openschichtplaner5-api

The REST API behind OpenSchichtplaner5 — a pip-installable FastAPI service over libopenschichtplaner5 (sp5lib), serving shift-planning data from the original Schichtplaner5 FoxPro .DBF files or the SQLite/PostgreSQL mirror: auth/2FA with JWT sessions, employees, schedule, absences, reports/exports (incl. the original personnel table via GET /api/personnel-table and statistics over a free evaluation period), leave-entitlement administration (POST /api/leave-entitlements/forfeit, annual close), notifications (SSE), webhooks, iCal feeds and more. All Soll/Ist/demand figures are computed by the sp5lib calculation facade — the API carries no arithmetic of its own.

Import name: the distribution is openschichtplaner5-api, but the importable package is sp5api (mirroring libopenschichtplaner5 → sp5lib). It was extracted from the main app's backend/api/ with full git history.

📖 Detailed documentation lives in the project wiki — endpoint overview, auth & permissions, full environment reference, deployment guides. Architecture notes are in docs/architecture.md.

What's inside

Permissions

Roles (Leser < Planer < Admin) gate read/write access; on top of that the granular 5USER rights of the original (WDUTIES, WABSENCES, WOVERTIMES, WNOTES, WDEVIATION, WCYCLEASS, WSWAPONLY, ADDEMPL, WPAST) are enforced per write route and exposed as a permissions object on GET /api/auth/me; WPAST=0 blocks every write with a date in the past, bulk routes included. The built-in Admin account and the last remaining administrator cannot be demoted or deleted. Full details: wiki → Auth & Permissions.

Installation

pip install openschichtplaner5-api

Prefer containers? See Docker below — the image runs standalone without a local Python environment.

Running

SP5_DB_PATH=/path/to/SP5/Daten python -m uvicorn sp5api.main:app --host 0.0.0.0 --port 8000

Key environment variables

Everything else (SPA serving, CORS, rate limits, brute-force lockout, SMTP, logging, password policy …) is covered by the wiki → environment reference and the main app's .env.example.

Docker

Production-ready multi-stage image (slim runtime, non-root, HEALTHCHECK on /api/health, EXPOSE 8000; SP5_DB_PATH=/app/data, SP5_BACKEND_DIR=/app/backend):

# local operation: DBF dir + optional .env (SECRET_KEY!), see docker-compose.yml
SP5_DBF_DIR=/path/to/SP5/Daten docker compose up --build

Each release publishes a multi-arch image (amd64+arm64) to the GitHub Container Registry, so no local build is needed. Standalone — mount the DBF directory to /app/data and pass the environment via .env (at minimum SECRET_KEY, see Key environment variables):

echo "SECRET_KEY=$(openssl rand -hex 32)" > .env

# published image (no build required):
docker run -d --name sp5-api -p 8000:8000 \
  -v /path/to/SP5/Daten:/app/data \
  --env-file .env \
  ghcr.io/mschabhuettl/openschichtplaner5-api:latest   # or :1.6.0

# or build it locally instead:
docker build -t openschichtplaner5-api .
docker run -d --name sp5-api -p 8000:8000 \
  -v /path/to/SP5/Daten:/app/data \
  --env-file .env \
  openschichtplaner5-api

# health check
curl http://localhost:8000/api/health
# → {"status":"healthy","checks":{"db":"ok",...},"version":"1.2.0",...}

The build installs the library from PyPI by default (libopenschichtplaner5[postgres]==1.7.0). To build against a different state, override the build arg:

docker build --build-arg LIB_SOURCE=git+https://github.com/mschabhuettl/libopenschichtplaner5.git@main .

PostgreSQL mode (standalone)

With DB_BACKEND=postgresql and DATABASE_URL set, the API runs against the PostgreSQL mirror. The initial schema is created automatically by the ORM (sp5lib) on first connect — no Alembic tree is required. The Alembic migration tree lives in the main application repo (backend/alembic) and is intentionally not bundled with this package (a copy would drift), so the startup auto-migration is skipped with an info log when no alembic.ini is found. To apply future schema migrations in a standalone deployment, point SP5_BACKEND_DIR at a directory containing alembic.ini + alembic/ (e.g. a checkout of the main app's backend/).

Development

python3 -m venv .venv && . .venv/bin/activate
pip install -e ".[dev]"
ruff check .
pytest

To develop against a local clone of the library instead of the PyPI release:

pip install -e "../libopenschichtplaner5[postgres]"

data/ and api/data/ at the repo root are runtime-state seeds (skills, wishes, notification settings) used by the test suite — the same layout the main app keeps under backend/, resolved via SP5_BACKEND_DIR. tests/fixtures/ holds the DBF fixture database.

Releasing

Tag vX.Y.Z and push — the release workflow builds sdist+wheel and publishes to PyPI via Trusted Publishing (OIDC).

License

MIT — see LICENSE.