Issues from mx-bluesky testing 2024-12-04

[rtuck99]

python ./utility_scripts/deploy/deploy_mx_bluesky.py i03
  File "/scratch/ws/mx-bluesky/.venv/lib/python3.11/site-packages/git/remote.py", line 895, in _get_fetch_info_from_stderr
    proc.wait(stderr=stderr_text)
  File "/scratch/ws/mx-bluesky/.venv/lib/python3.11/site-packages/git/cmd.py", line 834, in wait
    raise GitCommandError(remove_password_if_present(self.args), status, errstr)
git.exc.GitCommandError: Cmd('git') failed due to: exit code(1)
  cmdline: git fetch -v -- origin refs/tags/*:refs/tags/*

occurs when have a tag that differs from those in origin

• deploy script should print something it took ages with the slow fs, see Deploy script no longer gives console output when running dls_dev_env.sh #242 - we should fix the script as k8s is still some way off from production. stdout, stderr redirection fixed on beamline/local dev but also (possibly related) issue with borked pip install and script not exiting cleanly Beamline fixes from testing 2024-12-04 #700
• deploy_hyperion_tok8s.sh ensure_version_py() doesn't work properly if a non-k8s installation has already been performed as .venv already exists but without setuptools_scm - we should check for this and install it Beamline fixes from testing 2024-12-04 #700
• The warning that we are installing the wrong version to k8s isn't very prominent Beamline fixes from testing 2024-12-04 #700
• deploy_hyperion_to_k8s.sh references the wrong folders due to various hyperion/mx-bluesky renaming issues Beamline fixes from testing 2024-12-04 #700
• Various confusions in k8s deployment script due to version with or without v prefix, should standardise on supplying without and then adding it where necessary 700 fix deployment issues #701
• The deployed container image no longer starts, with log message:

usage: mx-bluesky [-h] [-v]
mx-bluesky: error: unrecognized arguments: --external-callbacks

#701

• 'HyperionLoadCentreCollect' object has no attribute 'context' - fixed on beamline
• rotation nexus files are reported in ispyb from the wrong (i.e. previous) rotation
• GPU processing not enabled (/dev/shm not set up) but is waiting for gpu processing
• Changing the omega_start value doesn't appear to generate the desired results as nexus files now look wrong with the change, needs more investigation with looking at actual smargon motor values from test data to figure out what changes are actually needed. Rotation start omega uses the wrong sign #247

[rtuck99]

Testing results from 2024-12-05:

• Ran into issue where gridscan fails with

File "/dls_sw/i03/software/bluesky/mx-bluesky_v1.4.1/dodal/src/dodal/devices/fast_grid_scan.py", line 107, in grid_position_to_motor_position
    raise IndexError(f"{grid_position} is outside the bounds of the grid")

https://ispyb.diamond.ac.uk/dc/visit/cm37235-5/dcg/13537858
Suggest that this be investigated along with #221 as to whether we should really be subtracting this value

• hyperion does not move beamstop into position, gda changes that introduced call to context.aperture_scatterguard_scintillator.moveToRobotLoad() are insufficient and don't seem to actually move the beamstop
  created Hyperion should check the beamstop is in position before collection #698
• Chi was not reset from previous position of 30 before doing a gridscan, if chi is zero then robot_load_then_xray_centre thinks nothing needs to be done even if current chi is non-zero move_phi_chi_omega does not move if angles are 0 Beamline fixes from testing 2024-12-04 #700
• Screen and collect is screening, but not collecting, possibly request never gets as far as hyperion? Suspect this is because self.sampleProvider.setSample(sample) in robot_load_and_centre/hyperion.py is never called when using load_centre_collect - the sampleProvider sample is then ultimately used by UDCRunner to fetch the next instruction but this is always null because it is never set. Screen and collect screens but subsequent collections are not made #703
• Zocalo sometimes returns hits with low counts on non-diffracting samples preventing a skip - fix applied on beamline to skip hits with total count < 3 Beamline fixes from testing 2024-12-04 #700