Clean up `SparkStreamingDataSet`

[deepyaman]

Right now, it's a copy-pasted version of SparkDataSet with minimal modifications. This results in some unused arguments, etc. in the YAML config (e.g. filepath: ignore).