Using AI to help with code writing

#C-Sharp #C# #AI #ChatGPT

I recently started using ChatGPT and my own local Ollama server.  The code is generally pretty good, and requires some tweaking, as you'd suspect.  For example, here is a code snippet of ChatGPT trying to show how to perform Name Entity Recognition.

using Microsoft.SemanticKernel;
using Microsoft.SemanticKernel.ChatCompletion;
#pragma warning disable SKEXP0070

// Create two kernels for each model
var intentKernel = Kernel.CreateBuilder()
                   .AddOllamaChatCompletion(
                       modelId: "llama3.2:latest",
                       endpoint: new Uri("http://192.168.2.19:11434"))
                   .Build();

var nerKernel = Kernel.CreateBuilder()
                   .AddOllamaChatCompletion(
                       modelId: "ner-model:latest",
                       endpoint: new Uri("http://192.168.2.19:11434"))
                   .Build();

var aiIntentService = intentKernel.GetRequiredService<IChatCompletionService>();
var aiNerService = nerKernel.GetRequiredService<IChatCompletionService>();

var chatHistory = new ChatHistory();

while (true)
{
   // Step 1: Get user prompt and classify intent
   Console.WriteLine("Your prompt:");
   var userQuestion = Console.ReadLine();

   var intentPrompt = $"Classify the intent of the following question into categories like 'financial', 'person', 'sports', 'weather', 'news', or 'general', but only return the intent, nothing else: {userQuestion}";
   chatHistory.Add(new ChatMessageContent(AuthorRole.User, intentPrompt));

   // Stream the intent classification response using the intent model
   Console.WriteLine("AI Intent Classification Response:");
   var intentResponse = "";
   await foreach (var item in aiIntentService.GetStreamingChatMessageContentsAsync(chatHistory))
   {
       Console.Write(item.Content);
       intentResponse += item.Content;
   }
   chatHistory.Add(new ChatMessageContent(AuthorRole.Assistant, intentResponse.Trim()));
   Console.WriteLine();

   // Step 2: If the intent is relevant, extract entities using the NER model
   if (intentResponse.Trim().ToLower() == "sports")
   {
       // Construct the NER prompt for sports-related entities
       var nerPrompt = $"Identify key entities in the following question: \"{userQuestion}\". " +
                       "Return entities in the format: EntityType: Entity. " +
                       "For example, 'Team: Chicago Bears', 'Date: Yesterday'.";

       // Use the NER kernel for the NER request
       var nerChatHistory = new ChatHistory();
       nerChatHistory.Add(new ChatMessageContent(AuthorRole.User, nerPrompt));

       // Stream the NER response using the NER model
       Console.WriteLine("AI NER Response:");
       var nerResponse = "";
       await foreach (var item in aiNerService.GetStreamingChatMessageContentsAsync(nerChatHistory))
       {
           Console.Write(item.Content);
           nerResponse += item.Content;
       }
       chatHistory.Add(new ChatMessageContent(AuthorRole.Assistant, nerResponse.Trim()));
       Console.WriteLine();

       // Parse entities (optional)
   }
   else
   {
       Console.WriteLine($"No relevant entities to extract for intent: {intentResponse}");
   }

   Console.WriteLine();
}
 

This code works in the basic sense, but it needs to be tweaked even more.  I fed it a copy of my code up to this point, and it did not directly integrate with my project.  I had to re-work it quite a bit.  It made for a great start, but it still took me a few hours to get it to work the way it needed to.