[Biopython-dev] testseq update (6/13/17)

[Peter Cock]

I commented on GitHub:
https://github.com/biopython/biopython/pull/1269#issuecomment-308071782

I wonder if the complicated import tricks you are trying in order to
be able to do "from Scripts import testseq" or similar could be
avoided by including Scripts/ in sys.path in the same way that
run_tests.py ensures our complied C modules can be imported from the
build/ subdirectory?

Peter


On Tue, Jun 13, 2017 at 8:30 AM, Adil Iqbal <aiqbal85 at gmail.com> wrote:
> TL:DR - Importing issue solved by 'manually_import' function. It can be used
> to solve similar problems in the future. Decided to remove non-Bio doctest
> for problematic versions of python. Open to input/suggestions/criticisms.
> Thank you.
>
> Hello all,
>
> I have solved the 'testseq' importing issue.
>
> Since I could not import 'testseq' directly, I wrote a function that could
> manually import 'testseq' into the 'test_testseq' unit test using the
> built-in module 'importlib'. The function, named 'manually_import,' is
> generalized to work for any future unit tests for modules outside the 'Bio'
> package. You can see the code in isolation here:
> https://github.com/Adil-Iqbal/Personal-Projects/blob/master/Test%20Sequence/manually_import.py
>
> If ever such an issue should arise again, please direct the contributor to
> "Tests/test_testseq.py"
> They can import the 'manually_import' function directly from there.
>
> Than the issue arose that the 'run_tests.py' file could not import
> 'Scripts.testseq' prior to python 3.3. I decided to try using a permutation
> of the 'manually_import' code to solve the problem and it worked! With one
> caveat, the doctests within 'testseq.py' were subject to the same import
> restrictions and copying the 'manually_import' code again within the
> doctests seemed excessive. The solution that worked in the end was to remove
> doctests from outside of the 'Bio' package in python versions prior to 3.3.
>
> Some of you have noted that I am very 'commited' to finishing my work on
> this project. I've downloaded and am now running flake8 tests locally. I'm
> looking into integrating flake8 directly into my IDE. I've also downloaded
> virtualenv, and am looking into other solutions for local testing as well
> (though it's not currently set up). The problem will definitely be addressed
> before I begin work on reverting the seeding. Thank you for helping. It
> means a lot.
>
> Best,
> Adil
>
> _______________________________________________
> Biopython-dev mailing list
> Biopython-dev at mailman.open-bio.org
> http://mailman.open-bio.org/mailman/listinfo/biopython-dev