I am trying to submit beams wordcount python example to a remote spark cluster on emr running yarn as its resource manager. According to the spark documentation this needs to be