-
Notifications
You must be signed in to change notification settings - Fork 13
8000
/
Copy pathtokenizers.py
549 lines (476 loc) · 15.7 KB
/
tokenizers.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
import re
try:
import deepsmiles
_has_deepsmiles = True
except ImportError as err:
_has_deepsmiles = False
DEEPSMILES_ERR = err
try:
import selfies
_has_selfies = True
except ImportError as err:
_has_selfies = False
SELFIES_ERR = err
try:
import smizip
_has_smizip = True
except ImportError as err:
_has_smizip = False
SMIZIP_ERR = err
try:
import atomInSmiles as AIS
_has_AIS = True
except ImportError as err:
_has_AIS = False
AIS_ERR = err
try:
import safe
_has_SAFE = True
except ImportError as err:
_has_SAFE = False
SAFE_ERR = err
class SMILESTokenizerChEMBL:
"""Deals with the tokenization and untokenization of SMILES."""
GRAMMAR = "SMILES"
def __init__(self, start_token="GO", end_token="EOS"):
self.REGEXPS = {
"brackets": re.compile(r"(\[[^\]]*\])"),
"2_ring_nums": re.compile(r"(%\d{2})"),
"brcl": re.compile(r"(Br|Cl)"),
}
self.REGEXP_ORDER = ["brackets", "2_ring_nums", "brcl"]
self.start_token = start_token
self.end_token = end_token
def tokenize(self, data, with_begin_and_end=False):
"""Tokenizes a SMILES string."""
def split_by(data, regexps):
if not regexps:
return list(data)
regexp = self.REGEXPS[regexps[0]]
splitted = regexp.split(data)
tokens = []
for i, split in enumerate(splitted):
if i % 2 == 0:
tokens += split_by(split, regexps[1:])
else:
tokens.append(split)
return tokens
tokens = split_by(data, self.REGEXP_ORDER)
if with_begin_and_end:
tokens = [self.start_token] + tokens + [self.end_token]
return tokens
def untokenize(self, tokens, **kwargs):
"""Untokenizes a SMILES string."""
smi = ""
for token in tokens:
if token == self.end_token:
break
if token != self.start_token:
smi += token
return smi
class SMILESTokenizerGuacaMol:
"""Deals with the tokenization and untokenization of SMILES."""
GRAMMAR = "SMILES"
def __init__(self, start_token="GO", end_token="EOS"):
self.REGEXPS = {
"brackets": re.compile(r"(\[[^\]]*\])"),
"brcl": re.compile(r"(Br|Cl)"),
}
self.REGEXP_ORDER = ["brackets", "brcl"]
self.start_token = start_token
self.end_token = end_token
def tokenize(self, data, with_begin_and_end=False):
"""Tokenizes a SMILES string."""
def split_by(data, regexps):
if not regexps:
return list(data)
regexp = self.REGEXPS[regexps[0]]
splitted = regexp.split(data)
tokens = []
for i, split in enumerate(splitted):
if i % 2 == 0:
tokens += split_by(split, regexps[1:])
else:
tokens.append(split)
return tokens
tokens = split_by(data, self.REGEXP_ORDER)
if with_begin_and_end:
tokens = [self.start_token] + tokens + [self.end_token]
return tokens
def untokenize(self, tokens, **kwargs):
"""Untokenizes a SMILES string."""
smi = ""
for token in tokens:
if token == self.end_token:
break
if token != self.start_token:
smi += token
return smi
class SMILESTokenizerEnamine:
"""Deals with the tokenization and untokenization of SMILES."""
GRAMMER = "SMILES"
def __init__(self, start_token="GO", end_token="EOS"):
smiles_atoms = [
"H",
"He",
"Li",
"Be"
8000
;,
"B",
"C",
"N",
"O",
"F",
"Ne",
"Na",
"Mg",
"Al",
"Si",
"P",
"S",
"Cl",
"Ca",
"Mn",
"Fe",
"Zn",
"Se",
"Br",
"Pd",
"Ag",
"Cd",
"I",
"Hg",
"b",
"c",
"n",
"o",
"p",
"s",
]
special_tokens = [
"<pad>",
"$",
"^",
"#",
"%",
"+",
"[",
"]",
"(",
")",
"-",
".",
"@",
"*",
":",
">",
"/",
"\\",
"0",
"1",
"2",
"3",
"4",
"5",
"6",
"7",
"8",
"9",
"=",
]
sorted_atoms = sorted(smiles_atoms, key=len, reverse=True)
special_tokens = [re.escape(token) for token in special_tokens]
regex_pattern = "|".join(special_tokens + sorted_atoms)
regex_pattern = r"(" + regex_pattern + ")"
smiles_regex = re.compile(regex_pattern)
self.REGEXP = re.compile(smiles_regex)
self.start_token = start_token
self.end_token = end_token
def tokenize(self, data, with_begin_and_end=False):
"""Tokenizes a SMILES string."""
def split_by(data, regexp):
if not regexp:
return list(data)
splitted = regexp.split(data)
tokens = []
for split in splitted:
tokens.append(split)
return list(filter(None, regexp.split(data)))
tokens = split_by(data, self.REGEXP)
if with_begin_and_end:
tokens = [self.start_token] + tokens + [self.end_token]
return tokens
def untokenize(self, tokens):
"""Untokenizes a SMILES string."""
smi = ""
for token in tokens:
if token == self.end_token:
break
if token != self.start_token:
smi += token
return smi
class DeepSMILESTokenizer:
"""Deals with the tokenization and untokenization of SMILES."""
GRAMMAR = "deepSMILES"
def __init__(
self,
start_token="GO",
end_token="EOS",
rings=True,
branches=True,
compress=False,
):
if not _has_deepsmiles:
raise RuntimeError(
"DeepSMILES library not found, please install with pip install deepsmiles."
) from DEEPSMILES_ERR
self.converter = deepsmiles.Converter(rings=rings, branches=branches)
self.run_compression = compress
self.REGEXPS = {
"brackets": re.compile(r"(\[[^\]]*\])"),
"brcl": re.compile(r"(Br|Cl)"),
}
self.REGEXP_ORDER = ["brackets", "brcl"]
self.start_token = start_token
self.end_token = end_token
def tokenize(self, data, with_begin_and_end=False):
"""Tokenizes a SMILES string via conversion to deepSMILES."""
data = self.converter.encode(data)
if self.run_compression:
data = self.compress(data)
def split_by(data, regexps):
if not regexps:
return list(data)
regexp = self.REGEXPS[regexps[0]]
splitted = regexp.split(data)
tokens = []
for i, split in enumerate(splitted):
if i % 2 == 0:
tokens += split_by(split, regexps[1:])
else:
tokens.append(split)
return tokens
tokens = split_by(data, self.REGEXP_ORDER)
if with_begin_and_end:
tokens = [self.start_token] + tokens + [self.end_token]
return tokens
def untokenize(self, tokens, convert_to_smiles=True):
"""Untokenizes a deepSMILES string followed by conversion to SMILES."""
smi = ""
for token in tokens:
if token == "$":
break
if token != "^":
smi += token
if convert_to_smiles:
try:
if self.run_compression:
smi = self.decompress(smi)
smi = self.converter.decode(smi)
except Exception: # deepsmiles.DecodeError doesn't capture IndexError?
smi = None
return smi
def compress(self, dsmi):
"""Compresses a deepSMILES string.
> compress("C)C")
'C)1C'
> compress("C)))C")
'C)3C'
> compress("C))))))))))C")
'C)10C'
"""
compressed = []
N = len(dsmi)
i = 0
while i < N:
x = dsmi[i]
compressed.append(x)
if x == ")":
start = i
while i + 1 < N and dsmi[i + 1] == ")":
i += 1
compressed.append(str(i + 1 - start))
i += 1
return "".join(compressed)
def decompress(self, cdsmi):
"""Decompresses a compressed deepSMILES string.
> decompress("C)1C")
'C)C'
> decompress("C)3C")
'C)))C'
> decompress("C)10C")
'C))))))))))C'
> decompress("C)C")
Traceback (most recent call last):
...
ValueError: A number should follow the parenthesis in C)C
> decompress("C)")
Traceback (most recent call last):
...
ValueError: A number should follow the parenthesis in C)
"""
decompressed = []
N = len(cdsmi)
i = 0
while i < N:
x = cdsmi[i]
if x == ")":
start = i
while i + 1 < N and cdsmi[i + 1].isdigit():
i += 1
if i == start:
raise ValueError(
f"A number should follow the parenthesis in {cdsmi}"
)
number = int(cdsmi[start + 1 : i + 1])
decompressed.append(")" * number)
else:
decompressed.append(x)
i += 1
return "".join(decompressed)
class SELFIESTokenizer:
"""Deals with the tokenization and untokenization of SMILES."""
GRAMMAR = "SELFIES"
def __init__(self, start_token="GO", end_token="EOS"):
if not _has_selfies:
raise RuntimeError(
"SELFIES library not found, please install with pip install selfies ."
) from SELFIES_ERR
self.start_token = start_token
self.end_token = end_token
def tokenize(self, data, with_begin_and_end=False):
"""Tokenizes a SMILES string via conversion to SELFIES."""
data = selfies.encoder(data)
tokens = list(selfies.split_selfies(data))
if with_begin_and_end:
tokens = [self.start_token] + tokens + [self.end_token]
return tokens
def untokenize(self, tokens, convert_to_smiles=True):
"""Untokenizes a SELFIES string followed by conversion to SMILES."""
smi = ""
for token in tokens:
if token == self.start_token:
break
if token != self.end_token:
smi += token
if convert_to_smiles:
try:
smi = selfies.decoder(smi)
except Exception:
smi = None
return smi
class AISTokenizer:
"""Deals with the tokenization and untokenization of SMILES."""
GRAMMAR = "AIS"
def __init__(self, start_token="GO", end_token="EOS"):
if not _has_AIS:
raise RuntimeError(
"atomInSmiles library not found, please install with pip install atomInSmiles."
) from AIS_ERR
self.start_token = start_token
self.end_token = end_token
def tokenize(self, data, with_begin_and_end=False):
"""Tokenizes a SMILES string via conversion to atomInSmiles."""
data = AIS.encode(data)
tokens = data.split(" ")
if with_begin_and_end:
tokens = [self.start_token] + tokens + [self.end_token]
return tokens
def untokenize(self, tokens, convert_to_smiles=True):
"""Untokenizes an atomInSmiles string followed by conversion to SMILES."""
smi = ""
for token in tokens:
if token == self.end_token:
smi = smi.rstrip()
break
if token != self.start_token:
smi += token + " "
if convert_to_smiles:
try:
smi = AIS.decode(smi)
except Exception:
smi = ""
return smi
class SAFETokenizer:
"""Deals with the tokenization and untokenization of SMILES."""
GRAMMAR = "SAFE"
def __init__(self, start_token="GO", end_token="EOS"):
if not _has_SAFE:
raise RuntimeError(
"SAFE library not found, please install with pip install safe-mol."
) from SAFE_ERR
self.start_token = start_token
self.end_token = end_token
def tokenize(self, data, with_begin_and_end=False):
"""Tokenizes a SMILES string via conversion to atomInSmiles."""
data = safe.encode(data)
tokens = safe.split(data)
if with_begin_and_end:
tokens = [self.start_token] + tokens + [self.end_token]
return tokens
def untokenize(self, tokens, convert_to_smiles=True):
"""Untokenizes an atomInSmiles string followed by conversion to SMILES."""
smi = ""
for token in tokens:
if token == self.end_token:
break
if token != self.start_token:
smi += token
if convert_to_smiles:
try:
smi = safe.decode(smi)
except Exception:
smi = None
return smi
class SmiZipTokenizer:
"""Deals with the tokenization and untokenization of SmiZipped SMILES."""
GRAMMAR = "SmiZip"
def __init__(self, ngrams, start_token="GO", end_token="EOS"):
if not _has_smizip:
raise RuntimeError(
"smizip library not found, please install with pip install smizip."
) from SMIZIP_ERR
self.zipper = smizip.SmiZip(ngrams)
self.start_token = start_token
self.end_token = end_token
def tokenize(self, data, with_begin_and_end=False):
"""Tokenizes a SMILES string via conversion to SmiZip tokens."""
tokens = self.zipper.zip(data, format=1) # format=1 returns the tokens
if with_begin_and_end:
tokens = [self.start_token] + tokens + [self.end_token]
return tokens
def untokenize(self, tokens, convert_to_smiles=True):
"""Join the SmiZip tokens to create a SMILES string."""
ntokens = []
for token in tokens:
if token == self.end_token:
break
if token == self.start_token:
continue
ntokens.append(token)
if convert_to_smiles:
smi = "".join(ntokens)
else:
smi = ",".join(ntokens)
return smi
class AsciiSMILESTokenizer:
"""Deals with the tokenization and untokenization of SMILES.
Uses ASCII characters as tokens.
"""
def __init__(self, start_token="^", end_token="$"):
self.start_token = start_token
self.end_token = end_token
def tokenize(self, data, with_begin_and_end=True):
"""Tokenizes a SMILES string."""
tokens = list(data)
if with_begin_and_end:
tokens = [self.start_token] + tokens + [self.end_token]
return tokens
def untokenize(self, tokens):
"""Untokenizes a SMILES string."""
smi = ""
for token in tokens:
if token == self.end_token:
break
if token != self.start_token:
smi += token
return smi