Example conversion to the new device manager

[tpoliaw]

Most devices are a straightforward swap to the new decorator. The main things to
be aware of:

• Devices are no longer cached anywhere.
  a, b = synchrotron.build(), synchrotron.build() will result in two synchrotrons.
  Adding @cache might work if the function takes no other devices but given
  most devices aren't hashable it probably won't work otherwise.
• devices.build_all() will only build one of each device, so two devices
  depending on the same third device will be passed the same instance.
• If a device factory calls another device factory itself (rather than using a
  parameter), it will create a separate instance.

Other things specific to i03's module

• The weird empty method for the ophyd_v1 device (eiger). If there was a
  post_create method, that would be where it was. Given ophyd v1 support is not
  meant to be long-lived, this will do for now.
• The path_provider is no longer set as a global and needs to be passed in when
  the methods that require them are called. When using the device manager, this
  is automatic, eg panda.build() will automatically create the path provider
  via the fixture. BlueApi will pass its own path_provider which would be used
  instead.
• The undulator and undulator_dcm now take their configuration path as a
  required argument rather than using the global if not set. This will be taken
  from the fixture unless overridden. When the two devices are built at the same
  time (eg undulator_dcm.build() which builds undulator as a dependency)
  both will use the same configuration path.

The CLI has already been updated so this can be tested via

dodal connect i03 -n devices

with or without -s to mock things as before. The -n devices is currently
needed to determine what the device manager is called. If we standardise on a
convention for this, that can be removed.

[DominicOram]

I think this is generally ok. I have made DiamondLightSource/mx-bluesky#1459 which fixes most of the tests on the mx-bluesky side I believe.

[DominicOram]

Could you take a look at the tests in test_device_instantiation.py?

[codecov]

Codecov Report

❌ Patch coverage is 98.50746% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 99.10%. Comparing base (9ed11d4) to head (c360687).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
...plans/configure_arm_trigger_and_disarm_detector.py	50.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1684      +/-   ##
==========================================
- Coverage   99.13%   99.10%   -0.03%     
==========================================
  Files         279      279              
  Lines       10504    10510       +6     
==========================================
+ Hits        10413    10416       +3     
- Misses         91       94       +3     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[rtuck99]

I've updated the docs and made the test_device_instantiation.py cover i03 again, everything else looks fine, will need DiamondLightSource/mx-bluesky#1459 as well when merged

[DominicOram]

With the manager itself merged I will tidy this up now to make it ready for merging

Stale

[DominicOram]

We should remove these docstrings as they're lies