RELEASE_NOTES update prior delivery of v2.0-stable

Allow tesseract 3.02.01 to be used.
Even 3.02.01 fails in few cases (see issue #28). I decided to allow this
version anyway because 3.02.02 is not yet available for some widespread
linux distributions

OCRmyPDF.sh

@@ -41,11 +41,11 @@ Usage: OCRmyPDF.sh  [-h] [-v] [-g] [-k] [-d] [-c] [-i] [-o dpi] [-f] [-l languag
 -f : Force to OCR the whole document, even if some page already contain font data 
      (which should not be the case for PDF files built from scnanned images) 
 -l : Set the language of the PDF file in order to improve OCR results (default "eng")
-     Any language supported by tesseract is supported.
+     Any language supported by tesseract is supported (Tesseract uses 3-character ISO 639-2 language codes)
+     Multiple languages may be specified, separated by '+' characters.
 -C : Pass an additional configuration file to the tesseract OCR engine.
      (this option can be used more than once)
      Note 1: The configuration file must be available in the "tessdata/configs" folder of your tesseract installation
-     Note 2: The folder "./tess-cfg" contains useful tesseract configuration files
 inputfile  : PDF file to be OCRed
 outputfile : The PDF/A file that will be generated 
 --------------------------------------------------------------------------------------
@@ -111,8 +111,7 @@ done
 # Remove the optional arguments parsed above.
 shift $((OPTIND-1))
 
-# Check if the number of mandatory parameters
-# provided is as expected
+# Check if the number of mandatory parameters provided is as expected
 if [ "$#" -ne "2" ]; then
 	echo "Exactly two mandatory argument shall be provided ($# arguments provided)"
 	usage
@@ -141,20 +140,36 @@ cd "`dirname $0`"
 ! command -v pdfimages > /dev/null && echo "Please install poppler-utils. Exiting..." && exit $EXIT_MISSING_DEPENDENCY
 ! command -v pdftoppm > /dev/null && echo "Please install poppler-utils. Exiting..." && exit $EXIT_MISSING_DEPENDENCY
 ! command -v pdffonts > /dev/null && echo "Please install poppler-utils. Exiting..." && exit $EXIT_MISSING_DEPENDENCY
-! command -v pdftk > /dev/null && echo "Please install pdftk. Exiting..." && exit $EXIT_MISSING_DEPENDENCY
 [ $PREPROCESS_CLEAN -eq 1 ] && ! command -v unpaper > /dev/null && echo "Please install unpaper. Exiting..." && exit $EXIT_MISSING_DEPENDENCY
 ! command -v tesseract > /dev/null && echo "Please install tesseract and tesseract-data. Exiting..." && exit $EXIT_MISSING_DEPENDENCY
-! command -v python2 > /dev/null && echo "Please install python v2.x, and the python libraries: reportlab, lxml. Exiting..." && exit $EXIT_MISSING_DEPENDENCY
+! command -v python2 > /dev/null && echo "Please install python v2.x. Exiting..." && exit $EXIT_MISSING_DEPENDENCY
+! python2 -c 'import lxml' 2>/dev/null && echo "Please install the python library lxml. Exiting..." && exit $EXIT_MISSING_DEPENDENCY
+! python2 -c 'import reportlab' 2>/dev/null && echo "Please install the python library reportlab. Exiting..." && exit $EXIT_MISSING_DEPENDENCY
 ! command -v gs > /dev/null && echo "Please install ghostcript. Exiting..." && exit $EXIT_MISSING_DEPENDENCY
 ! command -v java > /dev/null && echo "Please install java. Exiting..." && exit $EXIT_MISSING_DEPENDENCY
 
 
-# ensure tesseract v3.02.02 or newer is installed
+
+# ensure the right tesseract version is installed
 # older versions are known to produce malformed hocr output and should not be used
-tessversion=`tesseract -v 2>&1 | grep "tesseract"`
-if [ $VERBOSITY -ge $LOG_WARN -a $((`echo ${tessversion} | sed s/[^0-9]//g`-30202)) -lt 0 ]; then
-	echo "Warning: Please use tesseract 3.02.02 or newer. Older versions are known to produce invalid hocr output (installed version: ${tessversion})"
-fi
+# Even 3.02.01 fails in few cases (see issue #28). I decided to allow this version anyway because
+# 3.02.02 is not yet available for some widespread linux distributions
+reqtessversion="3.02.01"
+tessversion=`tesseract -v 2>&1 | grep "tesseract" | sed s/[^0-9.]//g`
+tesstooold=$(echo "`echo $tessversion | sed s/[.]//2`-`echo $reqtessversion | sed s/[.]//2` < 0" | bc)
+[ "$tesstooold" -eq "1" ] \
+	&& echo "Please install tesseract ${reqtessversion} or newer (currently installed version is ${tessversion})" && exit $EXIT_MISSING_DEPENDENCY
+
+# ensure the right GNU parallel version is installed
+# older version do not support -q flag (required to escape special characters)
+reqparallelversion="20130222"
+parallelversion=`parallel --minversion 0`
+! parallel --minversion "$reqparallelversion" > /dev/null \
+	&& echo "Please install GNU parallel ${reqparallelversion} or newer (currently installed version is ${parallelversion})" && exit $EXIT_MISSING_DEPENDENCY
+
+# ensure pdftoppm is provided by poppler-utils, not the older xpdf version
+! pdftoppm -v 2>&1 | grep -q 'Poppler' && echo "Please remove xpdf and install poppler-utils. Exiting..." && $EXIT_MISSING_DEPENDENCY
+
 
 
 # Display the version of the tools if log level is LOG_DEBUG
@@ -171,9 +186,6 @@ if [ $VERBOSITY -ge $LOG_DEBUG ]; then
 	pdftoppm -v
 	pdffonts -v
 	echo "--------------------------------"
-	echo "pdftk version:"
-	pdftk --version
-	echo "--------------------------------"
 	echo "unpaper version:"
 	unpaper --version
 	echo "--------------------------------"
@@ -193,24 +205,44 @@ fi
 
 
 
-# Initialize path to temporary files
-today=$(date +"%Y%m%d_%H%M")
-fld=$(basename "$FILE_INPUT_PDF" | sed 's/[.][^.]*//')
-TMP_FLD="${TMP}/$today.filename.$fld"
+# check if the languages passed to tesseract are all supported
+for currentlan in `echo "$LAN" | sed 's/+/ /g'`; do
+	if ! tesseract --list-langs 2>&1 | grep "^$currentlan\$" > /dev/null; then
+		echo "The language \"$currentlan\" is not supported by tesseract."
+		tesseract --list-langs 2>&1 | tr '\n' ' '; echo
+		echo "Exiting..."
+		exit $EXIT_BAD_ARGS
+	fi
+done
+
+
+
+# Initialize path to temporary files using mktemp
+# Goal: save tmp file in a sub-folder of the $TMPDIR environment variable (or in "/tmp" if unset)
+# Unfortunately, Linux mktemp is not compatible with FreeBSD/OSX mktemp
+# Linux version requires no arg
+# FreeBSD requires '-t prefix' to be used so that $TMPDIR is taken into account
+# But in Linux '-t template' is handled differently than in FreeBSD
+# Therefore different calls must be used for Linux and for FreeBSD
+prefix="$(date +"%Y%m%d_%H%M").filename.$(basename "$FILE_INPUT_PDF" | sed 's/[.][^.]*$//')"	# prefix made of date, time and pdf file name without extension
+TMP_FLD=`mktemp -d 2>/dev/null || mktemp -d -t "${prefix}" 2>/dev/null`				# try Linux syntax first, if it fails try FreeBSD/OSX			
+if [ $? -ne 0 ]; then
+	if [ -z "$TMPDIR" ]; then
+		echo "Could not create folder for temporary files. Please ensure you have sufficient right and \"/tmp\" exists"
+	else
+		echo "Could not create folder for temporary files. Please ensure you have sufficient right and \"$TMPDIR\" exists"
+	fi
+	exit $EXIT_FILE_ACCESS_ERROR
+fi
+[ $VERBOSITY -ge $LOG_DEBUG ] && echo "Created temporary folder: \"$TMP_FLD\""
+
 FILE_TMP="${TMP_FLD}/tmp.txt"						# temporary file with a very short lifetime (may be used for several things)
 FILE_PAGES_INFO="${TMP_FLD}/pages-info.txt"				# for each page: page #; width in pt; height in pt
-FILE_OUTPUT_PDF_CAT="${TMP_FLD}/ocred.pdf"				# concatenated OCRed PDF files
 FILE_VALIDATION_LOG="${TMP_FLD}/pdf_validation.log"			# log file containing the results of the validation of the PDF/A file
 
-# Create tmp folder
-[ $VERBOSITY -ge $LOG_DEBUG ] && echo "Creating temporary folder: \"$TMP_FLD\""
-rm -r -f "${TMP_FLD}"
-mkdir -p "${TMP_FLD}"
 
 
-
-
-# get the size of each pdf page (width / height) in pt (inch*72)
+# get the size of each pdf page (width / height) in pt (i.e. inch/72)
 [ $VERBOSITY -ge $LOG_DEBUG ] && echo "Input file: Extracting size of each page (in pt)"
 ! identify -format "%w %h\n" "$FILE_INPUT_PDF" > "$FILE_TMP" \
 	&& echo "Could not get size of PDF pages. Exiting..." && exit $EXIT_BAD_INPUT_FILE
@@ -219,23 +251,18 @@ sed '/^$/d' "$FILE_TMP" | awk '{printf "%04d %s\n", NR, $0}' > "$FILE_PAGES_INFO
 numpages=`tail -n 1 "$FILE_PAGES_INFO" | cut -f1 -d" "`
 
 # process each page of the input pdf file
-parallel -q -k --halt-on-error 1 "$OCR_PAGE" "$FILE_INPUT_PDF" "{}" "$numpages" "$TMP_FLD" \
+parallel --gnu -q -k --halt-on-error 1 "$OCR_PAGE" "$FILE_INPUT_PDF" "{}" "$numpages" "$TMP_FLD" \
 	"$VERBOSITY" "$LAN" "$KEEP_TMP" "$PREPROCESS_DESKEW" "$PREPROCESS_CLEAN" "$PREPROCESS_CLEANTOPDF" "$OVERSAMPLING_DPI" \
 	"$PDF_NOIMG" "$TESS_CFG_FILES" "$FORCE_OCR" < "$FILE_PAGES_INFO"
 ret_code="$?"
 [ $ret_code -ne 0 ] && exit $ret_code 
 
-# concatenate all pages
-[ $VERBOSITY -ge $LOG_DEBUG ] && echo "Output file: Concatenating all pages"
-! pdftk "${TMP_FLD}/"*ocred*.pdf cat output "$FILE_OUTPUT_PDF_CAT" \
-	&& echo "Could not concatenate individual PDF pages (\"${TMP_FLD}/*-ocred.pdf\") to one file. Exiting..." && exit $EXIT_OTHER_ERROR
-
-# convert the pdf file to match PDF/A format
-[ $VERBOSITY -ge $LOG_DEBUG ] && echo "Output file: Converting to PDF/A" 
+# concatenate all pages and convert the pdf file to match PDF/A format
+[ $VERBOSITY -ge $LOG_DEBUG ] && echo "Output file: Concatenating all pages to the final PDF/A file" 
 ! gs -dQUIET -dPDFA -dBATCH -dNOPAUSE -dUseCIEColor \
 	-sProcessColorModel=DeviceCMYK -sDEVICE=pdfwrite -sPDFACompatibilityPolicy=2 \
-	-sOutputFile="$FILE_OUTPUT_PDFA" "$FILE_OUTPUT_PDF_CAT" 1> /dev/null 2> /dev/null \
-	&& echo "Could not convert PDF file \"$FILE_OUTPUT_PDF_CAT\" to PDF/A. Exiting..." && exit $EXIT_OTHER_ERROR
+	-sOutputFile="$FILE_OUTPUT_PDFA" "${TMP_FLD}/"*ocred*.pdf 1> /dev/null 2> /dev/null \
+	&& echo "Could not concatenate all pages to the final PDF/A file. Exiting..." && exit $EXIT_OTHER_ERROR
 
 # validate generated pdf file (compliance to PDF/A)
 [ $VERBOSITY -ge $LOG_DEBUG ] && echo "Output file: Checking compliance to PDF/A standard" 

README.md

@@ -11,12 +11,14 @@ Main features
 - Generates a searchable PDF/A file from a PDF file containing only images
 - Places OCRed text accurately below the image to ease copy / paste
 - Keeps the exact resolution of the original embedded images
-    - or if requested oversample the images before OCRing so as to get better results 
+    - or if requested oversamples the images before OCRing so as to get better results 
 - If requested deskews and / or clean the image before performing OCR
 - Validates the generated file against the PDF/A specification using jhove
 - Provides debug mode to enable easy verification of the OCR results
 - Processes several pages in parallel if more than one CPU core is available
 
+For details: please consult the release notes
+
 Motivation
 ----------
 
@@ -35,13 +37,14 @@ I found many, but none of them were really satisfying.
 Install
 -------
 
-Download OCRmyPDF here: https://github.com/fritz-hh/OCRmyPDF/tags
+Download OCRmyPDF here: https://github.com/fritz-hh/OCRmyPDF/releases
 
 Copy the file in onto your linux/unix machine and extract it.
 
 Run: "sh ./OCRmyPDF.sh -h" to get the script usage
 
-If not yet installed, the script will notify you about dependencies that need to be installed
+If not yet installed, the script will notify you about dependencies that need to be installed.
+The script requires specific versions of the dependencies. Older version than the ones mentioned in the release notes are likely not to be compatible to OCRmyPDF.
 
 Support
 -------
@@ -49,7 +52,7 @@ Support
 In case you detect an issue, please:
 
 - Check if your issue is already known
-- if no problem report exists on github, please create one here: https://github.com/fritz-hh/OCRmyPDF/issues
+- If no problem report exists on github, please create one here: https://github.com/fritz-hh/OCRmyPDF/issues
 - Describe your problem thoroughly
 - Append the console output of the script when running the debug mode (-g option)
 - If possible provide your input PDF file as well as the content of the temporary folder (using a file sharing service like www.file-upload.net)
@@ -57,4 +60,9 @@ In case you detect an issue, please:
 Press & Media
 -------------
 
-- c't 1-2014, page 59: Detailed presentation of OCRmyPDF v1.0 in the leading german IT magazine (c't)
+- c't 1-2014, page 59: Detailed presentation of OCRmyPDF v1.0 in the leading german IT magazine c't (http://www.heise.de/ct/inhalt/2014/1/58/)
+
+Disclaimer
+----------
+
+The software is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

RELEASE_NOTES.md

@@ -5,6 +5,81 @@ Please always read this file before installing the package
 
 Download software here: https://github.com/fritz-hh/OCRmyPDF/tags
 
+v2.0-stable (2014-01-25):
+=======
+
+New features
+------------
+
+- Check if the language(s) passed using the -l option is supported by tesseract (fixes #60)
+
+Changes
+-------
+
+- Allow OCRmyPDF to be used with tesseract 3.02.01, even though OCR might fail for few PDF file (see issue #28). Rationale: For some linux distribution, no newer version than tesseract 3.02.01 is available
+
+Fixes
+-----
+
+- More robust algorithm for checking the version of the installed tesseract package
+
+Tested with
+-----------
+
+- Operating system: FreeBSD 9.1
+- Dependencies:
+   - parallel 20130222
+   - poppler-utils 0.22.2
+   - ImageMagick 6.8.0-7 2013-03-30
+   - Unpaper 0.3
+   - tesseract 3.02.02
+   - Python 2.7.3
+   - ghoscript (gs): 9.06
+   - java: openjdk version "1.7.0_17"
+
+v2.0-rc2 (2014-01-16):
+=======
+   
+New features
+------------
+
+- None
+
+Changes
+-------
+
+- Size reduction of final PDF file: (fixes #50)
+   - Support for monochrome (Black&White) images (massive size reduction in final PDF: >80%)
+   - Reduced size of grayscale images (by 13% on test PDF file)
+- Preventing fi, fl ligatures does not require anymore to pass an additional config file to tesseract using the -C option (fixes #58)
+- Location of temporary folder according to content of environment variable TMPDIR.
+- Dependency to pdftk removed
+- Check for compatible versions of dependencies: (fixes #51)
+   - parallel and tesseract
+   - python libraries reportlab and lxml
+
+Fixes
+-----
+
+- Improved portability with various shells (dash, bash, tcsh) and OS (FreeBSD, MAC OSX, Linux) (fixes #59)
+- Corrected bug in case the input PDF file contains a space character (fixes #48)
+- Prevent spurious error message in case there is no image in a PDF page
+- Prevent collision of temporary folder names (fixes #57)
+
+Tested with
+-----------
+
+- Operating system: FreeBSD 9.1
+- Dependencies:
+   - parallel 20130222
+   - poppler-utils 0.22.2
+   - ImageMagick 6.8.0-7 2013-03-30
+   - Unpaper 0.3
+   - tesseract 3.02.02
+   - Python 2.7.3
+   - ghoscript (gs): 9.06
+   - java: openjdk version "1.7.0_17"
+
 v2.0-rc1 (2014-01-07):
 ====
 
@@ -23,7 +98,7 @@ Changes
 -------
 
 - In debug mode: The debug page is now placed after the respective "normal" page
-- Reduced disk space usage if -d (deskew) or -c (cleanup) options are not selected
+- Reduced disk space usage in temporary folder if -d (deskew) or -c (cleanup) options are not selected
 - New file src/config.sh containing various configuration parameters
 - Documentation of the tesseract config file "tess-cfg/no_ligature" improved 
 - Improved consistency of the temporary file names
@@ -42,6 +117,7 @@ Tested with
 
 - Operating system: FreeBSD 9.1
 - Dependencies:
+   - parallel 20130222
    - poppler-utils 0.22.2
    - ImageMagick 6.8.0-7 2013-03-30
    - Unpaper 0.3

src/config.sh

@@ -1,11 +1,22 @@
+#####################################################################################
+# The following parameters might be changed by the user
+#####################################################################################
+
+DEFAULT_DPI=300				# dpi value used as fall back if the page dpi cannot be determined
+
+#####################################################################################
+# Do NOT change the following parameters
+#####################################################################################
+
 TOOLNAME="OCRmyPDF"
-VERSION="v2.0-rc1"
+VERSION="v2.0-stable"
 
 # possible exit codes
 EXIT_BAD_ARGS="1"
 EXIT_BAD_INPUT_FILE="2"
 EXIT_MISSING_DEPENDENCY="3"
 EXIT_INVALID_OUPUT_PDFA="4"
+EXIT_FILE_ACCESS_ERROR="5"
 EXIT_OTHER_ERROR="15"
 
 # possible log levels
@@ -16,10 +27,6 @@ LOG_DEBUG="3"				# debug level logging
 
 # various paths
 SRC="./src"				# location of the source folder (except source of external tools like jhove)
-TMP="./tmp"				# location of the temporary files (one sub-folder will be created per PDF file to be processed)
 OCR_PAGE="$SRC/ocrPage.sh"		# path to the script aimed at OCRing one page
 JHOVE="./jhove/bin/JhoveApp.jar"	# java SW for validating the final PDF/A
 JHOVE_CFG="./jhove/conf/jhove.conf"	# location of the jhove config file
-
-# other 
-DEFAULT_DPI=300				# dpi value used as fall back if the page dpi cannot be determined

src/hocrTransform.py

@@ -1,4 +1,5 @@
-#!/usr/bin/python
+#!/usr/local/bin/python2
+# coding: utf-8
 ##############################################################################
 # Copyright (c) 2013-14: fritz-hh from Github (https://github.com/fritz-hh)
 #
@@ -6,12 +7,95 @@
 # Initial version by Jonathan Brinley, jonathanbrinley@gmail.com
 ##############################################################################
 from reportlab.pdfgen.canvas import Canvas
+from reportlab.pdfgen.pdfimages import PDFImage
 from reportlab.lib.units import inch
 from lxml import etree as ElementTree
 from PIL import Image
 import re, sys
 import argparse
 
+
+def monkeypatch_method(cls):
+	'''
+	Override a class method at runtime.
+
+	Rationale:
+	https://mail.python.org/pipermail/python-dev/2008-January/076194.html
+	'''
+	def decorator(func):
+		setattr(cls, func.__name__, func)
+		return func
+	return decorator
+
+
+@monkeypatch_method(PDFImage)
+def PIL_imagedata(self):
+	'''
+	Add ability to output greyscale and 1-bit PIL images without conversion to RGB.
+
+	The upstream Python 2.7 version of reportlab converts 1-bit PIL images to RGB
+	instead of saving them in a lower BPP format.  They have since added the following
+	fix to their Python 3.3 branch, but it has not been back-ported.
+
+	https://bitbucket.org/rptlab/reportlab/commits/177ddcbe4df6f9b461dac62612df9b8da3966a5d
+	'''
+	image = self.image
+	if image.format == 'JPEG':
+		fp = image.fp
+		fp.seek(0)
+		return self._jpg_imagedata(fp)
+
+	from reportlab.lib.utils import import_zlib
+	from reportlab import rl_config
+	from reportlab.pdfbase.pdfutils import _AsciiBase85Encode, _chunker
+
+	self.source = 'PIL'
+	zlib = import_zlib()
+	if not zlib:
+		return
+
+	bpc = 8
+	# Use the colorSpace in the image
+	if image.mode == 'CMYK':
+		myimage = image
+		colorSpace = 'DeviceCMYK'
+		bpp = 4
+	elif image.mode == '1':
+		myimage = image
+		colorSpace = 'DeviceGray'
+		bpp = 1
+		bpc = 1
+	elif image.mode == 'L':
+		myimage = image
+		colorSpace = 'DeviceGray'
+		bpp = 1
+	else:
+		myimage = image.convert('RGB')
+		colorSpace = 'RGB'
+		bpp = 3
+	imgwidth, imgheight = myimage.size
+
+	# this describes what is in the image itself
+	# *NB* according to the spec you can only use the short form in inline images
+
+	imagedata = ['BI /W %d /H %d /BPC %d /CS /%s /F [%s/Fl] ID' %
+				 (imgwidth, imgheight, bpc, colorSpace, rl_config.useA85 and '/A85 ' or '')]
+
+	# use a flate filter and, optionally, Ascii Base 85 to compress
+	raw = myimage.tostring()
+	rowstride = (imgwidth * bpc * bpp + 7) / 8
+	assert len(raw) == rowstride * imgheight, "Wrong amount of data for image"
+	data = zlib.compress(raw)  # this bit is very fast...
+
+	if rl_config.useA85:
+		# ...sadly this may not be
+		data = _AsciiBase85Encode(data)
+	# append in blocks of 60 characters
+	_chunker(data, imagedata)
+	imagedata.append('EI')
+	return (imagedata, imgwidth, imgheight)
+
+
 class hocrTransform():
 	"""
 	A class for converting documents from the hOCR format.
@@ -24,20 +108,21 @@ class hocrTransform():
 
 		self.hocr = ElementTree.ElementTree()
 		self.hocr.parse(hocrFileName)
- 
+
 		# if the hOCR file has a namespace, ElementTree requires its use to find elements
 		matches = re.match('({.*})html', self.hocr.getroot().tag)
 		self.xmlns = ''
 		if matches:
 			self.xmlns = matches.group(1)
-			
+
 		# get dimension in pt (not pixel!!!!) of the OCRed image
+		self.width, self.height = None, None
 		for div in self.hocr.findall(".//%sdiv[@class='ocr_page']"%(self.xmlns)):
 			coords = self.element_coordinates(div)
 			self.width = self.px2pt(coords[2]-coords[0])
 			self.height = self.px2pt(coords[3]-coords[1])
 			break # there shouldn't be more than one, and if there is, we don't want it
-			
+
 		# no width and heigh definition in the ocr_image element of the hocr file
 		if self.width is None:
 			print("No page dimension found in the hocr file")
@@ -54,7 +139,7 @@ class hocrTransform():
 			return self._get_element_text(body).encode('utf-8') # XML gives unicode
 		else:
 			return ''
-  
+
 	def _get_element_text(self, element):
 		"""
 		Return the textual content of the element and its children
@@ -67,7 +152,7 @@ class hocrTransform():
 		if element.tail is not None:
 			text = text + element.tail
 		return text
-    
+
 	def element_coordinates(self, element):
 		"""
 		Returns a tuple containing the coordinates of the bounding box around
@@ -80,13 +165,24 @@ class hocrTransform():
 				coords = matches.group(1).split()
 				out = (int(coords[0]),int(coords[1]),int(coords[2]),int(coords[3]))
 		return out
-    
+
 	def px2pt(self, pxl):
 		"""
 		Returns the length in pt given length in pxl
 		"""
 		return float(pxl)/self.dpi*inch
-			
+
+	def replace_unsupported_chars(self, str):
+		"""
+		Given an input string, returns the corresponding string that:
+		- is available in the helvetica facetype
+		- does not contain any ligature (to allow easy search in the PDF file)
+		"""		
+		# The 'u' before the character to replace indicates that it is a unicode character
+		str=str.replace(u"fl","fl")
+		str=str.replace(u"fi","fi")
+		return str
+		
 	def to_pdf(self, outFileName, imageFileName, showBoundingboxes, fontname="Helvetica"):
 		"""
 		Creates a PDF file with an image superimposed on top of the text.
@@ -97,13 +193,13 @@ class hocrTransform():
 		"""
 		# create the PDF file
 		pdf = Canvas(outFileName, pagesize=(self.width, self.height), pageCompression=1) # page size in points (1/72 in.)
-		
+
 		# draw bounding box for each paragraph
 		pdf.setStrokeColorRGB(0,1,1)	# light blue for bounding box of paragraph
 		pdf.setFillColorRGB(0,1,1)	# light blue for bounding box of paragraph
 		pdf.setLineWidth(0)		# no line for bounding box
 		for elem in self.hocr.findall(".//%sp[@class='%s']" % (self.xmlns, "ocr_par")):
-		
+
 			elemtxt=self._get_element_text(elem).rstrip()
 			if len(elemtxt) == 0:
 				continue
@@ -113,12 +209,12 @@ class hocrTransform():
 			y1=self.px2pt(coords[1])
 			x2=self.px2pt(coords[2])
 			y2=self.px2pt(coords[3])
-			
+
 			# draw the bbox border
 			if showBoundingboxes == True:
-				pdf.rect(x1, self.height-y2, x2-x1, y2-y1, fill=1)		
-		
-		
+				pdf.rect(x1, self.height-y2, x2-x1, y2-y1, fill=1)
+
+
 		# check if element with class 'ocrx_word' are available
 		# otherwise use 'ocr_line' as fallback
 		elemclass="ocr_line"
@@ -133,6 +229,9 @@ class hocrTransform():
 		for elem in self.hocr.findall(".//%sspan[@class='%s']" % (self.xmlns, elemclass)):
 
 			elemtxt=self._get_element_text(elem).rstrip()
+			
+			elemtxt=self.replace_unsupported_chars(elemtxt)
+			
 			if len(elemtxt) == 0:
 				continue
 
@@ -141,7 +240,7 @@ class hocrTransform():
 			y1=self.px2pt(coords[1])
 			x2=self.px2pt(coords[2])
 			y2=self.px2pt(coords[3])
-			
+
 			# draw the bbox border
 			if showBoundingboxes == True:
 				pdf.rect(x1, self.height-y2, x2-x1, y2-y1, fill=0)
@@ -152,7 +251,7 @@ class hocrTransform():
 
 			# set cursor to bottom left corner of bbox (adjust for dpi)
 			text.setTextOrigin(x1, self.height-y2)
-       
+
 			# scale the width of the text to fill the width of the bbox
 			text.setHorizScale(100*(x2-x1)/pdf.stringWidth(elemtxt, fontname, fontsize))
 
@@ -162,13 +261,14 @@ class hocrTransform():
 
 		# put the image on the page, scaled to fill the page
 		if imageFileName != None:
-			im = Image.open(imageFileName)		
+			im = Image.open(imageFileName)
 			pdf.drawInlineImage(im, 0, 0, width=self.width, height=self.height)
-			
+
 		# finish up the page and save it
 		pdf.showPage()
 		pdf.save()
-  
+
+
 if __name__ == "__main__":
 	parser = argparse.ArgumentParser(description='Convert hocr file to PDF')
 	parser.add_argument('-b', '--boundingboxes', action="store_true", default=False, help='Show bounding boxes borders')
@@ -181,5 +281,5 @@ if __name__ == "__main__":
 	hocr = hocrTransform(args.hocrfile, args.resolution)
 	hocr.to_pdf(args.outputfile, args.image, args.boundingboxes)
 
-	
-	
+
+

src/ocrPage.sh

@@ -37,13 +37,14 @@ FORCE_OCR="${14}"			# Force to OCR, even if the page already contains fonts
 # Output:  A file containing the characteristics of the embedded image. File structure:
 #          DPI=<dpi>
 #          COLOR_SPACE=<colorspace>
+#          DEPTH=<colordepth>
 # 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 curImgInfo nbImg curImg propCurImg widthCurImg heightCurImg colorspaceCurImg dpi
+	local page widthPDF heightPDF curImgInfo nbImg curImg propCurImg widthCurImg heightCurImg colorspaceCurImg depthCurImg dpi
 
 	# page number
 	page="$1"
@@ -58,25 +59,26 @@ getImgInfo() {
 	
 	
 	# check if the page already contains fonts (which should not be the case for PDF based on scanned files
-	[ `pdffonts -f $page -l $page ${FILE_INPUT_PDF} | wc -l` -gt 2 ] && echo "Page $page: Page already contains font data !!!" && return 1
+	[ `pdffonts -f $page -l $page "${FILE_INPUT_PDF}" | wc -l` -gt 2 ] && echo "Page $page: Page already contains font data !!!" && return 1
 
 	
 	# extract raw image from pdf file to compute resolution
 	# unfortunately this image can have another orientation than in the pdf...
 	# so we will have to extract it again later using pdftoppm
 	pdfimages -f $page -l $page -j "$FILE_INPUT_PDF" "$curOrigImg" 1>&2	
-	# count number of extracted images
-	nbImg=`ls -1 "$curOrigImg"* | wc -l`
+	# count number of extracted images 
+	nbImg=$((`ls -1 "$curOrigImg"* 2>/dev/null | wc -l`))
 	if [ $nbImg -ne "1" ]; then
 		[ $VERBOSITY -ge $LOG_WARN ] && echo "Page $page: Expecting exactly 1 image covering the whole page (found $nbImg). Cannot compute dpi value."
 		return 2
 	fi
 	# Get characteristics of the extracted image
-	curImg=`ls -1 "$curOrigImg"*`
-	propCurImg=`identify -format "%w %h %[colorspace]" "$curImg"`
+	curImg=`ls -1 "$curOrigImg"* 2>/dev/null`
+	propCurImg=`identify -format "%w %h %[colorspace] %[depth]" "$curImg"`
 	widthCurImg=`echo "$propCurImg" | cut -f1 -d" "`
 	heightCurImg=`echo "$propCurImg" | cut -f2 -d" "`
 	colorspaceCurImg=`echo "$propCurImg" | cut -f3 -d" "`
+	depthCurImg=`echo "$propCurImg" | cut -f4 -d" "`
 	[ $VERBOSITY -ge $LOG_DEBUG ] && echo "Page $page: Size ${heightCurImg}x${widthCurImg} (in pixel)"	
 
 	# compute the resolution of the image (making the assumption that x & y resolution are equal)
@@ -87,6 +89,7 @@ getImgInfo() {
 	# save the image characteristics
 	echo "DPI=$dpi" > "$curImgInfo"
 	echo "COLOR_SPACE=$colorspaceCurImg" >> "$curImgInfo"
+	echo "DEPTH=$depthCurImg" >> "$curImgInfo"
 	
 	return 0
 }
@@ -109,24 +112,26 @@ curImgInfo="$TMP_FLD/${page}.orig-img-info.txt"			# Detected characteristics of
 
 
 # auto-detect the characteristics of the embedded image
+depthCurImg="8"			# default color depth
+colorspaceCurImg="sRGB"		# default color space
+dpi=$DEFAULT_DPI		# default resolution
+
 getImgInfo "$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
+if ([ "$ret_code" -eq "1" ] && [ "$FORCE_OCR" -eq "0" ]); then
 	echo "Page $page: Exiting... (Use the -f option to force OCRing, even though fonts are available in the input file)" && exit $EXIT_BAD_INPUT_FILE
-elif [ "$ret_code" -eq "1" -a "$FORCE_OCR" -eq "1" ]; then
-	colorspaceCurImg="sRGB"
-	dpi=$DEFAULT_DPI
+elif ([ "$ret_code" -eq "1" ] && [ "$FORCE_OCR" -eq "1" ]); then
 	[ $VERBOSITY -ge $LOG_WARN ] && echo "Page $page: OCRing anyway, assuming a default resolution of $dpi dpi"
 # in case the page contains more than one image, warn the user but go on with default parameters
 elif [ "$ret_code" -eq "2" ]; then
-	colorspaceCurImg="sRGB"
-	dpi=$DEFAULT_DPI
 	[ $VERBOSITY -ge $LOG_WARN ] && echo "Page $page: Continuing anyway, assuming a default resolution of $dpi dpi"
 else
 	# read the image characteristics from the file
 	dpi=`cat "$curImgInfo" | grep "^DPI=" | cut -f2 -d"="`
 	colorspaceCurImg=`cat "$curImgInfo" | grep "^COLOR_SPACE=" | cut -f2 -d"="`
+	depthCurImg=`cat "$curImgInfo" | grep "^DEPTH=" | cut -f2 -d"="`
 fi
 
 # perform oversampling if the resolution is not sufficient to get good OCR results
@@ -137,10 +142,13 @@ elif [ "$dpi" -lt "200" ]; then
 	[ $VERBOSITY -ge $LOG_WARN ] && echo "Page $page: Low image resolution detected ($dpi dpi). If needed, please use the \"-o\" to try to get better OCR results." 
 fi
 	
-# Identify if page image should be saved as ppm (color) or pgm (gray)
-ext="ppm"
-opt=""
-if [ "$colorspaceCurImg" = "Gray" ]; then
+# Identify if page image should be saved as ppm (color), pgm (gray) or pbm (b&w)
+ext="ppm" 	# by default (color image) the extension of the extracted image is ppm
+opt=""		# by default (color image) no option as to be passed to pdftoppm
+if [ "$colorspaceCurImg" = "Gray" ] && [ "$depthCurImg" = "1" ]; then		# if monochrome (b&w)
+	ext="pbm"
+	opt="-mono"
+elif [ "$colorspaceCurImg" = "Gray" ]; then					# if gray
 	ext="pgm"
 	opt="-gray"
 fi
@@ -201,7 +209,7 @@ 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 [ $KEEP_TMP -eq 0 ]; then
-	rm -f "$curOrigImg"*.*
+	rm -f "$curOrigImg"*
 	rm -f "$curHocr"
 	rm -f "$curImgPixmap"
 	rm -f "$curImgPixmapDeskewed"
@@ -209,4 +217,4 @@ if [ $KEEP_TMP -eq 0 ]; then
 	rm -f "$curImgInfo"
 fi
 
-exit 0
+exit 0

tess-cfg/README.md

@@ -1,12 +0,0 @@
-##############################################################################
-# Readme
-#
-# Copyright (c) 2013-14: fritz-hh from Github (https://github.com/fritz-hh)
-##############################################################################
-
-The file(s) located in this folder are tesseract configuration files.
-(Tesseract configuration files enable to tune the behaviour of tesseract)
-
-If needed, these files should be copied to the "tessdata/configs" folder of your tesseract installation.
-
-To request OCRmyPDF.sh to use a configuration file, please use the -C option

tess-cfg/no_ligature

@@ -1,9 +0,0 @@
-##############################################################################
-# tesseract config file provided for OCRmyPDF
-#
-# prevents tesseract to detect ligatures, as ligatures are not displayed correctly in the final PDF file
-# but are replaced by a black square
-#
-# Copyright (c) 2013-14: fritz-hh from Github (https://github.com/fritz-hh)
-##############################################################################
-tessedit_char_blacklist fifl