From 4ce249e6eddb7b40b6e4869e13b62898620ab08f Mon Sep 17 00:00:00 2001
From: fritz-hh <julien.pfefferkorn@googlemail.com>
Date: Mon, 22 Apr 2013 22:50:34 +0200
Subject: [PATCH] OCRmyPDF.sh: new debug option (-g) added

---
 OCRmyPDF.sh | 31 +++++++++++++++++++++++--------
 1 file changed, 23 insertions(+), 8 deletions(-)

diff --git a/OCRmyPDF.sh b/OCRmyPDF.sh
index 5cb14990..56c8c9e3 100644
--- a/OCRmyPDF.sh
+++ b/OCRmyPDF.sh
@@ -12,10 +12,14 @@ tesseract engine)
 Copyright: fritz from NAS4Free forum
 Version: $VERSION
 
-Usage: OCRmyPDF.sh  [-h] [-v] [-k] [-d] [-c] [-i] [-l language] inputfile
+Usage: OCRmyPDF.sh  [-h] [-v] [-g] [-k] [-d] [-c] [-i] [-l language] inputfile
 
 -h : Display this help message
 -v : Increase the verbosity (this option can be used more than once)
+-g : Activate debug mode:
+     - Generates a PDF file containing each page twice (once with the image, once without the image
+       but with the OCRed text as well as the detected bounding boxes detected during OCR)
+     - Set the verbosity to the highest possible 
 -k : Do not delete the temporary files
 -d : Deskew each page before performing OCR
 -c : Clean each page before performing OCR
@@ -49,12 +53,14 @@ KEEP_TMP="0"			# do not delete the temporary files (default)
 PREPROCESS_DESKEW="0"		# 0=no, 1=yes (deskew image)
 PREPROCESS_CLEAN="0"		# 0=no, 1=yes (clean image to improve OCR)
 PREPROCESS_CLEANINPDF="0"	# 0=no, 1=yes (put cleaned image in final PDF)
+DEBUG_MODE="0"			# 0=no, 1=yes (generates each PDF page twice, with and without image)
 
 # Parse optional command line arguments
-while getopts ":hvkdcil:" opt; do
+while getopts ":hvgkdcil:" opt; do
 	case $opt in
 		h) usage ; exit 0 ;;
 		v) VERBOSITY=$(($VERBOSITY+1)) ;;
+		g) VERBOSITY="10"; DEBUG_MODE="1" ;;
 		k) KEEP_TMP="1" ;;
 		d) PREPROCESS_DESKEW="1" ;;		
 		c) PREPROCESS_CLEAN="1" ;;
@@ -136,10 +142,11 @@ while read pageSize ; do
 	[ $VERBOSITY -ge $LOG_INFO ] && echo "Processing page $page"
 	
 	# create the name of the required file
-	curOrigImg="$tmp/${page}_Image"		# original image available in the current PDF page 
-						# (the image file may have a different orientation than in the pdf file)
-	curHocr="$tmp/$page.hocr"		# hocr file to be generated by the OCR SW for the current page
-	curOCRedPDF="$tmp/${page}-ocred.pdf"	# PDF file containing the image + the OCRed text for the current page
+	curOrigImg="$tmp/${page}_Image"			# original image available in the current PDF page 
+							# (the image file may have a different orientation than in the pdf file)
+	curHocr="$tmp/$page.hocr"			# hocr file to be generated by the OCR SW for the current page
+	curOCRedPDF="$tmp/${page}-ocred.pdf"		# PDF file containing the image + the OCRed text for the current page
+	curOCRedPDFDebug="$tmp/${page}-debug-ocred.pdf"	# PDF file containing data required to find out if OCR worked correctly
 	
 	[ $VERBOSITY -ge $LOG_DEBUG ] && echo "Page $page: Computing embedded image resolution"
 	# get width / height of PDF page
@@ -208,15 +215,22 @@ while read pageSize ; do
 	mv "$curHocr.html" "$curHocr"
 
 	# embed text and image to new pdf file
-	[ $VERBOSITY -ge $LOG_DEBUG ] && echo "Page $page: Embedding text in PDF"
 	if [ "$PREPROCESS_CLEANINPDF" -eq "1" ]; then
 		image4finalPDF="$curImgPixmapClean"
 	else
 		image4finalPDF="$curImgPixmapDeskewed"	
 	fi
+	[ $VERBOSITY -ge $LOG_DEBUG ] && echo "Page $page: Embedding text in PDF"
 	! python hocrTransform.py -r $dpi -i "$image4finalPDF" "$curHocr" "$curOCRedPDF" \
 		&& echo "Could not create PDF file from \"$curHocr\". Exiting..." && exit $EXIT_OTHER_ERROR
 	
+	# if requested generate special debug PDF page with visible OCR text
+	if [ $DEBUG_MODE -eq "1" ] ; then
+		[ $VERBOSITY -ge $LOG_DEBUG ] && echo "Page $page: Embedding text in PDF (debug page)"
+		! python hocrTransform.py -b -r $dpi "$curHocr" "$curOCRedPDFDebug" \
+			&& echo "Could not create PDF file from \"$curHocr\". Exiting..." && exit $EXIT_OTHER_ERROR	
+	fi
+	
 	# delete temporary files created for the current page
 	# to avoid using to much disk space in case of PDF files having many pages
 	if [ $KEEP_TMP -eq 0 ]; then
@@ -229,6 +243,7 @@ while read pageSize ; do
 
 	# go to next page of the pdf
 	cpt=$(($cpt+1))
+	
 done < "$FILE_SIZE_PAGES"
 
 
@@ -262,7 +277,7 @@ grep -i "Status.*not valid" "$FILE_VALIDATION_LOG" && pdf_valid=0
 grep -i "Status.*Not well-formed" "$FILE_VALIDATION_LOG" && pdf_valid=0
 [ $pdf_valid -eq 1 ] && echo "Output file: The generated PDF/A file is VALID" \
 	|| echo "Output file: The generated PDF/A file is INVALID"
-
+[ $VERBOSITY -ge $LOG_DEBUG ] && cat "$FILE_VALIDATION_LOG"