Re: [Help] Failed to upgrade bcftools to version 1.5

2017-07-28 Thread Andreas Tille
Hi Graham,

On Fri, Jul 28, 2017 at 07:27:36PM +0200, Graham Inggs wrote:
> On 28 July 2017 at 16:30, Sascha Steinbiss  wrote:
> > Once I'm at it I can also take a shot at bcftools. Never worked on that
> > before though, let's see how it goes.
> 
> Please take note you'll need to backtract in git until version
> 1.4.1-2, add the patches, and upload 1.4.1-3.  I don't know any good
> way to do this, perhaps Andreas knows what to do.

I'm not a Git expert but

   git format-patch 1.4.1-2

should give a set of patches that could be used for quilt after minimal
editiing (+ quilt refresh)

Kind regards

  Andreas. 

-- 
http://fam-tille.de



CI for trimmomatic

2017-07-28 Thread Nadiya Sitdykova
Hello Andreas,

test-suite for trimmomatic package is ready.

Regards, Nadiya


Re: CI for ray

2017-07-28 Thread Nadiya Sitdykova
Hello all,

Graham, thank you for your advice and those examples. It really makes sense
to get rid of ssh dependency for a whole package, since it could be used on
the same host.

Andreas, I removed ssh from dependencies and added mention by Graham line
to the run-unit-test.

Regards, Nadiya

On Thu, Jul 27, 2017 at 11:46 PM, Graham Inggs  wrote:

> Hi Nadiya
>
> On 28 July 2017 at 05:35, Nadiya Sitdykova  wrote:
> > 
> --
> > The value of the MCA parameter "plm_rsh_agent" was set to a path
> > that could not be found:
> >
> >   plm_rsh_agent: ssh : rsh
> >
> > Please either unset the parameter, or check that the path is correct
> > 
> --
> >
> > So I also added ssh to dependencies and in fixed the problem.
>
> I recall this seeing this message in other packages in debian-science
> and debichem.  An alternate solution would be to add the following
> line near the beginning of debian/tests/run-unit-test:
>
> export OMPI_MCA_plm_rsh_agent=/bin/false
>
> I believe that ssh is only required when MPI processes execute on
> different hosts.  It should not be required when the MPI processes all
> execute on the same host, like a buildd or a CI tester.
>
> codesearch.debian.net has many examples:
> https://codesearch.debian.net/search?q=path%3Adebian%2F*+
> OMPI_MCA_plm_rsh_agent
> and more specifically for autopkgtests:
> https://codesearch.debian.net/search?q=path%3Adebian%
> 2Ftests%2F*+OMPI_MCA_plm_rsh_agent
>
> Regards
> Graham
>


Re: [Help] Failed to upgrade bcftools to version 1.5

2017-07-28 Thread Graham Inggs
On 28 July 2017 at 16:30, Sascha Steinbiss  wrote:
> Once I'm at it I can also take a shot at bcftools. Never worked on that
> before though, let's see how it goes.

Please take note you'll need to backtract in git until version
1.4.1-2, add the patches, and upload 1.4.1-3.  I don't know any good
way to do this, perhaps Andreas knows what to do.  Otherwise, it can
wait until Monday when I get back, but it would be nice to have
bcftools uploaded before python-pysam.



Re: [Help] Failed to upgrade bcftools to version 1.5

2017-07-28 Thread Sascha Steinbiss
Hi all,

[...]
>> The remaining bugs blocking the transition are augustus (bug #869126) and
>> bcftools (bug #868958).  Sascha has already pushed the augustus changes to
>> git, and it just needs an upload.  The two fixes for bcftools are in the bug
>> report, someone should review them and upload.
>
> Sascha, will you care for this?

Good timing, indeed I was going to finish up the augustus package
tonight. I took some time to finally include all the little executables
that come with the main tool, and I wanted to take another close look to
make sure they all have man pages etc.

Once I'm at it I can also take a shot at bcftools. Never worked on that
before though, let's see how it goes.

Cheers
Sascha



signature.asc
Description: OpenPGP digital signature


Re: [Help] Failed to upgrade bcftools to version 1.5

2017-07-28 Thread Andreas Tille
Hi Graham,

On Fri, Jul 28, 2017 at 03:23:40PM +0200, Graham Inggs wrote:
> On 25/07/2017 11:20, Andreas Tille wrote:
> > So how will we deal with #865006 and issues in python-pysam?
> 
> I made a few minor changes to python-pysam 0.11.2.2 (which builds against
> bcftools 1.4.1) as it stands in git, and managed to get it to build and pass
> its tests.  If there are no objections, I can push my changes and upload on
> Monday.

There are definitely no objections! :-)
If you push soon I can upload tomorrow - on Sunday I'll start traveling
and will not be able to upload before I'll reach DebConf venue
 
> The remaining bugs blocking the transition are augustus (bug #869126) and
> bcftools (bug #868958).  Sascha has already pushed the augustus changes to
> git, and it just needs an upload.  The two fixes for bcftools are in the bug
> report, someone should review them and upload.
> 
> The patches can be grabbed from the Ubuntu package:
> https://launchpad.net/ubuntu/+source/bcftools/1.4.1-2ubuntu2

Sascha, will you care for this?

Kind regards

  Andreas. 

-- 
http://fam-tille.de



Re: [Help] Failed to upgrade bcftools to version 1.5

2017-07-28 Thread Graham Inggs

On 25/07/2017 11:20, Andreas Tille wrote:

So how will we deal with #865006 and issues in python-pysam?


I made a few minor changes to python-pysam 0.11.2.2 (which builds 
against bcftools 1.4.1) as it stands in git, and managed to get it to 
build and pass its tests.  If there are no objections, I can push my 
changes and upload on Monday.


The remaining bugs blocking the transition are augustus (bug #869126) 
and bcftools (bug #868958).  Sascha has already pushed the augustus 
changes to git, and it just needs an upload.  The two fixes for bcftools 
are in the bug report, someone should review them and upload.


The patches can be grabbed from the Ubuntu package:
https://launchpad.net/ubuntu/+source/bcftools/1.4.1-2ubuntu2

Have a good weekend!



Bug#869969: ITP: gfapy -- flexible and extensible software library for handling sequence graphs

2017-07-28 Thread Sascha Steinbiss
Package: wnpp
Severity: wishlist
Owner: Sascha Steinbiss 

* Package name: gfapy
  Version : 1.0.0
  Upstream Author : Giorgio Gonnella
* URL : https://github.com/ggonnella/gfapy
* License : ISC
  Programming Lang: Python
  Description : flexible and extensible software library for handling 
sequence graphs

The Graphical Fragment Assembly (GFA) are formats for the representation of
sequence graphs, including assembly, variation and splicing graphs. Two
versions of GFA have been defined (GFA1 and GFA2) and several sequence
analysis programs have been adopting the formats as an interchange format,
which allow the user to easily combine different sequence analysis tools.

This library implements the GFA1 and GFA2 specification. It is possible to
create a Gfa object from a file in the GFA format or from scratch, to
enumerate the graph elements (segments, links, containments, paths and header
lines), to traverse the graph (by traversing all links outgoing from or
incoming to a segment), to search for elements (e.g. which links connect two
segments) and to manipulate the graph (e.g. to eliminate a link or a segment
or to duplicate a segment distributing the read counts evenly on the copies).

The GFA format can be easily extended by users by defining own custom tags
and record types. In Gfapy, it is easy to write extensions modules, which
allow one to define custom record types and datatypes for the parsing and
validation of custom fields. The custom lines can be connected, using
references, to each other and to lines of the standard record types.



Re: CI for ray

2017-07-28 Thread Graham Inggs
Hi Nadiya

On 28 July 2017 at 05:35, Nadiya Sitdykova  wrote:
> --
> The value of the MCA parameter "plm_rsh_agent" was set to a path
> that could not be found:
>
>   plm_rsh_agent: ssh : rsh
>
> Please either unset the parameter, or check that the path is correct
> --
>
> So I also added ssh to dependencies and in fixed the problem.

I recall this seeing this message in other packages in debian-science
and debichem.  An alternate solution would be to add the following
line near the beginning of debian/tests/run-unit-test:

export OMPI_MCA_plm_rsh_agent=/bin/false

I believe that ssh is only required when MPI processes execute on
different hosts.  It should not be required when the MPI processes all
execute on the same host, like a buildd or a CI tester.

codesearch.debian.net has many examples:
https://codesearch.debian.net/search?q=path%3Adebian%2F*+OMPI_MCA_plm_rsh_agent
and more specifically for autopkgtests:
https://codesearch.debian.net/search?q=path%3Adebian%2Ftests%2F*+OMPI_MCA_plm_rsh_agent

Regards
Graham