webpeel 0.21.76 → 0.21.78

package/dist/core/domain-extractors.js

@@ -5114,14 +5114,38 @@ async function carsComExtractor(html, url) {
                 const mileage = v.mileage ? `${Number(v.mileage).toLocaleString()} mi` : '';
                 const bodyStyle = v.bodyStyle || '';
                 const fuelType = v.fuelType || '';
+                const sellerZip = v.seller?.zip || '';
                 if (title && title !== 'Used  ') {
-                    listings.push({ title, price, mileage, bodyStyle, fuelType, url: cardLink });
+                    listings.push({ title, price, mileage, bodyStyle, fuelType, url: cardLink, sellerZip });
                 }
             }
             catch { /* skip malformed */ }
         });
         if (listings.length === 0)
             return null; // Let pipeline handle it
+        // Extract dealer names from page HTML (text_style:"small", font_color:"grey")
+        const dealerPattern = /"text":"([^"]{3,50})","on_click_interactions":\[\],"text_style":"small","font_color":"grey/g;
+        const dealerNames = [];
+        let _dm;
+        while ((_dm = dealerPattern.exec(html)) !== null) {
+            const name = _dm[1];
+            if (!name.match(/^\d|^Used|^New|mi\)|^Review|^\$/))
+                dealerNames.push(name);
+        }
+        // Extract locations: "City, ST (X mi)" (e.g., "Ridgefield, NJ (8 mi)")
+        const locPattern = /([A-Z][a-z]+(?:\s[A-Z][a-z]+)*,\s[A-Z]{2}\s\(\d+\s*mi\))/g;
+        const locationList = [];
+        let _lm;
+        while ((_lm = locPattern.exec(html)) !== null) {
+            locationList.push(_lm[1]);
+        }
+        // Match dealers and locations to listings (they appear in page order)
+        for (let i = 0; i < listings.length; i++) {
+            if (i < dealerNames.length)
+                listings[i].dealer = dealerNames[i];
+            if (i < locationList.length)
+                listings[i].location = locationList[i];
+        }
         const priceRange = [minPrice && `$${minPrice}`, maxPrice && `$${maxPrice}`].filter(Boolean).join(' – ');
         const header = [
             `# 🚗 Cars.com — ${keyword || 'Vehicle Search'}`,
@@ -5139,9 +5163,16 @@ async function carsComExtractor(html, url) {
                 l.price,
                 l.mileage,
                 l.bodyStyle,
-                l.url && `[→](https://www.cars.com${l.url})`,
             ].filter(Boolean);
-            return parts.join(' · ');
+            const line = parts.join(' · ');
+            const details = [];
+            if (l.location)
+                details.push(`📍 ${l.location}`);
+            if (l.dealer)
+                details.push(`🏪 ${l.dealer}`);
+            if (l.url)
+                details.push(`🔗 [View listing](https://www.cars.com${l.url})`);
+            return line + (details.length ? '\n   ' + details.join(' · ') : '');
         });
         return {
             domain: 'cars.com',
@@ -5988,36 +6019,24 @@ async function googleFlightsExtractor(_html, url) {
     unique.sort((a, b) => a.price - b.price);
     // Helper: get airline booking URL
     function getAirlineBookingUrl(airline, from, to, dateStr) {
-        // Parse date from "Sat, Apr 4" → "2026-04-04" or "04/04/2026" etc.
-        const months = { Jan: '01', Feb: '02', Mar: '03', Apr: '04', May: '05', Jun: '06', Jul: '07', Aug: '08', Sep: '09', Oct: '10', Nov: '11', Dec: '12' };
-        let isoDate = '';
-        let mmddDate = '';
-        const dm = dateStr.match(/(\w{3})\s+(\d{1,2})/);
-        if (dm) {
-            const mon = months[dm[1]] || '01';
-            const day = dm[2].padStart(2, '0');
-            const year = new Date().getFullYear(); // current year
-            isoDate = `${year}-${mon}-${day}`;
-            mmddDate = `${mon}/${day}/${year}`;
-        }
         const fromUp = from.toUpperCase();
         const toUp = to.toUpperCase();
-        // Deep links with pre-filled origin, destination, date, 1 passenger
+        // Airline homepages — direct booking pages (deep links don't work without session/cookies)
         const urlMap = {
-            'United': `https://www.united.com/ual/en/us/flight-search/book-a-flight/results/rev?f=${fromUp}&t=${toUp}&d=${isoDate}&tt=1&at=1&sc=7&px=1&taxng=1&newHP=True&clm=7&st=bestmatches&tqp=A`,
-            'Delta': `https://www.delta.com/flight-search/book-a-flight?departure=${fromUp}&arrival=${toUp}&departureDate=${mmddDate}&paxCount=1&tripType=ONE_WAY`,
-            'JetBlue': `https://www.jetblue.com/booking/flights?from=${fromUp}&to=${toUp}&depart=${isoDate}&is498=true&is498=${true}&pax=1`,
-            'American': `https://www.aa.com/booking/find-flights?origin=${fromUp}&destination=${toUp}&departureDate=${isoDate}&tripType=OneWay&pax=1`,
-            'Spirit': `https://www.spirit.com/book/flights?origin=${fromUp}&destination=${toUp}&departureDate=${isoDate}&adults=1&tripType=oneWay`,
-            'Frontier': `https://www.flyfrontier.com/booking/flights?origin=${fromUp}&destination=${toUp}&date=${isoDate}&adults=1`,
-            'Southwest': `https://www.southwest.com/air/booking/select.html?originationAirportCode=${fromUp}&destinationAirportCode=${toUp}&departureDate=${isoDate}&adultPassengersCount=1&tripType=oneway`,
-            'Breeze': `https://www.flybreeze.com/home?from=${fromUp}&to=${toUp}&depart=${isoDate}&pax=1`,
-            'Alaska': `https://www.alaskaair.com/shopping/flights?A=${fromUp}&B=${toUp}&DT=${isoDate}&FT=ow&C=1`,
-            'Hawaiian': `https://www.hawaiianairlines.com/book/results?origin=${fromUp}&destination=${toUp}&departDate=${isoDate}&adults=1&tripType=ow`,
-            'Sun Country': `https://www.suncountry.com/booking/search?from=${fromUp}&to=${toUp}&depart=${isoDate}&passengers=1`,
-            'Avelo': `https://www.aveloair.com/book?from=${fromUp}&to=${toUp}&depart=${isoDate}`,
+            'United': `https://www.united.com`,
+            'Delta': `https://www.delta.com`,
+            'JetBlue': `https://www.jetblue.com`,
+            'American': `https://www.aa.com`,
+            'Spirit': `https://www.spirit.com`,
+            'Frontier': `https://www.flyfrontier.com`,
+            'Southwest': `https://www.southwest.com`,
+            'Breeze': `https://www.flybreeze.com`,
+            'Alaska': `https://www.alaskaair.com`,
+            'Hawaiian': `https://www.hawaiianairlines.com`,
+            'Sun Country': `https://www.suncountry.com`,
+            'Avelo': `https://www.aveloair.com`,
         };
-        return urlMap[airline] || `https://www.google.com/travel/flights?q=${encodeURIComponent(`${airline} flights ${from} to ${to} ${dateStr}`)}`;
+        return urlMap[airline] || `https://www.google.com/travel/flights?q=${encodeURIComponent(`${airline} flights ${fromUp} to ${toUp} ${dateStr}`)}`;
     }
     // Parse route from URL
     const u = new URL(url);
@@ -6037,11 +6056,12 @@ async function googleFlightsExtractor(_html, url) {
         md.push(`🛫 ${f.fromAirport} → ${f.toAirport} · ${f.duration} · ${f.stops}`);
         if (f.bags)
             md.push(`🧳 ${f.bags}`);
-        md.push(`🔗 [Book on ${f.airline}](${bookingUrl})`);
+        md.push(`🔍 [See price on Google Flights](${url})`);
+        md.push(`🛒 [Book on ${f.airline}](${bookingUrl})`);
         md.push('');
     }
     md.push('---');
-    md.push(`📌 *Prices change frequently. [View live prices on Google Flights](${url})*`);
+    md.push(`📌 *All prices verified via [Google Flights](${url}). Click "See price" to confirm, then book directly with the airline.*`);
     return {
         domain: 'google.com/travel/flights',
         type: 'flights',

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "webpeel",
-  "version": "0.21.76",
+  "version": "0.21.78",
   "description": "Fast web fetcher for AI agents - stealth mode, crawl mode, page actions, structured extraction, PDF parsing, smart escalation from simple HTTP to headless browser",
   "author": "Jake Liu",
   "license": "AGPL-3.0-only",