Package: faissR
Type: Package
Title: FAISS-Backed Nearest Neighbours, kNN Models, and k-Means
Version: 0.99.47
Authors@R: person(given = "Stefano",
    family = "Cacciatore",
    role = c("aut", "cre"),
    email = "tkcaccia@gmail.com",
    comment = c(ORCID = "0000-0001-7052-7156"))
Description: Native nearest-neighbour search, k-nearest neighbour
        prediction, and k-means helpers for large dimensionality
        reduction workflows in high-throughput biological data
        analysis, including single-cell, flow cytometry, imaging, and
        mass spectrometry applications. Functional native builds
        require the FAISS C++ library, including CPU-only indexes, and
        can optionally use FAISS GPU indexes with NVIDIA cuVS
        integration or direct RAPIDS cuVS/CUDA indexes on NVIDIA
        machines. CPU-only systems do not need NVIDIA libraries, but
        NVIDIA CUDA/cuVS libraries are mandatory when a GPU-enabled
        build is explicitly requested. Unsupported WebAssembly and
        Bioconductor staging platforms may build diagnostic-only stubs
        that report the unavailable native functionality. Supervised
        kNN models use knn(Xtrain, Ytrain) and predict(), or
        knn(Xtrain, Ytrain, Xtest) for immediate prediction. Class
        probabilities are returned through predict(type = "prob") for
        kNN classification models.
License: MIT + file LICENSE
URL: https://github.com/tkcaccia/faissR
BugReports: https://github.com/tkcaccia/faissR/issues
biocViews: Software, Infrastructure, GPU, SingleCell, FlowCytometry,
        MassSpectrometry, ImagingMassSpectrometry, Proteomics,
        Clustering, DimensionReduction, Classification,
        FunctionalPrediction
Encoding: UTF-8
Depends: R (>= 4.5.0)
Imports: Biobase, methods, Rcpp (>= 1.1.0)
Suggests: BiocNeighbors, BiocSingular, BiocStyle, S4Vectors,
        SingleCellExperiment, SummarizedExperiment, knitr, Matrix,
        float, rmarkdown, scater, scRNAseq, scran, scuttle, testthat,
        withr
LinkingTo: Rcpp (>= 1.1.0)
SystemRequirements: C++20, Fortran on Unix, FAISS C++ development
        library (Debian/Ubuntu: libfaiss-dev, libblas-dev,
        liblapack-dev; RPM distributions: faiss-devel, blas-devel,
        lapack-devel where available; macOS: static FAISS and OpenMP
        supplied through the R-macos recipes/toolchain or a compatible
        local prefix; Windows: an Rtools-compatible FAISS CPU build),
        complete LP64 BLAS and LAPACK with single-precision routines on
        Linux and Windows. Optional CUDA and cuVS libraries are used
        only when a GPU-enabled build is explicitly requested.
VignetteBuilder: knitr
RoxygenNote: 8.1.0
Config/roxygen2/version: 8.1.0
Config/testthat/edition: 3
Config/pak/sysreqs: nvidia-cuda-dev libfaiss-dev
Repository: https://biocstaging.r-universe.dev
Date/Publication: 2026-09-22 15:08:10 UTC
RemoteUrl: https://github.com/BiocStaging/faissR
RemoteRef: HEAD
RemoteSha: 82ed27005c5292b15f3c8499ac61047c37240c5b
NeedsCompilation: yes
Packaged: 2026-09-22 19:13:23 UTC; root
Author: Stefano Cacciatore [aut, cre] (ORCID:
    <https://orcid.org/0000-0001-7052-7156>)
Maintainer: Stefano Cacciatore <tkcaccia@gmail.com>