← Back to all sparks
R

rvMF

INFRA · APIS
Velocity0.0

Fast Generation of von Mises-Fisher Distributed Pseudo-Random Vectors

A fast von Mises-Fisher sampler that ships a citation more often than a feature.

directional-statisticsrandom-generationrcppcranmethods-package
Current state
rvMF generates pseudo-random vectors from the von Mises-Fisher distribution, the standard model for directional data on a sphere. Its released surface has not changed materially in the window: the three most recent tags cover a dropped dependency, a citation update, and a C++ linking correction. The February 2026 release removes the scModels import.
Where it's heading
This is a single-purpose sampler that reached its intended scope early and has been maintained rather than extended since. The visible work is dependency reduction and paper-and-citation housekeeping, which is the normal shape for a method package whose implementation is described in a published paper and whose correctness argument lives there rather than in the changelog.
Prediction
More of the same: occasional releases to keep the C++ layer building and the dependency list short. Nothing in the notes points toward new distributions or new sampling methods.

Recent moves

  1. 6mo ago

    Drops the scModels dependency

    A one-line release removing the scModels import. It shortens the dependency chain for anyone installing rvMF, but nothing about the sampler itself changes.

    View source ↗
  2. 2y ago

    Adds citation for the Statistics and Computing paper

    A documentation release marking the publication of the paper describing the sampler's implementation, with the Description field and citation information updated to match. The package gains a formal reference; its behaviour does not change.

    View source ↗
  3. 2y ago

    C++ function linkage corrected

    Housekeeping: a corrected C++ function linkage, a typo fix, and an ORCID identifier added. Internal only.

    View source ↗