Seal codex-
uploaded images from two dates in three bulk imports.
getting matches for images from one date (04/11) including other seal images uploaded not by me and my images from the same date but no matches at all for the second date images (07/11).
Hi @oliviasanders1 welcome!
That’s expected behavior. Codex works by showing you potential matches of the same species seen at the same location you selected, not just within your own uploads.
It’s a possibility that Hotspotter correctly determined that no match was found since it only displays results with a match score higher than 0.
Do you have a link to the Sightings in question so I can take a closer look?
Hi Thanks for your response!
I dont think its a coincidence as it seems for all of my individuals uploaded from my second date (07/11) have no matches, yet my individuals from (04/11) seem to have matched with something.
It seems like there is some sort of error occuring
I have uplaoded about 100 individuals for both dates
Here are some of seal ID’s for both dates :
07/11-
Codex ID : HaGrA-218
Codex ID : HaGrA-206
Codex ID : HaGrA-148
04/11- these are the ones that are displaying matches
Codex ID : HaGrA-004
Thanks for taking a closer look!
Sent that over.
Thankyou.
1 Like
Thanks for that!
I wanted to let you know I only found one Sighting for Nov 7 in Codex. I believe this set of Sightings was uploaded as Nov 6 despite the date in the spreadsheet. This can happen before you confirm the upload when the workflow allows you to make manual changes to your data before submission.
I see a few Sightings that were stuck in identification and I’ll need to work with my teammates to help get those moving again. I hope to have an update soon to verify matches are behaving.
Thanks for your patience! I’ll post any new updates in the Community thread so we don’t split the conversation.
Anastasia
Sorry for the churn; can you also send me the spreadsheet for your other import? It looks like duplicate Individual records were created for each seal between the two imports and I’m trying to figure out what caused it.
Hi,
Do you think that i may have to delete and reupload the Nov 7 date again?
No, don’t delete anything yet.
I know you’re just wondering about the match results, but the more I dig into this, the more I see other issues surfacing and I’m working out if and how they’re related to what you’re experiencing. I just wanted to be clear that I’m still working toward fixing that in addition to the other things I’m finding.
Can you also email us the spreadsheets for the other imports you worked on? I’m trying to determine if something in the metadata is contributing to the duplicate marked individuals.
I’m also waiting to hear back from a teammate regarding the date discrepancies between what you and I see.
That’s great Thankyou!
I did send over the other excel spreadsheets on email. I’ll resend incase you didn’t receive
1 Like
Thank you! I need to spend more time testing and I’ll follow up with you as soon as I have an update.
Thanks for your patience!
Some of the sightings had “Identification failed” error messages or were stuck in identification for days so we ended up fixing a bug in WBIA where heat maps errors were preventing identification from completing and restarting the server. That appears to have corrected the issues we saw. For example, here are a couple of the seals from the first import you sent that were previously stuck in identification and are now showing a list of match candidates:
- https://seals.wildme.org/match-results/ff0ea95c-4904-4bde-86cd-3c26717c76f9
- https://seals.wildme.org/match-results/7ed6f37d-4fd7-47cc-bdaa-ad5cd25f85bf
Regarding the differences in the dates on the Sightings, that’s a bug where it will adjust the date displayed based on the time zone of the user when a specific time for the Sighting isn’t entered. Right now, we’re only prioritizing the fixes that will let you see your match results.
Feel free to re-run identification on any Sighting where you previously saw no match candidates (find it in the three-dot menu on your Sighting page).
If it still says no match candidates were found, we are confident that those results are accurate now that WBIA is completing its ID jobs.
I’m still looking into why you have so many duplicate individual records. I’ll have to follow up on that next week when I’m back online.
I’m linking to the PR for the bug fixes uncovered during this research: Some bug fixes by naknomum · Pull Request #968 · WildMeOrg/houston · GitHub
Thankyou so much for taking the time to sort this for me!
Apart from a few indivudals from 04/11 the identification all seems to be working and I am getting matches.
1 Like
To follow up on the duplicate pages for your Individuals, did you create a manual Sighting report before bulk importing these? I ask because Codex doesn’t let users create new Individuals from a bulk import and all of the seals in your spreadsheets should have been new Individuals to Codex. I’m trying to figure out how it happened because I’ve not yet been able to reproduce it.
For example, Sealb1_1 has two individual pages even though it’s the same seal and both were uploaded by you:
- https://seals.wildme.org/individuals/b6d6c666-9cb2-44eb-abf8-e55f231d7470
- https://seals.wildme.org/individuals/43be919e-8656-4d95-a122-816042916540
What we should see is two Sightings under one Individual page, not one Sighting for two pages of the same individual.