An AI-powered web application that helps insurance professionals generate professional, high-quality insurance reports in Italian. The application leverages Google's Gemini 2.5 model to process uploaded documents and produce structured reports that can be reviewed, edited, and downloaded as DOCX files.

• Multi-Format Document Support: Upload images (PNG, JPEG), PDFs, DOCX, XLSX, TXT, and EML files
• Intelligent Document Processing: Automatic OCR for images, text extraction from documents, and data extraction from spreadsheets
• AI-Powered Report Generation: Uses Gemini 2.5 to generate professional Italian insurance reports based on uploaded documents
• Report Preview & Editing: Clean web interface to preview and edit generated reports before download
• Professional DOCX Export: Download reports as formatted Word documents with proper styling and structure
• Admin Panel: Administrative interface for system monitoring, report management, and configuration
• Rate Limiting: Built-in protection against abuse with configurable rate limits
• Authentication: Basic HTTP authentication for admin access

• Cargo Surveyors
• Claims Handlers
• Insurance Surveyors
• Legal professionals working with insurance claims
• Any professional requiring standardized insurance reports in Italian

• Python 3.10 or higher
• Google Gemini API key
• (Optional) PostgreSQL database for production deployments

1. Clone the repository:

   git clone https://github.com/andreasalomone/perito.git cd report-ai

2. Create a virtual environment:

   python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

4. Set up environment variables: Create a .env file in the root directory with the following variables:

   GEMINI_API_KEY=your_gemini_api_key_here FLASK_SECRET_KEY=your_secret_key_here BASIC_AUTH_USERNAME=admin BASIC_AUTH_PASSWORD=your_secure_password DATABASE_URL=sqlite:///instance/project.db # Or PostgreSQL URL for production LOG_LEVEL=INFO PORT=5000

Configuration is managed through environment variables and the core/config.py file. Key settings include:

• File Upload Limits:

  • Max file size: 25 MB (configurable via MAX_FILE_SIZE_MB)
  • Max total upload size: 100 MB (configurable via MAX_TOTAL_UPLOAD_SIZE_MB)

• LLM Settings:

  • Model: gemini-2.5-pro
  • Temperature: 0.5
  • Max tokens: 64000
  • Retry attempts: 3

• DOCX Generation:

  • Font: Times New Roman
  • Normal font size: 12pt
  • Heading font size: 12pt
  • Line spacing: 1.5

See core/config.py for all available configuration options.

python app.py

The application will be available at http://localhost:5000

python run_server.py

Initialize the database schema:

flask init-db

1. Upload Documents: Navigate to the home page and upload one or more documents related to your insurance claim
2. Generate Report: The AI will process your documents and generate a draft report in Italian
3. Review & Edit: Review the generated report and make any necessary edits in the preview interface
4. Download: Download the final report as a DOCX file

Access the admin panel at /admin/login using your configured credentials. The admin panel provides:

• System monitoring and statistics
• Report history and details
• AI control and configuration
• Template management

report-ai/ ├── admin/ # Admin panel routes, templates, and services ├── assets/ # Static assets (logos, favicons) ├── core/ # Core configuration, database, and models ├── docs/ # Documentation and coding guidelines ├── instance/ # Database files (SQLite) ├── scripts/ # Utility scripts ├── static/ # CSS and static files ├── templates/ # HTML templates ├── tests/ # Test suite ├── uploads/ # User-uploaded files (temporary) ├── app.py # Main Flask application ├── document_processor.py # Document processing logic ├── docx_generator.py # DOCX generation logic ├── llm_handler.py # LLM integration and prompt management ├── run_server.py # Production server entry point └── requirements.txt # Python dependencies 

This project follows strict coding guidelines (see docs/rules/):

• SOLID Principles: Single Responsibility, Open/Closed, Liskov Substitution, Interface Segregation, Dependency Inversion
• KISS: Keep It Simple, Stupid
• DRY: Don't Repeat Yourself
• YAGNI: You Ain't Gonna Need It
• Python: PEP 8 compliance, type hints, docstrings
• Testing: Unit tests, integration tests, and E2E tests

pytest

Run with coverage:

pytest --cov=. --cov-report=html

Install development dependencies:

pip install -r requirements-dev.txt

• Backend: Flask (async), Hypercorn
• AI/ML: Google Gemini 2.5 API
• Database: SQLite (development) / PostgreSQL (production)
• Document Processing: PyMuPDF, python-docx, openpyxl, Pillow
• Authentication: Flask-HTTPAuth
• Rate Limiting: Flask-Limiter
• Configuration: Pydantic Settings

• GET / - Home page with document upload interface
• POST /upload - Upload documents for processing
• POST /generate - Generate report from uploaded documents
• GET /report/<report_id> - View generated report
• POST /report/<report_id>/edit - Edit report content
• GET /download/<report_id> - Download report as DOCX
• /admin/* - Admin panel routes

[Add your license information here]

[Add contribution guidelines here]

For issues, questions, or feature requests, please create an issue on GitHub.

Built with Google Gemini 2.5 for intelligent report generation.