Method predict

tools/infer_doc.py:421–776

Predicts the layout parsing result for the given input. Args: input (Union[str, list[str], np.ndarray, list[np.ndarray]]): Input image path, list of image paths, numpy array of an image, or list of

(
        self,
        input: Union[str, list[str], np.ndarray, list[np.ndarray]],
        use_doc_orientation_classify: Union[bool, None] = False,
        use_doc_unwarping: Union[bool, None] = False,
        use_layout_detection: Union[bool, None] = None,
        use_chart_recognition: Union[bool, None] = None,
        layout_threshold: Optional[Union[float, dict]] = None,
        layout_nms: Optional[bool] = None,
        layout_unclip_ratio: Optional[Union[float, Tuple[float, float],
                                            dict]] = None,
        layout_merge_bboxes_mode: Optional[str] = None,
        use_queues: Optional[bool] = None,
        prompt_label: Optional[Union[str, None]] = None,
        format_block_content: Union[bool, None] = None,
        repetition_penalty: Optional[float] = None,
        temperature: Optional[float] = None,
        top_p: Optional[float] = None,
        min_pixels: Optional[int] = None,
        max_pixels: Optional[int] = None,
        max_new_tokens: Optional[int] = None,
        merge_layout_blocks: Optional[bool] = None,
        markdown_ignore_labels: Optional[list[str]] = None,
        **kwargs,
    )

        return parsing_res_lists, table_res_lists, imgs_in_doc

    def predict(
        self,
        input: Union[str, list[str], np.ndarray, list[np.ndarray]],
        use_doc_orientation_classify: Union[bool, None] = False,
        use_doc_unwarping: Union[bool, None] = False,
        use_layout_detection: Union[bool, None] = None,
        use_chart_recognition: Union[bool, None] = None,
        layout_threshold: Optional[Union[float, dict]] = None,
        layout_nms: Optional[bool] = None,
        layout_unclip_ratio: Optional[Union[float, Tuple[float, float],
                                            dict]] = None,
        layout_merge_bboxes_mode: Optional[str] = None,
        use_queues: Optional[bool] = None,
        prompt_label: Optional[Union[str, None]] = None,
        format_block_content: Union[bool, None] = None,
        repetition_penalty: Optional[float] = None,
        temperature: Optional[float] = None,
        top_p: Optional[float] = None,
        min_pixels: Optional[int] = None,
        max_pixels: Optional[int] = None,
        max_new_tokens: Optional[int] = None,
        merge_layout_blocks: Optional[bool] = None,
        markdown_ignore_labels: Optional[list[str]] = None,
        **kwargs,
    ) -> PaddleOCRVLResult:
        """
        Predicts the layout parsing result for the given input.

        Args:
            input (Union[str, list[str], np.ndarray, list[np.ndarray]]): Input image path, list of image paths,
                                                                        numpy array of an image, or list of numpy arrays.
            use_doc_orientation_classify (Optional[bool]): Whether to use document orientation classification.
            use_doc_unwarping (Optional[bool]): Whether to use document unwarping.
            layout_threshold (Optional[float]): The threshold value to filter out low-confidence predictions. Default is None.
            layout_nms (bool, optional): Whether to use layout-aware NMS. Defaults to False.
            layout_unclip_ratio (Optional[Union[float, Tuple[float, float]]], optional): The ratio of unclipping the bounding box.
                Defaults to None.
                If it's a single number, then both width and height are used.
                If it's a tuple of two numbers, then they are used separately for width and height respectively.
                If it's None, then no unclipping will be performed.
            layout_merge_bboxes_mode (Optional[str], optional): The mode for merging bounding boxes. Defaults to None.
            use_queues (Optional[bool], optional): Whether to use queues. Defaults to None.
            prompt_label (Optional[Union[str, None]], optional): The label of the prompt in ['ocr', 'formula', 'table', 'chart']. Defaults to None.
            format_block_content (Optional[bool]): Whether to format the block content. Default is None.
            repetition_penalty (Optional[float]): The repetition penalty parameter used for VL model sampling. Default is None.
            temperature (Optional[float]): Temperature parameter used for VL model sampling. Default is None.
            top_p (Optional[float]): Top-p parameter used for VL model sampling. Default is None.
            min_pixels (Optional[int]): The minimum number of pixels allowed when the VL model preprocesses images. Default is None.
            max_pixels (Optional[int]): The maximum number of pixels allowed when the VL model preprocesses images. Default is None.
            max_new_tokens (Optional[int]): The maximum number of new tokens. Default is None.
            merge_layout_blocks (Optional[bool]): Whether to merge layout blocks. Default is None.
            markdown_ignore_labels (Optional[list[str]]): The list of ignored markdown labels. Default is None.
            **kwargs (Any): Additional settings to extend functionality.

        Returns:
            PaddleOCRVLResult: The predicted layout parsing result.
        """
        model_settings = self.get_model_settings(