SDE - share computation common to drift and diffusion terms

Hi,

Thanks for the great library!

I'm solving a SDE of the form: $$d\mathbf{X}(t) = (\mathbf{u}(t, \mathbf{X}(t)) + (\nabla \cdot \mathbf{K})(t, \mathbf{X}(t)))dt + \mathbf{V}(t, \mathbf{X}(t)) \cdot d\mathbf{W}(t)$$ where $\mathbf{X}$ is the position vector, $\mathbf{u}$ the velocity vector field, $\mathbf{K}$ is the diffusivity tensor field and $\mathbf{K}=\frac{1}{2} \mathbf{V} \cdot \mathbf{V}^T$.

As you can see, both the drift and diffusion terms depend on the diffusivity tensor field $\mathbf{K}$. In pratice, I'm computing it from the velocity field $\mathbf{u}$.
Is there a way to avoid precomputing $\mathbf{K}$ across the entire domain and instead compute it locally (ie. using only a small neighborhood arount $\mathbf{X}(t)$ ) at each integration step, without having to recompute it separately for both the drift and diffusion terms?

Vadim

No file, test, or entry point is named. Start by tracing how drift and diffusion terms are evaluated for stochastic differential equations, then determine the API and validation needed to share locally computed diffusivity data without precomputing the full domain.