core a07568e..dc3ffe4
Release: v2.13.2
- 📦 v2.13.2
- workspace: ensure PIL.Image.open FD gets closed
- 📦 v2.13.1
- assert_file_grp_cardinality: allow optional "msg" keyword param
dinglehopper c3709e2..7930ecd
- 🔧 dinglehopper: Add PyCharm code style config
ocrd_anybaseocr 962617d..9abe7bf
Release: v1.0.0
- 📦 v1.0.0
- Merge pull request #66 from OCR-D/file-ids-and-such
ocrd_calamari 027fcd7..210c126
Release: v0.0.7
ocrd_cis fdd7345..5bb73c3
Release: v0.0.10
- Trainining -> Training
- typos
- Fix lgtm alert
- fix 4a1d2e39d32
- set pcGtsId to METS file ID everywhere
- remove old entry point for ocropy recognition
- ocrolib: ensure closing FDs after mkstemp and PIL.Image.open
- binarize: zoom wrt. DPI to avoid inverse edges (and gain speed) at high DPI
- processors: use make_file_id and assert_file_grp_cardinality
- ocropy processors: put derived images under output fileGrp…
ocrd_keraslm be69c97..52cb4ce
Release: v0.4.0
- 📦 v0.4.0
- wrapper: set pcGtsId to METS file ID
- ocrd cli: use make_file_id and assert_file_grp_cardinality
ocrd_olena 50ecb0b..54792be
Release: v1.2.1
- 📦 1.2.1
- fix handling input fileGrps where PAGE is mixed with derived images: (#68)
ocrd_pc_segmentation db0128f..5e433e3
Release: v0.1.3
- use make_file_id and assert_file_grp_cardinality
ocrd_repair_inconsistencies ba27aba..8182063
- Merge pull request #10 from kba/file-ids-and-such
ocrd_segment 06fb5ba..7596e9c
Release: v0.1.0
- 📦 v0.1.0
- Merge remote-tracking branch 'origin/typo_conent_content'
- replace_original: update docstring
- extract_pages: no pageId for COCO, input_file.pageId for json
- revert to old 3-output-file-grp logic in extract_pages.py
- extract_regions, extract_pages: pass pageId
- Use 'make_file_id' and 'assert_file_grp_cardinality' throughout
ocrd_tesserocr 1159fbf..49e93b9
Release: v0.9.1
ocrd_typegroups_classifier 94a3630..3dddd94
Release: v0.1.1
- 📦 v0.1.0
- use make_file_id/assert_file_grp_cardinality
- fix test
- heatmap.py --help: grad_cam -> heatmap
- Merge remote-tracking branch 'seuretm/master' into HEAD
ocrd_wrap 2cfffed..6d92c02
Release: v0.1.0
- ocrd-preprocess-image: avoid passing commas from parameter value to image file IDs
- fix spaces in makefile (ht @kba)
- set pcGtsId to METS file ID everywhere
- put derived images under output fileGrp…
- ensure mkstemp fd gets closed afterwards
- ensure PIL.Image gets closed afterwards
opencv-python 83acfc9..e4893f2
Release: 42
- update README [skip ci]
- add ccache and cmake to aarch64 Dockerfile (requires rebuild)
- aarch64 needs to be rebuilt with cmake
- test aarch64 build
- drop Python 3.5 support
- use switch instead
- force qt uninstall
- unlink and reinstall
- just use the raw url directly
- no auto update, use extract
- direct link
- 5.13.2
- force Qt 5.13
- add aarch64 Dockerfile
- another typo
- fix typo in version
- Add note about manylinux2014
- OpenCV 4.4.0
- add minimum supported Python version
- fix the travis USER issue
- fix twine upload command
- fix sdist upload
- fix min numpy version for 3.6
- reduce verbosity
- patch only on linux
- fix windows builds
- update licenses
- update license info
- wrong dir
- update patch
- update patch
- update patch
- copy fonts
- set QT_QPA_FONTDIR
- add fonts to the package
- azure-cli should work now
- use py27 compatible version
- use pip
- Install azure-cli from Ubuntu repo because latest update broke batch upload
- fix patch location
- auditwheel is installed under py37
- try to whitelist libxcb.so.1
- fix patch
- change path
- qt5.15 again
- fix typo, try without xcb helpers
- add ld lib path
- qt 5.13 test
sbb_textline_detector eba5f54..a7c92c9
workflow-configuration 7edcb90..7c89037
- avoid citation notice for GNU parallel