Making PDFs Searchable Using OCR (CLI)

[ernz]

This has taken be the entire evening to figure out. It appears that when it comes to linux, there are dozens of paid-for options or dead FOSS projects.

Lots of old tutorials reference tesseract, hocr, gs conversions, cuneiform, scantailor. I have tried many MANY combinations of all of these and they always have something wrong. Resolutions are changed, the overlaid hocr mappings are off or the font size if huge, etc. Many of the how-to's are now 8+ years old and not of any use any more.

I finally settled on https://github.com/fritz-hh/OCRmyPDF

It's a beautilfully written python extension with lots of cli flexibility. It does have lots of dependencies, but it works exceptionally well!

[TheFu]

gscan2pdf didn't work for you?
It worked on my 10.04 box, but I haven't tried it on the fresh 14.04 install.