Sanscript.js
Introduction
Sanscript is a transliteration library for Indian languages. It supports the most popular Indian scripts and several different romanization schemes. Although Sanscript focuses on Sanskrit transliteration, it has partial support for other languages and is easy to extend. This is primarily based on https://github.com/sanskrit/sanscript.js. Some of the fork changes are merged and also some of the logic from https://sanskritdocuments.org/
Setup
This project is presently not published in npm. Two alternatives are available
Usage
Sanscript is simple to use:
var output = Sanscript.t(input, from, to);
Here, from
and to
are the names of different schemes. In Sanscript, the word "scheme" refers to both scripts and romanizations. These schemes are of two types:
- Brahmic schemes, which are abugidas. All Indian scripts are Brahmic schemes.
- Roman schemes, which are alphabets. All romanizations are Roman schemes.
By default, Sanscript supports the following Brahmic schemes:
bengali
devanagari
gujarati
gurmukhi
kannada
malayalam
oriya
tamil
telugu
grantha
grantamil
and the following Roman schemes:
hk
(Harvard-Kyoto)iast
(International Alphabet of Sanskrit Transliteration)iso15919
(ISO 15919)itrans
(ITRANS)slp1
(Sanskrit Library Phonetic Basic)velthuis
(Velthuis)wx
(WX)
Disabling transliteration
When Sanscript sees the token ##
, it toggles the transliteration state:
Sanscript.t('ga##Na##pa##te', 'hk', 'devanagari'); // गNaपte
Sanscript.t('ध##र्म##क्षेत्रे', 'devanagari', 'hk'); // dhaर्मkSetre
When Sanscript sees the token \
, it disables transliteration on the character that immediately follows. \
is used for ITRANS compatibility; we recommend always using ##
instead.
Sanscript.t('a \\a', 'itrans', 'devanagari'); // अ a
Sanscript.t('\\##aham', 'itrans', 'devanagari'); // ##अहम्
Transliterating to lossy schemes
A lossy scheme does not have the letters needed to support lossless translation. For example, Bengali is a lossy scheme because it uses ব
for both ba
and va
. In future releases, Sanscript might let you choose how to handle lossiness. For the time being, it makes some fairly bad hard-coded assumptions. Corrections and advice are always welcome.
Transliteration options
You can tweak the transliteration function by passing an options
object:
<script src="node_modules/@sanskrit-coders/sanscript/sanscript.js"></script>
var output = Sanscript.t(input, from, to, options);
options
maps options to values. Currently, these options are supported:
skip_sgml
- If true, transliterate SGML tags as if they were ordinary words (<b>iti</b>
→<ब्>इति</ब्>
). Defaults tofalse
.syncope
- If true, use Hindi-style transliteration (ajay
→अजय
). In linguistics, this behavior is known as schwa syncope. Defaults tofalse
.
Contributing
Adding new schemes
Adding a new scheme is simple:
Sanscript.addBrahmicScheme(schemeName, schemeData);
Sanscript.addRomanScheme(schemeName, schemeData);
For help in creating schemeData
, see the comments on the addBrahmicScheme
and addRomanScheme
functions.
Testing
We use qunit. Just install the dev dependencies with yarn
and then open test/index.html to see the tests run.
Publishing to npm
npm publish --access public