Can you describe what the issue is you’re experiencing?
The IDs I entered are all in Encounter.locationID: Gulf of Alaska. I want the algorithm to search for matches in the eastern Pacific Ocean.
Can you provide steps on how to reproduce what you’re experiencing?
I uploaded all the photos and excel successfully. Waited for detection. When I “send to identification” I selected Pacific Ocean and highlight all the subregions and subsubregions and pressed send. I selected the results and the algorithm only detected matches in Gulf of Alaska. When I attempt this manually, I can get much larger search area. Should I change the Encounter.locationID to Pacific Ocean?
I’ll need more time to look into this. Selecting multiple location IDs from the bulk import table should check for matches in all of the selected regions. I spot-checked a few of your matches and can see the matching task ID was set to the following locations despite the match page saying that it only checked against 200 candidates:
Hi Anastasia,
Any progress in understanding this issue? I am very eager to see if we have any matches with other whales in the Eastern Pacific!
Thanks!
Thank you for the update. I hope it gets resolved. Doing matches one by one is not possible. The matching never seems to complete and return any result. It stays in processing mode. I don’t know if that is a bug or the system being overloaded.
Kiku
Hi Anastasia, Thank you for letting me know. Is there a workaround, even it it means that I have to do each one individually? I’m trying to get some results this week. Is there any other way to do this?
Anastasia, Did I mention that I am entered in as Lauren Wild or lawild on this job? I am helping her perform this upload. I am very sorry I forgot to mention it!! It is under her username that I am having the trouble but I think I had this happen under my noaa.swfsc account as well.
I think I have solved this. While we were looking for technical eplanations as to why the other areas were not being considered…it turns out our only Eastern Pacific sperm whale sightings are from the Gulf of Alaska. The other locations don’t have any data submitters (yet). Most of our sperm whale data from elsewhere thus far is from the Gulf of Mexico, Caribbean, and Atlantic and would not have been included.
I see. Thank you. Then there is nothing wrong with the search and I am searching against my own entries.
So, could you please confirm:
When I do a bulk search on Eastern Pacific, all the subregions of the Eastern Pacific will be included or do I still have to highlight all the subregions?
There are thousands of sperm whales registered by the Dalhousie/Whitehead Lab by Ana Eguiguren. Are these in the Pacific Ocean? Why are these not included in the Eastern Pacific? What about the Hawaiian Islands?
When I conduct a manual search on a single encounter, a list of regions appears. There is PacificOcean, and Pacific Ocean, and I noticed that I get different results with each. These do not correspond with the list of regions available for the bulk search. Why are they not using the same regions and structure? For ex. when I select PacificOcean it returns only images that have been tagged as specifically “PacificOcean” not or “Pacific Ocean”. This seems to me like a user generated field.
Thank you for clarifying my confusion. I have to be confident where my searches are searching!
The sub-regions have to be manually selected in order to be included in the search.
Those encounters are in the Galapagos Islands and Pacific Ocean location IDs. “Eastern Pacific” is a sub-region of Pacific Ocean, so one option could be to select Pacific Ocean and all of its sub-regions for broader match candidates.
When users enter an unsupported location ID in their bulk imports, Flukebook doesn’t currently validate that field, so it accepts whatever is entered as the location ID. This will be addressed in the upcoming 10.8.0 Wildbook update so users will be prompted to correct errors with the location ID before they can commit the import. In the meantime, we’re gradually getting through location ID cleanups in Wildbooks to correct issues like the one you described.
