"Geekdocs Class" Course Website Template

This repository is a self-contained template to create a course website:

• using Hugo static site generator,
• hosted on GitHub,
• deployed to GitHub Pages hosting,
• using the continuous integration available through GitHub Actions,
• with (optionally) a custom domain name.

It builds on the Geekdocs theme for Hugo, and is inspired by Kevin Lin's Just-the-Class Jekyll template based on the Just-the-Docs theme for Jekyll. This template uses files from both of these projects under their respective MIT licenses.

It is designed to work well both on desktop and mobile — and to make it easy to create a course website with a schedule, announcements, staff listings, and modules. It is also designed to be easy to customize and extend.

Screenshots

Offline Site Generation

By default, the site is built and deployed by GitHub Actions.
If you’d like to hand someone a ZIP file or open the site from a USB stick, run this one-liner:

env HUGO_RELATIVEURLS=true \
    HUGO_CANONIFYURLS=false \
    HUGO_UGLYURLS=true \
    hugo -b "/" --cleanDestinationDir -d ./offline-website

Why this works

Setting	What it does
-b "/"	Generates root-relative URLs (/exam/problem01/stories.html).
HUGO_RELATIVEURLS=true	Rewrites those root-relative URLs so they become truly relative to the current page, inserting the correct number of ../ hops (stories.html, ../../about.html, …).
HUGO_CANONIFYURLS=false	Stops Hugo from “helpfully” turning those relative links back into absolute ones.
HUGO_UGLYURLS=true	Writes every page as a single .html file (/exam/problem01/problem.html) instead of /exam/problem01/index.html, so double-clicking works without a web server.
--cleanDestinationDir	Empties offline-website/ before each build so no stale files linger.
-d ./offline-website	Sends the finished, self-contained site to that folder; you can zip it or open index.html directly.

The result is a bundle you can drop on a USB stick or email: every link, image, CSS file, and script resolves correctly when opened at
file:///…/offline-website/index.html, even several levels deep in the directory tree.

Thanks to o3 (ChatGPT, OpenAI) for the troubleshooting and explanation!