def main():
global board_size
board_size = 3
global t
t = ttt.TTT(board_size)
console.log(t.get_board())
display_board(t, board_size, None, None)
def test_1(self):
tic = ttt.TTT()
assert len(tic.board) == 9
assert all([s == '.' for s in tic.board])
assert tic.state is ttt.game_state(tic)
assert tic.display(
) == '. | . | .\n---------\n. | . | .\n---------\n. | . | .'
assert tic._nelem('.') == 9
assert tic._nelem('x') == 0
assert tic._nelem('o') == 0
self.assertRaises(AssertionError, lambda: tic._nelem('p'))
assert tic.state is ttt.GameStates.unfinished
tic.board = ['x', 'x', 'x', 'x', 'x']
assert tic.state is ttt.GameStates.invalid
tic.board = ['x'] * 9
assert tic.state is ttt.GameStates.invalid
tic.board = ['o'] * 9
assert tic.state is ttt.GameStates.invalid
tic.board = ['x', 'x', '.', '.', '.', '.', '.', '.', '.']
assert tic.state is ttt.GameStates.invalid
tic.board = ['x', 'o', 'o', '.', '.', '.', '.', '.', '.']
assert tic.state is ttt.GameStates.invalid
tic.board = ['x', 'x', 'x', 'o', 'o', '.', '.', '.', '.']
assert tic.state is ttt.GameStates.x_wins
tic.board = ['o', 'x', 'x', 'o', 'o', '.', 'x', 'x', 'o']
assert tic.state is ttt.GameStates.o_wins
tic.board = ['x', 'o', 'x', 'x', 'o', 'o', 'o', 'x', '.']
assert tic.state is ttt.GameStates.draw
# tic.board = ['o', 'x', 'o', 'x', 'o', '.', 'x', 'x', '.']
# assert tic.state is ttt.GameStates.o_wins
tac = ttt.TTT()
tac.play('NE')
assert tac.state is ttt.GameStates.unfinished
print(tac.display())
def tournament(pls, n=1000):
"""
Play tournament of players (everybody against everybody as long as they play opposite same side)
"""
print("Entering tournament with players:")
for p in pls:
print(" " + p.name + " playing " + {1: "x", 4: "o"}[p.pl])
print()
for pair in itertools.combinations(pls, 2):
if pair[0].pl == pair[1].pl:
# Invalid pairing
continue
p1, p2 = sorted(pair, key=lambda x: x.pl)
nx, no, nd = 0, 0, 0
for i in range(n):
t = ttt.TTT()
p1.initGame()
p2.initGame()
while True:
m = p1.move(t, p="greedy")
w = t.move(m, p1.pl)
if not w is None:
break
m = p2.move(t, p="greedy")
w = t.move(m, p2.pl)
if not w is None:
break
if w == 1:
nx += 1
elif w == -1:
nd += 1
else:
no += 1
print("Performance of " + p1.name + " (x) against " + p2.name + " (o)")
print(" x wins %6.2f %% and o wins %6.2f %%" %
(nx / n * 100, no / n * 100))
print(" draws %6.2f %%" % (nd / n * 100))
print(" x looses %6.2f %% and o looses %6.2f %%" %
((n - nx - nd) / n * 100, (n - no - nd) / n * 100))
print()
episode = max(episode, 1) # to avoid divide by zero
odds = 10 / episode
odds = max(odds, 0.01)
odds = min(odds, 1)
return odds
controller = ttt.TTT_vsRandoAI()
qlrn = qlearning.TFQLearning(controller,
comp_randact,
neural.neural,
future_discount=0.5)
(player, results) = qlrn.runEpisodes(100000)
#Play vs human
controller = ttt.TTT()
done = False
while not done:
state = controller.reset()
controller.printBoard()
while controller._isActive:
moves = player.computeQState(state)
move = np.argmax(moves)
print(f"AI played move {move}")
state, _, _ = controller.step(move)
controller.printBoard()
if controller._isActive:
move = int(input("Enter your move [0-8]: "))
state, _, _ = controller.step(move)
assert (not controller._isActive)
controller.printBoard()
def init():
global t
global login_sys
t = ttt.TTT(3)
login_sys = login.Login_Sys(2)
return json.dumps({'result': 'OK'})
def train(pls, rounds):
# Rewards for win, loose, and draw
rw, rl, rd = 1., 0., 0.5
print("Entering training with players:")
for p in pls:
print(" " + p.name + " playing " + {1: "x", 4: "o"}[p.pl])
print()
for pair in itertools.combinations(pls, 2):
if pair[0].pl == pair[1].pl:
# Invalid pairing
continue
p1, p2 = sorted(pair, key=lambda x: x.pl)
print("Training rounds: ", p1, " vs. ", p2)
for j in range(rounds):
# Training can begin
if not silent:
print("Game no. %5d" % j)
t = ttt.TTT()
p1.initGame()
p2.initGame()
# Make moves until victory
while True:
m = p1.move(t)
w = t.move(m, p1.pl)
if not w is None:
break
m = p2.move(t)
w = t.move(m, p2.pl)
if not w is None:
break
if not silent:
t.printHistory()
# Determine winner
if (w == 1):
if not silent: print("Winner is 'x'")
p1.update(rw)
p2.update(rl)
elif (w == 4):
if not silent: print("Winner is 'o'")
p1.update(rl)
p2.update(rw)
elif w == -1:
if not silent: print("Draw")
p1.update(rd)
p2.update(rd)
else:
raise (ValueError("OOpps: " + str(w)))
if not silent:
print()
print("-" * 80)
print()