Crate edgeparse_core 

EdgeParse Core Library

High-performance PDF-to-structured-data extraction engine. Implements a 20-stage processing pipeline for extracting text, tables, images, and semantic structure from PDF documents.

Modules

api

API module — Configuration and public types.

hybrid

Geometric hybrid triage + merge (Docling HTTP behind feature = "hybrid"). Hybrid extraction mode — geometric triage + completeness-aware merge.

models

Data model types for EdgeParse.

output

Output generators — TOC, JSON, Markdown, HTML, Text, CSV, Annotated PDF.

pdf

PDF loading layer — document loading, text extraction, line extraction.

pipeline

20-stage processing pipeline orchestrator.

tagged

Tagged PDF processor.

utils

Shared utility functions.

Structs

ExtractSession

Intermediate state between extract (plan) and assemble (finish).

Enums

EdgePdfError

Top-level error type for EdgeParse operations.

OcrAssembleMode

How assemble should obtain OCR for candidates.

Functions

assemble

Run the pipeline after injecting OCR-derived tables from session candidates.

convert

Main entry point: convert a PDF file to structured data.

convert_bytes

Convert a PDF from an in-memory byte slice to structured data.

extract_session

Extract pages, chunks, and OCR candidates without recognizing text.

Type Aliases

ProgressFn

Progress callback: (phase, done, total).