Анализ структуры и вёрстки PDF-документов

★ 1.3K

PDF Document Layout Analysis is a Docker microservice by HURIDOCS for intelligent PDF layout parsing: OCR, segmentation, and classification of page parts. Reach for it when a PDF needs more than "extract the text" — you want to understand its structure: identify titles, paragraphs, pictures, tables, and formulas, determine the correct reading order, and convert the document to Markdown or HTML (with automatic translation via Ollama). It is especially useful on complex multi-column and scanned documents where block order and types matter for downstream RAG or archiving. Built on Clean Architecture principles, it offers both a friendly Gradio web interface and a REST API for integration. Its focus is analyzing page structure and layout with reading-order detection, not plain flat OCR (for "document → text/JSON" there are OCRmyPDF, PaddleOCR, text-extract-api): it is a layout-understanding component you place before content extraction.