lightdb

Computationally focused database using pluggable store + indexer

Provided Stores

• Yahoo's HaloDB (https://github.com/yahoo/HaloDB) - Preferred for performance
• MapDB (https://mapdb.org)
• Facebook's RocksDB (https://rocksdb.org)
• Redis (https://redis.io)

Provided Indexers

• Apache Lucene (https://lucene.apache.org) - Most featureful
• SQLite (https://www.sqlite.org) - Fastest
• H2 (https://h2database.com)
• DuckDB (https://duckdb.org) - Experimental

1.0 TODO

• More performance improvements to SQLite integration
  • HikariCP for connection pooling
  • Property transactions without autocommit
• New benchmarks
  • JMH
  • Better parallel processing
  • Automated generation of performance bar charts
  • Shell script to run all benchmarks
• Rewrite
  • Cross-Platform
  • Transactions
  • Single listener to capture all events on a Collection
  • No AbstractCollection
  • DB.collection and / Model separation
  • Remove Pagination for better streaming "SearchResults"
  • No IndexSupport, simply instantiated index and pass ref to collection
    • Support for multiple indexes on the same collection
  • Generic unit tests with multiple implementations
  • Document classes
  • MDoc support
• New Integrations
  • Add Redis support
  • Support H2 as another SQL database

1.1 TODO

• Add btree supported indexes in Store
• Support SQL as a Store as well as Indexer
• Support Lucene as a Store as well as Indexer
• Provide a Scala.js implementation of Store using Web Storage API
• Provide a Scala.js implementation using IndexedDB
• ScalaNative support