ocrguru 0.1.0

CLI wrapper for Docling OCR engines

ocrguru

ocrguru is a lightweight, extensible CLI tool that wraps the powerful Docling OCR pipeline.
Process scanned PDFs or images in one command, choose your OCR engine, and export clean text in Markdown, JSON, or hOCR—all without the usual setup fuss.

---

📖 Table of Contents

1. ✨ Features
2. 🚀 Installation
3. ⚡ Quick Start
4. 🛠️ CLI Reference
5. 🎯 Examples
6. 📂 Project Structure
7. 🤝 Contributing
8. ✅ Testing
9. 📜 License

---

✨ Features

• Multiple Engines

  • RapidOCR (default ONNX-based)
  • EasyOCR (PyTorch-powered)
  • Tesseract (Python wrapper or CLI)

• Input Formats

  • .pdf, .png, .jpg/.jpeg, .tiff, .bmp

• Output Formats

  • Markdown (.md) – human-friendly
  • JSON (.json) – full coordinates & metadata
  • hOCR (.html) – preserve layout & styling

• Zero-Config Defaults

  • RapidOCR models auto-download from Hugging Face & cache locally

• Cross-Platform

  • Works on Windows, macOS, and Linux

• Extensible Codebase

  • Core logic lives in core.py
  • CLI interface in cli.py
  • Easily add new engines or pipeline options

---

🚀 Installation

From PyPI:

pip install ocrguru

From your GitHub clone:

git clone https://github.com/yourusername/ocrguru.git
cd ocrguru
pip install .

Note:

• docling and huggingface-hub will install automatically.
• For GPU-accelerated EasyOCR, install PyTorch with CUDA support:

  pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cu118
  

---

⚡ Quick Start

Perform OCR on a PDF using the default RapidOCR engine, export to Markdown:

docling-ocr   --engine rapidocr   --input ./scanned_document.pdf   --format md   --output ./scanned_document.md

No extra flags needed!

---

🛠️ CLI Reference

Usage: docling-ocr [OPTIONS]

Options:
  -e, --engine [easyocr|tesseract_py|tesseract_cli|rapidocr]
                                  OCR engine (default: rapidocr)
  -i, --input PATH                 Input file path (.pdf, image)
  -f, --format [md|json|html]      Output format (default: md)
  -o, --output PATH                Output file path
  -h, --help                       Show this message and exit

---

🎯 Examples

1. EasyOCR on an Image

docling-ocr   --engine easyocr   --input invoice.jpg   --format json   --output invoice.json

2. Tesseract CLI on PDF → hOCR

docling-ocr   --engine tesseract_cli   --input contract.pdf   --format html   --output contract.hocr.html

3. Batch Processing (RapidOCR default)

for pdf in reports/*.pdf; do
  out="${pdf%.pdf}.md"
  docling-ocr     --input "$pdf"     --output "$out"
done

---

📂 Project Structure

ocrguru/
├── src/
│   └── ocrguru/
│       ├── cli.py        # CLI entry point
│       └── core.py       # OCR conversion logic
├── tests/                # pytest test suite
├── pyproject.toml        # build & metadata
└── README.md             # this file

---

🤝 Contributing

We welcome your ideas and pull requests!

1. Fork the repo & create a feature branch
2. Install dev dependencies:

   pip install -e .[test]
   

3. Write tests in tests/ and implement your feature in src/ocrguru/
4. Run the test suite:

   pytest
   

5. Open a pull request against main

Please adhere to PEP 8 and write clear commit messages.

---

✅ Testing

We use pytest for automated testing. Coverage reporting is encouraged:

pytest --cov=ocrguru

Ensure new features include corresponding tests.

---

📜 License

Released under the MIT License. See LICENSE for full text.

---

❤️ Happy OCR’ing with ocrguru! ❤️