brianguenter/FastDifferentiation.jl

add function to generate sparse matrix with correct sparsity pattern for Jacobian or Hessian

Closed this issue · 1 comments

The user can't know what the sparsity pattern will be in general so can't use in_place option with sparse arrays except by calling make function with in_place=false and then using resulting matrix.

Write a function that doesn't generate code, just evaluates sparsity pattern and returns a spase array that matches the function so people can use in_place more conveniently.

figured out how to do this trivially without a new function. Just needs a documentation update.