-
Notifications
You must be signed in to change notification settings - Fork 1
/
prepro_sentens.py
202 lines (181 loc) · 7.92 KB
/
prepro_sentens.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
import argparse
import json
import os
import numpy as np
from collections import Counter
from tqdm import tqdm
basic='file'
negative=open(basic+'/BingLiuList/negative-words.txt','r').read().split('\n')
positive=open(basic+'/BingLiuList/positive-words.txt','r').read().split('\n')
negation=open(basic+'/negation','r').read().split('\n')
pos_dic=json.load(open(basic+'/pos_dic.json','r'))
def get_args():
parser = argparse.ArgumentParser()
# home = os.path.expanduser("")
source_dir = os.path.join('db', "cut_test_on_300")
target_dir = os.path.join('prepro_data', "cut_test_on_300/")
train_negative_num=0
glove_vec_size=100
glove_dir = os.path.join(basic, "", "glove")
parser.add_argument('-s', "--source_dir", default=source_dir)
parser.add_argument('-n', "--train_negative_num",type=int, default=train_negative_num)
parser.add_argument('-t', "--target_dir", default=target_dir)
parser.add_argument("--train_name", default='train_data')
parser.add_argument("--train_ratio", default=0.9, type=int)
parser.add_argument("--glove_corpus", default="6B")
parser.add_argument("--glove_dir", default=glove_dir)
parser.add_argument("--glove_vec_size", default=glove_vec_size, type=int)
return parser.parse_args()
def main():
args = get_args()
prepro(args)
def prepro(args):
prepro_each(args, 'val_train_val', out_name='val_train_val')
prepro_each(args, 'val_val', out_name='val_val')
prepro_each(args, 'val_train_train', out_name='val_train_train')
# prepro_together(args,['train','val'],out_name='train')
def get_word2vec(args, word_counter):
glove_path = os.path.join(args.glove_dir, "glove.{}.{}d.txt".format(args.glove_corpus, args.glove_vec_size))
sizes = {'6B': int(4e5), '42B': int(1.9e6), '840B': int(2.2e6), '2B': int(1.2e6)}
total = sizes[args.glove_corpus]
word2vec_dict = {}
with open(glove_path, 'r', encoding='utf-8') as fh:
for line in tqdm(fh, total=total):
array = line.lstrip().rstrip().split(" ")
word = array[0]
vector = list(map(float, array[1:]))
if word in word_counter:
word2vec_dict[word] = vector
elif word.capitalize() in word_counter:
word2vec_dict[word.capitalize()] = vector
elif word.lower() in word_counter:
word2vec_dict[word.lower()] = vector
elif word.upper() in word_counter:
word2vec_dict[word.upper()] = vector
print("{}/{} of word vocab have corresponding vectors in {}".format(len(word2vec_dict), len(word_counter), glove_path))
return word2vec_dict
def save(args, data, shared, data_type):
data_path = os.path.join(args.target_dir, "data_{}.json".format(data_type))
shared_path = os.path.join(args.target_dir, "shared_{}_{}.json".format(data_type,args.glove_vec_size))
json.dump(data, open(data_path, 'w'))
json.dump(shared, open(shared_path, 'w'))
def prepro_each(args, data_type, start_ratio=0.0, stop_ratio=1.0, out_name="default", in_path=None):
haoruopeng_feature = np.load(basic + '/haoruoprng_cut_test_on_300/{}.npy'.format(data_type)).tolist()
source_path=os.path.join(args.source_dir,'{}.json'.format(data_type))
dataset=json.load(open(source_path, 'r'))
if data_type == 'train':
candidtate_num = args.train_negative_num + 1
context_move=1
else:
candidtate_num = 2
context_move = 0
import nltk
sent_tokenize = nltk.sent_tokenize
def word_tokenize(tokens):
return [token.replace("''", '"').replace("``", '"').replace('-',' ') for token in nltk.word_tokenize(tokens)]
q, cq,rx, rcx= [], [],[],[]
q_pos, q_sem, x_pos, x_sem,q_neg,x_neg = [], [], [], [], [], []
x, cx = [], []
answerss = []
p = []
word_counter, char_counter, lower_word_counter = Counter(), Counter(), Counter()
def procee_context(context):
return context.replace("``", ' ').replace('-',' ').replace("''", ' ').replace(' ',' ')
for aii, article in enumerate(tqdm(dataset)):
sents=[]
for i in range(1,5):
sents.append(procee_context(article[i+context_move]))
assert len(sents)==4
sents_words=list(map(word_tokenize, sents))
sents_words_pos= list(map (nltk.pos_tag,sents_words))
x_semi = []
x_negi = []
x_posi = []
for i,s in enumerate(sents_words):
x_semii = []
x_negii = []
x_posii = []
for j,word in enumerate(s):
if word.lower() in negative:
x_semii.append(-1)
elif word in positive:
x_semii.append(1)
else:
x_semii.append(0)
if word.lower() in negation:
x_negii.append(1)
else:
x_negii.append(0)
x_posii.append(pos_dic[sents_words_pos[i][j][1]])
x_semi.append(x_semii)
x_negi.append(x_negii)
x_posi.append(x_posii)
# xi = [process_tokens(tokens) for tokens in xi]
cxi = [[list(xijk) for xijk in xij] for xij in sents_words]
x.append(sents_words)
cx.append(cxi)
for xij in sents_words:
for xijk in xij:
word_counter[xijk] += candidtate_num
lower_word_counter[xijk.lower()] += candidtate_num
for xijkl in xijk:
char_counter[xijkl] += candidtate_num
rxi=[aii]
if data_type == 'train':
right_id=0
else:
right_id=int(article[7])-1
for q_id in range(candidtate_num):
p.append(article[0])
ending = article[q_id+context_move+5]
ending = ending.replace("''", ' ')
ending = ending.replace("``", ' ').replace('-', ' ').replace(' ', ' ')
qi=word_tokenize(ending)
# q_posi = [p[1] for p in nltk.pos_tag(qi)]
q_posi = [pos_dic[p[1]] for p in nltk.pos_tag(qi)]
q_semi=[]
q_negi=[]
for ai in qi:
if ai.lower() in negative:
q_semi.append(-1)
elif ai in positive:
q_semi.append(1)
else:
q_semi.append(0)
if ai.lower() in negation:
q_negi.append(1)
else:
q_negi.append(0)
# qi = process_tokens(qi)
cqi = [list(qij) for qij in qi]
# answer
if q_id==right_id:
answer=1
else :answer=0
answerss.append(answer)
for qij in qi:
word_counter[qij] += 1
lower_word_counter[qij.lower()] += 1
for qijk in qij:
char_counter[qijk] += 1
q.append(qi)
cq.append(cqi)
rx.append(rxi)
rcx.append(rxi)
q_pos.append(q_posi)
x_pos.append(x_posi)
x_sem.append(x_semi)
q_sem.append(q_semi)
x_neg.append(x_negi)
q_neg.append(q_negi)
word2vec_dict = get_word2vec(args, word_counter)
lower_word2vec_dict = get_word2vec(args, lower_word_counter)
data = {'q': q, 'cq': cq, 'answerss': answerss,'*x': rx, '*cx': rcx, 'q_pos':q_pos,\
'q_neg':q_neg,'q_sem':q_sem,'x_pos':x_pos,'x_sem':x_sem,'x_neg':x_neg,'haoruopeng_feature':haoruopeng_feature,'p': p,}
shared = {'x': x, 'cx': cx,
'word_counter': word_counter, 'char_counter': char_counter, 'lower_word_counter': lower_word_counter,
'word2vec': word2vec_dict, 'lower_word2vec': lower_word2vec_dict}
print("saving ...")
save(args, data, shared, out_name)
if __name__ == "__main__":
main()