source: trunk/gsdl/perllib/classify/AZList.pm@ 425

# classifier plugin for sorting alphabetically

package AZList;

use sorttools;

sub new {
    my ($class, @options) = @_;

    if (!defined @options || !scalar @options) {
    die "AZList used with no metadata name to classify by\n";
    }

    return bless {
    'list'=>{},
    'metaname' => $options[0]
    }, $class;
}

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

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

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

    my $doc_OID = $doc_obj->get_OID();
    my $metavalue = $doc_obj->get_metadata_element ($doc_obj->get_top_section(),
                            $self->{'metaname'});

    # if this document doesn't contain the metadata element we're
    # sorting by we won't include it in this classification
    if (defined $metavalue && $metavalue =~ /\w/) {
    if ($self->{'metaname'} eq 'Creator') {
        &sorttools::format_string_name_english (\$metavalue);
    } else {
        &sorttools::format_string_english (\$metavalue);
    }
    if (defined $self->{'list'}->{$doc_OID}) {
        print STDERR "WARNING: AZList::classify called multiple times for $doc_OID\n";
    } 
    $self->{'list'}->{$doc_OID} = $metavalue;
    }
}

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

    my @classlist = sort {$self->{'list'}->{$a} cmp $self->{'list'}->{$b};} keys %{$self->{'list'}};

    return $self->splitlist (\@classlist);
}

sub get_entry {
    my $self = shift (@_);
    my ($title, $classifytype) = @_;
    
    # organise into classification structure
    my %classifyinfo = ('classifytype'=>$classifytype,
            'Title'=>$title,
            'contains'=>[]);

    return \%classifyinfo;
}

# splitlist takes an ordered list of classifications (@$classlistref) and splits it
# up into alphabetical sub-sections.
sub splitlist {
    my $self = shift (@_);
    my ($classlistref) = @_;
    my $classhash = {};

    # top level
    my $classifyinfo = $self->get_entry ($self->{'metaname'}, "AZList");

    # don't need to do any splitting if there are less than 20 classifications
    if ((scalar @$classlistref) <= 20) {
    foreach $subOID (@$classlistref) {
        push (@{$classifyinfo->{'contains'}}, {'OID'=>$subOID});
    }
    return $classifyinfo;
    }
    
    # first split up the list into separate A-Z and 0-9 classifications
    foreach $classification (@$classlistref) {
    my $title = $self->{'list'}->{$classification};
    $title =~ s/^(.).*$/$1/;
    $title =~ tr/[a-z]/[A-Z]/;
    if ($title =~ /^[0-9]$/) {$title = '0-9';}
    elsif ($title !~ /^[A-Z]$/) {
        print STDERR "AZList: WARNING $classification has badly formatted title ($title)\n";
    }
    $classhash->{$title} = [] unless defined $classhash->{$title};
    push (@{$classhash->{$title}}, $classification);
    }
    $classhash = $self->compactlist ($classhash);

    my @tmparr = ();
    foreach $subsection (sort keys (%$classhash)) {
    push (@tmparr, $subsection);
    }
    
    # if there's a 0-9 section it will have been sorted to the beginning
    # but we want it at the end
    if ($tmparr[0] eq '0-9') {
    shift @tmparr;
    push (@tmparr, '0-9');
    }

    foreach $subclass (@tmparr) {
    my $tempclassify = $self->get_entry($subclass, "AZList");
    foreach $subsubOID (@{$classhash->{$subclass}}) {
        push (@{$tempclassify->{'contains'}}, {'OID'=>$subsubOID});
    }
    push (@{$classifyinfo->{'contains'}}, $tempclassify);
    }

    return $classifyinfo;
}

sub compactlist {
    my $self = shift (@_);
    my ($classhashref) = @_;
    my $compactedhash = {};
    my @currentOIDs = ();
    my $currentfirstletter = "";
    my $currentlastletter = "";
    my $lastkey = "";

    # minimum and maximum documents to be displayed per page.
    # the actual maximum will be max + (min-1).
    # the smallest sub-section is a single letter at present
    # so in this case there may be many times max documents
    # displayed on a page.
    my $min = 10; 
    my $max = 30;

    foreach $subsection (sort keys %$classhashref) {
    if ($subsection eq '0-9') {
        @{$compactedhash->{$subsection}} = @{$classhashref->{$subsection}};
        next;
    }
    $currentfirstletter = $subsection if $currentfirstletter eq "";
    if ((scalar (@currentOIDs) < $min) ||
        ((scalar (@currentOIDs) + scalar (@{$classhashref->{$subsection}})) <= $max)) {
        push (@currentOIDs, @{$classhashref->{$subsection}});
        $currentlastletter = $subsection;
    } else {

        if ($currentfirstletter eq $currentlastletter) {
        @{$compactedhash->{$currentfirstletter}} = @currentOIDs;
        $lastkey = $currentfirstletter;
        } else {
        @{$compactedhash->{"$currentfirstletter-$currentlastletter"}} = @currentOIDs;
        $lastkey = "$currentfirstletter-$currentlastletter";
        } 
        if (scalar (@{$classhashref->{$subsection}}) >= $max) {
        $compactedhash->{$subsection} = $classhashref->{$subsection};
        @currentOIDs = ();
        $currentfirstletter = "";
        } else {
        @currentOIDs = @{$classhashref->{$subsection}};
        $currentfirstletter = $subsection;
        $currentlastletter = $subsection;
        }
    }
    }

    # add final OIDs to last sub-classification if there aren't many otherwise
    # add final sub-classification
    if (scalar (@currentOIDs) < $min) {
    my ($newkey) = $lastkey =~ /^(.)/;
    @currentOIDs = (@{$compactedhash->{$lastkey}}, @currentOIDs);
    delete $compactedhash->{$lastkey};
    @{$compactedhash->{"$newkey-$currentlastletter"}} = @currentOIDs;   
    } else {
    if ($currentfirstletter eq $currentlastletter) {
        @{$compactedhash->{$currentfirstletter}} = @currentOIDs;
    } else {
        @{$compactedhash->{"$currentfirstletter-$currentlastletter"}} = @currentOIDs;
    } 
    }

    return $compactedhash;
}

1;