Browser AI VNC - Shipping Line Tracking Automation

A comprehensive Docker-based solution for AI-driven shipping line tracking automation with built-in evaluation capabilities.

🎯 Overview

This project provides two main components:

1. Automation Container: A browser automation environment with VNC access and REST API
2. Evaluation Container: Automated testing and evaluation of the automation container

🚀 Quick Start

Run Full Evaluation Suite

# Run both automation and evaluation containers
docker compose up --build

# The evaluation container will automatically test the automation container

# Access points during testing:
# - VNC Viewer: localhost:5900 (password: secret)
# - Web VNC: http://localhost:6080
# - API: http://localhost:5001 (configurable via API_PORT env var)

📁 Project Structure

browseraivnc/
├── src/                        # Automation scripts
│   ├── api.py                  # Flask API for automation
│   └── main-simple-test.py     # Visual automation demo
├── evaluator/                  # Evaluation container
│   ├── src/
│   │   └── evaluator.py        # Simple evaluation script
│   ├── Dockerfile             # Evaluation container image
│   └── requirements.txt       # Evaluation dependencies
├── docker-compose.yml          # Full evaluation setup
└── README.md                   # This file

🤖 Automation Container

Features

• VNC Access: Visual browser automation with remote desktop
• REST API: Programmatic access to automation functions
• Caching: Intelligent in-memory caching for repeated requests
• Multi-Browser: Support for Chrome, Firefox, and Chromium

API Endpoints

Health Check

GET /health

Track Shipping

POST /track
Content-Type: application/json

{
  "booking_id": "SINI25432400",
  "force_fresh": false
}

Example API Usage

import requests

# Health check
response = requests.get("http://localhost:5001/health")
print(response.json())

# Track a booking
payload = {"booking_id": "SINI25432400"}
response = requests.post("http://localhost:5001/track", json=payload)
print(response.json())

🧪 Evaluation Container

Purpose

The evaluation container automatically tests the automation container to ensure:

• ✅ Tracking returns correct voyage number and arrival date
• ✅ Caching improves performance on repeat requests
• ✅ System works reliably

Evaluation Tests

1. Fresh Tracking Test: Tests initial tracking request and verifies correct data (YM MANDATE 0096W, 2025-02-28)
2. Cached Tracking Test: Tests repeat request is faster and uses cache

Running Evaluation

# Run full evaluation suite
docker compose up --build

# View evaluation logs
docker logs evaluation_container

Evaluation Report Example

🧪 Starting Simple Evaluation
========================================
⏳ Waiting for API to be ready...
✅ API is ready

🚢 Testing Fresh Request...
✅ Fresh request passed (12.3s)

💾 Testing Cached Request...
✅ Cached request passed (2.1s)

📊 Results: 2/2 tests passed
🎉 All tests passed!

🔧 Configuration

Environment Variables

Automation Container

• DISPLAY: X11 display (default: :99)
• SCREEN_WIDTH: Screen width (default: 1920)
• SCREEN_HEIGHT: Screen height (default: 1080)
• BROWSER: Browser choice (chrome/firefox/chromium)

Evaluation Container

• AUTOMATION_API_URL: API URL (default: http://automation:5000)
• PYTHONUNBUFFERED: Python output buffering (default: 1)

Volume Mounts

None - all evaluation output is shown in console logs

🐳 Docker Commands

Build and Run

# Build automation container
docker compose build

# Run automation container
docker compose up -d

# Run evaluation suite
docker compose up --build

# Stop all containers
docker compose down

Debugging

# View automation container logs
docker logs automation_container

# View evaluation container logs
docker logs evaluation_container

# Execute commands in running container
docker exec -it automation_container bash

📋 Assignment Context

This project was developed for the AI-based Shipping Line Tracking Assignment with the following requirements:

Assignment Goals

1. Natural Language Processing: Use AI to interpret tracking requests
2. Process Persistence: Cache and reuse automation steps
3. Adaptability: Handle different booking IDs without hardcoding
4. Repeatability: Consistent results across multiple runs

Sample Booking ID

• SINI25432400 (Test case with voyage: YM MANDATE 0096W, arrival: 2025-02-28)

Expected Outputs

• Voyage Number: "YM MANDATE 0096W"
• Arrival Date: "2025-02-28"
• Execution Time: Performance metrics
• Cache Status: Whether cached data was used

🔍 Troubleshooting

Common Issues

1. Container won't start

   # Check port conflicts
   docker ps -a
   
   # Clean up containers
   docker system prune -f

2. VNC connection fails

   # Check VNC logs
   docker logs automation_container | grep vnc
   
   # Restart container
   docker-compose restart

3. API not responding

   # Check API health
   curl http://localhost:5001/health
   
   # Check container logs
   docker logs automation_container | grep Flask

4. Evaluation fails

   # Check evaluation logs
   docker logs evaluation_container
   
   # Verify network connectivity
   docker network ls

📊 Metrics and Monitoring

Monitoring Commands

# Monitor container resources
docker stats

# Check disk usage
docker system df

# Monitor API requests
docker logs automation_container | grep -E "(GET|POST)"