fix: Fetch content using httpx rather than having feedparser do this, since we can't set a timeout with feedparser. This also uses a consistent user agent header when validating versus fetching the feed.

Author: kevinrenskers

example/uv.lock

@@ -1,6 +1,6 @@
 [project]
 name = "django-rss-filter"
-version = "0.6.0"
+version = "0.7.0"
 description = "Filter public RSS feeds, remove articles that contain certain keywords or categories."
 authors = [
     {name = "Kevin Renskers", email = "[email protected]"},

pyproject.toml

@@ -0,0 +1,7 @@
+# HTTP "User-Agent" header to send to servers when downloading feeds.
+# If you are embedding django-rss-filter in a larger application, you should
+# change this to your application name and URL.
+#
+# import rssfilter
+# rssfilter.USER_AGENT = "MyApp/1.0 +http://example.com/"
+USER_AGENT = "django-rss-filter/1.0 +https://github.com/loopwerk/django-rss-filter"

rssfilter/__init__.py

@@ -7,6 +7,7 @@
 from django.urls import reverse
 from django.utils import timezone
 
+from . import USER_AGENT
 from .settings import RSS_FILTER_CACHE_SECONDS
 from .utils import filter_feed, validate_feed
 
@@ -29,8 +30,9 @@ def clean(self):
         # Make sure we have a valid feed.
         # Let's assume it's valid if it's already in FeedCache.
         if not FeedCache.objects.filter(feed_url=self.feed_url).exists():
-            if not validate_feed(self.feed_url):
-                raise ValidationError({"feed_url": "This doesn't seem to be a valid RSS or Atom feed"})
+            valid, message = validate_feed(self.feed_url)
+            if not valid:
+                raise ValidationError({"feed_url": message})
 
     def get_filtered_feed_body(self) -> str:
         five_mins_ago = timezone.now() - timedelta(seconds=RSS_FILTER_CACHE_SECONDS)
@@ -59,7 +61,7 @@ def get_feed_body(self) -> str:
         if self.cache_date and self.cache_date > five_mins_ago:
             return self.feed_body
 
-        r = httpx.get(self.feed_url, follow_redirects=True)
+        r = httpx.get(self.feed_url, follow_redirects=True, timeout=2, headers={"User-Agent": USER_AGENT})
         self.feed_body = r.text
         self.cache_date = timezone.now()
         self.save()

rssfilter/models.py

@@ -1,16 +1,30 @@
 import feedparser
+import httpx
 from feedgen.feed import FeedGenerator
+from httpx import ConnectError, ConnectTimeout
 
+from . import USER_AGENT
 
-def validate_feed(feed_url: str) -> bool:
+
+def validate_feed(feed_url: str) -> tuple[bool, str]:
     try:
-        feed = feedparser.parse(feed_url)
+        # Fetch content using httpx rather than having feedparser do this,
+        # since we can't set a timeout with feedparser. It also makes sure
+        # that validating and then fetching the feed is done in a consistent
+        # manner.
+        r = httpx.get(feed_url, follow_redirects=True, timeout=2, headers={"User-Agent": USER_AGENT})
+
+        feed = feedparser.parse(r.text)
         version = feed.get("version", "")
         if not version:
-            return False
-        return True
+            return False, "This doesn't seem to be a valid RSS or Atom feed"
+        return True, ""
     except ValueError:
-        return False
+        return False, "This doesn't seem to be a valid RSS or Atom feed"
+    except ConnectTimeout:
+        return False, "Couldn't load the URL due to a connection timeout"
+    except ConnectError:
+        return False, "Couldn't load the URL due to a connection error"
 
 
 def filter_feed(feed_body: str, filtered_words: str, filtered_categories: str) -> str: