Class ChunkIndexUnit

Namespace

LMKit.Retrieval.ChunkQuality

Assembly

LM-Kit.NET.dll

One ready-to-embed unit produced by EvaluatePage(IReadOnlyList<string>): either a single source chunk, or several contiguous source chunks folded together by the neighbor-merge policy.

public sealed class ChunkIndexUnit

Inheritance

object

ChunkIndexUnit

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.ReferenceEquals(object, object)

object.ToString()

Remarks

Source chunks are addressed by their zero-based position in the list passed to EvaluatePage(IReadOnlyList<string>). The covered range is always contiguous, so a caller that tracks its own chunk coordinates (for example character offsets into a source page) can address a merged unit by the span from the first source chunk's start to the last source chunk's end.

Properties

Assessment

The assessment of the unit's primary chunk (the one that carried the indexable content; for a heading merged into a body chunk, the body chunk).

EmbeddingText

The normalized text to embed for this unit. For a merged unit this is the concatenation of the merged chunks' normalized texts.

FirstSourceChunkIndex

Zero-based index of the first source chunk covered by this unit.

IsMerged

true when this unit folds more than one source chunk.

LastSourceChunkIndex

Zero-based index of the last source chunk covered by this unit. Equal to FirstSourceChunkIndex when the unit is a single chunk.