Update app.py

app.py

@@ -42,7 +42,7 @@ def retrieve(query, vectorstore, top_k=5):
 class ChatRequest(BaseModel):
     message: str
     system_message: str = """You are Dan Infalt, a friendly public land deer hunting expert specializing in targeting mature bucks in pressured areas, but 
-    don’t worry, you won’t take yourself too seriously. You respond in a conversational matter. And you have dry humor you mix in every once in a while.
+    don’t worry, you won’t take yourself too seriously. You respond in a conversational matter but still direct. You have dry humor you mix in every once in a while.
     You focus on buck bedding, terrain reading, and aggressive yet calculated mobile tactics. Your blue-collar, no-nonsense approach 
     emphasizes deep scouting, strategic access, and minimalist setups. Through The Hunting Beast, you teach hunters how to kill big bucks 
     using terrain, wind, and thermals. You speak from firsthand experience, keeping your advice practical and to the point. Provide detailed 
@@ -103,20 +103,20 @@ async def chat(request: ChatRequest):
             results = retrieve(request.message, vectorstore, top_k=5)
             formatted_results = "\n\n".join(results)
 
-            rag_prompt = textwrap.dedent(f"""You have access to the following relevant information retrieved based on the user's query:
+            rag_prompt = f"""You have access to the following relevant information retrieved based on the user's query:
             
             {formatted_results}
 
             Using the information above, answer the user's query as accurately as possible:
 
             User's Query: {request.message}
-            """)
+            """
             
             # remove the unfformatted user message     
             del request.chat_history[-1]
             
             # add the user message with RAG data
-            request.chat_history.append({"role": "user", "parts": [{"text": rag_prompt}]})
+            request.chat_history.append({"role": "user", "parts": [{"text": textwrap.dedent(rag_prompt)}]})
 
             response = client.models.generate_content(
                 model="gemini-2.0-flash",