Knowledge Graph Experiment

This project builds a multi-level supply chain graph with product reviews for root cause analysis using LangChain and Neo4j.

Setup

Prerequisites

• Python 3.11
• Conda (Miniconda or Anaconda)
• Neo4j database (can be run via Docker)

Installation

1. Create and activate the conda environment:

conda create -n ai python=3.11 -y
conda activate ai

2. Install dependencies:

pip install -r requirements.txt

3. Set up environment variables: Copy the .env file and add your OpenAI API key:

# Edit .env file and add your OpenAI API key
OPENAI_API_KEY=your-actual-api-key-here

4. Start Neo4j (if using Docker):

docker-compose up -d

Running the Script

conda activate ai
python main.py

Features

• User Intent Understanding: Understands knowledge graph requirements
• File Discovery: Automatically discovers and selects relevant files
• Schema Design: Designs schema for structured data (CSV)
• Entity Extraction: Extracts entities from unstructured data (Markdown)
• Graph Construction: Builds the knowledge graph in Neo4j

Configuration

The main configuration options are in the .env file:

• NEO4J_URI: Neo4j connection URI
• NEO4J_USER: Neo4j username
• NEO4J_PASSWORD: Neo4j password
• OPENAI_API_KEY: OpenAI API key
• OPENAI_BASE_URL: (Optional) Custom OpenAI-compatible endpoint
• DATA_DIR: Directory containing data files

Project Structure

kg-experiment/
├── main.py           # Main pipeline script
├── requirements.txt  # Python dependencies
├── .env             # Environment variables
├── docker-compose.yaml # Neo4j Docker setup
├── data/            # Data directory for CSV and Markdown files
└── notebooks/       # Jupyter notebooks for experiments