numeraire.testing.check_estimator

numeraire.testing.check_estimator(estimator: Any, view_factory: Callable[[], Any], *, splitter: Any | None = None, min_train: int | None = None, forecast_kwargs: dict[str, Any] | None = None) → None

Run the full conformance suite; raise AssertionError on the first violation.

Parameters:

• estimator – Anything with fit(view) -> Model (a Estimator).

• view_factory – A deterministic zero-argument callable returning an equivalent view each call — synthetic data built with a fixed seed. Determinism is required because several checks rebuild the view to compare outputs.

• splitter – Forwarded to check_engine_roundtrip() to size the walk-forward run for the fixture (e.g. a small min_train for a short synthetic view).

• min_train – Forwarded to check_engine_roundtrip() to size the walk-forward run for the fixture (e.g. a small min_train for a short synthetic view).

• forecast_kwargs – Forwarded to check_engine_roundtrip() to size the walk-forward run for the fixture (e.g. a small min_train for a short synthetic view).