-
Notifications
You must be signed in to change notification settings - Fork 0
/
data_process.py
292 lines (265 loc) · 11.1 KB
/
data_process.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
# import tools
import os
import re
import json
import pandas as pd
from lxml import etree
from tqdm import tqdm
import settings
# search_pages_dir = "C:/Users/YL/Desktop/googleSearch/"
search_pages_dir = "/home/zfj/research-data/user_profiling/googleSearch/"
title_list2Id={ "Other(其他)": 0,
"B-Professor(教授)": 1,
"I-Professor(教授)": 2,
"B-Researcher(研究员)": 3,
"I-Researcher(研究员)": 4,
"B-Associate Professor(副教授)": 5,
"I-Associate Professor(副教授)": 6,
"B-Assistant Professor(助理教授)": 7,
"I-Assistant Professor(助理教授)": 8,
"B-Professorate Senior Engineer(教授级高级工程师)": 9,
"I-Professorate Senior Engineer(教授级高级工程师)": 10,
"B-Engineer(工程师)": 11,
"I-Engineer(工程师)": 12,
"B-Lecturer(讲师)": 13,
"I-Lecturer(讲师)": 14,
"B-Senior Engineer(高级工程师)": 15,
"I-Senior Engineer(高级工程师)": 16,
"B-Ph.D(博士生)": 17,
"I-Ph.D(博士生)": 18,
"B-Associate Researcher(副研究员)": 19,
"I-Associate Researcher(副研究员)": 20,
"B-Assistant Researcher(助理研究员)": 21,
"I-Assistant Researcher(助理研究员)": 22,
"B-Student(学生)": 23,
"I-Student(学生)": 24
}
def load_data(file):
lines = open(file,encoding='utf-8').readlines()
data=[]
for l in lines:
d = json.loads(l)
data.append(d)
return data
def r_excel_list(path):
result=[]
datas = pd.read_excel(path)
for i, row in datas.iterrows():
result.append(dict(row))
return result
def extract_google_page(file):
html = open(file,encoding='utf-8').read()
html = etree.HTML(html)
link_list =[]
items = html.xpath('//div[@class="ZINbbc xpd O9g5cc uUPGi"]')
for item in items:
try:
href = item.xpath('.//a')[0].attrib['href']
start = href.find('http')
end = href.find('&sa')
href = href[start:end]
title = item.xpath('.//div[@class="BNeawe vvjwJb AP7Wnd"]')[0].text
content = item.xpath('.//div[@class="BNeawe s3v9rd AP7Wnd"]')[1].text
link ={'href':href,'title':title,'content':content}
link_list.append(link)
except:
pass
return link_list
#加载给定id的两个搜索列表页，结果合并在一起
def get_search_list(id):
link_list = []
# file = '/DATA/disk1/model_data/wll_data/kaiyu/ccks_numberone/CCKS2021_Aminer_profiling_googlesearch/data/' + id+ '_s1.html'
file = search_pages_dir + id + '_s1.html'
# print("file s1 exists", os.path.exists(file))
if os.path.exists(file):
link_list = extract_google_page(file)
# file = '/DATA/disk1/model_data/wll_data/kaiyu/ccks_numberone/CCKS2021_Aminer_profiling_googlesearch/data/' + id+ '_s2.html'
file = search_pages_dir + id+ '_s2.html'
if os.path.exists(file):
link_list.extend(extract_google_page(file))
# print("link list", link_list)
return link_list
# 生成性别验证集dev
def get_gender_data(file):
data = r_excel_list(file)
json_result = []
for d in tqdm(data):
id = d['id']
name = d['name']
org = d['org']
gender = d['gender']
link_list = get_search_list(id)
text = name + '[SEP]' + org + '; ' # + '; '
if link_list:
for link in link_list:
if link['content']:
text += link['content'] + '; '
json_result.append({
'id': id,
'name': name,
'org': org,
'gender': gender,
'text': text
})
print(len(json_result))
with open("data/gender_dev.json", 'w', encoding="utf-8") as f:
json.dump(json_result, f, ensure_ascii=False, indent=4)
def contain_words(text,words): #判断文本中是否包含一组词语的至少一个
text = text.lower()
for word in words:
word = word.lower()
if text.find(word)>=0:
return text
return ''
# 生成职称验证集dev
def get_title_data(file):
data = r_excel_list(file)
json_result = []
words = { 'Professor', 'Researcher', 'Engineer ','Lecturer', 'Ph.D', 'Research ', 'Student'}
for d in tqdm(data):
id = d['id']
name = d['name']
org = d['org']
title = d['title']
link_list = get_search_list(id)
text = name + '; ' + org + '; '
if link_list:
for link in link_list:
if link['content']:
word = contain_words(link['content'], words)
text += word + '; '
text = text.replace('\n', ' ').replace('\t', ' ')
json_result.append({
'id': id,
'name': name,
'org': org,
'title': title,
'text': text
})
print(len(json_result))
with open("data/dev.json", 'w', encoding="utf-8") as f:
json.dump(json_result, f, ensure_ascii=False, indent=4)
# 生成测试集（性别、职称）test
def get_gender_test():
data = r_excel_list(r'data/raw/new_test.xlsx')
result = {}
json_result = []
words = {'Professor', 'Researcher', 'Engineer ', 'Lecturer', 'Ph.D', 'Research ', 'Student'}
for d in tqdm(data):
id = d['id']
result[id] = d
result[id]['id'] = id
result[id]['name'] = d['name']
result[id]['org'] = d['org']
result[id]['gender'] = ''
result[id]['title'] = ''
link_list = get_search_list(id)
result[id]['gender_text'] = d['name'] + '[SEP]' + d['org'] + '; ' # + '; '
result[id]['title_text'] = d['name'] + '; ' + d['org'] + '; '
if link_list:
for link in link_list:
if link['content']:
result[id]['gender_text'] += link['content'] + '; '
word = contain_words(link['content'], words)
result[id]['title_text'] += word + '; '
result[id]['title_text'] = result[id]['title_text'].replace('\n', ' ').replace('\t', ' ')
for key in ['gender', 'title']: # ['homepage','lang','gender','title']
homepage_data = load_data('data/raw/ground_truth.json')
for d in homepage_data:
result[d['id']][key] = d[key]
for id in result:
title_data = {"id": result[id]['id'], "name": result[id]['name'], "org": result[id]['org'], "gender": result[id]['gender'], "title": result[id]['title'],
"gender_text": result[id]['gender_text'], "title_text": result[id]['title_text']}
json_result.append(title_data)
with open("data/test.json", 'w', encoding="utf-8") as f:
json.dump(json_result, f, ensure_ascii=False, indent=4)
# 生成训练数据（性别、职称）train
def get_train_data():
json_result = []
data = r_excel_list(r'data/raw/new_train.xlsx')
words = {'Professor', 'Researcher', 'Engineer ', 'Lecturer', 'Ph.D', 'Research ', 'Student'}
for d in tqdm(data):
id = d['id']
name = d['name']
org = d['org']
gender=d['gender']
title = d['title']
link_list = get_search_list(id)
gender_text = name+'[SEP]'+org
title_text = name + '; ' + org + '; '
if link_list:
for link in link_list:
if link['content']:
gender_text += link['content'] + '; '
t = contain_words(link['content'], words)
title_text += t + '; '
gender_text = gender_text[:240].replace("\n", " ").replace("\r", " ").replace("\t", " ")
title_text = title_text.replace('\n', ' ').replace('\t', ' ')
title_data = {"id": id, "name": name, "org": org, "gender": gender, "title": title,
"gender_text": gender_text, "title_text": title_text}
json_result.append(title_data)
print(len(json_result))
with open("data/train.json", 'w', encoding="utf-8") as f:
json.dump(json_result, f, ensure_ascii=False, indent=4)
def find_head_idx(source, target):
head_idx=[]
target_len = len(target)
for i in range(len(source)):
if source[i: i + target_len] == target:
head_idx.append(i)
return head_idx
# 生成职称训练数据（用于序列标注）(train dev test)
def get_title_tag(tokenizer = None):
'''
生成职称文本对应序列标注
:param tokenizer:
:return:
'''
train_data = json.load(open('data/dev.json')) # 使用上方处理后的json文件
title2tis = json.load(open(r'data/title2tis.json'))
# for data in train_data:
# Researcher(研究员) 和 Research(研究员) 合并
result = []
title_list = ["Professor(教授)","Ph.D(博士生)","Associate Professor(副教授)","Assistant Professor(助理教授)",
"Engineer(工程师)","Senior Engineer(高级工程师)","Professorate Senior Engineer(教授级高级工程师)","Lecturer(讲师)",
"Researcher(研究员)","Associate Researcher(副研究员)","Assistant Researcher(助理研究员)","Student(学生)"]
total={}
for data in tqdm(train_data):
# text_lists = []
# text_tags = []
text_list = re.sub('[\u4e00-\u9fa5]', '', data['text']).split("; ") # 去除字符串之中的中文
# text_list = data['text'].split("; ")
for text in text_list:
if text != '':
text_tokens = tokenizer.tokenize(text)
if len(text_tokens) < 10: continue # 过滤太短的句子
# text_lists.append(text)
label = [0] * len(text_tokens)
if data['title'] != "Other(其他)":
for ti_num in title_list:
for tis in title2tis[ti_num]:
title = tokenizer.tokenize(tis)
title_head_idx = find_head_idx(source=text_tokens, target=title) # 返回主语在text_tokens第一个位置的下标
if len(title_head_idx) > 0:
for idx in title_head_idx:
label[idx] = title_list2Id[r'B-' + ti_num]
if len(title) > 1:
label[idx + 1:idx + len(title)] = (title_list2Id[r'I-' + ti_num],) * (len(title) - 1)
if label == [0] * len(text_tokens):
continue
title_data = {"id": data["id"], "title": data["title"], "text": text, "text_tags": label}
result.append(title_data)
print(len(result))
with open("data/title_dev.json", 'w', encoding="utf-8") as f:
json.dump(result, f, ensure_ascii=False, indent=4)
if __name__ == '__main__':
# get_gender_data(r'data/raw/new_dev.xlsx')
# get_title_data(r'data/raw/new_dev.xlsx')
# get_gender_test()
# get_train_data()
from pathlib import Path
from transformers import BertTokenizer
root_path = Path(os.path.abspath(os.path.dirname(__file__)))
bert_model_dir = root_path / 'pretrain_models/bert_base_cased'
tokenizer = BertTokenizer(vocab_file=os.path.join(bert_model_dir, 'vocab.txt'), do_lower_case=True) # 大小写不敏感
get_title_tag(tokenizer)