Here's an updated PR from #958 (new branch with squashed commits)
This PR contains several building blocks related to the secondary indexing capability. It is an incomplete solution and is not wired up to a query planner at this time.

Contains:

• Redesigned the secondary index table structure (design)
• Implementation of JOIN, PARTIAL, and FULL secondary index types
• Integration test to verify 18 permutations of secondary indices are built correctly during ingest: (Accumulo, HBase) * (Numeric, Temporal, Text) * (JOIN, PARTIAL, FULL)
• Implementation of secondary index query method for Accumulo and HBase (to be called by the query planner based on the CBO’s “index decision”)
• Integration test for secondary index queries (again tests all 18 permutations listed above)
• Removes Lucene dependencies and fuzzy text matching classes