Builds off Roger Whitson's work in Markov-Tweet. Updated to automate and simplify the process of emulating a twitter user's style as well as replacing now-defunct libraries. Also, allows for Arabic text.

To use:

• pip install tweepy
• Update the values in 'DO_ME_twitter_keys' and save as keys.py
• Update Markov_Tweet.py with the handle you wish to emulate.
• python Markov_Tweet.py

Essentially, this bot differs from the original in that it's designed to emulate existing Twitter users effortlessly. No need to compile training data or multi-step procedures. Given the handle, it will:

• Autodump the last 3240 tweets of that user in csv
• Remove retweets+mentions
• Convert it into a text file
• Feed that into a MCMC model
• Tweet a single status out to your account.

This can be modified to emulate a single user, skipping the new tweetdumps and conversions, by simply commenting out lines 19-20. You can add python Markov_Tweet.py to your CronJob to run regularly as a bot after 1 run. (Please don't tweetdump daily)

TODO:- Get a better damn MCMC model. Too little data and a broken model makes for very repetitive, nonsense tweets.