PredictModel

The BinaryToText parameter is optional.

You must specify either the Content parameter with the document's URL or the Body parameter with the document's base64-encoded content. If you use the Body parameter, you must also set BinaryToText to true.

PDF files are limited to 20 MB and 10 pages. For prediction types other than long-document extraction, the service processes only the first page.

The following describes the fields in the Data object for long-document information extraction.

originalFileUrl     The URL of the original file.
predictFile         An array of image URLs parsed from the document that are used for prediction.
data                The detailed prediction results.
angle               The rotation angle of the image. This field is returned only when `NeedRotate` is `true`. 0 indicates upright, 90 indicates rotated to the right, 180 indicates upside down, and 270 indicates rotated to the left.
content             A summary of the text blocks recognized in the image.
height              The height of the image after algorithmic correction.
width               The width of the image after algorithmic correction.
orgHeight           The height of the original image.
orgWidth            The width of the original image.
prism_wnum          The number of recognized text blocks, which is the size of the `prism_wordsInfo` array.

The following describes the fields in the prism_wordsInfo array:

angle                The angle of the text block. This angle only affects the `width` and `height` values. If the angle is -90, 90, -270, or 270, you must swap the `width` and `height` values.
height               The height of the text block.
width                The width of the text block.
pos                  The coordinates of the four points of the bounding rectangle for the text block, arranged clockwise: top-left, top-right, bottom-right, and bottom-left. If `NeedRotate` is `true` and the `angle` of the parent object is not 0, these coordinates are accurate only after you correct the image based on the parent `angle`.
word                 The text content of the block.
tableId              This field exists if `OutputTable` is `true` and the text block is within a table. It indicates the ID of the table.
tableCellId          This field exists if `OutputTable` is `true` and the text block is within a table. It indicates the ID of the table cell.

The following describes the fields for single-character information (charInfo):

word                  The character.
x                     The x-coordinate of the character's top-left corner.
y                     The y-coordinate of the character's top-left corner.
w                     The width of the character.
h                     The height of the character.

The following describes the fields in the prism_tablesInfo array:

tableId            The ID of the table, which corresponds to the `tableId` in the `prism_wordsInfo` object.
xCellSize          The number of cells along the x-axis (columns).
yCellSize          The number of cells along the y-axis (rows).

The following describes the fields for cell information (cellInfos), which includes the spatial topology of cells within the entire table:

tableCellId        The ID of the table cell, which corresponds to the `tableCellId` in the `prism_wordsInfo` object.
word               The text content of the cell.
xsc                `xsc` (xStartCell): The starting column index of the cell. The index is 0-based.
xec                `xec` (xEndCell): The ending column index of the cell. The index is 0-based. If `xsc` and `xec` are both 0, the cell spans one column and is in the first column.
ysc                `ysc` (yStartCell): The starting row index of the cell. The index is 0-based.
yec                `yec` (yEndCell): The ending row index of the cell. The index is 0-based.
pos                The coordinates of the four corners of the cell, arranged clockwise: top-left (X,Y), top-right (X,Y), bottom-right (X,Y), and bottom-left (X,Y).