Update feed_to_llm_v2.py

feed_to_llm_v2.py

@@ -62,16 +62,19 @@ def feed_articles_to_gpt_with_links(information, question):
     print("LLM Response Content:", response_content)
 
     # Use regex to extract Sources field
+    # Extract sources from the response content
     sources_pattern = r"Sources:\s*([\d,]+)"
     sources_match = re.search(sources_pattern, response_content)
+    
     if sources_match:
-        # Extract the list of article numbers
-        source_numbers = sources_match.group(1).split(',')
-        used_article_num = [int(num.strip()) - 1 for num in source_numbers]
+        # Split the matched source numbers and filter out any empty strings
+        source_numbers = [num.strip() for num in sources_match.group(1).split(',') if num.strip()]
+        used_article_num = [int(num) - 1 for num in source_numbers]
     else:
         print("No sources found in the response.")
         return response_content, [], [], []
 
+
     # Prepare links, titles, and domains for the cited articles
     links = [f"https://tobaccowatcher.globaltobaccocontrol.org/articles/{uuid}/" for uuid in uuids]
     titles = [titles for score, contents, uuids, titles, domains in information]