def test_push_valid(self):
ebuffer = ExperienceReplayBuffer(self.BUFFER_SIZE)
self.assertEqual(ebuffer.len, 0)
ebuffer.push(self.sarsa_one)
self.assertEqual(ebuffer.len, 1)
def test_sample_no_padding_no_sample_less_no_pop(self):
ebuffer = ExperienceReplayBuffer(self.BUFFER_SIZE)
self.assertEqual(ebuffer.len, 0)
ebuffer.push(self.sarsa_one)
ebuffer.push(self.sarsa_two)
batch = ebuffer.sample(self.BUFFER_SIZE)
self.assertEqual(len(batch), self.BUFFER_SIZE)
for act, exp in zip(batch, [self.sarsa_one, self.sarsa_two]):
np.testing.assert_array_equal(act.state, exp.state)
self.assertEqual(ebuffer.len, self.BUFFER_SIZE)
batch_two = ebuffer.sample(1)
self.assertEqual(len(batch), self.BUFFER_SIZE)
act = batch_two[0]
exp = self.sarsa_one
np.testing.assert_array_equal(act.state, exp.state)
ebuffer.push(self.sarsa_three)
batch_three = ebuffer.sample(1)
act = batch_three[0]
exp = self.sarsa_two
np.testing.assert_array_equal(act.state, exp.state)
def test_sample_no_padding_no_less_pop(self):
ebuffer = ExperienceReplayBuffer(self.BUFFER_SIZE)
self.assertEqual(ebuffer.len, 0)
ebuffer.push(self.sarsa_one)
ebuffer.push(self.sarsa_two)
batch = ebuffer.sample_and_pop(self.BUFFER_SIZE)
for act, exp in zip(batch, [self.sarsa_one, self.sarsa_two]):
np.testing.assert_array_equal(act.state, exp.state)
self.assertEqual(ebuffer.len, 0)
def test_sample_padding_no_less_pop(self):
ebuffer = ExperienceReplayBuffer(self.BUFFER_SIZE)
pad(ebuffer,
2,
state_col_dim=1,
action_col_dim=1,
reward_col_dim=1,
next_state_col_dim=1)
self.assertEqual(ebuffer.len, self.BUFFER_SIZE)
batch = ebuffer.sample_and_pop(self.BUFFER_SIZE)
for b in batch:
np.testing.assert_array_equal(b.state, np.array([0.0]))
self.assertEqual(ebuffer.len, 0)
ebuffer.push(self.sarsa_one)
self.assertEqual(ebuffer.len, 1)
batch = ebuffer.sample_and_pop(1)
np.testing.assert_array_equal(batch[0].state, self.sarsa_one.state)
self.assertEqual(ebuffer.len, 0)
def test_sample_padding_less_pop(self):
ebuffer = ExperienceReplayBuffer(self.BUFFER_SIZE)
pad(ebuffer,
2,
state_col_dim=1,
action_col_dim=1,
reward_col_dim=1,
next_state_col_dim=1)
self.assertEqual(ebuffer.len, self.BUFFER_SIZE)
try:
ebuffer.sample_and_pop(self.BUFFER_SIZE + 1)
self.assertEqual(1, 0)
except ValueError:
self.assertEqual(1, 1)
except Exception:
self.assertEqual(1, 0)
batch = ebuffer.sample_and_pop(self.BUFFER_SIZE + 1, sample_less=True)
self.assertEqual(ebuffer.len, 0)
for b in batch:
np.testing.assert_array_equal(b.state, np.array([0.0]))
def test_sample_no_padding_less_pop(self):
ebuffer = ExperienceReplayBuffer(self.BUFFER_SIZE)
self.assertEqual(ebuffer.len, 0)
ebuffer.push(self.sarsa_one)
try:
ebuffer.sample_and_pop(self.BUFFER_SIZE)
self.assertEqual(1, 0)
except ValueError:
self.assertEqual(1, 1)
except Exception:
self.assertEqual(1, 0)
batch = ebuffer.sample_and_pop(self.BUFFER_SIZE, sample_less=True)
self.assertEqual(ebuffer.len, 0)
np.testing.assert_array_equal(batch[0].state, self.sarsa_one.state)
def test_sample_no_padding_less_no_pop(self):
ebuffer = ExperienceReplayBuffer(self.BUFFER_SIZE)
self.assertEqual(ebuffer.len, 0)
ebuffer.push(self.sarsa_one)
ebuffer.push(self.sarsa_two)
try:
ebuffer.sample(self.BUFFER_SIZE + 1)
self.assertEqual(1, 0)
except ValueError:
self.assertEqual(1, 1)
except Exception:
self.assertEqual(1, 0)
batch = ebuffer.sample(self.BUFFER_SIZE + 1, sample_less=True)
self.assertEqual(len(batch), self.BUFFER_SIZE)
for act, exp in zip(batch, [self.sarsa_one, self.sarsa_two]):
np.testing.assert_array_equal(act.state, exp.state)