Lightnews — Scholar-powered news

Henrik Bengtsson @henrikbengtsson.bsky.social · 2h

There's also cran.r-project.org/package=ntfy, which doesn't even require sign-up and it can be self-hosted too

ntfy: Lightweight Wrapper to the 'ntfy.sh' Service

The 'ntfy' (pronounce: notify) service is a simple HTTP-based pub-sub notification service. It allows you to send notifications to your phone or desktop via scripts from any computer, entirely without...

cran.r-project.org

Reposted by Henrik Bengtsson

Calle Börstell @cborstell.bsky.social · 3d

Something like this!

Tweaked the calculation to something I had originally intended and included 3-letter endings only.

–arp is very Skåne

I like the distribution of the –red/–röd/–ryd endings, of the same origin

Code: github.com/borstell/maps

A map of Sweden in 10 different panels, each showing the log odds frequency of different 3-letter place name endings in various shades of red (darker means higher prevalence; gray means no data).

1 2

Henrik Bengtsson @henrikbengtsson.bsky.social · 5d

This resonates well with me. Please tell me there's also something for:

Everyone folds their cardboard boxes before putting them in the recycle bin

1 6

Henrik Bengtsson @henrikbengtsson.bsky.social · 5d

Nice.

-arp?

PS. Do you share your code for this somewhere?

1 2

Reposted by Henrik Bengtsson

Calle Börstell @cborstell.bsky.social · 7d

Distributions of endings in Swedish place names

#RStats #DataViz

A map of Sweden in 10 different panels, each showing the proportional frequency of different place name endings in various shades of blue

2 10 43

Reposted by Henrik Bengtsson

R Consortium @rconsortium.bsky.social · 6d

🚨 Big news for the R Community!

The Sovereign Tech Fund has invested $450,000 in the R Foundation to strengthen R’s sustainability, security, and modernization.

Details: r-consortium.org/posts/sovere...

#rstats #opensource #datascience

The Sovereign Tech Fund invests $450,000 in R Foundation to Enhance R’s Sustainability and Security – R Consortium

R Foundation embarking on a crucial project to modernize R’s core infrastructure, improve its maintainability, and fortify its supply chain.

r-consortium.org

2 8 50

Henrik Bengtsson @henrikbengtsson.bsky.social · 6d

I'd like to bring the attention also to long-serving, zero-dependency {gtools} on CRAN;

x <- c("8", "10", "1", "40", "9A", "21A", "21B")
> gtools::mixedsort(x)
[1] "1" "8" "9A" "10" "21A" "21B" "40"
> gtools::mixedorder(x)
[1] 3 1 5 2 6 7 4

cran.r-project.org/package=gtools

#RStats

2 1 10

Henrik Bengtsson @henrikbengtsson.bsky.social · 7d

Another real-world #HPC cluster confirmed for future.batchtools.futureverse.org - built-in template for PBS/TORQUE

plan(batchtools_torque)

works with PBSPro scheduler. Big thanks to @eliocamp.mastodon.social.ap.brid.gy - req'd some xtra efforts bc special R_LIBS_USER location

#RStats #parallel

4

Henrik Bengtsson @henrikbengtsson.bsky.social · 9d

"tarball size threshold for CRAN increased to 10 MB (as it already is in the policies)"

github.com/r-devel/r-sv...

#RStats

tarball size threshold for CRAN increased to 10 MB (as it already is … · r-devel/r-svn@d0b709c

…in the policies) git-svn-id: https://svn.r-project.org/R/trunk@88883 00db46b3-68df-0310-9c12-caf00c1e9a41

github.com

3

Reposted by Henrik Bengtsson

Angelo D’Ambrosio @adambrosiomd.bsky.social · 11d

I wish there was static compilation or JIT of function and objects of dependency packages, so that one would not wait to get all dependencies to use just a few functions

1 1

Reposted by Henrik Bengtsson

Jeroen Ooms @jeroenooms.bsky.social · 11d

Super excited for the new release of #rstats V8 which now also works in WebR by running js/wasm natively in web worker (thanks to @gws.phd). Will release some cool new packages soon to show the potential of WASM in R ✨

1 2 19

Henrik Bengtsson @henrikbengtsson.bsky.social · 18d

@ericmigi.com two Qs:

1. Is there a public forum to discuss bugs etc? I don't want to waste my or others time with duplicate reports. Think GitHub issues.

2. is there another way to report on bugs and issues than through the app? I'd like to be able to write a proper bug report on a real keyboard?

Reposted by Henrik Bengtsson

Johannes Rainer @jorainer.bsky.social · 21d

Was really fun to present our #SpectriPy @bioconductor.bsky.social package for combined #rstats & #Python analysis of #MassSpectrometry data today at #EuroBioC2025

also available on Zenodo: doi.org/10.5281/zeno...

SpectriPy: Enhancing Cross-Language Mass Spectrometry Data Analysis with R and Python

Mass spectrometry (MS) is a key technology used across multiple fields, including biomedical research and life sciences. Technological advancements result in increasingly large and complex data sets a...

doi.org

1 2 14

Henrik Bengtsson @henrikbengtsson.bsky.social · 21d

A #Slurm user just confirmed that "yay it works. Pretty sick!"

Thanks to excellent feedback from several users, it'll soon be even easier to distribute #rstats code via #HPC job schedulers using future.batchtools

#parallel #futureverse

Henrik Bengtsson @henrikbengtsson.bsky.social · 24d

If anyone else is following this, we've moved over to github.com/futureverse/..., where progress has already been made

1 6

Henrik Bengtsson @henrikbengtsson.bsky.social · 22d

#RStats #Positron Q: Is it possible to programmatically restart R running in Position Console?

I want to add support for it as startup::restart(), and if not possible (e.g. RStudio Console), at least produce an informative error message

github.com/HenrikBengts...

1 1

Reposted by Henrik Bengtsson

Emily Riederer @emilyriederer.bsky.social · 23d

Really insightful post from Julie Tibshirani (spotted in LinkedIn, can't find on Bsky) reflecting on #rstats 's unique governance structure and what can be learned for other languages

jtibs.substack.com/p/if-all-the...

If all the world were a monorepo

The R ecosystem and the case for extreme empathy in software maintenance

jtibs.substack.com

7 49 130

Henrik Bengtsson @henrikbengtsson.bsky.social · 23d

Not sure I understand the "cheating" part, but your approach seems 100% valid.

WASM and webR have really grown on me over the summer. So much potential...

1 1

Henrik Bengtsson @henrikbengtsson.bsky.social · 24d

If anyone else is following this, we've moved over to github.com/futureverse/..., where progress has already been made

1

Henrik Bengtsson @henrikbengtsson.bsky.social · 27d

Ideally there would be a single teeny installation tool `install-r` that would walk you through the installation on different OSes and give instructions what's missing and how to fix them. Sometimes it might be "Ask admin to install 'xyz'" and sometimes "Do you want to install 'pqr' now?"

1

Henrik Bengtsson @henrikbengtsson.bsky.social · 27d

I think we (the #RStats community) could do better when it comes to instructions, or better, tools for installing R on #Linux.

A major thing we cld clarify is the diff in installing as (a) root/admin/sudo vs (b) a non-privileged user. Some can only do the latter and it takes much more knowledge

1 2

Henrik Bengtsson @henrikbengtsson.bsky.social · 27d

Awesome (for real) - we're making progress here, and it'll help lots of others.

So, y'day I actually update the develop version of future.batchtools (>= 0.20.0-9015) to handle this case. Can you please retry after updating:

remotes::install_github("futureverse/future.batchtools", ref="develop")

1

Henrik Bengtsson @henrikbengtsson.bsky.social · 28d

It could also be that you're affected by **batchtools** (sic!) bug github.com/mlr-org/batc.... I you call:

$ sbatch --wrap="hostname"
Submitted batch job 849746

do you get more output than that one line? (That could explain where "runtime" comes from)

Make batch.id robust to warning messages from sbatch by bwcompton · Pull Request #314 · mlr-org/batchtools

I ran into a crazy bug today: getJobStatus gave me batch.id = "that". It turns out that when I requested a large amount of memory, sbatch returned this um, helpful message: sbatch: INFO: ...

github.com

1

Henrik Bengtsson @henrikbengtsson.bsky.social · 28d

What do you get?

> plan(future.batchtools::batchtools_slurm)
> f <- future(42)
> f$config$reg$status
> f$config$reg$status # need to call it a few times

> library(future)
> plan(future.batchtools::batchtools_slurm)
> f <- future(42)
> f$config$reg$status
> f$config$reg$status
Key: <job.id>
job.id def.id submitted started done error mem.used resource.id batch.id
<int> <int> <num> <num> <num> <char> <num> <int> <char>
1: 1 1 1757451949 NA NA <NA> NA 1 849716
log.file job.hash job.name
<char> <char> <char>
1: <NA> job454af4bd10d2eebe2308c0f9584a2584 <NA>
>

1 1

Henrik Bengtsson @henrikbengtsson.bsky.social · 28d

Thanks a bunch. That '--jobs=runtime' seems weird - it should be a job id number, not "runtime". I'm pulling that "job id" from batchtools - need to dig deeper.

BTW, if you have a GitHub account, I'm happy to continue the discussion in a new issue at github.com/futureverse/...

1

Henrik Bengtsson @henrikbengtsson.bsky.social · 28d

Thank you! Others reported on this in past, but I couldn't reproduce. I've now done a deep study on Slurm and I might have found the culprit (future.batchtools.futureverse.org/news/index.h...). Could you please retry?

remotes::install_github("futureverse/future.batchtools", ref="develop")

Changelog

future.batchtools.futureverse.org

1