Update app.py

app.py

@@ -18,7 +18,7 @@ def load_llm():
     )
     llm = Llama(
         model_path=model_path,
-        n_ctx=512,               
+        n_ctx=512,
         n_threads=max(2, os.cpu_count() or 2),
         n_gpu_layers=0,
         n_batch=32,
@@ -29,30 +29,36 @@ def load_llm():
 SYSTEM_PROMPT = "به فارسی، روشن و کوتاه پاسخ بده (حداکثر ۲ جمله)."
 
 def respond(message, history):
-   
-    messages = [{"role": "system", "content": SYSTEM_PROMPT}]
-    for user, assistant in history:
-        messages.append({"role": "user", "content": user})
-        messages.append({"role": "assistant", "content": assistant})
-    messages.append({"role": "user", "content": message})
+    # تاریخچه می‌تونه tuples یا messages باشه؛ هر دو رو پشتیبانی می‌کنیم
+    msgs = [{"role": "system", "content": SYSTEM_PROMPT}]
+    if history and isinstance(history[0], dict) and "role" in history[0]:
+        msgs.extend(history)  # already messages format
+    else:
+        for user, assistant in (history or []):
+            msgs.append({"role": "user", "content": user})
+            msgs.append({"role": "assistant", "content": assistant})
+    msgs.append({"role": "user", "content": message})
 
     llm = load_llm()
+    print(">> gen start")  # برای لاگ
     stream = llm.create_chat_completion(
-        messages=messages,      
-        max_tokens=60,
+        messages=msgs,
+        max_tokens=64,
         temperature=0.4,
         top_p=0.9,
         stream=True,
     )
-
     partial = ""
     for chunk in stream:
         delta = chunk["choices"][0]["delta"].get("content", "")
-        partial += delta
-        yield partial
+        if delta:
+            partial += delta
+            yield partial
+    print(">> gen done")   # برای لاگ
 
 demo = gr.ChatInterface(
     fn=respond,
+    type="messages",  # هشدار Gradio هم برطرف می‌شود
     title="چت‌بات خیلی ساده (CPU رایگان)",
     description="SmolLM2-135M (GGUF) با llama.cpp روی CPU.",
 )