Skip to content

README.md

Latest commit

 

History

History
211 lines (144 loc) · 4.67 KB

README.md

File metadata and controls

211 lines (144 loc) · 4.67 KB

Julius Baer Agentic AI AML Platform - Backend

FastAPI backend for real-time AML monitoring with AI-powered document analysis.

Current Status

✅ Ready to Run - No external dependencies required!

  • Using in-memory mock data (no MongoDB needed yet)
  • Using simulated AI responses (no Groq API needed yet)
  • Fully functional REST API
  • CORS configured for frontend integration

Features

  • Dashboard API: Summary statistics, alerts, transaction volumes
  • Alert Management: Detailed alert information, remediation
  • Transaction Analysis: Volume trends, historical data
  • Audit Trail: Complete activity logging
  • Mock AI Agents: Simulated multi-agent analysis

Quick Start

Prerequisites

  • Python 3.9+ installed
  • pip package manager

Installation

  1. Navigate to the backend directory:
cd backend
  1. Create a virtual environment:
python -m venv venv

# On Windows:
venv\Scripts\activate

# On Mac/Linux:
source venv/bin/activate
  1. Install dependencies:
pip install -r requirements.txt
  1. Run the server:
python main.py

Or using uvicorn directly:

uvicorn main:app --reload --host 0.0.0.0 --port 8000
  1. Open your browser:

API Endpoints

Alerts

  • GET /api/alerts/summary - Dashboard KPIs and statistics
  • GET /api/alerts/active - List of active alerts
  • GET /api/alerts/{alert_id} - Detailed alert information
  • POST /api/alerts/{alert_id}/remediate - Mark alert as remediated

Transactions

  • GET /api/transactions/volume - Transaction volume trend
  • GET /api/transactions/history/{client_id} - Client transaction history

Audit

  • GET /api/audit-trail/{alert_id} - Audit trail for specific alert

System

  • GET / - API information
  • GET /health - Health check

Project Structure

backend/
├── main.py                      # FastAPI application entry point
├── config.py                    # Configuration settings
├── requirements.txt             # Python dependencies
├── models/
│   └── schemas.py              # Pydantic models
├── services/
│   ├── database.py             # Database service (mock)
│   └── mock_data.py            # Mock data for development
└── api/
    └── routes/
        ├── alerts.py           # Alert endpoints
        ├── transactions.py     # Transaction endpoints
        └── audit.py            # Audit trail endpoints

Testing the API

Using curl

# Get dashboard summary
curl http://localhost:8000/api/alerts/summary

# Get active alerts
curl http://localhost:8000/api/alerts/active

# Get specific alert details
curl http://localhost:8000/api/alerts/ALT-788

# Get transaction volume
curl http://localhost:8000/api/transactions/volume

Using the Interactive Docs

Visit http://localhost:8000/docs to use the built-in Swagger UI for testing all endpoints.

Connecting to Frontend

The backend is configured to accept requests from http://localhost:3000 by default.

To test with the frontend:

  1. Start the backend: python main.py (runs on port 8000)
  2. Start the frontend: cd frontend && npm run dev (runs on port 3000)
  3. Frontend will automatically connect to the backend API

Next Steps: Adding Real Services

1. MongoDB Integration

When ready to add MongoDB:

# Install MongoDB dependencies
pip install motor pymongo

# Update .env file
MONGODB_URL=mongodb://localhost:27017
DATABASE_NAME=julius_baer_aml

Then uncomment the MongoDB code in services/database.py

2. Groq API Integration

When ready to add Groq AI:

# Install Groq
pip install groq

# Update .env file
GROQ_API_KEY=your_actual_groq_api_key

Create agent files in agents/ directory for real AI analysis.

3. WebSocket Support

For real-time alerts:

pip install websockets

Add WebSocket endpoint in main.py for live updates.

Environment Variables

Create a .env file (optional for now):

CORS_ORIGINS=http://localhost:3000

Development Tips

  • The API auto-reloads when you make changes (using --reload flag)
  • Check logs in the terminal for debugging
  • Use /docs endpoint to explore and test all APIs
  • Mock data is defined in services/mock_data.py - customize as needed

Production Deployment

When ready for production:

  1. Add MongoDB connection
  2. Add Groq API integration
  3. Add authentication/authorization
  4. Configure environment variables
  5. Use production ASGI server (Gunicorn + Uvicorn)
  6. Set up proper logging and monitoring

License

Proprietary - Julius Baer