Spark on Kubeflow Pipelines TODO Simple spark-app block using ResourceOp with success condition Convert ipynb into VolumeOp and attach into previous ResourceOp