From aa569a8ed13c0db16c88a2533d360281b2161ab8 Mon Sep 17 00:00:00 2001
From: wasrusgen <vasrusgen@gmail.com>
Date: Mon, 11 May 2026 16:34:04 +0300
Subject: [PATCH] dns: switch to Playwright (Qrator JS challenge); ozon: fix
 false-positive antibot detector
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- DNS: использовали httpx + proxy_pool но Qrator кидал 401 даже с residential
  → теперь Playwright + residential — браузер сам решает JS challenge
- OZON: теперь проверяем только <title>='Доступ ограничен' (точная), а не подстроку '/robotcheck/'
---
 backend-py/app/parsers/dns.py  | 49 +++++++++++++++-------------------
 backend-py/app/parsers/ozon.py |  9 ++++---
 r.json                         |  0
 3 files changed, 28 insertions(+), 30 deletions(-)
 create mode 100644 r.json

diff --git a/backend-py/app/parsers/dns.py b/backend-py/app/parsers/dns.py
index 50fd6c4..a512f0a 100644
--- a/backend-py/app/parsers/dns.py
+++ b/backend-py/app/parsers/dns.py
@@ -17,6 +17,7 @@ import httpx
 from bs4 import BeautifulSoup
 
 from .. import proxy_pool
+from . import playwright_engine
 
 log = logging.getLogger("zov.parser.dns")
 
@@ -44,41 +45,35 @@ _HEADERS = {
 _PRICE_RE = re.compile(r"(\d[\d\s]*)\s*₽")
 
 
-def search_dns(query: str, limit: int = 1, timeout: float = 12.0,
-               max_retries: int = 2) -> list[dict[str, Any]]:
-    """Поиск товара на DNS по строке запроса.
+def search_dns(query: str, limit: int = 1, timeout: float = 30.0,
+               max_retries: int = 1) -> list[dict[str, Any]]:
+    """Поиск на DNS через Playwright + residential proxy.
 
-    Использует Proxy6-пул если PROXY6_TOKEN задан, иначе ходит напрямую.
-    DNS защищён Qrator — без прокси скорее всего 401.
-
-    Возвращает список результатов (топ-N) или пустой при ошибке.
+    DNS защищён Qrator (JS challenge) — обычный HTTP не пройдёт даже с прокси.
+    Playwright решает challenge автоматически (как реальный браузер).
     """
     url = f"{_SEARCH_URL}?q={quote_plus(query)}"
     log.info("DNS search: %s", url)
 
-    last_err = None
+    html = None
     for attempt in range(max_retries + 1):
-        try:
-            with proxy_pool.proxied_client(timeout=timeout, headers=_HEADERS,
-                                            follow_redirects=True) as client:
-                resp = client.get(url)
-        except httpx.HTTPError as e:
-            last_err = e
-            log.warning("DNS request failed (attempt %d): %s", attempt + 1, e)
-            continue
+        html = playwright_engine.fetch_page(
+            url,
+            wait_selector="a[href*='/product/']",
+            wait_ms=5000,
+            timeout_ms=int(timeout * 1000),
+        )
+        if html:
+            break
 
-        if resp.status_code == 200:
-            text = resp.text
-            if "qrator" in text.lower() or "challenge" in text.lower() or "captcha" in text.lower():
-                log.warning("DNS Qrator/captcha on attempt %d, rotating proxy", attempt + 1)
-                continue
-            return _parse_search_html(text, limit=limit)
+    if not html:
+        log.warning("DNS: no HTML for query=%r", query)
+        return []
+    if "qrator" in html.lower()[:5000]:
+        log.warning("DNS: Qrator block for query=%r", query)
+        return []
 
-        log.warning("DNS returned status=%s on attempt %d", resp.status_code, attempt + 1)
-
-    log.warning("DNS gave up after %d attempts for query=%r (last_err=%s)",
-                max_retries + 1, query, last_err)
-    return []
+    return _parse_search_html(html, limit=limit)
 
 
 def _parse_search_html(html: str, limit: int) -> list[dict[str, Any]]:
diff --git a/backend-py/app/parsers/ozon.py b/backend-py/app/parsers/ozon.py
index 8defe65..34a2fbc 100644
--- a/backend-py/app/parsers/ozon.py
+++ b/backend-py/app/parsers/ozon.py
@@ -40,9 +40,12 @@ def search_ozon(query: str, limit: int = 3, timeout: float = 30.0,
     if not html:
         log.warning("OZON: no HTML for query=%r", query)
         return []
-    # Реальный anti-bot — это редирект на /robotcheck/ или специальная страница
-    if "/robotcheck/" in html or "Доступ ограничен" in html[:5000]:
-        log.warning("OZON: anti-bot block for query=%r", query)
+    # OZON показывает "Доступ ограничен" как <title>, либо редирект на /antibot/
+    import re as _re
+    title_m = _re.search(r"<title>(.*?)</title>", html, _re.IGNORECASE)
+    page_title = title_m.group(1) if title_m else ""
+    if "доступ ограничен" in page_title.lower() or "/antibot/" in html[:5000]:
+        log.warning("OZON: anti-bot page (title=%r)", page_title)
         return []
 
     return _parse_html(html, limit=limit)
diff --git a/r.json b/r.json
new file mode 100644
index 0000000..e69de29