#!/bin/bash
set -e

self=`realpath $0`
self_dir=`dirname "$self"`

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/*.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/models.tar.gz' 'sbb_binarization'
  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'
}
. $self_dir/qurator_data_lib.sh
handle_data

if [ -n "$1" ]; then
  sub_images=`echo "$@" | sed 's/Dockerfile-//'`
else
  sub_images=`ls -1 Dockerfile-core* | sed 's/Dockerfile-//'`
  sub_images="$sub_images `ls -1 Dockerfile-* | sed 's/Dockerfile-//'`"
fi
for sub_image in $sub_images; do
  docker build --cache-from=my_ocrd_workflow-$sub_image -t my_ocrd_workflow-$sub_image -f Dockerfile-$sub_image .
done