source: gsdl/trunk/perllib/plugins/BasePlugin.pm@ 16557

###########################################################################
#
# BasePlugin.pm -- base class for all the import plugins
# A component of the Greenstone digital library software
# from the New Zealand Digital Library Project at the 
# University of Waikato, New Zealand.
#
# Copyright (C) 1999-2005 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 BasePlugin;

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

use File::Basename;

use multiread;
use encodings;
use unicode;
use textcat;
use doc;
eval "require diagnostics"; # some perl distros (eg mac) don't have this
use ghtml;
use gsprintf 'gsprintf';

use PrintInfo;

BEGIN {
    @BasePlugin::ISA = ( 'PrintInfo' );
}

our $encoding_list =
    [ { 'name' => "ascii",
    'desc' => "{BasePlugin.encoding.ascii}" },
      { 'name' => "utf8",
    'desc' => "{BasePlugin.encoding.utf8}" },
      { 'name' => "unicode",
    'desc' => "{BasePlugin.encoding.unicode}" } ];
      

my $e = $encodings::encodings;
foreach my $enc (sort {$e->{$a}->{'name'} cmp $e->{$b}->{'name'}} keys (%$e)) 
{
    my $hashEncode =
    {'name' => $enc,
     'desc' => $e->{$enc}->{'name'}};
    
    push(@{$encoding_list},$hashEncode);
}

our $encoding_plus_auto_list = 
    [ { 'name' => "auto",
    'desc' => "{BasePlugin.filename_encoding.auto}" },
       { 'name' => "auto-language-analysis",
    'desc' => "{BasePlugin.filename_encoding.auto_language_analysis}" }, # textcat
      { 'name' => "auto-filesystem-encoding",
    'desc' => "{BasePlugin.filename_encoding.auto_filesystem_encoding}" }, # locale
      { 'name' => "auto-fl",
    'desc' => "{BasePlugin.filename_encoding.auto_fl}" }, # locale followed by textcat
      { 'name' => "auto-lf",
    'desc' => "{BasePlugin.filename_encoding.auto_lf}" } ]; # texcat followed by locale 

push(@{$encoding_plus_auto_list},@{$encoding_list});

my $arguments =
    [ { 'name' => "process_exp",
    'desc' => "{BasePlugin.process_exp}",
    'type' => "regexp",
    'deft' => "",
    'reqd' => "no" },
      { 'name' => "no_blocking",
    'desc' => "{BasePlugin.no_blocking}",
    'type' => "flag",
    'reqd' => "no"},
      { 'name' => "block_exp",
    'desc' => "{BasePlugin.block_exp}",
    'type' => "regexp",
    'deft' => "",
    'reqd' => "no" },
      { 'name' => "associate_ext",
    'desc' => "{BasePlugin.associate_ext}",
    'type' => "string",
    'reqd' => "no" },
      { 'name' => "associate_tail_re",
    'desc' => "{BasePlugin.associate_tail_re}",
    'type' => "string",
    'reqd' => "no" },
      { 'name' => "use_as_doc_identifier",
    'desc' => "{BasePlugin.use_as_doc_identifier}",
    'type' => "string",
    'reqd' => "no" ,
    'deft' => "" } ,
     { 'name' => "no_cover_image",
    'desc' => "{BasePlugin.no_cover_image}",
    'type' => "flag",
    'reqd' => "no" },
      { 'name' => "filename_encoding",
    'desc' => "{BasePlugin.filename_encoding}",
    'type' => "enum",
    'deft' => "auto",
    'list' => $encoding_plus_auto_list,
    'reqd' => "no" },
      { 'name' => "smart_block",
        'desc' => "{common.deprecated}. {BasePlugin.smart_block}",
        'type' => "flag",
        'reqd' => "no",
    'hiddengli' => "yes" } # deprecated, but leave in for old collections

      
      ];


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


sub new {

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

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

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

    if ($self->{'smart_block'}) {
    print STDERR "WARNING: -smart_block option has been deprecated and is no longer useful\n";
    }
    $self->{'smart_block'} = undef;

    my $plugin_name = (defined $pluginlist->[0]) ? $pluginlist->[0] : $class;
    $self->{'plugin_type'} = $plugin_name;

    $self->{'num_processed'} = 0;
    $self->{'num_not_processed'} = 0;
    $self->{'num_blocked'} = 0;
    $self->{'num_archives'} = 0;
    $self->{'cover_image'} = 1; # cover image is on by default
    $self->{'cover_image'} = 0 if ($self->{'no_cover_image'});
    #$self->{'option_list'} = $hashArgOptLists->{"OptList"};
    
    my $associate_ext = $self->{'associate_ext'};
    if ((defined $associate_ext) && ($associate_ext ne "")) {

    my $associate_tail_re = $self->{'associate_tail_re'};
    if ((defined $associate_tail_re) && ($associate_tail_re ne "")) {
        my $outhandle = $self->{'outhandle'};
        print $outhandle "Warning: can only specify 'associate_ext' or 'associate_tail_re'\n";
        print $outhandle "         defaulting to 'associate_tail_re'\n";
    }
    else {
        my @exts = split(/,/,$associate_ext);

        my @exts_bracketed = map { $_ = "(?:\\.$_)" } @exts;
        my $associate_tail_re = join("|",@exts_bracketed);
        $self->{'associate_tail_re'} = $associate_tail_re;
    }

    delete $self->{'associate_ext'};
    }

    return bless $self, $class;

}

# initialize BasePlugin options
# if init() is overridden in a sub-class, remember to call BasePlugin::init()
sub init {
    my $self = shift (@_);
    my ($verbosity, $outhandle, $failhandle) = @_;

    # verbosity is passed through from the processor
    $self->{'verbosity'} = $verbosity;

    # as are the outhandle and failhandle
    $self->{'outhandle'} = $outhandle if defined $outhandle;
    $self->{'failhandle'} = $failhandle;
#    $self->SUPER::init(@_);
    
    # set process_exp and block_exp to defaults unless they were
    # explicitly set

    if ((!$self->is_recursive()) and 
    (!defined $self->{'process_exp'}) || ($self->{'process_exp'} eq "")) {

    $self->{'process_exp'} = $self->get_default_process_exp ();
    if ($self->{'process_exp'} eq "") {
        warn ref($self) . " Warning: Non-recursive plugin has no process_exp\n";
    }
    }

    if ((!defined $self->{'block_exp'}) || ($self->{'block_exp'} eq "")) {
    $self->{'block_exp'} = $self->get_default_block_exp ();
    }

}

sub begin {
    my $self = shift (@_);
    my ($pluginfo, $base_dir, $processor, $maxdocs) = @_;
}

sub end {
    # potentially called at the end of each plugin pass 
    # import.pl only has one plugin pass, but buildcol.pl has multiple ones

    my ($self) = shift (@_);
}

sub deinit {
    # called only once, after all plugin passes have been done

    my ($self) = @_;
}

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

    $self->{'incremental'} = $incremental;
}

# this function should be overridden to return 1
# in recursive plugins
sub is_recursive {
    my $self = shift (@_);

    return 0; 
}

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

    return "";
}

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

    return "";
}

# default implementation is to do nothing
sub store_block_files {
    
    my $self =shift (@_);
    my ($filename_full_path, $block_hash) = @_;

}

# put files to block into hash 
sub use_block_expressions {

    my $self =shift (@_);
    my ($filename_full_path, $block_hash) = @_;

    if ($self->{'block_exp'} ne "" && $filename_full_path =~ /$self->{'block_exp'}/) {
    $block_hash->{'file_blocks'}->{$filename_full_path} = 1;
    }

}

#default implementation is to block a file with same name as this, but extension jpg or JPG, if cover_images is on.
sub block_cover_image
{
    my $self =shift;
    my ($filename, $block_hash) = @_;

    if ($self->{'cover_image'}) {
    my $coverfile = $filename;
    $coverfile =~ s/\.[^\\\/\.]+$/\.jpg/;
    if (!-e $coverfile) {
        $coverfile =~ s/jpg$/JPG/;
    }   
    if (-e $coverfile) {
        $block_hash->{'file_blocks'}->{$coverfile} = 1;
    } 
    }

    return;
}


# discover all the files that should be blocked by this plugin
# check the args ...
sub file_block_read {

    my $self = shift (@_);  
    my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $gli) = @_;
    # Keep track of filenames with same root but different extensions
    # Used to support -associate_ext and the more generalised
    # -associate_tail_re
    my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);

    my $associate_tail_re = $self->{'associate_tail_re'};
    if ((defined $associate_tail_re) && ($associate_tail_re ne "")) {
    
    my ($file_prefix,$file_ext) 
        = &util::get_prefix_and_tail_by_regex($filename_full_path,$associate_tail_re);
    
    if ((defined $file_prefix) && (defined $file_ext)) {
        my $shared_fileroot = $block_hash->{'shared_fileroot'};
        if (!defined $shared_fileroot->{$file_prefix}) {
        my $file_prefix_rec = { 'tie_to'  => undef, 
                        'exts'    => {} };
        $shared_fileroot->{$file_prefix} = $file_prefix_rec;
        }
        
        my $file_prefix_rec = $shared_fileroot->{$file_prefix};

        if ($self->can_process_this_file($filename_full_path)) {
        # This is the document the others should be tied to
        $file_prefix_rec->{'tie_to'} = $file_ext;
        }
        else {
        if ($file_ext =~ m/$associate_tail_re$/) {
            # this file should be associated to the main one
            $file_prefix_rec->{'exts'}->{$file_ext} = 1;
        }
        }

    }
    }

    # check block expressions
    $self->use_block_expressions($filename_full_path, $block_hash) unless $self->{'no_blocking'};

    # now check whether we are actually processing this
    if (!-f $filename_full_path || !$self->can_process_this_file($filename_full_path)) {
    return undef; # can't recognise
    }
   
    $self->store_block_files($filename_full_path, $block_hash) unless $self->{'no_blocking'};

    # block the cover image if there is one
    if ($self->{'cover_image'}) {
    $self->block_cover_image($filename_full_path, $block_hash) unless $self->{'no_blocking'};
    }
       
    return 1;
}

# plugins that rely on more than process_exp (eg XML plugins) can override this method
sub can_process_this_file {
    my $self = shift(@_);
    my ($filename) = @_;

    if ($self->{'process_exp'} ne "" && $filename =~ /$self->{'process_exp'}/) {
    return 1;
    }
    return 0;
    
}

# just converts path as is to utf8.
sub filepath_to_utf8 {
    my $self = shift (@_);  
    my ($file, $file_encoding) = @_;
    my $filemeta = $file;

    my $filename_encoding = $self->{'filename_encoding'}; # filename encoding setting

##  print STDERR "**** User chose filename encoding setting: $filename_encoding\n";
    
    # Whenever filename-encoding is set to any of the auto settings, we
    # check if the filename is already in UTF8. If it is, then we're done.
    if($filename_encoding =~ m/auto/) {
    if(&unicode::check_is_utf8($filemeta)) 
    {
##      print STDERR "**** It is already UTF8\n";
        $filename_encoding = "utf8";
        return $filemeta;
    } 
    }
    
    # Auto setting, but filename is not utf8
    if ($filename_encoding eq "auto") 
    {
    # try textcat
    $filename_encoding = $self->textcat_encoding($filemeta);
    
    # check the locale next
    $filename_encoding = $self->locale_encoding() if $filename_encoding eq "undefined";
    

    # now try the encoding of the document, if available
    if ($filename_encoding eq "undefined" && defined $file_encoding) {
        $filename_encoding = $file_encoding;
    }

    }

    elsif ($filename_encoding eq "auto-language-analysis") 
    {   
    $filename_encoding = $self->textcat_encoding($filemeta);

    # now try the encoding of the document, if available
    if ($filename_encoding eq "undefined" && defined $file_encoding) {
        $filename_encoding = $file_encoding;
    } 
    }

    elsif ($filename_encoding eq "auto-filesystem-encoding") 
    {
    # try locale
    $filename_encoding = $self->locale_encoding();
    }

    elsif ($filename_encoding eq "auto-fl") 
    {
    # filesystem-encoding (locale) then language-analysis (textcat)
    $filename_encoding = $self->locale_encoding();
    
    # try textcat
    $filename_encoding = $self->textcat_encoding($filemeta) if $filename_encoding eq "undefined";
        
    # else assume filename encoding is encoding of file content, if that's available
    if ($filename_encoding eq "undefined" && defined $file_encoding) {
        $filename_encoding = $file_encoding;
    }
    }
    
    elsif ($filename_encoding eq "auto-lf") 
    {
    # language-analysis (textcat) then filesystem-encoding (locale)
    $filename_encoding = $self->textcat_encoding($filemeta);
    
    # guess filename encoding from encoding of file content, if available
    if ($filename_encoding eq "undefined" && defined $file_encoding) {
        $filename_encoding = $file_encoding;
    }

    # try locale
    $filename_encoding = $self->locale_encoding() if $filename_encoding eq "undefined";
    }

##    print STDERR "**** filename_encoding selected: $filename_encoding \n";
        
    # if still undefined, use utf8 as fallback
    if ($filename_encoding eq "undefined") {
    $filename_encoding = "utf8";
    }

    # if the filename encoding is set to utf8 but it isn't utf8 already--such as when
    # 1. the utf8 fallback is used, or 2. if the system locale is used and happens to
    # be always utf8 (in which case the filename's encoding is also set as utf8 even 
    # though the filename need not be if it originates from another system)--in such
    # cases attempt to make the filename utf8 to match.
    if($filename_encoding eq "utf8" && !&unicode::check_is_utf8($filemeta)) {
##  print STDERR "**** BEFORE utf8 conversion: $filemeta\n";
    &unicode::ensure_utf8(\$filemeta);
##  print STDERR "**** AFTER utf8 conversion: $filemeta\n";
    }


    # convert non-unicode encodings to utf8
    if ($filename_encoding !~ m/(?:ascii|utf8|unicode)/) {
    $filemeta = &unicode::unicode2utf8(
      &unicode::convert2unicode($filename_encoding, \$filemeta)
    );
    }

    print "*** filename encoding found: $filename_encoding\n";
    print "*** utf8 encoded filename: $filemeta\n";

    return $filemeta;
}

# gets the filename with no path, converts to utf8, and then dm safes it.
#filename_encoding set by user
sub filename_to_utf8_metadata
{
    my $self = shift (@_);  
    my ($file, $file_encoding) = @_;

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

    my ($filemeta) = $file =~ /([^\\\/]+)$/; # getting the tail of the filepath (skips all string parts containing slashes upto the end)
    $filemeta = $self->filepath_to_utf8($filemeta, $file_encoding);

    my $dmsafe_filemeta = &ghtml::dmsafe($filemeta);

    return $dmsafe_filemeta;

}

sub locale_encoding {
    my $self = shift(@_);
    
    if (!defined $self->{'filesystem_encoding'}) {
    $self->{'filesystem_encoding'} = $self->get_filesystem_encoding();
    $self->{'filesystem_encoding'} = "undefined" if !defined $self->{'filesystem_encoding'};
    }

    print "filename encoding determined based on locale: " . $self->{'filesystem_encoding'} . "\n";
    return $self->{'filesystem_encoding'}; # can be the string "undefined"
}

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

    # analyse filenames without extensions and digits (and trimmed of surrounding 
    # whitespace), so that irrelevant chars don't confuse textcat
    my $strictfilemeta = $filemeta;
    $strictfilemeta =~ s/\.[^\.]+$//g;
    $strictfilemeta =~ s/\d//g;
    $strictfilemeta =~ s/^\s*//g;
    $strictfilemeta =~ s/\s*$//g;
    
##    print STDERR "**** strict filename is |$strictfilemeta|\n";
    my $filename_encoding = $self->encoding_from_language_analysis($strictfilemeta);
    if(!defined $filename_encoding) {
    $filename_encoding = "undefined";
    }

##    print STDERR "**** textcat found filename encoding: " . $file_textcat_encoding_map{$strictfilemeta} . "\n";
    return $filename_encoding; # can be the string "undefined"
}

# performs textcat
sub encoding_from_language_analysis {
    my $self = shift(@_);
    my ($text) = @_;

    my $outhandle = $self->{'outhandle'};
    my $best_encoding = undef;
   
    # get the language/encoding of the file using textcat
    $self->{'textcat'} = new textcat() unless defined($self->{'textcat'});
    #my $results = $self->{'textcat'}->classify(\$text);
    my $results = $self->{'textcat'}->classify_cached(\$text);


    if (scalar @$results < 0) { 
    print STDERR "**** Textcat returned 0 results\n";
    return undef;
    }
    
    print STDERR "**** TEXTCAT RESULTS for $text: ";
    print STDERR join(",", @$results);
    print STDERR "\n";

    # We have some results, we choose the first
    my ($language, $encoding) = $results->[0] =~ /^([^-]*)(?:-(.*))?$/;
    
    $best_encoding = $encoding;
    if (!defined $best_encoding) {
##  print STDERR "**** Textcat cannot determine encoding of filename: it's undefined.\n";
    return undef;
    } 
        
    if (defined $best_encoding && $best_encoding =~ m/^iso_8859/ && &unicode::check_is_utf8($text)) {
    # the text is valid utf8, so assume that's the real encoding (since textcat is based on probabilities)
##  print STDERR "*** Filename turns out to be UTF8\n";
    $best_encoding = 'utf8';
    }
    
    
    # check for equivalents where textcat doesn't have some encodings...
    # eg MS versions of standard encodings
    if (defined $best_encoding && $best_encoding =~ /^iso_8859_(\d+)/) {
##  print STDERR "**** best_encoding is ISO_8859: $best_encoding\n";

    my $iso = $1; # which variant of the iso standard?
    # iso-8859 sets don't use chars 0x80-0x9f, windows codepages do
    if ($text =~ /[\x80-\x9f]/) {
##      print STDERR "**** best_encoding is some windows value: $best_encoding\n";
        # Western Europe
        if ($iso == 1 or $iso == 15) { $best_encoding = 'windows_1252' }
        elsif ($iso == 2) {$best_encoding = 'windows_1250'} # Central Europe
        elsif ($iso == 5) {$best_encoding = 'windows_1251'} # Cyrillic
        elsif ($iso == 6) {$best_encoding = 'windows_1256'} # Arabic
        elsif ($iso == 7) {$best_encoding = 'windows_1253'} # Greek
        elsif ($iso == 8) {$best_encoding = 'windows_1255'} # Hebrew
        elsif ($iso == 9) {$best_encoding = 'windows_1254'} # Turkish
##      print STDERR "**** best_encoding windows value: $best_encoding\n";
    }
    }
    
    if (defined $best_encoding && $best_encoding !~ /^(ascii|utf8|unicode)$/ &&
    !defined $encodings::encodings->{$best_encoding}) 
    {
    if ($self->{'verbosity'}) { 
        gsprintf($outhandle, "BasePlugin: {ReadTextFile.unsupported_encoding}\n", $text, $best_encoding, "undef");
    }
##  print STDERR "***** unsupported encoding: $best_encoding. Setting it to undefined.\n";  
    $best_encoding = undef;
    }
##    print STDERR "**** language: $language\n" if defined $language;
##    print STDERR "**** encoding: $best_encoding\n" if defined $encoding;
    
    return $best_encoding;
}

# uses locale
sub get_filesystem_encoding {

    my $self = shift(@_);

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

    eval {
    use POSIX qw(locale_h);
    
    # With only one parameter, setlocale retrieves the 
    # current value
    my $current_locale = setlocale(LC_CTYPE);
    
    if ($current_locale =~ m/^.*\.(.*?)$/) {
        my $char_encoding = lc($1);
        if ($char_encoding =~ m/^(iso)(8859)(\d{1,2})$/) {
        $char_encoding = "$1\_$2\_$3";
        }

        $char_encoding =~ s/-/_/g;
        $char_encoding =~ s/^utf_8$/utf8/;
        
        if ($char_encoding =~ m/^\d+$/) {
        if (defined $encodings::encodings->{"windows_$char_encoding"}) {
            $char_encoding = "windows_$char_encoding";
        }
        elsif (defined $encodings::encodings->{"dos_$char_encoding"}) {
            $char_encoding = "dos_$char_encoding";
        }
        }
        
        if (($char_encoding =~ m/(?:ascii|utf8|unicode)/) 
        || (defined $encodings::encodings->{$char_encoding})) {
        $filesystem_encoding = $char_encoding;
        }
        else {
        print $outhandle "Warning: Unsupported character encoding '$char_encoding' from locale '$current_locale'\n";
        }
    }
    

    };
    if ($@) {
    print $outhandle "$@\n";
    print $outhandle "Warning: Unable to establish locale.  Will assume filesytem is UTF-8\n";
    
    }
    return $filesystem_encoding;
}

# is there ever only one Source? Sometimes this will be called twice, for images etc that are converted.
sub set_Source_metadata {
    my $self = shift (@_);  
    my ($doc_obj, $filename_no_path, $file_encoding) = @_;

    my $top_section = $doc_obj->get_top_section();
    
    # UTF-8 version of filename
    my $filemeta = $self->filename_to_utf8_metadata($filename_no_path, $file_encoding);
    $doc_obj->set_utf8_metadata_element($top_section, "Source", $filemeta);

}
     
sub add_OID {

    my $self = shift (@_);  
    my ($doc_obj) = @_;

    # See if a metadata field is specified as the field
    if ((defined $self->{'use_as_doc_identifier'}) && ($self->{'use_as_doc_identifier'} ne "")) {
    my $metadata_doc_id = $self->{'use_as_doc_identifier'};

    # Consider "tidying" up metadata_doc_id to be something
    # suitable in a URL
    # Could even support a user specified plugin RE for this.

    my $top_section = $doc_obj->get_top_section();
    my $oid = $doc_obj->get_metadata_element($top_section,$metadata_doc_id);
##  print STDERR "**** oid = $oid\n";
        $doc_obj->set_OID($oid);
    }
    # See if there is a plugin-specific set_OID function...
    elsif (defined ($self->can('set_OID'))) {
    # it will need $doc_obj to set the Identifier metadata...
    $self->set_OID(@_); # pass through any extra arguments supplied
    } else {
    # use the default set_OID() in doc.pm
    $doc_obj->set_OID();
    }
}



# The BasePlugin read_into_doc_obj() function. This function does all the
# right things to make general options work for a given plugin.  It doesn't do anything with the file other than setting reads in
# a file and sets up a slew of metadata all saved in doc_obj, which
# it then returns as part of a tuple (process_status,doc_obj)
#
# Much of this functionality used to reside in read, but it was broken
# down into a supporting routine to make the code more flexible.  
#
# recursive plugins (e.g. RecPlug) and specialized plugins like those
# capable of processing many documents within a single file (e.g.
# GMLPlug) will normally want to implement their own version of
# read_into_doc_obj()
#
# Note that $base_dir might be "" and that $file might 
# include directories

# currently blocking has been done before it gets here - does this affect secondary plugin stuff??
sub read_into_doc_obj {
    my $self = shift (@_);  
    my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;

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

    # should we move this to read? What about secondary plugins?
    print STDERR "<Processing n='$file' p='$self->{'plugin_type'}'>\n" if ($gli);
    print $outhandle "$self->{'plugin_type'} processing $file\n"
        if $self->{'verbosity'} > 1;

    my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
    # create a new document
    my $doc_obj = new doc ($filename_full_path, "indexed_doc");
    my $top_section = $doc_obj->get_top_section();

    # this should look at the plugin option too...
    $doc_obj->set_OIDtype ($processor->{'OIDtype'}, $processor->{'OIDmetadata'});    
    $doc_obj->add_utf8_metadata($top_section, "Plugin", "$self->{'plugin_type'}");
    $doc_obj->add_utf8_metadata($top_section, "FileSize", (-s $filename_full_path));
 
    $self->set_Source_metadata($doc_obj, $filename_no_path);

    # plugin specific stuff - what args do we need here??
    unless (defined ($self->process($pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli))) {
    print STDERR "<ProcessingError n='$file'>\n" if ($gli);
    return -1;
    }
    
    # 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();
    # can we merge these two methods??
    $self->add_associated_files($doc_obj, $filename_full_path);
    $self->extra_metadata ($doc_obj, $section, $metadata);
    $self->auto_extract_metadata($doc_obj);

    # if we haven't found any Title so far, assign one
    # this was shifted to here from inside read()
    $self->title_fallback($doc_obj,$section,$filename_no_path);
    
    $self->add_OID($doc_obj);
    
    return (1,$doc_obj);
}

sub add_dummy_text {
    my $self = shift(@_);
    my ($doc_obj, $section) = @_;

    # add NoText metadata so we can hide this dummy text in format statements
    $doc_obj->add_metadata($section, "NoText", "1");
    $doc_obj->add_text($section, &gsprintf::lookup_string("{BasePlugin.dummy_text}",1));
    
}

# does nothing. Can be overridden by subclass
sub auto_extract_metadata {
    my $self = shift(@_);
    my ($doc_obj) = @_;
}

# adds cover image, associate_file options stuff. Should be called by sub class
# read_into_doc_obj
sub add_associated_files {
    my $self = shift(@_);
    # whatis filename??
    my ($doc_obj, $filename) = @_;
    
    # add in the cover image
    if ($self->{'cover_image'}) {
    $self->associate_cover_image($doc_obj, $filename);
    }
   

}

# implement this if you are extracting metadata for other documents
sub metadata_read {
    my $self = shift (@_);
    my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $extrametakeys, $extrametadata, $processor, $maxdocs, $gli) = @_;
    
    # can we process this file??
    my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
    return undef unless $self->can_process_this_file($filename_full_path);

    return 1; # we recognise the file, but don't actually do anything with it
}


# The BasePlugin read() function. This function calls read_into_doc_obj()
# to ensure all the right things to make general options work for a
# given plugin are done. It then 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.  
#
# recursive plugins (e.g. RecPlug) and specialized plugins like those
# capable of processing many documents within a single file (e.g.
# GMLPlug) might want to implement their own version of read(), but
# more likely need to implement their own version of read_into_doc_obj()
#
# Return number of files processed, undef if can't recognise, -1 if can't
# process

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

    # can we process this file??
    my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
    return undef unless $self->can_process_this_file($filename_full_path);
    
    my ($process_status,$doc_obj) = $self->read_into_doc_obj(@_);
   
    if ((defined $process_status) && ($process_status == 1)) {

    # process the document
    $processor->process($doc_obj);

    $self->{'num_processed'} ++;
    undef $doc_obj; 
    }
    # delete any temp files that we may have created
    $self->clean_up_after_doc_obj_processing();

    # if process_status == 1, then the file has been processed.
    return $process_status;

}

# returns undef if file is rejected by the plugin
sub process {
    my $self = shift (@_);
    my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;

    gsprintf(STDERR, "BasePlugin::process {common.must_be_implemented}\n") && die "\n";

    return undef; # never gets here
}

# overwrite this method to delete any temp files that we have created
sub clean_up_after_doc_obj_processing {
    my $self = shift(@_);

}

# write_file -- used by ConvertToPlug, for example in post processing
#
# where should this go, is here the best place??
sub utf8_write_file {
    my $self = shift (@_);
    my ($textref, $filename) = @_;
    
    if (!open (FILE, ">$filename")) {
    gsprintf(STDERR, "ConvertToPlug::write_file {ConvertToPlug.could_not_open_for_writing} ($!)\n", $filename);
     die "\n";
     }
    print FILE $$textref;
    
    close FILE;
}


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

    my $file_derived_title = $file;
    $file_derived_title =~ s/_/ /g;
    $file_derived_title =~ s/\..*?$//;

    return $file_derived_title;
}


sub title_fallback
{
    my $self = shift (@_);
    my ($doc_obj,$section,$file) = @_;

    if (!defined $doc_obj->get_metadata_element ($section, "Title") or $doc_obj->get_metadata_element($section, "Title") eq "") {

    my $file_derived_title = $self->filename_to_utf8_metadata($self->filename_based_title($file));
    if (!defined $doc_obj->get_metadata_element ($section, "Title")) {
        $doc_obj->add_utf8_metadata ($section, "Title", $file_derived_title);
    }
    else {
        $doc_obj->set_utf8_metadata ($section, "Title", $file_derived_title);
    } 
    }
        
}
  
# add any extra metadata that's been passed around from one
# plugin to another.
# extra_metadata uses add_utf8_metadata so it expects metadata values
# to already be in utf8
sub extra_metadata {
    my $self = shift (@_);
    my ($doc_obj, $cursection, $metadata) = @_;

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

    foreach my $field (keys(%$metadata)) {
    # $metadata->{$field} may be an array reference
    if ($field eq "gsdlassocfile_tobe") {
        # 'gsdlassocfile_tobe' is artificially introduced metadata
        # that is used to signal that certain additional files should
        # be tied to this document.  Useful in situations where a
        # metadata pass in the plugin pipeline works out some files
        # need to be associated with a document, but the document hasn't
        # been formed yet.
        my $equiv_form = "";
        foreach my $gaf (@{$metadata->{$field}}) {
        my ($full_filename,$mimetype) = ($gaf =~ m/^(.*):(.*):$/);
        my ($tail_filename) = ($full_filename =~ /^.*[\/\\](.+?)$/);
        my $filename = $full_filename;
        $doc_obj->associate_file($full_filename,$tail_filename,$mimetype);

        # work out extended tail extension (i.e. matching tail re)

        my ($file_prefix,$file_extended_ext) 
            = &util::get_prefix_and_tail_by_regex($tail_filename,$associate_tail_re);
        my ($pre_doc_ext) = ($file_extended_ext =~ m/^(.*)\..*$/);

        my ($doc_ext) = ($tail_filename =~ m/^.*\.(.*)$/);
        my $start_doclink = "<a href=\"_httpprefix_/collect/[collection]/index/assoc/{Or}{[parent(Top):archivedir],[archivedir]}/$tail_filename\">";
        my $srcicon = "_icon".$doc_ext."_";
        my $end_doclink = "</a>";

        my $assoc_form = "$start_doclink\{If\}{$srcicon,$srcicon,$doc_ext\}$end_doclink";

        if (defined $pre_doc_ext) {
            # for metadata such as [mp3._edited] [mp3._full] ...
            $doc_obj->add_utf8_metadata ($cursection, "$doc_ext.$pre_doc_ext", $assoc_form); 
        }

        # for multiple metadata such as [mp3.assoclink]
        $doc_obj->add_utf8_metadata ($cursection, "$doc_ext.assoclink", $assoc_form); 
        
        $equiv_form .= " $assoc_form";  
        }
        $doc_obj->add_utf8_metadata ($cursection, "equivlink", $equiv_form); 
    }
    elsif (ref ($metadata->{$field}) eq "ARRAY") {
        map { 
        $doc_obj->add_utf8_metadata ($cursection, $field, $_); 
        } @{$metadata->{$field}};
    } else {
        $doc_obj->add_utf8_metadata ($cursection, $field, $metadata->{$field}); 
    }
    }
}


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

    $stats->{'num_processed'} += $self->{'num_processed'};
    $stats->{'num_not_processed'} += $self->{'num_not_processed'};
    $stats->{'num_archives'} += $self->{'num_archives'};

}

sub associate_cover_image {
    my $self = shift;
    my ($doc_obj, $filename) = @_;

    $filename =~ s/\.[^\\\/\.]+$/\.jpg/;
    if (exists $self->{'covers_missing_cache'}->{$filename}) {
    # don't stat() for existence eg for multiple document input files
    # (eg SplitPlug)
    return;
    }

    my $top_section=$doc_obj->get_top_section();

    if (-e $filename) {
        $doc_obj->associate_file($filename, "cover.jpg", "image/jpeg");
    $doc_obj->add_utf8_metadata($top_section, "hascover",  1);
    } else {
    my $upper_filename = $filename;
    $upper_filename =~ s/jpg$/JPG/;
    if (-e $upper_filename) {
        $doc_obj->associate_file($upper_filename, "cover.jpg",
                     "image/jpeg");
        $doc_obj->add_utf8_metadata($top_section, "hascover",  1);
    } else {
        # file doesn't exist, so record the fact that it's missing so
        # we don't stat() again (stat is slow)
        $self->{'covers_missing_cache'}->{$filename} = 1;
    }
    }

}


# Overridden by exploding plugins (eg. ISISPlug)
sub clean_up_after_exploding
{
    my $self = shift(@_);
}



1;