[gecode-users] Python wrappers

[Ellis Breen]

Interesting... That looks more like a code-generation approach to MiniZinc,
which again could make deployment a little more difficult than straight
interpretation (I intend to run Minizinc programs on numerous Buildbot
slaves), but should result in more efficient code. I guess this should be
automatable.

In my current setup, feeding Minizinc into a static application would be
much preferable, so I guess it's the mzn2fzn route for now.

Perhaps the Numberjack API could be extended to support Flatzinc for
applicable solvers?

I guess this might be considered overkill, as it defines its own high-level
interface. Maybe I need to look at using Boost::Python to talk to
Numberjack, and code my constraint descriptions in that.

Ellis

2009/10/29 Christian Schulte <cschulte at kth.se>

> Hi,
>
> There is a direct MiniZinc to Gecode interface. However, my take on the
> paper is that it is a little confusing (mildly put) in places. I am also
> not
> sure whether that work extends beyond paper:
>        http://www.dmi.unipg.it/CILC08/27_DovierCiprianoMauro.pdf
>
> Christian
>
> --
> Christian Schulte, www.ict.kth.se/~cschulte/
>
>
> -----Original Message-----
> From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On Behalf
> Of Guido Tack
> Sent: Thursday, October 29, 2009 9:11 PM
> To: Ellis Breen
> Cc: users at gecode.org
> Subject: Re: [gecode-users] Python wrappers
>
> Hi Ellis,
>
> Ellis Breen wrote:
> > I'm currently looking for a free-as-in-beer constraint programming
> > library with bindings in both C/C++ and Python. It seems Logilab
> > started a Python wrapper for Gecode but never completed it. Are
> > there any rival implementations? Would it be non-trivial to use SWIG
> > for this?
>
> we're currently implementing an interface to the Numberjack Python
> library (http://4c110.ucc.ie/numberjack/home).  It's work in progress,
> but we expect enough progress for it to be usable soon.  The interface
> is generated using SWIG.
>
> > Apparently Cassowary also supports Python but I can't seem to find
> > the bindings.
> >
> > Also, has anyone attempted to integrate the Minizinc to Flatzinc
> > compiler into a project containing Gecode, to allow interpretation
> > of Minizinc scripts without using an intermediary program?
>
> That's not so easy, because mzn2fzn is written in Mercury, and
> currently only available as a stand-alone program.  It's should be
> possible to wrap the Mercury code in a C library, but that hasn't been
> done yet.
>
> Cheers,
>        Guido
>
>
> _______________________________________________
> Gecode users mailing list
> users at gecode.org
> https://www.gecode.org/mailman/listinfo/gecode-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ps.uni-sb.de/pipermail/users/attachments/20091029/fd792887/attachment.htm>