-
Notifications
You must be signed in to change notification settings - Fork 6
/
utils.py
executable file
·317 lines (248 loc) · 11.4 KB
/
utils.py
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
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
import os
import numpy as np
import h5py
import json
import torch
import cv2
from scipy.misc import imread, imresize
from tqdm import tqdm
from collections import Counter
from random import seed, choice, sample
import pandas as pd
import dicom
import json
import re
import nltk
nltk.download('punkt')
nltk.download('stopwords')
from nltk.tokenize import word_tokenize
from nltk.corpus import stopwords
def iterate_csv(dataframe, word_freq, max_len, stopword=False):
dic = {}
image_paths = []
image_report = []
image_files = os.listdir('/data/medg/misc/interpretable-report-gen/cache/images/')
stop_words = stopwords.words('english')
my_new_stop_words = ['the','and','to','of','was','with','a','on','in','for','name',
'is','patient','s','he','at','as','or','one','she','his','her','am',
'were','you','pt','pm','by','be','had','your','this','date',
'from','there','an','that','p','are','have','has','h','but','o',
'namepattern','which','every','also','should','if','it','been','who','during', 'x']
stop_words.extend(my_new_stop_words)
stop_words = set(stop_words)
for idx, row in tqdm(dataframe.iterrows(), total=dataframe.shape[0]):
if (str(row['dicom_id']) + '.png') in image_files:
reports = []
path = os.path.join('/data/medg/misc/interpretable-report-gen/cache/images', str(row['dicom_id']) + '.png')
report = row['text']
if report:
tokens = word_tokenize(report)
if stopword:
filtered_tokens = [w for w in tokens if not w in stop_words]
else:
filtered_tokens = tokens
word_freq.update(filtered_tokens)
if len(filtered_tokens) <= max_len:
reports.append(filtered_tokens)
if len(reports) == 0:
continue
image_paths.append(path)
image_report.append(reports)
dic['images'] = image_paths
dic['report'] = image_report
dic['label'] = dataframe.as_matrix(columns=dataframe.columns[6:])
return dic
def create_input_files(dataset, captions_per_image, min_word_freq, output_folder,
max_len=100):
assert dataset in {'mimiccxr'}
# load data into three set
train = pd.read_csv('/data/medg/misc/liuguanx/TieNet/split/train.tsv', sep='\t')
val = pd.read_csv('/data/medg/misc/liuguanx/TieNet/split/val.tsv', sep='\t')
test = pd.read_csv('/data/medg/misc/liuguanx/TieNet/split/test.tsv', sep='\t')
# Read image paths and reports for each image
word_freq = Counter()
train_dict = iterate_csv(train,word_freq,max_len)
train_image_paths = train_dict['images']
train_image_captions = train_dict['report']
val_dict = iterate_csv(val,word_freq,max_len)
val_image_paths = val_dict['images']
val_image_captions = val_dict['report']
test_dict = iterate_csv(test,word_freq,max_len)
test_image_paths = test_dict['images']
test_image_captions = test_dict['report']
# Sanity check
assert len(train_image_paths) == len(train_image_captions)
assert len(val_image_paths) == len(val_image_captions)
assert len(test_image_paths) == len(test_image_captions)
# Create word map
words = [w for w in word_freq.keys() if word_freq[w] > min_word_freq]
word_map = {k: v + 1 for v, k in enumerate(words)}
word_map['<unk>'] = len(word_map) + 1
word_map['<start>'] = len(word_map) + 1
word_map['<end>'] = len(word_map) + 1
word_map['<pad>'] = 0
# Create a base/root name for all output files
base_filename = dataset + '_' + str(captions_per_image) + '_cap_per_img_' + str(min_word_freq) + '_min_word_freq'
np.save(os.path.join(output_folder, 'TRAIN' + '_LABELS_' + base_filename + '.npy'), train_dict['label'])
np.save(os.path.join(output_folder, 'VAL' + '_LABELS_' + base_filename + '.npy'), val_dict['label'])
np.save(os.path.join(output_folder, 'TEST' + '_LABELS_' + base_filename + '.npy'), test_dict['label'])
# Save word map to a JSON
with open(os.path.join(output_folder, 'WORDMAP_' + base_filename + '.json'), 'w') as j:
json.dump(word_map, j)
# Sample captions for each image, save images to HDF5 file, and captions and their lengths to JSON files
seed(123)
for impaths, imcaps, split in [(train_image_paths, train_image_captions, 'TRAIN'),
(val_image_paths, val_image_captions, 'VAL'),
(test_image_paths, test_image_captions, 'TEST')]:
with h5py.File(os.path.join(output_folder, split + '_IMAGES_' + base_filename + '.hdf5'), 'a') as h:
# Make a note of the number of captions we are sampling per image
h.attrs['captions_per_image'] = captions_per_image
# Create dataset inside HDF5 file to store images
images = h.create_dataset('images', (len(impaths), 3, 256, 256), dtype='uint8')
print("\nReading %s images and captions, storing to file...\n" % split)
enc_captions = []
caplens = []
for i, path in enumerate(tqdm(impaths)):
# Sample captions
if len(imcaps[i]) < captions_per_image:
captions = imcaps[i] + [choice(imcaps[i]) for _ in range(captions_per_image - len(imcaps[i]))]
else:
captions = sample(imcaps[i], k=captions_per_image)
# Sanity check
assert len(captions) == captions_per_image
# Read images
img = imread(impaths[i])
if len(img.shape) == 2:
img = img[:, :, np.newaxis]
img = np.concatenate([img, img, img], axis=2)
img = imresize(img, (256, 256))
img = img.transpose(2, 0, 1)
assert img.shape == (3, 256, 256)
assert np.max(img) <= 255
# Save image to HDF5 file
images[i] = img
for j, c in enumerate(captions):
# Encode captions
enc_c = [word_map['<start>']] + [word_map.get(word, word_map['<unk>']) for word in c] + [
word_map['<end>']] + [word_map['<pad>']] * (max_len - len(c))
# Find caption lengths
c_len = len(c) + 2
enc_captions.append(enc_c)
caplens.append(c_len)
# Sanity check
assert images.shape[0] * captions_per_image == len(enc_captions) == len(caplens)
# Save encoded captions and their lengths to JSON files
with open(os.path.join(output_folder, split + '_CAPTIONS_' + base_filename + '.json'), 'w') as j:
json.dump(enc_captions, j)
with open(os.path.join(output_folder, split + '_CAPLENS_' + base_filename + '.json'), 'w') as j:
json.dump(caplens, j)
def init_embedding(embeddings):
"""
Fills embedding tensor with values from the uniform distribution.
:param embeddings: embedding tensor
"""
bias = np.sqrt(3.0 / embeddings.size(1))
torch.nn.init.uniform_(embeddings, -bias, bias)
def load_embeddings(emb_file, word_map):
"""
Creates an embedding tensor for the specified word map, for loading into the model.
:param emb_file: file containing embeddings (stored in GloVe format)
:param word_map: word map
:return: embeddings in the same order as the words in the word map, dimension of embeddings
"""
# Find embedding dimension
with open(emb_file, 'r') as f:
emb_dim = len(f.readline().split(' ')) - 1
vocab = set(word_map.keys())
# Create tensor to hold embeddings, initialize
embeddings = torch.FloatTensor(len(vocab), emb_dim)
init_embedding(embeddings)
# Read embedding file
print("\nLoading embeddings...")
for line in open(emb_file, 'r'):
line = line.split(' ')
emb_word = line[0]
embedding = list(map(lambda t: float(t), filter(lambda n: n and not n.isspace(), line[1:])))
# Ignore word if not in train_vocab
if emb_word not in vocab:
continue
embeddings[word_map[emb_word]] = torch.FloatTensor(embedding)
return embeddings, emb_dim
def clip_gradient(optimizer, grad_clip):
"""
Clips gradients computed during backpropagation to avoid explosion of gradients.
:param optimizer: optimizer with the gradients to be clipped
:param grad_clip: clip value
"""
for group in optimizer.param_groups:
for param in group['params']:
if param.grad is not None:
param.grad.data.clamp_(-grad_clip, grad_clip)
def save_checkpoint(data_name, epoch, epochs_since_improvement, encoder, decoder, jointlearner, encoder_optimizer, decoder_optimizer, jointlearner_optimizer,
bleu4, best_bleu4, is_best, dest_dir):
"""
Saves model checkpoint.
:param data_name: base name of processed dataset
:param epoch: epoch number
:param epochs_since_improvement: number of epochs since last improvement in BLEU-4 score
:param encoder: encoder model
:param decoder: decoder model
:param encoder_optimizer: optimizer to update encoder's weights, if fine-tuning
:param decoder_optimizer: optimizer to update decoder's weights
:param bleu4: validation BLEU-4 score for this epoch
:param is_best: is this checkpoint the best so far?
"""
state = {'epoch': epoch,
'epochs_since_improvement': epochs_since_improvement,
'bleu-4': bleu4,
'best_bleu': best_bleu4,
'encoder': encoder,
'decoder': decoder,
'jointlearner': jointlearner,
'encoder_optimizer': encoder_optimizer,
'decoder_optimizer': decoder_optimizer,
'jointlearner_optimizer': jointlearner_optimizer}
filename = 'checkpoint_' + data_name + '.pth.tar'
if is_best:
torch.save(state, os.path.join(dest_dir, 'BEST_' + filename))
filename = os.path.join(dest_dir, filename)
torch.save(state, filename)
class AverageMeter(object):
"""
Keeps track of most recent, average, sum, and count of a metric.
"""
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 adjust_learning_rate(optimizer, shrink_factor):
"""
Shrinks learning rate by a specified factor.
:param optimizer: optimizer whose learning rate must be shrunk.
:param shrink_factor: factor in interval (0, 1) to multiply learning rate with.
"""
print("\nDECAYING learning rate.")
for param_group in optimizer.param_groups:
param_group['lr'] = param_group['lr'] * shrink_factor
print("The new learning rate is %f\n" % (optimizer.param_groups[0]['lr'],))
def accuracy(scores, targets, k):
"""
Computes top-k accuracy, from predicted and true labels.
:param scores: scores from the model
:param targets: true labels
:param k: k in top-k accuracy
:return: top-k accuracy
"""
batch_size = targets.size(0)
_, ind = scores.topk(k, 1, True, True)
correct = ind.eq(targets.view(-1, 1).expand_as(ind))
correct_total = correct.view(-1).float().sum() # 0D tensor
return correct_total.item() * (100.0 / batch_size)