Qavren.Edge.Ingestion
1.0.0
Prefix Reserved
dotnet add package Qavren.Edge.Ingestion --version 1.0.0
NuGet\Install-Package Qavren.Edge.Ingestion -Version 1.0.0
<PackageReference Include="Qavren.Edge.Ingestion" Version="1.0.0" />
<PackageVersion Include="Qavren.Edge.Ingestion" Version="1.0.0" />
<PackageReference Include="Qavren.Edge.Ingestion" />
paket add Qavren.Edge.Ingestion --version 1.0.0
#r "nuget: Qavren.Edge.Ingestion, 1.0.0"
#:package Qavren.Edge.Ingestion@1.0.0
#addin nuget:?package=Qavren.Edge.Ingestion&version=1.0.0
#tool nuget:?package=Qavren.Edge.Ingestion&version=1.0.0
Qavren.Edge.Ingestion
The document ingestion pipeline of the Qavren.Edge suite (sub-project 3): AddIngestion(), the
document model, plain-text and Markdown extractors, three chunkers (plain, heading-scoped,
token window), a token budget, xxHash128 content hashing for incremental re-indexing, a
cancellable and checkpointed runner, and the chunk writer into Qavren.Edge.VectorData.
dotnet add package Qavren.Edge.Ingestion
builder.UseQavrenEdge(edge => edge
.UseSqliteNative()
.AddSqlite(o => o.DatabaseName = "notes.db")
.AddOnnxEmbeddings()
.AddVectorStore()
.AddIngestion(migrationVersion: 10)); // claims versions 10 AND 11
var pipeline = sp.GetRequiredService<IIngestionPipeline>();
var result = await pipeline.RunAsync(
IngestionSource.Folder(@"C:\docs", "*.md", recursive: true),
options: new IngestionRunOptions { Budget = IngestionBudget.Background });
Console.WriteLine($"{result.Outcome}: +{result.ChunksAdded} -{result.ChunksRemoved}");
Satellites: .Pdf (PdfPig, Apache-2.0, opt-in), .OpenXml (DOCX), .Onnx (the real tokenizer
and a resource-aware throttle), .DataIngestion (the Microsoft.Extensions.DataIngestion shim).
Error codes 6000-6299.
Sub-project README: ingestion/README.md. MIT licensed. https://github.com/qavren-oss/qavren-edge
| 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
- Markdig (>= 1.3.2)
- Microsoft.Extensions.DependencyInjection.Abstractions (>= 10.0.12)
- Microsoft.Extensions.Logging.Abstractions (>= 10.0.12)
- Microsoft.Extensions.Options (>= 10.0.12)
- Microsoft.ML.Tokenizers (>= 2.0.0)
- Qavren.Edge.VectorData (>= 1.0.0)
- System.IO.Hashing (>= 10.0.12)
NuGet packages (4)
Showing the top 4 NuGet packages that depend on Qavren.Edge.Ingestion:
| Package | Downloads |
|---|---|
|
Qavren.Edge.Ingestion.OpenXml
DOCX text extraction for Qavren.Edge.Ingestion, over DocumentFormat.OpenXml. Outline-level heading detection, run merging, and an OpenXmlPartReader streaming path above 8 MiB. Separate from the core because the dependency is 14.66 MB. |
|
|
Qavren.Edge.Ingestion.Pdf
PDF text extraction for Qavren.Edge.Ingestion, over PdfPig. Page-at-a-time and stream-only, with scanned-page detection. PdfPig is Apache-2.0, which is why this extractor is a separate package. |
|
|
Qavren.Edge.Ingestion.Onnx
Bridges Qavren.Edge.Ingestion to the ONNX embedding stack: a chunk tokenizer over IEdgeTokenizer, a throttle over IEdgeResourceMonitor, and a chunk budget derived from the resolved EmbeddingPreset. |
|
|
Qavren.Edge.Ingestion.DataIngestion
Microsoft.Extensions.DataIngestion interop for Qavren.Edge.Ingestion, in both directions. Prerelease only: it references a prerelease abstractions package and ships on that package's cadence, so the stable core is never blocked by it. |
GitHub repositories
This package is not used by any popular GitHub repositories.
| Version | Downloads | Last Updated |
|---|---|---|
| 1.0.0 | 564 | 9/25/2026 |
| 1.0.0-rc.1 | 93 | 9/24/2026 |
| 0.1.0-preview.1 | 93 | 9/13/2026 |