Project: Graph-RAG Based Agentic System

This project aims to build a knowledge graph from product documentation. Interns will work over a 5-week timeline to parse documentation, ingest it into a Neo4j graph database, identify semantic relationships, and build a retriever to query the graph and expose it through an agentic chat interface.

Project Goals

• Parse Markdown documentation into structured JSON data.
• Construct a hierarchical knowledge graph in Neo4j.
• Enrich the graph with additional semantic links between documentation nodes.
• Develop a Python-based retriever to query the knowledge graph based on user input.
• (Optional) Create a simple Streamlit UI for interacting with the retriever.

Technology Stack

• Programming Language: Python
• Graph Database: Neo4j
• User Interface (Optional): Streamlit
• Documentation Format: Markdown

Getting Started

1. Ensure you have Python (3.10+) and Pip installed.
2. Set up a local Neo4j instance
3. Clone this repository.
4. Refer to timeline.md for task breakdown and details.
5. The product documentation to be processed is located in the /docs directory.

Project Structure

• /docs: Contains the raw product documentation in Markdown format.
• timeline.md: Outlines the project tasks, assignments, and deadlines.
• README.md: This file - provides an overview of the project.