A handy package for kickstarting AI contests

We proudly provide a Python framework that makes the building of an environment for adversarial search agents easy!

A lot of fun functionalities are provided an easily usable out of the box!

Implementing a new game setup from scratch? Painless!
Playing against remote agents? Easy!
Generating and collecting data about played games? Worriless!
Organizing a contest for a large range of agents? Quick and reliable!

Installation

The package is publicly available on PyPI.

We strongly encourage the use of a virtual environment:

    $ python3 -m venv venv

    $ source venv/bin/[activate|Activate.ps1]

    (venv)$ pip install seahorse

Running the example

Initial state	Intermediate state	Final state

An example implementation featuring the famous game of tic-tac-toe is available on the github repository under the branch with its name:

git clone https://github.com/corail-research/seahorse/ \
--branch tictactoe \
--single-branch tictactoe

Learn on how to launch the example and recreate one of your own through our tutorials

Main contributors

We are an enthusiastic team of M.Sc candidates led by Pr. Quentin Cappart at Polytechnique Montréal. The package was originally developed in the context of a introductory course to artificial intelligence given to undergrad computer and software engineering students.