You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Talking with Massif-1 raised the issue of how we are supposed to put things in TaskGroups ('DataCollectionGroups').
Apparently a lot of unnecessary TaskGroups are created and end up in ISPyB.
Taskgroups would seem to be used basically for organising display, both the UI queue display (in the Qt version) and for e.g. EXI.
Massif-1 seems to prefer a single task group per mounted sample, grouping mesh scan, centring, characterisation, acquisition and potentially multiple sweeps and multiple crystals in a single group.
Looking at usage, the Qt interface seems to create a TaskGroup for every Sample.
The web interface on the other hand seems to create a new TaskGroup for every major task (characterisation, data collection, workflow, XRF, ...) or other queue item.
In addition, the Gphl Workflow adds a TaskGroup for both (multisweep) DataCollection and (multisweep) Characterisation, AbstractXrayCentring adds a TaskGroup (I probably coded that one as well) and Charaterisation.handleDiffractionPlan adds a TaskGroup.
So, the question is: Should we not harmonize the use of TaskGroups between Web and Qt interfaces? How should the TaskGroup structure of the queue be? Should the additional TaskGroups used by the GPhL workflow (to group characterisation and acquisition sweeps generated by the workflow) be removed?
The text was updated successfully, but these errors were encountered:
Talking with Massif-1 raised the issue of how we are supposed to put things in TaskGroups ('DataCollectionGroups').
Apparently a lot of unnecessary TaskGroups are created and end up in ISPyB.
Taskgroups would seem to be used basically for organising display, both the UI queue display (in the Qt version) and for e.g. EXI.
Massif-1 seems to prefer a single task group per mounted sample, grouping mesh scan, centring, characterisation, acquisition and potentially multiple sweeps and multiple crystals in a single group.
Looking at usage, the Qt interface seems to create a TaskGroup for every Sample.
The web interface on the other hand seems to create a new TaskGroup for every major task (characterisation, data collection, workflow, XRF, ...) or other queue item.
In addition, the Gphl Workflow adds a TaskGroup for both (multisweep) DataCollection and (multisweep) Characterisation, AbstractXrayCentring adds a TaskGroup (I probably coded that one as well) and Charaterisation.handleDiffractionPlan adds a TaskGroup.
So, the question is: Should we not harmonize the use of TaskGroups between Web and Qt interfaces? How should the TaskGroup structure of the queue be? Should the additional TaskGroups used by the GPhL workflow (to group characterisation and acquisition sweeps generated by the workflow) be removed?
The text was updated successfully, but these errors were encountered: