fixes #40 and code cleanup

src/ocrPage.sh

@@ -34,16 +34,16 @@ FORCE_OCR="${14}"			# Force to OCR, even if the page already contains fonts
 # Param 2: PDF page width in pt
 # Param 3: PDF page height in pt
 # Param 4: temporary file path (Path of the file in which the output should be written)
-# Output: A file (<pagenum>-img-characteristics.txt) containing the characteristics of the embedded image
-#          Structure of the file:
-#          <dpi> <colorspace>
+# Output:  A file containing the characteristics of the embedded image. File structure:
+#          DPI=<dpi>
+#          COLOR_SPACE=<colorspace>
 # Returns:
 #       - 0: if no error occurs
 #       - 1: in case the page already contains fonts (which should be the case for PDF generated from scanned pages) 
 #       - 2: in case the page contains more than one image
 ##################################
 getImgInfo() {
-	local page widthPDF heightPDF curImgCharacteristics nbImg curImg propCurImg widthCurImg heightCurImg colorspaceCurImg dpi
+	local page widthPDF heightPDF curImgInfo nbImg curImg propCurImg widthCurImg heightCurImg colorspaceCurImg dpi
 
 	# page number
 	page="$1"
@@ -51,7 +51,7 @@ getImgInfo() {
 	widthPDF="$2"
 	heightPDF="$3"
 	# path of the file in which the output should be written
-	curImgCharacteristics="$4"
+	curImgInfo="$4"
 
 	
 	[ $VERBOSITY -ge $LOG_DEBUG ] && echo "Page $page: Size ${heightPDF}x${widthPDF} (h*w in pt)"
@@ -85,7 +85,8 @@ getImgInfo() {
 	dpi=`echo "scale=0;$dpi/1" | bc`
 	
 	# save the image characteristics
-	echo "$dpi $colorspaceCurImg" > $curImgCharacteristics
+	echo "DPI=$dpi" > "$curImgInfo"
+	echo "COLOR_SPACE=$colorspaceCurImg" >> "$curImgInfo"
 	
 	return 0
 }
@@ -104,11 +105,11 @@ curOrigImg="$TMP_FLD/${page}.orig-img"				# original image available in the curr
 curHocr="$TMP_FLD/${page}.hocr"					# hocr file to be generated by the OCR SW for the current page
 curOCRedPDF="$TMP_FLD/${page}.ocred.pdf"			# PDF file containing the image + the OCRed text for the current page
 curOCRedPDFDebug="$TMP_FLD/${page}.ocred.todebug.pdf"		# PDF file containing data required to find out if OCR worked correctly
-curImgCharacteristics="$TMP_FLD/${page}.orig-img-info.txt"	# Detected characteristics of the embedded image
+curImgInfo="$TMP_FLD/${page}.orig-img-info.txt"			# Detected characteristics of the embedded image
 
 
 # auto-detect the characteristics of the embedded image
-getImgInfo "$page" "$widthPDF" "$heightPDF" "$curImgCharacteristics"
+getImgInfo "$page" "$widthPDF" "$heightPDF" "$curImgInfo"
 ret_code="$?"
 # in case the page contains text do not OCR, unless the FORCE_OCR flag is set
 if [ "$ret_code" -eq "1" -a "$FORCE_OCR" -eq "0" ]; then
@@ -124,12 +125,11 @@ elif [ "$ret_code" -eq "2" ]; then
 	[ $VERBOSITY -ge $LOG_WARN ] && echo "Page $page: Continuing anyway, assuming a default resolution of $dpi dpi"
 else
 	# read the image characteristics from the file
-	dpi=`cat "$curImgCharacteristics" | cut -f1 -d" "`
-	colorspaceCurImg=`cat "$curImgCharacteristics" | cut -f2 -d" "`
+	dpi=`cat "$curImgInfo" | grep "^DPI=" | cut -f2 -d"="`
+	colorspaceCurImg=`cat "$curImgInfo" | grep "^COLOR_SPACE=" | cut -f2 -d"="`
 fi
 
-# perform oversampling if the resolution is not big enough
-# to get good OCR results
+# perform oversampling if the resolution is not sufficient to get good OCR results
 if [ "$dpi" -lt "$OVERSAMPLING_DPI" ]; then
 	[ $VERBOSITY -ge $LOG_WARN ] && echo "Page $page: Low image resolution detected ($dpi dpi). Performing oversampling ($OVERSAMPLING_DPI dpi) to try to get better OCR results." 
 	dpi="$OVERSAMPLING_DPI"
@@ -148,7 +148,7 @@ curImgPixmap="$TMP_FLD/$page.$ext"
 curImgPixmapDeskewed="$TMP_FLD/$page.deskewed.$ext"
 curImgPixmapClean="$TMP_FLD/$page.cleaned.$ext"
 
-# extract current page as image with right orientation and resolution
+# extract current page as image with correct orientation and resolution
 [ $VERBOSITY -ge $LOG_DEBUG ] && echo "Page $page: Extracting image as $ext file (${dpi} dpi)"
 ! pdftoppm -f $page -l $page -r $dpi $opt "$FILE_INPUT_PDF" > "$curImgPixmap" \
 	&& echo "Could not extract page $page as $ext from \"$FILE_INPUT_PDF\". Exiting..." && exit $EXIT_OTHER_ERROR
@@ -206,7 +206,7 @@ if [ $KEEP_TMP -eq 0 ]; then
 	rm -f "$curImgPixmap"
 	rm -f "$curImgPixmapDeskewed"
 	rm -f "$curImgPixmapClean"
-	rm -f "$curImgCharacteristics"
+	rm -f "$curImgInfo"
 fi
 
 exit 0