ORCID publications

The website now merges ORCID works with Google Scholar conference entries. See PUBLICATIONS.md for the current refresh commands, duplicate matching rules and automatic workflow.

ORCID source

The ORCID importer writes _data/orcid_publications.json, fetched from the public ORCID record configured in author.orcid in _config.yml. Only public works can appear. Edit publication metadata and visibility on ORCID to update the website.

Run python3 scripts/sync_orcid.py to refresh locally. The script uses only the Python standard library. To import a previously downloaded works response, use python3 scripts/sync_orcid.py --source /path/to/works.json.

After the changes are pushed to master, Sync publications from ORCID and Scholar runs daily at 06:23 UTC, and can also be run from GitHub’s Actions tab. It commits changed data and explicitly requests a Pages rebuild, because commits made with GITHUB_TOKEN do not trigger the usual automatic build. This matches this repository’s existing Pages configuration (master, root, branch publishing). The workflow needs Contents and Pages write permissions and an unprotected branch, or branch rules that allow this bot to push. GitHub may disable scheduled workflows on inactive public repositories; re-enable the workflow in Actions if that happens. No ORCID token or repository secret is needed.

Works are sorted newest first within each category. ORCID’s preferred record is used for each work group, and duplicate DOIs are removed. Journal articles and reviews go into Journal articles; preprints have their own section. Conference work types, posters and lectures go into Conference presentations and posters. Copernicus EGU/EMS meeting DOI patterns are also classified as conference abstracts, since ORCID imports these as preprints. The metadata does not identify whether each abstract was a talk or poster, so those are labelled conference abstracts. Supplementary materials and other work types appear in Other research outputs. Preprints and their published versions remain separate works.

An unsuccessful fetch leaves the cached data untouched. Existing _publications files and their URLs remain available, with the old list as a fallback if the ORCID data file is absent.

References: ORCID public data and GitHub Pages build API.