Before the first lesson

The labs run from a clone of the repository. Set it up once, before lesson 1, so the first lab is a walk through a running viewer rather than an install.

Install

Follow Getting started in the docs. In short: install uv, clone the repository, sync the workspace from the committed lock file and run the test suite once.

git clone https://github.com/MetaCoding-io/SemanticSeams.git
cd SemanticSeams
uv sync --all-packages
source .venv/bin/activate
pytest -q

A green suite means every Turtle and TriG file parses, every example passes the SHACL shapes, and every canned query returns the rows it should. If something fails here, fix it before the course: the labs assume this baseline.

Start the viewer

python3 -m uvicorn viewer.app:app --port 8000

Open http://localhost:8000. You should see the chat app's architecture view with a green health badge. Lesson 1's lab starts here.

What each module needs

Module	Needs	Notes
A (lessons 1 to 4)	The clone, the viewer	Lesson 3 regenerates the chat app with seams generate; nothing is run in Docker.
B (lessons 5 to 8)	A text editor	Chapters 1 to 7 of the expense tutorial run in process. No Docker, no network service.
C (lessons 9 to 11)	Docker from lesson 11	Chapter 8 starts the stack with docker compose. Lesson 9's language-model step is optional and needs a provider command; the hand-written path needs nothing.
D (lesson 12, capstone)	Depends on the capstone you choose	The generator capstone needs only Python. The domain capstone ends in docker compose up.

Keep a working file

The expense tutorial grows one file, expenses.trig, over eleven chapters, and the course's labs follow it. Keep your copy outside examples/ (for example work/expenses.trig) so that git status stays clean and you can diff your file against the chapter checkpoints in examples/expenses/. How checkpoints work explains the diff.

If you are teaching this

The instructor page has timings, which labs need Docker or a network, and what to do when a lab and the repository disagree.