Issue trying to run linkam_plan for i22: prepare about ophyd-async HOT 2 CLOSED

I've now looked into this, and discovered the source of the problem.

When we ran the I22 experiment in november, we discovered that if we tried to set num=0 into all detectors, that the detectors would understand they need to take images forever. This turned out to not be the case for the Pilatus, so we modified the definition of a Pilatus (https://github.com/bluesky/ophyd-async/blob/main/src/ophyd_async/epics/areadetector/controllers/pilatus_controller.py) so that it set the number of triggers to 999,999.

The behaviour then was, we create the Flyer and prepare it, which means we arm all the detector controllers:

By default, in the above code snippet, num=0 would be passed to all the detector controllers for this process. The tetramm in the directory_provider branch of dodal would raise an Exception if it ever got something that wasn't num=0: https://github.com/DiamondLightSource/dodal/blob/directory_provider/src/dodal/devices/tetramm.py#L130-L131

Now, we have changed the logic such that the number of frames comes from the plan, and such that num=0 is no longer passed. This was done in #96, however we failed to consider that the Tetramm would now raise this exception, as now a num was being passed to it.

A PR has been raised for the dodal repository, where this issue really ought to belong. But at the time of testing this out on Tuesday, it seemed like an ophyd_async issue.

Perhaps we ought to reconsider, at a higher level, if the PR I've proposed is the right thing to do (i.e. get the tetramm to ignore num or if we need to rethink how controllers work. Maybe we need a new subclass of controllers, e.g. ContinuousControllers which don't care about number of frames and DiscreteControllers which do? Also, we should probably remove this condition: https://github.com/bluesky/ophyd-async/blob/main/src/ophyd_async/epics/areadetector/controllers/pilatus_controller.py#L41.

from ophyd-async.

As an aside, we also need to consider the changes needed to the linkam plan in the i22-bluesky repository, now that the number of frames actually come from the plan (and is not assumed to be 0). I'm hoping that the changes are minimal but this needs further discussion.

from ophyd-async.