Twittov

Twittov generates nonsense tweets by applying the Markov chain model to your Twitter history. In a nutshell, it splits up text into smaller chunks and then tries to reform these chunks in a correct order. This is never completely accurate, so you end up with some spectacular nonsense. (Incidentally, this is the algorithm used by many spam generators...) If this interests you, other folks have done a much better job of explaining it in detail.

Releases

• January 2011: There's a new script and frontend at http://www.twittov.com/.

• November 2010: Because of various changes in the Twitter API, this code is outdated. I am working on a much more efficient OAuth version.

• twittov.py 0.9.2 (Latest)

  Proper parameter parsing, options for quiet and verbose mode, specifying a cache file, and forcing a cache update.

• twittov.py 0.9.1

  Writing to cache only when necessary. Added a triple() generator to iterate over tweets easily.

• twittov.py 0.9.0

  Woo, finished. Let me know if there are any problems.

How to use twittov.py

Download one of the twittov.py releases from above. You need Python 2.6.2 and the BeautifulSoup module. Navigate to the folder where you downloaded twittov.py and run it:

Usage: twittov.py [options] username
Options:
  -h, --help            show this help message and exit
  -q, --quiet           Don't print status messages to stdout.
  -v, --verbose         Print all messages to stdout.
  -r RANDOMNESS, --randomness=RANDOMNESS
                        Sets the randomness of the output. Must be an integer.
                        Default is 15.
  -l NUMWORDS, --length=NUMWORDS
                        Sets the *minimum* output length, in number of words.
                        NUMWORDS must be a positive integer. Default is 1.
  -c FILE, --cache-file=FILE
                        Sets the cache file. By default, we save to
                        .twittov.cache
  -f, --force-cache-update
                        Force download all tweets and update cache, even if
                        username is already in cache.

Introducing @twittov

If you'd rather not run a python script, just send a tweet to @twittov, and he'll respond with nonsense generated from your feed. If you have very few replies, or all your replies are unique, @twittov won't have enough data to produce proper nonsense. Instead, he'll respond with a joke (sometimes half).

Edit: I'm afraid Twittov's time has come and I no longer have the time to keep running the script and blocking spammers. Thanks everyone for tweeting @twittov when he was breathing!

Notes/Todo

• Please don't hesitate to contact me about bugs or suggestions @yaymukund, or mhl008 [at] gmail [dot] com.
• Don't forget to chmod +x twittov.py, so you can execute it.
• The length is a *minimum* length, and you'll often get Tweets that are longer than length. With the Markov algorithm, it's tough to pin down a precise length.
• If the randomness is too high, twittov may break (it just prints an error message and exits).