[Biopython-dev] Testings on Win 7 x64

[Peter Cock]

On Thu, Apr 24, 2014 at 5:34 PM, Manlio Calvi <manlio.calvi at gmail.com> wrote:
> OK I launched the tests on all my serpent's nest :)
>
> As some changes are made on .git folders and files and seeing how
> badly git on windows manage those I erased the biopython folders and
> started from scratch, that's the results:
>
> Python 3.4 x64 NO error (yay!):
> The only note here is the fact that bio.trieand KDtree get skipped
> because the C module refuse to compile.

It seems none of the compiled C code is getting tested :(
http://lists.open-bio.org/pipermail/biopython-dev/2014-April/011270.html

> Python 3.3 (WinPython 3.3.3.3 x64 2 errors (one may be temporary):
> CDAO test fail with (an issue we already seen):
>
> ======================================================================
> ERROR: test_write_0 (test_Phylo_CDAO.WriterTests)
> Write and re-parse the phylogenies in test.cdao.
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File ".\test_Phylo_CDAO.py", line 62, in test_write
>     CDAOIO.write([t1], outfile)
>   File "C:\Users\xxx\3.3 - Biopython
> test\WinPython-64bit-3.3.3.3\python-3.3.3.amd64\biopython\build\lib.win-amd64-3.3\Bio\Phylo\CDAOIO.py",
> line 74, in write
>     return Writer(trees).write(handle, plain=plain, **kwargs)
>   File "C:\Users\xxx\3.3 - Biopython
> test\WinPython-64bit-3.3.3.3\python-3.3.3.amd64\biopython\build\lib.win-amd64-3.3\Bio\Phylo\CDAOIO.py",
> line 268, in write
>     handle.write('@prefix %s: <%s> .\n' % (k, v))
> TypeError: 'str' does not support the buffer interface
> ---------------------------------------
>
> Also during the test this line popped over and over
>
> file://C:\Users\xxx\3.3 - Biopython
> test\WinPython-64bit-3.3.3.3\python-3.3.3.amd64\biopython\Tests\CDAO\test.cdao/tree0000001#edge_annotation0000071
> does not look like a valid URI, trying to serialize this will break.
>
> --------------------------------

See this thread - no replies yet so maybe we'd better note this as
a GitHub issue,
http://lists.open-bio.org/pipermail/biopython-dev/2014-April/011294.html


>
> The other isssue is related with NCBI qblast but being NCBI probably
> was I hit a server down.

I've noticed the NCBI blast website is currently very slow :(

> ######################################
>
> Python 2.7 x64 NO issues!
>

Excellent - we are making progress (slowly).

>
> Adding to that I hit a "strangeness" I don't understand because the
> mentioned scipy module IS installed, the test pass but not with
> "flying colours":
> test_CodonAlign ... test_CodonAlign.py:156: UserWarning: Importing
> scipy.linalg.expm failed. Skip testing ML method for dN/dS estimation
>   warnings.warn('Importing scipy.linalg.expm failed. Skip testing ML
> method for dN/dS estimation')
> test_CodonAlign.py:164: UserWarning: Importing scipy.optimize.minimize
> failed. Skip testing ML method for dN/dS estimation
>   warnings.warn('Importing scipy.optimize.minimize failed. Skip
> testing ML method for dN/dS estimation')
> ok

The CodonAlign stuff is new, this might need looking at:
https://github.com/biopython/biopython/pull/259
http://lists.open-bio.org/pipermail/biopython-dev/2014-April/011278.html

> Second a more general signalation, I've seen some test depends on PIL
> but PIL library get discontinued and substituted by one called Pillow
> that is said compatible with it predecessor (but I'm not sure how much
> it is). This is a more general issue because require an scan of the
> entire codebase. Looking around seems to me the NLTK package is in a
> situation like this, too.

That probably refers to test_GraphicsBitmaps.py which is trying to
check if ReportLab will be able to make PNG files. If ReportLab
can now use Pillow or PIL, we should allow for that...
https://github.com/python-imaging/Pillow

(As an aside ReportLab's Python 3 support is still quite new -
but I don't recall seeing any problems from Biopython's usage
of it.)

Peter