用python生成一段五子棋人机对战代码

以下是一个简单的五子棋人机对战代码，使用了minimax算法和alpha-beta剪枝实现：

import numpy as np

# 定义棋盘大小
SIZE = 15

# 定义棋子类型
EMPTY = 0
BLACK = 1
WHITE = 2

# 定义搜索深度
SEARCH_DEPTH = 3

# 定义评估函数参数
SCORES = {
    (0, 0): 7, (0, 1): 35, (0, 2): 800, (0, 3): 15000, (0, 4): 800000,
    (1, 0): 7, (1, 1): 15, (1, 2): 400, (1, 3): 1800, (1, 4): 100000,
    (2, 0): 7, (2, 1): 15, (2, 2): 200, (2, 3): 1000, (2, 4): 50000,
    (3, 0): 7, (3, 1): 15, (3, 2): 100, (3, 3): 500, (3, 4): 25000,
    (4, 0): 7, (4, 1): 15, (4, 2): 50, (4, 3): 200, (4, 4): 10000
}

# 初始化棋盘
board = np.zeros((SIZE, SIZE), dtype=int)

# 判断是否有五子连珠
def is_win(board, player):
    for i in range(SIZE):
        for j in range(SIZE):
            if board[i][j] == player:
                # 水平方向
                if j + 4 < SIZE and board[i][j+1:j+5].tolist() == [player] * 4:
                    return True
                # 竖直方向
                elif i + 4 < SIZE and board[i+1:i+5, j].tolist() == [player] * 4:
                    return True
                # 左上到右下方向
                elif i + 4 < SIZE and j + 4 < SIZE and np.diag(board[i+1:i+5, j+1:j+5]).tolist() == [player] * 4:
                    return True
                # 右上到左下方向
                elif i - 4 >= 0 and j + 4 < SIZE and np.diag(np.fliplr(board[i-3:i+1, j+1:j+5])).tolist() == [player] * 4:
                    return True
    return False

# 评估函数
def evaluate(board, player):
    score = 0
    for i in range(SIZE):
        for j in range(SIZE):
            if board[i][j] == player:
                for dx, dy in [(0, 1), (1, 0), (1, 1), (1, -1)]:
                    cnt = 0
                    for k in range(1, 5):
                        x, y = i + k * dx, j + k * dy
                        if x < 0 or x >= SIZE or y < 0 or y >= SIZE or board[x][y] != player:
                            break
                        cnt += 1
                    score += SCORES.get((cnt, player), 0)
    return score

# 极大极小值搜索
def search(board, player, depth, alpha, beta):
    if depth == 0:
        return None, evaluate(board, player)
    best_move = None
    if player == BLACK:
        best_score = -np.inf
        for i in range(SIZE):
            for j in range(SIZE):
                if board[i][j] == EMPTY:
                    board[i][j] = BLACK
                    _, score = search(board, WHITE, depth-1, alpha, beta)
                    board[i][j] = EMPTY
                    if score > best_score:
                        best_score = score
                        best_move = (i, j)
                    alpha = max(alpha, best_score)
                    if alpha >= beta:
                        return best_move, best_score
    else:
        best_score = np.inf
        for i in range(SIZE):
            for j in range(SIZE):
                if board[i][j] == EMPTY:
                    board[i][j] = WHITE
                    _, score = search(board, BLACK, depth-1, alpha, beta)
                    board[i][j] = EMPTY
                    if score < best_score:
                        best_score = score
                        best_move = (i, j)
                    beta = min(beta, best_score)
                    if alpha >= beta:
                        return best_move, best_score
    return best_move, best_score

# 人机对战
def play():
    player = BLACK
    while True:
        if player == BLACK:
            row, col = map(int, input('Your move (row, col): ').split())
            if board[row][col] != EMPTY:
                print('Invalid move!')
                continue
            board[row][col] = BLACK
        else:
            print('AI is thinking...')
            move, _ = search(board, WHITE, SEARCH_DEPTH, -np.inf, np.inf)
            row, col = move
            board[row][col] = WHITE
            print(f'AI moves ({row}, {col})')
        print(board)
        if is_win(board, player):
            print('You win!' if player == BLACK else 'AI wins!')
            break
        if np.count_nonzero(board) == SIZE * SIZE:
            print('Draw!')
            break
        player = WHITE if player == BLACK else BLACK

# 开始游戏
play()

使用该代码进行人机对战时，可以通过输入行列坐标来下棋子，例如：

Your move (row, col): 7 7

AI会自动思考并下出最优的落子位置，并输出到控制台上，例如：

AI is thinking...
AI moves (8, 8)

可以通过修改`SEARCH_DEPTH`来调整搜索深度，从而影响AI的棋力。但要注意，搜索深度过大会导致计算时间过长，影响游戏体验。