#!/usr/bin/python
#-*- coding:utf-8 -*-
"""
author:dasuda
"""
import urllib2
import re
import socket
import threading
findIP = [] #獲取的原始IP數(shù)據(jù)
IP_data = [] #拼接端口后的IP數(shù)據(jù)
IP_data_checked = [] #檢查可用性后的IP數(shù)據(jù)
findPORT = [] #IP對應(yīng)的端口
available_table = [] #可用IP的索引
def getIP(url_target):
 patternIP = re.compile(r'(?<=<td>)[\d]{1,3}\.[\d]{1,3}\.[\d]{1,3}\.[\d]{1,3}')
 patternPORT = re.compile(r'(?<=<td>)[\d]{2,5}(?=</td>)')
 print "now,start to refresh proxy IP..."
 for page in range(1,4):
  url = 'http://www.xicidaili.com/nn/'+str(page)
  headers = {"User-Agent": "Mozilla/5.0 (Windows NT 10.0; WOW64)"}
  request = urllib2.Request(url=url, headers=headers)
  response = urllib2.urlopen(request)
  content = response.read()
  findIP = re.findall(patternIP,str(content))
  findPORT = re.findall(patternPORT,str(content))
  #assemble the ip and port
  for i in range(len(findIP)):
   findIP[i] = findIP[i] + ":" + findPORT[i]
   IP_data.extend(findIP)
  print('get page', page)
 print "refresh done!!!"
 #use multithreading
 mul_thread_check(url_target)
 return IP_data_checked
def check_one(url_check,i):
 #get lock
 lock = threading.Lock()
 #setting timeout
 socket.setdefaulttimeout(8)
 try:
  ppp = {"http":IP_data[i]}
  proxy_support = urllib2.ProxyHandler(ppp)
  openercheck = urllib2.build_opener(proxy_support)
  urllib2.install_opener(openercheck)
  request = urllib2.Request(url_check)
  request.add_header('User-Agent',"Mozilla/5.0 (Windows NT 10.0; WOW64)")
  html = urllib2.urlopen(request).read()
  lock.acquire()
  print(IP_data[i],'is OK')
  #get available ip index
  available_table.append(i)
  lock.release()
 except Exception as e:
  lock.acquire()
  print('error')
  lock.release()
def mul_thread_check(url_mul_check):
 threads = []
 for i in range(len(IP_data)):
  #creat thread...
  thread = threading.Thread(target=check_one, args=[url_mul_check,i,])
  threads.append(thread)
  thread.start()
  print "new thread start",i
 for thread in threads:
  thread.join()
 #get the IP_data_checked[]
 for error_cnt in range(len(available_table)):
  aseemble_ip = {'http': IP_data[available_table[error_cnt]]}
  IP_data_checked.append(aseemble_ip)
 print "available proxy ip:",len(available_table)

patternIP = re.compile(r'(?<=<td>)[\d]{1,3}\.[\d]{1,3}\.[\d]{1,3}\.[\d]{1,3}')
patternPORT = re.compile(r'(?<=<td>)[\d]{2,5}(?=</td>)')
...
findIP = re.findall(patternIP,str(content))
findPORT = re.findall(patternPORT,str(content))

['ip1:port1','ip2:port2',....]