rho-pdf-diff

Package forked from https://github.com/JoshData/pdf-diff

Installation

This package relies on the system dependency pdftotext. After that is installed, simply install using setup.py with `pip install -e .[dev,test].

Usage

Script

There is an included script for generating a text diff between two PDF documents, rho-pdf-diff. The simplest usage is:

rho-pdf-diff <old-pdf-path> <new-pdf-path> <output-path>

To see additional options run rho-pdf-diff --help. There are a number of flags for disabling some filters which attempt to remove unimportant diffs.

Library

The main workflow for creating diff images from raw documents can be found in the source of the rho-pdf-diff script. This essentially involves creating a Document object for both of the old and new PDFs, and passing these to initialize a DiffResult. DiffResult also can take a sequence of DiffFilter objects, which are essentially rules for using information available at the Document level for removing unwanted diffs from being included in the final result.

After the text diff is created by the logic in DiffResult, a searchable PDF containing the side-by-side documents with diffs marked is generated using the DiffImageWriter class. Note: the page alignment step is currently very slow for long documents with many diffs.