usr bin env python encoding utf-8 twitter-juick py Created by Alexey K

#!/usr/bin/env python
# encoding: utf-8
"""
twitter-juick.py

Created by Alexey Karguine on 2009-02-25.
Copyright (c) 2009 alexeykarguine@me.com. All rights reserved.
"""

import sys
import os
import xml.dom.minidom
import urllib
import xmpp
import cPickle
import logging
import daemonize
import time
import re

RSS = 'http://twitter.com/statuses/user_timeline/username.xml'

FROM_GMAIL_ID = "user@jabber.srv/twitter2juick"
GMAIL_PASS = "passwd"
GTALK_SERVER = "jabber.srv"
TO_GMAIL_ID = 'juick@juick.com'

TimeOut = 60

# Logging
curr_dir = os.getcwd()
logfile = os.path.join(curr_dir, 'twitter2juick.log')
errfile = os.path.join(curr_dir, 'twitter2juick_crash.log')
logging.basicConfig(filename=logfile, level=logging.INFO, 
        format="%(asctime)s [%(levelname)s] %(message)s")

# Pickle
curr_dir = os.getcwd()
lastseenfn = os.path.join(curr_dir, 'lastid.pickle')

# regular for tags
pattern = re.compile('^#')


def xmppConnect():
	jid=xmpp.protocol.JID(FROM_GMAIL_ID)
	cl=xmpp.Client(jid.getDomain(),debug=[])
	if not cl.connect((GTALK_SERVER,5222)):
	    raise IOError('Can not connect to server.')
	if not cl.auth(jid.getNode(),GMAIL_PASS,resource=jid.getResource()):
	    raise IOError('Can not auth with server.')
	return cl
	
def getText(nodelist):
	rc = ""
	for node in nodelist:
		if node.nodeType == node.TEXT_NODE:
			rc = rc + node.data
	return rc

def getValue(node, name):
	"""Returns text value of node content"""
	return getText(node.getElementsByTagName(name)[0].childNodes)

def getLastseen():
	logging.debug('File: %s' % lastseenfn)
	try:
		pfile = open(lastseenfn, 'rb')
		try:
			lastseen = cPickle.load(pfile)
		except EOFError:
			logging.warn( "Pickle empty, passing.")
			lastseen = 0
	except IOError:
		logging.warn("Pickle file not present.")
		lastseen = 0
	return lastseen

def saveLastseen(lastseen):
	pfile = open(lastseenfn, 'wb')
	cPickle.dump(lastseen, pfile, 2)
	pfile.close()

def loop():
	logging.debug('In loop')
	lastid = getLastseen()

	logging.debug('get last: %d' % lastid)

	if lastid:
		URL = RSS+'?since_id='+str(lastid)
	else:
		URL = RSS
		
	logging.debug('URL: %s' % URL)
	try:
		data = urllib.urlopen(URL)
	except:
		logging.exception("Can't get RSS")
		return
		
	try:
		dom = xml.dom.minidom.parse(data)
	except:
		logging.exception("Can't parse XML")
		return
		
	statuses = dom.getElementsByTagName('status')
	statuses.reverse()
	logging.debug('Got rss, len: %s' % len(statuses))
	
	try:
		cl = xmppConnect()
	except:
		logging.exception("Can't connect to GTalk")
		return
		
	logging.debug('Connected to jabber')
	
	for status in statuses:
		logging.debug('In loop cycle')
		text = pattern.sub('*', getValue(status, 'text'))
		id_ = int(getValue(status, 'id'))
		is_reply = getValue(status, 'in_reply_to_screen_name')
		if len(is_reply) == 0:
			logging.info( "id: %s text: %s" % (id_, text) )
			cl.send(xmpp.Message(TO_GMAIL_ID, text))
		lastid = id_
		logging.debug('ID: %d' % id_)
		
	saveLastseen(lastid)
	cl.disconnect()
	logging.debug('EOF loop')

def main():
	daemonize.daemonize(stderr=errfile)
	while True:
		try:
			logging.debug('Starting new cycle')
			loop()
		except:
			logging.exception('main')
		finally:
			time.sleep(TimeOut)

if __name__ == '__main__':
	main()