Daft.ie MCP Server

import * as cheerio from 'cheerio'; import { dublinAreas } from '../data/locations.js'; import type { ParsedPriceResult, ParsedBedsResult } from '../types.js'; // Helper function to parse price strings (e.g., "€2,500 per month", "€500 per week") export const parsePrice = (priceString: string | undefined): ParsedPriceResult => { if (!priceString || typeof priceString !== 'string') return { value: null, type: 'unknown' }; const priceStrNormalized = priceString.toLowerCase(); // Check for "Price on Application" or similar non-numeric prices if (priceStrNormalized.includes('price on application') || priceStrNormalized.includes('contact agent')) { return { value: null, type: 'on_application' }; } // Regex to capture amount and an optional period (month/week), including variations like "p/m", "p/w" const match = priceStrNormalized.match( /(?:€\s*)?([\d,]+(?:\.\d{2})?)\s*(?:(?:per\s*)?(month|week|mth|wk|pm|p\/m|pw|p\/w|perweek))?/i ); if (match && match[1]) { // If we at least found a number that looks like an amount let amount = parseFloat(match[1].replace(/,/g, '')); if (isNaN(amount)) return { value: null, type: 'unknown' }; const periodWord = match[2] ? match[2].toLowerCase() : undefined; if (periodWord) { // If a period was explicitly found if ( periodWord.startsWith('week') || periodWord.startsWith('wk') || periodWord === 'pw' || periodWord === 'p/w' || periodWord === 'perweek' ) { amount = Math.round((amount * 52) / 12); // Convert weekly to monthly } // Otherwise, assume monthly (includes 'month', 'mth', 'pm', 'p/m') return { value: amount, type: 'numeric' }; } else if (priceStrNormalized.includes('€')) { // No explicit period, but has '€', assume it's a direct (likely monthly) price return { value: amount, type: 'numeric' }; } // If no period and no euro symbol, it's ambiguous, fall through to stricter checks or return null } // Fallback for simple numbers that might be prices but didn't match the main regex structure // (e.g. "1500 per month" where "per month" wasn't captured by the optional group logic if it's too far) // This section needs to be careful not to misinterpret address numbers. const stricterMatchForStandaloneNumbers = priceStrNormalized.match( /^([\d,]+(?:\.\d{2})?)\s*(?:per\s*)?(month|week|mth|wk|pm|p\/m|pw|p\/w|perweek)$/i ); if (stricterMatchForStandaloneNumbers && stricterMatchForStandaloneNumbers[1]) { let amount = parseFloat(stricterMatchForStandaloneNumbers[1].replace(/,/g, '')); if (isNaN(amount)) return { value: null, type: 'unknown' }; const periodWord = stricterMatchForStandaloneNumbers[2] ? stricterMatchForStandaloneNumbers[2].toLowerCase() : 'month'; if ( periodWord.startsWith('week') || periodWord.startsWith('wk') || periodWord === 'pw' || periodWord === 'p/w' || periodWord === 'perweek' ) { amount = Math.round((amount * 52) / 12); } return { value: amount, type: 'numeric' }; } // Final check for very simple numeric strings that might be prices if they are large enough // or if they were missed by above logic but are clearly prices. const simpleNumericOnly = priceString.replace(/[^0-9.]/g, ''); if (simpleNumericOnly.length > 0 && simpleNumericOnly === priceString.replace(/[^0-9€\s.,]/g, '')) { const potentialPrice = parseFloat(simpleNumericOnly); if (!isNaN(potentialPrice)) { if ( priceStrNormalized.includes('€') || priceStrNormalized.includes('per month') || priceStrNormalized.includes('per week') ) { if (priceStrNormalized.includes('€')) { return { value: potentialPrice, type: 'numeric' }; } } if ( priceStrNormalized.toLowerCase().includes('dublin') && priceStrNormalized.includes(potentialPrice.toString()) && potentialPrice < 100 ) { return { value: null, type: 'unknown' }; } if ( potentialPrice < 100 && !priceStrNormalized.includes('€') && !priceStrNormalized.match(/(month|week|mth|wk|pm|pw|p\/m|p\/w)/i) ) { return { value: null, type: 'unknown' }; // Small number, no indicators } if (potentialPrice >= 100) { return { value: potentialPrice, type: 'numeric' }; // Allow larger numbers if they are purely numeric } } } return { value: null, type: 'unknown' }; }; // Helper function to parse bed strings (e.g., "1 Bed", "2 Beds", "Studio") // Returns an object { min: number, max: number } or null export const parseBeds = (bedString: string | undefined): ParsedBedsResult | null => { if (!bedString || typeof bedString !== 'string') return null; const bedStrNormalized = bedString.toLowerCase(); if (bedStrNormalized.includes('studio')) { return { min: 1, max: 1, isStudio: true }; } const rangeMatch = bedStrNormalized.match(/(\d+)\s*(?:to|-)\s*(\d+)\s*bed/i); if (rangeMatch && rangeMatch[1] && rangeMatch[2]) { const minBeds = parseInt(rangeMatch[1], 10); const maxBeds = parseInt(rangeMatch[2], 10); if (!isNaN(minBeds) && !isNaN(maxBeds)) { return { min: Math.min(minBeds, maxBeds), max: Math.max(minBeds, maxBeds) }; } } const singleMatch = bedStrNormalized.match(/(\d+)\s*bed/i); if (singleMatch && singleMatch[1]) { const num = parseInt(singleMatch[1], 10); if (!isNaN(num)) { return { min: num, max: num }; } } return null; }; export const extractLatLng = (html: string, selector: string): { lat: number; lng: number } | null => { const $ = cheerio.load(html); const link = $(selector).attr('href'); if (link) { const match = link.match(/q=loc:([\d.-]+)\+([\d.-]+)/) || link.match(/viewpoint=([\d.-]+),([\d.-]+)/); if (match && match[1] && match[2]) { const lat = parseFloat(match[1]); const lng = parseFloat(match[2]); if (!isNaN(lat) && !isNaN(lng)) { return { lat, lng }; } } } return null; }; export const slugify = (str: string): string => { if (typeof str !== 'string') return ''; return str .trim() .toLowerCase() .replace(/\s+/g, '-') .replace(/[^a-z0-9-]/g, ''); }; export const formatBer = (ber: string | undefined): string | undefined => { if (!ber) return undefined; const cleanedBer = ber.replace('BER ', '').trim(); if (cleanedBer === 'SI_666') { return 'Exempt'; } return cleanedBer; }; export const generateDaftLocationSlug = (locationString: string): string => { if (typeof locationString !== 'string' || locationString.trim() === '') { return ''; } const parts = locationString .split(',') .map((p) => slugify(p.trim())) .filter((p) => p); if (parts.length === 2) { // Handles "Carrigaline, Cork" -> "carrigaline-cork" return parts.join('-'); } const singlePartSlug = slugify(locationString); // Handles "Dublin 2" -> "dublin-2-dublin" if (/^dublin-\d+$/.test(singlePartSlug)) { return `${singlePartSlug}-dublin`; } // Handles areas within Dublin like "Sandymount" -> "sandymount-dublin" if (dublinAreas.includes(singlePartSlug)) { return `${singlePartSlug}-dublin`; } // Default for single names like "cork" or "galway" return singlePartSlug; };