update domains pattern

olp966220 · Dec 20, 2024 · dc20476 · dc20476
1 parent 6dfcdc6
commit dc20476
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/subscribe/crawl.py b/subscribe/crawl.py
@@ -1519,7 +1519,7 @@ def get_links(url: str, prefix: str) -> list[str]:
             return {}
 
         separator = r'<h2 id="\d+" tabindex="-1">'
-        address_regex = r'<a href="(https?://[^\s]+)" target="_blank" rel="noreferrer">前往注册</a>'
+        address_regex = r'<a href="(https?://[^\s]+)" target="_blank" rel="noreferrer nofollow">前往注册</a>'
         coupon_regex = r"使用优惠码(?:\s+)?(?:<code>)?([^\r\n\s]+)(?:</code>(?:[\r\n\s]+)?)?0(?:\s+)?元购买"
 
         tasks = [[x, separator, address_regex, coupon_regex] for x in sorted(articles)]