def test_explicity_difficulty_setting(self):
# Set task difficulty explicitly and confirm target and lure size is fixed.
self.assertEqual(PointToTargetContent.difficulty_range, 3)
content = PointToTargetContent(difficulty=0)
self.assertEqual(content.target_sprite.width, 0.2)
self.assertEqual(content.lure_sprite.width, 0.1)
def test_step(self):
content = PointToTargetContent()
step_size = 180 * 60
for i in range(step_size):
x = np.random.uniform(low=-1.0, high=1.0)
y = np.random.uniform(low=-1.0, high=1.0)
local_focus_pos = [x, y]
reward, done, info = content.step(local_focus_pos)
self.assertGreaterEqual(reward, 0)
if i == (step_size - 1):
# At the last frame, done becomes True
self.assertTrue(done)
else:
# Otherwise done is False
self.assertFalse(done)
def test_reset(self):
content = PointToTargetContent()
env = Environment(content)
obs = env.reset()
image = obs['screen']
angle = obs['angle']
self.assertTrue(type(image) is np.ndarray)
self.assertEqual(image.shape, (128, 128, 3))
self.assertEqual(len(angle), 2)
def test_plus_target_confolict(self):
# Check conflict with random location and plus target region
content = PointToTargetContent()
for i in range(100):
for j in range(4):
target_width = 0.1
pos = content.quadrants[j].get_random_location(target_width)
x = pos[0]
y = pos[1]
# Check whether random location pos is not inside the plus marker
inside_plus_marker = x > -0.15 and x < 0.15 and y > -0.15 and y < 0.15
self.assertFalse(inside_plus_marker)
def test_step(self):
content = PointToTargetContent()
env = Environment(content)
action = np.array([0.0, 0.0])
obs, reward, done, _ = env.step(action)
image = obs['screen']
angle = obs['angle']
self.assertTrue(type(image) is np.ndarray)
self.assertEqual(image.shape, (128, 128, 3))
self.assertEqual(len(angle), 2)
self.assertTrue(type(reward) is int)
self.assertTrue(type(done) is bool)