Skip to content
New issue

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

#

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

Already on GitHub? # to your account

Cleanup silhouette variables #31

Merged
merged 4 commits into from
Oct 12, 2022
Merged

Cleanup silhouette variables #31

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
084a86f
cleanup sil
adamgayoso Oct 11, 2022
e469c50
cleanup sil
adamgayoso Oct 11, 2022
6c05648
cleanup sil
adamgayoso Oct 11, 2022
e968695
Update src/scib_metrics/utils/_silhouette.py
adamgayoso Oct 12, 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
12 changes: 5 additions & 7 deletions src/scib_metrics/utils/_silhouette.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,9 +55,8 @@ def _intra_cluster_distances_block(i: int, input: _IntraClusterData) -> jnp.ndar
return intra_dist


def _nearest_cluster_distances(X: np.ndarray, labels: np.ndarray):
def _nearest_cluster_distances(X: np.ndarray, labels: np.ndarray, unique_labels: np.ndarray):
"""Calculate the mean nearest-cluster distance for observation i."""
unique_labels = jnp.unique(labels)
inter_dist = jnp.array(np.inf * np.ones((X.shape[0],)))
label_combinations = jnp.array([list(i) for i in list(itertools.combinations(unique_labels, 2))])
for i in range(len(label_combinations)):
Expand All @@ -74,9 +73,9 @@ def _inter_values(subset_a: np.ndarray, subset_b: np.ndarray) -> Union[jnp.ndarr
return values_a, values_b


def _nearest_cluster_distance_block(inter_dist: jnp.ndarray, input: _InterClusterData) -> jnp.ndarray:
label_a = input.label_combos[inter_dist, 0]
label_b = input.label_combos[inter_dist, 1]
def _nearest_cluster_distance_block(combo_ind: int, input: _InterClusterData) -> jnp.ndarray:
label_a = input.label_combos[combo_ind, 0]
label_b = input.label_combos[combo_ind, 1]
label_mask_a = input.labels == label_a
label_mask_b = input.labels == label_b
subset_a = input.data[label_mask_a]
Expand Down Expand Up @@ -109,7 +108,6 @@ def silhouette_samples(X: np.ndarray, labels: np.ndarray) -> np.ndarray:
"""
if X.shape[0] != labels.shape[0]:
raise ValueError("X and labels should have the same number of samples")

intra_dist = _intra_cluster_distances(X, labels)
inter_dist = _nearest_cluster_distances(X, labels)
inter_dist = _nearest_cluster_distances(X, labels, np.unique(labels))
return jax.device_get((inter_dist - intra_dist) / jnp.maximum(intra_dist, inter_dist))