Why niltest?

When behavior is copied across implementation, tests, mocks, and a wiki, every change requires synchronization. niltest brings small behavioral examples back beside the implementation.

  • See representative inputs and results where the function is defined
  • Develop without a database or external API by returning fixed examples
  • Check the real implementation with those same cases

The core pattern in 30 seconds

Add @scenario, declare cases inside if expect:, then continue with an ordinary Python implementation.

python 
import niltest
from niltest import Mode, expect, scenario

niltest.configure(mode=Mode.TEST)  # @scenarioより前

@scenario("配送料")
def shipping_fee(subtotal: int, premium: bool = False) -> int:
    if expect:
        expect.case(
            "プレミアム会員は無料",
            given={"subtotal": 1_000, "premium": True},
            returns=0,
        )

    return 0 if premium or subtotal >= 5_000 else 500

A good fit when you want to

  • Share representative behavior for functions that call APIs or databases
  • Reduce drift between specifications and tests
  • Validate random IDs or tokens by type and conditions
  • Make business rules and small Python libraries easier to understand