Update app.py

app.py

@@ -6,11 +6,12 @@ API_URL = "http://0.0.0.0:8000/v1/chat/completions"
 def stream_completion(message, history, max_tokens, temperature, top_p, beta):
     """Gradio callback: stream the assistant’s reply token-by-token."""
     # -------- build OpenAI-style message list (no system prompt) -------------
-    messages = [{"role": "user", "content": u}         # past user turns
-                if i % 2 == 0 else                    # even idx → user
-                {"role": "assistant", "content": u}    # odd  idx → assistant
-                for i, (u, _) in enumerate(sum(([h[0], h[1]] for h in history), []))
-                if u]                                  # drop empty strings
+    for user_msg, assistant_msg in history:
+        if user_msg:        # past user turn
+            messages.append({"role": "user", "content": user_msg})
+        if assistant_msg:   # past assistant turn
+            messages.append({"role": "assistant", "content": assistant_msg})
+
     messages.append({"role": "user", "content": message})
 
     payload = {