feat: Add pgvector backend with LlamaIndex adapter#233

Open

esafwan wants to merge 2 commits into

kimi/vector-store-pgvector-backend

Contributor

esafwan commented Mar 30, 2026

Summary

Implements PostgreSQL/pgvector backend using LlamaIndex adapter.

Implementation

PGVectorBackend - Full KnowledgeBackend implementation
LlamaIndex integration - Uses llama-index-vector-stores-postgres
Connection management - Frappe config integration with defaults
Feature support - Metadata filtering and hybrid search ready

Files

huf/ai/knowledge/backends/pgvector_backend.py
huf/ai/knowledge/tests/test_pgvector_backend.py

Dependencies

Optional: llama-index-vector-stores-postgres, psycopg2-binary

Related

First production-ready vector backend beyond SQLite.

esafwan added 2 commits

March 28, 2026 11:04


          docs: add HUF Memory System design documents

8de517d

- PRD: Programmable Memory & Learning Layer
- DocType designs: Memory Record, Policy, Profile
- Tech specs: Capture/Retrieval, Storage Architecture
- Profiles: Travel Planning, Programming, Documentation
- Implementation plan and project tracking


          feat(knowledge): add pgvector backend with LlamaIndex adapter

3ef17e2

- PGVectorBackend implementation using llama-index-vector-stores-postgres
- Support for add_chunks, search, health_check operations
- Capability flags for filters and hybrid search

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet