爬取失信人员（只需输入姓名就能获取）

来源

代码：

 1 import requests
 2 import json
 3 import pandas as pd
 4 def listname(iname):
 5     name = []
 6     cardNum = []
 7     areaName = []
 8     for i in range(0, 100):
 9         url = 'https://sp0.baidu.com/8aQDcjqpAAV3otqbppnN2DJv/api.php'
10         para = {
11             'pn': str(i * 10),
12             'resource_id': '6899',
13             'query': '失信被执行人名单',
14             'iname': iname,
15             'rn': '10',
16             'ie': 'utf-8',
17             'oe': 'utf-8',
18             'format': 'json',
19             'cb': 'cb=jQuery110204669623007201096_1546820149157'
20         }
21         head = {
22             "Host": "sp0.baidu.com",
23             "Accept": "*/*",
24             "Connection": "keep-alive",
25             "Accept-Encoding": "gzip, deflate, br",
26             'Accept-Language': 'zh-CN,zh;q=0.8',
27             'Referer': 'https://www.baidu.com/s?wd=%E5%A4%B1%E4%BF%A1%E4%BA%BA',
28             'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/68.0.3440.75 Safari/537.36'
29         }
30         rep = requests.get(url, headers=head, params=para)
31         rep_str = rep.content.decode()
32         rep_json = json.loads(rep_str[7:-2])
33         rep_data = rep_json['data']
34         for each in rep_data:
35             for data in each['result']:
36                 name.append(data['iname'])
37                 cardNum.append(data['cardNum'])
38                 areaName.append(data['areaName'])
39     result_all = pd.DataFrame({'Name': iname, 'IDcard': cardNum, 'Area': areaName})
40     result_signal = result_all.drop_duplicates()
41     result = result_signal.reset_index(drop=True)
42     return result
43 print(listname('XXX'))

 

posted @ 2020-10-24 21:28  博二爷  阅读(452)  评论(0)  编辑  收藏  举报