wedb_embed
wedb_embed is a high-performance, disk-backed, multi-model storage engine built on top of the Fjall LSM-Tree engine. It provides zero-cost abstractions for complex Redis-compatible data structures, transaction capabilities, and metadata-aware key composition.
🚀 Features
- Multi-Model Data Structures:
- String: Raw KV, numerical float/int operations (
incr,decr,increx,getdel,getset). - Bitmap & Bitfield: Segmented LSB bitmap allocation, arbitrary-width signed/unsigned bitfield manipulation (
bitfield,bitfield_read_only). - Hash: Subkey encoding with versioning, per-field TTL expiration (
hexpire,hgetdel,hrandfield). - List: Double-ended deque with fast index-based access, trimming, pop, and atomic move (
lmove,lmove_m). - Set & Sorted Set (ZSet): Fast member lookup, score range iterators, set algebra (
sdiffcard,sunioncard). - Stream: Stream entry indexing, consumer group metadata, and PEL tracking.
- RedisJSON: Native JSON tree parsing with JSONPath query execution using
sonic-rs. - Probabilistic Data Structures: Scalable Bloom filters, Cuckoo filters, T-Digest estimators, and HyperLogLog.
- TimeSeries & Geospatial: Chunked timeseries compression and 2D Geohash spatial distance indexing.
- String: Raw KV, numerical float/int operations (
- Microsecond Read/Write Performance: Powered by Fjall LSM-Tree partition routing and memory-mapped block caches.
- KeyComposer: Namespace and database isolation, slot calculation (
CRC16), and internal key prefixing. - TTL Engine: High-resolution millisecond key and subkey expiration state machines.
📦 Installation
Add wedb_embed to your Cargo.toml:
🛠️ Usage Example
use ;
📜 License
Licensed under either of Apache License, Version 2.0 or MIT License at your option.