Source for the CTTIR group website, published at https://cttir.github.io/website/.
Built with Hugo (extended) and the Coder theme loaded as a Hugo Module. Deployed to GitHub Pages by GitHub Actions on every push to main.
hugo mod get -u
hugo server -DThen open http://localhost:1313/website/.
GitHub Actions workflow .github/workflows/hugo.yml builds and publishes the site. After first push to main, in the GitHub repository: Settings → Pages → Build and deployment → Source: GitHub Actions.
static/images/cttir-logo.png— 2×2 black tile with white "C T / I R" lettering. Used as landing-hero logo and favicon source.static/images/favicon.ico(optional, generated from the logo).
The About page includes a repository/tag graph below Scope. Its Hugo shortcode
reads data/software.yaml (public CTTIR repositories only), project membership
from data/projects.yaml, and proposed workflow membership from
data/workflows.yaml. Add cross-cutting topics in data/repository_tags.yaml.
Update the catalogue when public repositories are added; private repositories
and private project identifiers must not be included. All repositories remain
visible by default, including those without project or workflow assignments.
Checkboxes choose which tag families form graph groups. Separate facet filters use OR within each family and AND across families; search also matches tags and descriptions. The matching list provides keyboard access to repository details, and a server-rendered list remains available without JavaScript. Planned workflow membership describes editorial scope, not a tested package dependency.
The graph uses a locally served copy of vis-network 9.1.9, matching the tutorials
network, with its MIT license in static/js/vendor/. It makes no runtime GitHub
API or graph CDN requests. Rendering and behaviour live in
layouts/shortcodes/repository-network.html, assets/js/repository-network.js
and assets/css/custom.css.
Before publishing, build with hugo --minify; check the complete inventory,
search, each grouping family, combined filters, empty results, reset, repository
links, zoom, keyboard access, and narrow-screen layout.
MIT — see LICENSE.
Where applicable, LLM tools were used for narrowly defined, non-authorial tasks: copyediting, prose refinement, Markdown/LaTeX formatting, boilerplate scaffolding (such as CI configurations and build scripts), and code refactoring. Tools included Chat AI, the KISSKI (GWDG) service, and self-hosted Mistral Small (24B, Apache-2.0) served through a local vLLM architecture on an Ubuntu Linux workstation with Intel processors and NVIDIA GPUs. The self-hosted model ran locally, with no data sent to third parties.
All scientific claims, methodological choices, analyses, interpretations and conclusions remain the authors' own. Where applicable, LLM-generated material must be reviewed and revised before inclusion, and references must be verified against their DOI, arXiv ID or ISBN.
Local software setup: vLLM with Python/PyTorch, plus R and RStudio.