Decisions with Chat Completions & Responses APIs

// Import the required NuGet configuration packages
#r "nuget: Microsoft.Extensions.Configuration, 10.0.10"
#r "nuget: Microsoft.Extensions.Configuration.Json, 10.0.10"
#r "nuget: System.Text.Json, 10.0.10"

using Microsoft.Extensions.Configuration.Json;
using Microsoft.Extensions.Configuration;
using System.IO;
using System;

// Load the configuration settings from the local.settings.json and secrets.settings.json files
// The secrets.settings.json file is used to store sensitive information such as API keys
var configurationBuilder = new ConfigurationBuilder()
    .SetBasePath(Directory.GetCurrentDirectory())
    .AddJsonFile("local.settings.json", optional: true, reloadOnChange: true)
    .AddJsonFile("secrets.settings.json", optional: true, reloadOnChange: true);
var config = configurationBuilder.Build();

// IMPORTANT: You ONLY NEED either Azure OpenAI or OpenAI connection info, not both.
// Azure OpenAI Connection Info
var azureOpenAIEndpoint = config["AzureOpenAI:Endpoint"];
var azureOpenAIAPIKey = config["AzureOpenAI:APIKey"];
var azureOpenAIModelDeploymentName = config["AzureOpenAI:ModelDeploymentName"];
// OpenAI Connection Info 
var openAIAPIKey = config["OpenAI:APIKey"];
var openAIModelId = config["OpenAI:ModelId"];
// AzureOpenAI or OpenAI selection flag
var useAzureOpenAI = bool.Parse(config["AzureOpenAI:UseAzureOpenAI"] ?? "true");

// Import the Microdoft Extensions AI NuGet Packages
#r "nuget: Microsoft.Extensions.AI, 10.8.3"
#r "nuget: Microsoft.Extensions.AI.Abstractions, 10.8.3"
#r "nuget: Microsoft.Extensions.AI.OpenAI, 10.8.3"
// Import Azure & OpenAI NuGet Packages
#r "nuget: Azure.Identity, 1.21.0"
#r "nuget: OpenAI, 2.12.0"

using Azure;
using Microsoft.Extensions.AI;
using Microsoft.Extensions.Configuration;
using OpenAI;
using System.ClientModel;
using System.ComponentModel;
using System.Text.Json;


// Create the IChatClient based on the selected service
IChatClient chatClient;

// Create a new MEAI ChatClient instance
if (useAzureOpenAI)
{
    Console.WriteLine("Using Azure OpenAI Service");

    var apiKeyCredential = new ApiKeyCredential(azureOpenAIAPIKey!);

    var azureOpenAIClient = new OpenAIClient(
        apiKeyCredential,
        new OpenAIClientOptions
        {
            Endpoint = new Uri($"{azureOpenAIEndpoint!.TrimEnd('/')}/openai/v1")
        });

    chatClient = azureOpenAIClient.GetChatClient(azureOpenAIModelDeploymentName).AsIChatClient();
}
else
{
    Console.WriteLine("Using OpenAI Service");

    var apiKeyCredential = new ApiKeyCredential(azureOpenAIAPIKey);
    var openAIClient = new OpenAIClient(apiKeyCredential);

    // #pragma warning disable OPENAI001
    chatClient = openAIClient.GetChatClient(openAIModelId).AsIChatClient();
}

// Simple prompt to list some decision frameworks this GenAI LLM is familiar with 
// LLMs are trained on a diverse range of data and can provide insights on a wide range of topics like decision frameworks
// SLMS (smaller LLMs) are trained on a more specific range of data and may not provide insights on all topics
var simpleDecisionPrompt = """
Provide a list of 5 decision frameworks that can help improve the quality of decisions.

Output Format Instructions:
When generating Markdown, do not use any headings higher than ###. 
Avoid # and ## headers. Use only ###, ####, or lower-level headings if necessary. 
All top-level section headers should start at ### or lower. 
Never use ---, ***, or ___ for horizontal lines. There should be no horizontal lines in the output.
For separation, use extra extra spacing. Do not any render horizontal lines.
""";

// Execute the prompt against the AI model
var simpleDecisionPromptResponse = await chatClient.GetResponseAsync(simpleDecisionPrompt);
var simpleDecisionPromptResponseText = simpleDecisionPromptResponse.Text;

// Display the response string as Markdown
simpleDecisionPromptResponseText.DisplayAs("text/markdown");

// Set the overall system prompt to behave like a decision intelligence assistant (persona)
var systemPrompt = """
You are a Decision Intelligence assistant. 
Assist the user in exploring options, reasoning through decisions, problem-solving, and applying systems thinking to various scenarios. 
Provide structured, logical, and comprehensive advice.

Output Format Instructions:
When generating Markdown, do not use any headings higher than ###. 
Avoid # and ## headers. Use only ###, ####, or lower-level headings if necessary. 
All top-level section headers should start at ### or lower. 
Never use ---, ***, or ___ for horizontal lines. There should be no horizontal lines in the output.
For separation, use extra extra spacing. Do not any render horizontal lines.

Format the response using only a Markdown table. Only return a Markdown table. 
Do not enclose the table in triple backticks.
""";

// Simple instruction prompt to list 5 (five) decision frameworks this GenAI LLM is familiar with
var simpleDecisionPrompt = """
Provide five Decision Frameworks that can help improve the quality of decisions.
""";

// Create a new chat history object with proper system and user message roles
var chatMessages = new List<ChatMessage>();
// Add system and user messages to the chat history
var systemMessage = new ChatMessage(ChatRole.System, systemPrompt);
var userMessage = new ChatMessage(ChatRole.User, simpleDecisionPrompt);
chatMessages.Add(systemMessage);
chatMessages.Add(userMessage);


// Execute the chat messages against the AI model
var chatHistoryResponse = await chatClient.GetResponseAsync(chatMessages);
var chatHistoryResponseText = chatHistoryResponse.Text;

// Display the response string as Markdown
chatHistoryResponseText.DisplayAs("text/markdown");

// Capture the Assistant response and add it to the chat history
// This will persist the full conversation for future interactions
var assistantChatMessage = new ChatMessage(ChatRole.Assistant, chatHistoryResponseText);

// Add the assistant message to the decisoin chat history
chatMessages.Add(assistantChatMessage);

// Note: No reference is made to what previous frameworks were listed.
// Note: Previous context is maintained by the MEAI ChatHistory object 
var simpleDecisionPromptFollowupQuestionPartTwo = """
Which of the 5 decision frameworks listed above is best suited for the military intelligence community?  
Think carefully step by step about what decision frameworks are needed to answer the query.  
Select only the single best framework. 
""";

// Add User message to the chat history
var userFollowupChatMessage = new ChatMessage(ChatRole.User, simpleDecisionPromptFollowupQuestionPartTwo);
chatMessages.Add(userFollowupChatMessage);

// Execute the chat messages against the AI model
var chatHistoryResponseMilitaryIntelligence = await chatClient.GetResponseAsync(chatMessages);
var chatHistoryResponseMilitaryIntelligenceText = chatHistoryResponseMilitaryIntelligence.Text;

// Display the response string as Markdown
chatHistoryResponseMilitaryIntelligenceText.DisplayAs("text/markdown");

// Set the overall system prompt to behave like a decision intelligence assistant (persona)
var systemPromptDecisionConversation = """
You are a Decision Intelligence assistant. 
Assist the user in exploring options, reasoning through decisions, problem-solving, and applying systems thinking to various scenarios. 
Provide structured, logical, and comprehensive advice.
""";

// User prompt to start a decision conversation about buying a fitness tracker
var userPromptDecisionConversationTurn1 = """
I am considering buying a fitness tracker but I am not sure which features really matter. 
Could you explain the main considerations? 
""";

var assistantResponseDecisionConversationTurn1 = """
Absolutely! Key considerations for choosing a fitness tracker often include: 
- Fitness Features (e.g., step counting, heart-rate monitoring, workout tracking)
- Health Monitoring (like sleep tracking or advanced ECG capabilities)
- Battery Life
- Comfort and Fit
- Compatibility with your smartphone's operating system
- Price Range
- Knowing which of these you value most will help narrow down your options.
""";

var userPromptDecisionConversationTurn2 = """
Got it. Battery life and sleep tracking are both important to me. 
What's a typical battery life range for trackers, and how good is their sleep tracking?
""";

var assistantResponseDecisionConversationTurn2 = """
Different fitness trackers have varying battery lives, typically anywhere from 2-3 days (for more feature-rich or bright-display devices) 
up to about 10-14 days (for more minimalist trackers). 

Regarding sleep tracking:
- Many modern devices can track different sleep stages (light, deep, REM).
- High-end devices even track breathing rate and oxygen saturation during the night.
- Some also offer insights about sleep quality and tips for improvement.

If battery life is a priority, you may want to consider trackers with a more conservative display and fewer power-hungry features.
""";

var userPromptDecisionConversationTurn3 = """
Are there any popular models you would recommend for that balance of battery life and sleep tracking?
""";

var assistantResponseDecisionConversationTurn3 = """
Popular options that strike a balance between robust sleep tracking and good battery life include:
- Fitbit Inspire 3 or Fitbit Charge 5: Both have solid sleep tracking, multi-day battery, and strong app support.
- Garmin Vivosmart 5: Offers decent fitness and sleep insights, plus better-than-average battery life.
- Xiaomi Mi Band 7: Budget-friendly, surprisingly robust health metrics, and a long-lasting battery.

Each varies in price and features, so consider your budget and the specific metrics you’d like to track.
""";

var userPromptDecisionConversationTurn4 = """
That is great infor. 
I also have a friend who said I should consider an Apple Watch. 
Does that have comparable battery life and sleep tracking?
""";

var assistantResponseDecisionConversationTurn4 = """
The Apple Watch is known for its extensive feature set, including advanced health monitoring, 
but its battery life tends to be shorter—usually around 18-24 hours on a single charge 
(or possibly a bit more in low-power mode). While its sleep tracking features have improved over time, 
you might find you need to charge it daily or almost daily, which could conflict with your goal of 24/7 tracking. 
""";

// Create a new chat history with a multi-turn conversation about buying a fitness tracker
// Notice how the conversation flow uses different roles (user and assistant) to simulate a realistic decision-making dialogue
var chatHistoryMessages = new List<ChatMessage>();

chatHistoryMessages.Add(new ChatMessage(ChatRole.System, systemPromptDecisionConversation));
chatHistoryMessages.Add(new ChatMessage(ChatRole.User, userPromptDecisionConversationTurn1));
chatHistoryMessages.Add(new ChatMessage(ChatRole.Assistant, assistantResponseDecisionConversationTurn1));
chatHistoryMessages.Add(new ChatMessage(ChatRole.User, userPromptDecisionConversationTurn2));
chatHistoryMessages.Add(new ChatMessage(ChatRole.Assistant, assistantResponseDecisionConversationTurn2));
chatHistoryMessages.Add(new ChatMessage(ChatRole.User, userPromptDecisionConversationTurn3));
chatHistoryMessages.Add(new ChatMessage(ChatRole.Assistant, assistantResponseDecisionConversationTurn3));
chatHistoryMessages.Add(new ChatMessage(ChatRole.User, userPromptDecisionConversationTurn4));
chatHistoryMessages.Add(new ChatMessage(ChatRole.Assistant, assistantResponseDecisionConversationTurn4));

var userPromptDecisionConversationFinalDecision = """
Thank you for all of that information, I think I'll go with a Fitbit. 
The Fitbit Inspire 3 seems like a good fit for my budget and needs. 

Any quick final thoughts or insights on my decision? 
""";

// Add User message to the chat history
chatHistoryMessages.Add(new ChatMessage(ChatRole.User, userPromptDecisionConversationFinalDecision));

// Execute the chat messages against the AI model
var fitnessTrackerResponse = await chatClient.GetResponseAsync(chatHistoryMessages);
var fitnessTrackerResponseText = fitnessTrackerResponse.Text;

// Add the response to the chat history (Chat Messages)
chatHistoryMessages.Add(new ChatMessage(ChatRole.Assistant, fitnessTrackerResponseText));

// Display the response string as Markdown
fitnessTrackerResponseText.DisplayAs("text/markdown");

// Print the number of chat interactions and the chat history (turns)
Console.WriteLine("Number of decision chat interactions: " + chatHistoryMessages.Count());

// Change this to a string builder and show as markdown
var stringBuilderChatHistory = new StringBuilder();
foreach (var message in chatHistoryMessages)
{
    // add a new line for each message
    stringBuilderChatHistory.AppendLine($"**{message.Role.ToString().ToUpper()}**:");
    stringBuilderChatHistory.Append($"{message.Text.Replace("#", string.Empty)}");
    stringBuilderChatHistory.AppendLine("\n");
}

// Display the chat history as Markdown
stringBuilderChatHistory.ToString().DisplayAs("text/markdown");

var userPromptSummarizeTheDecisionChat = """
Briefly summarize the chat conversation about buying a fitness tracker.
""";

chatHistoryMessages.Add(new ChatMessage(ChatRole.User, userPromptSummarizeTheDecisionChat));

// Execute the chat messages against the AI model
var fitnessTrackerDecisionSummaryResponse = await chatClient.GetResponseAsync(chatHistoryMessages);
var fitnessTrackerDecisionSummaryResponseText = fitnessTrackerDecisionSummaryResponse.Text;

// Display the response string as Markdown
fitnessTrackerDecisionSummaryResponseText.DisplayAs("text/markdown");

// Optimize the chat history for any future decision interactions
// Prompt to look at the chat history and provide recommendations for optimizing decision-making
var userPromptOptimizeTheDecisionChat = """
Based on the chat conversation so far...
Provide some recommendations on how the decision interactions could be optimized for decision-making.
What are some considerations for the AI to help the use make even better decisions?
""";

// Remove the last message to avoid redundancy
chatHistoryMessages.RemoveAt(chatHistoryMessages.Count - 1);

chatHistoryMessages.Add(new ChatMessage(ChatRole.User, userPromptOptimizeTheDecisionChat));

// Execute the chat messages against the AI model
var decisionOptimizationResponse = await chatClient.GetResponseAsync(chatHistoryMessages);
var decisionOptimizationResponseText = decisionOptimizationResponse.Text;

// Display the response string as Markdown
decisionOptimizationResponseText.DisplayAs("text/markdown");