Table exctraction from PDF is advertised but completely absent

[riccardomalavolti]

Version 0.1.3

docker run --rm -i markdown:latest < ~/example.pdf > output.md

where example.pdf is a native PDF (not a scanned document).

markitdown extracts the text but there's no sign of tables, the output is simply interleaved by newlines.

[bjfk2006]

Looking at the source code, you can see that pdf is still using pdfminer. You can see the effect of converting pdf to md. Don't have too high expectations. Now the ocr model is used to realize the conversion of text, tables and formulas.

[emcf]

If you're still looking to accurately extract the tables from PDF check out this library