DOIdownloader

📝⬇️ DOIdownloader: You give it DOIs, it gives you the article PDFs.

It is surprisingly tricky to reliably obtain the full PDF of a scientific publication given its DOI. This Python package aims to do just that: you give it a list of DOIs, and it will download the full-text PDFs (or other formats if no PDF is available), taking care of much of the complexity. It ensures that lookups to different domains can happen asynchronously (i.e., one slow website won't stall all your other downloads).

DOIdownloader gives precedence to the publisher-formatted version (the so-called ‘Version of Record’), and will try downloading an open access pre- or postprint if you cannot access the publisher version. Importantly, DOIdownloader only tries downloading through routes that are widely considered to be legal. In more concrete terms, we do not download from Sci-Hub or similar platforms.

Installation

The package can be installed with pip:

pip install git+https://github.com/rafguns/doidownloader.git

Basic usage: command-line

The easiest way to get started is from the command-line. If you have a plain-text file of DOIs named dois.txt, you can download their PDFs as follows:

python -m doidownloader

This will download the results to a SQLite database named doi-fulltexts.db in the same directory. (You may notice that this also created a file called robots.txt. This is used to keep track of how long we should wait between calls to the same domain.)

Advanced usage: Python

Here's an example of how to use this from within Python:

import sqlite3
import doidownloader

# SQLite database where results will be stored
con = sqlite3.connect("somedois.db")
doidownloader.db.prepare_tables(con)
# List of DOIs to search for
dois_to_find = ["10.1108/JCRPP-02-2020-0025", "10.23860/JMLE-2020-12-3-1"]

async with doidownloader.DOIDownloader() as client:
    await save_fulltexts_from_dois(dois, con, client)

See __main__.py for an example of how to keep track of crawl delays per domain.