kaboom-engine 0.2.0

Core game engine for the Kaboom card game.

Kaboom Engine

The project implements the full game logic including turns, reactions, powers, and endgame rules, designed to be used as a reusable simulation engine, AI environment, or UI backend.

---

Features

• Complete Kaboom game rules
• Deterministic turn engine
• Reaction resolution system
• Card power mechanics
• Kaboom endgame logic
• Deck reshuffle handling
• Fully enumerated action space
• Action-based engine architecture
• Event-based results
• Fully tested core logic

The engine is designed so it can be used for:

• CLI or graphical game clients
• AI agents
• game simulations
• multiplayer servers
• reinforcement learning environments

---

Installation

pip install kaboom-engine

OR clone the repository:

git clone https://github.com/Arnav-Ajay/kaboom-core.git
cd kaboom-core

Install locally:

pip install -e .

---

Basic Usage

from kaboom.game.engine import GameEngine

engine = GameEngine(game_id=0, num_players=4, hand_size=4)
state = engine.state

players = state.players
current_player = state.current_player()

# draw a card
engine.draw_card(current_player.id)

# discard
engine.discard_card(current_player.id)

---

Running Simulations

The engine exposes the complete legal action space of the current game state.

import random

from kaboom.game.engine import GameEngine
from kaboom.game.turn import apply_action
from kaboom.game.validators import get_valid_actions

engine = GameEngine(game_id=0, num_players=2, hand_size=4)
state = engine.state

while state.phase != GamePhase.GAME_OVER:

    player = state.current_player()

    # skip inactive players (kaboom caller)
    if not player.active:
        state.advance_turn()
        continue

    actions = get_valid_actions(state)

    if not actions:
        break

    action = random.choice(actions)

    apply_action(state, action)
    

---

Architecture

The engine follows a modular architecture based on action execution.

GameState
    |
Action
    | (Draw / Discard / Replace / UsePower / CallKaboom / Reaction Actions)
    |
apply_action()
    |
ActionResult

---

Action System

All player interactions are represented as Action objects.

Core actions include:

Draw
Discard
Replace
UsePower
CallKaboom
CloseReaction
ReactDiscardOwnCard
ReactDiscardOwnCards
ReactDiscardOtherCard
ReactDiscardOtherCards

The function:

get_valid_actions(state)

returns the full list of legal actions for the current state.

---

Project Structure

kaboom/
    cards/
    players/
    powers/
    game/
        actions.py
        game_state.py
        phases.py
        reaction.py
        results.py
        turn.py
        validators.py

tests/
examples/

Key modules:

kaboom/cards      → card definitions
kaboom/players    → player state and memory
kaboom/powers     → card power implementations
kaboom/game       → engine logic, actions, validation

---

Testing

The engine includes a full pytest test suite.

Run tests:

pytest

Coverage includes:

• card scoring
• deck creation
• player management
• power mechanics
• turn actions
• reaction logic
• Kaboom endgame rules
• full game initialization

---

Future Improvements

Planned enhancements:

• CLI interface
• graphical UI
• AI player agents
• multiplayer networking
• replay and event logging
• Gym-compatible environment for reinforcement learning

---

License

This project is licensed under the MIT License.

---

Author

Arnav Ajay