Combinatorial Games. Episode 2: Tic-Tac-Toe Problems in Leetcode and Solve the Game using Minimax

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

# AC
from typing import List

class Solution:
    def tictactoe(self, moves: List[List[int]]) -> str:
        board = [[0] * 3 for _ in range(3)]
        for idx, xy in enumerate(moves):
            player = 1 if idx % 2 == 0 else -1
            board[xy[0]][xy[1]] = player

        turn = 0
        row, col = [0, 0, 0], [0, 0, 0]
        diag1, diag2 = False, False
        for r in range(3):
            for c in range(3):
                turn += board[r][c]
                row[r] += board[r][c]
                col[c] += board[r][c]
                if r == c:
                    diag1 += board[r][c]
                if r + c == 2:
                    diag2 += board[r][c]

        oWin = any(row[r] == 3 for r in range(3)) or any(col[c] == 3 for c in range(3)) or diag1 == 3 or diag2 == 3
        xWin = any(row[r] == -3 for r in range(3)) or any(col[c] == -3 for c in range(3)) or diag1 == -3 or diag2 == -3

        return "A" if oWin else "B" if xWin else "Draw" if len(moves) == 9 else "Pending"

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53

# AC
from typing import List

class Solution:
    def checkWin(self, r: int, c: int) -> bool:
        north = self.getConnectedNum(r, c, -1, 0)
        south = self.getConnectedNum(r, c, 1, 0)

        east = self.getConnectedNum(r, c, 0, 1)
        west = self.getConnectedNum(r, c, 0, -1)

        south_east = self.getConnectedNum(r, c, 1, 1)
        north_west = self.getConnectedNum(r, c, -1, -1)

        north_east = self.getConnectedNum(r, c, -1, 1)
        south_west = self.getConnectedNum(r, c, 1, -1)

        if (north + south + 1 >= 3) or (east + west + 1 >= 3) or \
                (south_east + north_west + 1 >= 3) or (north_east + south_west + 1 >= 3):
            return True
        return False

    def getConnectedNum(self, r: int, c: int, dr: int, dc: int) -> int:
        player = self.board[r][c]
        result = 0
        i = 1
        while True:
            new_r = r + dr * i
            new_c = c + dc * i
            if 0 <= new_r < 3 and 0 <= new_c < 3:
                if self.board[new_r][new_c] == player:
                    result += 1
                else:
                    break
            else:
                break
            i += 1
        return result

    def tictactoe(self, moves: List[List[int]]) -> str:
        self.board = [[0] * 3 for _ in range(3)]
        for idx, xy in enumerate(moves):
            player = 1 if idx % 2 == 0 else -1
            self.board[xy[0]][xy[1]] = player

        # only check last move
        r, c = moves[-1]
        win = self.checkWin(r, c)
        if win:
            return "A" if len(moves) % 2 == 1 else "B"

        return "Draw" if len(moves) == 9 else "Pending"

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

# AC
from typing import List

class Solution:

    def convertCell(self, c:str):
        return 1 if c == 'X' else -1 if c == 'O' else 0

    def validTicTacToe(self, board: List[str]) -> bool:
        turn = 0
        row, col = [0, 0, 0], [0, 0, 0]
        diag1, diag2 = False, False
        for r in range(3):
            for c in range(3):
                turn += self.convertCell(board[r][c])
                row[r] += self.convertCell(board[r][c])
                col[c] += self.convertCell(board[r][c])
                if r == c:
                    diag1 += self.convertCell(board[r][c])
                if r + c == 2:
                    diag2 += self.convertCell(board[r][c])

        xWin = any(row[r] == 3 for r in range(3)) or any(col[c] == 3 for c in range(3)) or diag1 == 3 or diag2 == 3
        oWin = any(row[r] == -3 for r in range(3)) or any(col[c] == -3 for c in range(3)) or diag1 == -3 or diag2 == -3
        if (xWin and turn == 0) or (oWin and turn == 1):
            return False
        return (turn == 0 or turn == 1) and (not xWin or not oWin)

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38

# AC
class TicTacToe:

    def __init__(self, n:int):
        """
        Initialize your data structure here.
        :type n: int
        """
        self.row, self.col, self.diag1, self.diag2, self.n = [0] * n, [0] * n, 0, 0, n

    def move(self, row:int, col:int, player:int) -> int:
        """
        Player {player} makes a move at ({row}, {col}).
        @param row The row of the board.
        @param col The column of the board.
        @param player The player, can be either 1 or 2.
        @return The current winning condition, can be either:
                0: No one wins.
                1: Player 1 wins.
                2: Player 2 wins.
        """
        if player == 2:
            player = -1

        self.row[row] += player
        self.col[col] += player
        if row == col:
            self.diag1 += player
        if row + col == self.n - 1:
            self.diag2 += player

        if self.n in [self.row[row], self.col[col], self.diag1, self.diag2]:
            return 1
        if -self.n in [self.row[row], self.col[col], self.diag1, self.diag2]:
            return 2
        return 0

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58

class ConnectNGame:

    PLAYER_A = 1
    PLAYER_B = -1
    AVAILABLE = 0
    RESULT_TIE = 0
    RESULT_A_WIN = 1
    RESULT_B_WIN = -1

    def __init__(self, N:int = 3, board_size:int = 3):
        assert N <= board_size
        self.N = N
        self.board_size = board_size
        self.board = [[ConnectNGame.AVAILABLE] * board_size for _ in range(board_size)]
        self.gameOver = False
        self.gameResult = None
        self.currentPlayer = ConnectNGame.PLAYER_A
        self.remainingPosNum = board_size * board_size
        self.actionStack = []

    def move(self, r: int, c: int) -> int:
        """

        :param r:
        :param c:
        :return: None: game ongoing
        """
        assert self.board[r][c] == ConnectNGame.AVAILABLE
        self.board[r][c] = self.currentPlayer
        self.actionStack.append((r, c))
        self.remainingPosNum -= 1
        if self.checkWin(r, c):
            self.gameOver = True
            self.gameResult = self.currentPlayer
            return self.currentPlayer
        if self.remainingPosNum == 0:
            self.gameOver = True
            self.gameResult = ConnectNGame.RESULT_TIE
            return ConnectNGame.RESULT_TIE
        self.currentPlayer *= -1

    def undo(self):
        if len(self.actionStack) > 0:
            lastAction = self.actionStack.pop()
            r, c = lastAction
            self.board[r][c] = ConnectNGame.AVAILABLE
            self.currentPlayer = ConnectNGame.PLAYER_A if len(self.actionStack) % 2 == 0 else ConnectNGame.PLAYER_B
            self.remainingPosNum += 1
            self.gameOver = False
            self.gameResult = None
        else:
            raise Exception('No lastAction')

    def getAvailablePositions(self) -> List[Tuple[int, int]]:
        return [(i,j) for i in range(self.board_size) for j in range(self.board_size) if self.board[i][j] == ConnectNGame.AVAILABLE]

    def getStatus(self) -> Tuple[Tuple[int, ...]]:
        return tuple([tuple(self.board[i]) for i in range(self.board_size)])

1
2
3
4
5
6
7
8

class Strategy(ABC):

    def __init__(self):
        super().__init__()

    @abstractmethod
    def action(self, game: ConnectNGame) -> Tuple[int, Tuple[int, int]]:
        pass

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38

class MinimaxStrategy(Strategy):
    def action(self, game: ConnectNGame) -> Tuple[int, Tuple[int, int]]:
        self.game = copy.deepcopy(game)
        result, move = self.minimax()
        return result, move

    def minimax(self) -> Tuple[int, Tuple[int, int]]:
        game = self.game
        bestMove = None
        assert not game.gameOver
        if game.currentPlayer == ConnectNGame.PLAYER_A:
            ret = -math.inf
            for pos in game.getAvailablePositions():
                move = pos
                result = game.move(*pos)
                if result is None:
                    assert not game.gameOver
                    result, oppMove = self.minimax()
                game.undo()
                ret = max(ret, result)
                bestMove = move if ret == result else bestMove
                if ret == 1:
                    return 1, move
            return ret, bestMove
        else:
            ret = math.inf
            for pos in game.getAvailablePositions():
                move = pos
                result = game.move(*pos)
                if result is None:
                    assert not game.gameOver
                    result, oppMove = self.minimax()
                game.undo()
                ret = min(ret, result)
                bestMove = move if ret == result else bestMove
                if ret == -1:
                    return -1, move
            return ret, bestMove

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58

class CountingMinimaxStrategy(Strategy):
    def action(self, game: ConnectNGame) -> Tuple[int, Tuple[int, int]]:
        self.game = copy.deepcopy(game)
        self.dpMap = {}
        result, move = self.minimax(game.getStatus())
        return result, move

    def minimax(self, gameStatus: Tuple[Tuple[int, ...]]) -> Tuple[int, Tuple[int, int]]:
        # print(f'Current {len(strategy.dpMap)}')

        if gameStatus in self.dpMap:
            return self.dpMap[gameStatus]

        game = self.game
        bestMove = None
        assert not game.gameOver
        if game.currentPlayer == ConnectNGame.PLAYER_A:
            ret = -math.inf
            for pos in game.getAvailablePositions():
                move = pos
                result = game.move(*pos)
                if result is None:
                    assert not game.gameOver
                    result, oppMove = self.minimax(game.getStatus())
                    self.dpMap[game.getStatus()] = result, oppMove
                else:
                    self.dpMap[game.getStatus()] = result, move
                game.undo()
                ret = max(ret, result)
                bestMove = move if ret == result else bestMove
            self.dpMap[gameStatus] = ret, bestMove
            return ret, bestMove
        else:
            ret = math.inf
            for pos in game.getAvailablePositions():
                move = pos
                result = game.move(*pos)

                if result is None:
                    assert not game.gameOver
                    result, oppMove = self.minimax(game.getStatus())
                    self.dpMap[game.getStatus()] = result, oppMove
                else:
                    self.dpMap[game.getStatus()] = result, move
                game.undo()
                ret = min(ret, result)
                bestMove = move if ret == result else bestMove
            self.dpMap[gameStatus] = ret, bestMove
            return ret, bestMove


if __name__ == '__main__':
    tic_tac_toe = ConnectNGame(N=3, board_size=3)
    strategy = CountingMinimaxStrategy()
    strategy.action(tic_tac_toe)
    print(f'Game States Number {len(strategy.dpMap)}')

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41

class AlphaBetaStrategy(Strategy):
    def action(self, game: ConnectNGame) -> Tuple[int, Tuple[int, int]]:
        self.game = game
        result, move = self.alpha_beta(self.game.getStatus(), -math.inf, math.inf)
        return result, move

    def alpha_beta(self, gameStatus: Tuple[Tuple[int, ...]], alpha:int=None, beta:int=None) -> Tuple[int, Tuple[int, int]]:
        game = self.game
        bestMove = None
        assert not game.gameOver
        if game.currentPlayer == ConnectNGame.PLAYER_A:
            ret = -math.inf
            for pos in game.getAvailablePositions():
                move = pos
                result = game.move(*pos)
                if result is None:
                    assert not game.gameOver
                    result, oppMove = self.alpha_beta(game.getStatus(), alpha, beta)
                game.undo()
                alpha = max(alpha, result)
                ret = max(ret, result)
                bestMove = move if ret == result else bestMove
                if alpha >= beta or ret == 1:
                    return ret, move
            return ret, bestMove
        else:
            ret = math.inf
            for pos in game.getAvailablePositions():
                move = pos
                result = game.move(*pos)
                if result is None:
                    assert not game.gameOver
                    result, oppMove = self.alpha_beta(game.getStatus(), alpha, beta)
                game.undo()
                beta = min(beta, result)
                ret = min(ret, result)
                bestMove = move if ret == result else bestMove
                if alpha >= beta or ret == -1:
                    return ret, move
            return ret, bestMove

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47

class AlphaBetaDPStrategy(Strategy):
    def action(self, game: ConnectNGame) -> Tuple[int, Tuple[int, int]]:
        self.game = game
        self.alphaBetaStack = [(-math.inf, math.inf)]
        result, move = self.alpha_beta_dp(self.game.getStatus())
        return result, move

    @lru_cache(maxsize=None)
    def alpha_beta_dp(self, gameStatus: Tuple[Tuple[int, ...]]) -> Tuple[int, Tuple[int, int]]:
        alpha, beta = self.alphaBetaStack[-1]
        game = self.game
        bestMove = None
        assert not game.gameOver
        if game.currentPlayer == ConnectNGame.PLAYER_A:
            ret = -math.inf
            for pos in game.getAvailablePositions():
                move = pos
                result = game.move(*pos)
                if result is None:
                    assert not game.gameOver
                    self.alphaBetaStack.append((alpha, beta))
                    result, oppMove = self.alpha_beta_dp(game.getStatus())
                    self.alphaBetaStack.pop()
                game.undo()
                alpha = max(alpha, result)
                ret = max(ret, result)
                bestMove = move if ret == result else bestMove
                if alpha >= beta or ret == 1:
                    return ret, move
            return ret, bestMove
        else:
            ret = math.inf
            for pos in game.getAvailablePositions():
                move = pos
                result = game.move(*pos)
                if result is None:
                    assert not game.gameOver
                    self.alphaBetaStack.append((alpha, beta))
                    result, oppMove = self.alpha_beta_dp(game.getStatus())
                    self.alphaBetaStack.pop()
                game.undo()
                beta = min(beta, result)
                ret = min(ret, result)
                bestMove = move if ret == result else bestMove
                if alpha >= beta or ret == -1:
                    return ret, move
            return ret, bestMove