created working reinforcement learning model

minimax/minimaxAlgo.py

@@ -1,56 +1,38 @@
 import random
-from copy import deepcopy
 from math import inf
 
 from utilities.constants import GREEN, WHITE
 
 
-class MiniMax():
+class MiniMax:
 
-    def AI(self, board, depth, maxPlayer, gameManager):
-        if depth == 0 or board.winner() is not None:
-            return board.scoreOfTheBoard(), board
+    def AI(self, depth, maxPlayer, gameManager):
+        if depth == 0 or gameManager.board.winner() is not None:
+            return gameManager.board.scoreOfTheBoard(), gameManager.board
 
-        if maxPlayer:
+        if type(maxPlayer) == int:
             maxEval = -inf
             bestMove = None
-            for move in self.getAllMoves(board, maxPlayer):
-                evaluation = self.AI(move, depth - 1, False, gameManager)[0]
+            for move in gameManager.board.getAllMoves(maxPlayer):
+                evaluation = self.AI(depth - 1, False, gameManager)[0]
                 maxEval = max(maxEval, evaluation)
                 if maxEval > evaluation:
                     bestMove = move
                 if maxEval == evaluation:
+                    # bestMove = move
                     bestMove = bestMove if random.choice([True, False]) else move
             return maxEval, bestMove
         else:
             minEval = inf
             bestMove = None
             colour = WHITE if gameManager.turn == GREEN else GREEN
-            for move in self.getAllMoves(board, colour):
-                evaluation = self.AI(move, depth - 1, True, gameManager)[0]
+            for move in gameManager.board.getAllMoves(colour):
+                evaluation = self.AI(depth - 1, True, gameManager)[0]
                 minEval = min(minEval, evaluation)
                 if minEval < evaluation:
                     bestMove = move
                 if minEval == evaluation:
+                    # bestMove = move
                     bestMove = bestMove if random.choice([True, False]) else move
 
             return minEval, bestMove
-
-    def _simulateMove(self, piece, move, board, skip):
-        board.move(piece, move[0], move[1])
-        if skip:
-            board.remove(skip)
-
-        return board
-
-    def getAllMoves(self, board, colour):
-        moves = []
-
-        for piece in board.getAllPieces(colour):
-            validMoves = board.getValidMoves(piece)
-            for move, skip in validMoves.items():
-                tempBoard = deepcopy(board)
-                tempPiece = tempBoard.getPiece(piece.row, piece.col)
-                newBoard = self._simulateMove(tempPiece, move, tempBoard, skip)
-                moves.append(newBoard)
-        return moves