Spark-based feature extractor

This release uses Apache Spark for performing feature extraction on very large sets of volumes, in parallel, potentially across many computing nodes.