When doing spot maps on Carcharias taurus in Sharkbook we have always had multiple results for both Modified Groth and i3s, but this shark has only produced 3 results for Modified Groth and no results for i3s. Can you please check?
It is weird - it works sometimes, and not others. Here is another example where no i3S results came up, and we only had a handful of modified groth results:
I found out this happens when the I3S jobs fails. The workaround for now is to delete the spot map and resubmit it to run both algorithms again. Let me know if that doesn’t resolve it for these two encounters.
In the past, we had to delete the encounter to start a new spot map. Has that bug been fixed now so that hopefully I just need to try and delete the spot map only rather than the whole encounter?
However it does appear to be happening more often than not now so perhaps there is something else going on that is making i3s fail?
The ability to delete a spot map from an encounter has been available for several years. I shared a walkthrough with screenshots of how to do it in your recent email thread with your international collaborators.
Can you be more specific with how much often it’s happening now? Additional example links will help us with troubleshooting.
Here is another example where only 1 result has come through for modified groth and no i3s results. I have this bug happening on a number of sharks - do you want me to share all the links?
For context, I3S would have to demonstrate a pretty significant failure rate to for us to divert resources to updating it as our team is pretty resource-constrained.
As long as redoing the spot map is getting the I3S results to display on the second run, this is the process that you’ll continue to follow.
With my earlier questions, I’m trying get a better understanding of how many of these jobs are failing compared to when you first reported it. For example, if this used to happen 1-2 times a week across multiple uploads vs now you see errors with every single upload, that’s quantifiable information we can use to determine what the appropriate next steps are.
I think we can safely say that the i3S issue is happening more than not. I have never seen this before, but this week most of the spot map reviews do not produce i3s results and only produce a handful of modified groth results.
I have also tested this fault by going into old encounters and the same issue occurs. I would recommend that you try adding a spot map for this encounter, and I am sure you will find the same issue. Sharkbook: Wildbook for Sharks
The problem for me is that unfortunately I need to combine all 3 algorithms for photo ID of this species - I cannot rely on MiewID alone. The biggest datasets I am trying to clean up are the adult sharks (e.g. at Julian Rocks) for which i3s is essential for my photo ID.
We have fixes that are waiting additional testing before deploying and also the deleted spot maps still appearing in search results. I hope to have a status update for you soon.