def test_get_default_q_value(self):
state = 0
action = 0
expected_q_value = 0
qv = learn.QValues(num_states=1, num_actions=1)
q_value = qv.get(state, action)
self.assertEqual(q_value, expected_q_value)
def test_set_q_value(self):
state = 0
action = 0
q_value = 10
qv = learn.QValues(num_states=1, num_actions=1)
qv.set(state, action, q_value)
set_q_value = qv.get(state, action)
self.assertEqual(set_q_value, q_value)
def test_get_max_action_index(self):
q_values = [[1, 2], [4, 3]]
qv = learn.QValues(num_states=2, num_actions=2)
for state, actions in enumerate(q_values):
for action, value in enumerate(actions):
qv.set(state, action, value)
actions = [qv.get_max_action(state) for state in [0, 1]]
self.assertEqual(actions, [1, 0])
def test_initial_num_actions(self):
qv = learn.QValues(num_actions=1)
self.assertEqual(qv.num_actions, 1)
def test_initial_num_states(self):
qv = learn.QValues(num_states=1)
self.assertEqual(qv.num_states, 1)
def test_default_num_actions(self):
qv = learn.QValues()
self.assertEqual(qv.num_actions, 0)