On match results page, in the yellow bar at the top, the ID of the animal flickers from CH00005 to CH0006, eventually remaining as the latter; although in repeating the same steps for the same encounter, I have also seen it remove the ID from the yellow bar altogether. Wildbook for Carnivores
The target image also changes to what appears to be one from the updated ID in the yellow bar - CH00006 (when it stays as CH00006 in the yellow bar). So the target image and the name in the yellow bar are the same. But it’s not the individual I selected match results for (in this case, CH00005).
What did you expect to happen?
That the annotation I selected in the encounter record to view match results for, is the one I see as the target image in the match results and also the one named in the yellow bar at the top.
What are some steps we could take to reproduce the issue? Per above
I am seeing the behavior as designed here. When you submit/import an Encounter, by design all found annotations from photos are combined into a single, consolidated matching task, even if we split one of the Annotations off later due to multiple detected annotations in an Encounter. This consolidated task is then attached to all related annotations (e.g., click "match results:). This gives you a single view into everything that was found and what each potentially matched. However, if you click “Start match” on a single Annotation after that, you will get the expected single-annotation match results attached to that annotation thereafter. Please let me know if subsequent matching has more than one annotation in the results.
Splitting the ID tasks with the annotations and Encounters is a good idea…it’s a clean line too…but it also means that the full context of matching an new Encounter will be lost (or at least split into multiple locations). It will also require a new UI to track the split from one ID job into two ID jobs (e.g., how does the user know that they need there are now two ID jobs they need to review?).
I believe our Next Generation of Wildbook will be handling more complex use cases like this.
I haven’t had any additional reports of this from Jeanette. I just viewed a couple of example encounters, including the original one that was reported above and the issue is not recurring. I’d say this can be marked as resolved and if we run into it again, I’ll put in a new report.
