forked from realyanyang/disambiguation
-
Notifications
You must be signed in to change notification settings - Fork 0
/
final_subpipe.py
644 lines (574 loc) · 28.3 KB
/
final_subpipe.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
#!/usr/bin/env python3
# -*- encoding: utf-8 -*-
'''
@File : final_subpipe.py
@Time : 2019/11/28 20:05:43
@Author : Yan Yang
@Contact : [email protected]
@Desc : None
'''
# .::::.
# .::::::::.
# :::::::::::
# ..:::::::::::'
# '::::::::::::'
# .::::::::::
# '::::::::::::::..
# ..::::::::::::.
# ``::::::::::::::::
# ::::``:::::::::' .:::.
# ::::' ':::::' .::::::::.
# .::::' :::: .:::::::'::::.
# .:::' ::::: .:::::::::' ':::::.
# .::' :::::.:::::::::' ':::::.
# .::' ::::::::::::::' ``::::.
# ...::: ::::::::::::' ``::.
# ```` ':. ':::::::::' ::::..
# '.:::::' ':'````..
# 美女保佑 永无BUG
import os
from collections import defaultdict
import numpy as np
import tqdm
import torch
from triplet_model import TripletModel
import pandas as pd
from utils import load_json, load_pickle, save_json, save_pickle, clean_name, get_name_index, SK_MLP
import time
from utils import get_coauthor_v2, get_year_diff_v2, get_venue_score_v2, get_org_score_v2, get_key_word_num_v2, get_relative_year_feature
from utils import get_org_with_set_score, get_venue_with_set_score, get_keywords_with_set_score
import matplotlib.pyplot as plt
import seaborn as sns
from utils import TextToVec
from multiprocessing import Pool
import math
from nltk.corpus import stopwords
import re
import string
from nltk.tokenize import word_tokenize
from nltk.stem import WordNetLemmatizer
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.metrics.pairwise import cosine_similarity
sns.set()
TEST_PUB_PATH = './final_dir/data/cna_test_pub.json'
TEST_UNASS_PATH = './final_dir/data/cna_test_unass_competition.json'
TEST_FEATURE_DIR = './final_dir/feature'
FINAL_DIR = './final_dir'
RESULT_SAVE_DIR = './final_dir/save'
NEW_DATA_DIR = './new-data' # original info, for test
STACK_MODEL_DIR_v2 = './stack_model_v2'
RANDOM_SEED = 1129
np.random.seed(RANDOM_SEED)
os.makedirs(TEST_FEATURE_DIR, exist_ok=True)
os.makedirs(RESULT_SAVE_DIR, exist_ok=True)
BASE_COLS = [
'coauthors_count', 'coauthors_count_by_all_count', 'coauthors_count_by_this_coauthor_count',
'this_paper_coauthor_count', 'this_paper_coathor_count_by_all_coauthor', 'this_paper_coauthor_count_by_this_paper_coauthor_count',
'min_diff', 'max_diff', 'mean_diff', 'meadian_diff', 'min_max_avg_diff', 'is_in_range',
'this_year_count', 'this_year_count_by_all_year', 'is_in_cate_range', 'before_one', 'before_two',
'later_one', 'later_two', 'venue_max_score', 'venue_mean_score', 'venue_max_add_score',
'venue_mean_add_score', 'venue_is_match', 'venue_score_add_score', 'org_max_score', 'org_mean_score',
'org_max_add_score', 'org_mean_add_score', 'org_is_match', 'org_score_add_score', 'org_year_abs_diff',
'keywords_max_score', 'keywords_mean_score', 'rela_year_diff', 'rela_coauthor_count',
'rela_coauthor_count_by1', 'rela_coauthor_count_by2', 'rela_org_max_score',
'rela_org_mean_score', 'rela_org_max_add_score', 'rela_org_mean_add_score',
'rela_org_score_add_score', 'rela_venue_max_score', 'rela_venue_mean_score',
'rela_venue_max_add_score', 'rela_venue_mean_add_score', 'rela_venue_score_add_score',
'rela_keyword_max_score', 'rela_keyword_mean_score'
]
# length: 50 !
SET_INFO_COLS = [
'org_set_count', 'org_set_count_by_all_count',
'org_set_count_by_this_count', 'venue_word_count', 'venue_word_count_by_all_count',
'venue_word_count_by_this_venue_count', 'keyword_count', 'keyword_count_by_all_count',
'keyword_count_by_this_keyword_count'
]
# length: 9 !
TITLE_COLS = [
'title'
]
np.random.seed(RANDOM_SEED)
def get_features(aid_pid_pair, pid_info_dict, aid_author_info_dict, aid_year_info_dict, aid_venue_dict, aid_org_year_list, aid_keywords_dict, aid_year_all_info_dict, org_info_set, aid_venue_set, aid_keywords_set):
feature = [
*get_coauthor_v2(aid_pid_pair, aid_author_info_dict, pid_info_dict),
*get_year_diff_v2(aid_year_info_dict, pid_info_dict),
*get_venue_score_v2(aid_venue_dict, pid_info_dict),
*get_org_score_v2(aid_pid_pair, aid_org_year_list, pid_info_dict),
*get_key_word_num_v2(aid_keywords_dict, pid_info_dict),
*get_relative_year_feature(aid_pid_pair, aid_year_all_info_dict, pid_info_dict),
*get_org_with_set_score(aid_pid_pair, pid_info_dict, org_info_set),
*get_venue_with_set_score(aid_venue_set, pid_info_dict),
*get_keywords_with_set_score(aid_keywords_set, pid_info_dict),
]
return feature
def gen_base_feature(index, multi_size):
# process test data and save in pickle
# testdatafeatures --> {pid-with-index: {candidate-aids: [...], data: [[xxx], [xxx], [xxx]...]}}
test_unass = load_json(TEST_UNASS_PATH)
test_pub = load_json(TEST_PUB_PATH)
# whole_author_profile_pub = load_json(WHOLE_AUTHOR_PROFILE_PUB_PATH)
aid2yearinfo = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2yearinfo.pkl'))
aid2coauthor = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2coauthor.pkl'))
aid2venue = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2venue.pkl'))
aid2keywords = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2keywords.pkl'))
aid2year = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2year.pkl'))
aid2orgwithyear = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2orgwithyear.pkl'))
name2aids = load_pickle(os.path.join(NEW_DATA_DIR, 'name2aids.pkl'))
# aid2pids = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2pids.pkl'))
aid2orgset = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2orgset.pkl'))
aid2venueset = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2venueset.pkl'))
aid2keywordsset = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2keywordsset.pkl'))
name_map = load_json(os.path.join(FINAL_DIR, 'name.different.modified.json'))
original_name = [pair[0] for pair in name_map]
changed_name = [pair[1] for pair in name_map]
name_map2 = load_json(os.path.join(FINAL_DIR, 'name.different.2.modified.json'))
original_name2 = [pair[0] for pair in name_map2]
changed_name2 = [pair[1] for pair in name_map2]
single_range = math.ceil(len(test_unass) / multi_size)
start = index * single_range
end = (index + 1) * single_range if (index + 1) * single_range < len(test_unass) else len(test_unass)
testdatafeatures = {}
all_authors_name = list(name2aids.keys())
print('Gen test features ...')
for pid_with_index in tqdm.tqdm(test_unass[start:end]):
inner_dict = {}
now_pid, index = pid_with_index.split('-')
author_name = test_pub[now_pid]['authors'][int(index)]['name']
author_name = clean_name(author_name)
if pid_with_index == 'ToCcabLT-1':
author_name = 'junliang_wang'
if pid_with_index == 'cVvvcFzj-1':
author_name = 'xiaojun_liu'
if author_name in original_name2:
name_index = original_name2.index(author_name)
author_name = changed_name2[name_index]
elif author_name in original_name:
name_index = original_name.index(author_name)
author_name = changed_name[name_index]
else:
index = get_name_index(author_name, all_authors_name)
author_name = all_authors_name[index]
if isinstance(author_name, str):
candidate_aids = name2aids[author_name]
elif isinstance(author_name, list):
candidate_aids = []
for name in author_name:
candidate_aids.extend(name2aids[name].tolist())
candidate_aids = np.array(candidate_aids)
else:
raise ValueError("check author name ! ! !")
inner_dict['candidate-aids'] = candidate_aids
data = []
for aid in candidate_aids:
new_pair = (aid, pid_with_index)
pid_info_dict = test_pub[now_pid]
aid_author_info_dict = aid2coauthor[aid]
aid_year_info_dict = aid2year[aid]
aid_venue_dict = aid2venue[aid]
aid_org_year_list = aid2orgwithyear[aid]
aid_keywords_dict = aid2keywords[aid]
aid_year_all_info_dict = aid2yearinfo[aid]
org_info_set = aid2orgset[aid]
aid_venue_set = aid2venueset[aid]
aid_keywords_set = aid2keywordsset[aid]
data.append(get_features(new_pair, pid_info_dict, aid_author_info_dict, aid_year_info_dict, aid_venue_dict, aid_org_year_list, aid_keywords_dict, aid_year_all_info_dict, org_info_set, aid_venue_set, aid_keywords_set))
data = np.array(data)
inner_dict['data'] = data
testdatafeatures[pid_with_index] = inner_dict
# save_pickle(testdatafeatures, os.path.join(TEST_FEATURE_DIR, 'u6uRzaff-5.pkl'))
return testdatafeatures
def emb_pair_to_distance(text_model_name, mission, original_emb, save_path):
"""
original_emb shape: [(emb0, meb1), ...]
type: numpy.ndarray
"""
if mission != 'title' and mission != 'abstract':
raise ValueError('mission value error')
triplet_model = TripletModel()
triplet_model.load_state_dict(torch.load(os.path.join('./text-model', text_model_name)))
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
triplet_model = triplet_model.to(device)
original_emb0 = np.stack([pair[0].tolist() for pair in original_emb])
original_emb1 = np.stack([pair[1].tolist() for pair in original_emb])
original_emb0 = np.expand_dims(original_emb0, axis=1)
original_emb1 = np.expand_dims(original_emb1, axis=1)
original_emb0 = torch.from_numpy(original_emb0).to(device).to(torch.float)
original_emb1 = torch.from_numpy(original_emb1).to(device).to(torch.float)
triplet_model.eval()
with torch.no_grad():
emb0 = triplet_model.get_emb(original_emb0)
emb1 = triplet_model.get_emb(original_emb1)
emb_sidtance = torch.sqrt(torch.sum(torch.pow((emb0 - emb1), 2), dim=1))
emb_sidtance = emb_sidtance.cpu().numpy()
df = pd.DataFrame(data=emb_sidtance, columns=[mission])
df.to_pickle(save_path)
def gen_title_feature():
aid2titlevec = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2titlevec.pkl'))
test_unass = load_json(TEST_UNASS_PATH)
test_pub = load_json(TEST_PUB_PATH)
name2aids = load_pickle(os.path.join(NEW_DATA_DIR, 'name2aids.pkl'))
texttovec = TextToVec()
name_map = load_json(os.path.join(FINAL_DIR, 'name.different.modified.json'))
original_name = [pair[0] for pair in name_map]
changed_name = [pair[1] for pair in name_map]
name_map2 = load_json(os.path.join(FINAL_DIR, 'name.different.2.modified.json'))
original_name2 = [pair[0] for pair in name_map2]
changed_name2 = [pair[1] for pair in name_map2]
all_authors_name = list(name2aids.keys())
# test_title_feature --> {pid-with-index: {candidate-aids: [...], data: [(emb0, meb1), ...]}}
test_title_feature = {}
print('Gen title emb pair ...')
for pid_with_index in tqdm.tqdm(test_unass):
inner_dict = {}
now_pid, index = pid_with_index.split('-')
author_name = test_pub[now_pid]['authors'][int(index)]['name']
author_name = clean_name(author_name)
if pid_with_index == 'ToCcabLT-1':
author_name = 'junliang_wang'
if pid_with_index == 'cVvvcFzj-1':
author_name = 'xiaojun_liu'
if author_name in original_name2:
name_index = original_name2.index(author_name)
author_name = changed_name2[name_index]
elif author_name in original_name:
name_index = original_name.index(author_name)
author_name = changed_name[name_index]
else:
index = get_name_index(author_name, all_authors_name)
author_name = all_authors_name[index]
if isinstance(author_name, str):
candidate_aids = name2aids[author_name]
elif isinstance(author_name, list):
candidate_aids = []
for name in author_name:
candidate_aids.extend(name2aids[name].tolist())
candidate_aids = np.array(candidate_aids)
else:
raise ValueError("check author name !!!")
inner_dict['candidate-aids'] = candidate_aids
info = test_pub[now_pid].get('title')
if info is None:
emb = np.zeros(300)
else:
emb = texttovec.get_vec(info)
data = []
for aid in candidate_aids:
emb_pair = (aid2titlevec[aid], emb)
data.append(emb_pair)
inner_dict['data'] = data
test_title_feature[pid_with_index] = inner_dict
save_pickle(test_title_feature, os.path.join(TEST_FEATURE_DIR, 'test-title-emb-pair-name-clean-2.pkl'))
print('Gen title distance ...')
test_title_emb_pair = load_pickle(os.path.join(TEST_FEATURE_DIR, 'test-title-emb-pair-name-clean-2.pkl'))
test_unass = load_json(TEST_UNASS_PATH)
title_emb_pair = []
for pid_with_index in tqdm.tqdm(test_unass):
for pair in test_title_emb_pair[pid_with_index]['data']:
title_emb_pair.append(pair)
emb_pair_to_distance(
'tm.title.1.checkpoint.pth', 'title', title_emb_pair,
os.path.join(TEST_FEATURE_DIR, 'test-title-distance-df-name-clean-2.pkl')
)
def predict(models):
test_unass = load_json(TEST_UNASS_PATH)
testdatafeatures = load_pickle(os.path.join(TEST_FEATURE_DIR, 'testdatafeatures-withsetinfo.pkl'))
title_feature_df = pd.read_pickle(os.path.join(TEST_FEATURE_DIR, 'test-title-distance-df.pkl'))
title_feature = title_feature_df.values
models_loaded = []
for model_info in models:
model = {
'model': load_pickle(model_info['model']),
'ss': load_pickle(model_info['ss']),
'cols': model_info['cols'],
'score': model_info['score']
}
models_loaded.append(model)
scores = [model_info['score'] for model_info in models_loaded]
weights = [score / sum(scores) for score in scores]
weights = np.array(weights).reshape(1, len(models_loaded))
print(weights)
submission = defaultdict(list)
for pid_with_index in tqdm.tqdm(test_unass):
candidate_aids = testdatafeatures[pid_with_index]['candidate-aids']
data = testdatafeatures[pid_with_index]['data']
data_length = len(candidate_aids)
title_data = title_feature[:data_length]
title_feature = title_feature[data_length:]
data = np.concatenate((data, title_data), axis=1)
default_cols = BASE_COLS + SET_INFO_COLS + TITLE_COLS
df = pd.DataFrame(data=data, columns=default_cols)
inner_data = np.zeros((len(candidate_aids), len(models_loaded)))
for num, model_info in enumerate(models_loaded):
model = model_info['model']
ss = model_info['ss']
data = df[model_info['cols']].values
data = ss.transform(data)
output = model.predict_proba(data)
inner_data[:, num] = output
final_output = np.sum((inner_data * weights), axis=1)
predict_author = candidate_aids[np.argmax(final_output)]
submission[predict_author].append(pid_with_index.split('-')[0])
save_json(submission, os.path.join(FINAL_DIR, 'result-top3models.json'))
def see_year_distribution():
# test_pub = load_json(TEST_PUB_PATH)
# # test_pub = load_json('./data2/cna_data/whole_author_profile_pub.json')
# year_count = []
# for pid in test_pub:
# year = test_pub[pid].get('year', '0')
# if year == '':
# year = 0
# else:
# year = int(year)
# if year <= 1500 or year >= 2100:
# year = 0
# if year != 0:
# year_count.append(year)
# df = pd.DataFrame(data=year_count, columns=['year'])
# plt.figure(figsize=(10, 5))
# sns.countplot(x='year', data=df)
# plt.xticks(rotation='vertical')
# plt.savefig(os.path.join(FINAL_DIR, 'test.year.png'))
diff_year = []
df_dict = load_pickle('./final_dir/feature/testdatafeatures-withsetinfo.pkl')
for pid_with_index in df_dict:
data = df_dict[pid_with_index]['data']
default_cols = BASE_COLS + SET_INFO_COLS
df = pd.DataFrame(data=data, columns=default_cols)
diff_year.extend(df['max_diff'].values.tolist())
df = pd.DataFrame(data=diff_year, columns=['diff_year'])
plt.figure(figsize=(15, 5))
sns.countplot(x='diff_year', data=df)
plt.xticks(rotation='vertical')
plt.savefig(os.path.join(FINAL_DIR, 'test.max.diff.year.png'))
def multi_gen_base_feature(multi_size):
result = []
p = Pool(multi_size)
for index in range(multi_size):
result.append(p.apply_async(gen_base_feature, args=(index, multi_size)))
print('Process %d start' % index)
p.close()
p.join()
testdatafeatures = {}
for sub_dict in result:
testdatafeatures.update(sub_dict.get())
save_pickle(testdatafeatures, os.path.join(TEST_FEATURE_DIR, 'testdatafeatures-withsetinfo-name-clean-2.pkl'))
def save_time(model):
test_unass = load_json(TEST_UNASS_PATH)
testdatafeatures = load_pickle(os.path.join(TEST_FEATURE_DIR, 'testdatafeatures-withsetinfo-name-clean-2.pkl'))
title_feature_df = pd.read_pickle(os.path.join(TEST_FEATURE_DIR, 'test-title-distance-df-name-clean-2.pkl'))
title_feature = title_feature_df.values
models_loaded = {
'model': load_pickle(model['model']),
'ss': load_pickle(model['ss']),
'cols': model['cols'],
'score': model['score'],
'name': model['name'],
}
print(models_loaded['name'])
model_result = {}
for pid_with_index in tqdm.tqdm(test_unass):
inner_dict = {}
candidate_aids = testdatafeatures[pid_with_index]['candidate-aids']
data = testdatafeatures[pid_with_index]['data']
data_length = len(candidate_aids)
title_data = title_feature[:data_length]
title_feature = title_feature[data_length:]
data = np.concatenate((data, title_data), axis=1)
default_cols = BASE_COLS + SET_INFO_COLS + TITLE_COLS
df = pd.DataFrame(data=data, columns=default_cols)
model = models_loaded['model']
ss = models_loaded['ss']
data = df[models_loaded['cols']].values
data = ss.transform(data)
output = model.predict_proba(data)
inner_dict['candidate-aids'] = candidate_aids
inner_dict['result-score'] = output
model_result[pid_with_index] = inner_dict
save_pickle(model_result, os.path.join(RESULT_SAVE_DIR, 'name.clean.2.%s.result.score.pkl' % models_loaded['name']))
def get_coauthor_count_for_enhence(aid_pid_pair, aid_author_info_dict, pid_info_dict):
index = int(aid_pid_pair[1].split('-')[1])
authors = pid_info_dict['authors']
authors = [clean_name(item['name']) for item in authors]
authors.pop(index)
count = 0
for author_name in authors:
if author_name in aid_author_info_dict.keys():
count += 1
return count
def get_org_score_for_enhence(aid_pid_pair, aid_org_year_list, pid_info_dict, my_stopwords, num_pattern, remove_punctuation, lemmatizer):
def clean_text(str_info):
str_lower = str_info.lower().strip()
result = str_lower.translate(remove_punctuation)
result = num_pattern.sub('', result)
tokens = word_tokenize(result)
result = [word for word in tokens if word not in my_stopwords]
result = [lemmatizer.lemmatize(word) for word in result]
result = [word for word in result if len(word) > 1]
return ' '.join(list(set(result)))
def get_score(corpus):
vectorizer = CountVectorizer()
org_vectors = vectorizer.fit_transform(corpus)
score = cosine_similarity(org_vectors[0], org_vectors[1])
return score.item()
index = int(aid_pid_pair[1].split('-')[1])
author = pid_info_dict['authors'][index]
org = author.get('org', '').lower()
if org == '':
return 0
aid_org = [pair[0] for pair in aid_org_year_list if pair[0] != '']
if aid_org == []:
return 0
aid_org_clean = list(map(clean_text, aid_org))
org_clean = clean_text(org)
corpus_list = [[org_clean, target_org] for target_org in aid_org_clean]
scores = list(map(get_score, corpus_list))
return np.max(scores)
def get_mean_final_score(model_result_paths):
print('Mean score begin ...')
test_unass = load_json(TEST_UNASS_PATH)
# aid2coauthor = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2coauthor.pkl'))
# test_pub = load_json(TEST_PUB_PATH)
# aid2orgwithyear = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2orgwithyear.pkl'))
# title_feature_df = pd.read_pickle(os.path.join(TEST_FEATURE_DIR, 'test-title-distance-df.pkl'))
# title_feature = title_feature_df.values
# org_text_process_dict = {
# 'my_stopwords': set(stopwords.words('english')),
# 'num_pattern': re.compile(r'\d+'),
# 'remove_punctuation': str.maketrans(string.punctuation, ' '*len(string.punctuation)),
# 'lemmatizer': WordNetLemmatizer(),
# }
result_dict_list = [load_pickle(path) for path in model_result_paths]
submission = defaultdict(list)
# count = 0
# problem_pids = []
for pid_with_index in tqdm.tqdm(test_unass):
candidate_aids = result_dict_list[0][pid_with_index]['candidate-aids']
inner_data = np.zeros((len(candidate_aids), len(result_dict_list)))
for num, result_dict in enumerate(result_dict_list):
data = result_dict[pid_with_index]['result-score']
inner_data[:, num] = data
final_output = np.mean(inner_data, axis=1)
predict_author = candidate_aids[np.argmax(final_output)]
submission[predict_author].append(pid_with_index.split('-')[0])
save_json(submission, os.path.join(FINAL_DIR, 'name-clean-2-mean-result-%d.json' % len(result_dict_list)))
# if np.max(final_output) < 0.5:
# print('-'*50)
# print(pid_with_index)
# if np.max(final_output) < 0.5:
# count += 1
# problem_pids.append(pid_with_index)
# coauthor_info = []
# for aid in candidate_aids:
# coauthor_info.append(get_coauthor_count_for_enhence((aid, pid_with_index), aid2coauthor[aid], test_pub[pid_with_index.split('-')[0]]))
# coauthor_info_array = np.array(coauthor_info)
# if np.max(coauthor_info_array) > 0:
# predict_author = candidate_aids[np.argmax(coauthor_info_array)]
# else:
# org_info = []
# for aid in candidate_aids:
# org_info.append(get_org_score_for_enhence((aid, pid_with_index), aid2orgwithyear[aid], test_pub[pid_with_index.split('-')[0]], **org_text_process_dict))
# org_info_array = np.array(org_info)
# if np.max(org_info_array) > 0.5:
# predict_author = candidate_aids[np.argmax(org_info_array)]
# else:
# predict_author = candidate_aids[np.argmax(final_output)]
# else:
# predict_author = candidate_aids[np.argmax(final_output)]
# submission[predict_author].append(pid_with_index.split('-')[0])
# save_json(problem_pids, os.path.join(FINAL_DIR, 'problem.pids.3.json'))
# save_json(submission, os.path.join(FINAL_DIR, 'name-clean-2-enhance-mean-result-%d.json' % len(result_dict_list)))
def check_name():
problem_pids = load_json(os.path.join(FINAL_DIR, 'problem.pids.3.json'))
name2aids = load_pickle(os.path.join(NEW_DATA_DIR, 'name2aids.pkl'))
test_pub = load_json(TEST_PUB_PATH)
all_authors_name = list(name2aids.keys())
name_map = []
for pid_with_index in tqdm.tqdm(problem_pids):
now_pid, index = pid_with_index.split('-')
author_name_no_clean = test_pub[now_pid]['authors'][int(index)]['name']
author_name = clean_name(author_name_no_clean)
if pid_with_index == 'ToCcabLT-1':
author_name = 'junliang_wang'
if pid_with_index == 'cVvvcFzj-1':
author_name = 'xiaojun_liu'
index = get_name_index(author_name, all_authors_name)
author_name_inlist = all_authors_name[index]
# if author_name_inlist != author_name:
name_map.append((pid_with_index, author_name_no_clean, author_name, author_name_inlist))
name_map = list(set(name_map))
print(len(name_map))
save_json(name_map, os.path.join(FINAL_DIR, 'name.different.3.json'))
def find_name(name):
name2aids = load_pickle(os.path.join(NEW_DATA_DIR, 'name2aids.pkl'))
aid2orgset = load_pickle(os.path.join(NEW_DATA_DIR, 'aid2orgset.pkl'))
all_authors_name = list(name2aids.keys())
for name_in_list in all_authors_name:
if name in name_in_list:
print('-'*50)
print(name_in_list)
for aid in name2aids[name_in_list]:
# print(aid2orgset[aid])
if 'beijing' in aid2orgset[aid]:
print(aid2orgset[aid])
if __name__ == "__main__":
models = [
{ # 0.85926333738039 original
'model': os.path.join(STACK_MODEL_DIR_v2, 'sm-191125-nosetinfo-extend3-sample11.pkl'),
'ss': os.path.join(STACK_MODEL_DIR_v2, 'standardscaler-last1year-nosetinfo-extend3-sample11.pkl'),
'cols': BASE_COLS,
'score': 0.85926333738039,
'name': 'sm-191125-nosetinfo-extend3-sample11.pkl',
},
{ # 0.858031834386063 with set info
'model': os.path.join(STACK_MODEL_DIR_v2, 'test-2-sm-191127-withsetinfo-sample11.pkl'),
'ss': os.path.join(STACK_MODEL_DIR_v2, 'standardscaler-last1year-withsetinfo-sample11.pkl'),
'cols': BASE_COLS + SET_INFO_COLS,
'score': 0.858031834386063,
'name': 'test-2-sm-191127-withsetinfo-sample11.pkl',
},
{ # 0.856180351089599 with set info
'model': os.path.join(STACK_MODEL_DIR_v2, 'sm-191127-withsetinfo-11.pkl'),
'ss': os.path.join(STACK_MODEL_DIR_v2, 'standardscaler-last1year-withsetinfo-11.pkl'),
'cols': BASE_COLS + SET_INFO_COLS,
'score': 0.856180351089599,
'name': 'sm-191127-withsetinfo-11.pkl',
},
{ # 0.855763586778158 with set info and title info
'model': os.path.join(STACK_MODEL_DIR_v2, 'sm-191128-withsetinfo-title-11-norf.pkl'),
'ss': os.path.join(STACK_MODEL_DIR_v2, 'standardscaler-last1year-withsetinfo-title-11.pkl'),
'cols': BASE_COLS + SET_INFO_COLS + TITLE_COLS,
'score': 0.855763586778158,
'name': 'sm-191128-withsetinfo-title-11-norf.pkl',
},
{ # 0.85364791527539
'model': os.path.join(STACK_MODEL_DIR_v2, 'sm-191126-withsetinfo-sample11.pkl'),
'ss': os.path.join(STACK_MODEL_DIR_v2, 'standardscaler-last1year-withsetinfo-sample11.pkl'),
'cols': BASE_COLS + SET_INFO_COLS,
'score': 0.85364791527539,
'name': 'sm-191126-withsetinfo-sample11.pkl',
},
{ # 0.855538436984147
'model': os.path.join(STACK_MODEL_DIR_v2, 'sm-191128-withsetinfo-title-11.pkl'),
'ss': os.path.join(STACK_MODEL_DIR_v2, 'standardscaler-last1year-withsetinfo-title-11.pkl'),
'cols': BASE_COLS + SET_INFO_COLS + TITLE_COLS,
'score': 0.855538436984147,
'name': 'sm-191128-withsetinfo-title-11.pkl',
},
]
# see_year_distribution()
process_size = 22
multi_gen_base_feature(process_size)
gen_title_feature()
for i in range(len(models)): # 建议手动多进程,节省时间
save_time(models[i]) # 无法代码直接多进程,会与模型的多进程冲突
# save_time(models[5])
model_result_paths = [
os.path.join(RESULT_SAVE_DIR, 'name.clean.2.sm-191125-nosetinfo-extend3-sample11.pkl.result.score.pkl'),
os.path.join(RESULT_SAVE_DIR, 'name.clean.2.sm-191126-withsetinfo-sample11.pkl.result.score.pkl'),
os.path.join(RESULT_SAVE_DIR, 'name.clean.2.sm-191127-withsetinfo-11.pkl.result.score.pkl'),
os.path.join(RESULT_SAVE_DIR, 'name.clean.2.sm-191128-withsetinfo-title-11-norf.pkl.result.score.pkl'),
os.path.join(RESULT_SAVE_DIR, 'name.clean.2.sm-191128-withsetinfo-title-11.pkl.result.score.pkl'),
os.path.join(RESULT_SAVE_DIR, 'name.clean.2.test-2-sm-191127-withsetinfo-sample11.pkl.result.score.pkl'),
]
get_mean_final_score(model_result_paths)
# check_name()