Higher order interpolation on regular grids #243
Description
Is your feature request related to a problem? Please describe.
I have my data sampled on a regular grid in the parameter space.
Describe the solution you'd like
I am using a regular grid interpolation, see RegularGridInterpolator: Interpolation on a regular or rectilinear grid in arbitrary dimensions. Supported are “linear”, “nearest”, “slinear”, “cubic”, “quintic” and “pchip”. It would be nice to have this available through EZyRB directly.
Describe alternatives you've considered
The LinearNDInterpolator which is wrapped here is much slower.
For 1D data, the interp1d class is used, but in EZyRB it allows for the default linear interpolation only.
Additional context
#166 suggested new wrappers. Beside the RBF interpolation, I found that the LinearNDInterpolator is wrapped here.
It does not perform well, because the interpolant is constructed by triangulating the input data with Qhull. Moreover, it allows only for linear interpolation. The RegularGridInterpolator takes advantage of the grid structure and allows for higher order interpolation (scipy 1.10 or higher required).