雪球网数据爬取

 1 import requests
 2 import json
 3 import pymysql
 4 
 5 class mysql_conn(object):
 6     # 魔术方法, 初始化, 构造函数
 7     def __init__(self):
 8         self.db = pymysql.connect(host='127.0.0.1', user='root', password='abc123', port=3306, database='py1011')
 9         self.cursor = self.db.cursor()
10     # 执行modify(修改)相关的操作
11     def execute_modify_mysql(self, sql):
12         self.cursor.execute(sql)
13         self.db.commit()
14     # 魔术方法, 析构化 ,析构函数
15     def __del__(self):
16         self.cursor.close()
17         self.db.close()
18 
19 
20 headers = {
21     'Cookie': 'aliyungf_tc=AQAAALoQF3p02gsAUhVFebQ3uBBNZn+H; xq_a_token=584d0cf8d5a5a9809761f2244d8d272bac729ed4; xq_a_token.sig=x0gT9jm6qnwd-ddLu66T3A8KiVA; xq_r_token=98f278457fc4e1e5eb0846e36a7296e642b8138a; xq_r_token.sig=2Uxv_DgYTcCjz7qx4j570JpNHIs; _ga=GA1.2.516718356.1534295265; _gid=GA1.2.1050085592.1534295265; u=301534295266356; device_id=f5c21e143ce8060c74a2de7cbcddf0b8; Hm_lvt_1db88642e346389874251b5a1eded6e3=1534295265,1534295722; Hm_lpvt_1db88642e346389874251b5a1eded6e3=1534295722',
22     'User-Agent': 'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36',
23 }
24 url = 'https://xueqiu.com/v4/statuses/public_timeline_by_category.json?since_id=-1&max_id=-1&count=10&category=111'
25 
26 response = requests.get(url,headers=headers)
27 
28 res_dict = json.loads(response.text)
29 
30 list_lsit = res_dict['list']
31 
32 db ={}
33 for list_item_dict in list_lsit:
34     data_dict = json.loads(list_item_dict['data'])
35 
36     db['id'] = data_dict['id']
37     db['title'] = data_dict['title']
38     db['description'] = data_dict['description']
39     db['target'] = data_dict['target']
40     try:
41         sql = 'insert into xueqiu (uid,title,description,target) values ("{id}","{title}","{description}","{traget}")'.fromart(**db)
42         mc = mysql_conn()
43         mc.execute_modify_mysql(sql)
44     except:
45         pass

 

转载于:https://www.cnblogs.com/daihao9527/p/9484820.html