Update app.py

app.py

@@ -1,41 +1,60 @@
-In this app, you can explore the outputs of a gpt-4 turbo LLM.
-"""
-
-theme = gr.themes.Default(primary_hue="green")
-
-with gr.Blocks(css="""#col_container { margin-left: auto; margin-right: auto;}
-                #chatbot {height: 520px; overflow: auto;}""",
-              theme=theme) as demo:
-    gr.HTML(title)
-    gr.HTML("""<h3 align="center">This app provides you full access to GPT-4.1 mini (1M token limit). You don't need any OPENAI API key.</h3>""")
-    gr.HTML("""
-<div style="padding: 12px; margin-bottom: 16px; border-radius: 6px; background-color: #e6f4ea; text-align: center;">
-    🚀 <b>Try our new ChatAnnotator</b> powered by Cohere's <b>Command-A</b> model! 
-    It offers a much-improved interface allowing you to <b>highlight and annotate specific errors</b> directly within chatbot responses, and even prompt immediate corrections. 
-    <a href="https://chatannotator.com" target="_blank">Click here to check it out!</a>
-</div>
-""")
-
-    with gr.Column(elem_id="col_container", visible=not DISABLED) as main_block:
-        chatbot = gr.Chatbot(elem_id='chatbot')
-        inputs = gr.Textbox(placeholder="Hi there!", label="Type an input and press Enter")
-        state = gr.State([])
-        with gr.Row():
-            with gr.Column(scale=7):
-                b1 = gr.Button(visible=not DISABLED)
-            with gr.Column(scale=3):
-                server_status_code = gr.Textbox(label="Status code from OpenAI server")
-
-        with gr.Accordion("Parameters", open=False):
-            top_p = gr.Slider(minimum=0, maximum=1.0, value=1.0, step=0.05, interactive=True, label="Top-p (nucleus sampling)")
-            temperature = gr.Slider(minimum=0, maximum=5.0, value=1.0, step=0.1, interactive=True, label="Temperature")
-            chat_counter = gr.Number(value=0, visible=False, precision=0)
-
-    inputs.submit(reset_textbox, [], [inputs, b1], queue=False)
-    inputs.submit(predict, [inputs, top_p, temperature, chat_counter, chatbot, state], 
-                  [chatbot, state, chat_counter, server_status_code, inputs, b1])
-    b1.click(reset_textbox, [], [inputs, b1], queue=False)
-    b1.click(predict, [inputs, top_p, temperature, chat_counter, chatbot, state], 
-             [chatbot, state, chat_counter, server_status_code, inputs, b1])
-
-    demo.queue(max_size=10, default_concurrency_limit=NUM_THREADS, api_open=False).launch(share=False)
+import gradio as gr
+import os
+import sys
+import json
+import requests
+import random
+
+MODEL = "gpt-4.1-mini"
+API_URL = os.getenv("API_URL")
+DISABLED = os.getenv("DISABLED") == 'True'
+OPENAI_API_KEYS = os.getenv("OPENAI_API_KEYS").split(',')
+NUM_THREADS = int(os.getenv("NUM_THREADS"))
+
+def exception_handler(exception_type, exception, traceback):
+    print(f"{exception_type.__name__}: {exception}")
+sys.excepthook = exception_handler
+sys.tracebacklimit = 0
+
+def predict(inputs, top_p, temperature, chat_counter, chatbot, history, request: gr.Request):
+    payload = {
+        "model": MODEL,
+        "messages": [{"role": "user", "content": inputs}],
+        "temperature": temperature,
+        "top_p": top_p,
+        "n": 1,
+        "stream": False,
+        "presence_penalty": 0,
+        "frequency_penalty": 0,
+    }
+    OPENAI_API_KEY = random.choice(OPENAI_API_KEYS)
+    headers = {
+        "Content-Type": "application/json",
+        "Authorization": f"Bearer {OPENAI_API_KEY}",
+    }
+
+    try:
+        response = requests.post(API_URL, headers=headers, json=payload)
+        response.raise_for_status()
+        data = response.json()
+
+        reply = data['choices'][0]['message']['content']
+
+        history.append(inputs)
+        history.append(reply)
+        chat_counter += 1
+
+        chatbot = [(history[i], history[i+1]) for i in range(0, len(history) - 1, 2)]
+        return chatbot, history, chat_counter, str(response.status_code), gr.update(interactive=False), gr.update(interactive=False)
+
+    except Exception as e:
+        print(f"Error: {e}")
+        return chatbot, history, chat_counter, f"Error: {e}", gr.update(interactive=True), gr.update(interactive=True)
+
+def reset_textbox():
+    return gr.update(value='', interactive=False), gr.update(interactive=False)
+
+title = """<h1 align="center">GPT-4.1 mini: Research Preview (Short-Term Availability)</h1>"""
+if DISABLED:
+    title = """<h1 align="center" style="color:red">This app has reached OpenAI's usage limit. Please check back tomorrow.</h1>"""
+description = """Language models can be conditioned to act like dialogue agents through a conversational prompt that typically takes the form: