def test__sample__size_specified__variance_of_generated_equal_to_specified_variance(
self):
gamma_model = utils.build_gamma()
samples = gamma_model.sample(10000)
assert_array_almost_equal(np.atleast_2d(np.var(samples, axis=1)),
gamma_model.covariance,
decimal=2)
def test__sample__size_specified__mean_of_generated_equal_to_specified_mean(
self):
gamma_model = utils.build_gamma()
samples = gamma_model.sample(1000000)
assert_array_almost_equal(np.mean(samples, axis=1),
gamma_model.mean,
decimal=2)
def test__sample__size_specified__return_horizontally_stacked_random_vectors(
self):
gamma_model = utils.build_gamma()
size = 1000
samples = gamma_model.sample(size)
assert samples.shape[0] == 1
assert samples.shape[1] == size
def test__from_gamma__cov_mean_dim_must_be_same(self):
gamma = utils.build_gamma()
expected_cov_type = CovarianceType.full
gaussian = convert_to_gaussian(gamma, expected_cov_type)
assert_equal(gaussian.dim, gamma.dim)
assert_array_equal(gaussian.covariance, gamma.covariance)
assert_array_equal(gaussian.mean, gamma.mean)
def test__sample__size_is_zero__raise_error(self):
gamma_model = utils.build_gamma()
with raises(
Exception,
match=
"<size> is negative or equal to zero. Must be integer greater than zero"
):
gamma_model.sample(0)
def test__sample__size_is_less_than_zero__raise_error(self):
gamma_model = utils.build_gamma()
for negative_number in uniform(-100, 0, 100):
with raises(
Exception,
match=
"<size> is negative or equal to zero. Must be integer greater than zero"
):
gamma_model.sample(negative_number)
def test__noise_type__always__gamma(self):
gamma_model = utils.build_gamma()
assert gamma_model.noise_type is NoiseType.gamma
def test__dim__always__one(self):
# only 1-Dim gamma process is supported at this time
gamma_model = utils.build_gamma()
assert gamma_model.dim == 1
def test__covariance_type__always__full(self):
gamma_model = utils.build_gamma()
assert gamma_model.covariance_type is CovarianceType.full