5.4. Combining Pydantic and semver

According to its homepage, Pydantic “enforces type hints at runtime, and provides user friendly errors when data is invalid.”

To work with Pydantic>2.0, use the following steps:

  1. Derive a new class from Version first and add the magic methods __get_pydantic_core_schema__() and __get_pydantic_json_schema__() like this:

    from typing import Annotated, Any, Callable
    from pydantic import GetJsonSchemaHandler
    from pydantic_core import core_schema
    from pydantic.json_schema import JsonSchemaValue
    from semver import Version
    class _VersionPydanticAnnotation:
        def __get_pydantic_core_schema__(
            _source_type: Any,
            _handler: Callable[[Any], core_schema.CoreSchema],
        ) -> core_schema.CoreSchema:
            def validate_from_str(value: str) -> Version:
                return Version.parse(value)
            from_str_schema = core_schema.chain_schema(
            return core_schema.json_or_python_schema(
                serialization = core_schema.to_string_ser_schema(),
        def __get_pydantic_json_schema__(
            cls, _core_schema: core_schema.CoreSchema, handler: GetJsonSchemaHandler
        ) -> JsonSchemaValue:
            return handler(core_schema.str_schema())
    ManifestVersion = Annotated[Version, _VersionPydanticAnnotation]
  2. Create a new model (in this example MyModel) and derive it from pydantic.BaseModel:

    import pydantic
    class MyModel(pydantic.BaseModel):
        version: _VersionPydanticAnnotation
  3. Use your model like this:

    model = MyModel.parse_obj({"version": "1.2.3"})

    The attribute model.version will be an instance of Version. If the version is invalid, the construction will raise a pydantic_core.ValidationError.