Switch to SqlVector<float> for embeddings

Updated the application to use SQL Server's native vector data type (`SqlVector<float>`) for embeddings, replacing the previous `float[]` or `string` representations.

- Updated `.editorconfig` with new code style preferences and diagnostic rule severities.
- Modified `DocumentChunk.cs` to use `SqlVector<float>` for the `Embedding` property.
- Updated migrations and `ApplicationDbContextModelSnapshot` to reflect the new `SqlVector<float>` type.
- Replaced `AddAzureSql` with `AddSqlServer` in `Program.cs` and removed `UseVectorSearch`.
- Adjusted `DocumentService` and `VectorSearchService` to handle `SqlVector<float>` and updated vector search logic.
- Removed the `EFCore.SqlServer.VectorSearch` package and upgraded EF Core to `10.0.0-rc.1`.
- Made minor adjustments to OpenAPI configuration and dependency management.

.editorconfig

@@ -82,6 +82,7 @@ csharp_style_prefer_local_over_anonymous_function = true:silent
 csharp_style_prefer_extended_property_pattern = true:suggestion
 csharp_style_implicit_object_creation_when_type_is_apparent = true:silent
 csharp_style_prefer_tuple_swap = true:silent
+csharp_style_prefer_simple_property_accessors = true:suggestion
 
 # Field preferences
 dotnet_style_readonly_field = true:suggestion
@@ -299,4 +300,7 @@ dotnet_diagnostic.IDE0010.severity = none
 dotnet_diagnostic.IDE0072.severity = none
 
 # IDE0305: Simplify collection initialization
-dotnet_diagnostic.IDE0305.severity = none
+dotnet_diagnostic.IDE0305.severity = none
+
+# CA1873: Avoid potentially expensive logging
+dotnet_diagnostic.CA1873.severity = none