[Bioc-devel] New BioC 3.21 settings for R 4.5.0

[Kasper Daniel Hansen]

Thanks for the clarification Herve.

There is probably no doubt that we have packages with too many imports.
However, like Kristian mentions, one of the original reasoning behind the
creation of Bioconductor and our build system with release and devel
branch, was the typical larger dependency graph in our corner of the
world (compared to CRAN). Ie. this was specifically one of the reasons that
Bioc was created in the first place, as opposed to building on CRAN. So I
would not be surprised if we eventually need to depart from CRAN's setting
on this. Having said all that, Herve's comment that it is just about direct
imports, is obviously super important.

Best,
Kasper

On Wed, Nov 13, 2024 at 3:35 AM Kristian Ullrich <ullrich using evolbio.mpg.de>
wrote:

> Hi Lluís,
>
> Thank you for the nice code snippet to get these values.
>
> As mentioned earlier, if only the imports are counted in a non recursive
> manner, the threshold is fine.
>
> Best regards
>
> Kristian
>
> --
> Dr. Kristian Ullrich, Dipl.-Bio.
> Max Planck Institute
> For Evolutionary Biology
>
> Scientific IT group
> August Thienemann Str. 2
> 24306 Plön
> Germany
> +49 4522 763 313
> ullrich using evolbio.mpg.de
>
> --
> CONFIDENTIALITY NOTICE:
> The contents of this email message and any attachments are intended solely
> for the addressee(s) and may contain confidential and/or privileged
> information and may be legally protected from disclosure. If you are not
> the intended recipient, you are hereby notified that any use,
> dissemination, copying, or storage of this message or its attachments is
> strictly prohibited.
>
> > On 12. Nov 2024, at 22:58, Lluís Revilla <lluis.revilla using gmail.com>
> wrote:
> >
> > Hi Kristian,
> >
> > I think the threshold was decided because it matches CRAN's, but it
> could be modified.
> > But let's check the numbers regarding this threshold:
> > Approximately 20% of Bioconductor packages are affected with this
> threshold (see code below).
> > The median of strong dependencies is just 11 (9 if we exclude those
> above 20 dependencies) and the third quartile is at 17.
> > The software package that depends on more packages has 88 strong direct
> dependencies.
> >
> > Best,
> >
> > Lluís
> >
> > BiocManager::version()
> > ## [1] '3.20'
> > ap = available.packages( repos = BiocManager::repositories()["BioCsoft"])
> > ## 'getOption("repos")' replaces Bioconductor standard repositories, see
> > ## 'help("repositories", package = "BiocManager")' for details.
> > ## Replacement repositories:
> > ##     CRAN: https://ftp.cixug.es/CRAN
> > pd = tools::package_dependencies(rownames(ap), db = ap, which = "strong")
> > sum(lengths(pd) >= 20)
> > ## [1] 407
> > length(pd)
> > ## [1] 2201
> >
> >
> >
> > On Tue, 12 Nov 2024 at 06:52, Kristian Ullrich <ullrich using evolbio.mpg.de
> <mailto:ullrich using evolbio.mpg.de>> wrote:
> >> Dear Hervé and Jennifer,
> >>
> >> Thank you for clarification so quickly.
> >>
> >> Best regards
> >>
> >> Kristian
> >>
> >> --
> >> Dr. Kristian Ullrich, Dipl.-Bio.
> >> Max Planck Institute
> >> For Evolutionary Biology
> >>
> >> Scientific IT group
> >> August Thienemann Str. 2
> >> 24306 Plön
> >> Germany
> >> +49 4522 763 313
> >> ullrich using evolbio.mpg.de <mailto:ullrich using evolbio.mpg.de>
> >>
> >> --
> >> CONFIDENTIALITY NOTICE:
> >> The contents of this email message and any attachments are intended
> solely for the addressee(s) and may contain confidential and/or privileged
> information and may be legally protected from disclosure. If you are not
> the intended recipient, you are hereby notified that any use,
> dissemination, copying, or storage of this message or its attachments is
> strictly prohibited.
> >>
> >> > On 12. Nov 2024, at 03:40, Hervé Pagès <hpages.on.github using gmail.com
> <mailto:hpages.on.github using gmail.com>> wrote:
> >> >
> >> > Hi Kristian,
> >> >
> >> > On 11/11/24 10:58, Kristian Ullrich wrote:
> >> >> Dear Jennifer,
> >> >>
> >> >> I do understand the reasoning behind the
> _R_CHECK_EXCESSIVE_IMPORTS_=20. However, the threshold needs to be
> discussed in my opinion.
> >> >>
> >> >> Given that a lot of bioconductor packages e.g. rely on the
> Biostrings and GenomicRanges package (rank 9 and 10 / 2289), the Biostrings
> package itself has currently 24 dependencies, which as a consequence would
> mean that all packages relying on Biostrings will automatically be
> classified as excessive import packages.
> >> > IIUC the count is only on the number of packages that you import i.e.
> if you import Biostrings, that count for 1.
> >> >
> >> > Note that unlike the Depends relationship, the Imports relationship
> is not transitive. That is, if you import Biostrings and Biostrings imports
> IRanges then you don't get IRanges's namespace in your package's namespace,
> unless you explicitly request it.
> >> >
> >> >>
> >> >> To be honest, I would then no longer understand the meaningfulness
> of such a “note flag” if one Bioconductor package uses a TOP10 package as
> an import and should therefore be considered vulnerable?
> >> >>
> >> >> Have you checked in Bioc 3.20, how many packages from the 2289
> packages would fall under this "vulnerable" category?
> >> > We'll see but hopefully it's about direct imports only. Also IIUC
> this is only going to be a NOTE. And we can always step back if things go
> wrong.
> >> >
> >> > Best,
> >> >
> >> > H.
> >> >
> >> >>
> >> >> Best regards and thank you in anticipation
> >> >>
> >> >> Kristian
> >> > --
> >> > Hervé Pagès
> >> >
> >> > Bioconductor Core Team
> >> > hpages.on.github using gmail.com <mailto:hpages.on.github using gmail.com>
> <mailto:hpages.on.github using gmail.com <mailto:hpages.on.github using gmail.com>>
> >>
> >>
> >>         [[alternative HTML version deleted]]
> >>
> >> _______________________________________________
> >> Bioc-devel using r-project.org <mailto:Bioc-devel using r-project.org> mailing list
> >> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>
>
>         [[alternative HTML version deleted]]
>
> _______________________________________________
> Bioc-devel using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>


-- 
Best,
Kasper

	[[alternative HTML version deleted]]