Skip to content

Changelog

All notable changes to Quiver will be documented in this file. The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Added

  • Support for custom distance functions
  • Improved error handling for DuckDB operations
  • Additional benchmarks for hybrid search

Fixed

  • Memory leak in batch processing
  • Race condition in concurrent searches

[1.0.0] - 2025-01-15

Added

  • Initial release of Quiver! 🎉
  • HNSW-based vector search
  • DuckDB integration for metadata storage
  • Hybrid search (vector + metadata)
  • Search with negative examples
  • Faceted search
  • Multi-vector search
  • Batch processing for vector additions
  • Persistence and recovery
  • Backup and restore functionality
  • Encryption for data at rest
  • HTTP API with Fiber
  • Comprehensive documentation
  • Benchmarks

[0.9.0] - 2024-12-10

Added

  • Beta release for early adopters
  • All core functionality implemented
  • Performance optimizations
  • Comprehensive test suite

Known Issues

  • Some edge cases in hybrid search not fully handled
  • Documentation needs improvement

[0.8.0] - 2024-11-15

Added

  • Alpha release for internal testing
  • Basic vector search functionality
  • Simple metadata storage
  • Preliminary API

How to Upgrade

Upgrading from 0.9.x to 1.0.0

  1. Update your Go dependencies:
go get -u github.com/TFMV/quiver@v1.0.0
  1. Update your code to use the new API:
// Old API
idx, err := quiver.New(config)

// New API
logger, _ := zap.NewDevelopment()
idx, err := quiver.New(config, logger)
  1. Migrate your data:
// Load from old format
oldIdx, _ := quiver.LoadLegacy("./old_data")

// Save in new format
oldIdx.Save("./new_data")

// Load with new API
newIdx, _ := quiver.Load(config, logger)

Upgrading from 0.8.x to 0.9.0

  1. Update your Go dependencies:
go get -u github.com/TFMV/quiver@v0.9.0
  1. Rebuild your index (data format changed):
// No automatic migration available
// You'll need to rebuild your index

Release Schedule

Version Expected Date Focus
1.1.0 2025-03-15 Performance improvements, additional distance metrics
1.2.0 2025-05-15 Enhanced backup strategies, cloud storage integration
1.3.0 2025-07-15 Distributed search, sharding capabilities
2.0.0 2025-10-15 Major architecture improvements, new search algorithms