Skip to content

Delving into HTML parsing and Markov chains with a Twitter Bot parody of herbal tea descriptions.

Notifications You must be signed in to change notification settings

eqfinney/verbal-infusions

Repository files navigation

Verbal Infusions

Author: Emily Quinn Finney

This package includes code used to support the infrastructure of a Twitter Bot (https://twitter.com/verbalinfusions) I wrote during my time at the Recurse Center (https://www.recurse.com/).

Features/Details

The package includes the following modules:

  • create_status.py: implements a Markov chain from which herbal tea parody tweets are generated.
  • hteaml_parser.py: parses a corpus of HTML using Python's Beautiful Soup and re packages. Apologies for the pun are not included.
  • twitter_bot.py: the infrastructure for the Twitter Bot, written using the Tweepy package that wraps the Twitter API (http://www.tweepy.org/).

Running Tests

The modules test_*.py are pytest modules that test the code corresponding to their names. To test the code, you will need pytest installed. Then type the following into the terminal:

pytest -v -s test_{$MODULE_NAME}.py

Last edited 04/05/2018

About

Delving into HTML parsing and Markov chains with a Twitter Bot parody of herbal tea descriptions.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages