Getting Started with Hubtel Semantic Kernel Sdk

This SDK allows you to interact with lLms using Microsoft Semantic Kernel library.

In order to get started with Hubtel Semantic Kernel Sdk, do the following activities

Install-Package Hubtel.SemanticKernel.Sdk

Install-Package Hubtel.SemanticKernel.Sdk is a command for the NuGet Package Manager Console in Visual Studio. This command installs the Hubtel.SemanticKernel.Sdk package into the current project.

    Install-Package Hubtel.SemanticKernel.Sdk

Add to appsettings.json

• The OpenAIModelName property is set to "gpt-3.5-turbo-0301". This suggests that the SemanticKernel component uses the GPT-3.5 Turbo model provided by OpenAI. GPT-3.5 Turbo is a powerful language model developed by OpenAI that can generate human-like text based on the input it's given.

• The OpenAIAPIKey property is set to "{Your_API_Key}". This is a placeholder for the API key you would need to interact with OpenAI's API. You would replace {Your_API_Key} with your actual OpenAI API key. This key is used to authenticate your application's requests to the OpenAI API.

• In summary, this configuration is used to set up and authenticate the connection between the SemanticKernel component and the OpenAI API.

  "SemanticKernel": {
    "OpenAIModelName" : "gpt-3.5-turbo-0301",
    "OpenAIAPIKey" : "{Your_API_Key}"
  }

Add to IServiceCollection in Program.cs or Startup.cs

• services.AddHubtelSemanticKernelSdk(Configuration) is a method call to AddHubtelSemanticKernelSdk on the services object. The services object is of type IServiceCollection, which is a framework-provided interface that is used to register services to the application's dependency injection container during the startup configuration phase.

• The AddHubtelSemanticKernelSdk method is likely an extension method that registers the Hubtel Semantic Kernel SDK services into the application's dependency injection container. This allows the SDK's services to be injected into other parts of the application where they are needed.

• The Configuration parameter being passed to the method is likely an instance of IConfiguration, another framework-provided interface that represents the application's configuration properties. This object is typically used to access configuration settings from sources like environment variables, command line arguments, or configuration files like appsettings.json. In this context,the AddHubtelSemanticKernelSdk method uses the Configuration object to access settings specific to the Hubtel Semantic Kernel SDK.

    services.AddHubtelSemanticKernelSdk(Configuration);

Inject and consume service methods

• This is a C# class named AIController, which is a part of an ASP.NET Core application. This class is decorated with the [ApiController] and [Route("[controller]")] attributes, indicating that it's a controller that handles HTTP requests in a RESTful API.

• The AIController class has a private readonly field _kernel of type ISemanticKernel. This field is initialized via the constructor, suggesting that the ISemanticKernel implementation is injected into this controller using dependency injection.

• In the constructor, the InitializeKernelBuilder method of _kernel is called with a large string argument. This string appears to be instructions for an AI model, specifying its behavior and the context it should use when processing input.

• The AIController class has a single public method, GetSummary, which is an asynchronous method that returns a Task<IActionResult>. This method is decorated with the [HttpGet("GetSummary/{query}")] attribute, indicating that it handles HTTP GET requests at the path /AI/GetSummary/{query}, where AI is the controller name inferred from the [controller] placeholder in the class-level route attribute, and {query} is a route parameter.

• Inside the GetSummary method, the Prompt method of _kernel is called with the query parameter, and the result is awaited. The Prompt method likely sends the query to the AI model and returns its response. The response is then wrapped in an Ok result (HTTP 200 OK), which is a common way to return successful responses in ASP.NET Core APIs.

[ApiController]
[Route("[controller]")]
public class AIController : ControllerBase
{
    private readonly ISemanticKernel _kernel;

    public AIController(ISemanticKernel kernel)
    {
        _kernel = kernel;
        
        _kernel.InitializeKernelBuilder(@"You are an expert in English Grammar.
               I will give you some words to correct the spelling or the sentence based on a particular context. 
               Do not explain anything. Just give me the correct version of the word as close to english as possible. 
               If you cannot find any suitable match just output  'NO_MATCH'
               The context you are to use is an ecommerce inventory for groceries.
               The output should be a json array of strings. 

               {{$input}}");
        
    }
    
    
    [HttpGet("GetSummary/{query}")]
    public async Task<IActionResult> GetSummary(string query)
    {
        var response = await _kernel.Prompt(query);
        return Ok(response);
    }
}