Skip to main content

​
Define workflow inputs

Define the inputs to a workflow by adding parameters to the function definition. 
from prefect import flow


@flow
def my_workflow(name: str = "World"):
    return f"Hello, {name}!"
Pass arguments to the function call to provide inputs to the workflow. 
my_workflow(name="Marvin")

​
Validate inputs

By default, Prefect will validate the types of inputs to your workflow. 
from prefect import flow


@flow
def my_workflow(name: str = "World"):
    return f"Hello, {name}!"

my_workflow(name=1) # This will fail with a type validation error
To turn off type validation, set the validate_parameters parameter to False: 
from prefect import flow


@flow(validate_parameters=False)
def my_workflow(name: str = "World"):
    return f"Hello, {name}!"

my_workflow(name=1)
If you use pydantic, parameters typed with a BaseModel subclass will be coerced to the appropriate types and validated. 
from prefect import flow
from pydantic import BaseModel


class Model(BaseModel):
    a: int
    b: str


@flow
def flow_that_validates_parameters(model: Model): ...

flow_that_validates_parameters(
    model={"a": "WRONG", "b": "fine"}
)