import { load } from "cheerio";

// ─── HTML Extraction ─────────────────────────────────────────────────────────

/**
 * Extract readable text from HTML using Cheerio.
 * Removes scripts, styles, nav, ads, and other non-content elements,
 * then tries to find the main content area (article, main, .content, etc.)
 * before falling back to the full body text.
 */
export function stripHtml(html: string): string {
  const $ = load(html);

  // Remove non-content elements
  $("script, style, noscript, template, iframe, svg").remove();
  $("nav, footer, aside, header, menu, button, form").remove();
  $("[role='navigation'], [role='banner'], [role='complementary']").remove();
  $(".advertisement, .ad, .ads, .cookie-banner, .newsletter").remove();

  // Try to find the main content area
  let $content = $("article, main, [role='main'], .content, #content, .post, .entry, .article").first();

  // Fallback to body or html
  if (!$content.length) {
    $content = $("body, html").first();
  }

  // Get text — cheerio decodes HTML entities automatically
  let text = $content.text();

  // Normalize whitespace
  text = text
    .replace(/\u00A0/g, " ")        // non-breaking space → regular space
    .replace(/\r\n/g, "\n")
    .replace(/\r/g, "\n")
    .replace(/[ \t]*\n[ \t]*/g, "\n")
    .replace(/\n{3,}/g, "\n\n")
    .replace(/[ \t]+/g, " ")
    .trim();

  return text;
}

export function normalizeUrl(url: string): string {
  try {
    const u = new URL(url);
    u.hostname = u.hostname.toLowerCase();
    u.hash = "";
    let path = u.pathname;
    if (path.endsWith("/") && path.length > 1) {
      path = path.slice(0, -1);
    }
    u.pathname = path;
    return u.toString();
  } catch {
    return url;
  }
}

export function isInternalUrl(url: string): boolean {
  try {
    const u = new URL(url);
    let hostname = u.hostname.toLowerCase();

    // Strip IPv6 brackets so "[::1]" → "::1"
    if (hostname.startsWith("[") && hostname.endsWith("]")) {
      hostname = hostname.slice(1, -1);
    }

    if (hostname === "localhost" || hostname === "0.0.0.0" || hostname === "::1" || hostname === "::") return true;
    if (hostname.startsWith("127.")) return true;
    if (hostname.startsWith("10.")) return true;
    if (hostname.startsWith("192.168.")) return true;
    if (hostname.startsWith("172.")) {
      const parts = hostname.split(".");
      const second = parseInt(parts[1], 10);
      if (second >= 16 && second <= 31) return true;
    }
    if (hostname.startsWith("169.254.")) return true;
    if (hostname.startsWith("fc") || hostname.startsWith("fd")) return true;
    if (u.protocol === "file:") return true;

    return false;
  } catch {
    return false;
  }
}

export function urlContainsSensitiveParam(url: string): boolean {
  try {
    const u = new URL(url);
    const sensitive = ["token", "key", "api_key", "apikey", "secret", "password", "auth", "credential"];
    for (const param of sensitive) {
      if (u.searchParams.has(param)) {
        const value = u.searchParams.get(param) ?? "";
        if (value.length > 4) return true;
      }
    }
    return false;
  } catch {
    return false;
  }
}