W.H.A.L.E.S.

Chess AI web application using minimax and deep learning on backend.

Live on Heroku

API

API docs available here

Backend

Backend setup

Install the following dependencies:

On macOS using Homebrew, that looks like this:

$ brew install git pyenv pipenv
$ pyenv install 3.6.8

Clone this repository:

$ git clone https://github.com/raxod502/cs121-whales.git
$ cd cs121-whales

Create a virtual environment for the backend:

$ pipenv install --python "$(pyenv prefix 3.6.8)"/bin/python

Download the models:

$ make download-models

Run the server:

$ make run-server-dev

Code is live-reloaded. To test in production mode using a real WSGI server:

$ make run-server-prod-test

In either case, the backend is now running on localhost:5000, or whichever port is printed on the command line.

Backend usage

Install HTTPie. On macOS, that looks like this:

$ brew install httpie

This makes it easy to test HTTP requests.

$ http GET localhost:5000/not-found
$ http GET localhost:5000/api/v1 command=unknown_command
$ http GET localhost:5000/api/v1 command=list_models
$ http GET localhost:5000/api/v1 command=get_move model=random pgn="1. e4 e5 2. Qh5 Nc6 3. Bc4 Nf6 *"

General tips

Developing documentation

Install Grip. On macOS, that looks like this:

$ brew install grip

Then you can render the README locally, so you can correct formatting errors before pushing to GitHub:

$ grip

Or you can render another Markdown document:

$ grip doc/api-spec.md

Typically the server runs on localhost:6419 if that port is available.

Vendored files

We vendor a forked version of chessboard.js from here, minified online.