In the context of Spark Streaming, which method allows you to define a function that executes on the output of every micro-batch?