1765 Simplify gridscan parameter hierarchy

[rtuck99]

Part 2 of a fix for #1765

Requires:

• 1765 Refactor grid scan params #1787
• mx-bluesky#1765 Refactor grid parameters dodal#2112

Link to dodal PR (if required): #XXX
(remember to update pyproject.toml with the dodal commit tag if you need it for tests to pass!)

Below is the parameter class Hierarchy for GenericGrid before this refactor:

Here is the parameter class Hierarchy of DiffractionExperimentWithSample after the refactor (GenericGrid and related subclasses are now completely gone)

A note on changes

The internal parameter model has changed significantly and is now simpler.

The GenericGrid and subclasses portion of the parameter hierarchy is being removed and replaced with GridScanParams and GridDetectionParams.

detect_grid_and_do_gridscan now only requires DiffractionExperimentWithSample and GridDetectionParams
the grid detection plan now return GridScanParams
DiffractionExperimentWithSample and GridScanParams is all that is needed to perform a grid scan.

The grid detection and grid scan plans are now properly genericised with type parameter TParameter denoting the concrete parameter type extending DiffractionExperimentWithSample, so that it can be passed to the BeamlineSpecific portions and also deserialized in the callbacks.

DiffractionExperimentWithSample hierarchy now concentrates mainly on general-purpose diffraction parameters such as exposure timings, transmission fraction, and sample metadata.

DetectorParams is now decoupled from the gridscan parameter model, meaning it is no longer constructed when a property on the parameter model. Instead it is explicitly constructed in the experiment plan from the parameters, and passed in explicitly into the gridscan plan and to the callbacks.

This means that many subclasses can be removed from the parameter model, and it is possible to run 2d, 3d, gridscans and grid detection in any combination without requiring modifications to the parameter model.

The rotation scan parameter model has not been changed; it still has a detector_params property.

Instructions to reviewer on how to test:

1. Do thing x
2. Confirm thing y happens

Checks for reviewer

• Would the PR title make sense to a user on a set of release notes

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.07%. Comparing base (b15560d) to head (0ef0ae4).

Additional details and impacted files

@@                        Coverage Diff                         @@
##           1765_refactor_grid_scan_params    #1786      +/-   ##
==================================================================
- Coverage                           93.10%   93.07%   -0.03%     
==================================================================
  Files                                 164      163       -1     
  Lines                                9046     8958      -88     
==================================================================
- Hits                                 8422     8338      -84     
+ Misses                                624      620       -4     

Components	Coverage Δ
i24 SSX	77.16% <ø> (ø)
hyperion	98.76% <100.00%> (-0.02%)	⬇️
other	98.21% <100.00%> (+0.06%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[DominicOram]

Generally this all looks a lot cleaner than before. I only noticed the one glaring thing but obviously should be looked at properly by someone in MX

[DominicOram]

Must: Doesn't this risk filling the detector params with the old energy before we change it below?

[rtuck99]

This should be ok, since this will only get populated if we didn't specify it, in which case it will be the current energy, and set_energy_plan won't set the energy (since it takes it from the original parameters demand_energy_ev, which will be None in this case)

In practice as the comment says we always specify an energy. This whole workflow only exists for the standalone robot load plan which is no longer exposed. I have to say I don't really like it as it makes the whole plan much more complicated, and am tempted to remove it.