Am 2014-05-14 11:52, schrieb Nathann Cohen:
> Now, it would be COOL if this function could cache its boolean/integer
> answers (for we ask the same questions very often), but caching the
> matrices is a really (really) bad idea.
>
> The return type is determined by an argument of orthogonal_array,
> existence=True, and we only want to cache answers to that.
>
> 1) Is there a way to obtain this behaviour with @cached_function ?
> 2) If not, do you think that implementing this is a nice idea ? (I am
> worried of making the standard use case slower)
+1 to implement filtering what is cached and what not (if it does not
slow down the code significantly; from which I think is not the case).
FYI: I have a similar problem at the moment: I performing an operation
on symbolic expressions. This is done recursively and I want to cache it
to improve the speed of my function. The problem is, that I ran out of
memory, since the input is sometimes very large, so I want to cache only
"smaller" parts. For me "smaller" is whenever operator!=operator.mul of
the symbolic input.
I wrote a workaround to delete out the not wanted elements of cache
directly (every couple of "iterations"), but this is not a very nice
solution. So I'd appreciate if cached_function can do that for me.
Maybe one could call the function then by:
@cached_function(filter=lambda s: s.operator() != operator.mul)
def my_function(s):
pass
Maybe there should be input_filter (=filter) and output_filter instead.
Daniel
--
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 post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
