GraphRAG Studio — initial commit: multimodal RAG system with KG visualization

Full-stack application for document-to-knowledge-graph pipeline:
- Backend: FastAPI + LangGraph ReAct agent + DeepSeek + MinerU parsing
- Frontend: React 19 + Vite + D3.js + shadcn/ui
- Pipeline: MinerU parsing → LangExtract entity extraction → KG building

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

backend/CLAUDE.md

@@ -0,0 +1,28 @@
+# Backend — GraphRAG Studio API
+
+## 路径
+
+```
+F:\GraphRAGAgent\backend\
+```
+
+## 启动命令
+
+```bash
+cd F:/GraphRAGAgent/backend
+.venv/Scripts/python.exe -m uvicorn main:app --host 0.0.0.0 --port 8000 --reload
+```
+
+## 接口测试
+
+服务启动后，运行：
+
+```bash
+.venv/Scripts/python.exe tests/test_api.py
+```
+
+## API 文档
+
+- Swagger UI：http://localhost:8000/docs
+- ReDoc：http://localhost:8000/redoc
+- 健康检查：http://localhost:8000/api/v1/health