mirror of
https://github.com/metafy-social/python-scripts.git
synced 2025-02-22 08:12:13 +00:00
commit
92b14c34d9
5
scripts/MiniMaxAlgo/README.md
Normal file
5
scripts/MiniMaxAlgo/README.md
Normal file
|
@ -0,0 +1,5 @@
|
||||||
|
## MiniMax Algorithm
|
||||||
|
|
||||||
|
It's implemented in such way that the AI minimizes it's loss and maximizes it's winning chances.
|
||||||
|
|
||||||
|

|
102
scripts/MiniMaxAlgo/TicTacToe.py
Normal file
102
scripts/MiniMaxAlgo/TicTacToe.py
Normal file
|
@ -0,0 +1,102 @@
|
||||||
|
import math
|
||||||
|
import time
|
||||||
|
from player import HumanPlayer, SmartComputerPlayer
|
||||||
|
|
||||||
|
|
||||||
|
class TicTacToe():
|
||||||
|
def __init__(self):
|
||||||
|
self.board = self.make_board()
|
||||||
|
self.current_winner = None
|
||||||
|
|
||||||
|
@staticmethod
|
||||||
|
def make_board():
|
||||||
|
return [' ' for _ in range(9)]
|
||||||
|
|
||||||
|
def print_board(self):
|
||||||
|
for row in [self.board[i*3:(i+1) * 3] for i in range(3)]:
|
||||||
|
print('| ' + ' | '.join(row) + ' |')
|
||||||
|
|
||||||
|
@staticmethod
|
||||||
|
def print_board_nums():
|
||||||
|
# 0 | 1 | 2
|
||||||
|
number_board = [[str(i) for i in range(j*3, (j+1)*3)] for j in range(3)]
|
||||||
|
for row in number_board:
|
||||||
|
print('| ' + ' | '.join(row) + ' |')
|
||||||
|
|
||||||
|
def make_move(self, square, letter):
|
||||||
|
if self.board[square] == ' ':
|
||||||
|
self.board[square] = letter
|
||||||
|
if self.winner(square, letter):
|
||||||
|
self.current_winner = letter
|
||||||
|
return True
|
||||||
|
return False
|
||||||
|
|
||||||
|
def winner(self, square, letter):
|
||||||
|
# check the row
|
||||||
|
row_ind = math.floor(square / 3)
|
||||||
|
row = self.board[row_ind*3:(row_ind+1)*3]
|
||||||
|
# print('row', row)
|
||||||
|
if all([s == letter for s in row]):
|
||||||
|
return True
|
||||||
|
col_ind = square % 3
|
||||||
|
column = [self.board[col_ind+i*3] for i in range(3)]
|
||||||
|
# print('col', column)
|
||||||
|
if all([s == letter for s in column]):
|
||||||
|
return True
|
||||||
|
if square % 2 == 0:
|
||||||
|
diagonal1 = [self.board[i] for i in [0, 4, 8]]
|
||||||
|
# print('diag1', diagonal1)
|
||||||
|
if all([s == letter for s in diagonal1]):
|
||||||
|
return True
|
||||||
|
diagonal2 = [self.board[i] for i in [2, 4, 6]]
|
||||||
|
# print('diag2', diagonal2)
|
||||||
|
if all([s == letter for s in diagonal2]):
|
||||||
|
return True
|
||||||
|
return False
|
||||||
|
|
||||||
|
def empty_squares(self):
|
||||||
|
return ' ' in self.board
|
||||||
|
|
||||||
|
def num_empty_squares(self):
|
||||||
|
return self.board.count(' ')
|
||||||
|
|
||||||
|
def available_moves(self):
|
||||||
|
return [i for i, x in enumerate(self.board) if x == " "]
|
||||||
|
|
||||||
|
|
||||||
|
def play(game, x_player, o_player, print_game=True):
|
||||||
|
|
||||||
|
if print_game:
|
||||||
|
game.print_board_nums()
|
||||||
|
|
||||||
|
letter = 'X'
|
||||||
|
while game.empty_squares():
|
||||||
|
if letter == 'O':
|
||||||
|
square = o_player.get_move(game)
|
||||||
|
else:
|
||||||
|
square = x_player.get_move(game)
|
||||||
|
if game.make_move(square, letter):
|
||||||
|
|
||||||
|
if print_game:
|
||||||
|
print(letter + ' makes a move to square {}'.format(square))
|
||||||
|
game.print_board()
|
||||||
|
print('')
|
||||||
|
|
||||||
|
if game.current_winner:
|
||||||
|
if print_game:
|
||||||
|
print(letter + ' wins!')
|
||||||
|
return letter # ends the loop and exits the game
|
||||||
|
letter = 'O' if letter == 'X' else 'X' # switches player
|
||||||
|
|
||||||
|
time.sleep(.8)
|
||||||
|
|
||||||
|
if print_game:
|
||||||
|
print('It\'s a tie!')
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
if __name__ == '__main__':
|
||||||
|
x_player = SmartComputerPlayer('X')
|
||||||
|
o_player = HumanPlayer('O')
|
||||||
|
t = TicTacToe()
|
||||||
|
play(t, x_player, o_player, print_game=True)
|
72
scripts/MiniMaxAlgo/player.py
Normal file
72
scripts/MiniMaxAlgo/player.py
Normal file
|
@ -0,0 +1,72 @@
|
||||||
|
import math
|
||||||
|
import random
|
||||||
|
|
||||||
|
|
||||||
|
class Player:
|
||||||
|
def __init__(self, letter):
|
||||||
|
self.letter = letter
|
||||||
|
|
||||||
|
def get_move(self, game):
|
||||||
|
pass
|
||||||
|
|
||||||
|
|
||||||
|
class HumanPlayer(Player):
|
||||||
|
def __init__(self, letter):
|
||||||
|
super().__init__(letter)
|
||||||
|
|
||||||
|
def get_move(self, game):
|
||||||
|
valid_square = False
|
||||||
|
val = None
|
||||||
|
while not valid_square:
|
||||||
|
square = input(self.letter + '\'s turn. Input move (0-9): ')
|
||||||
|
try:
|
||||||
|
val = int(square)
|
||||||
|
if val not in game.available_moves():
|
||||||
|
raise ValueError
|
||||||
|
valid_square = True
|
||||||
|
except ValueError:
|
||||||
|
print('Invalid square. Try again.')
|
||||||
|
return val
|
||||||
|
|
||||||
|
class SmartComputerPlayer(Player):
|
||||||
|
def __init__(self, letter):
|
||||||
|
super().__init__(letter)
|
||||||
|
|
||||||
|
def get_move(self, game):
|
||||||
|
if len(game.available_moves()) == 9:
|
||||||
|
square = random.choice(game.available_moves())
|
||||||
|
else:
|
||||||
|
square = self.minimax(game, self.letter)['position']
|
||||||
|
return square
|
||||||
|
|
||||||
|
def minimax(self, state, player):
|
||||||
|
max_player = self.letter # yourself
|
||||||
|
other_player = 'O' if player == 'X' else 'X'
|
||||||
|
|
||||||
|
# first we want to check if the previous move is a winner
|
||||||
|
if state.current_winner == other_player:
|
||||||
|
return {'position': None, 'score': 1 * (state.num_empty_squares() + 1) if other_player == max_player else -1 * (
|
||||||
|
state.num_empty_squares() + 1)}
|
||||||
|
elif not state.empty_squares():
|
||||||
|
return {'position': None, 'score': 0}
|
||||||
|
|
||||||
|
if player == max_player:
|
||||||
|
best = {'position': None, 'score': -math.inf} # each score should maximize
|
||||||
|
else:
|
||||||
|
best = {'position': None, 'score': math.inf} # each score should minimize
|
||||||
|
for possible_move in state.available_moves():
|
||||||
|
state.make_move(possible_move, player)
|
||||||
|
sim_score = self.minimax(state, other_player) # simulate a game after making that move
|
||||||
|
|
||||||
|
# undo move
|
||||||
|
state.board[possible_move] = ' '
|
||||||
|
state.current_winner = None
|
||||||
|
sim_score['position'] = possible_move # this represents the move optimal next move
|
||||||
|
|
||||||
|
if player == max_player: # X is max player
|
||||||
|
if sim_score['score'] > best['score']:
|
||||||
|
best = sim_score
|
||||||
|
else:
|
||||||
|
if sim_score['score'] < best['score']:
|
||||||
|
best = sim_score
|
||||||
|
return best
|
Loading…
Reference in New Issue
Block a user