文档转换器

方法：

• convert –
  转换
• convert_all –
  将所有内容转换
• initialize_pipeline –

  初始化所选格式的转换流水线。

属性：

• allowed_formats –
  允许的格式
• format_to_options –
• initialized_pipelines (Dict[Type[BasePipeline], BasePipeline]) –

allowed_formats = (
    allowed_formats
    if allowed_formats is not None
    else [e for e in InputFormat]
)

format_to_options = {format: _get_default_option(format=format) if (custom_option := get(format)) is None else _1RJYVPLzQgeyfor format in allowed_formats}

initialized_pipelines: Dict[
    Type[BasePipeline], BasePipeline
] = {}

convert(
    source: Union[Path, str, DocumentStream],
    headers: Optional[Dict[str, str]] = None,
    raises_on_error: bool = True,
    max_num_pages: int = maxsize,
    max_file_size: int = maxsize,
    page_range: PageRange = DEFAULT_PAGE_RANGE,
) -> ConversionResult

convert_all(
    source: Iterable[Union[Path, str, DocumentStream]],
    headers: Optional[Dict[str, str]] = None,
    raises_on_error: bool = True,
    max_num_pages: int = maxsize,
    max_file_size: int = maxsize,
    page_range: PageRange = DEFAULT_PAGE_RANGE,
) -> Iterator[ConversionResult]

initialize_pipeline(format: InputFormat)

基础: BaseModel

属性：

• assembled (AssembledUnit) –
• document (DoclingDocument) –
• errors (List[ErrorItem]) –
• input (InputDocument) –
• legacy_document –
• pages (列表[页面]) –
• status (ConversionStatus) –
• timings (Dict[str, ProfilingItem]) –

assembled: AssembledUnit = AssembledUnit()

document: DoclingDocument = _EMPTY_DOCLING_DOC

errors: List[ErrorItem] = []

input: InputDocument

legacy_document

pages: List[Page] = []

status: ConversionStatus = PENDING

timings: Dict[str, ProfilingItem] = {}

基类: str, Enum

属性：

• FAILURE –
  失败
• PARTIAL_SUCCESS –
  部分成功
• PENDING –
  待处理
• SKIPPED –
• STARTED –
• SUCCESS –

FAILURE = 'failure'

PARTIAL_SUCCESS = 'partial_success'

PENDING = 'pending'

SKIPPED = 'skipped'

STARTED = 'started'

SUCCESS = 'success'

基础: BaseModel

方法：

• set_optional_field_default –

属性：

• backend (类型[AbstractDocumentBackend]) –
• model_config –
• pipeline_cls (类型[基础流水线]) –
• pipeline_options (可选[PipelineOptions]) –

backend: Type[AbstractDocumentBackend]

model_config = ConfigDict(arbitrary_types_allowed=True)

pipeline_cls: Type[BasePipeline]

pipeline_options: Optional[PipelineOptions] = None

set_optional_field_default() -> FormatOption

基类: str, Enum

文档后端解析器支持的文档格式。

属性：

• ASCIIDOC –
• CSV –
• DOCX –
• HTML –
• IMAGE –
• JSON_DOCLING –
• MD –
• PDF –
• PPTX –
• XLSX –
• XML_JATS –
• XML_USPTO –

ASCIIDOC = 'asciidoc'

CSV = 'csv'

DOCX = 'docx'

HTML = 'html'

IMAGE = 'image'

JSON_DOCLING = 'json_docling'

MD = 'md'

PDF = 'pdf'

PPTX = 'pptx'

XLSX = 'xlsx'

XML_JATS = 'xml_jats'

XML_USPTO = 'xml_uspto'

基类: FormatOption

方法：

• set_optional_field_default –

属性：

• backend (类型[AbstractDocumentBackend]) –
• model_config –
• pipeline_cls (类型) –
• pipeline_options (可选[PipelineOptions]) –

backend: Type[AbstractDocumentBackend] = (
    DoclingParseV2DocumentBackend
)

model_config = ConfigDict(arbitrary_types_allowed=True)

pipeline_cls: Type = StandardPdfPipeline

pipeline_options: Optional[PipelineOptions] = None

set_optional_field_default() -> FormatOption

基础： FormatOption

方法：

• set_optional_field_default –

属性：

• backend (类型[AbstractDocumentBackend]) –
• model_config –
• pipeline_cls (类型) –
• pipeline_options (可选[PipelineOptions]) –

backend: Type[AbstractDocumentBackend] = (
    DoclingParseV2DocumentBackend
)

model_config = ConfigDict(arbitrary_types_allowed=True)

pipeline_cls: Type = StandardPdfPipeline

pipeline_options: Optional[PipelineOptions] = None

set_optional_field_default() -> FormatOption

基础: PaginatedPipeline

方法：

• download_models_hf –
• execute –
• get_default_options –
• get_ocr_model –
• get_picture_description_model –
• initialize_page –
  初始化页面
• is_backend_supported –
  后端是否支持

属性：

• build_pipe –
  构建管道
• enrichment_pipe –
• glm_model –
• keep_backend –
• keep_images –
  保持图像
• pipeline_options (PdfPipelineOptions) –

build_pipe = [
    PagePreprocessingModel(
        options=PagePreprocessingOptions(
            images_scale=images_scale
        )
    ),
    ocr_model,
    LayoutModel(
        artifacts_path=artifacts_path,
        accelerator_options=accelerator_options,
    ),
    TableStructureModel(
        enabled=do_table_structure,
        artifacts_path=artifacts_path,
        options=table_structure_options,
        accelerator_options=accelerator_options,
    ),
    PageAssembleModel(options=PageAssembleOptions()),
]

enrichment_pipe = [
    CodeFormulaModel(
        enabled=do_code_enrichment or do_formula_enrichment,
        artifacts_path=artifacts_path,
        options=CodeFormulaModelOptions(
            do_code_enrichment=do_code_enrichment,
            do_formula_enrichment=do_formula_enrichment,
        ),
        accelerator_options=accelerator_options,
    ),
    DocumentPictureClassifier(
        enabled=do_picture_classification,
        artifacts_path=artifacts_path,
        options=DocumentPictureClassifierOptions(),
        accelerator_options=accelerator_options,
    ),
    picture_description_model,
]

glm_model = ReadingOrderModel(options=ReadingOrderOptions())

keep_backend = True

keep_images = (
    generate_page_images
    or generate_picture_images
    or generate_table_images
)

pipeline_options: PdfPipelineOptions

download_models_hf(
    local_dir: Optional[Path] = None, force: bool = False
) -> Path

execute(
    in_doc: InputDocument, raises_on_error: bool
) -> ConversionResult

get_default_options() -> PdfPipelineOptions

get_ocr_model(
    artifacts_path: Optional[Path] = None,
) -> Optional[BaseOcrModel]

get_picture_description_model(
    artifacts_path: Optional[Path] = None,
) -> Optional[PictureDescriptionBaseModel]

initialize_page(
    conv_res: ConversionResult, page: Page
) -> Page

is_backend_supported(backend: AbstractDocumentBackend)

基础： FormatOption

方法：

• set_optional_field_default –

属性：

• backend (类型[AbstractDocumentBackend]) –
• model_config –
• pipeline_cls (类型) –
• pipeline_options (可选[PipelineOptions]) –

backend: Type[AbstractDocumentBackend] = (
    MsWordDocumentBackend
)

model_config = ConfigDict(arbitrary_types_allowed=True)

pipeline_cls: Type = SimplePipeline

pipeline_options: Optional[PipelineOptions] = None

set_optional_field_default() -> FormatOption

基础： FormatOption

方法：

• set_optional_field_default –

属性：

• backend (类型[AbstractDocumentBackend]) –
• model_config –
• pipeline_cls (类型) –
• pipeline_options (可选[PipelineOptions]) –

backend: Type[AbstractDocumentBackend] = (
    MsPowerpointDocumentBackend
)

model_config = ConfigDict(arbitrary_types_allowed=True)

pipeline_cls: Type = SimplePipeline

pipeline_options: Optional[PipelineOptions] = None

set_optional_field_default() -> FormatOption

基础： FormatOption

方法：

• set_optional_field_default –

属性：

• backend (类型[AbstractDocumentBackend]) –
• model_config –
• pipeline_cls (类型) –
• pipeline_options (可选[PipelineOptions]) –

backend: Type[AbstractDocumentBackend] = (
    MarkdownDocumentBackend
)

model_config = ConfigDict(arbitrary_types_allowed=True)

pipeline_cls: Type = SimplePipeline

pipeline_options: Optional[PipelineOptions] = None

set_optional_field_default() -> FormatOption

基础： FormatOption

方法：

• set_optional_field_default –

属性：

• backend (类型[AbstractDocumentBackend]) –
• model_config –
• pipeline_cls (类型) –
• pipeline_options (可选[PipelineOptions]) –

backend: Type[AbstractDocumentBackend] = AsciiDocBackend

model_config = ConfigDict(arbitrary_types_allowed=True)

pipeline_cls: Type = SimplePipeline

pipeline_options: Optional[PipelineOptions] = None

set_optional_field_default() -> FormatOption

基础： FormatOption

方法：

• set_optional_field_default –

属性：

• backend (类型[AbstractDocumentBackend]) –
• model_config –
• pipeline_cls (类型) –
• pipeline_options (可选[PipelineOptions]) –

backend: Type[AbstractDocumentBackend] = HTMLDocumentBackend

model_config = ConfigDict(arbitrary_types_allowed=True)

pipeline_cls: Type = SimplePipeline

pipeline_options: Optional[PipelineOptions] = None

set_optional_field_default() -> FormatOption

基础类: BasePipeline

简单模型管道.

该类目前用于生成直接的DoclingDocument输出的格式/后端。

方法：

• execute –
• get_default_options –
• is_backend_supported –
  后端是否支持

属性：

• build_pipe (List[Callable]) –
• enrichment_pipe (List[GenericEnrichmentModel[Any]]) –
• keep_images –
• pipeline_options –

build_pipe: List[Callable] = []

enrichment_pipe: List[GenericEnrichmentModel[Any]] = []

keep_images = False

pipeline_options = pipeline_options

execute(
    in_doc: InputDocument, raises_on_error: bool
) -> ConversionResult

get_default_options() -> PipelineOptions

is_backend_supported(backend: AbstractDocumentBackend)