第二周第二课时

运行结果

#具体代码
from bs4 import BeautifulSoup
import requests
import pymongo
import time
client = pymongo.MongoClient('localhost', 27017)
phonenumber = client['phonenumber']
phonenumber_sheet = phonenumber['phonenumber_sheet']


def get_urls(page_num):
    web_url = 'http://bj.58.com/shoujihao/1/pn{}/'.format(str(page_num))
    web_data = requests.get(web_url)
    time.sleep(1)
    soup = BeautifulSoup(web_data.text, 'lxml')
    detail_urls = soup.select('li > a.t')
    titles = soup.select('a.t > strong')
    for detail_url, title in zip(detail_urls, titles):
        data = {
            'detail_url': detail_url.get('href'),
            'title': title.get_text()
        }
        phonenumber_sheet.insert_one(data)
        print(data)


for page in range(1, 12):
    get_urls(page)