|
1
|
|
|
"""Functions that use Natural Language Processing. |
|
2
|
|
|
|
|
3
|
|
|
Word relationships found (via NLTK and other libraries) |
|
4
|
|
|
to find and generate related words. |
|
5
|
|
|
""" |
|
6
|
|
|
|
|
7
|
|
|
|
|
8
|
|
|
from __future__ import absolute_import |
|
9
|
|
|
|
|
10
|
|
|
import itertools |
|
11
|
|
|
|
|
12
|
|
|
from nltk.corpus import ( |
|
|
|
|
|
|
13
|
|
|
verbnet, |
|
14
|
|
|
wordnet, |
|
15
|
|
|
) |
|
16
|
|
|
|
|
17
|
|
|
from . import normalization |
|
18
|
|
|
|
|
19
|
|
|
|
|
20
|
|
|
def _get_synset_words(word): |
|
21
|
|
|
"""Simple helper wrapping the more involved get_synsets function. |
|
22
|
|
|
|
|
23
|
|
|
Args: |
|
24
|
|
|
word (str): The seed word. |
|
25
|
|
|
|
|
26
|
|
|
Returns: |
|
27
|
|
|
words (list): The list of NLTK words. |
|
28
|
|
|
""" |
|
29
|
|
|
res = get_synsets([word])[word] |
|
30
|
|
|
if not res: |
|
31
|
|
|
return [] |
|
32
|
|
|
res = res.values() |
|
33
|
|
|
words = list(normalization.flatten([l for l in res if l])) |
|
34
|
|
|
return words |
|
35
|
|
|
|
|
36
|
|
|
|
|
37
|
|
|
def print_all_synset_categories(): |
|
38
|
|
|
"""Print all domains and categories for research purposes. |
|
39
|
|
|
|
|
40
|
|
|
Returns: |
|
41
|
|
|
categories (list): A list of all wordnet synsets. |
|
42
|
|
|
""" |
|
43
|
|
|
categories = [] |
|
44
|
|
|
for synset in list(wordnet.all_synsets('n')): |
|
45
|
|
|
categories.append(synset) |
|
46
|
|
|
return categories |
|
47
|
|
|
|
|
48
|
|
|
|
|
49
|
|
|
def _get_lemma_names(sub_synset, use_definitions=False): |
|
|
|
|
|
|
50
|
|
|
results = [] |
|
51
|
|
|
if sub_synset(): |
|
52
|
|
|
for v in sub_synset(): |
|
53
|
|
|
if hasattr(v.lemma_names, '__call__'): |
|
54
|
|
|
results += v.lemma_names() |
|
55
|
|
|
else: |
|
56
|
|
|
results += v.lemma_names |
|
57
|
|
|
if use_definitions: |
|
58
|
|
|
results.append(v.definition.split()) |
|
59
|
|
|
return results |
|
60
|
|
|
|
|
61
|
|
|
|
|
62
|
|
|
def get_hyponyms(synset, use_definitions=False): |
|
63
|
|
|
"""Extract hyponyms from a synset. |
|
64
|
|
|
|
|
65
|
|
|
Args: |
|
66
|
|
|
synset (object): The synset instance. |
|
67
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
68
|
|
|
|
|
69
|
|
|
Returns: |
|
70
|
|
|
list: The results list. |
|
71
|
|
|
""" |
|
72
|
|
|
return _get_lemma_names(synset.hyponyms, use_definitions=use_definitions) |
|
73
|
|
|
|
|
74
|
|
|
|
|
75
|
|
|
def get_inst_hyponyms(synset, use_definitions=False): |
|
76
|
|
|
"""Extract instance hyponyms from a synset. |
|
77
|
|
|
|
|
78
|
|
|
Args: |
|
79
|
|
|
synset (object): The synset instance. |
|
80
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
81
|
|
|
|
|
82
|
|
|
Returns: |
|
83
|
|
|
list: The results list. |
|
84
|
|
|
""" |
|
85
|
|
|
return _get_lemma_names( |
|
86
|
|
|
synset.instance_hyponyms, use_definitions=use_definitions) |
|
87
|
|
|
|
|
88
|
|
|
|
|
89
|
|
|
def get_member_meronyms(synset, use_definitions=False): |
|
90
|
|
|
"""Extract meronyms from a synset. |
|
91
|
|
|
|
|
92
|
|
|
Args: |
|
93
|
|
|
synset (object): The synset instance. |
|
94
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
95
|
|
|
|
|
96
|
|
|
Returns: |
|
97
|
|
|
list: The results list. |
|
98
|
|
|
""" |
|
99
|
|
|
return _get_lemma_names( |
|
100
|
|
|
synset.member_meronyms, use_definitions=use_definitions) |
|
101
|
|
|
|
|
102
|
|
|
|
|
103
|
|
|
def get_substance_meronyms(synset, use_definitions=False): |
|
104
|
|
|
"""Extract substance meronyms from a synset. |
|
105
|
|
|
|
|
106
|
|
|
Args: |
|
107
|
|
|
synset (object): The synset instance. |
|
108
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
109
|
|
|
|
|
110
|
|
|
Returns: |
|
111
|
|
|
list: The results list. |
|
112
|
|
|
""" |
|
113
|
|
|
return _get_lemma_names( |
|
114
|
|
|
synset.substance_meronyms, use_definitions=use_definitions) |
|
115
|
|
|
|
|
116
|
|
|
|
|
117
|
|
|
def get_part_meronyms(synset, use_definitions=False): |
|
118
|
|
|
"""Extract part meronyms from a synset. |
|
119
|
|
|
|
|
120
|
|
|
Args: |
|
121
|
|
|
synset (object): The synset instance. |
|
122
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
123
|
|
|
|
|
124
|
|
|
Returns: |
|
125
|
|
|
list: The results list. |
|
126
|
|
|
""" |
|
127
|
|
|
return _get_lemma_names( |
|
128
|
|
|
synset.part_meronyms, use_definitions=use_definitions) |
|
129
|
|
|
|
|
130
|
|
|
|
|
131
|
|
|
def get_substance_holoynms(synset, use_definitions=False): |
|
132
|
|
|
"""Extract substance holoynms from a synset. |
|
133
|
|
|
|
|
134
|
|
|
Args: |
|
135
|
|
|
synset (object): The synset instance. |
|
136
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
137
|
|
|
|
|
138
|
|
|
Returns: |
|
139
|
|
|
list: The results list. |
|
140
|
|
|
""" |
|
141
|
|
|
return _get_lemma_names( |
|
142
|
|
|
synset.substance_holonyms, use_definitions=use_definitions) |
|
143
|
|
|
|
|
144
|
|
|
|
|
145
|
|
|
def get_topic_domains(synset, use_definitions=False): |
|
146
|
|
|
"""Extract topic domains from a synset. |
|
147
|
|
|
|
|
148
|
|
|
Args: |
|
149
|
|
|
synset (object): The synset instance. |
|
150
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
151
|
|
|
|
|
152
|
|
|
Returns: |
|
153
|
|
|
list: The results list. |
|
154
|
|
|
""" |
|
155
|
|
|
return _get_lemma_names( |
|
156
|
|
|
synset.topic_domains, use_definitions=use_definitions) |
|
157
|
|
|
|
|
158
|
|
|
|
|
159
|
|
|
def get_region_domains(synset, use_definitions=False): |
|
160
|
|
|
"""Extract region domains from a synset. |
|
161
|
|
|
|
|
162
|
|
|
Args: |
|
163
|
|
|
synset (object): The synset instance. |
|
164
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
165
|
|
|
|
|
166
|
|
|
Returns: |
|
167
|
|
|
list: The results list. |
|
168
|
|
|
""" |
|
169
|
|
|
return _get_lemma_names( |
|
170
|
|
|
synset.region_domains, use_definitions=use_definitions) |
|
171
|
|
|
|
|
172
|
|
|
|
|
173
|
|
|
def get_usage_domains(synset, use_definitions=False): |
|
174
|
|
|
"""Extract usage domains from a synset. |
|
175
|
|
|
|
|
176
|
|
|
Args: |
|
177
|
|
|
synset (object): The synset instance. |
|
178
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
179
|
|
|
|
|
180
|
|
|
Returns: |
|
181
|
|
|
list: The results list. |
|
182
|
|
|
""" |
|
183
|
|
|
return _get_lemma_names( |
|
184
|
|
|
synset.usage_domains, use_definitions=use_definitions) |
|
185
|
|
|
|
|
186
|
|
|
|
|
187
|
|
|
def get_attributes(synset, use_definitions=False): |
|
188
|
|
|
"""Extract attributes from a synset. |
|
189
|
|
|
|
|
190
|
|
|
Args: |
|
191
|
|
|
synset (object): The synset instance. |
|
192
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
193
|
|
|
|
|
194
|
|
|
Returns: |
|
195
|
|
|
list: The results list. |
|
196
|
|
|
""" |
|
197
|
|
|
return _get_lemma_names( |
|
198
|
|
|
synset.attributes, use_definitions=use_definitions) |
|
199
|
|
|
|
|
200
|
|
|
|
|
201
|
|
|
def get_entailments(synset, use_definitions=False): |
|
202
|
|
|
"""Extract entailments from a synset. |
|
203
|
|
|
|
|
204
|
|
|
Args: |
|
205
|
|
|
synset (object): The synset instance. |
|
206
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
207
|
|
|
|
|
208
|
|
|
Returns: |
|
209
|
|
|
list: The results list. |
|
210
|
|
|
""" |
|
211
|
|
|
return _get_lemma_names( |
|
212
|
|
|
synset.entailments, use_definitions=use_definitions) |
|
213
|
|
|
|
|
214
|
|
|
|
|
215
|
|
|
def get_causes(synset, use_definitions=False): |
|
216
|
|
|
"""Extract causes from a synset. |
|
217
|
|
|
|
|
218
|
|
|
Args: |
|
219
|
|
|
synset (object): The synset instance. |
|
220
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
221
|
|
|
|
|
222
|
|
|
Returns: |
|
223
|
|
|
list: The results list. |
|
224
|
|
|
""" |
|
225
|
|
|
if synset.causes(): |
|
226
|
|
|
return _get_lemma_names( |
|
227
|
|
|
synset.causes, use_definitions=use_definitions) |
|
228
|
|
|
|
|
229
|
|
|
|
|
230
|
|
|
def get_also_sees(synset, use_definitions=False): |
|
231
|
|
|
"""Extract also-sees from a synset. |
|
232
|
|
|
|
|
233
|
|
|
Args: |
|
234
|
|
|
synset (object): The synset instance. |
|
235
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
236
|
|
|
|
|
237
|
|
|
Returns: |
|
238
|
|
|
list: The results list. |
|
239
|
|
|
""" |
|
240
|
|
|
return _get_lemma_names( |
|
241
|
|
|
synset.also_sees, use_definitions=use_definitions) |
|
242
|
|
|
|
|
243
|
|
|
|
|
244
|
|
|
def get_verb_groups(synset, use_definitions=False): |
|
245
|
|
|
"""Extract verb groups from a synset. |
|
246
|
|
|
|
|
247
|
|
|
Args: |
|
248
|
|
|
synset (object): The synset instance. |
|
249
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
250
|
|
|
|
|
251
|
|
|
Returns: |
|
252
|
|
|
list: The results list. |
|
253
|
|
|
""" |
|
254
|
|
|
return _get_lemma_names( |
|
255
|
|
|
synset.verb_groups, use_definitions=use_definitions) |
|
256
|
|
|
|
|
257
|
|
|
|
|
258
|
|
|
def get_similartos(synset, use_definitions=False): |
|
259
|
|
|
"""Extract similar-tos from a synset. |
|
260
|
|
|
|
|
261
|
|
|
Args: |
|
262
|
|
|
synset (object): The synset instance. |
|
263
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
264
|
|
|
|
|
265
|
|
|
Returns: |
|
266
|
|
|
list: The results list. |
|
267
|
|
|
""" |
|
268
|
|
|
return _get_lemma_names( |
|
269
|
|
|
synset.similar_tos, use_definitions=use_definitions) |
|
270
|
|
|
|
|
271
|
|
|
|
|
272
|
|
|
def get_member_holoynms(synset, use_definitions=False): |
|
273
|
|
|
"""Extract member holonyms from a synset. |
|
274
|
|
|
|
|
275
|
|
|
Args: |
|
276
|
|
|
synset (object): The synset instance. |
|
277
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
278
|
|
|
|
|
279
|
|
|
Returns: |
|
280
|
|
|
list: The results list. |
|
281
|
|
|
""" |
|
282
|
|
|
return _get_lemma_names( |
|
283
|
|
|
synset.member_holonyms, use_definitions=use_definitions) |
|
284
|
|
|
|
|
285
|
|
|
|
|
286
|
|
|
def get_part_holoynms(synset, use_definitions=False): |
|
287
|
|
|
"""Extract part holonyms from a synset. |
|
288
|
|
|
|
|
289
|
|
|
Args: |
|
290
|
|
|
synset (object): The synset instance. |
|
291
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
292
|
|
|
|
|
293
|
|
|
Returns: |
|
294
|
|
|
list: The results list. |
|
295
|
|
|
""" |
|
296
|
|
|
return _get_lemma_names( |
|
297
|
|
|
synset.part_holonyms, use_definitions=use_definitions) |
|
298
|
|
|
|
|
299
|
|
|
|
|
300
|
|
|
def get_instance_hypernyms(synset, use_definitions=False): |
|
301
|
|
|
"""Extract instance hypernyms from a synset. |
|
302
|
|
|
|
|
303
|
|
|
Args: |
|
304
|
|
|
synset (object): The synset instance. |
|
305
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
306
|
|
|
|
|
307
|
|
|
Returns: |
|
308
|
|
|
list: The results list. |
|
309
|
|
|
""" |
|
310
|
|
|
return _get_lemma_names( |
|
311
|
|
|
synset.instance_hypernyms, use_definitions=use_definitions) |
|
312
|
|
|
|
|
313
|
|
|
|
|
314
|
|
|
def get_hypernyms(synset, use_definitions=False): |
|
315
|
|
|
"""Extract hypernyms from a synset. |
|
316
|
|
|
|
|
317
|
|
|
Args: |
|
318
|
|
|
synset (object): The synset instance. |
|
319
|
|
|
use_definitions (bool, optional): Extract definitions from the synset. |
|
320
|
|
|
|
|
321
|
|
|
Returns: |
|
322
|
|
|
list: The results list. |
|
323
|
|
|
""" |
|
324
|
|
|
return _get_lemma_names( |
|
325
|
|
|
synset.hypernyms, use_definitions=use_definitions) |
|
326
|
|
|
|
|
327
|
|
|
|
|
328
|
|
|
def get_verb_lemmas(verbs): |
|
329
|
|
|
"""Return verbnet lemmas for the given verbs. |
|
330
|
|
|
|
|
331
|
|
|
These verbs are stemmed before lookup to prevent empty results. |
|
332
|
|
|
|
|
333
|
|
|
Args: |
|
334
|
|
|
verbs (list) - The list of verbs (verbs) to reference. |
|
335
|
|
|
|
|
336
|
|
|
Returns: |
|
337
|
|
|
lemmas (list) - A list of lemmas for all verbs |
|
338
|
|
|
- these are not separated by verb. |
|
339
|
|
|
""" |
|
340
|
|
|
lemmas = [] |
|
341
|
|
|
for verb in normalization.stem_words(verbs): |
|
342
|
|
|
_lemmas = verbnet.classids(lemma=verb) |
|
343
|
|
|
lemmas += [l.split('-')[0] for l in _lemmas] |
|
344
|
|
|
return lemmas |
|
345
|
|
|
|
|
346
|
|
|
|
|
347
|
|
|
def get_word_synsets(word): |
|
348
|
|
|
"""Get all synsets for a word. |
|
349
|
|
|
|
|
350
|
|
|
Args: |
|
351
|
|
|
word (str): The word to lookup. |
|
352
|
|
|
|
|
353
|
|
|
Returns: |
|
354
|
|
|
object: The synset ring instance. |
|
355
|
|
|
""" |
|
356
|
|
|
return wordnet.synsets(word.encode('utf-8'), pos=None) |
|
357
|
|
|
|
|
358
|
|
|
|
|
359
|
|
|
def get_synset_definitions(word): |
|
360
|
|
|
"""Return all possible definitions for synsets in a word synset ring. |
|
361
|
|
|
|
|
362
|
|
|
Args: |
|
363
|
|
|
word (str): The word to lookup. |
|
364
|
|
|
|
|
365
|
|
|
Returns: |
|
366
|
|
|
definitions (list): The synset definitions list. |
|
367
|
|
|
""" |
|
368
|
|
|
definitions = [] |
|
369
|
|
|
synsets = get_word_synsets(word) |
|
370
|
|
|
for _synset in synsets: |
|
371
|
|
|
definitions.append(_synset.definition().split()) |
|
372
|
|
|
return definitions |
|
373
|
|
|
|
|
374
|
|
|
|
|
375
|
|
|
def get_synsets_definitions(words): |
|
376
|
|
|
"""Return all possible definitions for all synsets in the synset ring. |
|
377
|
|
|
|
|
378
|
|
|
Args: |
|
379
|
|
|
words (list): The list of words. |
|
380
|
|
|
|
|
381
|
|
|
Returns: |
|
382
|
|
|
sets (list): The synsets. |
|
383
|
|
|
""" |
|
384
|
|
|
return [get_synset_definitions(w) for w in words if w] |
|
385
|
|
|
|
|
386
|
|
|
|
|
387
|
|
|
def get_synsets(words, use_definitions=False, clean=False): |
|
388
|
|
|
"""Brute force loop on a synset ring to get all related words. |
|
389
|
|
|
|
|
390
|
|
|
You are expected to filter or remove any that are not relevant separately, |
|
391
|
|
|
if the resultant set is too long. |
|
392
|
|
|
The scoring module provides tools to filter based on pronunciation, |
|
393
|
|
|
but you can write your own and extend the functionality. |
|
394
|
|
|
|
|
395
|
|
|
Args: |
|
396
|
|
|
words (list): The list of words. |
|
397
|
|
|
use_definitions (bool, optional): Determine if definition words |
|
398
|
|
|
should also be extracted. |
|
399
|
|
|
clean (bool, optional): Determine if set should be de-duped, |
|
400
|
|
|
cleaned, etc... |
|
401
|
|
|
|
|
402
|
|
|
Returns: |
|
403
|
|
|
results (dict): The results dictionary. |
|
404
|
|
|
""" |
|
405
|
|
|
results = {} |
|
406
|
|
|
|
|
407
|
|
|
for word in words: |
|
408
|
|
|
synsets = get_word_synsets(word) |
|
409
|
|
|
|
|
410
|
|
|
key = {'synset_original': []} |
|
411
|
|
|
|
|
412
|
|
|
for synset in synsets: |
|
413
|
|
|
if hasattr(synset.lemma_names, '__call__'): |
|
414
|
|
|
key['synset_original'].append(synset.lemma_names()) |
|
415
|
|
|
else: |
|
416
|
|
|
key['synset_original'].append(synset.lemma_names) |
|
417
|
|
|
|
|
418
|
|
|
# More Specific *nyms (deep) |
|
419
|
|
|
key['hyponyms'] = get_hyponyms( |
|
420
|
|
|
synset, use_definitions=use_definitions) |
|
421
|
|
|
key['instance_hyponyms'] = get_inst_hyponyms( |
|
422
|
|
|
synset, use_definitions=use_definitions) |
|
423
|
|
|
key['member_meronyms'] = get_member_meronyms( |
|
424
|
|
|
synset, use_definitions=use_definitions) |
|
425
|
|
|
key['substance_meronyms'] = get_substance_meronyms( |
|
426
|
|
|
synset, use_definitions=use_definitions) |
|
427
|
|
|
key['part_meronyms'] = get_part_meronyms( |
|
428
|
|
|
synset, use_definitions=use_definitions) |
|
429
|
|
|
key['substance_holonyms'] = get_substance_holoynms( |
|
430
|
|
|
synset, use_definitions=use_definitions) |
|
431
|
|
|
|
|
432
|
|
|
# More Generic *nyms (shallow) |
|
433
|
|
|
key['member_holonyms'] = get_member_holoynms( |
|
434
|
|
|
synset, use_definitions=use_definitions) |
|
435
|
|
|
key['part_holonyms'] = get_part_holoynms( |
|
436
|
|
|
synset, use_definitions=use_definitions) |
|
437
|
|
|
key['instance_hypernyms'] = get_instance_hypernyms( |
|
438
|
|
|
synset, use_definitions=use_definitions) |
|
439
|
|
|
key['hypernyms'] = get_hypernyms( |
|
440
|
|
|
synset, use_definitions=use_definitions) |
|
441
|
|
|
|
|
442
|
|
|
# Other types |
|
443
|
|
|
key['topic_domains'] = get_topic_domains( |
|
444
|
|
|
synset, use_definitions=use_definitions) |
|
445
|
|
|
key['region_domains'] = get_region_domains( |
|
446
|
|
|
synset, use_definitions=use_definitions) |
|
447
|
|
|
key['usage_domains'] = get_usage_domains( |
|
448
|
|
|
synset, use_definitions=use_definitions) |
|
449
|
|
|
key['attributes'] = get_attributes( |
|
450
|
|
|
synset, use_definitions=use_definitions) |
|
451
|
|
|
key['entailments'] = get_entailments( |
|
452
|
|
|
synset, use_definitions=use_definitions) |
|
453
|
|
|
key['causes'] = get_causes( |
|
454
|
|
|
synset, use_definitions=use_definitions) |
|
455
|
|
|
key['also_sees'] = get_also_sees( |
|
456
|
|
|
synset, use_definitions=use_definitions) |
|
457
|
|
|
key['verb_groups'] = get_verb_groups( |
|
458
|
|
|
synset, use_definitions=use_definitions) |
|
459
|
|
|
key['similar_tos'] = get_similartos( |
|
460
|
|
|
synset, use_definitions=use_definitions) |
|
461
|
|
|
|
|
462
|
|
|
results[word] = key |
|
463
|
|
|
|
|
464
|
|
|
# 1. get words back |
|
465
|
|
|
# 2. flatten nested array |
|
466
|
|
|
# 3. split up words |
|
467
|
|
|
# 4. filter, clean, stem, uniquify |
|
468
|
|
|
|
|
469
|
|
|
for nlp_type in results: |
|
470
|
|
|
if clean: |
|
471
|
|
|
results[nlp_type] = sorted( |
|
472
|
|
|
normalization.uniquify( |
|
473
|
|
|
normalization.clean_sort( |
|
474
|
|
|
normalization.remove_stop_words( |
|
475
|
|
|
normalization.stem_words( |
|
476
|
|
|
normalization.remove_bad_words( |
|
477
|
|
|
list(itertools.chain( |
|
478
|
|
|
*results[nlp_type])))))))) |
|
479
|
|
|
|
|
480
|
|
|
return results |
|
481
|
|
|
|
This can be caused by one of the following:
1. Missing Dependencies
This error could indicate a configuration issue of Pylint. Make sure that your libraries are available by adding the necessary commands.
2. Missing __init__.py files
This error could also result from missing
__init__.pyfiles in your module folders. Make sure that you place one file in each sub-folder.