Update app.py

app.py

@@ -1,8 +1,8 @@
-#refer llama recipes for more info https://github.com/huggingface/huggingface-llama-recipes/blob/main/inference-api.ipynb
-#huggingface-llama-recipes : https://github.com/huggingface/huggingface-llama-recipes/tree/main
 import gradio as gr
-from openai import OpenAI
+from openai import OpenAI, APIError
 import os
+import tenacity
+import asyncio
 
 ACCESS_TOKEN = os.getenv("HF_TOKEN")
 
@@ -11,59 +11,88 @@ client = OpenAI(
     api_key=ACCESS_TOKEN,
 )
 
-def respond(
+# Retry logic with tenacity for handling API rate limits
+@tenacity.retry(wait=tenacity.wait_exponential(multiplier=1, min=4, max=10), stop=tenacity.stop_after_attempt(5))
+async def respond(
     message,
-    history: list[tuple[str, str]],
     system_message,
     max_tokens,
     temperature,
     top_p,
 ):
-    messages = [{"role": "system", "content": system_message}]
+    try:
+        # Only use the system message and the current message for the response
+        messages = [{"role": "system", "content": system_message},
+                    {"role": "user", "content": message}]
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+        response = ""
+        # Properly stream chat completions using dot notation
+        stream = client.chat.completions.create(
+            model="NousResearch/Hermes-3-Llama-3.1-8B",
+            max_tokens=max_tokens,
+            stream=True,
+            temperature=temperature,
+            top_p=top_p,
+            messages=messages,
+        )
 
-    messages.append({"role": "user", "content": message})
+        # Stream response and concatenate tokens
+        for chunk in stream:
+            if hasattr(chunk.choices[0].delta, 'content'):
+                token = chunk.choices[0].delta.content
+                response += token
 
-    response = ""
-    
-    for message in  client.chat.completions.create(
-        model="NousResearch/Hermes-3-Llama-3.1-8B",
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-        messages=messages,
-    ):
-        token = message.choices[0].delta.content
-        
-        response += token
-        yield response
-        
-chatbot = gr.Chatbot(height=600)
+        return response
+
+    except APIError as e:
+        # Handle both string and dict types of error bodies
+        error_details = e.body
+        if isinstance(error_details, dict):
+            error_type = error_details.get("type", "Unknown")
+            error_code = error_details.get("code", "Unknown")
+            error_param = error_details.get("param", "Unknown")
+            error_message = error_details.get("message", "An error occurred.")
+            error_str = f"{error_type}: {error_message} (code: {error_code}, param: {error_param})"
+        else:
+            error_str = f"Error: {error_details}"
+
+        print(f"APIError: {error_str}")
+        return error_str
+
+    except Exception as e:
+        print(f"Exception: {e}")
+        return "Error occurred. Please try again."
+
+
+# Async Gradio function to handle user input and response generation without history
+async def generate_response(message, system_message, max_tokens, temperature, top_p):
+    response = await respond(message, system_message, max_tokens, temperature, top_p)
+    return response
+
+
+def launch_app():
+    try:
+        demo = gr.Blocks()
+        with demo:
+            gr.Markdown("# Chatbot")
+            message = gr.Textbox(label="Message")
+            system_message = gr.Textbox(label="System message")
+            max_tokens = gr.Slider(minimum=1, maximum=2048, value=2048, step=1, label="Max new tokens")
+            temperature = gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature")
+            top_p = gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-P")
+            response = gr.Text(label="Response")
+
+            # Use the async version of generate_response without history
+            gr.Button("Generate Response").click(
+                generate_response,
+                inputs=[message, system_message, max_tokens, temperature, top_p],
+                outputs=[response],
+                show_progress=False,
+            )
+        demo.launch(show_error=True)
+    except KeyError as e:
+        print(f"Error: {e}")
+        print("Please try again.")
 
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-P",
-        ),
-        
-    ],
-    fill_height=True,
-    chatbot=chatbot,
-    theme="Nymbo/Alyx_Theme",
-)
 if __name__ == "__main__":
-    demo.launch()
+    launch_app()