Die Monte Carlo Tree Search für Tic-Tac-Toe (Code zu meinem Video https://youtu.be/CjldSexfOuU)

t3mc.py

# Code aus meinem YouTube-Video https://youtu.be/CjldSexfOuU

import random, copy

class Board:
    def __init__(self):
        self.board = [0,0,0,0,0,0,0,0,0]
        self.moves = []
        self.turn  = +1
        
    def __repr__(self):
        str, repr = "", "O.X"
        for i in [0,3,6]:
            str += "{} {} {}\n".format(repr[self.board[i]+1],repr[self.board[i+1]+1],repr[self.board[i+2]+1])
        return str[:-1]
    
    def list_moves(self):
        return [i for i in range(0,len(self.board)) if self.board[i] == 0]
    
    def make_move(self, pos):
        assert 0 <= pos <= 8 and self.board[pos] == 0
        self.board[pos] = self.turn
        self.moves += [pos]
        self.turn = -self.turn

    def undo_move(self):
        assert self.moves != []
        pos = self.moves.pop()
        self.board[pos] = 0
        self.turn = -self.turn
    
    def three_in_a_row(self):
        for row in [[0,1,2],[3,4,5],[6,7,8],[0,3,6],[1,4,7],[2,5,8],[0,4,8],[2,4,6]]:
            if abs(sum([self.board[i] for i in row])) == 3: return self.board[row[0]]
        return 0

def play_randomly(board):
    value = board.three_in_a_row()
    while value == 0:
        moves = board.list_moves()
        # print("moves =", moves)
        if moves == []: return 0
        random_move = moves[random.randrange(0,len(moves))]
        # print("move is", random_move)
        board.make_move(random_move)
        # print(board)
        value = board.three_in_a_row()
    return value

def simulate_plays(board, number=100):
    assert number >= 1
    count = [0,0,0]
    while number > 0:
        b = copy.deepcopy(board)
        count[play_randomly(b)+1] += 1
        number -= 1
    return count

def evaluate_moves(board, number=100):
    assert number >= 1
    moves = board.list_moves()
    values = []
    for move in moves:
        board.make_move(move)
        values += [simulate_plays(board,number)]
        board.undo_move()
    return values

def choose_best_move(board, number=100):
    assert number >= 1 and board.list_moves() != []
    moves  = board.list_moves()
    values = [r * board.turn for l,m,r in evaluate_moves(board,number)]
    max_val = max(values)
    max_idx = values.index(max_val)
    return moves[max_idx]

def play(board, number=100):
    assert number >= 1
    value = board.three_in_a_row()
    while value == 0:
        if board.list_moves() == []: return value
        move = choose_best_move(board, number)
        board.make_move(move)
        # print(str(board) + "\n---")
        value = board.three_in_a_row()
    return value

def tournament(rounds, number=100):
    count = [0,0,0]
    while rounds > 0:
        res = play(Board(),number)
        count[res+1] += 1
        rounds -= 1
    return count