Commit 8c79b4a9 authored by Jan Reimes's avatar Jan Reimes
Browse files

🎨 style(extraction): update pdf converter backend to use DoclingParseDocumentBackend

parent 6d060008

packages/3gpp-ai/threegpp_ai/operations/extraction.py

+1 −1
Original line number Diff line number Diff line
@@ -90,13 +90,13 @@ def _create_pdf_converter(vlm_options: VlmOptions | None = None) -> DocumentConv 
        # VLM pipeline with Granite Docling

        vlm_opts = VlmPipelineOptions(

            do_picture_description=vlm_options.enable_picture_description,

            picture_description_options=None,  # Use default Granite Docling

            images_scale=2.0,

            generate_picture_images=True,

            vlm_options=VlmConvertOptions.from_preset("granite_docling"),

        )

        pdf_format_option = FormatOption(

            pipeline_cls=VlmPipeline,

            backend=DoclingParseDocumentBackend,

            pipeline_options=vlm_opts,

        )

    return DocumentConverter(format_options={InputFormat.PDF: pdf_format_option})