File size: 17,110 Bytes
d08dd00
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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

import csv
import json
import os

from .examples import MultipleChoiceExample, TextExample, TokensExample


class DataProcessor:
    """Base class for data converters for sequence classification data sets."""

    def __init__(self, data_dir):
        self.data_dir = data_dir

    def get_examples(self, lang, mode):
        if mode == 'train':
            return self.get_train_examples(lang)
        elif mode == 'dev':
            return self.get_dev_examples(lang)
        elif mode == 'test':
            return self.get_test_examples(lang)

    def modes(self):
        return ['train', 'dev', 'test']

    def get_train_examples(self, lang):
        """Gets a collection of :class:`InputExample` for the train set."""
        raise NotImplementedError()

    def get_dev_examples(self, lang):
        """Gets a collection of :class:`InputExample` for the dev set."""
        raise NotImplementedError()

    def get_test_examples(self, lang):
        """Gets a collection of :class:`InputExample` for the test set."""
        raise NotImplementedError()

    def get_labels(self, lang):
        """Gets the list of labels for this data set."""
        raise NotImplementedError()

    @classmethod
    def read_csv(cls, input_file, quotechar=None):
        """Reads a tab separated value file."""
        with open(input_file, encoding='utf-8') as fp:
            return list(csv.reader(fp, delimiter=','))

    @classmethod
    def read_json(cls, input_file):
        """Reads a json file file."""
        with open(input_file, encoding='utf-8') as fp:
            return json.load(fp)

    @classmethod
    def readlines(cls, filepath):
        with open(filepath, encoding='utf-8') as fp:
            return fp.readlines()

    @classmethod
    def read_jsonl(cls, filepath):
        with open(filepath, 'r', encoding='utf-8') as fp:
            data = fp.readlines()
            data = list(map(lambda l: json.loads(l), data))
        return data


class IndicNLPHeadlines(DataProcessor):
    """Processor for the Headline Predction dataset"""

    def __init__(self, data_dir):
        self.data_dir = data_dir

    def get_train_examples(self, lang):
        """See base class."""
        fname = '{}/{}-train.json'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_json(fpath), 'train')

    def get_dev_examples(self, lang):
        '''See base class.'''
        fname = '{}/{}-valid.json'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_json(fpath), 'dev')

    def get_test_examples(self, lang):
        '''See base class.'''
        fname = '{}/{}-test.json'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_json(fpath), 'test')

    def get_labels(self, lang):
        """See base class."""
        return ['A', 'B', 'C', 'D']

    def _create_examples(self, items, set_type):
        """Creates examples for the training and dev sets."""
        examples = [
            MultipleChoiceExample(
                example_id=idx,
                question='',
                contexts=[item['content'], item['content'], item['content'],
                          item['content']],
                endings=[item['optionA'], item['optionB'], item['optionC'],
                         item['optionD']],
                label=item['correctOption'],
            )
            for idx, item in enumerate(items)
        ]
        return examples


class WikiCloze(DataProcessor):
    """Processor for Wiki Cloze QA dataset"""

    def __init__(self, data_dir):
        self.data_dir = data_dir

    def modes(self):
        return ['test']

    def get_test_examples(self, lang):
        """See base class."""
        fname = '{}.json'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_json(fpath)['cloze_data'], 'test')

    def get_labels(self, lang):
        """See base class."""
        return list(range(4))

    def _create_examples(self, items, set_type):
        """Creates examples for the training and dev sets."""
        examples = []
        for (i, item) in enumerate(items):
            if '' in [option.strip() for option in item['options']]:
                continue
            example = MultipleChoiceExample(
                example_id=i,
                question=item['question'].replace('<MASK>', '[MASK]'),
                contexts=[],
                endings=item['options'],
                label=item['options'].index(item['answer'])
            )
            examples.append(example)
        return examples


class IndicNLPGenre(DataProcessor):
    """Processor for the Article Genre Classification data set"""

    def __init__(self, data_dir):
        self.data_dir = data_dir

    def get_train_examples(self, lang):
        """See base class."""
        fname = '{}/{}-train.csv'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_csv(fpath), 'train')

    def get_dev_examples(self, lang):
        """See base class."""
        fname = '{}/{}-valid.csv'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_csv(fpath), 'dev')

    def get_test_examples(self, lang):
        fname = '{}/{}-test.csv'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_csv(fpath), 'test')

    def get_labels(self, lang):
        """See base class."""
        filename = '{}/{}-train.csv'.format(lang, lang)
        lines = self.read_csv(os.path.join(self.data_dir, filename))
        labels = map(lambda l: l[0], lines)
        labels = list(set(labels))
        return labels

    def _create_examples(self, lines, set_type):
        """Creates examples for the training and dev sets."""
        examples = []
        for (i, line) in enumerate(lines):
            example = TextExample(
                guid=('%s-%s' % (set_type, i)),
                text_a=line[1],
                label=line[0]
            )
            examples.append(example)
        return examples


class WikiNER(DataProcessor):

    def __init__(self, data_dir):
        self.data_dir = data_dir

    def get_examples(self, lang, mode):
        mode = 'valid' if mode == 'dev' else mode
        file_path = os.path.join(self.data_dir, lang, f'{mode}.txt')
        guid_index = 1
        examples = []
        with open(file_path, encoding='utf-8') as f:
            words = []
            labels = []
            for line in f:
                if line.startswith('-DOCSTART-') or line == '' or line == '\n':
                    if words:
                        example = TokensExample(
                            guid=f'{mode}-{guid_index}',
                            words=words,
                            labels=labels
                        )
                        examples.append(example)
                    guid_index += 1
                    words = []
                    labels = []
                else:
                    splits = line.split(' ')
                    words.append(splits[0])
                    if len(splits) > 1:
                        labels.append(splits[-1].replace('\n', ''))
                    else:
                        # Examples could have no label for mode = 'test'
                        labels.append('O')
            if words:
                example = TokensExample(
                    guid=f'{mode}-{guid_index}',
                    words=words,
                    labels=labels
                )
                examples.append(example)
        return examples

    def get_labels(self, lang):
        path = os.path.join(self.data_dir, lang, 'labels.txt')
        with open(path, 'r') as f:
            labels = f.read().splitlines()
        if 'O' not in labels:
            labels = ['O'] + labels
        return labels


class WikiSectionTitles(DataProcessor):
    """Processor for the Wikipedia Section Title Prediction dataset"""

    def __init__(self, data_dir):
        self.data_dir = data_dir

    def get_train_examples(self, lang):
        """See base class."""
        fname = '{}/{}-train.json'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_json(fpath), 'train')

    def get_dev_examples(self, lang):
        """See base class."""
        fname = '{}/{}-valid.json'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_json(fpath), 'dev')

    def get_test_examples(self, lang):
        """See base class."""
        fname = '{}/{}-test.json'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_json(fpath), 'test')

    def get_labels(self, lang):
        """See base class."""
        return ['titleA', 'titleB', 'titleC', 'titleD']

    def _create_examples(self, items, set_type):
        """Creates examples for the training and dev sets."""
        examples = [
            MultipleChoiceExample(
                example_id=idx,
                question='',
                contexts=[item['sectionText'], item['sectionText'],
                          item['sectionText'], item['sectionText']],
                endings=[item['titleA'], item['titleB'], item['titleC'],
                         item['titleD']],
                label=item['correctTitle'],
            )
            for idx, item in enumerate(items)
        ]
        return examples


class ManKiBaat(DataProcessor):
    """Processor for Man ki Baat dataset"""

    def __init__(self, data_dir):
        self.data_dir = data_dir

    def modes(self):
        return ['en', 'in']

    def get_examples(self, lang, mode):
        if mode == 'en':
            return self.get_examples_en(lang)
        elif mode == 'in':
            return self.get_examples_in(lang)

    def get_examples_en(self, lang):
        """Get examples of English language"""
        fname = 'en-{}/mkb.en'.format(lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.readlines(fpath), 'en')

    def get_examples_in(self, lang):
        """Get examples of the Indian language"""
        fname = 'en-{}/mkb.{}'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.readlines(fpath), 'in')

    def _create_examples(self, lines, set_type):
        """Creates examples for the training and dev sets."""
        examples = []
        for (i, line) in enumerate(lines):
            example = TextExample(
                guid=('%s-%s' % (set_type, i)),
                text_a=line,
                label=i
            )
            examples.append(example)
        return examples

    def get_labels(self, lang):
        # return dummy value greater than number of examples
        return list(range(10000))


class ACTSA(IndicNLPGenre):
    pass


class BBCNews(IndicNLPGenre):

    def get_dev_examples(self, lang):
        """See base class."""
        fname = '{}/{}-test.csv'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_csv(fpath), 'dev')


class INLTKHeadlines(IndicNLPGenre):
    pass


class SohamArticles(IndicNLPGenre):
    pass


class IITPMovies(IndicNLPGenre):
    pass


class IITProducts(IndicNLPGenre):
    pass


class AmritaParaphraseExact(IndicNLPGenre):

    def get_dev_examples(self, lang):
        """See base class."""
        fname = '{}/{}-test.csv'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_csv(fpath), 'dev')

    def get_labels(self, lang):
        """See base class."""
        filename = '{}/{}-train.csv'.format(lang, lang)
        lines = self.read_csv(os.path.join(self.data_dir, filename))
        labels = map(lambda l: l[2], lines)
        labels = list(set(labels))
        return labels

    def _create_examples(self, lines, set_type):
        """Creates examples for the training and dev sets."""
        examples = []
        for (i, line) in enumerate(lines):
            example = TextExample(
                guid=('%s-%s' % (set_type, i)),
                text_a=line[0],
                text_b=line[1],
                label=line[2]
            )
            examples.append(example)
        return examples


class AmritaParaphraseFuzzy(AmritaParaphraseExact):
    pass


class MidasDiscourse(DataProcessor):
    """Processor for the Article Genre Classification data set"""

    def __init__(self, data_dir):
        self.data_dir = data_dir

    def get_train_examples(self, lang):
        """See base class."""
        fname = '{}/train.json'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_json(fpath), 'train')

    def get_dev_examples(self, lang):
        """See base class."""
        fname = '{}/val.json'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_json(fpath), 'dev')

    def get_test_examples(self, lang):
        fname = '{}/test.json'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_json(fpath), 'test')

    def get_labels(self, lang):
        """See base class."""
        filename = '{}/train.json'.format(lang, lang)
        lines = self.read_json(os.path.join(self.data_dir, filename))
        labels = map(lambda l: l['Discourse Mode'], lines)
        labels = list(set(labels))
        return labels

    def _create_examples(self, lines, set_type):
        """Creates examples for the training and dev sets."""
        examples = []
        for (i, line) in enumerate(lines):
            example = TextExample(
                guid=('%s-%s' % (set_type, i)),
                text_a=line['Sentence'],
                label=line['Discourse Mode']
            )
            examples.append(example)
        return examples


class WNLI(DataProcessor):
    """Processor for the WNLI data set (GLUE version)."""

    def __init__(self, data_dir):
        self.data_dir = data_dir

    def get_train_examples(self, lang):
        """See base class."""
        fname = '{}/train.csv'.format(lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_csv(fpath), 'train')

    def get_dev_examples(self, lang):
        """See base class."""
        fname = '{}/dev.csv'.format(lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_csv(fpath), 'dev')

    def get_test_examples(self, lang):
        """See base class."""
        fname = '{}/dev.csv'.format(lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_csv(fpath), 'test')

    def get_labels(self, lang):
        """See base class."""
        return ['0', '1']

    def _create_examples(self, lines, set_type):
        """Creates examples for the training, dev and test sets."""
        examples = []
        for (i, line) in enumerate(lines):
            if i == 0:
                continue
            guid = "%s-%s" % (set_type, line[0])
            text_a = line[1]
            text_b = line[2]
            label = line[-1]
            examples.append(TextExample(guid=guid, text_a=text_a, text_b=text_b, label=label))
        return examples


class COPA(DataProcessor):
    """Processor for the Wikipedia Section Title Prediction dataset"""

    def __init__(self, data_dir):
        self.data_dir = data_dir

    def get_train_examples(self, lang):
        """See base class."""
        fname = '{}/train.jsonl'.format(lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_jsonl(fpath), 'train')

    def get_dev_examples(self, lang):
        """See base class."""
        fname = '{}/val.jsonl'.format(lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_jsonl(fpath), 'dev')

    def get_test_examples(self, lang):
        """See base class."""
        fname = '{}/val.jsonl'.format(lang, lang)
        fpath = os.path.join(self.data_dir, fname)
        return self._create_examples(self.read_jsonl(fpath), 'test')

    def get_labels(self, lang):
        """See base class."""
        return [0, 1]

    def _create_examples(self, items, set_type):
        """Creates examples for the training and dev sets."""
        examples = [
            MultipleChoiceExample(
                example_id=idx,
                question='',
                contexts=[item['premise'], item['premise']],
                endings=[item['choice1'], item['choice2']],
                label=item['label'],
            )
            for idx, item in enumerate(items)
        ]
        return examples