From c97b8dce3cf3a6efabc00645b924ca4dde593085 Mon Sep 17 00:00:00 2001
From: wasrusgen <vasrusgen@gmail.com>
Date: Mon, 11 May 2026 17:20:59 +0300
Subject: [PATCH] =?UTF-8?q?parsers:=20skip=20sponsored/ad=20URLs=20(cpc/sp?=
 =?UTF-8?q?onsored=3D1)=20=E2=80=94=20they=20expire=20in=202-3=20hours?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

User reported clicking matrix prices led to 'Произошла ошибка!' on OZON home page.
Cause: parsers captured /product/?sponsored=1&cpc=Jtiito95... links that died after few hours.

Fix:
- ozon.py: skip href with 'sponsored=1', '/promo/', 'cpc='. Strip query string from final URL.
- yamarket.py: skip 'sponsored=1', 'cpc=', 'advUuid' (Я.Маркет sponsored marker)
- citilink.py: strip query string from final URL (defensive)

Now matrix links go to canonical product pages that don't expire.
---
 backend-py/app/parsers/citilink.py | 4 +++-
 backend-py/app/parsers/ozon.py     | 8 ++++++--
 backend-py/app/parsers/yamarket.py | 3 +++
 3 files changed, 12 insertions(+), 3 deletions(-)

diff --git a/backend-py/app/parsers/citilink.py b/backend-py/app/parsers/citilink.py
index 53842d1..03d7b5a 100644
--- a/backend-py/app/parsers/citilink.py
+++ b/backend-py/app/parsers/citilink.py
@@ -62,7 +62,9 @@ def _parse_html(html: str, limit: int) -> list[dict[str, Any]]:
             continue
         seen_urls.add(product_id)
 
-        full_url = href if href.startswith("http") else f"{_BASE_URL}{href}"
+        # Финальный URL — БЕЗ query params (sponsored / cpc / tracking)
+        href_clean = href.split("?")[0]
+        full_url = href_clean if href_clean.startswith("http") else f"{_BASE_URL}{href_clean}"
 
         # Поднимаемся к родительской карточке — у Citilink CSS-in-JS, поэтому
         # ищем ближайший div, в котором есть и цена и название
diff --git a/backend-py/app/parsers/ozon.py b/backend-py/app/parsers/ozon.py
index 396c226..ff4d0c3 100644
--- a/backend-py/app/parsers/ozon.py
+++ b/backend-py/app/parsers/ozon.py
@@ -64,13 +64,17 @@ def _parse_html(html: str, limit: int) -> list[dict[str, Any]]:
             break
 
         href = link.get("href") or ""
-        # Нормализация URL — убираем query params для дедупа
+        # Пропускаем спонсорные ссылки — они истекают через 2-3 часа
+        if "sponsored=1" in href or "/promo/" in href or "cpc=" in href:
+            continue
+        # Чистим URL — убираем все query-параметры
         url_clean = href.split("?")[0]
         if url_clean in seen_urls:
             continue
         seen_urls.add(url_clean)
 
-        full_url = href if href.startswith("http") else f"{_BASE_URL}{href}"
+        # Финальный URL — БЕЗ query params (sponsored ссылки иначе через 2-3ч 404)
+        full_url = url_clean if url_clean.startswith("http") else f"{_BASE_URL}{url_clean}"
 
         # Поднимаемся до карточки — у OZON это обычно ближайший div с tile-* классом
         card = (
diff --git a/backend-py/app/parsers/yamarket.py b/backend-py/app/parsers/yamarket.py
index 00d7269..b1e653b 100644
--- a/backend-py/app/parsers/yamarket.py
+++ b/backend-py/app/parsers/yamarket.py
@@ -67,6 +67,9 @@ def _parse_html(html: str, limit: int) -> list[dict[str, Any]]:
         if len(results) >= limit:
             break
         href = link.get("href") or ""
+        # Пропускаем sponsored — их URL'ы с CPC-токенами истекают через несколько часов
+        if "sponsored=1" in href or "cpc=" in href or "advUuid" in href:
+            continue
         m_id = re.search(r"/card/[^/]+/(\d+)", href)
         if not m_id:
             continue