Backgammon in Prolog

A classic implementation of the game Backgammon, written in Prolog. This project focuses on game logic, player interactions, and board representation using declarative programming principles.

Getting started

1. Clone the repository

git clone https://github.com/Francesinhas/backgammon-in-prolog.git

2. Create an environment

python3 -m venv .venv && source .venv/bin/activate

3. Install Prolog

• For Windows: Download and install SWI-Prolog
• For macOS: brew install swi-prolog
• For Linux (Debian/Ubuntu): sudo apt-get install swi-prolog

4. Set Up Python Environment

pip install pyswip
pip install pygame

5. Run the game

python3 python_frontend/app.py

6. (Optional - for development purposes) VS Code Setup

• Install the "VSC-Prolog" extension by "arthwang" for VS Code. This extension provides syntax highlighting and query support.

Project Structure

Prolog backend

• We'll start by defining the board representation and basic predicates in game_rules.pl.
• Then, we'll handle state transitions in state_manager.pl.
• Finally, we'll create a simple AI in ai.pl.

Python interface

• A simple python user interface has been created.

Workflow

1. We'll define the board using the Tesauro encoding. We'll represent the board as 24 points, each with a color and a number of checkers. Also, we have the bar and off for each player.

Plan

• Prolog logic
  • implement project basic structure
  • check logic in game rules
  • check updated logic in state manager
  • implement AI
• link .pl files together #test
• connect to python #important
  • define prolog interface
  • start implementing python gui

Improvements

• possible issues:
  • valid_move - target computation
    • why compute target if To is the target? #q
  • state_manager
    • doesnt check if moves are valid, should it? #q
    • move_from_bar calls apply_move(Player, bar, To). bar doesn't look like a valid argument
      • implement the logic for placing the piece directly there or in another rule
  • [ ]
• relevant logic:
  • checking the dice numbers when figuring if a move or bearing can be made (not in game_rules yet)
  • [ ]
• nice to have:
  • state_manager:
    • don t assert back a number of poins after a move if it drops under 0

GUI communication with prolog

→ init_game() → show_board() → while not winner: → player_turn() → roll_dice → get_input → validate/apply → update board → check_winner → ai_turn() → choose_move → apply → update board → check_winner → show_winner()

Tests

• check if capturing a piece actually puts it on the bar
  • check if the bar state is sent to state_to_list
• check perfom move from bar