[Debian-med-packaging] Qiime building.

[Tim Booth]

Hi,

On Tue, 2012-08-14 at 21:08 +0100, Andreas Tille wrote:
> On Tue, Aug 14, 2012 at 06:24:40PM +0100, Tim Booth wrote:
> > As for the self-tests, I get 13 failures but not all of these actually need to be fixed.
> > In general, if Qiime doesn't work with, say, the standard R version, the preferred fix
> > would be to patch Qiime rather than to bundle the legacy R if at all possible.
> > Also, the latest version of Qiime has added several new tools which are not yet packaged.
> 
> Is Qiime actually that picky about versions?

The self test script reports all dependency versions and issues a
warning if it's not the version it expects.  But aside from that, Qiime
will try to run with whatever it has.
Unfortunately there are no functional self-tests so there's no
systematic way to see if Qiime works with a given set of tool versions.
 
> > matplotlib is not quite the right version (wontfix)
> > ParsInsert is missing (*)
> 
> You mean this one: http://sourceforge.net/projects/parsinsert/ ?

Yes.  I had a quick look and it seems a nice simple build - "make"
yields one binary.

> > R is 2.14 not 2.12 (wontfix)
> > blast is 2.2.25 not 2.2.22 (wontfix)
> > clearcut is missing (*)

I've not actually tried to track these things down yet - was still
working on fixing the existing functionality for the latest Qiime+Ubuntu
combo.  Ditto for the other packages below.

> 
> > cluster_jobs_fp not set up (you need to do this for your cluster)
> > mothur is missing (rebuilding it just now)
> 
> You mean just for your PPA right?

I mean on my own machine - the build takes quite a while (and actually
just failed again as I made a dumb typo).  Then I'll push to my PPA and
SVN.

> > numpy is not 1.5.1 (won't fix)
> > pplacer is missing (*)
> 
> Found a publication: http://www.biomedcentral.com/1471-2105/11/538/
> Download URL ?
> 
> > python is 2.7.3 not 2.7.2 (won't fix)
> > raxmlHPC is missing
> 
> We do have just plain raxml.  Should we in addition package a HPC version?
> 
> > rtax is missing (*)
> 
> URL???
> 
> > uclust is missing (needs manual install due to non-free-ness)
> 
> Any volunteer to nag the authors?

The non-free-ness of uClust is deliberate, possibly due to contractual
issues.  I think a more sensible approach would be to use an alternative
default cluster algorithm (Qiime supports this already) and so sideline
uClust.  Discussing this option with BL users and Qiime developers is on
my oh-so-long TODO list.

Cheers,

TIM


-- 
Tim Booth <tbooth at ceh.ac.uk>
NERC Environmental Bioinformatics Centre 

Centre for Ecology and Hydrology
Maclean Bldg, Benson Lane
Crowmarsh Gifford
Wallingford, England
OX10 8BB 

http://nebc.nerc.ac.uk
+44 1491 69 2705
-- 
This message (and any attachments) is for the recipient only. NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.