def test_identity_matrix_representation(data_random):
matrix = data_random
basis = Identity()
basis.fit(matrix)
np.testing.assert_allclose(matrix.T, basis.matrix_representation())
def test_prefit_basis(data_random):
data = data_random
basis = Identity()
basis.fit(data)
# This data should be ignored during the fit
data_to_ignore = nan * data_random
model = SensorSelector(basis=basis)
model.fit(data_to_ignore, prefit_basis=True)
assert not any(isnan(model.get_selected_sensors()))
def test_update_n_basis_modes_refit(data_random):
data = data_random
n_basis_modes = 5
model = SensorSelector(basis=Identity(n_basis_modes=n_basis_modes))
model.fit(data)
assert model.basis_matrix_.shape[1] == n_basis_modes
model.update_n_basis_modes(n_basis_modes + 1, data)
assert model.basis_matrix_.shape[1] == n_basis_modes + 1
def test_prefit_basis(data_binary_classification):
x, y, _ = data_binary_classification
basis = Identity().fit(x)
model_prefit = SSPOC(basis=basis)
model_prefit.fit(x, y, prefit_basis=True, quiet=True)
model_standard = SSPOC().fit(x, y, quiet=True)
np.testing.assert_allclose(model_prefit.sensor_coef_,
model_standard.sensor_coef_)
def test_update_n_basis_modes_errors(data_random):
data = data_random
n_basis_modes = 5
model = SensorSelector(basis=Identity(n_basis_modes=n_basis_modes))
model.fit(data)
with pytest.raises(ValueError):
model.update_n_basis_modes(0)
with pytest.raises(ValueError):
model.update_n_basis_modes("5")
with pytest.raises(ValueError):
model.update_n_basis_modes(data.shape[0] + 1)
# Need to pass x when increasing n_basis_modes beyond capacity
# of the original basis
with pytest.raises(ValueError):
model.update_n_basis_modes(n_basis_modes + 1)
"""Unit tests for basis classes"""
import numpy as np
import pytest
from sklearn.exceptions import NotFittedError
from pysensors.basis import Identity
from pysensors.basis import RandomProjection
from pysensors.basis import SVD
@pytest.mark.parametrize("basis", [Identity(), SVD(), RandomProjection()])
def test_not_fitted(basis):
with pytest.raises(NotFittedError):
basis.matrix_representation()
def test_identity_matrix_representation(data_random):
matrix = data_random
basis = Identity()
basis.fit(matrix)
np.testing.assert_allclose(matrix.T, basis.matrix_representation())
@pytest.mark.parametrize("basis", [SVD, RandomProjection])
def test_matrix_representation(basis, data_random):
data = data_random
n_features = data.shape[1]
n_components = 5
pytest.lazy_fixture("data_vandermonde"),
pytest.lazy_fixture("data_random")
],
)
def test_get_all_sensors(data):
x = data
max_sensors = x.shape[1]
model = SensorSelector()
model.fit(x)
assert len(model.get_all_sensors()) == max_sensors
@pytest.mark.parametrize(
"basis",
[Identity(), SVD(), RandomProjection(n_basis_modes=5)])
def test_basis_compatibility(data_vandermonde, basis):
x = data_vandermonde
model = SensorSelector(basis=basis)
model.fit(x)
check_is_fitted(model)
def test_n_sensors(data_random):
# Check for bad inputs
with pytest.raises(ValueError):
model = SensorSelector(n_sensors=0)
with pytest.raises(ValueError):
model = SensorSelector(n_sensors=5.4)
with pytest.raises(ValueError):