[mlir][sparse] use consistent type for COO object and sparse tensor storage

There was a slightly mismatch between the double COO and actual numerical
type in the final sparse tensor storage (due to external formats always
using double). This minor revision removes that inconsistency by using a
properly typed COO and casting during the "add" method instead. This also
prepares alternative ways of initializing the COO object.

Reviewed By: gussmith23

Differential Revision: https://reviews.llvm.org/D107310