Skip to contents

Generates the kNN graph with additional distances

rs_sc_knn_w_dist.Rd

[Experimental] This function is a wrapper over the Rust-based generation of the approximate nearest neighbours.

Usage

rs_sc_knn_w_dist(embd, knn_params, validate_index, verbose, seed)

Arguments

embd

Numerical matrix. The embedding matrix to use to generate the kNN graph.

knn_params

List. The kNN parameters defined by params_sc_neighbours().

validate_index

Boolean. If you want to validate the index via an exhaustive search in a subset of cells.

verbose

Integer. 0L - quiet; 1L - normal verbosity; 2L - detailed verbosity.

seed

Integer. Seed for reproducibility purposes.

Value

A list with:

  • indices - An integer matrix (cells x k) representing the indices (0-indexed!) of the approximate nearest neighbours.

  • dist - A numerical matrix (cells x k) representing the distances to the nearest neighbours.

  • dist_metric - String representing the used distance metric.

Rows the search left short are padded by repeating their last neighbour and distance.