TIL about https://github.blog/news-insights/product-news/supercharging-github-actions-with-job-summaries/ for exporting Markdown from #GitHubActions, eg nice error logs. You never know where some hobby side project will take you!

#RStats question : is there a method to get around the issue that roxygen2::roxygenize triggers compiling C/C++ code when only doc generation is needed ?
ref : https://github.com/r-lib/roxygen2/issues/771

@josi @jimgar my only interaction with the registry was when i had to fork this repo that build multi-architecture docker images https://github.com/open-webui/open-webui/blob/main/.github/workflows/docker-build.yaml
lot of bizarre env variables and scopes to setup

@HydrePrever l idée d un petit bateaux avait clairement ses limites même si j imagine c était pour éviter quelque chose comme le desastre du Mavi Marmara

@tylermorganwall next meson and ninja based build systems ?😆
This is great since cmake covers more than enough

@tylermorganwall Great ! there are Bioconductor packages that are also there for building static or dynamic libs for package usage (Rhtslib being a prominent one). There is a package called BiocCmake that basically downloads cmake binary, are you planning to make such package for CRAN ?

@josi the registration form is not working for me for some reason

Introducing Portable Linux R Binary Packages! 📦

We're excited to preview this solution for complex system library dependencies, particularly for those in diverse Linux environments.

Learn about portable binary packages and how to install them in our blog post: https://posit.co/blog/introducing-portable-linux-r-binary-packages/

#RStats #Linux

@HydrePrever jolie vue, c est out ca ?

@yokofakun far as i know with no dedicated library implementing the specs does this, probably would have to import to plink2 bpgen format and merge using --pmerge (may break if multi-allelelic sites are not properly merged) or implement it yourself using c/c++ libraries providing the readers

#Rstats hivemind :
are there packages out there that implemented directly or indirectly the idea of persistent vectors/data structure https://hypirion.com/musings/understanding-persistent-vector-pt-1 ?
Are similar ideas already used in base R ?

htslib/samtools/bcftools 1.22 released
https://github.com/samtools/samtools/releases/tag/1.22

https://github.com/samtools/bcftools/releases/tag/1.22

https://github.com/samtools/htslib/releases/tag/1.22

Are you using Ublock Origin ? Add "google.com##.hdzaWe" without quotes to your Ublock Origin My Filters to block the google AI overview and apply changes.

#htslib #Bioinformatics #GenomicsIO
@yokofakun any idea what is the fastest method to get the nth bcf record using htslib or bcftools.h without explicit loops? (trying out something similar to your project https://github.com/lindenb/rbcf but with ALTREP)
I guess it is possible to guess which block to (lazy) parse if one know the blocks offsets and number of records per block

@simon in that setup, unclear why only python (other than the llms being better at python vs say R)

@louisaslett will do

@simon is the sandbox in the browser or somewhere else ?

@eddelbuettel yes it is done in Bioconductor to make the versioning of system packages uniform, this is the main reason for me too (otherwise i mostly use r2u, many thanks for that !). I can few packages on CRAN have install.libs.R files https://github.com/search?q=org%3Acran%20install.libs.R&type=code

#Rstats now in understand why there is no package on CRAN or Bioconductor that just provides libgsl.a and the headers

It was so so good to work with Monash Uni students on their first steps in C++ programming for R applications! 🚀✨ They made it! And now, they're ready to code! 🤖👾

https://arp.numbat.space/week11/

#rstats #rcpp #armadillo