🧾 Invoice Tracker

A sophisticated AI-powered invoice tracking system with OCR capabilities, intelligent categorization, and advanced RAG (Retrieval-Augmented Generation) analysis.

✨ Features

• 📄 OCR Processing: Extract data from invoice images using Google Vision API
• 🤖 AI Categorization: Automatic invoice categorization using Gemini AI
• 🔍 Advanced RAG: Chain-of-thought reasoning for intelligent spending analysis
• 📊 Vector Search: ChromaDB integration for semantic invoice search
• 🎨 Modern UI: Clean Gradio interface with multiple tabs
• 🔄 Duplicate Detection: Smart duplicate removal with Hebrew/multilingual support
• 📈 Analytics: Comprehensive spending insights and trends
• 💾 Database: Persistent storage with ChromaDB vector database

🛠️ Technology Stack

• Backend: Python 3.8+
• AI/ML: Google Gemini, Google Vision API
• Database: ChromaDB (Vector Database)
• Frontend: Gradio
• OCR: Google Cloud Vision API
• Environment: python-dotenv

📦 Installation

1. Clone the repository

   git clone https://github.com/yourusername/invoice-tracker.git
   cd invoice-tracker

2. Install dependencies

   pip install -r requirements.txt

3. Set up environment variables Create a .env file in the src folder:

   GEMINI_API_KEY=your_gemini_api_key_here
   GOOGLE_CLOUD_PROJECT=your_google_cloud_project_id

4. Get API Keys

   • Gemini API: Get from Google AI Studio
   • Google Cloud Vision: Set up at Google Cloud Console

🚀 Usage

1. Start the application

   cd src
   python main.py

2. Open your browser to the displayed URL (usually http://localhost:7860)

3. Upload invoices via the "📄 Upload Invoices" tab

4. Ask questions using the "🤖 AI Insights" tab with natural language:

   • "What's my total spending this month?"
   • "Which vendor do I spend the most on?"
   • "Show me my grocery expenses"

📁 Project Structure

invoice-tracker/
├── src/
│   ├── main.py                 # Application entry point
│   ├── services/
│   │   ├── ocr_service.py      # OCR processing
│   │   ├── categorization_service.py  # AI categorization
│   │   ├── rag_service.py      # RAG with chain-of-thought
│   │   └── chroma_service.py   # ChromaDB operations
│   ├── ui/
│   │   └── gradio_interface.py # Web interface
│   ├── utils/
│   │   ├── invoice_processor.py # Invoice processing
│   │   └── cleanup_duplicates.py # Duplicate detection
│   ├── data/
│   │   ├── invoices.json       # Invoice storage
│   │   └── chroma_db/          # Vector database
│   └── .env                    # Environment variables (not in repo)
├── requirements.txt            # Dependencies
├── .gitignore                 # Git ignore rules
└── README.md                  # This file

🔧 Features Overview

📄 OCR Processing

• Extracts text from invoice images
• Identifies vendors, amounts, dates, invoice numbers
• Supports multiple image formats (PNG, JPG, PDF)

🤖 AI Categorization

• Automatic spending category assignment
• Smart vendor recognition
• Context-aware categorization

🔍 Advanced RAG

• Chain-of-Thought Reasoning: Structured 4-step analysis
• Vector Similarity Search: Find relevant invoices using ChromaDB
• Enhanced Context: Combines statistics with query-specific data
• Natural Language Queries: Ask questions in plain English

📊 Analytics

• Total spending analysis
• Category breakdowns with percentages
• Monthly/yearly trends
• Top vendor identification
• Spending pattern insights

🔄 Duplicate Detection

• Multi-criteria duplicate detection
• Hebrew/multilingual support
• Semantic vendor matching
• Date format normalization
• Invoice number matching

🧠 Chain-of-Thought RAG Process

The RAG system follows a structured reasoning approach:

1. Understanding the Question

   • Analyze what information is needed
   • Identify relevant time periods and categories

2. Data Analysis Approach

   • Select relevant data points
   • Plan necessary calculations

3. Calculations & Insights

   • Perform step-by-step analysis
   • Identify patterns and anomalies

4. Final Answer

   • Provide clear, actionable insights
   • Include specific numbers and recommendations

🛠️ API Configuration

Gemini API Setup

1. Go to Google AI Studio
2. Create a new API key
3. Add to .env file as GEMINI_API_KEY

Google Cloud Vision Setup

1. Create project at Google Cloud Console
2. Enable Vision API
3. Create service account and download JSON key
4. Set GOOGLE_APPLICATION_CREDENTIALS environment variable

📝 Example Queries

• Total Spending: "What's my total spending this year?"
• Category Analysis: "How much do I spend on groceries vs restaurants?"
• Vendor Analysis: "Which vendor do I spend the most money on?"
• Trends: "Show me my spending trends over the last 6 months"
• Specific Items: "Find all invoices for technology purchases"

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Google AI for Gemini API
• ChromaDB for vector database capabilities
• Gradio for the web interface
• Google Cloud Vision for OCR capabilities

🐛 Troubleshooting

Common Issues

1. RAG not working: Check GEMINI_API_KEY in .env file
2. OCR failing: Verify Google Cloud credentials
3. ChromaDB errors: Ensure chromadb package is installed
4. Import errors: Check Python path and package installations

Support

If you encounter issues:

1. Check the console output for error messages
2. Verify all environment variables are set correctly
3. Ensure all dependencies are installed
4. Check API key permissions and quotas