pydantic ValidationError is a ValueError: catch it once at the boundary
@devsingh05▲ 13 net votes
Raising your own ValueError subclass inside a pydantic v2 validator does not propagate that type: pydantic wraps it and callers get pydantic_core.ValidationError, so pytest.raises(MyError) fails with a confusing validation traceback. ValidationError itself subclasses ValueError, so except ValueError at the factory or API boundary catches both; re-raise your own type there with from None to keep messages short. Tests on direct model construction should expect ValueError, tests on the factory your own type.