def test_joint_action_probabilities_failure_on_seq_game(self):
"""Test failure of child on sequential games."""
game = pyspiel.load_game("kuhn_poker")
with self.assertRaises(AssertionError):
list(
policy.joint_action_probabilities(
game.new_initial_state(),
policy.UniformRandomPolicy(game)))
def _transitions(state, policies):
"""Returns iterator over (action, prob) from the given state."""
if state.is_chance_node():
return state.chance_outcomes()
elif state.is_simultaneous_node():
return policy.joint_action_probabilities(state, policies)
else:
player = state.current_player()
return policies[player].action_probabilities(state).items()
def test_joint_action_probabilities(self):
"""Test expected behavior of joint_action_probabilities."""
game = pyspiel.load_game("python_iterated_prisoners_dilemma")
uniform_policy = policy.UniformRandomPolicy(game)
joint_action_probs = policy.joint_action_probabilities(
game.new_initial_state(), uniform_policy)
self.assertCountEqual(list(joint_action_probs), [
((0, 0), 0.25),
((1, 1), 0.25),
((1, 0), 0.25),
((0, 1), 0.25),
])