so you want to do something with a markov bot and tweeter....

tweetmarkov.py

#!/usr/bin/env python
# 
import twitter
from pymarkovchain import MarkovChain
from pprint import pprint


TW_CONSUMER_KEY='xxx'
TW_CONSUMER_SECRET='yyy'
TW_ACCESS_TOKEN='aaa'
TW_TOKEN_SECRET='bbb'

if __name__ == "__main__":
  api = twitter.Api(consumer_key=TW_CONSUMER_KEY,
                      consumer_secret=TW_CONSUMER_SECRET,
                      access_token_key=TW_ACCESS_TOKEN,
                      access_token_secret=TW_TOKEN_SECRET)
  tweets = api.GetUserTimeline(screen_name="RikerGoogling",count=200)
  #pprint(tweets)
  # or someplace. Expect a scary error if this file is E_NOENT
  # then the library makes the file. /shrug
  mc = MarkovChain("/var/tmp/markov.riker") 
  # don't worry about a bunch of None 
  [ mc.generateDatabase(t.__dict__['text']) for t in tweets ]
  print mc.generateStringWithSeed("rudy")
  print mc.generateStringWithSeed("rudy")
  print mc.generateStringWithSeed("picard")
  print mc.generateStringWithSeed("picard")
  print mc.generateStringWithSeed("planet")
  print mc.generateStringWithSeed("planet")
  print mc.generateStringWithSeed("planet")
  print mc.generateStringWithSeed("planet")