Add FrameReceiver controllers, move FP controllers to odin_data.py

[jsouter]

Addressing #11 (see also odin-detector/odin-data#346), using the generic OdinDataAdapter as the basis for DAQController and DAQAdapterController which are inherited by FrameProcessor(/Receiver)Controller and FrameProcessor(/Receiver)AdapterController.

edit: WIP, can break down further and add a FrameReceiverDecoderController

[codecov]

Codecov Report

Attention: Patch coverage is 93.87755% with 3 lines in your changes missing coverage. Please review.

Project coverage is 82.30%. Comparing base (d00e26e) to head (29e7c3a).
Report is 2 commits behind head on main.

Files with missing lines	Patch %	Lines
src/odin_fastcs/odin_controller.py	33.33%	2 Missing ⚠️
src/odin_fastcs/odin_data.py	97.72%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #33      +/-   ##
==========================================
+ Coverage   81.38%   82.30%   +0.91%     
==========================================
  Files          10        9       -1     
  Lines         360      390      +30     
==========================================
+ Hits          293      321      +28     
- Misses         67       69       +2     

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

[GDYendell]

This looks like the right sort of thing. I think I am OK with having fp and fr in one file as they are all very small classes.

Could you change DAQ -> OdinData in the class names and move FrameProcessorAdapterController above FrameProcessorController so that they are in hierarchical order?

[jsouter]

This looks like the right sort of thing. I think I am OK with having fp and fr in one file as they are all very small classes.

Could you change DAQ -> OdinData in the class names and move FrameProcessorAdapterController above FrameProcessorController so that they are in hierarchical order?

Have done that, had to overload init as the forward reference needed for _subcontroller_cls needed to be moved into function scope

[jsouter]

Added the decoder sub controller, can add in a few tests if needed.

[GDYendell]

had to overload init as the forward reference needed for _subcontroller_cls needed to be moved into function scope

Ah I didn't think about that. In that case it is not worth it. The AdapterControllers can go after.

[GDYendell]

Do we know where this decoder name parameter is coming from? That doesn't appear in the API, so I am not sure why odin-fastcs is creating that.

ERROR:root:Update loop failed for api/0.1/fr/1/status/decoder/name:
name not found in response:
{'response': 'OdinDatatAdapter GET error: Invalid path: 1/status/decoder/name'}

[jsouter]

Do we know where this decoder name parameter is coming from? That doesn't appear in the API, so I am not sure why odin-fastcs is creating that.

ERROR:root:Update loop failed for api/0.1/fr/1/status/decoder/name:
name not found in response:
{'response': 'OdinDatatAdapter GET error: Invalid path: 1/status/decoder/name'}

Will look into it, I know that config/hdf/file/name had to be changed to "prefix" because of some interaction between fastcs and odin that makes "name" an invalid name for a parameter.

edit: EigerFrameDecoder::get_status() (and the dummy decoders) add "name" to the status_msg, will test if either changing that to another key fixes it or if it needs to be formally added to the API somewhere.

[jsouter]

Okay yeah, changing the key to "dummy_name" as a quick test in DummyUDPFrameDecoder makes the signal work properly. I suppose we could either remove the parameter from odin, rename it, or add some logic in odin-fastcs to do special handling of attributes called "name"

[GDYendell]

Ah so name is visible when requesting metadata, but not otherwise. That is why it creates the parameter, but then can't read it from the server.

I suggest we rename to type or class, but there is no way to enforce that across all decoders. I think the odin server should error on start up, rather than on requesting the parameter.

[GDYendell]

It would be good to add

• unit tests that create the controllers, pass in some dummy OdinParameters and make sure _process_parameters does the right thing, like test_fp_process_parameters and test_fp_create_plugin_sub_controllers.

• an introspection test that takes a dumped server response, parses it and makes sure the parameters are as expected. See dump_server_response.py and test_introspection.py.

[GDYendell]

Could you squash this into one commit and force push, and mark as ready for review if you are happy with it?

[GDYendell]

Thanks @jsouter !

[GDYendell]

This didn't break things. main is broken because the CI happened to pick up the latest pydantic (2.9.0) on merge.

See pydantic/pydantic#10345 and epics-containers/pvi#137.