catlog22
|
31cc060837
|
feat: Add vector embeddings for core memory semantic search
- Add memory_chunks table for storing chunked content with embeddings
- Create Python embedder script (memory_embedder.py) using CodexLens fastembed
- Add TypeScript bridge (memory-embedder-bridge.ts) for Python interop
- Implement content chunking with paragraph/sentence-aware splitting
- Add vectorSimilarity dimension to clustering (weight 0.3)
- New CLI commands: ccw memory embed, search, embed-status
- Extend core-memory MCP tool with embed/search/embed_status operations
Clustering improvement: max relevance 0.388 → 0.809 (+109%)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
2025-12-20 13:09:43 +08:00 |
|