This is a project that will attempt to use reinforcement learning and neural networks to learn how to play the card game hearts.

Next few steps for project are:

• Add tests
• Make sure that correct number of wins are happening for certain sets of starting cards
• Install Keras
• Add some NN capabilities and tests

Current assumptions / changes to the rules made to simplify the gameplay:

• No shooting for the moon
• No passing card across the table from each other
• Variable number of players (as few as 2 players)
• Subsets of the full deck can be used (as few as 2 cards per player)

NOTE: The 'display_utils.py' file was for an initial attempt to read the screen of Ubuntu's Hearts program in order to play it. It was extremely slow at playing even a single game (a few seconds per game) and the complexity of the game could not be turned down sufficiently, so I opted to simply create a python framework for the game myself (now in game.py). This original attempt made the following assumptions:

• Using Ubuntu 16.04
• Using the Hearts game that comes as standard
• Using original sized window
• Set the background of the game to uniform green (b'\x00\xff\x00')