import { getDomain } from "tldts";

/**
 * Extract eTLD+1 domain from a URL string.
 * Returns null for invalid URLs.
 */
export function extractDomain(url: string): string | null {
  try {
    const normalized = url.startsWith("http") ? url : "https://" + url;
    const parsed = getDomain(normalized);
    return parsed ?? null;
  } catch {
    return null;
  }
}

/**
 * Get displayable domain from a URL string.
 */
export function getDisplayDomain(url: string): string {
  try {
    const normalized = url.startsWith("http") ? url : "https://" + url;
    const { hostname } = new URL(normalized);
    return hostname;
  } catch {
    return url;
  }
}

/**
 * LV-2: true when the URL's hostname contains non-ASCII characters — a
 * potential homograph (Cyrillic "аpple.com" etc.) — OR a punycode (xn--)
 * label, which is the URL parser's normalized form of the same risk.
 * Matching/keying is NOT affected (tldts keeps Unicode labels distinct,
 * preventing cross-matching); this only drives a non-blocking UI warning so a
 * user who saved or imported such a URL can see it isn't the ASCII site.
 * Accepts bare hostnames as well as full URLs.
 */
export function detectHomograph(url: string): boolean {
  try {
    const hostname = new URL(url.includes("://") ? url : `https://${url}`).hostname;
    return (
      // biome-ignore lint/suspicious/noControlCharactersInRegex: intentional — detects non-ASCII hostnames
      /[^\x00-\x7F]/.test(hostname) ||
      hostname.split(".").some((l) => l.startsWith("xn--"))
    );
  } catch {
    return false;
  }
}