Labs

Hands-on exercises with complete runnable code and a zero-setup environment.

Open any lab in GitHub Codespaces or Gitpod (or docker compose up

locally) and the required services are already running — no DB connections, no

cloud, no devops.

Available Labs#

How a lab is structured#

lab-XX-name/
  lab.json          manifest: module link, infra needed, languages, run/test commands
  README.md         instructions
  python/
    stub.py         the file you implement (TODOs)
    solution.py     reference solution
    test_lab.py     run to validate your work

To validate your work:

bash
cd labs/<lab>/python
python -m unittest test_lab.py     # green OK = done

The platform layer (labs that need persistence)#

Labs with infra in their lab.json (e.g. lab-03) use real Postgres and

Redis — but you never write a connection string. Import the ready handles:

python
from labkit import db, cache

db is Postgres (db.query, db.queryone, db.execute, db.query_count);

cache is Redis (cache.get_json, cache.set_json, cache.delete, ...).

The services and credentials are wired by the environment — see

`.devcontainer/` and `tooling/python/labkit`.

What Makes a Good Lab Session#

1. Run the solution unmodified first — see the baseline behaviour
2. Implement the stub — make test_lab.py go green
3. Change one parameter at a time — observe the effect
4. Explain the results to yourself — connect to the theory in the parent module