OntoSource

An web-based platform for ontology engineering and knowledge graph exploration - intuitive editing, visualization, and reasoning, powered by OWLAPY.

Prerequisites

• Conda (Anaconda or Miniconda)
• Node.js 18+
• npm or pnpm

Quick Start

Run the entire project with a single command:

./start.sh

This script will:

• Create the conda environment if it doesn't exist
• Install all dependencies
• Start both backend and frontend services
• Display all service URLs

Press Ctrl+C to stop all services.

Manual Setup

Backend (FastAPI)

# Create conda environment with Python 3.13.7
conda create -n ontosource python=3.13.7 -y

# Activate the environment
conda activate ontosource

# Install Python dependencies
pip install -r requirements.txt

# Navigate to services directory
cd services

# Run the API server
uvicorn main:app --reload

The API will be available at http://localhost:8000. API documentation at http://localhost:8000/docs.

Frontend (Next.js)

# Navigate to webapp directory
cd webapp

# Install dependencies
npm install
# or
pnpm install

# Run development server
npm run dev
# or
pnpm dev

The webapp will be available at http://localhost:3000.

Features

• Ontology Generation: AI-powered ontology creation from natural language descriptions
• Import & Export: Support for OWL/RDF formats (.owl, .rdf, .xml, .ttl, .n3, .nt) -️ Interactive Editing: Full CRUD operations for classes, properties, individuals, and axioms
• Knowledge Graph Visualization: Interactive graph rendering with react-flow
• Classical Reasoning: Inference of instances, sub/super classes, and entity types
• Neural Reasoning: Embedding-based reasoning (EBR)
• Multi-Ontology Management: Upload, store, and switch between multiple ontologies
• Entity Explorer: Browse and analyze ontology entities with detailed summaries

API Endpoints

• /generate - AI-powered ontology generation from text
• /ontology - Upload, list, and manage ontologies
• /modify - Create and edit ontology entities and axioms
• /reasoning - Perform logical reasoning (instances, sub/super classes, types)
• /neural-reasoning - Neural embedding-based reasoning

For full API documentation, visit http://localhost:8000/docs when running the backend.