llm_ocr_py: AI Document Extractor Toolkit

A modular Python toolkit for intelligent document processing using OCR and LLMs. Extract structured data from PDFs, Excel files, images, and more with extensible pipelines.

Features

• Modular Architecture: Chain of responsibility pattern for flexible document processing
• Multi-Engine OCR: Support for Tesseract and PaddleOCR with layout preservation
• LLM Integration: Async processing with rate limiting for API calls
• Concurrent Processing: Multi-threaded extraction for high throughput
• Extensible: Easy to add new document types and processing methods

Quick Start

Prerequisites

• Python 3.13+
• Tesseract OCR
• Poppler (for PDF processing)
• PaddleOCR (optional, for advanced OCR tasks)

Installation

1. Clone the repository:

   git clone https://github.com/Prajwal-Prathiksh/llm_ocr_py.git
   cd llm_ocr_py

2. Install dependencies (using uv for best experience):

   uv sync
   source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install system dependencies:

   • Linux: sudo apt-get install tesseract-ocr poppler-utils
   • Windows: Use the provided Setup.ps1 script or install via Scoop

Usage

from src.document_processor import DocumentProcessor

processor = DocumentProcessor()
result = processor.process("path/to/document.pdf")
print(result)

For detailed tutorials, see the tutorials/ directory.

Testing

Run the test suite to ensure everything is set up correctly:

python -m pytest

Project Structure

• src/: Core modules
  • document_processor/: Processing pipelines for different document types
  • llms/: LLM client integrations
  • concurrency_utils.py: Multi-threading utilities
• tests/: Unit tests
• tutorials/: Jupyter notebooks with examples
• assets/: OCR training data and test files

Contributing

Contributions welcome! Please see the tutorials for API details and submit PRs for new features.

License

GPL v3 License - see LICENSE for details.