GlotScript

• GlotScript-Tool: determines the script (writing system) of input text using ISO 15924.

• GlotScript-Resource: provides a resource displaying the writing systems for various languages.

GlotScript Resource

What writing system is each language written in?

See metadata folder.

GlotScript Tool

Detect the script (writing system) of text based on ISO 15924.

• Unicode version: 15.0.0
• The codes were sourced from Wikipedia ISO_15924.
• Unicode ranges were extracted from Unicode Character Database.

Special codes

• Zinh code is the Unicode script property value of characters that may be used with multiple scripts, and that inherit their script from a preceding base character. In some cases, we opted to integrate parts of the Zinh code (e.g. ARABIC FATHATAN..ARABIC HAMZA BELOW, ARABIC LETTER SUPERSCRIPT ALEF) into a different block.
• Zyyy code is the Unicode script for "Common" characters.
• Zzzz code is for Unicode script for "uncoded" script.

Install from pip

pip3 install GlotScript

Install from git

pip3 install GlotScript@git+https://github.com/cisnlp/GlotScript

Usage: Script Detection

from GlotScript import get_script_predictor
sp = get_script_predictor()

OR

from GlotScript import sp

sp('これは日本人です')
>> ('Hira', 0.625, {'details': {'Hira': 0.625, 'Hani': 0.375}, 'tie': False, 'interval': 0.25})

sp('This is Latin')[:1]
>> ('Latn', 1.0)

sp('මේක සිංහල')[0]
>> 'Sinh'

sp('𝄞𝄫  𒊕𒀸')
>> ('Xsux', 0.5, {'details': {'Xsux': 0.5, 'Zyyy': 0.5}, 'tie': True, 'interval': 0.0})

Usage: Script Separation

from GlotScript import separate_script

sent = "Hello Salut سلام 你好 こんにちは שלום مرحبا"
separate_script(sent)
>> {
   "Latn":"Hello Salut     ",
   "Hebr":"     שלום ",
   "Arab":"  سلام    مرحبا",
   "Hani":"   你好   ",
   "Hira":"    こんにちは  "
}

Exploring Unicode Blocks: Related Sources

Click to Exapand

• List of Unicode characters - Wikipedia
• Lightweight Plain-Text Editor for macOS - CotEditor
• The Cygwin Terminal – terminal emulator for Cygwin, MSYS, and WSL - mintty
• ISO_15924 Wikipedia
• Unicode Character Database (Blocks) - Unicode
• Unicode Character Database (Scripts) - Unicode
• A free, web-based font editor, focusing on font design hobbyists. - Glyphr-Studio-1
• Kotlin - JetBrains
• UNIX-like reverse engineering framework and command-line toolset - radare2
• FreeOrion Game
• DOMinator - Firefox
• SHSans-derived CJK font family - glow-sans
• Unicode Subset Bitfields - Microsoft
• Stops - FAIR NLLB FB
• Gradient Boosting on Decision Trees - catboost
• Blender
• Unicode Wikipedia

Citation

If you use any part of this library in your research, please cite it using the following BibTex entry.

@article{kargaran2023glotscript,
title        = {GlotScript: A Resource and Tool for Low Resource Writing System Identification},
author       = {Kargaran, Amir Hossein and Yvon, Fran{\c{c}}ois and Sch{\"u}tze, Hinrich},
year         = 2023,
journal      = {arXiv preprint arXiv:2309.13320}
}