Skip to content

Extractors

Extractors for structured information extraction from text using LLMs.

Modules:

Name Description
base

Core extraction function extract_from_text and related types.

aggregation_utils

Aggregation functions for combining structured outputs across multiple extractions.

chunking

ChunkingExtractor for running extraction over document chunks.

chunking_utils

Tokenization and chunk-iteration utilities used by ChunkingExtractor.

conditional

ConditionalUnionExtractor for multi-pass extraction with chat history.

repeat

RepeatingExtractor for repeated extraction with majority-vote aggregation.

union

UnionExtractor for multi-pass extraction with per-pass parameter overrides.

multi_pass

MultiPassExtractorWithChunking for multi-pass extraction with per-pass parameter overrides on document chunking.