parsers: better image extraction — real product photos in report cards

CITILINK:
- Now reads data-src / data-original / srcset / src in priority order
- srcset → picks largest size variant (last in comma-list)
- Filters only _next/static/images (placeholder) and 'placeholder' in URL
- Accepts cs.citilink.ru / c.citilink.ru / images.citilink.ru product photos

ЯНДЕКС.МАРКЕТ:
- Collects all img attrs (data-src, data-original, srcset, data-srcset, src)
- Prefers avatars.mds.yandex.net (real product CDN), skips yastatic (icons/logos)
- Auto-appends /300x300 suffix to avatars.mds URLs without size

ENRICH_ONE (aggregator):
- Image picked by source priority: yamarket > wb > ozon > citilink > dns
- Yamarket photos are cleanest (avatars.mds.yandex.net)
- WB has product photos via basket-XX.wbbasket.ru

backend-py/app/parsers/__init__.py

@@ -86,10 +86,18 @@ def enrich_one(query: str, sources: tuple = DEFAULT_SOURCES) -> dict[str, Any]:
 
     # Агрегация
     prices = [i["price_min_rub"] for i in items.values() if i and i.get("price_min_rub")]
-    images = [i["image_url"]     for i in items.values() if i and i.get("image_url")]
     ratings = [i["rating"]       for i in items.values() if i and i.get("rating")]
     reviews = [i["reviews_count"] for i in items.values() if i and i.get("reviews_count")]
 
+    # Фото — выбираем по приоритету источника (качество фото различается)
+    image_priority = ("yamarket", "wb", "ozon", "citilink", "dns")
+    image_url = None
+    for src in image_priority:
+        i = items.get(src)
+        if i and i.get("image_url"):
+            image_url = i["image_url"]
+            break
+
     # Я.Маркет даёт количество магазинов
     stores = None
     if items.get("yamarket") and items["yamarket"].get("stores_count"):
@@ -107,7 +115,7 @@ def enrich_one(query: str, sources: tuple = DEFAULT_SOURCES) -> dict[str, Any]:
         **{src: items.get(src) for src in fetchers.keys()},
         "price_min_rub": min(prices) if prices else None,
         "price_max_rub": max(prices) if prices else None,
-        "image_url": images[0] if images else None,
+        "image_url": image_url,
         "rating_max": max(ratings) if ratings else None,
         "reviews_total": sum(reviews) if reviews else None,
         "stores_count": stores,

backend-py/app/parsers/citilink.py

@@ -133,20 +133,37 @@ def _extract_card(card, url: str) -> dict[str, Any] | None:
     if not title or len(title) < 10:
         return None
 
-    # Картинка — ищем реальное товарное фото (НЕ placeholder/SSR-иконку)
+    # Картинка — пробуем разные источники: src, data-src, srcset
     img_url = None
     for img_el in card.find_all("img"):
-        src = img_el.get("src") or img_el.get("data-src") or ""
-        if not src or "data:image" in src:
-            continue
-        if src.startswith("//"):
-            src = "https:" + src
-        # Filter placeholders: rstatic.citilink.ru/_next/static/images/... всегда заглушка
-        if "_next/static/images" in src:
-            continue
-        # Реальные товарные фото — на c.citilink.ru или main.citilink.ru
-        img_url = src
-        break
+        # Источники в порядке приоритета
+        candidates = []
+        for attr in ("data-src", "data-original", "data-srcset", "srcset", "src"):
+            val = img_el.get(attr) or ""
+            if not val:
+                continue
+            if attr in ("srcset", "data-srcset"):
+                # Берём самый большой размер (последний в srcset)
+                parts = val.split(",")
+                if parts:
+                    largest = parts[-1].strip().split(" ")[0]
+                    candidates.append(largest)
+            else:
+                candidates.append(val)
+
+        for src in candidates:
+            if not src or "data:image" in src:
+                continue
+            if src.startswith("//"):
+                src = "https:" + src
+            # Отсеиваем placeholder'ы Next.js (всегда заглушки)
+            if "_next/static/images" in src or "placeholder" in src.lower():
+                continue
+            # Реальные товарные фото — обычно на cs.citilink.ru / c.citilink.ru / images.citilink.ru
+            img_url = src
+            break
+        if img_url:
+            break
 
     # Рейтинг
     rating = None

backend-py/app/parsers/yamarket.py

@@ -138,20 +138,40 @@ def _extract_card(card, link_el, url: str) -> dict[str, Any] | None:
         except ValueError:
             pass
 
-    # Картинка (исключаем placeholder'ы)
+    # Картинка — приоритет: avatars.mds.yandex.net (реальные товарные фото)
     img_url = None
+    candidates = []
     for img_el in card.find_all("img"):
-        src = img_el.get("src") or img_el.get("data-src") or ""
+        for attr in ("data-src", "data-original", "srcset", "data-srcset", "src"):
+            v = img_el.get(attr) or ""
+            if not v:
+                continue
+            if attr in ("srcset", "data-srcset"):
+                # Берём самый большой размер
+                parts = v.split(",")
+                if parts:
+                    candidates.append(parts[-1].strip().split(" ")[0])
+            else:
+                candidates.append(v)
+    # Чистим и приоритизируем
+    for src in candidates:
         if not src or "data:image" in src:
-            srcset = img_el.get("srcset") or ""
-            if srcset:
-                src = srcset.split(",")[0].strip().split(" ")[0]
+            continue
         if src.startswith("//"):
             src = "https:" + src
-        if not src or "yastatic" in src or "_next/static" in src:
+        if "yastatic" in src:  # это иконки/логотипы
             continue
-        img_url = src
-        break
+        # avatars.mds.yandex.net — высший приоритет
+        if "avatars.mds.yandex" in src:
+            img_url = src
+            break
+        # Любой полноценный http(s) URL — fallback
+        if src.startswith("http") and not img_url:
+            img_url = src
+
+    # Если URL не содержит размерного суффикса — добавим /300x300
+    if img_url and "avatars.mds.yandex" in img_url and not re.search(r"/\d+x\d+/?$|/orig/?$", img_url):
+        img_url = img_url.rstrip("/") + "/300x300"
 
     # Рейтинг
     rating = None