Update app.py

app.py

@@ -1,6 +1,4 @@
 import os
-import time
-import json
 import gradio as gr
 from openai import OpenAI
 
@@ -10,11 +8,10 @@ API_HOSTS = {
 }
 
 MODELS_INFO = {
-    "gpt-3.5-turbo": {"input_price": "0.0035", "output_price": "0.0105", "features": "Default model, equivalent to gpt-3.5-turbo-0125"},
-    "o1-preview": {"input_price": "0.105", "output_price": "0.42", "features": "Powerful preview reasoning model"},
-    "gpt-4o": {"input_price": "0.0175", "output_price": "0.07", "features": "Cheaper and faster GPT-4O; supports image fee"},
-    "gpt-4-turbo": {"input_price": "0.07", "output_price": "0.21", "features": "Multimodal with image recognition, function tools support"},
-    "gpt-4o-ca": {"input_price": "0.01", "output_price": "0.04", "features": "Cheap CA variant, but limited stability/prior daily limit"}
+    "gpt-3.5-turbo": {"input_price": "0.0035", "output_price": "0.0105", "features": "Default, fast, affordable"},
+    "gpt-4o": {"input_price": "0.0175", "output_price": "0.07", "features": "Cheaper & faster GPT-4O"},
+    "gpt-4-turbo": {"input_price": "0.07", "output_price": "0.21", "features": "Multimodal, tool use"},
+    "gpt-4o-ca": {"input_price": "0.01", "output_price": "0.04", "features": "CA variant, daily free limit"},
 }
 
 def create_client(host):
@@ -37,11 +34,15 @@ def get_model_card(model_name):
 def respond(user, history, host_choice, model_name, temperature, top_p, max_tokens, sys_prompt):
     history = history or []
     if not user.strip():
-        return history + [("", "⚠️ Please enter a message.")]
+        yield history + [("", "⚠️ Please enter a message.")]
+        return
+
     try:
         client = create_client(API_HOSTS[host_choice])
     except Exception as e:
-        return history + [("", f"❌ {e}")]
+        yield history + [("", f"❌ {e}")]
+        return
+
     messages = [{"role": "system", "content": sys_prompt or "You are a helpful assistant."}]
     for u, a in history:
         messages.append({"role": "user", "content": u})
@@ -49,45 +50,64 @@ def respond(user, history, host_choice, model_name, temperature, top_p, max_toke
     messages.append({"role": "user", "content": user})
 
     try:
-        resp = client.chat.completions.create(
+        stream = client.chat.completions.create(
             model=model_name,
             messages=messages,
             temperature=temperature,
             top_p=top_p,
-            max_tokens=max_tokens
+            max_tokens=max_tokens,
+            stream=True
         )
-        out = resp.choices[0].message.content.strip() or "No response received."
+
+        partial = ""
+        history.append((user, partial))
+        yield history
+
+        for chunk in stream:
+            delta = chunk.choices[0].delta.content or ""
+            partial += delta
+            history[-1] = (user, partial)
+            yield history
+
     except Exception as e:
         err = str(e)
         if "429" in err:
             out = (
-                "🚫 Daily quota reached for the selected model (429).\n"
-                "Please try again after 00:00 (China time) or switch models/hosts."
+                "🚫 Daily quota reached for this model.\n"
+                "Please try again after 00:00 China time or switch model/host."
             )
         else:
             out = f"❌ API Error: {e}"
-    history.append((user, out))
-    return history
+        history.append((user, out))
+        yield history
 
-with gr.Blocks(title="ChatAnywhere-powered Chatbot", theme=gr.themes.Soft()) as demo:
-    gr.Markdown("##  ChatAnywhere Chatbot\nPowered by the ChatAnywhere API — Use wisely!")
+with gr.Blocks(title="ChatAnywhere Realtime Chatbot", theme=gr.themes.Soft()) as demo:
+    gr.Markdown("## 💬 ChatAnywhere Realtime Chatbot\nPowered by GPT-5 via ChatAnywhere API")
     with gr.Row():
         with gr.Column(scale=3):
-            chat = gr.Chatbot(label="Chat", height=500, show_copy_button=True, render_markdown=True)
-            msg = gr.Textbox(placeholder="Type your message...", lines=2)
-            clear = gr.Button("Clear")
+            chat = gr.Chatbot(label="Conversation", height=500, show_copy_button=True, render_markdown=True)
+            with gr.Row():
+                msg = gr.Textbox(placeholder="Type your message...", lines=2, scale=4)
+                send = gr.Button("Send", scale=1)
+                clear = gr.Button("Clear", scale=1)
         with gr.Column(scale=1):
             host = gr.Radio(list(API_HOSTS.keys()), value="Domestic", label="API Host")
             model = gr.Dropdown(list(MODELS_INFO.keys()), value="gpt-3.5-turbo", label="Model")
             model_card = gr.Markdown(get_model_card("gpt-3.5-turbo"))
             temperature = gr.Slider(0.0, 1.5, value=0.7, step=0.05, label="Temperature")
             top_p = gr.Slider(0.05, 1.0, value=1.0, step=0.05, label="Top-p")
-            max_tokens = gr.Slider(64, 8192, value=512, step=64, label="Max Tokens")
+            max_tokens = gr.Slider(64, 4096, value=512, step=64, label="Max Tokens")
             sys_prompt = gr.Textbox(label="System Prompt (optional)", lines=2)
-    msg.submit(respond, [msg, chat, host, model, temperature, top_p, max_tokens, sys_prompt], chat)
+
     model.change(lambda m: get_model_card(m), model, model_card)
-    clear.click(lambda: [], None, chat)
+
+    send.click(respond, [msg, chat, host, model, temperature, top_p, max_tokens, sys_prompt], chat)
+    msg.submit(respond, [msg, chat, host, model, temperature, top_p, max_tokens, sys_prompt], chat)
+
+    send.click(lambda _: "", msg, msg)
     msg.submit(lambda _: "", msg, msg)
 
+    clear.click(lambda: [], None, chat)
+
 if __name__ == "__main__":
     demo.launch()