I have a generic DataManager (need for implementing subclasses) looking like:

class DataManager():
    @task
    def download_data(self, source: str, destination: str):
        pass

And I tried to use that in a flow like:

@task
def get_manager():
    return DataManager()

with Flow("Flow") as flow:
    manager = get_manager()
    manager.download_data(...)

This fails during build with

AttributeError: 'FunctionTask' object has no attribute 'download_data'

because of the object

manager

I understand this is because of my definition of the Manager object, and I could remove the

@task

from my library's functions, and use them as usual python function inside my flow But I'm not sure if there's already a standardized way to implement tasks in a python package ? I haven't found anything in the docs yet