Contribute

Datafeeder is a monorepo: a FastAPI backend, an Angular frontend, Airflow ELT DAGs, and shared Python libraries, managed together with uv (Python workspace) and npm/nx (frontend).

See also the Architecture page for how the pieces fit together.

Setting up a dev environment

See the prerequisites and quick-start pages. In short:

make install-python   # installs Python deps via uv, writes AIRFLOW_UID into .env
make up                # backend deps + docker compose (georchestra, airflow, geoserver, geonetwork)
make run-backend       # backend, with auto-reload
cd apps/frontend && npm install && npm start

make help lists every available target.

Running the tests

make test-libs             # libs/data_manipulation, pytest
make test-backend          # apps/backend, pytest
make test-backend-coverage # apps/backend, pytest with html/term coverage report

Frontend tests (from apps/frontend/):

npm run test:ut      # unit tests (Vitest)
npm run test:e2e     # end-to-end tests (Cypress)
npm run test:e2e:ci  # headless mode, used by CI

Linting and formatting

make fix-and-check-all-python   # ruff lint --fix, format, then check --verbose

Frontend formatting/linting is run via npm run format (see apps/frontend/package.json).

Keeping the frontend API client in sync

The frontend's TypeScript API client is generated from the backend's OpenAPI schema and is not hand-written. Any change to a backend route or model must be followed by regenerating it — see Regenerating the API client.

AI agent skills

This repository ships domain-specific instructions for AI coding agents under .agents/skills/ (Airflow DAGs, Angular components, FastAPI endpoints, frontend/API sync, geospatial data handling, Figma-to-code, Tailwind design systems). agents.md at the repository root is the entry point.