uroman-fast 0.1.0b0

Fast Python wrapper for uroman-rs universal romanizer

uroman-fast

Fast Python wrapper for uroman-rs, a high-performance universal romanizer that converts text from any Unicode script to Latin characters.

Features

• High performance — Rust-powered implementation via PyO3
• Universal support — Handles Japanese, Arabic, Devanagari, Chinese, Korean, and 100+ other scripts
• Optional language hints — ISO 639-3 codes improve romanization quality
• Zero dependencies — No Python runtime dependencies beyond the extension

Installation

From PyPI

pip install uroman-fast

From source

git clone https://github.com/YOUR_USERNAME/uroman-fast.git
cd uroman-fast
pip install -e .

Requirements

• Python ≥ 3.8
• Rust ≥ 1.85 (for building from source; install via rustup)

Usage

from uroman_fast import romanize

# Basic usage
print(romanize("こんにちは、世界！"))
# konnichiha, shijie!

# With optional language code for better romanization
print(romanize("ユーロマン", lang="jpn"))
# yuuroman

# Arabic
print(romanize("مرحبا", lang="ara"))
# marhaba

# Devanagari (Hindi)
print(romanize("नमस्ते", lang="hin"))
# namaste

API

romanize(text: str, lang: str | None = None) -> str

• text — Input text in any Unicode script
• lang — Optional ISO 639-3 language code (e.g., "jpn", "ara", "hin", "cmn")
• Returns — Romanized text in Latin script

Building from source

# Clone the repository
git clone https://github.com/YOUR_USERNAME/uroman-fast.git
cd uroman-fast

# Ensure Rust 1.85+ is installed
rustup update

# Install in editable mode (development)
pip install -e .

# Or build a wheel
pip install maturin
maturin build --release

License

MIT — see LICENSE for details.

Acknowledgments

• uroman-rs by Stellanomia — the underlying romanization engine
• uroman — the original Perl implementation