public class NamedAggregators
This class wraps a map of named aggregators. Spark expects that all accumulators be declared
before a job is launched. Beam allows aggregators to be used and incremented on the fly.
We create a map of named aggregators and instantiate in the the spark context before the job
is launched. We can then add aggregators on the fly in Spark.