Identification phase blocked

Bug Reports
,
1

Hello Anastasia,
I hope you are doing well.

Sorry to bother you, but we are currently analyzing our Steno bredanensis photos on Flukebook and are encountering identification issues.

For most of our bulk imports, the identification phase does not complete, and we are therefore unable to relaunch it. We also tried running the identification process image by image, but this did not change anything. The issue has been persisting for several days.

Would it be possible to force the identification process for the following bulk imports, please?

Thank you very much in advance for your help.

Best regards,

2

Hi @rebeca

Thanks for these examples! It looks like the PIEv2 algorithm is failing to return matches. I’ll need a little more time to dig into why it’s misbehaving before restarting these matches. Thanks for your patience.

3

Quick update: there’s a misconfiguration issue with PIEv2 and Steno bredanensis. This should be resolved soon.

4

Thank you Anastasia ^^

5

We have a fix ready. We are just waiting for the current queue of identification jobs to complete so we can test the fix on your imports.

6

Thank you very much ! , but aren’t the identifications currently blocked?

7

We’ve deployed the fix and re-sent the missing PIEv2 jobs from your imports back through identification. They are currently in process. Thanks for your patience.

8

Quick followup: I verified identification has completed for each import. I did notice that this one shows ID is only 99% complete, but I looked through all 38 pages of encounters and verified that they’re all marked as complete: https://www.flukebook.org/react/bulk-import-task?id=9c4d2f99-b8d5-40ba-8796-18c1ca550dad

This is a known issue we’re working on, but I wanted to give you a heads up that these should all be ready for your review. Happy holidays!

9

Thank you so much Anastasia !

1 Like