-
Notifications
You must be signed in to change notification settings - Fork 2.5k
/
utils.py
1004 lines (853 loc) · 43.5 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
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
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
from __future__ import absolute_import, division, print_function
import json
import logging
import math
import collections
from io import open
from os import path as osp
from tqdm import tqdm
import bs4
from bs4 import BeautifulSoup as bs
from transformers.models.bert.tokenization_bert import BasicTokenizer, whitespace_tokenize
from torch.utils.data import Dataset
from lxml import etree
from markuplmft.data.tag_utils import tags_dict
logger = logging.getLogger(__name__)
class StrucDataset(Dataset):
"""Dataset wrapping tensors.
Each sample will be retrieved by indexing tensors along the first dimension.
Arguments:
*tensors (*torch.Tensor): tensors that have the same size of the first dimension.
page_ids (list): the corresponding page ids of the input features.
cnn_feature_dir (str): the direction where the cnn features are stored.
token_to_tag (torch.Tensor): the mapping from each token to its corresponding tag id.
"""
def __init__(self, *tensors):
tensors = tuple(tensor for tensor in tensors)
assert all(len(tensors[0]) == len(tensor) for tensor in tensors)
self.tensors = tensors
def __getitem__(self, index):
output = [tensor[index] for tensor in self.tensors]
return tuple(item for item in output)
def __len__(self):
return len(self.tensors[0])
class SRCExample(object):
r"""
The Containers for SRC Examples.
Arguments:
doc_tokens (list[str]): the original tokens of the HTML file before dividing into sub-tokens.
qas_id (str): the id of the corresponding question.
tag_num (int): the total tag number in the corresponding HTML file, including the additional 'yes' and 'no'.
question_text (str): the text of the corresponding question.
orig_answer_text (str): the answer text provided by the dataset.
all_doc_tokens (list[str]): the sub-tokens of the corresponding HTML file.
start_position (int): the position where the answer starts in the all_doc_tokens.
end_position (int): the position where the answer ends in the all_doc_tokens; NOTE that the answer tokens
include the token at end_position.
tok_to_orig_index (list[int]): the mapping from sub-tokens (all_doc_tokens) to origin tokens (doc_tokens).
orig_to_tok_index (list[int]): the mapping from origin tokens (doc_tokens) to sub-tokens (all_doc_tokens).
tok_to_tags_index (list[int]): the mapping from sub-tokens (all_doc_tokens) to the id of the deepest tag it
belongs to.
"""
# the difference between T-PLM and H-PLM is just add <xx> and </xx> into the
# original tokens and further-tokenized tokens
def __init__(self,
doc_tokens,
qas_id,
tag_num, # <xx> ?? </xx> is counted as one tag
question_text=None,
html_code=None,
orig_answer_text=None,
start_position=None, # in all_doc_tokens
end_position=None, # in all_doc_tokens
tok_to_orig_index=None,
orig_to_tok_index=None,
all_doc_tokens=None,
tok_to_tags_index=None,
xpath_tag_map=None,
xpath_subs_map=None,
):
self.doc_tokens = doc_tokens
self.qas_id = qas_id
self.tag_num = tag_num
self.question_text = question_text
self.html_code = html_code
self.orig_answer_text = orig_answer_text
self.start_position = start_position
self.end_position = end_position
self.tok_to_orig_index = tok_to_orig_index
self.orig_to_tok_index = orig_to_tok_index
self.all_doc_tokens = all_doc_tokens
self.tok_to_tags_index = tok_to_tags_index
self.xpath_tag_map = xpath_tag_map
self.xpath_subs_map = xpath_subs_map
def __str__(self):
return self.__repr__()
def __repr__(self):
"""
s = ""
s += "qas_id: %s" % self.qas_id
s += ", question_text: %s" % (
self.question_text)
s += ", doc_tokens: [%s]" % (" ".join(self.doc_tokens))
if self.start_position:
s += ", start_position: %d" % self.start_position
if self.end_position:
s += ", end_position: %d" % self.end_position
"""
s = "[INFO]\n"
s += f"qas_id ({type(self.qas_id)}): {self.qas_id}\n"
s += f"tag_num ({type(self.tag_num)}): {self.tag_num}\n"
s += f"question_text ({type(self.question_text)}): {self.question_text}\n"
s += f"html_code ({type(self.html_code)}): {self.html_code}\n"
s += f"orig_answer_text ({type(self.orig_answer_text)}): {self.orig_answer_text}\n"
s += f"start_position ({type(self.start_position)}): {self.start_position}\n"
s += f"end_position ({type(self.end_position)}): {self.end_position}\n"
s += f"tok_to_orig_index ({type(self.tok_to_orig_index)}): {self.tok_to_orig_index}\n"
s += f"orig_to_tok_index ({type(self.orig_to_tok_index)}): {self.orig_to_tok_index}\n"
s += f"all_doc_tokens ({type(self.all_doc_tokens)}): {self.all_doc_tokens}\n"
s += f"tok_to_tags_index ({type(self.tok_to_tags_index)}): {self.tok_to_tags_index}\n"
s += f"xpath_tag_map ({type(self.xpath_tag_map)}): {self.xpath_tag_map}\n"
s += f"xpath_subs_map ({type(self.xpath_subs_map)}): {self.xpath_subs_map}\n"
s += f"tree_id_map ({type(self.tree_id_map)}): {self.tree_id_map}\n"
return s
class InputFeatures(object):
r"""
The Container for the Features of Input Doc Spans.
Arguments:
unique_id (int): the unique id of the input doc span.
example_index (int): the index of the corresponding SRC Example of the input doc span.
page_id (str): the id of the corresponding web page of the question.
doc_span_index (int): the index of the doc span among all the doc spans which corresponding to the same SRC
Example.
tokens (list[str]): the sub-tokens of the input sequence, including cls token, sep tokens, and the sub-tokens
of the question and HTML file.
token_to_orig_map (dict[int, int]): the mapping from the HTML file's sub-tokens in the sequence tokens (tokens)
to the origin tokens (all_tokens in the corresponding SRC Example).
token_is_max_context (dict[int, bool]): whether the current doc span contains the max pre- and post-context for
each HTML file's sub-tokens.
input_ids (list[int]): the ids of the sub-tokens in the input sequence (tokens).
input_mask (list[int]): use 0/1 to distinguish the input sequence from paddings.
segment_ids (list[int]): use 0/1 to distinguish the question and the HTML files.
paragraph_len (int): the length of the HTML file's sub-tokens.
start_position (int): the position where the answer starts in the input sequence (0 if the answer is not fully
in the input sequence).
end_position (int): the position where the answer ends in the input sequence; NOTE that the answer tokens
include the token at end_position (0 if the answer is not fully in the input sequence).
token_to_tag_index (list[int]): the mapping from sub-tokens of the input sequence to the id of the deepest tag
it belongs to.
is_impossible (bool): whether the answer is fully in the doc span.
"""
def __init__(self,
unique_id,
example_index,
page_id,
doc_span_index,
tokens,
token_to_orig_map,
token_is_max_context,
input_ids,
input_mask,
segment_ids,
paragraph_len,
start_position=None,
end_position=None,
token_to_tag_index=None,
is_impossible=None,
xpath_tags_seq=None,
xpath_subs_seq=None
):
self.unique_id = unique_id
self.example_index = example_index
self.page_id = page_id
self.doc_span_index = doc_span_index
self.tokens = tokens
self.token_to_orig_map = token_to_orig_map
self.token_is_max_context = token_is_max_context
self.input_ids = input_ids
self.input_mask = input_mask
self.segment_ids = segment_ids
self.paragraph_len = paragraph_len
self.start_position = start_position
self.end_position = end_position
self.token_to_tag_index = token_to_tag_index
self.is_impossible = is_impossible
self.xpath_tags_seq = xpath_tags_seq
self.xpath_subs_seq = xpath_subs_seq
def html_escape(html):
r"""
replace the special expressions in the html file for specific punctuation.
"""
html = html.replace('"', '"')
html = html.replace('&', '&')
html = html.replace('<', '<')
html = html.replace('>', '>')
html = html.replace(' ', ' ')
return html
def get_xpath4tokens(html_fn: str, unique_tids: set):
xpath_map = {}
tree = etree.parse(html_fn, etree.HTMLParser())
nodes = tree.xpath('//*')
for node in nodes:
tid = node.attrib.get("tid")
if int(tid) in unique_tids:
xpath_map[int(tid)] = tree.getpath(node)
xpath_map[len(nodes)] = "/html"
xpath_map[len(nodes) + 1] = "/html"
return xpath_map
def get_xpath_and_treeid4tokens(html_code, unique_tids, max_depth):
unknown_tag_id = len(tags_dict)
pad_tag_id = unknown_tag_id + 1
max_width = 1000
width_pad_id = 1001
pad_x_tag_seq = [pad_tag_id] * max_depth
pad_x_subs_seq = [width_pad_id] * max_depth
def xpath_soup(element):
xpath_tags = []
xpath_subscripts = []
tree_index = []
child = element if element.name else element.parent
for parent in child.parents: # type: bs4.element.Tag
siblings = parent.find_all(child.name, recursive=False)
para_siblings = parent.find_all(True, recursive=False)
xpath_tags.append(child.name)
xpath_subscripts.append(
0 if 1 == len(siblings) else next(i for i, s in enumerate(siblings, 1) if s is child))
tree_index.append(next(i for i, s in enumerate(para_siblings, 0) if s is child))
child = parent
xpath_tags.reverse()
xpath_subscripts.reverse()
tree_index.reverse()
return xpath_tags, xpath_subscripts, tree_index
xpath_tag_map = {}
xpath_subs_map = {}
for tid in unique_tids:
element = html_code.find(attrs={'tid': tid})
if element is None:
xpath_tags = pad_x_tag_seq
xpath_subscripts = pad_x_subs_seq
xpath_tag_map[tid] = xpath_tags
xpath_subs_map[tid] = xpath_subscripts
continue
xpath_tags, xpath_subscripts, tree_index = xpath_soup(element)
assert len(xpath_tags) == len(xpath_subscripts)
assert len(xpath_tags) == len(tree_index)
if len(xpath_tags) > max_depth:
xpath_tags = xpath_tags[-max_depth:]
xpath_subscripts = xpath_subscripts[-max_depth:]
xpath_tags = [tags_dict.get(name, unknown_tag_id) for name in xpath_tags]
xpath_subscripts = [min(i, max_width) for i in xpath_subscripts]
# we do not append them to max depth here
xpath_tags += [pad_tag_id] * (max_depth - len(xpath_tags))
xpath_subscripts += [width_pad_id] * (max_depth - len(xpath_subscripts))
xpath_tag_map[tid] = xpath_tags
xpath_subs_map[tid] = xpath_subscripts
return xpath_tag_map, xpath_subs_map
def read_squad_examples(input_file, root_dir, is_training, tokenizer, simplify=False, max_depth=50):
r"""
pre-process the data in json format into SRC Examples.
Arguments:
split_flag:
attention_width:
input_file (str): the inputting data file in json format.
root_dir (str): the root directory of the raw WebSRC dataset, which contains the HTML files.
is_training (bool): True if processing the training set, else False.
tokenizer (Tokenizer): the tokenizer for PLM in use.
method (str): the name of the method in use, choice: ['T-PLM', 'H-PLM', 'V-PLM'].
simplify (bool): when setting to Ture, the returned Example will only contain document tokens, the id of the
question-answers, and the total tag number in the corresponding html files.
Returns:
list[SRCExamples]: the resulting SRC Examples, contained all the needed information for the feature generation
process, except when the argument simplify is setting to True;
set[str]: all the tag names appeared in the processed dataset, e.g. <div>, <img/>, </p>, etc..
"""
with open(input_file, "r", encoding='utf-8') as reader:
input_data = json.load(reader)["data"]
def is_whitespace(c):
if c == " " or c == "\t" or c == "\r" or c == "\n" or ord(c) == 0x202F:
return True
return False
def html_to_text_list(h):
tag_num, text_list = 0, []
for element in h.descendants:
if (type(element) == bs4.element.NavigableString) and (element.strip()):
text_list.append(element.strip())
if type(element) == bs4.element.Tag:
tag_num += 1
return text_list, tag_num + 2 # + 2 because we treat the additional 'yes' and 'no' as two special tags.
def e_id_to_t_id(e_id, html):
t_id = 0
for element in html.descendants:
if type(element) == bs4.element.NavigableString and element.strip():
t_id += 1
if type(element) == bs4.element.Tag:
if int(element.attrs['tid']) == e_id:
break
return t_id
def calc_num_from_raw_text_list(t_id, l):
n_char = 0
for i in range(t_id):
n_char += len(l[i]) + 1
return n_char
def word_tag_offset(html):
cnt, w_t, t_w, tags, tags_tids = 0, [], [], [], []
for element in html.descendants:
if type(element) == bs4.element.Tag:
content = ' '.join(list(element.strings)).split()
t_w.append({'start': cnt, 'len': len(content)})
tags.append('<' + element.name + '>')
tags_tids.append(element['tid'])
elif type(element) == bs4.element.NavigableString and element.strip():
text = element.split()
tid = element.parent['tid']
ind = tags_tids.index(tid)
for _ in text:
w_t.append(ind)
cnt += 1
assert cnt == len(w_t)
w_t.append(len(t_w))
w_t.append(len(t_w) + 1)
return w_t
def subtoken_tag_offset(html, s_tok):
w_t = word_tag_offset(html)
s_t = []
unique_tids = set()
for i in range(len(s_tok)):
s_t.append(w_t[s_tok[i]])
unique_tids.add(w_t[s_tok[i]])
return s_t, unique_tids
examples = []
all_tag_list = set()
total_num = sum([len(entry["websites"]) for entry in input_data])
with tqdm(total=total_num, desc="Converting websites to examples") as t:
for entry in input_data:
domain = entry["domain"]
for website in entry["websites"]:
# Generate Doc Tokens
page_id = website["page_id"]
curr_dir = osp.join(root_dir, domain, page_id[0:2], 'processed_data')
html_fn = osp.join(curr_dir, page_id + '.html')
html_file = open(html_fn).read()
html_code = bs(html_file, "html.parser")
raw_text_list, tag_num = html_to_text_list(html_code) # 字符列表及标签数
doc_tokens = []
char_to_word_offset = []
page_text = ' '.join(raw_text_list)
prev_is_whitespace = True
for c in page_text:
if is_whitespace(c):
prev_is_whitespace = True
else:
if prev_is_whitespace:
doc_tokens.append(c)
else:
doc_tokens[-1] += c
prev_is_whitespace = False
char_to_word_offset.append(len(doc_tokens) - 1)
doc_tokens.append('no')
char_to_word_offset.append(len(doc_tokens) - 1)
doc_tokens.append('yes')
char_to_word_offset.append(len(doc_tokens) - 1)
tag_list = []
assert len(doc_tokens) == char_to_word_offset[-1] + 1, (len(doc_tokens), char_to_word_offset[-1])
if simplify:
for qa in website["qas"]:
qas_id = qa["id"]
example = SRCExample(doc_tokens=doc_tokens, qas_id=qas_id, tag_num=tag_num)
examples.append(example)
t.update(1)
else:
# Tokenize all doc tokens
# tokenize sth like < / >
tok_to_orig_index = []
orig_to_tok_index = []
all_doc_tokens = []
for (i, token) in enumerate(doc_tokens):
orig_to_tok_index.append(len(all_doc_tokens))
if token in tag_list:
sub_tokens = [token]
else:
sub_tokens = tokenizer.tokenize(token)
for sub_token in sub_tokens:
tok_to_orig_index.append(i)
all_doc_tokens.append(sub_token)
# Generate extra information for features
tok_to_tags_index, unique_tids = subtoken_tag_offset(html_code, tok_to_orig_index)
xpath_tag_map, xpath_subs_map = get_xpath_and_treeid4tokens(html_code,
unique_tids,
max_depth=max_depth)
assert tok_to_tags_index[-1] == tag_num - 1, (tok_to_tags_index[-1], tag_num - 1)
# Process each qas, which is mainly calculate the answer position
for qa in website["qas"]:
qas_id = qa["id"]
question_text = qa["question"]
start_position = None
end_position = None
orig_answer_text = None
if is_training:
if len(qa["answers"]) != 1:
raise ValueError(
"For training, each question should have exactly 1 answer.")
answer = qa["answers"][0]
orig_answer_text = answer["text"]
if answer["element_id"] == -1:
num_char = len(char_to_word_offset) - 2
else:
num_char = calc_num_from_raw_text_list(e_id_to_t_id(answer["element_id"], html_code),
raw_text_list)
answer_offset = num_char + answer["answer_start"]
answer_length = len(orig_answer_text) if answer["element_id"] != -1 else 1
start_position = char_to_word_offset[answer_offset]
end_position = char_to_word_offset[answer_offset + answer_length - 1]
# Only add answers where the text can be exactly recovered from the
# document. If this CAN'T happen it's likely due to weird Unicode
# stuff so we will just skip the example.
#
# Note that this means for training mode, every example is NOT
# guaranteed to be preserved.
actual_text = " ".join([w for w in doc_tokens[start_position:(end_position + 1)]
if (w[0] != '<' or w[-1] != '>')])
cleaned_answer_text = " ".join(whitespace_tokenize(orig_answer_text))
if actual_text.find(cleaned_answer_text) == -1:
logger.warning("Could not find answer of question %s: '%s' vs. '%s'",
qa['id'], actual_text, cleaned_answer_text)
continue
example = SRCExample(
doc_tokens=doc_tokens,
qas_id=qas_id,
tag_num=tag_num,
question_text=question_text,
html_code=html_code,
orig_answer_text=orig_answer_text,
start_position=start_position,
end_position=end_position,
tok_to_orig_index=tok_to_orig_index,
orig_to_tok_index=orig_to_tok_index,
all_doc_tokens=all_doc_tokens,
tok_to_tags_index=tok_to_tags_index,
xpath_tag_map=xpath_tag_map,
xpath_subs_map=xpath_subs_map,
)
examples.append(example)
t.update(1)
return examples, all_tag_list
def convert_examples_to_features(examples, tokenizer, max_seq_length, doc_stride, max_query_length, is_training,
cls_token='[CLS]', sep_token='[SEP]', pad_token=0,
sequence_a_segment_id=0, sequence_b_segment_id=1,
cls_token_segment_id=0, pad_token_segment_id=0,
mask_padding_with_zero=True, max_depth=50):
r"""
Converting the SRC Examples further into the features for all the input doc spans.
Arguments:
examples (list[SRCExample]): the list of SRC Examples to process.
tokenizer (Tokenizer): the tokenizer for PLM in use.
max_seq_length (int): the max length of the total sub-token sequence, including the question, cls token, sep
tokens, and documents; if the length of the input is bigger than max_seq_length, the input
will be cut into several doc spans.
doc_stride (int): the stride length when the input is cut into several doc spans.
max_query_length (int): the max length of the sub-token sequence of the questions; the question will be truncate
if it is longer than max_query_length.
is_training (bool): True if processing the training set, else False.
cls_token (str): the cls token in use, default is '[CLS]'.
sep_token (str): the sep token in use, default is '[SEP]'.
pad_token (int): the id of the padding token in use when the total sub-token length is smaller that
max_seq_length, default is 0 which corresponding to the '[PAD]' token.
sequence_a_segment_id: the segment id for the first sequence (the question), default is 0.
sequence_b_segment_id: the segment id for the second sequence (the html file), default is 1.
cls_token_segment_id: the segment id for the cls token, default is 0.
pad_token_segment_id: the segment id for the padding tokens, default is 0.
mask_padding_with_zero: determine the pattern of the returned input mask; 0 for padding tokens and 1 for others
when True, and vice versa.
Returns:
list[InputFeatures]: the resulting input features for all the input doc spans
"""
pad_x_tag_seq = [216] * max_depth
pad_x_subs_seq = [1001] * max_depth
unique_id = 1000000000
features = []
for (example_index, example) in enumerate(tqdm(examples, desc="Converting examples to features")):
xpath_tag_map = example.xpath_tag_map
xpath_subs_map = example.xpath_subs_map
query_tokens = tokenizer.tokenize(example.question_text)
if len(query_tokens) > max_query_length:
query_tokens = query_tokens[0:max_query_length]
tok_start_position = None
tok_end_position = None
if is_training:
tok_start_position = example.orig_to_tok_index[example.start_position]
if example.end_position < len(example.doc_tokens) - 1:
tok_end_position = example.orig_to_tok_index[example.end_position + 1] - 1
else:
tok_end_position = len(example.all_doc_tokens) - 1
(tok_start_position, tok_end_position) = _improve_answer_span(
example.all_doc_tokens, tok_start_position, tok_end_position, tokenizer,
example.orig_answer_text)
# The -3 accounts for [CLS], [SEP] and [SEP]
max_tokens_for_doc = max_seq_length - len(query_tokens) - 3
# We can have documents that are longer than the maximum sequence length.
# To deal with this we do a sliding window approach, where we take chunks
# of the up to our max length with a stride of `doc_stride`.
_DocSpan = collections.namedtuple( # pylint: disable=invalid-name
"DocSpan", ["start", "length"])
doc_spans = []
start_offset = 0
while start_offset < len(example.all_doc_tokens):
length = len(example.all_doc_tokens) - start_offset
if length > max_tokens_for_doc:
length = max_tokens_for_doc
doc_spans.append(_DocSpan(start=start_offset, length=length))
if start_offset + length == len(example.all_doc_tokens):
break
start_offset += min(length, doc_stride)
for (doc_span_index, doc_span) in enumerate(doc_spans):
tokens = []
token_to_orig_map = {}
token_is_max_context = {}
segment_ids = []
token_to_tag_index = []
# CLS token at the beginning
tokens.append(cls_token)
segment_ids.append(cls_token_segment_id)
token_to_tag_index.append(example.tag_num)
# Query
tokens += query_tokens
segment_ids += [sequence_a_segment_id] * len(query_tokens)
token_to_tag_index += [example.tag_num] * len(query_tokens)
# SEP token
tokens.append(sep_token)
segment_ids.append(sequence_a_segment_id)
token_to_tag_index.append(example.tag_num)
# Paragraph
for i in range(doc_span.length):
split_token_index = doc_span.start + i
token_to_orig_map[len(tokens)] = example.tok_to_orig_index[split_token_index]
token_to_tag_index.append(example.tok_to_tags_index[split_token_index])
is_max_context = _check_is_max_context(doc_spans, doc_span_index,
split_token_index)
token_is_max_context[len(tokens)] = is_max_context
tokens.append(example.all_doc_tokens[split_token_index])
segment_ids.append(sequence_b_segment_id)
paragraph_len = doc_span.length
# SEP token
tokens.append(sep_token)
segment_ids.append(sequence_b_segment_id)
token_to_tag_index.append(example.tag_num)
input_ids = tokenizer.convert_tokens_to_ids(tokens)
# The mask has 1 for real tokens and 0 for padding tokens. Only real
# tokens are attended to.
input_mask = [1 if mask_padding_with_zero else 0] * len(input_ids)
# Zero-pad up to the sequence length.
while len(input_ids) < max_seq_length:
input_ids.append(pad_token)
input_mask.append(0 if mask_padding_with_zero else 1)
segment_ids.append(pad_token_segment_id)
token_to_tag_index.append(example.tag_num)
assert len(input_ids) == max_seq_length
assert len(input_mask) == max_seq_length
assert len(segment_ids) == max_seq_length
assert len(token_to_tag_index) == max_seq_length
span_is_impossible = False
start_position = None
end_position = None
if is_training:
# For training, if our document chunk does not contain an annotation
# we throw it out, since there is nothing to predict.
doc_start = doc_span.start
doc_end = doc_span.start + doc_span.length - 1
out_of_span = False
if not (tok_start_position >= doc_start and
tok_end_position <= doc_end):
out_of_span = True
if out_of_span:
span_is_impossible = True
start_position = 0
end_position = 0
else:
doc_offset = len(query_tokens) + 2
start_position = tok_start_position - doc_start + doc_offset
end_position = tok_end_position - doc_start + doc_offset
xpath_tags_seq = [xpath_tag_map.get(tid, pad_x_tag_seq) for tid in token_to_tag_index] # ok
xpath_subs_seq = [xpath_subs_map.get(tid, pad_x_subs_seq) for tid in token_to_tag_index] # ok
# we need to get extended_attention_mask
features.append(
InputFeatures(
unique_id=unique_id,
example_index=example_index,
page_id=example.qas_id[:-5],
doc_span_index=doc_span_index,
tokens=tokens,
token_to_orig_map=token_to_orig_map,
token_is_max_context=token_is_max_context,
input_ids=input_ids,
input_mask=input_mask,
segment_ids=segment_ids,
paragraph_len=paragraph_len,
start_position=start_position,
end_position=end_position,
token_to_tag_index=token_to_tag_index,
is_impossible=span_is_impossible,
xpath_tags_seq=xpath_tags_seq,
xpath_subs_seq=xpath_subs_seq,
))
unique_id += 1
return features
# ---------- copied ! --------------
def _improve_answer_span(doc_tokens, input_start, input_end, tokenizer, orig_answer_text):
tok_answer_text = " ".join(tokenizer.tokenize(orig_answer_text))
for new_start in range(input_start, input_end + 1):
for new_end in range(input_end, new_start - 1, -1):
text_span = " ".join([w for w in doc_tokens[new_start:(new_end + 1)]
if w[0] != '<' or w[-1] != '>'])
if text_span == tok_answer_text:
return new_start, new_end
return input_start, input_end
# ---------- copied ! --------------
def _check_is_max_context(doc_spans, cur_span_index, position):
best_score = None
best_span_index = None
for (span_index, doc_span) in enumerate(doc_spans):
end = doc_span.start + doc_span.length - 1
if position < doc_span.start:
continue
if position > end:
continue
num_left_context = position - doc_span.start
num_right_context = end - position
score = min(num_left_context, num_right_context) + 0.01 * doc_span.length
if best_score is None or score > best_score:
best_score = score
best_span_index = span_index
return cur_span_index == best_span_index
RawResult = collections.namedtuple("RawResult",
["unique_id", "start_logits", "end_logits"])
def write_predictions(all_examples, all_features, all_results, n_best_size, max_answer_length, do_lower_case,
output_prediction_file, output_tag_prediction_file,
output_nbest_file, verbose_logging, tokenizer):
r"""
Compute and write down the final results, including the n best results.
Arguments:
all_examples (list[SRCExample]): all the SRC Example of the dataset; note that we only need it to provide the
mapping from example index to the question-answers id.
all_features (list[InputFeatures]): all the features for the input doc spans.
all_results (list[RawResult]): all the results from the models.
n_best_size (int): the number of the n best buffer and the final n best result saved.
max_answer_length (int): constrain the model to predict the answer no longer than it.
do_lower_case (bool): whether the model distinguish upper and lower case of the letters.
output_prediction_file (str): the file which the best answer text predictions will be written to.
output_tag_prediction_file (str): the file which the best answer tag predictions will be written to.
output_nbest_file (str): the file which the n best answer predictions including text, tag, and probabilities
will be written to.
verbose_logging (bool): if true, all of the warnings related to data processing will be printed.
"""
logger.info("Writing predictions to: %s" % output_prediction_file)
logger.info("Writing nbest to: %s" % output_nbest_file)
example_index_to_features = collections.defaultdict(list)
for feature in all_features:
example_index_to_features[feature.example_index].append(feature)
unique_id_to_result = {}
for result in all_results:
unique_id_to_result[result.unique_id] = result
_PrelimPrediction = collections.namedtuple( # pylint: disable=invalid-name
"PrelimPrediction",
["feature_index", "start_index", "end_index", "start_logit", "end_logit", "tag_ids"])
all_predictions = collections.OrderedDict()
all_tag_predictions = collections.OrderedDict()
all_nbest_json = collections.OrderedDict()
for (example_index, example) in enumerate(all_examples):
features = example_index_to_features[example_index]
prelim_predictions = []
for (feature_index, feature) in enumerate(features):
result = unique_id_to_result[feature.unique_id]
start_indexes = _get_best_indexes(result.start_logits, n_best_size)
end_indexes = _get_best_indexes(result.end_logits, n_best_size)
# if we could have irrelevant answers, get the min score of irrelevant
for start_index in start_indexes:
for end_index in end_indexes:
# We could hypothetically create invalid predictions, e.g., predict
# that the start of the span is in the question. We throw out all
# invalid predictions.
if start_index >= len(feature.tokens):
continue
if end_index >= len(feature.tokens):
continue
if start_index not in feature.token_to_orig_map:
continue
if end_index not in feature.token_to_orig_map:
continue
if not feature.token_is_max_context.get(start_index, False):
continue
if end_index < start_index:
continue
length = end_index - start_index + 1
if length > max_answer_length:
continue
tag_ids = set(feature.token_to_tag_index[start_index: end_index + 1])
prelim_predictions.append(
_PrelimPrediction(
feature_index=feature_index,
start_index=start_index,
end_index=end_index,
start_logit=result.start_logits[start_index],
end_logit=result.end_logits[end_index],
tag_ids=list(tag_ids)))
prelim_predictions = sorted(
prelim_predictions,
key=lambda x: (x.start_logit + x.end_logit),
reverse=True)
_NbestPrediction = collections.namedtuple( # pylint: disable=invalid-name
"NbestPrediction", ["text", "start_logit", "end_logit", "tag_ids"])
seen_predictions = {}
nbest = []
for pred in prelim_predictions:
if len(nbest) >= n_best_size:
break
feature = features[pred.feature_index]
if pred.start_index > 0: # this is a non-null prediction
tok_tokens = feature.tokens[pred.start_index:(pred.end_index + 1)]
orig_doc_start = feature.token_to_orig_map[pred.start_index]
orig_doc_end = feature.token_to_orig_map[pred.end_index]
orig_tokens = example.doc_tokens[orig_doc_start:(orig_doc_end + 1)]
tok_text = " ".join(tok_tokens)
# De-tokenize WordPieces that have been split off.
tok_text = tok_text.replace(" ##", "")
tok_text = tok_text.replace("##", "")
# Clean whitespace
tok_text = tok_text.strip()
tok_text = " ".join(tok_text.split())
orig_text = " ".join(orig_tokens)
final_text = _get_final_text(tok_text, orig_text, do_lower_case, verbose_logging)
if final_text in seen_predictions:
continue
seen_predictions[final_text] = True
else:
final_text = ""
seen_predictions[final_text] = True
nbest.append(
_NbestPrediction(
text=final_text,
start_logit=pred.start_logit,
end_logit=pred.end_logit,
tag_ids=pred.tag_ids))
# In very rare edge cases we could have no valid predictions. So we
# just create a nonce prediction in this case to avoid failure.
if not nbest:
nbest.append(
_NbestPrediction(text="empty", start_logit=0.0, end_logit=0.0, tag_ids=[-1]))
assert len(nbest) >= 1
total_scores = []
best_non_null_entry = None
for entry in nbest:
total_scores.append(entry.start_logit + entry.end_logit)
if not best_non_null_entry:
if entry.text:
best_non_null_entry = entry
probs = _compute_softmax(total_scores)
nbest_json = []
for (i, entry) in enumerate(nbest):
output = collections.OrderedDict()
output["text"] = entry.text
output["probability"] = probs[i]
output["start_logit"] = entry.start_logit
output["end_logit"] = entry.end_logit
output["tag_ids"] = entry.tag_ids
nbest_json.append(output)
assert len(nbest_json) >= 1
best = nbest_json[0]["text"].split()
best = ' '.join([w for w in best
if (w[0] != '<' or w[-1] != '>')
and w != "<end-of-node>"
and w != tokenizer.sep_token
and w != tokenizer.cls_token])
all_predictions[example.qas_id] = best
all_tag_predictions[example.qas_id] = nbest_json[0]["tag_ids"]
all_nbest_json[example.qas_id] = nbest_json
with open(output_prediction_file, "w") as writer:
writer.write(json.dumps(all_predictions, indent=4) + "\n")
with open(output_nbest_file, "w") as writer:
writer.write(json.dumps(all_nbest_json, indent=4) + "\n")
with open(output_tag_prediction_file, 'w') as writer:
writer.write(json.dumps(all_tag_predictions, indent=4) + '\n')
return
def _get_final_text(pred_text, orig_text, do_lower_case, verbose_logging=False):
def _strip_spaces(text):
ns_chars = []
ns_to_s_map = collections.OrderedDict()
for (i, c) in enumerate(text):
if c == " ":
continue
ns_to_s_map[len(ns_chars)] = i
ns_chars.append(c)
ns_text = "".join(ns_chars)
return ns_text, ns_to_s_map
# We first tokenize `orig_text`, strip whitespace from the result
# and `pred_text`, and check if they are the same length. If they are
# NOT the same length, the heuristic has failed. If they are the same
# length, we assume the characters are one-to-one aligned.
tokenizer = BasicTokenizer(do_lower_case=do_lower_case)
tok_text = " ".join(tokenizer.tokenize(orig_text))
start_position = tok_text.find(pred_text)
if start_position == -1:
if verbose_logging:
logger.info(
"Unable to find text: '%s' in '%s'" % (pred_text, orig_text))
return orig_text
end_position = start_position + len(pred_text) - 1
(orig_ns_text, orig_ns_to_s_map) = _strip_spaces(orig_text)
(tok_ns_text, tok_ns_to_s_map) = _strip_spaces(tok_text)
if len(orig_ns_text) != len(tok_ns_text):
if verbose_logging:
logger.info("Length not equal after stripping spaces: '%s' vs '%s'",
orig_ns_text, tok_ns_text)
return orig_text
# We then project the characters in `pred_text` back to `orig_text` using
# the character-to-character alignment.
tok_s_to_ns_map = {}
for (i, tok_index) in tok_ns_to_s_map.items():
tok_s_to_ns_map[tok_index] = i
orig_start_position = None
if start_position in tok_s_to_ns_map:
ns_start_position = tok_s_to_ns_map[start_position]
if ns_start_position in orig_ns_to_s_map:
orig_start_position = orig_ns_to_s_map[ns_start_position]
if orig_start_position is None:
if verbose_logging:
logger.info("Couldn't map start position")
return orig_text
orig_end_position = None
if end_position in tok_s_to_ns_map:
ns_end_position = tok_s_to_ns_map[end_position]
if ns_end_position in orig_ns_to_s_map:
orig_end_position = orig_ns_to_s_map[ns_end_position]
if orig_end_position is None:
if verbose_logging:
logger.info("Couldn't map end position")
return orig_text
output_text = orig_text[orig_start_position:(orig_end_position + 1)]
return output_text
def _get_best_indexes(logits, n_best_size):
"""Get the n-best logits from a list."""
index_and_score = sorted(enumerate(logits), key=lambda x: x[1], reverse=True)
best_indexes = []
for i in range(len(index_and_score)):
if i >= n_best_size:
break
best_indexes.append(index_and_score[i][0])
return best_indexes
def _compute_softmax(scores):
"""Compute softmax probability over raw logits."""
if not scores:
return []
max_score = None
for score in scores:
if max_score is None or score > max_score:
max_score = score
exp_scores = []
total_sum = 0.0
for score in scores:
x = math.exp(score - max_score)
exp_scores.append(x)
total_sum += x