Module entity_extraction 

Entity extraction for automatic identity linking

Provides lightweight Named Entity Recognition (NER) for:

• @mentions (e.g., @ronaldo, @acme-corp)
• Email addresses
• URLs with domain extraction
• Capitalized names (simple heuristic)
• Known identity aliases (database lookup)

Invariants

• Extraction never panics on any input
• Empty/whitespace input returns empty results
• Duplicate mentions are deduplicated with count
• Results are sorted by first occurrence position

Performance

• Regex patterns are compiled once (lazy_static)
• Single pass through text for pattern matching
• Bounded output: max 100 entities per text

Structs

ExtractedEntity

An extracted entity from text

ExtractionConfig

Configuration for entity extraction

ExtractionResult

Result of entity extraction

Enums

ExtractedEntityType

Type of extracted entity

Functions

auto_link_memory

Auto-link entities found in a memory’s content to identities.

extract_entities

Extract entities from text content.