// Lexicon limits come in two units — maxGraphemes (what a reader calls a
// character) and maxLength (UTF-8 bytes). String.length is neither.

const segmenter = new Intl.Segmenter("en", { granularity: "grapheme" });
const encoder = new TextEncoder();

export function graphemeLength(text: string): number {
	let n = 0;
	for (const _ of segmenter.segment(text)) n++;
	return n;
}

export function utf8Length(text: string): number {
	return encoder.encode(text).length;
}

export interface TextLimit {
	graphemes: number;
	bytes: number;
}

export function withinLimit(text: string, limit: TextLimit): boolean {
	return (
		graphemeLength(text) <= limit.graphemes && utf8Length(text) <= limit.bytes
	);
}

// Grapheme boundaries, not code units: a code-unit slice splits surrogate pairs
// and detaches combining marks. The byte pass only binds on emoji-dense text.
export function truncateToLimit(text: string, limit: TextLimit): string {
	let graphemes = [...segmenter.segment(text)].map((s) => s.segment);
	if (graphemes.length > limit.graphemes) {
		graphemes = graphemes.slice(0, limit.graphemes);
	}
	let out = graphemes.join("");
	while (graphemes.length > 0 && utf8Length(out) > limit.bytes) {
		graphemes.pop();
		out = graphemes.join("");
	}
	return out;
}