use python to track friends’ status on xiaonei

Although I am not very interested in xiaonei.com. I just want to write a program to know the status of my friends and classmates. Through this program, you can receive the status infomation of your friends in gtalk. The post_status section (to login to a gtalk account and send a message using xmpp) of the code is referred from solrex.(the exact post is here)

To use this program, you need to install the python-xmpp library, if you are using debian or ubuntu, you can install it by sudo apt-get install python-xmpp. Also it means that you need to have two gtalk accounts, you can use the not usually used one to send the message to the usually used one. Notice that these two accounts need to be in friend list. Maybe there are better ways.

 #!/usr/bin/python
 # -*- coding: UTF-8 -*-
 #
 # Author: ideal (idealities AT gmail.com)
 # Date:   2008-11-04
 # Homepage: http://qdhedu.com/ideal
 
 from sgmllib import SGMLParser
 import urllib
 import urllib2
 import xmpp
 
 def post_status(sendto = 'username@gmail.com', message = ''): #put your usually used account here in sendto
     login = 'bjtulinux' #your not usually used account
     passwd = '' #and the password
     talk = xmpp.Client('gmail.com', debug=[])
     talk.connect(server = ('talk.google.com', 5223))
     talk.auth(login, passwd, 'python')
     talk.send(xmpp.Message(sendto, message))
 
 def get_status(email = 'idealities@gmail.com', passwd = ''): #your account and password for xiaonei.com
     url = 'http://xiaonei.com/Login.do'
     vals = {'email': email, 'password': passwd}
     useragent = 'Mozilla/5.0 (Windows; U; Windows NT 5.2; zh-CN; rv:1.9.0.3) Gecko/2008092417 Firefox/3.0.3'
 
     import cookielib
     cook = cookielib.CookieJar()
     opener = urllib2.build_opener(urllib2.HTTPCookieProcessor(cook))
     opener.addheaders = [('User-Agent', useragent)]
     urllib2.install_opener(opener)
     sock = urllib2.urlopen(url, urllib.urlencode(vals))
     text = sock.read()
     sock.close()
     return text
 
 class Status(SGMLParser):
     def reset(self):
         self.message = []
         self.isInBlogDiv = False
         self.isInStatusDiv = False
         self.isInH4 = False
         self.isInH4A = False
         self.isInDate = False
         self.count = 0
         SGMLParser.reset(self)
 
     def start_div(self, attrs):
         if self.isInBlogDiv or self.isInStatusDiv:
             self.count += 1
         for key, value in attrs:
             if key == "class":
                 if value == "feed feed-blog text-story expand bold-feed ":
                     self.isInBlogDiv = True
                 elif value == "feed feed-status text-story expand ":
                     self.isInStatusDiv = True
 
     def end_div(self):
         if self.isInBlogDiv or self.isInStatusDiv:
             self.count -= 1
         if self.count == -1:
             if self.isInBlogDiv:
                 self.isInBlogDiv = False
             elif self.isInStatusDiv:
                 self.isInStatusDiv = False
             self.count = 0
 
     def start_h4(self, attrs):
         if self.isInBlogDiv or self.isInStatusDiv:
             self.isInH4 = True
 
     def end_h4(self):
         if self.isInH4:
             self.isInH4 = False
 
     def start_a(self, attrs):
         if self.isInH4:
             self.isInH4A = True
 
     def end_a(self):
         if self.isInH4A:
             self.isInH4A = False
 
     def start_span(self, attrs):
         if self.isInH4:
             for key, value in attrs:
                 if key == "class" and value == "date":
                     self.isInDate = True
 
     def end_span(self):
         if self.isInDate:
             self.isInDate = False
 
     def handle_data(self, text):
         if self.isInH4A:
             if text != "回复":
                 self.message.append(text)
         elif self.isInDate:
             self.message.append(text + "n")
         elif self.isInH4:
             self.message.append(text)
 
     def output(self):
         """Return processed HTML as a single string"""
         return "".join(self.message)
 
 if __name__ == '__main__':
     html = get_status()
     msg = Status()
     msg.feed(html)
     post_status(message = msg.output())
     msg.close()

You can put it in cron and let it run every n hours.