-
Notifications
You must be signed in to change notification settings - Fork 46
/
process_raw.py
200 lines (167 loc) · 9.99 KB
/
process_raw.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
# -*- coding: utf-8 -*-
"""
@author: alexyang
@contact: alex.yang0326@gmail.com
@file: process_raw.py
@time: 2019/1/5 17:05
@desc: process raw data
"""
import os
import codecs
import pandas as pd
import xml.etree.ElementTree as ET
from sklearn.model_selection import train_test_split
def process_xml(file_path, is_train_file, save_folder):
content_term, aspect_term, sentiment_term, start, end = list(), list(), list(), list(), list() # data for aspect term
content_cate, aspect_cate, sentiment_cate = list(), list(), list() # data for aspect category
polarity = {'negative': 0, 'neutral': 1, 'positive': 2}
tree = ET.parse(file_path)
root = tree.getroot()
for sentence in root:
text = sentence.find('text').text.lower()
for asp_terms in sentence.iter('aspectTerms'):
for asp_term in asp_terms.iter('aspectTerm'):
if asp_term.get('polarity') in polarity:
_text = text
_start = int(asp_term.get('from'))
_end = int(asp_term.get('to'))
_aspect = asp_term.get('term').lower()
_sentiment = polarity[asp_term.get('polarity')]
if _start > 0 and text[_start - 1] != ' ':
_text = text[:_start] + ' ' + text[_start:]
_start += 1
_end += 1
if _end < len(_text) and _text[_end] != ' ':
_text = _text[:_end] + ' ' + _text[_end:]
if _text[_start:_end] != _aspect:
raise Exception('{}=={}=={}'.format(_text, _text[_start:_end], _aspect))
content_term.append(_text)
aspect_term.append(_aspect)
sentiment_term.append(_sentiment)
start.append(_start)
end.append(_end)
for asp_cates in sentence.iter('aspectCategories'):
for asp_cate in asp_cates.iter('aspectCategory'):
if asp_cate.get('polarity') in polarity:
content_cate.append(text)
aspect_cate.append(asp_cate.get('category'))
sentiment_cate.append(polarity[asp_cate.get('polarity')])
if not os.path.exists(os.path.join(save_folder, 'term')):
os.makedirs(os.path.join(save_folder, 'term'))
if not is_train_file:
test_data = {'content': content_term, 'aspect': aspect_term, 'sentiment': sentiment_term,
'from': start, 'to': end}
test_data = pd.DataFrame(test_data, columns=test_data.keys())
test_data.to_csv(os.path.join(save_folder, 'term/test.csv'), index=None)
else:
train_content, valid_content, train_aspect, valid_aspect, train_senti, valid_senti, train_start, valid_start, \
train_end, valid_end = train_test_split(content_term, aspect_term, sentiment_term, start, end, test_size=0.1)
train_data = {'content': train_content, 'aspect': train_aspect, 'sentiment': train_senti,
'from': train_start, 'to': train_end}
train_data = pd.DataFrame(train_data, columns=train_data.keys())
train_data.to_csv(os.path.join(save_folder, 'term/train.csv'), index=None)
valid_data = {'content': valid_content, 'aspect': valid_aspect, 'sentiment': valid_senti,
'from': valid_start, 'to': valid_end}
valid_data = pd.DataFrame(valid_data, columns=valid_data.keys())
valid_data.to_csv(os.path.join(save_folder, 'term/valid.csv'), index=None)
if len(content_cate) > 0:
if not os.path.exists(os.path.join(save_folder, 'category')):
os.makedirs(os.path.join(save_folder, 'category'))
if not is_train_file:
test_data = {'content': content_cate, 'aspect': aspect_cate, 'sentiment': sentiment_cate}
test_data = pd.DataFrame(test_data, columns=test_data.keys())
test_data.to_csv(os.path.join(save_folder, 'category/test.csv'), index=None)
else:
train_content, valid_content, train_aspect, valid_aspect, \
train_senti, valid_senti = train_test_split(content_cate, aspect_cate, sentiment_cate, test_size=0.1)
train_data = {'content': train_content, 'aspect': train_aspect, 'sentiment': train_senti}
train_data = pd.DataFrame(train_data, columns=train_data.keys())
train_data.to_csv(os.path.join(save_folder, 'category/train.csv'), index=None)
valid_data = {'content': valid_content, 'aspect': valid_aspect, 'sentiment': valid_senti}
valid_data = pd.DataFrame(valid_data, columns=valid_data.keys())
valid_data.to_csv(os.path.join(save_folder, 'category/valid.csv'), index=None)
def process_twitter(file_path, is_train_file, save_folder):
polarity = {'-1': 0, '0': 1, '1': 2}
content, aspect, sentiment, start, end = list(), list(), list(), list(), list()
with codecs.open(file_path, 'r', encoding='utf8')as reader:
lines = reader.readlines()
for i in range(0, len(lines), 3):
_content = lines[i].strip().lower()
_aspect = lines[i+1].strip().lower()
_sentiment = lines[i+2].strip().lower()
_start = _content.find('$t$')
_end = _start + len(_aspect)
content.append(_content.replace('$t$', _aspect))
aspect.append(_aspect)
sentiment.append(polarity[_sentiment])
start.append(_start)
end.append(_end)
if not os.path.exists(save_folder):
os.makedirs(save_folder)
if not is_train_file:
test_data = {'content': content, 'aspect': aspect, 'sentiment': sentiment, 'from': start, 'to': end}
test_data = pd.DataFrame(test_data, columns=test_data.keys())
test_data.to_csv(os.path.join(save_folder, 'test.csv'), index=None)
else:
train_content, valid_content, train_aspect, valid_aspect, train_senti, valid_senti, train_start, valid_start, \
train_end, valid_end = train_test_split(content, aspect, sentiment, start, end, test_size=0.1)
train_data = {'content': train_content, 'aspect': train_aspect, 'sentiment': train_senti,
'from': train_start, 'to': train_end}
train_data = pd.DataFrame(train_data, columns=train_data.keys())
train_data.to_csv(os.path.join(save_folder, 'train.csv'), index=None)
valid_data = {'content': valid_content, 'aspect': valid_aspect, 'sentiment': valid_senti,
'from': valid_start, 'to': valid_end}
valid_data = pd.DataFrame(valid_data, columns=valid_data.keys())
valid_data.to_csv(os.path.join(save_folder, 'valid.csv'), index=None)
def process_fsauor(file_path, save_path):
folder = os.path.dirname(save_path)
if not os.path.exists(folder):
os.makedirs(folder)
aspect_names = ['location_traffic_convenience', 'location_distance_from_business_district',
'location_easy_to_find', 'service_wait_time', 'service_waiters_attitude',
'service_parking_convenience', 'service_serving_speed', 'price_level',
'price_cost_effective', 'price_discount', 'environment_decoration', 'environment_noise',
'environment_space', 'environment_cleaness', 'dish_portion', 'dish_taste', 'dish_look',
'dish_recommendation', 'others_overall_experience', 'others_willing_to_consume_again']
polarity = {-1: 0, 0: 1, 1: 2}
content, aspect, sentiment = list(), list(), list()
raw_data = pd.read_csv(file_path, header=0, index_col=0)
for _, row in raw_data.iterrows():
text = row['content']
for aspect_name in aspect_names:
if row[aspect_name] != -2:
content.append(text)
aspect.append(aspect_names)
sentiment.append(polarity[row[aspect_name]])
csv_data = {'content': content, 'aspect': aspect, 'sentiment': sentiment}
csv_data = pd.DataFrame(csv_data, columns=csv_data.keys())
csv_data.to_csv(save_path, index=0)
def process_bdci(file_path, is_train_file, save_folder):
if not os.path.exists(save_folder):
os.makedirs(save_folder)
raw_data = pd.read_csv(file_path, header=0, index_col=0)
if is_train_file:
polarity = {-1: 0, 0: 1, 1: 2}
raw_data = raw_data[['content', 'subject', 'sentiment_value']]
raw_data['sentiment_value'].replace(polarity, inplace=True)
raw_data.rename(columns={'subject': 'aspect', 'sentiment_value': 'sentiment'})
train_data, valid_data = train_test_split(raw_data, test_size=0.1)
train_data.to_csv(os.path.join(save_folder, 'train.csv'), index=None)
valid_data.to_csv(os.path.join(save_folder, 'valid.csv'), index=None)
else:
raw_data[['content']].to_csv(os.path.join(save_folder, 'test.csv'), index=None)
if __name__ == '__main__':
process_xml('./raw_data/semeval14_laptop/Laptop_Train_v2.xml', is_train_file=True, save_folder='./data/laptop')
process_xml('./raw_data/semeval14_laptop/Laptops_Test_Gold.xml', is_train_file=False, save_folder='./data/laptop')
process_xml('./raw_data/semeval14_restaurant/Restaurants_Train_v2.xml', is_train_file=True,
save_folder='./data/restaurant')
process_xml('./raw_data/semeval14_restaurant/Restaurants_Test_Gold.xml', is_train_file=False,
save_folder='./data/restaurant')
process_twitter('./raw_data/twitter/train.txt', is_train_file=True, save_folder='./data/twitter')
process_twitter('./raw_data/twitter/test.txt', is_train_file=False, save_folder='./data/twitter')
# process_fsauor('./raw_data/fsauor2018/train.csv', save_path='./data/fsauor/train.csv')
# process_fsauor('./raw_data/fsauor2018/valid.csv', save_path='./data/fsauor/valid.csv')
# process_fsauor('./raw_data/fsauor2018/test.csv', save_path='./data/fsauor/test.csv')
#
# process_bdci('./raw_data/bdci18_car/train.csv', is_train_file=True, save_folder='./data/bdci')
# process_bdci('./raw_data/bdci18_car/test.csv', is_train_file=True, save_folder='./data/bdci')