DonatelloPyzza 1.8.0

A simple and educational game to learn algorithmic where you can develop the behavior of a turtle to move in a maze.

DonatelloPyzza

DonatelloPyzza is a simple and educational game to help beginners learn algorithmic in Python in high school and university.

A turtle can move through a grid and touch each cells until it finds the pizza. This game can be used at several levels:

• for young beginners: they can hard-code a path to help the turtle find its pizza in small mazes.
• for beginners: they can develop intuitive heuristics to find the pizza
• for intermediate or advanced developers: they can develop a complex path finding method or AI-based solutions.

Installation

pip install donatellopyzza

Getting started

First, import the right modules to run the game:

from donatellopyzza import Game
from donatellopyzza import Action
from donatellopyzza import Feedback

Game is a class that you will use to create a game instance.

Let's create the game and its environment:

# specify the name of the maze
__ENVIRONMENT__ = "maze"
# display the interface (or not)
__GUI__ = True

game = Game(__ENVIRONMENT__, __GUI__)
# returns a turtle that execute actions on its environment
turtle = game.start()

Once the game has started, you get a turtle instance which you can move around the board. To do this, the following instruction can be used:

feedback, _ = turtle.execute(Action.MOVE_FORWARD)
print(feedback)

You can use the feedback from the execute() method to see what happened after your action.

Finally, let's run your turtle:

python yourfile.py

Learning the rules: actions and feedbacks

Action and Feedback define the different actions and feedbacks types. You can use the following actions in your code:

Action.MOVE_FORWARD -> make your turtle go one step forward
Action.TURN_RIGHT -> your turtle will turn on its right
Action.TURN_LEFT -> on its left
Action.TOUCH -> the turtle will touch the cell in front of it to know its type

Depending on your action, the game can provide you one of the following feedback:

Feedback.COLLISION -> you just tried to walk in a wall !
Feedback.MOVED -> you successfully moved
Feedback.MOVED_ON_PIZZA -> your turtle is on the pizza (congratulation!)
Feedback.TOUCHED_WALL -> you just touched a wall
Feedback.TOUCHED_NOTHING -> the touched cell is empty (no wall, no pizza, you can walk on it)
Feedback.TOUCHED_PIZZA -> the turtle touched the pizza

Generating and save your own mazes

Maze is a class used to generate and save new mazes. You can retrieve saved maze by their names as indicated in example files. A new maze is generated (and saved) as follow:

from donatellopyzza import MazeGenerator

generator = MazeGenerator()
maze = generator.create_maze(10, 10)
fn = "test"
maze.save(maze, filename=fn)

For more details, you can find several complete examples of the game loop in the examples folder on the github repository of this project. You can also look for the documentation at this link

Have fun!

What's new

• 2024-09-27 (v1.8) Add a RLGame class facilitating the use of RL algorithms.
• 2024-09-25 (v1.7) Add a reward function in order to play the game with reinforcement learning. Add the possibility to change the color of each cell for debugging.
• 2023-04-09 (v1.6) Integration of an algorithm assessment class. Allows to evaluate automatically a solution on several mazes.
• 2023-04-07 (v1.5) Integration of a maze generator from Alexandru Văleanu (https://github.com/AlexandruValeanu/Mazify).
• 2023-01-17 (v1.2) Initial release

Roadmap

• make several tests for this package
• make tutorials to help beginners use this package
• make a more formal documentation
• promote this game through a website
• add a RLGame class to facilitate the use of RL algorithm for solving the game
• add a reward in the step() method to allow reinforcment learning
• adapt the GUI to resize it depending on the number of cells
• add a test infrastructure to validate users' algorithm on several mazes
• make possible to the user to select the difficulty of the maze when generating it
• debug the GUI
• add a gridworld generator