Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 94.97%. Comparing base (eece064) to head (18dbbeb).
Report is 3 commits behind head on main.

@@            Coverage Diff             @@
##             main     #799      +/-   ##
==========================================
- Coverage   94.99%   94.97%   -0.03%     
==========================================
  Files         115      115              
  Lines        4719     4698      -21     
==========================================
- Hits         4483     4462      -21     
  Misses        236      236              

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

@DominicOram at the moment I only see the Slits device used at i22 and p38, where I know the scientists and their ways of working and for now they don't write tests themselves. And the device is just in the devices/i22 folder and any beamline can use the single Slits device however they want.

With SixSlits the whole device will fail to connect and throw errors.

that is a valid point, might need a decorator 'allow fail'? I am not sure if the slits are used independently like what you're describing now at i22

If we go this way I don't think it should be called six_slits: we'd still be able to address individual slits with e.g. inject by doing inject = ("slits.1") (NB: Might not work for device vectors? Would this work with inject("slits[1]")?)

at the latest version it's a SlitsCollection. testing out ipython behavior of device factor shortly

@DominicOram what is a plan that you'd like to define in ipython and then run it that the inject behavior would be used?

@DominicOram what is a plan that you'd like to define in ipython and then run it that the inject behaviour would be used?

Something like:

def slit_scan(slit: Slits = inject("slits_1"), i0: TetrammDetector = inject("i0"):
    yield from bps.scan([i0], slit.x, 0, 100)

ok how do I get the beamline module to be referred to with inject?

I don't see this in tutorials

search also does not provide information on the use of dodal in ipython https://diamondlightsource.github.io/dodal/main/search.html?q=ipython

ok how do I get the beamline module to be referred to with inject?

It's a blueapi-ism, assuming that there is a device with that name in the blueapi context that will fetchable when the pydantic model is deserialised: regardless of what beamline module is loaded. It's not particularly well developed or documented, I'm hoping there's something in Pydantic 2 that we'll be able to leverage for the same effect but I will have to look after I get back.

@DominicOram on re-reading your initial comment I see that the granularity of skip_device decorator is decreased as a result of this PR, closing the PR then