R Observatory: Metrics for rwig version 0.3.0

Balamuta, James Joseph

doi:10.5281/zenodo.21843040

rwig

0.3.0Package record

Wasserstein Index Generation (WIG) Model

0packages depend

2.4Kdownloads / year

91.6%test coverage

13/13flavors OK

Overview

About

Maintained by Fangzhou XieCRAN page ↗GitHub ↗Website ↗Issues ↗

Efficient implementation of several Optimal Transport algorithms in Fangzhou Xie (2025) doi:10.48550/arXiv.2504.08722 and the Wasserstein Index Generation (WIG) model in Fangzhou Xie (2020) doi:10.1016/j.econlet.2019.108874.

Changes in 0.3.0

Performance work driven by perf profiles; all solver outputs are checked

against frozen outputs of version 0.2.0 (relative differences below 1e-11)

and the Julia-derived reference solutions in the tests.

wdl()/wig() on the CPU now predict the document barycenters with the same batched Gibbs-kernel iteration the training uses, on a whole batch of documents at once. Before, inference solved one barycenter per document and, with the default method = "auto", did so in the log domain, which cost about 20 times the whole training (2000 NYT headlines, 1 epoch: 33 minutes, now 33 seconds). method, threshold and n_threads of the barycenter control are documented as unused by WDL.
The log-stabilized sinkhorn() and barycenter() exponentiate with a vectorised rational approximation (1-2 ulp of exp()) instead of one libm call per matrix entry, and log sinkhorn() no longer computes a column soft-min per iteration whose only use, the column term of the error, is identically zero. 1000 x 1000 log Sinkhorn, 1000 iterations: 19 s to 8 s on one thread.

Show all changes in 0.3.0Hide the rest of the changes in 0.3.0

Worker threads (n_threads) spin briefly between the short parallel sections of an iteration instead of sleeping; 12 threads now give about 5x on the forward passes where they gave 3x.
The backward passes of the log-stabilized sinkhorn() and barycenter() (with_grad = TRUE) no longer materialize the M x N softmax matrices: the forward pass keeps its row and column soft-mins (M + N numbers per iteration), from which each adjoint product is one fused sweep over the cost matrix, split over the threads by column blocks. 1000 x 1000 log Sinkhorn with gradient, 1000 iterations: 15.6 s to 13.2 s on one thread, 3.8 s to 2.1 s on 12; 800 x 800 log barycenter with gradients: 37.9 s to 33.9 s and 10.9 s to 5.2 s.
Kernel products with up to 4 right-hand columns (the "parallel" barycenter()) use one dsymv/dgemv per column rather than a level-3 call that repacks the N x N kernel every time: 800 x 800 barycenter with 4 sources, 0.79 s to 0.34 s.
CUDA wdl(): the per-document cublasDgemv launches of the weight adjoint are one strided-batched GEMM (2000 headlines: 5.6 s to 4.5 s).

From NEWS.md in the release

Install

R (CRAN)pak (dev)Ubuntu (r2u)Fedora (COPR)openSUSE (OBS)

install.packages("rwig", repos = "https://cloud.r-project.org/")

Installs the current CRAN release. Prebuilt binaries on Windows and macOS.

if (!requireNamespace("pak", quietly = TRUE))
  install.packages("pak", repos = "https://cloud.r-project.org/")
pak::pak("fangzhou-xie/rwig")

Development version from fangzhou-xie/rwig. Installs pak first if it is not already available.

Release

# enable cranapt (r2u) if it is not already set up, then install
grep -rqs r2u.stat.illinois.edu /etc/apt/sources.list /etc/apt/sources.list.d/ \
  || curl -fsSL https://raw.githubusercontent.com/eddelbuettel/r2u/master/inst/scripts/add_cranapt_noble.sh | sudo bash
sudo apt install -y r-cran-rwig

Fast Ubuntu binaries via r2u (no compilation). The cranapt setup runs once and is skipped when it is already enabled. We list this channel from recorded r2u download activity rather than from a live index of the repository, so treat it as last known rather than guaranteed.

bspm (used by r2u) also needs make: sudo apt install make

# enable the iucar/cran COPR if it is not already present, then install
ls /etc/yum.repos.d/*iucar*cran*.repo >/dev/null 2>&1 \
  || sudo dnf copr enable -y iucar/cran
sudo dnf install -y R-CRAN-rwig

RPMs from the iucar/cran COPR, rebuilt from CRAN.

For install.packages() integration in R, also: sudo dnf install R-CoprManager

# add the autoCRAN OBS repo (Tumbleweed) if it is missing, then install
zypper -q lr | grep -qi autoCRAN \
  || sudo zypper addrepo https://download.opensuse.org/repositories/devel:/languages:/R:/autoCRAN/openSUSE_Tumbleweed/devel:languages:R:autoCRAN.repo
sudo zypper install -y R-rwig

RPMs auto-built from CRAN on openSUSE OBS (Tumbleweed). We list this channel from recorded autoCRAN download activity rather than from a live index of the repository, so treat it as last known rather than guaranteed.

Health

CRAN checks

13OK

CRAN check page ↗

Slowest check: 3.7 min · r-devel-windows-x86_64

Status changes

OK13 OK2026-08-04
NOTE12 OK, 1 NOTE2026-08-01
OK13 OK2026-05-02
NOTE10 OK, 1 NOTE2026-04-25
ERROR6 OK, 1 NOTE, 1 ERROR2026-04-22

Advisories

none known

How this is measured·How advisories are found

Documentation

In release 0.3.0

READMEYes · 339 wordsVignettesYes · 9 · dynamicNEWSYes · 100% structured · mentions 0.3.0Citation fileNoWebsiteYes · pkgdown

Help pages: 9 (1 internal)includes a package overview page
Examples that run: 100%6 run in a plain check
Exported functions with examples: 86%
Documented parameters: 59%
Return-value docs: 100%
Help pages citing references: 78%

Vignettes (9)

How these are measured

Downloads

2.4K

CRAN downloads in the past year

Rank #9,178 · ~6/day · ~197/mo

Releases R versions

Daily download trend is not available in this view yet.

45730 days

1K90 days

2.4K1 year

Compare downloads with other packages →

Also on151 r2u2 autocran

How downloads are counted

Repository

fangzhou-xie/rwig ↗Account: fangzhou-xie

1Stars

0Forks

0Open issues

0Open PRs

0Releases

58Commits

2Commit identities

58 commits · Last activity 2026-10-07

Responsiveness

not measured

Stars over time

2025-06-11 · 12026-07-07 · 1

How these are measured

PracticesBeta

11 development-tooling and community-health practices detected across 6 families in the upstream repository

Checks run against github.com/fangzhou-xie/rwig on 2026-10-18.

Continuous integration (2)

GitHub ActionsR CMD check workflow

The R CMD check workflow runs on Linux, macOS and Windows, and on R-devel

Docs source (4)

README.Rmdpkgdown

GitHub Pages site last deployed 2026-10-07 · The repository has a site deploy workflow

pkgdown site last built 2026-10-07

Reproducibility and dev environment (1)

data-raw/

CRAN release process (2)

cran-comments.mdCRAN-SUBMISSION

Show all practicesHide 2 practices

Lint, format, editor (1)

RStudio project

Git structural (1)

.gitattributes

How this is detected·Detection ruleset v3 (2026-09-26)

AI coding toolsBeta

This repository shows signs of AI coding tools.

Claude Codefirst seen 2026-03-06
Named in .Rbuildignore: .claude
Named in .gitignore: .claude
Commits: at least 25 crediting it in the message

Models named in its commits

Claude Fable 5.125 commits · Sep to Oct 2026

Counts are commits, not lines of code. Commits by a tool's accounts count once across all of its accounts. Commits crediting it are the largest count among its credit lines, a minimum.How this is detected·Detection ruleset v4 (updated 2026-09-27)

Dependencies

Declared dependencies

9 external dependencies (excludes base and recommended)

Installing rwig adds 10 packages (7 with compiled code, 16 MB of source tarballs). Heaviest: stringi 11 MB, Rcpp 1.8 MB, rlang 766 KB.

Depends (1)

R >= 4.2

Imports (7)

utilsstatsrlang cli word2vec tokenizers stopwords

LinkingTo (0)

none

Suggests (4)

RhpcBLASctl knitr rmarkdown tinytest

Enhances (0)

none

Dependency license breakdown →

Reverse dependencies

0direct

0indirect

Nothing depends on this yet.

How these are counted

Code & Tests

Datasets

People & History

People (1)

Fangzhou Xie

Releases (3)

Pick two releases to compare their code metrics.

0.3.0Latest
record
2026-10-07 · current release · diff ↗
0.2.0
record
2026-09-20 · diff ↗
R
R 4.6.0 released · 2026-04-24
0.1.0
record
2026-04-21
R
R 4.5.0 released · 2025-04-11

Package metadata

First published: 2026-04-21
Total releases: 3 / 1 yrs
License: MIT + file LICENSE OSI
Minimum R: ≥ 4.2
Bundled data: 343 KB / 1 file
Download size: 478 KB

Cite

Cite this package

Run in R for the authors' preferred citation:

citation("rwig")

Format

Xie, F. (2026). rwig: Wasserstein Index Generation (WIG) Model (Version 0.3.0) [Computer software]. https://doi.org/10.32614/CRAN.package.rwig

This package has no citation file, so citation() returns this entry built from DESCRIPTION.

Cite the R Observatory

For a number measured here: a download total, a coverage figure, an archival date.

Format

APA

Balamuta, J. J. (2026). R Observatory: Metrics for rwig version 0.3.0 [Data set]. HJJB, LLC. Data release v2026-10-21. https://doi.org/10.5281/zenodo.21843040

From data release v2026-10-21, which the citation names so these numbers can be found later. More on citing and the projects behind them.

Report a problem with this page →