Encamina.Enmarcha.SemanticKernel.Abstractions 8.1.8-preview-04

Semantic Kernel - Abstractions

This project mainly contains abstractions used by other ENMARCHA NuGet packages. Additionally, it includes extension methods and utilities primarily related to Semantic Kernel.

Setup

Nuget package

First, install NuGet. Then, install Encamina.Enmarcha.SemanticKernel.Abstractions from the package manager console:

PM> Install-Package Encamina.Enmarcha.SemanticKernel.Abstractions

.NET CLI:

Install .NET CLI. Next, install Encamina.Enmarcha.SemanticKernel.Abstractions from the .NET CLI:

dotnet add package Encamina.Enmarcha.SemanticKernel.Abstractions

How to use

In addition to the abstractions such as interfaces or abstract classes, which have their implementations in other ENMARCHA NuGets (for example, IMemoryManager/MemoryManager), there are some utilities that can be used directly.

ILengthFunctions

// Entry point
var tokenCount = ILengthFunctions.LengthByTokenCount("This text contains 5 tokens");
// tokenCount = 5

The above code returns the number of tokens in the text using the GPT-3 tokenizer.

SemanticKernelOptions

// ...
  "SemanticKernelOptions": {
    "ChatModelName": "gpt-35-turbo", // Name (sort of a unique identifier) of the model to use for chat
    "ChatModelDeploymentName": "gpt-35-turbo", // Model deployment name on the LLM (for example OpenAI) to use for chat
    "EmbeddingsModelName": "text-embedding-ada-002", // Name (sort of a unique identifier) of the model to use for embeddings
    "EmbeddingsModelDeploymentName": "text-embedding-ada-002", // Model deployment name on the LLM (for example OpenAI) to use for embeddings
    "Endpoint": "https://your-url.openai.azure.com/", // Uri for an LLM resource (like OpenAI). This should include protocol and hostname.
    "Key": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx", // Key credential used to authenticate to an LLM resource
    // ....
  },
// ...

The above section is taken from a JSON file that is typically added to the appsettings.json. This provides the necessary configurations for Semantic Kernel to operate.

Once these configurations are defined, they can be added to the application's configuration.

var builder = WebApplication.CreateBuilder(new WebApplicationOptions
{
    // ...
});

// ...

// Or others configuration providers...
builder.Configuration.AddJsonFile("appsettings.json", optional: true, reloadOnChange: true);

builder.Services.AddOptions<SemanticKernelOptions>().Bind(builder.Configuration.GetSection(nameof(SemanticKernelOptions)))
    .ValidateDataAnnotations()
    .ValidateOnStart();

SKContextExtensions

This class provides some static methods to obtain information/functionality about SKContext.

// ...

var resultContext = await kernel.RunAsync(contextVariables);

// Extensions method from SKContextExtensions
resultContext.ValidateAndThrowIfErrorOccurred();
var maxWordCount = resultContext.GetContextVariable("dummycontextvariable");