https://github.com/efcore/EFCore.SqlServer.VectorSearch

https://github.com/efcore/EFCore.SqlServer.VectorSearch

Last synced: about 2 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/efcore/EFCore.SqlServer.VectorSearch
• Owner: efcore
• License: mit
• Created: 2024-05-24T17:52:37.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2024-11-08T17:08:55.000Z (2 months ago)
• Last Synced: 2024-11-08T18:20:26.458Z (2 months ago)
• Language: C#
• Size: 35.2 KB
• Stars: 11
• Watchers: 2
• Forks: 1
• Open Issues: 3
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-semantickernel - EFCore.SqlServer.VectorSearch

README

        
# EFCore.SqlServer.VectorSearch

> [!IMPORTANT]  

> This plugin is in prerelease status, and the APIs described below are likely to change before the final release.

> Vector Functions are in Public Preview. Learn the details about vectors in Azure SQL here: https://aka.ms/azure-sql-vector-public-preview

This Entity Framework Core plugin provides integration between EF and Vector Support in Azure SQL Database, allowing LINQ to be used to perform vector similarity search, and seamless insertion/retrieval of vector data.

To use the plugin, reference the [EFCore.SqlServer.VectorSearch](https://www.nuget.org/packages/EFCore.SqlServer.VectorSearch) nuget package, and enable the plugin by adding `UseVectorSearch()` to your `UseSqlServer()` config as follows:

```c#

builder.Services.AddDbContext(options =>

  options.UseSqlServer("", o => o.UseVectorSearch()));

```

Once the plugin has been enabled, add an ordinary `float[]` property to the .NET type being mapped with EF:

```c#

public class Product

{

    public int Id { get; set; }

    public float[] Embedding { get; set; }

}

```

Finally, configure the property to be mapped as a vector by letting EF Core know using the `HasColumnType` method. Use the `vector` type and specify the number of dimension that your vector will have:

```c#

protected override void OnModelCreating(ModelBuilder modelBuilder)

{

    modelBuilder.Entity().Property(p => p.Embedding).HasColumnType("vector(3)");

}

```

That's it - you can now perform similarity search in LINQ queries! For example, to get the top 5 most similar products:

```c#

var someVector = new[] { 1f, 2f, 3f };

var products = await context.Products

    .OrderBy(p => EF.Functions.VectorDistance("cosine", p.Embedding, vector))

    .Take(5)

    .ToArrayAsync();

```

A full sample using EF Core and vectors is available here:

https://github.com/Azure-Samples/azure-sql-db-vector-search/tree/main/EF-Core

Ideas? Issues? Let us know on the [issues page](https://github.com/efcore/EFCore.SqlServer.VectorSearch/issues).