From d0cabc3f13147a7a337c16eb177f4612675b2c32 Mon Sep 17 00:00:00 2001
From: JOUNGWOOK KWON <elikwon@JOUNGWOOKui-MacBookAir.local>
Date: Mon, 30 Mar 2026 13:44:53 +0900
Subject: [PATCH] =?UTF-8?q?fix:=20og:image=EC=97=90=EC=84=9C=20=ED=94=8C?=
 =?UTF-8?q?=EB=9E=AB=ED=8F=BC=20=EB=A1=9C=EA=B3=A0(Google=EB=89=B4?=
 =?UTF-8?q?=EC=8A=A4=20=EB=93=B1)=20=ED=95=84=ED=84=B0=EB=A7=81?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- _is_platform_logo(): 로고/아이콘/기본이미지 패턴 감지
- Google 뉴스 URL인 경우 실제 기사 URL로 리다이렉트 추적
- 로고 이미지 걸러지면 Pexels 폴백으로 진행

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---
 bots/publisher_bot.py | 30 +++++++++++++++++++++++++++---
 1 file changed, 27 insertions(+), 3 deletions(-)

diff --git a/bots/publisher_bot.py b/bots/publisher_bot.py
index 117d39d..d3f3193 100644
--- a/bots/publisher_bot.py
+++ b/bots/publisher_bot.py
@@ -206,10 +206,32 @@ def build_json_ld(article: dict, blog_url: str = '') -> str:
     return f'<script type="application/ld+json">\n{json.dumps(schema, ensure_ascii=False, indent=2)}\n</script>'
 
 
+def _is_platform_logo(image_url: str) -> bool:
+    """플랫폼 로고/아이콘 이미지인지 판별 — 대표 이미지로 부적합"""
+    skip_patterns = [
+        'logo', 'icon', 'avatar', 'banner', '/ad/',
+        'google.com/images/branding', 'googlenews', 'google-news',
+        'facebook.com', 'twitter.com', 'naver.com/favicon',
+        'default_image', 'placeholder', 'noimage', 'no-image',
+        'og-default', 'share-default', 'sns_', 'common/',
+    ]
+    url_lower = image_url.lower()
+    return any(p in url_lower for p in skip_patterns)
+
+
 def _fetch_og_image(url: str) -> str:
     """원본 기사 URL에서 og:image 메타태그 크롤링"""
     if not url or not url.startswith('http'):
         return ''
+    # Google 뉴스 리다이렉트인 경우 실제 기사 URL 추출 시도
+    if 'news.google.com' in url:
+        try:
+            resp = requests.head(url, timeout=10, allow_redirects=True,
+                                 headers={'User-Agent': 'Mozilla/5.0'})
+            if resp.url and 'news.google.com' not in resp.url:
+                url = resp.url
+        except Exception:
+            pass
     try:
         resp = requests.get(url, timeout=10, headers={
             'User-Agent': 'Mozilla/5.0 (compatible; BlogBot/1.0)',
@@ -220,15 +242,17 @@ def _fetch_og_image(url: str) -> str:
         # og:image
         og = soup.find('meta', property='og:image')
         if og and og.get('content', '').startswith('http'):
-            return og['content']
+            if not _is_platform_logo(og['content']):
+                return og['content']
         # twitter:image
         tw = soup.find('meta', attrs={'name': 'twitter:image'})
         if tw and tw.get('content', '').startswith('http'):
-            return tw['content']
+            if not _is_platform_logo(tw['content']):
+                return tw['content']
         # 본문 첫 번째 큰 이미지
         for img in soup.find_all('img', src=True):
             src = img['src']
-            if src.startswith('http') and not any(x in src.lower() for x in ['logo', 'icon', 'avatar', 'banner', 'ad']):
+            if src.startswith('http') and not _is_platform_logo(src):
                 return src
     except Exception as e:
         logger.warning(f"og:image 크롤링 실패 ({url}): {e}")