爬取失信人

爬取失信人信息

import requests as r
import MySQLdb
conn = MySQLdb.connect(
    host = "****",
    user = "root",
    password = "****",
    port = 3306,
    charset = "utf8",
    db = "spider"
)
cursor = conn.cursor()

class TrustBreaker:

    '''
        人名  iname
        身份证号 cardNum
        执行法院  gistUnit
        省份  areaNameNew
        案号  gistId
        生效法律文书确定的义务  duty
        被执行人的履行情况   performance
        失信被执行人行为具体情形  disruptTypeName
        发布时间    publishDate
    '''
    @classmethod
    def get_breaker(self):
        url = "https://sp0.baidu.com/8aQDcjqpAAV3otqbppnN2DJv/api.php?"
        headers = {
            'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/78.0.3904.17 Safari/537.36'
        }
        # p = 0
        # while 1:
        #     page = p * 10
        #     p += 1
        for i in range(0,50):
            page = i*10
            params = {
                'resource_id': 6899,
                'query': '失信人',
                'pn': str(page),
                'rn': str(10),
                'ie': 'utf-8',
                'oe': 'utf-8'
            }
            res = r.request(method="GET",url=url,headers=headers,params=params).json()
            data_list = res["data"]
            for i in data_list:
                for j in i["result"]:
                    iname = j["iname"]
                    print(iname)
                    cardnum = j["cardNum"]
                    gistunit = j["gistUnit"]
                    areanamenew = j["areaNameNew"]
                    gistid = j["gistId"]
                    duty = j["duty"]
                    performance = j["performance"]
                    disrupttypename = j["disruptTypeName"]
                    publishdate = j["publishDate"]
                    sql = 'insert into breaker_spider(iname,cardnum,gistunit,areanamenew,gistid,duty,performance,disrupttypename,publishdate) values(%s,%s,%s,%s,%s,%s,%s,%s,%s) '
                    cursor.execute(sql,(iname,cardnum,gistunit,areanamenew,gistid,duty,performance,disrupttypename,publishdate))
                    conn.commit()
        cursor.close()
        conn.close()

if __name__ == '__main__':
    TrustBreaker.get_breaker()