Skip to main content

​
Parameters

ParameterTypeDefaultDescription
configDoclingLoaderConfigRequiredConfiguration object for Docling loading behavior

​
Functions

​
__init__

Initialize the Docling loader. Parameters:
  • config (DoclingLoaderConfig): Configuration object specifying extraction mode, chunking strategy, and other processing options

​
_create_converter

Create and configure a DocumentConverter instance with OCR and pipeline options. Returns:
  • DocumentConverter: Configured DocumentConverter instance

​
_create_pdf_pipeline_options

Create PDF pipeline options with OCR configuration. Returns:
  • PdfPipelineOptions: Configured PdfPipelineOptions instance

​
_create_ocr_options

Create OCR options based on configured backend. Returns:
  • Union[RapidOcrOptions, TesseractCliOcrOptions]: Configured OCR options instance

​
_create_chunker

Create and configure a chunker instance based on config. Returns:
  • Optional[any]: Configured chunker instance or None if chunking is not available

​
_is_url

Check if the source is a URL. Parameters:
  • source (Union[str, Path]): Source to check
Returns:
  • bool: True if source is a URL

​
_validate_source

Validate a source path or URL. Parameters:
  • source (Union[str, Path]): File path or URL to validate
Returns:
  • Union[str, Path]: Validated source

​
_convert_document

Convert a single document using Docling. Parameters:
  • source (Union[str, Path]): Path or URL to the document
Returns:
  • Optional[DoclingDocument]: DoclingDocument instance or None if conversion failed

​
_handle_conversion_error

Handle conversion errors using base class method. Parameters:
  • source (Union[str, Path]): Source that failed conversion
  • error (Exception): Error that occurred
Returns:
  • None: Returns None for document conversion failures

​
_extract_markdown

Extract document as markdown. Parameters:
  • dl_doc (DoclingDocument): DoclingDocument instance
  • source (Union[str, Path]): Original source path/URL
  • document_id (str): Document ID from base class
Returns:
  • List[Document]: List containing a single Document with markdown content

​
_extract_chunks

Extract document as semantic chunks. Parameters:
  • dl_doc (DoclingDocument): DoclingDocument instance
  • source (Union[str, Path]): Original source path/URL
  • document_id (str): Document ID from base class
Returns:
  • List[Document]: List of Documents, one per chunk

​
_add_docling_metadata

Add Docling-specific metadata to the metadata dict. Parameters:
  • metadata (dict): Metadata dictionary to update
  • dl_doc (DoclingDocument): DoclingDocument instance

​
load

Load and process documents from the given source(s). Parameters:
  • source (Union[str, Path, List[Union[str, Path]]]): Single file path/URL or list of file paths/URLs
Returns:
  • List[Document]: List of processed Document objects

​
aload

Asynchronously load and process documents. Parameters:
  • source (Union[str, Path, List[Union[str, Path]]]): Single file path/URL or list of file paths/URLs
Returns:
  • List[Document]: List of processed Document objects

​
batch

Load documents from multiple sources. Parameters:
  • sources (List[Union[str, Path]]): List of file paths/URLs
Returns:
  • List[Document]: List of processed Document objects from all sources

​
abatch

Asynchronously load documents from multiple sources with parallel processing. Parameters:
  • sources (List[Union[str, Path]]): List of file paths/URLs
Returns:
  • List[Document]: List of processed Document objects from all sources

​
get_supported_extensions

Get list of file extensions supported by Docling. Returns:
  • List[str]: List of supported extensions including dot (e.g., ‘.pdf’)

​
can_load

Check if this loader can handle the given source. Parameters:
  • source (Union[str, Path]): File path or URL to check
Returns:
  • bool: True if the source can be loaded, False otherwise
⌘I