did we figure out when in the airflow rev history this was added?
unrelated to your diff, but I actually think function name is a little misleading. This is actually just executing one step in the normal case or just a few coalesced steps rather than the entire pipeline
Actually one test case to think about. An Airflow operator that executes N steps and the last one skips. Or we should constrain this so that we only execute a single step, which might be the right thing
good call. the underlying implementations are annoyingly a little different. in DockerOperator we're expected to implement execute which gets passed the Airflow context, whereas PythonOperator you're expected to provide a callable, and Airflow calls you with kwargs - fortunately they dump context into kwargs so I can get these at least a little closer in implementation
I created a DagsterSkipMixin wrapper over SkipMixin to gather this logic in one place, lmk if this is along the lines of what you were thinking here