feat: harden admin security, add AI trip planner, map view, and SEO/notification improvements

Security:
- requireAdmin() session check added to every admin-only server action
  (previously relied only on middleware path matching, which Next.js
  Server Actions don't reliably respect)
- Real Prisma + bcrypt admin auth, replacing hardcoded credentials; split
  into an Edge-safe auth.config.ts (used by proxy.ts) and the full
  Prisma-backed auth.ts (route handler, server actions, server components)
- Removed hardcoded fallback secret on the Instagram sync cron endpoint
- Honeypot field + per-IP rate limiting on contact/business-submission
  forms and the analytics events endpoint

Features:
- AI trip planner (/plan-olustur, /plan/[id]) backed by DeepSeek, grounded
  to only recommend isLocalApproved listings, with a deterministic
  link-injection fallback for anything the model doesn't format as markdown
- Interactive Leaflet/OpenStreetMap view on category listing pages
- Telegram notifications for new contact messages and business submissions

SEO:
- Brand-consistent favicon/apple-icon/PWA icons and default Open Graph/
  Twitter share images, generated via next/og (replacing default Next.js
  placeholders)
- BreadcrumbList structured data on category and listing detail pages
- Fixed two remaining raw <img> tags to use next/image

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

lib/markdown.ts

@@ -0,0 +1,44 @@
+// Lightweight safe Markdown to HTML parsing function — shared by blog posts and AI itinerary results.
+export function renderMarkdownToHtml(md: string): string {
+  if (!md) return ''
+  let html = md
+    .replace(/&/g, '&')
+    .replace(/</g, '&lt;')
+    .replace(/>/g, '&gt;')
+
+  // Bold
+  html = html.replace(/\*\*(.*?)\*\*/g, '<strong>$1</strong>')
+  html = html.replace(/__(.*?)__/g, '<strong>$1</strong>')
+
+  // Italic
+  html = html.replace(/\*(.*?)\*/g, '<em>$1</em>')
+  html = html.replace(/_(.*?)_/g, '<em>$1</em>')
+
+  // Headings
+  html = html.replace(/^### (.*?)$/gm, '<h4 class="text-base font-heading font-bold text-pine mt-6 mb-2 lowercase">$1</h4>')
+  html = html.replace(/^## (.*?)$/gm, '<h3 class="text-lg font-heading font-extrabold text-pine mt-8 mb-3 lowercase">$1</h3>')
+  html = html.replace(/^# (.*?)$/gm, '<h2 class="text-xl font-heading font-extrabold text-pine mt-10 mb-4 lowercase">$1</h2>')
+
+  // Bullet Lists
+  html = html.replace(/^\* (.*?)$/gm, '<li class="ml-4 list-disc text-sm text-ink/80 leading-relaxed">$1</li>')
+  html = html.replace(/^- (.*?)$/gm, '<li class="ml-4 list-disc text-sm text-ink/80 leading-relaxed">$1</li>')
+
+  // Blockquotes
+  html = html.replace(/^&gt; (.*?)$/gm, '<blockquote class="border-l-2 border-turquoise/40 pl-4 italic text-sm text-shutter my-3">$1</blockquote>')
+
+  // Links
+  html = html.replace(/\[(.*?)\]\((.*?)\)/g, '<a href="$2" class="text-turquoise hover:underline" target="_blank" rel="noopener">$1</a>')
+
+  // Paragraphs
+  const blocks = html.split(/\n\n+/)
+  html = blocks.map(block => {
+    const trimmed = block.trim()
+    if (!trimmed) return ''
+    if (trimmed.startsWith('<h') || trimmed.startsWith('<li') || trimmed.startsWith('<ul') || trimmed.startsWith('<ol') || trimmed.startsWith('<blockquote')) {
+      return trimmed
+    }
+    return `<p class="leading-relaxed mb-4 text-sm sm:text-base text-ink/80 font-medium">${trimmed.replace(/\n/g, '<br/>')}</p>`
+  }).join('\n')
+
+  return html
+}