Crate flow_utils 

flow-utils

Shared algorithms and utilities for flow cytometry crates.

This crate provides high-performance implementations of common algorithms used across multiple flow cytometry crates, including:

• Kernel Density Estimation (KDE): FFT-accelerated KDE with GPU support
• Clustering: K-means, DBSCAN, Gaussian Mixture Model
• PCA: Principal Component Analysis for dimensionality reduction

Features

• gpu: Enable GPU acceleration for KDE (requires burn and cubecl)

Re-exports

pub use kde::KernelDensity;

pub use kde::KernelDensity2D;

pub use kde::KdeError;

pub use kde::KdeResult;

pub use clustering::KMeans;

pub use clustering::KMeansConfig;

pub use clustering::KMeansResult;

pub use clustering::Dbscan;

pub use clustering::DbscanConfig;

pub use clustering::DbscanResult;

pub use clustering::Gmm;

pub use clustering::GmmConfig;

pub use clustering::GmmResult;

pub use clustering::ClusteringError;

pub use clustering::ClusteringResult;

pub use pca::Pca;

pub use pca::PcaError;

pub use pca::PcaResult;

Modules

clustering

Clustering algorithms module

common

Common utilities for flow-utils

kde

Kernel Density Estimation (KDE) module

pca

Principal Component Analysis (PCA) module