mirror of
https://github.com/docling-project/docling.git
synced 2026-05-17 13:10:38 +00:00
geoHeil
5b1df788ef
* ci: tighten pre-commit guardrails Signed-off-by: Georg Heiler <georg.kf.heiler@gmail.com> * ci: validate pre-commit guardrail changes Signed-off-by: Georg Heiler <georg.kf.heiler@gmail.com> * ci: switch hook validation to prek Signed-off-by: Georg Heiler <georg.kf.heiler@gmail.com> * ci: exempt active slim plan from max-lines Signed-off-by: Georg Heiler <georg.kf.heiler@gmail.com> * ci: move max-lines config under github Signed-off-by: Georg Heiler <georg.kf.heiler@gmail.com> * ci: fail on uncovered tach modules Signed-off-by: Georg Heiler <georg.kf.heiler@gmail.com> * ci: ignore generated docs in max-lines check Signed-off-by: Georg Heiler <georg.kf.heiler@gmail.com> * ci: clarify local validation tasks Signed-off-by: Georg Heiler <georg.kf.heiler@gmail.com> * docs: refine agent instructions Signed-off-by: Georg Heiler <georg.kf.heiler@gmail.com> * ci: replace mypy with ty (cherry picked from commit 382afbde8f00abfaeba95ea9c8e9cc603f27a2d9) Signed-off-by: Georg Heiler <georg.kf.heiler@gmail.com> * ci: replace justfile with makefile Signed-off-by: Georg Heiler <georg.kf.heiler@gmail.com> --------- Signed-off-by: Georg Heiler <georg.kf.heiler@gmail.com>
33 lines
842 B
Python
33 lines
842 B
Python
from docling_core.types.doc.page import BoundingRectangle, TextCell
|
|
|
|
from docling.utils.layout_postprocessor import LayoutPostprocessor
|
|
|
|
|
|
def _text_cell(index: int) -> TextCell:
|
|
return TextCell(
|
|
index=index,
|
|
rect=BoundingRectangle(
|
|
r_x0=0,
|
|
r_y0=0,
|
|
r_x1=1,
|
|
r_y1=0,
|
|
r_x2=1,
|
|
r_y2=1,
|
|
r_x3=0,
|
|
r_y3=1,
|
|
),
|
|
text=str(index),
|
|
orig=str(index),
|
|
from_ocr=False,
|
|
)
|
|
|
|
|
|
def test_sort_cells_uses_native_cell_index_order() -> None:
|
|
processor = object.__new__(LayoutPostprocessor)
|
|
cells = [_text_cell(3), _text_cell(1), _text_cell(2)]
|
|
|
|
sorted_cells = processor._sort_cells(cells)
|
|
|
|
assert [cell.index for cell in sorted_cells] == [1, 2, 3]
|
|
assert [cell.index for cell in cells] == [3, 1, 2]
|