def play(strategy, verbose=False, allow_unavailable_action=True):
"""Plays a single game, using a provided strategy.
Args:
strategy: A function that takes as argument a state and a list of available
actions and returns an action from the list.
allow_unavailable_action: Boolean, whether strategy is passed all actions
or just the available ones.
verbose: If true, prints game states, actions and scores.
Returns:
score, experiences where score is the final score and experiences is the
list Experience instances that represent the collected experience.
"""
game = Game()
state = game.state().copy()
game_over = game.game_over()
experiences = []
while not game_over:
if verbose:
print("Score:", game.score())
game.print_state()
old_state = state
next_action = strategy(
old_state,
range(4) if allow_unavailable_action else game.available_actions())
if game.is_action_available(next_action):
reward = game.do_action(next_action)
state = game.state().copy()
game_over = game.game_over()
if verbose:
print("Action:", ACTION_NAMES[next_action])
print("Reward:", reward)
experiences.append(
Experience(old_state, next_action, reward, state, game_over,
False, game.available_actions()))
else:
experiences.append(
Experience(state, next_action, 0, state, False, True,
game.available_actions()))
if verbose:
print("Score:", game.score())
game.print_state()
print("Game over.")
return game.score(), experiences
def play(strategy, verbose=False, allow_unavailable_action=True):
"""Plays a single game, using a provided strategy.
Args:
strategy: A function that takes as argument a state and a list of available
actions and returns an action from the list.
allow_unavailable_action: Boolean, whether strategy is passed all actions
or just the available ones.
verbose: If true, prints game states, actions and scores.
Returns:
score, experiences where score is the final score and experiences is the
list Experience instances that represent the collected experience.
"""
game = Game()
state = game.state().copy()
game_over = game.game_over()
experiences = []
while not game_over:
if verbose:
print("Score:", game.score())
game.print_state()
old_state = state
next_action = strategy(
old_state, range(4) if allow_unavailable_action
else game.available_actions())
if game.is_action_available(next_action):
reward = game.do_action(next_action)
state = game.state().copy()
game_over = game.game_over()
if verbose:
print("Action:", ACTION_NAMES[next_action])
print("Reward:", reward)
experiences.append(Experience(old_state, next_action, reward, state,
game_over, False, game.available_actions()))
else:
experiences.append(Experience(state, next_action, 0, state, False, True,
game.available_actions()))
if verbose:
print("Score:", game.score())
game.print_state()
print("Game over.")
return game.score(), experiences
def test_available_actions():
state = np.array([[[1, 2, 0], [1, 2, 0], [1, 2, 0]],
[[1, 2, 0], [1, 2, 0], [1, 2, 0]],
[[1, 2, 0], [1, 2, 0], [1, 2, 0]]])
game = Game(state=state)
actions = game.available_actions()
# All actions except left is available
assert actions == [1, 2, 3, 4, 5]
def test_available_actions_none_available():
state = np.array([[[1, 2, 3], [5, 6, 7], [1, 2, 3]],
[[8, 9, 10], [11, 12, 13], [8, 9, 10]],
[[14, 15, 16], [17, 18, 19], [14, 15, 16]]])
game = Game(state=state)
actions = game.available_actions()
# All actions except left is available
assert actions == []
assert game.game_over()
def test_available_actions():
state = np.array([[1, 2, 3, 0],
[1, 2, 3, 0],
[1, 2, 3, 0],
[1, 2, 3, 0]])
game = Game(state=state)
actions = game.available_actions()
# All actions except left is available
assert actions == [1, 2, 3]
def test_available_actions_none_available():
state = np.array([[1, 2, 3, 4],
[5, 6, 7, 8],
[1, 2, 3, 4],
[5, 6, 7, 8]])
game = Game(state=state)
actions = game.available_actions()
# All actions except left is available
assert actions == []
assert game.game_over()
def test_available_actions_none_available():
state = np.array([[1, 2, 3, 4],
[5, 6, 7, 8],
[1, 2, 3, 4],
[5, 6, 7, 8]])
game = Game(state=state)
actions = game.available_actions()
# All actions except left is available
assert actions == []
assert game.game_over()
