// customMarkdownGrammar.js

// --- Custom tags with CM6-compatible class names ---
export const customTags = {
	pageLine: "pageLine", // .cm-pageLine
	snippetLine: "snippetLine", // .cm-snippetLine
	columnSplit: "columnSplit", // .cm-columnSplit
	snippetBreak: "snippetBreak", // .cm-snippetBreak
	inlineBlock: "inline-block", // .cm-inline-block
	block: "block", // .cm-block
	emoji: "emoji", // .cm-emoji
	superscript: "superscript", // .cm-superscript
	subscript: "subscript", // .cm-subscript
	definitionTerm: "dt-highlight", // .cm-dt-highlight
	definitionDesc: "dd-highlight", // .cm-dd-highlight
	injection: "injection", // .cm-injection
};

// --- Tokenizer function ---
export function tokenizeCustomMarkdown(text) {
	const tokens = [];
	const lines = text.split("\n");

	// Track multi-line blocks
	let inBlock = false;
	let blockStart = 0;

	lines.forEach((lineText, lineNumber) => {
		// --- Page / snippet lines ---
		if (/\\page/.test(lineText)) tokens.push({ line: lineNumber, type: customTags.pageLine });
		if (/\\snippet/.test(lineText)) tokens.push({ line: lineNumber, type: customTags.snippetLine });
		if (/^\\column(?:break)?$/.test(lineText)) tokens.push({ line: lineNumber, type: customTags.columnSplit });
		if (/\\snippet/.test(lineText)) tokens.push({ line: lineNumber, type: customTags.snippetBreak });

		// --- Emoji ---
		if (/:\w+?:/.test(lineText)) tokens.push({ line: lineNumber, type: customTags.emoji });

		// --- Superscript / Subscript ---
		if (/\^\^/.test(lineText)) tokens.push({ line: lineNumber, type: customTags.subscript });
		if (/\^/.test(lineText)) tokens.push({ line: lineNumber, type: customTags.superscript });

		// --- Definition lists ---
		if (/::/.test(lineText)) {
			tokens.push({ line: lineNumber, type: customTags.definitionDesc });
			tokens.push({ line: lineNumber, type: customTags.definitionTerm });
		}

		// Track ranges already marked for injections
		const injectionRanges = [];

		if (line.includes("{") && line.includes("}")) {
			const regex = /{[^{}]*}/gm;
			let match;
			while ((match = regex.exec(line)) != null) {
				codeMirror?.markText(
					{ line: lineNumber, ch: match.index },
					{ line: lineNumber, ch: match.index + match[0].length },
					{ className: "injection" },
				);
				injectionRanges.push([match.index, match.index + match[0].length]);
			}
		}

		// Now mark inline blocks, but skip overlapping injection ranges
		if (line.includes("{{") && line.includes("}}")) {
			const regex = /{{[^{}]*}}/gm;
			let match;
			while ((match = regex.exec(line)) != null) {
				const start = match.index,
					end = match.index + match[0].length;
				const overlaps = injectionRanges.some(([iStart, iEnd]) => start < iEnd && end > iStart);
				if (!overlaps) {
					codeMirror?.markText(
						{ line: lineNumber, ch: start },
						{ line: lineNumber, ch: end },
						{ className: "inline-block" },
					);
				}
			}
		}

		// --- Multi-line blocks `{{…}}` --- only start/end lines
		if (lineText.trimLeft().startsWith("{{") && !lineText.trimLeft().endsWith("}}")) {
			inBlock = true;
			blockStart = lineNumber;
			tokens.push({ line: lineNumber, type: customTags.block });
		}
		if (lineText.trimLeft().startsWith("}}") && inBlock) {
			tokens.push({ line: lineNumber, type: customTags.block });
			inBlock = false;
		}
	});

	return tokens;
}