menna2150/Medical-RAG-system

Medical decision-support RAG app for Egypt using FastAPI, FAISS, BioBERT embeddings, and a React/Vite frontend. # MedRAG-EG — Medical Decision-Support RAG System for Egyptian Doctors > ⚠️ **Decision support only.** Final diagnoses and prescriptions must be made by a licensed physician. > > medical-rag-system-kpy8.ver… A production-ready Retrieval-Augmented Generation (RAG) system that takes patient symptoms and returns ranked differential diagnoses with evidence, recommended tests, treatments, and **medications available in the Egyptian market** (with EGP price ranges). --- ## 1. System Architecture ``` ┌────────────────────────────────────────────────────────────────────────┐ │ REACT FRONTEND (Vite + Tailwind) │ │ SymptomForm → /analyze → DiagnosisCards + MedicationTable │ └──────────────────────────────┬─────────────────────────────────────────┘ │ HTTPS / Axios ▼ ┌────────────────────────────────────────────────────────────────────────┐ │ FASTAPI BACKEND (Python 3.11) │ │ │ │ ┌──────────────┐ ┌──────────────┐ ┌─────────────────────────┐ │ │ │ Query Proc. │ → │ Embedding │ → │ FAISS Vector Retrieval │ │ │ │ (EN/AR norm) │ │ (BioBERT) │ │ (top-k chunks) │ │ │ └──────────────┘ └──────────────┘ └────────────┬────────────┘ │ │ │ │ │ ┌──────────────────────────────────────────────────────▼────────────┐ │ │ │ Reasoner (LLM) -- grounded prompt with retrieved evidence │ │ │ │ → Differential dx (top 3–5) with confidence + reasoning │ │ │ └────────────┬───────────────────────────────────────────────────────┘ │ │ ▼ │ │ ┌──────────────────────┐ ┌──────────────────────────────────────┐ │ │ │ Drug Matcher (EG) │ → │ Safety / Post-processing │ │ │ │ generic→brand+price │ │ con …