Quick fix for sudden stops in document classification

[luckow]

In January, I noticed that automatic classification for documents had stopped on two of my Paperless instances. I didn't have time to fix it, so I resorted to manual interaction.
Now, I looked into the logs and found the following message: System status detected a possible problem while loading the classifier: Classifier file does not exist (yet). Re-training may be pending.
This specific error made it easy to search the internet for a solution.

In the official docs, I found information about Managing the Automatic matching algorithm.
Additionally, the Cloudron documentation provided the syntax for using the console: Cloudron Paperless-NGX Docs.

Fixing the issue is actually quite simple:

• Open the console of your paperless app and run

cd /app/code/src
python3 manage.py document_create_classifier

• Wait a few seconds/minutes
• For new documents, automatic classification should work as expected.
• For older documents, there is an action item reprocess which does not work as expected.

The reason why document classification stopped in the first place remains unknown.