mirror of
https://github.com/qurator-spk/ocrd-galley.git
synced 2025-06-09 06:39:53 +02:00
📝 Document the functions
This commit is contained in:
parent
2204aee104
commit
e3a1afbc93
1 changed files with 24 additions and 4 deletions
|
@ -9,6 +9,8 @@ fi
|
|||
|
||||
|
||||
remove_filegrp() {
|
||||
# Remove the given file group from the workspace
|
||||
|
||||
filegrp_use=$1
|
||||
mets=$2
|
||||
|
||||
|
@ -21,18 +23,23 @@ remove_filegrp() {
|
|||
}
|
||||
|
||||
do_binarization() {
|
||||
# Binarize the images
|
||||
|
||||
remove_filegrp OCR-D-IMG-BIN mets.xml
|
||||
ocrd-kraken-binarize -l $LOG_LEVEL \
|
||||
-m mets.xml -I OCR-D-IMG -O OCR-D-IMG-BIN
|
||||
}
|
||||
|
||||
do_fontident() {
|
||||
# Identify fonts in the images
|
||||
|
||||
network=`python3 -c "import ocrd_typegroups_classifier, os; print(os.path.join(os.path.dirname(ocrd_typegroups_classifier.__file__), 'models', 'classifier.tgc'))"`
|
||||
ocrd_typegroups_classifier_parameters="
|
||||
{
|
||||
\"network\": \"$network\",
|
||||
\"stride\": 143
|
||||
}"
|
||||
|
||||
remove_filegrp OCR-D-FONTIDENT mets.xml
|
||||
ocrd-typegroups-classifier -l $LOG_LEVEL \
|
||||
-m mets.xml -I OCR-D-IMG -O OCR-D-FONTIDENT \
|
||||
|
@ -47,6 +54,8 @@ do_fontident() {
|
|||
}
|
||||
|
||||
do_linesegmentation() {
|
||||
# Segment the lines in the binarized images
|
||||
|
||||
remove_filegrp OCR-D-SEG-REGION mets.xml
|
||||
remove_filegrp OCR-D-SEG-LINE mets.xml
|
||||
#ocrd-ocropy-segment -l $LOG_LEVEL \
|
||||
|
@ -68,6 +77,8 @@ do_linesegmentation() {
|
|||
}
|
||||
|
||||
do_ocr() {
|
||||
# Perform OCR on the segmented lines
|
||||
|
||||
ocrd_tesserocr_recognize_parameters='{ "model": "eng" }' # TODO mods:language + fontident → model
|
||||
remove_filegrp OCR-D-OCR-TESS mets.xml
|
||||
ocrd-tesserocr-recognize -l $LOG_LEVEL \
|
||||
|
@ -80,7 +91,11 @@ do_ocr() {
|
|||
}
|
||||
|
||||
page_fix_xml() {
|
||||
# XXX core does not produce valid XML https://github.com/OCR-D/core/issues/242
|
||||
# Fix the PAGE XML generated by OCR-D core
|
||||
#
|
||||
# XXX core does not produce valid XML (See https://github.com/OCR-D/core/issues/242), fix it by setting the correct
|
||||
# PAGE XML version. This makes PAGE Viewer open the file.
|
||||
|
||||
filegrp=$1
|
||||
|
||||
local file
|
||||
|
@ -90,6 +105,8 @@ page_fix_xml() {
|
|||
}
|
||||
|
||||
page_validate_xml() {
|
||||
# Validate all PAGE XML against the XML schema
|
||||
|
||||
filegrp=$1
|
||||
|
||||
local file
|
||||
|
@ -99,8 +116,11 @@ page_validate_xml() {
|
|||
}
|
||||
|
||||
page_fix_image_references() {
|
||||
# Make image references relative to the PAGE XML file. The rest of OCR-D probably isn't going to like it, but it
|
||||
# is a. correct and b. makes PAGE Viewer open the image file automatically.
|
||||
# Make image references relative to the PAGE XML file
|
||||
#
|
||||
# The rest of OCR-D probably isn't going to like it, but it is a. correct and b. makes PAGE Viewer open the image file
|
||||
# automatically.
|
||||
|
||||
filegrp=$1
|
||||
|
||||
local file
|
||||
|
@ -116,7 +136,7 @@ do_binarization
|
|||
do_linesegmentation
|
||||
do_ocr
|
||||
page_fix_xml OCR-D-OCR-TESS
|
||||
page_validate_xml OCR-D-OCR-TESS # This also makes sure PAGE Viewer can open it
|
||||
page_validate_xml OCR-D-OCR-TESS
|
||||
page_fix_image_references OCR-D-OCR-TESS
|
||||
|
||||
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue