331 fix and improve tests

[d-perl]

Fixes #331
Improves the test infrastructure for dodal:

• adds fixtures to mock config file paths so that devices relying on them can be created in CI
• runs device instantiation tests on all beamlines and simulated versions thereof

Instructions to reviewer on how to test:

1. Run new tests

Checks for reviewer

• Would the PR title make sense to a scientist on a set of release notes
• If a new device has been added does it follow the standards

[DiamondJoseph]

Leaving as comment to not block anything, but I have some reservations.
We're not making use of configuration files (at least, currently) and the testing rightly makes that optional- but it still runs the tests twice for each non-mx beamline when we have no conditional devices, but if we begin adding optional beamlines which rely on things other than "am I the live beamline or not", e.g. devices which are sometimes removed from the beamline, devices which might be used in one profile but not others, we'll be unable to test that without completely rewriting these tests again.

[DiamondJoseph]

Is all of this required for the tests?
BeamLine BL03S?

[d-perl]

I could chop some of it out, good point

[DiamondJoseph]

Not really extensible for @skip_device using other conditions that the value of beamline.BL.

Maybe testing the configurations when @skip_device is !x when it is default x should be in test_beamline, and this only test only the default case?

[d-perl]

I'm not sure I totally understood the second part of this comment but I agree with the first - instead I included skipped devices in these tests

[d-perl]

Thanks @DiamondJoseph, I basically agree with your concerns.

I updated the tests to only run once for each beamline, but create skipped devices instead.

Unfortunately we do use data from these config files in instantiating several devices, so it was only because they were being skipped that tests could run at all. Even more annoyingly, because the default files are on the DLS filesystem they might only break in CI.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (7c73805) 89.97% compared to head (0029972) 92.89%.
Report is 13 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #333      +/-   ##
==========================================
+ Coverage   89.97%   92.89%   +2.92%     
==========================================
  Files          83       83              
  Lines        3171     3182      +11     
==========================================
+ Hits         2853     2956     +103     
+ Misses        318      226      -92     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[noemifrisina]

Thanks

[DiamondJoseph]

@dperl-dls Still getting failing tests #332 rebased on top of main

I've seen it with p45 and i04, so there's still some state not being cleaned up properly.

=========================== short test summary info ============================
FAILED tests/beamlines/unit_tests/test_i24.py::test_device_creation - AssertionError: assert 'BL04I-MO-VGON-01:' == 'BL24I-MO-VGON-01:'
  
  - BL24I-MO-VGON-01:
  ?   ^
  + BL04I-MO-VGON-01:
  ?   ^
==== 1 failed, 378 passed, 5 skipped, 29 deselected, 64 warnings in 14.22s =====