db6

Unified database with pluggable storage engines (Memory / BTree / LSM).

• KV: Key-Value store
• SQL: Parser + Planner + Executor
• FTS: Full-Text Search with BM25 ranking
• Msgq: Message Queue (sync & async)

Installation

cargo add db6

Quick Start

cargo build
cargo test
cargo run

Storage Engines

Engine	Backend	Features	Use Case
Memory	BTreeMap (in-memory)	ORDER BY, range scan	Prototyping, cache
BTree	Disk BTree	Full transactions, persistence	General purpose
LSM	LSM-tree	Bloom filter, WAL, high write throughput	Write-heavy workloads

Key-Value API

use db6::{KvEngine, KvStore};

let engine = KvEngine::new("memory")?;
engine.put(1, b"key", b"value")?;
let val = engine.get(1, b"key")?;
engine.delete(1, b"key")?;

SQL

use db6::{parse, Executor, ResultSet};

let stmts = parse("SELECT * FROM users WHERE age > 18")?;
let mut exec = Executor::new(engine);
let result = exec.execute(&stmts[0])?;

Fluent Query API

use db6::Db;

let mut db = Db::new("memory")?;
let rows = db.select("name, email")
    .from("users")
    .filter("age > 18")
    .execute()?;

Full-Text Search

use db6::{FtsIndex, FtsTokenizer, CjkTokenizer};

let mut index = FtsIndex::new(engine);
index.insert(1, "database system")?;
let results = index.search("data")?;
let ranked = index.search_bm25("data")?;

Message Queue

use db6::msgq::Msgq;

let msgq = Msgq::new("memory")?;
let mut queue = msgq.queue("tasks");
queue.enqueue(b"work".to_vec(), 30)?;
let msg = queue.dequeue(0)?;
queue.ack(&msg.unwrap().id)?;

Publish

./pub.sh <new_version>
# e.g. ./pub.sh 4.14.0

Version History

Phase	Version	Features
v0.x	0.x	StorageEngine trait + 3 engines
v1.0+	1.x	FTS (inverted index + BM25)
v2.0+	2.x	SQL (parser, planner, executor)
v3.0+	3.x	Fluent API + LSM enhancements
v4.0+	4.x	Message Queue (sync + async)

Related Projects

• btree6
• lsm5
• sp6

License

MIT