标签:爬虫
用python爬虫算一下demo大师网站的总创收。。。。。。


#!/usr/bin/env python
#coding:utf-8
import requests
import json
from bs4 import BeautifulSoup
def demodashi(url):
response = requests.get(url)
html = response.text
html = json.loads(html)
totalPage = html['totalPage']
currentPage = html['currentPage']
data = html['data']
demouserdict = {}
for demo in data:
demouserdict[demo['ID']] = demo
pircesumpage = 0
for id in demouserdict:
downnum = demouserdict[id]['DOWN_NUM']
usrurl = 'http://www.demodashi.com/demo/%s.html' % id
header={
'Accept': 'text / html, application / xhtml + xml, application / xml;q = 0.9, image / webp, * / *;q = 0.8',
'Accept - Language': 'zh - CN, zh;q = 0.8',
'Cookie': 'userIdInfoCookie = XYq4JJPg8Pgw8LEUMhD;userSessionInfoCookie = 1ABE726154AB6AA9E8E771AD3188ED8C;JSESSIONID = BBAA6DCFC9E2CE2568AEEBF38EB3F101;Hm_lvt_27ece87947b753bb8e85404fdffce737 = 1521615577, 1521716172, 1522033433, 1522042544;Hm_lpvt_27ece87947b753bb8e85404fdffce737 = 1522042544',
'User - Agent': 'Mozilla / 5.0(Windows NT 6.1;WOW64) AppleWebKit / 537.36(KHTML, likeGecko) Chrome / 55.0.2883.87UBrowser / 6.2.3964 .2Safari / 537.36',
}
response = requests.get(usrurl,headers=header)
html = response.text
soup = BeautifulSoup(html,'lxml')
try:
pirce = soup.find('div',class_='sj-dm dsj-l').b.string
except:
pirce = 0
pircesumpage += int(pirce)*int(downnum)
return pircesumpage,totalPage
def bankuai(url):
totalPage = demodashi(url)[1]
totalpirce = 0
for page in range(0,totalPage):
url = 'http://www.demoku.com/portal/demo/getDemoList4page?queryString=&toPage=%s&orderBy=&type=&typeId=&ptype=&to1=demo&to2=buy&userId=' % page
totalpirce += demodashi(url)[0]
return totalpirce
if __name__ == '__main__':
'''
全部 :http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=&type=&typeId=&ptype=&to1=demo&to2=buy&userId=
全部 (热门、推荐、最新)
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=downLoadMax&type=&typeId=&ptype=&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=avgMarkMax&type=&typeId=&ptype=&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=uploadNew&type=&typeId=&ptype=&to1=demo&to2=buy&userId=
大数据
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=downLoadMax&type=1&typeId=2&ptype=2&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=avgMarkMax&type=1&typeId=2&ptype=2&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=uploadNew&type=1&typeId=2&ptype=2&to1=demo&to2=buy&userId=
移动技术
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=downLoadMax&type=1&typeId=9&ptype=9&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=avgMarkMax&type=1&typeId=9&ptype=9&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=uploadNew&type=1&typeId=9&ptype=9&to1=demo&to2=buy&userId=
web系列
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=downLoadMax&type=1&typeId=12&ptype=12&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=avgMarkMax&type=1&typeId=12&ptype=12&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=uploadNew&type=1&typeId=12&ptype=12&to1=demo&to2=buy&userId=
前端开发
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=downLoadMax&type=1&typeId=26&ptype=26&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=avgMarkMax&type=1&typeId=26&ptype=26&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=uploadNew&type=1&typeId=26&ptype=26&to1=demo&to2=buy&userId=
系统与编程
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=downLoadMax&type=1&typeId=31&ptype=31&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=avgMarkMax&type=1&typeId=31&ptype=31&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=uploadNew&type=1&typeId=31&ptype=31&to1=demo&to2=buy&userId=
互联网应用
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=downLoadMax&type=1&typeId=47&ptype=47&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=avgMarkMax&type=1&typeId=47&ptype=47&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=uploadNew&type=1&typeId=47&ptype=47&to1=demo&to2=buy&userId=
方案设计
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=downLoadMax&type=1&typeId=50&ptype=50&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=avgMarkMax&type=1&typeId=50&ptype=50&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=uploadNew&type=1&typeId=50&ptype=50&to1=demo&to2=buy&userId=
产品设计
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=downLoadMax&type=1&typeId=52&ptype=52&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=avgMarkMax&type=1&typeId=52&ptype=52&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=uploadNew&type=1&typeId=52&ptype=52&to1=demo&to2=buy&userId=
UI设计
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=downLoadMax&type=1&typeId=54&ptype=54&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=avgMarkMax&type=1&typeId=54&ptype=54&to1=demo&to2=buy&userId=
http://www.demodashi.com/portal/demo/getDemoList4page?queryString=&toPage=0&orderBy=uploadNew&type=1&typeId=54&ptype=54&to1=demo&to2=buy&userId=
'''
urllist = ['http://www.demoku.com/portal/demo/getDemoList4page?queryString=&toPage=1&orderBy=&type=&typeId=&ptype=&to1=demo&to2=buy&userId=']
for url in urllist:
print 'demo大师网站总收入: ' + str(bankuai(url)) 内容如果侵犯隐私,联系我会马上删除。。。。。 (961769710)
标签:爬虫
原文地址:http://blog.51cto.com/9399369/2091177