aigrok 0.3.3

A Python package for document processing and analysis with LLM integration and OCR capabilities

AIGrok

AIGrok is an advanced document analysis tool that uses AI to process, analyze, and extract information from documents.

Features

• 📄 Process multiple document formats (PDF, TXT, etc.)
• 🤖 Advanced AI-powered analysis with OpenAI and Ollama support
• 🔍 Smart content extraction with OCR capabilities
• 💾 Result caching for performance
• 🛠️ Extensible provider system
• 📊 Comprehensive reporting
• 📝 Structured output support
• 🔄 Automatic model discovery
• 📈 Code coverage tracking

Quick Start

1. Install AIGrok:

   pip install aigrok
   

2. Set up your configuration:

   aigrok config init
   

3. Process a document:

   aigrok process document.pdf
   

Documentation

For detailed documentation, please see the docs directory:

• CLI Reference - Command-line interface guide
• API Documentation - API reference
• Configuration - Configuration guide
• Deployment - Deployment instructions
• Architecture - System architecture
• Contributing - Contribution guidelines

Requirements

• Python 3.9+ (including 3.12)
• One of:
  • Ollama with vision models installed
  • OpenAI API key for GPT-4 Vision
• 4GB RAM minimum
• Internet connection for API access

Installation

From PyPI

pip install aigrok

From Source

git clone https://github.com/yourusername/aigrok.git
cd aigrok
pip install -e .

Basic Usage

from aigrok import process_document

# Process a document
result = process_document("document.pdf")
print(result.text)

# Process with custom prompt
result = process_document(
    "document.pdf",
    prompt="Extract main topics",
    model="llama2-vision"
)

Contributing

We welcome contributions! Please see our Contributing Guide for details.

License

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

Support

• 📚 Documentation
• 💬 Discussions
• 🐛 Issue Tracker

Acknowledgments

• Thanks to all contributors
• Built with Ollama
• Inspired by the need for better document analysis tools
• 90% of this project was written by AI using Cursor or Windsurf with Anthropic's Claude.