Workflow Engine

A dynamic workflow execution system with auto-discovery of handlers and a visual workflow builder.

Features

• 🔍 Auto-Discovery: Automatically discover and register handlers from codebase
• 🎯 Type-Safe: Pydantic schemas for validation and IDE support
• 🖥️ Visual Builder: Drag-and-drop workflow builder interface
• 🔄 Dynamic Execution: Execute workflows with context sharing between steps
• 📊 Monitoring: Detailed execution tracking and results
• 🛡️ Secure: Registry pattern prevents arbitrary code execution
• 🚀 Extensible: Easy to add new handlers and action types

Quick Start

1. Install dependencies:

   pip install -r requirements.txt

2. Initialize database:

   python scripts/init_db.py
   python scripts/seed_data.py

3. Start the server:

   python app/main.py

4. Open the workflow builder: Navigate to http://localhost:8000

Architecture

• /handlers/: Handler implementations organized by domain
• /api/: REST API endpoints for CRUD operations
• /core/: Core engine with registry and executor
• /client/: Web-based workflow builder interface

Adding Custom Handlers

1. Create a new handler class inheriting from BaseHandler
2. Use decorators to define metadata and parameters
3. The system will auto-discover and register your handler

Example:

@handler(name="my_handler", category="custom")
class MyHandler(BaseHandler):
    @parameter("input_text", ParameterType.STRING, "Input text to process")
    @output("result", ParameterType.STRING, "Processed result")
    async def execute(self, params):
        # Your logic here
        return ActionResult(success=True, data={"result": "processed"})

API Endpoints

• GET /api/handlers/ - List available handlers
• GET /api/workflows/ - List workflows
• POST /api/workflows/ - Create workflow
• POST /api/workflows/{id}/execute - Execute workflow
• GET /health - Health check

Docker Deployment

docker-compose up -d

License

MIT License