On Fri, 25 Apr 2025 at 02:31, Matthew Samuel <matthemat...@gmail.com> wrote:
>
> I'm reaching out because I'm running into issues that should be pretty
> elementary if I were doing it right. For example, the best representation of
> linear combinations of Schubert polynomials is as a dict of coefficients of
> basis elements, but there are some times when they need to be left
> uncombined, but then you should be able to add to the uncombined terms and
> combine with the cominable ones. sympy does automatically do that, for
> example,
>
> >>> DSx([3,4,1,2]) + DSx([4,1,3,2],"z")
> DSx((3, 4, 1, 2), y) + DSx((4, 1, 3, 2), z)
>
> This is sympy.Add, not a DoubleSchubertAlgebraElement, because the sets of
> coefficient variables, y and z, are different. This is what I want. When we
> do this, however,
>
> >>> DSx([3,4,1,2]) + DSx([4,1,3,2],"z") + DSx([3,4,1,2])
> 2*DSx((3, 4, 1, 2), y) + DSx((4, 1, 3, 2), z)
>
> This looks correct, however that 2*DSx((3, 4, 1, 2), y) is a sympy.Mul
> object, when what I want is for it to be internally represented as
> {(3,4,1,2): 2}, which it is not.
>
> I feel like this is similar to the Poly class being a subclass of Basic
> instead of Expr, but I don't know if I want to be that rigid. And combining
> terms as Expr alone is inefficient. These can have hundreds or thousands of
> terms after multiplying, and the hashmap of keys is the most efficient way to
> combine coefficients.
My first thought is that you should probably use something more like
Poly rather than Expr here. Combining as Expr alone will be
inefficient.
Does a typical calculation here use a closed finite set of DSx elements?
If so you could use e.g. a sparse matrix to represent the linear combination.
> I guess this is a question? What's the most elegant way to get that 2 into
> that dict?
If you are working with Expr then you probably just want to use
something like Expr.replace.
--
Oscar
--
You received this message because you are subscribed to the Google Groups
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sympy+unsubscr...@googlegroups.com.
To view this discussion visit
https://groups.google.com/d/msgid/sympy/CAHVvXxSYsJ_YNEn1O38DcJbFx4g7ojtmg%3D_xp6qHGhCbWyeGPQ%40mail.gmail.com.
