Hi everyone 🤗 We're currently in the progress of publishing some modules for metabolomics QC based on MetaProViz. Since MetaProViz has no conda/bioconda package, we build our own Docker image for it, currently hosted at ghcr.io/saezlab/metaproviz. We were told this needs to move to an official nf-core-owned registry, and that only an official nf-core core team member can do that. Could anybody help us out, or point us to the right person? Thanks so much!

You can ask on #C09H6NYHR9T

Though fair warning: we will probably ask you to package on bioconda instead 😅 nf-co.re/docs/…/bioconda

Thanks! Since all our modules are based on our Bioconductor R-package, conda does not make sense. Given, we have an R-package as our basis, we need the Docker image. So how could we proceed here?

I'm pretty sure most if not all bioconductor packages are already on bioconductor

Thanks I will look into it. Is there a ways that we can continue for now to submit the docker image, so we can start on the review in the meanwhile? hence, could someone move it into the official registry?

Yes every bioconductor package is already in bioconda

Normally it is quite quick @Christina Schmidt to do, so we would suggest take a day or two first before we faff with custom images

Ok, we will look into it and if there are any issues come back to you

Just got Claude to check the container you supplied. Looks like 2 of the packages are very new and haven't made it to bioconda yet, but should do soon (

bioconductor-metaproviz

,

bioconductor-omnipathr

). Then there is one core R dep that's missing (

r-qcc

) that'll need to be packaged on conda-forge.

Thanks for all the infos. The version is because the newer version is on bioconductor devel branch. I think in october should be the next merge into main.

Ah so you need something that's ahead of bioconductor? But your docker image doesn't have that, right?

Thanks! We will get back to the PR next week as Giulia is not avaliable .

Hi, I’m back now 🤗 Thank you so much for helping out! I’ll have a look at the PRs, and if there are any issues, I’ll come back to you.

Bioconda is usually hours / a day or so. But yeah bioconductor being so slow is a problem. It's a shame that there's not a better way around that..

Thanks Phil, that’s a good idea! Is it realistic that the Bioconda recipe for

bioconductor-metaproviz

is merged and built soon? We know Bioconda is the better solution. But if it will take longer, we would really appreciate being allowed to use the custom image for a short period of time, because the timing is critical for us. We need to publish the modules by the end of the year. I saw that your PR for OmnipathR 4.0.0 (bioconda/bioconda-recipes#69442) is merged, thank you! Would it be possible to update the draft PR you already started, “Add bioconductor-metaproviz 4.0.0” (bioconda/bioconda-recipes#69443)? Or should we open a new one? Here is what would change compared to the current draft: • source: the GitHub tag

v4.99.0

(commit

f76eacf0973161bc204c66f33440898e62033ce1

),

<https://github.com/saezlab/MetaProViz/archive/refs/tags/v4.99.0.tar.gz>

, instead of the Bioconductor 4.0.0 tarball • version: 4.99.0 instead of 4.0.0 •

r-qcc

dependency removed, the new version does not import it any more, so the conda-forge PR for it is no longer needed for us • two new imports:

r-ggnewscale

and

r-withr

, both are on conda-forge One question: after the Bioconductor release, do we need to update the recipe ourselves (switch back to the Bioconductor URL and bump the version), or does the Bioconda bulk update handle it? We saw

bioconductor-skeleton update-all-packages

in the bulk docs, but are not sure how it treats a recipe with a GitHub source. Thanks again for all the help!

Updated, does this look ok? github.com/bioconda/bioconda-recipes/pull/69443

Thank you so much!

Merged!

Anytime for you

Thank you both! 🙏