Issue
When a classic single submission happens (and possibly also the newer react/api submission), it appears that the Task tree created when IA pipeline is initiated gets "disconnected" between detection/embedding steps and final identification Tasks.
This creates the scenario where if the user waits for match results based on this initial (detection) Task ID, they will never see the final results appear.
User experience
If the user waits on the encounter page before looking for match results, it appears that eventually the link for results will reflect the proper ident-based result Task id, and the user will get to the correct match results Task id, waiting for completion.
However, if the user jumps to results too soon, they end up getting directed to the detection Task id, which will never connect to the actual desired ident results.
Issue
When a classic single submission happens (and possibly also the newer react/api submission), it appears that the Task tree created when IA pipeline is initiated gets "disconnected" between detection/embedding steps and final identification Tasks.
This creates the scenario where if the user waits for match results based on this initial (detection) Task ID, they will never see the final results appear.
User experience
If the user waits on the encounter page before looking for match results, it appears that eventually the link for results will reflect the proper ident-based result Task id, and the user will get to the correct match results Task id, waiting for completion.
However, if the user jumps to results too soon, they end up getting directed to the detection Task id, which will never connect to the actual desired ident results.