This repo contains a streamlit app for introducing and teaching example GraphRAG patterns.
This sample app uses the classic Northwind database. We will provide a script for loading the data from raw source and setting up the text embeddings and vector index. Until then you can use neo4j.dump
file to automatically setup the database with vectors and all.
-
Create a
secrets.toml
file usingsecrets.toml.example
as a templatecp .streamlit/secrets.toml.example .streamlit/secrets.toml vi .streamlit/secrets.toml
fill in the below credentials. You can use a blank sandbox or any other Neo4j instance.
# OpenAI OPENAI_API_KEY = "sk-..." # NEO4J NEO4J_URI = "neo4j+s://<xxxxx>.databases.neo4j.io" NEO4J_USERNAME = "neo4j" NEO4J_PASSWORD = "<password>"
-
Install requirements (recommended in an isolated python virtual environment):
pip install -r requirements.txt
-
Run app:
streamlit run Home.py --server.port=80