qurator-spk/ocrd-galley
1
0
Fork 0
mirror of https://github.com/qurator-spk/ocrd-galley.git synced 2025-10-31 19:24:12 +01:00
Code Issues Projects Releases Wiki Activity

♻ Remove file groups using "ocrd workspace remove-group"

Browse source
This commit is contained in:
Gerber, Mike 2020-02-07 14:26:20 +01:00
parent c8039db686
commit d7a2aac44b
1 changed files with 9 additions and 23 deletions
Download patch file Download diff file Expand all files Collapse all files

32
my_ocrd_workflow
View file

@ -7,20 +7,6 @@ if [ "$LOG_LEVEL" = "DEBUG" -o "$LOG_LEVEL" = "TRACE" ]; then
fi
remove_filegrp() {
# Remove the given file group from the workspace
filegrp_use=$1
mets=$2
xmlstarlet ed --inplace \
-N mets=http://www.loc.gov/METS/ \
-d "//mets:fileGrp[@USE='$filegrp_use']" $mets
# XXX See also https://github.com/OCR-D/core/issues/245
# XXX This should also delete the files (after checking if they are indeed inside the workspace) and the directory
}
do_validate() {
validate_options="
--skip dimension
@ -42,8 +28,8 @@ do_binarization() {
# Binarize the images
ocrd_olena_binarize_paramters='{"impl": "sauvola-ms-split"}'
remove_filegrp OCR-D-IMG-BINPAGE mets.xml
remove_filegrp OCR-D-IMG-BIN mets.xml
ocrd workspace remove-group -rf OCR-D-IMG-BINPAGE
ocrd workspace remove-group -rf OCR-D-IMG-BIN
ocrd-olena-binarize -l $LOG_LEVEL \
-m mets.xml -I OCR-D-IMG -O OCR-D-IMG-BINPAGE \
-p "$ocrd_olena_binarize_paramters"
@ -52,8 +38,8 @@ do_binarization() {
do_linesegmentation_tesserocr() {
# Segment the lines in the binarized images
remove_filegrp OCR-D-SEG-REGION mets.xml
remove_filegrp OCR-D-SEG-LINE mets.xml
ocrd workspace remove-group -rf OCR-D-SEG-REGION
ocrd workspace remove-group -rf OCR-D-SEG-LINE
ocrd-tesserocr-segment-region -l $LOG_LEVEL \
-m mets.xml -I OCR-D-IMG-BINPAGE -O OCR-D-SEG-REGION
ocrd-tesserocr-segment-line -l $LOG_LEVEL \
@ -67,8 +53,8 @@ do_linesegmentation_sbb() {
# TODO: Check that this works with the RGB images
ocrd_sbb_textline_detector_parameters='{"model": "/var/lib/textline_detection"}'
remove_filegrp OCR-D-SEG-REGION mets.xml
remove_filegrp OCR-D-SEG-LINE mets.xml
ocrd workspace remove-group -rf OCR-D-SEG-REGION
ocrd workspace remove-group -rf OCR-D-SEG-LINE
ocrd-sbb-textline-detector -l $LOG_LEVEL \
-m mets.xml -I OCR-D-IMG-BINPAGE -O OCR-D-SEG-LINE \
-p "$ocrd_sbb_textline_detector_parameters"
@ -78,7 +64,7 @@ do_ocr() {
# Perform OCR on the segmented lines
ocrd_tesserocr_recognize_parameters='{ "model": "GT4HistOCR_2000000" }' # TODO mods:language + fontident → model
remove_filegrp OCR-D-OCR-TESS mets.xml
ocrd workspace remove-group -rf OCR-D-OCR-TESS
ocrd-tesserocr-recognize -l $LOG_LEVEL \
-m mets.xml -I OCR-D-SEG-LINE -O OCR-D-OCR-TESS \
-p "$ocrd_tesserocr_recognize_parameters"
@ -89,7 +75,7 @@ do_ocr_calamari() {
"checkpoint": "/var/lib/calamari-models/GT4HistOCR/2019-07-22T15:49+0200/*.ckpt.json",
"textequiv_level": "line"
}'
remove_filegrp OCR-D-OCR-CALAMARI mets.xml
ocrd workspace remove-group -rf OCR-D-OCR-CALAMARI
ocrd-calamari-recognize -l $LOG_LEVEL \
-m mets.xml -I OCR-D-SEG-LINE -O OCR-D-OCR-CALAMARI \
-p "$ocrd_calamari_recognize_parameters"
@ -153,7 +139,7 @@ for ocr_filegrp in OCR-D-OCR-CALAMARI OCR-D-OCR-TESS; do
#do_validate
if ocrd workspace list-group | grep -q OCR-D-GT-PAGE; then
remove_filegrp $ocr_filegrp-EVAL mets.xml
ocrd workspace remove-group -rf $ocr_filegrp-EVAL
ocrd-dinglehopper -m mets.xml -I OCR-D-GT-PAGE,$ocr_filegrp -O $ocr_filegrp-EVAL
fi