Jungfrau testing notes

With @jacob720 and @adaudon

- Names of devices in composite need to match names of devices in device module . Specifically, we had name mismatches between `RotationScanComposite` in `beamlines/i24/jungfrau_commissioning` and the device names in dodal's i24 module
- Similarly, the metadata writer kept trying to do `data.get("detector") when the device name is currently `commissioning_jungfrau`. Also name was wrong when injecting jungfrau on the dark plans.
- Currently hardcoding det y to an outdated value. Need to ask scientists where to get this value from - I think they should set it on some file which we read via the config server
- JF rotations: Need an extra parameter when calling zebra setup: `ttl_input_for_detector_to_use=composite.zebra.mapping.outputs.TTL_JUNGFRAU`
- Need to review and merge https://github.com/DiamondLightSource/mx-bluesky/pull/1638. It contains some fixes and (hopefully) puts data into ispyb
- Need to change `RotationExperiment` for when numtracker is being used. In the common version, we have a `_detector_params_impl` function which does stuff with the storage directory parameter. We need to make sure we get the storage directory from bluesky doc "instrument_session" (provided by numtracker). We need to make sure we have this info before we try to use storage directory in this function
- `RotationScanComposite` has xbpm_feedback, but it's not currently being used in plan, nor does it exist in i24's device module
- After ophyd-async sets some JF signals (eg pedestal loops), the IOC appears to turn pedestal mode OFF after a short delay. This caused a race condition and meant pedestal mode was being turned off after we explicitly turn it on. Long term, need to investigate why this happens from the IOC. Short term, need to add an async sleep(0.3s) in JF prepare just before we turn on pedestal mode
- For `gui_run_jf_rotation_scan`: Needed to rename `det_distance_mm` to `detector_distance_mm` and give the function a `MsgGenerator` return type hint