On Fri, Mar 23, 2012 at 07:33:01PM -0700, David Roe wrote:
> Unfortunately, whenever I write a Python class I've also been very
> frustrated trying to get the class I'm writing to inherit from the
> appropriate element_class. As a consequence I try to write most of my
> elements in Cython. :-)
>
> The inheritance is set up to work correctly if you follow the paradigm of
> just defining an ElementMethods class on the category (see basically any
> category in Sage) and throwing methods into it. You could just inherit
> from Rings().element_class (for example), but this doesn't work very well
> if you want to use one element class for elements of parents in different
> categories (e.g. Z/NZ for both prime and composite N).
>
> Frankly, this is one of my biggest frustrations with the current category
> framework. But maybe I'm just missing something. :-)
I think you are :-) At least, I very much hope so.
In principle all a parent P needs to do is to setup its category C in
__init__, and have P.Element be (or point to) the base class for its
elements. Then, P.element_class will inherit both from P.Element, and
from C.element_class. Note that C can typically depend on the
arguments passed down to __init__.
Of course, the construction of elements itself should use
P.element_class, not P.Element. This is what the default
implementation of _element_constructor_ does, so usually you need not
worry about it.
sage: TestSuite(P).run()
will complain if things are not done right.
For plenty of examples, have a look at:
sage: A = Sets().example()
sage: A??
and friends.
Cheers,
Nicolas
--
Nicolas M. Thiéry "Isil" <nthi...@users.sf.net>
http://Nicolas.Thiery.name/
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
