###########################################################################
#
# PagedImgPlug.pm -- plugin for sets of images and OCR text that
#  make up a document
# A component of the Greenstone digital library software
# from the New Zealand Digital Library Project at the 
# University of Waikato, New Zealand.
#
# Copyright (C) 1999 New Zealand Digital Library Project
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
#
###########################################################################

# PagedImgPlug
# processes sequences of images, with optional OCR text
#
# This plugin takes *.item files, which contain metadata and lists of image 
# files, and produces a document containing sections, one for each page.
# The files should be named something.item, then you can have more than one 
# book in a directory. You will need to create these files, one for each
# document/book.
#
#There are two formats for the item files: a plain text format, and an xml 
#format. You can use either format, and can have both formats in the same 
#collection if you like. If you use the plain format, you must not start the 
#file off with <PagedDocument>

#### PLAIN FORMAT
# The format of the xxx.item file is as follows:
# The first lines contain any metadata for the whole document
# <metadata-name>metadata-value
# eg.
# <Title>Snail farming
# <Date>19230102
# Then comes a list of pages, one page per line, each line has the format
#
# pagenum:imagefile:textfile:r
#
# page num and imagefile are required. pagenum is used for the Title 
# of the section, and in the display is shown as page <pagenum>. 
# imagefile is the image for the page. textfile is an optional text 
# file containing the OCR (or any) text for the page - this gets added 
# as the text for the section. r is optional, and signals that the image 
# should be rotated 180deg. Eg use this if the image has been made upside down.
# So an example item file looks like:
# <Title>Snail farming
# <Date>19960403
# 1:p1.gif:p1.txt:
# 2:p2.gif::
# 3:p3.gif:p3.txt:
# 3b:p3b.gif:p3b.txt:r
# The second page has no text, the fourth page is a back page, and 
# should be rotated.
# 

#### XML FORMAT
# The xml format looks like the following
#<PagedDocument>
#<Metadata name="Title">The Title of the entire document</Metadata>
#<Page pagenum="1" imgfile="xxx.jpg" txtfile="yyy.jpg">
#<Metadata name="Title">The Title of this page</Metadata>
#</Page>
#... more pages
#</PagedDocument>
#PagedDocument contains a list of Pages, Metadata and PageGroups. Any metadata 
#that is not inside another tag will belong to the document.
#Each Page has a pagenum (not used at the moment), an imgfile and/or a txtfile.
#These are both optional - if neither is used, the section will have no content.
#Pages can also have metadata associated with them.
#PageGroups can be introduced at any point - they can contain Metadata and Pages and other PageGroups. They are used to introduce hierarchical structure into the document.
#For example
#<PagedDocument>
#<PageGroup>
#<Page>
#<Page>
#</PageGroup>
#<Page>
#</PagedDocument>
#would generate a structure like
#X
#--X
#  --X
#  --X
#--X
#PageGroup tags can also have imgfile/textfile metadata if you like - this way they get some content themselves.

#Currently the XML structure doesn't work very well with the paged document type, unless you use numerical Titles for each section.
#There is still a bit of work to do on this format:
#* enable other text file types, eg html, pdf etc
#* make the document paging work properly
#* add pagenum as Title unless a Title is present?

# All the supplemetary image amd text files should be in the same folder as 
# the .item file.
#
# To display the images instead of the document text, you can use [srcicon] 
# in the DocumentText format statement.
# For example, 
#
# format DocumentText "<center><table width=_pagewidth_><tr><td>[srcicon]</td></tr></table></center>"
#
# To have it create thumbnail size images, use the '-thumbnail' option.
# To have it create medium size images for display, use the '-screenview'
# option. As usual, running 
# 'perl -S pluginfo.pl PagedImgPlug' will list all the options.

# If you want the resulting documents to be presented with a table of 
# contents, use '-documenttype hierarchy', otherwise they will have 
# next and previous arrows, and a goto page X box. 

# If you have used -screenview, you can also use [screenicon] in the format
# statement to display the smaller image.  Here is an example  that switches 
# between the two: 
#
# format DocumentText "<center><table width=_pagewidth_><tr><td>{If}{_cgiargp_ eq full,<a href='_httpdocument_&d=_cgiargd_&p=small'>Switch to small version.</a>,<a href='_httpdocument_&d=_cgiargd_&p=full'>Switch to fullsize version</a>}</td></tr><tr><td>{If}{_cgiargp_ eq full,<a href='_httpdocument_&d=_cgiargd_&p=small' title='Switch to small version'>[srcicon]</a>,<a href='_httpdocument_&d=_cgiargd_&p=full' title='Switch to fullsize version'>[screenicon]</a>}</td></tr></table></center>"
#
# Additional metadata can be added into the .item files, alternatively you can 
# use normal metadata.xml files, with the name of the xxx.item file as the 
# FileName (only for document level metadata).

package PagedImgPlug;

use XMLPlug;
use strict;
no strict 'refs'; # allow filehandles to be variables and viceversa

sub BEGIN {
    @PagedImgPlug::ISA = ('XMLPlug');
}

my $type_list =
    [ { 'name' => "paged",
        'desc' => "{PagedImgPlug.documenttype.paged}" },
      { 'name' => "hierarchy",
        'desc' => "{PagedImgPlug.documenttype.hierarchy}" } ];

my $arguments =
    [ { 'name' => "process_exp",
	'desc' => "{BasPlug.process_exp}",
	'type' => "string",
	'deft' => &get_default_process_exp(),
	'reqd' => "no" },
      { 'name' => "block_exp",
	'desc' => "{BasPlug.block_exp}",
	'type' => "string",
	'deft' => &get_default_block_exp(),
	'reqd' => "no" },
      { 'name' => "title_sub",
	'desc' => "{HTMLPlug.title_sub}",
	'type' => "string", 
	'deft' => "" },
      { 'name' => "noscaleup",
	'desc' => "{ImagePlug.noscaleup}",
	'type' => "flag",
	'reqd' => "no" },
      { 'name' => "thumbnail",
	'desc' => "{PagedImgPlug.thumbnail}",
	'type' => "flag",
	'reqd' => "no" },
      { 'name' => "thumbnailsize",
	'desc' => "{ImagePlug.thumbnailsize}",
	'type' => "int",
	'deft' => "100",
	'range' => "1,",
	'reqd' => "no" },
      { 'name' => "thumbnailtype",
	'desc' => "{ImagePlug.thumbnailtype}",
	'type' => "string",
	'deft' => "gif",
	'reqd' => "no" },
      { 'name' => "screenview",
	'desc' => "{PagedImgPlug.screenview}",
	'type' => "flag",
	'reqd' => "no" },
      { 'name' => "screenviewsize",
	'desc' => "{PagedImgPlug.screenviewsize}",
	'type' => "int",
        'deft' => "500",
        'range' => "1,",
	'reqd' => "no" },
      { 'name' => "screenviewtype",
	'desc' => "{PagedImgPlug.screenviewtype}",
	'type' => "string",
	'deft' => "jpg",
	'reqd' => "no" },
      { 'name' => "converttotype",
	'desc' => "{ImagePlug.converttotype}",
	'type' => "string",
	'deft' => "",
	'reqd' => "no" },
      { 'name' => "minimumsize",
	'desc' => "{ImagePlug.minimumsize}",
	'type' => "int",
	'deft' => "100",
	'range' => "1,",
	'reqd' => "no" },
      { 'name' => "headerpage",
	'desc' => "{PagedImgPlug.headerpage}",
	'type' => "flag",
	'reqd' => "no" },
      { 'name' => "documenttype",
	'desc' => "{PagedImgPlug.documenttype}",
	'type' => "enum",
	'list' => $type_list,
	'deft' => "paged",
	'reqd' => "no" } ];


my $options = { 'name'     => "PagedImgPlug",
		'desc'     => "{PagedImgPlug.desc}",
		'abstract' => "no",
		'inherits' => "yes",
		'args'     => $arguments };

sub new {
    my ($class) = shift (@_);
    my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
    push(@$pluginlist, $class);

    if(defined $arguments){ push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});}
    if(defined $options) { push(@{$hashArgOptLists->{"OptList"}},$options)};
    
    my $self = new XMLPlug($pluginlist, $inputargs, $hashArgOptLists);
    
    return bless $self, $class;
}

sub get_default_process_exp {
    my $self = shift (@_);

    return q^\.item$^;
}

sub get_doctype {
    my $self = shift(@_);
    
    return "PagedDocument";
}


# want to block everything except the .item ones 
# but instead we will block images and txt files
sub get_default_block_exp {
    my $self = shift (@_);
    
    return q^(?i)(\.jpe?g|\.gif|\.png|\.tif?f|\.te?xt|~)$^
}

# Create the thumbnail and screenview images, and discover the Image's
# size, width, and height using the convert utility.
sub process_image {
    my $self = shift (@_);
    my $filename = shift (@_); # filename with full path
    my $srcfile = shift (@_); # filename without path
    my $doc_obj = shift (@_);
    my $section = shift (@_); #the current section
    my $rotate = shift (@_); # whether to rotate the image or not
    $rotate = 0 unless defined $rotate;
    
    # check that the image file exists!!
    if (!-f $filename) {
	print "PagedImgPlug: ERROR: File $filename does not exist, skipping\n";
	return 0;
    }

    my $top=0;
    if ($section eq $doc_obj->get_top_section()) {
	$top=1;
    }
    my $verbosity = $self->{'verbosity'};
    my $outhandle = $self->{'outhandle'};

    # check the filename is okay
    return 0 if ($srcfile eq "" || $filename eq "");
    
    my $minimumsize = $self->{'minimumsize'};
    if (defined $minimumsize && (-s $filename < $minimumsize)) {
        print $outhandle "PagedImgPlug: \"$filename\" too small, skipping\n"
	    if ($verbosity > 1);
    }

    # Convert the image to a new type (if required), and rotate if required.
    my $converttotype = $self->{'converttotype'};
    my $originalfilename = "";  # only set if we do a conversion
    my $type = "unknown";
    my $converted = 0;
    my $rotated=0;

    if ($converttotype ne "" && $filename !~ /$converttotype$/) {
	$converted=1;
	$originalfilename = $filename;
	my $filehead = &util::get_tmp_filename();
	$filename = $filehead . ".$converttotype";
	my $n = 1;
	while (-e $filename) {
	    $filename = "$filehead$n\.$converttotype";
	    $n++;
	}
	$self->{'tmp_filename1'} = $filename;

	my $rotate_option = "";
	if ($rotate eq "r") {
	    $rotate_option = "-rotate 180 ";
	}
	
	my $command = "convert -verbose \"$originalfilename\" $rotate_option \"$filename\"";
	print $outhandle "CONVERT: $command\n" if ($verbosity > 2);
	my $result = '';
	$result = `$command`;
	print $outhandle "CONVERT RESULT = $result\n" if ($verbosity > 2);

	$type = $converttotype;
    } elsif ($rotate eq "r") {
	$rotated=1;
	$originalfilename = $filename;
	$filename  = &util::get_tmp_filename();
	
	my $command = "convert \"$originalfilename\" -rotate 180 \"$filename\"";
	print $outhandle "ROTATE: $command\n" if ($verbosity > 2);
	my $result = '';
	$result = `$command`;
	print $outhandle "ROTATE RESULT = $result\n" if ($verbosity > 2);
	
    } 
    

    # Add the image metadata
    my $file; # the new file name
    my $id = $srcfile;
    $id =~ s/\.([^\.]*)$//;  # the new file name without an extension
    if ($converted) {
	# we have converted the image
	# add on the new extension
	$file .= "$id.$converttotype";
    } else {
	$file = $srcfile;
    }
    
    my $url =$file; # the new file name prepared for a url
    my $srcurl = $srcfile;
    $url =~ s/ /%20/g;
    $srcurl =~ s/ /%20/g;
    
    $doc_obj->add_metadata ($section, "Image", $url);

    # Also want to set filename as 'Source' metadata to be
    # consistent with other plugins
    $doc_obj->add_metadata ($section, "Source", $srcurl);

    my ($image_type, $image_width, $image_height, $image_size) 
	= &identify($filename, $outhandle, $verbosity);

    $doc_obj->add_metadata ($section, "ImageType",   $image_type);
    $doc_obj->add_metadata ($section, "ImageWidth",  $image_width);
    $doc_obj->add_metadata ($section, "ImageHeight", $image_height);
    $doc_obj->add_metadata ($section, "ImageSize",   $image_size);
    $doc_obj->add_metadata ($section, "FileFormat", "PagedImg");
    
    if ($type eq "unknown" && $image_type) {
	$type = $image_type;
    }

    if ($top) {
	$doc_obj->add_metadata ($section, "srclink", 
				"<a href=\"_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/[Image]\">");
	$doc_obj->add_metadata ($section, "srcicon", "<img src=\"_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/[Image]\">");

    } else {
	$doc_obj->add_metadata ($section, "srclink", 
				"<a href=\"_httpprefix_/collect/[collection]/index/assoc/[parent(Top):assocfilepath]/[Image]\">");
	$doc_obj->add_metadata ($section, "srcicon", "<img src=\"_httpprefix_/collect/[collection]/index/assoc/[parent(Top):assocfilepath]/[Image]\">");

    }
    $doc_obj->add_metadata ($section, "/srclink", "</a>");


    # Add the image as an associated file
    $doc_obj->associate_file($filename,$file,"image/$type",$section);
    print $outhandle "associating file $filename as name $file\n" if ($verbosity > 2);

    if ($self->{'thumbnail'}) {
	# Make the thumbnail image
	my $thumbnailsize = $self->{'thumbnailsize'} || 100;
	my $thumbnailtype = $self->{'thumbnailtype'} || 'gif';
	
	my $filehead = &util::get_tmp_filename();
	my $thumbnailfile = $filehead . ".$thumbnailtype";
	my $n=1;
	while (-e $thumbnailfile) {
	    $thumbnailfile = $filehead . $n . ".$thumbnailtype";
	    $n++;
	}
	
	$self->{'tmp_filename2'} = $thumbnailfile;
	
	# Generate the thumbnail with convert
	my $command = "convert  -verbose -geometry $thumbnailsize" 
	    . "x$thumbnailsize \"$filename\" \"$thumbnailfile\"";
	print $outhandle "THUMBNAIL: $command\n" if ($verbosity > 2);
	my $result = '';
	$result = `$command 2>&1` ;
	print $outhandle "THUMB RESULT: $result\n" if ($verbosity > 2);
	
	# Add the thumbnail as an associated file ...
	if (-e "$thumbnailfile") { 
	    $doc_obj->associate_file("$thumbnailfile", $id."thumb.$thumbnailtype", "image/$thumbnailtype",$section);
	    $doc_obj->add_metadata ($section, "ThumbType", $thumbnailtype);
	    $doc_obj->add_metadata ($section, "Thumb", $id."thumb.$thumbnailtype");
	    if ($top) {
		$doc_obj->add_metadata ($section, "thumbicon", "<img src=\"_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/[Thumb]\" width=[ThumbWidth] height=[ThumbHeight]>");
	    } else {
		$doc_obj->add_metadata ($section, "thumbicon", "<img src=\"_httpprefix_/collect/[collection]/index/assoc/[parent(Top):assocfilepath]/[Thumb]\" width=[ThumbWidth] height=[ThumbHeight]>");
	    }
	}
	
	# Extract Thumnail metadata from convert output
	if ($result =~ m/[0-9]+x[0-9]+=>([0-9]+)x([0-9]+)/) {
	    $doc_obj->add_metadata ($section, "ThumbWidth", $1);
	    $doc_obj->add_metadata ($section, "ThumbHeight", $2);
	}
    }
    # Make a screen-sized version of the picture if requested
    if ($self->{'screenview'}) {
	
	# To do: if the actual image is smaller than the screenview size,
	# we should use the original !

	my $screenviewsize = $self->{'screenviewsize'} || 500;
	my $screenviewtype = $self->{'screenviewtype'} || 'jpeg';
	my $filehead = &util::get_tmp_filename();
	my $screenviewfilename = $filehead . ".$screenviewtype";
	my $n=1;
	while (-e $screenviewfilename) {
	    $screenviewfilename = "$filehead$n\.$screenviewtype";
	    $n++;
	}
	$self->{'tmp_filename3'} = $screenviewfilename;

	# make the screenview image
	my $command = "convert  -verbose -geometry $screenviewsize" 
	    . "x$screenviewsize \"$filename\" \"$screenviewfilename\"";
	print $outhandle "SCREENVIEW: $command\n" if ($verbosity > 2);
	my $result = "";
	$result = `$command  2>&1` ;
	print $outhandle "SCREENVIEW RESULT: $result\n" if ($verbosity > 3);

	# get screenview dimensions, size and type
        if ($result =~ m/[0-9]+x[0-9]+=>([0-9]+)x([0-9]+)/) {
	    $doc_obj->add_metadata ($section, "ScreenWidth", $1);
	    $doc_obj->add_metadata ($section, "ScreenHeight", $2);
	}elsif ($result =~ m/([0-9]+)x([0-9]+)/) {
	    #if the image hasn't changed size,  the previous regex doesn't match
	    $doc_obj->add_metadata ($section, "ScreenWidth", $1);
	    $doc_obj->add_metadata ($section, "ScreenHeight", $2);
	}
    
	#add the screenview as an associated file ...
	if (-e "$screenviewfilename") { 
	    $doc_obj->associate_file("$screenviewfilename", $id."sv.$screenviewtype",
				     "image/$screenviewtype",$section);
	    print $outhandle "associating screen file $screenviewfilename as name $id sv.$screenviewtype\n" if ($verbosity > 2);
	    
	    $doc_obj->add_metadata ($section, "ScreenType", $screenviewtype);
	    $doc_obj->add_metadata ($section, "Screen", $id."sv.$screenviewtype");

	    if ($top) {
		$doc_obj->add_metadata ($section, "screenicon", "<img src=\"_httpprefix_/collect/[collection]/index/assoc/[assocfilepath]/[Screen]\" width=[ScreenWidth] height=[ScreenHeight]>");
	    } else {
		$doc_obj->add_metadata ($section, "screenicon", "<img src=\"_httpprefix_/collect/[collection]/index/assoc/[parent(Top):assocfilepath]/[Screen]\" width=[ScreenWidth] height=[ScreenHeight]>");

	    }
	} else {
	    print $outhandle "PagedImgPlug: couldn't find \"$screenviewfilename\"\n";
	}
    }

    return $type;


}



# Discover the characteristics of an image file with the ImageMagick
# "identify" command.

sub identify { 
    my ($image, $outhandle, $verbosity) = @_;

    # Use the ImageMagick "identify" command to get the file specs
    my $command = "identify \"$image\" 2>&1";
    print $outhandle "$command\n" if ($verbosity > 2);
    my $result = '';
    $result = `$command`;
    print $outhandle "$result\n" if ($verbosity > 3);

    # Read the type, width, and height
    my $type =   'unknown';
    my $width =  'unknown';
    my $height = 'unknown';

    my $image_safe = quotemeta $image;
    if ($result =~ /^$image_safe (\w+) (\d+)x(\d+)/) {
	$type = $1;
	$width = $2;
	$height = $3;
    }
	
    # Read the size
    my $size = "unknown";
    if ($result =~ m/^.* ([0-9]+)b/) {
	$size = $1;
    } elsif ($result =~ m/^.* ([0-9]+)kb/) {
	$size = 1024 * $1;
    }

    print $outhandle "file: $image:\t $type, $width, $height, $size\n" 
	if ($verbosity > 3);

    # Return the specs
    return ($type, $width, $height, $size);
}


# The PagedImgPlug read() function. This function does all the right things
# to make general options work for a given plugin. It calls the process()
# function which does all the work specific to a plugin (like the old
# read functions used to do). Most plugins should define their own 
# process() function and let this read() function keep control.
#
# PagedImgPlug overrides read() because there is no need to read the actual 
# text of the file in, because the contents of the file is not text...
#
# Return number of files processed, undef if can't process
# Note that $base_dir might be "" and that $file might 
# include directories

sub read_into_doc_obj {
    my $self = shift (@_);
    my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
    my $outhandle   = $self->{'outhandle'};
    
    #check process and block exps, smart block, etc
    my ($block_status,$filename) = $self->read_block(@_);    
    return $block_status if ((!defined $block_status) || ($block_status==0));
    
    print $outhandle "PagedImgPlug processing \"$filename\"\n"
	if $self->{'verbosity'} > 1;
    print STDERR "<Processing n='$file' p='PagedImgPlug'>\n" if ($gli);
    
    # here we need to decide if we have an old text .item file, or a new xml 
    # .item file - for now the test is if the first non-empty line is 
    # <PagedDocument> then its xml
    my $xml_version = 0;
    open (ITEMFILE, $filename) || die "couldn't open $filename\n";
    
    my $backup_filename = "backup.item";
    open (BACKUP,">$backup_filename")|| die "couldn't write to $backup_filename\n";
    my $line = "";
    my $num = 0;
    $line = <ITEMFILE>;
    while ($line !~ /\w/) {
	$line = <ITEMFILE>;
    }
    chomp $line;
    if ($line =~ /<PagedDocument/) {
	$xml_version = 1;
    }
    close ITEMFILE;
    open (ITEMFILE, $filename) || die "couldn't open $filename\n";
    $line = <ITEMFILE>;
    $line =~ s/^\xEF\xBB\xBF//; # strip BOM
    $line =~ s/\x0B+//ig;
    $line =~ s/&/&amp;/g;
    print BACKUP ($line);
    #Tidy up the item file some metadata title contains \vt-vertical tab
    while ($line = <ITEMFILE>) {
	$line =~ s/\x0B+//ig;
	$line =~ s/&/&amp;/g;
	print BACKUP ($line);
    }
    close ITEMFILE;
    close BACKUP;
    &File::Copy::copy ($backup_filename, $filename);
    &util::rm($backup_filename);

    my $doc_obj;
    if ($xml_version) {
	$file =~ s/^[\/\\]+//; # $file often begins with / so we'll tidy it up
	$self->{'file'} = $file;
	$self->{'filename'} = $filename;
	$self->{'processor'} = $processor;
	$self->{'metadata'} = $metadata;

	eval {
	    $@ = "";
	    my $xslt = $self->{'xslt'};
	    if (defined $xslt && ($xslt ne "")) {
		# perform xslt
		my $transformed_xml = $self->apply_xslt($xslt,$filename);
		
		# feed transformed file (now in memory as string) into XML parser
		#$self->{'parser'}->parse($transformed_xml);
		$self->parse_string($transformed_xml);
	    }
	    else {
		#$self->{'parser'}->parsefile($filename);
		$self->parse_file($filename);
	    }
	};
	
	
	
	if ($@) {
	    
	    # parsefile may either croak somewhere in XML::Parser (e.g. because
	    # the document is not well formed) or die somewhere in XMLPlug or a
	    # derived plugin (e.g. because we're attempting to process a
	    # document whose DOCTYPE is not meant for this plugin). For the
	    # first case we'll print a warning and continue, for the second
	    # we'll just continue quietly
	    
	    print STDERR "**** XML Parse Error is: $@\n";
	    
	    my ($msg) = $@ =~ /Carp::croak\(\'(.*?)\'\)/;
	    if (defined $msg) {	
		my $outhandle = $self->{'outhandle'};
		my $plugin_name = ref ($self);
		print $outhandle "$plugin_name failed to process $file ($msg)\n";
	    }

	    # reset ourself for the next document
	    $self->{'section_level'}=0;
	    print STDERR "<ProcessingError n='$file'>\n" if ($gli);
	    return -1; # error during processing
	}
	$doc_obj = $self->{'doc_obj'};
    } else {
	my ($dir);
	($dir, $file) = $filename =~ /^(.*?)([^\/\\]*)$/;

	#process the .item file
	$doc_obj = $self->process_item($filename, $dir, $file, $processor);
	
    }
    
    if ($self->{'cover_image'}) {
	$self->associate_cover_image($doc_obj, $filename);
    }

    # include any metadata passed in from previous plugins 
    # note that this metadata is associated with the top level section
    my $section = $doc_obj->get_top_section();
    $self->extra_metadata ($doc_obj, $section, $metadata);
    #my $text="";
    # do plugin specific processing of doc_obj
    #unless (defined ($self->process(\$text, $pluginfo, $base_dir, $file, $metadata, $doc_obj))) {
	#print STDERR "<ProcessingError n='$file'>\n" if ($gli);
	#return -1;
    #}
    # do any automatic metadata extraction
    $self->auto_extract_metadata ($doc_obj);

    $self->{'num_processed'}++;
    return (1,$doc_obj);
}

sub read
{
    my $self = shift (@_);
    my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;

    my ($process_status,$doc_obj) = $self->read_into_doc_obj(@_);

    if ((defined $process_status) && ($process_status == 1)) {
	# process the document
	$processor->process($doc_obj);

	#if(defined($self->{'places_filename'})){
	#    &util::rm($self->{'places_filename'});
	#    $self->{'places_filename'} = undef;
	#}
	#$self->{'num_processed'} ++;
	undef $doc_obj; 
    }
    
    # clean up temporary files - we do this here instead of in  
    # process_image becuase associated files aren't actually copied 
    # until after process has been run.
    if (defined $self->{'tmp_filename1'} && 
	-e $self->{'tmp_filename1'}) {
	&util::rm($self->{'tmp_filename1'})
    }
    if (defined $self->{'tmp_filename2'} && 
	-e $self->{'tmp_filename2'}) {
	&util::rm($self->{'tmp_filename2'})
    }
    if (defined $self->{'tmp_filename3'} && 
	-e $self->{'tmp_filename3'}) {
      &util::rm($self->{'tmp_filename3'})
    }
    # if process_status == 1, then the file has been processed.
    return $process_status;
}

sub xml_start_tag {
    my $self = shift(@_);
    my ($expat, $element) = @_;
    $self->{'element'} = $element;
    
    my $doc_obj = $self->{'doc_obj'};
    if ($element eq "PagedDocument") {
	$self->{'current_section'} = $doc_obj->get_top_section();
    } elsif ($element eq "PageGroup" || $element eq "Page") {
	# create a new section as a child
	$self->{'current_section'} = $doc_obj->insert_section($doc_obj->get_end_child($self->{'current_section'}));
	$self->{'num_pages'}++;
	# assign pagenum as  what??
	my $pagenum = $_{'pagenum'}; #TODO!!
	if (defined $pagenum) {
	    $doc_obj->set_utf8_metadata_element($self->{'current_section'}, 'PageNum', $pagenum);
	}
	my ($imgfile) = $_{'imgfile'};
	if (defined $imgfile) {
	    $self->process_image($self->{'base_dir'}.$imgfile, $imgfile, $doc_obj, $self->{'current_section'});
	}
	my ($txtfile) = $_{'txtfile'};
	if (defined($txtfile)&& $txtfile ne "") {
	    $self->process_text ($self->{'base_dir'}.$txtfile, $txtfile, $doc_obj, $self->{'current_section'});
	} else {
	    # otherwise add in some dummy text
	    #create an empty text string so we don't break downstream plugins 
	    my $text = &gsprintf::lookup_string("{BasPlug.dummy_text}",1);
	    $doc_obj->add_utf8_text($self->{'current_section'}, $text);
	}
    } elsif ($element eq "Metadata") {
	$self->{'metadata_name'} = $_{'name'};
    }
}

sub xml_end_tag {
    my $self = shift(@_);
    my ($expat, $element) = @_;
    
    my $doc_obj = $self->{'doc_obj'};
    if ($element eq "Page" || $element eq "PageGroup") {
	# if Title hasn't been assigned, set PageNum as Title
	if (!defined $doc_obj->get_metadata_element ($self->{'current_section'}, "Title") && defined $doc_obj->get_metadata_element ($self->{'current_section'}, "PageNum" )) {
	    $doc_obj->add_utf8_metadata ($self->{'current_section'}, "Title", $doc_obj->get_metadata_element ($self->{'current_section'}, "PageNum" ));
	}
	# move the current section back to the parent
	$self->{'current_section'} = $doc_obj->get_parent_section($self->{'current_section'});
    } elsif ($element eq "Metadata") {
	
	$doc_obj->add_utf8_metadata ($self->{'current_section'}, $self->{'metadata_name'}, $self->{'metadata_value'});
	$self->{'metadata_name'} = "";
	$self->{'metadata_value'} = "";

    }
    # otherwise we ignore the end tag
}


sub xml_text {
    my $self = shift(@_);
    my ($expat) = @_; 

    if ($self->{'element'} eq "Metadata" && $self->{'metadata_name'}) {
   	$self->{'metadata_value'} .= $_;
    }
}

sub xml_doctype {
}

sub open_document {
    my $self = shift(@_);
    
    # create a new document
    $self->{'doc_obj'} = new doc ($self->{'filename'}, "indexed_doc");
    my $doc_obj = $self->{'doc_obj'};
    $doc_obj->set_OIDtype ($self->{'processor'}->{'OIDtype'});
    my ($dir, $file) = $self->{'filename'} =~ /^(.*?)([^\/\\]*)$/;
    $self->{'base_dir'} = $dir;
    $self->{'num_pages'} = 0;
    my $topsection = $doc_obj->get_top_section();
    if ($self->{'documenttype'} eq 'paged') {
	# set the gsdlthistype metadata to Paged - this ensures this document will
	# be treated as a Paged doc, even if Titles are not numeric
    
	$doc_obj->set_utf8_metadata_element ($topsection, "gsdlthistype", "Paged");
    } else {
	$doc_obj->set_utf8_metadata_element ($topsection, "gsdlthistype", "Hierarchy");
    }

    $doc_obj->add_metadata ($topsection, "Source", $file);
    if ($self->{'headerpage'}) {
	$doc_obj->add_text($topsection, &gsprintf::lookup_string("{BasPlug.dummy_text}"));
    }

}

sub close_document {
    my $self = shift(@_);
    my $doc_obj = $self->{'doc_obj'};
        
    $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "Plugin", "$self->{'plugin_type'}");
    $doc_obj->add_metadata($doc_obj->get_top_section(), "FileFormat", "PagedImg");
    
    # add numpages metadata
    $doc_obj->set_utf8_metadata_element ($doc_obj->get_top_section(), 'NumPages', $self->{'num_pages'});

    # add an OID
    $doc_obj->set_OID();
    
}

sub process_item {
    my $self = shift (@_);
    my ($filename, $dir, $file, $processor) = @_;

    my $doc_obj = new doc ($filename, "indexed_doc");
    $doc_obj->set_OIDtype ($processor->{'OIDtype'}, $processor->{'OIDmetadata'});
    my $topsection = $doc_obj->get_top_section();
    $doc_obj->add_utf8_metadata($topsection, "Plugin", "$self->{'plugin_type'}");
    $doc_obj->add_metadata($topsection, "FileFormat", "PagedImg");

    if ($self->{'documenttype'} eq 'paged') {
	# set the gsdlthistype metadata to Paged - this ensures this document will
	# be treated as a Paged doc, even if Titles are not numeric
	$doc_obj->set_utf8_metadata_element ($topsection, "gsdlthistype", "Paged");
    } else {
	$doc_obj->set_utf8_metadata_element ($topsection, "gsdlthistype", "Hierarchy");
    }

    $doc_obj->add_metadata ($topsection, "Source", $file);

    open (ITEMFILE, $filename) || die "couldn't open $filename\n";
    my $line = "";
    my $num = 0;
    while (defined ($line = <ITEMFILE>)) {
	next unless $line =~ /\w/;
	chomp $line;
	next if $line =~ /^#/; # ignore comment lines
	if ($line =~ /^<([^>]*)>\s*(.*?)\s*$/) {
	    $doc_obj->set_utf8_metadata_element ($topsection, $1, $2);
	    #$meta->{$1} = $2;
	} else {
	    $num++;
	    # line should be like page:imagefilename:textfilename:r - the r is optional -> means rotate the image 180 deg
	    $line =~ s/^\s+//; #remove space at the front
	    $line =~ s/\s+$//; #remove space at the end
	    my ($pagenum, $imgname, $txtname, $rotate) = split /:/, $line;
	    
	    # create a new section for each image file
	    my $cursection = $doc_obj->insert_section($doc_obj->get_end_child($topsection));
	    # the page number becomes the Title
	    $doc_obj->set_utf8_metadata_element($cursection, 'Title', $pagenum);
	    # process the image for this page if there is one
	    if (defined $imgname && $imgname ne "") {
		my $result1 = $self->process_image($dir.$imgname, $imgname, $doc_obj, $cursection, $rotate);
		
		if (!defined $result1)
		{
		    print "PagedImgPlug: couldn't process image \"$dir.$imgname\" for item \"$filename\"\n";
		}
	    }
	    # process the text file if one is there
	    if (defined $txtname && $txtname ne "") {
		my $result2 = $self->process_text ($dir.$txtname, $txtname, $doc_obj, $cursection);
		if (!defined $result2) {
		    print "PagedImgPlug: couldn't process text file \"$dir.$txtname\" for item \"$filename\"\n";
		}
	    } else {
		# otherwise add in some dummy text 
		$doc_obj->add_text($cursection, &gsprintf::lookup_string("{BasPlug.dummy_text}"));
	    }
	}
    }
    
    close ITEMFILE;

    # if we want a header page, we need to add some text into the top section, otherwise this section will become invisible
    if ($self->{'headerpage'}) {
	$doc_obj->add_text($topsection, &gsprintf::lookup_string("{BasPlug.dummy_text}"));
    }
    $file =~ s/\.item//i;
    $doc_obj->set_OID ();
    # add numpages metadata
    $doc_obj->set_utf8_metadata_element ($topsection, 'NumPages', "$num");
    return $doc_obj;
}

sub process_text {
    my $self = shift (@_);
    my ($fullpath, $file, $doc_obj, $cursection) = @_;
    
    # check that the text file exists!!
    if (!-f $fullpath) {
	print "PagedImgPlug: ERROR: File $fullpath does not exist, skipping\n";
	return 0;
    }

    # Do encoding stuff
    my ($language, $encoding) = $self->textcat_get_language_encoding ($fullpath);

    my $text="";
    &BasPlug::read_file($self, $fullpath, $encoding, $language, \$text);
    if (!length ($text)) {
        my $plugin_name = ref ($self);
        print "PagedImgPlug: ERROR: $fullpath contains no text\n" if $self->{'verbosity'};
        return 0;
    }

    # we need to escape the escape character, or else mg will convert into
    # eg literal newlines, instead of leaving the text as '\n'
    $text =~ s/\\/\\\\/g; # macro language
    $text =~ s/_/\\_/g; # macro language
    $text =~ s/</&lt;/g;
    $text =~ s/>/&gt;/g;
 
    # insert preformat tags and add text to document object
    $doc_obj->add_utf8_text($cursection, "<pre>\n$text\n</pre>");
    
    return 1;
}

# do plugin specific processing of doc_obj
sub process {
    my $self = shift (@_);
    my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj) = @_;
    my $outhandle = $self->{'outhandle'};
    
    return 1;
}

1;