Update app.py

app.py

@@ -1,53 +1,79 @@
+from openai import OpenAI  # Assuming Nvidia client is available in the same library, adjust if necessary
 import streamlit as st
 import os
-from openai import OpenAI
+from datetime import datetime
 
-# Initialize the Nvidia API client using API Key stored in Streamlit secrets
-client =OpenAI(
-  base_url = "https://integrate.api.nvidia.com/v1",
-  api_key = os.getenv("NVIDIA_API_KEY")
+# Initialize Nvidia client
+client = OpenAI(
+    base_url="https://integrate.api.nvidia.com/v1",  # Nvidia API endpoint
+    api_key=os.environ["NVIDIA_API_KEY"]  # Nvidia API Key from Streamlit secrets
 )
 
-# Define Streamlit app layout
-st.title("AWS Well-Architected Review")
-st.write("Get recommendations for optimizing your AWS architecture.")
+st.title("ChatGPT-like clone with Nvidia Nemotron 70B Model")
 
+# Sidebar with instructions and Clear Session button
+with st.sidebar:
+    # Instruction
+    st.markdown("### Instructions 🤖\nThis is a basic chatbot. Ask anything, and the AI will try to help you! The app is supported by Yiqiao Yin.")
+
+    # Add a section to ask the user for the response length
+    st.markdown("#### Select the desired length of the AI response:")
+    response_length = st.radio(
+        "How detailed do you want the response to be?",
+        ('Efficient', 'Medium', 'Academic')
+    )
+
+    # Set max_tokens based on user selection
+    if response_length == 'Efficient':
+        max_tokens = 100
+    elif response_length == 'Medium':
+        max_tokens = 600
+    else:  # 'Academic'
+        max_tokens = 1024
+
+    # Clear
+    if st.button("Clear Session"):
+        st.session_state.clear()
+    st.write(f"Copyright © 2010-{datetime.now().year} Present Yiqiao Yin")
+
+# Initialize session state variables if not already present
+if "nvidia_model" not in st.session_state:
+    st.session_state["nvidia_model"] = "nvidia/llama-3.1-nemotron-70b-instruct"
 
 if "messages" not in st.session_state:
-    st.session_state.messages = [
-        {"role": "system", "content": "You are an assistant that provides recommendations based on AWS Well-Architected Review best practices. Focus on the 5 pillars: Operational Excellence, Security, Reliability, Performance Efficiency, and Cost Optimization."}
-    ]
-
-# User input for AWS architecture description
-architecture_input = st.text_area("Describe your AWS architecture:")
-
-# Button to submit the input
-if st.button("Get Recommendations"):
-    if architecture_input:
-        # Add user input to the conversation
-        st.session_state.messages.append({"role": "user", "content": architecture_input})
-        
-        with st.chat_message("assistant"):
-            with st.spinner("Generating recommendations..."):
-                # Create Nvidia completion request with conversation history
-                stream = client.chat.completions.create(
-                    model="nvidia/llama-3.1-nemotron-70b-instruct",  # Nvidia model name
-                    messages=st.session_state.messages,  # Include all messages in the API call
-                    temperature=0.5,
-                    top_p=0.7,
-                    max_tokens=1024,
-                    stream=True,
-                )
-
-                response_chunks = []
-                for chunk in stream:
-                    print('chunk-' + chunk.choices[0].delta.content)
-                    if chunk.choices[0].delta.content is not None:
-                        response_chunks.append(chunk.choices[0].delta.content)
-                        print(chunk.choices[0].delta.content)
-                response = "".join(response_chunks)
-                
-                # Display the response as recommendations
-                st.markdown(f"**Recommendations:**\n\n{response}")
-                # Add response to conversation history
-                st.session_state.messages.append({"role": "assistant", "content": response})
+    # Adding the initial system message
+    st.session_state.messages = [{"role": "system", "content": "You are a helpful assistant."}]
+
+# Render the chat history
+for message in st.session_state.messages:
+    with st.chat_message(message["role"]):
+        st.markdown(message["content"])
+
+# Get new user input
+if prompt := st.chat_input("What is up?"):
+    # Add user message to the session state
+    st.session_state.messages.append({"role": "user", "content": prompt})
+    with st.chat_message("user"):
+        st.markdown(prompt)
+
+    # Display assistant's message while waiting for the response
+    with st.chat_message("assistant"):
+        with st.spinner("The assistant is thinking... Please wait."):
+            # Create Nvidia completion request with full conversation history
+            stream = client.chat.completions.create(
+                model=st.session_state["nvidia_model"],
+                messages=st.session_state.messages,  # Include all previous messages in the API call
+                temperature=0.5,
+                top_p=0.7,
+                max_tokens=max_tokens,
+                stream=True,
+            )
+            response_chunks = []
+            for chunk in stream:
+                if chunk.choices[0].delta.content is not None:
+                    response_chunks.append(chunk.choices[0].delta.content)
+            response = "".join(response_chunks)
+            st.markdown(response)
+
+    # Store the assistant response in the session state
+    st.session_state.messages.append({"role": "assistant", "content": response})