An R package to extract text from PDFs
pdftext provides tools to extract text from PDFs, whether they have an embedded text layer and can be extracted with, e.g., pdftools::pdf_text; or are purely image-based PDFs that require optical character recognition (OCR).
The main function of pdftext, pdf_to_txt, gets the text either by extracting from the text layer (if available) or by OCR (if image-based) and writes the text to a file. The helper function load_text will read one of those text files and return a list, with each element of the list a single page (as done with pdftools::pdf_text). The "workhorse" functions for generating images for OCR and doing the OCR, are exposed so they can be used easily in other contexts.
pdftext depends on four applications that must be installed and on the user's $PATH:
- tesseract, which is used for OCR, can be installed following the directions here;
- convert, which is used to convert PDFs to images, is part of ImageMagick; and
- unpaper, which is used to prep page images for OCR, can be installed with a package manager like
apt(Linux Debian distros) orhomebrew(OS X), or by forking the repo and compiling. - poppler (
pdftotext), to test if the "PDF" is actually a PDF.
Install these four applications and you're one step away from having a working package. The last step is to set a couple of options (optionally), as indicated in the startup message when pdftext is loaded.