Introduction | Function overview | Prerequisites and Setup | Java version | Package options | Custom JAR setup | Quick Start | Core Function: fastFindKNN() | Output structure | Algorithm choice: ANN vs. exact | Distance metrics | Euclidean (default) | Cosine | Dot-product (ANN only) | Skipping distance computation | HNSW-DiskANN Tuning | Parameter reference | High-recall configuration | Product Quantization (pq.subspaces) | Thread management | Graph Construction | Shared Nearest-Neighbor graph | K-Nearest Neighbor graph | Community detection | Full Single-Cell Workflow | Using a real SingleCellExperiment | Comparing with BiocNeighbors | Drop-in BiocNeighbors integration | Timing comparison | Troubleshooting | Common errors | Debugging with verbose mode | Checking the Java runtime | Using a safe wrapper | Session Information
Package: jvecfor 1.1.0
Anestis Gkanogiannis
jvecfor: Fast K-Nearest Neighbor Search for Single-Cell Analysis
Drop-in replacement for BiocNeighbors::findKNN using the jvecfor Java library, which builds on the jvector library to leverage the Java Vector API for portable SIMD acceleration across AVX2, AVX-512, and ARM NEON hardware. jvecfor/jvector implements HNSW-DiskANN approximate search and VP-tree exact search. The package achieves approximately 2x speedup over Annoy-based search at n >= 50K cells while returning output structurally identical to BiocNeighbors, making it suitable for seamless integration into existing Bioconductor single-cell workflows. Convenience wrappers delegate shared nearest-neighbor (SNN) and k-nearest-neighbor (KNN) graph construction to the bluster package.
Authors:
jvecfor_1.1.0.tar.gz
jvecfor_1.1.0.zip(r-4.7)jvecfor_1.1.0.zip(r-4.6)jvecfor_1.1.0.zip(r-4.5)
jvecfor_1.1.0.tgz(r-4.6-any)jvecfor_1.1.0.tgz(r-4.5-any)
jvecfor_1.1.0.tar.gz(r-4.7-any)jvecfor_1.1.0.tar.gz(r-4.6-any)
jvecfor_1.1.0.tgz(r-4.6-emscripten)
manual.pdf |manual.html✨
DESCRIPTION |NEWS
card.svg |card.png
jvecfor/json (API)
| # Install 'jvecfor' in R: |
| install.packages('jvecfor', repos = c('https://bioc.r-universe.dev', 'https://cloud.r-project.org')) |
Bug tracker:https://github.com/gkanogiannis/jvecfor/issues
On BioConductor:jvecfor-1.1.0(bioc 3.24)jvecfor-1.0.0(bioc 3.23)
singlecellgraphandnetworkclusteringclassification
Last updated from:df5d2ab8ae. Checks:1 NOTE, 6 OK, 3 ERROR. Indexed: yes.
| Target | Result | Time | Files | Syslog |
|---|---|---|---|---|
| bioc-checks | NOTE | 177 | ||
| linux-devel-x86_64 | OK | 223 | ||
| source / vignettes | OK | 229 | ||
| linux-release-x86_64 | OK | 256 | ||
| macos-release-arm64 | OK | 103 | ||
| macos-oldrel-arm64 | OK | 117 | ||
| windows-devel | ERROR | 125 | ||
| windows-release | ERROR | 148 | ||
| windows-oldrel | ERROR | 124 | ||
| wasm-release | OK | 191 |
Exports:buildIndexfastFindKNNfastMakeKNNGraphfastMakeSNNGraphfindKnnFromIndexjvecfor_setupJvecforParamshow
Dependencies:abindassortheadbeachmatBHBiocGenericsBiocNeighborsBiocParallelblustercliclustercodetoolscpp11data.tableDelayedArrayformatRfutile.loggerfutile.optionsgenericsglueigraphIRangeslambda.rlatticelifecyclemagrittrMatrixMatrixGenericsmatrixStatspkgconfigprocessxpsR6RcpprlangS4ArraysS4VectorssnowSparseArrayvctrsXVector
Readme and manuals
Help Manual
| Help page | Topics |
|---|---|
| jvecfor: Fast K-Nearest Neighbor Search for Single-Cell Analysis | jvecfor-package jvecfor |
| Fast K-Nearest Neighbor Search | fastFindKNN |
| Build a K-Nearest Neighbor Graph | fastMakeKNNGraph |
| Build a Shared Nearest-Neighbor Graph | fastMakeSNNGraph |
| Install a Custom jvecfor JAR | jvecfor_setup |
| JvecforIndex: BiocNeighbors Index Class for jvecfor | JvecforIndex-class |
| JvecforParam: BiocNeighbors Parameter Class for jvecfor | buildIndex,JvecforParam-method findKnnFromIndex,JvecforIndex-method JvecforParam JvecforParam-class show,JvecforParam-method |
