[Biopython-dev] 1.62b test coverage report

[Ben Fulton]

Fixed the following:

I had installed Mafft version 6.850-1 from apt-get, which apparently is
more than a year old and doesn't work. The tests ran after I installed it
from source.

I had not gotten a path set up properly for XXMotif; once I did the tests
all ran.

The DiAlign tests passed after I removed the precondition checks.

Did not fix:

The site http://www.rubic.rdg.ac.uk/~mab/software.html is down, and I can't
find anywhere else to install the PopGen software from.


So with all of those modifications, I ran coverage against the latest code
from GitHub. Results are once again available on my website,
http://benfulton.net/BioPython162_Coverage , and the following issues
remain:

EmbossPhylipNew - skipped, too hard to install
Fasttree - error, apparently a versioning issue
PopGen_FDist and PopGen_DFdist - skipped, unavailable
Prank - failed, recent versions of the tool have some kind of output change




On Fri, Aug 2, 2013 at 8:00 AM, Peter Cock <p.j.a.cock at googlemail.com>wrote:

> On Fri, Aug 2, 2013 at 10:31 AM, Peter Cock <p.j.a.cock at googlemail.com>
> wrote:
> >
> >> The NCBI_BLAST error involves rpsblast not being in the install.
> >> Version 2.2.25-7 using apt-get.
> >
> > I believe this is down to an NCBI stupidity with binary name
> > clashes, both the old 'legacy' C BLAST and the new C++
> > BLAST+ suite have a binary called rpsblast.
> >
> > Our test code copes with this by searching the path and checking
> > each rpsblast binary found - looking for the new version only.
> >
> > However, Debian policy is to resolve ambiguities like this with
> > a unilateral renaming - in this case I *think* they called the new
> > binary rpsblast+ instead. Can you confirm that? I don't have
> > access to a Debian machine right now.
>
> Certainly this was their plan and was done on Bio-Linux,
> http://lists.debian.org/debian-med/2011/05/msg00025.html
>
> > So, strictly speaking the Biopython test is correct - you don't
> > have the new rpsblast installed. However, it would be more
> > helpful if we also checked for the Debian alias rpsblast+ too.
> >
> > That shouldn't be too complicated to do - especially if you
> > could rerun the tests using Biopython from git for me?
>
> This commit is now on our master branch,
>
>
> https://github.com/biopython/biopython/commit/148b681a66061cc03d70f940a2efdede29adc64a
>
> Thanks,
>
> Peter
>