/**
 * SEO indexability checker — determines if a page is crawlable and indexable.
 *
 * Checks:
 * 1. X-Robots-Tag response header (noindex, nofollow, noarchive)
 * 2. <meta name="robots"> tag (noindex, nofollow)
 * 3. Canonical URL (self-canonical vs cross-canonical)
 * 4. rel="nofollow" on outbound links
 * 5. hreflang tags (multilingual)
 * 6. Open Graph completeness (og:title, og:description, og:image)
 * 7. Structured data presence (JSON-LD)
 * 8. Meta description
 * 9. Title tag length (< 60 chars recommended)
 * 10. URL structure (query params, fragments, excessive depth)
 */

import type { Page } from 'playwright';

export interface SeoIssue {
  rule: string;
  severity: 'critical' | 'warning' | 'info';
  message: string;
  value?: string;
}

export interface SeoCheckResult {
  url: string;
  isIndexable: boolean;
  canonicalUrl?: string;
  isSelfCanonical?: boolean;
  hreflang: Record<string, string>; // lang → url
  issues: SeoIssue[];
  score: number; // 0-100
  checkedAt: string;
}

export async function checkSeo(page: Page, url: string, responseHeaders: Record<string, string> = {}): Promise<SeoCheckResult> {
  const issues: SeoIssue[] = [];
  const checkedAt = new Date().toISOString();
  let isIndexable = true;

  // 1. X-Robots-Tag
  const robotsHeader = responseHeaders['x-robots-tag'] ?? '';
  if (/noindex/i.test(robotsHeader)) {
    isIndexable = false;
    issues.push({ rule: 'x-robots-noindex', severity: 'critical', message: 'X-Robots-Tag: noindex found in response header', value: robotsHeader });
  }

  // 2-9. Page-level checks
  const pageData = await page.evaluate(() => {
    const meta = (name: string) => document.querySelector(`meta[name="${name}"]`)?.getAttribute('content') ?? '';
    const og = (prop: string) => document.querySelector(`meta[property="og:${prop}"]`)?.getAttribute('content') ?? '';
    const canonical = document.querySelector('link[rel="canonical"]')?.getAttribute('href') ?? '';
    const hreflang: Record<string, string> = {};
    document.querySelectorAll('link[rel="alternate"][hreflang]').forEach((el: any) => {
      hreflang[el.getAttribute('hreflang')] = el.getAttribute('href') ?? '';
    });
    const jsonLd = Array.from(document.querySelectorAll('script[type="application/ld+json"]')).map((s: any) => s.textContent ?? '');
    return {
      robotsMeta: meta('robots'),
      description: meta('description'),
      title: document.title ?? '',
      canonical,
      hreflang,
      ogTitle: og('title'),
      ogDescription: og('description'),
      ogImage: og('image'),
      hasJsonLd: jsonLd.length > 0,
      jsonLd,
    };
  }).catch(() => ({
    robotsMeta: '', description: '', title: '', canonical: '', hreflang: {},
    ogTitle: '', ogDescription: '', ogImage: '', hasJsonLd: false, jsonLd: [],
  }));

  // 2. Meta robots
  if (/noindex/i.test(pageData.robotsMeta)) {
    isIndexable = false;
    issues.push({ rule: 'meta-robots-noindex', severity: 'critical', message: 'Meta robots noindex found', value: pageData.robotsMeta });
  }

  // 3. Canonical
  let isSelfCanonical: boolean | undefined;
  if (pageData.canonical) {
    isSelfCanonical = pageData.canonical === url || pageData.canonical.replace(/\/$/, '') === url.replace(/\/$/, '');
    if (!isSelfCanonical) {
      issues.push({ rule: 'cross-canonical', severity: 'warning', message: 'Cross-canonical points elsewhere', value: pageData.canonical });
    }
  } else {
    issues.push({ rule: 'missing-canonical', severity: 'warning', message: 'No canonical link tag' });
  }

  // 4. Title
  if (!pageData.title) {
    issues.push({ rule: 'missing-title', severity: 'critical', message: 'Page has no title tag' });
  } else if (pageData.title.length > 60) {
    issues.push({ rule: 'title-too-long', severity: 'warning', message: `Title ${pageData.title.length} chars (>60)`, value: pageData.title });
  }

  // 5. Meta description
  if (!pageData.description) {
    issues.push({ rule: 'missing-description', severity: 'warning', message: 'Missing meta description' });
  } else if (pageData.description.length > 160) {
    issues.push({ rule: 'description-too-long', severity: 'info', message: `Description ${pageData.description.length} chars (>160)` });
  }

  // 6. OG tags
  if (!pageData.ogTitle) issues.push({ rule: 'missing-og-title', severity: 'info', message: 'Missing og:title' });
  if (!pageData.ogImage) issues.push({ rule: 'missing-og-image', severity: 'info', message: 'Missing og:image' });

  // 7. Structured data
  if (!pageData.hasJsonLd) issues.push({ rule: 'no-structured-data', severity: 'info', message: 'No JSON-LD structured data found' });

  // 8. URL depth
  try {
    const depth = new URL(url).pathname.split('/').filter(Boolean).length;
    if (depth > 5) issues.push({ rule: 'deep-url', severity: 'info', message: `URL depth ${depth} (>5 levels)`, value: url });
  } catch {}

  const criticals = issues.filter(i => i.severity === 'critical').length;
  const warnings = issues.filter(i => i.severity === 'warning').length;
  const score = Math.max(0, 100 - criticals * 20 - warnings * 5 - issues.filter(i => i.severity === 'info').length * 2);

  return {
    url,
    isIndexable,
    canonicalUrl: pageData.canonical || undefined,
    isSelfCanonical,
    hreflang: pageData.hreflang,
    issues,
    score,
    checkedAt,
  };
}