-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathalign.py
150 lines (130 loc) · 4.31 KB
/
align.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
import os
import nltk
import copy
import random
import time
random.seed(1)
filenames = []
cnn_filenames = os.listdir(r'./cnn_stories_tokenized/')
for cnn_name in cnn_filenames:
filenames.append(r'./cnn_stories_tokenized/'+cnn_name)
dm_filenames = os.listdir(r'./dm_stories_tokenized/')
for dm_name in dm_filenames:
filenames.append(r'./dm_stories_tokenized/'+dm_name)
random.shuffle(filenames)
print("Total files:")
print(len(filenames))
print("")
file1 = open(r'complex.txt', 'w')
file2 = open(r'simple.txt', 'w')
from sentence_transformers import SentenceTransformer, util
model = SentenceTransformer('all-mpnet-base-v2', device='cuda')
model.max_seq_length = 128
start = time.time()
num_bigger_than_0_8 = 0
num_smaller_than_0_6 = 0
other = 0
total_src_sentences = 0
total_tgt_sentences = 0
def takeSecond(elem):
return elem[1]
for idx, name in enumerate(filenames):
if idx % 100000 == 0:
print(idx)
file = open(name, 'r')
src = []
tgt = []
lines = file.readlines()
signal = False
for line in lines:
if line.strip() == "":
continue
if line.strip() == "@highlight":
signal = True
continue
if signal:
tgt.append(line.strip())
signal = False
continue
src_sentences = nltk.sent_tokenize(line.strip())
for src_sentence in src_sentences:
src.append(src_sentence)
total_src_sentences += len(src)
total_tgt_sentences += len(tgt)
file.close()
if len(src) == 0 or len(tgt) == 0:
continue
embeddings1 = model.encode(src, convert_to_tensor=True)
embeddings2 = model.encode(tgt, convert_to_tensor=True)
cosine_scores = util.pytorch_cos_sim(embeddings1, embeddings2)
for j in range(len(tgt)):
str_src = ""
str_tgt = tgt[j]
tmp_dict = []
for i in range(len(src)):
tmp_dict.append([src[i], cosine_scores[i][j].item(), i])
tmp_dict.sort(key=takeSecond, reverse=True)
if tmp_dict[0][1] >= 0.8:
str_src = str_src + tmp_dict[0][0]
num_bigger_than_0_8 += 1
elif tmp_dict[0][1] < 0.6:
num_smaller_than_0_6 += 1
continue
else:
other += 1
str_src_list = [tmp_dict[0][2]]
tmp_str_src_list = [tmp_dict[0][2]]
for i in range(1, len(tmp_dict)):
tmp_str_src_list.append(tmp_dict[i][2])
tmp_str_src_list.sort()
tmp_str_src = ""
for element in tmp_str_src_list:
tmp_str_src = tmp_str_src + src[element] + " "
# print(tmp_str_src)
tmp_embeddings1 = model.encode(tmp_str_src, convert_to_tensor=True)
tmp_embeddings2 = model.encode(str_tgt, convert_to_tensor=True)
tmp_cosine_scores = util.pytorch_cos_sim(tmp_embeddings1, tmp_embeddings2)
# print(tmp_cosine_scores[0][0].item())
# print("")
if tmp_cosine_scores[0][0].item() < 0.7 or len(str_src_list) >= 2:
# print("result:")
for element in str_src_list:
str_src = str_src + src[element] + " "
# print(str_src)
break
else:
str_src_list = copy.deepcopy(tmp_str_src_list)
print(str_src.strip())
print(str_tgt.strip())
file1.write(str_src.strip())
file1.write("\n")
file2.write(str_tgt.strip())
file2.write("\n")
# for j in range(len(tgt)):
# signal = False
# str_src = ""
# str_tgt = tgt[j]
# for i in range(len(src)):
# if cosine_scores[i][j] >= 0.6:
# signal = True
# str_src = str_src + src[i] + " "
# # print("")
# # print(src[i])
# # print(tgt[j])
# # print(cosine_scores[i][j])
# # print("")
# if not signal:
# continue
# file1.write(str_src)
# file1.write("\n")
# file2.write(str_tgt)
# file2.write("\n")
end = time.time()
print("CPU time")
print(end-start)
print("total_src_sentences")
print(total_src_sentences)
print("total_tgt_sentences")
print(total_tgt_sentences)
file1.close()
file2.close()