Implement variable resolution

We have data of the resolution of the instruments as a function of wavelength per order

Add the option to take this as an input and figure out the correct sampling scheme to have a variable resolution. if not, at least do it by order