-
Notifications
You must be signed in to change notification settings - Fork 0
/
pseudo_source_data_iterator.py
263 lines (218 loc) · 9.7 KB
/
pseudo_source_data_iterator.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
import numpy
import gzip
import shuffle
from util import load_dict
def fopen(filename, mode='r'):
if filename.endswith('.gz'):
return gzip.open(filename, mode)
return open(filename, mode)
class PseudoSourceTextIterator:
"""Bitext iterator processing parallel data,
as well as parallel data with a pseudo-source."""
def __init__(self, source, target,
pseudo_source, pseudo_target,
source_dicts, target_dict,
batch_size=128,
maxlen=100,
n_words_source=-1,
n_words_target=-1,
skip_empty=False,
shuffle_each_epoch=False,
sort_by_length=True,
use_factor=False,
maxibatch_size=20,
noise=False):
if shuffle_each_epoch:
self.source_orig = source
self.target_orig = target
self.pseudo_source_orig = pseudo_source
self.pseudo_target_orig = pseudo_target
self.source, self.target = shuffle.main([self.source_orig, self.target_orig], temporary=True)
self.pseudo_source, self.pseudo_target = shuffle.main([self.pseudo_source_orig, self.pseudo_target_orig], temporary=True)
else:
self.source = fopen(source, 'r')
self.target = fopen(target, 'r')
self.pseudo_source = fopen(pseudo_source, 'r')
self.pseudo_target = fopen(pseudo_target, 'r')
self.source_dicts = []
for source_dict in source_dicts:
self.source_dicts.append(load_dict(source_dict))
self.target_dict = load_dict(target_dict)
# batch size in divided by two: one part
# is for parallel data, the other is for
# the pseudo-source data.
self.batch_size = batch_size/2
self.maxlen = maxlen
self.skip_empty = skip_empty
self.use_factor = use_factor
self.pseudo_src_noise = noise
self.n_words_source = n_words_source
self.n_words_target = n_words_target
if self.n_words_source > 0:
for d in self.source_dicts:
for key, idx in d.items():
if idx >= self.n_words_source:
del d[key]
if self.n_words_target > 0:
for key, idx in self.target_dict.items():
if idx >= self.n_words_target:
del self.target_dict[key]
self.shuffle = shuffle_each_epoch
self.sort_by_length = sort_by_length
self.source_buffer = []
self.target_buffer = []
self.pseudo_source_buffer = []
self.pseudo_target_buffer = []
self.k = batch_size * maxibatch_size
self.end_of_data = False
def __iter__(self):
return self
def __len__(self):
return sum([1 for _ in self])
def reset(self):
# clear buffers for new epoch
self.source_buffer = []
self.target_buffer = []
self.pseudo_source_buffer = []
self.pseudo_target_buffer = []
if self.shuffle:
self.source, self.target = shuffle.main([self.source_orig, self.target_orig], temporary=True)
self.pseudo_source, self.pseudo_target = shuffle.main([self.pseudo_source_orig, self.pseudo_target_orig], temporary=True)
else:
self.source.seek(0)
self.target.seek(0)
self.pseudo_source.seek(0)
self.pseudo_target.seek(0)
def next(self):
if self.end_of_data:
self.end_of_data = False
self.reset()
raise StopIteration
source = []
target = []
pseudo_source = []
pseudo_target = []
# fill buffer, if it's empty
assert len(self.source_buffer) == len(self.target_buffer), 'Buffer size mismatch! (real data)'
assert len(self.pseudo_source_buffer) == len(self.pseudo_target_buffer), 'Buffer size mismatch! (pseudo data)'
if len(self.source_buffer) == 0:
for ss in self.source:
ss = ss.split()
tt = self.target.readline().split()
if self.skip_empty and (len(ss) == 0 or len(tt) == 0):
continue
if len(ss) > self.maxlen or len(tt) > self.maxlen:
continue
self.source_buffer.append(ss)
self.target_buffer.append(tt)
if len(self.source_buffer) == self.k:
break
if len(self.source_buffer) == 0 or len(self.target_buffer) == 0:
self.end_of_data = False
self.reset()
raise StopIteration
# sort by target buffer
if self.sort_by_length:
tlen = numpy.array([len(t) for t in self.target_buffer])
tidx = tlen.argsort()
_sbuf = [self.source_buffer[i] for i in tidx]
_tbuf = [self.target_buffer[i] for i in tidx]
self.source_buffer = _sbuf
self.target_buffer = _tbuf
else:
self.source_buffer.reverse()
self.target_buffer.reverse()
if len(self.pseudo_source_buffer) == 0:
for ss in self.pseudo_source:
ss = ss.split()
tt = self.pseudo_target.readline().split()
if self.skip_empty and (len(ss) == 0 or len(tt) == 0):
continue
if len(ss) > self.maxlen or len(tt) > self.maxlen:
continue
self.pseudo_source_buffer.append(ss)
self.pseudo_target_buffer.append(tt)
if len(self.pseudo_source_buffer) == self.k:
break
if len(self.pseudo_source_buffer) == 0 or len(self.pseudo_target_buffer) == 0:
self.end_of_data = False
self.reset()
raise StopIteration
# sort by target buffer
if self.sort_by_length:
tlen = numpy.array([len(t) for t in self.pseudo_target_buffer])
tidx = tlen.argsort()
_sbuf = [self.pseudo_source_buffer[i] for i in tidx]
_tbuf = [self.pseudo_target_buffer[i] for i in tidx]
self.pseudo_source_buffer = _sbuf
self.pseudo_target_buffer = _tbuf
else:
self.pseudo_source_buffer.reverse()
self.pseudo_target_buffer.reverse()
try:
# actual work here
while True:
# read from source file and map to word index
try:
ss = self.source_buffer.pop()
except IndexError:
break
tmp = []
for w in ss:
if self.use_factor:
w = [self.source_dicts[i][f] if f in self.source_dicts[i] else 1 for (i,f) in enumerate(w.split('|'))]
else:
w = [self.source_dicts[0][w] if w in self.source_dicts[0] else 1]
tmp.append(w)
ss = tmp
# read from source file and map to word index
tt = self.target_buffer.pop()
tt = [self.target_dict[w] if w in self.target_dict else 1
for w in tt]
if self.n_words_target > 0:
tt = [w if w < self.n_words_target else 1 for w in tt]
# read from source file and map to word index (pseudo source)
try:
pss = self.pseudo_source_buffer.pop()
except IndexError:
break
if self.pseudo_src_noise:
# integrate noise in pseudo data (target copies)
# drop words
import random
p_drop = 0.1
pss = [s for s in pss if random.random() > p_drop]
# word permutations: a word can't be further
# than k words from its initial position.
k = 3
check_perm = [True for s in pss]
for i in range(len(pss)):
if check_perm[i] and random.random() < 0.1:
choose = [ii for ii in range(i-k, i+k+1) if ii >= 0 and ii < len(pss) and check_perm[ii]]
c = random.choice(choose)
# word permutation
pss[c], pss[i] = pss[i], pss[c]
check_perm[i] = check_perm[c] = False
tmp = []
for w in pss:
w = [self.source_dicts[0][w] if w in self.source_dicts[0] else 1]
tmp.append(w)
pss = tmp
# read from source file and map to word index (pseudo target)
ptt = self.pseudo_target_buffer.pop()
ptt = [self.target_dict[w] if w in self.target_dict else 1
for w in ptt]
if self.n_words_target > 0:
ptt = [w if w < self.n_words_target else 1 for w in ptt]
source.append(ss)
target.append(tt)
pseudo_source.append(pss)
pseudo_target.append(ptt)
if len(source) >= self.batch_size or \
len(target) >= self.batch_size or \
len(pseudo_source) >= self.batch_size or \
len(pseudo_target) >= self.batch_size:
break
except IOError:
self.end_of_data = True
return source, target, pseudo_source, pseudo_target