The files within the docs folder are publicly displayed on the Image section's website.
To include your latest publications from Google Scholar, update the list of author ids included in the "env" file. To find your author id, navigate to your google scholar page and extract the id from the "user" query parameter in the URL. "user=<your_id>". Then, include a comma followed by your author id in the env > AUTHOR list, i.e ",<your_id>".
The workflow fetches publications through the SerpApi Google Scholar Author API. The repository needs a secret named SERPAPI_API_KEY (Settings → Secrets and variables → Actions, or gh secret set SERPAPI_API_KEY); the free plan covers the monthly usage. If the fetch returns no publications, the workflow fails without changing the file.
The monthly workflow merges newly scraped publications into docs/_data/publications.yml instead of replacing the file, so entries that the scraper no longer returns (for example older publications) are kept.
Every SerpApi search counts against the monthly quota, so a run costs one search per author: the results are sorted by publication date, and the fetch stops as soon as a page ends on a title that is already stored. A run that finds nothing new therefore costs 11 searches. The fetch also leaves the citation ids it paid for in updated-citation-ids.json, which the enrichment step reuses instead of requesting every author profile again. Enrichment fills in complete author lists for entries whose Google Scholar author list is truncated ("..."), first from Crossref and, for at most 25 entries per run, from the SerpApi citation endpoint. A Google Scholar profile that is private returns no articles; the run then prints a warning naming the author id instead of dropping that author silently.
The homepage shows the first five entries of docs/_data/publications.yml, i.e. the most recent publications by year in the order they appear in the file. The monthly merge keeps the newest scrape at the top of the file. Ordering within the same year is not meaningful: the scraper shuffles entries before sorting them by year.
Student projects are listed on the "Student Projects" page from the data file docs/_data/student_projects.yml. To publish a project, append an entry with a title, description, supervisor, contact e-mail, levels, topics, and status ("open" or "taken"). levels and topics are YAML lists, so a project can target several levels and cover several topics. The optional courses list contains recommended courses to have completed before the project (e.g. courses: [Image analysis]); it is shown on the card. Put the card cover image in docs/assets/img/projects/ and reference it from the entry as assets/img/projects/<file>. Cards are shuffled on each page load. Changes take effect once the change is merged to the main branch.
The status, level, and topic filters are generated automatically from the entries, and the level and topic filters allow selecting several values at once. Use consistent spelling (e.g. always "Master Thesis") so projects group correctly.
The current entries were imported from the Projects4Students Trello board: all projects to the left of the "Supervisors to the right do not accept projects" divider are listed. Levels come from the board's color labels, and topics were derived from the project descriptions.
Follow the jekyll website tutorial to develop and compile the pages locally.
To run the site locally, we can use docker.
Build the image:
docker build -t image-site docs
Serve the site:
docker run --rm -p 4000:4000 -v "$PWD/docs:/site" -w /site image-site bundle exec jekyll serve --host 0.0.0.0 --destination /tmp/_site
The site is then available at http://localhost:4000/image-website/ (the new page is at http://localhost:4000/image-website/student-projects/).