smueller/youtube-music
1
0
Fork 0
mirror of https://github.com/th-ch/youtube-music.git synced 2026-01-12 02:51:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat(synced-lyrics): thai romanization (#3618)

Browse Source 
Co-authored-by: Angelos Bouklis <me@arjix.dev>
Co-authored-by: JellyBrick <shlee1503@naver.com>
This commit is contained in:
hax0r31337
2025-09-05 15:30:39 +08:00
committed by GitHub
parent fd68c204f6
commit 588b84ecd0
3 changed files with 62 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

75
src/plugins/synced-lyrics/renderer/utils.tsx
View File

@ -1,15 +1,11 @@
import { render } from 'solid-js/web';
import KuromojiAnalyzer from 'kuroshiro-analyzer-kuromoji';
import Kuroshiro from 'kuroshiro';
import { romanize as esHangulRomanize } from 'es-hangul';
import hanja from 'hanja';
import pinyin from 'tiny-pinyin';
import * as pinyin from 'tiny-pinyin';
import { romanize as romanizeThaiFrag } from '@dehoist/romanize-thai';
import { lazy } from 'lazy-var';
import { detect } from 'tinyld';
import { waitForElement } from '@/utils/wait-for-element';
@ -155,26 +151,9 @@ const hasKorean = (lines: string[]) =>
const hasChinese = (lines: string[]) =>
lines.some((line) => /[\u4E00-\u9FFF]+/.test(line));
export const romanize = async (line: string) => {
const lang = detect(line);
const handlers: Record<string, (line: string) => Promise<string> | string> = {
ja: romanizeJapanese,
ko: romanizeHangul,
zh: romanizeChinese,
};
const NO_OP = (l: string) => l;
const handler = handlers[lang] ?? NO_OP;
line = await handler(line);
if (hasJapanese([line])) line = await romanizeJapanese(line);
if (hasKorean([line])) line = romanizeHangul(line);
if (hasChinese([line])) line = romanizeChinese(line);
return line;
};
// https://en.wikipedia.org/wiki/Thai_(Unicode_block)
const hasThai = (lines: string[]) =>
lines.some((line) => /[\u0E00-\u0E7F]+/.test(line));
export const romanizeJapanese = async (line: string) =>
(await kuroshiro.get()).convert(line, {
@ -190,3 +169,47 @@ export const romanizeChinese = (line: string) => {
pinyin.convertToPinyin(match, ' ', true),
);
};
const thaiSegmenter = Intl.Segmenter.supportedLocalesOf('th').includes('th')
? new Intl.Segmenter('th', { granularity: 'word' })
: null;
export const romanizeThai = (line: string) => {
if (!thaiSegmenter) return romanizeThaiFrag(line);
const segments = Array.from(thaiSegmenter.segment(line));
const latin = segments
.map((segment) =>
segment.isWordLike
? romanizeThaiFrag(segment.segment)
: segment.segment.trim(),
)
.join(' ')
.trim();
return latin;
};
const handlers: Record<string, (line: string) => Promise<string> | string> = {
ja: romanizeJapanese,
ko: romanizeHangul,
zh: romanizeChinese,
th: romanizeThai,
};
export const romanize = async (line: string) => {
const lang = detect(line);
const handler = handlers[lang];
if (handler) {
return handler(line);
}
// fallback
if (hasJapanese([line])) line = await romanizeJapanese(line);
if (hasKorean([line])) line = romanizeHangul(line);
if (hasChinese([line])) line = romanizeChinese(line);
if (hasThai([line])) line = romanizeThai(line);
return line;
};