-
Notifications
You must be signed in to change notification settings - Fork 0
/
analysis.py
165 lines (132 loc) · 6.65 KB
/
analysis.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
"""Rough script to experiment with a trained model."""
from model import QuickThoughtsModel
import tensorflow as tf
from create_tfrecords import parse_single_example
import glob
import numpy as np
import re
from tensorflow.contrib.data import padded_batch_and_drop_remainder
FLAGS = tf.flags.FLAGS
tf.flags.DEFINE_integer('batch_size', 300, 'Batch Size')
tf.flags.DEFINE_integer('embedding_size', 100, 'Size of sentence embedding')
tf.flags.DEFINE_integer("context_size", 1, "Prediction context size")
tf.flags.DEFINE_integer('wembedding_size', 100, 'Word embeddings dimension size')
tf.flags.DEFINE_string('tfrecords_input', 'data/tfrecords/', 'Path to directory where tfrecords will be read from.')
tf.flags.DEFINE_string('logdir', 'logdir', 'Path to logdir.')
tf.flags.DEFINE_string('vocab_file', 'data/vocab_100000.txt', 'Path to vocab file')
tf.flags.DEFINE_string('checkpoint_to_load', 'YOUR CHECKPOINT HERE', 'Path to checkpoint file')
PADDED_SHAPES = ({'ids': [FLAGS.max_sentence_length], 'mask': [FLAGS.max_sentence_length]})
vocab = []
for i, word in enumerate(tf.gfile.FastGFile(FLAGS.vocab_file)):
vocab.append(word.decode("utf-8").strip())
def num_tfrecords(files):
"""Count tfrecords in data"""
return sum(sum((1 for _ in tf.python_io.tf_record_iterator(str(file_)))) for file_ in files)
def ids_to_sentence(ids):
"""Convert list of vocab indexes into sentence"""
return ' '.join([vocab[i] for i in ids])
def sentence_to_ids(sentence, vocab):
"""Convert a sentence string to a list of ints"""
sentence = sentence.lower()
sentence = ''.join(re.findall('[a-z ]', sentence))
def word_to_idx(word):
if word not in vocab: return 0
else: return vocab.index(word)
ids = [word_to_idx(w) for w in sentence.split()[:FLAGS.max_sentence_length]]
mask_len = len(ids)
while len(ids) < FLAGS.max_sentence_length:
ids.append(0)
mask = [int(i < mask_len) for i in range(FLAGS.max_sentence_length)]
return ids, mask
def construct_encoding_dictionary():
"""Build a dictionary of sentence:encoding pairs"""
sentence_encoding = {}
dev_graph = tf.Graph()
with dev_graph.as_default():
train_files = glob.glob('data/tfrecords/*train.tfrecord')
train_data = tf.data.TFRecordDataset(train_files) \
.map(parse_single_example) \
.apply(padded_batch_and_drop_remainder(FLAGS.batch_size, PADDED_SHAPES))
training_iterator = train_data.make_initializable_iterator()
global_step = tf.Variable(0, trainable=False)
model = QuickThoughtsModel()
sentences = training_iterator.get_next()
encoding = model.encode(sentences['ids'], sentences['mask'])
loss = model.loss(encoding)
opt = tf.train.AdamOptimizer(0.001)
gradients = opt.compute_gradients(loss)
metric_summary = tf.summary.merge_all()
with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
step = opt.apply_gradients(gradients, global_step=global_step)
dev_saver = tf.train.Saver(var_list=tf.global_variables())
devSess = tf.Session()
dev_saver.restore(devSess, FLAGS.checkpoint_to_load)
devSess.run(training_iterator.initializer)
i =0
while True:
i+=1
print('Collecting {}'.format(i))
try:
ids_, mask_, loss_, encoding_ = devSess.run([sentences['ids'], sentences['mask'], loss, encoding])
except tf.errors.OutOfRangeError:
break
sents = [(ids_to_sentence(x[:sum(y)]), z) for (x, y, z) in zip(ids_, mask_, encoding_)]
sentence_encoding.update(sents)
return sentence_encoding
def custom_search(sentence_encoding, vocab):
"""Encodes a custom sentence and searches against other sentences for closest match."""
dev_graph = tf.Graph()
with dev_graph.as_default():
train_files = glob.glob('data/tfrecords/*train.tfrecord')
train_data = tf.data.TFRecordDataset(train_files) \
.map(parse_single_example) \
.apply(padded_batch_and_drop_remainder(FLAGS.batch_size, PADDED_SHAPES))
training_iterator = train_data.make_initializable_iterator()
global_step = tf.Variable(0, trainable=False)
model = QuickThoughtsModel()
ids_ph = tf.placeholder(dtype=tf.int32, shape=[FLAGS.batch_size ,FLAGS.max_sentence_length], name='ids')
mask_ph = tf.placeholder(dtype=tf.int32, shape=[FLAGS.batch_size, FLAGS.max_sentence_length], name='ids_mask')
encoding = model.encode(ids_ph, mask_ph)
loss = model.loss(encoding)
opt = tf.train.AdamOptimizer(0.001)
gradients = opt.compute_gradients(loss)
metric_summary = tf.summary.merge_all()
with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
step = opt.apply_gradients(gradients, global_step=global_step)
dev_saver = tf.train.Saver(var_list=tf.global_variables())
devSess = tf.Session()
dev_saver.restore(devSess, FLAGS.checkpoint_to_load)
devSess.run(training_iterator.initializer)
while True:
try:
sentence = input('\nEnter custom text to search:')
except Exception as e:
print('Failed, try again "{}"'.format(e))
ids, mask = sentence_to_ids(sentence, vocab)
ids = [ids] * FLAGS.batch_size
mask = [mask] * FLAGS.batch_size
encoding_ = devSess.run(encoding, {ids_ph: ids, mask_ph: mask})
print("Similar sentences to: '{}'".format(sentence))
print_closest_matches_to_sentence(encoding_, sentence_encoding)
def print_closest_matches_to_sentence(encoding, sentence_encodings):
"""Print the closest matches to the given sentence encoding"""
keys = sentence_encodings.keys()
mat = np.array(sentence_encodings.values())
mat_sim = np.matmul(encoding, mat.T)[0]
similar_ids = np.argsort(mat_sim)[::-1]
for id in similar_ids[:5]:
print("\t{}: {}".format(mat_sim[id], keys[id]))
def print_similar_sentences(sentence_encoding):
"""Given the sentence:encoding dict, print out the most similar sentences from the dict."""
keys = sentence_encoding.keys()
mat = np.array(sentence_encoding.values())
mat_sim = np.matmul(mat, mat.T)
for row in range(10):
print('\n\nOriginal sentence: {}'.format(keys[row]))
similar_ids = np.argsort(mat_sim[row])[::-1]
for id in similar_ids[:5]:
print("\t{}: {}".format(mat_sim[row][id], keys[id]))
if __name__ == '__main__':
sentence_encoding = construct_encoding_dictionary()
print_similar_sentences(sentence_encoding)
custom_search(sentence_encoding, vocab)