Wikipedia Knowledge Graph RAG System

An intelligent question-answering system powered by knowledge graphs, combining Neo4j database and Wikipedia data for efficient Retrieval-Augmented Generation (RAG).

Core Features

• 🔍 Smart Retrieval - Hybrid vector and graph-based search
• 📚 Knowledge Construction - Automated Wikipedia entity relationship extraction
• 🧠 Context Aware - Multi-turn conversation memory
• 🔄 Real-time Updates - Dynamic Wikipedia data synchronization

Prerequisites

• Python 3.8+
• LangChain 0.3.7
• Neo4j Database
• OpenAI API access

Environment Variables

Create a .env file with:

AURA_INSTANCENAME=your_instance
NEO4J_URI=your_uri
NEO4J_USERNAME=your_username
NEO4J_PASSWORD=your_password
NEO4J_DATABASE=your_database
OPENAI_API_KEY=your_key
OPENAI_ENDPOINT=your_endpoint

Installation

pip install -r requirements.txt

Usage

python wikipedia_kg_rag.py

Example queries:

# Simple query
"How did the Roman empire fall?"

# Query with chat history
"When did he become the first emperor?"

Architecture

graph TD
    A[Wikipedia Data] -->|Extract| B[Data Processor]
    B -->|Store| C[Neo4j Graph]
    B -->|Embed| D[Vector Store]
    E[User Query] -->|Input| F[Query Engine]
    F -->|Search| C
    F -->|Match| D
    C -->|Context| G[Response Generator]
    D -->|Similar Docs| G
    G -->|Generate| H[Final Answer]

Loading