Update app.py

app.py

@@ -3,10 +3,9 @@ import requests
 import json
 import os
 
-API_KEY = os.getenv('API_KEY') 
+API_KEY = os.getenv('API_KEY')
 INVOKE_URL = "https://api.nvcf.nvidia.com/v2/nvcf/pexec/functions/0e349b44-440a-44e1-93e9-abe8dcb27158"
 FETCH_URL_FORMAT = "https://api.nvcf.nvidia.com/v2/nvcf/pexec/status/"
-
 headers = {
     "Authorization": f"Bearer {API_KEY}",
     "Accept": "application/json",
@@ -15,16 +14,10 @@ headers = {
 
 BASE_SYSTEM_MESSAGE = "I carefully provide accurate, factual, thoughtful, nuanced answers and am brilliant at reasoning."
 
-def clear_chat(chat_history_state, chat_message):
-    print("Clearing chat...")
-    chat_history_state = []
-    chat_message = ''
-    return chat_history_state, chat_message
-
 def user(message, history, system_message=None):
     print(f"User message: {message}")
     history = history or []
-    if system_message:  # Check if a system message is provided and should be added
+    if system_message:
         history.append({"role": "system", "content": system_message})
     history.append({"role": "user", "content": message})
     return history
@@ -37,8 +30,7 @@ def call_nvidia_api(history, max_tokens, temperature, top_p):
         "max_tokens": max_tokens,
         "stream": False
     }
-
-    print(f"Payload enviado: {payload}")  # Imprime o payload enviado
+    print(f"Payload enviado: {payload}")
 
     session = requests.Session()
     response = session.post(INVOKE_URL, headers=headers, json=payload)
@@ -47,61 +39,55 @@ def call_nvidia_api(history, max_tokens, temperature, top_p):
         request_id = response.headers.get("NVCF-REQID")
         fetch_url = FETCH_URL_FORMAT + request_id
         response = session.get(fetch_url, headers=headers)
-    
-    response.raise_for_status()
-    response_body = response.json()
+        response.raise_for_status()
+        response_body = response.json()
+        print(f"Payload recebido: {response_body}")
 
-    print(f"Payload recebido: {response_body}")  # Imprime o payload recebido
-
-    if response_body["choices"]:
-        assistant_message = response_body["choices"][0]["message"]["content"]
-        history.append({"role": "assistant", "content": assistant_message})
-    
-    return history
+        if response_body["choices"]:
+            assistant_message = response_body["choices"][0]["message"]["content"]
+            history.append({"role": "assistant", "content": assistant_message})
+            return history
 
 def chat(history, system_message, max_tokens, temperature, top_p, top_k, repetition_penalty):
     print("Starting chat...")
-    updated_history = call_nvidia_api(history, max_tokens, temperature, top_p)
+    updated_history = user(None, history, system_message)
+    updated_history = call_nvidia_api(updated_history, max_tokens, temperature, top_p)
     return updated_history, ""
 
 def update_chatbot(message, chat_history, system_message, max_tokens, temperature, top_p):
     print("Updating chatbot...")
-    if not chat_history or (chat_history and chat_history[-1]["role"] != "user"):
-        chat_history = user(message, chat_history, system_message if not chat_history else None)
-    else:
-        chat_history = user(message, chat_history)
-    chat_history, _ = call_nvidia_api(chat_history, max_tokens, temperature, top_p)
+    chat_history = user(message, chat_history, system_message if not chat_history else None)
+    chat_history = call_nvidia_api(chat_history, max_tokens, temperature, top_p)
     return chat_history
 
-system_msg = gr.Textbox(BASE_SYSTEM_MESSAGE, label="System Message", placeholder="System prompt.", lines=5)
-max_tokens = gr.Slider(20, 1024, label="Max Tokens", step=20, value=1024)
-temperature = gr.Slider(0.0, 1.0, label="Temperature", step=0.1, value=0.2)
-top_p = gr.Slider(0.0, 1.0, label="Top P", step=0.05, value=0.7)
-
 with gr.Blocks() as demo:
     with gr.Row():
         with gr.Column():
             gr.Markdown("LLAMA 2 70B Free Demo")
-            description="""
+            description = """
 <div style="text-align: center; font-size: 1.5em; margin-bottom: 20px;">
     <strong>Explore the Capabilities of LLAMA 2 70B</strong>
 </div>
-<p>Llama 2 is a large language AI model capable of generating text and code in response to prompts.
-</p>
-<p> <strong>How to Use:</strong></p>
+<p>Llama 2 is a large language AI model capable of generating text and code in response to prompts.</p>
+<p><strong>How to Use:</strong></p>
 <ol>
     <li>Enter your <strong>message</strong> in the textbox to start a conversation or ask a question.</li>
     <li>Adjust the parameters in the "Additional Inputs" accordion to control the model's behavior.</li>
     <li>Use the buttons below the chatbot to submit your query, clear the chat history, or perform other actions.</li>
 </ol>
-<p> <strong>Powered by NVIDIA's cutting-edge AI API, LLAMA 2 70B offers an unparalleled opportunity to interact with an AI model of exceptional conversational ability, accessible to everyone at no cost.</strong></p>
-<p> <strong>HF Created by:</strong> @artificialguybr (<a href="https://twitter.com/artificialguybr">Twitter</a>)</p>
-<p> <strong>Discover more:</strong> <a href="https://artificialguy.com">artificialguy.com</a></p>
+<p><strong>Powered by NVIDIA's cutting-edge AI API, LLAMA 2 70B offers an unparalleled opportunity to interact with an AI model of exceptional conversational ability, accessible to everyone at no cost.</strong></p>
+<p><strong>HF Created by:</strong> @artificialguybr (<a href="https://twitter.com/artificialguybr">Twitter</a>)</p>
+<p><strong>Discover more:</strong> <a href="https://artificialguy.com">artificialguy.com</a></p>
 """
-
             gr.Markdown(description)
+
             chat_history_state = gr.State([])
 
+            system_msg = gr.Textbox(BASE_SYSTEM_MESSAGE, label="System Message", placeholder="System prompt.", lines=5)
+            max_tokens = gr.Slider(20, 1024, label="Max Tokens", step=20, value=1024)
+            temperature = gr.Slider(0.0, 1.0, label="Temperature", step=0.1, value=0.2)
+            top_p = gr.Slider(0.0, 1.0, label="Top P", step=0.05, value=0.7)
+
             chatbot = gr.ChatInterface(
                 fn=lambda message, history: update_chatbot(message, history, system_msg.value, max_tokens.value, temperature.value, top_p.value),
                 additional_inputs=[system_msg, max_tokens, temperature, top_p],
@@ -114,9 +100,6 @@ with gr.Blocks() as demo:
                 chat_history_state.value = []
                 chatbot.textbox.value = ""
 
-            chatbot.clear(
-                fn=clear_chat,
-                outputs=[chat_history_state, chatbot.textbox]
-            )
+            chatbot.clear(fn=clear_chat, outputs=[chat_history_state, chatbot.textbox])
 
 demo.launch()