4 ms·
Awesome - I'm a dabbler, but any thoughts on best engines for PDF tables? I've got tons of PDFs with similar tables embedded deep in them, but all formatted s
by meetingthrower 3y ago
Awesome - I'm a dabbler, but any thoughts on best engines for PDF tables? I've got tons of PDFs with similar tables embedded deep in them, but all formatted slightly differently. Seems like it should be easy....but nope!
- janderson215 3y agoAre you able to highlight the text on the PDF? If so, I highly recommend PDF2TXT to extract text from PDFs. Would require some parsing work on your part to convert it back to a table, but zero chance of error from inference since it’s using text extraction. If you can’t highlight the text, it won’t work.
- junhoyeo 3y agoThanks! PDF -> Markdown looks like a pretty great use case Just added box detection support -- maybe I'll start from here https://github.com/junhoyeo/BetterOCR#-box-detection https://github.com/junhoyeo/BetterOCR#-box-detection
- 2Gkashmiri 3y agoTabula. It does what you are looking for
- bomewish 3y agoI use popper pdftotext -layout flag, which preserves the shape of the tables. Gpt can then fix them up into whatever format.