ckanext-doi

Gives every CKAN dataset a DOI (Digital Object Identifier) through DataCite, so it can be cited in publications.

The DOI is reserved as soon as the dataset is created, and registered with DataCite - resolvable at https://doi.org/<prefix>/<id> - once the dataset is active and public. A draft or private dataset keeps its DOI unregistered: you can quote it in a paper under embargo, it starts resolving the day the dataset is published. Its metadata (title, creators, dates, size, formats...) is sent again to DataCite whenever it changes.

Fork of NaturalHistoryMuseum/ckanext-doi, maintained by Phen-ICS (CNRS/IGBMC) for the FAIR3R catalogue. Published on PyPI as ckanext-fair3r-doi; the plugin is still called doi.

What this fork changes

• CKAN 2.12 (SQLAlchemy 2.0, Python 3.12), with upstream's Alembic migration for its table (ckan db upgrade -p doi).
• Test mode links work: in test mode, DOI links go to DataCite's test resolver (handle.test.datacite.org), since test DOIs don't exist on doi.org.
• Cleaner metadata: ISO 8601 dates, the size in a readable unit (and left out rather than "0 kb"), contributors kept apart from creators, and the DOI pointing at ckanext.doi.site_url when set.
• Clearer errors: a public dataset without an author is a form error on the author field, not a server error.
• Only datasets get DOIs: other package types (harvest sources, pages...) are left alone.
• Organization bulk actions (make public, private, delete) carry on past a dataset that fails.
• ckan doi update-doi skips private and deleted datasets and carries on past errors, instead of stopping at the first one.
• French translation, and CI that tests, builds and publishes to PyPI on every merge.

Requirements

• CKAN 2.12.
• A DataCite repository account: a test account (DataCite Fabrica test) for development, a live one to register real DOIs. Each has its own credentials and prefix.

Installation

1. Install the package in CKAN's virtualenv:

   pip install ckanext-fair3r-doi
   

2. Enable the plugin in ckan.ini:

   ckan.plugins = ... doi
   

3. Configure your DataCite account (see Configuration):

   ckanext.doi.account_name = ABC.DEFG
   ckanext.doi.account_password = <password>
   ckanext.doi.prefix = 10.1234
   ckanext.doi.publisher = My Institute
   ckanext.doi.test_mode = true
   

4. Create its table:

   ckan -c /path/to/ckan.ini db upgrade -p doi
   

5. Restart CKAN. Make a dataset public: the page shows "DataCite DOI created" and the DOI appears in its citation.

Configuration

Which fields go to DataCite

DataCite Metadata Schema 4.5 (4.2 with an older datacite library). The required ones:

Everything else (description, dates, sizes, formats, subjects...) is mapped in lib/metadata.py. Other plugins can add to or change it through the IDoi interface (below) - this is how ckanext-fair3r sends its FDF metadata (ORCID, ROR, genes, organisms...).

Commands

Templates

Citation snippets, for your own templates:

{% snippet "doi/snippets/package_citation.html", pkg_dict=g.pkg_dict %}
{% snippet "doi/snippets/resource_citation.html", pkg_dict=g.pkg_dict, res=res %}

Extending: the IDoi interface

A plugin implementing ckanext.doi.interfaces.IDoi can change the metadata before it is sent. Both methods run after this extension's own, in plugin order.

Tests

docker exec -u ckan -it ckan-app pytest --ckan-ini=/plugins/ckanext-doi/test.ini /plugins/ckanext-doi/ckanext/doi/tests

The DataCite API is mocked: no network or account needed. Always pass --ckan-ini: the fixtures empty the database they run against, and refuse to run unless its name contains test.

Credits and license

Originally written by the Natural History Museum, London. Released under the GNU General Public License v3.0 or later - see LICENSE.