1147 commits

This branch

This branch

All branches

Author	SHA1	Message	Date
cneud	558867eb24	fix typo	2025-10-01 00:04:07 +02:00
Robert Sachunsky	3aa7ad04fa	📝 update changelog	2025-09-30 23:14:52 +02:00
Robert Sachunsky	f0de1adabf	rm loky dependency	2025-09-30 23:12:18 +02:00
Robert Sachunsky	7daec392b9	Dockerfile: fix up CUDA installation for mixed TF/Torch	2025-09-30 22:10:45 +02:00
Robert Sachunsky	ad129ed46c	CI: remove OS from model cache keys	2025-09-30 22:05:53 +02:00
Robert Sachunsky	c86e59f481	CI: update model key, split up cache restore/save	2025-09-30 22:03:46 +02:00
Robert Sachunsky	a3d8197930	makefile: update model URL	2025-09-30 21:50:21 +02:00
Robert Sachunsky	61b20cc83d	tests: switch from subtests to parametrize, use --isolate everywhere to free CUDA memory in between	2025-09-30 19:20:35 +02:00
Robert Sachunsky	375e0263d4	CNN-RNN OCR model: switch to 20250930 version (compatible with TF 2.12 on CPU as well)	2025-09-30 19:16:50 +02:00
Robert Sachunsky	b21051db21	ProcessPoolExecutor: shutdown during del() instead of atexit()	2025-09-30 19:16:00 +02:00
Robert Sachunsky	08c8c26028	indent extremely long lines	2025-09-30 03:52:19 +02:00
Robert Sachunsky	f857ee7b51	simplify	2025-09-30 02:26:00 +02:00
Robert Sachunsky	c0137c29ad	try to fix the failed outsourcing of utils_ocr	2025-09-30 02:23:43 +02:00
Robert Sachunsky	13f85b0d5c	Merge branch 'main' into loky-with-shm-for-175-rebuilt	2025-09-30 02:07:20 +02:00
cneud	070dafca75	remove duplicate LICENSE	2025-09-29 22:17:27 +02:00
cneud	53c1ca11fc	Update README.md	2025-09-29 22:15:17 +02:00
Robert Sachunsky	758602403e	replace loky with concurrent.futures.ProcessPoolExecutor (faster)	2025-09-29 17:48:22 +02:00
Robert Sachunsky	0366707136	get_smallest_skew: do not pass logger	2025-09-29 17:48:22 +02:00
Robert Sachunsky	b94c96fcbb	find_num_col: exit early if empty (avoiding exceptions)	2025-09-29 17:48:22 +02:00
Robert Sachunsky	04c3d7dd1b	get_smallest_skew: avoid shm if no ProcessPoolExecutor is passed	2025-09-29 17:48:22 +02:00
Robert Sachunsky	0662ece536	do_work_of_slopes*: use shm also in non-light mode(s)	2025-09-29 17:48:22 +02:00
Robert Sachunsky	31f240c3b8	do_image_rotation, do_work_of_slopes_new_curved: pass arrays via shared memory	2025-09-29 17:48:22 +02:00
Robert Sachunsky	8be2c79771	Revert "deskewing with faster multiprocessing" ... This reverts commit 5db3e9fa64.	2025-09-29 17:48:22 +02:00
Robert Sachunsky	abf5c0f845	get_smallest_skew: when shifting search range of rotation angle, compare resulting (maximum) variances instead of blindly assuming the new range is better	2025-09-29 17:48:22 +02:00
Robert Sachunsky	dc0caad512	writer: use @type='heading' instead of 'header'	2025-09-29 17:48:22 +02:00
Robert Sachunsky	f458e3ece0	writer: SeparatorRegion needs SeparatorRegionType (not ImageRegionType)	2025-09-29 17:48:22 +02:00
Robert Sachunsky	4337d62985	contours: rename 'pixel' → 'label' for clarity	2025-09-29 17:48:22 +02:00
Robert Sachunsky	5b16c2fc00	avoid pulling unused 'image_page_rotated' through functions	2025-09-29 17:48:22 +02:00
Robert Sachunsky	5bff2d156a	use box2rect instead of crop_image_inside_box when no image needed	2025-09-29 17:48:22 +02:00
Robert Sachunsky	9b5182c1c0	utils: introduce box2rect and box2slice	2025-09-29 17:48:19 +02:00
Robert Sachunsky	bca2ae3d78	get_marginals: exit early if no peaks found to avoid spurious overlap mask	2025-09-29 17:47:51 +02:00
Robert Sachunsky	235539a350	filter_contours_without_textline_inside: avoid removing from identical lists twice	2025-09-29 17:47:51 +02:00
Robert Sachunsky	11e143afee	polygon2contour: avoid overflow	2025-09-29 17:47:51 +02:00
Robert Sachunsky	7a9e8256ee	increase dilatation: textregions/lines (5→6), seplines (0→1)	2025-09-29 17:47:51 +02:00
Robert Sachunsky	f3faa29528	refactor shapely converisons into contour2polygon / polygon2contour, also handle heterogeneous geometries	2025-09-29 17:47:51 +02:00
Robert Sachunsky	0650274ffa	move dilate_*_contours to .utils.contour, rename dilate_textregions_contours_textline_version → dilate_textline_contours	2025-09-29 17:47:47 +02:00
Robert Sachunsky	a433c73628	filter_contours_area_of_image*: also ensure validity here	2025-09-29 17:46:50 +02:00
Robert Sachunsky	17bcf1af71	rename *lines_xml → *seplines for clarity	2025-09-29 17:46:50 +02:00
Robert Sachunsky	e730725da3	check_any_text_region_in_model_one_is_main_or_header_light: return original instead of resampled contours	2025-09-29 17:46:50 +02:00
Robert Sachunsky	7b51fd6624	avoid creating invalid polygons via rounding	2025-09-29 17:46:50 +02:00
Robert Sachunsky	41cc38c51a	get_textregion_contours_in_org_image_light: no back rotation, drop slope_first (always 0)	2025-09-29 17:46:48 +02:00
Robert Sachunsky	afba70c920	separate_lines/do_work_of_slopes: skip if crop is empty	2025-09-29 17:44:39 +02:00
Robert Sachunsky	66b2bce8b9	return_boxes_of_images_by_order_of_reading_new: log any exceptions	2025-09-29 17:44:36 +02:00
Robert Sachunsky	b48c41e68f	return_boxes_of_images_by_order_of_reading_new: simplify, avoid changing dtype during np.append	2025-09-29 17:42:53 +02:00
Robert Sachunsky	09ece86f0d	dilate_textregions_contours: simplify (via shapely's Polygon.buffer()), ensure validity	2025-09-29 17:42:53 +02:00
kba	9d8b858dfc	remove docs/eynollah-layout, superseded by docs/model.md and docs/usage.md	2025-09-29 16:01:29 +02:00
kba	2bcd20ebc7	reference the now-merged training tools in README.md	2025-09-29 15:21:42 +02:00
kba	ce02a3553b	🔥 remove obsolete versions of the training document	2025-09-29 15:18:21 +02:00
kba	6d379782ab	📝 align former upstream train.md with wiki train.md syntactically	2025-09-29 15:11:02 +02:00
kba	52a7c93319	add documentation on training eynollah from sbb_pixelwise_segmentation wiki	2025-09-29 15:05:05 +02:00