diff --git a/my_ocrd_workflow b/my_ocrd_workflow index f9451c4..b806b97 100755 --- a/my_ocrd_workflow +++ b/my_ocrd_workflow @@ -22,6 +22,11 @@ remove_filegrp() { # XXX This should also delete the files (after checking if they are indeed inside the workspace) and the directory } +do_validate() { + ocrd workspace validate --page-strictness lax mets.xml + # XXX ocrd-tesserocr INCONSISTENCY in TextRegion → use "--page-strictness lax" for now +} + do_binarization() { # Binarize the images @@ -47,10 +52,6 @@ do_fontident() { # XXX Check if ocrd-typegroups-classifier uses the whole image # 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 Unspecified USE category 'FONTIDENT' in fileGrp 'OCR-D-FONTIDENT' - # XXX File 'OCR-D-FONTIDENT_OCR-D-IMG_0002' does not manifest any physical page. - # XXX Won't download remote image } do_linesegmentation() { @@ -61,14 +62,12 @@ do_linesegmentation() { #ocrd-ocropy-segment -l $LOG_LEVEL \ # -m mets.xml -I OCR-D-IMG-BIN -O OCR-D-SEG-LINE # XXX ocrd-ocropy-segment throws an exception for buerger_gedichte_1778.ocrd - #ocrd workspace validate mets.xml ocrd-tesserocr-segment-region -l $LOG_LEVEL \ -m mets.xml -I OCR-D-IMG-BIN -O OCR-D-SEG-REGION - #ocrd workspace validate mets.xml ocrd-tesserocr-segment-line -l $LOG_LEVEL \ -m mets.xml -I OCR-D-SEG-REGION -O OCR-D-SEG-LINE - #ocrd workspace validate mets.xml + # XXX compare ocrd-tesserocr-segment* vs tesseract native } @@ -80,9 +79,6 @@ do_ocr() { ocrd-tesserocr-recognize -l $LOG_LEVEL \ -m mets.xml -I OCR-D-SEG-LINE -O OCR-D-OCR-TESS \ -p <(echo $ocrd_tesserocr_recognize_parameters) - #ocrd workspace validate mets.xml - # XXX INCONSISTENCY in TextRegion ID 'dummy' - # (The whitespace is different) } page_fix_xml() { @@ -129,10 +125,17 @@ page_fix_image_references() { do_fontident do_binarization +do_validate + do_linesegmentation +page_validate_xml OCR-D-SEG-REGION +page_validate_xml OCR-D-SEG-LINE +do_validate + do_ocr -page_fix_xml OCR-D-OCR-TESS -page_validate_xml OCR-D-OCR-TESS +do_validate + +page_fix_xml OCR-D-OCR-TESS # XXX is it necessary anymore? page_fix_image_references OCR-D-OCR-TESS