[Biopython-dev] [biopython] pairwise2: Allow insertion after deletion and performance improvements (#782)

Sat Mar 19 01:48:30 UTC 2016

Hi Markus,
Great to hear you are willing to work further on this module.
I would suggest to write the documentation first, not for Bio.pairwise2 as it currently is, but for Bio.pairwise2 as you would like it to be. Then we can discuss on the mailing list, and once we agree on the design of the module we can make further changes to the code. This will make it easier for Biopython users and developers to try and test the code while it's under development.You may consider also that a module for pairwise alignment may fit better under Bio.Align rather than as a standalone module. That would also allow us to develop a pairwise alignment module under Bio.Align while keeping Bio.pairwise2 untouched. Bio.pairwise2 can then be deprecated over time.
-Michiel

    On Wednesday, March 16, 2016 7:53 PM, PD Dr. Markus Piotrowski <Markus.Piotrowski at ruhr-uni-bochum.de> wrote:

   Dear Michiel,

 I'm happy to work further on the module in the future. E.g. I think there are still some performance gains possible when using *score_only* (and I wonder if I will be able to implement Miller and Myers algorithm which works in linear space).
 Also, some of the interesting parameters like *penalize_end_gaps* and *score_only* should really be mentioned in the doctest part.
 I also agree that the user interface is quite unpythonic (I'm still not sure if I have understood how the decoding is going on...).
 Regarding the cpairwise2module.c I have to say that this is the first time I coded in C (it took me 2 days to until I was able to compile it under Windows, with at least one day installing/de-installing/re-installing Microsofts Visual Studio), thus I tried to introduce as few changes as possible. Suggestions are very welcome.

 Markus

 Am 16.03.2016 um 05:41 schrieb mdehoon:

 @peterjc , @MarkusPiotrowski : Thanks for your contribution.
 I think that the pairwise2 code has not had a dedicated maintainer since 2005, and while we have done some maintenance on it in the past, I don't think we currently have a real expert for this code among the Biopython developers. So I would suggest to accept the changes by @MarkusPiotrowski as is, as he has looked at the code in much detail and is likely the most familiar with the code now. I also think that the user interface to pairwise2 is rather unpythonic, with arguments encoded in the function names, and it currently lacks documentation. I would be in favor of a further revision of this module (after accepting the current changes) to make it more intuitive and pythonic to users. @MarkusPiotrowski if you'd like to further revise this module, let's discuss on the mailing list. The only comment on the submitted code is that some functions in cpairwise2module.c (for example calc_affine_penalty) could be declared static if they are not used outside of this file. —
 You are receiving this because you were mentioned.
 Reply to this email directly or view it on GitHub     

_______________________________________________
Biopython-dev mailing list
Biopython-dev at mailman.open-bio.org
http://mailman.open-bio.org/mailman/listinfo/biopython-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.open-bio.org/pipermail/biopython-dev/attachments/20160319/f9de801b/attachment.html>