Bengali Romanizer

A learner-facing romanizer for standard Bangladeshi Bengali. It uses an expert-built pronunciation dictionary instead of deriving pronunciation from spelling alone.

Installation

pip install bengali-romanizer

Usage

>>> import bengali_romanizer
>>> bengali_romanizer.romanize('বাংলা')
'bangla'
>>> bengali_romanizer.romanize('নমস্কার')
'nômoshkar'
>>> bengali_romanizer.romanize('ধন্যবাদ')
'dhonnobad'
>>> bengali_romanizer.romanize('ভক্তি')
'bhokti'
>>> bengali_romanizer.romanize('আন্দোলন')
'andolon'
>>> bengali_romanizer.romanize('প্রাচীন')
'prachin'

The library romanizes a word only when its pronunciation is unambiguous in the bundled dictionary. It leaves unknown words and context-dependent homographs in Bengali rather than teaching a guessed pronunciation.

Reading the output

Most letters have their familiar Latin values. ô is the vowel in British English lot, æ is the vowel in cat, ng is the final sound in sing, and sh, ch, and chh distinguish Bengali consonants. Dots under ṭ and ḍ mark retroflex consonants. a never implies a long vowel.

Pronunciations come from Google's native-speaker-transcribed Bengali pronunciation dictionary, licensed under CC BY 4.0. The dictionary targets the standard language used in Bangladesh and was built for accurate pronunciation by beginning learners and speech systems.

Documentation

The API reference is generated from the public Python API and deployed by GitHub Actions.

Lines of Code