Project

Bresse is a library for use LLM (Large Language Model) like chess AI. This library takes advantage of the fact that LLMs can reproduce a format representing a chess game called PGN. You will find various tools to facilitate experiments, PGN modifications during inference and various useful information.

Installation

This project uses Poetry for dependency management. To install the dependencies, run the following command:

poetry install

Usage

To run the project, use the following commands:

python src

Contribution

To install the development dependencies, run:

poetry install --dev

To add a new dependency, use:

poetry add <dependency>

For development-specific dependencies, use:

poetry add --group dev <dependency>

Structure

├── README.md         # The file you are currently reading
├── htmlcov           # The coverage report folder
├── pyproject.toml    # The poetry configuration file
├── ruff.toml         # The ruff configuration file (linter, formatter)
├── scripts           # Scripts useful for the project (no CI/CD)
├── src               # The source code folder
│   ├── __init__.py   # Can add global variables
│   ├── __main__.py   # The entry point of the project
└── tests             # The tests folder (pytest)

Tests

To run the tests, use the following command:

pytest

There are test who tests user cases, this tests use real inference of LLM and will use your .env file to get the API key. This tests cost money and is disabled per default, so be careful when running them.

pytest -m "costly or not costly"