SANS (ZOOM & LARMOR): rocking scan

As an instrument scientist on ZOOM, I would like to be able to perform a rocking scan.

This scan involves:
- Changing a motor (angle) of the sample in the beam
- Measuring from two distinct regions-of-interest on the detector (call these measurements `A` & `B`). For clarity these regions-of-interest will be provided as numpy arrays of spectrum ids.
* Scientifically, A and B correspond to a region on the left-hand side and the right-hand side of the detector. The aim of this scan is to center a crystal in the rotation axis, aiming to get the same intensity on both the left and right hand side of the detector.
- No time-of-flight or wavelength chopping is required for this scan. The ROIs are purely spectrum numbers.
- As the angle is varied, both `A` and `B` will trace out non-overlapping, but similarly-shaped, gaussians. In other words the two gaussians are expected to have similar fit parameters (height, width, background), **except** that the center (`x0`) parameter will be different.
- We want to find the location where the "A" peak and "B" peak are equal to one another, or in other words, a point equidistant from the centre of peak A and the centre of peak B.
* This is achievable with two independent Gaussian fits, for A and B, and then taking the average position of the `x0` parameter of the two fits in an instrument-level plan wrapper (check using historical data linked below - but this has been confirmed with RD as a sensible approach)
* Reducer(s) need to be written/extended to expose _multiple_ regions of interest simultaneously, derived from the same DAE data. We should use the spectra-map under the hood as the number of pixels in both regions of interest may be large.

### Acceptance criteria
- [ ] Implement
- [ ] Test on ZOOM or another SANS beamline