Local LLM Chat for OpenMetadata

A VS Code extension that brings AI-powered data discovery to your IDE using your choice of LLM provider. Chat with your OpenMetadata catalog using OpenAI, Ollama, or any custom endpoint.

Key Features

• Multiple LLM Providers: Choose between OpenAI, Ollama, or custom OpenAI-compatible endpoints
• Complete Privacy: Use local models with Ollama for offline, private data analysis
• Flexible Configuration: Switch between providers without code changes
• Natural Language Search: Ask questions like "show me customer tables" or search by keywords
• AI-Powered Insights: Get intelligent analysis of your datasets and data quality
• Interactive Data Lineage: Visualize upstream and downstream table relationships
• Column Details: Explore table schemas with expandable column information

Requirements

1. OpenMetadata Server: Running locally at http://localhost:8585

   • Use the OpenMetadata Docker setup
   • Load sample data for testing

2. LLM Provider (choose one):

   • OpenAI: Get an API key from platform.openai.com
   • Ollama: Install Ollama and pull a model (e.g., ollama pull llama2)
   • Custom: Any OpenAI-compatible API endpoint

Installation & Setup

Option 1: Build from Source

1. Clone and Install

cd /path/to/your/workspace
git clone <your-repo-url> local-llm-chat-vscode-openmetadata
cd local-llm-chat-vscode-openmetadata
npm install

2. Configure Your LLM Provider

Open VS Code settings (Ctrl+, or Cmd+,) and configure your preferred provider:

For OpenAI:

{
  "openmetadataExplorer.llm.provider": "openai",
  "openmetadataExplorer.llm.openai.apiKey": "sk-your-api-key-here",
  "openmetadataExplorer.llm.openai.model": "gpt-4o",
  "openmetadataExplorer.openmetadataUrl": "http://localhost:8585",
  "openmetadataExplorer.openmetadataAuthToken": "YOUR_BOT_TOKEN"
}

For Ollama:

{
  "openmetadataExplorer.llm.provider": "ollama",
  "openmetadataExplorer.llm.ollama.endpoint": "http://localhost:11434",
  "openmetadataExplorer.llm.ollama.model": "llama2",
  "openmetadataExplorer.openmetadataUrl": "http://localhost:8585",
  "openmetadataExplorer.openmetadataAuthToken": "YOUR_BOT_TOKEN"
}

For Custom Endpoint (e.g., LM Studio, LocalAI):

{
  "openmetadataExplorer.llm.provider": "custom",
  "openmetadataExplorer.llm.custom.endpoint": "http://localhost:1234/v1",
  "openmetadataExplorer.llm.custom.apiKey": "optional-api-key",
  "openmetadataExplorer.llm.custom.model": "model-name",
  "openmetadataExplorer.openmetadataUrl": "http://localhost:8585",
  "openmetadataExplorer.openmetadataAuthToken": "YOUR_BOT_TOKEN"
}

3. Get OpenMetadata Bot Token

1. Open http://localhost:8585 and login (default: admin/admin)
2. Go to Settings → Bots
3. Click Add Bot with these details:
   • Name: vscode-llm-bot
   • Description: Bot for VS Code LLM extension
4. Click Generate Token and copy the JWT token (starts with eyJ)
5. Assign Data Consumer role to the bot

4. Run in Debug Mode

1. Press F5 to launch the extension in a new VS Code window
2. Look for OPEN METADATA panel at the bottom
3. Verify connection and start searching

Option 2: Package and Install

1. Build the Extension

npm run compile
npm run package

2. Install the VSIX

• Open Command Palette (Ctrl+Shift+P or Cmd+Shift+P)
• Type: Extensions: Install from VSIX...
• Select the generated .vsix file
• Reload when prompted

3. Configure as shown in Option 1, step 2

How to Use

Search Your Data

• Keyword Search: Type table names like "customer" or "orders"
• Natural Language: Ask questions like "show me customer data"
• Browse Results: Click on tables to see column details and AI insights

View Data Lineage

1. Search for any table
2. Click View Lineage on the table card
3. Use the interactive graph:
   • Click + buttons to expand upstream/downstream relationships
   • Click - buttons to collapse connections
   • Drag nodes to reposition them
   • Zoom with mouse wheel

Example Searches

• customer - Find customer-related tables
• orders - Discover transaction data
• sales - Locate revenue tables
• product - Find catalog information

LLM Provider Comparison

Recommended Models

OpenAI:

• gpt-4o - Best quality, faster than GPT-4
• gpt-4 - High quality, slower
• gpt-3.5-turbo - Fast and cheap
• o1-preview - Advanced reasoning
• o1-mini - Cost-effective reasoning

Ollama:

• llama2 - General purpose, 7B parameters
• mistral - Fast and capable
• codellama - Optimized for code/data
• llama3 - Latest Llama model

Custom:

• Any OpenAI-compatible API endpoint works
• Examples: LM Studio, LocalAI, vLLM, Text Generation WebUI

Development

Build Commands

# Development build with watch
npm run watch

# Production build
npm run compile

# Package for distribution
npm run package

Project Structure

src/
├── extension.ts                 # Extension entry point
├── OpenMetadataExplorerProvider.ts  # Main provider
├── services/
│   ├── OpenAIService.ts        # OpenAI integration
│   ├── LocalLLMService.ts      # Ollama/Custom integration
│   ├── UnifiedLLMService.ts    # LLM orchestrator
│   ├── OpenMetadataService.ts  # OpenMetadata API
│   └── LineageService.ts       # Data lineage
└── webview/
    ├── App.tsx                  # Main React app
    └── components/              # React components

Key Service Files

UnifiedLLMService.ts

• Orchestrates between OpenAI, Ollama, and custom endpoints
• Handles provider selection and initialization
• Provides unified interface for all LLM operations

OpenAIService.ts

• Direct integration with OpenAI API
• Supports chat completions format
• Handles API key validation

LocalLLMService.ts

• OpenAI-compatible format for Ollama
• Fallback to legacy formats for compatibility
• Flexible response parsing

Configuration Reference

OpenMetadata Settings

LLM Provider Settings

OpenAI Settings

Ollama Settings

Custom Endpoint Settings

Troubleshooting

OpenAI Issues

• "API key invalid": Check your API key at platform.openai.com
• "Rate limit exceeded": Wait a moment or upgrade your OpenAI plan
• "Model not found": Verify the model name (gpt-4o, gpt-4, etc.)

Ollama Issues

• "Connection refused": Make sure Ollama is running (ollama serve)
• "Model not found": Pull the model first (ollama pull llama2)
• Slow responses: Use a smaller model or upgrade your hardware

OpenMetadata Issues

• "Connection failed": Verify OpenMetadata is running at the configured URL
• "Authentication failed": Check your bot token is valid and not expired
• "No tables found": Ensure sample data is loaded in OpenMetadata

Current Status

Version 1.0.0 - Initial Release

• OpenAI, Ollama, and custom endpoint support
• Natural language search with AI insights
• Interactive data lineage visualization
• Professional UI optimized for developers

Planned Features

• Column-level lineage relationships
• Data quality monitoring integration
• Advanced search filters and exports
• Streaming responses for better UX
• Multi-turn conversations

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Thanks to the Ollama team for making local LLMs accessible
• LM Studio for providing an excellent local inference platform
• The VS Code extension API team for comprehensive documentation

🙋‍♂️ Get Involved

If you encounter any issues or have questions:

• 🐛 Report bugs
• 💡 Request features
• ⭐ Star the repo if you find it useful!

☕ Support the Project

If you like this project, support further development with a repost or coffee:

📬 Contact

• 🧑‍💻 Markus Begerow
• 💾 GitHub
• ✉️ Twitter

Privacy Notice: This extension operates entirely locally. No data is sent to external servers unless you explicitly configure it to use a remote API endpoint.