fix: recipe import from unsupported websites and external URL deletion

- Enable wild mode in recipe scraper (supported_only=False) to work with any
  website that uses schema.org structured data, not just officially supported sites
- Fix storage service to skip deletion of external URLs (imported recipe images)
  instead of treating them as local file paths

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

packages/api/scripts/scrape_recipe.py

@@ -2,7 +2,7 @@
 """
 Recipe scraper script using the recipe-scrapers library.
 This script is called by the Node.js API to scrape recipes from URLs.
-Uses wild mode (supported_only=False) to work with any website, not just officially supported ones.
+Uses wild mode (supported_only=False) to work with any website that uses schema.org structured data.
 """
 
 import sys
@@ -52,8 +52,8 @@ def scrape_recipe(url):
         html = fetch_html(url)
 
         # Use scrape_html to scrape the recipe
-        # Works with officially supported websites
+        # supported_only=False enables wild mode for any website with schema.org data
-        scraper = scrape_html(html, org_url=url)
+        scraper = scrape_html(html, org_url=url, supported_only=False)
 
         # Extract recipe data with safe extraction
         recipe_data = {

packages/api/src/services/storage.service.ts

@@ -51,6 +51,11 @@ export class StorageService {
   }
 
   async deleteFile(fileUrl: string): Promise<void> {
+    // Skip deletion if this is an external URL (from imported recipes)
+    if (fileUrl.startsWith('http://') || fileUrl.startsWith('https://')) {
+      return;
+    }
+
     if (storageConfig.type === 'local') {
       const basePath = storageConfig.localPath || './uploads';
       const filePath = path.join(basePath, fileUrl.replace('/uploads/', ''));