Skip to content

CHANGELOG.md

Latest commit

 

History

History
189 lines (129 loc) · 6.88 KB

CHANGELOG.md

File metadata and controls

189 lines (129 loc) · 6.88 KB
Raw

Changelog

All notable changes to the RAPTOR AI Framework will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Planned

  • Community feedback integration
  • Performance optimizations
  • Extended documentation and tutorials
  • Additional examples

[Aigle 0.3] - 2026-06

🚀 Community Beta — Modular Platform Release

Added

  • Modular deployment system: 21 independent Docker Compose modules under Aigle/0.3/deployment/modules/, managed by a single build entry point (build.py / deploy.sh wrapper) with dependency-ordered start, per-module lifecycle (--stop/--delete/--restart/--build), status and logs
  • Hybrid search: OpenSearch BM25 + Qdrant vector retrieval with RRF fusion and cross-encoder re-ranking (module 17)
  • Graph database & GraphRAG: Neo4j entity/temporal-fact graph with LLM-powered graph query and reasoning, including relation date normalization (date_utils) (modules 19–20)
  • Agent Protocol (A2A): agent discovery, orchestration, and multi-agent RAG pipeline over vector / keyword / GraphRAG / TKG agents (module 21)
  • Video Search 2.0: video-centric search endpoint — multi-recall (BM25 + Vector + GraphRAG + TKG) → RRF fusion → cross-encoder rerank → per-video aggregation with time-coded segments
  • Branch-aware multi-tenancy: branch_id isolation propagated through upload, media processing, indexing, search, and agents
  • Demo frontend: React + Vite web UI (file upload, natural-language video search, upload history) with Docker deployment (Aigle/0.3/raptor-demo-frontend/)
  • Authentication module rework: restructured FastAPI app (app/ package) with group and account management, permission endpoint used by the API Gateway, SMTP e-mail notifications, updated Keycloak realm
  • Custom MLflow image (raptor/mlflow:0.3) in module 07
  • Build & source maintenance guide: Aigle/0.3/BUILD.md

Changed

  • GPU stack upgraded for Blackwell (sm_120 / RTX 50-series): CUDA base image 12.6 → 12.8, PyTorch 2.7.1 cu126 → cu128 wheels, PaddlePaddle GPU 3.0/3.1 → 3.3.0 (cu129); shared raptor/media-worker:0.3 base image for modules 09–12
  • SeaweedFS upgraded 3.96 → 4.32; LakeFS SDK calls moved off the event loop (thread pool); index/graph status sync added to asset management (module 04)
  • API Gateway permission model: JWT verification + permission check via authentication module /auth/permission (Keycloak UMA removed)
  • Configuration templates: per-modality model selection (VIDEO/AUDIO/DOCUMENT_INFERENCE_MODEL, IMAGE/VIDEO_VLM_MODEL_PATH), SMTP block, memory/timeout tuning keys; removed per-service *_GPU_COUNT, GATEWAY_KEYCLOAK_*, TKG_AGENT_URL, RERANKER_AGENT_URL
  • search API payload_schema now supports "contextual" only

Known Limitations

  • Module 14 (monitoring) is incomplete
  • Module 11 (video-processing) requires high GPU VRAM
  • Docker Compose deployment is not intended for production scale (Kubernetes planned for v1.0)

[Aigle 0.2] - 2026-02

Community Beta

  • Kafka-based multi-modal processing pipelines (audio / video / image / document)
  • MLflow model lifecycle management and Ollama model registration
  • Qdrant vector search APIs (video / audio / document / image)
  • Redis cluster caching, evaluation & testing API on DHT infrastructure

[Aigle 0.1.0-beta] - 2025-10-22

🎉 First Release

This is the first community beta release of the RAPTOR AI Framework by DHT Taiwan Team.

Added

Core Features

  • Multi-Modal Content Processing: Video, audio, image, and text analysis
  • Semantic Search Engine: Vector-based similarity search with context understanding
  • AI-Powered Metadata Generation: Automatic tagging and classification
  • LLM Orchestration Framework: Flexible integration with multiple language models
  • Content Intelligence Pipeline: Extract insights from unstructured media
  • Entity Recognition System: Identify people, places, objects, and concepts
  • Configuration Management: Flexible configuration for different deployment scenarios
  • Logging and Monitoring: Comprehensive observability and metrics

Documentation

  • Comprehensive README with project overview
  • Detailed RELEASE_PROCEDURE for future releases
  • CONTRIBUTING guidelines for community contributors
  • CODE_OF_CONDUCT for community standards
  • Apache 2.0 LICENSE file
  • Initial API documentation structure

Repository Structure

  • /Aigle/0.1/ - Source code for first release
  • /docs/ - Documentation directory
  • /examples/ - Example code and tutorials
  • .github/ - GitHub issue templates and workflows

Community

  • GitHub issue templates for bug reports and feature requests
  • Pull request template
  • Community communication channels setup (GitHub, Telegram, Instagram, X)

Known Limitations

This is a beta release. The following are known limitations:

  • Limited documentation coverage for advanced features
  • Some features may not be fully optimized for production use
  • API may change in future releases based on community feedback

Breaking Changes

None (first release)

Security

  • All dependencies audited for known vulnerabilities
  • Secure default configurations implemented

Contributors

Special thanks to the DHT Taiwan Team for developing this first release.

Links

Version History Summary

Version Codename Release Date Type Status
0.1.0-beta Aigle 2025-10-22 Beta Current

How to Read This Changelog

Version Format

[Codename Major.Minor.Patch-stage] - YYYY-MM-DD

Change Categories

  • Added: New features
  • Changed: Changes in existing functionality
  • Deprecated: Soon-to-be removed features
  • Removed: Removed features
  • Fixed: Bug fixes
  • Security: Security-related changes

Symbols

  • 🎉 Major release or milestone
  • 🚀 New feature
  • 🐛 Bug fix
  • 📝 Documentation
  • ⚡ Performance improvement
  • 🔒 Security fix
  • ⚠️ Breaking change
  • 🗑️ Deprecation

Upgrade Guide

To Aigle 0.1.0-beta

This is the first release, so no upgrade is necessary.

For future upgrades, we'll provide detailed migration guides here.

Feedback and Contributions

We welcome your feedback on this release! Please:

  1. Report bugs: GitHub Issues
  2. Request features: GitHub Issues
  3. Contribute: See CONTRIBUTING.md
  4. Discuss: Join our community channels

Maintained by DHT Taiwan Team

For more information, visit https://dhtsolution.com/