Archive lets you manage your Hive using source control. You write tables and views in a natural way as HiveQL SELECT statements, then scaffold the HiveQL with a Python dependency hierarchy to enable Archive to facilitate operating on your Hive.
This approach arose out of frustration with developing an elaborate warehouse by cutting and pasting database definitions from a git repository into Qubole's QPal. The abstractions likely leak all over the place and it may be awkward to those that come from a database administration background, but it serves the practical purpose of letting developers build and manage a Hive without having to do something as sophisticated as build a SQLAlchemy backend for HiveQL.
Archive is under active development at Radico, but is not ready for drop-in use by others. At this time, it should be thought of more as a thought experiment than a production-hardened tool.
