Sonograms of Eastern Wood-Pewee not appearing on Species Stats tab

[Dot-Dash]

On the Species Stats tab, if I select Eastern Wood-Pewee in the dropdown at the top "Which bird would you like to explore for the dates...", on the right-hand side, instead of a sonogram for one of the recordings, I get "RECORDING NOT AVAILABLE :(" .

This consistently happens when selecting any of the available recordings. For example,
Eastern_Wood-Pewee-76-2025-08-27-birdnet-18:13:53.mp3. There are at least 20 recordings in that list. As far as I can tell, all of them display the error message when selected.

This does not happen when selecting any of the other 40-odd species that have been recorded. For those species, the sonogram(s) are shown as expected.

I know this seems really random. I tried restarting core services as suggested in the troubleshooting guide. There was no change.

To Reproduce
Steps to reproduce the behavior:

1. Go to Species Stats
2. From the bird dropdown, select Eastern Wood-Pewee.
3. See error message RECORDING NOT AVAILABLE :(

Expected behavior
Sonogram is displayed.

Screenshots

Additional context
System installed about 10 days ago. I have not Excluded or Whitelisted any Species. Note that the sonogram of the Eastern Wood-Pewee does appear as expected in the Best Recordings tab, and the Recordings tab.

Your build
Raspberry Pi 4 Model B (4GB)
Kernel Version 6.12.34 +rpt-rpi-v8 (SMP) aarch64
Distro Name Debian GNU/Linux 12 (bookworm)

Code or log snippets
none

Paste your 
log output 
or 
code here

[alexbelgium]

Hi, you should check if the files actually exists. Perhaps the disk cleaner removed them. Currently it removes files without keeping a minimum number per species. I have done a PO to propose this feature which could be reactivated if this is the issue you faced

[Dot-Dash]

So, yes, many of the files do not exist. A few do exist, and for those, the spectrogram appears normally! Apparently only those Eastern Wood Pewee recordings with confidence of about 93% or higher exist. Others are gone. That's a small percentage of the original recordings which is why I didn't stumble on them earlier. However, recordings of other species with lower confidence levels still exist.

My 256 GB storage is only about 3% used, so I wouldn't expect files to be deleted for space yet.

Does anyone know why this particuar species is filtered so heavily? This does not seem to be a very common species locally and is way down the list of common recordings.

If the files are purged, then they should no longer appear in the GUI, right?

Just for reference, the bird recordings are stored in the path: ~/BirdSongs/Extracted/By_Date/, then in folders named YYYY_MM_DD, then in folders with the species name.

Here are the Eastern Wood Peewe files I found for one particular day:

Eastern_Wood-Pewee-93-2025-08-24-birdnet-16:03:00.mp3
Eastern_Wood-Pewee-93-2025-08-24-birdnet-16:03:00.mp3.png
Eastern_Wood-Pewee-93-2025-08-24-birdnet-16:03:06.mp3
Eastern_Wood-Pewee-93-2025-08-24-birdnet-16:03:06.mp3.png
Eastern_Wood-Pewee-93-2025-08-24-birdnet-16:03:21.mp3
Eastern_Wood-Pewee-93-2025-08-24-birdnet-16:03:21.mp3.png
Eastern_Wood-Pewee-93-2025-08-24-birdnet-16:03:33.mp3
Eastern_Wood-Pewee-93-2025-08-24-birdnet-16:03:33.mp3.png
Eastern_Wood-Pewee-93-2025-08-24-birdnet-16:04:06.mp3
Eastern_Wood-Pewee-93-2025-08-24-birdnet-16:04:06.mp3.png
Eastern_Wood-Pewee-94-2025-08-24-birdnet-16:02:39.mp3
Eastern_Wood-Pewee-94-2025-08-24-birdnet-16:02:39.mp3.png
Eastern_Wood-Pewee-94-2025-08-24-birdnet-16:03:30.mp3
Eastern_Wood-Pewee-94-2025-08-24-birdnet-16:03:30.mp3.png
Eastern_Wood-Pewee-96-2025-08-24-birdnet-16:03:15.mp3
Eastern_Wood-Pewee-96-2025-08-24-birdnet-16:03:15.mp3.png

[alexbelgium]

What are you Disk Management settings in the advanced options ?

[Dot-Dash]

Purge is selected.
Purge threshold (Disk Used %) 92
Number of files to keep for each species: 10

[alexbelgium]

Then indeed it keeps 10 files per species, starting from the highest confidence, except for the files in the last 7 days which explain the 6 additional files

You should probably set it to something like 200 (which means around 700mo per species if you have an average mp3 of 3.5mo)

[Dot-Dash]

Okay, sounds reasonable. I did that. But I think the important question is why the GUI on the Species Stats page is listing files that no longer exist, and then giving an error message when they can't be found.

[alexbelgium]

Actually, checking if files exists would greatly slow down the code, I guess it was implemented this way on purpose

[Nachtzuster]

Okay, sounds reasonable. I did that. But I think the important question is why the GUI on the Species Stats page is listing files that no longer exist, and then giving an error message when they can't be found.

As far as I know, on the original fork it could happen that those files did not exist due to errors in the processing 'pipeline', that is probably why this code is making such a fuss about it.
now, missing files, very likely means they just got cleaned up, so a less dramatic message is probably in order 😄