Get the publisher (or any variables) from a DOI via DataCite

datapublisher_from_doi.py

import xml.etree.ElementTree as ET

import requests

# curl -LH "Accept: application/vnd.datacite.datacite+xml" https://doi.org/10.34894/KEIQRD

def get_datapublisher_from_doi(doi):
    """Get the publisher from the DOI.

    Arguments
    ---------
    doi: str
        The DOI to find the publisher for.

    Returns
    -------
    str:
        The publisher.

    """

    r = requests.get(
        f"https://doi.org/{doi}",
        headers= {"Accept": "application/vnd.datacite.datacite+xml"},
        allow_redirects=True
    )

    if r.status_code != 200:
        raise ValueError("DOI not found")

    tree = ET.fromstring(r.content)

    node_pub = tree.find("{http://datacite.org/schema/kernel-4}publisher")
    if node_pub is not None:
        return node_pub.text

    return None


def test_datapublisher_retrieval():

    assert get_datapublisher_from_doi("10.34894/KEIQRD") == "DataverseNL"