I think the following should work:
class MyObject:
def __classcall__(cls, arg):
if isinstance(arg, special):
return typecall(MyObject_specific_case, arg)
else:
return typecall(MyObject, arg)
plus the same __init__ you had before. I haven't checked it though....
David
On Fri, May 13, 2022 at 11:38 AM Antoine Leudière <
antoine.leudi...@gmail.com> wrote:
> Hi,
>
> I have a class `MyObject` and a class `MyObject_specific_case` that
> inherits `MyObject`:
>
> ```
> class MyObject:
> def __init__(self, arg):
> # Do stuff
>
> class MyObject_specific_case:
> def __init__(self, arg):
> # Do stuff
> super().__init__(arg)
> ```
>
> It is easy to decide if the object is a 'specific case' by looking at `arg`.
> Therefore, I would like some sort of mechanism that autonomously decides to
> instantiate `MyObject` or `MyObject_specific_case` depending on the value
> of `arg`. The ideal interface would be:
>
> ```
> sage: my_object = MyObject(gen)
> sage: # Assume arg corresponds to the specific case:
> sage: isinstance(my_object, MyObject_specific_case)
> True
> sage: isinstance(my_object, MyObject)
> True
> sage: # Assume arg does not correspond to the specific case:
> sage: isinstance(my_object, MyObject_specific_case)
> False
> sage: isinstance(my_object, MyObject)
> True
> ```
>
> My question is: what is the idiomatic way to do this?
>
> Naively, one could use a function that takes `arg` as argument, but we
> would want both the function and the class to be named `MyObject`, so
> there is a collision. From what I understand, Sage typically uses a class
> along the lines of `UniqueRepresentation` or `UniqueFactory`. However,
> those seem to do way more than what I ask (
> https://doc.sagemath.org/html/en/reference/structure/sage/structure/unique_representation.html,
>
> https://doc.sagemath.org/html/en/reference/structure/sage/structure/factory.html),
> and I am a bit lost.
>
> P.-S. : In real life, `MyObject` is `FiniteDrinfeldModule` and `
> MyObject_specific_case` is `FiniteDrinfeldModule_rank_two` (see
> https://trac.sagemath.org/ticket/33713).
>
> --
> You received this message because you are subscribed to the Google Groups
> "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to sage-devel+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/sage-devel/92c5b512-f235-474b-ab18-6f4e26dacbf9n%40googlegroups.com
> <https://groups.google.com/d/msgid/sage-devel/92c5b512-f235-474b-ab18-6f4e26dacbf9n%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/sage-devel/CAChs6_kdkdHDMj0EZ-mKwLVi42%2ByRv_Ns7u%3Dbh0Ghw3X3vFK7A%40mail.gmail.com.
