From 05b99e9ed43ee42ef9ad7b80f9282fb6598ee38e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Antoni=20Nu=C3=B1ez=20Romeu?= <antoni.nunez@ficosa.com>
Date: Thu, 16 Jul 2026 15:19:07 +0200
Subject: [PATCH] feat: working scraping for Promofarma, DocMorris, and Primor

- Fixed Primor URL format and selectors
- Created simple N8N webhook that calls scraper API
- All 3 sources now working:
  - Promofarma: 20 products
  - DocMorris: 10 products
  - Primor: 20 products
- Total: 130 products in database
- N8N webhook at /webhook/scrape working correctly
---
 apps/parapharmacy-api/src/scraper.js | 34 +++++++++++++------
 n8n/workflow-sdk-simple.js           | 51 ++++++++++++++++++++++++++++
 2 files changed, 75 insertions(+), 10 deletions(-)
 create mode 100644 n8n/workflow-sdk-simple.js

diff --git a/apps/parapharmacy-api/src/scraper.js b/apps/parapharmacy-api/src/scraper.js
index ca09f55..9838d83 100644
--- a/apps/parapharmacy-api/src/scraper.js
+++ b/apps/parapharmacy-api/src/scraper.js
@@ -28,12 +28,12 @@ const SOURCES = {
   },
   primor: {
     name: 'Primor',
-    searchUrl: (q) => `https://www.primor.eu/search?s=${encodeURIComponent(q)}`,
+    searchUrl: (q) => `https://www.primor.eu/catalogsearch/result/?q=${encodeURIComponent(q)}`,
     selectors: {
-      product: '.product-item, .product-miniature',
-      name: '.product-title, h3',
-      price: '.price',
-      link: 'a.product-title',
+      product: 'form.product-item',
+      name: 'form.product-item',
+      price: 'form.product-item',
+      link: 'a',
       image: 'img'
     }
   }
@@ -74,7 +74,10 @@ async function scrapeSource(browser, source, query) {
         'a[href*="/p/"]',
         '[class*="Product"]',
         'article',
-        '[role="listitem"]'
+        '[role="listitem"]',
+        '.product-item',
+        '.product',
+        '[data-product]'
       ];
 
       for (const sel of selectors) {
@@ -85,8 +88,8 @@ async function scrapeSource(browser, source, query) {
             count: els.length,
             samples: Array.from(els).slice(0, 3).map(el => ({
               tag: el.tagName,
-              class: el.className?.substring(0, 100),
-              html: el.outerHTML?.substring(0, 300)
+              class: el.className?.substring(0, 150),
+              html: el.outerHTML?.substring(0, 500)
             }))
           };
         }
@@ -110,8 +113,19 @@ async function scrapeSource(browser, source, query) {
 
       cards.forEach(card => {
         // Try data attributes first (Promofarma style)
-        const name = card.getAttribute('data-name') || card.querySelector(selectors.name)?.innerText?.trim();
-        const priceStr = card.getAttribute('data-pvp') || card.querySelector(selectors.price)?.innerText?.trim();
+        let name = card.getAttribute('data-name');
+        let priceStr = card.getAttribute('data-pvp');
+
+        // If no data attributes, try to extract from content
+        if (!name) {
+          const nameEl = card.querySelector('[class*="name"], h3, h2, .product-name');
+          name = nameEl?.innerText?.trim();
+        }
+        if (!priceStr) {
+          const priceEl = card.querySelector('[class*="price"], .price');
+          priceStr = priceEl?.innerText?.trim();
+        }
+
         const linkEl = card.querySelector(selectors.link);
         const imgEl = card.querySelector(selectors.image);
 
diff --git a/n8n/workflow-sdk-simple.js b/n8n/workflow-sdk-simple.js
new file mode 100644
index 0000000..fd5e81b
--- /dev/null
+++ b/n8n/workflow-sdk-simple.js
@@ -0,0 +1,51 @@
+import { workflow, node, trigger } from '@n8n/workflow-sdk';
+
+const webhook = trigger({
+  type: 'n8n-nodes-base.webhook',
+  version: 2,
+  config: {
+    name: 'Webhook',
+    position: [240, 300],
+    parameters: {
+      httpMethod: 'POST',
+      path: 'scrape',
+      responseMode: 'lastNode'
+    }
+  },
+  output: [{}]
+});
+
+const callApi = node({
+  type: 'n8n-nodes-base.httpRequest',
+  version: 4.2,
+  config: {
+    name: 'Call Scraper API',
+    position: [440, 300],
+    parameters: {
+      method: 'POST',
+      url: 'http://parapharmacy-api:3002/api/scrape',
+      sendBody: true,
+      specifyBody: 'json',
+      jsonBody: '={{ JSON.stringify($json.body || { queries: ["capricare"], sources: ["promofarma"] }) }}'
+    }
+  },
+  output: [{}]
+});
+
+const response = node({
+  type: 'n8n-nodes-base.code',
+  version: 2,
+  config: {
+    name: 'Response',
+    position: [640, 300],
+    parameters: {
+      jsCode: `return [{ json: $input.first().json }];`
+    }
+  },
+  output: [{}]
+});
+
+export default workflow('scrape-simple', 'Parapharmacy Scraper Simple')
+  .add(webhook)
+  .to(callApi)
+  .to(response);