Exemplos de check_for_victory em Python

Exemplo n.º 1

Arquivo: minimax.py Projeto: Bluther/rota

def _optimal_placement_helper(state, depth=config.MAX_DEPTH, alpha=-inf, beta=inf):
	'''
	Helper function for optimal_placement's recursion.
	Returns the max/min value taking that path can produce.
	'''

	global wincount
	global losscount
	global zerocount

	state.winner = gamestate.check_for_victory(state)
	#Return the score of a win, modified by the depth (to encourage continuing)
	if state.winner != 0:
		if state.winner > 0:
			wincount += 1
			return config.PLAYERWIN - depth
		elif state.winner < 0:
			losscount += 1
			return config.COMPWIN + depth

	if state.num_pieces >= (config.MAX_PIECES * 2):
		return _optimal_move_helper(state, depth)

	available_spots = _all_possible_placements(state)

	if state.turn == 0:
		#Player's turn
		v = alpha
		for spot in [x - 1 for x in available_spots]:
			#Make next iteration's state
			copystate = copy.deepcopy(state)
			copystate.board[spot] = 'p'
			copystate.num_pieces += 1
			copystate.turn = 1

			#Use new state in the next iteration
			v_prime = _optimal_placement_helper(copystate, depth-1, v, beta)
			if v_prime > v:
				v = v_prime
			if v > beta:
				return beta
		return v
	else:
		#Computer's turn
		v = beta
		for spot in [x - 1 for x in available_spots]:
			#Make next iteration's state
			copystate = copy.deepcopy(state)
			copystate.board[spot] = 'c'
			copystate.num_pieces += 1
			copystate.turn = 0

			#Use new state in the next iteration
			v_prime = _optimal_placement_helper(copystate, depth-1, alpha, v)
			if v_prime < v:
				v = v_prime
			if v < alpha:
				return alpha
		return v

Exemplo n.º 2

Arquivo: minimax.py Projeto: Bluther/rota

def _optimal_move_helper(state, depth=config.MAX_DEPTH, alpha=-inf, beta=inf):
	
	state.winner = gamestate.check_for_victory(state)

	global wincount
	global losscount
	global zerocount
	#Max depth, just return 0. See config note for this.
	if state.winner != 0:
		if state.winner > 0:
			wincount += 1
			return config.PLAYERWIN - depth
		elif state.winner < 0:
			losscount += 1
			return config.COMPWIN + depth
	elif depth <= 0:
		zerocount += 1
		return 0

	available_moves = _all_possible_moves(state)

	'''
	This loop makes a copy of the state for each possible destination,
		moves the piece to that destination, and updates the statedict
		to the proper value for that move.
	'''
	if state.turn == 0:
		v = alpha
		for move in available_moves:
			#source is the source piece's location, adjusted to be 0-indexed
			source = move[0]
			#dests is a list of destinations source can move to, adjusted to be 0-indexed
			dests = move[1:]
			for dest in dests:
				#constructing the next iteration's state
				copystate = copy.deepcopy(state)
				copystate.board[source - 1] = '-'
				copystate.board[dest - 1] = 'p'
				copystate.turn = 1

				#running the next iteration
				v_prime = _optimal_move_helper(copystate, depth-1, v, beta)
				if v_prime > v:
					v = v_prime
				if v > beta:
					return beta
		return v

	else:
		v = inf
		for move in available_moves:
			#source is the source piece's location, adjusted to be 0-indexed
			source = move[0]
			#dests is a list of destinations source can move to, adjusted to be 0-indexed
			dests = move[1:]
			for dest in dests:
				#constructing the next iteration's state
				copystate = copy.deepcopy(state)
				copystate.board[source - 1] = '-'
				copystate.board[dest - 1] = 'c'
				copystate.turn = 0

				#Running the next iteration
				v_prime = _optimal_move_helper(copystate, depth-1, alpha, v)
				if v_prime < v:
					v = v_prime
				if v < alpha:
					return alpha
		return v