I am still :ablobcatsweatsip: ing on Nostr: I've been nervous about running my OCR script on my image collection even though it's ...

I've been nervous about running my OCR script on my image collection even though it's versioned and backed up. I think what I'm going to do is add a before/after verification pass to ensure that image content was not modified and that the composition of folders remains the same.

Given how it works, it's incredibly unlikely that content would be tampered with. If it is however, that would be hard to catch and verify, so I just don't want to take any chances.

so, before: for each folder, calculate hashes for each file.
after: run over the renamed images. Make sure the number of hashes per folder is the same. Make sure none of those hashes in the set are different.