Camelot VS unstructured-inference

I had good repeated success extracting tables from PDFs using Camelot (Python, https://github.com/camelot-dev/camelot)

I found this today and it is working well https://github.com/camelot-dev/camelot

here is the corresponding bug report in git: https://github.com/camelot-dev/camelot/issues/339

Try https://github.com/camelot-dev/camelot. It seems like this should work on your case.

Yeah, I’m pretty sure the Remarkable OCR is not up to these kinds of tasks unfortunately. If you know some coding you could write something that’d likely work well in Python using for ex. this for receiving the mail attachment and this for converting the PDF to CSV. This is in case you’d write your data as a table on the Remarkable, which I guess is preferable to writing something like (0.5, 8.4, -0.3). If you’d rather do it that way, there are other more suitable OCR tools like this one. The checkbox use-case in the comment above would also be possible by modifying this approach. DM if you’d like to discuss further work.

Looks like it's a bit in-progress: https://github.com/camelot-dev/camelot/pull/209

"Update docs" isn't checked, and that's what I was going on.

Tesseract is widely known to be "meh" at this point.

If you look at RAG frameworks as one example they'll typically use/support a variety of implementations. Tesseract is almost always supported but it's rarely ideal with projects like Unstructured[0] and DocTR[1] being preferred. By leveraging more-or-less SOTA vision models[2][3] they embarrass Tesseract.

I haven't compared them to the Apple Vision framework but they're absolutely better than Tesseract and potentially even Apple Vision.

[0] - https://github.com/Unstructured-IO/unstructured-inference

[1] - https://github.com/mindee/doctr

[2] - https://github.com/mindee/doctr#models-architectures

[3] - https://github.com/Unstructured-IO/unstructured-inference#mo...