COSIE.utils

Functions

compute_knn_graph	Construct a k-nearest neighbors (k-NN) graph based on an input feature matrix.
compute_neighborhood_embedding	Compute the neighborhood embedding of each cell based on its spatial k-nearest neighbor graph.
construct_knn_graph_hnsw	Efficiently compute approximate k-nearest neighbor (k-NN) graph using the hnswlib library.
nn_approx	Efficiently find approximate K-nearest neighbors using the Annoy library.
setup_seed	Set the random seed for Python, NumPy, and PyTorch to ensure reproducibility.