What exactly does a spark task contain? Does it have Javabytecode that acts on each partition of data? Can it have more than one operation like map, filter, filter, etc?