-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathDataSet.py
153 lines (133 loc) · 6.75 KB
/
DataSet.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
#!/usr/bin/env python
# encoding: utf-8
'''
@author: slade
@file: DataSet.py
@time: 2020/9/14 19:58
@desc:
'''
import pandas as pd
from collections import Counter
import numpy as np
import json
import gensim
from utils import gen_word_set, convert_word2id, convert_seq2bow, load_vocab
from Config import Config
from random import shuffle, choice
from tqdm import tqdm
class Dataset(object):
def __init__(self):
self.config = Config
self._sequenceLength = Config.sequenceLength # 每条输入的序列处理为定长
self._embeddingSize = Config.embeddingSize
self._batchSize = Config.batchSize
self._vocab_map = load_vocab(Config.vocab_path)
self._vocab_size = len(self._vocab_map)
def _genTrainEvalData(self, data_map, rate):
"""
生成训练集和验证集
"""
title, title_len, review_pos, review_pos_len, review_neg, review_neg_len = data_map['title'], data_map[
'title_len'], \
data_map['review_pos'], data_map[
'review_pos_len'], \
data_map['review_neg'], data_map[
'review_neg_len']
index = list(range(len(title)))
unzip_all_data = []
for idx in tqdm(index):
unzip_all_data.append([title[idx], review_pos[idx], 1])
for item in review_neg[idx]:
unzip_all_data.append([title[idx], item, 0])
trainIndex = int(len(unzip_all_data) * rate)
index = list(range(len(unzip_all_data)))
shuffle(index)
trainData = np.array(unzip_all_data)[index[:trainIndex]]
evalData = np.array(unzip_all_data)[index[trainIndex:]]
return trainData, evalData
def dataGen(self, file_path="./data/title_reviews.txt"):
"""
初始化训练集和验证集
"""
data_map = {'title': [], 'title_len': [], 'review_pos': [], 'review_pos_len': [], 'review_neg': [],
'review_neg_len': []}
with open(file_path, encoding='utf8') as f:
for line in tqdm(f.readlines()):
spline = line.strip().split('\t')
if len(spline) != 6:
continue
title, pos, neg1, neg2, neg3, neg4 = spline
negs = [neg1, neg2, neg3, neg4]
cur_arr, cur_len = [], []
# only 4 negative sample
# for each in negs:
# cur_arr.append(convert_word2id(each, self._vocab_map))
# each_len = len(each) if len(each) < Config.sequenceLength else Config.sequenceLength
# cur_len.append(each_len)
each = choice(negs)
cur_arr.append(convert_word2id(each, self._vocab_map))
each_len = len(each) if len(each) < Config.sequenceLength else Config.sequenceLength
cur_len.append(each_len)
data_map['title'].append(convert_word2id(title, self._vocab_map))
data_map['title_len'].append(
len(title) if len(title) < Config.sequenceLength else Config.sequenceLength)
data_map['review_pos'].append(convert_word2id(pos, self._vocab_map))
data_map['review_pos_len'].append(
len(pos) if len(pos) < Config.sequenceLength else Config.sequenceLength)
data_map['review_neg'].append(cur_arr)
data_map['review_neg_len'].append(cur_len)
return self._genTrainEvalData(data_map, Config.rate)
def _multiGenTrainEvalData(self, data_map, rate):
"""
生成训练集和验证集
"""
title, title_len, review_pos, review_pos_len, review_neg, review_neg_len = data_map['title'], data_map[
'title_len'], \
data_map['review_pos'], data_map[
'review_pos_len'], \
data_map['review_neg'], data_map[
'review_neg_len']
index = list(range(len(title)))
unzip_all_data = []
for idx in tqdm(index):
unzip_all_data.append([title[idx], review_pos[idx], 1])
for item in review_neg[idx]:
unzip_all_data.append([title[idx], item, 0])
trainIndex = int(len(unzip_all_data) * rate)
index = list(range(len(unzip_all_data)))
shuffle(index)
trainData = np.array(unzip_all_data)[index[:trainIndex]]
evalData = np.array(unzip_all_data)[index[trainIndex:]]
return trainData, evalData
def multiDataGen(self, file_path="./data/title_reviews.txt"):
"""
初始化训练集和验证集
"""
data_map = {'title': [], 'title_len': [], 'review_pos': [], 'review_pos_len': [], 'review_neg': [],
'review_neg_len': []}
with open(file_path, encoding='utf8') as f:
for line in tqdm(f.readlines()):
spline = line.strip().split('\t')
if len(spline) != 6:
continue
title, pos, neg1, neg2, neg3, neg4 = spline
negs = [neg1, neg2, neg3, neg4]
cur_arr, cur_len = [], []
# only 4 negative sample
for each in negs:
cur_arr.append(convert_word2id(each, self._vocab_map))
each_len = len(each) if len(each) < Config.sequenceLength else Config.sequenceLength
cur_len.append(each_len)
each = choice(negs)
cur_arr.append(convert_word2id(each, self._vocab_map))
each_len = len(each) if len(each) < Config.sequenceLength else Config.sequenceLength
cur_len.append(each_len)
data_map['title'].append(convert_word2id(title, self._vocab_map))
data_map['title_len'].append(
len(title) if len(title) < Config.sequenceLength else Config.sequenceLength)
data_map['review_pos'].append(convert_word2id(pos, self._vocab_map))
data_map['review_pos_len'].append(
len(pos) if len(pos) < Config.sequenceLength else Config.sequenceLength)
data_map['review_neg'].append(cur_arr)
data_map['review_neg_len'].append(cur_len)
return self._genTrainEvalData(data_map, Config.rate)