Merge 7b5f593709 into fc911f3734
GitHub
commit
201624959d
@ -0,0 +1 @@
|
||||
.git
|
||||
@ -1,70 +0,0 @@
|
||||
FROM ubuntu:22.04
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG OCRD_VERSION_MINIMUM="2.47.0"
|
||||
ENV LC_ALL=C.UTF-8 LANG=C.UTF-8
|
||||
ENV PIP_DEFAULT_TIMEOUT=120
|
||||
|
||||
|
||||
RUN echo "APT::Acquire::Retries \"3\";" > /etc/apt/apt.conf.d/80-retries && \
|
||||
apt-get update && \
|
||||
apt-get install -y \
|
||||
build-essential \
|
||||
curl \
|
||||
git \
|
||||
xz-utils \
|
||||
pkg-config \
|
||||
# For add-apt-repository:
|
||||
software-properties-common \
|
||||
# XML utils
|
||||
libxml2-utils \
|
||||
xmlstarlet \
|
||||
# OCR-D uses ImageMagick for pixel density estimation
|
||||
imagemagick \
|
||||
# pyenv builds
|
||||
# TODO: builder container?
|
||||
libz-dev \
|
||||
libssl-dev \
|
||||
libbz2-dev \
|
||||
liblzma-dev \
|
||||
libncurses-dev \
|
||||
libffi-dev \
|
||||
libreadline-dev \
|
||||
libsqlite3-dev \
|
||||
libmagic-dev \
|
||||
&& \
|
||||
apt-get clean && \
|
||||
rm -rf /var/lib/apt/lists/*
|
||||
|
||||
|
||||
# Set up OCR-D logging
|
||||
RUN echo "setOverrideLogLevel(os.getenv('LOG_LEVEL', 'INFO'))" >/etc/ocrd_logging.py
|
||||
|
||||
|
||||
# Install pyenv
|
||||
# TODO: do not run as root
|
||||
# TODO: does just saying "3.7" work as intended?
|
||||
ENV HOME=/root
|
||||
ENV PYENV_ROOT=/usr/local/share/pyenv
|
||||
ENV PATH=$PYENV_ROOT/shims:$PYENV_ROOT/bin:$PATH
|
||||
RUN \
|
||||
git clone --depth=1 https://github.com/yyuu/pyenv.git $PYENV_ROOT && \
|
||||
pyenv install 3.7 && \
|
||||
pyenv global 3.7 && \
|
||||
pyenv rehash && \
|
||||
pip install -U pip wheel && \
|
||||
pip install setuptools
|
||||
|
||||
# Install pip installable-stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
"ocrd >= ${OCRD_VERSION_MINIMUM}"
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
WORKDIR /data
|
||||
|
||||
# Default command
|
||||
CMD ['ocrd']
|
||||
@ -1,70 +0,0 @@
|
||||
FROM nvidia/cuda:12.1.0-cudnn8-runtime-ubuntu22.04
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG OCRD_VERSION_MINIMUM="2.47.0"
|
||||
ENV LC_ALL=C.UTF-8 LANG=C.UTF-8
|
||||
ENV PIP_DEFAULT_TIMEOUT=120
|
||||
|
||||
|
||||
RUN echo "APT::Acquire::Retries \"3\";" > /etc/apt/apt.conf.d/80-retries && \
|
||||
apt-get update && \
|
||||
apt-get install -y \
|
||||
build-essential \
|
||||
curl \
|
||||
git \
|
||||
xz-utils \
|
||||
pkg-config \
|
||||
# For add-apt-repository:
|
||||
software-properties-common \
|
||||
# XML utils
|
||||
libxml2-utils \
|
||||
xmlstarlet \
|
||||
# OCR-D uses ImageMagick for pixel density estimation
|
||||
imagemagick \
|
||||
# pyenv builds
|
||||
# TODO: builder container?
|
||||
libz-dev \
|
||||
libssl-dev \
|
||||
libbz2-dev \
|
||||
liblzma-dev \
|
||||
libncurses-dev \
|
||||
libffi-dev \
|
||||
libreadline-dev \
|
||||
libsqlite3-dev \
|
||||
libmagic-dev \
|
||||
&& \
|
||||
apt-get clean && \
|
||||
rm -rf /var/lib/apt/lists/*
|
||||
|
||||
|
||||
# Set up OCR-D logging
|
||||
RUN echo "setOverrideLogLevel(os.getenv('LOG_LEVEL', 'INFO'))" >/etc/ocrd_logging.py
|
||||
|
||||
|
||||
# Install pyenv
|
||||
# TODO: do not run as root
|
||||
# TODO: does just saying "3.7" work as intended?
|
||||
ENV HOME=/root
|
||||
ENV PYENV_ROOT=/usr/local/share/pyenv
|
||||
ENV PATH=$PYENV_ROOT/shims:$PYENV_ROOT/bin:$PATH
|
||||
RUN \
|
||||
git clone --depth=1 https://github.com/yyuu/pyenv.git $PYENV_ROOT && \
|
||||
pyenv install 3.7 && \
|
||||
pyenv global 3.7 && \
|
||||
pyenv rehash && \
|
||||
pip install -U pip wheel && \
|
||||
pip install setuptools
|
||||
|
||||
# Install pip installable-stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
"ocrd >= ${OCRD_VERSION_MINIMUM}"
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
WORKDIR /data
|
||||
|
||||
# Default command
|
||||
CMD ['ocrd']
|
||||
@ -1,18 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core:$GIT_COMMIT
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG DINGLEHOPPER_VERSION="0.9.2"
|
||||
|
||||
|
||||
# Build pip installable stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
"dinglehopper == $DINGLEHOPPER_VERSION"
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
# Default command
|
||||
CMD ["ocrd-dinglehopper"]
|
||||
@ -1,18 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core-cuda12.1:$GIT_COMMIT
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG EYNOLLAH_VERSION="0.3.0"
|
||||
|
||||
|
||||
# Build pip installable stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
"eynollah == ${EYNOLLAH_VERSION}"
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
# Default command
|
||||
CMD ["ocrd-eynollah-segment"]
|
||||
@ -1,19 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core-cuda12.1:$GIT_COMMIT
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG OCRD_ANYBASEOCR_VERSION="1.8.2"
|
||||
|
||||
|
||||
|
||||
# Build pip installable stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
"ocrd_anybaseocr == ${OCRD_ANYBASEOCR_VERSION}"
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
# Default command
|
||||
CMD ["ocrd-anybaseocr-crop"]
|
||||
@ -1,20 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core-cuda12.1:$GIT_COMMIT
|
||||
|
||||
|
||||
# XXX https://github.com/OCR-D/core/issues/642
|
||||
#ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG OCRD_CALAMARI_VERSION="1.0.5"
|
||||
|
||||
# Build pip installable stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
"ocrd_calamari == $OCRD_CALAMARI_VERSION"
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
# Default command
|
||||
CMD ["ocrd-calamari-recognize"]
|
||||
@ -1,22 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core-cuda12.1:$GIT_COMMIT
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
|
||||
|
||||
# Build pip installable stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
# Resolve conflicts early:
|
||||
'tensorflow-gpu == 1.15.*' \
|
||||
'calamari-ocr == 0.3.5' \
|
||||
# Now the real stuff:
|
||||
'ocrd_calamari == 0.0.7'
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
# Default command
|
||||
RUN ln -s ocrd-calamari-recognize /usr/local/bin/ocrd-calamari-recognize03
|
||||
CMD ["ocrd-calamari-recognize"]
|
||||
@ -1,19 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core:$GIT_COMMIT
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG OCRD_CIS_VERSION="0.1.5"
|
||||
|
||||
|
||||
# Build pip installable stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
# Now the real stuff:
|
||||
"https://github.com/cisocrgroup/ocrd_cis/archive/v${OCRD_CIS_VERSION}.tar.gz"
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
# Default command
|
||||
CMD ["ocrd-cis-ocropy-segment"]
|
||||
@ -1,24 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core:$GIT_COMMIT
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG OCRD_FILEFORMAT_VERSION="0.5.0"
|
||||
|
||||
|
||||
RUN apt-get update && \
|
||||
apt-get install -y \
|
||||
git \
|
||||
openjdk-11-jdk-headless \
|
||||
wget \
|
||||
unzip \
|
||||
&& \
|
||||
apt-get clean && rm -rf /var/lib/apt/lists/*
|
||||
RUN git clone --depth 1 --branch v${OCRD_FILEFORMAT_VERSION} https://github.com/OCR-D/ocrd_fileformat.git && \
|
||||
cd ocrd_fileformat/ && \
|
||||
git submodule update --init && \
|
||||
make install-fileformat install PREFIX=/usr/local && \
|
||||
cd .. && rm -rf ocrd_fileformat/
|
||||
|
||||
|
||||
# Default command
|
||||
CMD ['ocrd-fileformat-transform']
|
||||
@ -1,39 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core:$GIT_COMMIT
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG OCRD_OLENA_VERSION="1.3.0"
|
||||
|
||||
|
||||
# Build ocrd_olena
|
||||
RUN apt-get update && \
|
||||
apt-get install -y \
|
||||
imagemagick \
|
||||
&& \
|
||||
apt-get clean && rm -rf /var/lib/apt/lists/*
|
||||
|
||||
# Install olena from .deb
|
||||
RUN curl -sSL --retry 3 -O https://qurator-data.de/~mike.gerber/olena_2.1.0+ocrd-git+2-ubuntu22.04/olena-bin_2.1.0+ocrd-git+2_amd64.deb && \
|
||||
dpkg -i --force-depends olena-bin_2.1.0+ocrd-git+2_amd64.deb && \
|
||||
rm -f olena-bin_2.1.0+ocrd-git+2_amd64.deb && \
|
||||
apt-get update && \
|
||||
apt-get -f install -y && \
|
||||
apt-get clean && rm -rf /var/lib/apt/lists/* && \
|
||||
if ! scribo-cli sauvola --help >/dev/null 2>&1; then echo "Olena/scribo is not installed correctly" >&2; exit 1; fi
|
||||
RUN curl -sSL --retry 3 -o ocrd_olena.tar.gz https://github.com/OCR-D/ocrd_olena/archive/v${OCRD_OLENA_VERSION}.tar.gz && \
|
||||
mkdir ocrd_olena && \
|
||||
tar xvz -C ocrd_olena --strip-components=1 -f ocrd_olena.tar.gz && \
|
||||
cd ocrd_olena && \
|
||||
sed -i 's/^install: deps/install:/' Makefile && \
|
||||
${PIP_INSTALL} ocrd && \
|
||||
apt install xmlstarlet && \
|
||||
make install PREFIX=/usr/local && \
|
||||
cd .. && rm -rf ocrd_olena ocrd_olena.tar.gz
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
# Default command
|
||||
CMD ['ocrd-olena-binarize']
|
||||
@ -1,19 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core:$GIT_COMMIT
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG OCRD_SEGMENT_VERSION="0.1.22"
|
||||
|
||||
|
||||
# Build pip installable stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
# Now the real stuff:
|
||||
"ocrd-segment == ${OCRD_SEGMENT_VERSION}"
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
# Default command
|
||||
CMD ["ocrd-segment-extract-regions"]
|
||||
@ -1,31 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core:$GIT_COMMIT
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG TESSDATA_BEST_VERSION="4.0.0"
|
||||
ARG OCRD_TESSEROCR_VERSION="0.17.0"
|
||||
ENV TESSDATA_PREFIX /usr/local/share/tessdata
|
||||
|
||||
|
||||
# Install Leptonica and Tesseract.
|
||||
# TODO: Review if alex-p's repo is still necessary on jammy (jammy has 4.1.1,
|
||||
# alex-p has 4.1.3, but not for jammy.)
|
||||
# RUN add-apt-repository ppa:alex-p/tesseract-ocr && \
|
||||
RUN apt-get update && \
|
||||
apt-get install -y \
|
||||
tesseract-ocr \
|
||||
libtesseract-dev \
|
||||
&& \
|
||||
apt-get clean && rm -rf /var/lib/apt/lists/*
|
||||
|
||||
|
||||
# Build pip installable stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
"ocrd_tesserocr == ${OCRD_TESSEROCR_VERSION}"
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
# Default command
|
||||
CMD ["ocrd-tesserocr-recognize"]
|
||||
@ -1,18 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core:$GIT_COMMIT
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG OCRD_WRAP_VERSION="0.1.7"
|
||||
|
||||
|
||||
# Build pip installable stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
"ocrd_wrap == ${OCRD_WRAP_VERSION}"
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
# Default command
|
||||
CMD ["ocrd-preprocess-image"]
|
||||
@ -1,19 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core-cuda12.1:$GIT_COMMIT
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG SBB_BINARIZATION_VERSION="0.1.0"
|
||||
|
||||
|
||||
# Build pip installable stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
# Now the real stuff:
|
||||
"sbb_binarization == $SBB_BINARIZATION_VERSION"
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
# Default command
|
||||
CMD ["ocrd-sbb-binarize"]
|
||||
@ -1,20 +0,0 @@
|
||||
ARG GIT_COMMIT="latest"
|
||||
FROM quratorspk/ocrd-galley-core-cuda12.1:$GIT_COMMIT
|
||||
|
||||
ARG PIP_INSTALL="pip install --no-cache-dir"
|
||||
ARG SBB_TEXTLINE_DETECTOR_COMMIT="c4df3d6"
|
||||
|
||||
|
||||
# Build pip installable stuff
|
||||
RUN ${PIP_INSTALL} \
|
||||
# https://github.com/qurator-spk/sbb_textline_detection/issues/50
|
||||
"h5py < 3" \
|
||||
https://github.com/qurator-spk/sbb_textline_detector/archive/$SBB_TEXTLINE_DETECTOR_COMMIT.tar.gz
|
||||
|
||||
|
||||
# Check pip dependencies
|
||||
RUN pip check
|
||||
|
||||
|
||||
# Default command
|
||||
CMD ["ocrd-sbb-textline-detector"]
|
||||
@ -1,33 +0,0 @@
|
||||
#!/bin/bash
|
||||
set -e
|
||||
|
||||
self=`realpath $0`
|
||||
self_dir=`dirname "$self"`
|
||||
|
||||
|
||||
|
||||
if [ -n "$1" ]; then
|
||||
sub_images=""
|
||||
for arg in "$@"; do
|
||||
arg_sub_image=`echo "$arg" | sed 's/Dockerfile-//'`
|
||||
NL=$'\n'
|
||||
sub_images+="$NL$arg_sub_image"
|
||||
done
|
||||
else
|
||||
sub_images=`ls -1 Dockerfile-core* | sed 's/Dockerfile-//'`
|
||||
sub_images="$sub_images `ls -1 Dockerfile-* | sed 's/Dockerfile-//'`"
|
||||
fi
|
||||
echo "Building:"
|
||||
echo "$sub_images"
|
||||
echo
|
||||
|
||||
|
||||
# Update base images if we build a core image
|
||||
if echo "$sub_images" | grep -q core; then
|
||||
docker pull ubuntu:22.04
|
||||
docker pull nvidia/cuda:12.1.0-cudnn8-runtime-ubuntu22.04
|
||||
fi
|
||||
|
||||
for sub_image in $sub_images; do
|
||||
docker build --cache-from=quratorspk/ocrd-galley-$sub_image -t quratorspk/ocrd-galley-$sub_image -f Dockerfile-$sub_image .
|
||||
done
|
||||
@ -0,0 +1,4 @@
|
||||
#!/bin/sh
|
||||
set -ex
|
||||
|
||||
docker build . -t ocrd_trocr:latest -f Dockerfile-ocrd_trocr
|
||||
@ -1 +0,0 @@
|
||||
test-core.sh
|
||||
@ -1,14 +0,0 @@
|
||||
#!/bin/sh
|
||||
set -ex
|
||||
|
||||
test_id=`basename $0`
|
||||
cd `mktemp -d /tmp/$test_id-XXXXX`
|
||||
|
||||
# Prepare test workspace
|
||||
wget https://qurator-data.de/examples/actevedef_718448162.first-page+binarization+segmentation.zip
|
||||
unzip actevedef_718448162.first-page+binarization+segmentation.zip
|
||||
cd actevedef_718448162.first-page+binarization+segmentation
|
||||
|
||||
# Run tests
|
||||
ocrd-cis-ocropy-segment -I OCR-D-IMG-BIN -O TEST-CIS-OCRPY-SEGMENT
|
||||
# TODO -recognize
|
||||
@ -0,0 +1,24 @@
|
||||
#!/bin/sh
|
||||
set -e
|
||||
|
||||
count_ok=0
|
||||
count_failed=0
|
||||
|
||||
for test in tests/*.sh; do
|
||||
echo "== $test"
|
||||
$test && result=$? || result=$?
|
||||
|
||||
if [[ $result = 0 ]]; then
|
||||
echo "✔"
|
||||
count_ok=$((count_ok+1))
|
||||
else
|
||||
echo "❌"
|
||||
count_failed=$((count_failed+1))
|
||||
fi
|
||||
echo
|
||||
done
|
||||
|
||||
echo "$count_ok ok, $count_failed failed"
|
||||
if [[ $count_failed -gt 0 ]]; then
|
||||
exit 1
|
||||
fi
|
||||
@ -0,0 +1,16 @@
|
||||
#!/bin/sh
|
||||
set -ex
|
||||
|
||||
test_id=`basename $0`
|
||||
cd `mktemp -d /tmp/$test_id-XXXXX`
|
||||
|
||||
# Prepare processors
|
||||
|
||||
# Prepare test workspace
|
||||
wget https://qurator-data.de/examples/actevedef_718448162.first-page.zip
|
||||
unzip actevedef_718448162.first-page.zip
|
||||
cd actevedef_718448162.first-page
|
||||
|
||||
# Run tests
|
||||
ocrd-anybaseocr-binarize -I OCR-D-IMG -O OCR-D-BIN -P operation_level page -P threshold 0.3
|
||||
ocrd-anybaseocr-deskew -I OCR-D-BIN -O OCR-D-DESKEW -P maxskew 5.0 -P skewsteps 20 -P operation_level page
|
||||
@ -0,0 +1,30 @@
|
||||
#!/bin/sh
|
||||
set -ex
|
||||
|
||||
test_id=`basename $0`
|
||||
cd `mktemp -d /tmp/$test_id-XXXXX`
|
||||
|
||||
OCRD_CIS_OCROPY_MODEL=fraktur.pyrnn.gz
|
||||
|
||||
# Prepare processors
|
||||
ocrd resmgr download ocrd-cis-ocropy-recognize $OCRD_CIS_OCROPY_MODEL
|
||||
|
||||
# Prepare test workspace
|
||||
wget https://qurator-data.de/examples/actevedef_718448162.first-page.zip
|
||||
unzip actevedef_718448162.first-page.zip
|
||||
cd actevedef_718448162.first-page
|
||||
|
||||
# XXX ocrd-cis-ocropy-segment wasn't happy with the binarized input (no
|
||||
# "binarized" AlternativeImage?!), so we do it here again
|
||||
ocrd-skimage-binarize -I OCR-D-IMG -O OCR-D-IMG-BIN
|
||||
|
||||
# Run tests
|
||||
ocrd-cis-ocropy-segment \
|
||||
-I OCR-D-IMG-BIN -O TEST-CIS-OCROPY-SEG-LINE \
|
||||
-P level-of-operation page
|
||||
test "$(grep TextLine TEST-CIS-OCROPY-SEG-LINE/*.xml | wc -l)" -gt 50
|
||||
|
||||
ocrd-cis-ocropy-recognize \
|
||||
-I TEST-CIS-OCROPY-SEG-LINE -O TEST-CIS-OCROPY-OCR \
|
||||
-P model $OCRD_CIS_OCROPY_MODEL
|
||||
test "$(grep Unicode TEST-CIS-OCROPY-OCR/*.xml | wc -l)" -gt 50
|
||||
@ -0,0 +1,45 @@
|
||||
processor_images = {
|
||||
"ocrd": "ocrd/all:maximum",
|
||||
"ocrd-olena-binarize": "ocrd/all:maximum",
|
||||
"ocrd-sbb-binarize": "ocrd/all:maximum",
|
||||
"ocrd-sbb-textline-detector": "ocrd/all:maximum",
|
||||
"ocrd-calamari-recognize": "ocrd/all:maximum",
|
||||
"ocrd-calamari-recognize03": "ocrd/all:maximum",
|
||||
"ocrd-tesserocr-segment-region": "ocrd/all:maximum",
|
||||
"ocrd-tesserocr-segment-line": "ocrd/all:maximum",
|
||||
"ocrd-tesserocr-recognize": "ocrd/all:maximum",
|
||||
"ocrd-dinglehopper": "ocrd/all:maximum",
|
||||
"ocrd-cis-ocropy-clip": "ocrd/all:maximum",
|
||||
"ocrd-cis-ocropy-resegment": "ocrd/all:maximum",
|
||||
"ocrd-cis-ocropy-segment": "ocrd/all:maximum",
|
||||
"ocrd-cis-ocropy-deskew": "ocrd/all:maximum",
|
||||
"ocrd-cis-ocropy-denoise": "ocrd/all:maximum",
|
||||
"ocrd-cis-ocropy-binarize": "ocrd/all:maximum",
|
||||
"ocrd-cis-ocropy-dewarp": "ocrd/all:maximum",
|
||||
"ocrd-cis-ocropy-recognize": "ocrd/all:maximum",
|
||||
"ocrd-fileformat-transform": "ocrd/all:maximum",
|
||||
"ocrd-segment-extract-pages": "ocrd/all:maximum",
|
||||
"ocrd-segment-extract-regions": "ocrd/all:maximum",
|
||||
"ocrd-segment-extract-lines": "ocrd/all:maximum",
|
||||
"ocrd-segment-from-masks": "ocrd/all:maximum",
|
||||
"ocrd-segment-from-coco": "ocrd/all:maximum",
|
||||
"ocrd-segment-repair": "ocrd/all:maximum",
|
||||
"ocrd-segment-evaluate": "ocrd/all:maximum",
|
||||
"ocrd-preprocess-image": "ocrd/all:maximum",
|
||||
"ocrd-skimage-normalize": "ocrd/all:maximum",
|
||||
"ocrd-skimage-denoise-raw": "ocrd/all:maximum",
|
||||
"ocrd-skimage-binarize": "ocrd/all:maximum",
|
||||
"ocrd-skimage-denoise": "ocrd/all:maximum",
|
||||
"ocrd-eynollah-segment": "ocrd/all:maximum",
|
||||
"ocrd-anybaseocr-binarize": "ocrd/all:maximum",
|
||||
"ocrd-anybaseocr-crop": "ocrd/all:maximum",
|
||||
"ocrd-anybaseocr-deskew": "ocrd/all:maximum",
|
||||
|
||||
# non OCR-D CLI
|
||||
"ocr-transform": "ocrd/all:maximum",
|
||||
"dinglehopper": "ocrd/all:maximum",
|
||||
"dinglehopper-extract": "ocrd/all:maximum",
|
||||
|
||||
# specialized images
|
||||
"ocrd-trocr-recognize": "ocrd_trocr",
|
||||
}
|
||||
@ -1,40 +0,0 @@
|
||||
sub_images = {
|
||||
"ocrd": "core",
|
||||
"ocrd-olena-binarize": "ocrd_olena",
|
||||
"ocrd-sbb-binarize": "sbb_binarization",
|
||||
"ocrd-sbb-textline-detector": "sbb_textline_detector",
|
||||
"ocrd-calamari-recognize": "ocrd_calamari",
|
||||
"ocrd-calamari-recognize03": "ocrd_calamari03",
|
||||
"ocrd-tesserocr-segment-region": "ocrd_tesserocr",
|
||||
"ocrd-tesserocr-segment-line": "ocrd_tesserocr",
|
||||
"ocrd-tesserocr-recognize": "ocrd_tesserocr",
|
||||
"ocrd-dinglehopper": "dinglehopper",
|
||||
"ocrd-cis-ocropy-clip": "ocrd_cis",
|
||||
"ocrd-cis-ocropy-resegment": "ocrd_cis",
|
||||
"ocrd-cis-ocropy-segment": "ocrd_cis",
|
||||
"ocrd-cis-ocropy-deskew": "ocrd_cis",
|
||||
"ocrd-cis-ocropy-denoise": "ocrd_cis",
|
||||
"ocrd-cis-ocropy-binarize": "ocrd_cis",
|
||||
"ocrd-cis-ocropy-dewarp": "ocrd_cis",
|
||||
"ocrd-cis-ocropy-recognize": "ocrd_cis",
|
||||
"ocrd-fileformat-transform": "ocrd_fileformat",
|
||||
"ocrd-segment-extract-pages": "ocrd_segment",
|
||||
"ocrd-segment-extract-regions": "ocrd_segment",
|
||||
"ocrd-segment-extract-lines": "ocrd_segment",
|
||||
"ocrd-segment-from-masks": "ocrd_segment",
|
||||
"ocrd-segment-from-coco": "ocrd_segment",
|
||||
"ocrd-segment-repair": "ocrd_segment",
|
||||
"ocrd-segment-evaluate": "ocrd_segment",
|
||||
"ocrd-preprocess-image": "ocrd_wrap",
|
||||
"ocrd-skimage-normalize": "ocrd_wrap",
|
||||
"ocrd-skimage-denoise-raw": "ocrd_wrap",
|
||||
"ocrd-skimage-binarize": "ocrd_wrap",
|
||||
"ocrd-skimage-denoise": "ocrd_wrap",
|
||||
"ocrd-eynollah-segment": "eynollah",
|
||||
"ocrd-anybaseocr-crop": "ocrd_anybaseocr",
|
||||
"ocrd-anybaseocr-deskew": "ocrd_anybaseocr",
|
||||
"ocrd-trocr-recognize": "ocrd_trocr",
|
||||
|
||||
# non OCR-D CLI
|
||||
"ocr-transform": "ocrd_fileformat",
|
||||
}
|
||||
Loading…
Reference in New Issue