I wonder if the following is possible... invoking a dagster run as if it were an API call (a very slow one) Suppose I have:

@asset 
def data()->pd.DataFrame:
    ... 

@asset 
def clean_data(data)->pd.DataFrame:
    ... 

@asset 
def predictions(clean_data, model)
    ...

It'd be nice to be able to "call" this pipeline with some arbitrary input for data I envision a jupyter notebook where

data

is some pandas dataframe (generated by some process) - would there be a way to leverage the pipelines defined in dagster to calculate the

predictions

on this data?

the graphql api might help you here https://docs.dagster.io/concepts/dagit/graphql