API Reference lextract v#0.1.2

Aligns extracted entities back to their positions in source text.

Represents the quality/status of text alignment.

Represents a document with extracted entities and relationships.

Annotates documents with extractions using LLMs.

Represents a character position interval in text.

Integrates semantic text chunking with tokenization for document processing.

Configuration for extraction operations using NimbleOptions for validation.

Represents an input document for information extraction.

Main error aggregator for LeXtract using Splode.

Error class for external service and resource failures.

Error for LLM annotation/inference failures.

Error for template file I/O failures.

Error for tokenizer loading failures.

Error class for validation and format errors.

Error for configuration validation failures.

Error for unknown or invalid format types.

Error for invalid template structure or content.

Error class for processing pipeline failures.

Error for text alignment failures.

Error for text chunking failures.

Error for JSON/YAML parsing failures.

Error for extraction resolution failures.

Error for tokenization failures.

Error class for unexpected errors.

Catch-all error for truly unexpected failures.

Represents a training example for extraction with input text and expected output.

Represents a single extracted entity with alignment information.

Handles detection and parsing of different text formats (JSON, YAML).

Enumeration of supported format types for parsing LLM outputs.

Prompt generation for LLM extraction.

Parses LLM output into Extraction structs.

Generate and manage NimbleOptions schemas from extraction examples.

Analyzes example extractions to infer schema information.

Represents a chunk of text from a document, used for processing long documents.

Represents a token position interval.

Tokenization wrapper using Hugging Face Tokenizers library.