Convenience methods for CartesianIndices?

[tlnagy]

I find myself running something like the following a lot:

sample_points = collect(hcat([[point.I...] for point in findall(outline)]...));
eval_points = collect(hcat([[point.I...] for point in findall(hole)]...));

itp = interpolate(Shepard(), sample_points, data[findall(outline)]))
vals = ScatteredInterpolation.evaluate(itp, eval_points)

What do you think about adding support for CartesianIndices so that collections of them can be passed directly to interpolate and evaluate?

Also thanks for the great package! I'm really happy to not have call out to Python to do scattered interpolations any more.

[eljungsk]

Hi,

Thanks for the feedback!

I'm not fully sure that I understand your use case. If you have a collection I of CartesianIndex into an AbstractArray A, don't you get an Array that you can use as input to interpolate and evaluate by doing A[I]?