Build Spark pipelines that read and update Neo4j.
Recent revisions are kept. Restoring one creates a new revision.