def test_sample_constant_distribution(self):
"""Returns a constant sample distribution."""
constant_dist = Distribution(dist_type="constant", constant=1)
sample = constant_dist.sample(5)
expected_sample = constant_dist.constant*ones(5)
assert all(sample == expected_sample)
def test_sample_None_distribution(self):
"""Returns a None sample distribution."""
None_dist = Distribution(dist_type=None)
sample = None_dist.sample(size=10)
expected_sample = full(10, None)
assert all(sample == expected_sample)
def test_sample_numpy_distribution(self):
"""Returns a numpy sample distribution."""
numpy_dist = Distribution(
dist_type="numpy",
dist_name="beta",
a=1, b=1
)
sample = numpy_dist.sample(5)
expected = random.default_rng(seed=numpy_dist.seed).beta(1, 1, 5)
assert all(sample == expected)
def test_sample_weights_distribution_data(
self,
fixture_weights_distribution
):
"""Returns a weights sample distribution from a data array."""
weights_dist = Distribution(
dist_type="weights",
data=pytest.data_2D
)
sample = weights_dist.sample(5)
expected_sample = random.default_rng(int(time())).choice(
a=pytest.data_2D[:, 0],
p=pytest.data_2D[:, 1],
size=5
)
assert all(sample == expected_sample)
def test_sample_empirical_distribution_file(
self,
fixture_empirical_distribution
):
"""Returns an empirical sample distribution from a file."""
empirical_dist = Distribution(
dist_type="empirical",
data=pytest.data_1D,
kernel="gaussian",
bandwidth=0.1
)
sample = empirical_dist.sample(5)
expected_sample = KernelDensity(
kernel="gaussian",
bandwidth=0.1
).fit(pytest.data_1D.reshape(-1, 1)).sample(
n_samples=5, random_state=int(time())
).flatten()
assert all(sample == expected_sample)
