feat(synced-lyrics): thai romanization

[hax0r31337]

Support of Thai romanization from @dehoist/romanize-thai

[ArjixWasTaken]

i also suggest to expand the language detection to entire lyrics instead of a single line
as the language detection are mistaking japanese sentence with no kana to mandarin

that is a really hard problem to solve, since there are songs that mix japanese/korean/chinese

[hax0r31337]

i also suggest to expand the language detection to entire lyrics instead of a single line
as the language detection are mistaking japanese sentence with no kana to mandarin

that is a really hard problem to solve, since there are songs that mix japanese/korean/chinese

well, I never thought there are cases like that

[Copilot]

Pull Request Overview

This PR adds Thai language romanization support to the synced lyrics plugin by integrating the @dehoist/romanize-thai library. The implementation includes Thai text detection, word segmentation using the Intl.Segmenter API, and fallback handling.

• Adds Thai romanization functionality with intelligent word segmentation
• Refactors the romanization logic to use a cleaner handler-based approach
• Integrates the @dehoist/romanize-thai external dependency

Reviewed Changes

Copilot reviewed 2 out of 3 changed files in this pull request and generated 2 comments.

File	Description
src/plugins/synced-lyrics/renderer/utils.tsx	Implements Thai romanization function and refactors existing romanization logic
package.json	Adds the @dehoist/romanize-thai dependency

Files not reviewed (1)

• pnpm-lock.yaml: Language not supported

[Copilot]

[nitpick] The ternary operation could be more readable by extracting the logic into a separate function or using explicit if-else blocks for better clarity.

Suggested change

	.map((segment) => segment.isWordLike ? romanizeThaiFrag(segment.segment) : segment.segment.trim())
	.join(' ')
	.trim();
	return latin;
	};
	.map(processThaiSegment)
	.join(' ')
	.trim();
	return latin;
	};
	const processThaiSegment = (segment: { isWordLike: boolean; segment: string }): string => {
	return segment.isWordLike ? romanizeThaiFrag(segment.segment) : segment.segment.trim();
	};

[Copilot]

Consider why .trim() is only applied to non-word segments. This asymmetry might lead to inconsistent spacing behavior and should be documented or applied consistently.

Suggested change

	.map((segment) => segment.isWordLike ? romanizeThaiFrag(segment.segment) : segment.segment.trim())
	.map((segment) => romanizeThaiFrag(segment.segment).trim())

[ArjixWasTaken]

'cause word's can't have spaces in them, dumb AI

[ArjixWasTaken]

i also suggest to expand the language detection to entire lyrics instead of a single line
as the language detection are mistaking japanese sentence with no kana to mandarin

that is a really hard problem to solve, since there are songs that mix japanese/korean/chinese

well, I never thought there are cases like that

One solution might be to detect what percentage of the lyrics is japanese, if it's above a threshold of like 95%, then we assume all chinese characters are actually kanji