Versioned according to Semantic Versioning.
Changed:
- Update Dockerfile, #96
1.5.0 - 2023-10-20
Added:
- Support for
--mets-server-url
, #95
1.4.0 - 2023-10-12
Changed:
- 🔥 relicense as GPLv2, #90
- Update olena to include OCR-D/olena#8
- Adapt to OCR-D/core#1079
- Require
ocrd >= 2.55.2
Fixed:
- Use
OcrdFile.local_filename
consistently after OCR-D/core#1079 merged, #88
Removed:
- We no longer use Travis for CI, #65, #91
1.3.0 - 2022-05-19
Changed:
- Reduce
winsize
for non-Sauvola algorithms, #84, #85 - Accumulate files to be added to the mets and use
ocrd workspace bulk-add
, #77, #85 - Properly support METS not-named
mets.xml
1.2.6 - 2021-12-08
Changed:
- only mkdir output_file_grp if needed
- simplify file iterator by delegating to bashlib input-files
1.2.5 - 2021-11-11
Changed:
- on
make install
, check for existingscribo-cli
in PATH - differentiate
make install
vsmake install-tools
1.2.4 - 2021-09-22
Fixed:
- Docker: skip autoremove (which kills needed rtlibs)
1.2.3 - 2020-10-13
Fixed:
- pass correct
k
value forniblack
- always build against ImageMagick 6 (not 7)
Changed:
- also crop AlternativeImage to page border if not already
- remove temporary images from cropping afterwards
- use updated
ocrd workspace
CLI for tests ocrd-olena-binarize:
1.2.2 - 2020-09-30
Fixed:
- typo in 1.2.1 error
1.2.1 - 2020-08-21
Fixed:
- images may be mixed with PAGE-XML in
input_file_grp
, OCR-D/core#505 OCR-D/spec#164 make install
: update pip and ocrd- CI setup
1.2.0 - 2020-07-31
Changed:
- use
sauvola-ms-split
as defaultimpl
(method) - default to
win-size=0
which now equals odd DPI value of image - set
pcGtsId
in PAGE-XML, ht @mikegerber, #63 - images are written to the same
output_file_grp
as the PAGE-XML, OCR-D/core#505
Added:
- parameter
dpi
to override image meta-data DPI
1.1.10 - 2020-06-17
Changed:
- Use
ocrd log
for logging - Implement
--overwrite
1.1.9 - 2020-06-12
Fixed:
- Check whether ImageMagick/GraphicsMagick and Boost are installed before compilation, #53
1.1.8 - 2020-06-12
Fixed:
- Use explicit
-format
inidentify
call, #57
1.1.7 - 2020-05-27
Fixed:
- Use install-time Python binary when needed
Changed:
- Ignore
Page/PrintSpace
(as opposed toPage/Border
)
1.1.6 - 2020-04-22
Fixed:
- Quote all arguments in
scribo-cli
to allow file names with whitespace - Pass on different
k
values at each scale forsauvola-ms*
Changed:
- Skip
AlternativeImage/@comments=binarized
in PAGE input - Update
repo/olena
to upstream master - Update
repo/assets
to upstream master
1.1.5 - 2020-03-19
Fixed:
- ignore/pass
file://
prefixes in absolute pathnames - disable Qt when building Olena
1.1.4 - 2020-03-06
Changed:
- Update
repo/olena
to upstream master - Update
repo/assets
to upstream master
Fixed:
- ignore/pass
file://
prefixes in absolute pathnames - compensate for different
k
defaults for singh/niblackimpl
- pass on correct
k
values for sauvola-ms-{fg,split}impl
1.1.3 - 2019-02-25
Changed:
- Update
repo/olena
to most recent upstream make build-olena
configured not to build documentation
- Use bash from env instead of fixed shebang
- Ensure minimum bash version
Changed:
- Update olena upstream repo to include gitignore rules
- Ensure ocrd-tool.json is reinstalled if changed, #26
Changed:
- Remove the patches and use patched OCR-D fork as git submodule
- Deployment to DockerHub as
ocrd/all:minimum
,ocrd/all:medium
,ocrd/all:maximum
Fixed:
- add namespace prefix to MetadataItem/Label(s), if any
- use proper logging format, delegate to bashlib constants
- pass input parameters to algorithms
- respect -g option (page_id)
- when AlternativeImage exists already, append after last
- date: Use explicit formatting for iso8601 compliant date w/o TZ (kba/fix-created-date)
- much improved olena build process
Changed
- negate and convert output images in scribo itself (without extra ImageMagick CLI)
- improved docker support, 2 docker builds: pure scribo and OCR-D binarization wrapper
- overwrite image file IDs if necessary (bertsky/add-imagefile-force)
First release