[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: How to deal with tools used by BioPython and PyCogent



Hi all,

I would agree with Eric, "suggests" seems like a better match than
"recommends" for the command line tools which Biopython has
wrapper code for. Perhaps the only exception where I might suggest
a full "recommends" would be the NCBI BLAST+ binaries,
https://packages.debian.org/sid/ncbi-blast+ but even these will
not be used by everyone using Biopython.

The related question of should they be included as Build-Depends
just because there is a unit test which would exercise them is again
more policy driven.

Andreas has uncovered a couple of gotchas where we'd not yet
noticed a behaviour change in a recent tool release (e.g. EMBOSS),
so there is certainly merit in including these at build time (thank
you again for flagging these issues!).

However, for anyone doing "apt-get build-deps python-biopython"
as a precursor to building the latest Biopython from source, having
things like EMBOSS and NCBI BLAST+ installed is not required.

Does that help?

Peter

(So that's two votes for "suggests" from Biopython developers,
but this is of course up to you guys as the Debian packaging
experts.)

On Tue, May 6, 2014 at 1:00 AM, Eric Talevich <eric.talevich@gmail.com> wrote:
> Hi Andreas,
>
> Thanks for your work on this. I personally thought of these additional tools
> wrapped by Biopython as more worthy of "Suggests" rather than "Recommends",
> based on my understanding of the policy manual
> (https://www.debian.org/doc/debian-policy/ch-relationships.html):
> * Recommends - "should list packages that would be found together with this
> one in all but unusual installations"
> * Suggests - "tells the packaging system and the user that the listed
> packages are related to this one and can perhaps enhance its usefulness, but
> that installing this one without them is perfectly reasonable"
>
> Most biologists would only use a subset of these packages in their own work;
> some of the programs are borderline obsolete and the Biopython wrappers are
> only maintained for backward compatibility. There are a lot of these
> command-line tools, and they are prone to cluttering the namespace once
> installed (some packages may install multiple scripts), even though Debian
> Med packaging generally ensures there are no actual collisions.
>
> Cheers,
> Eric
>
> On Sun, Apr 27, 2014 at 12:58 AM, Andreas Tille <andreas@an3as.eu> wrote:
>>
>> Hi,
>>
>> as you might have noticed by my SVN commits for BioPython I have done
>> some effort to run all possible tests of the testsuite shipped with
>> BioPython.  I also tried to get autopkgtest working since I consider all
>> possible tests we can provide as an important service for letting
>> scientist create reproducible results.
>>
>> For making all those tests we need to Build-Depend from a lot of our
>> Debian Med maintained packages - nice to see that we have quite a good
>> coverage of all those tools.  (The situation is a bit different for
>> PyCogent, but I will report about this in a different mail.)
>>
>> Now I'm wondering how to deal with all these tools used by BioPython and
>> PyCogent which are called in the according test suites and mentioned in
>> Build-Depends.  I personally think that we should add these to
>> Recommends since we should make sure that all functionalities are
>> supported once a user is installing BioPython / PyCogent.  However, some
>> people might argue that their system might be bloated by packages they
>> are not really using.  I personally would not share this opinion but I'm
>> also no user of these toolsets and thus I wonder what you as potential
>> user might think about this.
>>
>> If I do not get any opinions I will upload with Recommends identical
>> to Build-Depends.
>>
>> Kind regards
>>
>>        Andreas.
>>
>> --
>> http://fam-tille.de
>>
>>
>> --
>> To UNSUBSCRIBE, email to debian-med-REQUEST@lists.debian.org
>> with a subject of "unsubscribe". Trouble? Contact
>> listmaster@lists.debian.org
>> Archive: https://lists.debian.org/20140427075859.GC2151@an3as.eu
>>
>


Reply to: