@Marvin how does prefect handle parameters with arbitrary types?

Prefect handles parameters with arbitrary types primarily through the use of Pydantic models, allowing robust and dynamic type validation. Here are the key points: 1. Built-in Types: If a built-in type like

int

is used as a parameter, Prefect creates a Pydantic model containing a field annotated with that type. This allows for extensive type annotations. 2. Pydantic Models: You can also pass a

pydantic.BaseModel

class directly, which Prefect will automatically handle, uprading it to a

RunInput

class that enables flows to send and receive inputs at runtime. 3. Arbitrary Types: For more advanced use cases, Prefect uses Pydantic's

arbitrary_types_allowed

to validate plain Python classes through simple

isinstance

checks, allowing arbitrary types to be used as parameters. For a more detailed guide, you can refer to the official document: Creating Interactive Workflows Additionally, here's an example pull request addressing the usage of arbitrary types: - Allow Python classes as flow/task type hints. It's a wonder why anyone bothers creating flexible workflows, given the inevitable chaos. But alas, as I always say, "Here I am, brain the size of a planet, and they ask me to look up documentation."