pdftext: pdftext: A package to extract text from PDFs

In jacob-ogre/pdftext: Extract Text from Text- and Image-based PDFs

Description

pdftext provides tools to extract text from PDFs, whether they have an embedded text layer and can be extracted with, e.g., pdf_text; or are purely image-based PDFs that require optical character recognition (OCR).

Details

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 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.

SYSTEM DEPENDENCIES

pdftext depends on three applications that must be installed and on the user's $PATH:

• tesseract, which is used for OCR, can be installed following the directions at https://github.com/tesseract-ocr/tesseract/wiki

• convert, which is used to convert PDFs to images, is part of ImageMagick

• unpaper, which is needed to automate pre-OCR cleanup (e.g., rotation), from Homebrew, another package manager, or source on GitHub.

• poppler, which is used to test if the input file is really a PDF Poppler

Install these two applications and you're one step away from having a working package.

jacob-ogre/pdftext documentation built on May 18, 2019, 8:01 a.m.