v2024-07-01
core c5b5580..79c61e3
Release: v2.66.1
- 📦 v2.66.1
- 📝 changelog
- GHA Docker: build docker.io first, then tag ghcr.io
- Dockerfile.cuda*: adapt to #1225 cc6ea57
- 📦 v2.66.0
- 📝 changelog
- Merge remote-tracking branch 'origin/download-file-no-absolute-urls'
- 📝 changelog
- Merge remote-tracking branch 'bertsky/param-preset-resolve-anew'
- 📝 changelog
- Merge remote-tracking branch 'bertsky/workspace-clean'
- 📝 changelog
- Merge remote-tracking branch 'origin/utilize-ps-proxy-to-ms'
- 📝 changelog
- Merge remote-tracking branch 'bertsky/add-docker-tf1'
- 📝 changelog
- Merge remote-tracking branch 'origin/fix-ci-no37'
- 📝 changelog
- Merge remote-tracking branch 'origin/ocrd-logging-debug-true'
- 📝 changelog
- cli.workspace.find --undo-download: add --keep-files
- cli.workspace.find --undo-download: only if .url exists
- 📝 changelog
- Merge branch 'master' of https://github.com/OCR-D/core
- Merge branch 'rm-coverage'
- 📝 changelog
- Merge pull request #1225 from bertsky/docker-editable
- 📝 changelog;
- Merge pull request #1227 from OCR-D/fix-ocrd-file-remove-url
- fix CD by fetching tags, too
- 📦 v2.65.0
- 📝 changelog
- ci: fix integration test
- Merge branch 'master' into test-workflow
- make network-integration-test: disable ocrd_all test
- disable ocrd all test in core
- ci: disable scrutinizer build
- Remove ocrd_all-tests from core makefile
- make ocrd all tests callable from Makefile
- merge master
- Merge branch 'test-workflow' of github.com:OCR-D/core into test-workflow merge master
- Make make assets in Dockerfile skipable
- remove duplicates
- Add a test for workflow run in ocrd_all
dinglehopper f8e3108..bc5818d
Release: v0.9.6
- ✔ GitHub Actions: Update used actions
- ⚙ pre-commit: Update hooks
- 🐛 Fix reading plain text files
- 📦 v0.9.6
- Revert "✔ Test on Python 3.13"
- 🐛 GHA: Install possible shapely build requirements (if building from source)
- Merge pull request #111 from stweil/typos
- 🐛 GHA: Install possible lxml build requirements (if building from source)
- ✔ Test on Python 3.13
- 🐛 Fix Python 3.12 support by requiring ocrd >= 2.65.0
- ⚙ pre-commit: Update hooks
- ✔ Test using empty files
- ⚙ pre-commit: Update hooks
- 🧹 tests: Move comment out of the code (bad style + weird formatting)
- ⚙ cli: Annotate types in process_dir()
- ⚙ pre-commit: Update hooks
- 🧹 Make from_text_segment()'s textequiv_level keyword-only
- 🧹 Make process_dir() keyword arguments keyword-only
- ✒ README-DEV: Releasing a new version
- 📦 v0.9.5
- ⚙ pre-commit: Add mypy dependencies
- 🐛 Check that we always get a valid ALTO namespace (satifies mypy)
- 🎨 Reformat (Black)
- ⚙ pre-commit: Update hooks
- Merge branch 'master' of https://github.com/qurator-spk/dinglehopper
- 🐛 Fix word segmentation with uniseg 0.8.0
- 🚧 GitLab CI Test: Depend on child pipeline
- 🚧 GitLab CI Test: Push after pulling
ocrd_detectron2 a8402d8..1f56273
Release: v0.1.8
- patch -f instead of -t
- suppress question if patch detectron2#5010 already applied
ocrd_fileformat ba79de9..fb769ff
Release: v0.10.0
- CI: add Dockerhub CD
- update textract2page again
- forgot to add AWS-Textract→PAGE
- update ocr-fileformat (with newer textract2page)
- Merge pull request #53 from OCR-D/osx-cp
ocrd_froc 45d5dcd..42f1ce0
Release: v0.6.1
- 📦 v0.6.1
- Merge pull request #16 from stweil/typos
- 📦 v0.6.0
- 📝 changelog
- Merge branch 'main' into pr-8-feedback
- add parameter min_score_style to filter low-confidence font classification results
- don't replace textstyle, only fontFamily if overwrite_style
- add parameter "overwrite_text" (default: false): whether to add or replace TextEquiv
- rename parameter network -> model
- rename parameter replace_textstyle -> overwrite_style
ocrd_keraslm 472197f..ea79b2a
Release: v0.4.3
- test: with initLogging
- test: allow running with published model_dta_full instead of training model_dta_test
- improve/update readme
- wrapper.rate: improve/update docstring
- wrapper.rate: use resolve_resource for model file path, split off setup()
- save training history (metrics), add cmd to print them
- train: allow passing single file for val_data, too
- fix continuing from chkpt
- update readme
- add model_dta_full.h5 ref
- chlog
- 📦 0.4.3
- generate: add option --variants (nr of nbest seqs)
- plot_context_embeddings_projection: add years
- test: allow passing directory for data, too
ocrd_kraken bdbe6fc..a6160ce
Release: v0.4.1
ocrd_neat 2f7d01c..06c8b38
Release: v0.0.1
- character normalization based on aletheia mapping
- Merge pull request #13 from qurator-spk/fix-ppn-xpath
ocrd_pagetopdf 24f77d3..7c5ab70
Release: v1.1.0
ocrd_segment 3993139..de824e9
Release: v0.1.24
- 📦 v0.1.24
- repair: add params spread / spread_level, update/improve docs
- repair: validate/repair polygons with 0 px tolerance
- project/repair join_polygons: fix rare case of adjacent rings
- from_masks: skip if no seg file
- 📦 v0.1.23
- Merge pull request #67 from OCR-D/project-parent
ocrd_tesserocr ed73d96..d23992b
Release: v0.19.1
- 📦 v0.19.1
- 📦 v0.19.0
- Update submodule tesseract
- Merge pull request #213 from OCR-D/osx-cp
- Merge pull request #212 from stweil/tesseract
- Merge pull request #211 from stweil/libcurl-dev
- model resources: fix frak2021 description
- recognize models: use _best instead of _fast
- sanitise loggers: no points, but warnings
- update changelog
- recognize: more robust polygon handling
- Merge pull request #208 from bertsky/subclass-docstrings
- Merge pull request #205 from stweil/update
- Merge pull request #203 from stweil/build
sbb_binarization b89ec49..978f425
Release: v0.1.0
- document tested Python+TF2 versions
- remove Python 3.11 from test
- Update README.md
- Update tensorflow version requirements
workflow-configuration bd149f8..eeea260
Release: 0.1.3
- ocrd-make -X: no --cleanup to prevent race for --bf
- ocrd-make: add JOBDB option
- refactor remote/parallel opts
- ocrd-make -X: fix passing on parallel options
- ocrd-make METSSERV sequentially: fix exitcode
- ocrd-make PAGEWISE: fix enumerating pages
- ocrd-make: have page-parallel preclude document-parallel (but still use parallel)
- ocrd-make: have page-parallel preclude document-parallel (to prevent oversubscription)
- add FAILRETRY variable, logic and docs
- Update README.md
- CI: less verbose ocrd_logging.conf
- CI: override prefix
- ocrd-import: replace mkfifo / bulk-add with METS Server
- ocrd-import: attempt to convert all non-XML and non-image files
- add CI/CD
- add Docker
- add tests
- ocrd-make: add PAGEWISE and METSSERV modes
- add workflow with model variable
- ocrd-make info|show: delegate other make options, too
- fix exitcode logging in sequential case
- add TIMEOUT (processor timeout) and FAILDUMMY (ocrd-dummy fallback)
- ocrd-make: add -X for computation on remote host (with OCR-D installed)
- toolrecipe: revert f3083f85 (:fire: no more support for multiple output fileGrps)
- update readme