Skip to content

readme.md

Latest commit

 

History

History
40 lines (25 loc) · 1.45 KB

readme.md

File metadata and controls

40 lines (25 loc) · 1.45 KB

Neo4j Knowledge Graph API

This directory contains examples for hosting and querying knowledge graphs stored in Neo4j databases via OpenAI-compatible API endpoints.

Files

  • atlas_api_server_demo.py: General-purpose API server demo that sets up an OpenAI-compatible endpoint for querying a Neo4j knowledge graph with LLM-powered retrieval.

  • atlas_api_server_demo_wiki.py: API server demo specifically configured for Wikipedia knowledge graph hosted in Neo4j.

  • atlas_api_server_demo_pes2o.py: API server demo specifically configured for PES2O (Papers with Entities, Subjects, and Ontologies) knowledge graph.

  • atlas_api_server_demo_cc.py: API server demo specifically configured for CommonCrawl knowledge graph with customized retrieval settings.

  • atlas_api_client_demo.py: Client example that demonstrates how to query multiple knowledge graph API servers (Wiki, PES2O, CC) using the OpenAI client interface.

Quick Start

Server Setup

  1. Configure your Neo4j connection and API keys in config.ini
  2. Start a knowledge graph API server:
python atlas_api_server_demo_wiki.py    # For Wikipedia KG
# or
python atlas_api_server_demo_cc.py      # For CommonCrawl KG
# or
python atlas_api_server_demo_pes2o.py   # For PES2O KG

Client Usage

Query the running API servers:

python atlas_api_client_demo.py

The client will send queries to all configured knowledge graph servers and return enriched responses.