Merge in RR/cv-analysis from new_pyinfra to master
Squashed commit of the following:
commit f7a01a90aad1c402ac537de5bdf15df628ad54df
Author: Julius Unverfehrt <julius.unverfehrt@iqser.com>
Date: Wed Jul 27 10:40:59 2022 +0200
fix typo
commit ff4d549fac5b612c2d391ae85823c5eca1e91916
Author: Julius Unverfehrt <julius.unverfehrt@iqser.com>
Date: Wed Jul 27 10:34:04 2022 +0200
adjust build scripts for new pyinfra
commit ecd70f60d46406d8b6cc7f36a1533d706c917ca8
Author: Julius Unverfehrt <julius.unverfehrt@iqser.com>
Date: Wed Jul 27 09:42:55 2022 +0200
simplify logging by using default configurations
commit 20193c14c940eed2b0a7a72058167e26064119d0
Author: Julius Unverfehrt <julius.unverfehrt@iqser.com>
Date: Tue Jul 26 17:16:57 2022 +0200
tidy-up, refactor config logic to not dependent on external files
commit d8069cd4d404a570bb04a04278161669d1c83332
Author: Isaac Riley <Isaac.Riley@iqser.com>
Date: Tue Jul 26 15:14:59 2022 +0200
update pyinfra
commit c3bc11037cca9baf016043ab997c566f5b4a2586
Author: Isaac Riley <Isaac.Riley@iqser.com>
Date: Tue Jul 26 15:09:14 2022 +0200
repair tests
commit 6f4e4f2863ee16ae056c1d432f663858c5f10221
Author: Isaac Riley <Isaac.Riley@iqser.com>
Date: Tue Jul 26 14:52:38 2022 +0200
updated server logic to work with new pyinfra; update scripts for pyinfra as submodule
commit 2a18dba81de5ee84d0bdf0e77f478693e8d8aef4
Author: Isaac Riley <Isaac.Riley@iqser.com>
Date: Tue Jul 26 14:10:41 2022 +0200
formatting
commit d87ce9328de9aa2341228af9b24473d5e583504e
Author: Isaac Riley <Isaac.Riley@iqser.com>
Date: Tue Jul 26 14:10:11 2022 +0200
make server logic compatible with new pyinfra
41 lines
2.1 KiB
Python
41 lines
2.1 KiB
Python
import cv2
|
|
import numpy as np
|
|
import pytest
|
|
|
|
from cv_analysis.figure_detection.text import (
|
|
remove_primary_text_regions,
|
|
apply_threshold_to_image,
|
|
)
|
|
from test.utils.utils import powerset
|
|
|
|
|
|
@pytest.mark.parametrize("error_tolerance", [0.07])
|
|
@pytest.mark.parametrize("background_color", [255, 220])
|
|
class TestFindPrimaryTextRegions:
|
|
def test_blank_page_stays_blank(self, background, error_tolerance):
|
|
result_page = remove_primary_text_regions(background)
|
|
|
|
np.testing.assert_equal(result_page, apply_threshold_to_image(background))
|
|
|
|
@pytest.mark.parametrize("image_size", [(200, 200), (500, 500), (800, 800)])
|
|
def test_page_without_text_keeps_images(self, page_with_images, error_tolerance):
|
|
result_page = remove_primary_text_regions(page_with_images)
|
|
np.testing.assert_equal(result_page, apply_threshold_to_image(page_with_images))
|
|
|
|
@pytest.mark.parametrize("font_scale", [1, 1.5, 2])
|
|
@pytest.mark.parametrize("font_style", [cv2.FONT_HERSHEY_SIMPLEX, cv2.FONT_HERSHEY_COMPLEX])
|
|
@pytest.mark.parametrize("text_types", powerset(["body", "header", "caption"]))
|
|
def test_page_with_only_text_gets_text_removed(self, page_with_text, error_tolerance):
|
|
result_page = remove_primary_text_regions(page_with_text)
|
|
relative_error = np.sum(result_page != apply_threshold_to_image(page_with_text)) / result_page.size
|
|
assert relative_error <= error_tolerance
|
|
|
|
@pytest.mark.parametrize("image_size", [(200, 200), (500, 500), (800, 800)])
|
|
@pytest.mark.parametrize("font_scale", [1, 1.5, 2])
|
|
@pytest.mark.parametrize("font_style", [cv2.FONT_HERSHEY_SIMPLEX, cv2.FONT_HERSHEY_COMPLEX])
|
|
@pytest.mark.parametrize("text_types", powerset(["body", "header", "caption"]))
|
|
def test_page_with_images_and_text_keeps_images(self, page_with_images_and_text, error_tolerance):
|
|
result_page = remove_primary_text_regions(page_with_images_and_text)
|
|
relative_error = np.sum(result_page != apply_threshold_to_image(page_with_images_and_text)) / result_page.size
|
|
assert relative_error <= error_tolerance
|