Update flask_app.py

flask_app.py

@@ -165,14 +165,21 @@ def model_inference(retriever, question, llm):
     match = re.findall(r"Answer:\s*(.*)", response, re.DOTALL)
     final_answer = "Answer:\n        " + match[-1].strip() if match else "Answer:\n        The answer is not found in the context provided."
 
-    # Extract only the first relevant reference
+    
+    # Extract document title by removing file extension and trailing numbers/versions
+    def extract_title(file_name):
+        base_name = os.path.splitext(file_name)[0]  # Remove file extension
+        title = re.sub(r'\s*(final\d*|v\d+|rev\d+)$', '', base_name, flags=re.IGNORECASE)
+        return title.strip()
+    
     references = (
-        f"Source: {retrieved_docs[0].metadata['source']}, Page: {retrieved_docs[0].metadata.get('page_label', 'Unknown')}"
+        f"Source: {extract_title(retrieved_docs[0].metadata['source'])}"
         if retrieved_docs else "No references found."
     )
-
+    
     return f"{final_answer}\nReferences: {references}\nContext: {context}"
 
+
 # # Example usage:
 # formatted_output = model_inference(retriever, "What is the main purpose of the Khyber Pakhtunkhwa Arms Act, 2013?", llm)