aa

gurt/api.py

@@ -218,17 +218,21 @@ async def get_ai_response(cog: 'GurtCog', message: discord.Message, model_name:
         # --- 1. Build System Prompt ---
         # This now includes dynamic context like mood, facts, etc.
         system_prompt_text = await build_dynamic_system_prompt(cog, message)
-        # Create a LangChain prompt template
-        # LangchainAgent default prompt includes placeholders for chat_history and agent_scratchpad
-        # We just need to provide the system message part.
-        # Note: The exact structure might depend on the agent type used by LangchainAgent.
-        # Assuming a standard structure:
+        # Create a LangChain prompt template using explicit message types
+        # This might be more robust against parsing issues in the system prompt text itself.
         prompt_template = ChatPromptTemplate.from_messages([
-            ("system", system_prompt_text),
-            MessagesPlaceholder(variable_name="chat_history"), # Standard placeholder
-            ("user", "{input}"), # User input placeholder
-            MessagesPlaceholder(variable_name="agent_scratchpad"), # Tool execution placeholder
+            SystemMessage(content=system_prompt_text), # Pass the generated system prompt text here
+            MessagesPlaceholder(variable_name="chat_history"),
+            HumanMessage(content="{input}"),
+            MessagesPlaceholder(variable_name="agent_scratchpad"),
         ])
+        # Validate the template's input variables explicitly
+        expected_vars = {"input", "chat_history", "agent_scratchpad"}
+        if set(prompt_template.input_variables) != expected_vars:
+             logger.error(f"Prompt template validation failed! Expected variables: {expected_vars}, Got: {prompt_template.input_variables}")
+             # Handle error appropriately, maybe raise or return error response
+             raise ValueError("Prompt template validation failed.")
+
 
         # --- 2. Prepare Tools ---
         # Wrap tools from TOOL_MAPPING to include the 'cog' instance
@@ -741,4 +745,4 @@ async def get_internal_ai_json_response(
         except Exception as log_e:
             logger.error(f"Error logging internal API call: {log_e}", exc_info=True)
 
-    return final_parsed_data
+    return final_parsed_data