Dice up JS strings by script, emoji and bidi direction. Hand ported from Pango with the help of SheenBidi and emoji-segmenter, both of which are compiled and distributed in WebAssembly form.
A useful text-stack component if you want to do your own font selection and shaping in JS, or if you need more power to customize the display of different scripts or emojis in a browser.
The module itself is a promise since WASM loads asynchronously:
const {emoji, script, bidi} = await require('itemizer');
Each of the values on the resolved object have the same API but iterate different things. The emoji iterator stops at boundaries between text and emoji sequences:
const str = 'I ๐๐ผ proper i18n support';
let last = 0;
for (const {i, isEmoji} of emoji(str)) {
console.log(str.slice(last, i)); // logs 3 times
last = i;
}
The bidi iterator has the same API:
const str = 'Latin is common แแฝแแแแแถแแแขแแ ';
let last = 0;
for (const {i, dir} of bidi(str, 0 /* base level (ex. 1 for RTL) */)) {
console.log(str.slice(last, i)); // logs 2 times
last = i;
}
and so does the script iterator's API:
const str = 'Latin is common แแฝแแแแแถแแแขแแ ';
for (const {i, script} of script(str)) {
console.log(script); // logs 2 times
}
Check out the repo and run
$ node test.js
- Get ragel from your package manager
- Get WASI (sysroot version)
- Get LLVM Clang 9+ from your package manager
- Open the makefile and set
RL
,WASI_SYSROOT
CXX
according to steps 1, 2, 3 make