properly feed in conversation history

vibe_bot/main.py

@@ -416,22 +416,19 @@ async def handle_chat(
         user_id=str(ctx.author.id), current_message=message, max_context=5
     )
 
+    prompts = [{"role": "user", "content": message}]
+
     if context:
-        user_message = f"\n\nRelevant conversation history:\n{context}\n\n{message}"
-    else:
-        user_message = message
+        prompts = context + prompts
 
-    logger.info(user_message)
+    logger.info(prompts)
 
-    system_prompt_edit = (
-        "Keep your responses somewhat short, limited to 500 words or less. "
-        f"{system_prompt}"
-    )
+    system_prompt_edit = f"{system_prompt}\nKeep your responses under 2-3 sentences."
 
     try:
-        bot_response = llama_wrapper.chat_completion_instruct(
+        bot_response = llama_wrapper.chat_completion_with_history(
             system_prompt=system_prompt_edit,
-            user_prompt=user_message,
+            prompts=prompts,  # type: ignore
             openai_url=CHAT_ENDPOINT,
             openai_api_key=CHAT_ENDPOINT_KEY,
             model=CHAT_MODEL,