def test_partial_equals_full():
X1, _ = generate_data(10, 3, seed=1)
X2 = X1 * np.random.normal(0, 1, X1.shape)
Xs = [X1, X2]
gcca = GCCA()
projs_full = gcca.fit(Xs).transform(Xs)
projs_partial = gcca.partial_fit(Xs, reset=True).transform(Xs)
assert_almost_equal(np.abs(projs_full), np.abs(projs_partial))
def test_fit_elbows():
X, _ = generate_data(10, 3)
Xs = [X, X]
gcca = GCCA(n_elbows=2)
_ = gcca.fit_transform(Xs)
assert_equal(gcca.ranks_[0], 4)
def test_max_ranks():
X, _ = generate_data(10, 3)
Xs = [X, X]
gcca = GCCA(n_elbows=2, max_rank=True)
projs = gcca.fit_transform(Xs)
assert_equal(gcca.ranks_[0], 4)
def test_fit_view_idx():
n = 2
Xs = _get_Xs(n)
gcca = GCCA(fraction_var=0.9).fit(Xs)
projs = [gcca.transform(Xs[i], view_idx=i) for i in range(n)]
dists = _compute_dissimilarity(projs)
# Checks up to 7 decimal points
assert_almost_equal(np.zeros((n, n)), dists)
def test_partial_multistep():
X1, _ = generate_data(10, 3, seed=1)
X2 = X1 * np.random.normal(0, 1, X1.shape)
Xs = [X1, X2]
gcca = GCCA()
projs_full = gcca.partial_fit(Xs).transform(Xs)
projs_partial = gcca.partial_fit(Xs[0], reset=True,
multiview_step=False).partial_fit(
Xs[1],
reset=False,
multiview_step=True).transform(Xs)
assert_almost_equal(np.abs(projs_full), np.abs(projs_partial))
def test_fit_sv_tolerance():
n = 2
Xs = _get_Xs(n)
projs = GCCA(sv_tolerance=1).fit_transform(Xs)
dists = _compute_dissimilarity(projs)
# Checks up to 7 decimal points
assert_almost_equal(np.zeros((n, n)), dists)
def test_fit_tall():
n = 2
Xs = _get_Xs(n)
projs = GCCA(fraction_var=0.9, tall=True).fit_transform(Xs)
dists = _compute_dissimilarity(projs)
# Checks up to 7 decimal points
assert_almost_equal(np.zeros((n, n)), dists)
# averages from 2x3 windows.
# Load the data
Xs, y = load_UCImultifeature(views=[1, 2, 3])
# Inspect the dataset
print(f'There are {len(Xs)} views.')
print(f'There are {Xs[0].shape[0]} observations')
print(f'The feature sizes are: {[X.shape[1] for X in Xs]}')
###############################################################################
# Embed Views
# ^^^^^^^^^^^
# Create GCCA object and embed the
gcca = GCCA()
Xs_latents = gcca.fit_transform(Xs)
print(f'The feature sizes are: {[X.shape[1] for X in Xs_latents]}')
###############################################################################
# Plot the first two views against each other
# -------------------------------------------
# The top three dimensions from the latents spaces of the profile correlation
# and pixel average views are plotted against each other. However, their latent
# spaces are influenced the the Karhunen-Love coefficients, not plotted.
crossviews_plot(Xs_latents[[0, 2]],
dimensions=[0, 1, 2],
labels=y,
cmap='Set1',
def test_multiview_step():
X1, _ = generate_data(10, 3, seed=1)
gcca = GCCA()
with pytest.warns(UserWarning):
projs_partial = gcca.partial_fit(X1)
def test_no_fit(Xs={"Xs": mat_good}, err=RuntimeError):
with pytest.raises(err):
np.random.seed(1)
GCCA().transform(**Xs)
def test_bad_inputs(Xs, params, err):
with pytest.raises(err):
np.random.seed(1)
GCCA(**params).fit(**Xs)
def test_n_jobs():
X, _ = generate_data(10, 3)
Xs = [X, X]
gcca = GCCA(n_elbows=2, n_jobs=-1)
_ = gcca.fit_transform(Xs)