How coderdojo-linz/ai-workshops works
Project type and purpose
• This is a multi-purpose AI Projects monorepo containing:
– A standalone Java demo (Gps-Data).
– A collection of data-driven exercises (CSV files + Markdown guides).
– A “playground” area holding:
• IaC (Terraform) to provision cloud roles and resources.
• A Next.js workshop UI with API routes that serve exercises and support a chat interface.
– A CI/CD workflow for building and publishing Docker images.
File structure
– Top-level components:
• Gps-Data/ (Java application)
• exercises/ (CSV data + guide markdowns for interactive data exercises)
• playground/
– iac/ (Terraform configs for Azure roles)
– workshop-ui/ (Next.js frontend + API, Dockerfile, build-utils)
– CI/CD: .github/workflows/build-publish-docker.yaml builds/pushes the Docker container(s).
– Configs: .gitignore, tsconfig.json, Dockerfile, Terraform .tf files.
Key insights from README
– “playground” is for non-production AI experiments (workshop UI, IaC).
– The exercises folder is static content used by the Next.js API and by offline guides.
Main components & relationships
a. Components:
1. Java Demo (Gps-Data) – a standalone CLI or library.
2. Exercises Data Store – raw CSV files + Markdown guides under exercises/.
3. Infra Provisioner – Terraform code in playground/iac.
4. Workshop UI – Next.js app in playground/workshop-ui.
5. CI/CD Pipeline – GitHub Actions in .github/workflows.
b. Interactions:
• The Next.js Workshop UI reads exercise metadata and CSV data from the exercises/ folder at build-time or runtime via API routes.
• The UI’s API (src/app/api/exercises and /chat) serves JSON to the React frontend and proxies requests to an LLM (system prompt + user chat).
• IaC (Terraform) provisions cloud roles/storage that the Workshop UI or backend could consume (e.g., storing logs or data).
• CI/CD builds the Workshop UI Docker image and publishes it to a container registry.
• The Java demo is independent; it’s built and run locally via its own build tool (e.g., Maven/Gradle) if integrated.
c. Architectural patterns:
• Monorepo layout separating concerns by folder.
• Next.js server-side API routes + React client for a full-stack single-page app.
• Infrastructure as Code (Terraform) for reproducible cloud setup.
• GitHub Actions for CI/CD and Docker-based deployment.
d. Key technologies:
• Java 17 (Gps-Data)
• Node.js/TypeScript, Next.js, React (Workshop UI)
• Terraform (Azure provider)
• GitHub Actions, Docker