The TwitterBot is able to recommend resources based on keywords extracted from twitter status updates using the Twitter Stream and REST API. This Twitter application is used as a distribution channel for recommendations provided by the EEXCESS recommender system. Amongst others, the bot is able to follow the current stream of tweets and can reply with a recommendation after identifying a relevant tweet.

Check out this [abstract] (https://github.com/EEXCESS/Twitter/blob/master/Abstract-TwitterBot.pdf?raw=true) for more information.

• Python3 required

• Setup database (e.g. MySql)
• Add python dependencies:
  • PyMySql
  • peewee
  • rdflib
  • twitter
  • requests
• Checkout source code
• Create Twitter App with write permissions
• Fill in your information in the Config.py file in the package de.uop.twitterbot
• Run CreateTables.py to create all necessary tables in the database
• Use KeywordUtil.py to modify the keyword set
• Modify Recommender.py if you want to use another recommender
• Run RunTwitterBot.py to start start the bot

The sources are released under The MIT License (MIT).