aitherboard/src/lib/services/content/opengraph-fetcher.ts

/**
 * OpenGraph metadata fetcher service
 * Fetches OpenGraph metadata from URLs and caches results
 */

export interface OpenGraphData {
  title?: string;
  description?: string;
  image?: string;
  url?: string;
  siteName?: string;
  type?: string;
  cachedAt: number;
}

const CACHE_DURATION = 7 * 24 * 60 * 60 * 1000; // 7 days
const CACHE_KEY_PREFIX = 'opengraph_';

/**
 * Fetch OpenGraph metadata from a URL
 * Uses a CORS proxy if needed, caches results in localStorage
 */
export async function fetchOpenGraph(url: string): Promise<OpenGraphData | null> {
  // Check cache first
  const cached = getCachedOpenGraph(url);
  if (cached && Date.now() - cached.cachedAt < CACHE_DURATION) {
    return cached;
  }

  try {
    // Try to fetch the page HTML
    // Note: Direct fetch may fail due to CORS, so we'll use a simple approach
    // In production, you might want to use a backend proxy or service
    const response = await fetch(url, {
      method: 'GET',
      headers: {
        'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
        'User-Agent': 'Mozilla/5.0 (compatible; aitherboard/1.0)'
      },
      mode: 'cors',
      cache: 'no-cache'
    });

    if (!response.ok) {
      throw new Error(`HTTP ${response.status}`);
    }

    const html = await response.text();
    const ogData = parseOpenGraph(html, url);
    
    // Cache the result
    if (ogData) {
      cacheOpenGraph(url, ogData);
    }
    
    return ogData;
  } catch (error) {
    console.warn('Failed to fetch OpenGraph data:', error);
    // Return cached data even if expired, or null
    return cached || null;
  }
}

/**
 * Parse OpenGraph metadata from HTML
 */
function parseOpenGraph(html: string, url: string): OpenGraphData | null {
  const og: Partial<OpenGraphData> = {
    cachedAt: Date.now()
  };

  // Extract OpenGraph meta tags
  const ogTitleMatch = html.match(/<meta\s+property=["']og:title["']\s+content=["']([^"']+)["']/i) ||
                        html.match(/<meta\s+content=["']([^"']+)["']\s+property=["']og:title["']/i);
  if (ogTitleMatch) {
    og.title = decodeHtmlEntities(ogTitleMatch[1]);
  }

  const ogDescriptionMatch = html.match(/<meta\s+property=["']og:description["']\s+content=["']([^"']+)["']/i) ||
                              html.match(/<meta\s+content=["']([^"']+)["']\s+property=["']og:description["']/i);
  if (ogDescriptionMatch) {
    og.description = decodeHtmlEntities(ogDescriptionMatch[1]);
  }

  const ogImageMatch = html.match(/<meta\s+property=["']og:image["']\s+content=["']([^"']+)["']/i) ||
                     html.match(/<meta\s+content=["']([^"']+)["']\s+property=["']og:image["']/i);
  if (ogImageMatch) {
    og.image = ogImageMatch[1];
    // Make image URL absolute if relative
    if (og.image && !og.image.startsWith('http')) {
      try {
        const baseUrl = new URL(url);
        og.image = new URL(og.image, baseUrl).href;
      } catch {
        // Invalid URL, keep as is
      }
    }
  }

  const ogUrlMatch = html.match(/<meta\s+property=["']og:url["']\s+content=["']([^"']+)["']/i) ||
                  html.match(/<meta\s+content=["']([^"']+)["']\s+property=["']og:url["']/i);
  if (ogUrlMatch) {
    og.url = ogUrlMatch[1];
  } else {
    og.url = url;
  }

  const ogSiteNameMatch = html.match(/<meta\s+property=["']og:site_name["']\s+content=["']([^"']+)["']/i) ||
                          html.match(/<meta\s+content=["']([^"']+)["']\s+property=["']og:site_name["']/i);
  if (ogSiteNameMatch) {
    og.siteName = decodeHtmlEntities(ogSiteNameMatch[1]);
  }

  const ogTypeMatch = html.match(/<meta\s+property=["']og:type["']\s+content=["']([^"']+)["']/i) ||
                  html.match(/<meta\s+content=["']([^"']+)["']\s+property=["']og:type["']/i);
  if (ogTypeMatch) {
    og.type = ogTypeMatch[1];
  }

  // Fallback to regular meta tags if OpenGraph not available
  if (!og.title) {
    const titleMatch = html.match(/<title[^>]*>([^<]+)<\/title>/i);
    if (titleMatch) {
      og.title = decodeHtmlEntities(titleMatch[1].trim());
    }
  }

  if (!og.description) {
    const metaDescriptionMatch = html.match(/<meta\s+name=["']description["']\s+content=["']([^"']+)["']/i) ||
                                  html.match(/<meta\s+content=["']([^"']+)["']\s+name=["']description["']/i);
    if (metaDescriptionMatch) {
      og.description = decodeHtmlEntities(metaDescriptionMatch[1]);
    }
  }

  // Return null if we have no useful data
  if (!og.title && !og.description && !og.image) {
    return null;
  }

  return og as OpenGraphData;
}

/**
 * Decode HTML entities
 */
function decodeHtmlEntities(text: string): string {
  const textarea = document.createElement('textarea');
  textarea.innerHTML = text;
  return textarea.value;
}

/**
 * Get cached OpenGraph data
 */
function getCachedOpenGraph(url: string): OpenGraphData | null {
  if (typeof window === 'undefined') return null;
  
  try {
    const key = CACHE_KEY_PREFIX + url;
    const cached = localStorage.getItem(key);
    if (cached) {
      return JSON.parse(cached) as OpenGraphData;
    }
  } catch (error) {
    console.warn('Error reading cached OpenGraph data:', error);
  }
  
  return null;
}

/**
 * Cache OpenGraph data
 */
function cacheOpenGraph(url: string, data: OpenGraphData): void {
  if (typeof window === 'undefined') return;
  
  const key = CACHE_KEY_PREFIX + url;
  try {
    localStorage.setItem(key, JSON.stringify(data));
  } catch (error) {
    console.warn('Error caching OpenGraph data:', error);
    // If storage is full, try to clear old entries
    try {
      clearOldCacheEntries();
      localStorage.setItem(key, JSON.stringify(data));
    } catch {
      // Give up
    }
  }
}

/**
 * Clear old cache entries to free up space
 */
function clearOldCacheEntries(): void {
  if (typeof window === 'undefined') return;
  
  const now = Date.now();
  const keysToRemove: string[] = [];
  
  for (let i = 0; i < localStorage.length; i++) {
    const key = localStorage.key(i);
    if (key && key.startsWith(CACHE_KEY_PREFIX)) {
      try {
        const data = JSON.parse(localStorage.getItem(key) || '{}') as OpenGraphData;
        if (now - data.cachedAt > CACHE_DURATION) {
          keysToRemove.push(key);
        }
      } catch {
        keysToRemove.push(key);
      }
    }
  }
  
  keysToRemove.forEach(key => localStorage.removeItem(key));
}