def test_extract_minimal(graph):
x = brownian()
y = brownian()
X = {'x': x, 'y': y}
features, graph = extract(X, transformers='minimal', return_graph=True)
print(list(features.columns))
print(graph.optimized.outputs)
assert features.shape == (x.shape[0], len(minimal['series-to-attribute']) * 2 * 2)
np.testing.assert_almost_equal(
features.loc[:, 'Mean(Input(x)){0}'],
Mean().transform(x).values[:, 0, 0]
)
def test_extract_custom(graph):
x = brownian()
y = brownian()
X = {'x': x, 'y': y}
transformers = [Mean(), Variance()]
features, graph = extract(X, transformers=transformers, return_graph=True)
print(list(features.columns))
print(graph.optimized.outputs)
assert features.shape == (x.shape[0], len(transformers) * 2 * 2)
np.testing.assert_almost_equal(
features.loc[:, 'Mean(Input(x)){0}'],
Mean().transform(x).values[:, 0, 0]
)
def test_transform(graph):
x = brownian()
y = brownian()
result = graph.transform({'x': x, 'y': y}, return_dataframe=False)
assert len(result) == 3
np.testing.assert_almost_equal(
result[graph.outputs[0]].values,
Mean().transform(x).values
)
np.testing.assert_almost_equal(
result[graph.outputs[1]].values,
PowerSpectralDensity().transform(y).values
)
np.testing.assert_almost_equal(
result[graph.outputs[2]].values,
Variance().transform(Add().transform(x, y)).values
)
def test_transform_to_dataframe(graph):
x = brownian()
y = brownian()
result = graph.transform({'x': x, 'y': y}, return_dataframe=True)
print(graph.outputs)
assert result.shape == (10, 4)
def x():
return brownian()
def y():
return brownian()
