Ocrd

189 verbose addNamed ">=0.4.23 <0.5.0" is a valid semver range for xml2js
190 silly addNameRange { name: 'xml2js', range: '>=0.4.23 <0.5.0', hasData: false }
191 silly mapToRegistry name xml2js
192 silly mapToRegistry using default registry
193 silly mapToRegistry registry https://registry.npmjs.org/
194 silly mapToRegistry uri https://registry.npmjs.org/xml2js
195 verbose addNameRange registry:https://registry.npmjs.org/xml2js not in flight; fetching
196 silly resolveWithNewModule merge1.2.1 checking installable status
197 silly cache add args [ 'merge1.2.1', null ]
198 verbose cache add spec merge1.2.1
199 silly cache add parsed spec Result {
199 silly cache add raw: 'merge1.2.1',
199 silly cache add scope: null,
199 silly cache add name: 'merge',
199 silly cache add rawSpec: '1.2.1',
199 silly cache add spec: '1.2.1',
199 silly cache add type: 'version' }
200 silly addNamed merge1.2.1
201 verbose addNamed "1.2.1" is a plain semver version for merge
202 silly mapToRegistry name merge
203 silly mapToRegistry using default registry
204 silly mapToRegistry registry https://registry.npmjs.org/
205 silly mapToRegistry uri https://registry.npmjs.org/merge
206 verbose addNameVersion registry:https://registry.npmjs.org/merge not in flight; fetching
207 verbose get https://registry.npmjs.org/xml2js not expired, no request
208 silly addNameRange number 2 { name: 'xml2js', range: '>=0.4.23 <0.5.0', hasData: true }
209 silly addNameRange versions [ 'xml2js',

> * 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
> * :package: 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](https://github.com/OCR-D/ocrd_kraken) [bdbe6fc](https://github.com/OCR-D/ocrd_kraken/commits/bdbe6fc)..[a6160ce](https://github.com/OCR-D/ocrd_kraken/commits/a6160ce)

> * update assets
> * :package: v0.4.2
> * generators: do not require input length > window size
> * train: allow passing directory for training data, too
> * add checkpointing, allow continuing from ckpt
> * suppress TF gibberish

[ocrd_tesserocr](https://github.com/OCR-D/ocrd_tesserocr) [08a020f](https://github.com/OCR-D/ocrd_tesserocr/commits/08a020f)..[ed73d96](https://github.com/OCR-D/ocrd_tesserocr/commits/ed73d96)

> * :package: v0.4.1
> * Update CHANGELOG.md
> * getLogger per method

[ocrd_kraken](https://github.com/OCR-D/ocrd_kraken) [cea90bd](https://github.com/OCR-D/ocrd_kraken/commits/cea90bd)..[0a839ad](https://github.com/OCR-D/ocrd_kraken/commits/0a839ad)

> * :package: v0.4.0
> * wrapper: set pcGtsId to METS file ID
> * ocrd cli: use make_file_id and assert_file_grp_cardinality

[ocrd_olena](https://github.com/OCR-D/ocrd_olena) [50ecb0b](https://github.com/OCR-D/ocrd_olena/commits/50ecb0b)..[54792be](https://github.com/OCR-D/ocrd_olena/commits/54792be)

> * Update requirements.txt

[ocrd_repair_inconsistencies](https://github.com/qurator-spk/ocrd_repair_inconsistencies) [6897996](https://github.com/qurator-spk/ocrd_repair_inconsistencies/commits/6897996)..[ba27aba](https://github.com/qurator-spk/ocrd_repair_inconsistencies/commits/ba27aba)

> * 🧹 examples/fix-ocr-d-gt-page.sh: rmdir after ocrd workspace remove-group is not necessary anymore
> * 🧹 examples/fix-ocr-d-gt-page.sh: rmdir after ocrd workspace remove-group is not necessary anymore
> * ✨ Support TextRegions nested in Table/GraphicRegions

[ocrd_segment](https://github.com/OCR-D/ocrd_segment) [f8a8daa](https://github.com/OCR-D/ocrd_segment/commits/f8a8daa)..[06fb5ba](https://github.com/OCR-D/ocrd_segment/commits/06fb5ba)