#coding=utf-8
import json
import requests
import re
class spprec(object):
\tdef __init__(self, url):
\t\tself.session = requests.session()
\t\tself.session.headers.update({
\t\t\t'User-Agent':'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/48.0.2564.116 Safari/537.36'
\t\t})
\t\tself.url = url
\t\tself.post_data = {}
\tdef getInfo(self, num):
\t\tif len(self.post_data) == 0:
\t\t\tres = self.session.get(self.url)
\t\telse:
\t\t\tself.post_data['__EVENTARGUMENT'] = num
\t\t\tself.post_data["__EVENTTARGET"] = "MoreInfoList1$Pager"
\t\t\tres =
self.session.post(url=self.url, data=self.post_data)
\t\tresult = re.findall('<input type="hidden" name="([^"]+)" id="([^"]+)" value="([^"]+)" />', res.text, re.S)
\t\tpost_data = {}
\t\tfor item in result:
\t\t\tpost_data.update({item[0]: item[2]})
\t\tself.post_data = post_data
\t\tresult = re.findall('<a href="(/sczw/InfoDetail/Default.aspx[^"]+)" target="_blank" title="([^"]+)"', res.text, re.S)
\t\treturn result
obj = spprec("
http://www.spprec.com/sczw/jyfwpt/005001/005001001/MoreInfo.aspx?CategoryNum=005001001")
for i in xrange(1, 5):
\tdata = obj.getInfo(i)
\tfor item in data:
\t\tprint item[1]
缩进全没了,请自行替换吧,\t 对应 4 个空格或者一个 tab