app.py

import streamlit as st
import os
from openai import OpenAI

# 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")
)

# Define Streamlit app layout
st.title("AWS Well-Architected Review")
st.write("Get recommendations for optimizing your AWS architecture.")


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-3.1-70b-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:
                    if chunk.choices[0].delta.content is not None:
                        response_chunks.append(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})