https://github.com/bertrandom/costanzabot

Bot generated from a markov chain of everything George Costanza says in Seinfeld
https://github.com/bertrandom/costanzabot

Last synced: 23 days ago
JSON representation

Bot generated from a markov chain of everything George Costanza says in Seinfeld

• Host: GitHub
• URL: https://github.com/bertrandom/costanzabot
• Owner: bertrandom
• Created: 2014-05-09T07:09:52.000Z (over 10 years ago)
• Default Branch: master
• Last Pushed: 2014-05-09T07:36:33.000Z (over 10 years ago)
• Last Synced: 2023-04-17T14:00:28.477Z (over 1 year ago)
• Language: Ruby
• Homepage:
• Size: 676 KB
• Stars: 4
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# costanzabot

Bot generated from a markov chain of everything George Costanza says in Seinfeld

## Usage

`./test.rb` will spit out an example tweet.

## How this particular sausage was made

or "instructions for Future Bert if he wants to do this again".

This was mostly just an excuse to play with [twitter_ebooks](https://github.com/mispy/twitter_ebooks).

First, start with [seinfeld-scripts](https://github.com/colinpollock/seinfeld-scripts).

Next, patch it with this [fix](https://github.com/cschep/seinfeld-scripts/blob/d34dd1deac1b71c760f59bf810494b51e5b87042/download.sh).

```

mkdir scripts

./download.sh scripts

```

Edit `scripts/01.shtml` and change `pc: 101, season 1, episode 1 (Pilot)
` to `pc: 101, season 1, episode 0 (Pilot)
`.

```

./run.sh

```

This will create an sqlite database called `seinfeld.db`.

Extract everything that George says into a text file:

```

sqlite3 seinfeld.db "SELECT sentence.utterance_id, sentence.text from sentence, utterance WHERE utterance.speaker = 'GEORGE' AND sentence.utterance_id = utterance.id;" > export.txt

```

Move export.txt to the `seed` directory of this repo.

Join the utterances into single lines of text with the parsing script:

`./parse.rb > costanza.txt`

Create a skeleton for the bot.

`ebooks new costanzabot`

Move `costanza.txt` into the `corpus` directory.

Consume the lines into a Markov model:

```

ebooks consume corpus/costanza.txt

```