def test_opt_2():
f = lambda x, y, z: x**2 + y**2 + z**2
opt = Optimizer(f)
opt.descent()
np.testing.assert_allclose(opt.global_optimizer,
np.array([0, 0, 0]),
atol=1e-3)
def test_descent_2_verbose_2():
f = lambda x, y, z: x**2 + y**2 + z**2
opt = Optimizer(f, quadratic_matrix=np.eye(3), tuning=True, verbose=2)
opt.descent()
optimal_point = opt.global_optimizer
assert abs(optimal_point[0]) < 1e-3
assert abs(optimal_point[1]) < 1e-3
assert abs(optimal_point[2]) < 1e-3
def test_descent_1():
f = lambda x, y, z: x**2 + y**2 + (z - 2)**2
opt = Optimizer(f)
opt.descent()
optimal_point = opt.global_optimizer
assert abs(optimal_point[0]) < 1e-3
assert abs(optimal_point[1]) < 1e-3
assert abs(optimal_point[2] - 2) < 1e-3
def test_init_value_error_sampling():
f = lambda x: x
sampling = False
covariance_matrix = np.eye(2)
with pytest.raises(ValueError):
opt = Optimizer(f=f,
tuning=sampling,
quadratic_matrix=covariance_matrix)
def test_init_quad_value_error_2():
f = lambda x, y: x + y
init_points = [1, 2]
sampling = True
cov_matrix = np.eye(3)
with pytest.raises(ValueError):
opt = Optimizer(f=f,
tuning=sampling,
init_points=init_points,
quadratic_matrix=cov_matrix)
def test_init_quad_type_error():
f = lambda x: x
init_points = [1]
sampling = True
cov_matrix = [1, 2]
with pytest.raises(TypeError):
opt = Optimizer(f=f,
tuning=sampling,
init_points=init_points,
quadratic_matrix=cov_matrix)
def test_init_dimensions():
f = lambda x: x
init_points = [1, 2]
sampling = True
cov_matrix = np.eye(2)
opt = Optimizer(f=f,
tuning=sampling,
init_points=init_points,
quadratic_matrix=cov_matrix)
assert opt.vectorize == True
assert opt.dimension == 2
def test_init_quad_value_error():
f = lambda x, y: x + y
init_points = [1, 2]
sampling = True
cov_matrix = np.array([1, 2])
print(len(cov_matrix.shape))
with pytest.raises(ValueError):
opt = Optimizer(f=f,
tuning=sampling,
init_points=init_points,
quadratic_matrix=cov_matrix)
def test_init_without_sampling():
f = lambda x: x
max_iter = 10
init_points = [1]
tolerance = 1
random_restarts = 2
beta_1 = 0.5
beta_2 = 0.1
step_size = 1e-2
epsilon = 1e-4
tuning = False
verbose = 0
opt = Optimizer(f=f,
max_iter=max_iter,
init_points=init_points,
tolerance=tolerance,
random_restarts=random_restarts,
beta_1=beta_1,
beta_2=beta_2,
step_size=step_size,
epsilon=epsilon,
tuning=tuning,
verbose=verbose)
def test_init_type_error_init_points():
f = lambda x, y: x + y
init_points = 'hey'
with pytest.raises(TypeError):
opt = Optimizer(f=f, init_points=init_points)
def test_init_type_error_tol():
f = lambda x: x
tol = 'hey'
with pytest.raises(TypeError):
opt = Optimizer(f=f, tolerance=tol)
def test_descenterror_cov():
f = lambda x, y, z: x**2 + y**2 + z**2
with pytest.raises(ValueError):
opt = Optimizer(f, quadratic_matrix=None, tuning=True, verbose=2)
opt.descent()
optimal_point = opt.global_optimizer
def test_init_value_error_init_points():
f = lambda x, y: x + y
init_points = [1]
with pytest.raises(ValueError):
opt = Optimizer(f=f, init_points=init_points)
def test_init_epochs_value_error():
f = lambda x: x
epochs = -1
with pytest.raises(ValueError):
opt = Optimizer(f=f, max_epochs=epochs)
def test_init_epochs_type_error():
f = lambda x: x
epochs = 'hey'
with pytest.raises(TypeError):
opt = Optimizer(f=f, max_epochs=epochs)
def test_init_type_error_beta2():
f = lambda x: x
beta_2 = 'hey'
with pytest.raises(TypeError):
opt = Optimizer(f=f, beta_2=beta_2)
def test_init_type_error_sampling():
f = lambda x: x
sampling = 'hey'
with pytest.raises(TypeError):
opt = Optimizer(f=f, tuning=sampling)
def test_init_type_error_lr():
f = lambda x: x
lr = 'hey'
with pytest.raises(TypeError):
opt = Optimizer(f=f, step_size=lr)
def test_init_type_error_restarts():
f = lambda x: x
restarts = 'hey'
with pytest.raises(TypeError):
opt = Optimizer(f=f, random_restarts=restarts)
def test_init_type_error_iter():
f = lambda x: x
max_iter = 'hey'
with pytest.raises(TypeError):
opt = Optimizer(f=f, max_iter=max_iter)
def test_init_verbose_value_error():
f = lambda x: x
verbose = -1
with pytest.raises(ValueError):
opt = Optimizer(f=f, verbose=verbose)
def test_init_verbose_type_error():
f = lambda x: x
verbose = 'hey'
with pytest.raises(TypeError):
opt = Optimizer(f=f, verbose=verbose)
def test_init_value_error_tol():
f = lambda x: x
tol = -1
with pytest.raises(ValueError):
opt = Optimizer(f=f, tolerance=tol)
def test_produce_random_sample():
f = lambda x: x**2
Opt = Optimizer(f)
init_points = Opt.produce_random_points()
assert init_points.shape[0] == 1
def test_init_value_error_restarts():
f = lambda x: x
restarts = -1
with pytest.raises(ValueError):
opt = Optimizer(f=f, random_restarts=restarts)
def test_dimensionality_2():
f = lambda x, y: x + y
Opt = Optimizer(f)
init_points = Opt.produce_random_points()
assert init_points.shape[0] == 2
def test_init_value_error_lr():
f = lambda x: x
lr = -1
with pytest.raises(ValueError):
opt = Optimizer(f=f, step_size=lr)
def test_produce_random_sample_2():
f = lambda x, y, z: x * y + z
opt = Optimizer(f)
init_points = opt.produce_random_points()
assert init_points.shape[0] == 3
def test_init_value_error_beta1():
f = lambda x: x
beta_1 = -1
with pytest.raises(ValueError):
opt = Optimizer(f=f, beta_1=beta_1)
def test_init_without_list():
f = lambda x: x
init_point = 2
opt = Optimizer(f=f, init_points=init_point)
