source: trunk/gsdl/perllib/mgbuildproc.pm@ 3665

###########################################################################
#
# mgbuildproc.pm -- 
# 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.
#
###########################################################################

# This document processor outputs a document
# for mg to process


package mgbuildproc;

use classify;
use doc;
use docproc;
use util;

BEGIN {
    @ISA = ('docproc');
}


sub new {
    my ($class, $collection, $source_dir, $build_dir,
    $verbosity, $outhandle) = @_;
    my $self = new docproc ();

    # outhandle is where all the debugging info goes
    # output_handle is where the output of the plugins is piped
    # to (i.e. mg, gdbm etc.)
    $outhandle = STDERR unless defined $outhandle;

    $self->{'collection'} = $collection;
    $self->{'source_dir'} = $source_dir;
    $self->{'build_dir'} = $build_dir;
    $self->{'verbosity'} = $verbosity;
    $self->{'classifiers'} = [];
    $self->{'mode'} = "text";
    $self->{'assocdir'} = $build_dir;
    $self->{'dontgdbm'} = {};
    $self->{'index'} = "section:text";
    $self->{'indexexparr'} = [];
    $self->{'output_handle'} = "STDOUT";
    $self->{'num_docs'} = 0;
    $self->{'num_sections'} = 0;
    $self->{'num_bytes'} = 0;
    $self->{'num_processed_bytes'} = 0;
    $self->{'store_text'} = 1;
    $self->{'outhandle'} = $outhandle;
    
    #used by browse interface
    $self->{'doclist'} = [];
    
    $self->{'indexing_text'} = 0;

    return bless $self, $class;
}

sub reset {
    my $self = shift (@_);
    
    $self->{'num_docs'} = 0;
    $self->{'num_sections'} = 0;
    $self->{'num_processed_bytes'} = 0;
    $self->{'num_bytes'} = 0;
}

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

    return $self->{'num_docs'};
}

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

    return $self->{'num_sections'};
}

# num_bytes is the actual number of bytes in the collection
# this is normally the same as what's processed during text compression
sub get_num_bytes {
    my $self = shift (@_);

    return $self->{'num_bytes'};
}

# num_processed_bytes is the number of bytes actually passed
# to mg for the current index
sub get_num_processed_bytes {
    my $self = shift (@_);

    return $self->{'num_processed_bytes'};
}

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

    $self->{'output_handle'} = $handle;
}

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

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

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

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

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

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

sub set_index {
    my $self = shift (@_);
    my ($index, $indexexparr) = @_;

    $self->{'index'} = $index;
    $self->{'indexexparr'} = $indexexparr if defined $indexexparr;
}

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

    return $self->{'index'};
}

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

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

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

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

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

    return $self->{'indexing_text'};
}

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

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

sub get_doc_list {
    my $self = shift(@_);
    
    return @{$self->{'doclist'}};
}


sub process {
    my $self = shift (@_);
    my $method = $self->{'mode'};

    $self->$method(@_);
}

# use 'Paged' if document has no more than 2 levels
# and each section at second level has a number for
# Title metadata
sub get_document_type {
    my $self = shift (@_);
    my ($doc_obj) = @_;

    my $thistype = "VList";
    my $childtype = "VList";
    my $title;
    my @tmp = ();
    
    my $section = $doc_obj->get_top_section ();
    my $first = 1;
    while (defined $section) {
    @tmp = split /\./, $section;
    if (scalar(@tmp) > 1) {
        return ($thistype, $childtype);
    }
    if (!$first) {
        $title = $doc_obj->get_metadata_element ($section, "Title");
        if (!defined $title || $title !~ /^\d+$/) {
        return ($thistype, $childtype);
        }
    }
    $first = 0;
    $section = $doc_obj->get_next_section($section);
    }
    if ($doc_obj->get_text_length ($doc_obj->get_top_section())) {
    $thistype = "Paged";
    } else {
    $thistype = "Invisible";
    }
    $childtype = "Paged";
    return ($thistype, $childtype);
}

sub assoc_files {
    my $self = shift (@_);
    my ($doc_obj, $archivedir) = @_;
    my ($afile);
    
    foreach my $assoc_file (@{$doc_obj->get_assoc_files()}) {
    # if assoc file starts with a slash, we put it relative to the assoc
    # dir, otherwise it is relative to the HASH... directory
    if ($assoc_file->[1] =~ m@^[/\\]@) {
        $afile = &util::filename_cat($self->{'assocdir'},$assoc_file->[1]);
    } else {
        $afile = &util::filename_cat($self->{'assocdir'}, $archivedir, $assoc_file->[1]);
    }
    &util::hard_link ($assoc_file->[0], $afile);
    }
}

sub infodb {
    my $self = shift (@_);
    my ($doc_obj, $filename) = @_;
    my $handle = $self->{'output_handle'};
#    $handle = "main::STDOUT";

    my $doctype = $doc_obj->get_doc_type();

    # only output this document if it is one to be indexed
    return if ($doctype ne "indexed_doc");

    my ($archivedir) = $filename =~ /^(.*?)(?:\/|\\)[^\/\\]*$/;
    $archivedir = "" unless defined $archivedir;
    $archivedir =~ s/\\/\//g;
    $archivedir =~ s/^\/+//;
    $archivedir =~ s/\/+$//;

    $self->assoc_files ($doc_obj, $archivedir);

    # this is another document
    $self->{'num_docs'} += 1 unless ($doctype eq "classification");

    # is this a paged or a hierarchical document
    my ($thistype, $childtype) = $self->get_document_type ($doc_obj);

    my $section = $doc_obj->get_top_section ();
    my $doc_OID = $doc_obj->get_OID();
    my $first = 1;
    my $url = "";
    while (defined $section) {
    # update a few statistics
    $self->{'num_bytes'} += $doc_obj->get_text_length ($section);
    $self->{'num_sections'} += 1 unless ($doctype eq "classification");

    # output the section name
    if ($section eq "") { print $handle "[$doc_OID]\n"; }
    else { print $handle "[$doc_OID.$section]\n"; }

    # output the fact that this document is a document (unless doctype
    # has been set to something else from within a plugin
    my $dtype = $doc_obj->get_metadata_element ($section, "doctype");
    if (!defined $dtype || $dtype !~ /\w/) {
        print $handle "<doctype>doc\n";
    }

    # output whether this node contains text
    if ($doc_obj->get_text_length($section) > 0) {
        print $handle "<hastxt>1\n";
    } else {
        print $handle "<hastxt>0\n";
    }

    # output all the section metadata
    my $metadata = $doc_obj->get_all_metadata ($section);
    foreach my $pair (@$metadata) {
        my ($field, $value) = (@$pair);

        if ($field ne "Identifier" && $field !~ /^gsdl/ && 
        defined $value && $value ne "") {

        # escape problematic stuff
        $value =~ s/\\/\\\\/g;
        $value =~ s/\n/\\n/g;
        $value =~ s/\r/\\r/g;
        if ($value =~ /-{70,}/) {
            # if value contains 70 or more hyphens in a row we need
            # to escape them to prevent txt2db from treating them
            # as a separator
            $value =~ s/-/&\#045;/gi;
        }

        # special case for URL metadata
        if ($field =~ /^URL$/i) {
                    $url .= "[$value]\n";
                    if ($section eq "") {$url .= "<section>$doc_OID\n";}
                    else {$url .= "<section>$doc_OID.$section\n";}
                    $url .= '-' x 70 . "\n";
        }

        if (!defined $self->{'dontgdbm'}->{$field}) {
            print $handle "<$field>$value\n";
        }
        }
    }
    
    # output archivedir if at top level
    if ($section eq $doc_obj->get_top_section()) {
        print $handle "<archivedir>$archivedir\n";
    }
    
    # output document display type
    if ($first) {
        print $handle "<thistype>$thistype\n";
    }

    # output a list of children
    my $children = $doc_obj->get_children ($section);
    if (scalar(@$children) > 0) {
        print $handle "<childtype>$childtype\n";
        print $handle "<contains>";
        my $firstchild = 1;
        foreach my $child (@$children) {
        print $handle ";" unless $firstchild;
        $firstchild = 0;
        if ($child =~ /^.*?\.(\d+)$/) {
            print $handle "\".$1";
        } else {
            print $handle "\".$child";
        }
#       if ($child eq "") { print $handle "$doc_OID"; }
#       elsif ($section eq "") { print $handle "$doc_OID.$child"; }
#       else { print $handle "$doc_OID.$section.$child"; }
        }
        print $handle "\n";
    }

    # output the matching document number
    print $handle "<docnum>$self->{'num_sections'}\n";

    print $handle '-' x 70, "\n";

    
    # output a database entry for the document number
    print $handle "[$self->{'num_sections'}]\n";
    if ($section eq "") { print $handle "<section>$doc_OID\n"; }
    else { print $handle "<section>$doc_OID.$section\n"; }
    print $handle '-' x 70, "\n";
    
        # output entry for url
        if ($url ne "") {
            print $handle $url;
        }

    $first = 0;
    $section = $doc_obj->get_next_section($section);
    }

    #add this document to the browse structure
    push(@{$self->{'doclist'}},$doc_obj->get_OID()) 
    unless ($doctype eq "classification");

    # classify this document
    &classify::classify_doc ($self->{'classifiers'}, $doc_obj);

}

sub find_paragraphs {
    $_[1] =~ s/(<p\b)/\cC$1/gi;
}

sub filter_text {
    # $self->filter_text ($field, $new_text);
    # don't want to do anything for this version, however,
    # in a particular collection you might want to override
    # this method to post-process certain fields depending on
    # the field, or whether we are outputting it for indexing
}

sub text {
    my $self = shift (@_);
    my ($doc_obj) = @_;
    my $handle = $self->{'output_handle'};
    my $indexed_doc = 1;

    # only output this document if it is one to be indexed
    return if ($doc_obj->get_doc_type() ne "indexed_doc");

    # see if this document belongs to this subcollection
    foreach my $indexexp (@{$self->{'indexexparr'}}) {
    $indexed_doc = 0;
    my ($field, $exp, $options) = split /\//, $indexexp;
    if (defined ($field) && defined ($exp)) {
        my ($bool) = $field =~ /^(.)/;
        $field =~ s/^.// if $bool eq '!';
        if ($field =~ /^filename$/i) {
        $field = $doc_obj->get_source_filename();
        } else {
        $field = $doc_obj->get_metadata_element($doc_obj->get_top_section(), $field);
        }
        next unless defined $field;
        if ($bool eq '!') {
        if ($options =~ /^i$/i) {
            if ($field !~ /$exp/i) {$indexed_doc = 1; last;}
        } else {
            if ($field !~ /$exp/) {$indexed_doc = 1; last;}
        }
        } else {
        if ($options =~ /^i$/i) {
            if ($field =~ /$exp/i) {$indexed_doc = 1; last;}
        } else {
            if ($field =~ /$exp/) {$indexed_doc = 1; last;}
        }
        }
    }
    }

    # this is another document
    $self->{'num_docs'} += 1;

    # get the parameters for the output
    my ($level, $fields) = split (/:/, $self->{'index'});
    $fields =~ s/\ball\b/Title,Creator,text/;
    $fields =~ s/\btopall\b/topTitle,topCreator,toptext/;

    my $doc_section = 0; # just for this document
    my $text = "";
    my $text_extra = "";

    # get the text for this document
    my $section = $doc_obj->get_top_section();
    while (defined $section) {
    # update a few statistics
    $doc_section++;
    $self->{'num_sections'} += 1;

    if ($indexed_doc) {
        $self->{'num_bytes'} += $doc_obj->get_text_length ($section);
        foreach my $field (split (/,/, $fields)) {
        # only deal with this field if it doesn't start with top or
        # this is the first section
        my $real_field = $field;
        if (!($real_field =~ s/^top//) || ($doc_section == 1)) {
            my $new_text = "";
            if ($real_field eq "text") {
            $new_text = $doc_obj->get_text ($section) if $self->{'store_text'};
            $self->{'num_processed_bytes'} += length ($new_text);
            $new_text =~ s/[\cB\cC]//g;
            $self->find_paragraphs($new_text);
            
            } else {
            my $first = 1;
            foreach $meta (@{$doc_obj->get_metadata ($section, $real_field)}) {
                $meta =~ s/[\cB\cC]//g;
                $self->{'num_processed_bytes'} += length ($meta);
                $new_text .= "\cC" unless $first;
                $new_text .= $meta if $self->{'store_text'};
                $first = 0;
            }
            }
            
            # filter the text
            $self->filter_text ($field, $new_text);

            $text .= "$new_text\cC";
        }
        }
    }
    
    if ($level eq "document") { $text_extra .= "\cB"; }
    else { $text .= "\cB"; }
    
    $section = $doc_obj->get_next_section($section);
    }

    print $handle "$text$text_extra";
}

1;