Graph RAG Reporting

This guide demonstrates how to leverage the Perseus client for building a powerful Graph RAG (Retrieval Augmented Generation) reporting application. You'll learn to convert a PDF document into a Markdown file using an LLM, construct a rich knowledge graph from its content, and then use this graph to generate insightful, context-aware reports.

Download the SDK

First, get the Perseus SDK and navigate to the example folder for Graph RAG reporting:

git clone https://github.com/Lettria/perseus-client.git
cd perseus-client/examples/advanced/graph-rag-reporting

This folder contains all the necessary templates and scripts to follow along with this tutorial.

Setup Environment

Ensure you have the required tools installed:

• Docker & Docker Compose 🐳
• Python 3.8+ 🐍

Copy the template environment file and fill in your credentials:

cp template.env .env

Install Dependencies & Start Services

Install Python dependencies

pip install -r requirements.txt

Then, start the necessary Docker services:

docker compose up -d

⏳ The embedder service may take a few minutes to fully boot on first run, as it needs to download the underlying model.

Convert PDF to Markdown

Transform your PDF document into Markdown format for easier processing:

python pdf_to_markdown.py assets/LOREAL_Rapport_Annuel_2024.pdf

This step uses an LLM to extract and structure the text content.

Index the Document

Create a knowledge graph from the Markdown file:

python index.py assets/LOREAL_Rapport_Annuel_2024.md

This step extracts entities and relationships from the document and stores them in a structured graph.

Generate a Report

Query your knowledge graph to generate context-aware reports:

python report.py "What are the main activities of L'Oréal?"

The output is a detailed, structured answer leveraging the graph’s connections.

Cleaning Up 🧹

When you’re done, stop and remove the Docker containers:

docker compose down

This frees up system resources and ensures a clean environment for next time.