def test_policy_with_heuristic(self):
board_size = (3, 3)
def heuristic(board_state, max_score, min_score, is_max):
return -min_score
players = {
"p1":
dnbpy.Level3MinimaxPolicy(board_size,
3,
random_state=0,
heuristic=heuristic),
"p2":
dnbpy.Level2HeuristicPolicy(board_size, random_state=0)
}
game = dnbpy.Game(board_size, ["p1", "p2"])
while not game.is_finished():
current_player = game.get_current_player()
opp_player = "p2" if current_player == "p1" else "p1"
edge = players[current_player].select_edge(
game.get_board_state(), game.get_score(current_player),
game.get_score(opp_player))
game.select_edge(edge, current_player)
self.assertEqual(5, game.get_score("p1"))
self.assertEqual(4, game.get_score("p2"))
def test_random_vs_level2(self):
board_size = (2, 2)
players = {
0: dnbpy.RandomPolicy(random_state=0),
1: dnbpy.Level2HeuristicPolicy(board_size, random_state=0)
}
result = dnbpy.duel(board_size, players)
self.assertEqual({'lost': 3, 'won': 19, 'tied': 2}, result[1])
board_size = (3, 3)
players = {
0: dnbpy.RandomPolicy(random_state=0),
1: dnbpy.Level2HeuristicPolicy(board_size, random_state=0)
}
result = dnbpy.duel(board_size, players)
self.assertEqual({'lost': 0, 'won': 48, 'tied': 0}, result[1])
def test_level2_vs_level3_variable_depth(self):
board_size = (3, 3)
players = {
0: dnbpy.Level2HeuristicPolicy(board_size, random_state=0),
1: dnbpy.Level3MinimaxPolicy(board_size, random_state=0)
}
result = dnbpy.duel(board_size, players)
self.assertEqual({'tied': 0, 'lost': 7, 'won': 41}, result[1])
def test_level2_vs_level3update_alpha(self):
board_size = (3, 3)
players = {
0:
dnbpy.Level2HeuristicPolicy(board_size, random_state=0),
1:
dnbpy.Level3MinimaxPolicy(board_size,
3,
random_state=0,
update_alpha=True)
}
result = dnbpy.duel(board_size, players)
self.assertEqual({'tied': 0, 'lost': 13, 'won': 35}, result[1])
def test_select_edge(self):
policy = dnbpy.Level2HeuristicPolicy(board_size=(3, 3))
"""
*---------*---------*---------*
| |
| $L2 | 5 6
| |
*---------*---------* 9 *
|
10 11 12 |
|
*---------* 15 * 16 *
|
17 18 19 |
|
* 21 * 22 *---------*
"""
board_state = [
1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0,
0, 1
]
edge = policy.select_edge(board_state)
self.assertEqual(5, edge)
"""
*---------*---------*---------*
| | |
| $L2 | $L2 | 6
| | |
*---------*---------* 9 *
|
10 11 12 |
|
*---------* 15 * 16 *
|
17 18 19 |
|
* 21 * 22 *---------*
"""
board_state = [
1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0,
0, 1
]
edge = policy.select_edge(board_state)
self.assertTrue(edge not in [6, 9, 10, 11, 16, 19])
def test_policy_with_no_depth(self):
board_size = (3, 3)
players = {
"p1": dnbpy.Level3MinimaxPolicy(board_size, random_state=0),
"p2": dnbpy.Level2HeuristicPolicy(board_size, random_state=0)
}
game = dnbpy.Game(board_size, ["p1", "p2"])
while not game.is_finished():
current_player = game.get_current_player()
opp_player = "p2" if current_player == "p1" else "p1"
edge = players[current_player].select_edge(
game.get_board_state(), game.get_score(current_player),
game.get_score(opp_player))
game.select_edge(edge, current_player)
self.assertEqual(7, game.get_score("p1"))
self.assertEqual(2, game.get_score("p2"))
def play():
print("DNBPy - Play")
num_players = int(input("How many players?: "))
if num_players < 2:
print("Error: there must be at least two players")
sys.exit(0)
players = []
for n in range(num_players):
player = input("player {} name: ".format(n + 1))
players.append(player)
if len(set(players)) == 1:
print("Error: player names must be unique")
sys.exit(0)
board_rows = int(input("Number of board rows: "))
if board_rows < 1:
print("Error: there must be at least one row")
sys.exit(0)
board_cols = int(input("Number of board columns: "))
if board_cols < 1:
print("Error: there must be at least one column")
sys.exit(0)
minimax_depth = None
if "$L3" in players:
val = input("Minimax depth (leave empty for variable depth): ")
if len(val.strip()) > 0:
minimax_depth = int(val)
if minimax_depth < 1:
print("Error: minimax depth must be greater than 0")
sys.exit(0)
num_playouts = 100
if "$mcts" in players:
num_playouts = int(input("Number of playouts: "))
if num_playouts < 1:
print("Error: number of playouts must be greater than 0")
sys.exit(0)
print("preparing game...")
board_size = (board_rows, board_cols)
game = dnbpy.Game(board_size, players)
print(game)
computer_players = {
"$random":
dnbpy.RandomPolicy(),
"$L1":
dnbpy.Level1HeuristicPolicy(board_size=board_size),
"$L2":
dnbpy.Level2HeuristicPolicy(board_size=board_size),
"$L3":
dnbpy.Level3MinimaxPolicy(board_size=board_size,
depth=minimax_depth,
update_alpha=True),
"$mcts":
dnbpy.MCTSPolicy(board_size=board_size, num_playouts=num_playouts)
}
while not game.is_finished():
current_player = game.get_current_player()
if current_player in computer_players:
# get the first player that isn't the current player
opp_player = [p for p in players if p != current_player][0]
move = computer_players[current_player].select_edge(
game.get_board_state(), game.get_score(current_player),
game.get_score(opp_player))
game.select_edge(move, current_player)
print("player %s selects edge %s" % (current_player, move))
else:
try:
move = int(
input("{} select your move: ".format(current_player)))
game.select_edge(move, current_player)
except Exception:
print("illegal move selection.. select again")
print(game)
