source: gsdl/trunk/perllib/classify.pm@ 15703

###########################################################################
#
# classify.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.
#
###########################################################################

# functions to handle classifiers

package classify;

require util;
require AllList;

use gsprintf;
use strict; no strict 'subs';
use unbuildutil;


sub gsprintf
{
    return &gsprintf::gsprintf(@_);
}


my $oid_to_clids = {};

sub load_classifier_for_info {
    my ($classifier) = shift @_;

    # find the classifier
    my $customclassname;
    if (defined($ENV{'GSDLCOLLECTION'}))
    {
    $customclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "custom", $ENV{'GSDLCOLLECTION'},
                                              "perllib", "classify", "${classifier}.pm");
    }
    my $colclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "perllib", "classify", "${classifier}.pm");
    my $mainclassname = &util::filename_cat($ENV{'GSDLHOME'}, "perllib", "classify", "${classifier}.pm");

    if (defined($customclassname) && -e $customclassname) { require $customclassname; }
    elsif (-e $colclassname) { require $colclassname; }
    elsif (-e $mainclassname) { require $mainclassname; }
    else { 
    &gsprintf(STDERR, "{classify.could_not_find_classifier}\n", $classifier) && die "\n";
    }
    my ($classobj);
    my $options = "-gsdlinfo";
    eval ("\$classobj = new \$classifier([],[$options])");
    die "$@" if $@;

    return $classobj;
}

sub load_classifiers {
    my ($classify_list, $build_dir, $outhandle) = @_;
    my @classify_objects = ();
    my $classify_number  = 1;

     my $colclassdir = &util::filename_cat($ENV{'GSDLCOLLECTDIR'},"perllib/classify");
     unshift (@INC, $colclassdir);
    
    foreach my $classifyoption (@$classify_list) {

    # get the classifier name
    my $classname = shift @$classifyoption;
    next unless defined $classname;

    # find the classifier
        my $customclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "custom", $ENV{'GSDLCOLLECTION'},
                                                  "perllib", "classify", "${classname}.pm");
        my $colclassname = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "perllib", "classify", "${classname}.pm");
        my $mainclassname = &util::filename_cat($ENV{'GSDLHOME'}, "perllib", "classify", "${classname}.pm");

    if (-e $customclassname) { require $customclassname; }
    elsif (-e $colclassname) { require $colclassname; }
    elsif (-e $mainclassname) { require $mainclassname; }
    else { &gsprintf(STDERR, "{classify.could_not_find_classifier}\n", $classname) && die "\n";
           # die "ERROR - couldn't find classifier \"$classname\"\n";
        }

    # create the classify object
    my ($classobj);

    my @newoptions;

    # do these first so they can be overriden by user supplied options
    push @newoptions, "-builddir", "$build_dir" if ($build_dir);
    push @newoptions, "-outhandle", "$outhandle" if ($outhandle);
    push @newoptions, "-verbosity", "2";

    # backwards compatability hack: if the classifier options are 
    # in "x=y" format, convert them to parsearg ("-x y") format.
    my ($opt, $key, $value);
    foreach $opt (@$classifyoption) {
        # if ($opt =~ /^(\w+)=(.*)$/) {
        #   push @newoptions, "-$1", $2;
        # } else {
        push @newoptions, $opt;
        #}
    }

    map { $_ = "\"$_\""; } @newoptions;
    my $options .= join (",", @newoptions);


    eval ("\$classobj = new \$classname([],[$options])");
    die "$@" if $@;

    $classobj->set_number($classify_number);
    $classify_number ++;

        # add this object to the list
    push (@classify_objects, $classobj);
    }

    my ($classobj);
    eval ("\$classobj = new AllList()");
    die "$@" if $@;
    push (@classify_objects, $classobj);

    return \@classify_objects;
}

# init_classifiers resets all the classifiers and readys them to process
# the documents.
sub init_classifiers {
    my ($classifiers) = @_;
    
    foreach my $classobj (@$classifiers) {
    $classobj->init();
    }
}



# takes a hashref containing the metadata for a gdbmfile entry, and extracts 
# the childrens numbers (from the 'contains' entry).    
# assumes format is ".1;".2;".3
sub get_children {  
    my ($doc_db_hash) = @_; 

    my $children = undef;

    my $contains = $doc_db_hash->{'contains'};
    if (defined ($contains)) {
    $contains =~ s/\@$//;  #remove trailing @
    $contains =~ s/^\"\.//; #remove initial ".
    @$children = split /\;\"\./, $contains;
    }

    return $children;
}

    
sub recurse_sections {
    my ($doc_obj, $children, $parentoid, $parentsection, $gdbm_recs) = @_;

    return if (!defined $children);

    foreach my $child (sort { $a <=> $b} @$children) {
    $doc_obj->create_named_section("$parentsection.$child");
    my $doc_db_rec = $gdbm_recs->{"$parentoid.$child"};
    my $doc_db_hash = db_rec_to_hash($doc_db_rec);

    # get child's children
    my $newchildren = &get_children($doc_db_hash); 

    # add content for current section
    add_section_content($doc_obj, "$parentsection.$child", $doc_db_hash);

    # process all the children if there are any
    if (defined ($newchildren))
    {
        recurse_sections($doc_obj, $newchildren, "$parentoid.$child", 
                 "$parentsection.$child", $gdbm_recs);
    }
    }
}                        


sub add_section_content {
    my ($doc_obj, $cursection, $doc_db_hash) = @_;
 
    foreach my $key (keys %$doc_db_hash) {
    #don't need to store these metadata
    next if $key =~ /(thistype|childtype|contains|docnum|doctype|classifytype)/i;
    # but do want things like hastxt and archivedir
    my @items = split /@/, $doc_db_hash->{$key};
    map {$doc_obj->add_metadata ($cursection, $key, $_); } @items;

    }
}


# gets all the metadata from a gdbm file entry, and puts it into a hashref
sub db_rec_to_hash {
    
    my ($gdb_str_ref) = @_;

    my $hashref = {};

    my @entries = split(/\n/, $gdb_str_ref);
    foreach my $entry (@entries) {
    my($key, $value) = ($entry =~ /^<([^>]*)>(.*?)$/ );
    $hashref->{$key} .= '@' if defined $hashref->{$key};
    $hashref->{$key} .= $value;
    
    }
   
    return $hashref;
}                     


sub reconstruct_doc_objs_metadata
{
    my ($fulldbname) = @_;

#   tie %gdbm_recs, 'GDBM_File', $fulldbname, &GDBM_WRCREAT, 0640;

    my %gdbm_recs;
    &unbuildutil::read_gdbm($fulldbname,\%gdbm_recs);


    # dig out top level doc sections
    my %top_sections = ();
    my %top_docnums = ();
    foreach my $key ( keys %gdbm_recs )
    {
    my $md_rec = $gdbm_recs{$key};
    my $md_hash = db_rec_to_hash($md_rec);

    if ((defined $md_hash->{'doctype'}) && ($md_hash->{'doctype'} eq "doc")) {
        next if ($key =~ m/\./);
        $top_sections{$key} = $md_hash;
        $top_docnums{$key} = $md_hash->{'docnum'};
    }
    }

    # for greenstone document objects based on metadata in gdbm file
    my @all_docs = ();
    # we need to make sure the documents were processed in the same order as
    # before, so sort based on their docnums
    foreach my $oid ( sort { $top_docnums{$a} <=> $top_docnums{$b} } keys %top_sections )
    {
    my $doc_db_hash = $top_sections{$oid};

    my $doc_obj = new doc();
    $doc_obj->set_OID($oid);
    my $top = $doc_obj->get_top_section();
        add_section_content ($doc_obj, $top, $doc_db_hash);
        my $children = &get_children($doc_db_hash);
        recurse_sections($doc_obj, $children, $oid, $top, \%gdbm_recs);

    push(@all_docs,$doc_obj);
    }    

#    untie %gdbm_recs;

    return \@all_docs;   
}





# classify_doc lets each of the classifiers classify a document
sub classify_doc {
    my ($classifiers, $doc_obj) = @_;
    
    foreach my $classobj (@$classifiers) {
    my $title = $classobj->{'title'};
    $classobj->classify($doc_obj);
    }
}


# output_classify_info outputs all the info needed for the classification
# to the database
sub output_classify_info
{
    my ($classifiers, $infodb_handle, $remove_empty_classifications, $gli) = @_;

    $gli = 0 unless defined $gli;

    # create a classification containing all the info
    my $classifyinfo = { 'classifyOID'=> 'browse',
             'contains' => [] };

    # get each of the classifications
    my $next_classify_num = 1;
    foreach my $classifier (@$classifiers)
    {
    my $classifier_info = $classifier->get_classify_info($gli);
    $classifier_info->{'classifyOID'} = "CL$next_classify_num" unless defined($classifier_info->{'classifyOID'});
        print STDERR "*** outputting information for classifier: $classifier_info->{'classifyOID'}\n";

    push(@{$classifyinfo->{'contains'}}, $classifier_info);
    $next_classify_num++;
    }

    &print_classify_info ($infodb_handle, $classifyinfo, "", $remove_empty_classifications);
}


sub print_classify_info
{
    my ($infodb_handle, $classifyinfo, $OID, $remove_empty_classifications) = @_;

    $OID =~ s/^\.+//; # just for good luck

    # book information is printed elsewhere
    return if (defined ($classifyinfo->{'OID'}));
 
    # don't want empty classifications
    return if (&check_contents ($classifyinfo, $remove_empty_classifications) == 0 && $remove_empty_classifications);
   
    $OID = $classifyinfo->{'classifyOID'} if defined ($classifyinfo->{'classifyOID'});

    my %classify_infodb = ();
    $classify_infodb{"doctype"} = [ "classify" ];
    $classify_infodb{"hastxt"} = [ "0" ];
    $classify_infodb{"childtype"} = [ $classifyinfo->{'childtype'} ]
    if defined $classifyinfo->{'childtype'};
    $classify_infodb{"Title"} = [ $classifyinfo->{'Title'} ]
    if defined $classifyinfo->{'Title'};
    $classify_infodb{"numleafdocs"} = [ $classifyinfo->{'numleafdocs'} ]
    if defined $classifyinfo->{'numleafdocs'};
    $classify_infodb{"thistype"} = [ $classifyinfo->{'thistype'} ]
    if defined $classifyinfo->{'thistype'};
    $classify_infodb{"parameters"} = [ $classifyinfo->{'parameters'} ]
    if defined $classifyinfo->{'parameters'};
    $classify_infodb{"supportsmemberof"} = [ $classifyinfo->{'supportsmemberof'} ]
    if defined $classifyinfo->{'supportsmemberof'};
    
    my $contains_text = "";
    my $mdoffset_text = "";
    
    my $next_subOID = 1;
    my $first = 1;
    foreach my $tempinfo (@{$classifyinfo->{'contains'}}) {
    # empty contents were made undefined by clean_contents()
    next unless defined $tempinfo;
    
    if (!defined ($tempinfo->{'classifyOID'}) ||
        $tempinfo->{'classifyOID'} ne "oai") {
        $contains_text .= ";" unless $first;
    }
    $mdoffset_text .= ";" unless $first;
    $first = 0;
    
    if (defined ($tempinfo->{'classifyOID'})) {
        if ($tempinfo->{'classifyOID'} ne "oai") {
        $contains_text .= $tempinfo->{'classifyOID'};
        }

            # Extra code for incremental building.
            # We need to store a listing of the classifiers each DOI is in
            my $clids = [];
            #rint STDERR "==1. Recording reverse lookup for " . $tempinfo->{'classifyOID'} . "==\n";
            if(defined($oid_to_clids->{$tempinfo->{'classifyOID'}})) {
              #rint STDERR "Found existing array!\n";
              $clids = $oid_to_clids->{$tempinfo->{'classifyOID'}};
            }
            #rint STDERR "Appended $OID to \"" . join(";", @{$clids}) . "\"\n";
            push(@{$clids}, $OID);
            $oid_to_clids->{$tempinfo->{'classifyOID'}} = $clids;
            #rint STDERR "Result: \"" . join(";", @{$clids}) . "\"\n";
            
        &print_classify_info ($infodb_handle, $tempinfo, $tempinfo->{'classifyOID'},
                  $remove_empty_classifications);
    } elsif (defined ($tempinfo->{'OID'})) {
        $contains_text .= $tempinfo->{'OID'};
        $mdoffset_text .= $tempinfo->{'offset'} if (defined ($tempinfo->{'offset'}));

            
              # note: we don't want to print the contents of the books
              # Extra code for incremental building.
              # We need to store a listing of the classifiers each DOI is in
              my $clids = [];
              #rint STDERR "==2. Recording reverse lookup for " . $tempinfo->{'OID'} . "==\n";
              if(defined($oid_to_clids->{$tempinfo->{'OID'}})) {
                #rint STDERR "Found existing array!\n";
                $clids = $oid_to_clids->{$tempinfo->{'OID'}};
              }
              #rint STDERR "Appended $OID to \"" . join(";", @{$clids}) . "\"\n";
              push(@{$clids}, $OID);
              $oid_to_clids->{$tempinfo->{'OID'}} = $clids;
              #rint STDERR "Result: \"" . join(";", @{$clids}) . "\"\n";


        } else {
        
        # Supress having top-level node in Collage classifier
        # so no bookshelf icon appears, top-level, along with the
        # applet
        
        if (!defined ($tempinfo->{'Title'}) || $tempinfo->{'Title'} ne "Collage") {
            $contains_text .= "\".$next_subOID"; 
        }

                # Extra code for incremental building.
                # We need to store a listing of the classifiers each DOI is in
                my $clids = [];
                #rint STDERR "==3. Recording reverse lookup for $OID.$next_subOID==\n";
                if(defined($oid_to_clids->{$OID . "." . $next_subOID})) {
                  #rint STDERR "Found existing array!\n";
                  $clids = $oid_to_clids->{$OID . "." . $next_subOID};
                }
                #rint STDERR "Appended $OID to \"" . join(";", @{$clids}) . "\"\n";
                push(@{$clids}, $OID);
                $oid_to_clids->{$OID . "." . $next_subOID} = $clids;
                #rint STDERR "Result: \"" . join(";", @{$clids}) . "\"\n";
        
        &print_classify_info ($infodb_handle, $tempinfo, "$OID.$next_subOID",
                      $remove_empty_classifications);
        $next_subOID++;
        }
    }
    
    $classify_infodb{"contains"} = [ $contains_text ];
    $classify_infodb{"mdtype"} = [ $classifyinfo->{'mdtype'} ]
    if defined $classifyinfo->{'mdtype'};
    $classify_infodb{"mdoffset"} = [ $mdoffset_text ]
    if ($mdoffset_text !~ m/^;+$/);
    
    &dbutil::write_infodb_entry($infodb_handle, $OID, \%classify_infodb);
}


sub check_contents {
    my ($classifyinfo,$remove_empty_classifications) = @_;
    $remove_empty_classifications = 0 unless ($remove_empty_classifications);
    my $num_leaf_docs = 0;
    my $sub_num_leaf_docs = 0;

    return $classifyinfo->{'numleafdocs'} if (defined $classifyinfo->{'numleafdocs'});

    foreach my $content (@{$classifyinfo->{'contains'}}) {
    if (defined $content->{'OID'}) {
        # found a book
        $num_leaf_docs ++;
    } elsif (($sub_num_leaf_docs = &check_contents ($content,$remove_empty_classifications)) > 0) {
        # there's a book somewhere below
        $num_leaf_docs += $sub_num_leaf_docs;
    } else {
        if ($remove_empty_classifications){
        # section contains no books so we want to remove 
        # it from its parents contents
        $content = undef;
        }
    }
    }

    $classifyinfo->{'numleafdocs'} = $num_leaf_docs;
    return $num_leaf_docs;
}

1;