[Biopython-dev] biopython-1.51 test log, sage inclusion

[Peter]

On Mon, Aug 17, 2009 at 4:42 PM, Marshall Hampton<mhampton at d.umn.edu> wrote:
>
> Hi,
>
> I am preparing biopython-1.51 for inclusion as an optional package for Sage
> (www.sagemath.org).  I ran the test suite and got 8 errors; I am not sure if
> these are all expected.

I wouldn't have expected any failures.

> The KDTree ones I have seen before, but some look new.  My test log is
> available at:
>
> http://sage.math.washington.edu/home/mhampton/biopython-1.51-testlog.txt
>
> in case anyone wants to take a look.

This one should be simple: test_EMBOSS.py
ValueError: Disagree on file ig IntelliGenetics/VIF_mase-pro.txt in
genbank format: 16 vs 1 records
This is a known regression in EMBOSS 6.1.0 which will be fixed
in their next release. Can you check this by running embossversion?

The others are all ImportErrors (e.g. cannot import name _CKDTree)
I rather suspect you are running the test suite BEFORE compiling
the C extensions, and that this may similarly affect Bio.Restriction.

> Biopython has been available in Sage for several years as an optional
> package, but I would like to make it a standard component. This has
> has become much more likely since the clean-up of Numeric and
> mx-texttools dependencies.

Cool.

> I think the only real issue is setting up some testing during the Sage
> package installation, which is my motivation for  really understanding
> the test failures.

I don't know anything about your test framework, but surely other
packages (e.g. NumPy) have a similar requirement (compile
before test) so this should be fixable.

Regards,

Peter