Share via

Using the Vector Store Text Search (Preview)

  • Article

Warning

The Semantic Kernel Text Search functionality is in preview, and improvements that require breaking changes may still occur in limited circumstances before release.

Overview

The Vector Store Text Search implementation uses the Vector Store Connectors to retrieve search results. This means you can use Vector Store Text Search with any Vector Store which Semantic Kernel supports and any implementation of Microsoft.Extensions.VectorData.Abstractions.

Limitations

See the limitations listed for the Vector Store connector you are using.

Getting started

The sample below shows how to use an in-memory vector store to create a VectorStoreTextSearch and use it to perform a text search.

using Microsoft.Extensions.VectorData;
using Microsoft.SemanticKernel.Connectors.InMemory;
using Microsoft.SemanticKernel.Connectors.OpenAI;
using Microsoft.SemanticKernel.Data;
using Microsoft.SemanticKernel.Embeddings;

// Create an embedding generation service.
var textEmbeddingGeneration = new OpenAITextEmbeddingGenerationService(
        modelId: TestConfiguration.OpenAI.EmbeddingModelId,
        apiKey: TestConfiguration.OpenAI.ApiKey);

// Construct an InMemory vector store.
var vectorStore = new InMemoryVectorStore();
var collectionName = "records";

// Get and create collection if it doesn't exist.
var recordCollection = vectorStore.GetCollection<TKey, TRecord>(collectionName);
await recordCollection.CreateCollectionIfNotExistsAsync().ConfigureAwait(false);

// TODO populate the record collection with your test data
// Example https://github.com/microsoft/semantic-kernel/blob/main/dotnet/samples/Concepts/Search/VectorStore_TextSearch.cs

// Create a text search instance using the InMemory vector store.
var textSearch = new VectorStoreTextSearch<DataModel>(recordCollection, textEmbeddingGeneration);

// Search and return results as TextSearchResult items
var query = "What is the Semantic Kernel?";
KernelSearchResults<TextSearchResult> textResults = await textSearch.GetTextSearchResultsAsync(query, new() { Top = 2, Skip = 0 });
Console.WriteLine("\n--- Text Search Results ---\n");
await foreach (TextSearchResult result in textResults.Results)
{
    Console.WriteLine($"Name:  {result.Name}");
    Console.WriteLine($"Value: {result.Value}");
    Console.WriteLine($"Link:  {result.Link}");
}

Coming soon

More coming soon.

Coming soon

More coming soon.

Next steps

The following sections of the documentation show you how to:

  1. Create a plugin and use it for Retrieval Augmented Generation (RAG).
  2. Use text search together with function calling.
  3. Learn more about using vector stores for text search.

Text Search Abstractions Text Search Plugins Text Search Function Calling Text Search with Vector Stores