Skip to content

TST use global_dtype in sklearn/cluster/tests/test_affinity_propagation.py #22667

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 13 commits into from
Nov 3, 2022
Merged

TST use global_dtype in sklearn/cluster/tests/test_affinity_propagation.py #22667

Changes from all commits
Commits
Show all changes
13 commits
Select commit Hold shift + click to select a range
542f2c7
TST Adapt test_affinity_propagation.py to test implementations on 32b…
jjerphan Mar 3, 2022
76df15d
Merge branch 'main' into tst/test_affinity_propagation-32bit
jjerphan Mar 18, 2022
d1fed63
TST Use global_dtype
jjerphan Mar 18, 2022
4272be8
TST Do not copy on when casting
jjerphan Mar 23, 2022
168eba5
Merge branch 'main' into tst/test_affinity_propagation-32bit
jjerphan Mar 24, 2022
88780f9
Apply review comments
jjerphan Mar 24, 2022
3ee4572
DOC Improve comment regarding dtype preservation
jjerphan Mar 24, 2022
fe5b107
Apply review comments
jjerphan Mar 29, 2022
6684b79
DOC
jjerphan May 17, 2022
30b8a3e
TST Review comments
jjerphan Jun 10, 2022
337fa79
Merge branch 'main' into tst/test_affinity_propagation-32bit
jjerphan Jun 10, 2022
7978f2a
Merge branch 'main' into tst/test_affinity_propagation-32bit
jjerphan Oct 21, 2022
0ea2b5e
TST adding global_dtype to 2 addtional tests
glemaitre Nov 3, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
55 changes: 33 additions & 22 deletions sklearn/cluster/tests/test_affinity_propagation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,10 +29,14 @@
random_state=0,
)

# TODO: AffinityPropagation must preserve dtype for its fitted attributes
# and test must be created accordingly to this new behavior.
# For more details, see: https://github.com/scikit-learn/scikit-learn/issues/11000

def test_affinity_propagation(global_random_seed):

def test_affinity_propagation(global_random_seed, global_dtype):
"""Test consistency of the affinity propagations."""
S = -euclidean_distances(X, squared=True)
S = -euclidean_distances(X.astype(global_dtype, copy=False), squared=True)
preference = np.median(S) * 10
cluster_centers_indices, labels = affinity_propagation(
S, preference=preference, random_state=global_random_seed
Expand Down Expand Up @@ -108,11 +112,12 @@ def test_affinity_propagation_precomputed_with_sparse_input():
AffinityPropagation(affinity="precomputed").fit(csr_matrix((3, 3)))


def test_affinity_propagation_predict(global_random_seed):
def test_affinity_propagation_predict(global_random_seed, global_dtype):
# Test AffinityPropagation.predict
af = AffinityPropagation(affinity="euclidean", random_state=global_random_seed)
labels = af.fit_predict(X)
labels2 = af.predict(X)
X_ = X.astype(global_dtype, copy=False)
labels = af.fit_predict(X_)
labels2 = af.predict(X_)
assert_array_equal(labels, labels2)


Expand All @@ -131,23 +136,23 @@ def test_affinity_propagation_predict_error():
af.predict(X)


def test_affinity_propagation_fit_non_convergence():
def test_affinity_propagation_fit_non_convergence(global_dtype):
# In case of non-convergence of affinity_propagation(), the cluster
# centers should be an empty array and training samples should be labelled
# as noise (-1)
X = np.array([[0, 0], [1, 1], [-2, -2]])
X = np.array([[0, 0], [1, 1], [-2, -2]], dtype=global_dtype)

# Force non-convergence by allowing only a single iteration
af = AffinityPropagation(preference=-10, max_iter=1, random_state=82)

with pytest.warns(ConvergenceWarning):
af.fit(X)
assert_array_equal(np.empty((0, 2)), af.cluster_centers_)
assert_allclose(np.empty((0, 2)), af.cluster_centers_)
assert_array_equal(np.array([-1, -1, -1]), af.labels_)


def test_affinity_propagation_equal_mutual_similarities():
X = np.array([[-1, 1], [1, -1]])
def test_affinity_propagation_equal_mutual_similarities(global_dtype):
X = np.array([[-1, 1], [1, -1]], dtype=global_dtype)
S = -euclidean_distances(X, squared=True)

# setting preference > similarity
Expand Down Expand Up @@ -178,10 +183,10 @@ def test_affinity_propagation_equal_mutual_similarities():
assert_array_equal([0, 0], labels)


def test_affinity_propagation_predict_non_convergence():
def test_affinity_propagation_predict_non_convergence(global_dtype):
# In case of non-convergence of affinity_propagation(), the cluster
# centers should be an empty array
X = np.array([[0, 0], [1, 1], [-2, -2]])
X = np.array([[0, 0], [1, 1], [-2, -2]], dtype=global_dtype)

# Force non-convergence by allowing only a single iteration
with pytest.warns(ConvergenceWarning):
Expand All @@ -195,8 +200,10 @@ def test_affinity_propagation_predict_non_convergence():
assert_array_equal(np.array([-1, -1, -1]), y)


def test_affinity_propagation_non_convergence_regressiontest():
X = np.array([[1, 0, 0, 0, 0, 0], [0, 1, 1, 1, 0, 0], [0, 0, 1, 0, 0, 1]])
def test_affinity_propagation_non_convergence_regressiontest(global_dtype):
X = np.array(
[[1, 0, 0, 0, 0, 0], [0, 1, 1, 1, 0, 0], [0, 0, 1, 0, 0, 1]], dtype=global_dtype
)
af = AffinityPropagation(affinity="euclidean", max_iter=2, random_state=34)
msg = (
"Affinity propagation did not converge, this model may return degenerate"
Expand All @@ -208,17 +215,17 @@ def test_affinity_propagation_non_convergence_regressiontest():
assert_array_equal(np.array([0, 0, 0]), af.labels_)


def test_equal_similarities_and_preferences():
def test_equal_similarities_and_preferences(global_dtype):
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The previous test (test_affinity_propagation_non_convergence_regressiontest) should raise the warning for the different dtypes I think?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

and test_affinity_propagation_poredict_non_convergence as well.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The previous test (test_affinity_propagation_non_convergence_regressiontest) should raise the warning for the different dtypes I think?

Which difference of dtypes are you referring to?

# Unequal distances
X = np.array([[0, 0], [1, 1], [-2, -2]])
X = np.array([[0, 0], [1, 1], [-2, -2]], dtype=global_dtype)
S = -euclidean_distances(X, squared=True)

assert not _equal_similarities_and_preferences(S, np.array(0))
assert not _equal_similarities_and_preferences(S, np.array([0, 0]))
assert not _equal_similarities_and_preferences(S, np.array([0, 1]))

# Equal distances
X = np.array([[0, 0], [1, 1]])
X = np.array([[0, 0], [1, 1]], dtype=global_dtype)
S = -euclidean_distances(X, squared=True)

# Different preferences
Expand Down Expand Up @@ -251,10 +258,14 @@ def test_affinity_propagation_random_state():


@pytest.mark.parametrize("centers", [csr_matrix(np.zeros((1, 10))), np.zeros((1, 10))])
def test_affinity_propagation_convergence_warning_dense_sparse(centers):
"""Non-regression, see #13334"""
def test_affinity_propagation_convergence_warning_dense_sparse(centers, global_dtype):
"""
Check that having sparse or dense `centers` format should not
influence the convergence.
Non-regression test for gh-13334.
"""
rng = np.random.RandomState(42)
X = rng.rand(40, 10)
X = rng.rand(40, 10).astype(global_dtype, copy=False)
y = (4 * rng.rand(40)).astype(int)
ap = AffinityPropagation(random_state=46)
ap.fit(X, y)
Expand All @@ -265,11 +276,11 @@ def test_affinity_propagation_convergence_warning_dense_sparse(centers):


# FIXME; this test is broken with different random states, needs to be revisited
def test_affinity_propagation_float32():
def test_correct_clusters(global_dtype):
# Test to fix incorrect clusters due to dtype change
# (non-regression test for issue #10832)
X = np.array(
[[1, 0, 0, 0], [0, 1, 1, 0], [0, 1, 1, 0], [0, 0, 0, 1]], dtype="float32"
[[1, 0, 0, 0], [0, 1, 1, 0], [0, 1, 1, 0], [0, 0, 0, 1]], dtype=global_dtype
)
afp = AffinityPropagation(preference=1, affinity="precomputed", random_state=0).fit(
X
Expand Down