graphaide 0.4.4

Multi-agentic knowledge graph construction and reasoning system

GraphAide

GraphAide is an advanced multi-agentic query and reasoning system that constructs knowledge graphs (KG) from diverse sources and enables querying and reasoning over the resulting graphs. It harnesses both knowledge graphs and LLMs to rapidly develop domain-specific digital assistants.

Technical Summary

GraphAide v2 is a multi-agent system that uses LangGraph for workflow orchestration, LLMs for entity/relationship extraction, Neo4j for graph storage, and ChromaDB for vector embeddings.

The codebase follows a factory pattern: ModelFactory, GraphDBFactory, VectorStoreFactory, and AgentFactory create and inject dependencies into specialized agents. Agents communicate through shared state objects (KGGenerationState) that track nodes, edges, and messages as data flows through the pipeline.

Project Structure:

• src/graphgen/v2/ - Core modules (factories, state, workflows)
• src/graphgen/v2/agents/ - Agent implementations (extractor, loaders, visualizers)
• templates/ - Prompt templates for each agent
• tests/notebooks/ - Demo notebooks and test data

Tested Models:

Provider	LLM	Embedding
LMStudio (local)	google/gemma-3-27b:2	text-embedding-nomic-embed-text-v1.5
AWS Bedrock	anthropic.claude-3-5-sonnet-20241022-v2:0	amazon.titan-embed-text-v1
OpenAI	gpt-4o	text-embedding-3-small

Quick Start

📚 Documentation:

• QUICKSTART.md - Installation, setup, and first examples
• ARCHITECTURE.md - Deep technical overview (factories, agents, state, workflows)
• REFERENCE.md - Complete CLI/API reference and feature matrix
• CLAUDE.md - Guide for Claude AI Code (developers only)

Installation

# From project root
pip install .

# With Docker
docker compose up -d

Deployment Scripts

Linux/Mac/WSL

./graphaide_docker_build.sh 2.0.1              # Build production + dev images
./start-graphaide-prod.sh                      # Start production services
./start-graphaide-dev.sh                       # Start development services

Windows (PowerShell)

.\scripts\windows\graphaide_docker_build.ps1 2.0.1        # Build images
.\scripts\windows\start-graphaide-prod.ps1                # Start production
.\scripts\windows\start-graphaide-dev.ps1                 # Start development

Note: For Windows users, we recommend using WSL (Windows Subsystem for Linux) with the bash scripts above. Native PowerShell scripts are available in scripts/windows/ for reference.

Usage

Ways to Run GraphAide:

Mode	Description
CLI (Host)	graphaide extract, graphaide query, etc.
CLI (Docker)	Run inside container via docker exec
Python API	Import WorkflowManager for programmatic access
REST API	FastAPI endpoints at http://localhost:8000

Getting Started:

• Start with QUICKSTART.md for 5-minute setup
• Run graphaide extract document.pdf to test locally
• See tests/notebooks/demoGraphAide.ipynb for advanced examples

Support

Please reach out to [e-mail] for any questions.

Authors and acknowledgment

Please reach out to [e-mail] for any questions.

Citation

If you use GraphAide in your research, please cite:

@inproceedings{purohit2024graphaide,
  title={GraphAide: Advanced Graph-Assisted Query and Reasoning System},
  author={Purohit, Sumit and Chin, George and Mackey, Patrick S and Cottam, Joseph A},
  booktitle={2024 IEEE International Conference on Big Data (BigData)},
  pages={3485--3493},
  year={2024},
  organization={IEEE}
}

The research described in this paper is partially supported by the Resilience Through Data Driven, Intelligently Designed Control (RD2C) Initiative at Pacific Northwest National Laboratory (PNNL) and the United States federal government. Pacific Northwest National Laboratory is a multiprogram national laboratory operated for the US Department of Energy (DOE) by Battelle Memorial Institute under Contract No. DEAC05-76RL01830. PNNL Information Release PNNL-SA205147.

License

Please refer LICENSE and DISCLAIMER files for details.

Project status

Active development