Update handler.py

- Main() function -Now fetches from both Bleeping Computer and Hacker News, combines the articles.

- Handles Hacker News' 'GMT' timezone by converting it to '+0000', allowing correct date parsing.

- After processing all articles, filters links to include only those from Bleeping Computer (based on URL) before sending to DynamoDB.

Author: Deadpool2000

lambdas/security-newsletter/src/handler.py

@@ -21,11 +21,23 @@ def main(event, _):
     """
     logging.info("Event: %s", event)
 
-    latest_articles = get_latest_article_with_timezone(fetch_bleeping_computer_rss())
+    # Fetch articles from both feeds
+    bleeping_articles = fetch_bleeping_computer_rss()
+    hacker_articles = fetch_hacker_news_rss()
+
+    # Combine articles from both feeds
+    all_articles = bleeping_articles + hacker_articles
+
+    # Get today's articles from the combined list
+    latest_articles = get_latest_article_with_timezone(all_articles)
 
     logging.info("Latest articles: %s", latest_articles)
-    links = [article["link"] for article in latest_articles]
-    return publish_message_to_table(links)
+    
+    # Extract links from Bleeping Computer articles only
+    bleeping_links = [article["link"] for article in latest_articles 
+                      if "bleepingcomputer.com" in article["link"]]
+    
+    return publish_message_to_table(bleeping_links)
 
 
 def get_latest_article_with_timezone(articles, timezone_str="UTC"):
@@ -37,48 +49,58 @@ def get_latest_article_with_timezone(articles, timezone_str="UTC"):
     todays_articles = []
     for article in articles:
         date_str = article["published"]
-        parsed_date = datetime.strptime(date_str, "%a, %d %b %Y %H:%M:%S %z")
-        formatted_date = parsed_date.isoformat()
-        pub_date = datetime.fromisoformat(formatted_date.replace("Z", "+00:00"))
-        pub_date = pub_date.astimezone(tz)
+        # Handle 'GMT' timezone suffix
+        if date_str.endswith(' GMT'):
+            date_str = date_str.replace(' GMT', ' +0000')
+        try:
+            parsed_date = datetime.strptime(date_str, "%a, %d %b %Y %H:%M:%S %z")
+        except ValueError as e:
+            logging.error("Error parsing date '%s': %s", date_str, e)
+            continue
+        pub_date = parsed_date.astimezone(tz)
         if pub_date.date() == today:
             todays_articles.append(article)
-
     return todays_articles
 
 
 def fetch_bleeping_computer_rss(feed_url="https://www.bleepingcomputer.com/feed/"):
     """Fetch articles from the Bleeping Computer RSS feed."""
-    # Parse the RSS feed
     feed = feedparser.parse(feed_url)
-
     if feed.bozo:
         raise ValueError(f"Error parsing feed: {feed.bozo_exception}")
-
-    # Extract articles
     articles = []
     for entry in feed.entries:
-        articles.append(
-            {
-                "title": entry.title,
-                "link": entry.link,
-                "published": entry.published,
-                "summary": entry.summary,
-            }
-        )
-
+        articles.append({
+            "title": entry.title,
+            "link": entry.link,
+            "published": entry.published,
+            "summary": entry.summary,
+        })
     return articles
 
+def fetch_hacker_news_rss(feed_url="https://feeds.feedburner.com/TheHackersNews"):
+    """Fetch articles from The Hacker News RSS feed."""
+    feed = feedparser.parse(feed_url)
+    if feed.bozo:
+        raise ValueError(f"Error parsing feed: {feed.bozo_exception}")
+    h_articles = []
+    for entry in feed.entries:
+        h_articles.append({
+            "title": entry.title,
+            "link": entry.link,
+            "published": entry.published,
+            "summary": entry.summary,
+        })
+    return h_articles
+
 
 def publish_message_to_table(links: str):
     """
-    Sends a message to the SQS queue.
+    Sends a message to the DynamoDB table.
     Returns a dictionary with the message and the message ID.
     """
-    logging.info("Sending message to SQS queue")
-
+    logging.info("Sending message to DynamoDB table")
     dynamodb_client = boto3.client("dynamodb")
-
     for link in links:
         logging.info("Link: %s", link)
         response = dynamodb_client.put_item(
@@ -92,5 +114,4 @@ def publish_message_to_table(links: str):
             },
         )
         logging.info("Response: %s", response)
-
     return {"message": "Message has been logged to DynamoDB!", "links": links}