Look into uncertainty propagation and usage for flux calculation in resample

We could add uncertainty handling for the Interpolator resamples (currently spline and linear). We can also look into using the uncertainties as weighting for the final flux calculation of the resampling, but I'm not sure when and where this would be appropriate.

Start by locating the Interpolator resampling implementation and the spline and linear paths. Determine how uncertainties are represented during resampling and where the final flux calculation occurs. The work is done when an agreed approach for uncertainty propagation and any appropriate uncertainty weighting is implemented and covered by relevant tests.

Written by the indexing model from the issue text.