Adding AI

Subclassing Player

Creating a new type of player is as simple as subclassing the Player object and implementing the sense and plan methods. An example is shown below.

class HumanPlayer(Player):
    def __init__(self):
        super().__init__()

    def sense(self, track, keys):
        # do something with pressed keys and the track
        return percepts

    def plan(self, percepts):
        # use the percepts as gotten by `sense()` and return an acceleration_command and rotation_command in range
        # [-1, 1]
        return acceleration_command, rotation_command

Example AI

An example AI can be found in player. Look for the NaiveAi() class. It uses linear distance sensors, which can be found in the same module.

Controlling AI

Below concept demonstrates how to control the flow of AI players during the game.

from src.game import Engine, Environment
from src.player import Ai
import pygame


def main():
    track = Environment('assen')
    game_engine = Engine(track, [Ai()])
    game_engine.bind_action(pygame.K_9, lambda: game_engine.add_player(Ai()))
    game_engine.bind_action(pygame.K_0, lambda: new_player(game_engine))
    game_engine.bind_action(pygame.K_q, lambda: game_engine.remove_all_players(keep=[game_engine.get_best_player()]))
    game_engine.play(False)


def new_player(game_engine):
    best_player = game_engine.get_best_player()
    baby = best_player.give_birth()
    game_engine.add_player(baby)


if __name__ == "__main__":
    main()