-
Notifications
You must be signed in to change notification settings - Fork 0
/
create_dataset_galician.py
491 lines (397 loc) · 20.6 KB
/
create_dataset_galician.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
# imports
import nltk
nltk.download('wordnet')
nltk.download('omw-1.4')
nltk.download('propbank')
nltk.download('treebank')
from nltk.corpus import propbank
from nltk.corpus import wordnet as wn
from datasets import Dataset, ClassLabel, Sequence, Features, Value, load_from_disk
from collections import defaultdict
import pandas as pd
# functions
def ctg_to_conll_dict(file, conll_dict=None, gal_to_propbank_dict=None):
if conll_dict==None and gal_to_propbank_dict==None:
conll_dict = defaultdict(dict)
gal_to_propbank_dict = defaultdict(dict)
elif conll_dict==None or gal_to_propbank_dict==None:
return 'Error! Must provide either both or neither conll_dict and gal_to_propbank_dict.'
else:
conll_dict = conll_dict
gal_to_propbank_dict = gal_to_propbank_dict
with open(file, 'r') as f:
lines = f.readlines()
sent_id = None
for line in lines:
if 'sent_id' in line:
if sent_id != None:
conll_dict[sent_id]['sent'] = sent
conll_dict[sent_id]['sent_data'] = sent_data_dict
line = line.strip().split(' ')
line = line[-1].split('-')
sent_id = int(line[-1])
sent_data_dict = defaultdict(dict)
sent = None
word_index = None
word_form = None
lemma = None
upos = None
xpos = None
feats = None
head = None
deprel = None
deps = None
misc = None
elif '#' in line:
line = line.strip().split('=')
sent = line[-1]
else:
line = line.split('\t')
if len(line) < 2:
continue
else:
word_index = line[0]
word_form = line[1]
lemma = line[2]
upos = line[3]
xpos = line[4]
feats = line[5]
head = line[6]
deprel = line[7]
deps = line[8]
misc = line[9].strip()
sent_data_dict[word_index]['word_form'] = word_form
sent_data_dict[word_index]['lemma'] = lemma
sent_data_dict[word_index]['upos'] = upos
sent_data_dict[word_index]['xpos'] = xpos
sent_data_dict[word_index]['feats'] = feats
sent_data_dict[word_index]['head'] = head
sent_data_dict[word_index]['deprel'] = deprel
sent_data_dict[word_index]['deps'] = deps
sent_data_dict[word_index]['misc'] = misc
sent_data_dict[word_index]['role'] = '_'
if upos == 'VERB':
lemma_synsets = wn.synsets(lemma, lang='glg')
rolesets = []
arg_sets = []
for synset in lemma_synsets:
split_synset = str(synset).split("'")
split_split = split_synset[1].split('.')
syn = split_split[0] + '.' + split_split[-1] #'garner.01'
try:
roleset = propbank.roleset(syn)
rolesets.append(syn)
args = dict()
for role in roleset.findall('roles/role'):
arg_num = 'A' + role.attrib['n']
args[arg_num] = role.attrib['descr']
arg_sets.append(args)
except:
continue # no match
gal_to_propbank_dict[lemma]['rolesets'] = rolesets
gal_to_propbank_dict[lemma]['arg_sets'] = arg_sets
if len(rolesets) == 0:
sent_data_dict[word_index]['role'] = 'undefined.01'
sent_data_dict[word_index]['args'] = {'A0' : 'describer',
'A1' : 'thing defined',
'A2' : 'attribute'}
elif len(rolesets) == 1:
sent_data_dict[word_index]['role'] = rolesets[0]
sent_data_dict[word_index]['args'] = arg_sets[0]
else:
sent_data_dict[word_index]['role'] = 'see gal_to_propbank_dict'
sent_data_dict[word_index]['args'] = {'A0' : 'describer',
'A1' : 'thing defined',
'A2' : 'attribute'}
f.close()
return conll_dict, gal_to_propbank_dict
def ctg_conll_add_args(conll_dictionary):
for sent_id in conll_dictionary.keys(): # each sent
sent = conll_dictionary[sent_id]['sent_data']
verbs = []
for word_num in sent:
base = sent[word_num]
if base['upos'] == 'VERB':
verbs.append(word_num)
for word_num in sent:
base = sent[word_num]
head = base['head']
xpos = base['xpos']
deprel = base['deprel']
if word_num in verbs:
verb_number = verbs.index(word_num)
base['arg'] = 'r' + str(verb_number) + ':root'
else:
if head in verbs:
verb_number = verbs.index(head)
possible_args = sent[head]['args'].keys()
if deprel == 'obl':
if xpos == 'NCFP000' and 'A0' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg0'
elif xpos == 'NCMP000' and 'A1' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg1'
elif xpos == 'NCFS000' or xpos == 'NCMS000' and 'A2' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg2'
elif xpos == 'NCMS000' and 'A0' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg0'
else:
base['arg'] = '_'
elif deprel == 'nsubj':
if 'A0' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg0'
else:
base['arg'] = '_'
elif deprel == 'obj':
if 'A1' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg1'
else:
base['arg'] = '_'
elif deprel == 'iobj':
if 'A2' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg2'
else:
base['arg'] = '_'
else:
base['arg'] = '_'
else:
base['arg'] = '_'
def treegal_to_conll_dict(file, conll_dict=None, gal_to_propbank_dict=None):
if conll_dict==None and gal_to_propbank_dict==None:
conll_dict = defaultdict(dict)
gal_to_propbank_dict = defaultdict(dict)
elif conll_dict==None or gal_to_propbank_dict==None:
return 'Error! Must provide either both or neither conll_dict and gal_to_propbank_dict.'
else:
conll_dict = conll_dict
gal_to_propbank_dict = gal_to_propbank_dict
with open(file, 'r') as f:
lines = f.readlines()
sent_id = None
for line in lines:
if 'sent_id' in line:
if sent_id != None:
conll_dict[sent_id]['sent'] = sent
conll_dict[sent_id]['sent_data'] = sent_data_dict
line = line.strip().split(' ')
sent_id = int(line[-1])
sent_data_dict = defaultdict(dict)
sent = None
word_index = None
word_form = None
lemma = None
upos = None
xpos = None
feats = None
head = None
deprel = None
deps = None
misc = None
elif '#' in line:
line = line.strip().split('=')
sent = line[-1]
else:
line = line.split('\t')
if len(line) < 2:
continue
else:
word_index = line[0]
word_form = line[1]
lemma = line[2]
upos = line[3]
xpos = line[4]
feats = line[5]
head = line[6]
deprel = line[7]
deps = line[8]
misc = line[9].strip()
sent_data_dict[word_index]['word_form'] = word_form
sent_data_dict[word_index]['lemma'] = lemma
sent_data_dict[word_index]['upos'] = upos
sent_data_dict[word_index]['xpos'] = xpos
sent_data_dict[word_index]['feats'] = feats
sent_data_dict[word_index]['head'] = head
sent_data_dict[word_index]['deprel'] = deprel
sent_data_dict[word_index]['deps'] = deps
sent_data_dict[word_index]['misc'] = misc
sent_data_dict[word_index]['role'] = '_'
if upos == 'VERB':
# print('VERB')
lemma_synsets = wn.synsets(lemma, lang='glg')
rolesets = []
arg_sets = []
for synset in lemma_synsets:
split_synset = str(synset).split("'")
split_split = split_synset[1].split('.')
syn = split_split[0] + '.' + split_split[-1] #'garner.01'
try:
roleset = propbank.roleset(syn)
rolesets.append(syn)
args = dict()
for role in roleset.findall('roles/role'):
arg_num = 'A' + role.attrib['n']
args[arg_num] = role.attrib['descr']
arg_sets.append(args)
except:
continue # no match
gal_to_propbank_dict[lemma]['rolesets'] = rolesets
gal_to_propbank_dict[lemma]['arg_sets'] = arg_sets
if len(rolesets) == 0:
sent_data_dict[word_index]['role'] = 'undefined.01'
sent_data_dict[word_index]['args'] = {'A0' : 'describer',
'A1' : 'thing defined',
'A2' : 'attribute'}
elif len(rolesets) == 1:
sent_data_dict[word_index]['role'] = rolesets[0]
sent_data_dict[word_index]['args'] = arg_sets[0]
else:
sent_data_dict[word_index]['role'] = 'see gal_to_propbank_dict'
sent_data_dict[word_index]['args'] = {'A0' : 'describer',
'A1' : 'thing defined',
'A2' : 'attribute'}
f.close()
return conll_dict, gal_to_propbank_dict
def treegal_conll_add_args(conll_dictionary):
for sent_id in conll_dictionary.keys(): # each sent
sent = conll_dictionary[sent_id]['sent_data']
verbs = []
for word_num in sent:
base = sent[word_num]
if base['upos'] == 'VERB':
verbs.append(word_num)
for word_num in sent:
base = sent[word_num]
head = base['head']
xpos = base['xpos']
deprel = base['deprel']
if word_num in verbs:
verb_number = verbs.index(word_num)
base['arg'] = 'r' + str(verb_number) + ':root'
else:
if head in verbs:
verb_number = verbs.index(head)
possible_args = sent[head]['args'].keys()
if deprel == 'obl':
if xpos == 'Zgms' and 'A0' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg0'
elif xpos == 'Scfs' or xpos == 'Tnfs' and 'A1' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg1'
elif xpos == 'Infp' and 'A2' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg2'
else:
base['arg'] = '_'
elif deprel == 'nsubj':
if 'A0' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg0'
else:
base['arg'] = '_'
elif deprel == 'obj':
if 'A1' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg1'
else:
base['arg'] = '_'
elif deprel == 'iobj':
if 'A2' in possible_args:
base['arg'] = 'r' + str(verb_number) + ':arg2'
else:
base['arg'] = '_'
else:
base['arg'] = '_'
else:
base['arg'] = '_'
def write_to_conllu(conll_dict, file_name):
with open(file_name, 'w') as f:
f.write('# global.columns = ID FORM LEMMA UPOS XPOS FEATS HEAD DEPREL DEPS MISC ROLE ARG\n') # ARG= root#:arg
for sent_id in conll_dict:
base = conll_dict[sent_id]['sent_data']
f.write('\n')
f.write(f'# sent_id = {sent_id}\n')
f.write(f'# sent = {conll_dict[sent_id]["sent"]}\n')
for word_id in base:
word_base = base[word_id]
f.write(f'{word_id} {word_base["word_form"]} {word_base["lemma"]} {word_base["upos"]} {word_base["xpos"]} {word_base["feats"]} {word_base["head"]} {word_base["deprel"]} {word_base["deps"]} {word_base["misc"]} {word_base["role"]} {word_base["arg"]}\n')
f.close()
def import_data_from_conllu(file_path, ddict=None):
sent_count = 0
if ddict:
data_dict = ddict
else:
data_dict = {'tokens' : [],
'tags' : []
}
with open(file_path, 'r') as f:
lines = f.readlines()
toks = []
labels = []
for line in lines:
if '# sent ' in line:
if sent_count != 0:
if toks not in data_dict['tokens']:
data_dict['tokens'].append(toks)
data_dict['tags'].append(labels)
toks = []
labels = []
sent_count += 1
elif line.startswith('#') or len(line.split()) <2:
continue
else:
line = line.split()
tok_idx = line[0]
tok = line[1]
arg = line[-1]
try:
int_tok_idx = int(tok_idx)
toks.append(tok)
if arg == '_':
labels.append('O')
else:
labels.append(arg)
except:
continue
f.close()
return data_dict
if __name__ == "__main__":
# process CTG data & add arguments
try:
ctg_conll_dict_dev, ctg_gal_to_propbank_dict_dev = ctg_to_conll_dict('data/gl_ctg-ud-dev.conllu')
ctg_conll_dict_test, ctg_gal_to_propbank_dict_test = ctg_to_conll_dict('data/gl_ctg-ud-test.conllu',
conll_dict=ctg_conll_dict_dev,
gal_to_propbank_dict=ctg_gal_to_propbank_dict_dev)
ctg_conll_dict_all, ctg_gal_to_propbank_dict_all = ctg_to_conll_dict('data/gl_ctg-ud-train.conllu',
conll_dict=ctg_conll_dict_test,
gal_to_propbank_dict=ctg_gal_to_propbank_dict_test)
ctg_conll_add_args(ctg_conll_dict_all)
except:
print('Error! Cannot find CTG files. Please double check you have downloaded all files and that they are stored correctly.')
print('Files should be stored as: "data/gl_ctg-ud-dev.conllu", "data/gl_ctg-ud-test.conllu", and "data/gl_ctg-ud-train.conllu".')
# process TreeGal data & add arguments
try:
treegal_conll_dict_test, treegal_gal_to_propbank_dict_test = treegal_to_conll_dict('data/gl_treegal-ud-test.conllu')
treegal_conll_dict_all, treegal_gal_to_propbank_dict_all = treegal_to_conll_dict('data/gl_treegal-ud-train.conllu',
conll_dict=treegal_conll_dict_test,
gal_to_propbank_dict=treegal_gal_to_propbank_dict_test)
treegal_conll_add_args(treegal_conll_dict_all)
except:
print('Error! Cannot find TreeGal files. Please double check you have downloaded all files and that they are stored correctly.')
print('Files should be stored: "data/gl_treegal-ud-test.conllu", and "data/gl_treegal-ud-train.conllu".')
# write new data dictionaries to conllu files
write_to_conllu(ctg_conll_dict_all, 'data/ctg_conll_dict_all.conllu')
write_to_conllu(treegal_conll_dict_all, 'data/treegal_conll_dict_all.conllu')
# import conllu files, join CTG and TreeGal data, and remove duplicate sentences
ctg_data = import_data_from_conllu('data/ctg_conll_dict_all.conllu')
all_data = import_data_from_conllu('data/ctg_conll_dict_all.conllu', ddict=ctg_data) # fin_new_dict =
# assign sentence IDs
all_sent_ids = [i for i in range(len(all_data['tokens']))]
all_data['ids'] = all_sent_ids
# ensure data combined correctly
assert len(all_data['tags']) == len(all_data['tokens'])
assert len(all_data['ids']) == len(all_data['tags'])
# define labels
all_labels = sorted(set(label for labels in all_data['tags'] for label in labels))
# convert to Dataset object, split into train & test sets, and convert labels to ClassLabel
gal_ds = Dataset.from_dict(all_data)
split_gal_ds = gal_ds.train_test_split(test_size=0.2, shuffle=True)
final_gal_ds = split_gal_ds.cast_column('tags', Sequence(ClassLabel(names=all_labels)))
# save Dataset object locally
final_gal_ds.save_to_disk('data/final_gal_ds.hf')
# to load from local run command
# final_gal_ds = load_from_disk('data/final_gal_ds.hf')