n-hop neighbors in spatial graph

Description of feature

I think it would be useful to add a function to derive n-hop neighbors.
In some cases, that's better than just using a larger radius since it takes gaps in the tissue properly into account.
For instance, in the case of core needle biopsies, we often have situations like this one, where

(A) = cell of interest
(B) = cell of same tissue section within radius $r_1$, that's reachable via 3-hop neighborhood based on radius $r_2$.
(C) = cell on other tissue section (-> distances are not meaningful with respect to cell A) that's also within $r_1$, but cannot be reached via 3-hop neighborhood based on radius $r_2$.

IMO this would be useful irrespective of the method (radius, delauney), so probably a postprocessing function?

The issue names no file, test, or entry point; start by reviewing Squidpy's existing radius and Delaunay spatial-neighborhood implementations. Use the A/B/C example to define n-hop reachability, and consider how a postprocessing function can work independently of the graph-construction method.