Hi, Should flows always be in a separate standalone file or can they be dynamically created within a function or a class and registered as a Deployment. I'm trying something like this just want some inputs on whether it would work or not or what is the right way to dynamically create flows. I have a process which tries to convert data extracted from different loaders and store it in a db. These loaders and destination db are dynamically configured by the users so Im trying something like this. Any help , guidance or advice is appreciated.

class PrefectExecutor: 
   ## appropriate init method 
 
   def setup(self):
        """
        Create deployment flows for all the loaders
        """
        for loader in self.loaders:
            print(loader.config.id)
            self._build_and_deploy_loader(loader)

    def _build_and_deploy_loader(self, loader):
        try:
            from prefect import flow
            from prefect.deployments import Deployment
            from prefect.filesystems import LocalFileSystem
        except ImportError:
            print(
                """
                Prefect is not found. Install prefect with "pip install prefect==2.10.21"
                """
            )

        @flow
        def process_job(parameters: dict):
            print(parameters)
            stack = Stack(etl=loader, destination_db = self.destination_db)
            stack.etl.run(**parameters)

        deployment = Deployment.build_from_flow(
            process_job, name=str(loader.config.id), storage=LocalFileSystem.load("local")
        )
        deployment.apply()

hey @Sam Joel - have you checked out the `.serve()` method at all?

@Nate Yeah I checked out the serve but underneath isn't the serve method also creating a deployment ? and the main doubt I had is that the flow uses in memory variables from its parent class and when I create a deployment it tries to store the code to a filesystem how does the remote flow execution work in that scenario. Since it might not have access to the in memory variables . Can you please shed some light on this ?

> isn't the serve method also creating a deployment yes! > the flow uses in memory variables from its parent class I'm not quite sure what you mean by that, but I can't think of a scenario where you can't get the objects your flow needs by importing and / or accepting them via flow arguments - do you have a different situation?