VMXm grid scans: Current status

[olliesilvester]

There is a lengthy stack of unmerged PR's with all the required for VMXm grid scans - the latestone is here. VMXm's live GDA has some hacky code to optionally run gridscans through BlueAPI. In /dls_sw/i02-1/software/daq_configuration/domain/domain.properties there is an option gda.gridscan.blueapi.enabled. If this is set to true then BlueAPI will be used for grid scans (after a GDA server restart). BlueAPI is hosted on the beamline cluster and has its configuration here.

The code is largely untested and unstable. There are a number of known issues and probably lots of unknown ones.

Next steps

1. Schedule testing time and run grid scans and try and everything working (incl grid scans). There will likely be lots of little hotfixes needed before anything works
2. Adjust the PR's linked above to address hotfixes
3. Get PR's reviewed + merged then redeploy and test. Repeat this until we have something stable.

[olliesilvester]

Known issues:

• VMXm GDA code to trigger BlueAPI needs tidying. The existing code isn't merged into gerrit yet, and some parameters are being hardcoded instead of correctly passed in
• Resolution calculation from bluesky seems dodgy - seems like we were sometimes getting negative resolutions
• check_topup_and_wait_if_necessary wasn't working in do_fgs.py for some unknown reason. Code got stuck on this, I had to comment it out in the live code for gridscans to work.
• gridscan ispyb callback for activity gated start doesn't work as is, since we need a data document to fill in some of the bits. Need to move this to an activity_gated_event so that the data document exists when this function is ran.
• Need to allow sample id to be None - it seemed to work okay after allowing this
• There are some parameters (eg beamsize) which GDA is currently passing to BlueAPI. We should add the vmxm beamsize device to dodal and let bluesky handle this. In general, we should try to reduce the parameters sent to bluesky and put in logic to let bluesky get these parameters
• Common but intermittent issue where Eiger gets stuck when arming during the step which logs "eiger arming: Setting eiger camera PVs...". I'm not sure about the cause, Eiger IOC logs looked okay. For the next round of testing, I suggest making set_cam_pvs eiger function waiting + logging after each PV set so then we can work out which specific PV it's actually getting stuck on. Arming this bit synchronously might actually fix the issue.

Also, GDA is currently taking snapshots and passing in the locations to Bluesky. Eventually bluesky should do the snapshots itself, since we already have plans for this.

Hotfixes I will try and do today or tomorrow:

• vmxm plan actually needs entry point + msggenerator type hint for blueapi
• dy and dx in mm need /1000, not *1000
• Need to pass in box size width and height to parameters
• Eiger settings are wrong. we need to be using EIGER2_X_9M_SIZE in DetectorParams and in the DETECTOR_SIZE_PER_BEAMLINE map for i02-1. Also need to make sure roi mode is always off
• Need to pass through omega_starts_deg
  EDIT: above hotfixes are done, but I haven't updated tests to match