// 🚀 Full SEO + Broken Link + 404 + Accessibility + Image Alt CSV Export
// Run with: node seo_full_audit.js

const { Builder, By } = require("selenium-webdriver");
const chrome = require("selenium-webdriver/chrome");
const axios = require("axios");
const xml2js = require("xml2js");
const fs = require("fs");
const path = require("path");

// CSV file for Image Alt issues
const csvPath = path.join(__dirname, "image_alt_issues.csv");
fs.writeFileSync(csvPath, "Page URL,Image Src,Alt Text,Issue Type\n", "utf8");

// ==========================
// 1️⃣ Fetch URLs from sitemap.xml
// ==========================
async function getUrlsFromSitemap(sitemapUrl) {
  try {
    const res = await axios.get(sitemapUrl);
    const parsed = await xml2js.parseStringPromise(res.data);
    return parsed.urlset.url.map((u) => u.loc[0]);
  } catch (err) {
    console.error("❌ Failed to load sitemap:", err.message);
    return [];
  }
}

// ==========================
// 2️⃣ Check HTTP Status
// ==========================
async function checkLinkStatus(url) {
  try {
    const res = await axios.get(url, {
      timeout: 10000,
      validateStatus: () => true,
    });

    if (
      res.status === 200 &&
      (
        /404/i.test(res.data.match(/<title[^>]*>(.*?)<\/title>/)?.[1] ?? "")
      )
    ) {
      return "Soft 404";
    }

    return res.status;
  } catch (err) {
    return err.response ? err.response.status : "❌ No Response";
  }
}

// ==========================
// 3️⃣ Main SEO + Accessibility + Image Alt Audit
// ==========================
async function checkSEO(url, siteDomain) {
  const options = new chrome.Options();
  options.addArguments("--headless", "--no-sandbox", "--disable-gpu");

  const driver = await new Builder()
    .forBrowser("chrome")
    .setChromeOptions(options)
    .build();

  try {
    const pageStatus = await checkLinkStatus(url);
    if (pageStatus === 404 || pageStatus === "Soft 404") {
      console.log(`\n🚫 ${url} → ❌ Page not found (${pageStatus})`);
      return;
    }

    await driver.get(url);
    const pageSource = await driver.getPageSource();

    // Basic SEO Elements
    const title = await driver.getTitle();
    const descElem = await driver.findElements(By.css('meta[name="description"]'));
    const canonicalElem = await driver.findElements(By.css('link[rel="canonical"]'));
    const robotsElem = await driver.findElements(By.css('meta[name="robots"]'));
    const viewportElem = await driver.findElements(By.css('meta[name="viewport"]'));
    const charset = await driver.findElements(By.css('meta[charset]'));
    const htmlTag = await driver.findElement(By.css("html"));
    const langAttr = await htmlTag.getAttribute("lang").catch(() => "");
    const h1Tags = await driver.findElements(By.css("h1"));
    const h2Tags = await driver.findElements(By.css("h2"));

    // Meta Description
    let descContent = descElem.length > 0 ? await descElem[0].getAttribute("content") : "";
    const descLength = descContent.length;
    const descStatus =
      descLength === 0
        ? "❌ Missing"
        : descLength < 50
          ? `⚠️ Too short (${descLength})`
          : descLength > 160
            ? `⚠️ Too long (${descLength})`
            : "✅ Perfect";

    // Title length check
    const titleLength = title.length;
    const titleStatus =
      titleLength === 0
        ? "❌ Missing"
        : titleLength < 30
          ? `⚠️ Too short (${titleLength})`
          : titleLength > 65
            ? `⚠️ Too long (${titleLength})`
            : "✅ Perfect";

    // Canonical
    const canonicalURL =
      canonicalElem.length > 0 ? await canonicalElem[0].getAttribute("href") : "❌ Missing";

    // 🖼️ Image Accessibility Audit
    const imgs = await driver.findElements(By.css("img"));
    let missingAlt = 0;
    let emptyAlt = 0;
    let duplicateAlt = [];
    const altTextMap = new Map();

    for (const img of imgs) {
      const src = await img.getAttribute("src");
      const alt = (await img.getAttribute("alt"))?.trim() ?? null;

      if (alt === null) {
        missingAlt++;
        fs.appendFileSync(csvPath, `"${url}","${src}","","Missing Alt"\n`, "utf8");
        continue;
      }

      if (alt === "") {
        emptyAlt++;
        fs.appendFileSync(csvPath, `"${url}","${src}","(empty)","Empty Alt"\n`, "utf8");
      }

      if (altTextMap.has(alt)) {
        altTextMap.set(alt, altTextMap.get(alt) + 1);
      } else {
        altTextMap.set(alt, 1);
      }
    }

    for (const [altText, count] of altTextMap.entries()) {
      if (altText && count > 1) {
        duplicateAlt.push({ altText, count });
        fs.appendFileSync(
          csvPath,
          `"${url}","","${altText}","Duplicate Alt (${count} times)"\n`,
          "utf8"
        );
      }
    }

    // Detect tracking & schema tags
    const hasGTM = pageSource.includes("googletagmanager.com/gtm.js");
    const hasClarity = pageSource.includes("clarity.ms/tag");
    const hasFBPixel = pageSource.includes("fbevents.js") || pageSource.includes("fbq(");
    const hasAnalytics = pageSource.includes("www.googletagmanager.com/gtag/js");

    const ogTags = await driver.findElements(By.css("meta[property^='og:']"));
    const twitterTags = await driver.findElements(By.css("meta[name^='twitter:']"));
    const schemaScripts = await driver.findElements(By.css('script[type="application/ld+json"]'));

    // Links check
    const anchorTags = await driver.findElements(By.css("a[href]"));
    const brokenLinks = [];
    for (const a of anchorTags) {
      const href = await a.getAttribute("href");
      if (!href || href.startsWith("#") || href.startsWith("mailto:")) continue;

      const fullUrl = href.startsWith("http")
        ? href
        : `${siteDomain}${href.startsWith("/") ? href : `/${href}`}`;

      if (fullUrl.includes(siteDomain)) {
        const status = await checkLinkStatus(fullUrl);
        if (status === 404 || status === "Soft 404" || status === "❌ No Response") {
          brokenLinks.push({ link: fullUrl, status });
        }
      }
    }

    // Lazy loading check
    const images = await driver.findElements(By.css("img, video, iframe"));
    const lazyLoadCount = await Promise.all(
      images.map(async (img) => {
        const loading = await img.getAttribute("loading");
        return loading === "lazy";
      })
    );
    const lazyLoaded = lazyLoadCount.filter((v) => v).length;

    // Console Summary
    console.log(`\n🔍 Checking: ${url}`);
    console.log("-------------------------------------------");
    console.log("Title:", titleStatus);
    console.log("Meta Description:", descStatus);
    console.log("Canonical URL:", canonicalURL);
    console.log("Meta Robots:", robotsElem.length > 0 ? "✅ Found" : "⚠️ Missing");
    console.log("Viewport:", viewportElem.length > 0 ? "✅ Found" : "⚠️ Missing");
    console.log("Charset:", charset.length > 0 ? "✅ Found" : "❌ Missing");
    console.log("HTML lang:", langAttr ? `✅ ${langAttr}` : "⚠️ Missing");
    console.log("H1 Tags:", h1Tags.length > 0 ? `✅ ${h1Tags.length}` : "❌ Missing");
    console.log("H2 Tags:", h2Tags.length > 0 ? `ℹ️ ${h2Tags.length}` : "⚠️ None");
    console.log("Images:", imgs.length);
    console.log(
      "Missing Alt:",
      missingAlt > 0 ? `❌ ${missingAlt}` : "✅ None"
    );
    console.log(
      "Empty Alt:",
      emptyAlt > 0 ? `⚠️ ${emptyAlt}` : "✅ None"
    );
    console.log(
      "Duplicate Alt:",
      duplicateAlt.length > 0 ? `⚠️ ${duplicateAlt.length}` : "✅ None"
    );
    console.log("Lazy Loaded Images:", lazyLoaded > 0 ? `✅ ${lazyLoaded}` : "⚠️ None");
    console.log("Open Graph Tags:", ogTags.length > 0 ? "✅ Found" : "⚠️ Missing");
    console.log("Twitter Tags:", twitterTags.length > 0 ? "✅ Found" : "⚠️ Missing");
    console.log("Schema Markup:", schemaScripts.length > 0 ? "✅ Found" : "⚠️ Missing");
    console.log("Google Analytics:", hasAnalytics ? "✅ Found" : "⚠️ Missing");
    console.log("GTM:", hasGTM ? "✅ Found" : "⚠️ Missing");
    console.log("Clarity:", hasClarity ? "✅ Found" : "⚠️ Missing");
    console.log("Facebook Pixel:", hasFBPixel ? "✅ Found" : "⚠️ Missing");

    if (brokenLinks.length > 0) {
      console.log("\n❌ Broken Links:");
      brokenLinks.forEach((b) => console.log(`   → ${b.link} [${b.status}]`));
    } else {
      console.log("✅ No broken links found.");
    }

  } catch (err) {
    console.error(`❌ Error on ${url}:`, err.message);
  } finally {
    await driver.quit();
  }
}

// ==========================
// 4️⃣ Run Full Site Audit
// ==========================
(async () => {
  const sitemapUrl = "https://rapharehab.ca/sitemap.xml";
  const siteDomain = "http://localhost:3000";

  console.log("📄 Fetching URLs from sitemap...");
  const urls = await getUrlsFromSitemap(sitemapUrl);

  if (urls.length === 0) {
    console.error("❌ No URLs found in sitemap.");
    return;
  }

  console.log(`✅ Found ${urls.length} URLs in sitemap.`);
  console.log("🚀 Starting Full SEO + Accessibility + Broken Link Audit...");

  for (const url of urls) {
    await checkSEO(url, siteDomain);
  }

  console.log("\n✅ Full SEO Audit Completed!");
  console.log(`📁 CSV Report: ${csvPath}`);
})();