Update app.py

app.py

@@ -1,6 +1,7 @@
 import gradio as gr
 from openai import OpenAI
 import os
+import time
 
 HOSTS = {
     "Domestic (Lower Latency)": "https://api.chatanywhere.tech/v1",
@@ -8,38 +9,86 @@ HOSTS = {
 }
 
 def create_client(base_url):
-    return OpenAI(
-        api_key=os.getenv("OPENAI_API_KEY"),
-        base_url=base_url
-    )
+    api_key = os.getenv("OPENAI_API_KEY")
+    if not api_key:
+        raise ValueError("Missing API key. Please set OPENAI_API_KEY in your environment.")
+    return OpenAI(api_key=api_key, base_url=base_url)
 
 def respond_stream(user_message, history, host_choice, temperature):
     history = history or []
+
+    # Empty message handling
+    if not user_message.strip():
+        yield history + [("", "⚠️ Please enter a message before sending.")]
+        return
+
+    # Prepare conversation context
     messages = [{"role": "system", "content": "You are GPT-5, a helpful, friendly, and concise assistant."}]
     for human, ai in history:
         messages.append({"role": "user", "content": human})
         messages.append({"role": "assistant", "content": ai})
     messages.append({"role": "user", "content": user_message})
 
-    client = create_client(HOSTS[host_choice])
+    # API Host handling
+    try:
+        client = create_client(HOSTS[host_choice])
+    except ValueError as e:
+        yield history + [(user_message, f"❌ {e}")]
+        return
+    except Exception as e:
+        yield history + [(user_message, f"❌ Failed to initialize client: {e}")]
+        return
+
+    # API request with streaming
     try:
         with client.chat.completions.stream(
             model="gpt-5",
             messages=messages,
-            temperature=temperature
+            temperature=temperature,
+            timeout=30  # network timeout
         ) as stream:
             partial = ""
+            last_update_time = time.time()
+
             for event in stream:
                 if event.type == "message.delta" and event.delta.content:
                     partial += event.delta.content
-                    yield history + [(user_message, partial)]
+
+                    # Update UI at most 5 times per second
+                    if time.time() - last_update_time > 0.2:
+                        yield history + [(user_message, partial)]
+                        last_update_time = time.time()
+
             history.append((user_message, partial))
             yield history
+
     except Exception as e:
-        yield history + [(user_message, f"❌ Error: {e}")]
+        # Retry with alternate host
+        try:
+            alt_host = HOSTS["Overseas"] if host_choice == "Domestic (Lower Latency)" else HOSTS["Domestic (Lower Latency)"]
+            client = create_client(alt_host)
+
+            with client.chat.completions.stream(
+                model="gpt-5",
+                messages=messages,
+                temperature=temperature,
+                timeout=30
+            ) as stream:
+                partial = ""
+                for event in stream:
+                    if event.type == "message.delta" and event.delta.content:
+                        partial += event.delta.content
+                        yield history + [(user_message, partial)]
+
+                history.append((user_message, partial))
+                yield history
+
+        except Exception as e2:
+            yield history + [(user_message, f"❌ Both API hosts failed.\nHost1 Error: {e}\nHost2 Error: {e2}")]
 
+# UI Layout
 with gr.Blocks(title="GPT-5 Chatbot", theme=gr.themes.Soft()) as demo:
-    gr.Markdown("<h1 style='text-align:center'>💬 GPT-5 Chatbot</h1><p style='text-align:center'>Fast, interactive, and pretty</p>")
+    gr.Markdown("<h1 style='text-align:center'>💬 GPT-5 Chatbot</h1><p style='text-align:center'>Fast, resilient, and beautiful</p>")
 
     with gr.Row():
         with gr.Column(scale=3):