Merge pull request #182 from R0m1k3/claude/rebuild-search-amazon-01LBnzN1wfNMe1EG9HcpMCdK

Claude/rebuild search amazon 01 l bnz n1wf n me1 eg9 hcp m cd k

app/services/improved_search_service.py

@@ -625,7 +625,7 @@ async def search_products(
     for site in active_sites:
         matched = False
 
-        # Normalize domain for comparison (remove www., e-, etc.)
+        # Normalize domain for comparison (remove www., lowercase, etc.)
         normalized_domain = site.domain.lower().replace("www.", "").strip()
 
         for key in SITE_CONFIGS.keys():
@@ -639,16 +639,48 @@ async def search_products(
                 logger.debug(f"✓ Matched {site.domain} → {key} (exact)")
                 break
 
-            # 2. Key contains domain or domain contains key
+            # 2. Normalize punctuation (. vs - vs nothing) and compare
+            # e.leclerc → eleclerc, e-leclerc.com → eleclecrcom
+            domain_no_punct = normalized_domain.replace("-", "").replace(".", "")
+            key_no_punct = normalized_key.replace("-", "").replace(".", "")
+
+            # Exact match without punctuation
+            if domain_no_punct == key_no_punct:
+                site_keys.append(key)
+                matched = True
+                logger.debug(f"✓ Matched {site.domain} → {key} (normalized punctuation - exact)")
+                break
+
+            # Contains match without punctuation (handles .com, .fr suffixes)
+            # eleclerc in eleclecrcom → True
+            if domain_no_punct in key_no_punct or key_no_punct in domain_no_punct:
+                site_keys.append(key)
+                matched = True
+                logger.debug(f"✓ Matched {site.domain} → {key} (normalized punctuation - contains)")
+                break
+
+            # 3. Key contains domain or domain contains key
             if normalized_key in normalized_domain or normalized_domain in normalized_key:
                 site_keys.append(key)
                 matched = True
                 logger.debug(f"✓ Matched {site.domain} → {key} (contains)")
                 break
 
-            # 3. Remove prefixes like "e-" and try again
-            domain_without_prefix = normalized_domain.replace("e-", "").replace("la-", "")
-            key_without_prefix = normalized_key.replace("e-", "").replace("la-", "")
+            # 4. Remove prefixes like "e-", "e.", "la-", "la." and try again
+            domain_without_prefix = (
+                normalized_domain
+                .replace("e-", "")
+                .replace("e.", "")
+                .replace("la-", "")
+                .replace("la.", "")
+            )
+            key_without_prefix = (
+                normalized_key
+                .replace("e-", "")
+                .replace("e.", "")
+                .replace("la-", "")
+                .replace("la.", "")
+            )
 
             if domain_without_prefix == key_without_prefix:
                 site_keys.append(key)
@@ -656,6 +688,24 @@ async def search_products(
                 logger.debug(f"✓ Matched {site.domain} → {key} (without prefix)")
                 break
 
+            # 5. Last resort: fuzzy match on core name (remove all punct + common prefixes)
+            domain_core = (
+                domain_no_punct
+                .replace("e", "", 1)  # Remove first 'e' if present
+                .replace("la", "", 1)  # Remove first 'la' if present
+            )
+            key_core = (
+                key_no_punct
+                .replace("e", "", 1)
+                .replace("la", "", 1)
+            )
+
+            if len(domain_core) > 5 and domain_core in key_core:
+                site_keys.append(key)
+                matched = True
+                logger.debug(f"✓ Matched {site.domain} → {key} (fuzzy core)")
+                break
+
         if not matched:
             logger.warning(f"⚠️ No config found for site: {site.domain} (id={site.id})")
 

app/services/parsers/fnac_parser.py

@@ -80,11 +80,30 @@ class FnacParser(BaseParser):
                 if not self.filter_by_query(title, query):
                     continue
 
-                # Extract image
+                # Extract image - try in link first, then parent
                 image_url = None
                 parent = link.find_parent(["article", "li", "div"])
-                if parent:
-                    # Try picture > source first
+
+                # PRIORITY 1: Try in the link itself
+                picture = link.find("picture")
+                if picture:
+                    source = picture.find("source")
+                    if source and source.get("srcset"):
+                        image_url = source.get("srcset").split(",")[0].split()[0]
+                    if not image_url:
+                        img = picture.find("img")
+                        if img:
+                            image_url = self._get_image_src(img)
+
+                # Try img directly in link
+                if not image_url:
+                    img = link.find("img")
+                    if img:
+                        image_url = self._get_image_src(img)
+
+                # PRIORITY 2: Try in parent container
+                if not image_url and parent:
+                    # Try picture > source
                     picture = parent.find("picture")
                     if picture:
                         source = picture.find("source")
@@ -101,8 +120,10 @@ class FnacParser(BaseParser):
                         if img:
                             image_url = self._get_image_src(img)
 
-                    if image_url:
-                        image_url = self.make_absolute_url(image_url)
+                if image_url:
+                    image_url = self.make_absolute_url(image_url)
+                else:
+                    self.logger.warning(f"No image found for Fnac product: {title[:40]}...")
 
                 # Extract price
                 price = None

app/services/parsers/generic_parser.py

@@ -50,6 +50,8 @@ class GenericParser(BaseParser):
         Uses:
         - product_selector: CSS selector for product links
         - product_image_selector: CSS selector for images (optional)
+
+        If configured selector fails, tries common fallback selectors.
         """
         soup = BeautifulSoup(html, 'html.parser')
         products = []
@@ -62,8 +64,44 @@ class GenericParser(BaseParser):
 
         links = soup.select(product_selector)
 
+        # If configured selector fails, try common fallback selectors
         if not links:
-            self.logger.warning(f"No products found for {self.site_key} with selector: {product_selector}")
+            self.logger.warning(f"No products with configured selector: {product_selector}")
+            self.logger.info("Trying fallback selectors...")
+
+            # Common e-commerce selectors (ordered by specificity)
+            fallback_selectors = [
+                # Product-specific URLs
+                "a[href*='/produit']",
+                "a[href*='/products/']",  # Shopify
+                "a[href*='/product']",
+                "a[href*='/p/']",
+                "a[href*='/item']",
+                # Common class patterns
+                "article a[href]",
+                ".product a[href]",
+                ".product-card a[href]",
+                ".product-item a[href]",
+                "[class*='product'] a[href]",
+                # Shopify specific
+                "a.product-card__link",
+                "a[href*='/collections/']",
+                # Generic structure
+                "article[class*='product'] a",
+                "div[class*='product'] a",
+                "li[class*='product'] a",
+                # Very generic (last resort)
+                "a[class*='product']",
+            ]
+
+            for fallback in fallback_selectors:
+                links = soup.select(fallback)
+                if links and len(links) >= 3:  # At least 3 links to be credible
+                    self.logger.info(f"✓ Found {len(links)} links with fallback: {fallback}")
+                    break
+
+        if not links:
+            self.logger.warning(f"No products found for {self.site_key} even with fallbacks")
             return []
 
         self.logger.info(f"Found {len(links)} product links for {self.site_name}")