readme.md

Hello, I'm SUCATA.

A subtitle translator powered by Open-Source Artificial Intelligence models.

---

📖 Table of Contents

• Overview
• Features
• Project Structure
• Getting Started
  • Prerequisites
  • Installation
  • Usage
• Testing
• Important Note
• Project Roadmap
• Contributing
• License
• Acknowledgments

---

🌟 Overview

Sucata is an open-source tool designed to extract and translate subtitles from .mkv files, as well as process .srt, .ass, and .ssa subtitle files directly. Featuring a user-friendly interface, Sucata leverages AI language models, such as Llama, to deliver high-quality translations while preserving the context and style of the original dialogues.

---

⚙️ Features

• Subtitle Extraction:

  • Compatible with subtitle tracks in MKV files.
  • Track selection via GUI for enhanced usability.

• Intelligent Translation:

  • Support for .srt, .ass, and .ssa files.
  • Contextual adaptation of slang, cultural expressions, and emotional tones.

• Graphical User Interface:

  • Built with Tkinter, offering simplicity and accessibility.

• Multi-Language Support:

  • Supports multiple languages, including Arabic, Bengali, English, French, German, Hindi, Indonesian, Japanese, Korean, Mandarin Chinese, Marathi, Portuguese, Brazilian Portuguese, Russian, Spanish, Tamil, Telugu, Turkish, Urdu, Vietnamese, and Western Punjabi.

---

📂 Project Structure

└── sucata/
    ├── app.py                # Main project
    ├── fonts/                # Utilized fonts
    │   ├── FKGroteskNeueTrial-Bold.otf
    │   ├── FKGroteskNeueTrial-Regular.otf
    │   └── Horizon.otf
    ├── img/                  # Project images
    │   ├── sucata_hello.png
    │   ├── sucata_icon.ico
    │   ├── sucata_preview.jpeg
    │   └── kofi_pt-BR.png
    ├── requirements.txt      # Project dependencies
    └── README.md             # Universal Readme
    └── README-pt-BR.md       # Portuguese Brazilian Readme

---

🚀 Getting Started

🛠️ Prerequisites

• Python: Requires Python 3.9 or later.
• Pip: Python's package manager.
• External Tools:
  • mkvextract and mkvmerge for MKV file handling.
• A Hugging Face account (optional): Needed to access certain AI models like Llama.

📦 Installation

1. Clone the Repository:

   git clone https://github.com/pedronalis/sucata.git

2. Navigate to the Directory:

   cd sucata

3. Install Dependencies:

   pip install -r requirements.txt

4. Configure the Model (Optional):

   • To use Llama or Qwen Models, request access from Hugging Face.
   • If you don’t have access, use an alternative open-source model.

---

🪄 Usage

1. Run the program:

   python app.py

2. In the interface:
   • Select an MKV or subtitle file (.srt/.ass/.ssa).
   • Choose the source language.
   • Click Start Translation and monitor the progress in the log.

---

🧪 Testing

• Basic:
  • Use a small .srt file to test translations.
• Advanced:
  • Test with MKV files containing multiple subtitle tracks.

---

📝 Important Note

This program does not replace a professional translator and does not guarantee 100% perfect translations. While it uses advanced Artificial Intelligence models, errors or contextual inaccuracies may occur in some translations.

To improve results, you can customize the prompt inside the app.py file to fit your desired language and style. This can help the AI produce translations better tailored to your needs.

---

📅 Project Roadmap

• Initial Release: Subtitle translation and track extraction.
• Implement support for additional languages.
• Add support for batch translations.
• Improve model optimization with fine-tuning.

---

🤝 Contributing

• Fork the Repository and make your improvements!
• Submit a Pull Request with your changes.
• Report Bugs or suggest enhancements.

---

📜 License

This project is licensed under the GNU AGPL.

Feel free to contribute! ❤️

---

🙏 Acknowledgments

• Hugging Face: For supporting open-source models.
• Tkinter Community: For accessible documentation.
• All contributors and testers helping to improve the project.

---