Agentic Knowledge Graph Workshop

This workshop will cover the basics of knowledge graph (KG) construction using a multi-agent workflow.

Prerequisites

• Python 3.12

Setup

1. Install python libraries in a local virtual environment:

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

2. Set up Neo4j

2.1 Neo4j Desktop on your local machine

Install Neo4j Desktop:

1. Download and install Neo4j Desktop from https://neo4j.com/download-center/
2. Open Neo4j Desktop and create a new instance (a graph database instance)
   • remember to use the same password you set in the .env file!

Install required plugins:

• install "APOC" for standard procedures and functions
• install "GenAI" for AI capabilities

Edit neo4j.conf (for older versions of Desktop):

dbms.security.procedures.unrestricted=apoc.*
dbms.security.procedures.allowlist=apoc.*,genai.*

Check the procedures are available:

SHOW PROCEDURES WHERE name CONTAINS "genai.vector"

2.2 Neo4j Aura in the cloud

Sign-up for Neo4j Aura:

1. Visit: http://console.neo4j.io
2. Click the "Create instance" button on the "Instances" page
3. Optional: name the instance
4. Click "Create instance" on the bottom of the page
5. Copy the password someplace safe (for use in a .env file later)
6. Click "Download and continue"
7. Wait for the instance to be provisioned

3. Create a copy of the environment variables file:

cp .env.template .env

Make two changes:

• Replace NEO4J_PASSWORD with the password for your database
• Replace OPENAI_API_KEY with the OpenAI API key provided for the workshop
• Replace NEO4J_IMPORT_DIR with the location of the data directory of the cloned repository