.Net: Fixed sorting for negative DOT product (#13071)

### Motivation and Context

When using negative dot product scoring, the more negative the value,
the closer the vector is to the query vector.

### Description

Fixed the sorting used for negative dot product distance calculations

### Contribution Checklist

- [x] The code builds clean without any errors or warnings
- [x] The PR follows the [SK Contribution
Guidelines](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md)
and the [pre-submission formatting
script](https://github.com/microsoft/semantic-kernel/blob/main/CONTRIBUTING.md#development-scripts)
raises no violations
- [x] All unit tests pass, and I have added new tests where possible
- [x] I didn't break anyone 😄

---------

Co-authored-by: westey <[email protected]>

Author: yorek

dotnet/src/VectorData/SqlServer/SqlServerCommandBuilder.cs

@@ -657,8 +657,8 @@ private static string Map(PropertyModel property)
         DistanceFunction.CosineDistance => ("COSINE", "ASC"),
         // A value of 0 indicates that the vectors are identical, while larger values indicate greater dissimilarity.
         DistanceFunction.EuclideanDistance => ("EUCLIDEAN", "ASC"),
-        // A value closer to 0 indicates higher similarity, while more negative values indicate greater dissimilarity.
-        DistanceFunction.NegativeDotProductSimilarity => ("DOT", "DESC"),
+        // Smaller numbers indicate more similar vectors
+        DistanceFunction.NegativeDotProductSimilarity => ("DOT", "ASC"),
         _ => throw new NotSupportedException($"Distance function {name} is not supported.")
     };
 }

dotnet/test/VectorData/SqlServer.ConformanceTests/Support/SqlServerTestEnvironment.cs

@@ -1,7 +1,6 @@
 // Copyright (c) Microsoft. All rights reserved.
 
 using Microsoft.Extensions.Configuration;
-using Microsoft.SemanticKernel.Connectors.SqlServer;
 
 namespace SqlServer.ConformanceTests.Support;
 
@@ -17,7 +16,7 @@ internal static class SqlServerTestEnvironment
             .AddJsonFile(path: "testsettings.json", optional: true)
             .AddJsonFile(path: "testsettings.development.json", optional: true)
             .AddEnvironmentVariables()
-            .AddUserSecrets<SqlServerVectorStore>()
+            .AddUserSecrets<SqlServerFixture>()
             .Build();
 
         return configuration.GetSection("SqlServer")["ConnectionString"];

dotnet/test/VectorData/VectorData.ConformanceTests/VectorSearch/VectorSearchDistanceFunctionComplianceTests.cs

@@ -24,7 +24,7 @@ public virtual Task DotProductSimilarity()
 
     [ConditionalFact]
     public virtual Task NegativeDotProductSimilarity()
-        => this.SimpleSearch(DistanceFunction.NegativeDotProductSimilarity, -1, 1, 0, [1, 2, 0]);
+        => this.SimpleSearch(DistanceFunction.NegativeDotProductSimilarity, -1, 1, 0, [0, 2, 1]);
 
     [ConditionalFact]
     public virtual Task EuclideanDistance()

dotnet/test/VectorData/Weaviate.ConformanceTests/VectorSearch/WeaviateVectorSearchDistanceFunctionComplianceTests.cs

@@ -1,6 +1,5 @@
 // Copyright (c) Microsoft. All rights reserved.
 
-using Microsoft.Extensions.VectorData;
 using VectorData.ConformanceTests.VectorSearch;
 using Weaviate.ConformanceTests.Support;
 using Xunit;
@@ -15,13 +14,6 @@ public class WeaviateVectorSearchDistanceFunctionComplianceTests_NamedVectors(We
     public override Task DotProductSimilarity() => Assert.ThrowsAsync<NotSupportedException>(base.DotProductSimilarity);
 
     public override Task EuclideanDistance() => Assert.ThrowsAsync<NotSupportedException>(base.EuclideanDistance);
-
-    /// <summary>
-    /// Tests vector search using <see cref="DistanceFunction.NegativeDotProductSimilarity"/>, computing -(u · v) as a distance metric per Weaviate's convention.
-    /// Expects scores of -1 (exact match), 1 (opposite), and 0 (orthogonal), sorted ascending ([0, 2, 1]), with lower scores indicating closer matches.
-    /// <see href="https://weaviate.io/developers/weaviate/config-refs/distances#available-distance-metrics"/>.
-    /// </summary>
-    public override Task NegativeDotProductSimilarity() => this.SimpleSearch(DistanceFunction.NegativeDotProductSimilarity, -1, 1, 0, [0, 2, 1]);
 }
 
 public class WeaviateVectorSearchDistanceFunctionComplianceTests_UnnamedVector(WeaviateDynamicDataModelNamedVectorsFixture fixture)
@@ -32,11 +24,4 @@ public class WeaviateVectorSearchDistanceFunctionComplianceTests_UnnamedVector(W
     public override Task DotProductSimilarity() => Assert.ThrowsAsync<NotSupportedException>(base.DotProductSimilarity);
 
     public override Task EuclideanDistance() => Assert.ThrowsAsync<NotSupportedException>(base.EuclideanDistance);
-
-    /// <summary>
-    /// Tests vector search using <see cref="DistanceFunction.NegativeDotProductSimilarity"/>, computing -(u · v) as a distance metric per Weaviate's convention.
-    /// Expects scores of -1 (exact match), 1 (opposite), and 0 (orthogonal), sorted ascending ([0, 2, 1]), with lower scores indicating closer matches.
-    /// <see href="https://weaviate.io/developers/weaviate/config-refs/distances#available-distance-metrics"/>.
-    /// </summary>
-    public override Task NegativeDotProductSimilarity() => this.SimpleSearch(DistanceFunction.NegativeDotProductSimilarity, -1, 1, 0, [0, 2, 1]);
 }