Add an element to the section

Add a section to the document

Determine document type from file extension

Process document using Qwen2.5-VL Args: file_path: Path to document file max_concurrent: Maximum num

Convert to markdown representation

Read image file and encode as base64

Completely reconstruct tables using python-docx Args: content: Original markdown content tables: Tab

Call vision model API to get markdown representation of document Args: image_content: Base64 encoded image

Convert complex DataFrames to markdown tables Handles cases where the standard to_markdown might fail Args:

Convert a python-docx table to markdown Args: table: python-docx table object Returns:

Detect and format tables directly from the content when docx2python fails Args: content: Markdown content

Remove page marker headings from the markdown content Args: content: Markdown content with page markers

Convert python-docx table to markdown table format Args: table: python-docx Table object Re

Configure API settings for all instances Args: api_key: API key for the vision model base_url: Base

Create enhanced Gradio UI with batch processing

Factory method to get appropriate processor for a file

Process a single file

Convert to dictionary representation

Convert document to markdown format

Call vision model API with multiple images Args: image_contents: List of base64 encoded images page_

Convert pandas DataFrame to markdown table

Extract text from PDF using PyPDF2

Extract tables directly from python-docx and insert them as proper markdown tables Args: doc: python-docx Docume

Process PDF document with parallel page processing Args: file_path: Path to PDF file document: Docum

Process PDF document with multi-image batches for improved context Args: file_path: Path to PDF file

Process tables extracted by docx2python and replace mammoth's table output with better markdown tables Args:

Detect document type from file extension

Process all files in directory

Initialize the document processor

Initialize vision document processor Args: api_key: API key for the vision model base_url: Base URL

Initialize PDF processor Args: use_vision_for_complex: Whether to use vision processing for complex PDFs

Process multiple pages in a single API call Args: temp_dir: Temporary directory page_indices: List o

Process a single page in the PDF Args: temp_dir: Temporary directory page_index: Index of the page

Get direct markdown if available, otherwise generate from structure Returns: str: Markdown representation of the

Process a document and return structured content Args: file_path: Path to the document file

Process code file

Process text file

Process Excel file and convert to markdown Args: file_path: Path to Excel file Returns:

Process Jupyter Notebook

Process DOCX document with improved table handling

Process markdown file

Process PDF document

Process PowerPoint presentation

Process multiple documents and return summary

Save as markdown file

Convert to dictionary representation

Convert to dictionary representation

Convert to markdown representation