Update app.py

app.py

@@ -125,17 +125,22 @@ class MultiAgentRAG:
         return rewritten_query.strip()
 
     def generation_agent(self, query, retrieved_content):
-        generation_prompt = f"""
-        Using the following information retrieved from the knowledge base:
-
-        {retrieved_content}
-
-        Provide a comprehensive answer to the question below:
-
-        Question: {query}
-        Answer:
-        """
-        input_ids = self.tokenizer.encode(generation_prompt, return_tensors="pt").to(self.model.device)
+        conversation = [
+            {"role": "system", "content": "You are a knowledgeable assistant with access to a comprehensive database."},
+            {"role": "user", "content": f"""
+            I need you to answer my question and provide related information in a specific format.
+            I have provided five relatable json files {retrieved_content}, choose the most suitable chunks for answering the query.
+            RETURN ONLY SOLUTION without additional comments, sign-offs, retrived chunks, refrence to any Ticket or extra phrases. Be direct and to the point.
+            IF THERE IS NO ANSWER RELATABLE IN RETRIEVED CHUNKS, RETURN "NO SOLUTION AVAILABLE".
+            DO NOT GIVE REFRENCE TO ANY CHUNKS OR TICKETS,BE ON POINT.
+            
+            Here's my question:
+            Query: {query}
+            Solution==>
+            """}
+        ]
+        
+        input_ids = self.tokenizer.encode(self.tokenizer.apply_chat_template(conversation, tokenize=False), return_tensors="pt").to(self.model.device)
         return self.generate_response_with_timeout(input_ids)
 
     def run_multi_agent_rag(self, query):