I’ve actually tried a number of times to search through the documentation, and it has some good info, but some stuff (like your low-level API example above) I couldn’t find anywhere. Most importantly I haven’t been able to find any summary of the options for the Variogram method, just some occasional examples of “GaussianVariogram” and “SphericalVariogram” and “range”, without elucidation.
(The farthest I’ve gotten are [DocStrings]((Docstrings · KrigingEstimators.jl), which doesn’t list the Variogram options, just gives a few examples; and KrigingEstimators, which is apparently a dead end.)
Is there a more explicit documentation page that I’ve missed?
Also, I won’t know the data properties in advance, so the interpolation must be done without human intervention each time. (Interpolation is done by an automated “walker” through the data which interpolates thousands of times in a row; the data environment changes.) Can the Variogram parameters (range? etc.?) be determined directly from the data in each array of samples? In particular, I’m comparing the Kriging interpolation performance vs. Cubic Splines; the spline interpolator acts directly on the data without needing prior knowledge of the data properties.