Update agent_engine.py

agent_engine.py

@@ -19,38 +19,12 @@ HF_API_TOKEN = os.environ.get("HF_API_TOKEN", "")
 logging.basicConfig(level=logging.INFO, format="%(asctime)s %(levelname)s %(message)s")
 
 
-def safe_chat(system_prompt: str, history: list, temperature: float = 0.7) -> str:
-    """Call HF inference API with timing and error handling."""
-    start = perf_counter()
-    payload = {
-        "inputs": [{"role": "system", "content": system_prompt}] + history,
-        "parameters": {"max_new_tokens": 300, "temperature": temperature}
-    }
-    headers = {"Authorization": f"Bearer {HF_API_TOKEN}"} if HF_API_TOKEN else {}
-    try:
-        resp = requests.post(
-            f"https://api-inference.huggingface.co/models/{CHAT_MODEL}",
-            json=payload,
-            headers=headers,
-            timeout=60
-        )
-        if resp.status_code == 200:
-            data = resp.json()
-            text = data[0].get('generated_text', '').strip()
-        elif resp.status_code == 503:
-            logging.warning("Model loading, retrying...")
-            sleep(15)
-            return safe_chat(system_prompt, history, temperature)
-        else:
-            logging.error(f"HF error {resp.status_code}: {resp.text}")
-            text = f"⚠️ API Error {resp.status_code}"
-    except Exception as e:
-        logging.error(f"safe_chat exception: {e}")
-        text = f"⚠️ System Error: {e}"
-    elapsed = perf_counter() - start
-    logging.info(f"safe_chat: {elapsed:.3f}s for '{system_prompt[:30]}...'")
-    return text
-
+chat_history = "\n".join([f"{msg['role'].capitalize()}: {msg['content']}" for msg in history])
+full_prompt = f"{system_prompt}\n\n{chat_history}\n\nAssistant:"
+payload = {
+    "inputs": full_prompt,
+    "parameters": {"max_new_tokens": 300, "temperature": temperature}
+}
 
 def step_turn(conversation: list, turn: int, topic: str, params: dict) -> list:
     """Advance one turn of the multi-agent conversation."""