• Expand key size to support future versioning
• queries run on snapshots
• basic support for inserts, supported using transactions (completely revamped!)
• faster ingestion of datasets
• load more ontologies by default (BrickTag, BrickUse)
• ?x ?y ?z queries supported
• general performance improvements, bug fixes, improved error messages

Better k8 packaging and BOSSWAVE permissions

Supports multiple buildings and is better packaged for Kubernetes/etc

Currently only builds for Linux systems

• lexer pool for queries was causing syntactically incorrect queries to reappear when subsequent correct queries were submitted
• weird bug in btree incorrectly detected duplicates; doesn't seem to have affected other results
• added sample container/service definitions

This release should fix most of the correctness problems we were seeing (using the tests existing as of gtfierro/brick_database_eval@c427432)

Some work has also been done to reduce intermediate allocations, meaning the queries should run a little faster.

New in this release:

• all operators implemented except for unqualified ?s ?p ?o triples
• HTTPS support for query frontend
• graph exploration draft frontend

Previous choices in how the internal query state was handled made it difficult to implement predicate variables, and the code for generating the list of final tuples was complex and a bit slow due to an influx of small object allocations.

The new structure is slightly changed, but the core query logic remains untouched. The internal Hod operators now take a *queryContext pointer that provides some nice higher level functions for implementing the operators.

The goal is to net 2 benefits from this change:

• easier implementation of predicate variable operator queries
• easier to integrate object pools to reduce allocation during query execution