I spent literally hours trying to make layout-parser.readthedocs.io/e work in virtualenv so that I can OCR text from newspaper article scans

Eventually, I got it to work (with lots of pinning) and thought: this will be a pain to productionize for my server! …but ultimately I realized all packages I needed were actually already packaged in nixpkgs! πŸ˜…

Now with a flake.nix like gist.github.com/stapelberg/45b, I have a reproducible (stable!) way of running this program πŸ₯³

screenshot of a newspaper article and how my flake.nix can be run to run the demo code for layoutparser + tesseract
0

If you have a fediverse account, you can quote this note from your own instance. Search https://mas.to/users/zekjur/statuses/114819160798929039 on your instance and quote it. (Note that quoting is not supported in Mastodon.)