def test_time_delay_transform_matches_input(data_random):
x = data_random
observables = TimeDelay(delay=2, n_delays=4)
observables.fit(x)
y = observables.transform(x)
assert_array_equal(y[1], x[[9, 7, 5, 3, 1]].flatten())
def test_time_delay_inverse(data_random):
x = data_random
delay = 2
n_delays = 3
n_deleted_rows = delay * n_delays
observables = TimeDelay(delay=delay, n_delays=n_delays)
y = observables.fit_transform(x)
# First few rows of x are deleted which don't have enough
# time history
assert_array_equal(observables.inverse(y), x[n_deleted_rows:])
def test_simulate_with_time_delay(data_2D_superposition):
x = data_2D_superposition
observables = TimeDelay()
model = Koopman(observables=observables)
model.fit(x)
n_steps = 10
n_consumed_samples = observables.n_consumed_samples
x_pred = model.simulate(x[:n_consumed_samples + 1], n_steps=n_steps)
assert_allclose(x[n_consumed_samples + 1:n_consumed_samples + n_steps + 1],
x_pred)
def test_score_complex_data(data_random_complex):
x = data_random_complex
model = Koopman().fit(x)
with pytest.raises(ValueError):
model.score(x, cast_as_real=False)
@pytest.mark.parametrize(
"observables",
[
Identity(),
Polynomial(),
TimeDelay(),
pytest.lazy_fixture("data_custom_observables"),
],
)
def test_observables_integration(data_random, observables):
x = data_random
model = Koopman(observables=observables).fit(x)
check_is_fitted(model)
y = model.predict(x)
assert y.shape[1] == x.shape[1]
@pytest.mark.parametrize(
"observables",
[
def test_time_delay_output_shape(data_random, delay, n_delays):
x = data_random
y = TimeDelay(delay=delay, n_delays=n_delays).fit_transform(x)
assert y.shape == (x.shape[0] - delay * n_delays,
(n_delays + 1) * x.shape[1])
from pykoopman.observables import Polynomial
from pykoopman.observables import TimeDelay
@pytest.fixture
def data_small():
t = linspace(0, 5, 10)
return stack((t, t**2), axis=1)
@pytest.mark.parametrize(
"observables",
[
Identity(),
Polynomial(),
TimeDelay(),
pytest.lazy_fixture("data_custom_observables"),
],
)
def test_if_fitted(observables, data_random):
x = data_random
with pytest.raises(NotFittedError):
observables.transform(x)
with pytest.raises(NotFittedError):
observables.inverse(x)
with pytest.raises(NotFittedError):
observables.get_feature_names()
observables.fit(x)