source: gsdl/trunk/perllib/basebuildproc.pm@ 15695

###########################################################################
#
# basebuildproc.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 indexing (should be 
# implemented by subclass) and storing in the database

package basebuildproc;

eval {require bytes};

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

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

sub new()
  {
    my ($class, $collection, $source_dir, $build_dir, $keepold, $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, database etc.)
    $outhandle = STDERR unless defined $outhandle;

    $self->{'collection'} = $collection;
    $self->{'source_dir'} = $source_dir;
    $self->{'build_dir'}  = $build_dir;
    $self->{'keepold'}    = $keepold;
    $self->{'verbosity'}  = $verbosity;
    $self->{'outhandle'}  = $outhandle;

    $self->{'classifiers'} = [];
    $self->{'mode'} = "text";
    $self->{'assocdir'} = $build_dir;
    $self->{'dontdb'} = {};

    $self->{'index'} = "section:text";
    $self->{'indexexparr'} = [];

    my $found_num_data = 0;
    my $buildconfigfile = undef;

    if ($keepold) {
    # For incremental building need to seed num_docs etc from values
    # stored in build.cfg (if present)
      print STDERR "Keepold!\n";
    $buildconfigfile = &util::filename_cat($build_dir, "build.cfg");
      print STDERR "Build cfg: $buildconfigfile\n";
    if (-e $buildconfigfile) {
        $found_num_data = 1;
    }
    else {
        # try the index dir
        $buildconfigfile = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, 
                           "index", "build.cfg");
            print STDERR "Index cfg: $buildconfigfile\n";
        if (-e $buildconfigfile) {
        $found_num_data = 1;
        }
    }

    }
    #else
    #  {
    #    print STDERR "Removeold!\n";
    #  }

    if ($found_num_data)
      {
        #print STDERR "Found_Num_Data!\n";
    my $buildcfg = &colcfg::read_build_cfg($buildconfigfile);
    $self->{'starting_num_docs'}     = $buildcfg->{'numdocs'};
        #print STDERR "- num_docs:     $self->{'starting_num_docs'}\n";
    $self->{'starting_num_sections'} = $buildcfg->{'numsections'};
        #print STDERR "- num_sections: $self->{'starting_num_sections'}\n";
    $self->{'starting_num_bytes'}    = $buildcfg->{'numbytes'};
        #print STDERR "- num_bytes:    $self->{'starting_num_bytes'}\n";
    }
    else
      {
        #print STDERR "NOT Found_Num_Data!\n";
        $self->{'starting_num_docs'}     = 0;
    $self->{'starting_num_sections'} = 0;
    $self->{'starting_num_bytes'}    = 0;
      }

    $self->{'output_handle'} = "STDOUT";
    $self->{'num_docs'}      = $self->{'starting_num_docs'};
    $self->{'num_sections'}  = $self->{'starting_num_sections'};
    $self->{'num_bytes'}     = $self->{'starting_num_bytes'};

    $self->{'num_processed_bytes'} = 0;
    $self->{'store_text'} = 1;

    # what level (section/document) the database - indexer intersection is
    $self->{'db_level'} = "section";
    #used by browse interface
    $self->{'doclist'} = [];

    $self->{'indexing_text'} = 0;

    return bless $self, $class;

}

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

    $self->{'num_docs'}      = $self->{'starting_num_docs'};
    $self->{'num_sections'}  = $self->{'starting_num_sections'};
    $self->{'num_bytes'}     = $self->{'starting_num_bytes'};
    
    $self->{'num_processed_bytes'} = 0;
}

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

    $self->{'num_docs'}      = 0;
    $self->{'num_sections'}  = 0;
    $self->{'num_bytes'}     = 0;
    
    $self->{'num_processed_bytes'} = 0;
}

sub is_incremental_capable
{
    # By default we return 'no' as the answer
    # Safer to assume non-incremental to start with, and then override in
    # inherited classes that are.

    return 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 get_mode {
    my $self = shift (@_);

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

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

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

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

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

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

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

sub set_index_languages {
    my $self = shift (@_);
    my ($lang_meta, $langarr) = @_;
    $self->{'lang_meta'} = $lang_meta;
    $self->{'langarr'} = $langarr;
}

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'}};
}

# the standard database level is section, but you may want to change it to document
sub set_db_level {
    my $self= shift (@_);
    my ($db_level) = @_;

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

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

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



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

    # Keep some statistics relating to metadata sets used and
    # frequency of particular metadata fields within each set

    # Union of metadata prefixes and frequency of fields
    # (both scoped for this document alone, and across whole collection)
    
    if ($field =~ m/^(.+)\.(.*)$/) {
    my $prefix = $1;
    my $core_field = $2;

    $self->{'doc_mdprefix_fields'}->{$prefix}->{$core_field}++;
    $self->{'mdprefix_fields'}->{$prefix}->{$core_field}++;
    }
    elsif ($field =~ m/^[[:upper:]]/) {
    # implicit 'ex' metadata set

    $self->{'doc_mdprefix_fields'}->{'ex'}->{$field}++;
    $self->{'mdprefix_fields'}->{'ex'}->{$field}++;
    }

}


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

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

    # only output this document if it is a "indexed_doc" or "info_doc" (database only) document
    return if ($doctype ne "indexed_doc" && $doctype ne "info_doc");

    my $archivedir = "";

    if (defined $filename)
    {
    # doc_obj derived directly from file

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

    $archivedir = $dir;

    # resolve the final filenames of the files associated with this document
    $self->assoc_files ($doc_obj, $archivedir);
    }
    else
    {
    # doc_obj reconstructed from database (has metadata, doc structure but no text)
    my $top_section = $doc_obj->get_top_section();
    $archivedir = $doc_obj->get_metadata_element($top_section,"archivedir");
    }


    #GRB: moved 1/06/2004 from GRB01062004
    #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);
    #GRB: end of moved block

    # 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 = "";

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

    while (defined $section)
    {
    my $section_OID = $doc_OID;
    if ($section ne "")
    {
        $section_OID = $doc_OID . "." . $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
    print $handle "[$section_OID]\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 reconstructed from database file then no need to 
    # explicitly add <hastxt> as this is preserved as metadata when
    # the database file is loaded in

    if (defined $filename)
    {
        # doc_obj derived directly from file
        if ($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";
            $url .= "<section>$section_OID\n";
                    $url .= '-' x 70 . "\n";
        }

        if (!defined $self->{'dontdb'}->{$field}) {
            print $handle "<$field>$value\n";

            if ($section eq "")
            {
            $self->infodb_metadata_stats($field);
            }
        }
        }
    }

    if ($section eq "")
    {
        my $doc_mdprefix_fields = $self->{'doc_mdprefix_fields'};

        foreach my $prefix (keys %$doc_mdprefix_fields)
        {
        print $handle "<metadataset>$prefix\n";

        foreach my $field (keys %{$doc_mdprefix_fields->{$prefix}})
        {
            my $val = $doc_mdprefix_fields->{$prefix}->{$field};

            print $handle "<metadatalist-$prefix>$field\n";
            print $handle "<metadatafreq-$prefix-$field>$val\n";
        }

        }
    }

    # If doc_obj reconstructed from database file then no need to 
    # explicitly add <archivedir> as this is preserved as metadata when
    # the database file is loaded in

    if (defined $filename)
    {
        # output archivedir if at top level
        if ($section eq $doc_obj->get_top_section()) {
        print $handle "<archivedir>$archivedir\n";
        }
    }

    # output document display type
    if ($first) {
        print $handle "<thistype>$thistype\n";
    }


    if ($self->{'db_level'} eq "document") {
        # doc num is num_docs not num_sections
        # output the matching document number
        print $handle "<docnum>$self->{'num_docs'}\n";
            
    } else {
        # 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";
            }
        }
        print $handle "\n";
        }
        #output the matching doc number
        print $handle "<docnum>$self->{'num_sections'}\n";
        
    } 
    
    print $handle '-' x 70, "\n";

    
    # output a database entry for the document number
    if ($self->{'db_level'} eq "document") {
        print $handle "[$self->{'num_docs'}]\n";
        print $handle "<section>$doc_OID\n";
    }
    else {
        print $handle "[$self->{'num_sections'}]\n";
        print $handle "<section>$section_OID\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);
    last if ($self->{'db_level'} eq "document"); # if no sections wanted, only add the docs
    }

    #GRB01062004: see code above moved from here
}


sub text {
    my $self = shift (@_);
    my ($doc_obj) = @_;
    
    my $handle = $self->{'outhandle'};
    print $handle "basebuildproc::text function must be implemented in sub classes\n";
    die "\n";
}

# should the document be indexed - according to the subcollection and language
# specification.
sub is_subcollection_doc {
    my $self = shift (@_);
    my ($doc_obj) = @_;
    
    my $indexed_doc = 1;
    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 $bool eq '!';
        my @metadata_values;
        if ($field =~ /^filename$/i) {
        push(@metadata_values, $doc_obj->get_source_filename());
        }
        else {
        @metadata_values = @{$doc_obj->get_metadata($doc_obj->get_top_section(), $field)};
        }
        next unless @metadata_values;
        foreach my $metadata_value (@metadata_values) {
        if ($bool eq '!') {
            if ($options =~ /^i$/i) {
            if ($metadata_value !~ /$exp/i) {$indexed_doc = 1; last;}
            } else {
            if ($metadata_value !~ /$exp/) {$indexed_doc = 1; last;}
            }
        } else {
            if ($options =~ /^i$/i) {
            if ($metadata_value =~ /$exp/i) {$indexed_doc = 1; last;}
            } else {
            if ($metadata_value =~ /$exp/) {$indexed_doc = 1; last;}
            }
        }
        }

        last if ($indexed_doc == 1);
    }
    }
    
    # if this doc is so far in the sub collection, and we have lang info, 
    # now we check the languages to see if it matches
    if($indexed_doc && defined $self->{'lang_meta'}) {
    $indexed_doc = 0;
    my $field = $doc_obj->get_metadata_element($doc_obj->get_top_section(), $self->{'lang_meta'});
    if (defined $field) {
        foreach my $lang (@{$self->{'langarr'}}) {
        my ($bool) = $lang =~ /^(.)/;
        if ($bool eq '!') {
            $lang =~ s/^.//;
            if ($field !~ /$lang/) {
            $indexed_doc = 1; last;
            }
        } else {
            if ($field =~ /$lang/) {
            $indexed_doc = 1; last;
            }
        }
        }
    } 
    }
    return $indexed_doc;
    
}

# use 'Paged' if document has no more than 2 levels
# and each section at second level has a number for
# Title metadata
# also use Paged if gsdlthistype metadata is set to Paged
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 $gsdlthistype = $doc_obj->get_metadata_element ($section, "gsdlthistype");
    if (defined $gsdlthistype) {
    if ($gsdlthistype eq "Paged") {
        $childtype = "Paged";
        if ($doc_obj->get_text_length ($doc_obj->get_top_section())) {
        $thistype = "Paged";
        } else {
        $thistype = "Invisible";
        }
        
        return ($thistype, $childtype);
    } elsif ($gsdlthistype eq "Hierarchy") {
        return ($thistype, $childtype); # use VList, VList
    }
    }
    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()}) {
      #rint STDERR "Processing associated file - copy " . $assoc_file->[0] . " to " . $assoc_file->[1] . "\n";
    # if assoc file starts with a slash, we put it relative to the assoc
    # dir, otherwise it is relative to the HASH... directory
    if ($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);
    }
}