def test_gen_1d_gauss_range():
res = 180
range_values = (-10, res - 11)
rfs, centers = sim.generate_1d_gaussian_rfs(1,
res,
range_values,
random_tuning=False)
sim_data = sim.generate_1d_rf_responses(rfs, np.array([-10]), res,
range_values, 0)
assert sim_data[0, ] > 0
range_values = (10, res + 10)
rfs, centers = sim.generate_1d_gaussian_rfs(1,
res,
range_values,
random_tuning=False)
sim_data = sim.generate_1d_rf_responses(rfs, np.array([10]), res,
range_values, 0)
assert sim_data[0, ] > 0
def test_gen_1D_gauss_shape():
n_vox = 10
res = 180
rfs, centers = sim.generate_1d_gaussian_rfs(n_vox, res, (0, res - 1))
assert rfs.shape == (n_vox, res)
assert centers.size == n_vox
sim_data = sim.generate_1d_rf_responses(rfs, np.array([0, 10, 20]), res,
(0, res - 1))
assert sim_data.shape == (n_vox, 3)
# Test number of data dimensions
def test_data_dimensions():
x = np.random.rand(5, 10, 2)
s = InvertedEncoding()
with pytest.raises(ValueError):
s.fit(x, np.random.rand(5))
# Define some data to use in the following tests.
n, dim = 297, 9
n_ = n // dim
y = np.repeat(np.linspace(0, 180-(180/dim), dim), n_)
voxel_rfs, _ = generate_1d_gaussian_rfs(dim, 180, (0, 179),
random_tuning=False)
X = generate_1d_rf_responses(voxel_rfs, y, 180, (0, 179),
trial_noise=0.25).transpose()
X2 = generate_1d_rf_responses(voxel_rfs, y, 180, (0, 179),
trial_noise=0.25).transpose()
# Test if valid data can be fit.
def test_can_fit_data():
Invt_model = InvertedEncoding()
Invt_model.fit(X, y)
# Test if valid data can be fit in circular space.
def test_can_fit_circular_space():
s = InvertedEncoding(6, 5, 'circular', range_stop=360)
s.fit(X, y)