File Coverage

blib/lib/Text/Mining/Corpus/Document.pm

Criterion	Covered	Total	%
statement	3	3	100.0
branch			n/a
condition			n/a
subroutine	1	1	100.0
pod			n/a
total	4	4	100.0

package Text::Mining::Corpus::Document;
use base qw(Text::Mining::Base);
use Class::Std;
use Class::Std::Utils;
use Text::Mining::Parser;

use warnings;
use strict;
use Carp;

use version; our $VERSION = qv('0.0.8');

{
	my %document_id_of           : ATTR( :set :get );
	my %document_type_of         : ATTR( :set :get :default<> );
	my %corpus_id_of             : ATTR( :init_arg :set :get :default<> );
	my %corpus_name_of           : ATTR( :set :get :default<> );
	my %submitted_by_user_id_of  : ATTR( :set :get :default<> );
	my %document_title_of        : ATTR( :set :get :default<> );
	my %document_url_of          : ATTR( :set :get :default<> );
	my %document_path_of         : ATTR( :set :get :default<> );
	my %file_name_of             : ATTR( :init_arg :set :get :default<> );
	my %file_type_of             : ATTR( :init_arg :set :get :default<> );
	my %bytes_of                 : ATTR( :set :get :default<> );
	my %enter_date_of            : ATTR( :set :get :default<> );
	my %exit_date_of             : ATTR( :set :get :default<> );

	sub BUILD {
		my ($self, $ident, $arg_ref) = @_;

		return;
	}

	sub START {
		my ($self, $ident, $arg_ref) = @_;

		#print "  FILE TYPE: ", $self->get_file_type(), "\n\n";

		if    (defined $arg_ref->{document_id})        { $self->_get_document($arg_ref); }
		elsif (defined $arg_ref->{file_name})          { $self->insert( $arg_ref ); }

		return;
	}

	sub parse {
		my ( $self, $arg_ref )  = @_;
		$arg_ref->{file_type}   = $self->get_document_type();
		$arg_ref->{file_name}   = $self->get_file_name();
		$arg_ref->{document_id} = $self->get_document_id();

		my $parser = Text::Mining::Parser->new( $arg_ref );

		print STDERR $parser->version(), "\n";
		print STDERR $parser->stats(), "\n";
		print STDERR $parser->parse(), "\n";

		my $document_id = $self->get_document_id();
		return $document_id;
	}

#	sub get_id                    { my ($self)  = @_; return $id_of{ident $self}; }
#	sub get_document_id           { my ($self)  = @_; return $id_of{ident $self}; }
#	sub get_submitted_document_id { my ($self)  = @_; return $id_of{ident $self}; }
#	sub get_corpus_id             { my ($self)  = @_; return $corpus_id_of{ident $self}; }
#	sub get_submitted_by_user_id  { my ($self)  = @_; return $submitted_by_user_id_of{ident $self}; }
#	sub get_document_url          { my ($self)  = @_; return $document_url_of{ident $self}; }
#	sub get_document_path         { my ($self)  = @_; return $document_path_of{ident $self}; }
#	sub get_file_name    { my ($self)  = @_; return $file_name_of{ident $self}; }
#	sub get_enter_date            { my ($self)  = @_; return $enter_date_of{ident $self}; }
#	sub get_exit_date             { my ($self)  = @_; return $exit_date_of{ident $self}; }

	sub _get_document {
		my ($self, $arg_ref) = @_;
		my $ident = ident $self;

		my $sql  = "select document_id, document_type_id, corpus_id, document_path, document_file_name, bytes, enter_date ";
		   $sql .= "from documents ";
	   	   $sql .= "where document_id = '$arg_ref->{document_id}'";

		($document_id_of{$ident},
	   	 $document_type_of{$ident},
	   	 $corpus_id_of{$ident},
	   	 $document_path_of{$ident},
	   	 $file_name_of{$ident},
	   	 $bytes_of{$ident},
	   	 $enter_date_of{$ident}) = $self->library()->sqlexec($sql, '@');
	}

	sub all {
		my ($self)  = @_;
		my (@documents);

		my $sql  = "select document_id from documents order by document_id asc";
		my $documents = $self->library()->sqlexec( $sql, '\@@' );

		foreach my $document (@$documents) { push @documents, Text::Librarian::Document->new({ document_id => $document->[0] }); }

		return \@documents;
	}

	sub display_all {
		my ($self, $c, $root_url)     = @_;
		my @switch                    = (1, 0);
		my @classes                   = ('rowB', 'rowA');
		my $documents                      = Text::Librarian::Document->all();
		my ($html, $switch, $class);

		$html                        .= " \n";  \n";  \n";  \n";  \n";  \n";  \n";  \n";  \n";  \n";  \n";  \n";  \n"; 




109




 





 





 





 





 





 



 		$html                        .= "	




110




 





 





 





 





 





 



 		$html                        .= "		   




111




 





 





 





 





 





 



 		$html                        .= "		Name 




112




 





 





 





 





 





 



 		$html                        .= "		Description 




113




 





 





 





 





 





 



 		$html                        .= "		Path 




114




 





 





 





 





 





 



 		$html                        .= "	




115




 





 





 





 





 





 



 		foreach my $document (@$documents) { 




116




 





 





 





 





 





 



 			$switch               = $switch[$switch]; 




117




 





 





 





 





 





 



 			$class                = $classes[$switch]; 




118




 





 





 





 





 





 



 		        $html                .= "	




119




 





 





 





 





 





 



 			$html                .= "		 [X] 




120




 





 





 





 





 





 



 			$html                .= "		  " . $document->get_name() . "  




121




 





 





 





 





 





 



 			$html                .= "	         " . $document->get_desc() . "




122




 





 





 





 





 





 



 			$html                .= "	         " . $document->get_path() . "




123




 





 





 





 





 





 



 			$html                .= "	




124




 





 





 





 





 





 



 		} 




125




 





 





 





 





 





 



 		$html                        .= " \n";
		$html                        .= "
		$html                        .= "
		$html                        .= "
		$html                        .= "
		$html                        .= "
		$html                        .= "
		foreach my $document (@$documents) {
			$switch               = $switch[$switch];
			$class                = $classes[$switch];
		        $html                .= "
			$html                .= "
			$html                .= "
			$html                .= "
			$html                .= "
			$html                .= "
		}
		$html                        .= "
		return $html;
	}

	sub update {
		my ( $self, $arg_ref )   = @_;
		my $ident       = ident $self;
		my @updates     = ();

		if ( defined $arg_ref->{corpus_id} )          { $self->set_desc( $arg_ref->{corpus_id} ); push @updates, "corpus_id = '" . $self->_html_to_sql( $arg_ref->{corpus_id} ) . "'"; }
		if ( defined $arg_ref->{document_path} )      { $self->set_path( $arg_ref->{document_path} ); push @updates, "document_path = '" . $self->_html_to_sql( $arg_ref->{document_path} ) . "'"; }
		if ( defined $arg_ref->{file_name} )          { $self->set_file_name( $arg_ref->{file_name} ); push @updates, "file_name = '" . $self->_html_to_sql( $arg_ref->{file_name} ) . "'"; }
		if ( defined $arg_ref->{bytes} )              { $self->set_desc( $arg_ref->{bytes} ); push @updates, "bytes = '" . $self->_html_to_sql( $arg_ref->{bytes} ) . "'"; }

		my $sql  = "update documents set " . join( ', ', @updates ) . " where document_id = '$document_id_of{$ident}'";
		$self->library()->sqlexec($sql);
	}

	sub insert {
		my ($self, $arg_ref)  = @_;
		foreach ('corpus_id', 'document_title', 'document_path', 'file_name', 'bytes') { $arg_ref->{$_} = $self->_html_to_sql( $arg_ref->{$_} || '' ); }

		# Set doc_type_id : alpha - should be live or at least configured
		my %doc_types = ( txt => 1, xml => 2, pdf => 3 );
		$arg_ref->{document_type_id} = $doc_types{ $arg_ref->{file_type} };

		my $sql  = "insert into documents (document_type_id, corpus_id, document_title, document_path, document_file_name, bytes) ";
		   $sql .= "values ('$arg_ref->{document_type_id}', '$arg_ref->{corpus_id}', '$arg_ref->{document_title}', '$arg_ref->{document_path}', '$arg_ref->{file_name}', '$arg_ref->{bytes}') ";
		#print "\n", $sql, "\n\n";
		$self->library()->sqlexec($sql);

		   $sql  = "select LAST_INSERT_ID()";
		( $arg_ref->{document_id} ) = $self->library()->sqlexec($sql, '@');

		$self->_get_document( $arg_ref );
	}

	sub delete {
		my ( $self )  = @_;
		my $ident             = ident $self;

		$self->library()->sqlexec("delete from documents where document_id = '" . $self->get_document_id() . "'");
	}
}

1; # Magic true value required at end of module
__END__