Add option to only retry non-tagged files in classify

[nielstron]

This is picking up #1137 again.

Works locally, no idea how to unit test this :)

[nielstron]

This is now working on my server after I hacked it into the local settings.

[marcelklehr]

Hey @nielstron

Really cool that you're taking this on! I'll look into this soon

[marcelklehr]

This approach works in so far as it does indeed restrict the classify command to files that haven't been tagged with the recognize tag yet. However: The recognize tag is only assigned by the imagenet, movinet and musicnn classifiers, not by the face recognition classifier. So, when e.g. the classify command fails or is stopped during the imagenet classification, the images it has tagged so far will not be put through face recognition at all, because they have the tag from the imagenet classification run.

[nielstron]

Thanks for the clarification! Any suggestion on how to fix this?

[nielstron]

I see two options

1. Determine seperate whether the face classification pipeline has been run for a file and put it into the queue if not
2. always put the file in the face classification queue (which I fear to be quite expensive)

For 1 it would be good to know if such a tag exists/how to obtain it.

[marcelklehr]

Yeah, I agree with your assessment, for 1. you can find if a face detection exists for a file using the following method: https://github.com/nextcloud/recognize/blob/main/lib/Db/FaceDetectionMapper.php#L82

[nielstron]

If face detection for a file did not find a face, is the fact that we processed the file still stored somewhere? Or would this re-run classification on all files that don't contain a face?

[marcelklehr]

Ah, good question. There's no marker for whether a file has been through face recognition. It might make sense to introduce a new magic tag for this.

[nielstron]

Okay. This patch now skips running the no-face-recognition classification for tagged files. I don't think I will have time to contribute more elaborate features in the near future.

[marcelklehr]

Cool, this works well enough! Thank you!

[github-actions]

Hello there,
Thank you so much for taking the time and effort to create a pull request to our Nextcloud project.

We hope that the review process is going smooth and is helpful for you. We want to ensure your pull request is reviewed to your satisfaction. If you have a moment, our community management team would very much appreciate your feedback on your experience with this PR review process.

Your feedback is valuable to us as we continuously strive to improve our community developer experience. Please take a moment to complete our short survey by clicking on the following link: https://cloud.nextcloud.com/apps/forms/s/i9Ago4EQRZ7TWxjfmeEpPkf6

Thank you for contributing to Nextcloud and we hope to hear from you soon!

(If you believe you should not receive this message, you can add yourself to the blocklist.)