realcodespider

import requests
import re
from bs4 import BeautifulSoup
class :
    def __init__(self):
        pass

    def deals (self, html):

        soup = BeautifulSoup(html,'html.parser');
        #print (soup.a.select('.aw-user-name'))
        print (soup.prettify());
        for i in soup.find_all('div',class_ = "mod-head"):
            try :
                if i.div.p.a.get_text() == "钟俊威":
                    print(i.div.p.a.get_text() + i.find_next_sibling().div.get_text())
            except:
                pass
    def load(self, url):
        print(url);
        headers = {
            "Connection":"keep-alive",
            "User-Agent":"Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:61.0) Gecko/20100101 Firefox/61.0"
        }
        #request = requests.Request(url=url,headers=headers);

        try:
            response = requests.get(url,headers = headers)
            html = response.text;
        except:
            print("No");
        self.deals(html);
    def main(self):
        url = "https://www.ctguqmx.com/question/"
        for i in range(300 , 400):
            self.load("https://www.ctguqmx.com")
             self.load(url + str(i));
if __name__ == '__main__':
    spider = Spider();
    spider.main();