1. from Tkinter import *
  2.  
  3. from lxml.html import parse
  4. import re
  5.  
  6. curProxy = {}
  7. proxies = []
  8.  
  9. class App:
  10.  
  11.     def __init__(self, master):
  12.  
  13.         counter = "what"
  14.  
  15.         frame = Frame(master)
  16.         frame.pack()
  17.  
  18.         self.hi_there = Button(frame, text="Scrape", command=self.scrape)
  19.         self.hi_there.pack(side=LEFT)
  20.  
  21.         self.scrollbar = Scrollbar(frame, orient=VERTICAL)
  22.  
  23.         self.listbox = Listbox(frame, yscrollcommand=self.scrollbar.set)
  24.         self.scrollbar.config(command=self.listbox.yview)
  25.         self.scrollbar.pack(side=RIGHT, fill=Y)
  26.  
  27.         self.listbox.pack()
  28.  
  29.  
  30.  
  31.     def scrape(self):
  32.  
  33.         urls = ['http://www.proxy4free.com/page1.html']
  34.     ##                'http://www.proxy4free.com/page2.html',
  35.     ##                'http://proxynext.com/proxylist1.php',
  36.     ##                'http://proxynext.com/proxylist2.php']
  37.  
  38.         for url in urls:
  39.  
  40.             doc = parse(url).getroot()
  41.             for td in doc.cssselect('td'):
  42.  
  43.                 ipPattern = re.compile("\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}(?!(-|.|:))")
  44.                 ipMatch = ipPattern.match(td.text_content())
  45.  
  46.                 portPattern = re.compile("\d{2,4}(?!(-|.|:))")
  47.                 portMatch = portPattern.match(td.text_content())
  48.  
  49.                 if ipMatch:
  50.                     curProxy['ip'] = td.text_content()
  51.  
  52.                 elif portMatch:
  53.                     curProxy['port'] = td.text_content()
  54.                     print curProxy
  55.  
  56.                     proxies.append(curProxy)
  57.  
  58.         for i in range(20):
  59.             self.listbox.insert(END, proxies[i]['ip'] + ':' + proxies[i]['port'])
  60.  
  61. root = Tk()
  62. app = App(root)
  63. root.mainloop()