You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
62 lines
2.4 KiB
Bash
62 lines
2.4 KiB
Bash
#!/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
|
|
|
|
|
|
|
|
DATA_SUBDIR=data
|
|
get_from_annex() {
|
|
annex_get 'calamari-models/GT4HistOCR/2019-07-22T15_49+0200/*.ckpt*'
|
|
annex_get 'calamari-models/GT4HistOCR/2019-12-11T11_10+0100/*.ckpt*'
|
|
annex_get 'calamari-models/GT4HistOCR/2019-12-18T17_24+0100*/*.ckpt*'
|
|
annex_get 'mirror/github.com/Calamari-OCR/calamari_models/gt4histocr/*.ckpt*'
|
|
annex_get 'tesseract-models/GT4HistOCR/*.traineddata'
|
|
annex_get 'textline_detection/*.h5'
|
|
annex_get 'mirror/github.com/tesseract-ocr/tessdata_best/archive/4.0.0-repacked.tar.gz'
|
|
annex_get 'sbb_binarization/2021-03-09/*.h5'
|
|
annex_get 'eynollah/*.h5'
|
|
}
|
|
get_from_web() {
|
|
download_to 'https://qurator-data.de/calamari-models/GT4HistOCR/model.tar.xz' 'calamari-models/GT4HistOCR/2019-07-22T15_49+0200'
|
|
download_to 'https://qurator-data.de/calamari-models/GT4HistOCR/2019-12-11T11_10+0100/model.tar.xz' 'calamari-models/GT4HistOCR/2019-12-11T11_10+0100'
|
|
download_to 'https://qurator-data.de/tesseract-models/GT4HistOCR/models.tar' 'tesseract-models/GT4HistOCR'
|
|
download_to 'https://qurator-data.de/sbb_textline_detector/models.tar.gz' 'textline_detection'
|
|
download_to --strip-components 1 'https://qurator-data.de/sbb_binarization/2021-03-09/models.tar.gz' 'sbb_binarization/2021-03-09'
|
|
download_to --no-unpack 'https://qurator-data.de/mirror/github.com/tesseract-ocr/tessdata_best/archive/4.0.0-repacked.tar.gz' 'mirror/github.com/tesseract-ocr/tessdata_best/archive/4.0.0-repacked.tar.gz'
|
|
download_to --strip-components 1 'https://qurator-data.de/eynollah/models_eynollah.tar.gz' 'eynollah'
|
|
}
|
|
. $self_dir/qurator_data_lib.sh
|
|
#handle_data
|
|
|
|
|
|
|
|
# Update base images if we build a core image
|
|
if echo "$sub_images" | grep -q core; then
|
|
docker pull ubuntu:18.04
|
|
docker pull nvidia/cuda:10.0-cudnn7-runtime-ubuntu18.04
|
|
docker pull nvidia/cuda:10.1-cudnn7-runtime-ubuntu18.04
|
|
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
|