ExtractedEntity

Models & Types · AiDotNet.RetrievalAugmentedGeneration.Graph.Construction

Represents an entity extracted from text during knowledge graph construction.

For Beginners

When building a knowledge graph from text, the first step is identifying entities — the “things” mentioned in the text (people, places, organizations). Each extracted entity has:

  • Name: The text mention (“Albert Einstein”)
  • Label: The entity type (“PERSON”, “ORGANIZATION”, “LOCATION”)
  • Confidence: How sure we are this is a real entity (0.0 to 1.0)
  • Offsets: Where in the text this entity was found

Constructors

Constructor Summary
ExtractedEntity()

Properties

Property Summary
double Confidence { get; set; } Confidence score for this extraction (0.0 to 1.0).
int EndOffset { get; set; } End character offset in the source text.
int StartOffset { get; set; } Start character offset in the source text.
string Description { get; set; } Optional natural-language description of the entity (populated by LLM extraction; empty for the regex/heuristic path).
string Label { get; set; } Entity type label (e.g., PERSON, ORGANIZATION, LOCATION, CONCEPT).
string Name { get; set; } The entity’s surface form (name as it appears in text).