Thanks John for the answer. On Tue, 2022-06-07 at 08:41 -0700, John H Palmieri wrote: > Isn't this because you're using CachedRepresentation? From the documentation: > > Instances of a class have a *cached representation behavior* when several > instances constructed with the same arguments share the same memory > representation.
Yep, this is precisely the case. Sorry if I was not clear. What I meant is that this is a downside for me, and that I would like to prevent this behavior. Antoine > > On Tuesday, June 7, 2022 at 6:18:50 AM UTC-7 antoine....@gmail.com wrote: > > I thank all of you for your answers. And sorry for taking a bit long to > > reply! > > > > After many trys and errors, here is something that works for me: > > > > ``` > > > > from sage.structure.sage_object import SageObject > > from sage.structure.unique_representation import CachedRepresentation > > > > class MyInteger(CachedRepresentation, SageObject): > > > > @staticmethod > > def __classcall_private__(cls, n): > > if n % 2 == 0: > > return MyInteger_even(n) > > else: > > return cls.__classcall__(cls, n) > > > > def __init__(self, n): > > self.n = n > > > > > > class MyInteger_even(MyInteger): > > > > def __init__(self, n): > > self.n = n > > self.half = n // 2 > > ``` > > > > However, this has a major drawback, as all instances that have same data are > > references to one another. And I don't find any way to bypass that: > > > > ``` > > > > sage: from copy import deepcopy > > sage: s1 = MyInteger(2) > > sage: s2 = MyInteger(2) > > sage: s3 = deepcopy(s2) > > sage: s1 == s2 > > True > > sage: s1 is s2 > > True > > sage: s1 == s3 > > True > > sage: s1 is s3 > > True > > sage: s1.n > > 2 > > sage: s3.n = 5 > > sage: s1.n > > 5 > > sage: > > ``` > > > > I appreciate that this may not be a problem for classes representing sets > > and > > categories (like `EuclideanSpace`). But for classes representing children > > elements, this may cause problems. And this is the case for me, as I need > > this > > for `FiniteDrinfeldModule` and `FiniteDrinfeldModule_rank_two` > > (https://trac.sagemath.org/ticket/33713). > > > > Is there a way to avoid this? > > > > Kindest regards, > > Antoine > > > > On Sat, 2022-05-14 at 21:07 -0700, Nils Bruin wrote: > > > It's probably worth pointing out that __classcall_private__ is not a > > standard > > > python facility. It looks like you need > > > > > > class Foo(metaclass=ClasscallMetaclass): > > > > > > to make it work on Foo. > > > > > > On Saturday, 14 May 2022 at 20:21:48 UTC-7 Travis Scrimshaw wrote: > > > > For this you want to use __classcall_private__ as otherwise you would > > likely > > > > end up in an infinite loop when you try to construct the subclass. There > > are > > > > lots of examples of this in the Sage library code. > > > > > > > > Best, > > > > Travis > > > > > > > > > > > > On Sunday, May 15, 2022 at 5:42:49 AM UTC+9 Eric Gourgoulhon wrote: > > > > > Hi, > > > > > > > > > > Le samedi 14 mai 2022 à 00:07:02 UTC+2, David Roe a écrit : > > > > > > 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 > > > > > > > > > > > > > > > > > > > > > > > > > > > An alternative is to use __classcall_private__ > > > > > For an example, see the class EuclideanSpace in > > > > > src/sage/manifolds/differentiable/examples/euclidean.py > > > > > > > > > > EuclideanSpace(n) actually returns an instance of the subclass > > > > > EuclideanPlane if n = 2 or of the subclass Euclidean3dimSpace if n = > > > > > 3. > > > > > > > > > > Best wishes, > > > > > > > > > > Eric. > > > -- > > > You received this message because you are subscribed to a topic in the > > Google > > > Groups "sage-devel" group. > > > To unsubscribe from this topic, visit > > > https://groups.google.com/d/topic/sage-devel/PaUReuoxEXI/unsubscribe. > > > To unsubscribe from this group and all its topics, send an email to > > > sage-devel+...@googlegroups.com. > > > To view this discussion on the web visit > > > > > https://groups.google.com/d/msgid/sage-devel/a8bc696f-7887-41b2-8fe9-af3ac055eb71n%40googlegroups.com > > > > > . > > > > > -- > You received this message because you are subscribed to a topic in the Google > Groups "sage-devel" group. > To unsubscribe from this topic, visit > https://groups.google.com/d/topic/sage-devel/PaUReuoxEXI/unsubscribe. > To unsubscribe from this group and all its topics, 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/ba864814-83af-4f4f-b5d7-e08f6108b551n%40googlegroups.com > . -- 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/36cf3378b59a5ede016af22760a657af8a553ec9.camel%40gmail.com.
