/
textencoder.py
571 lines (448 loc) · 20.8 KB
/
textencoder.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
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
import collections
import logging
import multiprocessing
import os
import re
import tempfile
from itertools import chain
import cachetools
import torch
from fairseq.data import Dictionary
PAD = "<PAD>_"
EOS = "<EOS>_"
UNK = "<UNK>_"
RESERVED_TOKENS = ["<Lua_Heritage>", PAD, EOS, UNK]
PAD_ID = RESERVED_TOKENS.index(PAD) # Normally 1
EOS_ID = RESERVED_TOKENS.index(EOS) # Normally 2
UNK_ID = RESERVED_TOKENS.index(UNK) # Normally 3
_UNESCAPE_REGEX = re.compile(r"\\u|\\\\|\\([0-9]+);")
_ESCAPE_CHARS = set(u"\\_u;0123456789")
def _escape_token(token, alphabet=None):
"""Escape away underscores and OOV characters and append '_'.
This allows the token to be expressed as the concatenation of a list
of subtokens from the vocabulary. The underscore acts as a sentinel
which allows us to invertibly concatenate multiple such lists.
Args:
token: A unicode string to be escaped.
alphabet: A set of all characters in the vocabulary's alphabet.
Returns:
escaped_token: An escaped unicode string.
Raises:
ValueError: If the provided token is not unicode.
"""
if not isinstance(token, str):
raise ValueError("Expected string type for token, got %s" % type(token))
token = token.replace("\\", "\\\\").replace("_", "\\u")
if alphabet is not None:
chars = [c if c in alphabet and c != u"\n" else r"\%d;" % ord(c) for c in token]
token = ''.join(chars)
return token + "_"
def _unescape_token(escaped_token):
"""Inverse of _escape_token().
Args:
escaped_token: a unicode string
Returns:
token: a unicode string
"""
def match(m):
if m.group(1) is None:
return "_" if m.group(0) == "\\u" else "\\"
try:
return chr(int(m.group(1)))
except (ValueError, OverflowError) as _:
return "\u3013" # Unicode for undefined character.
trimmed = escaped_token[:-1] if escaped_token.endswith("_") else escaped_token
return _UNESCAPE_REGEX.sub(match, trimmed)
def _collect_counts_from_file(filename):
counter = collections.Counter()
with open(filename, 'r', encoding='utf-8') as f:
for line in f.readlines():
for word in line.strip().split():
counter[word] += 1
return counter
def _build_from_token_counts(args):
counter, min_count, count_threshold = args
dictionary = SubwordDictionary.build_from_token_counts(counter, min_count,
num_iterations=2, count_threshold=count_threshold)
return min_count, len(dictionary)
class _SubwordDictionaryFactory(object):
__INITIAL_MAX_SIZE = 16e3
def __init__(self, target_size, vocab_threads=1, custom_tokens=None, padding_factor=8, count_threshold=None):
self._approx_vocab_size = target_size
self._vocab_threads = vocab_threads
self._custom_tokens = [_escape_token(t) for t in custom_tokens] if custom_tokens is not None else []
self._padding_factor = padding_factor
self._count_threshold = count_threshold
self._logger = logging.getLogger("SubwordDictionary::Factory")
def build(self, files, tmp_path=None):
if tmp_path is None:
tmp_path = tempfile.gettempdir()
else:
os.makedirs(tmp_path, exist_ok=True)
token_counts_file = os.path.join(tmp_path, 'token_counts.dict')
if os.path.isfile(token_counts_file):
token_counts = self._load_token_counts(token_counts_file)
else:
token_counts = self._collect_token_counts(files)
self._save_token_counts(token_counts, token_counts_file)
target_size = self._approx_vocab_size
# Searching the minimum max_size
max_size = self.__INITIAL_MAX_SIZE
while True:
max_size, success = self._run_max_size_attempt(max_size, token_counts)
if success:
break
min_size = 1 if max_size == self.__INITIAL_MAX_SIZE else int(max_size / 2)
self._logger.info("Generating vocab file: min = %d, max = %d" % (min_size, max_size))
ret = self._build_to_target_size(target_size, token_counts, min_size, max_size)
# Pad to padding factor if necessary
if self._padding_factor > 1 and len(ret) % self._padding_factor > 0:
ret.force_length(len(ret) + (self._padding_factor - len(ret) % self._padding_factor))
return ret
def _collect_token_counts(self, files):
self._logger.info("Collecting counts BEGIN")
pool = multiprocessing.Pool(processes=min(os.cpu_count() or 1, 16))
try:
counts_array = pool.map(_collect_counts_from_file, files)
counts = counts_array[0]
for c in counts_array[1:]:
counts.update(c)
self._logger.info("Collecting counts END")
return counts
finally:
pool.terminate()
@staticmethod
def _load_token_counts(filename):
token_counts = {}
with open(filename, 'r', encoding='utf-8') as f:
for line in f:
count, token = line.strip().split(maxsplit=1)
token_counts[token] = int(count)
return token_counts
@staticmethod
def _save_token_counts(token_counts, filename):
with open(filename, 'w', encoding='utf-8') as f:
for token, count in token_counts.items():
f.write("%d %s\n" % (count, token))
def _run_max_size_attempt(self, max_size, token_counts):
pool = multiprocessing.Pool(processes=self._vocab_threads)
try:
max_size_candidates = [int(max_size * 2 ** x) for x in range(self._vocab_threads)]
self._logger.info("Vocabulary max_size attempt with candidates = %s" % str(max_size_candidates))
results = pool.map(_build_from_token_counts,
[(token_counts, x, self._count_threshold) for x in max_size_candidates])
for _max_size, vocab_size in results:
if vocab_size <= self._approx_vocab_size:
return _max_size, True
last_max_size, last_vocab_size = results[-1]
self._logger.info("Failed to identify Vocabulary max_size with last_max_size = %d, last_vocab_size = %d" %
(last_max_size, last_vocab_size))
return last_max_size * 2, False
finally:
pool.terminate()
def _build_to_target_size(self, target_size, token_counts, min_val, max_val, num_iterations=4):
"""Builds a SubwordTextEncoder that has `vocab_size` near `target_size`.
Uses simple recursive binary search to find a minimum token count that most
closely matches the `target_size`.
Args:
target_size: Desired vocab_size to approximate.
token_counts: A dictionary of token counts, mapping string to int.
min_val: An integer; lower bound for the minimum token count.
max_val: An integer; upper bound for the minimum token count.
num_iterations: An integer; how many iterations of refinement.
Returns:
A SubwordTextEncoder instance.
Raises:
ValueError: If `min_val` is greater than `max_val`.
"""
if min_val > max_val:
raise ValueError("Lower bound for the minimum token count "
"is greater than the upper bound.")
if target_size < 1:
raise ValueError("Target size must be positive.")
reserved_tokens = RESERVED_TOKENS + self._custom_tokens
def bisect(_min_val, _max_val):
"""Bisection to find the right size."""
present_count = (_max_val + _min_val) // 2
self._logger.info("Trying min_count %d" % present_count)
subtokenizer = SubwordDictionary.build_from_token_counts(
token_counts, present_count, num_iterations,
reserved_tokens=reserved_tokens, count_threshold=self._count_threshold)
# Being within 1% of the target size is ok.
is_ok = abs(len(subtokenizer) - target_size) * 100 < target_size
# If min_val == max_val, we can't do any better than this.
if is_ok or _min_val >= _max_val or present_count < 2:
return subtokenizer
if len(subtokenizer) > target_size:
other_subtokenizer = bisect(present_count + 1, _max_val)
else:
other_subtokenizer = bisect(_min_val, present_count - 1)
if other_subtokenizer is None:
return subtokenizer
if abs(len(other_subtokenizer) - target_size) < abs(len(subtokenizer) - target_size):
return other_subtokenizer
return subtokenizer
return bisect(min_val, max_val)
class SubwordDictionary(Dictionary):
class Factory(_SubwordDictionaryFactory):
pass
@classmethod
def build_from_token_counts(cls, token_counts, min_count, num_iterations=4,
reserved_tokens=None, count_threshold=None):
"""Train a SubwordTextEncoder based on a dictionary of word counts.
Args:
token_counts: a dictionary of Unicode strings to int.
min_count: an integer - discard subtokens with lower counts.
num_iterations: an integer. how many iterations of refinement.
reserved_tokens: List of reserved tokens. The global variable
`RESERVED_TOKENS` must be a prefix of `reserved_tokens`. If this
argument is `None`, it will use `RESERVED_TOKENS`.
count_threshold: if specified, skip all tokens with count < count_threshold
but still uses them for alphabet generation
Raises:
ValueError: if reserved is not 0 or len(RESERVED_TOKENS). In this case, it
is not clear what the space is being reserved for, or when it will be
filled in.
"""
logger = logging.getLogger("SubwordDictionary::build_from_token_counts")
if reserved_tokens is None:
reserved_tokens = RESERVED_TOKENS
else:
# There is not complete freedom in replacing RESERVED_TOKENS.
for default, proposed in zip(RESERVED_TOKENS, reserved_tokens):
if default != proposed:
raise ValueError("RESERVED_TOKENS must be a prefix of "
"reserved_tokens.")
sd = SubwordDictionary()
# Initialize the alphabet. Note, this must include reserved tokens or it can
# result in encoding failures.
alphabet_tokens = chain(token_counts.keys(), reserved_tokens)
sd._init_alphabet_from_tokens(alphabet_tokens)
# Bootstrap the initial list of subtokens with the characters from the
# alphabet plus the escaping characters.
sd._init_subtokens_from_list(list(sd._alphabet) + reserved_tokens)
# We build iteratively. On each iteration, we segment all the words,
# then count the resulting potential subtokens, keeping the ones
# with high enough counts for our new vocabulary.
if min_count < 1:
min_count = 1
for i in range(num_iterations):
logger.info("iteration {0} BEGIN".format(i + 1))
# Collect all substrings of the encoded token that break along current
# subtoken boundaries.
subtoken_counts = collections.defaultdict(int)
for token, count in token_counts.items():
if count_threshold is not None and count < count_threshold:
continue
escaped_token = _escape_token(token, sd._alphabet)
subtokens = sd._subtokens_of_escaped(escaped_token)
start = 0
for subtoken in subtokens:
last_position = len(escaped_token) + 1
for end in range(start + 1, last_position):
new_subtoken = escaped_token[start:end]
subtoken_counts[new_subtoken] += count
start += len(subtoken)
# Array of sets of candidate subtoken strings, by length.
len_to_subtoken_strings = []
for subtoken_string, count in subtoken_counts.items():
lsub = len(subtoken_string)
if count >= min_count:
while len(len_to_subtoken_strings) <= lsub:
len_to_subtoken_strings.append(set())
len_to_subtoken_strings[lsub].add(subtoken_string)
# Consider the candidates longest to shortest, so that if we accept
# a longer subtoken string, we can decrement the counts of its prefixes.
new_subtoken_strings = []
for lsub in range(len(len_to_subtoken_strings) - 1, 0, -1):
subtoken_strings = len_to_subtoken_strings[lsub]
for subtoken_string in subtoken_strings:
count = subtoken_counts[subtoken_string]
if count >= min_count:
# Exclude alphabet tokens here, as they must be included later,
# explicitly, regardless of count.
if subtoken_string not in sd._alphabet:
new_subtoken_strings.append((count, subtoken_string))
for l in range(1, lsub):
subtoken_counts[subtoken_string[:l]] -= count
# Include the alphabet explicitly to guarantee all strings are encodable.
new_subtoken_strings.extend((subtoken_counts.get(a, 0), a) for a in sd._alphabet)
new_subtoken_strings.sort(reverse=True)
# Reinitialize to the candidate vocabulary.
new_subtoken_strings = [subtoken for _, subtoken in new_subtoken_strings]
if reserved_tokens:
new_subtoken_strings = reserved_tokens + new_subtoken_strings
sd._init_subtokens_from_list(new_subtoken_strings)
logger.info("iteration %d END, vocab_size = %d" % (i + 1, len(sd)))
return sd
def __init__(self, subtokens=None):
# super().__init__() - DO NOT CALL
self.unk_word, self.pad_word, self.eos_word = UNK, PAD, EOS
self.symbols = []
self.indices = {}
self.count = None
self.pad_index = RESERVED_TOKENS.index(PAD)
self.eos_index = RESERVED_TOKENS.index(EOS)
self.unk_index = RESERVED_TOKENS.index(UNK)
self.nspecial = len(RESERVED_TOKENS)
self._cache = cachetools.LRUCache(maxsize=2 ** 20)
self._max_subtoken_len = 0
self._alphabet = set()
self._original_size = None
if subtokens is not None and len(subtokens) > 0:
self._init_subtokens_from_list(subtokens)
self._init_alphabet_from_tokens(subtokens)
def _init_subtokens_from_list(self, subtokens):
self.symbols = subtokens
self.indices = {s: i for i, s in enumerate(subtokens) if s}
# we remember the maximum length of any subtoken to avoid having to
# check arbitrarily long strings.
self._max_subtoken_len = max([len(s) for s in subtokens])
def _init_alphabet_from_tokens(self, tokens):
# Include all characters from all tokens in the alphabet to guarantee that
# any token can be encoded. Additionally, include all escaping characters.
self._alphabet = {c for token in tokens for c in token}
self._alphabet |= _ESCAPE_CHARS
@property
def original_size(self):
return self._original_size or len(self)
def force_length(self, new_length):
self._original_size = len(self)
count = new_length - self._original_size
if count < 0:
raise ValueError('new length (%d) must be greater than current length (%d)' % (new_length, len(self)))
for i in range(count):
self.symbols.append('')
def __getitem__(self, idx):
if idx < len(self.symbols):
return self.symbols[idx]
raise ValueError("invalid id %d" % idx)
def index(self, sym):
if sym in self.indices:
return self.indices[sym]
raise ValueError("unknown symbol '%s'" % sym)
def add_symbol(self, word, n=1):
raise NotImplementedError
def update(self, new_dict):
raise NotImplementedError
def finalize(self, threshold=-1, nwords=-1, padding_factor=8):
raise NotImplementedError
@classmethod
def language_tag(cls, lang):
return '[[%s]]' % lang
@classmethod
def size_of(cls, f):
if isinstance(f, str):
with open(f, 'r', encoding='utf-8') as fd:
return cls.size_of(fd)
size = 0
for _ in f:
size += 1
return size
@classmethod
def load(cls, f, ignore_utf_errors=False):
if isinstance(f, str):
try:
if not ignore_utf_errors:
with open(f, 'r', encoding='utf-8') as fd:
return cls.load(fd)
else:
with open(f, 'r', encoding='utf-8', errors='ignore') as fd:
return cls.load(fd)
except FileNotFoundError as e:
raise e
except Exception:
raise Exception("Incorrect encoding detected in {}, please "
"rebuild the dataset".format(f))
def unpack(s):
# Some vocab files wrap words in single quotes, but others don't
if (s.startswith("'") and s.endswith("'")) or (s.startswith("\"") and s.endswith("\"")):
return s[1:-1]
else:
return s
return cls(subtokens=[unpack(line.strip()) for line in f])
def save(self, f):
if isinstance(f, str):
with open(f, 'w', encoding='utf-8') as fd:
return self.save(fd)
for symbol in self.symbols:
print("'{}'".format(symbol), file=f)
def indexes_of(self, subtoken_ids):
indexes = []
i = 0
for j in range(len(subtoken_ids)):
### handle "empty" sub_tokens like UNK
if self[subtoken_ids[j]] == '':
subtoken_ids[j] = UNK_ID
_id = subtoken_ids[j]
if _id == self.eos():
break
elif _id != self.pad():
if j > 0:
if self[_id] == '_':
x = subtoken_ids[j - 1]
if self[x].endswith('_'):
continue
else:
indexes.append(i)
i += 1
elif self[_id].endswith('_'):
indexes.append(i)
i += 1
else:
indexes.append(i)
else:
if self[_id] == '_':
continue
elif self[_id].endswith('_'):
indexes.append(i)
i += 1
else:
indexes.append(i)
return indexes
def string(self, tensor, bpe_symbol=None, escape_unk=False):
if torch.is_tensor(tensor) and tensor.dim() == 2:
return '\n'.join(self.string(t) for t in tensor)
concatenated = "".join(self.tokens(tensor))
ret = []
for t in concatenated.split("_"):
if t:
unescaped = _unescape_token(t + "_")
if unescaped:
ret.append(unescaped)
return ' '.join(ret)
def tokens(self, subtoken_ids):
subtokens = []
for i in subtoken_ids:
if i == self.eos():
break
elif i != self.pad():
subtokens.append(self[i])
return subtokens
def tokenize(self, raw_text):
ret = []
for token in raw_text.strip().split():
ret.extend(self._subtokens_of(token))
return ret
@cachetools.cachedmethod(cache=lambda self: self._cache, key=lambda token: token)
def _subtokens_of(self, token):
return self._subtokens_of_escaped(_escape_token(token, self._alphabet))
def _subtokens_of_escaped(self, escaped_token):
# NOTE: This algorithm is greedy; it won't necessarily produce the "best"
# list of subtokens.
ret = []
start = 0
token_len = len(escaped_token)
while start < token_len:
for end in range(
min(token_len, start + self._max_subtoken_len), start, -1):
subtoken = escaped_token[start:end]
if subtoken in self.indices:
ret.append(subtoken)
start = end
break
else: # Did not break
# If there is no possible encoding of the escaped token then one of the
# characters in the token is not in the alphabet.
return [self.unk_string()]
return ret