Natural Language Processing tools and resources for Ethiopian languages.
AbugidaNLP is an open-source library focused on developing NLP tools tailored to Ethiopian languages. The library aims to support tasks like calendar system conversion, numeral conversion, and transliteration between Ethiopic and Latin scripts. As one of the first steps towards bridging the gap in NLP tools for underrepresented languages, AbugidaNLP aspires to grow into a comprehensive toolkit for linguists, researchers, and developers.
AbugidaNLP is currently in its infancy, offering:
- Transliteration: Convert text between Ethiopic and Latin scripts.
- Calendar System Conversion: Seamlessly convert dates between Ethiopic and Gregorian calendars.
- Numeral System Conversion: Translate numbers between Ge'ez and Hindu-Arabic numeral systems.
- Python 3.8 or higher
- pip
Install AbugidaNLP using pip:
pip install abugida
Here’s how to get started with AbugidaNLP:
from abugida import ScriptConverter # Import the ScriptConverter class from AbugidaNLP
# Initialize the ScriptConverter
converter = ScriptConverter()
# Forward transliteration: Convert Ethiopic script to Latin script
result_fwd = converter.transliterate("በመተባበራችን", "fwd")
print(result_fwd) # Output: bemetebaberacn
# Backward transliteration: Convert Latin script to Ethiopic script
result_bwd = converter.transliterate("merejawoc", "bwd")
print(result_bwd) # Output: መረጃዎች
- Import the library: The
ScriptConverter
class handles script conversions. - Initialize: Create an instance of the
ScriptConverter
to use its methods. - Transliteration:
- Use the
"fwd"
mode for Ethiopic to Latin conversion. - Use the
"bwd"
mode for Latin to Ethiopic conversion.
- Use the
Documentation for AbugidaNLP is under construction. Stay tuned for detailed usage examples and API references.
We welcome contributions to AbugidaNLP! Whether it's improving documentation, adding new features, or fixing bugs, your help is valuable.
- Fork the repository.
- Create a new branch for your feature/bugfix.
- Submit a pull request.
This project is licensed under the MIT License - see the LICENSE file for details.
Special thanks to the following resources for their inspiration and insights:
For questions or support, please open an issue.