Add ocrad 0.5, OCR program implemented as filter.

GNU Ocrad is an OCR (Optical Character Recognition) program implemented
as a filter and based on a feature extraction method. It reads a bitmap
image in pbm format and outputs text in ISO-8859-1 (Latin-1) charset.
Also includes a layout analyser able to separate the columns or blocks
of text normally found on printed pages.
It can be used as a stand-alone console application, or as a backend to
other programs.

PR:		59232
Submitted by:	Ulrich Spoerlein <q@uni.de>

11 files changed, 350 insertions, 0 deletions

diff --git a/graphics/ocrad/Makefile b/graphics/ocrad/Makefile
new file mode 100644
index 000000000000..29d747d48b03
--- /dev/null
+++ b/graphics/ocrad/Makefile
@@ -0,0 +1,33 @@
+# New ports collection makefile for:	graphics/ocrad
+# Date created:				12.11.2003
+# Whom:					Ulrich Spoerlein <q@uni.de>
+#
+# $FreeBSD$
+#
+
+PORTNAME=	ocrad
+PORTVERSION=	0.5
+CATEGORIES=	graphics
+MASTER_SITES=	${MASTER_SITE_GNU}
+MASTER_SITE_SUBDIR=	${PORTNAME}
+
+MAINTAINER=	q@uni.de
+COMMENT=	OCR program implemented as filter
+
+USE_BZIP2=	yes
+HAS_CONFIGURE=	yes
+USE_GETOPT_LONG=	yes
+
+CONFIGURE_ARGS=	--prefix=${PREFIX}
+MAKE_ENV=	CPPFLAGS="${CPPFLAGS}" LDFLAGS="${LDFLAGS}"
+
+DOCS=		AUTHORS ChangeLog NEWS README TODO
+INFO=		ocrad
+
+.if !defined(NOPORTDOCS)
+post-install:
+	@${MKDIR} ${DOCSDIR}
+	cd ${WRKSRC} && ${INSTALL_DATA} ${DOCS} ${DOCSDIR}
+.endif
+
+.include <bsd.port.mk>
diff --git a/graphics/ocrad/distinfo b/graphics/ocrad/distinfo
new file mode 100644
index 000000000000..f91e66017830
--- /dev/null
+++ b/graphics/ocrad/distinfo
@@ -0,0 +1 @@
+MD5 (ocrad-0.5.tar.bz2) = 75bdfda680ddeede5dafa523a16c7191
diff --git a/graphics/ocrad/files/patch-Makefile.in b/graphics/ocrad/files/patch-Makefile.in
new file mode 100644
index 000000000000..56d780e5a694
--- /dev/null
+++ b/graphics/ocrad/files/patch-Makefile.in
@@ -0,0 +1,45 @@
+--- Makefile.in.orig	Sat Oct 18 01:29:16 2003
++++ Makefile.in	Sun Nov 16 18:18:58 2003
+@@ -4,13 +4,14 @@
+ 
+ DISTNAME = ocrad-0.5
+ 
+-CXX = g++
+-INSTALL = install
+-INSTALL_PROGRAM = $(INSTALL)
+-INSTALL_DATA = $(INSTALL) -m 644
+-SHELL = /bin/sh
+-CXXFLAGS = -Wall -W -O2
+-LDFLAGS =
++CXX?= g++
++INSTALL?= install
++INSTALL_PROGRAM?= $(INSTALL)
++INSTALL_DATA?= $(INSTALL) -m 644
++SHELL?= /bin/sh
++CXXFLAGS?= -Wall -W -O2
++CPPFLAGS?=
++LDFLAGS?=
+ 
+ objs = common.o rectangle.o iso_8859_1.o bitmap.o block.o blockmap.o \
+        profile.o feats.o character.o recognize1.o \
+@@ -22,10 +23,10 @@
+ all : ocrad
+ 
+ ocrad : $(objs)
+-	$(CXX) $(LDFLAGS) $(CXXFLAGS) -o ocrad $(objs)
++	$(CXX) $(LDFLAGS) $(CXXFLAGS) $(CPPFLAGS) -o ocrad $(objs)
+ 
+ ocradp : $(objs)
+-	$(CXX) $(LDFLAGS) $(CXXFLAGS) -pg -o ocradp $(objs)
++	$(CXX) $(LDFLAGS) $(CXXFLAGS) $(CPPFLAGS) -pg -o ocradp $(objs)
+ 
+ %.o : %.cc
+ 	$(CXX) $(CXXFLAGS) -c -o $@ $<
+@@ -42,6 +43,7 @@
+ textline.o   : block.h character.h iso_8859_1.h textline.h
+ textblock.o  : block.h character.h textline.h textblock.h
+ main.o       : block.h blockmap.h bitmap.h character.h textline.h textblock.h
++	$(CXX) $(CXXFLAGS) $(CPPFLAGS) -c main.cc
+ 
+ 
+ install : all install-info
diff --git a/graphics/ocrad/files/patch-bitmap.cc b/graphics/ocrad/files/patch-bitmap.cc
new file mode 100644
index 000000000000..4768997c17b1
--- /dev/null
+++ b/graphics/ocrad/files/patch-bitmap.cc
@@ -0,0 +1,34 @@
+--- bitmap.cc.orig	Sun Nov 16 17:28:14 2003
++++ bitmap.cc	Sun Nov 16 17:33:30 2003
+@@ -24,12 +24,11 @@
+ #include "rectangle.h"
+ #include "bitmap.h"
+ 
+-
+ namespace {
+ 
+ char pbm_getrawbyte( FILE * f ) throw( Bitmap::Error )
+   {
+-  int ch = std::getc( f );
++  int ch = getc( f );
+ 
+   if( ch == EOF )
+     throw Bitmap::Error( "end-of-file reading pbm file.\n" );
+@@ -58,7 +57,7 @@
+   char ch;
+   int i = 0;
+ 
+-  do ch = pbm_getc( f ); while( std::isspace( ch ) );
++  do ch = pbm_getc( f ); while( isspace( ch ) );
+   if( !std::isdigit( ch ) )
+     throw Bitmap::Error( "junk in pbm file where an integer should be.\n" );
+   do { i = (i * 10) + (ch - '0'); ch = pbm_getc( f ); }
+@@ -71,7 +70,7 @@
+   {
+   char ch;
+ 
+-  do ch = pbm_getc( f ); while( std::isspace( ch ) );
++  do ch = pbm_getc( f ); while( isspace( ch ) );
+ 
+   if( ch == '0' ) return false;
+   if( ch == '1' ) return true;
diff --git a/graphics/ocrad/files/patch-character.h b/graphics/ocrad/files/patch-character.h
new file mode 100644
index 000000000000..5dacbaf8555a
--- /dev/null
+++ b/graphics/ocrad/files/patch-character.h
@@ -0,0 +1,10 @@
+--- character.h.orig	Sun Nov 16 17:40:28 2003
++++ character.h	Sun Nov 16 17:40:44 2003
+@@ -64,6 +64,7 @@
+   void swap_guesses( int i, int j ) throw();
+   const Guess & guess( int i ) const throw();
+   int guesses() const throw() { return _guess.size(); }
++#undef isalnum
+   bool isalnum() const throw();
+ 
+   void join( Character & c ) throw();
diff --git a/graphics/ocrad/files/patch-configure b/graphics/ocrad/files/patch-configure
new file mode 100644
index 000000000000..802fa316e8ee
--- /dev/null
+++ b/graphics/ocrad/files/patch-configure
@@ -0,0 +1,29 @@
+--- configure.orig	Mon Aug 25 00:07:09 2003
++++ configure	Wed Nov 12 22:50:24 2003
+@@ -13,7 +13,7 @@
+ while true ; do
+ 
+ 	# Break out if there are no more args
+-	if [ $# == 0 ]; then break ; fi
++	if [ $# = 0 ]; then break ; fi
+ 
+ 	# Get the first arg, and shuffle
+ 	option=$1
+@@ -48,7 +48,7 @@
+ 
+ # Find the source files, if location was not specified.
+ srcdirtext=
+-if [ x${srcdir} == x ]; then
++if [ x${srcdir} = x ]; then
+ 	srcdirtext="or . or .." ; srcdir=.
+ 	if [ ! -r ${srcdir}/${srctrigger} ] ; then srcdir=.. ; fi
+ 	if [ ! -r ${srcdir}/${srctrigger} ] ; then
+@@ -65,7 +65,7 @@
+ fi
+ 
+ # Set srcdir to . if that's what it is.
+-if [ $(pwd) == $(cd ${srcdir} ; pwd) ] ; then srcdir=. ; fi
++if [ $(pwd) = $(cd ${srcdir} ; pwd) ] ; then srcdir=. ; fi
+ 
+ # write variables to config file.
+ rm -f Makefile
diff --git a/graphics/ocrad/files/patch-iso_8859_1.h b/graphics/ocrad/files/patch-iso_8859_1.h
new file mode 100644
index 000000000000..9d07257db560
--- /dev/null
+++ b/graphics/ocrad/files/patch-iso_8859_1.h
@@ -0,0 +1,14 @@
+--- iso_8859_1.h.orig	Sun Nov 16 17:14:19 2003
++++ iso_8859_1.h	Sun Nov 16 17:33:51 2003
+@@ -100,6 +100,11 @@
+   static unsigned char base_letter( unsigned char ch ) throw();
+   static unsigned char compose( unsigned char base_letter,
+                                 unsigned char accent ) throw();
++#undef isalnum
++#undef isalpha
++#undef isupper
++#undef isvowel
++#undef toupper
+   static bool isalnum( unsigned char ch ) throw();
+   static bool isalpha( unsigned char ch ) throw();
+   static bool isupper( unsigned char ch ) throw();
diff --git a/graphics/ocrad/files/patch-main.cc b/graphics/ocrad/files/patch-main.cc
new file mode 100644
index 000000000000..e0497bf2e2f6
--- /dev/null
+++ b/graphics/ocrad/files/patch-main.cc
@@ -0,0 +1,21 @@
+--- main.cc.orig	Sat Oct 18 01:27:29 2003
++++ main.cc	Sun Nov 16 18:16:41 2003
+@@ -28,6 +28,7 @@
+ #include <cstring>
+ #include <vector>
+ #include <getopt.h>
++#include <libgen.h>
+ #include "common.h"
+ #include "rectangle.h"
+ #include "bitmap.h"
+@@ -289,8 +290,8 @@
+     {
+     if( infile == stdin )
+       {
+-      std::ungetc( std::getc( infile ), infile );
+-      if( std::feof( infile ) || std::ferror( infile ) ) infile = 0;
++      ungetc( getc( infile ), infile );
++      if( feof( infile ) || ferror( infile ) ) infile = 0;
+       }
+     while( infile != stdin )
+       {
diff --git a/graphics/ocrad/files/patch-textline.cc b/graphics/ocrad/files/patch-textline.cc
new file mode 100644
index 000000000000..804a281ab3fa
--- /dev/null
+++ b/graphics/ocrad/files/patch-textline.cc
@@ -0,0 +1,146 @@
+--- textline.cc.orig	Sun Nov 16 17:39:59 2003
++++ textline.cc	Sun Nov 16 17:44:01 2003
+@@ -33,7 +33,7 @@
+   for( ; end < characters(); ++end )
+     {
+     Character & c = character( end );
+-    if( c.guesses() && std::isspace( c.guess(0).ch ) ) break;
++    if( c.guesses() && isspace( c.guess(0).ch ) ) break;
+     }
+   return end;
+   }
+@@ -264,11 +264,11 @@
+     if( c1.guesses() == 1 )
+       {
+       unsigned char ch = c1.guess( 0 ).ch;
+-      if( std::isspace( ch ) ) { begin = i + 1 ; continue; }
++      if( isspace( ch ) ) { begin = i + 1 ; continue; }
+       if( ch != 'c' && ch != 'o' && ch != 's' && ch != 'u' && ch != 'v' &&
+           ch != 'w' && ch != 'x' && ch != 'z' ) continue;
+       if( 4 * c1.height() > 5 * mean_height() )
+-        { c1.only_guess( std::toupper( ch ), 0 ); continue; }
++        { c1.only_guess( toupper( ch ), 0 ); continue; }
+       if( 5 * c1.height() < 4 * mean_height() ) continue;
+       for( int j = begin; j < characters(); ++j ) if( j != i )
+         {
+@@ -276,12 +276,12 @@
+         if( c2.guesses() >= 1 )
+           {
+           unsigned char ch2 = c2.guess( 0 ).ch;
+-          if( std::isspace( ch2 ) ) break;
+-          if( ( std::isalpha( ch2 ) && 4 * c1.height() > 5 * c2.height() ) ||
+-              ( ( std::isupper( ch2 ) || ch2 == 'l' ) && ch2 != 'B' &&
++          if( isspace( ch2 ) ) break;
++          if( ( isalpha( ch2 ) && 4 * c1.height() > 5 * c2.height() ) ||
++              ( ( isupper( ch2 ) || ch2 == 'l' ) && ch2 != 'B' &&
+                 ( c1.height() >= c2.height() ||
+                   Ocrad::similar( c1.height(), c2.height(), 10 ) ) ) )
+-            { c1.insert_guess( 0, std::toupper( ch ), 1 ); break; }
++            { c1.insert_guess( 0, toupper( ch ), 1 ); break; }
+           }
+         }
+       }
+@@ -310,7 +310,7 @@
+     if( c1.guesses() >= 1 )
+       {
+       unsigned char ch = c1.guess( 0 ).ch;
+-      if( std::isspace( ch ) ) { begin = i + 1 ; continue; }
++      if( isspace( ch ) ) { begin = i + 1 ; continue; }
+       if( ch <= 127 || c1.block_vector().size() < 2 ) continue;
+       int chb = ISO_8859_1::base_letter( ch );
+       if( chb != 'o' && chb != 'u' ) continue;
+@@ -323,11 +323,11 @@
+           unsigned char ch2 = c2.guess( 0 ).ch;
+           int ch2b = ISO_8859_1::base_letter( ch2 );
+           if( !ch2b && ch2 > 127 ) continue;
+-          if( std::isspace( ch2 ) ) break;
+-          if( ( std::isalpha( ch2 ) && 4 * b1.height() > 5 * c2.height() ) ||
+-              ( std::isupper( ch2 ) && Ocrad::similar( b1.height(), c2.height(), 10 ) ) ||
+-              ( std::isalpha( ch2b ) && 4 * c1.height() > 5 * c2.height() ) ||
+-              ( std::isupper( ch2b ) && Ocrad::similar( c1.height(), c2.height(), 10 ) ) )
++          if( isspace( ch2 ) ) break;
++          if( ( isalpha( ch2 ) && 4 * b1.height() > 5 * c2.height() ) ||
++              ( isupper( ch2 ) && Ocrad::similar( b1.height(), c2.height(), 10 ) ) ||
++              ( isalpha( ch2b ) && 4 * c1.height() > 5 * c2.height() ) ||
++              ( isupper( ch2b ) && Ocrad::similar( c1.height(), c2.height(), 10 ) ) )
+             { c1.insert_guess( 0, ISO_8859_1::toupper( ch ), 1 ); break; }
+           }
+         }
+@@ -341,7 +341,7 @@
+     if( c1.guesses() >= 1 )
+       {
+       unsigned char ch = c1.guess( 0 ).ch;
+-      if( std::isspace( ch ) ) { begin = i + 1 ; continue; }
++      if( isspace( ch ) ) { begin = i + 1 ; continue; }
+       if( ch != 'o' && ch != 'O' && ch != 'l' ) continue;
+       for( int j = begin; j < characters(); ++j ) if( j != i )
+         {
+@@ -349,8 +349,8 @@
+         if( c2.guesses() >= 1 )
+           {
+           unsigned char ch2 = c2.guess( 0 ).ch;
+-          if( std::isspace( ch2 ) ) break;
+-          if( std::isdigit( ch2 ) )
++          if( isspace( ch2 ) ) break;
++          if( isdigit( ch2 ) )
+             {
+             if( Ocrad::similar( c1.height(), c2.height(), 10 ) )
+               c1.insert_guess( 0, (ch == 'l') ? '1' : '0', c1.guess(0).value + 1 );
+@@ -372,7 +372,7 @@
+       Character & c2 = character( i );
+       if( !c2.guesses() ) continue;
+       unsigned char ch = c2.guess( 0 ).ch;
+-      if( !std::isalnum( ch ) && ch != '.' && ch != '|' ) continue;
++      if( !isalnum( ch ) && ch != '.' && ch != '|' ) continue;
+       switch( ch )
+         {
+         case 'g': case 'j': case 'p': case 'q': case 'y':
+@@ -426,14 +426,14 @@
+       if( i < characters() - 1 && character( i + 1 ).guesses() )
+         rch = character( i + 1 ).guess( 0 ).ch;
+       if( ISO_8859_1::isupper( rch ) &&
+-          ( !lch || ISO_8859_1::isupper( lch ) || std::isspace( lch ) ) )
++          ( !lch || ISO_8859_1::isupper( lch ) || isspace( lch ) ) )
+         { c.insert_guess( 0, 'I', 1 ); continue; }
+       if( ch == 'l' ) continue;
+       if( ISO_8859_1::isalpha( lch ) || ISO_8859_1::isalpha( rch ) )
+         { c.insert_guess( 0, 'l', 1 ); continue; }
+-      if( rch == '|' && ( !lch || std::isspace( lch ) ) &&
++      if( rch == '|' && ( !lch || isspace( lch ) ) &&
+           i < characters() - 2 && character( i + 2 ).guesses() &&
+-          std::isalpha( character( i + 2 ).guess( 0 ).ch ) )
++          isalpha( character( i + 2 ).guess( 0 ).ch ) )
+         { c.insert_guess( 0, 'l', 1 ); continue; }
+       }
+     }
+@@ -475,7 +475,7 @@
+     if( c.guesses() )
+       {
+       unsigned char ch = c.guess( 0 ).ch;
+-      if( std::isspace( ch ) ) { begin = i + 1 ; continue; }
++      if( isspace( ch ) ) { begin = i + 1 ; continue; }
+       if( i == begin && ch == 'a' && c.guesses() == 2 &&
+           c.guess( 1 ).ch == 'Q' && 4 * c.height() > 5 * mean_height() )
+         c.swap_guesses( 0, 1 );
+@@ -501,7 +501,7 @@
+     if( c1.guesses() )
+       {
+       unsigned char ch = c1.guess(0).ch;
+-      if( std::isspace( ch ) ) { begin = i + 1 ; continue; }
++      if( isspace( ch ) ) { begin = i + 1 ; continue; }
+       if( c1.guesses() != 2 || ch != 'B' || c1.guess(1).ch != 'a' ) continue;
+       if( 4 * c1.height() > 5 * mean_height() ) continue;
+       for( int j = begin; j < characters(); ++j ) if( j != i )
+@@ -510,9 +510,9 @@
+         if( c2.guesses() >= 1 )
+           {
+           unsigned char ch2 = c2.guess(0).ch;
+-          if( std::isspace( ch2 ) ) break;
+-          if( ( std::isalpha( ch2 ) && 5 * c1.height() < 4 * c2.height() ) ||
+-              ( std::islower( ch2 ) &&
++          if( isspace( ch2 ) ) break;
++          if( ( isalpha( ch2 ) && 5 * c1.height() < 4 * c2.height() ) ||
++              ( islower( ch2 ) &&
+                 ( c1.height() <= c2.height() ||
+                   Ocrad::similar( c1.height(), c2.height(), 10 ) ) ) )
+             { c1.swap_guesses( 0, 1 ); break; }
diff --git a/graphics/ocrad/pkg-descr b/graphics/ocrad/pkg-descr
new file mode 100644
index 000000000000..844adfc1ce83
--- /dev/null
+++ b/graphics/ocrad/pkg-descr
@@ -0,0 +1,9 @@
+GNU Ocrad is an OCR (Optical Character Recognition) program implemented
+as a filter and based on a feature extraction method. It reads a bitmap
+image in pbm format and outputs text in ISO-8859-1 (Latin-1) charset.
+Also includes a layout analyser able to separate the columns or blocks
+of text normally found on printed pages.
+It can be used as a stand-alone console application, or as a backend to
+other programs.
+
+WWW: http://www.gnu.org/software/ocrad/ocrad.html
diff --git a/graphics/ocrad/pkg-plist b/graphics/ocrad/pkg-plist
new file mode 100644
index 000000000000..c3269db7f09f
--- /dev/null
+++ b/graphics/ocrad/pkg-plist
@@ -0,0 +1,8 @@
+@comment $FreeBSD$
+bin/ocrad
+%%PORTDOCS%%%%DOCSDIR%%/AUTHORS
+%%PORTDOCS%%%%DOCSDIR%%/ChangeLog
+%%PORTDOCS%%%%DOCSDIR%%/NEWS
+%%PORTDOCS%%%%DOCSDIR%%/README
+%%PORTDOCS%%%%DOCSDIR%%/TODO
+%%PORTDOCS%%@dirrm %%DOCSDIR%%