Hi! I am trying to understand how flyte converts the Python workflow into DAG that can be run on airflow / k8s? Any pointers on how Python code is parsed? Is the DAG represented as a proto? Will like to look into running parts of DAG on AWS lambda.
09/10/2023, 3:24 AM
It does not run it on airflow
It runs it on its own scheduler engine
The intermediate representation is in protobuf
09/10/2023, 1:50 PM
Got it. Thanks. Any code pointers where this intermediate protobuf representation is generated?