scib_metrics.nearest_neighbors.NeighborsResults#

class scib_metrics.nearest_neighbors.NeighborsResults(indices, distances)[source]#

Nearest neighbors results data store.

distances#

Array of distances to the nearest neighbors.

indices#

Array of indices of the nearest neighbors. Self should always be included here; however, some approximate algorithms may not return the self edge.

Attributes table#

`knn_graph_connectivities`	Compute connectivities using the UMAP approach.
`knn_graph_distances`	Return the sparse weighted adjacency matrix.
`n_neighbors`	Number of neighbors.
`n_samples`	Number of samples (cells).
`indices`
`distances`

Subset down to n neighbors.

NeighborsResults.knn_graph_connectivities#

Compute connectivities using the UMAP approach.

Connectivities (similarities) are computed from distances using the approach from the UMAP method, which is also used by scanpy.

NeighborsResults.knn_graph_distances#: Return the sparse weighted adjacency matrix.

NeighborsResults.subset_neighbors(n)[source]#

Subset down to n neighbors.