rohingyalanguage/rohingya-hanifi-rohingyalish-english

Rohingya Hanifi–Rohingyalish–English Lexicon

A multilingual lexical dataset from RohingyaLanguage.org connecting English dictionary headwords with Rohingyalish (Latin-script Rohingya) and Hanifi Rohingya script.

Dataset summary

• 15,926 validated rows
• Based on 6,510 English dictionary entries
• Languages: English and Rohingya (rhg)
• Scripts: Rohingyalish/Latin and Hanifi Rohingya
• Hanifi forms are generated using the same rule-based converter used by RohingyaLanguage.org's Script Converter.
• Every published row passed an exact Rohingyalish → Hanifi → Rohingyalish round-trip check.
• Rows that produced converter warnings, duplicates, or non-exact round trips were excluded.

Files and loading

The train split contains 15,926 rows in four UTF-8 JSONL shards under data/. All seven columns are available in the Hugging Face Dataset Viewer. The Hanifi and Rohingyalish text is preserved from the source export.

from datasets import load_dataset


dataset = load_dataset("rohingyalanguage/rohingya-hanifi-rohingyalish-english", split="train")

Source export: GitHub dataset folder on the hf-dataset-export branch.

Fields

Field	Description
id	Dataset row identifier
english	English dictionary headword or phrase
rohingyalish	Rohingya in Latin/Rohingyalish orthography
hanifi	Hanifi Rohingya transliteration generated by the project converter
part_of_speech	Part-of-speech/origin notation from the source dictionary
source_url	Corresponding dictionary page on RohingyaLanguage.org
validation	Validation status; current release uses exact_roundtrip

Source project

• Website: https://rohingyalanguage.org/
• Dictionary: https://rohingyalanguage.org/tools/dictionary/
• Script Converter: https://rohingyalanguage.org/tools/script-converter/
• Hanifi Unicode Reference: https://rohingyalanguage.org/resources/hanifi-rohingya-unicode/
• Online Rohingya Keyboard: https://rohingyalanguage.org/tools/rohingya-keyboard/

Validation methodology

The source dictionary was flattened so that each English–Rohingyalish translation is one row. Rohingyalish text was converted to Hanifi using the project's rule-based converter. Rows containing unknown/unmapped characters were removed. The generated Hanifi was then converted back to Rohingyalish using the inverse converter. Only rows whose normalized round-trip output exactly matched the normalized source Rohingyalish form were retained.

This confirms converter consistency. It does not mean every Hanifi form has been independently reviewed by a human linguist. Hanifi tone-marking conventions may vary among writers.

Intended uses

This dataset may be useful for:

• Rohingya NLP and low-resource language research
• Hanifi/Rohingyalish transliteration
• Search and normalization systems
• Dictionary and educational applications
• OCR post-processing and text normalization experiments
• Language-model evaluation or data preparation

Citation / project attribution

For attribution, credit RohingyaLanguage.org and link to this dataset and the CC BY 4.0 license. Indicate if you made changes.

Source website:

https://rohingyalanguage.org/

Maintainer identity

Dataset preparation and project maintenance: Abdul Aziz (AB Aziz), ORCID 0009-0003-8386-3740, at RohingyaLanguage.org. This credit describes the project and derived-data work; it does not replace attribution to the underlying source dictionary.

Archived 500-entry sample

A separate, reproducible 500-entry sample is archived on Zenodo: 10.5281/zenodo.22886278. That DOI identifies the sample release rather than the full 15,926-row dataset hosted here.

Suggested citation for the archived sample:

RohingyaLanguage.org, & Aziz, A. (2026). Rohingya Hanifi–Rohingyalish–English Lexicon: 500-Entry Sample (Version 1.0.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.22886278

License

This dataset is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0).

The publisher states that they have full redistribution rights for the underlying RohingyaLanguage.org dictionary data.