|
|
|
|
#!/bin/bash
|
|
|
|
|
|
|
|
|
|
set -e # Abort on error
|
|
|
|
|
set -x
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
remove_filegrp() {
|
|
|
|
|
filegrp_use=$1
|
|
|
|
|
mets=$2
|
|
|
|
|
|
|
|
|
|
xmlstarlet ed --inplace \
|
|
|
|
|
-N mets=http://www.loc.gov/METS/ \
|
|
|
|
|
-d "//mets:fileGrp[@USE='$filegrp_use']" $mets
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
do_fontident() {
|
|
|
|
|
ocrd_typegroups_classifier_parameters='
|
|
|
|
|
{
|
|
|
|
|
"network": "/home/mike/devel/OCR-D/monorepo/ocrd_typegroups_classifier/ocrd_typegroups_classifier/models/classifier.tgc",
|
|
|
|
|
"stride":143
|
|
|
|
|
}'
|
|
|
|
|
remove_filegrp OCR-D-FONTIDENT mets.xml
|
|
|
|
|
ocrd-typegroups-classifier -l DEBUG \
|
|
|
|
|
-m mets.xml -I OCR-D-IMG -O OCR-D-FONTIDENT \
|
|
|
|
|
-p <(echo $ocrd_typegroups_classifier_parameters)
|
|
|
|
|
# XXX does DEFAULT have any meaning? /buerger_gedichte_1778.ocrd does not have
|
|
|
|
|
# any DEFAULT, yet -I DEFAULT seems to work for ocrd-typegroups-classifier
|
|
|
|
|
#ocrd workspace validate mets.xml
|
|
|
|
|
# XXX <error>Unspecified USE category 'FONTIDENT' in fileGrp 'OCR-D-FONTIDENT'</error>
|
|
|
|
|
# XXX <error>File 'OCR-D-FONTIDENT_OCR-D-IMG_0002' does not manifest any physical page.</error>
|
|
|
|
|
# XXX <notice>Won't download remote image <OCR-D-IMG/OCR-D-IMG_0002></notice>
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
do_linesegmentation() {
|
|
|
|
|
remove_filegrp OCR-D-SEG-LINE mets.xml
|
|
|
|
|
ocrd-ocropy-segment -l DEBUG \
|
|
|
|
|
-m mets.xml -I OCR-D-IMG -O OCR-D-SEG-LINE
|
|
|
|
|
#ocrd workspace validate mets.xml
|
|
|
|
|
|
|
|
|
|
# XXX This leaves copies of the images at the top level of the workspace, because it "downloads" the "remote" files.
|
|
|
|
|
# Clean it up.
|
|
|
|
|
find . -maxdepth 1 -name "OCR-D-IMG*" -type f -exec rm -v {} \;
|
|
|
|
|
|
|
|
|
|
# XXX ocrd-tesserocr-segment-line does not seem to produce any line segmentation
|
|
|
|
|
# XXX mv {ocrd-ocropy-segment,-line}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
do_ocr() {
|
|
|
|
|
ocrd_tesserocr_recognize_parameters='{ "model": "eng" }' # TODO mods:language + fontident → model
|
|
|
|
|
remove_filegrp OCR-D-OCR-TESS mets.xml
|
|
|
|
|
ocrd-tesserocr-recognize -l DEBUG \
|
|
|
|
|
-m mets.xml -I OCR-D-SEG-LINE -O OCR-D-OCR-TESS \
|
|
|
|
|
-p <(echo $ocrd_tesserocr_recognize_parameters)
|
|
|
|
|
#ocrd workspace validate mets.xml
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
page_validate_xml() {
|
|
|
|
|
filegrp=$1
|
|
|
|
|
|
|
|
|
|
local file
|
|
|
|
|
for file in `ocrd workspace find -G $filegrp`; do
|
|
|
|
|
xmllint --noout --schema `dirname $0`/xsd/pagecontent.2018-07-15.xsd $file
|
|
|
|
|
done
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
do_fontident
|
|
|
|
|
do_linesegmentation
|
|
|
|
|
do_ocr
|
|
|
|
|
page_validate_xml OCR-D-OCR-TESS # This also makes sure PAGE Viewer can open it
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# XXX Multiple calls create multiple identical mets:agent elements
|
|
|
|
|
# XXX Global -l DEBUG
|
|
|
|
|
|
|
|
|
|
# vim:tw=120:
|
