Struct TsrTableInput 

pub struct TsrTableInput {
    pub page: u32,
    pub crop_pdf_pt_bbox: [f32; 4],
    pub render_dpi: f32,
    pub structure_tokens: Vec<String>,
    pub cell_bboxes: Vec<Vec<f32>>,
}

Input for extract_tables_with_structure_mem: one cropped table region plus the raw structure-recovery output for it.

The structure tokens and bboxes are typically produced by an external table-structure recognition model (e.g. SLANet on PaddleOCR) running on a rendered crop of the page. pdf-inspector uses the structure to lay out the cells and pulls the cell text from the native PDF — no OCR involved.

Fields

page: u32

0-indexed page number where the crop was taken from.

crop_pdf_pt_bbox: [f32; 4]

Crop bbox on the page, [x1, y1, x2, y2] in PDF points with top-left origin, relative to the visible page box — the frame of a rendered page image (see extract_text_in_regions_mem).

render_dpi: f32

DPI the crop image was rendered at (e.g. 200.0). Used to convert cell bboxes from image-pixels back to PDF points.

structure_tokens: Vec<String>

Raw structure tokens emitted by the TSR model, in document order. See [tables::structured::parse_structure] for the accepted grammar.

cell_bboxes: Vec<Vec<f32>>

One bbox per cell (in document order, parallel to the cell open-tags in structure_tokens). May be 4-element [x1,y1,x2,y2] or 8-element 4-corner polygon, in crop image-pixel space.

Trait Implementations

impl Clone for TsrTableInput

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for TsrTableInput

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.