def run_test():
board = Board()
bot1 = Bot(1, board)
bot2 = Bot(2, board)
while not board.get_winner():
print board
if not board.can_be_played():
break
bot1.act()
print board
if not board.can_be_played():
break
bot2.act()
print board
bot1.update_strategy()
bot2.update_strategy()
print 'bot %d wins' % board.get_winner()
def run_test():
board = Board()
bot1 = Bot(1, board)
bot2 = Bot(2, board)
while not board.get_winner():
print board
if not board.can_be_played():
break
bot1.act()
print board
if not board.can_be_played():
break
bot2.act()
print board
bot1.update_strategy()
bot2.update_strategy()
print "bot %d wins" % board.get_winner()
class App(gtk.Window):
def __init__(self):
super(App, self).__init__()
self.set_title('Tic-Tac-Toe')
self.connect('destroy', gtk.main_quit)
self.set_position(gtk.WIN_POS_CENTER)
self.set_default_size(250,250)
self.show()
self.init_widgets()
self.game_count = 0
self.start_new_game()
def init_widgets(self):
vbox = gtk.VBox()
self.add(vbox)
self.button_position = {}
self.buttons = {}
for i in xrange(3):
hbox = gtk.HBox()
vbox.add(hbox)
for j in xrange(3):
button = gtk.Button()
self.button_position[button] = (i,j)
hbox.add(button)
button.connect('clicked', self.game_button_clicked)
button.show()
self.buttons[i, j] = button
hbox.show()
hbox = gtk.HBox()
restart_button = gtk.Button('restart')
restart_button.connect('clicked', self.start_new_game)
restart_button.show()
hbox.add(restart_button)
self.score_label= gtk.Label('0 - 0')
self.score_label.show()
hbox.add(self.score_label)
vbox.add(hbox)
hbox.show()
vbox.show()
def draw_board_state(self):
for i in xrange(3):
for j in xrange(3):
button = self.buttons[i,j]
if self.board.board[i][j] == 0:
label = ' '
elif self.board.board[i][j] == 1:
label = 'X'
else:
label = '0'
button.set_label(label)
def start_new_game(self, widget=None):
for button in self.button_position:
button.set_label(' ')
self.board = Board()
if self.game_count % 2 == 0:
self.bot = Bot(1, self.board)
self.bot.act()
self.draw_board_state()
self.player_color = 2
else:
self.bot = Bot(2, self.board)
self.player_color = 1
self.game_count += 1
def game_button_clicked(self, button):
position = self.button_position[button]
if position not in self.board.get_possible_actions():
return
self.board.move(self.player_color, self.button_position[button])
self.draw_board_state()
if not self.board.can_be_played():
return
self.bot.act()
self.draw_board_state()
if not self.board.get_winner() != 0:
return
from bot import Bot bot = Bot() bot.act() bot.end()
class App(gtk.Window):
def __init__(self):
super(App, self).__init__()
self.set_title('Tic-Tac-Toe')
self.connect('destroy', gtk.main_quit)
self.set_position(gtk.WIN_POS_CENTER)
self.set_default_size(250, 250)
self.show()
self.init_widgets()
self.game_count = 0
self.start_new_game()
def init_widgets(self):
vbox = gtk.VBox()
self.add(vbox)
self.button_position = {}
self.buttons = {}
for i in xrange(3):
hbox = gtk.HBox()
vbox.add(hbox)
for j in xrange(3):
button = gtk.Button()
self.button_position[button] = (i, j)
hbox.add(button)
button.connect('clicked', self.game_button_clicked)
button.show()
self.buttons[i, j] = button
hbox.show()
hbox = gtk.HBox()
restart_button = gtk.Button('restart')
restart_button.connect('clicked', self.start_new_game)
restart_button.show()
hbox.add(restart_button)
self.score_label = gtk.Label('0 - 0')
self.score_label.show()
hbox.add(self.score_label)
vbox.add(hbox)
hbox.show()
vbox.show()
def draw_board_state(self):
for i in xrange(3):
for j in xrange(3):
button = self.buttons[i, j]
if self.board.board[i][j] == 0:
label = ' '
elif self.board.board[i][j] == 1:
label = 'X'
else:
label = '0'
button.set_label(label)
def start_new_game(self, widget=None):
for button in self.button_position:
button.set_label(' ')
self.board = Board()
if self.game_count % 2 == 0:
self.bot = Bot(1, self.board)
self.bot.act()
self.draw_board_state()
self.player_color = 2
else:
self.bot = Bot(2, self.board)
self.player_color = 1
self.game_count += 1
def game_button_clicked(self, button):
position = self.button_position[button]
if position not in self.board.get_possible_actions():
return
self.board.move(self.player_color, self.button_position[button])
self.draw_board_state()
if not self.board.can_be_played():
return
self.bot.act()
self.draw_board_state()
if not self.board.get_winner() != 0:
return
def selenium_driver(selenium_url):
mobile_emulation = {"deviceName": "iPhone 5"}
chrome_options = webdriver.ChromeOptions()
chrome_options.add_argument('--incognito')
chrome_options.add_argument(
"start-maximized") # open Browser in maximized mode
chrome_options.add_argument("disable-infobars") # disabling infobars
chrome_options.add_argument("--disable-extensions") # disabling extensions
chrome_options.add_argument("--no-sandbox")
chrome_options.add_argument(
"--disable-dev-shm-usage") # overcome limited resource problems
chrome_options.add_experimental_option("excludeSwitches",
['enable-automation'])
chrome_options.add_experimental_option("mobileEmulation", mobile_emulation)
selenium_driver = webdriver.Remote(
command_executor=selenium_url,
desired_capabilities=chrome_options.to_capabilities())
return selenium_driver
bot = Bot(selenium_local_session=False)
selenium_url = "http://%s:%d/wd/hub" % (os.environ.get('SELENIUM',
'selenium'), 4444)
bot.set_selenium_remote_session(selenium_url=selenium_url,
selenium_driver=selenium_driver(selenium_url))
bot.act('http://web:5000')
# bot.end()
class Test:
def __init__(self):
self.sample_batch_size = 32
self.episodes = 10
self.env = PivitEnv(6)
self.state_size = 6 #self.env.observation_space.shape[0]
self.action_size = 6 #self.env.action_space.n
self.dqnagent = DQNAgent(self.state_size)
self.minimax = Bot()
def testDQN(self):
"""Testar DQN jogando contra si"""
dqn_points = 0
for index_episode in range(self.episodes):
state = self.env.reset()
#state = np.reshape(state, [1, self.state_size])
done = False
index = 0
while not done:
action = self.dqnagent.act(state, index % 2)
if action == None:
break
next_state, reward, done, _ = self.env.step(action)
if reward >= 0:
print("DQNAgent made a valid move")
reward = 1
self.dqnagent.memorize(change_colors(state), action, reward,
change_colors(next_state), done)
dqn_points += reward
state = next_state
index += 1
if index == 100:
break
self.dqnagent.replay(self.sample_batch_size)
return dqn_points
def testMinimaxDQN(self):
"""Testar DQN jogando contra MinMax"""
dqn_points = 0
minmax_points = 0
for index_episode in range(self.episodes):
state = self.env.reset()
#state = np.reshape(state, [1, self.state_size])
done = False
index = 0
while not done:
action = self.minimax.act(box_to_board(state), 1, 0)
if action == None:
break
next_state, reward, done, _ = self.env.step(action)
self.dqnagent.memorize(state, action, reward + 1, next_state,
done)
minmax_points += reward
state = next_state
action = self.dqnagent.act(state, 1)
if action == None:
break
next_state, reward, done, _ = self.env.step(action)
if reward >= 0:
print("DQNAgent made a valid move")
reward = 1
self.dqnagent.memorize(change_colors(state), action, reward,
change_colors(next_state), done)
dqn_points += reward
state = next_state
index += 1
if index == 100:
break
self.dqnagent.replay(self.sample_batch_size)
return minmax_points, dqn_points
def testMinimax(self):
try:
for index_episode in range(self.episodes):
state = self.env.reset()
done = False
index = 0
while not done:
action = self.minimax.act(box_to_board(state), 1,
index % 2)
if action == None:
break
next_state, reward, done, _ = self.env.step(action)
self.dqnagent.memorize(state, action, reward, next_state,
done)
state = next_state
index += 1
if index == 100:
break
print("Episode", index_episode, "Number of moves:", index + 1)
self.dqnagent.replay(self.sample_batch_size)
finally:
self.dqnagent.save_model()
