def test_that_lastimprovement_returns_delta_2(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.15, 0.15, 0.15))
p._score = [75, 1000, 0]
self.assertEqual(p.last_improvement(), -1000)
def test_that_lastimprovement_is_inf_if_not_enough_scores2(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.15, 0.15, 0.15))
p._score = [0]
self.assertEqual(p.last_improvement(), float('inf'))
def test_that_parameters_are_correctly_assigned(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.15, 0.15, 0.15))
self.assertEqual(p._parameters.omega(), 0.15)
self.assertEqual(p._parameters.phip(), 0.15)
self.assertEqual(p._parameters.phig(), 0.15)
def test_that_lastimprovement_is_calculated(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.15, 0.15, 0.15))
p._score = [0, 5]
self.assertEqual(p.last_improvement(), 5)
def test_that_last_movement_returns_distance_2d(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.15, 0.15, 0.15))
p._position = [np.array([0, 0]), np.array([6, 0])]
self.assertEqual(p.last_movement(), 6)
def test_that_lastimprovement_returns_infinity_if_only_one_score(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.15, 0.15, 0.15))
p._score = [0]
self.assertEqual(p.last_improvement(), float("inf"))
def test_that_particle_is_initialized_with_velocity(self, mock):
mock.return_value = np.array([666])
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.5, 0.5, 0.5))
self.assertTrue(mock.called)
self.assertEqual(p.velocity(), np.array([666]))
self.assertEqual(len(p._velocity), 1)
def test_that_swarm_has_to_be_created_with_positive_number_of_particles_2(
self):
with self.assertRaises(ValueError):
Swarm(swarm_size=0,
bounds=Bounds(np.array([1]), np.array([2])),
parameters=PsoParameters(0.15, 0.15, 0.15),
minimum_improvement=10e-8,
minimum_step=10e-8)
def test_that_velocity_returns_the_current_velocity_2(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.15, 0.15, 0.15))
p._velocity = list(range(10))
self.assertEqual(p.velocity(), p._velocity[-1])
self.assertEqual(p.velocity(), 9)
self.assertEqual(len(p._velocity), 10)
def test_that_position_returns_the_current_position_2(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.15, 0.15, 0.15))
p._position = list(range(10))
self.assertEqual(p.position(), p._position[-1])
self.assertEqual(p.position(), 9)
self.assertEqual(len(p._position), 10)
def test_best_score(self, mock_score):
s = Swarm(swarm_size=10,
bounds=Bounds(np.array([1]), np.array([2])),
parameters=PsoParameters(0.15, 0.15, 0.15),
minimum_improvement=10e-8,
minimum_step=10e-8)
mock_score.side_effect = list(reversed(range(10)))
self.assertEqual(s.best_score(), 9)
def test_movement_2(self, mock):
s = Swarm(swarm_size=10,
bounds=Bounds(np.array([1]), np.array([2])),
parameters=PsoParameters(0.15, 0.15, 0.15),
minimum_improvement=10e-8,
minimum_step=10e-8)
mock.return_value = 0
self.assertFalse(s.still_moving())
def test_that_swarm_can_update_the_scores_of_all_particles_if_scores_equal_swarmsize(
self):
s = Swarm(swarm_size=10,
bounds=Bounds(np.array([1]), np.array([2])),
parameters=PsoParameters(0.15, 0.15, 0.15),
minimum_improvement=10e-8,
minimum_step=10e-8)
with self.assertRaises(ValueError):
s.update_scores(list(range(9)))
def test_that_best_score_returns_the_highest_score(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.15, 0.15, 0.15))
p._score = list(range(10))
p._position = list(reversed(range(10)))
self.assertEqual(len(p._score), 10)
self.assertEqual(len(p._position), 10)
self.assertEqual(p.best_score(), 9)
def test_that_best_position_can_be_calculated_if_same_number_of_scores(
self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.15, 0.15, 0.15))
p._score = list(range(9))
p._position = list(reversed(range(10)))
with self.assertRaises(ValueError):
p.best_position()
def test_that_update_score_increases_the_amount_of_stored_scores(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.15, 0.15, 0.15))
self.assertEqual(len(p._score), 0)
p.update_score(666)
self.assertEqual(len(p._score), 1)
self.assertEqual(p.best_score(), 666)
def test_that_swarm_is_iterable(self):
s = Swarm(swarm_size=10,
bounds=Bounds(np.array([1]), np.array([2])),
parameters=PsoParameters(0.15, 0.15, 0.15),
minimum_improvement=10e-8,
minimum_step=10e-8)
particles = [p for p in s]
self.assertEqual(len(particles), 10)
def test_that_swarm_can_update_the_scores_of_all_particles(self, mock):
s = Swarm(swarm_size=10,
bounds=Bounds(np.array([1]), np.array([2])),
parameters=PsoParameters(0.15, 0.15, 0.15),
minimum_improvement=10e-8,
minimum_step=10e-8)
s.update_scores(list(range(10)))
self.assertTrue(mock.called)
self.assertEqual(mock.call_count, 10)
def test_that_update_velocity_is_appleid_to_all_particles(
self, mock_update, mock_position):
s = Swarm(swarm_size=10,
bounds=Bounds(np.array([1]), np.array([2])),
parameters=PsoParameters(0.15, 0.15, 0.15),
minimum_improvement=10e-8,
minimum_step=10e-8)
mock_position.return_value = 666
s.update_velocity()
self.assertTrue(mock_update.called)
self.assertEqual(mock_update.call_count, 10)
def test_that_maximum_number_of_threads_is_limited_by_swarm_size(
self, mock_update, mock_size):
s = Swarm(swarm_size=1,
bounds=Bounds(np.array([1]), np.array([2])),
parameters=PsoParameters(0.15, 0.15, 0.15),
minimum_improvement=10e-8,
minimum_step=10e-8)
ex = Executor(DummyObjectiveFunction(), threads=10)
mock_size.return_value = 1
ex.calculate_scores(s)
self.assertTrue(mock_size.called)
self.assertEqual(mock_update.call_count, 1)
def test_that_bounds_are_correctly_assigned(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.5, 0.5, 0.5))
self.assertEqual(p._bounds.lower(), np.array([1]))
self.assertEqual(p._bounds.upper(), np.array([2]))
def test_that_particle_is_initialized_without_score(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.5, 0.5, 0.5))
self.assertEqual(len(p._score), 0)
def test_that_omega_must_be_between_0_and_1_5(self):
p = PsoParameters(omega=1, phip=0.5, phig=0.5)
self.assertEqual(p.omega(), 1)
self.assertEqual(p.phip(), 0.5)
self.assertEqual(p.phig(), 0.5)
def test_that_phip_must_be_between_0_and_1_2(self):
with self.assertRaises(ValueError):
PsoParameters(omega=0.5, phip=1.1, phig=0.5)
def test_that_best_score_can_be_calculated_if_scores_not_empty(self):
p = Particle(Bounds(np.array([1]), np.array([2])),
PsoParameters(0.15, 0.15, 0.15))
with self.assertRaises(ValueError):
p.best_score()
