Unit Tests¶

Unit tests verify a single function or class in isolation — without a database, network, or browser. They are the base of the test pyramid: the most numerous, the fastest, the most precise in their error messages.

What this level verifies¶

Backend business logic: pure calculations and engine rules (VPD via Tetens, GDD accumulation, EC budget), adapter logic (GBIF/Perenual enrichment).
Frontend logic without a DOM: Redux slices (reducers, actions) and custom hooks that are testable as pure functions.

Tested areas at a glance¶

Area	Tested elements	Extent
Domain logic (backend)	VPD/GDD/EC calculations, phase engine, karenz gate, companion/crop rotation	extensive
Repositories (backend)	ArangoDB access, graph queries	extensive
Migrations	schema migration framework	extensive
Celery tasks	background jobs, retention/anonymization	moderate
Adapters	weather (DWD/Open-Meteo/OWM/NASA), GBIF, Perenual, PlantNet, Home Assistant, notifications	moderate
Frontend	Redux slices, custom hooks	moderate

Tooling & location¶

	Backend	Frontend
Tooling	pytest (`asyncio_mode = "auto"`)	vitest
Location	`src/backend/tests/unit/`	`src/frontend/src/test/store/`, `…/hooks/`
Dependencies	none external	none (no provider wrapper needed)

Running¶

# Backend
cd src/backend && pytest tests/unit/ -v

# Frontend
cd src/frontend && npm test

Full prerequisites and patterns are in the testing concept → Backend Tests and → Frontend Tests.

Conventions¶

Engine tests instantiate the class directly and mock no repositories.
Service tests mock the repository (AsyncMock(spec=…)).
Redux slice tests run entirely without React — call the reducer as a pure function.
Every new feature needs at least one unit test for its business logic.