/** HTML → markdown via turndown + GFM. */

import TurndownService from "turndown";
import { gfm } from "turndown-plugin-gfm";

export function normalizeWhitespace(text: string): string {
	if (!/[\r\t]| \n|\n |\n{3,}| {2}/u.test(text)) {
		return text.trim();
	}
	return text
		.replaceAll(/\r\n?/gu, "\n")
		.replaceAll(/[\t ]+/gu, " ")
		.replaceAll(/ *\n */gu, "\n")
		.replaceAll(/\n{3,}/gu, "\n\n")
		.trim();
}

export interface MarkdownOptions {
	removeImages?: boolean;
}

function createMarkdownService(removeImages: boolean): TurndownService {
	const turndown = new TurndownService({
		codeBlockStyle: "fenced",
		headingStyle: "atx",
		bulletListMarker: "-",
		emDelimiter: "_",
		strongDelimiter: "**",
	});
	turndown.use(gfm);
	turndown.remove(["script", "style", "noscript", "template"]);
	if (removeImages) {
		turndown.addRule("removeImages", { filter: "img", replacement: () => "" });
	}
	turndown.addRule("stableLinks", {
		filter: "a",
		replacement: (content, node) => {
			const href = (node as HTMLAnchorElement).getAttribute?.("href");
			if (!href) return content;
			const label = content.trim().replaceAll(/\s+/gu, " ");
			return label ? `[${label}](${href})` : href;
		},
	});
	return turndown;
}

const keepImagesService = createMarkdownService(false);
const removeImagesService = createMarkdownService(true);

function stripLargeElements(html: string): string {
	if (html.length < 40_000) return html;
	const hasTable = html.includes("<table");
	const hasList = html.includes("<ul") || html.includes("<ol");
	if (!hasTable && !hasList) return html;
	let trCount = 0;
	let liCount = 0;
	if (hasTable) {
		trCount = html.match(/<tr[\s>]/giu)?.length ?? 0;
	}
	if (hasList) {
		liCount = html.match(/<li[\s>]/giu)?.length ?? 0;
	}
	if (trCount < 20 && liCount < 100) return html;
	let result = html;
	if (trCount >= 20) {
		result = result.replaceAll(/<table[\s\S]*?<\/table>/giu, "\n\n");
	}
	if (liCount >= 100) {
		result = result.replaceAll(/<(ul|ol)[\s\S]*?<\/(ul|ol)>/giu, "\n\n[Long list]\n\n");
	}
	return result;
}

/** Drop inline data:/blob: images from HTML before turndown (Reddit/JS challenge pages). */
export function stripInlineMediaFromHtml(html: string): string {
	return html
		.replace(/<img\b[^>]*\bsrc=["']data:[^"']*["'][^>]*>/giu, "")
		.replace(/<img\b[^>]*\bsrc=["']blob:[^"']*["'][^>]*>/giu, "")
		.replace(/\bsrc=["']data:[^"']{200,}["']/giu, 'src=""')
		.replace(/\bhref=["']data:[^"']{200,}["']/giu, 'href=""');
}

/**
 * Sanitize extracted text/markdown before it enters the model context.
 * Strips data-URI embeds and long base64 blobs that blow up token counts.
 */
export function sanitizeForContext(text: string): string {
	let out = text;
	// Markdown image embeds with data: or blob:
	out = out.replace(/!\[[\s\S]*?]\(\s*data:[\s\S]*?\)/gi, "[inline image omitted]");
	out = out.replace(/!\[[\s\S]*?]\(\s*blob:[\s\S]*?\)/gi, "[inline image omitted]");
	out = out.replace(/<img\b[^>]*\bsrc=["']data:[\s\S]*?["'][^>]*>/gi, "");
	// Bare data: URLs (including truncated challenge pages)
	out = out.replace(
		/\bdata:(?:image|application|font|audio|video)\/[a-z0-9.+-]+;base64,[A-Za-z0-9+/=\s]{80,}/gi,
		"[base64 omitted]",
	);
	// Long base64 payloads without a data: prefix (must include +/ to avoid URLs)
	out = out.replace(/\b(?=[A-Za-z0-9+/]*[+/])[A-Za-z0-9+/]{200,}={0,2}\b/g, "[base64 omitted]");
	return out;
}

export function htmlToMarkdown(html: string, options: MarkdownOptions = {}): string {
	const service = options.removeImages ? removeImagesService : keepImagesService;
	const cleaned = stripInlineMediaFromHtml(stripLargeElements(html));
	return sanitizeForContext(normalizeWhitespace(service.turndown(cleaned)));
}

/** Strip tags for plain text fallback. */
export function htmlToText(html: string): string {
	return sanitizeForContext(
		normalizeWhitespace(
			html
				.replace(/<script[\s\S]*?<\/script>/giu, " ")
				.replace(/<style[\s\S]*?<\/style>/giu, " ")
				.replace(/<[^>]+>/gu, " "),
		),
	);
}