Add new :DATASET pv to capture records and pass the name to the client

[evalott100]

Closes #117

The ioc side of PandABlocks/PandABlocks-client#91

[codecov]

Codecov Report

Attention: Patch coverage is 88.29787% with 11 lines in your changes missing coverage. Please review.

Project coverage is 90.94%. Comparing base (8e25d64) to head (b5193ff).

Files	Patch %	Lines
src/pandablocks_ioc/ioc.py	70.00%	6 Missing ⚠️
src/pandablocks_ioc/_hdf_ioc.py	89.58%	3 Missing and 2 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #118      +/-   ##
==========================================
- Coverage   90.97%   90.94%   -0.04%     
==========================================
  Files           8        8              
  Lines        1285     1358      +73     
  Branches      206      213       +7     
==========================================
+ Hits         1169     1235      +66     
- Misses         78       83       +5     
- Partials       38       40       +2     

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

[coretl]

This overrides the on_update with something that updates the cache, but we still need the old behaviour of actually sending the capture value to the PandA. Maybe

PandABlocks-ioc/src/pandablocks_ioc/ioc.py

Lines 664 to 676 in 8e25d64

	if (
	"on_update" not in kwargs
	and "on_update_name" not in kwargs
	and record_creation_func in OUT_RECORD_FUNCTIONS
	):
	record_updater = _RecordUpdater(
	record_info,
	self._record_prefix,
	self._client,
	self._all_values_dict,
	labels if labels else None,
	)
	extra_kwargs["on_update"] = record_updater.update

needs updating so that it calls the user supplied on_update and sends the value to PandA? I haven't checked where else we call _create_record_info though...

[AlexanderWells-diamond]

Inside _create_record_info there is a check for whether on_update has been passed in as a kwarg. Looking at the diff viewer there definitely is a on_update currently checked in for the capture_record_name record.

[evalott100]

I think this should be fine, _create_record_info is only in ioc.py anyway

[coretl]

Inside _create_record_info there is a check for whether on_update has been passed in as a kwarg. Looking at the diff viewer there definitely is a on_update currently checked in for the capture_record_name record.

I'm confused, I can't see any mention of on_update in main:

PandABlocks-ioc/src/pandablocks_ioc/ioc.py

Lines 865 to 877 in 8e25d64

	capture_record_name = EpicsName(record_name + ":CAPTURE")
	labels, capture_index = self._process_labels(
	field_info.capture_labels, values[capture_record_name]
	)
	record_dict[capture_record_name] = self._create_record_info(
	capture_record_name,
	"Capture options",
	builder.mbbOut,
	int,
	PviGroup.CAPTURE,
	labels=labels,
	initial_value=capture_index,
	)

so that means it will prod the record_updater on caput which I assume pushes a value to PandA?

In the change here then a different on_update is supplied so no record_updater ever is made so no value is pushed to the PandA?

[evalott100]

I did a quick fix by creating the record updater in _make_pos_out _make_ext_out - the alternatives would leave me with annoying errors.

I made a new issue for solving the problem more generally #121