article-epub

Description

A command-line tool written in Python to convert scientific articles available as HTML into ePub form for reading on a supported e-reader. Uses a plugin system with a "recipe" for each supported scientific publisher. Takes an article URL, title, or (ideally) DOI as input.

Obviously, you need to be able to legally access any article you want to convert, e.g. via a university library.

Like most web scraping applications, the provided recipes are liable to break frequently.

Currently, the following publishers are supported:

ScienceDirect (Elsevier)
Springer
Wiley
Oxford
BioOne
Royal Society
PLoS ONE
National Institutes of Health
NRC Research Press
Taylor & Francis
Annual Reviews
Nature Publishing
University of Chicago Press

Dependencies

Linux environment required
Calibre (to access ebook-convert)
Firefox with headless support
Geckodriver installed somewhere in $PATH
Pandoc

Python packages (available with pip):
Selenium
BeautifulSoup
pypandoc

Installation

I plan to properly package the application for installation with pip in the future, but for now it has to be installed manually.

Make sure the dependencies listed above are installed.
Clone the repository:

sh git clone https://github.com/kenkellner/article-epub
Make article-epub.py executable and put it in your $PATH.
To allow importing the library for use in other Python scripts, copy (or symlink) the article_epub folder into your local Python 3 library (e.g. ~/.local/lib/python3.6/site-packages/).

Usage

usage: article-epub [-h] [-u URL] [-d DOI] [-t TITLE] [-o FILE] [-p]

optional arguments:
  -h, --help  show this help message and exit
  -u URL      URL of article
  -d DOI      DOI of article
  -t TITLE    Title of article
  -o FILE     Name of output file
  -p          List supported publishers