Releases: OCR-D/ocrd_all
v2022-06-03
Fixed:
- Work around issues with git's new
safe.directory
setting interfering withsudo git
, #321, #323, #324
Changed:
- Disable
cor-asv-ann
andocrd_keraslm
which require old tensorflow-gpu unless it is available, #325 - Dockerfile: use
$VIRTUAL_ENV
instead of hard-coded paths or$PREFIX
Added:
- Start building on Ubuntu 22.04 in addition to 20.04 and 18.04, #318
core ecdb840..60697e4
Release: v2.35.0
- 📦 v2.35.0
- 📝 changelog
- update bagit-profile and ocrd_tool.schema.yml from spec
- update spec
- 📦 v2.34.0
- 📝 changelog
- Merge remote-tracking branch 'bertsky/patch-12'
- 📝 changelog
- Merge branch 'resmgr-do-not-parse-dates'
- 📝 changelog
- Merge branch 'ocrd-log-stdin'
- 📝 changelog
- cleanup
- change tests according to previous commit
- change return type to list of tuples
- remove printing error message
- remove single newline from last commit
- resource_manager list_available: fix #854
- test_ocrd_mets: test new pageId regex
- Merge branch 'add-pageid-regex' of https://github.com/bertsky/core into add-pageid-regex
- 📝 changelog
- Merge branch 'add-pageid-regex'
- 📝 changelog
- adapt safe_filename without dots
- safe_filename: allow underscore, avoid dot
- Merge pull request #849 from tdoan2010/master
ocrd_detectron2 855fc3b..833319a
Release: v0.1.1
- table mode: relax image requirements
- fix typo
- add param operation_level, add level table
ocrd_keraslm 2101c8f..787341d
Release: v0.4.1
- Update test-python3 to large resource class in config.yml
ocrd_olena fce9933..dbef534
Release: v1.3.0
- 📦 v1.3.0
- install libtool in Docker
- Merge branch 'circleci-update-resource-class' into fix-default-winsize
- re-use METS basename for output
- use bulk-add for METS results
- use 1/6th DPI for non-Sauvola win-size default
ocrd_segment e9acc81..309b7f9
Release: v0.1.21
- forgot src file
- new processor replace-text
- repair/project: fix MST bridge polygons
- repair/project: use minimum spanning tree for shortest path search when joining polygons
- 📦 0.1.21
- Merge remote-tracking branch 'bertsky/extract-regions-coco'
- repair/project: join polygons directly instead of alphashape
- fixed parameter in save_image_file(), updated classes-parameter to type array
- Update ocrd_segment/ocrd-tool.json
- Update ocrd_segment/extract_regions.py
- added classes parameter to extract-region-processor; enables extraction of specific regions e.g. TableRegion
- tool.json: restrict numerical ranges
- extract-lines: reduce default min-* values
opencv-python 5de8d66..f2f924c
Release: 64
- Merge pull request #666 from asenyaev/asen/update_all_dependencies
workflow-configuration e50dee4..bb2a4e2
- ocrd-import: fix mkdir context
- ocrd-import: avoid . in IDs
- ocrd-import: ensure grayscale after PDF import
- ocrd-import: forgot to set errexit in subshell too
v2022-05-04
Changed:
make clean
: remove sub-venvs and all semaphores, #315
Removed:
ocrd_pc_segmentation
has been disabled since it's not used and not supported upstream, #317
core f9c2b6c..ecdb840
Release: v2.33.0
- 📦 v2.33.0
- 📝 changelog
- Merge remote-tracking branch 'stweil/master'
- 📝 changelog
- Merge remote-tracking branch 'mexthecat/issues_838'
- 📝 changelog
- Merge remote-tracking branch 'origin/remove-assets-server'
- 📝 changelog
- Merge remote-tracking branch 'bertsky/workspace-altimg-retrieve-existing'
- 📝 changelog
- workspace bulk_add: fix automatic file_id
- Simplify Makefile
- Remove version restrictions for pip (no longer needed)
- Always use apt --no-install-recommend
- Merge branch 'master' of https://github.com/OCR-D/core
- 📝 changelog
- workspace.remove_file_group: pass recursive
eynollah ae1990a..34a0617
Release: v0.0.11
- depend on tensorflow instead of tensorflow-gpu (#76)
- Merge pull request #74 from cneud/main
- require model command line option (fix #59) (#73)
- Update README.md
- Merge pull request #71 from mikegerber/fix/downgrade-patch-dim-log
- issue #55 resolved
- updating model directory
- Merge pull request #70 from qurator-spk/cneud-readme
ocrd_cis ca15800..2cdfa79 (rewind)
Release: v0.1.5
join_polygons (alpha shape): make more robust
remove Calamari dependency (not used, only CLI callout)
segment: for more robust bg separator detection, combine criteria of gradient maximum and percentile
segment: rewrite separator detection…
resegment: if method=lineest, then annotate baselines, too
resegment (ccomps/baseline): propagate/spread twice to catch diacritics/punctuation, too
re/segment: alpha shape: smaller α to avoid holes
segment: fix lines2regions non-continguous partitions
segment: annotate baselines, too
ocrd_detectron2 5bb90b1..855fc3b
Release: v0.1.1
- update requirements
- Fix sed command for MacOS
- make deps: ignore CUDA 11.2
ocrd_doxa 351eb0c..a95f8e7
Release: v0.0.2
- 📦 v0.0.2
- add missing param 'dpi'
ocrd_fileformat 9451a99..38f4edf
Release: v0.5.0
- 📦 v0.5.0
- update README
- run pages in parallel (2 concurrent jobs)
- make install: differentiate between install-tools (script+json) and deps (just OCRD)
- use bulk-add and update to core 2.30
- simplify and update to core 2.29
- update ocr-fileformat
- allow all supported transforms for from-to
ocrd_keraslm b11b51d..2101c8f
Release: v0.4.1
ocrd_segment f6c7f1c..e9acc81
Release: v0.1.20
- 📦 0.1.20
- extract_pages: fix interference between plotting and JSON extraction in b024b80aa
- alpha shape: make even more robust yet
- alpha shape: make more robust
- 📦 0.1.19
- repair (
sanitize
): rewrite…- repair (
plausibilize
): use true alpha shape instead of convex hull- project: reduce default alpha for alphashape
- project: add level-of-operation=table
- repair: add option 'simplify', generalize 'ensure_consistent'
- Shapely 1.8 compatibility
- 📦 0.1.18
- extract-lines: fix regression in 8bfeb9f9bf
opencv-python 0eb6c12..5de8d66
Release: 64
- Merge pull request #644 from nsait-linaro/windows-arm64
- Merge pull request #652 from asenyaev/asen/pinned_dependencies
- Merge pull request #637 from orbisvicis/master
sbb_binarization 0877199..e4c1eb2
Release: v0.0.9
- Merge pull request #37 from bertsky/patch-2
- importing types from click is enough
- Update cli.py
- fixing #30. Directory of model(s) is needed not model file name.
- Merge pull request #36 from qurator-spk/0.0.9
- improve usage instructions
- Merge pull request #35 from cneud/upgrade_to_tf2_with_tf1.compat_session
- Revert "Merge pull request #34 from qurator-spk/upgrade_to_tf2_with_compat_mode"
- Merge pull request #34 from qurator-spk/upgrade_to_tf2_with_compat_mode
- Merge pull request #31 from bertsky/factor-setup
workflow-configuration bef631c..e50dee4
- Modifiy cp statement (fix build on MacOS)
- ocrd-import: allow running in parallel
- ocrd-import: use bulk-add for speedup
v2022-03-30
core 6d359e7..f9c2b6c
Release: v2.32.0
- 📦 v2.32.0
- Merge remote-tracking branch 'bertsky/fix-recropped' into test-2.32
- Merge remote-tracking branch 'origin/identify-pil-fallback'
- ocrd zip bag: in-place not required
dinglehopper edc24cd..1febea8
- Merge pull request #66 from stweil/master
ocrd_anybaseocr 5bc6de9..94e5037
Release: v1.8.2
ocrd_detectron2 6daa4f6..5bb90b1
Release: v0.1.1
- make deps: ignore CUDA 10.0 and 11.0
ocrd_fileformat fbd0457..9451a99
Release: v0.4.1
- 📦 v0.4.1
- 📝 changelog
- Merge pull request #38 from bertsky/patch-1
ocrd_pagetopdf 7eb2e4e..6155605
Release: v1.0.0
- Merge pull request #19 from bertsky/fix-imagegrp
ocrd_segment 00f80bd..f6c7f1c
Release: v0.1.17
- extract-lines: fix typo
- fix regression in 7e1f189c..d41993df
- ocrd_tool.resources: better descriptions
- ocrd-tool: list known resources
- recognize: improve messages
opencv-python 65f12d9..0eb6c12
Release: 64
- Merge pull request #642 from asenyaev/asen/skip_gapi
- Merge pull request #641 from asenyaev/asen/revert_pre_downloaded_files
v2022-03-25
cor-asv-ann bea828a..1058657
Release: v0.1.12
- Ignore Python build artifacts
ocrd_anybaseocr 1a6734a..736e302
Release: v1.8.0
ocrd_cis b856f5b..2cdfa79
Release: v0.1.5
- revert e673544 (crashes OpenCV)
- fix Workspace.save_image_file args
- update to shapely 1.8
ocrd_olahd_client 128c3b8..9809d0c
- Merge pull request #4 from stweil/master
opencv-python 8880543..65f12d9
Release: 64
- Merge pull request #639 from asenyaev/asen/pre_downloaded_files
- Merge pull request #638 from asenyaev/asen/pylint_test
sbb_textline_detector 325389a..a78f40a
ocrd_tesserocr bfe71b7..4876c84
Release: v0.13.6
- ocrd_tool.resources: better descriptions
v2022-03-20
- Update kraken to include the work-in-progress segmentation, #288
- GitHub actions for
make all
cor-asv-ann f19a9b3..bea828a
Release: v0.1.12
- show warning if metric=Levenshtein-fast and confusion>0
core 1a7b81c..6d359e7
Release: v2.31.0
- 📦 v2.31.0
- Merge remote-tracking branch 'origin/anybaseocr-layout-analysis'
- Merge remote-tracking branch 'origin/wsval-missing-args'
- Merge remote-tracking branch 'bertsky/patch-5'
- OcrdMets.add_file: fix fileGrp regex exception msg
- Merge pull request #808 from ulb-sachsen-anhalt/doc/upd/install-deps
- Merge pull request #779 from ulb-sachsen-anhalt/feat/rfct/pytest4
dinglehopper 195354c..edc24cd
- ✔️ DroneCI: Build on Python 3.6 → 3.10
- 👷🏾♂️ Remove str() on Path objects
- ✔️ CircleCI: Stop testing using Python 3.5
- 🐛 Bump rapidfuzz dep to >= 2.0.5 (Fixes gh-65)
eynollah f0ac0bb..10f1ace
Release: v0.0.11
ocrd_anybaseocr e930c93..1a6734a
Release: v1.7.0
- 📦 v1.7.0
- README: explain resmgr download and pip install
- layout-analysis: fix parent fornew chapter/section
- update CHANGELOG
- fix/update README
- 📦 1.7.0
- test_dewarp: also when on CPU, use CROP as input
- dewarping: improve model loading and rename params…
- dewarping: fix image post-processing…
- dewarping: fix prepare_data…
- test_dewarp: mets.find_files is a generator now
- tests: fix initLogging
- tests: fix relative import
- add test for dewarping
- update requirements
- ocrd-tool (tiseg/layout-analysis): use SavedFormat instead of HDF5 by default
- layout-analysis: improve…
- Makefile: fix test dependencies; update to resmgr cwd semantics
tesseract 8b5571f8..c2a3efe2
Release: 5.1.0
- Create new release 5.1.0
- Remove unneeded include statements
- Fix typo in descriptions of thresholding parameters
- Fix linker flags for MSYS2 clang64 builds
- [cmake] Indent.
- Merge pull request #3752 from OgreTransporter/fix-issue-3751
- README: ALTO is supported since version 4.1.0
- Handle image and line regions in output formats ALTO, hOCR and text
- Merge pull request #3753 from stweil/clang-warnings
- Update codeql-analysis.yml
- Update Makefile.am
- Remove unused code.
- Simplify. Move related function from separate file.
- Merge pull request #3745 from egorpugin/main
- Fix warnings.
- Remove unused fwd.
- Remove unused ctors in macros.
- Fix list type.
- Remove in-class TESS_API.
- Ignore more build files.
- Fix build.
- Fix warnings.
- Remove unneeded dtor.
- Fix warnings.
- Shorten macros.
- Remove a local function from class TableRecognizer
- Remove some local functions from class ImageFind
- Remove unused functions ImageFind::ComposeRGB and ImageFind::ClipToByte
- Remove unused function ImageFind::ComputeRectangleColors
- Remove unused resolution parameters
- Remove some local functions from class LineFinder
- Fix comment
ocrd_calamari 76b34c5..ecb0d67
Release: v1.0.3
- ✔️ Handle PAGE XML namespace dynamically and without hardcoding
- ✔️ CircleCI: Use a test matrix to test on different Python versions
- ✔️ Move installing deps to Makefile
- ✔️ CircleCI: Test using the python:3.6.12 image
- Merge pull request #74 from bertsky/fix-tests-again
- git push
- Merge pull request #70 from bertsky/patch-2
ocrd_cis 529f7f5..b856f5b
Release: v0.1.5
- resegment: join_polygons: make equidistant points relative to estimated scale
- resegment: join_polygons: allow non-contiguous input, too
- resegment: fix method=baseline
- Merge pull request #77 from bertsky/fix-resegment
ocrd_segment ff1f477..00f80bd
Release: v0.1.17
- 📦 0.1.17
- project: use alpha-shape instead of convex hull
- forgot actual source file (86e9474edd)
- update README
- 📦 v0.1.16
- project: new processor
- repair/plausibilize: fix scope of result list (avoids deleting multiple times)
- 📦 v0.1.15
- Merge branch 'excel_gt_writer'
- plausibilize: both analyse & apply iff enabled
- 📦 v0.1.14
- repair: fix line split rule
- Merge branch 'master' of ssh://github.com/OCR-D/ocrd_segment
- repair: generalize plausibilization…
ocrd_tesserocr ac27465..bfe71b7
Release: v0.13.6
- ocrd-tool: list known resources
opencv-python 209d32e..8880543
Release: 64
- Merge pull request #632 from asenyaev/asen/fix_current_builds
workflow-configuration f5faf29..bef631c
- ocrd-import: delegate to ocrd logging
- ocrd-import: fix numerical retval
- ocrd-import: fix backing out of errors (again)
- ocrd-import: show error if not directory
- ocrd-import: fix histexpand
- ocrd-page-transform: rely on recent bashlib resource changes
v2022-02-03
New projects:
core 66cd13f9..836eb05e
Release: v2.30.0
- 📦 v2.30.0
- 📝 changelog
- Merge branch 'show-resource-subdir'
- 📝 changelog
- Merge branch 'bulk-add-from-stdin'
- 📝 changelog
- Merge branch 'simplify-bashlib'
- 📝 changelog
- Merge remote-tracking branch 'bertsky/patch-2'
- Merge remote-tracking branch 'bertsky/patch-1'
- 📝 changelog
- Merge branch 'master' into decorator-working-dir
- Merge remote-tracking branch 'origin/master' into decorator-working-dir
- Resolve.resolve_mets_arguments: simplify resolution logic, test all cases
- ocrd workspace CLI: Rely on the --mets*/--directory logic in Resolver
- make resolve_mets_arguments a method of Resolver, reintroduce --mets-basename DeprecactionWarning
- Merge branch 'master' into decorator-working-dir
- use resolve_mets_arguments in processor decorator, fix #693
- resolve_mets_arguments: handle --mets == mets.xml, use pathlib
- resolve_mets_arguments: handle mets&directory&!mets_basename
- extract --mets/--mets-basenme/--directory logic to utils.resolve_mets_arguments
tesseract 5d5cd77e..8b5571f8
Release: 5.0.1
- Merge pull request #3742 from stweil/robustness
- Fix old TODO (STATS::rangemax_)
- Additional fix for CI action autotools-macos for MacPorts
- Fix CI action autotools-macos for MacPorts
- Require cmake 3.10 or newer (#3724)
- Add new parameter curl_timeout for curl_easy_setop
- Add missing include file for std::max, std::min
- ci: Setting PKG_CONFIG_PATH is no longer needed for autoconf builds on MacOS
- autoconf: Set PKG_CONFIG_PATH required for Homebrew (macOS)
- Merge pull request #3716 from ffontaine/main
- Create new release 5.0.1
- Update documentation for release 5.0.1
- Merge pull request #3712 from stweil/cmake
- Fix build on non-x86 non-ARM with clang
- Merge pull request #3705 from stweil/cmake
- Update required packages for CI
- Merge pull request #3704 from stweil/cmake-fixes
- Use cmake policy CMP0074 only with version 3.12 or newer
- Update submodule googletest to release v1.11.0
- Remove unused include statement
- Add function ERRCODE::error with only 2 parameters
- Fix some compiler warnings (implicit float to double conversion)
- Remove unused forward declaration
- Merge pull request #3699 from stweil/bugfix
- Merge pull request #3697 from stweil/opt
- Catch nullptr in STATS::pile_count (fix isse #3694)
- Update generator for lookup tables to use TFloat instead of double
- Fix some clang compiler warnings
- Merge pull request #3690 from stweil/SPDX
- Add SPDX-License-Identifier to public include files
- Fix compiler warning [-Wsign-compare]
- Allow printing of bitfield with variadic templates
- Fix compiler warnings caused by empty statements
- Replace tprintf comment in API header
- Cmake install (#3685)
- clean up condition to detect MSCV
- fix lstm.cpp build with clang
- cmake: fix bcf7bbb6e23fcd7c081
- cmake: do not try to even configure training tools if pkg-config is not present.
- cmake: reformat with cmake-format and check with cmake-lint
- cmake: do not try to configure training tools without PkgConfig
- cmake: remove unused code
- CMake: Also check x86_64 and i686
- cmake: fix fb74c7eac1292fca6221d66ed4428d3a031e99dd
- Merge pull request #3681 from stweil/curl
- Update README.md
- Update unittest-disablelegacy.yml
- Update vcpkg-4.1.1.yml
- Fix broken msys2 build with gcc 11
- Update unittest-disablelegacy.yml
- Merge pull request #3669 from pkubaj/patch-1
- README: Latest stable version is 5.0.0
ocrd_cis 6f8a612..529f7f5
Release: v0.1.5
- compute_hlines/compute_separators_morph: early length filt by opening must be softer than final criterion
- compute_hlines: fix h/v kernel size
- ocropy.lines2regions: fall back to topological partitioning
- ocropy.lines2regions: improve splitting by separators
- ocropy.segment: improve within-region hmerge rule
- ocropy.lines2regions: fix textline assignment around existing regions
- ocrolib.morph: opt out of np.unique counts
cor-asv-ann 55e58f0..f19a9b3
Release: v0.1.12
- align: fix 'alignment' if only 1 seq survived
- align: fix conf for too long/short subseqs
- align: fix conf avg for all-empty lines
- update subrepos
- add processor align (purely rule-based line-level n-ary multi-OCR voting)
- add standalone CLI proc, update documentation
- {gen,vectorize}_lines: track filenames and add unsupervised (no GT)
- {seq2seq,scripts}.evaluate: add charmap option as well
- lib.alignment: fix static confusion count
- Merge remote-tracking branch 'origin/master' into HEAD
- scripts.compare: show confusion as well
- scripts.compare: ensure report contains line content (even for file lists)
dinglehopper 2497876..195354c
- Merge branch 'feat/compare-line-texts'
- ⬆️ dinglehopper: Require rapidfuzz >= 1.9.1
- 💩 ocrd-galley: Work around OCR-D/core#730
- Merge branch 'fix/ci-py310'
- Revert "⬆️ CircleCI: Add Python 3.10"
- ⬆️ CircleCI: Add Python 3.10
- ⬆️ CircleCI: Switch to the new cimg/python image
- Merge branch 'rapidfuzz'
eynollah e769f62..f0ac0bb
Release: v0.0.11
- 📦 v0.0.11
- Merge pull request #61 from qurator-spk/resource-directory
ocrd_fileformat 72139bc..fbd0457
Release: v0.4.0
ocrd_segment 8f180b6..ff1f477
Release: v0.1.13
- evaluate: aggregate per-page metrics
- evaluate: avoid clipped image and transparency
- evaluate: if detection not found, use gt for logging pageID
opencv-python f07201c..209d32e
Release: 62
- Merge pull request #617 from asenyaev/asen/update_dockerfile
- Merge pull request #598 from asenyaev/asen/4.5.5_release
- Merge pull request #596 from asenyaev/asen/config_version
- Merge pull request #595 from asenyaev/asen/python_limited_api
- Merge pull request #590 from cclauss/patch-4
- Merge pull request #580 from jreese/patch-1
- Merge pull request #579 from asenyaev/asen/fix_pylin_and_pycharm
- Merge pull request #577 from asenyaev/asen/fix_build_from_source
- Merge pull request #581 from asenyaev/asen/upgrade_auditwheel
- Merge pull request #582 from kianmeng/fix-typos
- Merge pull request #578 from asenyaev/asen/update_submodules_for_release
- Merge pull request #561 from andruekonst/patch-1
- Merge pull request #565 from cclauss/patch-2
- Merge pull request #563 from asenyaev/asen/python_loader_support
- Merge pull request #558 from cclauss/patch-2
- Merge pull request #556 from asenyaev/asen/upgrade_to_python_3_10
- Merge pull request #547 from asenyaev/asen/lapack_issue
- Merge pull request #555 from asenyaev/asen/fix_macos_m1
workflow-configuration 9296644..f5faf29
- ocrd-page-transform: add param mimetype (default to PAGE-XML)
- ocrd-page-transform: add option to pass xslt-params
- add more PAGE XSLT, add option for param mechanism
- add 2 more PAGE XSLs
- add ocrd-page-transform as processor
- ocrd-import: detect+allow ALTO-XML
- ocrd-import: fix filenames with white-space
- ocrd-import: add --skip option
- fix regression ...
v2021-12-13
core c4ca4fab..66cd13f9
Release: v2.29.0
- test_exif: remove platform-dependent skipif mark
- Merge remote-tracking branch 'upstream/master' into feat/rfct/pytest
- [test][rfct] switch to pytest style
- [app][dep] add uniform formatter
- [test][rfct] ocrd-exif param + skipif
- [app][dep] rather opencv-headless
- [test][rfct] exchange unittest with pytest
- 📦 v2.29.0
- Update CHANGELOG.md
- Merge branch 'restrict-fileGrp-USE' into release-2.29.0
- 📝 changelog
- Merge branch 'pr/744'
- Merge branch 'pr/753'
- Merge pull request #755 from ulb-sachsen-anhalt/feat/rfct/review-dep-opencv
ocrd_olena faf3170..fce9933
Release: v1.2.6
ocrd_segment bdc6771..8f180b6
Release: v0.1.13
- 📦 0.1.13
- evaluate: fix multi-matching and over/undersegmentation metrics, add pixel scores
- evaluate: add pixel-wise measures IoU / IoGT / IoDT…
- evaluate: also get over/undersegmentation and false DT/GT…
- evaluate: refactor and expose standalone CLI
- change pycocotools dependency to fork
- 📦 0.1.12
- evaluate: cleanup unrestricted maxDets
- evaluate: fix pycocotools match pair extraction, work around annotation id > 0 assumption
v2021-11-30
Changed:
- Re-enable
TESSDATA_PREFIX
, #261
core 9069a658..c4ca4fab
Release: v2.28.0
- 📦 v2.28.0
- Merge branch 'pr/748'
- Merge branch 'pr/739'
- Merge branch 'pr/743'
- Merge remote-tracking branch 'bertsky/workspace-download-wait' > * 📝 changelog
- metsHdr/agent/note: add only necessary namespace
- run_processor: add mets:notes with in/output fileGrp and params via mets:agent
- allow adding mets:note to mets:agent
- Merge pull request #737 from OCR-D/update-readme-docs
- 📦 v2.27.0
- Merge remote-tracking branch 'bertsky/workspace-remove-allow-empty-regex'
- Merge branch 'resmgr-undefined-resdict'
- Merge branch 'fix-workspace-remove-group-fnfe'
- Merge branch 'rename-group-id'
- remove undeclared requirement on six, fix #726
- bump spec to 3.14.0
- 📦 v2.26.1
- Fixed tesseract resource URL for resource 'configs'
ocrd_segment 34c72c0..bdc6771
Release: v0.1.11
- evaluate: basic mAP/AR functionality (based on pycocotools)
tesseract 3313bb79..924b746e
Release: 5.0.0
- Create new release 5.0.0
- Merge pull request #3664 from stweil/classify
- Merge pull request #3663 from stweil/clang7
- autobuild: Fix autogen.sh (reduce build time)
- Remove unused GenericVector::contains_index, UnicityTable::contains_id
- Remove member function GenericVector::contains
- Remove member function UnicityTable::contains
workflow-configuration 8d3dc73..af76e39
- PAGE transformation scripts: improve error handling vs options
- PAGE transformation scripts: cancel if input is not PAGE
- add PAGE transformation script for namespace update
- add PAGE transformation scripts for text projection
- in multi-input case, allow repeating the same input fileGrps
- update README
- ocrd-import: avoid additional . in file IDs/names
- add PAGE transformation script, add comments
- remove old repair/view rules
- add+wrap various PAGE transformation scripts
- ocrd-import: fix backing out of errors
- ocrd-import: allow upper case suffixes
v2021-10-12
cor-asv-ann 1b8c120..55e58f0
Release: v0.1.12
- lib.keras_train: adapt to Keras 2.3.*
core 271dcdf..9069a65
Release: v2.26.0
- 📝 changelog, describe --no-copy-files fix of #713 properly
- 📦 v2.26.0
- 📝 changelog
- Merge branch 'resmgr-tesseract-main'
- 📝 changelog
- Merge branch 'log-to-stderr'
- 📝 changelog
- Merge remote-tracking branch 'bertsky/ocrd-utils-scale'
- 📝 changelog
- Merge remote-tracking branch 'bertsky/fix-workspace-merge-filegrp'
- 📝 changelog
- makefile: reintroduce tests -> assets dependency
- Adapt tests to not use OcrdFile directly anymore
- Workspace.remove_file: Will always be a single file (because ID is unique)
- OcrdFile.str: don't raise if not connected to OcrdMets
- workspace: raise FNFE in download_file, remove mets:file only after recursion
- Workspace.download_url: Use a dummy METS instead of direct OcrdFile instantiation
- utils.get_local_filename: handle Kitodo/Goobi file URLs gracefully
- workspace validator: catch unresolveable files when doing multipage validation
- OcrdFile should never be instantiated w/o reference to METS
- eliminate OcrdFile() from OcrdWorkspace, #324
- tests/data/sample-features: PNG urls missed .png suffix in mets.xml
- tests/data/sample-features: missed foo0.png reference in mets.xml
- Merge pull request #706 from OCR-D/more-apidoc-fixes
- Merge pull request #705 from OCR-D/fix-exif-test
eynollah d784202..e769f62
Release: v0.0.10
- 📦 v0.0.10
- Merge remote-tracking branch 'vahidrezanezhad/main' into main
- 📦 v0.0.9
- Merge remote-tracking branch 'vahidrezanezhad/main' into main
- 📦 v0.0.8
- set pcGtsId before adding file to mets
ocrd_fileformat 86763d1..72139bc
Release: v0.3.0
- update ocr-fileformat
- 📦 v0.3.0
ocrd_olena 9db846c..faf3170
Release: v1.2.4
- Update CHANGELOG.md
- Dockerfile: skip apt autoremove
- auto-push to Dockerhub
ocrd_tesserocr c0e1440..b37a710
Release: v0.13.6
- 📦 0.13.6
- Merge pull request #180 from bertsky/no-music-by-default
- segment/recognize: add param find_staves (for pageseg_apply_music_mask)
- segment/recognize: do not find tables when already looking for cells
- 📦 0.13.5
opencv-python 86c3d2a..f07201c
Release: 56
- Merge pull request #545 from Skylion007/patch-1
- Merge pull request #515 from asmorkalov/as/docs_update_actions
workflow-configuration f1a858a..8d3dc73
- skip workspace backups when searching for METS files
v2021-07-22
tesseract c676d5bc..88d4028a
Release: 5.0.0-alpha-20210401
- Enable pragma for SIMD also when _OPENMP is defined
- Use SIMD instructions for DotProductNative
- Fix deadlock in lstmtraing. (#3488)
- Fix LSTMTrainerTest.BidiTest
- Fix IntSimdMatrixTest.AVX2
- Fix check for NEON on 32 bit ARM
- Catch another FP division by 0 (fixes issue #3483)
- Use CMAKE_SYSTEM_PROCESSOR to check for SIMD support options (#3484)
- Catch FP division by 0 (fixes issue #3483)
- Fix CMake Linux build (#3478)
- Bugfix OpenMP with Visual Studio (#3475)
- Fix vector resize with init for all elements (issue #3473) (#3474)
- Add --reset_learning_rate option to lstmtraining (#3470)
- Fix missing reset of best_error_history_ in LSTMTrainer::InitIterations() (#3469)
- CI: Replace g++-8 by g++-11 for MacOS
- Show layer specified learning rates with combine_tessdata -l (#3468)
- Fix permanently disabled optimizations in CMake (#3467)
- Null check for GetSourceYResolution (#3457)
- Merge pull request #3439 from amitdo/remove-var
- Merge pull request #3438 from amitdo/pango
- Merge pull request #3437 from amitdo/sauvola
- Remove scripts for training
- Fix some typos (found by codespell)
- Refactor ObjectCache::DeleteUnusedObjects with reverse iterator
- Add test for TessBaseAPI::ClearPersistentCache to unit tests
- Merge pull request #3431 from nagadomi/fix_segv
- travis: Use libleptonica-dev from Ubuntu focal instead of local build
- Simplify list of required packages for CI builds
- Fix comment in code
- Remove unneeded dependency autoconf-archive from CI builds
- Update submodule abseil to tagged release 20210324.1
- Update README.md
- Remove an arbitrary limit for the image size
- Fix compiler warning in binarization code (uninitialized local variable)
- Optimize code for binarization
- Fix compiler warning
- Fix autoconf build for latest MacOS (Intel and M1)
- Add conditional compilation with GRAPHICS_DISABLED
- Fix warning from autoconf
- Merge pull request #3422 from nickjwhite/lstmevalmoreinfo
- Postfix for #3418.
- Merge pull request #3418 from amitdo/thresholder
- Allow explicit casts of Image to Pix**.
- [misc] Fix variable type. Fixes warning.
- Remove unneeded header.
- Revert d6cdc52. Fixes #3412.
- Merge pull request #3410 from ulb-sachsen-anhalt/fix/fix-typo-icu
- Merge pull request #3408 from GerHobbelt/patch-2
- Merge pull request #3406 from Lucas-C/add-print-fonts-table-arg
- Update cmake-win64.yml
- Improved fix for issue #3405
- Revert "Fix double free. Closes #3405."
- Misc.
- Fix double free. Closes #3405.
- Remove unneeded ctor.
- Make FontSet just a vector.
- Properly init font set.
- Replace FontSet struct with vector. It may be improved further (remove pointer?).
- Remove unused headers.
- Fix issue #3404 (empty page regression)
- Remove two assertions which are triggered by fuzzing
- Fix some compiler warnings with GRAPHICS_DISABLED
- Fix memory leak (OSS-Fuzz issue 33220)
- Merge pull request #3401 from adaptech-cz/fix-wrong-param
- Fix old TODO for struct EDGEPT
- Add SVSemaphore destructor to avoid system objects leaks
- Replace pointer by value for ScrollView mutex
- Fix NULL pointer access (issue #3394)
- Merge pull request #3396 from zdenop/v5.clang.cmake.win
- Fix OSS-Fuzz issue 32142 (container-overflow write)
- Clean, format and optimize code in edgblob.cpp / edgblob.h
- Fix headers.
- Remove unneeded header.
- [edgblob] Replace unique ptr with vector. Fix possible index issues.
- Fix another use-after-free (issue #3394)
- Merge pull request #3395 from amitdo/amitdo-patch-1
- Fix use-after-free (issue #3394)
- Merge pull request #3393 from eighttails/fix_zero_division
- Fix some typos (most found by codespell)
- automake: Add build rule for fuzzer-api-512x256
- Update URLs for Google groups
- Fix memory leak (OSS-Fuzz issue #32246)
- Fix broken build for fuzzer
- Merge pull request #3389 from amitdo/authors
- lstmrecognizer.cpp: Call OutputStats() only when 'invert' is true (#3387)
- Change message from 'Found SSE' to 'Found SSE4.1'
- Add const attribute to some functions (API change)
- Simplify class REJ by replacing two std::bitset<16> by one std::bitset<32>
- Make more functions in class REJ inline
- Remove bits16.h and BITS16 data type
- Format public API files with modified rules for clang-format
- clang-format: Clean formatting rules
- Modernize and optimize list_rec a little bit
- Modernize code
- Remove mfdefs.cpp from CMakeLists.txt and Makefile.am
- Refactor microfeatures a bit.
- Replace microfeatures from oldlist to std::forward_list.
- Correct template args.
- Fix new / delete class mismatch
- Add missing TESS_API.
- Reorder list templates.
- Misc.
- [elist2] Convert macros to template. Remove source file macro ELIST2IZE.
- [elist] Rework macro into template and small macro. Move common iterator template into 'list_iterator.h'.
- Misc.
- Remove unused methods.
- [elist] Remove unused macros QUOTE_IT.
- [clist] Remove unused methods.
- Rework big clist macro into template and small macro. Remove unused macros QUOTE_IT and CLISTIZE (source file macro).
- Merge pull request #3382 from Shreeshrii/windows
- RAII: TessBaseAPI::GetIterator()
- RAII: TessBaseAPI::AnalyseLayout()
- Clean code for hOCR renderer
- Merge pull request #3380 from stweil/gridsearch
- Remove unnecessary pointer hasher for uset.
- Fix CI and update package list before installation
- Merge pull request #3376 from stweil/master
- Merge pull request #3375 from amitdo/viewer
- Create new pre-release 5.0.0-alpha-20210401
- Add Image::operator&=().
- Add Image::operator|=().
- Remove arg from Image::copy().
- Init variable.
- Add Image::isZero().
- Add Image::clone().
- Add Image::copy().
- Add new files ccstruct/image.cpp and ccstruct/image.h to Makefile
- Fix name conflict between class and function named Image
- [misc] Update header guard.
- Revert 4fa05b9147b957344bf94df1e0fb781acc2c84a6. Make a note.
- Basic usage of new Image class. Only pixDestroy is wrapped at the moment.
- Initial tesseract Image wrapper.
- Remove unused ifdef.
- Partially fix issue #3374
- Fix some compiler warnings
- Improve some structs further (fixes several CID issues)
- lstm: Include missing header
- Fix compilation for Tensorflow code
- Fix some compiler warnings
- Merge pull request #3368 from stweil/master
- Merge pull request #3367 from stweil/master
- Merge pull request #3366 from stweil/master
- Add more list tests
- Misc.
- Remove default locale setting in debug config. Any locale errors must be fixed separately (if any).
- Merge pull request #3365 from Shreeshrii/fix
- Replace some snprintf by std::to_string and modernize more code
- Merge pull request #3364 from stweil/master
- Merge pull request #3363 from Shreeshrii/split
- Don't convert for loop after '#pragma omp parallel' with clang-tidy
- Remove TessBaseAPI::SetThresholder (API change)
- Modernize code for OTSU and reduce public API further
- Merge pull request #3358 from stweil/master
- Merge pull request #3356 from Shreeshrii/ciupdate
- Merge pull request #3355 from eighttails/output_training_command_line
- Update submodule abseil to tagged release 20200923.3
- Fix some compiler warnings
- Merge pull request #3354 from stweil/master
- Merge pull request #3351 from stweil/master
- Use more smart pointers.
- Merge pull request #3350 from stweil/master
- Merge pull request #3349 from stweil/master
- Remove unused function UnicharIdArrayUtils::find_in
- Disable unused WriteMemoryCallback if libcurl is not used
- Use smart pointers.
- Misc.
- Fix double free caused by commit f33e80e (fixes issue #3348)
- Rename GenericVector::get
- Rename UnicityTable::get, UnicityTable::get_mutable
- Replace std::array by std::vector to avoid stack overflow
- Merge pull request #3345 from stweil/vector
- Merge pull request #3344 from stweil/vector
- Merge pull request #3342 from stweil/master
- Fix new alert reported by LGTM
- Merge pull request #3341 from stweil/vector
- Merge pull request #3340 from stweil/vector
- Partially revert "Replace more GenericVector by std::vector for src/ccutil"
- Partially revert "Replace remaining GenericVector by std::vector for src/training"
- Merge pull request #3338 from stweil/vector
- Merge pull request #3337 from stweil/vector
- Fix new alerts reported by LGTM
- Add missing reference & for loop iterator
- Make only locally used functions row_y_order and row_spacing_order static
- Merge pull request #3336 from stweil/vector
- Merge pull request #3335 from stweil/genericvector
- Merge pull request #3330 from Sintun/master
- Fix build system macro. Fixes #3334.
- Fix alerts reported by LGTM
- Fix removal of duplicates in StructuredTable::FindLinedStructure
- Merge pull request #3332 from stweil/vector