Files
dickendock/backend/src/tools/koepi/routes.js

208 lines
9.7 KiB
JavaScript

const express = require('express');
const https = require('https');
const { authenticate } = require('../../middleware/auth');
const router = express.Router();
// ── Cache ─────────────────────────────────────────────────────────────────────
const cache = {};
function cacheGet(k) { const e=cache[k]; if(!e||Date.now()>e.ex) { delete cache[k]; return null; } return e.d; }
function cacheSet(k,d,ms=60*60*1000) { cache[k]={d,ex:Date.now()+ms}; }
// ── marktguru Cookie (Standort Duisburg 47259) ────────────────────────────────
const MG_SETTINGS = encodeURIComponent(JSON.stringify({
location: { name:'Duisburg', uniqueName:'duisburg', longitude:6.7625, latitude:51.4332, zipCodes:['47259'] },
locationSource: 'manual', hidePageflipMenu: false, cookieAccept: true, userGroup: 35,
}));
// ── Gewünschte Händler ────────────────────────────────────────────────────────
const TARGET_RETAILERS = ['edeka','e center','rewe','netto','kaufland','penny','trinkgut'];
function isTargetRetailer(name) {
if (!name) return false;
const n = name.toLowerCase();
return TARGET_RETAILERS.some(r => n.includes(r));
}
// ── Angebote via Puppeteer von marktguru ─────────────────────────────────────
async function scrapeMarktguru() {
const cacheKey = 'koepi:mg';
const cached = cacheGet(cacheKey);
if (cached) return cached;
let puppeteer;
try { puppeteer = require('puppeteer-core'); }
catch(e) { return { offers:[], error:'puppeteer-core nicht installiert', scrapedAt: new Date().toISOString() }; }
const browser = await puppeteer.launch({
executablePath: '/usr/bin/chromium-browser',
args: ['--no-sandbox','--disable-setuid-sandbox','--disable-dev-shm-usage','--disable-gpu','--single-process'],
headless: true,
});
try {
const page = await browser.newPage();
await page.setUserAgent('Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36');
await page.setViewport({ width:1280, height:800 });
await page.setCookie({ name:'mg_user-settings', value:MG_SETTINGS, domain:'.marktguru.de', path:'/' });
await page.goto('https://www.marktguru.de/search/k%C3%B6nig%20pilsener?zipCode=47259', {
waitUntil: 'networkidle2', timeout: 30000,
});
await new Promise(r => setTimeout(r, 4000));
const offers = await page.evaluate(() => {
// Exakte Selektoren basierend auf marktguru HTML-Struktur
const cards = document.querySelectorAll('li.offer-list-item');
return Array.from(cards).map(card => {
const name = card.querySelector('h3')?.textContent?.trim() || '';
const brand = card.querySelector('dd.brand a')?.textContent?.trim() || '';
const price = card.querySelector('span.price')?.textContent?.trim() || '';
const oldPrice = card.querySelector('.price-bubble .old-price, .crossed')?.textContent?.trim() || '';
const retailer = card.querySelector('dd.retailer-name a')?.textContent?.trim() || '';
const dateRange= card.querySelector('dd.valid')?.textContent?.trim() || '';
const daysLeft = card.querySelector('dd.time-left span')?.textContent?.trim() || '';
const daysText = card.querySelector('dd.time-left')?.textContent?.trim() || '';
const validity = daysLeft ? `Noch ${daysLeft} Tag${daysLeft==='1'?'':'e'}` :
card.querySelector('dt.time-left')?.textContent?.includes('Brandneu') ? 'Neu' : '';
const descEl = card.querySelector('.info div, .info');
const description = descEl?.textContent?.trim() || '';
const img = card.querySelector('img.offer-list-item-img');
const badge = card.querySelector('.badge, .discount-badge')?.textContent?.trim() || '';
const address = card.querySelector('.retailer-address, .address, dd.address, [class*="address"]')?.textContent?.trim() || '';
// Bild: medium → large für bessere Qualität
const imgSrc = img?.src?.replace('/medium.webp', '/large.webp') || img?.src || null;
return { name, brand, price, oldPrice, retailer, dateRange, validity, description, badge, address, image: imgSrc };
});
});
// Filtern auf Ziel-Händler + nur König Pilsener
const filtered = offers.filter(o => {
const isKP = (o.name + ' ' + o.brand + ' ' + o.description).toLowerCase();
const hasKP = isKP.includes('pilsen') || isKP.includes('pilsener');
return hasKP && isTargetRetailer(o.retailer);
});
// Deduplizieren
const seen = new Set();
const unique = filtered.filter(o => {
const k = `${o.retailer}|${o.price}|${o.description?.slice(0,30)}`;
if (seen.has(k)) return false;
seen.add(k); return true;
});
const result = { offers: unique, total: unique.length, scrapedAt: new Date().toISOString(), source: 'marktguru.de' };
cacheSet(cacheKey, result);
return result;
} finally {
await browser.close();
}
}
// ── Prospekte von kaufda ──────────────────────────────────────────────────────
function fetchUrl(url) {
return new Promise((resolve, reject) => {
const req = https.get(url, {
headers: {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36',
'Accept': 'text/html', 'Accept-Language': 'de-DE,de;q=0.9',
}
}, res => {
let d = '';
res.on('data', c => d += c);
res.on('end', () => resolve({ status: res.statusCode, body: d }));
});
req.setTimeout(15000, () => { req.destroy(); reject(new Error('Timeout')); });
req.on('error', reject);
});
}
function extractNextData(html) {
const marker = 'application/json">';
const ni = html.indexOf('NEXT_DATA');
if (ni === -1) return null;
const st = html.indexOf(marker, ni) + marker.length;
const en = html.indexOf('</script>', st);
try { return JSON.parse(html.slice(st, en)); } catch { return null; }
}
async function scrapeProspekte() {
const cacheKey = 'koepi:prospekte';
const cached = cacheGet(cacheKey);
if (cached) return cached;
const { body } = await fetchUrl(
'https://www.kaufda.de/shelf?query=K%C3%B6nig+Pilsener&lat=51.3750&lng=6.7680&zip=47259&city=Duisburg'
);
const nd = extractNextData(body);
const contents = nd?.props?.pageProps?.pageInformation?.shelfContents?.contents || [];
const seen = new Map();
for (const item of contents) {
const c = item.content;
if (!c || item.contentType !== 'brochure') continue;
const id = c.contentId;
if (!id || seen.has(id)) continue;
const pub = c.publisher?.name || '';
const store = c.closestStore;
seen.set(id, {
id, publisher: pub,
title: c.title || '',
store: store?.name || pub,
city: store?.city || 'Duisburg',
street: store ? `${store.street||''} ${store.streetNumber||''}`.trim() : '',
zip: store?.zip || '47259',
image: c.brochureImages?.find(i => i.size === '260x270')?.url || c.brochureImage?.url || null,
validFrom: c.validFrom || null,
validTo: c.validUntil || null,
pageCount: c.pageCount || null,
url: `https://www.kaufda.de/webapp/brochure/${id}`,
isTarget: isTargetRetailer(pub),
badges: c.contentBadges?.map(b => b.name) || [],
});
}
const all = [...seen.values()];
const result = {
targeted: all.filter(b => b.isTarget),
others: all.filter(b => !b.isTarget),
scrapedAt: new Date().toISOString(),
};
cacheSet(cacheKey, result);
return result;
}
// ── Bild-Proxy ────────────────────────────────────────────────────────────────
router.get('/img', authenticate, (req, res) => {
const url = req.query.url;
if (!url || !url.startsWith('https://')) return res.status(400).end();
try {
const parsed = new URL(url);
const allowed = ['content-media.bonial.biz','publisher-media.bonial.biz','cdn.marktguru.de','www.marktguru.de'];
if (!allowed.some(h => parsed.hostname === h)) return res.status(403).end();
const r = https.get(url, { headers: {'User-Agent':'Mozilla/5.0','Referer':'https://www.marktguru.de/','Origin':'https://www.marktguru.de'} }, imgRes => {
res.setHeader('Content-Type', imgRes.headers['content-type'] || 'image/jpeg');
res.setHeader('Cache-Control', 'public, max-age=3600');
imgRes.pipe(res);
});
r.on('error', () => res.status(502).end());
} catch { res.status(400).end(); }
});
// ── Routes ────────────────────────────────────────────────────────────────────
router.get('/offers', authenticate, async (req, res) => {
try { res.json(await scrapeMarktguru()); }
catch(e) { console.error('KöPi offers:', e.message); res.status(500).json({ error: e.message }); }
});
router.get('/prospekte', authenticate, async (req, res) => {
try { res.json(await scrapeProspekte()); }
catch(e) { res.status(500).json({ error: e.message }); }
});
router.post('/clear-cache', authenticate, (req, res) => {
if (req.user?.role !== 'admin') return res.status(403).json({ error: 'Kein Zugriff' });
['koepi:mg','koepi:prospekte'].forEach(k => delete cache[k]);
res.json({ ok: true });
});
module.exports = router;