Document Classification System

An AI-powered web application that automatically classifies home loan application documents using AWS Textract for OCR and AWS Bedrock for machine learning classification.

🎯 Features

• Automatic Document Classification: Classifies 6 types of home loan documents
• AI-Powered OCR: Extracts text using AWS Textract with PyPDF fallback
• Machine Learning: Classification using AWS Bedrock (Claude) with offline fallback
• Web Interface: Modern, responsive UI with drag-and-drop file upload
• Model Evaluation: Comprehensive evaluation dashboard with metrics and visualizations
• Real-time Processing: Background task processing with progress tracking
• Cloud-Ready: Easy deployment to Render cloud platform

📋 Supported Document Types

• 📄 Government ID - Driver's License, Passport, National ID
• 💰 Payslip - Income statements, Pay stubs
• 🏦 Bank Statement - Account statements, Transaction records
• 💼 Employment Letter - Job verification, Employment confirmation
• ⚡ Utility Bill - Electric, Gas, Water, Internet bills
• 🐷 Savings Statement - Investment accounts, Savings records

🚀 Quick Start

Local Development

1. Clone the repository

   git clone <your-repo-url>
   cd document_classification_homeloan

2. Create virtual environment

   python3 -m venv venv_linux
   source venv_linux/bin/activate  # Linux/Mac
   # or
   venv_linux\Scripts\activate     # Windows

3. Install dependencies

   pip install -r requirements.txt

4. Set up environment variables (optional)

   cp .env.example .env
   # Edit .env with your AWS credentials

5. Run the application

   python main.py

6. Access the application

   • Main interface: http://localhost:8000
   • Model evaluation: http://localhost:8000/evaluation-standalone

Cloud Deployment (Render)

1. Push to GitHub
2. Deploy to Render - See DEPLOYMENT.md for detailed instructions
3. Set AWS credentials in Render environment variables (optional)

📊 Model Evaluation

The system includes a comprehensive evaluation interface:

• Batch Testing: Upload multiple labeled documents for testing
• Performance Metrics: Accuracy, Precision, Recall, F1-Score
• Confusion Matrix: Visual classification performance analysis
• Per-Class Metrics: Individual document type performance
• Export Results: Download evaluation reports

Access at: /evaluation-standalone

🔧 Architecture

document_classification_homeloan/
├── src/
│   ├── api/                    # FastAPI routes and endpoints
│   ├── classification/         # Document classification logic
│   ├── evaluation/            # Model evaluation system
│   ├── utils/                 # Configuration and utilities
│   └── web/                   # Web interface
│       ├── templates/         # HTML templates
│       └── static/            # CSS, JS, assets
├── data/                      # Data storage
│   ├── uploads/              # Uploaded files
│   └── results/              # Classification results
├── main.py                   # Application entry point
├── requirements.txt          # Python dependencies
├── render.yaml              # Render deployment config
└── README.md                # This file

🛠️ Technology Stack

• Backend: FastAPI, Python 3.11+
• AI/ML: AWS Textract, AWS Bedrock (Claude), scikit-learn
• Frontend: HTML5, Bootstrap 5, JavaScript
• Data Processing: pandas, numpy
• Visualization: matplotlib, seaborn
• Deployment: Render (native Python)

🔒 Security Features

• File type validation and size limits
• Content-based file validation using python-magic
• Secure environment variable handling
• HTTPS enforcement on deployment
• Input sanitization and validation

🎛️ Configuration

Key environment variables:

# AWS Configuration (optional)
AWS_ACCESS_KEY_ID=your_access_key
AWS_SECRET_ACCESS_KEY=your_secret_key
AWS_REGION=us-east-1

# Application Settings
ENVIRONMENT=production
LOG_LEVEL=INFO
MAX_FILE_SIZE=20971520  # 20MB

🔄 Fallback System

The system gracefully handles AWS service unavailability:

• Text Extraction: Falls back from AWS Textract to PyPDF
• Classification: Falls back from AWS Bedrock to rule-based classifier
• Offline Mode: Fully functional without AWS credentials

📈 Performance

• Processing Time: ~30 seconds per document (with AWS services)
• Accuracy: 95%+ with proper training data
• File Support: PDF documents up to 20MB
• Concurrent Users: Scalable with background task processing

🧪 Testing

The system includes comprehensive testing capabilities:

• Unit tests with pytest
• API endpoint testing
• UI component testing
• End-to-end workflow testing

Run tests:

pytest tests/

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests
5. Submit a pull request

📄 License

This project is licensed under the MIT License.

🙋‍♂️ Support

For deployment issues, see DEPLOYMENT.md or create an issue in the repository.

---

Live Demo: https://document-classification-system.onrender.com