def choose_action(self, board: Board):
while True:
print(board.pretty())
row = input('Row: ')
col = input('Col: ')
try:
board.pick((int(row) - 1, int(col) - 1))
break
except ValueError:
print(f'Bad action: {row, col}. Try again.')
pass
def play(self):
board = Board(*self.size)
random.shuffle(self.players)
players = itertools.cycle(self.players)
current_player = None
while not board.ended:
current_player = next(players)
current_player.choose_action(board)
current_player.end_round(0)
next(players).end_round(1)
def test_board_created_with_all_available_actions(self):
rows = 3
cols = 4
self.assertEqual(list(itertools.product(range(rows), range(cols))),
Board(rows, cols).available_actions())
def test_pick_shrinks_list_when_whole_row_is_picked(self):
board = Board(4, 1)
self.assertEqual(4, len(board._grid))
board.pick((2, 0))
self.assertEqual(2, len(board._grid))
def test_pick_same_cell_throws(self):
board = Board(2, 2)
board.pick((1, 1))
with self.assertRaisesRegex(ValueError, 'Invalid action'):
board.pick((1, 1))
def test_pick_removes_cell(self):
board = Board(2, 2)
self.assertEqual(4, len(board.available_actions()))
board.pick((1, 1))
self.assertEqual(3, len(board.available_actions()))
board.pick((1, 0))
self.assertEqual(2, len(board.available_actions()))
board.pick((0, 1))
self.assertEqual(1, len(board.available_actions()))
def choose_action(self, board: Board):
action = self.agent.choose_action(
board.available_actions(),
lambda x: self._copy_board_and_get_new_hash(x, board)
)
board.pick(action)
def _copy_board_and_get_new_hash(action, board: Board):
new_board = board.copy()
new_board.pick(action)
return new_board.hash