airbnb-finder/src/lib/airbnb/parsers/text-patterns.ts

/**
 * Text pattern parsers for extracting data from visible HTML text
 * Supports both German and English patterns
 */

// Main capacity pattern - "2 guests · 1 bedroom · 2 beds · 1 bath"
const CAPACITY_PATTERN = /(\d+)\s*(?:guests?|gäste?)\s*[·•]\s*(\d+)\s*(?:bedrooms?|schlafzimmer?)\s*[·•]\s*(\d+)\s*(?:beds?|betten?)\s*[·•]\s*(\d+(?:[.,]\d+)?)\s*(?:baths?|bäder?|bath)/i;

// Fallback patterns for different orders
const CAPACITY_FALLBACK_1 = /(\d+)\s*(?:guests?|gäste?)\s*[·•]\s*(\d+)\s*(?:bedrooms?|schlafzimmer?)/i;
const CAPACITY_FALLBACK_2 = /(\d+)\s*(?:beds?|betten?)\s*[·•]\s*(\d+)\s*(?:baths?|bäder?)/i;

// Rating patterns - "4.88 · 200 reviews" or "4,88 (200)" or "4,88 · 200 Bewertungen"
const RATING_PATTERN = /(\d+[.,]\d+)\s*(?:[·•\(]?\s*(\d+)\s*(?:reviews?|bewertungen)?\)?)/i;

// Host patterns - "Hosted by David" or "Gehostet von David"
const HOST_PATTERN = /(?:hosted by|gehostet von)\s+([^\n·•,]{2,40})/i;

// Price patterns - more flexible
const PRICE_PATTERN = /(?:€|EUR|\$)\s*(\d+(?:[.,]\d{0,2})?)|(\d+(?:[.,]\d{0,2})?)\s*(?:€|EUR|\$)/i;
const PRICE_PER_NIGHT = /(\d+(?:[.,]\d+)?)\s*[\/·]\s*(?:night|nacht|per\s*night)/i;

// Max guests patterns - "6 guests maximum" or "max. 6 Gäste" or "Up to 6 guests"
const MAX_GUESTS_PATTERN = /(?:max\.?|maximum|up to)\s*(\d+)\s*(?:guests?|gäste?)|(\d+)\s*(?:guests?|gäste?)\s*(?:maximum|max\.?)/i;

// Bedroom patterns
const BEDROOM_PATTERN = /(\d+)\s*(?:bedroom|schlafzimmer|bedrooms|schlafzimmer)/i;

// Bed patterns
const BED_PATTERN = /(\d+)\s*(?:bed|beds|bett|betten)/i;

// Bathroom patterns  
const BATHROOM_PATTERN = /(\d+(?:[.,]\d+)?)\s*(?:bath|baths|badezimmer|bäder)/i;

// Title in page - <title> tag
const TITLE_PATTERN = /<title>([^<]+)<\/title>/i;

// Location from address pattern
const LOCATION_PATTERN = /([A-Z][a-zA-ZäöüÄÖÜß]+(?:\s+[A-Z][a-zA-ZäöüÄÖÜß]+)*,\s*[A-Z][a-zA-ZäöüÄÖÜß]+)/i;

export interface CapacityFacts {
  guests: number;
  bedrooms: number;
  beds: number;
  bathrooms: number;
}

export interface RatingFacts {
  rating: number;
  reviewCount: number;
}

/**
 * Parse capacity facts from text like "2 guests · 1 bedroom · 2 beds · 1 bath"
 */
export function parseCapacityFacts(text: string): CapacityFacts | null {
  // Try main pattern first
  const match = text.match(CAPACITY_PATTERN);
  if (match) {
    return {
      guests: parseInt(match[1], 10),
      bedrooms: parseInt(match[2], 10),
      beds: parseInt(match[3], 10),
      bathrooms: parseFloat(match[4].replace(',', '.')),
    };
  }
  
  // Fallback: try to extract individual values
  const bedroomMatch = text.match(BEDROOM_PATTERN);
  const bedMatch = text.match(BED_PATTERN);
  const bathroomMatch = text.match(BATHROOM_PATTERN);
  const guestMatch = text.match(/(\d+)\s*(?:guests?|gäste?)/i);
  
  if (bedroomMatch || bedMatch || bathroomMatch || guestMatch) {
    return {
      guests: guestMatch ? parseInt(guestMatch[1], 10) : 0,
      bedrooms: bedroomMatch ? parseInt(bedroomMatch[1], 10) : 0,
      beds: bedMatch ? parseInt(bedMatch[1], 10) : 0,
      bathrooms: bathroomMatch ? parseFloat(bathroomMatch[1].replace(',', '.')) : 0,
    };
  }
  
  return null;
}

/**
 * Parse rating from text like "4.88 · 200 reviews"
 */
export function parseRating(text: string): RatingFacts | null {
  const match = text.match(RATING_PATTERN);
  if (!match) return null;
  
  const rating = parseFloat(match[1].replace(',', '.'));
  const reviewCount = match[2] ? parseInt(match[2], 10) : 0;
  
  if (isNaN(rating)) return null;
  
  return { rating, reviewCount };
}

/**
 * Extract title from HTML
 */
export function parseTitle(html: string): string | null {
  // Try <title> tag first
  const titleMatch = html.match(/<title>([^<]+)<\/title>/i);
  if (titleMatch) {
    // Clean up title - usually "Title - Airbnb" format
    let title = titleMatch[1].replace(/\s*[-|–]\s*Airbnb.*$/i, '').trim();
    if (title) return title;
  }
  
  // Try og:title
  const ogTitle = html.match(/<meta[^>]*property=["']og:title["'][^>]*content=["']([^"']+)["']/i);
  if (ogTitle) return ogTitle[1];
  
  return null;
}

/**
 * Parse host name from text like "Hosted by David"
 */
export function parseHost(text: string): string | null {
  const match = text.match(HOST_PATTERN);
  if (!match) return null;
  
  return match[1].trim();
}

/**
 * Parse price from text like "€ 150 / night"
 */
export function parsePriceFromText(text: string): number | null {
  const match = text.match(PRICE_PATTERN);
  if (!match) return null;
  
  const price = parseFloat(match[1].replace(',', '.'));
  return isNaN(price) ? null : price;
}

/**
 * Parse max guests from text like "6 guests maximum"
 */
export function parseMaxGuests(text: string): number | null {
  const match = text.match(MAX_GUESTS_PATTERN);
  if (!match) return null;
  
  // Pattern has two capture groups depending on word order
  const value = match[1] || match[2];
  return value ? parseInt(value, 10) : null;
}

/**
 * Extract all text content from HTML for pattern matching
 */
export function extractVisibleText(html: string): string {
  // Remove script and style tags
  let text = html.replace(/<script[^>]*>[\s\S]*?<\/script>/gi, ' ');
  text = text.replace(/<style[^>]*>[\s\S]*?<\/style>/gi, ' ');
  
  // Replace block elements with newlines
  text = text.replace(/<\/(div|p|br|li|tr|td|th|h[1-6]|section|article|header|footer)[^>]*>/gi, '\n');
  
  // Remove remaining tags
  text = text.replace(/<[^>]+>/g, ' ');
  
  // Decode HTML entities
  text = text
    .replace(/&nbsp;/g, ' ')
    .replace(/&amp;/g, '&')
    .replace(/&lt;/g, '<')
    .replace(/&gt;/g, '>')
    .replace(/&quot;/g, '"')
    .replace(/&#(\d+);/g, (_, num) => String.fromCharCode(parseInt(num, 10)));
  
  // Normalize whitespace
  text = text.replace(/\s+/g, ' ').trim();
  
  return text;
}