Update p38 to use device_factory

[DiamondJoseph]

Updates p38 beamline definition to use device_factory

Instructions to reviewer on how to test:

1. Check p38 beamline can still connect with dodal connect test

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
• If changing the API for a pre-existing device, ensure that any beamlines using this device have updated their Bluesky plans accordingly
• Have the connection tests for the relevant beamline(s) been run via dodal connect ${BEAMLINE}

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.68%. Comparing base (ffdd9e8) to head (f74c791).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #921   +/-   ##
=======================================
  Coverage   97.68%   97.68%           
=======================================
  Files         160      160           
  Lines        6600     6617   +17     
=======================================
+ Hits         6447     6464   +17     
  Misses        153      153           

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

[stan-dot]

dodal connect p38 fails but so it does on main, so it's a net improvement

[stan-dot]

better than main even if does not connect

[DiamondJoseph]

Reveals #925 when trying to connect- connects to all non-sim devices

high_pressure_xray_cell: NotConnected:
    all_valves_control: NotConnected:
        fast_valve_control: NotConnected:
            5: NotConnected:
                open: TypeError: BL38P-EA-HPXC-01:V5:OPENSEQ with inferred datatype int cannot be coerced to ValveOpenSeqRequest
            6: NotConnected:
                open: TypeError: BL38P-EA-HPXC-01:V6:OPENSEQ with inferred datatype int cannot be coerced to ValveOpenSeqRequest
        valve_control: NotConnected:
            1: NotConnected:
                open: TypeError: BL38P-EA-HPXC-01:V1:OPENSEQ with inferred datatype int cannot be coerced to ValveOpenSeqRequest
            3: NotConnected:
                open: TypeError: BL38P-EA-HPXC-01:V3:OPENSEQ with inferred datatype int cannot be coerced to ValveOpenSeqRequest
    pressure_transducers: NotConnected:
        3: NotConnected:
            slow_beckhoff_voltage_readout: NotConnected: ca://BL38P-EA-ADC-03:CH1
ppump: NotConnected:
    direction: TypeError: Differing datatypes: BL38P-EA-PUMP-01:INFO:DIR has 0, BL38P-EA-PUMP-01:SET:DIR has 3
    state: TypeError: Differing choices: BL38P-EA-PUMP-01:INFO:RUN has ('STOPped', 'STARTed'), BL38P-EA-PUMP-01:SET:RUN has ('STOP', 'START')

@stan-dot @barnettwilliam can I leave this with you to resolve the ppump and hpxc?

d3, d12, i0 I am investigating, as they also do not connect

e: i0, it are both disconnected from the synoptic screen on p38: I can only assume the IOC has not been started up since power loss.

[stan-dot]

ppump
idk where it is tbh

[stan-dot]

fixing this now

high_pressure_xray_cell: NotConnected:
    all_valves_control: NotConnected:
        fast_valve_control: NotConnected:
            5: NotConnected:
                open: TypeError: BL38P-EA-HPXC-01:V5:OPENSEQ with inferred datatype int cannot be coerced to ValveOpenSeqRequest
            6: NotConnected:
                open: TypeError: BL38P-EA-HPXC-01:V6:OPENSEQ with inferred datatype int cannot be coerced to ValveOpenSeqRequest
        valve_control: NotConnected:
            1: NotConnected:
                open: TypeError: BL38P-EA-HPXC-01:V1:OPENSEQ with inferred datatype int cannot be coerced to ValveOpenSeqRequest
            3: NotConnected:
                open: TypeError: BL38P-EA-HPXC-01:V3:OPENSEQ with inferred datatype int cannot be coerced to ValveOpenSeqRequest
    pressure_transducers: NotConnected:
        3: NotConnected:
            slow_beckhoff_voltage_readout: NotConnected: ca://BL38P-EA-ADC-03:CH1

[stan-dot]

slits return not connected at the moment, even though they should be mocked

[stan-dot]

I presume that the mock logic is not working for the slits

[callumforrester]

@DiamondJoseph and @stan-dot can you recommend a machine to run the system tests from? I need a RHEL8 one that can access the P38 PVs

[callumforrester]

System tests are currently failing for me

logs.txt

[stan-dot]

@callumforrester please include such long logs as a file not as a comment string

[callumforrester]

@stan-dot my bad, edited

[callumforrester]

System tests continuing to fail:
dodal-connect-p38-log.txt

[stan-dot]

I am not sure how to track what is the actual status of the p38 beamline. maybe some things don't connect because they are physically disconnected

that was the case for i20-1

[stan-dot]

discussed the connection issues with @barnettwilliam , some pressure jump device issues waiting for #812 (?)

[stan-dot]

and slits_N where N is from 1 to 6 also are not connected

caget times out:
caget BL38P-MO-FSWT-01:FILTER-101:STATUS_RBV

[stan-dot]

@DiamondJoseph have there been any issues with the mock not working with device_factory? that's what I suspect is happening

[DiamondJoseph]

@DiamondJoseph have there been any issues with the mock not working with device_factory? that's what I suspect is happening

That does appear to be the case. I know blueapi respects mock=True, but I thought the dodal connect tests did too. Leave it to me, I'll look at it tomorrow.

[stan-dot]

but I thought the dodal connect tests did too

fantastic... this was one of the edge cases that I got stuck at when doing device_factory back in October before you did your implementation

I was trying to make it work there too, and now we're witnessing a miss

this is an issue for all the 'update-ixx' PRs

[stan-dot]

this is the relevant issue

#925

[barnettwilliam]

Testing in P38 and excluding known errors for the mock devices, the expected devices except the pressure cell (d3, d12, i0m, and ppump) are connecting.

The pressure cell errors are down to the types returned by the OPENSEQ pvs being int and they are enums in dodal so the pressure jump dodal device should be corrected.

high_pressure_xray_cell: NotConnected:
    all_valves_control: NotConnected:
        fast_valve_control: NotConnected:
            5: NotConnected:
                open: TypeError: BL38P-EA-HPXC-01:V5:OPENSEQ with inferred datatype int cannot be coerced to ValveOpenSeqRequest
            6: NotConnected:
                open: TypeError: BL38P-EA-HPXC-01:V6:OPENSEQ with inferred datatype int cannot be coerced to ValveOpenSeqRequest
        valve_control: NotConnected:
            1: NotConnected:
                open: TypeError: BL38P-EA-HPXC-01:V1:OPENSEQ with inferred datatype int cannot be coerced to ValveOpenSeqRequest
            3: NotConnected:
                open: TypeError: BL38P-EA-HPXC-01:V3:OPENSEQ with inferred datatype int cannot be coerced to ValveOpenSeqRequest

[barnettwilliam]

Created #1034 to fix the pressure jump cell type issue.

[stan-dot]

@barnettwilliam after the latest merge we've got a branch conflict, could you please take a look?

[barnettwilliam]

Yes will do

[barnettwilliam]

@stan-dot conflicts are now resolved

[stan-dot]

the non-connecting devices are those that fail due to the bug with device_factory interacting with dodal connect

related ticket (?)
#973

@callumforrester I'd say this is bug-for-bug compatible with the main branch and ready to merge

two months have passed