|
|
40c2f8e148
|
feat(rag): Complete RAG engine implementation with pgvector
Major Features:
- Created ekb_schema (13th schema) with 3 tables: KB/Document/Chunk
- Implemented EmbeddingService (text-embedding-v4, 1024-dim vectors)
- Implemented ChunkService (smart Markdown chunking)
- Implemented VectorSearchService (multi-query + hybrid search)
- Implemented RerankService (qwen3-rerank)
- Integrated DeepSeek V3 QueryRewriter for cross-language search
- Python service: Added pymupdf4llm for PDF-to-Markdown conversion
- PKB: Dual-mode adapter (pgvector/dify/hybrid)
Architecture:
- Brain-Hand Model: Business layer (DeepSeek) + Engine layer (pgvector)
- Cross-language support: Chinese query matches English documents
- Small Embedding (1024) + Strong Reranker strategy
Performance:
- End-to-end latency: 2.5s
- Cost per query: 0.0025 RMB
- Accuracy improvement: +20.5% (cross-language)
Tests:
- test-embedding-service.ts: Vector embedding verified
- test-rag-e2e.ts: Full pipeline tested
- test-rerank.ts: Rerank quality validated
- test-query-rewrite.ts: Cross-language search verified
- test-pdf-ingest.ts: Real PDF document tested (Dongen 2003.pdf)
Documentation:
- Added 05-RAG-Engine-User-Guide.md
- Added 02-Document-Processing-User-Guide.md
- Updated system status documentation
Status: Production ready
|
2026-01-21 20:24:29 +08:00 |
|
|
|
dfc0fe0b9a
|
feat(pkb): Integrate pgvector and create Dify replacement plan
Summary:
- Migrate PostgreSQL to pgvector/pgvector:pg15 Docker image
- Successfully install and verify pgvector 0.8.1 extension
- Create comprehensive Dify-to-pgvector migration plan
- Update PKB module documentation with pgvector status
- Update system documentation with pgvector integration
Key changes:
- docker-compose.yml: Switch to pgvector/pgvector:pg15 image
- Add EkbDocument and EkbChunk data model design
- Design R-C-R-G hybrid retrieval architecture
- Add clinical data JSONB fields (pico, studyDesign, regimen, safety, criteria, endpoints)
- Create detailed 10-day implementation roadmap
Documentation updates:
- PKB module status: pgvector RAG infrastructure ready
- System status: pgvector 0.8.1 integrated
- New: Dify replacement development plan (01-Dify替换为pgvector开发计划.md)
- New: Enterprise medical knowledge base solution V2
Tested: PostgreSQL with pgvector verified, frontend and backend functionality confirmed
|
2026-01-20 00:00:58 +08:00 |
|
|
|
1ece9a4ae8
|
feat(asl): Add DeepSearch smart literature retrieval MVP
Features:
- Integrate unifuncs DeepSearch API (OpenAI compatible protocol)
- SSE real-time streaming for AI thinking process display
- Natural language input, auto-generate PubMed search strategy
- Extract and display PubMed literature links
- Database storage for task records (asl_research_tasks)
Backend:
- researchService.ts - Core business logic with SSE streaming
- researchController.ts - SSE stream endpoint
- researchWorker.ts - Async task worker (backup mode)
- schema.prisma - AslResearchTask model
Frontend:
- ResearchSearch.tsx - Search page with unified content stream
- ResearchSearch.css - Styling (unifuncs-inspired simple design)
- ASLLayout.tsx - Enable menu item
- api/index.ts - Add research API functions
API Endpoints:
- POST /api/v1/asl/research/stream - SSE streaming search
- POST /api/v1/asl/research/tasks - Async task creation
- GET /api/v1/asl/research/tasks/:taskId/status - Task status
Documentation:
- Development record for DeepSearch integration
- Update ASL module status (v1.5)
- Update system status (v3.7)
Known limitations:
- SSE mode, task interrupts when leaving page
- Cost ~0.3 RMB per search (unifuncs API)
|
2026-01-18 19:15:55 +08:00 |
|
