extension/src/content/detector.js

/**
 * Content script - Detects job postings on the current page.
 * Extracts schema.org/JSON-LD data and common ATS page patterns.
 */

(function () {
  "use strict";

  // Detect if current page has a job posting
  const jobData = detectJobPosting();
  if (jobData) {
    chrome.runtime.sendMessage({ type: "DETECT_JOB", data: jobData });
  }

  // Listen for extraction requests from background
  chrome.runtime.onMessage.addListener((message, sender, sendResponse) => {
    if (message.type === "EXTRACT_JOB") {
      const data = extractJobData();
      sendResponse({ data });
      return true;
    }
  });

  function detectJobPosting() {
    // 1. Check JSON-LD
    const jsonLd = extractJsonLd();
    if (jsonLd) return jsonLd;

    // 2. Check common ATS URL patterns
    const url = window.location.href.toLowerCase();
    const isJobPage =
      url.includes("/jobs/") ||
      url.includes("/careers/") ||
      url.includes("/positions/") ||
      url.includes("/job/") ||
      url.includes("greenhouse.io") ||
      url.includes("lever.co") ||
      url.includes("myworkdayjobs.com") ||
      url.includes("smartrecruiters.com") ||
      url.includes("ashbyhq.com");

    if (isJobPage) {
      return { detected: true, source: "url_pattern" };
    }

    // 3. Check meta tags
    const ogType = document.querySelector('meta[property="og:type"]');
    if (ogType && ogType.content === "article") {
      // Could be a job posting page
      const title = document.title.toLowerCase();
      if (
        title.includes("job") ||
        title.includes("career") ||
        title.includes("position") ||
        title.includes("hiring")
      ) {
        return { detected: true, source: "meta_tags" };
      }
    }

    return null;
  }

  function extractJsonLd() {
    const scripts = document.querySelectorAll('script[type="application/ld+json"]');
    for (const script of scripts) {
      try {
        const data = JSON.parse(script.textContent);
        if (data["@type"] === "JobPosting") {
          return normalizeJobPosting(data);
        }
        if (Array.isArray(data)) {
          for (const item of data) {
            if (item["@type"] === "JobPosting") {
              return normalizeJobPosting(item);
            }
          }
        }
        if (data["@graph"]) {
          for (const item of data["@graph"]) {
            if (item["@type"] === "JobPosting") {
              return normalizeJobPosting(item);
            }
          }
        }
      } catch (e) {
        // Invalid JSON
      }
    }
    return null;
  }

  function normalizeJobPosting(data) {
    const location = data.jobLocation;
    let locationStr = "";
    if (location) {
      if (typeof location === "string") {
        locationStr = location;
      } else if (location.address) {
        const addr = location.address;
        locationStr = [addr.addressLocality, addr.addressRegion, addr.addressCountry]
          .filter(Boolean)
          .join(", ");
      }
    }

    return {
      detected: true,
      source: "json_ld",
      title: data.title || "",
      company: data.hiringOrganization?.name || "",
      location: locationStr,
      description: data.description || "",
      employment_type: data.employmentType || "",
      salary: data.baseSalary || null,
      date_posted: data.datePosted || "",
      apply_url: data.url || window.location.href,
      remote: data.jobLocationType === "TELECOMMUTE",
    };
  }

  function extractJobData() {
    // Try JSON-LD first
    const jsonLd = extractJsonLd();
    if (jsonLd && jsonLd.title) return jsonLd;

    // Fallback: heuristic extraction from page content
    const data = {
      detected: true,
      source: "heuristic",
      title: "",
      company: "",
      location: "",
      description: "",
      apply_url: window.location.href,
    };

    // Try to get title from common selectors
    const titleSelectors = [
      'h1[class*="job"]',
      'h1[class*="title"]',
      'h1[data-qa*="job"]',
      ".job-title",
      ".posting-headline h2",
      '[data-automation="job-detail-title"]',
      "h1",
    ];
    for (const sel of titleSelectors) {
      const el = document.querySelector(sel);
      if (el && el.textContent.trim().length > 3) {
        data.title = el.textContent.trim();
        break;
      }
    }

    // Company name
    const companySelectors = [
      '[class*="company-name"]',
      '[data-qa*="company"]',
      ".company-name",
      'meta[property="og:site_name"]',
    ];
    for (const sel of companySelectors) {
      const el = document.querySelector(sel);
      if (el) {
        data.company = el.content || el.textContent?.trim() || "";
        if (data.company) break;
      }
    }

    // Location
    const locationSelectors = [
      '[class*="location"]',
      '[data-qa*="location"]',
      ".job-location",
    ];
    for (const sel of locationSelectors) {
      const el = document.querySelector(sel);
      if (el && el.textContent.trim()) {
        data.location = el.textContent.trim();
        break;
      }
    }

    // Description - get the main content area
    const descSelectors = [
      '[class*="job-description"]',
      '[class*="description"]',
      '[data-qa*="description"]',
      ".content-payload",
      "article",
      'main [class*="content"]',
    ];
    for (const sel of descSelectors) {
      const el = document.querySelector(sel);
      if (el && el.textContent.trim().length > 100) {
        data.description = el.textContent.trim().substring(0, 5000);
        break;
      }
    }

    return data;
  }
})();