Class DoclingDocumentParser

java.lang.Object

dev.langchain4j.data.document.parser.docling.DoclingDocumentParser

All Implemented Interfaces:

DocumentParser

public class DoclingDocumentParser
extends Object
implements DocumentParser

A DocumentParser backed by a Docling Serve instance.

The operation performed by Docling (convert, hierarchical chunking, hybrid chunking, ...) is controlled by the DocumentRequest template supplied to the DoclingDocumentParser.Builder. The template carries everything except the document source: the parser injects the source of the InputStream being parsed into a fresh copy of the template for each call, then routes the request to the matching Docling endpoint. How the resulting response is turned into a Document is controlled by the document extractor (for conversion requests) and the chunk extractor (for chunk requests); text-only variants are also available.

How Docling is actually called is controlled by the request executor. By default the parser calls the asynchronous convert/chunk endpoints and exposes both a blocking parse(InputStream) and a non-blocking parseAsync(InputStream).

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static final class	DoclingDocumentParser.Builder

Constructor Summary

Constructors

Constructor	Description
DoclingDocumentParser(ai.docling.serve.api.DoclingServeApi doclingClient)	Deprecated, for removal: This API element is subject to removal in a future version. use builder() with DoclingDocumentParser.Builder.doclingClient(DoclingServeApi) instead
DoclingDocumentParser(ai.docling.serve.api.DoclingServeApi doclingClient, ai.docling.serve.api.convert.request.options.ConvertDocumentOptions options)	Deprecated, for removal: This API element is subject to removal in a future version. use builder() with DoclingDocumentParser.Builder.doclingClient(DoclingServeApi) and a DoclingDocumentParser.Builder.documentRequest(DocumentRequest) carrying a ConvertDocumentRequest with the desired ConvertDocumentOptions instead

Method Summary

Modifier and Type	Method	Description
static DoclingDocumentParser.Builder	builder()
Document	parse(InputStream inputStream)	Parses a given InputStream into a Document.
CompletionStage<Document>	parseAsync(InputStream inputStream)	Parses the given InputStream without blocking, returning a stage that completes with the resulting Document.

Methods inherited from class Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

DoclingDocumentParser

@Deprecated(forRemoval=true)
public DoclingDocumentParser(ai.docling.serve.api.DoclingServeApi doclingClient)

Deprecated, for removal: This API element is subject to removal in a future version.

use builder() with DoclingDocumentParser.Builder.doclingClient(DoclingServeApi) instead

DoclingDocumentParser

@Deprecated(forRemoval=true)
public DoclingDocumentParser(ai.docling.serve.api.DoclingServeApi doclingClient,
 ai.docling.serve.api.convert.request.options.ConvertDocumentOptions options)

Deprecated, for removal: This API element is subject to removal in a future version.

use builder() with DoclingDocumentParser.Builder.doclingClient(DoclingServeApi) and a DoclingDocumentParser.Builder.documentRequest(DocumentRequest) carrying a ConvertDocumentRequest with the desired ConvertDocumentOptions instead

Method Details

builder

public static DoclingDocumentParser.Builder builder()

parse

public Document parse(InputStream inputStream)

Description copied from interface: DocumentParser

Parses a given InputStream into a Document. The specific implementation of this method will depend on the type of the document being parsed.

Note: This method does not close the provided InputStream - it is the caller's responsibility to manage the lifecycle of the stream.

Specified by:

parse in interface DocumentParser

Parameters:

inputStream - The InputStream that contains the content of the Document.

Returns:

The parsed Document.

parseAsync

public CompletionStage<Document> parseAsync(InputStream inputStream)

Parses the given InputStream without blocking, returning a stage that completes with the resulting Document. This is the asynchronous counterpart of parse(InputStream) and is intended for reactive or non-blocking pipelines.

Parameters:

inputStream - the document content to parse; the caller owns its lifecycle and this method does not close it

Returns:

a stage completing with the parsed Document