emboss-data recently grew from a slim 5 megabytes to a massive 305.
Closer inspection reveals the primary culprits to be large taxonomy
and gene ontology databases:

 63927095 ./usr/share/EMBOSS/data/TAXONOMY/names.dmp
 58897689 ./usr/share/EMBOSS/index/taxon.xtax
 53221137 ./usr/share/EMBOSS/data/TAXONOMY/nodes.dmp
 28885180 ./usr/share/EMBOSS/index/taxon.xid
 23876286 ./usr/share/EMBOSS/index/taxon.xup
 20110641 ./usr/share/EMBOSS/data/OBO/gene_ontology.1_2.obo
 13816644 ./usr/share/EMBOSS/index/go.xde
  8964472 ./usr/share/EMBOSS/index/taxon.xrnk
  8963308 ./usr/share/EMBOSS/index/taxon.xgc
  6047464 ./usr/share/EMBOSS/index/go.xnm
  5642511 ./usr/share/EMBOSS/index/taxon.xmgc
  2504535 ./usr/share/EMBOSS/index/go.xac
  2220728 ./usr/share/EMBOSS/index/go.xis
  1292180 ./usr/share/EMBOSS/index/go.xid
[...]
   437512 ./usr/share/EMBOSS/index/go.xns

Could you please consider splitting them out into a separate package
or two, which emboss-data would then merely recommend (or perhaps
suggest, if few users actually need them)?

Thanks!

Dear EMBOSS developers,

today I received the following bug report about the quantity of data shipped in
the Debian package for EMBOSS.

http://bugs.debian.org/682042

In Debian, one solution would be to transfer this data in a separate optional
package.  But before doing so, I would like to ask you if this data really
oughts to be distributed with EMBOSS ?  After all, for many other databases,
there are scripts to download and index the data after installation.  Will
EMBOSS 6.5 ship the taxonomy and gene ontology databases as well ?

Have a nice day,

forwarded 682042 emboss@lists.open-bio.org
thanks

Le Wed, Jul 18, 2012 at 11:42:12PM -0400, Aaron M. Ucko a écrit :

Dear Aaron,

I feel a bit ashamed that I did not realise.  If this is not solved upstream, I
will split as you suggested.

Have a nice day,

They are included in the release which appeared on 15th July
(announcement in preparation).

For developers the data is updated by rsync ... we could provide scripts
to upload and index the data at the end of installation though I found
in preparing the release that two ontologies had moved in the last year
so that is error-prone.

Some EMBOSS applications assume these databases are installed,
particularly EDAM and the NCBI taxonomy. EDAM is used for all metadata,
the taxonomy for organism searches in data retrieval. The Gene Ontology
is included in analysis of GO terms in metadata.

So if they are in an optional package ... some things will not work if
it is not installed. EDAM I would say is essential.

regards,

Peter Rice
EMBOSS Team

Charles Plessy <plessy@debian.org> writes:

No problem; thanks for looking into it now.  I'm glad to see that the
upstream developers understand my concern as well, and am fine with
leaving EDAM in the main emboss-data package if it's indeed essential;
at slightly under 1 megabyte, it falls (just) below the cutoff I
suggested.

Andreas Tille <andreas@an3as.eu> writes:

Hi, Andreas.  Although Cn3D++ comes from the same overarching tree
(NCBI's C++ Toolkit) as BLAST+, it is a separate project.  The packaging
work done for BLAST+ would make for a decent starting point, though.

Meanwhile, what ever happened with https://bugs.debian.org/682042 ? ;-)