Xfty.VectorDatabases
1.0.0-beta.7
See the version list below for details.
dotnet add package Xfty.VectorDatabases --version 1.0.0-beta.7
NuGet\Install-Package Xfty.VectorDatabases -Version 1.0.0-beta.7
<PackageReference Include="Xfty.VectorDatabases" Version="1.0.0-beta.7" />
<PackageVersion Include="Xfty.VectorDatabases" Version="1.0.0-beta.7" />
<PackageReference Include="Xfty.VectorDatabases" />
paket add Xfty.VectorDatabases --version 1.0.0-beta.7
#r "nuget: Xfty.VectorDatabases, 1.0.0-beta.7"
#:package Xfty.VectorDatabases@1.0.0-beta.7
#addin nuget:?package=Xfty.VectorDatabases&version=1.0.0-beta.7&prerelease
#tool nuget:?package=Xfty.VectorDatabases&version=1.0.0-beta.7&prerelease
Xfty.VectorDatabases
A bundled Xfty IValueExpression
for a vector-database record's embedding field - a fixed-length float[] of
independent random values, structurally a vector but not a semantically
meaningful embedding (see Deliberately out of scope
below).
dotnet add package Xfty.VectorDatabases
Usage
using Net.NowhereAtAll.Xfty.VectorDatabases;
// DocumentChunk here is illustrative - any record with a float[] field works.
new MasterTemplate<DocumentChunk>(x => x.Id)
{
[x => x.Embedding] = new RandomVectorExpression(KnownEmbeddingDimensions.OpenAiTextEmbedding3Small),
};
RandomVectorExpression(int dimensions, float min = -1f, float max = 1f, bool normalize = false) -
normalize: true produces a unit-length vector, since cosine-similarity
comparisons and several vector-DB schemas assume one.
KnownEmbeddingDimensions bundles named dimension constants for popular
embedding models (OpenAiTextEmbedding3Small, CohereEmbedV3, …) so a test
doesn't have to hardcode or look up the number.
Persisting a vector field
This package only produces the value. For actual persistence:
- pgvector, through the unmodified
Xfty.EntityFrameworkCoregateway - proven against a real Postgres container, zero new gateway code needed. - Qdrant, direct client or via
Microsoft.Extensions.VectorData- seeXfty.VectorDatabases.Qdrant/Xfty.VectorDatabases.MicrosoftExtensionsVectorData- both preview packages, read their own READMEs first.
Deliberately out of scope
Calling a real embedding API (OpenAI, Cohere, …) for a semantically
meaningful vector is not a goal of this package - it would break every
Xfty value expression's offline-and-fast contract (network latency, API
cost, a credential CI would have to hold). A test asserting a genuine
nearest-neighbor relationship needs vectors informed by its own domain and
should write its own expression for that.
Full documentation
- Vector database support (design) - the full rationale, including why calling a real embedding API is a non-goal
- Everything else
Xftydoes
| Product | Versions Compatible and additional computed target framework versions. |
|---|---|
| .NET | net10.0 is compatible. net10.0-android was computed. net10.0-browser was computed. net10.0-ios was computed. net10.0-maccatalyst was computed. net10.0-macos was computed. net10.0-tvos was computed. net10.0-windows was computed. |
-
net10.0
- Xfty (>= 1.0.0-beta.7)
NuGet packages
This package is not used by any NuGet packages.
GitHub repositories
This package is not used by any popular GitHub repositories.
| Version | Downloads | Last Updated |
|---|---|---|
| 1.0.0-beta.12 | 47 | 9/18/2026 |
| 1.0.0-beta.11 | 58 | 9/10/2026 |
| 1.0.0-beta.10 | 50 | 9/10/2026 |
| 1.0.0-beta.9 | 53 | 9/10/2026 |
| 1.0.0-beta.8 | 69 | 9/10/2026 |
| 1.0.0-beta.7 | 56 | 9/7/2026 |
| 1.0.0-beta.6 | 56 | 9/7/2026 |
| 1.0.0-beta.5 | 69 | 9/6/2026 |
| 1.0.0-beta.3 | 71 | 9/6/2026 |
| 1.0.0-beta.2 | 64 | 9/6/2026 |
| 1.0.0-beta.1 | 68 | 9/6/2026 |