Deploy and test v9.2.0

[DominicOram]

We want the latest version of Hyperion to work

[DominicOram]

Issues from testing 30/04/24

• Had to pip install main of blueapi due to it depending on dodal<1.21. I think we could have pinned to https://github.com/DiamondLightSource/blueapi/releases/tag/0.4.2-a1 instead but the release process didn't. Follow up in 819 add dodal and nexgen versions to release notes #1305 (comment)
• Had to run (fixed in Dependency issue with running hyperion entry point #1346):

    pip install "starlette<0.28.0,>=0.27.0"
    pip install stomp.py==8.1.0

• Issues with GLIB like Modify the deploy script to run on the control machine so that the correct versions of dependencies are installed #1103, despite being on pyzmq<25, could be a Python 3.11 thing? Fixed by doing the install on the control machine as specified in the issue - follow up on issue.
• Various hacking on the GDA side to get the new params to work
• Snapshots should be stored next to the data, not just in the visit directory, this folder also needs to be created if it doesn't exist
• Tolerance on the aperture/scatterguard is way too low dodal.devices.aperturescatterguard.InvalidApertureMove: ApertureScatterguard safe move is not yet defined for positions outside of LARGE, MEDIUM, SMALL, ROBOT_LOAD. Current aperture z (15.8008), outside of tolerance (0.00030000000000000003) from target (15.8). https://graylog.diamond.ac.uk/messages/graylog-daq-mx-hyperion_7/580201b0-06f7-11ef-80bd-f6e3d090082b - increased tolerance to 10 MRES steps. Follow up in Use tolerances in beamline parameters file for aperture scatterguard dodal#494
• Nexus files are incorrect frame size:

{'beamline': 'i03',
 'dc_comments': 'Hyperion: Xray centring - Diffraction grid scan of 30 by 22 '
                'images in 20.0 um by 20.0 um steps. Top left (px): [477,285], '
                'bottom right (px): [686,438].',
 'file': '/dls/i03/data/2024/cm37235-2/xraycentring/TestInsulin/ins_9/ins_9_52_master.h5',
 'ispyb_expected_images': '660',
 'ispyb_wavelength': '0.976292',
 'reason': 'HDF5 file '
           '/dls/i03/data/2024/cm37235-2/xraycentring/TestInsulin/ins_9/ins_9_52_master.h5 '
           'contains invalid pixel_mask: pixel_mask inconsistent with '
           'data_size (pixel_mask.shape=(2162, 2068) data_size=(4362, 4148))'}

• fixed by setting ROI correctly in GDA params and in Hyperion params

• Issue with rotation params not being sent from GDA correctly. Stopped for the day

[DominicOram]

Testing 02/05/2024

• Update GDA to use the new rotation params
• rotation plan was emitting the new params in the metadata, moved to old_params
• https://ispyb.diamond.ac.uk/dc/visit/cm37235-2/id/13771896 has no resolution info - passed resolution in from GDA and that worked, we should make an issue to actually calculate it in the rotation plan though - follow up in Calculate resolution if not passed in #1359
• Otherwise all looks fine

[DominicOram]

Also saw an issue where screen + collect wasn't working. This was due to experiment type not being passed in from GDA -> Hyperion correctly - follow up in #1360

[DominicOram]

Also couldn't get robot loads to work, follow up on #1348