Connect Four AI

A high-performance, perfect Connect Four solver written in Rust, with bindings for Python and WebAssembly.

This library can strongly solve any Connect Four position and determine the optimal move. For full details, performance benchmarks, and demos, please see the main GitHub Repository.

Key Features

• Perfect Solver: Implements an optimised negamax search, which utilises alpha-beta pruning and a transposition table to quickly converge on exact game outcomes.

• AI Player: Features an AI player with configurable difficulty. It can play perfectly by always choosing the optimal move, or can simulate a range of skill levels by probabilistically selecting moves based on their scores.

• Bitboard Representation: Uses a compact and efficient bitboard representation for game positions, allowing for fast move generation and evaluation.

• Embedded Opening Book: Includes a pre-generated opening book of depth 8, which is embedded directly into the binary for instant lookups of early-game solutions.

• Parallel Book Generator: A tool built with rayon for generating new, deeper opening books.

• Cross-Platform: Available as a Rust crate, Python package, and WebAssembly module for seamless integration into a wide range of projects.

Installation

To use the library, you can simply install it using pip:

pip install connect-four-ai

Example Usage

This is a basic example of how to use the Solver to find the score of a position:

from connect_four_ai import Solver, Position

# Creates a position from a sequence of 1-indexed moves
position = Position.from_moves("76461241141")

# Initialises and uses the Solver to calculate the exact score of the position
solver = Solver()
score = solver.solve(position)

print(score)    # Output: -1

License

This project is licensed under the MIT License. See the LICENSE file for details.