關住 公 縱 號 “  阿蒙課程分享    ”  獲得學習資料及趣味分享　

# -*- coding:utf-8 -*-  
# author:murongtiedan
# updatetime:2018/3/14
# 功能：爬蟲之模擬登入，urllib和requests都用了...  
問題背景：

在登入頁面點選登入之後頁面重定向了無數次，想獲取某次重定向前的頁面的 cookie 作為後來值應用

import urllib2
import requests
import re



headers = {
    'User-Agent': 'Mozilla/5.0 (Windows NT 6.2) AppleWebKit/535.11 (KHTML, like Gecko) Chrome/17.0.963.12 Safari/535.11'
 
,
    # 'Referer':'http://cas.oa.ops:8083/cas/login?service=http://10.40.10.165:40992/war-cwd-transaction-ser/sso/user/redirect'
}


def get_verify_data(IP,PORT):
    loginURL = "http://cas.oa.ops:8083/cas/login?service=http://" + IP + ":" + PORT + "/war-cwd-transaction-ser/sso/user/redirect"  # POST傳送到的網址
request = urllib2.Request(loginURL, headers
 
=headers)
    response = urllib2.urlopen(request)
    content = response.read()
    lt_pattern = re.compile(r'name="lt" value="(.*?)" />', re.S)  #re.S  使 . 匹配包括換行在內的所有字元
ex_pattern = re.compile(r'name="execution" value="(.*?)" />', re.S)  #re.S  使 . 匹配包括換行在內的所有字元
lt = re.findall(lt_pattern, content)
    execution = re.findall(ex_pattern, content)
    return 
 
{'lt':lt[0],'execution':execution[0]}


def get_cookie(*args):
    IP = args[0]
    PORT = args[1]
    if type(IP) != 'str':
        IP = str(IP)
    if type(PORT) != 'str':
        PORT = str(PORT)
    loginURL = "http://cas.oa.ops:8083/cas/login?service=http://"+ IP +":"+ PORT +"/war-cwd-transaction-ser/sso/user/redirect"  # POST傳送到的網址
verify_data = get_verify_data(IP,PORT)
    data = {
        "username": "xiatian",
        "password": "xiatian123",
        "lt": verify_data['lt'],
        "execution": verify_data['execution'],
        "_eventId": 'submit',
        'submit': '登入'
}
    s = requests.session()
    login = s.post(loginURL, data=data, headers=headers,allow_redirects= False)  # 傳送登入資訊，返回響應資訊（包含cookie）
    # print(login.headers)
cookie = s.get(login.headers['Location'], data=data, headers=headers, allow_redirects=False)  # 傳送登入資訊，返回響應資訊（包含cookie）
    # print(cookie.headers)
return {'cookie':cookie.headers['Set-Cookie'].split('=')[1].split(';')[0]}

print(get_cookie('10.40.10.165','40992'))