Add a workflow or provider that can stitch multiple reflectivity curves

[nvaytet]

In the Amor workflow, we have multiple reflectivity curves for multiple sample rotation angles. It is common practice to stitch these curves together to obtain a single reflectivity curve that covers a wide Q range. We should have a workflow that can achieve that: take in multiple files, compute I(Q) and scale each curve to stitch them together.

Open questions:

• Is the scaling of each curve just arbitrary or is it based on a physical process? I.e. do we just use some sort of iterative procedure to get the best overlap between curves, or do we need to compute the scaling based on the experiment parameters?
• Is the scaling just a single number applied to all Q bins, or is it Q-dependent?

[paracini]

The scaling is arbitrary and based on the best match of the two curves in the overlapping q range so an iterative procedure would work. Different facilities approach this differently, some merge the data in the overlap region and some leave all data points from both curves. I prefer this second option, without merging.

The scale is a single number applied to all Q bins.

It's useful to have in this procedure some diagnostics on what is the overlapping q range available in the stitching region and if this is below a certain threshold add a warning about possible issues arising from too few points.

A note for future reference on this topic: We can stitch data from AMOR because the wavelength $\Delta\lambda/\lambda$ is constant due to the double blind chopper which means $\Delta q/q$ is constant. That won't be the case on Estia and FREIA where we won't have constant $\Delta\lambda/\lambda$ and therefore $\Delta q/q$ which means overlapping regions will have different resolutions, we don't foresee to stitch in those cases