Identification bug

In which Wildbook did the issue occur?
What operating system were you using? (eg. MacOS 10.15.3)
What web browser were you using? (eg. Chrome 79)
What is your role on the site? (admin, researcher, etc)

What happened?

Hi there !

Sorry to spam but I decided to ask for your help today because I cannot go anymore with so much troubles getting Flukebook to work.

I’ve put on Flukebook another of my bulk import (Flukebook | Login) three weeks ago (on the 16/11/2023) and it still has some difficulties to show me the results of the identification process. I already relaunched it a few times trying different locations (all_locations or Caribbean Sea) but it doesn’t seem to work. Actually, the last identification job on this bulk import seems to be done but I cannot see the pictures when I click on each encounter page, I just see a white background with the green dotted square from detection but no pictures displayed…

Moreover, another of my problem is the fact that again, it seems that Flukebook is trying to match only within my own list of individuals and not others. Indeed, I rerun the identification process on this bulk import (Flukebook | Login) and it’s still displaying results against 1600 pictures instead of 20000ish so I don’t understand.

Thanks in advance for your time !
Bests, Louise :smile:

What did you expect to happen?
1-results of the identification process, with pictures
2-matches against more individuals

What are some steps we could take to reproduce the issue?
I’ve joined you by email the two bulk import I’m talking about in this forum.
If this is a bulk import report, send the spreadsheet to with the email subject line matching your bug report

Hi @lsimon

Thanks for sending in your spreadsheet! I think this may be related to a technical issue on our side and not the import itself. I hope to have more information on this later today.

I re-ran both of these through ID and everything’s processing normally so far. I know you already mentioned re-running ID on your own, but I was concerned about the missing match results, so I’m going to monitor until these are complete. I’ll follow up tomorrow as it’s the end of the workday here.

I need to wait until the ID results are finished to dive in further on the number of match candidates. Out of curiosity, what is the 20,000 number you referenced based on? The number of whales in s specific location ID?

Ok, fingers crossed then ! I have issues with my internet connection today so I won’t be able to double check with you until tomorrow.

It’s just the number of candidates I had on previous bulk imports when the program has run properly so I assumed it was all the sperm whales pics on Flukebook but I don’t really know …

No problem. One of the imports is already complete and the match results are loading correctly and the second import should be done with ID soon. I’ll follow up here when I have any updates.

Hi @Anastasia,

I’ve been taking a look today first thing in the morning and it seems that the ID has been working but only against my own pics although I presume you have been running it “all locations” right ?

I… don’t believe I did. :sweat:

I’m so sorry about that. I’m re-running ID for all locations on these.

Don’t worry, it’s already super kind of you to help me understand why it isn’t working ! :smile:
I will wait for those results and keep you update if it’s still matching against my own individuals (hopefully not :crossed_fingers: )

1 Like

Hi @Anastasia,

It seems that those two bulk import you resent to identification last week on the 2023/07/12 are still not showing any results. And for the tasks which have finished the ID, they are still matching against the wrong number of pics (even if it was rerun all-locations).

I was wondering if maybe the fact that I’m putting “Caribbean Sea” in the location-ID column of my datasheet can cause Flukebook to bug more than if I had put nothing (delete this column) :thinking:

This isn’t related to anything you’ve done! I’ve been working with my teammates on this and we think there was an issue with WBIA.

We’re testing a fix with this import and if WBIA behaves, we’ll apply it to the other import, too. Flukebook | Login

Flukebook | Login is complete now. I’m checking match results and they’re loading, but I’m seeing similar number for “against 1919 candidates” in the results even though they were run against all locations. That’s way too few. My teammates and I are still looking into why there are so few match candidates.

1 Like

Hi @Anastasia !

First, thanks a lot for your help !

It’s not only those two that are not displaying the right amount of candidates but all the bulk import that I’ve “newly” imported (within the last month). Here are the link to each of them :

I found really weird that it isn’t working because when I’ve started to put all my new bulk import on Flukebook and run the identification all-locations on the first one (Flukebook | Login), it was working properly (matching against the 26000ish individuals, on the 21th of November 2023). Here is an example to an encounter page results displaying the right number of candidates to match against:

Thanks again for your work,
Louise :smile:

1 Like

This is helpful; thank you! I may not have an update on this for you until tomorrow but I wanted to let you know we’re still working on it.

1 Like

Checking in to let you know this is still being worked on. Thanks for your patience.

1 Like

Thanks for hanging in there! It looks like there was a discrepancy in the location ID code for Caribbean Sea. We’ve corrected it and then I re-ran this bulk import to test whether more match candidates were available and saw 22,000 in the one I checked.

So now I’m re-running this import as well as these:

against all locations, so when they’re complete, you should see a bigger number of candidates they’ve matched against.

1 Like

Hi @Anastasia !

Wonderful, thanks a lot for your work and time :smile:
I will take a look at the results next week !

Have a great week-end :blush:

1 Like

So far, it’s looking good!

If you find any other imports that don’t have a big match candidate pool, feel free to re-send those through ID so you can see the correct number of results.

1 Like