Class MarkdownTextSplitter

Namespace

Mythosia.AI.Rag.Splitters

Assembly

Mythosia.AI.Rag.dll

Structure-aware Markdown splitter that understands heading hierarchy (H1–H6), preserves atomic blocks (code fences, tables), and prepends heading breadcrumbs to each chunk so that vector search retrieves contextually rich fragments.

public class MarkdownTextSplitter : ITextSplitter

Inheritance

object

MarkdownTextSplitter

Implements

ITextSplitter

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Constructors

MarkdownTextSplitter()

public MarkdownTextSplitter()

MarkdownTextSplitter(int, int)

public MarkdownTextSplitter(int chunkSize, int chunkOverlap = 200)

Parameters

chunkSize int

chunkOverlap int

Properties

ChunkOverlap

Number of overlapping characters carried from the previous chunk.

public int ChunkOverlap { get; set; }

Property Value

int

ChunkSize

Maximum characters per chunk (excluding the prepended breadcrumb).

public int ChunkSize { get; set; }

Property Value

int

IncludeHeadingBreadcrumb

When true, each chunk is prefixed with the heading path that leads to its content (e.g. "# Doc Title\n## Section\n### Sub-section\n\n"). This dramatically improves retrieval relevance. Default is true.

public bool IncludeHeadingBreadcrumb { get; set; }

Property Value

bool

MinSplitHeadingLevel

Minimum heading level that triggers a new section split. 1 = split on all headings (#–######), 2 = ignore H1, etc. Default: 1.

public int MinSplitHeadingLevel { get; set; }

Property Value

int

Methods

Split(RagDocument)

Splits a document into chunks. Implementations may split by character count, token count, sentence boundary, etc.

public IReadOnlyList<RagChunk> Split(RagDocument document)

Parameters

document RagDocument

The document to split.

Returns

IReadOnlyList<RagChunk>

An ordered list of chunks.