[Dune-devel] [GSoC2016] Proposal about the Python bindings for the DUNE grid interface

Andreas Dedner a.s.dedner at warwick.ac.uk
Wed Mar 23 11:32:06 CET 2016


Dear Michael.
One last remark concerning the build system:
There is already an approach available of how to use the dune build
system for modules shipping python code:
  https://gitlab.dune-project.org/quality/dune-python.git
You should definitely take a look at that.
Best
Andreas

On 23/03/16 05:26, Michaël Sghaïer wrote:
> Dear Andreas,
> dear Martin,
> 
> thank you a lot for your feedbacks, they have been very useful!
> 
> 2016-03-21 4:09 GMT-04:00 Martin Nolte <nolte at mathematik.uni-freiburg.de
> <mailto:nolte at mathematik.uni-freiburg.de>>:
> 
>     However, there is one aspect you seem to have hardly covered: the
>     build system. Apart from being able to find python using CMake --
>     which would not really be worth mentioning -- we will need to
>     generate quite a few python modules, e.g., to support the different
>     grid implementations. Therefore, I would suggest to have a look at
>     CMake, too.
> 
> 
> You are right, I forgot about the build system. I added two weeks to my
> schedule to take care about it (see the google doc of my proposal).
> 
> Following your comments on my google doc, I made some (big) changes,
> please take a look. I especially divided the project on the
> serial/parallel aspect of a grid implementation. Tell me if I am being
> realistic about my milestones and my timing or not.
> 
> By the way, you didn't answer my questions about dune-fempy. It don't
> bother me to have information about that later but I have to know if two
> weeks allocated to combine my bindings to dune-fempy are enough or not,
> thank you.
> 
>  2016-03-21 4:09 GMT-04:00 Martin
> Nolte <nolte at mathematik.uni-freiburg.de
> <mailto:nolte at mathematik.uni-freiburg.de>>:
> 
>     As Andreas already said, wrapping a few more methods might be
>     interesting, too. You could also have a look at the FieldMatrix,
>     especially at the operator[] (which returns a reference to a
>     FieldVector).
> 
> 
> Thanks for this suggestion, I am going to work on it! 
> 
> Best,
> 
> Michaël
> 
> 
> 
> _______________________________________________
> Dune-devel mailing list
> Dune-devel at dune-project.org
> http://lists.dune-project.org/mailman/listinfo/dune-devel
> 




More information about the Dune-devel mailing list