Implement evaluate_at for FCI

[dschwoerer]

I do have a version, that, if I remember correctly right now takes the RZ coordinates and check on the two adjecent R-Z slices, which index that point has, and then interpolate.

I think it would be better to do it field aligned, interpolate to the needed phi, based on R and R_forward and R_backward from the closest slice. That could be done 2nd order. However, to interpolate the data we would need the forward and backward slices of the data, so this sounds like it would need to be done using boutpp, or we would need to reimplement the XZ interpolation routine to get the parallel fields, which we probably want to avoid.