source: main/trunk/greenstone2/perllib/plugins/PDFPlugin.pm@ 21800

###########################################################################
#
# PDFPlugin.pm -- reasonably with-it pdf plugin
# A component of the Greenstone digital library software
# from the New Zealand Digital Library Project at the 
# University of Waikato, New Zealand.
#
# Copyright (C) 1999-2001 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.
#
###########################################################################
package PDFPlugin;

use ConvertBinaryFile;
use ReadTextFile;
use unicode;
use strict;
no strict 'refs'; # so we can use a var for filehandles (eg STDERR)

sub BEGIN {
    @PDFPlugin::ISA = ('ConvertBinaryFile', 'ReadTextFile');
}

my $convert_to_list =
    [ { 'name' => "auto",
    'desc' => "{ConvertBinaryFile.convert_to.auto}" },
      { 'name' => "html",
    'desc' => "{ConvertBinaryFile.convert_to.html}" },
      { 'name' => "text",
    'desc' => "{ConvertBinaryFile.convert_to.text}" },
      { 'name' => "pagedimg_jpg",
    'desc' => "{ConvertBinaryFile.convert_to.pagedimg_jpg}"},
      { 'name' => "pagedimg_gif",
    'desc' => "{ConvertBinaryFile.convert_to.pagedimg_gif}"},
      { 'name' => "pagedimg_png",
    'desc' => "{ConvertBinaryFile.convert_to.pagedimg_png}"}, 
      ];


my $arguments = 
    [
     { 'name' => "convert_to",
       'desc' => "{ConvertBinaryFile.convert_to}",
       'type' => "enum",
       'reqd' => "yes",
       'list' => $convert_to_list, 
       'deft' => "html" },   
     { 'name' => "process_exp",
       'desc' => "{BasePlugin.process_exp}",
       'type' => "regexp",
       'deft' => &get_default_process_exp(),
       'reqd' => "no" },
     { 'name' => "block_exp",
       'desc' => "{BasePlugin.block_exp}",
       'type' => "regexp",
       'deft' => &get_default_block_exp() },
     { 'name' => "metadata_fields",
       'desc' => "{HTMLPlugin.metadata_fields}",
       'type' => "string",
       'deft' => "" },
      { 'name' => "metadata_field_separator",
    'desc' => "{HTMLPlugin.metadata_field_separator}",
    'type' => "string",
    'deft' => "" },
     { 'name' => "noimages",
       'desc' => "{PDFPlugin.noimages}",
       'type' => "flag" },
     { 'name' => "allowimagesonly",
       'desc' => "{PDFPlugin.allowimagesonly}",
       'type' => "flag" },
     { 'name' => "complex",
       'desc' => "{PDFPlugin.complex}",
       'type' => "flag" },
     { 'name' => "nohidden",
       'desc' => "{PDFPlugin.nohidden}",
       'type' => "flag" },
     { 'name' => "zoom",
       'desc' => "{PDFPlugin.zoom}",
       'deft' => "2",
       'range' => "1,3", # actually the range is 0.5-3 
       'type' => "int" },
     { 'name' => "use_sections",
       'desc' => "{PDFPlugin.use_sections}",
       'type' => "flag" },
     { 'name' => "description_tags",
       'desc' => "{HTMLPlugin.description_tags}",
       'type' => "flag" }
     ];

my $options = { 'name'     => "PDFPlugin",
        'desc'     => "{PDFPlugin.desc}",
        'abstract' => "no",
        'inherits' => "yes",
        'srcreplaceable' => "yes", # Source docs in PDF can be replaced with GS-generated html      
        'args'     => $arguments };

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

    push(@$inputargs,"-title_sub");
    push(@$inputargs,'^(Page\s+\d+)?(\s*1\s+)?');

    push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
    push(@{$hashArgOptLists->{"OptList"}},$options);

    my @arg_array = @$inputargs;
    my $self = new ConvertBinaryFile($pluginlist, $inputargs, $hashArgOptLists);
    
    if ($self->{'info_only'}) {
    # don't worry about any options etc
    return bless $self, $class;
    }

    $self->{'filename_extension'} = "pdf";
    $self->{'file_type'} = "PDF";

    # these are passed through to gsConvert.pl by ConvertBinaryFile.pm
    my $zoom = $self->{"zoom"};
    $self->{'convert_options'} = "-pdf_zoom $zoom";
    $self->{'convert_options'} .= " -pdf_complex" if $self->{"complex"};
    $self->{'convert_options'} .= " -pdf_nohidden" if $self->{"nohidden"};
    $self->{'convert_options'} .= " -pdf_ignore_images" if $self->{"noimages"};
    $self->{'convert_options'} .= " -pdf_allow_images_only" if $self->{"allowimagesonly"};

    my $secondary_plugin_options = $self->{'secondary_plugin_options'};

    if (!defined $secondary_plugin_options->{'HTMLPlugin'}) {
    $secondary_plugin_options->{'HTMLPlugin'} = [];
    }
    if (!defined $secondary_plugin_options->{'TextPlugin'}) {
    $secondary_plugin_options->{'TextPlugin'} = [];
    }
    if (defined $self->{'convert_to'} && $self->{'convert_to'} =~ m/(pagedimage|pagedimg).*/i) {
    if (!defined $secondary_plugin_options->{'PagedImagePlugin'}){
        $secondary_plugin_options->{'PagedImagePlugin'} = [];
        my $pagedimg_options = $secondary_plugin_options->{'PagedImagePlugin'}; 
        push(@$pagedimg_options, "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');
        push(@$pagedimg_options, "-screenviewsize", "1000");
        push(@$pagedimg_options, "-enable_cache");
    }
    }
    my $html_options = $secondary_plugin_options->{'HTMLPlugin'};
    my $text_options = $secondary_plugin_options->{'TextPlugin'};
    my $pagedimg_options = $secondary_plugin_options->{'PagedImagePlugin'};

#    if ($self->{'input_encoding'} eq "auto") {
#   $self->{'input_encoding'} = "utf8";
#    }

    # if pdftohtml is always producing utf8, then htmlplug always needs this option
    push(@$html_options,"-input_encoding", "utf8");
    push(@$html_options,"-extract_language") if $self->{'extract_language'};

    push(@$html_options, "-processing_tmp_files");
    push(@$pagedimg_options, "-processing_tmp_files");

    # Instruct HTMLPlug (when eventually accessed through read_into_doc_obj) 
    # to extract these metadata fields from the HEAD META fields
    my $required_metadata;
    if (defined $self->{'metadata_fields'} && $self->{'metadata_fields'} =~ /\S/) {
    push(@$html_options,"-metadata_fields",$self->{'metadata_fields'});
    } else {
    push(@$html_options,"-metadata_fields","Title,GENERATOR,date,author<Creator>");
    }
    if (defined $self->{'metadata_field_separator'} && $self->{'metadata_field_separator'} =~ /\S/) {
    push(@$html_options,"-metadata_field_separator",$self->{'metadata_field_separator'});
    }
        
    if ($self->{'use_sections'} || $self->{'description_tags'}) {
    $self->{'description_tags'} = 1;
    push(@$html_options,"-description_tags");
    }

    # following title_sub removes "Page 1" added by pdftohtml, and a leading
    # "1", which is often the page number at the top of the page. Bad Luck
    # if your document title actually starts with "1 " - is there a better way?
    push(@$html_options , "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');
    push(@$text_options , "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');

    my $associate_tail_re = $self->{'associate_tail_re'};
    if ((defined $associate_tail_re) && ($associate_tail_re ne "")) {
    push(@$html_options, "-associate_tail_re", $associate_tail_re);
    push(@$text_options, "-associate_tail_re", $associate_tail_re);
    push(@$pagedimg_options, "-associate_tail_re", $associate_tail_re) if defined $pagedimg_options;
    }

    push(@$html_options, "-file_rename_method", "none");
    push(@$text_options, "-file_rename_method", "none");
    push(@$pagedimg_options, "-file_rename_method", "none") if defined $pagedimg_options;

    $self = bless $self, $class;
    $self->load_secondary_plugins($class,$secondary_plugin_options,$hashArgOptLists);
    return $self;
}

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

    return q^(?i)\.pdf$^;
}

# so we don't inherit HTMLPlug's block exp...
sub get_default_block_exp {
    return "";
}
    
sub convert_post_process
{
    my $self = shift (@_);
    my ($conv_filename) = @_;

    my $outhandle=$self->{'outhandle'};

    #$self->{'input_encoding'} = "utf8"; # The output is always in utf8 (is it?? it is for html, but what about other types?)
    #my ($language, $encoding) = $self->textcat_get_language_encoding ($conv_filename);

    # read in file ($text will be in utf8)
    my $text = "";
    # encoding will be utf8 for html files - what about other types? will we do this step for them anyway?
    $self->read_file ($conv_filename, "utf8", "", \$text);

    # Calculate number of pages based on <a ...> tags (we have a <a name=1> etc
    # for each page).  Metadata based on this calculation not set until process()
    # 
    # Note: this is done even if we are not breaking to document into pages as it might
    # be useful to give an indication of document length in browser through setting
    # num_pages as metadata.
    my @pages = ($text =~ /\<[Aa] name=\"?\w+\"?>/ig);
    my $num_pages = scalar(@pages);
    $self->{'num_pages'} = $num_pages;

    if ($self->{'use_sections'}
    && $self->{'converted_to'} eq "HTML") {

    print $outhandle "PDFPlugin: Calculating sections...\n";

    # we have "<a name=1></a>" etc for each page
    # it may be <A name=
    my @sections = split('<[Aa] name=', $text);

    my $top_section = "";

    if (scalar (@sections) == 1) { #only one section - no split!
        print $outhandle "PDFPlugin: warning - no sections found\n";
    } else {
        $top_section .= shift @sections; # keep HTML header etc as top_section
    }

    # handle first section specially for title? Or all use first 100...
    
    my $title = $sections[0];
    $title =~ s/^\"?\w+\"?>//; # specific for pdftohtml...
    $title =~ s/<\/([^>]+)><\1>//g; # (eg) </b><b> - no space
    $title =~ s/<[^>]*>/ /g;
    $title =~ s/(?:&nbsp;|\xc2\xa0)/ /g; # utf-8 for nbsp...
    $title =~ s/^\s+//s;
    $title =~ s/\s+$//;
    $title =~ s/\s+/ /gs;
    $title =~ s/^$self->{'title_sub'}// if ($self->{'title_sub'});
    $title =~ s/^\s+//s; # in case title_sub introduced any...
    $title = substr ($title, 0, 100);
    $title =~ s/\s\S*$/.../;


    if (scalar (@sections) == 1) { # no sections found
        $top_section .= $sections[0];
        @sections=();
    } else {
        $top_section .= "<!--<Section>\n<Metadata name=\"Title\">$title</Metadata>\n-->\n <!--</Section>-->\n";
    }

    # add metadata per section...
    foreach my $section (@sections) {
        # section names are not always just digits, may be like "outline"
        $section =~ s@^\"?(\w+)\"?></a>@@; # leftover from split expression...

        $title = $1; # Greenstone does magic if sections are titled digits
        if (! defined($title) ) {
        print STDERR "no title: $section\n";
        $title = " "; # get rid of the undefined warning in next line
        }
        my $newsection = "<!-- from PDFPlugin -->\n<!-- <Section>\n";
        $newsection .= "<Metadata name=\"Title\">" . $title
        . "</Metadata>\n--><p>\n";
        $newsection .= $section;
        $newsection .= "<!--</Section>-->\n";
        $section = $newsection;
    }

    $text=join('', ($top_section, @sections));
    }

    # turn any high bytes that aren't valid utf-8 into utf-8.
    unicode::ensure_utf8(\$text);

    # Write it out again!
    $self->utf8_write_file (\$text, $conv_filename);
}


# do plugin specific processing of doc_obj for HTML type
sub process {
    my $self = shift (@_);
    my ($pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;

    my $result = $self->process_type($base_dir,$file,$doc_obj);

    # fix up the extracted date metadata to be in Greenstone date format,
    # and fix the capitalisation of 'date'
    my $cursection = $doc_obj->get_top_section();
    foreach my $datemeta (@{$doc_obj->get_metadata($cursection, "date")}) {
    $doc_obj->delete_metadata($cursection, "date", $datemeta);

    # We're just interested in the date bit, not the time
    # some pdf creators (eg "Acrobat 5.0 Scan Plug-in for Windows")
    # set a /CreationDate, and set /ModDate to 000000000. pdftohtml
    # extracts the ModDate, so it is 0...
    $datemeta =~ /(\d+)-(\d+)-(\d+)/;
    my ($year, $month, $day) = ($1,$2,$3);
    if (defined($year) && defined($month) && defined($day)) {
        if ($year == 0) {next}
        if ($year < 100) {$year += 1900} # just to be safe
        if ($month =~ /^\d$/) {$month="0$month"} # single digit
        if ($day =~ /^\d$/) {$day="0$day"} # single digit
        my $date="$year$month$day";
        $doc_obj->add_utf8_metadata($cursection, "Date", $date);
    }
    }

    $doc_obj->add_utf8_metadata($cursection, "NumPages", $self->{'num_pages'});
    
    if ($self->{'use_sections'} && $self->{'converted_to'} eq "HTML") {
    # we explicitly make it a paged document, cos greenstone won't get it
    # right if any section has an empty title, or one with letters in it
    $doc_obj->set_utf8_metadata_element ($cursection, "gsdlthistype", "Paged");
    }

    return $result;
}

1;