dawsonia.table_detect.scipy_proj

Module Contents

Functions

table_detect_scipy_proj	get position and size for every tables Parameters
get_table_structure
projections_of_label_tables
simply_order	simplify the position of tables to avoid random position. Parameters
sort_insertion	Sort tables according to LL using Insertion Sort algorithm. Parameters
add_edge	Add the vertical or horizontal edges of a table Parameters
_get_position_old	determines the position of the maximums of a projection of an array (on x or y) therefore determine the position of the rows of a table
get_position	Detect peaks or local maxima in proj_x to find positions of the rows in the table.

Data

logger
__all__

API

dawsonia.table_detect.scipy_proj.logger

‘getLogger(…)’

dawsonia.table_detect.scipy_proj.__all__

(‘table_detect_scipy_proj’, ‘get_table_structure’, ‘projections_of_label_tables’, ‘get_position’, ‘s…

dawsonia.table_detect.scipy_proj.table_detect_scipy_proj(binary_tables: numpy.typing.NDArray[numpy.bool_], table_fmt, preproc_cfg: dawsonia.typing.PreprocConfig, original_image) → tuple[dawsonia.typing.TablePositions, dawsonia.typing.TableSizes, int]

get position and size for every tables Parameters

binary_tables : array 2D binary image of an empty one-page table table_fmt: TableFormat configs from table_formats file only_top_table : boolean if True return only 2 first tables Returns

list_pos3 : list Each element of this list is a list containing the informations of each box of a table. The format of the information is::

   [
        vertical position [int],
        horizontal position [int],
        width [int],
        height [int]
   ]

the position is defined in relation to the position on the whole page
and not in relation to the relative position of the table

l_size3 : list list of size of tables (same order than list_pos3

dawsonia.table_detect.scipy_proj.get_table_structure(table_fmt, preproc_cfg, original_image, label_tables, projections_xy, sensibility, l_size3, list_pos3) → None

dawsonia.table_detect.scipy_proj.projections_of_label_tables(label_tables, nb_labels, min_nb_pixels)

dawsonia.table_detect.scipy_proj.simply_order(tot_p)

simplify the position of tables to avoid random position. Parameters

tot_p : list list of position of every tables Returns

tot_p : list list of position of every tables this list is ordered

dawsonia.table_detect.scipy_proj.sort_insertion(L, LL)

Sort tables according to LL using Insertion Sort algorithm. Parameters

L : list same len than LL LL : list contains elements that can be sorted Returns

L : list contains L sorted elements

dawsonia.table_detect.scipy_proj.add_edge(proj: numpy.typing.NDArray) → numpy.typing.NDArray

Add the vertical or horizontal edges of a table Parameters

proj: array 1D horizontal or vertical projection of a binary table Returns

proj : array 1D horizontal or vertical projection of a binary table, edge add with the max value

dawsonia.table_detect.scipy_proj._get_position_old(proj_x: numpy.typing.NDArray) → tuple[list[int], list[int]]

determines the position of the maximums of a projection of an array (on x or y) therefore determine the position of the rows of a table

Parameters

proj_x: array 1D horizontal or vertical projection of a binary table

Returns

x_pos : list of int list of position of the middle between 2 lines. x_len : list of int list of distance between line

dawsonia.table_detect.scipy_proj.get_position(xs: numpy.typing.NDArray) → tuple[numpy.typing.NDArray, numpy.typing.NDArray]

Detect peaks or local maxima in proj_x to find positions of the rows in the table.

Parameters

proj_x: array 1D horizontal or vertical projection of a binary table

Returns

x_pos : array 1D position of the middle between 2 lines. x_len : array 1D distance between line