/**
 * Content Metadata Utilities
 * Extract metadata tags (dubbed, subtitled, etc.) from content titles
 */

/**
 * Extract metadata tags from a content title
 * Returns an array of tags like ["dubbed", "subtitled", "arabic"]
 */
export function extractContentMetadata(title: string): string[] {
  const tags: string[] = [];
  const lowerTitle = title.toLowerCase();

  // Arabic metadata terms
  if (/\bمدبلج\b/.test(title)) {
    tags.push('dubbed');
  }
  if (/\bمترجم\b/.test(title)) {
    tags.push('subtitled');
  }
  if (/\bعربي\b/.test(title)) {
    tags.push('arabic');
  }

  // English metadata terms
  if (/\bdubbed\b/i.test(lowerTitle)) {
    if (!tags.includes('dubbed')) tags.push('dubbed');
  }
  if (/\bsubtitled\b/i.test(lowerTitle) || /\bsub\b/i.test(lowerTitle)) {
    if (!tags.includes('subtitled')) tags.push('subtitled');
  }
  if (/\barabic\b/i.test(lowerTitle)) {
    if (!tags.includes('arabic')) tags.push('arabic');
  }

  return tags;
}

/**
 * Clean title by removing metadata tags for display/search
 * This is used to normalize titles for TMDb matching
 */
export function cleanTitleForMatching(title: string): string {
  // First, try to extract just the English/Latin portion of the title
  // This handles cases like "Red Sonja (2025) ريد سونيا" -> "Red Sonja (2025)"
  const cleanedTitle = extractLatinTitle(title);

  return cleanedTitle
    // Remove Arabic metadata
    .replace(/\bمدبلج\b/g, ' ')
    .replace(/\bمترجم\b/g, ' ')
    .replace(/\bعربي\b/g, ' ')
    // Remove English metadata
    .replace(/\bdubbed\b/gi, ' ')
    .replace(/\bsubtitled\b/gi, ' ')
    .replace(/\bsub\b/gi, ' ')
    .replace(/\barabic\b/gi, ' ')
    // Clean up spaces
    .replace(/\s+/g, ' ')
    .trim();
}

/**
 * Detect the primary script/language of a title
 * Returns the detected script type for search optimization
 */
export function detectTitleScript(title: string): 'latin' | 'arabic' | 'cjk' | 'cyrillic' | 'mixed' | 'unknown' {
  const hasLatin = /[a-zA-Z]/.test(title);
  const hasArabic = /[\u0600-\u06FF\u0750-\u077F\u08A0-\u08FF\uFB50-\uFDFF\uFE70-\uFEFF]/.test(title);
  const hasCJK = /[\u4E00-\u9FFF\u3040-\u309F\u30A0-\u30FF\uAC00-\uD7AF]/.test(title);
  const hasCyrillic = /[\u0400-\u04FF]/.test(title);

  // Count scripts present
  const scripts = [hasLatin, hasArabic, hasCJK, hasCyrillic].filter(Boolean).length;

  if (scripts === 0) return 'unknown';
  if (scripts > 1) return 'mixed';
  if (hasLatin) return 'latin';
  if (hasArabic) return 'arabic';
  if (hasCJK) return 'cjk';
  if (hasCyrillic) return 'cyrillic';

  return 'unknown';
}

/**
 * Get the TMDb language code for a detected script
 * Used for searching content in non-Latin languages
 */
export function getLanguageForScript(script: ReturnType<typeof detectTitleScript>): string | null {
  switch (script) {
    case 'arabic':
      return 'ar'; // Arabic
    case 'cjk':
      return 'zh'; // Chinese (could also be ja/ko, but zh is most common)
    case 'cyrillic':
      return 'ru'; // Russian (could also be uk/bg/etc)
    default:
      return null; // Use default language
  }
}

/**
 * Extract the Latin (English/Western) portion of a multilingual title
 * Handles cases like:
 * - "Red Sonja (2025) ريد سونيا" -> "Red Sonja (2025)"
 * - "Gladiator II مصارع 2" -> "Gladiator II"
 * - "قصة حب" -> "قصة حب" (pure Arabic, returns as-is for native search)
 * - "The Matrix" -> "The Matrix" (no change)
 */
export function extractLatinTitle(title: string): string {
  // Check if title has both Latin and non-Latin characters
  const hasLatin = /[a-zA-Z]/.test(title);
  const hasNonLatin = /[\u0600-\u06FF\u0750-\u077F\u08A0-\u08FF\uFB50-\uFDFF\uFE70-\uFEFF\u4E00-\u9FFF\u3040-\u309F\u30A0-\u30FF\uAC00-\uD7AF\u0400-\u04FF]/.test(title);

  // If title has both Latin and non-Latin characters, extract Latin portion
  if (hasLatin && hasNonLatin) {
    // Split by non-Latin script boundaries and extract Latin portions
    // Keep Latin letters, numbers, spaces, and common punctuation
    const latinParts = title
      .split(/[\u0600-\u06FF\u0750-\u077F\u08A0-\u08FF\uFB50-\uFDFF\uFE70-\uFEFF\u4E00-\u9FFF\u3040-\u309F\u30A0-\u30FF\uAC00-\uD7AF\u0400-\u04FF]+/)
      .map(part => part.trim())
      .filter(part => part.length > 0 && /[a-zA-Z]/.test(part));

    if (latinParts.length > 0) {
      // Join the Latin parts and clean up
      return latinParts.join(' ').replace(/\s+/g, ' ').trim();
    }
  }

  // Return original title if:
  // - It's purely Latin
  // - It's purely non-Latin (TMDb will need to search in original language)
  // - Extraction failed
  return title;
}

/**
 * Extract year from title
 * "Big World (2024)" -> 2024
 * "Big World 2023" -> 2023
 * "2024 Big World" -> 2024
 */
export function extractYearFromTitle(title: string): number | null {
  // Try parentheses format first: "Title (2024)" - most reliable
  const parenMatch = title.match(/\((\d{4})\)/);
  if (parenMatch) {
    const year = parseInt(parenMatch[1], 10);
    if (year >= 1900 && year <= new Date().getFullYear() + 1) {
      return year;
    }
  }

  // Try any 4-digit year pattern (19xx or 20xx)
  const yearMatch = title.match(/\b(19|20)\d{2}\b/);
  if (yearMatch) {
    const year = parseInt(yearMatch[0], 10);
    if (year >= 1900 && year <= new Date().getFullYear() + 1) {
      return year;
    }
  }

  return null;
}

/**
 * Remove year from title for cleaner TMDb search
 * "Big World (2024)" -> "Big World"
 * "2024 Big World" -> "Big World"
 * "Big World 2024" -> "Big World"
 */
export function stripYearFromTitle(title: string): string {
  return title
    // Remove year in parentheses: "Title (2024)" -> "Title"
    .replace(/\s*\(\d{4}\)\s*/g, ' ')
    // Remove year at start: "2024 Title" -> "Title"
    .replace(/^(19|20)\d{2}\s+/g, '')
    // Remove year at end: "Title 2024" -> "Title"
    .replace(/\s+(19|20)\d{2}$/g, '')
    // Remove standalone year in middle (careful not to remove from titles like "2001: A Space Odyssey")
    .replace(/\s+(19|20)\d{2}\s+/g, ' ')
    // Clean up extra spaces
    .replace(/\s+/g, ' ')
    .trim();
}