Hi everyone, how can I manage reading and writing parts of my DAG from/to database. Let me give you the context here: I am building a DAG that has two main abstract functionalities: 1) training a model and updating this model, and I am saving the model's weights unstructuredly into a database, now I want to my DAG to read the model weights from the updates and update it using the incoming new data, how can I architect my DAG is a good way to accomplish that?

What's the challenge you're facing here?

The challenge is that my DAQ should update a model once in a while. The model weights are saved in a database, so the DAQ should check if there's weights already saved in the database to be use while updating the model and then write the updated weights again to the database

def prior_model_weights(file_path: str) -> Weights | None:
   # check for weights -- load or return None

def trained_model(prior_model_weights: Weights | None, ...) -> Model:
   # train model -- check if weights is None or not ...

def write_model_file(trained_model: Model, file_path: str) -> dict:
  # write models

something like that shoudl work -- now if the

| None

annotation doesn't work, file a bug 🙂