Jobber/extractors/qajobsboard/src/application-link.ts

/**
 * Probe external application URLs (mainly LinkedIn) for expiry and hiring location.
 */

export interface ApplicationLinkProbe {
  expired: boolean;
  location?: string;
}

const LINKEDIN_JOB_RE =
  /^https?:\/\/(?:[a-z]+\.)?linkedin\.com\/jobs\/view\/\d+/i;

const EXPIRED_URL_RE = /expired_jd_redirect|unavailable|no longer available/i;
const EXPIRED_BODY_RE =
  /\bno longer accepting applications\b|\bjob you were looking for is no longer available\b|\bthis job is no longer available\b/i;

export function isLinkedInJobUrl(url: string | undefined): boolean {
  return Boolean(url?.trim() && LINKEDIN_JOB_RE.test(url.trim()));
}

export function parseIsoDate(value: string | undefined): Date | null {
  if (!value?.trim()) return null;
  const parsed = Date.parse(value);
  if (Number.isNaN(parsed)) return null;
  return new Date(parsed);
}

export function isPostingExpiredByDate(
  expiresAt: string | undefined,
  now: Date = new Date(),
): boolean {
  const expiry = parseIsoDate(expiresAt);
  if (!expiry) return false;
  return expiry.getTime() < now.getTime();
}

function extractLinkedInLocationFromHtml(html: string): string | undefined {
  const patterns = [
    /"addressLocality"\s*:\s*"([^"]+)"[^}]*"addressCountry"\s*:\s*"([^"]+)"/i,
    /"addressCountry"\s*:\s*"([^"]+)"[^}]*"addressLocality"\s*:\s*"([^"]+)"/i,
    /"jobLocation"\s*:\s*\{[^}]*"name"\s*:\s*"([^"]+)"/i,
  ];

  for (const pattern of patterns) {
    const match = html.match(pattern);
    if (!match) continue;
    if (match.length === 3) {
      const a = match[1]?.trim();
      const b = match[2]?.trim();
      if (a && b) {
        const countryLike =
          /\b(india|canada|united states|uk|united kingdom)\b/i;
        if (countryLike.test(b)) return `${a}, ${b}`;
        if (countryLike.test(a)) return `${b}, ${a}`;
        return `${a}, ${b}`;
      }
    }
    if (match[1]?.trim()) return match[1].trim();
  }

  const indiaCity = html.match(
    /\b(Bengaluru|Bangalore|Mumbai|Hyderabad|Pune|Chennai|Delhi|Gurgaon|Noida)[^<]{0,40}\bIndia\b/i,
  );
  if (indiaCity?.[0]) {
    return indiaCity[0].replace(/\s+/g, " ").trim();
  }

  const countryOnly = html.match(
    /\b(?:job\s+)?location[^<]{0,40}\b(India|Canada|United States|United Kingdom)\b/i,
  );
  if (countryOnly?.[1]) return countryOnly[1].trim();

  return undefined;
}

export async function probeApplicationLink(
  url: string,
  fetchImpl: typeof fetch = fetch,
): Promise<ApplicationLinkProbe | null> {
  if (!url?.trim()) return null;
  if (!isLinkedInJobUrl(url)) return null;

  const response = await fetchImpl(url, {
    redirect: "follow",
    headers: {
      Accept: "text/html,application/xhtml+xml",
      "User-Agent": "JobOps/1.0",
    },
  });

  const finalUrl = response.url ?? url;
  const html = await response.text();

  const expired =
    EXPIRED_URL_RE.test(finalUrl) ||
    EXPIRED_BODY_RE.test(html) ||
    response.status === 404;

  const location = expired ? undefined : extractLinkedInLocationFromHtml(html);

  return { expired, ...(location ? { location } : {}) };
}