Anders Logg wrote: > On Tue, Aug 18, 2009 at 10:36:54PM +0200, Kristian Oelgaard wrote: >> Quoting Anders Logg <[email protected]>: >> >>> On Tue, Aug 18, 2009 at 08:13:22PM +0100, Garth N. Wells wrote: >>> >>>>>>>>>>> changeset: 1612:de211a9d88bce134c4abc2b128283ad35415a529 >>>>>>>>>>> tag: tip >>>>>>>>>>> user: "Kristian Oelgaard <[email protected]>" >>>>>>>>>>> date: Tue Aug 18 17:57:36 2009 +0100 >>>>>>>>>>> files: ffc/fem/finiteelement.py >>>>>>>>>>> description: >>>>>>>>>>> Restrictions appears to be working for the Poisson demo. >>> Excellent! >>> >>>>> No, only interval, triangle, tetrahedron, quadrilateral and hexahedron >>>>> >>>>> is predefined in UFL. I assume simplices and restrictions currently only >>>>> supports a restriction to one topological dimension lower i.e., >>>>> tetrahedron -> triangle >>>>> triangle -> interval >>>>> >>>> We need a 'facet', which is a triangle in 3D and interval in 2D, so we >>>> write dimension-independent forms (at least for 2D and 3D). >>> That should be fairly easy to add to UFL (__getitem__ in finiteelement.py). >>> You can check if the input argument is facet (which needs to be >>> added), then look at self._cell and figure out the appropriate value >>> of index. >>> >>> Let me know if you have any trouble. >> OK, this functionality should be in place now. Do note that the built-in >> facet >> has no meaning on its own, and it only works with restrictions. I don't know >> how to make this implementation better at the moment (it's still work in >> progress) but feel free to have a look. >> >> Kristian > > Really nice. Now we just need a pair of demos (C++/Python). :-) >
We need to do some work to have it functioning through PyDOLFIN. I think the PyDOLFIN FunctionSpace needs to be extended. Any ideas Johan? Garth > -- > Anders > > > ------------------------------------------------------------------------ > > _______________________________________________ > FFC-dev mailing list > [email protected] > http://www.fenics.org/mailman/listinfo/ffc-dev _______________________________________________ DOLFIN-dev mailing list [email protected] http://www.fenics.org/mailman/listinfo/dolfin-dev
