Rag.NET.Chunking
1.0.0
dotnet add package Rag.NET.Chunking --version 1.0.0
NuGet\Install-Package Rag.NET.Chunking -Version 1.0.0
<PackageReference Include="Rag.NET.Chunking" Version="1.0.0" />
<PackageVersion Include="Rag.NET.Chunking" Version="1.0.0" />
<PackageReference Include="Rag.NET.Chunking" />
paket add Rag.NET.Chunking --version 1.0.0
#r "nuget: Rag.NET.Chunking, 1.0.0"
#:package Rag.NET.Chunking@1.0.0
#addin nuget:?package=Rag.NET.Chunking&version=1.0.0
#tool nuget:?package=Rag.NET.Chunking&version=1.0.0
Rag.NET.Chunking
Advanced chunking strategies for Rag.NET: embedding-based semantic chunking, token-aware windowing, late chunking, LLM proposition extraction, hierarchical merging and code-aware chunking — pick per corpus what the core package's recursive default cannot express.
Install
dotnet add package Rag.NET.Chunking
Install alongside the core pipeline package (dotnet add package Rag.NET), which supplies
the AddRagNet(...) builder the strategies register into.
Setup
Inside your AddRagNet(...) builder callback — one strategy at a time; the last
registration wins:
using Rag.NET.Chunking.Semantic;
rag.UseSemanticChunking();
Example
Semantic chunking groups sentences by embedding similarity; token-aware chunking counts real tokenizer tokens instead of characters:
using Rag.NET.Chunking.Semantic;
using Rag.NET.Chunking.TokenAware;
using Rag.NET.Models.Options;
// Tuned semantic chunking:
rag.UseSemanticChunking(new SemanticChunkingOptions
{
BreakpointPercentile = 0.25f, // lower = more, smaller chunks
MinChunkSize = 100, // characters; undersized groups merge with neighbours
MaxChunkSize = 1500, // characters; oversized groups split at sentences
});
// Or: token-aware windows sized for your embedding model.
rag.UseTokenAwareChunking(o =>
{
o.ModelName = "gpt-4"; // selects the tokenizer encoding
o.WindowSizeTokens = 256;
o.OverlapTokens = 32;
});
UseHierarchicalMerging, UsePropositionChunking, UseLateChunking (pair it with
Rag.NET.Embeddings.Onnx token embeddings) and UseCodeChunking follow the same shape.
Full guide
| 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
- Microsoft.Extensions.Logging.Abstractions (>= 10.0.12)
- Microsoft.ML.Tokenizers (>= 1.0.3)
- Microsoft.ML.Tokenizers.Data.Cl100kBase (>= 1.0.3)
- Rag.NET.Abstractions (>= 1.0.0)
NuGet packages (1)
Showing the top 1 NuGet packages that depend on Rag.NET.Chunking:
| Package | Downloads |
|---|---|
|
Rag.NET.Chunking.Templates
Domain-specific chunking templates for Rag.NET (Legal, Book, Academic Paper, Q&A Pairs, Email, Resume) |
GitHub repositories
This package is not used by any popular GitHub repositories.