Rehearse builds, migrations and API startup

In this lesson you will rehearse builds, migrations and API startup. Work with the Equipment Rental API in the downloadable lab. The reference is a complete solution with separate lesson checks, so you can inspect the answer, make a deliberate local change and verify its behavior.

Delivery needs a repeatable sequence: install or build, supply configuration, apply reviewed migrations, start the API, then check it. The Dockerfile installs the lock and copies only the runtime source and migrations. It runs the API as a non-root uid and does not embed a database URL, token key or .env file. The image command starts the app factory; it does not automatically migrate or create privileged users.

Treat schema changes as a separate controlled step. The local runner starts a fresh PostgreSQL container, applies the real migration chain and executes the lessons, then removes the test database. For an existing deployment, preserve data, verify backups and compatibility, apply migrations deliberately and retain a working rollback route. A previous image alone cannot reverse a schema change safely. Our migration downgrade refuses destructive deletion, so a rollback plan needs an appropriate forward repair or verified restore. Readiness, schema revision and a representative contract request together provide stronger evidence than a running container alone. This lesson rehearses locally and authorizes no Internet deployment.

Worked source: equipment/api.py, create_app.

Run lesson25, then python run_checks.py for the full suite. The startup check compares user counts before and after the app starts; no account is silently created. Inspect Dockerfile and .dockerignore. Optional local image build: docker build -t equipment-rental-learning:1.0.0 .; it still requires separate runtime configuration and a migrated disposable database.

def test_lesson25(db, engine):
    # Startup uses an already-migrated schema; it does not create tables or users.
    with engine.connect() as connection:
        before = connection.scalar(text("SELECT count(*) FROM users"))
    app = create_app(
        Settings(database_url=os.environ["LAB_DATABASE_URL"], jwt_secret=KEY)
    )
    with TestClient(app) as client:
        assert client.get("/equipment").status_code == 200
    with engine.connect() as connection:
        assert connection.scalar(text("SELECT count(*) FROM users")) == before

python run_checks.py -k lesson25