🧬 GraphRAG Medical Chatbot

Graph-Based Retrieval-Augmented Generation with Neo4j, LangChain & Google Gemini

---

📖 Overview

GraphRAG Medical Chatbot is a Retrieval-Augmented Generation (RAG) system designed to answer complex biological questions related to the BSG gene, associated diseases, proteins, and molecular interactions.

Unlike traditional RAG pipelines that rely exclusively on vector similarity search, this project leverages a Knowledge Graph (Neo4j) to retrieve structured biomedical relationships through dynamically generated Cypher queries.

The system combines:

• 🧠 Large Language Models (Google Gemini)
• 🗄 Graph database reasoning (Neo4j)
• 🔎 Optional hybrid fallback search
• 💬 Conversational memory via LangChain
• 🌐 Interactive UI built with Streamlit

This architecture enables precise, explainable, and relationship-aware answers to biomedical queries.

---

🚀 Key Features

🔗 Graph-Based Retrieval

• Converts natural language questions into Cypher queries
• Retrieves structured biomedical data from Neo4j
• Grounds responses in graph data to reduce hallucinations

🧠 LLM-Powered Answer Generation

• Uses Google Gemini (2.x Flash / Pro models) via langchain-google-genai
• Synthesizes structured graph results into human-readable explanations
• Deterministic Cypher generation (temperature = 0)

💬 Conversational Memory

• Maintains chat history across interactions
• Supports contextual follow-up questions

🌐 Interactive Interface

• Built with Streamlit
• Clean chat-based UI
• Sidebar with example biomedical queries

---

🛠️ Tech Stack

Component	Technology
Language	Python 3.10+
LLM	Google Gemini (via langchain-google-genai)
Graph DB	Neo4j
Orchestration	LangChain Core + Community
Frontend	Streamlit
Environment	python-dotenv

---

⚙️ Installation & Setup

1 Clone the Repository

git clone https://github.com/Ciccio1307/GraphRAG.git
cd GraphRAG

2 Create a Virtual Environment

python3 -m venv venv
source venv/bin/activate   # Windows: venv\Scripts\activate

3 Install Dependencies

pip install -r requirements.txt

4 Setup Environment Variables

Copy the example file and fill in your credentials:

cp .env.example .env

Then open .env and replace the placeholder values with your real credentials:

Variable	Where to get it
GOOGLE_API_KEY / GEMINI_API_KEY	Google AI Studio
NEO4J_URI / NEO4J_USER / NEO4J_PASSWORD	Your local Neo4j instance or AuraDB
GRAPHRAG_*_MODEL	Leave as-is or swap for another Gemini model

Never commit your .env file. It is listed in .gitignore and should stay local only.

▶️ Running the Application

Option A — Quick Start (Mac/Linux)
./start_demo.sh
Option B — Manual Start
streamlit run src/main_3.py