Closed
Description
This is to be re-assessed after a bit more real-life testing and improvement of the new history (ref #803).
Reasons for moving to a separate DB column:
- We could better define its structure, on the line of
# Option 0
DatasetHistoryItemType = dict[Literal["worfklowtask", "status", "parallelization"]]
history: list[DatasetHistoryItemType]
# Option 1
class DatasetHistoryItem(Basemodel):
worfklowtask: WorkflowTaskRead # TBD
status: WorkflowTaskStatusType
parallelization: dict # TBD
- Better scope separation between
Dataset.meta
andDataset.history
:- We should avoid "leaking" the full history of a dataset to the task (which we are currently doing)
- We could differentiate the current priority rules which are in-place for
meta
in some endpoints but in principle only apply tohistory