elementwise_add
"""
def sample_predictor_configs(self, program_config):
# cpu
before_num_ops = len(program_config.ops) + 2
config = self.create_inference_config(use_gpu=False)
yield config, ["fc"], (1e-5, 1e-5)
# for gpu
config = self.create_inference_config(use_gpu=True)
yield config, ["fc"], (1e-5, 1e-5)
# trt static_shape
config = self.create_trt_inference_config()
config.enable_tensorrt_engine(
max_batch_size=8,
workspace_size=102400,
min_subgraph_size=0,
precision_mode=paddle_infer.PrecisionType.Float32,
use_static=False,
use_calib_mode=False)
yield config, ['fc'], (1e-5, 1e-5)
def add_ignore_pass_case(self):
# Here we put some skip rules to avoid known bugs
def teller1(program_config, predictor_config):
# shape of bias should be [1, mul_y_shape[-1]] or [mul_y_shape[-1]]
x_shape = list(program_config.inputs["mul_x"].shape)
y_shape = list(program_config.weights["mul_y"].shape)
bias_shape = program_config.weights["bias"].shape
bias_shape = list(program_config.weights["bias"].shape)
if predictor_config.tensorrt_engine_enabled():
# TensorRT cann't handle all the situation of elementwise_add
# disable it until this problem fixed
predictor_config.exp_disable_tensorrt_ops(["elementwise_add"])
if bias_shape != [y_shape[-1]] and bias_shape != [1, y_shape[-1]]:
return True
return False
def teller2(program_config, predictor_config):
# TODO fuse has bug while axis != -1
axis = program_config.ops[1].attrs["axis"]
if axis != -1 and axis != program_config.ops[0].attrs[
"x_num_col_dims"]:
return True
return False
self.add_ignore_check_case(
teller1,
IgnoreReasons.PASS_ACCURACY_ERROR,
"The pass output has diff while shape of bias is not [out_size] or [1, out_size].",
)
self.add_ignore_check_case(
teller2,
IgnoreReasons.PASS_ACCURACY_ERROR,
"The pass output has diff while axis of elementwise_add is not -1.",
)
def is_program_valid(self, prog_config):
add_x_rank = prog_config.ops[0].attrs["x_num_col_dims"] + 1
add_y_rank = len(prog_config.weights["bias"].shape)
axis = prog_config.ops[1].attrs["axis"]
if add_x_rank == add_y_rank:
if axis != -1 or axis != 0:
return False
return True
def sample_program_config(self, draw):
# 1. Generate shape of input:X of mul
x_shape = draw(
st.lists(st.integers(min_value=1, max_value=4),
min_size=2,
max_size=4))
# 2. Generate attr:x_num_col_dims/y_num_col_dims of mul
x_num_col_dims = draw(
st.integers(min_value=1, max_value=len(x_shape) - 1))
y_num_col_dims = 1
# 3. Generate legal shape of input:Y of mul
y_shape = draw(
st.lists(st.integers(min_value=1, max_value=8),
min_size=2,
max_size=2))
y_shape[0] = int(np.prod(x_shape[x_num_col_dims:]))
# 4. Generate legal attr:axis of elementwise_add
mul_out_shape = x_shape[:x_num_col_dims] + y_shape[1:]
axis = draw(st.integers(min_value=-1, max_value=x_num_col_dims))
# 5. Generate legal shape of input:Y of elementwise_add
if axis >= 0:
max_bias_rank = x_num_col_dims + 1 - axis
bias_rank = draw(st.integers(min_value=1, max_value=max_bias_rank))
bias_shape = mul_out_shape[axis:axis + bias_rank]
else:
max_bias_rank = 1
bias_rank = draw(
st.integers(min_value=1, max_value=len(mul_out_shape)))
bias_shape = mul_out_shape[-1 * bias_rank:]
# 6. Random choose if use broadcast for elementwise_add, e.g [3, 4] -> [1, 4]
if draw(st.booleans()):
broadcast_dims = draw(st.integers(min_value=1, max_value=bias_rank))
for i in range(0, broadcast_dims):
bias_shape[i] = 1
# 7. Random choose if add a relu operator
has_relu = draw(st.booleans())
# Now we have all the decided parameters to compose a program
# shape of inputs/weights tensors: x_shape, y_shape, bias_shape...
# parameters of operators: x_num_col_dims, y_num_col_dims, axis...
# a random boolean value(has_relu) to decide if program include a relu op
# Here we will compose a program
# Still has some risks that the program is invalid or cause bug while running
# Use function `is_program_valid` to filter the invalid programs before running
# Use function `add_skip_pass_case` to ignore the programs even if they cause bug while runing
mul_op = OpConfig(
"mul",
inputs={
"X": ["mul_x"],
"Y": ["mul_y"]
},
outputs={"Out": ["mul_out"]},
x_num_col_dims=x_num_col_dims,
y_num_col_dims=y_num_col_dims,
)
add_op = OpConfig(
"elementwise_add",
inputs={
"X": ["mul_out"],
"Y": ["bias"]
},
outputs={"Out": ["add_out"]},
axis=axis,
)
ops = [mul_op, add_op]
if has_relu:
relu_op = OpConfig("relu",
inputs={"X": ["add_out"]},
outputs={"Out": ["relu_out"]})
ops.append(relu_op)
program_config = ProgramConfig(
ops=ops,
weights={
"mul_y": TensorConfig(shape=y_shape),
"bias": TensorConfig(shape=bias_shape),
},
inputs={
"mul_x": TensorConfig(shape=x_shape),
},
outputs=ops[-1].outputs["Out"],
)
return program_config
def test(self):
self.run_and_statis(quant=False,
max_examples=500,
passes=["fc_fuse_pass"])
if __name__ == "__main__":
unittest.main()
|
[
"numpy.prod",
"hypothesis.strategies.integers",
"program_config.TensorConfig",
"hypothesis.strategies.booleans",
"unittest.main",
"program_config.OpConfig"
] |
[((7681, 7696), 'unittest.main', 'unittest.main', ([], {}), '()\n', (7694, 7696), False, 'import unittest\n'), ((6363, 6516), 'program_config.OpConfig', 'OpConfig', (['"""mul"""'], {'inputs': "{'X': ['mul_x'], 'Y': ['mul_y']}", 'outputs': "{'Out': ['mul_out']}", 'x_num_col_dims': 'x_num_col_dims', 'y_num_col_dims': 'y_num_col_dims'}), "('mul', inputs={'X': ['mul_x'], 'Y': ['mul_y']}, outputs={'Out': [\n 'mul_out']}, x_num_col_dims=x_num_col_dims, y_num_col_dims=y_num_col_dims)\n", (6371, 6516), False, 'from program_config import TensorConfig, ProgramConfig, OpConfig\n'), ((6646, 6760), 'program_config.OpConfig', 'OpConfig', (['"""elementwise_add"""'], {'inputs': "{'X': ['mul_out'], 'Y': ['bias']}", 'outputs': "{'Out': ['add_out']}", 'axis': 'axis'}), "('elementwise_add', inputs={'X': ['mul_out'], 'Y': ['bias']},\n outputs={'Out': ['add_out']}, axis=axis)\n", (6654, 6760), False, 'from program_config import TensorConfig, ProgramConfig, OpConfig\n'), ((4643, 4676), 'numpy.prod', 'np.prod', (['x_shape[x_num_col_dims:]'], {}), '(x_shape[x_num_col_dims:])\n', (4650, 4676), True, 'import numpy as np\n'), ((4818, 4869), 'hypothesis.strategies.integers', 'st.integers', ([], {'min_value': '(-1)', 'max_value': 'x_num_col_dims'}), '(min_value=-1, max_value=x_num_col_dims)\n', (4829, 4869), True, 'import hypothesis.strategies as st\n'), ((5457, 5470), 'hypothesis.strategies.booleans', 'st.booleans', ([], {}), '()\n', (5468, 5470), True, 'import hypothesis.strategies as st\n'), ((5709, 5722), 'hypothesis.strategies.booleans', 'st.booleans', ([], {}), '()\n', (5720, 5722), True, 'import hypothesis.strategies as st\n'), ((6936, 7010), 'program_config.OpConfig', 'OpConfig', (['"""relu"""'], {'inputs': "{'X': ['add_out']}", 'outputs': "{'Out': ['relu_out']}"}), "('relu', inputs={'X': ['add_out']}, outputs={'Out': ['relu_out']})\n", (6944, 7010), False, 'from program_config import TensorConfig, ProgramConfig, OpConfig\n'), ((4121, 4158), 'hypothesis.strategies.integers', 'st.integers', ([], {'min_value': '(1)', 'max_value': '(4)'}), '(min_value=1, max_value=4)\n', (4132, 4158), True, 'import hypothesis.strategies as st\n'), ((4512, 4549), 'hypothesis.strategies.integers', 'st.integers', ([], {'min_value': '(1)', 'max_value': '(8)'}), '(min_value=1, max_value=8)\n', (4523, 4549), True, 'import hypothesis.strategies as st\n'), ((5040, 5089), 'hypothesis.strategies.integers', 'st.integers', ([], {'min_value': '(1)', 'max_value': 'max_bias_rank'}), '(min_value=1, max_value=max_bias_rank)\n', (5051, 5089), True, 'import hypothesis.strategies as st\n'), ((5507, 5552), 'hypothesis.strategies.integers', 'st.integers', ([], {'min_value': '(1)', 'max_value': 'bias_rank'}), '(min_value=1, max_value=bias_rank)\n', (5518, 5552), True, 'import hypothesis.strategies as st\n'), ((7213, 7240), 'program_config.TensorConfig', 'TensorConfig', ([], {'shape': 'y_shape'}), '(shape=y_shape)\n', (7225, 7240), False, 'from program_config import TensorConfig, ProgramConfig, OpConfig\n'), ((7266, 7296), 'program_config.TensorConfig', 'TensorConfig', ([], {'shape': 'bias_shape'}), '(shape=bias_shape)\n', (7278, 7296), False, 'from program_config import TensorConfig, ProgramConfig, OpConfig\n'), ((7359, 7386), 'program_config.TensorConfig', 'TensorConfig', ([], {'shape': 'x_shape'}), '(shape=x_shape)\n', (7371, 7386), False, 'from program_config import TensorConfig, ProgramConfig, OpConfig\n')]
|
import cv2
from tkinter import Tk
from tkinter.filedialog import askopenfilename
import numpy as np
import imutils
import threading
def main():
cap = cv2.VideoCapture(vid_path)
status1, previous_frame = cap.read()
total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
copy_frame = cv2.cvtColor(previous_frame, cv2.COLOR_BGR2GRAY)
fgbg = cv2.createBackgroundSubtractorMOG2()
hsv = np.zeros_like(previous_frame)
hsv[...,1] = 255
t = 20
red = 30
check_red = 1
start = 0
radiuce_up_limit =60
radiuce_low_limit = 30
i = 0
while(i < total_frames - 1):
ret, frame = cap.read()
i = i + 1
frame1 = frame.copy()
current_frame = cv2.cvtColor(frame1, cv2.COLOR_BGR2GRAY)
current_frame = cv2.GaussianBlur(current_frame, (var_blur,var_blur), 0)
# frame differening
frame_diff = cv2.absdiff(current_frame,copy_frame)
ret ,binary_image1 = cv2.threshold(frame_diff,3,255,cv2.THRESH_BINARY)
# Background Subtraction
binary_image3 = fgbg.apply(current_frame)
# combination of two methods
final_binary = cv2.bitwise_and(binary_image3,binary_image1)
lab_val = 255
n_labels, img_labeled, lab_stats, _ = \
cv2.connectedComponentsWithStats(final_binary, connectivity=8,
ltype=cv2.CV_32S)
if check_red == 1:
red = red +10
if red > radiuce_up_limit:
check_red =0
else:
red = red -10
if red == radiuce_low_limit:
check_red =1
if lab_stats[1:, 4].size > 2:
re = lab_stats[1:, 4].argsort()[-2:][::-1] + 1
largest_mask = np.zeros(final_binary.shape, dtype=np.uint8)
largest_mask[img_labeled == re[0]] = lab_val
cnts1 = cv2.findContours(largest_mask.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
cnts1 = cnts1[0] if imutils.is_cv2() else cnts1[1]
X1 = cnts1[0][0]
cX1 = X1[0][0]
cY1 = X1[0][1]
cv2.circle(frame, (cX1, cY1), red, (0, 255, 255), 3)
cv2.putText(frame,'Breathing',(10,40),cv2.FONT_HERSHEY_SIMPLEX,1,(0,255,255),1,cv2.LINE_AA)
cv2.imshow('Frame',frame)
else:
t = t+1
if t > 40:
if lab_stats[1:, 4].size > 0 and start == 1:
t = 0
cv2.putText(frame,'Not Breathing',(10,40),cv2.FONT_HERSHEY_SIMPLEX,1,(0,0,255),1,cv2.LINE_AA)
cv2.imshow('Frame',frame)
else:
cv2.circle(frame, (cX1, cY1), red, (0, 255, 255), 3)
cv2.putText(frame,'Breathing',(10,40),cv2.FONT_HERSHEY_SIMPLEX,1,(0,255,255),1,cv2.LINE_AA)
cv2.imshow('Frame',frame)
previous_frame = current_frame
k = cv2.waitKey(1) & 0xff
if k == 27:
break
cap.release()
cv2.destroyAllWindows()
Tk().withdraw()
vid_path = askopenfilename(filetypes =(("Video File", "*.mp4"),("Video File","*.avi"),("Video File", "*.flv"),("All Files","*.*")),
title = "Choose a video.")
no_of_threads = 1
var_blur = 3
thred = []
jobs = []
for i in range(0, no_of_threads):
thred = threading.Thread(target=main)
jobs.append(thred)
for j in jobs:
j.start()
for j in jobs:
j.join()
#
#
#
|
[
"cv2.createBackgroundSubtractorMOG2",
"imutils.is_cv2",
"cv2.imshow",
"cv2.destroyAllWindows",
"cv2.threshold",
"cv2.waitKey",
"tkinter.filedialog.askopenfilename",
"cv2.putText",
"cv2.circle",
"cv2.cvtColor",
"cv2.GaussianBlur",
"cv2.bitwise_and",
"numpy.zeros",
"cv2.connectedComponentsWithStats",
"tkinter.Tk",
"cv2.VideoCapture",
"threading.Thread",
"numpy.zeros_like",
"cv2.absdiff"
] |
[((3251, 3404), 'tkinter.filedialog.askopenfilename', 'askopenfilename', ([], {'filetypes': "(('Video File', '*.mp4'), ('Video File', '*.avi'), ('Video File', '*.flv'),\n ('All Files', '*.*'))", 'title': '"""Choose a video."""'}), "(filetypes=(('Video File', '*.mp4'), ('Video File', '*.avi'),\n ('Video File', '*.flv'), ('All Files', '*.*')), title='Choose a video.')\n", (3266, 3404), False, 'from tkinter.filedialog import askopenfilename\n'), ((174, 200), 'cv2.VideoCapture', 'cv2.VideoCapture', (['vid_path'], {}), '(vid_path)\n', (190, 200), False, 'import cv2\n'), ((322, 370), 'cv2.cvtColor', 'cv2.cvtColor', (['previous_frame', 'cv2.COLOR_BGR2GRAY'], {}), '(previous_frame, cv2.COLOR_BGR2GRAY)\n', (334, 370), False, 'import cv2\n'), ((382, 418), 'cv2.createBackgroundSubtractorMOG2', 'cv2.createBackgroundSubtractorMOG2', ([], {}), '()\n', (416, 418), False, 'import cv2\n'), ((429, 458), 'numpy.zeros_like', 'np.zeros_like', (['previous_frame'], {}), '(previous_frame)\n', (442, 458), True, 'import numpy as np\n'), ((3195, 3218), 'cv2.destroyAllWindows', 'cv2.destroyAllWindows', ([], {}), '()\n', (3216, 3218), False, 'import cv2\n'), ((3523, 3552), 'threading.Thread', 'threading.Thread', ([], {'target': 'main'}), '(target=main)\n', (3539, 3552), False, 'import threading\n'), ((754, 794), 'cv2.cvtColor', 'cv2.cvtColor', (['frame1', 'cv2.COLOR_BGR2GRAY'], {}), '(frame1, cv2.COLOR_BGR2GRAY)\n', (766, 794), False, 'import cv2\n'), ((819, 875), 'cv2.GaussianBlur', 'cv2.GaussianBlur', (['current_frame', '(var_blur, var_blur)', '(0)'], {}), '(current_frame, (var_blur, var_blur), 0)\n', (835, 875), False, 'import cv2\n'), ((932, 970), 'cv2.absdiff', 'cv2.absdiff', (['current_frame', 'copy_frame'], {}), '(current_frame, copy_frame)\n', (943, 970), False, 'import cv2\n'), ((1008, 1060), 'cv2.threshold', 'cv2.threshold', (['frame_diff', '(3)', '(255)', 'cv2.THRESH_BINARY'], {}), '(frame_diff, 3, 255, cv2.THRESH_BINARY)\n', (1021, 1060), False, 'import cv2\n'), ((1211, 1256), 'cv2.bitwise_and', 'cv2.bitwise_and', (['binary_image3', 'binary_image1'], {}), '(binary_image3, binary_image1)\n', (1226, 1256), False, 'import cv2\n'), ((1351, 1436), 'cv2.connectedComponentsWithStats', 'cv2.connectedComponentsWithStats', (['final_binary'], {'connectivity': '(8)', 'ltype': 'cv2.CV_32S'}), '(final_binary, connectivity=8, ltype=cv2.CV_32S\n )\n', (1383, 1436), False, 'import cv2\n'), ((3224, 3228), 'tkinter.Tk', 'Tk', ([], {}), '()\n', (3226, 3228), False, 'from tkinter import Tk\n'), ((1910, 1954), 'numpy.zeros', 'np.zeros', (['final_binary.shape'], {'dtype': 'np.uint8'}), '(final_binary.shape, dtype=np.uint8)\n', (1918, 1954), True, 'import numpy as np\n'), ((2308, 2360), 'cv2.circle', 'cv2.circle', (['frame', '(cX1, cY1)', 'red', '(0, 255, 255)', '(3)'], {}), '(frame, (cX1, cY1), red, (0, 255, 255), 3)\n', (2318, 2360), False, 'import cv2\n'), ((2373, 2479), 'cv2.putText', 'cv2.putText', (['frame', '"""Breathing"""', '(10, 40)', 'cv2.FONT_HERSHEY_SIMPLEX', '(1)', '(0, 255, 255)', '(1)', 'cv2.LINE_AA'], {}), "(frame, 'Breathing', (10, 40), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, \n 255, 255), 1, cv2.LINE_AA)\n", (2384, 2479), False, 'import cv2\n'), ((2477, 2503), 'cv2.imshow', 'cv2.imshow', (['"""Frame"""', 'frame'], {}), "('Frame', frame)\n", (2487, 2503), False, 'import cv2\n'), ((3113, 3127), 'cv2.waitKey', 'cv2.waitKey', (['(1)'], {}), '(1)\n', (3124, 3127), False, 'import cv2\n'), ((2146, 2162), 'imutils.is_cv2', 'imutils.is_cv2', ([], {}), '()\n', (2160, 2162), False, 'import imutils\n'), ((2685, 2792), 'cv2.putText', 'cv2.putText', (['frame', '"""Not Breathing"""', '(10, 40)', 'cv2.FONT_HERSHEY_SIMPLEX', '(1)', '(0, 0, 255)', '(1)', 'cv2.LINE_AA'], {}), "(frame, 'Not Breathing', (10, 40), cv2.FONT_HERSHEY_SIMPLEX, 1,\n (0, 0, 255), 1, cv2.LINE_AA)\n", (2696, 2792), False, 'import cv2\n'), ((2795, 2821), 'cv2.imshow', 'cv2.imshow', (['"""Frame"""', 'frame'], {}), "('Frame', frame)\n", (2805, 2821), False, 'import cv2\n'), ((2855, 2907), 'cv2.circle', 'cv2.circle', (['frame', '(cX1, cY1)', 'red', '(0, 255, 255)', '(3)'], {}), '(frame, (cX1, cY1), red, (0, 255, 255), 3)\n', (2865, 2907), False, 'import cv2\n'), ((2924, 3030), 'cv2.putText', 'cv2.putText', (['frame', '"""Breathing"""', '(10, 40)', 'cv2.FONT_HERSHEY_SIMPLEX', '(1)', '(0, 255, 255)', '(1)', 'cv2.LINE_AA'], {}), "(frame, 'Breathing', (10, 40), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, \n 255, 255), 1, cv2.LINE_AA)\n", (2935, 3030), False, 'import cv2\n'), ((3032, 3058), 'cv2.imshow', 'cv2.imshow', (['"""Frame"""', 'frame'], {}), "('Frame', frame)\n", (3042, 3058), False, 'import cv2\n')]
|
import numpy as np
import scipy.stats as stats
import scipy.special as spec
import util
class HMCParams:
def __init__(self, tau, tau_g, L, eta, mass, r_clip, grad_clip):
self.tau = tau
self.tau_g = tau_g
self.L = L
self.eta = eta
self.mass = mass
self.r_clip = r_clip
self.grad_clip = grad_clip
class GradClipCounter:
def __init__(self):
self.clipped_grad = 0
self.grad_accesses = 0
def zcdp_iters(epsilon, delta, params, n, compute_less_grad=False):
rho = (np.sqrt(epsilon - np.log(delta)) - np.sqrt(-np.log(delta)))**2
rho_l = 1 / (2 * params.tau**2 * n)
rho_g = 1 / (2 * params.tau_g**2 * n)
# print("rho_l: {}".format(rho_l))
# print("rho_g: {}".format(rho_g))
if compute_less_grad:
iters = int((rho - rho_g) / (rho_l + params.L * rho_g))
else:
iters = int(rho / (rho_l + (params.L + 1) * rho_g))
return iters
def adp_delta(k, epsilon, params, n, compute_less_grad=False):
tau_l = params.tau
tau_g = params.tau_g
L = params.L
grad_evals = k * L + 1 if compute_less_grad else k * (L + 1)
mu = k / (2 * tau_l**2 * n) + grad_evals / (2 * tau_g**2 * n)
term1 = spec.erfc((epsilon - mu) / (2 * np.sqrt(mu)))
term2 = np.exp(epsilon) * spec.erfc((epsilon + mu) / (2 * np.sqrt(mu)))
return (0.5 * (term1 - term2)).sum()
def adp_iters(epsilon, delta, params, n, compute_less_grad=False):
low_iters = zcdp_iters(epsilon, delta, params, n, compute_less_grad)
up_iters = max(low_iters, 1)
while adp_delta(up_iters, epsilon, params, n, compute_less_grad) < delta:
up_iters *= 2
while int(up_iters) - int(low_iters) > 1:
new_iters = (low_iters + up_iters) / 2
new_delta = adp_delta(new_iters, epsilon, params, n, compute_less_grad)
if new_delta > delta:
up_iters = new_iters
else:
low_iters = new_iters
if adp_delta(int(up_iters), epsilon, params, n, compute_less_grad) < delta:
return int(up_iters)
else:
return int(low_iters)
def hmc(problem, theta0, epsilon, delta, params, verbose=True, use_adp=True, compute_less_grad=False):
data = problem.data
n, data_dim = data.shape
dim = theta0.size
temp_scale = problem.temp_scale
tau = params.tau
tau_g = params.tau_g
L = params.L
eta = params.eta
mass = params.mass
r_clip = params.r_clip
grad_clip = params.grad_clip
if not use_adp:
iters = zcdp_iters(epsilon, delta, params, n, compute_less_grad)
else:
iters = adp_iters(epsilon, delta, params, n, compute_less_grad)
if verbose:
print("Iterations: {}".format(iters))
sigma = tau * np.sqrt(n)
chain = np.zeros((iters + 1, dim))
chain[0, :] = theta0
leapfrog_chain = np.zeros((iters * L, dim))
clipped_r = np.zeros(iters)
clipped_grad_counter = GradClipCounter()
accepts = 0
grad_noise_sigma = 2 * tau_g * np.sqrt(n) * grad_clip
def grad_fun(theta):
ll_grad, clips = problem.log_likelihood_grad_clipped(grad_clip, theta, data)
clipped_grad_counter.clipped_grad += clips
clipped_grad_counter.grad_accesses += 1
pri_grad = problem.log_prior_grad(theta)
return temp_scale * (ll_grad + stats.norm.rvs(size=dim, scale=grad_noise_sigma)) + pri_grad
if compute_less_grad:
grad = grad_fun(theta0)
llc = problem.log_likelihood_no_sum(theta0, data)
for i in range(iters):
current = chain[i, :]
#TODO: this assumes diagonal M
p = stats.norm.rvs(size=dim) * np.sqrt(mass)
p_orig = p.copy()
prop = current.copy()
if compute_less_grad:
grad_new = grad.copy()
else:
grad_new = grad_fun(current)
for j in range(L):
p += 0.5 * eta * (grad_new)# - 0.5 * grad_noise_sigma**2 * p / mass)
prop += eta * p / mass
leapfrog_chain[i * L + j] = prop
grad_new = grad_fun(prop)
p += 0.5 * eta * (grad_new)# - 0.5 * grad_noise_sigma**2 * p / mass)
llp = problem.log_likelihood_no_sum(prop, data)
r = llp - llc
d = np.sqrt(np.sum((current - prop)**2))
clip = d * r_clip
clipped_r[i] = np.sum(np.abs(r) > clip)
r = np.clip(r, -clip, clip)
lpp = problem.log_prior(prop)
lpc = problem.log_prior(current)
s = stats.norm.rvs(size=1, scale=sigma * d * 2 * r_clip)
dp = 0.5 * np.sum(p_orig**2 / mass) - 0.5 * np.sum(p**2 / mass)
dH = dp + temp_scale * (np.sum(r) + s) + lpp - lpc
u = np.log(np.random.rand())
if u < dH - 0.5 * (temp_scale * sigma * d * 2 * r_clip)**2:
chain[i + 1, :] = prop
if compute_less_grad:
grad = grad_new
llc = llp
accepts += 1
else:
chain[i + 1, :] = current
if verbose and (i + 1) % 100 == 0:
print("Iteration: {}".format(i + 1))
if verbose:
print("Gradient evals: {}".format(clipped_grad_counter.grad_accesses))
return util.MCMCResult(
problem, chain, leapfrog_chain, iters, accepts, np.sum(clipped_r) / n / iters,
np.sum(clipped_grad_counter.clipped_grad) / n / clipped_grad_counter.grad_accesses
)
# return (
# chain, leapfrog_chain, accepts, clipped_r, iters,
# clipped_grad_counter.clipped_grad, clipped_grad_counter.grad_accesses
# )
|
[
"numpy.clip",
"numpy.abs",
"numpy.sqrt",
"numpy.random.rand",
"numpy.log",
"scipy.stats.norm.rvs",
"numpy.exp",
"numpy.sum",
"numpy.zeros"
] |
[((2760, 2786), 'numpy.zeros', 'np.zeros', (['(iters + 1, dim)'], {}), '((iters + 1, dim))\n', (2768, 2786), True, 'import numpy as np\n'), ((2833, 2859), 'numpy.zeros', 'np.zeros', (['(iters * L, dim)'], {}), '((iters * L, dim))\n', (2841, 2859), True, 'import numpy as np\n'), ((2876, 2891), 'numpy.zeros', 'np.zeros', (['iters'], {}), '(iters)\n', (2884, 2891), True, 'import numpy as np\n'), ((1281, 1296), 'numpy.exp', 'np.exp', (['epsilon'], {}), '(epsilon)\n', (1287, 1296), True, 'import numpy as np\n'), ((2736, 2746), 'numpy.sqrt', 'np.sqrt', (['n'], {}), '(n)\n', (2743, 2746), True, 'import numpy as np\n'), ((4332, 4355), 'numpy.clip', 'np.clip', (['r', '(-clip)', 'clip'], {}), '(r, -clip, clip)\n', (4339, 4355), True, 'import numpy as np\n'), ((4449, 4501), 'scipy.stats.norm.rvs', 'stats.norm.rvs', ([], {'size': '(1)', 'scale': '(sigma * d * 2 * r_clip)'}), '(size=1, scale=sigma * d * 2 * r_clip)\n', (4463, 4501), True, 'import scipy.stats as stats\n'), ((2989, 2999), 'numpy.sqrt', 'np.sqrt', (['n'], {}), '(n)\n', (2996, 2999), True, 'import numpy as np\n'), ((3593, 3617), 'scipy.stats.norm.rvs', 'stats.norm.rvs', ([], {'size': 'dim'}), '(size=dim)\n', (3607, 3617), True, 'import scipy.stats as stats\n'), ((3620, 3633), 'numpy.sqrt', 'np.sqrt', (['mass'], {}), '(mass)\n', (3627, 3633), True, 'import numpy as np\n'), ((4217, 4246), 'numpy.sum', 'np.sum', (['((current - prop) ** 2)'], {}), '((current - prop) ** 2)\n', (4223, 4246), True, 'import numpy as np\n'), ((4652, 4668), 'numpy.random.rand', 'np.random.rand', ([], {}), '()\n', (4666, 4668), True, 'import numpy as np\n'), ((1255, 1266), 'numpy.sqrt', 'np.sqrt', (['mu'], {}), '(mu)\n', (1262, 1266), True, 'import numpy as np\n'), ((4302, 4311), 'numpy.abs', 'np.abs', (['r'], {}), '(r)\n', (4308, 4311), True, 'import numpy as np\n'), ((4521, 4547), 'numpy.sum', 'np.sum', (['(p_orig ** 2 / mass)'], {}), '(p_orig ** 2 / mass)\n', (4527, 4547), True, 'import numpy as np\n'), ((4554, 4575), 'numpy.sum', 'np.sum', (['(p ** 2 / mass)'], {}), '(p ** 2 / mass)\n', (4560, 4575), True, 'import numpy as np\n'), ((5212, 5229), 'numpy.sum', 'np.sum', (['clipped_r'], {}), '(clipped_r)\n', (5218, 5229), True, 'import numpy as np\n'), ((5251, 5292), 'numpy.sum', 'np.sum', (['clipped_grad_counter.clipped_grad'], {}), '(clipped_grad_counter.clipped_grad)\n', (5257, 5292), True, 'import numpy as np\n'), ((568, 581), 'numpy.log', 'np.log', (['delta'], {}), '(delta)\n', (574, 581), True, 'import numpy as np\n'), ((594, 607), 'numpy.log', 'np.log', (['delta'], {}), '(delta)\n', (600, 607), True, 'import numpy as np\n'), ((1331, 1342), 'numpy.sqrt', 'np.sqrt', (['mu'], {}), '(mu)\n', (1338, 1342), True, 'import numpy as np\n'), ((3311, 3359), 'scipy.stats.norm.rvs', 'stats.norm.rvs', ([], {'size': 'dim', 'scale': 'grad_noise_sigma'}), '(size=dim, scale=grad_noise_sigma)\n', (3325, 3359), True, 'import scipy.stats as stats\n'), ((4606, 4615), 'numpy.sum', 'np.sum', (['r'], {}), '(r)\n', (4612, 4615), True, 'import numpy as np\n')]
|
# -*- coding: utf-8 -*-
import numpy
from simmate.toolkit import Structure
from pymatgen.analysis.diffusion.neb.pathfinder import (
DistinctPathFinder,
MigrationHop as PymatgenMigrationHop,
IDPPSolver,
)
from typing import List
class MigrationImages(list):
"""
This class is just a list of structures for a diffusion pathway. It has
utility methods to help create these structures but otherwise behaves
exactly like a python list.
Note, this class is primarily used to generate inputs for calculations. If
you'd like more advanced features, you should represent your diffusion
pathway as a MigrationHop instead.As a rule of thumb: Only use this class
if you are manually creating your pathway from endpoint supercells or from
a set of supercell images.
All MigrationHop's can be converted to MigrationImages (using the
`from_migration_hop` method); but not all MigrationImages can be converted
to MigrationHops.
"""
def __init__(self, structures: List[Structure]):
# This init function does nothing except apply typing -- specifically,
# it says that it expects a list of structures.
super().__init__(structures)
def get_sum_structure(self, tolerance: float = 1e-3):
"""
Takes all structures and combines them into one. Atoms that are within
the given tolerance are joined into a single site.
This is primarily used to view a diffusing pathway within a single
structure -- as well as how the host lattice changes during diffusion.
If you are able to convert your pathway to a MigrationHop, the
MigrationHop.write_path() method is much faster and cleaner than this
method, so it should be preffered. Also, because there are many atoms
that are overlapping here, the output structure may cause programs
like VESTA to crash.
#### Parameters
- `tolerance`:
the angle and distance tolerance to consider fractional coordinates
as matching. Matching sites will be merged as 1 site in the final
sum structure.
"""
# OPTIMIZE: this is very inefficient. It's much faster to visualize
# structures with MigrationHop class because you know which atom is
# moving. Here, we need to treat all atoms as moving. We can also
# speed this up by only looking at diffusing species too.
final_coords = []
final_species = []
for structure in self: # recall self is a list of structures
for site in structure:
is_new = True
for coords in final_coords:
if all(
numpy.isclose(
site.frac_coords,
coords,
rtol=tolerance,
atol=tolerance,
)
):
is_new = False
break
if is_new:
final_coords.append(site.frac_coords)
final_species.append(site.specie)
structure = Structure(
lattice=structure.lattice,
species=final_species,
coords=final_coords,
)
return structure
@staticmethod
def get_nimages(
pathway_length: float,
min_image_step: float = 0.7,
require_midpoint: bool = True,
):
"""
Gives the desirable number of images (not including start/end structures).
This method helps generate a MigrationImages object, and typically is
not called directly. The other classmethods of MigrationImages call
this for you.
#### Parameters
- `pathway_length`:
The length of the pathway.
- `min_image_step`:
The minimum step distance for the diffusing atom between images.
The default is 0.7 Angstroms. For example, a path 2.8A long would
require at least 4 images for this default.
- `require_midpoint`:
Whether there should be an image at the midpoint. In other words,
whether the number of images should be odd. This is often important
if you expect the transition state to be at the midpoint and you are
not running CI-NEB. The default is True.
Returns
-------
- `nimages`:
The number of images to use for this pathway.
"""
# At a minimum, we want to have images be 0.7 angstroms apart, and
# with one additional image.
nimages = pathway_length // min_image_step + 1
# We also want an odd number of images. This ensures we have an image
# at exactly the midpoint, which is often necessary if we aren't
# running CI-NEB.
if require_midpoint and nimages % 2 == 0:
nimages += 1
# This is a float but it makes more sense to have an integer
return int(nimages)
@classmethod
def from_migration_hop(
cls,
migration_hop: PymatgenMigrationHop,
vacancy_mode: bool = True,
min_nsites: int = 80,
max_nsites: int = 240,
min_length: int = 10,
**kwargs,
):
"""
Creates a MigrationImages object from a MigrationHop object
#### Parameters
- `migration_hop`:
The MigrationHop object that should be converted.
- `vacancy_mode`:
Whether to use single-vacancy diffusion (True) or interstitial
diffusion (False). The default is True.
- `min_nsites`:
The minimum number of sites to have in the supercell structure.
The default is 80.
- `max_nsites`:
The maximum number of sites to have in the supercell structure.
The default is 240.
- `min_length`:
The minimum length for each vector in the supercell structure.
The default is 10 Angstroms.
- `**kwargs`:
Any arguments that are normally accepted by IDPPSolver
"""
# The third thing returned is the bulk_supercell which we don't need.
start_supercell, end_supercell, _ = migration_hop.get_sc_structures(
vac_mode=vacancy_mode,
min_atoms=min_nsites,
max_atoms=max_nsites,
min_length=min_length,
)
# calculate the number of images required
nimages = cls.get_nimages(migration_hop.length)
return cls.from_endpoints(
start_supercell,
end_supercell,
nimages=nimages,
**kwargs,
)
@classmethod
def from_endpoints(
cls,
structure_start: Structure,
structure_end: Structure,
nimages: int,
**kwargs,
):
"""
Creates a MigrationImages object from start and end supercell structures.
You do not need to specify the diffusing atom(s) as all sites are
linearly interpolated and then relaxed by IDPP.
#### Parameters
- `structure_start`:
The starting supercell of the diffusion pathway.
- `structure_end`:
The ending supercell of the diffusion pathway.
- `nimages`:
The number of desired images for the pathway. Note, if you know the
pathway length of your path, you can use the `get_nimages` static
method to get a logical number of images.
- `**kwargs`:
Any arguments that are normally accepted by IDPPSolver
"""
# Run IDPP relaxation on the images before returning them
idpp_solver = IDPPSolver.from_endpoints(
[structure_start, structure_end],
nimages=nimages,
**kwargs,
)
images = idpp_solver.run()
return cls(images)
@classmethod
def from_startend_sites(
cls,
structure: Structure,
site_start: int,
site_end: int,
**kwargs,
):
"""
Creates a MigrationImages object from a bulk structure and start/end
periodic sites of the diffusing atom.
For example, this would allow a diffusion pathway that goes from a site
at (0,0,0) to (1,1,1). Thus, symmetry and periodic boundry conditions
are considered.
Note, this method just creates a MigrationHop and then uses the
`from_migration_hop` method to make a MigrationImages object.
#### Parameters
- `structure`:
The bulk crystal structure (NOT the supercell).
- `site_start`:
The starting periodic site for this pathway.
- `site_end`:
The end periodic site for this pathway.
- `**kwargs`:
Any arguments that are normally accepted by `from_migration_hop`.
"""
# This information is all we need for a MigrationHop object
pathway = PymatgenMigrationHop(site_start, site_end, structure)
return cls.from_migration_hop(pathway, **kwargs)
@classmethod
def from_structure(
cls,
structure: Structure,
migrating_specie: str,
pathfinder_kwargs: dict = {},
**kwargs,
):
"""
Given a bulk crystal structure, this will find all symmetrically
unique pathways and return them as list of MigrationImages objects.
#### Parameters
- `structure`:
The bulk crystal structure (NOT the supercell).
- `migrating_specie`:
The identity of the diffusing ion (e.g. "Li" or "Li1+"). Note, only
provide oxidation state if you are using an oxidation-state decorated
structure.
- `pathfinder_kwargs`:
Any arguments that are normally accepted by DistinctPathFinder, but
given as a dictionary. The default is {}.
- `**kwargs`:
Any arguments that are normally accepted by `from_migration_hop`.
"""
# convert to the LLL reduced primitive cell to make it as cubic as possible
structure_lll = structure.get_sanitized_structure()
# Use pymatgen to find all the symmetrically unique pathways.
# NOTE: This only finds pathways up until the structure is percolating.
# If you are interested in longer pathways, then this script needs to
# be adjusted by passing additional kwargs
pathfinder = DistinctPathFinder(
structure_lll,
migrating_specie=migrating_specie,
**pathfinder_kwargs,
)
pathways = pathfinder.get_paths()
# Now go through each path and convert to a MigrationPath. We return
# these as a list of paths.
migration_paths = []
for pathway in pathways:
migration_path = cls.from_migration_hop(
migration_hop=pathway,
**kwargs,
)
migration_paths.append(migration_path)
return migration_paths
@classmethod
def from_dynamic(cls, migration_images):
"""
This is an experimental feature. The code here is a repurposing of
Structre.from_dynamic so consider making a general class for
from_dynamic methods.
"""
is_from_past_calc = False
# assume any list is in the MigrationHop format if there are more than
# two structures (i.e. there is at least one midpoint image)
if isinstance(migration_images, list) and len(migration_images) > 2:
migration_images_cleaned = migration_images
else:
raise Exception("Unknown format provided for migration_images input.")
migration_images_cleaned.is_from_past_calc = is_from_past_calc
return migration_images_cleaned
def as_dict(self):
return [s.as_dict() for s in self]
|
[
"simmate.toolkit.Structure",
"numpy.isclose",
"pymatgen.analysis.diffusion.neb.pathfinder.MigrationHop",
"pymatgen.analysis.diffusion.neb.pathfinder.IDPPSolver.from_endpoints",
"pymatgen.analysis.diffusion.neb.pathfinder.DistinctPathFinder"
] |
[((3195, 3280), 'simmate.toolkit.Structure', 'Structure', ([], {'lattice': 'structure.lattice', 'species': 'final_species', 'coords': 'final_coords'}), '(lattice=structure.lattice, species=final_species, coords=final_coords\n )\n', (3204, 3280), False, 'from simmate.toolkit import Structure\n'), ((7774, 7864), 'pymatgen.analysis.diffusion.neb.pathfinder.IDPPSolver.from_endpoints', 'IDPPSolver.from_endpoints', (['[structure_start, structure_end]'], {'nimages': 'nimages'}), '([structure_start, structure_end], nimages=nimages,\n **kwargs)\n', (7799, 7864), False, 'from pymatgen.analysis.diffusion.neb.pathfinder import DistinctPathFinder, MigrationHop as PymatgenMigrationHop, IDPPSolver\n'), ((9060, 9113), 'pymatgen.analysis.diffusion.neb.pathfinder.MigrationHop', 'PymatgenMigrationHop', (['site_start', 'site_end', 'structure'], {}), '(site_start, site_end, structure)\n', (9080, 9113), True, 'from pymatgen.analysis.diffusion.neb.pathfinder import DistinctPathFinder, MigrationHop as PymatgenMigrationHop, IDPPSolver\n'), ((10560, 10654), 'pymatgen.analysis.diffusion.neb.pathfinder.DistinctPathFinder', 'DistinctPathFinder', (['structure_lll'], {'migrating_specie': 'migrating_specie'}), '(structure_lll, migrating_specie=migrating_specie, **\n pathfinder_kwargs)\n', (10578, 10654), False, 'from pymatgen.analysis.diffusion.neb.pathfinder import DistinctPathFinder, MigrationHop as PymatgenMigrationHop, IDPPSolver\n'), ((2732, 2803), 'numpy.isclose', 'numpy.isclose', (['site.frac_coords', 'coords'], {'rtol': 'tolerance', 'atol': 'tolerance'}), '(site.frac_coords, coords, rtol=tolerance, atol=tolerance)\n', (2745, 2803), False, 'import numpy\n')]
|
# Copyright 2020 LMNT, Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
import numpy as np
import os
import torch
import torch.nn as nn
from torch.nn.parallel import DistributedDataParallel
from torch.utils.tensorboard import SummaryWriter
from tqdm import tqdm
from wavegrad.dataset import from_path as dataset_from_path
from wavegrad.model import WaveGrad
def _nested_map(struct, map_fn):
if isinstance(struct, tuple):
return tuple(_nested_map(x, map_fn) for x in struct)
if isinstance(struct, list):
return [_nested_map(x, map_fn) for x in struct]
if isinstance(struct, dict):
return { k: _nested_map(v, map_fn) for k, v in struct.items() }
return map_fn(struct)
class WaveGradLearner:
def __init__(self, model_dir, model, dataset, optimizer, params, *args, **kwargs):
os.makedirs(model_dir, exist_ok=True)
self.model_dir = model_dir
self.model = model
self.dataset = dataset
self.optimizer = optimizer
self.params = params
self.autocast = torch.cuda.amp.autocast(enabled=kwargs.get('fp16', False))
self.scaler = torch.cuda.amp.GradScaler(enabled=kwargs.get('fp16', False))
self.step = 0
self.is_master = True
beta = np.array(self.params.noise_schedule)
noise_level = np.cumprod(1 - beta)**0.5
noise_level = np.concatenate([[1.0], noise_level], axis=0)
self.noise_level = torch.tensor(noise_level.astype(np.float32))
self.loss_fn = nn.L1Loss()
self.summary_writer = None
def state_dict(self):
if hasattr(self.model, 'module') and isinstance(self.model.module, nn.Module):
model_state = self.model.module.state_dict()
else:
model_state = self.model.state_dict()
return {
'step': self.step,
'model': { k: v.cpu() if isinstance(v, torch.Tensor) else v for k, v in model_state.items() },
'optimizer': { k: v.cpu() if isinstance(v, torch.Tensor) else v for k, v in self.optimizer.state_dict().items() },
'params': dict(self.params),
'scaler': self.scaler.state_dict(),
}
def load_state_dict(self, state_dict):
if hasattr(self.model, 'module') and isinstance(self.model.module, nn.Module):
self.model.module.load_state_dict(state_dict['model'])
else:
self.model.load_state_dict(state_dict['model'])
self.optimizer.load_state_dict(state_dict['optimizer'])
self.scaler.load_state_dict(state_dict['scaler'])
self.step = state_dict['step']
def save_to_checkpoint(self, filename='weights'):
save_basename = f'{filename}-{self.step}.pt'
save_name = f'{self.model_dir}/{save_basename}'
link_name = f'{self.model_dir}/{filename}.pt'
torch.save(self.state_dict(), save_name)
if os.name == 'nt':
torch.save(self.state_dict(), link_name)
else:
if os.path.islink(link_name):
os.unlink(link_name)
os.symlink(save_basename, link_name)
def restore_from_checkpoint(self, filename='weights'):
try:
checkpoint = torch.load(f'{self.model_dir}/{filename}.pt')
self.load_state_dict(checkpoint)
return True
except FileNotFoundError:
return False
def train(self, max_steps=None):
device = next(self.model.parameters()).device
while True:
for features in tqdm(self.dataset, desc=f'Epoch {self.step // len(self.dataset)}') if self.is_master else self.dataset:
if max_steps is not None and self.step >= max_steps:
return
features = _nested_map(features, lambda x: x.to(device) if isinstance(x, torch.Tensor) else x)
loss = self.train_step(features)
if torch.isnan(loss).any():
raise RuntimeError(f'Detected NaN loss at step {self.step}.')
if self.is_master:
if self.step % 100 == 0:
self._write_summary(self.step, features, loss)
if self.step % len(self.dataset) == 0:
self.save_to_checkpoint()
self.step += 1
def train_step(self, features):
for param in self.model.parameters():
param.grad = None
audio = features['audio']
spectrogram = features['spectrogram']
N, T = audio.shape
S = 1000
device = audio.device
self.noise_level = self.noise_level.to(device)
with self.autocast:
s = torch.randint(1, S + 1, [N], device=audio.device)
l_a, l_b = self.noise_level[s-1], self.noise_level[s]
noise_scale = l_a + torch.rand(N, device=audio.device) * (l_b - l_a)
noise_scale = noise_scale.unsqueeze(1)
noise = torch.randn_like(audio)
noisy_audio = noise_scale * audio + (1.0 - noise_scale**2)**0.5 * noise
predicted = self.model(noisy_audio, spectrogram, noise_scale.squeeze(1))
loss = self.loss_fn(noise, predicted.squeeze(1))
self.scaler.scale(loss).backward()
self.scaler.unscale_(self.optimizer)
self.grad_norm = nn.utils.clip_grad_norm_(self.model.parameters(), self.params.max_grad_norm)
self.scaler.step(self.optimizer)
self.scaler.update()
return loss
def _write_summary(self, step, features, loss):
writer = self.summary_writer or SummaryWriter(self.model_dir, purge_step=step)
writer.add_audio('audio/reference', features['audio'][0], step, sample_rate=self.params.sample_rate)
writer.add_scalar('train/loss', loss, step)
writer.add_scalar('train/grad_norm', self.grad_norm, step)
writer.flush()
self.summary_writer = writer
def _train_impl(replica_id, model, dataset, args, params):
torch.backends.cudnn.benchmark = True
opt = torch.optim.Adam(model.parameters(), lr=params.learning_rate)
learner = WaveGradLearner(args.model_dir, model, dataset, opt, params, fp16=args.fp16)
learner.is_master = (replica_id == 0)
learner.restore_from_checkpoint()
learner.train(max_steps=args.max_steps)
def train(args, params):
dataset = dataset_from_path(args.data_dirs, params)
model = WaveGrad(params).cuda()
_train_impl(0, model, dataset, args, params)
def train_distributed(replica_id, replica_count, port, args, params):
os.environ['MASTER_ADDR'] = 'localhost'
os.environ['MASTER_PORT'] = str(port)
torch.distributed.init_process_group('nccl', rank=replica_id, world_size=replica_count)
device = torch.device('cuda', replica_id)
torch.cuda.set_device(device)
model = WaveGrad(params).to(device)
model = DistributedDataParallel(model, device_ids=[replica_id])
_train_impl(replica_id, model, dataset_from_path(args.data_dirs, params, is_distributed=True), args, params)
|
[
"wavegrad.model.WaveGrad",
"torch.nn.L1Loss",
"numpy.array",
"wavegrad.dataset.from_path",
"os.path.islink",
"torch.isnan",
"torch.utils.tensorboard.SummaryWriter",
"torch.randint",
"os.unlink",
"numpy.concatenate",
"torch.nn.parallel.DistributedDataParallel",
"torch.randn_like",
"torch.cuda.set_device",
"torch.device",
"os.makedirs",
"numpy.cumprod",
"torch.load",
"os.symlink",
"torch.distributed.init_process_group",
"torch.rand"
] |
[((6383, 6424), 'wavegrad.dataset.from_path', 'dataset_from_path', (['args.data_dirs', 'params'], {}), '(args.data_dirs, params)\n', (6400, 6424), True, 'from wavegrad.dataset import from_path as dataset_from_path\n'), ((6662, 6754), 'torch.distributed.init_process_group', 'torch.distributed.init_process_group', (['"""nccl"""'], {'rank': 'replica_id', 'world_size': 'replica_count'}), "('nccl', rank=replica_id, world_size=\n replica_count)\n", (6698, 6754), False, 'import torch\n'), ((6762, 6794), 'torch.device', 'torch.device', (['"""cuda"""', 'replica_id'], {}), "('cuda', replica_id)\n", (6774, 6794), False, 'import torch\n'), ((6797, 6826), 'torch.cuda.set_device', 'torch.cuda.set_device', (['device'], {}), '(device)\n', (6818, 6826), False, 'import torch\n'), ((6875, 6930), 'torch.nn.parallel.DistributedDataParallel', 'DistributedDataParallel', (['model'], {'device_ids': '[replica_id]'}), '(model, device_ids=[replica_id])\n', (6898, 6930), False, 'from torch.nn.parallel import DistributedDataParallel\n'), ((1409, 1446), 'os.makedirs', 'os.makedirs', (['model_dir'], {'exist_ok': '(True)'}), '(model_dir, exist_ok=True)\n', (1420, 1446), False, 'import os\n'), ((1798, 1834), 'numpy.array', 'np.array', (['self.params.noise_schedule'], {}), '(self.params.noise_schedule)\n', (1806, 1834), True, 'import numpy as np\n'), ((1897, 1941), 'numpy.concatenate', 'np.concatenate', (['[[1.0], noise_level]'], {'axis': '(0)'}), '([[1.0], noise_level], axis=0)\n', (1911, 1941), True, 'import numpy as np\n'), ((2029, 2040), 'torch.nn.L1Loss', 'nn.L1Loss', ([], {}), '()\n', (2038, 2040), True, 'import torch.nn as nn\n'), ((6964, 7026), 'wavegrad.dataset.from_path', 'dataset_from_path', (['args.data_dirs', 'params'], {'is_distributed': '(True)'}), '(args.data_dirs, params, is_distributed=True)\n', (6981, 7026), True, 'from wavegrad.dataset import from_path as dataset_from_path\n'), ((1853, 1873), 'numpy.cumprod', 'np.cumprod', (['(1 - beta)'], {}), '(1 - beta)\n', (1863, 1873), True, 'import numpy as np\n'), ((3376, 3401), 'os.path.islink', 'os.path.islink', (['link_name'], {}), '(link_name)\n', (3390, 3401), False, 'import os\n'), ((3438, 3474), 'os.symlink', 'os.symlink', (['save_basename', 'link_name'], {}), '(save_basename, link_name)\n', (3448, 3474), False, 'import os\n'), ((3561, 3606), 'torch.load', 'torch.load', (['f"""{self.model_dir}/{filename}.pt"""'], {}), "(f'{self.model_dir}/{filename}.pt')\n", (3571, 3606), False, 'import torch\n'), ((4825, 4874), 'torch.randint', 'torch.randint', (['(1)', '(S + 1)', '[N]'], {'device': 'audio.device'}), '(1, S + 1, [N], device=audio.device)\n', (4838, 4874), False, 'import torch\n'), ((5069, 5092), 'torch.randn_like', 'torch.randn_like', (['audio'], {}), '(audio)\n', (5085, 5092), False, 'import torch\n'), ((5650, 5696), 'torch.utils.tensorboard.SummaryWriter', 'SummaryWriter', (['self.model_dir'], {'purge_step': 'step'}), '(self.model_dir, purge_step=step)\n', (5663, 5696), False, 'from torch.utils.tensorboard import SummaryWriter\n'), ((6435, 6451), 'wavegrad.model.WaveGrad', 'WaveGrad', (['params'], {}), '(params)\n', (6443, 6451), False, 'from wavegrad.model import WaveGrad\n'), ((6837, 6853), 'wavegrad.model.WaveGrad', 'WaveGrad', (['params'], {}), '(params)\n', (6845, 6853), False, 'from wavegrad.model import WaveGrad\n'), ((3411, 3431), 'os.unlink', 'os.unlink', (['link_name'], {}), '(link_name)\n', (3420, 3431), False, 'import os\n'), ((4961, 4995), 'torch.rand', 'torch.rand', (['N'], {'device': 'audio.device'}), '(N, device=audio.device)\n', (4971, 4995), False, 'import torch\n'), ((4174, 4191), 'torch.isnan', 'torch.isnan', (['loss'], {}), '(loss)\n', (4185, 4191), False, 'import torch\n')]
|
# Import necessary packages here
from typing import List
import warnings
from datetime import datetime
import pandas as pd
import numpy as np
import matplotlib.dates as mdates
from matplotlib import rc, pyplot as plt
# ============================================================================
# ============================================================================
# Date: December 18, 2020
# Purpose: This file contains classes and functions necessary for
# plotting.
# Source Code Metadata
__author__ = "<NAME>"
__copyright__ = "Copyright 2020, Jon Webb Inc."
__version__ = "1.0"
# ============================================================================
# ============================================================================
def text_date_plot(dates: List[List[str]], y_data: List[List[float]],
line_colors: List[str], line_style: List[str],
line_weight: List[str], x_label: str, y_label: str,
dat_labels: List[str], label_pos: str, y_scale: str = 'LIN',
plot_name: str = 'NULL', save: bool = False,
label_font_size: int = 18, tick_font_size: int = 18,
style_name: str = 'default', title: str = 'NULL',
title_font_size: int = 24) -> None:
"""
:param dates: A list of lists, where each inner list contains a list of dates
as a text string in the format YYYY-MM-DD or YYYY/MM/DD
:param y_data: A list of lists containing y-axis data corresponding to the
list of lists in `dates`
:param line_colors: A list of line colors ,one for each curve.
Acceptable line color indicators can be found in documentation
for
matplotlib colors <https://matplotlib.org/3.1.0/gallery/color/named_colors.html>`_.
:param line_style: A list of line styles, one for each curve. Acceptable line
styles can be found in documentation for
`matplotlib style <https://matplotlib.org/3.1.0/gallery/lines_bars_and_markers/linestyles.html>`_.
:param line_weight: A list of line weights, one for each curve.
:param x_label: The x-axis label
:param y_label: The y-axis label
:param dat_labels: A list of labels, one for each curve
:param label_pos: The position of the label in the plot, examples might be
``upper left``, ``lower right``.
:param y_scale: 'LOG' or 'LIN' for logarithmic or linear scale
:param plot_name: The plot name and path-link, if the user wants to save the
plot. If not, the variable is defaulted to ``NULL``
:param save: True or False, defaulted to False
:param label_font_size: The font size for plot labels, defaulted to 18
:param tick_font_size: The tick font size, defaulted to 18
:param style_name: The plot style to be used. Acceptable styles can be
found at
`matplotlib styles <https://matplotlib.org/3.2.1/gallery/style_sheets/style_sheets_reference.html>`_.
defaulted to ``default``
:param title: The title of the plot to incorporate into the header. Defaulted to NULL
:param title_font_size: The font size for the tile, defaulted to 24
:return None:
This function utilizes the matplotlib
`subplots <https://matplotlib.org/3.3.3/api/_as_gen/matplotlib.pyplot.subplots.html>`_ functionality
to produce single plots of one or multiple data sets as a function of date. This function assumes that the
date string is in the format of a text string and not a Timestamp or datetime. This function also autonomusly
determines the appropriate date display format. If you desire plots as a
function of time you should use the ``text_time_plot`` function. The function can be used in the
following manner;
.. code-block:: python
> # Use stock data for example
> tickers = ['AAPL', 'WMT']
> data = yf.download(tickers, '2015-1-1')['Adj Close']
> # transform Timestamps to string
> dates = list(data.index.strftime('%Y-%m-%d'))
> date_list = [dates, dates]
> y_list = [list(data[tickers[0]]), list(data[tickers[1]])]
> colors = ['red', 'green']
> line_style = ['-', '-']
> weight = [1.0, 1.0]
> text_date_plot(date_list, y_list, colors, line_style, weight, 'Date',
'$', tickers, 'upper left')
.. image:: date.eps
:align: center
"""
# Adjust format for YYYY/MM/DD to YYYY-MM-DD
outer_list = []
for i in range(len(dates)):
inner_list = []
for j in range(len(dates[i])):
year = dates[i][j][0:4]
month = dates[i][j][5:7]
day = dates[i][j][8:10]
date_string = year + '-' + month + '-' + day
inner_list.append(datetime.strptime(date_string, '%Y-%m-%d'))
outer_list.append(inner_list)
# Determine time difference between min and max point
days = 0
for i in outer_list:
delta = (max(i) - min(i)).days
if delta > days:
days = delta
# Start plot
fig, td_plot = plt.subplots()
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
if y_scale.upper() == 'LOG':
td_plot.set_yscale('log')
if days <= 15:
myfmt = mdates.DateFormatter('%d')
td_plot.xaxis.set_major_locator(mdates.DayLocator())
elif days <= 180:
myfmt = mdates.DateFormatter('%b-%y')
td_plot.xaxis.set_major_locator(mdates.MonthLocator())
else:
myfmt = mdates.DateFormatter('%b-%y')
td_plot.xaxis.set_major_locator(plt.MaxNLocator(4))
td_plot.set_xlabel(x_label, fontsize=label_font_size)
td_plot.set_ylabel(y_label, fontsize=label_font_size)
if title != 'NULL':
td_plot.set_title(title, fontsize=title_font_size)
td_plot.xaxis.set_major_formatter(myfmt)
for i in range(len(outer_list)):
td_plot.plot(outer_list[i], y_data[i], color=line_colors[i],
label=dat_labels[i], linewidth=line_weight[i],
linestyle=line_style[i])
plt.legend(loc=label_pos)
if not save:
plt.show()
else:
plt.savefig(plot_name)
# ----------------------------------------------------------------------------
def two_d_line_matplot(x_data: List[List[float]], y_data: List[List[float]],
line_colors: List[str], line_style: List[str],
line_weight: List[str], x_label: str, y_label: str,
dat_labels: List[str], label_pos: str, x_scale: str = 'LIN',
y_scale: str = 'LIN', plot_name: str = 'NULL',
save: bool = False, label_font_size: int = 18,
tick_font_size: int = 18, style_name: str = 'default',
title: str = 'NULL', title_font_size: int = 24) -> None:
"""
:param x_data: A list of lists, where the inner lists contain data points
for the x-axis
:param y_data: A list of lists, where the inner lists contain data points
for the y-axis
:param line_colors: A list of line colors ,one for each curve.
Acceptable line color indicators can be found in documentation
for
matplotlib colors <https://matplotlib.org/3.1.0/gallery/color/named_colors.html>`_.
:param line_style: A list of line styles, one for each curve. Acceptable line
styles can be found in documentation for
`matplotlib style <https://matplotlib.org/3.1.0/gallery/lines_bars_and_markers/linestyles.html>`_.
:param line_weight: A list of line weights, one for each curve.
:param x_label: The label for the x-axis
:param y_label: The label for the y-axis
:param dat_labels: A list of labels, one for each curve
:param label_pos: The position of the label in the plot, examples might be
``upper left``, ``lower right``.
:param x_scale: LOG or LIN for logarithmic or linear, defaulted to LIN
:param y_scale: LOG or LIN for logarithmic or linear, defaulted to LIN
:param plot_name: The plot name and path-link, if the user wants to save the
plot. If not, the variable is defaulted to ``NULL``
:param save: True or False, defaulted to False
:param label_font_size: The font size for plot labels, defaulted to 18
:param tick_font_size: The tick font size, defaulted to 18
:param style_name: The plot style to be used. Acceptable styles can be
found at
`matplotlib styles <https://matplotlib.org/3.2.1/gallery/style_sheets/style_sheets_reference.html>`_.
defaulted to ``default``
:param title: The title of the plot to incorporate into the header. Defaulted to NULL
:param title_font_size: The font size for the tile, defaulted to 24
:return None:
This function utilizes the matplotlib
`subplots <https://matplotlib.org/3.3.3/api/_as_gen/matplotlib.pyplot.subplots.html>`_ functionality
to produce single plots of one or multiple data sets. This function will only produce line plots and
not scatter plots or a combination of both. The function can be used in the following manner;
.. code-block:: python
> x_dat = np.linspace(0, 10, 15)
> y1_dat = x_dat
> y2_dat = x_dat ** 2.0
> y3_dat = x_dat ** 3.0
> x_list = [x_dat, x_dat, x_dat]
> y_list = [y1_dat, y2_dat, y3_dat]
> colors = ['red', 'blue', 'black']
> line_style = ['-', '-', '--']
> labels = ['linear', 'squared', 'cubed']
> weight = [1, 2, 3]
> two_d_line_matplot(x_list, y_list, colors, line_style, weight, 'x-data',
'y-data', labels, 'upper left')
.. image:: line_plot.eps
:scale: 90%
:align: center
"""
# Error checking and warnings
if save and plot_name == 'NULL':
warnings.warn('if save is True then plot name cannot be NULL')
if len(x_data) != len(y_data):
warnings.warn('length of x list of lists is not the same as y list of lists, plot not printed')
return
if len(line_colors) != len(x_data):
warnings.warn('line colors list not the same length as data lists, plot not printed')
return
if len(line_style) != len(x_data):
warnings.warn('line_style list not the same length as data lists, plot not printed')
return
if len(line_weight) != len(x_data):
warnings.warn('line_weight list not the same length as data lists, plot not printed')
return
if y_scale != 'LOG' and y_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
if x_scale != 'LOG' and x_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
# begin plot
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
fig, td_plot = plt.subplots()
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
td_plot.set_xlabel(x_label, fontsize=label_font_size)
td_plot.set_ylabel(y_label, fontsize=label_font_size)
if title != 'NULL':
td_plot.set_title(title, fontsize=title_font_size)
if x_scale.upper() == 'LOG':
td_plot.set_xscale('log')
if y_scale.upper() == 'LOG':
td_plot.set_yscale('log')
for i in range(len(line_colors)):
td_plot.plot(x_data[i], y_data[i], color=line_colors[i],
label=dat_labels[i], linewidth=line_weight[i],
linestyle=line_style[i])
plt.legend(loc=label_pos)
if not save:
plt.show()
else:
plt.savefig(plot_name)
# ----------------------------------------------------------------------------
def two_d_scatter_matplot(x_data: List[List[float]], y_data: List[List[float]],
marker_colors: List[str], marker_style: List[str],
x_label: str, y_label: str, dat_labels: List[str],
label_pos: str, x_scale: str = 'LIN',
y_scale: str = 'LIN', plot_name: str = 'NULL',
save: bool = False, label_font_size: int = 18,
tick_font_size: int = 18, style_name: str = 'default',
title: str = 'NULL', title_font_size: int = 24) -> None:
"""
:param x_data: A list of lists, where the inner lists contain data points
for the x-axis
:param y_data: A list of lists, where the inner lists contain data points
for the y-axis
:param marker_colors: A list of line colors ,one for each curve.
Acceptable line color indicators can be found in documentation
for `matplotlib colors <https://matplotlib.org/3.1.0/gallery/color/named_colors.html>`_.
:param marker_style: A list of line styles, one for each curve. Acceptable line
styles can be found in documentation for `matplotlib style`_.
:param x_label: The label for the x-axis
:param y_label: The label for the y-axis
:param dat_labels: A list of labels, one for each curve
:param label_pos: The position of the label in the plot, examples might be
``upper left``, ``lower right``
:param x_scale: LOG or LIN for logarithmic or linear, defaulted to LIN
:param y_scale: LOG or LIN for logarithmic or linear, defaulted to LIN
:param plot_name: The plot name and path-link, if the user wants to save the
plot. If not, the variable is defaulted to ``NULL``
:param save: True or False, defaulted to False
:param label_font_size: The font size for plot labels, defaulted to 18
:param tick_font_size: The tick font size, defaulted to 18
:param style_name: The plot style to be used. Acceptable styles can be
found at
`matplotlib styles <https://matplotlib.org/3.2.1/gallery/style_sheets/style_sheets_reference.html>`_.
defaulted to ``default``
:param title: The title of the plot to incorporate into the header. Defaulted to NULL
:param title_font_size: The font size for the tile, defaulted to 24
:return None:
This function utilizes the matplotlib
`subplots <https://matplotlib.org/3.3.3/api/_as_gen/matplotlib.pyplot.subplots.html>`_ functionality
to produce single plots of one or multiple data sets. This function will only produce line plots and
not scatter plots or a combination of both. The function can be used in the following manner;
.. code-block:: python
> x_dat = np.linspace(0, 10, 15)
> y1_dat = x_dat
> y2_dat = x_dat ** 2.0
> y3_dat = x_dat ** 3.0
> x_list = [x_dat, x_dat, x_dat]
> y_list = [y1_dat, y2_dat, y3_dat]
> colors = ['red', 'blue', 'black']
> line_style = ['-', '-', '--']
> labels = ['linear', 'squared', 'cubed']
> weight = [1, 2, 3]
> two_d_scatter_matplot(x_list, y_list, colors, line_style, weight, 'x-data',
'y-data', labels, 'upper left')
.. image:: scatter_plot.eps
:align: center
"""
# Error checking and warnings
if save and plot_name == 'NULL':
warnings.warn('if save is True then plot name cannot be NULL')
if len(x_data) != len(y_data):
warnings.warn('length of x list of lists is not the same as y list of lists, plot not printed')
return
if len(marker_colors) != len(x_data):
warnings.warn('line colors list not the same length as data lists, plot not printed')
return
if len(marker_style) != len(x_data):
warnings.warn('line_style list not the same length as data lists, plot not printed')
return
if y_scale != 'LOG' and y_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
if x_scale != 'LOG' and x_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
# begin plot
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
fig, td_plot = plt.subplots()
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
td_plot.set_xlabel(x_label, fontsize=label_font_size)
td_plot.set_ylabel(y_label, fontsize=label_font_size)
if title != 'NULL':
td_plot.set_title(title, fontsize=title_font_size)
if x_scale.upper() == 'LOG':
td_plot.set_xscale('log')
if y_scale.upper() == 'LOG':
td_plot.set_yscale('log')
for i in range(len(marker_colors)):
td_plot.plot(x_data[i], y_data[i], color=marker_colors[i],
label=dat_labels[i], marker=marker_style[i],
linestyle=' ')
plt.legend(loc=label_pos)
if not save:
plt.show()
else:
plt.savefig(plot_name)
# ----------------------------------------------------------------------------
def two_d_scatter_line_matplot(x_data: List[List[float]], y_data: List[List[float]],
marker_colors: List[str], marker_style: List[str],
line_style: List[str], line_weight: List[str],
x_label: str, y_label: str, dat_labels: List[str],
label_pos: str, x_scale: str = 'LIN',
y_scale: str = 'LIN', plot_name: str = 'NULL',
save: bool = False, label_font_size: int = 18,
tick_font_size: int = 18, style_name: str = 'default',
title: str = 'NULL', title_font_size: int = 24) -> None:
"""
:param x_data: A list of lists, where the inner lists contain data points
for the x-axis
:param y_data: A list of lists, where the inner lists contain data points
for the y-axis
:param marker_colors: A list of line colors ,one for each curve.
Acceptable line color indicators can be found in documentation
for `matplotlib colors <https://matplotlib.org/3.1.0/gallery/color/named_colors.html>`_.
:param marker_style: A list of line styles, one for each curve. Acceptable line
styles can be found in documentation for `matplotlib style`_.
:param line_style: A list of line styles, one for each curve. Acceptable line
styles can be found in documentation for
`matplotlib style <https://matplotlib.org/3.1.0/gallery/lines_bars_and_markers/linestyles.html>`_.
:param line_weight: A list of line weights, one for each curve.
:param x_label: The label for the x-axis
:param y_label: The label for the y-axis
:param dat_labels: A list of labels, one for each curve
:param label_pos: The position of the label in the plot, examples might be
``upper left``, ``lower right``
:param x_scale: LOG or LIN for logarithmic or linear, defaulted to LIN
:param y_scale: LOG or LIN for logarithmic or linear, defaulted to LIN
:param plot_name: The plot name and path-link, if the user wants to save the
plot. If not, the variable is defaulted to ``NULL``
:param save: True or False, defaulted to False
:param label_font_size: The font size for plot labels, defaulted to 18
:param tick_font_size: The tick font size, defaulted to 18
:param style_name: The plot style to be used. Acceptable styles can be
found at
`matplotlib styles <https://matplotlib.org/3.2.1/gallery/style_sheets/style_sheets_reference.html>`_.
defaulted to ``default``
:param title: The title of the plot to incorporate into the header. Defaulted to NULL
:param title_font_size: The font size for the tile, defaulted to 24
:return None:
This function utilizes the matplotlib
`subplots <https://matplotlib.org/3.3.3/api/_as_gen/matplotlib.pyplot.subplots.html>`_ functionality
to produce single plots of one or multiple data sets overlaid with line plots. This function will only produce
line plots and not scatter plots or a combination of both. The function can be used in the following manner;
.. code-block:: python
> x_dat = np.linspace(0, 10, 15)
> y1_dat = x_dat
> y2_dat = x_dat ** 2.0
> y3_dat = x_dat ** 3.0
> x_list = [x_dat, x_dat, x_dat]
> y_list = [y1_dat, y2_dat, y3_dat]
> colors = ['red', 'blue', 'black']
> line_style = ['-', '-', '--']
> labels = ['linear', 'squared', 'cubed']
> weight = [1, 2, 3]
> marker_style = ['^', 'o', 'd']
> two_d_scatter_line_matplot(x_list, y_list, colors, marker_style,
line_style, weight, 'x-axis', 'y-axis',
labels, 'upper left', save=True, plot_name=plt_name)
.. image:: line_mark.eps
:align: center
"""
# Error checking and warnings
if save and plot_name == 'NULL':
warnings.warn('if save is True then plot name cannot be NULL')
if len(x_data) != len(y_data):
warnings.warn('length of x list of lists is not the same as y list of lists, plot not printed')
return
if len(marker_colors) != len(x_data):
warnings.warn('line colors list not the same length as data lists, plot not printed')
return
if len(marker_style) != len(x_data):
warnings.warn('line_style list not the same length as data lists, plot not printed')
return
if y_scale != 'LOG' and y_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
if x_scale != 'LOG' and x_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
# begin plot
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
fig, td_plot = plt.subplots()
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
if title != 'NULL':
td_plot.set_title(title, fontsize=title_font_size)
td_plot.set_xlabel(x_label, fontsize=label_font_size)
td_plot.set_ylabel(y_label, fontsize=label_font_size)
if x_scale.upper() == 'LOG':
td_plot.set_xscale('log')
if y_scale.upper() == 'LOG':
td_plot.set_yscale('log')
for i in range(len(marker_colors)):
td_plot.plot(x_data[i], y_data[i], color=marker_colors[i],
label=dat_labels[i], marker=marker_style[i],
linestyle=line_style[i], linewidth=line_weight[i])
plt.legend(loc=label_pos)
if not save:
plt.show()
else:
plt.savefig(plot_name)
# ----------------------------------------------------------------------------
def one_d_histogram_plot(data: List[List[float]], labels: List[List[str]],
x_label: str, y_label: str, colors: List[str],
edge_colors: List[str], shading: List[float],
label_pos: str, num_bins: int = 50, tick_font_size: int = 18,
label_font_size: str = 18, style_name: str = 'default',
save: bool = False, plot_name: str = 'NULL',
hist_type: str = 'bar', dens: bool = False,
title: str = 'NULL', title_font_size: int = 24) -> None:
"""
:param data: A list of lists containing data for one or multiple
distributions
:param labels: A list of labels, one for each distribution
:param x_label: The label for the x-axis
:param y_label: The label for the y-axis
:param colors: The fill colors for each ``bar`` plot. If a ``step`` plot
is selected, this input is irrelevant, but data must still be
passed to the function.
:param edge_colors: The colors for the edge of each bar or step plot
:param shading: The level of transparency for bar plot fill. a Value of
0 is invisible, 1 is the maximum color density
:param label_pos: Where in the plot, the labels for each curve are to be
placed. ``upper left`` or ``lower right`` are examples.
:param num_bins: The number of bins to be plotted, defaulted to 50
:param tick_font_size: The size for each tick, defaulted to 18
:param label_font_size: The size for printed font, defaulted to 18
:param style_name: The plot style to be used. Acceptable styles can be
found at
`matplotlib styles <https://matplotlib.org/3.2.1/gallery/style_sheets/style_sheets_reference.html>`_.
defaulted to ``default``
:param save: True or False, defaulted to False
:param plot_name: The plot name and path-link, if the user wants to save the
plot. If not, the variable is defaulted to ``NULL``
:param hist_type: {``bar``, ``barstacked``, ``step``, ``stepfilled``}
See
`histogram <https://matplotlib.org/3.1.1/api/_as_gen/matplotlib.pyplot.hist.html>`_
for more information.
:param dens: If True, the first element of the return tuple will be the counts
normalized to form a probability density, i.e., the area (or integral)
under the histogram will sum to 1
:param title: The title of the plot to incorporate into the header. Defaulted to NULL
:param title_font_size: The font size for the tile, defaulted to 24
:return:
This function utilizes the matplotlib
`subplots <https://matplotlib.org/3.3.3/api/_as_gen/matplotlib.pyplot.subplots.html>`_ functionality
to produce single phistogram plots or multiple overlaid plots. The function can be used in the following manner;
.. code-block:: python
> np.random.seed(19680801)
> x = np.random.normal(15.0, 3.0, 1000)
> y = np.random.normal(20.0, 3.0, 1000)
> data = [x, y]
> labels = ['one', 'two']
> colors = ['blue', 'green']
> edge_colors = ['black', 'black']
> alpha = [0.9, 0.2]
> x_label = 'x-axis'
> y_label = 'y-axis'
> one_d_histogram_plot(data, labels, x_label, y_label, colors, edge_colors,
alpha, 'upper left', num_bins=50, hist_type='step',
dens=True)
.. image:: hist1.eps
:align: center
The plot parameters can be changed to produce a normalized plot, only
showing the histogram outline with the following code.
.. code-block:: python
> np.random.seed(19680801)
> x = np.random.normal(15.0, 3.0, 1000)
> y = np.random.normal(20.0, 3.0, 1000)
> data = [x, y]
> labels = ['one', 'two']
> colors = ['black', 'red']
> edge_colors = ['black', 'red']
> alpha = [1.0, 1.0]
> x_label = 'x-axis'
> y_label = 'y-axis'
> one_d_histogram_plot(data, labels, x_label, y_label, colors, edge_colors,
alpha, 'upper left', num_bins=50)
.. image:: hist2.eps
:align: center
"""
if len(labels) != len(data):
warnings.warn("data list should be the same length as the labels list")
if len(labels) != len(colors):
warnings.warn("data list should be the same length as the colors list")
if len(labels) != len(edge_colors):
warnings.warn("labels list should be the same length as the edge_colors list")
if len(labels) != len(shading):
warnings.warn("labels list should be the same length as the shading list")
plt.tight_layout()
plt.gcf().subplots_adjust(bottom=0.15)
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
plt.xlabel(x_label, fontsize=label_font_size)
plt.ylabel(y_label, fontsize=label_font_size)
if title != 'NULL':
plt.title(title, fontsize=title_font_size)
for i in range(len(labels)):
plt.hist(data[i], bins=num_bins, color=colors[i], edgecolor=edge_colors[i],
alpha=shading[i], label=labels[i], histtype=hist_type, density=dens)
plt.legend(loc=label_pos)
if not save:
plt.show()
else:
plt.savefig(plot_name)
plt.close()
# ================================================================================
# ================================================================================
class MatPlotDataFrame:
"""
:param df: Dataframe containing columnar data to be plotted
This class will plot user specified data from a pandas dataframe
"""
def __init__(self, df: pd.DataFrame):
self.df = df
self.colors = ['lightgrey', 'deepskyblue', 'sandybrown',
'teal', 'limegreen', 'coral',
'hotpink', 'magenta', 'red',
'white', 'gold', 'darkgreen',
'turqoise', 'olive', 'orange',
'mediumvioletred', 'purple' , 'darkred']
self.styles = ['o' for i in range(len(self.colors))]
# --------------------------------------------------------------------------------
def scatter_plot_parse_column(self, x_header: str, y_header: str, parsing_header: str,
column_values: List[str], style_name: str='default',
marker_colors: List[str]=['None'], marker_style: List[str]=['None'],
fill_alpha: np.float32=0.7, edge_color: str='black', x_label: str='',
y_label: str='', title: str='', label_pos: str='upper right',
x_scale: str='LIN', y_scale: str='LIN', plot_name: str='NULL',
save: bool=False, label_font_size: int=18,
tick_font_size: int=18, title_font_size: int=24,
marker_size: int=35, marker_edge_width: np.float32=0.8,
grid: bool=False, grid_style='-', grid_color='grey') -> None:
"""
:param x_header: The title of the dataframe column containing the x-axis
data sets
:param y_header: The title of the dataframe column containing the y-axis
data sets
:param parsing_header: The title of the dataframe column containing the
values which will be used to parse the dataframe into
one or multiple data sets
:param column_values: The values contained in the parsing_header column
that will be used to parse the data set into
multiple data sets
:param style_name: The name of the matplotlib style that will be used to
format the plot. Defaulted to 'default'. Possible
styles can be found at :href
`styles<https://matplotlib.org/stable/api/style_api.html>`
:param marker_colors: A list of marker colors, where each marker color
corresponds to each data set. This parameter has a
default color lists that can accomodate 18 different
data sets. The user can override the default colors
with a list of their own. Potential colors can be
found at :href `colors<https://matplotlib.org/stable/gallery/color/named_colors.html>`
:param marker_style: A list of marker styles, where each marker style corresponds
to a data set. This parameter has a default list of 18 circle
marker styles that the user can override. Marker styles
can be found at :href `marker style<https://matplotlib.org/stable/api/markers_api.html>`
:param fill_apha: The density of the marker fill. Defaulted to 0.7
:param edge_color: The color of the line surrounding the marker
:param x_label: The x axis label,defaulted to ' '
:param y_label: The y axis label, defaulted to ' '
:param title: The plot title, defaulted to ' '
:param label_pos: The position of the legend in the plot. Defaulted to 'upper right'
:param x_scale: 'LOG' or 'LIN', defaulted to 'LIN'
:param y_scale: 'LOG' or 'LIN', defaulted to 'LIN'
:param plot_name: The name of the file containing the plot if the plot is to
be saved. Defaulted to 'NULL'
:param save: True if the plot is to be saved, False if the plot is to be
shown and not saved. Defaulted to False
:param label_font_size: The label font size, defaulted to 18
:param tick_font_size: The tick font size, defaulted to 18
:param title_font_size: The title font size, defaulted to 24
:param marker_size: The size of the marker, defaulted to 35
:param marker_edge_width: The thickness of the line outlining
each marker. Defaulted to 0.8
:param grid: True if a grid overlaid on the plot is desired, False if not
:param grid_color: Defaulted to 'grey'
:grid_style: Defaulted to '-'
This method will parse a dataframe column based on a user specified
value or list of values, and plot the data in a user specified
x and y axis column based on filter data. As an example, consider
a dataframe with the following columnar data structure.
.. code-block:: python
> length = 20
> x = np.linspace(0, length, num=length)
> linear = x
> squared = x ** 2.0
> lin = np.repeat('linear', length)
> sq = np.repeat('squared', length)
> # Combine arrays into one
> x = np.hstack((x, x))
> y = np.hstack((linear, squared))
> power = np.hstack((lin, sq))
> # Create dataframe
> dictionary = {'x': x, 'y': y, 'power': power}
> df = pd.DataFrame(dictionary)
> # Plot data
> obj = MatPlotDataFrame(df)
> parsing_header = 'power'
> column_values = ['linear', 'squared']
obj.scatter_plot_filter_column('x', 'y', parsing_header,
column_values,
marker_colors=['red', 'green'],
marker_style=['o', '^'],
label_pos='upper left')
.. image:: mat_scatter_test1.eps
:align: center
"""
df_list = [self.df[self.df[parsing_header] == col_val] for
col_val in column_values]
# Error checking
if marker_colors[0] == 'None':
marker_colors = self.colors
if len(marker_colors) < len(column_values):
msg1 = 'FATAL ERROR: The length of the marker color list must be as '
msg2 = 'large or larger than the size of the column values'
sys.exit(msg + ms2)
if marker_style[0] == 'None':
marker_style = self.styles
if len(marker_style) < len(column_values):
msg1 = 'FATAL ERROR: The length of the marker stye list must be as '
msg2 = 'large or larger than the size of the column values'
sys.exit(msg + ms2)
if save and plot_name == 'NULL':
warnings.warn('if save is True then plot name cannot be NULL')
if y_scale != 'LOG' and y_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
if x_scale != 'LOG' and x_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
# begin plot
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
fig, td_plot = plt.subplots()
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
td_plot.set_xlabel(x_label, fontsize=label_font_size)
td_plot.set_ylabel(y_label, fontsize=label_font_size)
if title != 'NULL':
td_plot.set_title(title, fontsize=title_font_size)
if x_scale.upper() == 'LOG':
td_plot.set_xscale('log')
if y_scale.upper() == 'LOG':
td_plot.set_yscale('log')
for i in range(len(df_list)):
td_plot.scatter(df_list[i][x_header], df_list[i][y_header],
label=column_values[i], marker=marker_style[i],
color=marker_colors[i], alpha=fill_alpha,
edgecolors=edge_color, s=marker_size,
linewidth=marker_edge_width)
plt.legend(loc=label_pos)
if grid:
plt.grid(color=grid_color, linestyle=grid_style)
if not save:
plt.show()
else:
plt.savefig(plot_name)
plt.close()
# --------------------------------------------------------------------------------
def scatter_plot_columns(self, x_headers: List[str], y_headers: List[str],
labels: List[str], style_name: str='default',
marker_colors: List[str]=['None'],
marker_style: List[str]=['None'], fill_alpha: np.float32=0.7,
edge_color: str='black', x_label: str='', y_label: str='',
title: str='', label_pos: str='upper right', x_scale: str='LIN',
y_scale: str='LIN', plot_name: str='NULL', save: bool=False,
label_font_size: int=18, tick_font_size: int=18,
title_font_size: int=24, marker_size: int=35,
marker_edge_width: np.float32=0.8, grid: bool=False,
grid_style='-', grid_color='grey'):
"""
:param x_headers: The title of the dataframe columns containing the x-axis
data sets
:param y_headers: The title of the dataframe columns containing the y-axis
data sets
:param labels: A list of the label names for each data set
:param style_name: The name of the matplotlib style that will be used to
format the plot. Defaulted to 'default'. Possible
styles can be found at :href
`styles<https://matplotlib.org/stable/api/style_api.html>`
:param marker_colors: A list of marker colors, where each marker color
corresponds to each data set. This parameter has a
default color lists that can accomodate 18 different
data sets. The user can override the default colors
with a list of their own. Potential colors can be
found at :href `colors<https://matplotlib.org/stable/gallery/color/named_colors.html>`
:param marker_style: A list of marker styles, where each marker style corresponds
to a data set. This parameter has a default list of 18 circle
marker styles that the user can override. Marker styles
can be found at :href `marker style<https://matplotlib.org/stable/api/markers_api.html>`
:param fill_apha: The density of the marker fill. Defaulted to 0.7
:param edge_color: The color of the line surrounding the marker
:param x_label: The x axis label,defaulted to ' '
:param y_label: The y axis label, defaulted to ' '
:param title: The plot title, defaulted to ' '
:param label_pos: The position of the legend in the plot. Defaulted to 'upper right'
:param x_scale: 'LOG' or 'LIN', defaulted to 'LIN'
:param y_scale: 'LOG' or 'LIN', defaulted to 'LIN'
:param plot_name: The name of the file containing the plot if the plot is to
be saved. Defaulted to 'NULL'
:param save: True if the plot is to be saved, False if the plot is to be
shown and not saved. Defaulted to False
:param label_font_size: The label font size, defaulted to 18
:param tick_font_size: The tick font size, defaulted to 18
:param title_font_size: The title font size, defaulted to 24
:param marker_size: The size of the marker, defaulted to 35
:param marker_edge_width: The thickness of the line outlining
each marker. Defaulted to 0.8
:param grid: True if a grid overlaid on the plot is desired, False if not
:param grid_color: Defaulted to 'grey'
:grid_style: Defaulted to '-'
This method will plot used defined dataframe columns for the x and
y axis of a 2-d plot as a scatter plot.
.. code-block:: python
> length = 20
> x = np.linspace(0, 20, num=20)
> linear = x
> squared = x ** 2.0
> # create dataframe
> dictionary = {'x': x, 'linear': linear, 'squared': squared}
> df = pd.DataFrame(dictionary)
> # plot data
> obj = MatPlotDataFrame(df)
> x_headers = ['x', 'x']
> y_headers = ['linear', 'squared']
> obj.scatter_plot_columns(x_headers, y_headers, y_headers,
x_label='x-axis', y_label='y-axis', title='Test',
style_name='default',marker_colors=['red', 'green'],
fill_alpha=0.7, marker_style=['o', '^'],
label_pos='upper left', grid=False, save=True,
plot_name=plt_name)
.. image:: mat_scatter_test2.eps
:align: center
"""
# Error checking
if marker_colors[0] == 'None':
marker_colors = self.colors
if len(x_headers) != len(y_headers):
sys.exit('FATAL ERROR: x and y arrays must be the same size')
if marker_style[0] == 'None':
marker_style = self.styles
if len(marker_style) < len(x_headers):
msg1 = 'FATAL ERROR: The length of the marker stye list must be as '
msg2 = 'large or larger than the size of the column values'
sys.exit(msg + ms2)
if save and plot_name == 'NULL':
warnings.warn('if save is True then plot name cannot be NULL')
if y_scale != 'LOG' and y_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
if x_scale != 'LOG' and x_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
# begin plot
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
fig, td_plot = plt.subplots()
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
td_plot.set_xlabel(x_label, fontsize=label_font_size)
td_plot.set_ylabel(y_label, fontsize=label_font_size)
if title != 'NULL':
td_plot.set_title(title, fontsize=title_font_size)
if x_scale.upper() == 'LOG':
td_plot.set_xscale('log')
if y_scale.upper() == 'LOG':
td_plot.set_yscale('log')
for i in range(len(x_headers)):
td_plot.scatter(self.df[x_headers[i]], self.df[y_headers[i]],
label=labels[i], marker=marker_style[i],
color=marker_colors[i], alpha=fill_alpha,
edgecolors=edge_color, s=marker_size,
linewidth=marker_edge_width)
plt.legend(loc=label_pos)
if grid:
plt.grid(color=grid_color, linestyle=grid_style)
if not save:
plt.show()
else:
plt.savefig(plot_name)
plt.close()
# --------------------------------------------------------------------------------
def line_plot_parse_column(self, x_header: str, y_header: str, parsing_header: str,
column_values: List[str], style_name: str='default',
line_colors: List[str]=['None'], line_weight: np.float32=2.0,
fill_alpha: np.float32=0.7, line_style: str='-', x_label: str='',
y_label: str='', title: str='', label_pos: str='upper right',
x_scale: str='LIN', y_scale: str='LIN', plot_name: str='NULL',
save: bool=False, label_font_size: int=18,
tick_font_size: int=18, title_font_size: int=24,
marker_size: int=35, marker_edge_width: np.float32=0.8,
grid: bool=False, grid_style='-', grid_color='grey') -> None:
"""
:param x_header: The title of the dataframe column containing the x-axis
data sets
:param y_header: The title of the dataframe column containing the y-axis
data sets
:param parsing_header: The title of the dataframe column containing the
values which will be used to parse the dataframe into
one or multiple data sets
:param column_values: The values contained in the parsing_header column
that will be used to parse the data set into
multiple data sets
:param style_name: The name of the matplotlib style that will be used to
format the plot. Defaulted to 'default'. Possible
styles can be found at :href
`styles<https://matplotlib.org/stable/api/style_api.html>`
:param line_colors: A list of line colors, where each marker color
corresponds to each data set. This parameter has a
default color lists that can accomodate 18 different
data sets. The user can override the default colors
with a list of their own. Potential colors can be
found at :href `colors<https://matplotlib.org/stable/gallery/color/named_colors.html>`
:param line_weight: The weight corresponding to the line thickness, defaulted to 2.0
:param fill_apha: The density of the marker fill. Defaulted to 0.7
:param x_label: The x axis label,defaulted to ' '
:param y_label: The y axis label, defaulted to ' '
:param title: The plot title, defaulted to ' '
:param label_pos: The position of the legend in the plot. Defaulted to 'upper right'
:param x_scale: 'LOG' or 'LIN', defaulted to 'LIN'
:param y_scale: 'LOG' or 'LIN', defaulted to 'LIN'
:param plot_name: The name of the file containing the plot if the plot is to
be saved. Defaulted to 'NULL'
:param save: True if the plot is to be saved, False if the plot is to be
shown and not saved. Defaulted to False
:param label_font_size: The label font size, defaulted to 18
:param tick_font_size: The tick font size, defaulted to 18
:param title_font_size: The title font size, defaulted to 24
:param marker_size: The size of the marker, defaulted to 35
:param marker_edge_width: The thickness of the line outlining
each marker. Defaulted to 0.8
:param grid: True if a grid overlaid on the plot is desired, False if not
:param grid_color: Defaulted to 'grey'
:grid_style: Defaulted to '-'
This method will parse a dataframe column based on a user specified
value or list of values, and plot the data in a user specified
x and y axis column based on filter data. As an example, consider
a dataframe with the following columnar data structure.
.. code-block:: python
> length = 20
> x = np.linspace(0, length, num=length)
> linear = x
> squared = x ** 2.0
> lin = np.repeat('linear', length)
> sq = np.repeat('squared', length)
> # Combine arrays into one
> x = np.hstack((x, x))
> y = np.hstack((linear, squared))
> power = np.hstack((lin, sq))
> # Create dataframe
> dictionary = {'x': x, 'y': y, 'power': power}
> df = pd.DataFrame(dictionary)
> # Plot data
> obj = MatPlotDataFrame(df)
> parsing_header = 'power'
> column_values = ['linear', 'squared']
obj.line_plot_filter_column('x', 'y', parsing_header,
column_values,
marker_colors=['red', 'green'],
marker_style=['o', '^'],
label_pos='upper left')
.. image:: line_scatter_test1.eps
:align: center
"""
df_list = [self.df[self.df[parsing_header] == col_val] for
col_val in column_values]
# Error checking
if line_colors[0] == 'None':
line_colors = self.colors
if len(line_colors) < len(column_values):
msg1 = 'FATAL ERROR: The length of the marker color list must be as '
msg2 = 'large or larger than the size of the column values'
sys.exit(msg + ms2)
if save and plot_name == 'NULL':
warnings.warn('if save is True then plot name cannot be NULL')
if y_scale != 'LOG' and y_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
if x_scale != 'LOG' and x_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
# begin plot
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
fig, td_plot = plt.subplots()
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
td_plot.set_xlabel(x_label, fontsize=label_font_size)
td_plot.set_ylabel(y_label, fontsize=label_font_size)
if title != 'NULL':
td_plot.set_title(title, fontsize=title_font_size)
if x_scale.upper() == 'LOG':
td_plot.set_xscale('log')
if y_scale.upper() == 'LOG':
td_plot.set_yscale('log')
for i in range(len(df_list)):
td_plot.plot(df_list[i][x_header], df_list[i][y_header],
label=column_values[i], linestyle=line_style,
color=line_colors[i], linewidth=line_weight)
plt.legend(loc=label_pos)
if grid:
plt.grid(color=grid_color, linestyle=grid_style)
if not save:
plt.show()
else:
plt.savefig(plot_name)
plt.close()
# --------------------------------------------------------------------------------
def line_plot_columns(self, x_headers: str, y_headers: str, labels: List[str],
style_name: str='default', line_colors: List[str]=['None'],
line_weight: np.float32=2.0, fill_alpha: np.float32=0.7,
line_style: str='-', x_label: str='', y_label: str='',
title: str='', label_pos: str='upper right', x_scale: str='LIN',
y_scale: str='LIN', plot_name: str='NULL', save: bool=False,
label_font_size: int=18, tick_font_size: int=18,
title_font_size: int=24, marker_size: int=35,
marker_edge_width: np.float32=0.8, grid: bool=False,
grid_style='-', grid_color='grey') -> None:
"""
:param x_headers: The title of the dataframe columns containing the x-axis
data sets
:param y_headers: The title of the dataframe columns containing the y-axis
data sets
:param labels: A list containing the name of each label
:param style_name: The name of the matplotlib style that will be used to
format the plot. Defaulted to 'default'. Possible
styles can be found at :href
`styles<https://matplotlib.org/stable/api/style_api.html>`
:param line_colors: A list of line colors, where each marker color
corresponds to each data set. This parameter has a
default color lists that can accomodate 18 different
data sets. The user can override the default colors
with a list of their own. Potential colors can be
found at :href `colors<https://matplotlib.org/stable/gallery/color/named_colors.html>`
:param line_weight: The weight corresponding to the line thickness, defaulted to 2.0
:param fill_apha: The density of the marker fill. Defaulted to 0.7
:param x_label: The x axis label,defaulted to ' '
:param y_label: The y axis label, defaulted to ' '
:param title: The plot title, defaulted to ' '
:param label_pos: The position of the legend in the plot. Defaulted to 'upper right'
:param x_scale: 'LOG' or 'LIN', defaulted to 'LIN'
:param y_scale: 'LOG' or 'LIN', defaulted to 'LIN'
:param plot_name: The name of the file containing the plot if the plot is to
be saved. Defaulted to 'NULL'
:param save: True if the plot is to be saved, False if the plot is to be
shown and not saved. Defaulted to False
:param label_font_size: The label font size, defaulted to 18
:param tick_font_size: The tick font size, defaulted to 18
:param title_font_size: The title font size, defaulted to 24
:param marker_size: The size of the marker, defaulted to 35
:param marker_edge_width: The thickness of the line outlining
each marker. Defaulted to 0.8
:param grid: True if a grid overlaid on the plot is desired, False if not
:param grid_color: Defaulted to 'grey'
:grid_style: Defaulted to '-'
This method will plot used defined dataframe columns for the x and
y axis of a 2-d plot as a line plot.
.. code-block:: python
> length = 20
> x = np.linspace(0, 20, num=20)
> linear = x
> squared = x ** 2.0
> # create dataframe
> dictionary = {'x': x, 'linear': linear, 'squared': squared}
> df = pd.DataFrame(dictionary)
> # plot data
> obj = MatPlotDataFrame(df)
> x_headers = ['x', 'x']
> y_headers = ['linear', 'squared']
> obj.line_plot_columns(x_headers, y_headers, y_headers,
x_label='x-axis', y_label='y-axis', title='Test',
style_name='default',marker_colors=['red', 'green'],
fill_alpha=0.7, marker_style=['o', '^'],
label_pos='upper left', grid=False, save=True,
plot_name=plt_name)
.. image:: line_scatter_test2.eps
:align: center
"""
# Error checking
if line_colors[0] == 'None':
line_colors = self.colors
if len(line_colors) < len(labels):
msg1 = 'FATAL ERROR: The length of the marker color list must be as '
msg2 = 'large or larger than the size of the column values'
sys.exit(msg + ms2)
if save and plot_name == 'NULL':
warnings.warn('if save is True then plot name cannot be NULL')
if y_scale != 'LOG' and y_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
if x_scale != 'LOG' and x_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
# begin plot
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
fig, td_plot = plt.subplots()
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
td_plot.set_xlabel(x_label, fontsize=label_font_size)
td_plot.set_ylabel(y_label, fontsize=label_font_size)
if title != 'NULL':
td_plot.set_title(title, fontsize=title_font_size)
if x_scale.upper() == 'LOG':
td_plot.set_xscale('log')
if y_scale.upper() == 'LOG':
td_plot.set_yscale('log')
for i in range(len(x_headers)):
td_plot.plot(self.df[x_headers[i]], self.df[y_headers[i]],
label=labels[i], linestyle=line_style,
color=line_colors[i], linewidth=line_weight)
plt.legend(loc=label_pos)
if grid:
plt.grid(color=grid_color, linestyle=grid_style)
if not save:
plt.show()
else:
plt.savefig(plot_name)
plt.close()
# --------------------------------------------------------------------------------
def timedate_plot_parse_column(self, x_header: str, y_header: str, parsing_header: str,
column_values: List[str], style_name: str='default',
line_colors: List[str]=['None'], line_weight: np.float32=2.0,
fill_alpha: np.float32=0.7, line_style: str='-', x_label: str='',
y_label: str='', title: str='', label_pos: str='upper right',
x_scale: str='LIN', y_scale: str='LIN', plot_name: str='NULL',
save: bool=False, label_font_size: int=18,
tick_font_size: int=18, title_font_size: int=24,
marker_size: int=35, marker_edge_width: np.float32=0.8,
grid: bool=False, grid_style='-', grid_color='grey'):
"""
:param x_header: The title of the dataframe column containing the x-axis
data sets. It is assumes that the x axis is the datetime
axis for this plot.
:param y_header: The title of the dataframe column containing the y-axis
data sets
:param parsing_header: The title of the dataframe column containing the
values which will be used to parse the dataframe into
one or multiple data sets
:param column_values: The values contained in the parsing_header column
that will be used to parse the data set into
multiple data sets
:param style_name: The name of the matplotlib style that will be used to
format the plot. Defaulted to 'default'. Possible
styles can be found at :href
`styles<https://matplotlib.org/stable/api/style_api.html>`
:param line_colors: A list of line colors, where each marker color
corresponds to each data set. This parameter has a
default color lists that can accomodate 18 different
data sets. The user can override the default colors
with a list of their own. Potential colors can be
found at :href `colors<https://matplotlib.org/stable/gallery/color/named_colors.html>`
:param line_weight: The weight corresponding to the line thickness, defaulted to 2.0
:param fill_apha: The density of the marker fill. Defaulted to 0.7
:param x_label: The x axis label,defaulted to ' '
:param y_label: The y axis label, defaulted to ' '
:param title: The plot title, defaulted to ' '
:param label_pos: The position of the legend in the plot. Defaulted to 'upper right'
:param x_scale: 'LOG' or 'LIN', defaulted to 'LIN'
:param y_scale: 'LOG' or 'LIN', defaulted to 'LIN'
:param plot_name: The name of the file containing the plot if the plot is to
be saved. Defaulted to 'NULL'
:param save: True if the plot is to be saved, False if the plot is to be
shown and not saved. Defaulted to False
:param label_font_size: The label font size, defaulted to 18
:param tick_font_size: The tick font size, defaulted to 18
:param title_font_size: The title font size, defaulted to 24
:param marker_size: The size of the marker, defaulted to 35
:param marker_edge_width: The thickness of the line outlining
each marker. Defaulted to 0.8
:param grid: True if a grid overlaid on the plot is desired, False if not
:param grid_color: Defaulted to 'grey'
:grid_style: Defaulted to '-'
This method will parse a dataframe column based on a user specified
value or list of values, and plot the data in a user specified
x and y axis column based on filter data. As an example, consider
a dataframe with the following columnar data structure.
.. code-block:: python
> length = 20
> x = np.linspace(0, length, num=length)
> linear = x
> squared = x ** 2.0
> lin = np.repeat('linear', length)
> sq = np.repeat('squared', length)
> # Combine arrays into one
> x = np.hstack((x, x))
> y = np.hstack((linear, squared))
> power = np.hstack((lin, sq))
> # Create dataframe
> dictionary = {'x': x, 'y': y, 'power': power}
> df = pd.DataFrame(dictionary)
> # Plot data
> obj = MatPlotDataFrame(df)
> parsing_header = 'power'
> column_values = ['linear', 'squared']
obj.line_plot_filter_column('x', 'y', parsing_header,
column_values,
marker_colors=['red', 'green'],
marker_style=['o', '^'],
label_pos='upper left')
.. image:: line_scatter_test1.eps
:align: center
"""
max_date = self.df[x_header].max()
min_date = self.df[x_header].min()
diff = (max_date - min_date) / np.timedelta64(1, 'D')
df_list = [self.df[self.df[parsing_header] == col_val] for
col_val in column_values]
df_list = [df.set_index(x_header) for df in df_list]
# Error checking
if line_colors[0] == 'None':
line_colors = self.colors
if len(line_colors) < len(column_values):
msg1 = 'FATAL ERROR: The length of the marker color list must be as '
msg2 = 'large or larger than the size of the column values'
sys.exit(msg + ms2)
if save and plot_name == 'NULL':
warnings.warn('if save is True then plot name cannot be NULL')
if y_scale != 'LOG' and y_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
if x_scale != 'LOG' and x_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
# begin plot
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
fig, td_plot = plt.subplots()
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
td_plot.set_xlabel(x_label, fontsize=label_font_size)
td_plot.set_ylabel(y_label, fontsize=label_font_size)
if title != 'NULL':
td_plot.set_title(title, fontsize=title_font_size)
if x_scale.upper() == 'LOG':
td_plot.set_xscale('log')
if y_scale.upper() == 'LOG':
td_plot.set_yscale('log')
if diff <= 2:
myfmt = mdates.DateFormatter('%H')
td_plot.xaxis.set_major_locator(plt.MaxNLocator(6))
elif diff <= 15:
myfmt = mdates.DateFormatter('%b-%d')
td_plot.xaxis.set_major_locator(plt.MaxNLocator(6))
elif diff <= 180:
myfmt = mdates.DateFormatter('%b-%Y')
td_plot.xaxis.set_major_locator(plt.MaxNLocator(5))
elif diff <= 2191:
myfmt = mdates.DateFormatter('%Y')
td_plot.xaxis.set_major_locator(plt.MaxNLocator(5))
else:
myfmt = mdates.DateFormatter('%Y')
td_plot.xaxis.set_major_locator(plt.MaxNLocator(5))
td_plot.xaxis.set_major_formatter(myfmt)
for i in range(len(df_list)):
td_plot.plot(df_list[i].index, df_list[i][y_header],
label=column_values[i], linestyle=line_style,
color=line_colors[i], linewidth=line_weight)
plt.legend(loc=label_pos)
if grid:
plt.grid(color=grid_color, linestyle=grid_style)
if not save:
plt.show()
else:
plt.savefig(plot_name)
plt.close()
# --------------------------------------------------------------------------------
def timedate_plot_columns(self, x_headers: str, y_headers: str, labels: List[str],
style_name: str='default',
line_colors: List[str]=['None'], line_weight: np.float32=2.0,
fill_alpha: np.float32=0.7, line_style: str='-', x_label: str='',
y_label: str='', title: str='', label_pos: str='upper right',
x_scale: str='LIN', y_scale: str='LIN', plot_name: str='NULL',
save: bool=False, label_font_size: int=18,
tick_font_size: int=18, title_font_size: int=24,
marker_size: int=35, marker_edge_width: np.float32=0.8,
grid: bool=False, grid_style='-', grid_color='grey'):
"""
:param x_headers: The title of the dataframe column containing the x-axis
data sets. It is assumes that the x axis is the datetime
axis for this plot.
:param y_headers: The title of the dataframe column containing the y-axis
data sets
:param labels: A list of the labels to use for each curve in the legend
:param style_name: The name of the matplotlib style that will be used to
format the plot. Defaulted to 'default'. Possible
styles can be found at :href
`styles<https://matplotlib.org/stable/api/style_api.html>`
:param line_colors: A list of line colors, where each marker color
corresponds to each data set. This parameter has a
default color lists that can accomodate 18 different
data sets. The user can override the default colors
with a list of their own. Potential colors can be
found at :href `colors<https://matplotlib.org/stable/gallery/color/named_colors.html>`
:param line_weight: The weight corresponding to the line thickness, defaulted to 2.0
:param fill_apha: The density of the marker fill. Defaulted to 0.7
:param x_label: The x axis label,defaulted to ' '
:param y_label: The y axis label, defaulted to ' '
:param title: The plot title, defaulted to ' '
:param label_pos: The position of the legend in the plot. Defaulted to 'upper right'
:param x_scale: 'LOG' or 'LIN', defaulted to 'LIN'
:param y_scale: 'LOG' or 'LIN', defaulted to 'LIN'
:param plot_name: The name of the file containing the plot if the plot is to
be saved. Defaulted to 'NULL'
:param save: True if the plot is to be saved, False if the plot is to be
shown and not saved. Defaulted to False
:param label_font_size: The label font size, defaulted to 18
:param tick_font_size: The tick font size, defaulted to 18
:param title_font_size: The title font size, defaulted to 24
:param marker_size: The size of the marker, defaulted to 35
:param marker_edge_width: The thickness of the line outlining
each marker. Defaulted to 0.8
:param grid: True if a grid overlaid on the plot is desired, False if not
:param grid_color: Defaulted to 'grey'
:grid_style: Defaulted to '-'
This method will parse a dataframe column based on a user specified
value or list of values, and plot the data in a user specified
x and y axis column based on filter data. As an example, consider
a dataframe with the following columnar data structure.
.. code-block:: python
> length = 20
> x = np.linspace(0, length, num=length)
> linear = x
> squared = x ** 2.0
> lin = np.repeat('linear', length)
> sq = np.repeat('squared', length)
> # Combine arrays into one
> x = np.hstack((x, x))
> y = np.hstack((linear, squared))
> power = np.hstack((lin, sq))
> # Create dataframe
> dictionary = {'x': x, 'y': y, 'power': power}
> df = pd.DataFrame(dictionary)
> # Plot data
> obj = MatPlotDataFrame(df)
> parsing_header = 'power'
> column_values = ['linear', 'squared']
obj.line_plot_filter_column('x', 'y', parsing_header,
column_values,
marker_colors=['red', 'green'],
marker_style=['o', '^'],
label_pos='upper left')
.. image:: line_scatter_test1.eps
:align: center
"""
diff = 0
for i in range(len(x_headers)):
max_date = self.df[x_headers[i]].max()
min_date = self.df[x_headers[i]].min()
delta = (max_date - min_date) / np.timedelta64(1, 'D')
if delta > diff:
diff = delta
# Error checking
if line_colors[0] == 'None':
line_colors = self.colors
if len(line_colors) < len(x_headers):
msg1 = 'FATAL ERROR: The length of the marker color list must be as '
msg2 = 'large or larger than the size of the column values'
sys.exit(msg + ms2)
if save and plot_name == 'NULL':
warnings.warn('if save is True then plot name cannot be NULL')
if y_scale != 'LOG' and y_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
if x_scale != 'LOG' and x_scale != 'LIN':
warnings.warn('y_scale must be set to LOG or LIN')
# begin plot
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
fig, td_plot = plt.subplots()
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
td_plot.set_xlabel(x_label, fontsize=label_font_size)
td_plot.set_ylabel(y_label, fontsize=label_font_size)
if title != 'NULL':
td_plot.set_title(title, fontsize=title_font_size)
if x_scale.upper() == 'LOG':
td_plot.set_xscale('log')
if y_scale.upper() == 'LOG':
td_plot.set_yscale('log')
if diff <= 2:
myfmt = mdates.DateFormatter('%H')
td_plot.xaxis.set_major_locator(plt.MaxNLocator(6))
elif diff <= 15:
myfmt = mdates.DateFormatter('%b-%d')
td_plot.xaxis.set_major_locator(plt.MaxNLocator(6))
elif diff <= 180:
myfmt = mdates.DateFormatter('%b-%Y')
td_plot.xaxis.set_major_locator(plt.MaxNLocator(5))
elif diff <= 2191:
myfmt = mdates.DateFormatter('%Y')
td_plot.xaxis.set_major_locator(plt.MaxNLocator(5))
else:
myfmt = mdates.DateFormatter('%Y')
td_plot.xaxis.set_major_locator(plt.MaxNLocator(5))
td_plot.xaxis.set_major_formatter(myfmt)
for i in range(len(x_headers)):
td_plot.plot(self.df[x_headers[i]], self.df[y_headers[i]],
label=labels[i], linestyle=line_style,
color=line_colors[i], linewidth=line_weight)
plt.legend(loc=label_pos)
if grid:
plt.grid(color=grid_color, linestyle=grid_style)
if not save:
plt.show()
else:
plt.savefig(plot_name)
plt.close()
# --------------------------------------------------------------------------------
def histogram_plot_parse_column(self, header: str, parsing_header: str,
column_values: List[str], x_label: str='',
y_label: str='', colors: List[str]=['None'],
edge_colors: List[str]=['None'],
shading: List[float]=['None'], label_pos: str='upper right',
num_bins: int = 50,
tick_font_size: int = 18, label_font_size: str = 18,
style_name: str = 'default', save: bool = False,
plot_name: str = 'NULL', hist_type: str = 'bar',
dens: bool = False, title: str = 'NULL',
title_font_size: int = 24) -> None:
"""
:param headers: A string representing the dataframe column that contains the
data to be parsed and plotted
:param parsing_header: A string representing the dataframe header that contains
key phrases that will be used to filter the dataframe
for specific data
:param column_values: The key phrases in the dataframe column described by the
`parsing_header` variable
:param x_label: The title for the x axis. Defaulted to ''
:param y_label: The title for the y axis. Defaulted to ''
:param colors: A list containing the colors that will be used to represent
each plot.
:param edge_colors: A list of line colors, where each marker color
corresponds to each data set. This parameter has a
default color lists that can accomodate 18 different
data sets. The user can override the default colors
with a list of their own. Potential colors can be
found at :href
`colors<https://matplotlib.org/stable/gallery/color/named_colors.html>`_
:param shading: The density of the fill for each plot, defaulted to 0.7
:param label_pos: The position of the ledgend in the plot. Defaulted to 'upper_right'
:param num_bins: The number of bins used to represent the histogram. Defaulted to 50
:param tick_font_size: The font size of the plot ticks. Defaulted to 18
:param label_font_size: The font size of plot labels. Defaulted to 18
:param style_name: The plot style, defaulted to 'default'. Acceptable styles can be
found at
`matplotlib styles <https://matplotlib.org/3.2.1/gallery/style_sheets/style_sheets_reference.html>`_.
:param save: True if the plot is to be saved, False if the plot is only to be
shown
:param plot_name: The name of the plot, if it is to be saved
:param hist_type: {``bar``, ``barstacked``, ``step``, ``stepfilled``}
See
`histogram <https://matplotlib.org/3.1.1/api/_as_gen/matplotlib.pyplot.hist.html>`_
for more information.
:param dens: If True, the first element of the return tuple will be the counts
normalized to form a probability density, i.e., the area (or integral)
under the histogram will sum to 1
:param title: The title of the plot to incorporate into the header. Defaulted to NULL
:param title_font_size: The font size for the tile, defaulted to 24
.. code-block:: python
> np.random.seed(19680801)
> x = np.random.normal(15.0, 3.0, 1000)
> y = np.random.normal(20.0, 3.0, 1000)
> data = [x, y]
> labels = ['one', 'two']
> one = np.repeat('one', len(x))
> two = np.repeat('two', len(x))
> x = np.hstack((x, y))
> y = np.hstack((one, two))
> dictionary = {'data': x, 'type': y}
> df = pd.DataFrame(dictionary)
> obj = MatPlotDataFrame(df)
> obj.histogram_plot_parse_column('data', 'type', labels, x_label='x-axis',
y_label='y-axis', shading=[0.9, 0.4], save=True,
.. image:: hist2.eps
:align: center
"""
if colors[0] == "None":
colors = self.colors
if edge_colors[0] == 'None':
edge_colors = np.repeat('black', len(column_values))
if shading[0] == "None":
shading = np.repeat(0.7, len(column_values))
df_list = [self.df[self.df[parsing_header] == col_val] for
col_val in column_values]
plt.tight_layout()
plt.gcf().subplots_adjust(bottom=0.15)
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
plt.xlabel(x_label, fontsize=label_font_size)
plt.ylabel(y_label, fontsize=label_font_size)
if title != 'NULL':
plt.title(title, fontsize=title_font_size)
if title != 'NULL':
plt.title(title, fontsize=title_font_size)
for i in range(len(column_values)):
plt.hist(df_list[i][header], bins=num_bins, color=colors[i], edgecolor=edge_colors[i],
alpha=shading[i], label=column_values[i], histtype=hist_type, density=dens)
plt.legend(loc=label_pos)
if not save:
plt.show()
else:
plt.savefig(plot_name)
plt.close()
# --------------------------------------------------------------------------------
def histogram_plot_columns(self, x_headers: List[str], labels: List[str],
x_label: str='',
y_label: str='', colors: List[str]=['None'],
edge_colors: List[str]=['None'],
shading: List[float]=['None'], label_pos: str='upper right',
num_bins: int = 50,
tick_font_size: int = 18, label_font_size: str = 18,
style_name: str = 'default', save: bool = False,
plot_name: str = 'NULL', hist_type: str = 'bar',
dens: bool = False, title: str = 'NULL',
title_font_size: int = 24) -> None:
"""
:param x_headers: A list of strings representing the dataframe columns to be
used for the x axis of a plot
:param labels: A list of labels, each label corresponding to each
histogram
:param x_label: The title for the x axis. Defaulted to ''
:param y_label: The title for the y axis. Defaulted to ''
:param colors: A list containing the colors that will be used to represent
each plot.
:param edge_colors: A list of line colors, where each marker color
corresponds to each data set. This parameter has a
default color lists that can accomodate 18 different
data sets. The user can override the default colors
with a list of their own. Potential colors can be
found at :href
`colors<https://matplotlib.org/stable/gallery/color/named_colors.html>`_
:param shading: The density of the fill for each plot, defaulted to 0.7
:param label_pos: The position of the ledgend in the plot. Defaulted to 'upper_right'
:param num_bins: The number of bins used to represent the histogram. Defaulted to 50
:param tick_font_size: The font size of the plot ticks. Defaulted to 18
:param label_font_size: The font size of plot labels. Defaulted to 18
:param style_name: The plot style, defaulted to 'default'. Acceptable styles can be
found at
`matplotlib styles <https://matplotlib.org/3.2.1/gallery/style_sheets/style_sheets_reference.html>`_.
:param save: True if the plot is to be saved, False if the plot is only to be
shown
:param plot_name: The name of the plot, if it is to be saved
:param hist_type: {``bar``, ``barstacked``, ``step``, ``stepfilled``}
See
`histogram <https://matplotlib.org/3.1.1/api/_as_gen/matplotlib.pyplot.hist.html>`_
for more information.
:param dens: If True, the first element of the return tuple will be the counts
normalized to form a probability density, i.e., the area (or integral)
under the histogram will sum to 1
:param title: The title of the plot to incorporate into the header. Defaulted to NULL
:param title_font_size: The font size for the tile, defaulted to 24
.. code-block:: python
> np.random.seed(19680801)
> x = np.random.normal(15.0, 3.0, 1000)
> y = np.random.normal(20.0, 3.0, 1000)
> data = [x, y]
> labels = ['one', 'two']
> one = np.repeat('one', len(x))
> two = np.repeat('two', len(x))
> x = np.hstack((x, y))
> y = np.hstack((one, two))
> dictionary = {'data': x, 'type': y}
> df = pd.DataFrame(dictionary)
> obj = MatPlotDataFrame(df)
> obj.histogram_plot_parse_column('data', 'type', labels, x_label='x-axis',
y_label='y-axis', shading=[0.9, 0.4], save=True,
.. image:: hist2.eps
:align: center
"""
if colors[0] == "None":
colors = self.colors
if edge_colors[0] == 'None':
edge_colors = np.repeat('black', len(labels))
if shading[0] == "None":
shading = np.repeat(0.7, len(labels))
plt.tight_layout()
plt.gcf().subplots_adjust(bottom=0.15)
plt.rcParams.update({'figure.autolayout': True})
plt.style.use(style_name)
rc('xtick', labelsize=tick_font_size)
rc('ytick', labelsize=tick_font_size)
plt.xlabel(x_label, fontsize=label_font_size)
plt.ylabel(y_label, fontsize=label_font_size)
if title != 'NULL':
plt.title(title, fontsize=title_font_size)
if title != 'NULL':
plt.title(title, fontsize=title_font_size)
for i in range(len(x_headers)):
plt.hist(self.df[x_headers[i]], bins=num_bins, color=colors[i],
edgecolor=edge_colors[i], alpha=shading[i], label=labels[i],
density=dens)
plt.legend(loc=label_pos)
if not save:
plt.show()
else:
plt.savefig(plot_name)
plt.close()
# ================================================================================
# ================================================================================
# eof
# TODO Create histogram version of plots
# TODO Repeat for Bokeh plots
|
[
"matplotlib.pyplot.grid",
"matplotlib.pyplot.hist",
"matplotlib.pyplot.ylabel",
"matplotlib.rc",
"matplotlib.dates.DayLocator",
"matplotlib.pyplot.xlabel",
"matplotlib.pyplot.style.use",
"matplotlib.pyplot.close",
"warnings.warn",
"matplotlib.pyplot.savefig",
"matplotlib.pyplot.MaxNLocator",
"matplotlib.dates.MonthLocator",
"matplotlib.pyplot.gcf",
"matplotlib.dates.DateFormatter",
"numpy.timedelta64",
"matplotlib.pyplot.title",
"matplotlib.pyplot.legend",
"matplotlib.pyplot.show",
"datetime.datetime.strptime",
"matplotlib.pyplot.rcParams.update",
"matplotlib.pyplot.tight_layout",
"matplotlib.pyplot.subplots"
] |
[((5218, 5232), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {}), '()\n', (5230, 5232), True, 'from matplotlib import rc, pyplot as plt\n'), ((5237, 5285), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (5256, 5285), True, 'from matplotlib import rc, pyplot as plt\n'), ((5290, 5315), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (5303, 5315), True, 'from matplotlib import rc, pyplot as plt\n'), ((5320, 5357), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (5322, 5357), False, 'from matplotlib import rc, pyplot as plt\n'), ((5362, 5399), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (5364, 5399), False, 'from matplotlib import rc, pyplot as plt\n'), ((6306, 6331), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (6316, 6331), True, 'from matplotlib import rc, pyplot as plt\n'), ((11101, 11149), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (11120, 11149), True, 'from matplotlib import rc, pyplot as plt\n'), ((11154, 11179), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (11167, 11179), True, 'from matplotlib import rc, pyplot as plt\n'), ((11199, 11213), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {}), '()\n', (11211, 11213), True, 'from matplotlib import rc, pyplot as plt\n'), ((11218, 11255), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (11220, 11255), False, 'from matplotlib import rc, pyplot as plt\n'), ((11260, 11297), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (11262, 11297), False, 'from matplotlib import rc, pyplot as plt\n'), ((11854, 11879), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (11864, 11879), True, 'from matplotlib import rc, pyplot as plt\n'), ((16347, 16395), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (16366, 16395), True, 'from matplotlib import rc, pyplot as plt\n'), ((16400, 16425), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (16413, 16425), True, 'from matplotlib import rc, pyplot as plt\n'), ((16445, 16459), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {}), '()\n', (16457, 16459), True, 'from matplotlib import rc, pyplot as plt\n'), ((16464, 16501), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (16466, 16501), False, 'from matplotlib import rc, pyplot as plt\n'), ((16506, 16543), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (16508, 16543), False, 'from matplotlib import rc, pyplot as plt\n'), ((17092, 17117), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (17102, 17117), True, 'from matplotlib import rc, pyplot as plt\n'), ((22170, 22218), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (22189, 22218), True, 'from matplotlib import rc, pyplot as plt\n'), ((22223, 22248), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (22236, 22248), True, 'from matplotlib import rc, pyplot as plt\n'), ((22268, 22282), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {}), '()\n', (22280, 22282), True, 'from matplotlib import rc, pyplot as plt\n'), ((22287, 22324), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (22289, 22324), False, 'from matplotlib import rc, pyplot as plt\n'), ((22329, 22366), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (22331, 22366), False, 'from matplotlib import rc, pyplot as plt\n'), ((22950, 22975), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (22960, 22975), True, 'from matplotlib import rc, pyplot as plt\n'), ((27969, 27987), 'matplotlib.pyplot.tight_layout', 'plt.tight_layout', ([], {}), '()\n', (27985, 27987), True, 'from matplotlib import rc, pyplot as plt\n'), ((28035, 28083), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (28054, 28083), True, 'from matplotlib import rc, pyplot as plt\n'), ((28088, 28113), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (28101, 28113), True, 'from matplotlib import rc, pyplot as plt\n'), ((28118, 28155), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (28120, 28155), False, 'from matplotlib import rc, pyplot as plt\n'), ((28160, 28197), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (28162, 28197), False, 'from matplotlib import rc, pyplot as plt\n'), ((28202, 28247), 'matplotlib.pyplot.xlabel', 'plt.xlabel', (['x_label'], {'fontsize': 'label_font_size'}), '(x_label, fontsize=label_font_size)\n', (28212, 28247), True, 'from matplotlib import rc, pyplot as plt\n'), ((28252, 28297), 'matplotlib.pyplot.ylabel', 'plt.ylabel', (['y_label'], {'fontsize': 'label_font_size'}), '(y_label, fontsize=label_font_size)\n', (28262, 28297), True, 'from matplotlib import rc, pyplot as plt\n'), ((28580, 28605), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (28590, 28605), True, 'from matplotlib import rc, pyplot as plt\n'), ((28687, 28698), 'matplotlib.pyplot.close', 'plt.close', ([], {}), '()\n', (28696, 28698), True, 'from matplotlib import rc, pyplot as plt\n'), ((5502, 5528), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%d"""'], {}), "('%d')\n", (5522, 5528), True, 'import matplotlib.dates as mdates\n'), ((6357, 6367), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (6365, 6367), True, 'from matplotlib import rc, pyplot as plt\n'), ((6386, 6408), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (6397, 6408), True, 'from matplotlib import rc, pyplot as plt\n'), ((10207, 10269), 'warnings.warn', 'warnings.warn', (['"""if save is True then plot name cannot be NULL"""'], {}), "('if save is True then plot name cannot be NULL')\n", (10220, 10269), False, 'import warnings\n'), ((10313, 10418), 'warnings.warn', 'warnings.warn', (['"""length of x list of lists is not the same as y list of lists, plot not printed"""'], {}), "(\n 'length of x list of lists is not the same as y list of lists, plot not printed'\n )\n", (10326, 10418), False, 'import warnings\n'), ((10472, 10562), 'warnings.warn', 'warnings.warn', (['"""line colors list not the same length as data lists, plot not printed"""'], {}), "(\n 'line colors list not the same length as data lists, plot not printed')\n", (10485, 10562), False, 'import warnings\n'), ((10620, 10709), 'warnings.warn', 'warnings.warn', (['"""line_style list not the same length as data lists, plot not printed"""'], {}), "(\n 'line_style list not the same length as data lists, plot not printed')\n", (10633, 10709), False, 'import warnings\n'), ((10768, 10858), 'warnings.warn', 'warnings.warn', (['"""line_weight list not the same length as data lists, plot not printed"""'], {}), "(\n 'line_weight list not the same length as data lists, plot not printed')\n", (10781, 10858), False, 'import warnings\n'), ((10923, 10973), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (10936, 10973), False, 'import warnings\n'), ((11028, 11078), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (11041, 11078), False, 'import warnings\n'), ((11905, 11915), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (11913, 11915), True, 'from matplotlib import rc, pyplot as plt\n'), ((11934, 11956), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (11945, 11956), True, 'from matplotlib import rc, pyplot as plt\n'), ((15598, 15660), 'warnings.warn', 'warnings.warn', (['"""if save is True then plot name cannot be NULL"""'], {}), "('if save is True then plot name cannot be NULL')\n", (15611, 15660), False, 'import warnings\n'), ((15704, 15809), 'warnings.warn', 'warnings.warn', (['"""length of x list of lists is not the same as y list of lists, plot not printed"""'], {}), "(\n 'length of x list of lists is not the same as y list of lists, plot not printed'\n )\n", (15717, 15809), False, 'import warnings\n'), ((15865, 15955), 'warnings.warn', 'warnings.warn', (['"""line colors list not the same length as data lists, plot not printed"""'], {}), "(\n 'line colors list not the same length as data lists, plot not printed')\n", (15878, 15955), False, 'import warnings\n'), ((16015, 16104), 'warnings.warn', 'warnings.warn', (['"""line_style list not the same length as data lists, plot not printed"""'], {}), "(\n 'line_style list not the same length as data lists, plot not printed')\n", (16028, 16104), False, 'import warnings\n'), ((16169, 16219), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (16182, 16219), False, 'import warnings\n'), ((16274, 16324), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (16287, 16324), False, 'import warnings\n'), ((17143, 17153), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (17151, 17153), True, 'from matplotlib import rc, pyplot as plt\n'), ((17172, 17194), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (17183, 17194), True, 'from matplotlib import rc, pyplot as plt\n'), ((21421, 21483), 'warnings.warn', 'warnings.warn', (['"""if save is True then plot name cannot be NULL"""'], {}), "('if save is True then plot name cannot be NULL')\n", (21434, 21483), False, 'import warnings\n'), ((21527, 21632), 'warnings.warn', 'warnings.warn', (['"""length of x list of lists is not the same as y list of lists, plot not printed"""'], {}), "(\n 'length of x list of lists is not the same as y list of lists, plot not printed'\n )\n", (21540, 21632), False, 'import warnings\n'), ((21688, 21778), 'warnings.warn', 'warnings.warn', (['"""line colors list not the same length as data lists, plot not printed"""'], {}), "(\n 'line colors list not the same length as data lists, plot not printed')\n", (21701, 21778), False, 'import warnings\n'), ((21838, 21927), 'warnings.warn', 'warnings.warn', (['"""line_style list not the same length as data lists, plot not printed"""'], {}), "(\n 'line_style list not the same length as data lists, plot not printed')\n", (21851, 21927), False, 'import warnings\n'), ((21992, 22042), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (22005, 22042), False, 'import warnings\n'), ((22097, 22147), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (22110, 22147), False, 'import warnings\n'), ((23001, 23011), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (23009, 23011), True, 'from matplotlib import rc, pyplot as plt\n'), ((23030, 23052), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (23041, 23052), True, 'from matplotlib import rc, pyplot as plt\n'), ((27531, 27602), 'warnings.warn', 'warnings.warn', (['"""data list should be the same length as the labels list"""'], {}), "('data list should be the same length as the labels list')\n", (27544, 27602), False, 'import warnings\n'), ((27646, 27717), 'warnings.warn', 'warnings.warn', (['"""data list should be the same length as the colors list"""'], {}), "('data list should be the same length as the colors list')\n", (27659, 27717), False, 'import warnings\n'), ((27766, 27844), 'warnings.warn', 'warnings.warn', (['"""labels list should be the same length as the edge_colors list"""'], {}), "('labels list should be the same length as the edge_colors list')\n", (27779, 27844), False, 'import warnings\n'), ((27889, 27963), 'warnings.warn', 'warnings.warn', (['"""labels list should be the same length as the shading list"""'], {}), "('labels list should be the same length as the shading list')\n", (27902, 27963), False, 'import warnings\n'), ((28330, 28372), 'matplotlib.pyplot.title', 'plt.title', (['title'], {'fontsize': 'title_font_size'}), '(title, fontsize=title_font_size)\n', (28339, 28372), True, 'from matplotlib import rc, pyplot as plt\n'), ((28414, 28562), 'matplotlib.pyplot.hist', 'plt.hist', (['data[i]'], {'bins': 'num_bins', 'color': 'colors[i]', 'edgecolor': 'edge_colors[i]', 'alpha': 'shading[i]', 'label': 'labels[i]', 'histtype': 'hist_type', 'density': 'dens'}), '(data[i], bins=num_bins, color=colors[i], edgecolor=edge_colors[i],\n alpha=shading[i], label=labels[i], histtype=hist_type, density=dens)\n', (28422, 28562), True, 'from matplotlib import rc, pyplot as plt\n'), ((28631, 28641), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (28639, 28641), True, 'from matplotlib import rc, pyplot as plt\n'), ((28660, 28682), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (28671, 28682), True, 'from matplotlib import rc, pyplot as plt\n'), ((36372, 36420), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (36391, 36420), True, 'from matplotlib import rc, pyplot as plt\n'), ((36429, 36454), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (36442, 36454), True, 'from matplotlib import rc, pyplot as plt\n'), ((36478, 36492), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {}), '()\n', (36490, 36492), True, 'from matplotlib import rc, pyplot as plt\n'), ((36501, 36538), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (36503, 36538), False, 'from matplotlib import rc, pyplot as plt\n'), ((36547, 36584), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (36549, 36584), False, 'from matplotlib import rc, pyplot as plt\n'), ((37342, 37367), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (37352, 37367), True, 'from matplotlib import rc, pyplot as plt\n'), ((37547, 37558), 'matplotlib.pyplot.close', 'plt.close', ([], {}), '()\n', (37556, 37558), True, 'from matplotlib import rc, pyplot as plt\n'), ((43499, 43547), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (43518, 43547), True, 'from matplotlib import rc, pyplot as plt\n'), ((43556, 43581), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (43569, 43581), True, 'from matplotlib import rc, pyplot as plt\n'), ((43605, 43619), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {}), '()\n', (43617, 43619), True, 'from matplotlib import rc, pyplot as plt\n'), ((43628, 43665), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (43630, 43665), False, 'from matplotlib import rc, pyplot as plt\n'), ((43674, 43711), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (43676, 43711), False, 'from matplotlib import rc, pyplot as plt\n'), ((44469, 44494), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (44479, 44494), True, 'from matplotlib import rc, pyplot as plt\n'), ((44674, 44685), 'matplotlib.pyplot.close', 'plt.close', ([], {}), '()\n', (44683, 44685), True, 'from matplotlib import rc, pyplot as plt\n'), ((50791, 50839), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (50810, 50839), True, 'from matplotlib import rc, pyplot as plt\n'), ((50848, 50873), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (50861, 50873), True, 'from matplotlib import rc, pyplot as plt\n'), ((50897, 50911), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {}), '()\n', (50909, 50911), True, 'from matplotlib import rc, pyplot as plt\n'), ((50920, 50957), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (50922, 50957), False, 'from matplotlib import rc, pyplot as plt\n'), ((50966, 51003), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (50968, 51003), False, 'from matplotlib import rc, pyplot as plt\n'), ((51629, 51654), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (51639, 51654), True, 'from matplotlib import rc, pyplot as plt\n'), ((51834, 51845), 'matplotlib.pyplot.close', 'plt.close', ([], {}), '()\n', (51843, 51845), True, 'from matplotlib import rc, pyplot as plt\n'), ((57108, 57156), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (57127, 57156), True, 'from matplotlib import rc, pyplot as plt\n'), ((57165, 57190), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (57178, 57190), True, 'from matplotlib import rc, pyplot as plt\n'), ((57214, 57228), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {}), '()\n', (57226, 57228), True, 'from matplotlib import rc, pyplot as plt\n'), ((57237, 57274), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (57239, 57274), False, 'from matplotlib import rc, pyplot as plt\n'), ((57283, 57320), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (57285, 57320), False, 'from matplotlib import rc, pyplot as plt\n'), ((57944, 57969), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (57954, 57969), True, 'from matplotlib import rc, pyplot as plt\n'), ((58149, 58160), 'matplotlib.pyplot.close', 'plt.close', ([], {}), '()\n', (58158, 58160), True, 'from matplotlib import rc, pyplot as plt\n'), ((64598, 64646), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (64617, 64646), True, 'from matplotlib import rc, pyplot as plt\n'), ((64655, 64680), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (64668, 64680), True, 'from matplotlib import rc, pyplot as plt\n'), ((64704, 64718), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {}), '()\n', (64716, 64718), True, 'from matplotlib import rc, pyplot as plt\n'), ((64727, 64764), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (64729, 64764), False, 'from matplotlib import rc, pyplot as plt\n'), ((64773, 64810), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (64775, 64810), False, 'from matplotlib import rc, pyplot as plt\n'), ((66163, 66188), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (66173, 66188), True, 'from matplotlib import rc, pyplot as plt\n'), ((66368, 66379), 'matplotlib.pyplot.close', 'plt.close', ([], {}), '()\n', (66377, 66379), True, 'from matplotlib import rc, pyplot as plt\n'), ((72353, 72401), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (72372, 72401), True, 'from matplotlib import rc, pyplot as plt\n'), ((72410, 72435), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (72423, 72435), True, 'from matplotlib import rc, pyplot as plt\n'), ((72459, 72473), 'matplotlib.pyplot.subplots', 'plt.subplots', ([], {}), '()\n', (72471, 72473), True, 'from matplotlib import rc, pyplot as plt\n'), ((72482, 72519), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (72484, 72519), False, 'from matplotlib import rc, pyplot as plt\n'), ((72528, 72565), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (72530, 72565), False, 'from matplotlib import rc, pyplot as plt\n'), ((73921, 73946), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (73931, 73946), True, 'from matplotlib import rc, pyplot as plt\n'), ((74126, 74137), 'matplotlib.pyplot.close', 'plt.close', ([], {}), '()\n', (74135, 74137), True, 'from matplotlib import rc, pyplot as plt\n'), ((79108, 79126), 'matplotlib.pyplot.tight_layout', 'plt.tight_layout', ([], {}), '()\n', (79124, 79126), True, 'from matplotlib import rc, pyplot as plt\n'), ((79182, 79230), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (79201, 79230), True, 'from matplotlib import rc, pyplot as plt\n'), ((79239, 79264), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (79252, 79264), True, 'from matplotlib import rc, pyplot as plt\n'), ((79273, 79310), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (79275, 79310), False, 'from matplotlib import rc, pyplot as plt\n'), ((79319, 79356), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (79321, 79356), False, 'from matplotlib import rc, pyplot as plt\n'), ((79365, 79410), 'matplotlib.pyplot.xlabel', 'plt.xlabel', (['x_label'], {'fontsize': 'label_font_size'}), '(x_label, fontsize=label_font_size)\n', (79375, 79410), True, 'from matplotlib import rc, pyplot as plt\n'), ((79419, 79464), 'matplotlib.pyplot.ylabel', 'plt.ylabel', (['y_label'], {'fontsize': 'label_font_size'}), '(y_label, fontsize=label_font_size)\n', (79429, 79464), True, 'from matplotlib import rc, pyplot as plt\n'), ((79883, 79908), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (79893, 79908), True, 'from matplotlib import rc, pyplot as plt\n'), ((80010, 80021), 'matplotlib.pyplot.close', 'plt.close', ([], {}), '()\n', (80019, 80021), True, 'from matplotlib import rc, pyplot as plt\n'), ((84532, 84550), 'matplotlib.pyplot.tight_layout', 'plt.tight_layout', ([], {}), '()\n', (84548, 84550), True, 'from matplotlib import rc, pyplot as plt\n'), ((84606, 84654), 'matplotlib.pyplot.rcParams.update', 'plt.rcParams.update', (["{'figure.autolayout': True}"], {}), "({'figure.autolayout': True})\n", (84625, 84654), True, 'from matplotlib import rc, pyplot as plt\n'), ((84663, 84688), 'matplotlib.pyplot.style.use', 'plt.style.use', (['style_name'], {}), '(style_name)\n', (84676, 84688), True, 'from matplotlib import rc, pyplot as plt\n'), ((84697, 84734), 'matplotlib.rc', 'rc', (['"""xtick"""'], {'labelsize': 'tick_font_size'}), "('xtick', labelsize=tick_font_size)\n", (84699, 84734), False, 'from matplotlib import rc, pyplot as plt\n'), ((84743, 84780), 'matplotlib.rc', 'rc', (['"""ytick"""'], {'labelsize': 'tick_font_size'}), "('ytick', labelsize=tick_font_size)\n", (84745, 84780), False, 'from matplotlib import rc, pyplot as plt\n'), ((84789, 84834), 'matplotlib.pyplot.xlabel', 'plt.xlabel', (['x_label'], {'fontsize': 'label_font_size'}), '(x_label, fontsize=label_font_size)\n', (84799, 84834), True, 'from matplotlib import rc, pyplot as plt\n'), ((84843, 84888), 'matplotlib.pyplot.ylabel', 'plt.ylabel', (['y_label'], {'fontsize': 'label_font_size'}), '(y_label, fontsize=label_font_size)\n', (84853, 84888), True, 'from matplotlib import rc, pyplot as plt\n'), ((85303, 85328), 'matplotlib.pyplot.legend', 'plt.legend', ([], {'loc': 'label_pos'}), '(loc=label_pos)\n', (85313, 85328), True, 'from matplotlib import rc, pyplot as plt\n'), ((85430, 85441), 'matplotlib.pyplot.close', 'plt.close', ([], {}), '()\n', (85439, 85441), True, 'from matplotlib import rc, pyplot as plt\n'), ((5569, 5588), 'matplotlib.dates.DayLocator', 'mdates.DayLocator', ([], {}), '()\n', (5586, 5588), True, 'import matplotlib.dates as mdates\n'), ((5628, 5657), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%b-%y"""'], {}), "('%b-%y')\n", (5648, 5657), True, 'import matplotlib.dates as mdates\n'), ((5747, 5776), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%b-%y"""'], {}), "('%b-%y')\n", (5767, 5776), True, 'import matplotlib.dates as mdates\n'), ((27992, 28001), 'matplotlib.pyplot.gcf', 'plt.gcf', ([], {}), '()\n', (27999, 28001), True, 'from matplotlib import rc, pyplot as plt\n'), ((36053, 36115), 'warnings.warn', 'warnings.warn', (['"""if save is True then plot name cannot be NULL"""'], {}), "('if save is True then plot name cannot be NULL')\n", (36066, 36115), False, 'import warnings\n'), ((36178, 36228), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (36191, 36228), False, 'import warnings\n'), ((36291, 36341), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (36304, 36341), False, 'import warnings\n'), ((37397, 37445), 'matplotlib.pyplot.grid', 'plt.grid', ([], {'color': 'grid_color', 'linestyle': 'grid_style'}), '(color=grid_color, linestyle=grid_style)\n', (37405, 37445), True, 'from matplotlib import rc, pyplot as plt\n'), ((37479, 37489), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (37487, 37489), True, 'from matplotlib import rc, pyplot as plt\n'), ((37516, 37538), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (37527, 37538), True, 'from matplotlib import rc, pyplot as plt\n'), ((43180, 43242), 'warnings.warn', 'warnings.warn', (['"""if save is True then plot name cannot be NULL"""'], {}), "('if save is True then plot name cannot be NULL')\n", (43193, 43242), False, 'import warnings\n'), ((43305, 43355), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (43318, 43355), False, 'import warnings\n'), ((43418, 43468), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (43431, 43468), False, 'import warnings\n'), ((44524, 44572), 'matplotlib.pyplot.grid', 'plt.grid', ([], {'color': 'grid_color', 'linestyle': 'grid_style'}), '(color=grid_color, linestyle=grid_style)\n', (44532, 44572), True, 'from matplotlib import rc, pyplot as plt\n'), ((44606, 44616), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (44614, 44616), True, 'from matplotlib import rc, pyplot as plt\n'), ((44643, 44665), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (44654, 44665), True, 'from matplotlib import rc, pyplot as plt\n'), ((50472, 50534), 'warnings.warn', 'warnings.warn', (['"""if save is True then plot name cannot be NULL"""'], {}), "('if save is True then plot name cannot be NULL')\n", (50485, 50534), False, 'import warnings\n'), ((50597, 50647), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (50610, 50647), False, 'import warnings\n'), ((50710, 50760), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (50723, 50760), False, 'import warnings\n'), ((51684, 51732), 'matplotlib.pyplot.grid', 'plt.grid', ([], {'color': 'grid_color', 'linestyle': 'grid_style'}), '(color=grid_color, linestyle=grid_style)\n', (51692, 51732), True, 'from matplotlib import rc, pyplot as plt\n'), ((51766, 51776), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (51774, 51776), True, 'from matplotlib import rc, pyplot as plt\n'), ((51803, 51825), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (51814, 51825), True, 'from matplotlib import rc, pyplot as plt\n'), ((56789, 56851), 'warnings.warn', 'warnings.warn', (['"""if save is True then plot name cannot be NULL"""'], {}), "('if save is True then plot name cannot be NULL')\n", (56802, 56851), False, 'import warnings\n'), ((56914, 56964), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (56927, 56964), False, 'import warnings\n'), ((57027, 57077), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (57040, 57077), False, 'import warnings\n'), ((57999, 58047), 'matplotlib.pyplot.grid', 'plt.grid', ([], {'color': 'grid_color', 'linestyle': 'grid_style'}), '(color=grid_color, linestyle=grid_style)\n', (58007, 58047), True, 'from matplotlib import rc, pyplot as plt\n'), ((58081, 58091), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (58089, 58091), True, 'from matplotlib import rc, pyplot as plt\n'), ((58118, 58140), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (58129, 58140), True, 'from matplotlib import rc, pyplot as plt\n'), ((63692, 63714), 'numpy.timedelta64', 'np.timedelta64', (['(1)', '"""D"""'], {}), "(1, 'D')\n", (63706, 63714), True, 'import numpy as np\n'), ((64280, 64342), 'warnings.warn', 'warnings.warn', (['"""if save is True then plot name cannot be NULL"""'], {}), "('if save is True then plot name cannot be NULL')\n", (64293, 64342), False, 'import warnings\n'), ((64405, 64455), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (64418, 64455), False, 'import warnings\n'), ((64518, 64568), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (64531, 64568), False, 'import warnings\n'), ((65227, 65253), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%H"""'], {}), "('%H')\n", (65247, 65253), True, 'import matplotlib.dates as mdates\n'), ((66218, 66266), 'matplotlib.pyplot.grid', 'plt.grid', ([], {'color': 'grid_color', 'linestyle': 'grid_style'}), '(color=grid_color, linestyle=grid_style)\n', (66226, 66266), True, 'from matplotlib import rc, pyplot as plt\n'), ((66300, 66310), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (66308, 66310), True, 'from matplotlib import rc, pyplot as plt\n'), ((66337, 66359), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (66348, 66359), True, 'from matplotlib import rc, pyplot as plt\n'), ((72035, 72097), 'warnings.warn', 'warnings.warn', (['"""if save is True then plot name cannot be NULL"""'], {}), "('if save is True then plot name cannot be NULL')\n", (72048, 72097), False, 'import warnings\n'), ((72160, 72210), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (72173, 72210), False, 'import warnings\n'), ((72273, 72323), 'warnings.warn', 'warnings.warn', (['"""y_scale must be set to LOG or LIN"""'], {}), "('y_scale must be set to LOG or LIN')\n", (72286, 72323), False, 'import warnings\n'), ((72982, 73008), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%H"""'], {}), "('%H')\n", (73002, 73008), True, 'import matplotlib.dates as mdates\n'), ((73976, 74024), 'matplotlib.pyplot.grid', 'plt.grid', ([], {'color': 'grid_color', 'linestyle': 'grid_style'}), '(color=grid_color, linestyle=grid_style)\n', (73984, 74024), True, 'from matplotlib import rc, pyplot as plt\n'), ((74058, 74068), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (74066, 74068), True, 'from matplotlib import rc, pyplot as plt\n'), ((74095, 74117), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (74106, 74117), True, 'from matplotlib import rc, pyplot as plt\n'), ((79505, 79547), 'matplotlib.pyplot.title', 'plt.title', (['title'], {'fontsize': 'title_font_size'}), '(title, fontsize=title_font_size)\n', (79514, 79547), True, 'from matplotlib import rc, pyplot as plt\n'), ((79592, 79634), 'matplotlib.pyplot.title', 'plt.title', (['title'], {'fontsize': 'title_font_size'}), '(title, fontsize=title_font_size)\n', (79601, 79634), True, 'from matplotlib import rc, pyplot as plt\n'), ((79691, 79863), 'matplotlib.pyplot.hist', 'plt.hist', (['df_list[i][header]'], {'bins': 'num_bins', 'color': 'colors[i]', 'edgecolor': 'edge_colors[i]', 'alpha': 'shading[i]', 'label': 'column_values[i]', 'histtype': 'hist_type', 'density': 'dens'}), '(df_list[i][header], bins=num_bins, color=colors[i], edgecolor=\n edge_colors[i], alpha=shading[i], label=column_values[i], histtype=\n hist_type, density=dens)\n', (79699, 79863), True, 'from matplotlib import rc, pyplot as plt\n'), ((79942, 79952), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (79950, 79952), True, 'from matplotlib import rc, pyplot as plt\n'), ((79979, 80001), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (79990, 80001), True, 'from matplotlib import rc, pyplot as plt\n'), ((84929, 84971), 'matplotlib.pyplot.title', 'plt.title', (['title'], {'fontsize': 'title_font_size'}), '(title, fontsize=title_font_size)\n', (84938, 84971), True, 'from matplotlib import rc, pyplot as plt\n'), ((85016, 85058), 'matplotlib.pyplot.title', 'plt.title', (['title'], {'fontsize': 'title_font_size'}), '(title, fontsize=title_font_size)\n', (85025, 85058), True, 'from matplotlib import rc, pyplot as plt\n'), ((85112, 85255), 'matplotlib.pyplot.hist', 'plt.hist', (['self.df[x_headers[i]]'], {'bins': 'num_bins', 'color': 'colors[i]', 'edgecolor': 'edge_colors[i]', 'alpha': 'shading[i]', 'label': 'labels[i]', 'density': 'dens'}), '(self.df[x_headers[i]], bins=num_bins, color=colors[i], edgecolor=\n edge_colors[i], alpha=shading[i], label=labels[i], density=dens)\n', (85120, 85255), True, 'from matplotlib import rc, pyplot as plt\n'), ((85362, 85372), 'matplotlib.pyplot.show', 'plt.show', ([], {}), '()\n', (85370, 85372), True, 'from matplotlib import rc, pyplot as plt\n'), ((85399, 85421), 'matplotlib.pyplot.savefig', 'plt.savefig', (['plot_name'], {}), '(plot_name)\n', (85410, 85421), True, 'from matplotlib import rc, pyplot as plt\n'), ((4913, 4955), 'datetime.datetime.strptime', 'datetime.strptime', (['date_string', '"""%Y-%m-%d"""'], {}), "(date_string, '%Y-%m-%d')\n", (4930, 4955), False, 'from datetime import datetime\n'), ((5698, 5719), 'matplotlib.dates.MonthLocator', 'mdates.MonthLocator', ([], {}), '()\n', (5717, 5719), True, 'import matplotlib.dates as mdates\n'), ((5817, 5835), 'matplotlib.pyplot.MaxNLocator', 'plt.MaxNLocator', (['(4)'], {}), '(4)\n', (5832, 5835), True, 'from matplotlib import rc, pyplot as plt\n'), ((65298, 65316), 'matplotlib.pyplot.MaxNLocator', 'plt.MaxNLocator', (['(6)'], {}), '(6)\n', (65313, 65316), True, 'from matplotlib import rc, pyplot as plt\n'), ((65363, 65392), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%b-%d"""'], {}), "('%b-%d')\n", (65383, 65392), True, 'import matplotlib.dates as mdates\n'), ((71567, 71589), 'numpy.timedelta64', 'np.timedelta64', (['(1)', '"""D"""'], {}), "(1, 'D')\n", (71581, 71589), True, 'import numpy as np\n'), ((73053, 73071), 'matplotlib.pyplot.MaxNLocator', 'plt.MaxNLocator', (['(6)'], {}), '(6)\n', (73068, 73071), True, 'from matplotlib import rc, pyplot as plt\n'), ((73118, 73147), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%b-%d"""'], {}), "('%b-%d')\n", (73138, 73147), True, 'import matplotlib.dates as mdates\n'), ((79135, 79144), 'matplotlib.pyplot.gcf', 'plt.gcf', ([], {}), '()\n', (79142, 79144), True, 'from matplotlib import rc, pyplot as plt\n'), ((84559, 84568), 'matplotlib.pyplot.gcf', 'plt.gcf', ([], {}), '()\n', (84566, 84568), True, 'from matplotlib import rc, pyplot as plt\n'), ((65437, 65455), 'matplotlib.pyplot.MaxNLocator', 'plt.MaxNLocator', (['(6)'], {}), '(6)\n', (65452, 65455), True, 'from matplotlib import rc, pyplot as plt\n'), ((65503, 65532), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%b-%Y"""'], {}), "('%b-%Y')\n", (65523, 65532), True, 'import matplotlib.dates as mdates\n'), ((73192, 73210), 'matplotlib.pyplot.MaxNLocator', 'plt.MaxNLocator', (['(6)'], {}), '(6)\n', (73207, 73210), True, 'from matplotlib import rc, pyplot as plt\n'), ((73258, 73287), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%b-%Y"""'], {}), "('%b-%Y')\n", (73278, 73287), True, 'import matplotlib.dates as mdates\n'), ((65577, 65595), 'matplotlib.pyplot.MaxNLocator', 'plt.MaxNLocator', (['(5)'], {}), '(5)\n', (65592, 65595), True, 'from matplotlib import rc, pyplot as plt\n'), ((65644, 65670), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%Y"""'], {}), "('%Y')\n", (65664, 65670), True, 'import matplotlib.dates as mdates\n'), ((65769, 65795), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%Y"""'], {}), "('%Y')\n", (65789, 65795), True, 'import matplotlib.dates as mdates\n'), ((73332, 73350), 'matplotlib.pyplot.MaxNLocator', 'plt.MaxNLocator', (['(5)'], {}), '(5)\n', (73347, 73350), True, 'from matplotlib import rc, pyplot as plt\n'), ((73399, 73425), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%Y"""'], {}), "('%Y')\n", (73419, 73425), True, 'import matplotlib.dates as mdates\n'), ((73524, 73550), 'matplotlib.dates.DateFormatter', 'mdates.DateFormatter', (['"""%Y"""'], {}), "('%Y')\n", (73544, 73550), True, 'import matplotlib.dates as mdates\n'), ((65715, 65733), 'matplotlib.pyplot.MaxNLocator', 'plt.MaxNLocator', (['(5)'], {}), '(5)\n', (65730, 65733), True, 'from matplotlib import rc, pyplot as plt\n'), ((65840, 65858), 'matplotlib.pyplot.MaxNLocator', 'plt.MaxNLocator', (['(5)'], {}), '(5)\n', (65855, 65858), True, 'from matplotlib import rc, pyplot as plt\n'), ((73470, 73488), 'matplotlib.pyplot.MaxNLocator', 'plt.MaxNLocator', (['(5)'], {}), '(5)\n', (73485, 73488), True, 'from matplotlib import rc, pyplot as plt\n'), ((73595, 73613), 'matplotlib.pyplot.MaxNLocator', 'plt.MaxNLocator', (['(5)'], {}), '(5)\n', (73610, 73613), True, 'from matplotlib import rc, pyplot as plt\n')]
|
import tempfile
from concurrent.futures import ProcessPoolExecutor, as_completed
import numpy as np
import pytest
import zarr
from dask.distributed import LocalCluster
from swyft import Dataset, DirectoryStore, Prior, Simulator
from swyft.store.simulator import SimulationStatus
PARAMS = ["z1", "z2"]
PRIOR = Prior.from_uv(lambda u: u * np.array([1 for _ in PARAMS]), len(PARAMS))
OUTPUT_SHAPE = (20, 20)
SIM_SHAPES = {"x": OUTPUT_SHAPE}
N_SIMULATIONS = 1000
BATCH_SIZE = 100
MAX_WORKERS = 4 # number of simultaneous processes acting on the store
def model(_):
""" Model with dummy parameters. Return random numbers in (0; 1]. """
return dict(x=-np.random.random(OUTPUT_SHAPE) + 1)
@pytest.fixture(scope="function")
def store():
simulator = Simulator(model, sim_shapes=SIM_SHAPES)
with tempfile.TemporaryDirectory() as tmpdir:
path = f"{tmpdir}/test_store"
yield DirectoryStore(path=path, params=PARAMS, simulator=simulator)
@pytest.fixture(scope="module")
def cluster():
return LocalCluster(n_workers=2, threads_per_worker=1)
def simulate(cluster, path="./cache", wait_for_results=True):
"""
Open store, sample simulation parameters and run the corresponding
simulations.
"""
simulator = Simulator(model=model, sim_shapes=SIM_SHAPES, cluster=cluster)
store = DirectoryStore(path=path, params=PARAMS, simulator=simulator)
dataset = Dataset(N_SIMULATIONS, PRIOR, store=store)
dataset.simulate(wait_for_results=wait_for_results, batch_size=BATCH_SIZE)
return dataset.indices
def read_from_store(path):
""" Extract data from the Zarr Directory store """
z = zarr.open(f"{path}/samples/pars")
x = zarr.open_group(f"{path}/samples/sims")
s = zarr.open_array(f"{path}/samples/simulation_status")
return z[:], {key: val[:] for key, val in x.items()}, s[:]
def test_concurrent_runs_waiting_for_results(cluster, store):
"""
Run several processes that access the same store to sample parameters and
to submit the corresponding simulations. The outcome of the simulations
is waited for within the processes, so when they return all outcome should
be written to the store.
"""
path = store._zarr_store.path
with ProcessPoolExecutor(max_workers=MAX_WORKERS) as executor:
futures = []
for i in range(MAX_WORKERS):
# each process grows and sample the same cache
future = executor.submit(
simulate,
cluster=cluster.scheduler_address,
path=path,
)
futures.append(future)
for future in as_completed(futures):
# processes are waiting for results, so all simulations should be finished
status = store.get_simulation_status(future.result())
assert np.all(status == SimulationStatus.FINISHED)
z, x, s = read_from_store(path)
# check shape of the parameter array
n_simulations, n_params = z.shape
# the real number of samples can differ slightly from the required value
assert n_simulations > 0.80 * N_SIMULATIONS and n_simulations < 1.20 * N_SIMULATIONS
assert n_params == len(PARAMS)
# check shape and values of the simulation array
assert x.keys() == SIM_SHAPES.keys()
for key, val in SIM_SHAPES.items():
assert x[key].shape == (n_simulations, *val)
assert np.all(x[key][:] > 0.0) # all simulation output has been updated
# check shape and values of the status array
assert s.shape == (n_simulations,)
assert np.all(s == SimulationStatus.FINISHED) # all simulations are done
def test_concurrent_run_without_waiting_for_results(cluster, store):
"""
Run several processes that access the same store to sample parameters and
to submit the corresponding simulations. The processes do not wait for the
simulations to be done, so when they return some simulations should still
be running.
"""
path = store._zarr_store.path
with ProcessPoolExecutor(max_workers=MAX_WORKERS) as executor:
futures = []
for i in range(MAX_WORKERS):
# each process grows and sample the same cache
future = executor.submit(
simulate,
cluster=cluster.scheduler_address,
path=path,
wait_for_results=False,
)
futures.append(future)
for future in as_completed(futures):
# processes are not waiting for results, so some simulations should still be running
status = store.get_simulation_status(future.result())
assert np.any(status == SimulationStatus.RUNNING)
z, x, s = read_from_store(path)
# check shape of the parameter array
n_simulations, n_params = z.shape
# the real number of samples can differ slightly from the required value
assert n_simulations > 0.80 * N_SIMULATIONS and n_simulations < 1.20 * N_SIMULATIONS
assert n_params == len(PARAMS)
# check shape of the simulation array
assert x.keys() == SIM_SHAPES.keys()
for key, val in SIM_SHAPES.items():
assert x[key].shape == (n_simulations, *val)
# check shape of the status array
assert s.shape == (n_simulations,)
# now explicitly wait for simulations
store.wait_for_simulations(indices=np.arange(n_simulations))
z, x, s = read_from_store(path)
for key, val in SIM_SHAPES.items():
assert np.all(x[key] > 0.0) # all simulation output has been updated
assert np.all(s == SimulationStatus.FINISHED) # all simulations are done
|
[
"tempfile.TemporaryDirectory",
"dask.distributed.LocalCluster",
"numpy.random.random",
"swyft.Simulator",
"numpy.any",
"swyft.Dataset",
"concurrent.futures.as_completed",
"numpy.array",
"zarr.open",
"swyft.DirectoryStore",
"zarr.open_group",
"pytest.fixture",
"concurrent.futures.ProcessPoolExecutor",
"numpy.all",
"zarr.open_array",
"numpy.arange"
] |
[((700, 732), 'pytest.fixture', 'pytest.fixture', ([], {'scope': '"""function"""'}), "(scope='function')\n", (714, 732), False, 'import pytest\n'), ((969, 999), 'pytest.fixture', 'pytest.fixture', ([], {'scope': '"""module"""'}), "(scope='module')\n", (983, 999), False, 'import pytest\n'), ((762, 801), 'swyft.Simulator', 'Simulator', (['model'], {'sim_shapes': 'SIM_SHAPES'}), '(model, sim_shapes=SIM_SHAPES)\n', (771, 801), False, 'from swyft import Dataset, DirectoryStore, Prior, Simulator\n'), ((1026, 1073), 'dask.distributed.LocalCluster', 'LocalCluster', ([], {'n_workers': '(2)', 'threads_per_worker': '(1)'}), '(n_workers=2, threads_per_worker=1)\n', (1038, 1073), False, 'from dask.distributed import LocalCluster\n'), ((1258, 1320), 'swyft.Simulator', 'Simulator', ([], {'model': 'model', 'sim_shapes': 'SIM_SHAPES', 'cluster': 'cluster'}), '(model=model, sim_shapes=SIM_SHAPES, cluster=cluster)\n', (1267, 1320), False, 'from swyft import Dataset, DirectoryStore, Prior, Simulator\n'), ((1333, 1394), 'swyft.DirectoryStore', 'DirectoryStore', ([], {'path': 'path', 'params': 'PARAMS', 'simulator': 'simulator'}), '(path=path, params=PARAMS, simulator=simulator)\n', (1347, 1394), False, 'from swyft import Dataset, DirectoryStore, Prior, Simulator\n'), ((1409, 1451), 'swyft.Dataset', 'Dataset', (['N_SIMULATIONS', 'PRIOR'], {'store': 'store'}), '(N_SIMULATIONS, PRIOR, store=store)\n', (1416, 1451), False, 'from swyft import Dataset, DirectoryStore, Prior, Simulator\n'), ((1650, 1683), 'zarr.open', 'zarr.open', (['f"""{path}/samples/pars"""'], {}), "(f'{path}/samples/pars')\n", (1659, 1683), False, 'import zarr\n'), ((1692, 1731), 'zarr.open_group', 'zarr.open_group', (['f"""{path}/samples/sims"""'], {}), "(f'{path}/samples/sims')\n", (1707, 1731), False, 'import zarr\n'), ((1740, 1792), 'zarr.open_array', 'zarr.open_array', (['f"""{path}/samples/simulation_status"""'], {}), "(f'{path}/samples/simulation_status')\n", (1755, 1792), False, 'import zarr\n'), ((3556, 3594), 'numpy.all', 'np.all', (['(s == SimulationStatus.FINISHED)'], {}), '(s == SimulationStatus.FINISHED)\n', (3562, 3594), True, 'import numpy as np\n'), ((5530, 5568), 'numpy.all', 'np.all', (['(s == SimulationStatus.FINISHED)'], {}), '(s == SimulationStatus.FINISHED)\n', (5536, 5568), True, 'import numpy as np\n'), ((811, 840), 'tempfile.TemporaryDirectory', 'tempfile.TemporaryDirectory', ([], {}), '()\n', (838, 840), False, 'import tempfile\n'), ((2241, 2285), 'concurrent.futures.ProcessPoolExecutor', 'ProcessPoolExecutor', ([], {'max_workers': 'MAX_WORKERS'}), '(max_workers=MAX_WORKERS)\n', (2260, 2285), False, 'from concurrent.futures import ProcessPoolExecutor, as_completed\n'), ((2630, 2651), 'concurrent.futures.as_completed', 'as_completed', (['futures'], {}), '(futures)\n', (2642, 2651), False, 'from concurrent.futures import ProcessPoolExecutor, as_completed\n'), ((3390, 3413), 'numpy.all', 'np.all', (['(x[key][:] > 0.0)'], {}), '(x[key][:] > 0.0)\n', (3396, 3413), True, 'import numpy as np\n'), ((4005, 4049), 'concurrent.futures.ProcessPoolExecutor', 'ProcessPoolExecutor', ([], {'max_workers': 'MAX_WORKERS'}), '(max_workers=MAX_WORKERS)\n', (4024, 4049), False, 'from concurrent.futures import ProcessPoolExecutor, as_completed\n'), ((4434, 4455), 'concurrent.futures.as_completed', 'as_completed', (['futures'], {}), '(futures)\n', (4446, 4455), False, 'from concurrent.futures import ProcessPoolExecutor, as_completed\n'), ((5455, 5475), 'numpy.all', 'np.all', (['(x[key] > 0.0)'], {}), '(x[key] > 0.0)\n', (5461, 5475), True, 'import numpy as np\n'), ((340, 371), 'numpy.array', 'np.array', (['[(1) for _ in PARAMS]'], {}), '([(1) for _ in PARAMS])\n', (348, 371), True, 'import numpy as np\n'), ((904, 965), 'swyft.DirectoryStore', 'DirectoryStore', ([], {'path': 'path', 'params': 'PARAMS', 'simulator': 'simulator'}), '(path=path, params=PARAMS, simulator=simulator)\n', (918, 965), False, 'from swyft import Dataset, DirectoryStore, Prior, Simulator\n'), ((2825, 2868), 'numpy.all', 'np.all', (['(status == SimulationStatus.FINISHED)'], {}), '(status == SimulationStatus.FINISHED)\n', (2831, 2868), True, 'import numpy as np\n'), ((4639, 4681), 'numpy.any', 'np.any', (['(status == SimulationStatus.RUNNING)'], {}), '(status == SimulationStatus.RUNNING)\n', (4645, 4681), True, 'import numpy as np\n'), ((5337, 5361), 'numpy.arange', 'np.arange', (['n_simulations'], {}), '(n_simulations)\n', (5346, 5361), True, 'import numpy as np\n'), ((661, 691), 'numpy.random.random', 'np.random.random', (['OUTPUT_SHAPE'], {}), '(OUTPUT_SHAPE)\n', (677, 691), True, 'import numpy as np\n')]
|
# @Author: <NAME>
# @Date: 2021-03-22 09:43:07
# @Last Modified by: <NAME>
# @Last Modified time: 2021-11-08 15:09:29
#!/usr/bin/env python
## based on: detectron2.modeling.roi_heads.box_head
## based on: detectron2.modeling.roi_heads.fast_rcnn
import torch
from torch import nn
import numpy as np
import logging
from typing import Dict, List, Tuple, Union
from fvcore.nn import giou_loss, smooth_l1_loss
import fvcore.nn.weight_init as weight_init
from torch import nn
from torch.nn import functional as F
from detectron2.config import configurable, get_cfg
from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm
from detectron2.modeling.box_regression import Box2BoxTransform
from detectron2.structures import Boxes, Instances
from detectron2.utils.events import get_event_storage
from detectron2.utils.registry import Registry
## these specific libraries are needed to alter the network architecture
from detectron2.modeling.roi_heads import ROI_HEADS_REGISTRY, StandardROIHeads, Res5ROIHeads
from detectron2.modeling.roi_heads.box_head import ROI_BOX_HEAD_REGISTRY
from detectron2.modeling.roi_heads.fast_rcnn import fast_rcnn_inference, _log_classification_stats
from detectron2.modeling.roi_heads.mask_head import ROI_MASK_HEAD_REGISTRY, BaseMaskRCNNHead
@ROI_BOX_HEAD_REGISTRY.register()
class FastRCNNConvFCHeadDropout(nn.Sequential):
"""
A head with several 3x3 conv layers (each followed by norm & relu) and then
several fc layers (each followed by relu).
"""
@configurable
def __init__(
self, input_shape: ShapeSpec, *, conv_dims: List[int], fc_dims: List[int], conv_norm="",
dropout_probability: float = 0.5,
):
"""
NOTE: this interface is experimental.
Args:
input_shape (ShapeSpec): shape of the input feature.
conv_dims (list[int]): the output dimensions of the conv layers
fc_dims (list[int]): the output dimensions of the fc layers
conv_norm (str or callable): normalization for the conv layers.
See :func:`detectron2.layers.get_norm` for supported types.
"""
super().__init__()
assert len(conv_dims) + len(fc_dims) > 0
self._output_size = (input_shape.channels, input_shape.height, input_shape.width)
self.conv_norm_relus = []
for k, conv_dim in enumerate(conv_dims):
conv = Conv2d(
self._output_size[0],
conv_dim,
kernel_size=3,
padding=1,
bias=not conv_norm,
norm=get_norm(conv_norm, conv_dim),
activation=nn.ReLU(),
)
self.add_module("conv{}".format(k + 1), conv)
self.conv_norm_relus.append(conv)
self._output_size = (conv_dim, self._output_size[1], self._output_size[2])
self.fcs = []
for k, fc_dim in enumerate(fc_dims):
if k == 0:
self.add_module("flatten", nn.Flatten())
dropout = nn.Dropout(p=dropout_probability)
self.add_module("dropout{}".format(k + 1), dropout)
fc = nn.Linear(int(np.prod(self._output_size)), fc_dim)
self.add_module("fc{}".format(k + 1), fc)
self.add_module("fc_relu{}".format(k + 1), nn.ReLU())
self.fcs.append(fc)
self._output_size = fc_dim
for layer in self.conv_norm_relus:
weight_init.c2_msra_fill(layer)
for layer in self.fcs:
weight_init.c2_xavier_fill(layer)
@classmethod
def from_config(cls, cfg, input_shape):
num_conv = cfg.MODEL.ROI_BOX_HEAD.NUM_CONV
conv_dim = cfg.MODEL.ROI_BOX_HEAD.CONV_DIM
num_fc = cfg.MODEL.ROI_BOX_HEAD.NUM_FC
fc_dim = cfg.MODEL.ROI_BOX_HEAD.FC_DIM
return {
"input_shape": input_shape,
"conv_dims": [conv_dim] * num_conv,
"fc_dims": [fc_dim] * num_fc,
"conv_norm": cfg.MODEL.ROI_BOX_HEAD.NORM,
"dropout_probability": cfg.MODEL.ROI_BOX_HEAD.DROPOUT_PROBABILITY,
}
def forward(self, x):
for layer in self:
x = layer(x)
return x
@property
@torch.jit.unused
def output_shape(self):
"""
Returns:
ShapeSpec: the output feature shape
"""
o = self._output_size
if isinstance(o, int):
return ShapeSpec(channels=o)
else:
return ShapeSpec(channels=o[0], height=o[1], width=o[2])
class FastRCNNOutputLayersDropout(nn.Module):
"""
Two linear layers for predicting Fast R-CNN outputs:
1. proposal-to-detection box regression deltas
2. classification scores
"""
@configurable
def __init__(
self,
input_shape: ShapeSpec,
*,
box2box_transform,
num_classes: int,
test_score_thresh: float = 0.0,
test_nms_thresh: float = 0.5,
test_topk_per_image: int = 100,
cls_agnostic_bbox_reg: bool = False,
smooth_l1_beta: float = 0.0,
box_reg_loss_type: str = "smooth_l1",
loss_weight: Union[float, Dict[str, float]] = 1.0,
softmaxes: bool = False,
dropout_probability: float = 0.5,
):
"""
NOTE: this interface is experimental.
Args:
input_shape (ShapeSpec): shape of the input feature to this module
box2box_transform (Box2BoxTransform or Box2BoxTransformRotated):
num_classes (int): number of foreground classes
test_score_thresh (float): threshold to filter predictions results.
test_nms_thresh (float): NMS threshold for prediction results.
test_topk_per_image (int): number of top predictions to produce per image.
cls_agnostic_bbox_reg (bool): whether to use class agnostic for bbox regression
smooth_l1_beta (float): transition point from L1 to L2 loss. Only used if
`box_reg_loss_type` is "smooth_l1"
box_reg_loss_type (str): Box regression loss type. One of: "smooth_l1", "giou"
loss_weight (float|dict): weights to use for losses. Can be single float for weighting
all losses, or a dict of individual weightings. Valid dict keys are:
* "loss_cls": applied to classification loss
* "loss_box_reg": applied to box regression loss
"""
super().__init__()
if isinstance(input_shape, int): # some backward compatibility
input_shape = ShapeSpec(channels=input_shape)
self.num_classes = num_classes
input_size = input_shape.channels * (input_shape.width or 1) * (input_shape.height or 1)
# prediction layer for num_classes foreground classes and one background class (hence + 1)
self.dropout1 = nn.Dropout(p=dropout_probability)
self.cls_score = nn.Linear(input_size, num_classes + 1)
num_bbox_reg_classes = 1 if cls_agnostic_bbox_reg else num_classes
box_dim = len(box2box_transform.weights)
self.dropout2 = nn.Dropout(p=dropout_probability)
self.bbox_pred = nn.Linear(input_size, num_bbox_reg_classes * box_dim)
nn.init.normal_(self.cls_score.weight, std=0.01)
nn.init.normal_(self.bbox_pred.weight, std=0.001)
for l in [self.cls_score, self.bbox_pred]:
nn.init.constant_(l.bias, 0)
self.box2box_transform = box2box_transform
self.smooth_l1_beta = smooth_l1_beta
self.test_score_thresh = test_score_thresh
self.test_nms_thresh = test_nms_thresh
self.test_topk_per_image = test_topk_per_image
self.box_reg_loss_type = box_reg_loss_type
if isinstance(loss_weight, float):
loss_weight = {"loss_cls": loss_weight, "loss_box_reg": loss_weight}
self.loss_weight = loss_weight
self.softmaxes = softmaxes
@classmethod
def from_config(cls, cfg, input_shape):
return {
"input_shape": input_shape,
"box2box_transform": Box2BoxTransform(weights=cfg.MODEL.ROI_BOX_HEAD.BBOX_REG_WEIGHTS),
# fmt: off
"num_classes" : cfg.MODEL.ROI_HEADS.NUM_CLASSES,
"cls_agnostic_bbox_reg" : cfg.MODEL.ROI_BOX_HEAD.CLS_AGNOSTIC_BBOX_REG,
"smooth_l1_beta" : cfg.MODEL.ROI_BOX_HEAD.SMOOTH_L1_BETA,
"test_score_thresh" : cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST,
"test_nms_thresh" : cfg.MODEL.ROI_HEADS.NMS_THRESH_TEST,
"test_topk_per_image" : cfg.TEST.DETECTIONS_PER_IMAGE,
"box_reg_loss_type" : cfg.MODEL.ROI_BOX_HEAD.BBOX_REG_LOSS_TYPE,
"loss_weight" : {"loss_box_reg": cfg.MODEL.ROI_BOX_HEAD.BBOX_REG_LOSS_WEIGHT},
"softmaxes" : cfg.MODEL.ROI_HEADS.SOFTMAXES,
"dropout_probability" : cfg.MODEL.ROI_BOX_HEAD.DROPOUT_PROBABILITY,
# fmt: on
}
def forward(self, x):
"""
Args:
x: per-region features of shape (N, ...) for N bounding boxes to predict.
Returns:
(Tensor, Tensor):
First tensor: shape (N,K+1), scores for each of the N box. Each row contains the
scores for K object categories and 1 background class.
Second tensor: bounding box regression deltas for each box. Shape is shape (N,Kx4),
or (N,4) for class-agnostic regression.
"""
if x.dim() > 2:
x = torch.flatten(x, start_dim=1)
x = self.dropout1(x)
scores = self.cls_score(x)
x = self.dropout2(x)
proposal_deltas = self.bbox_pred(x)
return scores, proposal_deltas
def losses(self, predictions, proposals):
"""
Args:
predictions: return values of :meth:`forward()`.
proposals (list[Instances]): proposals that match the features that were used
to compute predictions. The fields ``proposal_boxes``, ``gt_boxes``,
``gt_classes`` are expected.
Returns:
Dict[str, Tensor]: dict of losses
"""
scores, proposal_deltas = predictions
# parse classification outputs
gt_classes = (
cat([p.gt_classes for p in proposals], dim=0) if len(proposals) else torch.empty(0)
)
_log_classification_stats(scores, gt_classes)
# parse box regression outputs
if len(proposals):
proposal_boxes = cat([p.proposal_boxes.tensor for p in proposals], dim=0) # Nx4
assert not proposal_boxes.requires_grad, "Proposals should not require gradients!"
# If "gt_boxes" does not exist, the proposals must be all negative and
# should not be included in regression loss computation.
# Here we just use proposal_boxes as an arbitrary placeholder because its
# value won't be used in self.box_reg_loss().
gt_boxes = cat(
[(p.gt_boxes if p.has("gt_boxes") else p.proposal_boxes).tensor for p in proposals],
dim=0,
)
else:
proposal_boxes = gt_boxes = torch.empty((0, 4), device=proposal_deltas.device)
losses = {
"loss_cls": cross_entropy(scores, gt_classes, reduction="mean"),
"loss_box_reg": self.box_reg_loss(
proposal_boxes, gt_boxes, proposal_deltas, gt_classes
),
}
return {k: v * self.loss_weight.get(k, 1.0) for k, v in losses.items()}
def box_reg_loss(self, proposal_boxes, gt_boxes, pred_deltas, gt_classes):
"""
Args:
All boxes are tensors with the same shape Rx(4 or 5).
gt_classes is a long tensor of shape R, the gt class label of each proposal.
R shall be the number of proposals.
"""
box_dim = proposal_boxes.shape[1] # 4 or 5
# Regression loss is only computed for foreground proposals (those matched to a GT)
fg_inds = nonzero_tuple((gt_classes >= 0) & (gt_classes < self.num_classes))[0]
if pred_deltas.shape[1] == box_dim: # cls-agnostic regression
fg_pred_deltas = pred_deltas[fg_inds]
else:
fg_pred_deltas = pred_deltas.view(-1, self.num_classes, box_dim)[
fg_inds, gt_classes[fg_inds]
]
if self.box_reg_loss_type == "smooth_l1":
gt_pred_deltas = self.box2box_transform.get_deltas(
proposal_boxes[fg_inds],
gt_boxes[fg_inds],
)
loss_box_reg = smooth_l1_loss(
fg_pred_deltas, gt_pred_deltas, self.smooth_l1_beta, reduction="sum"
)
elif self.box_reg_loss_type == "giou":
fg_pred_boxes = self.box2box_transform.apply_deltas(
fg_pred_deltas, proposal_boxes[fg_inds]
)
loss_box_reg = giou_loss(fg_pred_boxes, gt_boxes[fg_inds], reduction="sum")
else:
raise ValueError(f"Invalid bbox reg loss type '{self.box_reg_loss_type}'")
# The reg loss is normalized using the total number of regions (R), not the number
# of foreground regions even though the box regression loss is only defined on
# foreground regions. Why? Because doing so gives equal training influence to
# each foreground example. To see how, consider two different minibatches:
# (1) Contains a single foreground region
# (2) Contains 100 foreground regions
# If we normalize by the number of foreground regions, the single example in
# minibatch (1) will be given 100 times as much influence as each foreground
# example in minibatch (2). Normalizing by the total number of regions, R,
# means that the single example in minibatch (1) and each of the 100 examples
# in minibatch (2) are given equal influence.
return loss_box_reg / max(gt_classes.numel(), 1.0) # return 0 if empty
def inference(self, predictions: Tuple[torch.Tensor, torch.Tensor], proposals: List[Instances]):
"""
Args:
predictions: return values of :meth:`forward()`.
proposals (list[Instances]): proposals that match the features that were
used to compute predictions. The ``proposal_boxes`` field is expected.
Returns:
list[Instances]: same as `fast_rcnn_inference`.
list[Tensor]: same as `fast_rcnn_inference`.
"""
boxes = self.predict_boxes(predictions, proposals)
scores = self.predict_probs(predictions, proposals)
image_shapes = [x.image_size for x in proposals]
return fast_rcnn_inference(
boxes,
scores,
image_shapes,
self.test_score_thresh,
self.test_nms_thresh,
self.test_topk_per_image,
self.softmaxes,
)
def predict_boxes_for_gt_classes(self, predictions, proposals):
"""
Args:
predictions: return values of :meth:`forward()`.
proposals (list[Instances]): proposals that match the features that were used
to compute predictions. The fields ``proposal_boxes``, ``gt_classes`` are expected.
Returns:
list[Tensor]:
A list of Tensors of predicted boxes for GT classes in case of
class-specific box head. Element i of the list has shape (Ri, B), where Ri is
the number of proposals for image i and B is the box dimension (4 or 5)
"""
if not len(proposals):
return []
scores, proposal_deltas = predictions
proposal_boxes = cat([p.proposal_boxes.tensor for p in proposals], dim=0)
N, B = proposal_boxes.shape
predict_boxes = self.box2box_transform.apply_deltas(
proposal_deltas, proposal_boxes
) # Nx(KxB)
K = predict_boxes.shape[1] // B
if K > 1:
gt_classes = torch.cat([p.gt_classes for p in proposals], dim=0)
# Some proposals are ignored or have a background class. Their gt_classes
# cannot be used as index.
gt_classes = gt_classes.clamp_(0, K - 1)
predict_boxes = predict_boxes.view(N, K, B)[
torch.arange(N, dtype=torch.long, device=predict_boxes.device), gt_classes
]
num_prop_per_image = [len(p) for p in proposals]
return predict_boxes.split(num_prop_per_image)
def predict_boxes(
self, predictions: Tuple[torch.Tensor, torch.Tensor], proposals: List[Instances]
):
"""
Args:
predictions: return values of :meth:`forward()`.
proposals (list[Instances]): proposals that match the features that were
used to compute predictions. The ``proposal_boxes`` field is expected.
Returns:
list[Tensor]:
A list of Tensors of predicted class-specific or class-agnostic boxes
for each image. Element i has shape (Ri, K * B) or (Ri, B), where Ri is
the number of proposals for image i and B is the box dimension (4 or 5)
"""
if not len(proposals):
return []
_, proposal_deltas = predictions
num_prop_per_image = [len(p) for p in proposals]
proposal_boxes = cat([p.proposal_boxes.tensor for p in proposals], dim=0)
predict_boxes = self.box2box_transform.apply_deltas(
proposal_deltas,
proposal_boxes,
) # Nx(KxB)
return predict_boxes.split(num_prop_per_image)
def predict_probs(
self, predictions: Tuple[torch.Tensor, torch.Tensor], proposals: List[Instances]
):
"""
Args:
predictions: return values of :meth:`forward()`.
proposals (list[Instances]): proposals that match the features that were
used to compute predictions.
Returns:
list[Tensor]:
A list of Tensors of predicted class probabilities for each image.
Element i has shape (Ri, K + 1), where Ri is the number of proposals for image i.
"""
scores, _ = predictions
num_inst_per_image = [len(p) for p in proposals]
probs = F.softmax(scores, dim=-1)
return probs.split(num_inst_per_image, dim=0)
@ROI_HEADS_REGISTRY.register()
class Res5ROIHeadsDropout(Res5ROIHeads):
def __init__(self, cfg, input_shape):
super().__init__(cfg, input_shape, box_predictor=FastRCNNOutputLayersDropout(cfg, 2048))
@ROI_HEADS_REGISTRY.register()
class StandardROIHeadsDropout(StandardROIHeads):
def __init__(self, cfg, input_shape):
super().__init__(cfg, input_shape, box_predictor=FastRCNNOutputLayersDropout(cfg, 1024))
@ROI_MASK_HEAD_REGISTRY.register()
class MaskRCNNConvUpsampleHeadDropout(BaseMaskRCNNHead, nn.Sequential):
"""
A mask head with several conv layers, plus an upsample layer (with `ConvTranspose2d`).
Predictions are made with a final 1x1 conv layer.
"""
@configurable
def __init__(self, input_shape: ShapeSpec, *, num_classes, conv_dims, conv_norm="",
dropout_probability: float = 0.5, **kwargs):
"""
NOTE: this interface is experimental.
Args:
input_shape (ShapeSpec): shape of the input feature
num_classes (int): the number of foreground classes (i.e. background is not
included). 1 if using class agnostic prediction.
conv_dims (list[int]): a list of N>0 integers representing the output dimensions
of N-1 conv layers and the last upsample layer.
conv_norm (str or callable): normalization for the conv layers.
See :func:`detectron2.layers.get_norm` for supported types.
"""
super().__init__(**kwargs)
assert len(conv_dims) >= 1, "conv_dims have to be non-empty!"
self.conv_norm_relus = []
cur_channels = input_shape.channels
for k, conv_dim in enumerate(conv_dims[:-1]):
conv = Conv2d(
cur_channels,
conv_dim,
kernel_size=3,
stride=1,
padding=1,
bias=not conv_norm,
norm=get_norm(conv_norm, conv_dim),
activation=nn.ReLU(),
)
self.add_module("mask_fcn{}".format(k + 1), conv)
self.conv_norm_relus.append(conv)
cur_channels = conv_dim
self.dropout1 = nn.Dropout(p=dropout_probability)
self.deconv = ConvTranspose2d(
cur_channels, conv_dims[-1], kernel_size=2, stride=2, padding=0
)
self.add_module("deconv_relu", nn.ReLU())
cur_channels = conv_dims[-1]
self.dropout2 = nn.Dropout(p=dropout_probability)
self.predictor = Conv2d(cur_channels, num_classes, kernel_size=1, stride=1, padding=0)
for layer in self.conv_norm_relus + [self.deconv]:
weight_init.c2_msra_fill(layer)
# use normal distribution initialization for mask prediction layer
nn.init.normal_(self.predictor.weight, std=0.001)
if self.predictor.bias is not None:
nn.init.constant_(self.predictor.bias, 0)
@classmethod
def from_config(cls, cfg, input_shape):
ret = super().from_config(cfg, input_shape)
conv_dim = cfg.MODEL.ROI_MASK_HEAD.CONV_DIM
num_conv = cfg.MODEL.ROI_MASK_HEAD.NUM_CONV
ret.update(
conv_dims=[conv_dim] * (num_conv + 1), # +1 for ConvTranspose
conv_norm=cfg.MODEL.ROI_MASK_HEAD.NORM,
input_shape=input_shape,
)
if cfg.MODEL.ROI_MASK_HEAD.CLS_AGNOSTIC_MASK:
ret["num_classes"] = 1
else:
ret["num_classes"] = cfg.MODEL.ROI_HEADS.NUM_CLASSES
ret["dropout_probability"] = cfg.MODEL.ROI_MASK_HEAD.DROPOUT_PROBABILITY
return ret
def layers(self, x):
for layer in self:
x = layer(x)
return x
|
[
"numpy.prod",
"torch.nn.ReLU",
"torch.nn.Dropout",
"detectron2.modeling.roi_heads.fast_rcnn._log_classification_stats",
"torch.nn.init.constant_",
"detectron2.layers.cross_entropy",
"fvcore.nn.smooth_l1_loss",
"torch.nn.functional.softmax",
"torch.arange",
"detectron2.layers.ShapeSpec",
"torch.nn.Flatten",
"detectron2.layers.cat",
"detectron2.modeling.roi_heads.fast_rcnn.fast_rcnn_inference",
"detectron2.layers.ConvTranspose2d",
"fvcore.nn.weight_init.c2_xavier_fill",
"detectron2.modeling.box_regression.Box2BoxTransform",
"fvcore.nn.weight_init.c2_msra_fill",
"detectron2.modeling.roi_heads.mask_head.ROI_MASK_HEAD_REGISTRY.register",
"detectron2.layers.Conv2d",
"detectron2.layers.nonzero_tuple",
"fvcore.nn.giou_loss",
"detectron2.modeling.roi_heads.box_head.ROI_BOX_HEAD_REGISTRY.register",
"torch.empty",
"torch.nn.init.normal_",
"torch.cat",
"detectron2.modeling.roi_heads.ROI_HEADS_REGISTRY.register",
"detectron2.layers.get_norm",
"torch.nn.Linear",
"torch.flatten"
] |
[((1340, 1372), 'detectron2.modeling.roi_heads.box_head.ROI_BOX_HEAD_REGISTRY.register', 'ROI_BOX_HEAD_REGISTRY.register', ([], {}), '()\n', (1370, 1372), False, 'from detectron2.modeling.roi_heads.box_head import ROI_BOX_HEAD_REGISTRY\n'), ((18533, 18562), 'detectron2.modeling.roi_heads.ROI_HEADS_REGISTRY.register', 'ROI_HEADS_REGISTRY.register', ([], {}), '()\n', (18560, 18562), False, 'from detectron2.modeling.roi_heads import ROI_HEADS_REGISTRY, StandardROIHeads, Res5ROIHeads\n'), ((18751, 18780), 'detectron2.modeling.roi_heads.ROI_HEADS_REGISTRY.register', 'ROI_HEADS_REGISTRY.register', ([], {}), '()\n', (18778, 18780), False, 'from detectron2.modeling.roi_heads import ROI_HEADS_REGISTRY, StandardROIHeads, Res5ROIHeads\n'), ((18973, 19006), 'detectron2.modeling.roi_heads.mask_head.ROI_MASK_HEAD_REGISTRY.register', 'ROI_MASK_HEAD_REGISTRY.register', ([], {}), '()\n', (19004, 19006), False, 'from detectron2.modeling.roi_heads.mask_head import ROI_MASK_HEAD_REGISTRY, BaseMaskRCNNHead\n'), ((6937, 6970), 'torch.nn.Dropout', 'nn.Dropout', ([], {'p': 'dropout_probability'}), '(p=dropout_probability)\n', (6947, 6970), False, 'from torch import nn\n'), ((6996, 7034), 'torch.nn.Linear', 'nn.Linear', (['input_size', '(num_classes + 1)'], {}), '(input_size, num_classes + 1)\n', (7005, 7034), False, 'from torch import nn\n'), ((7183, 7216), 'torch.nn.Dropout', 'nn.Dropout', ([], {'p': 'dropout_probability'}), '(p=dropout_probability)\n', (7193, 7216), False, 'from torch import nn\n'), ((7242, 7295), 'torch.nn.Linear', 'nn.Linear', (['input_size', '(num_bbox_reg_classes * box_dim)'], {}), '(input_size, num_bbox_reg_classes * box_dim)\n', (7251, 7295), False, 'from torch import nn\n'), ((7305, 7353), 'torch.nn.init.normal_', 'nn.init.normal_', (['self.cls_score.weight'], {'std': '(0.01)'}), '(self.cls_score.weight, std=0.01)\n', (7320, 7353), False, 'from torch import nn\n'), ((7362, 7411), 'torch.nn.init.normal_', 'nn.init.normal_', (['self.bbox_pred.weight'], {'std': '(0.001)'}), '(self.bbox_pred.weight, std=0.001)\n', (7377, 7411), False, 'from torch import nn\n'), ((10471, 10516), 'detectron2.modeling.roi_heads.fast_rcnn._log_classification_stats', '_log_classification_stats', (['scores', 'gt_classes'], {}), '(scores, gt_classes)\n', (10496, 10516), False, 'from detectron2.modeling.roi_heads.fast_rcnn import fast_rcnn_inference, _log_classification_stats\n'), ((14819, 14959), 'detectron2.modeling.roi_heads.fast_rcnn.fast_rcnn_inference', 'fast_rcnn_inference', (['boxes', 'scores', 'image_shapes', 'self.test_score_thresh', 'self.test_nms_thresh', 'self.test_topk_per_image', 'self.softmaxes'], {}), '(boxes, scores, image_shapes, self.test_score_thresh,\n self.test_nms_thresh, self.test_topk_per_image, self.softmaxes)\n', (14838, 14959), False, 'from detectron2.modeling.roi_heads.fast_rcnn import fast_rcnn_inference, _log_classification_stats\n'), ((15838, 15894), 'detectron2.layers.cat', 'cat', (['[p.proposal_boxes.tensor for p in proposals]'], {'dim': '(0)'}), '([p.proposal_boxes.tensor for p in proposals], dim=0)\n', (15841, 15894), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((17519, 17575), 'detectron2.layers.cat', 'cat', (['[p.proposal_boxes.tensor for p in proposals]'], {'dim': '(0)'}), '([p.proposal_boxes.tensor for p in proposals], dim=0)\n', (17522, 17575), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((18449, 18474), 'torch.nn.functional.softmax', 'F.softmax', (['scores'], {'dim': '(-1)'}), '(scores, dim=-1)\n', (18458, 18474), True, 'from torch.nn import functional as F\n'), ((20731, 20764), 'torch.nn.Dropout', 'nn.Dropout', ([], {'p': 'dropout_probability'}), '(p=dropout_probability)\n', (20741, 20764), False, 'from torch import nn\n'), ((20788, 20873), 'detectron2.layers.ConvTranspose2d', 'ConvTranspose2d', (['cur_channels', 'conv_dims[-1]'], {'kernel_size': '(2)', 'stride': '(2)', 'padding': '(0)'}), '(cur_channels, conv_dims[-1], kernel_size=2, stride=2, padding=0\n )\n', (20803, 20873), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((21003, 21036), 'torch.nn.Dropout', 'nn.Dropout', ([], {'p': 'dropout_probability'}), '(p=dropout_probability)\n', (21013, 21036), False, 'from torch import nn\n'), ((21063, 21132), 'detectron2.layers.Conv2d', 'Conv2d', (['cur_channels', 'num_classes'], {'kernel_size': '(1)', 'stride': '(1)', 'padding': '(0)'}), '(cur_channels, num_classes, kernel_size=1, stride=1, padding=0)\n', (21069, 21132), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((21320, 21369), 'torch.nn.init.normal_', 'nn.init.normal_', (['self.predictor.weight'], {'std': '(0.001)'}), '(self.predictor.weight, std=0.001)\n', (21335, 21369), False, 'from torch import nn\n'), ((3098, 3131), 'torch.nn.Dropout', 'nn.Dropout', ([], {'p': 'dropout_probability'}), '(p=dropout_probability)\n', (3108, 3131), False, 'from torch import nn\n'), ((3511, 3542), 'fvcore.nn.weight_init.c2_msra_fill', 'weight_init.c2_msra_fill', (['layer'], {}), '(layer)\n', (3535, 3542), True, 'import fvcore.nn.weight_init as weight_init\n'), ((3586, 3619), 'fvcore.nn.weight_init.c2_xavier_fill', 'weight_init.c2_xavier_fill', (['layer'], {}), '(layer)\n', (3612, 3619), True, 'import fvcore.nn.weight_init as weight_init\n'), ((4498, 4519), 'detectron2.layers.ShapeSpec', 'ShapeSpec', ([], {'channels': 'o'}), '(channels=o)\n', (4507, 4519), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((4553, 4602), 'detectron2.layers.ShapeSpec', 'ShapeSpec', ([], {'channels': 'o[0]', 'height': 'o[1]', 'width': 'o[2]'}), '(channels=o[0], height=o[1], width=o[2])\n', (4562, 4602), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((6646, 6677), 'detectron2.layers.ShapeSpec', 'ShapeSpec', ([], {'channels': 'input_shape'}), '(channels=input_shape)\n', (6655, 6677), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((7475, 7503), 'torch.nn.init.constant_', 'nn.init.constant_', (['l.bias', '(0)'], {}), '(l.bias, 0)\n', (7492, 7503), False, 'from torch import nn\n'), ((8156, 8221), 'detectron2.modeling.box_regression.Box2BoxTransform', 'Box2BoxTransform', ([], {'weights': 'cfg.MODEL.ROI_BOX_HEAD.BBOX_REG_WEIGHTS'}), '(weights=cfg.MODEL.ROI_BOX_HEAD.BBOX_REG_WEIGHTS)\n', (8172, 8221), False, 'from detectron2.modeling.box_regression import Box2BoxTransform\n'), ((9612, 9641), 'torch.flatten', 'torch.flatten', (['x'], {'start_dim': '(1)'}), '(x, start_dim=1)\n', (9625, 9641), False, 'import torch\n'), ((10369, 10414), 'detectron2.layers.cat', 'cat', (['[p.gt_classes for p in proposals]'], {'dim': '(0)'}), '([p.gt_classes for p in proposals], dim=0)\n', (10372, 10414), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((10438, 10452), 'torch.empty', 'torch.empty', (['(0)'], {}), '(0)\n', (10449, 10452), False, 'import torch\n'), ((10613, 10669), 'detectron2.layers.cat', 'cat', (['[p.proposal_boxes.tensor for p in proposals]'], {'dim': '(0)'}), '([p.proposal_boxes.tensor for p in proposals], dim=0)\n', (10616, 10669), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((11288, 11338), 'torch.empty', 'torch.empty', (['(0, 4)'], {'device': 'proposal_deltas.device'}), '((0, 4), device=proposal_deltas.device)\n', (11299, 11338), False, 'import torch\n'), ((11383, 11434), 'detectron2.layers.cross_entropy', 'cross_entropy', (['scores', 'gt_classes'], {'reduction': '"""mean"""'}), "(scores, gt_classes, reduction='mean')\n", (11396, 11434), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((12141, 12207), 'detectron2.layers.nonzero_tuple', 'nonzero_tuple', (['((gt_classes >= 0) & (gt_classes < self.num_classes))'], {}), '((gt_classes >= 0) & (gt_classes < self.num_classes))\n', (12154, 12207), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((12715, 12803), 'fvcore.nn.smooth_l1_loss', 'smooth_l1_loss', (['fg_pred_deltas', 'gt_pred_deltas', 'self.smooth_l1_beta'], {'reduction': '"""sum"""'}), "(fg_pred_deltas, gt_pred_deltas, self.smooth_l1_beta,\n reduction='sum')\n", (12729, 12803), False, 'from fvcore.nn import giou_loss, smooth_l1_loss\n'), ((16141, 16192), 'torch.cat', 'torch.cat', (['[p.gt_classes for p in proposals]'], {'dim': '(0)'}), '([p.gt_classes for p in proposals], dim=0)\n', (16150, 16192), False, 'import torch\n'), ((20930, 20939), 'torch.nn.ReLU', 'nn.ReLU', ([], {}), '()\n', (20937, 20939), False, 'from torch import nn\n'), ((21205, 21236), 'fvcore.nn.weight_init.c2_msra_fill', 'weight_init.c2_msra_fill', (['layer'], {}), '(layer)\n', (21229, 21236), True, 'import fvcore.nn.weight_init as weight_init\n'), ((21426, 21467), 'torch.nn.init.constant_', 'nn.init.constant_', (['self.predictor.bias', '(0)'], {}), '(self.predictor.bias, 0)\n', (21443, 21467), False, 'from torch import nn\n'), ((3373, 3382), 'torch.nn.ReLU', 'nn.ReLU', ([], {}), '()\n', (3380, 3382), False, 'from torch import nn\n'), ((13039, 13099), 'fvcore.nn.giou_loss', 'giou_loss', (['fg_pred_boxes', 'gt_boxes[fg_inds]'], {'reduction': '"""sum"""'}), "(fg_pred_boxes, gt_boxes[fg_inds], reduction='sum')\n", (13048, 13099), False, 'from fvcore.nn import giou_loss, smooth_l1_loss\n'), ((2654, 2683), 'detectron2.layers.get_norm', 'get_norm', (['conv_norm', 'conv_dim'], {}), '(conv_norm, conv_dim)\n', (2662, 2683), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((2712, 2721), 'torch.nn.ReLU', 'nn.ReLU', ([], {}), '()\n', (2719, 2721), False, 'from torch import nn\n'), ((3062, 3074), 'torch.nn.Flatten', 'nn.Flatten', ([], {}), '()\n', (3072, 3074), False, 'from torch import nn\n'), ((3227, 3253), 'numpy.prod', 'np.prod', (['self._output_size'], {}), '(self._output_size)\n', (3234, 3253), True, 'import numpy as np\n'), ((16445, 16507), 'torch.arange', 'torch.arange', (['N'], {'dtype': 'torch.long', 'device': 'predict_boxes.device'}), '(N, dtype=torch.long, device=predict_boxes.device)\n', (16457, 16507), False, 'import torch\n'), ((20479, 20508), 'detectron2.layers.get_norm', 'get_norm', (['conv_norm', 'conv_dim'], {}), '(conv_norm, conv_dim)\n', (20487, 20508), False, 'from detectron2.layers import Conv2d, ConvTranspose2d, ShapeSpec, batched_nms, cat, cross_entropy, nonzero_tuple, get_norm\n'), ((20537, 20546), 'torch.nn.ReLU', 'nn.ReLU', ([], {}), '()\n', (20544, 20546), False, 'from torch import nn\n')]
|
#thomas feiring model
import math
import numpy as np
import pandas as pd
#enter the year for which you need prediction starting 2019
year=2019
number_of_days=365
day=0
df = pd.read_csv('groundtruth.csv')
u=df['Mean']
X_t= u[0]
sd=df['St dev']
print("Month,Year,Inflow")
#lag -1 correlation
lag=df['co relation']
np.random.seed(9001)
for i in range(number_of_days):
rn=np.random.normal(0,1,1)[0]
z_t=(X_t-u[day])/sd[day]
z_t1=lag[day]*z_t+rn*math.sqrt(1-lag[day]*lag[day])
X_t1=u[(day+1)%365]+z_t1*sd[(day+1)%365]
print(day+1,",",year,",",X_t1)
if(day==364):
year=year+1
day=0
day=day+1
X_t=X_t1
|
[
"numpy.random.normal",
"math.sqrt",
"numpy.random.seed",
"pandas.read_csv"
] |
[((174, 204), 'pandas.read_csv', 'pd.read_csv', (['"""groundtruth.csv"""'], {}), "('groundtruth.csv')\n", (185, 204), True, 'import pandas as pd\n'), ((313, 333), 'numpy.random.seed', 'np.random.seed', (['(9001)'], {}), '(9001)\n', (327, 333), True, 'import numpy as np\n'), ((373, 398), 'numpy.random.normal', 'np.random.normal', (['(0)', '(1)', '(1)'], {}), '(0, 1, 1)\n', (389, 398), True, 'import numpy as np\n'), ((454, 488), 'math.sqrt', 'math.sqrt', (['(1 - lag[day] * lag[day])'], {}), '(1 - lag[day] * lag[day])\n', (463, 488), False, 'import math\n')]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.