Implement Spark Transformation Type

[utzwestermann]

We should build a Schedoscope-RDD that allows one to directly specify Spark pipelines outgoing from the dependencies of a view in the transformVia clause. Since all views are Hive table partitions, this should be doable using SQL-RDDs. More problematic is bringing the final RDD into the partition of the view itself.

[utzwestermann]

A basic Spark transformation has been implemented in Version 0.7.1. No inline Spark available yet tho.