xmltv/grab/ar tv_grab_ar,1.12,1.13

Mariano Cosentino <[email protected]>
Newsgroups gmane.comp.tv.xmltv.cvs
Message-ID <[email protected]> 
Update of /cvsroot/xmltv/xmltv/grab/ar
In directory sfp-cvs-1.v30.ch3.sourceforge.com:/tmp/cvs-serv8728/xmltv/grab/ar

Modified Files:
	tv_grab_ar 
Log Message:
added function to handle character conversion to utf-8
solved issue with descriptions
additional handling due to some changes on source format


Index: tv_grab_ar
===================================================================
RCS file: /cvsroot/xmltv/xmltv/grab/ar/tv_grab_ar,v
retrieving revision 1.12
retrieving revision 1.13
diff -C2 -d -r1.12 -r1.13
*** tv_grab_ar	22 Jun 2011 05:56:34 -0000	1.12
--- tv_grab_ar	12 Jul 2013 19:54:10 -0000	1.13
***************
*** 71,74 ****
--- 71,77 ----
  
  =head1 AUTHOR
+ 1.14 Mariano S. Cosentino, [email protected]
+ 1.13 Mariano S. Cosentino, [email protected]
+ 1.12 rmeden
  1.10 Mariano S. Cosentino, [email protected]
  1.9 Mariano S. Cosentino, [email protected]
***************
*** 90,97 ****
  
  # Author's Updates
  #   1.10 	Mariano Cosentino
  #		* 	fixed prototype warning
  #		* 	Removed default value for location, as the old channel ids are no compatible with the site's new format.
- # Author's Updates
  #   1.9 	Mariano Cosentino
  #		* 	The source website changed, I've adapted the script to the new format
--- 93,105 ----
  
  # Author's Updates
+ #   1.14        Mariano Cosentino
+ #               *       added function to handle character conversion to utf-8
+ #   1.13        Mariano Cosentino
+ #               *       solved issue with descriptions
+ #   1.12        rmeden
+ #               apply Karl's patches to remove switch statements
  #   1.10 	Mariano Cosentino
  #		* 	fixed prototype warning
  #		* 	Removed default value for location, as the old channel ids are no compatible with the site's new format.
  #   1.9 	Mariano Cosentino
  #		* 	The source website changed, I've adapted the script to the new format
***************
*** 146,151 ****
  use HTML::Entities; # parse entities
  use IO::File;
  
! use Encode qw(from_to is_utf8 _utf8_off encode);
  use utf8;
  
--- 154,160 ----
  use HTML::Entities; # parse entities
  use IO::File;
+ use Switch;
  
! use Encode qw(from_to is_utf8 _utf8_off encode decode);
  use utf8;
  
***************
*** 323,327 ****
      $w_args{OUTPUT} = $fh;
  }
! $w_args{encoding} = 'utf-8';
  my $writer = new XMLTV::Writer(%w_args);
  $writer->start($HEAD);
--- 332,336 ----
      $w_args{OUTPUT} = $fh;
  }
! $w_args{encoding} = 'ISO-8859-15';
  my $writer = new XMLTV::Writer(%w_args);
  $writer->start($HEAD);
***************
*** 686,690 ****
                  my ($channel_num) = ($opt->attr('onclick') =~ /.*sintonia=(\d+).*/i);
                  $channels{$channel_id}=$channel_name;
! 		my $coded_chan_name=encode("UTF-8",$channel_name);
  #		print $coded_chan_name;
                  push @ch_all, { 
--- 695,699 ----
                  my ($channel_num) = ($opt->attr('onclick') =~ /.*sintonia=(\d+).*/i);
                  $channels{$channel_id}=$channel_name;
! 		my $coded_chan_name=encodeco("utf-8",$channel_name);
  #		print $coded_chan_name;
                  push @ch_all, { 
***************
*** 834,840 ****
  # <a href="ficha.php?verFicha=224193" rel="#overlay_ficha" class="programaSemana" title="Parental Control" style="height:34px; top:-1px">
  
! 	my $title= encode("UTF-8",trim($celda->attr('title')));
  	if (! (defined $title)) {
! 		$title= encode("UTF-8",trim($celda->as_text));         
  	} 
          my ($top)=($celda->attr('style') =~ /top:\s*(-{0,1}\d+)\s*(px|)/i);
--- 843,849 ----
  # <a href="ficha.php?verFicha=224193" rel="#overlay_ficha" class="programaSemana" title="Parental Control" style="height:34px; top:-1px">
  
! 	my $title= encodeco("UTF-8",trim($celda->attr('title')));
  	if (! (defined $title)) {
! 		$title= encodeco("UTF-8",trim($celda->as_text));         
  	} 
          my ($top)=($celda->attr('style') =~ /top:\s*(-{0,1}\d+)\s*(px|)/i);
***************
*** 867,871 ****
  	if ($opt_get_details) {
  		my %TempDesc;
! 		my ($progindex)=($celda->attr('href') =~ /ficha\.php\?verFicha.(\d+)/i);
  		if (exists $Descriptions{$progindex}) {
  			#original format
--- 876,880 ----
  	if ($opt_get_details) {
  		my %TempDesc;
! 		my ($progindex)=($celda->attr('rev') =~ /ficha\.php\?verFicha.(\d+)/i);
  		if (exists $Descriptions{$progindex}) {
  			#original format
***************
*** 888,892 ****
  				my $Categoria = ($ficha_tree->look_down('_tag','h1'));
  				if (defined $Categoria) {
! 					$Categoria = encode("UTF-8",trim($Categoria->as_text));
  				} else {
  					$Categoria = "";
--- 897,906 ----
  				my $Categoria = ($ficha_tree->look_down('_tag','h1'));
  				if (defined $Categoria) {
! 					$Categoria = encodeco("UTF-8",trim($Categoria->as_text));
! #					 $valor =~ s/, /,/g;
! 
! #                                        my @algo =split(',', $Categoria);
! #                                        $Categoria = \@algo
! 
  				} else {
  					$Categoria = "";
***************
*** 903,907 ****
  				my $TituloEsp = ($ficha_tree->look_down('_tag','h2'));
  				if (defined $TituloEsp) {
! 					$TituloEsp = encode("UTF-8",trim($TituloEsp->as_text));
  				} else {
  					$TituloEsp = $title;
--- 917,921 ----
  				my $TituloEsp = ($ficha_tree->look_down('_tag','h2'));
  				if (defined $TituloEsp) {
! 					$TituloEsp = encodeco("UTF-8",trim($TituloEsp->as_text));
  				} else {
  					$TituloEsp = $title;
***************
*** 916,920 ****
  				my $sinopsis= ($SinopsisBox->look_down('_tag','p')) if (defined $SinopsisBox); 
  				if (defined $sinopsis) {
! 					$sinopsis=encode("UTF-8",trim($sinopsis->as_text));
  				} else {
  					$sinopsis = "";
--- 930,934 ----
  				my $sinopsis= ($SinopsisBox->look_down('_tag','p')) if (defined $SinopsisBox); 
  				if (defined $sinopsis) {
! 					$sinopsis=encodeco("UTF-8",trim($sinopsis->as_text));
  				} else {
  					$sinopsis = "";
***************
*** 932,983 ****
  					else {
  						my $valor;
! 						$valor=encode("UTF-8",trim($Propiedad->as_text));
! 						if ($PropertyField =~ m/Protagonista((s)?)/i ) {
! 							$valor =~ s/, /,/g;
! 							my @algo =split(',', $valor);
! 							$credits{actor} = \@algo
! 						}
! 						elsif ($PropertyField =~ m/Director((es)?)/i) { 
! 							$valor =~ s/, /,/g;
! 							my @algo =split(',', $valor);
! 							$credits{director} = \@algo
! 						}
! 						elsif ($PropertyField =~ m/Conductor((es)?)/i) { 
! 							$valor =~ s/, /,/g;
! 							my @algo =split(',', $valor);
! 							$credits{presenter} = \@algo
! 						}
! 						elsif ($PropertyField =~ m/G.nero/i) { 
! 							$ret{category} = $valor;
! 						}
! 						elsif ($PropertyField =~ m/(Duraci&oacute;n|Duración)/) { # html, utf-8 bytes, missing unicode chars
! 							$valor =~ /(\d*)\s(.*)(\n|\r)?/i;
! 							my $TempLength=$1;
! 							my $tempunit=$2;
! 							$tempunit=~ s/^\s//;
! 							$tempunit=~ s/\s$//;
! #							$tempunit=~ s/minutos?/minutes/;
! #							$tempunit=~ s/horas?/hours/;
! #							$tempunit=~ s/segundos?/seconds/;
! #							if ($tempunit=~ s/minutos?/i) { $TempLength = ($TempLength *60) }	
! #							if ($tempunit=~ s/horas?/i) { $TempLength = ($TempLength *60) }	
! 							$ret{length} = ($TempLength *60);
! 						}
! 						elsif ($PropertyField =~ m/Pa.s/i) { 
! 							$ret{country} = $valor;
! 						}
! 						elsif ($PropertyField =~ m/(A&ntilde;o|Ano)/) { 
! 							$ret{date} = $valor;
! 						}
! 						elsif ($PropertyField eq 'Título original') { 
! 							$ret{title2} = $valor;
! 						}
! 						elsif ($PropertyField =~ m/Clasificaci.n/i) { 
! 							$ret{rating} = $valor;
! 						}
! 						else { 
! 							print STDERR "ignoring useless data ... [$PropertyField]\n";
  						}
! 					$PropertyField = '';
  					}
  				}
--- 946,1001 ----
  					else {
  						my $valor;
! 						$valor=encodeco("UTF-8",trim($Propiedad->as_text));
! 						switch ($PropertyField) {
! 							case (m/Protagonista((s)?)/i) {
! 								$valor =~ s/, /,/g;
! 								my @algo =split(',', $valor);
! 								$credits{actor} = \@algo
! 							}
! 							case (m/Director((es)?)/i) { 
! 								$valor =~ s/, /,/g;
! 								my @algo =split(',', $valor);
! 								$credits{director} = \@algo
! 							}
! 							case (m/Conductor((es)?)/i) { 
! 								$valor =~ s/, /,/g;
! 								my @algo =split(',', $valor);
! 								$credits{presenter} = \@algo
! 							}
! 							case (m/G.nero/i) { 
! 								$ret{category} = $valor;
! 							}
! 							case (['Duraci&oacute;n','Duración']) { 
! 								$valor =~ /(\d*)\s(.*)(\n|\r)?/i;
! 								my $TempLength=$1;
! 								my $tempunit=$2;
! 								$tempunit=~ s/^\s//;
! 								$tempunit=~ s/\s$//;
! #								$tempunit=~ s/minutos?/minutes/;
! #								$tempunit=~ s/horas?/hours/;
! #								$tempunit=~ s/segundos?/seconds/;
! #								if ($tempunit=~ s/minutos?/i) { $TempLength = ($TempLength *60) }	
! #								if ($tempunit=~ s/horas?/i) { $TempLength = ($TempLength *60) }	
! 								$ret{length} = ($TempLength *60);
! 							}
! 							case (m/Pa.s/i) { 
! 								$ret{country} = $valor;
! 							}
! 							case (['A&ntilde;o','Ano']) { 
! 								$ret{date} = $valor;
! 							}
! 							case ('Título original') { 
! 								$ret{title2} = $valor;
! 							}
! 							case ([m/Clasificaci.n/i]) { 
! 								$ret{rating} = $valor;
! 							}
! 							case "" { 
! #								print "ignoring useless data ... [$PropertyField]\n" ;;
! #								%ret->{''} = $valor;
! 							}
! 							else {}
  						}
! 					$PropertyField ="";
  					}
  				}
***************
*** 1035,1039 ****
--- 1053,1069 ----
  
  
+ sub encodeco {
+  my($tempdummy,$temptexto)=@_;
+ #    print "$temptexto\n";
+ #    # decode the qouted-printable input
+ #    #$body = decode_qp( $body );
+ #    # decode to Perl's internal format
+ #    $temptexto = decode( 'iso-8859-1', $temptexto );
+ #    # encode to UTF-8
+ #    $temptexto = encode( 'utf-8', $temptexto );
+ #    print "$temptexto\n";
  
+     return $temptexto;
+ }
  
  
***************
*** 1064,1077 ****
  sub Clean_amp {
      my ($texto)=@_;
! 	Encode::from_to($texto, "iso-8859-1", "utf8");
! 	my $textoold=$texto;
! 	$texto =~ s/&amp;/&/g;
! 	$texto =~ s/&acute;/á/g;
! 	$texto =~ s/&aacute;/á/g;
! 	$texto =~ s/&eacute;/é/g;
! 	$texto =~ s/&iacute;/í/g;
! 	$texto =~ s/&oacute;/ó/g;
! 	$texto =~ s/&uacute;/ú/g;
! 	$texto =~ s/&ntilde;/ñ/g;
  
  
--- 1094,1108 ----
  sub Clean_amp {
      my ($texto)=@_;
! #	Encode::from_to($texto, "iso-8859-1", "utf8");
!         $texto=encodeco("utf-8",$texto);
! #	my $textoold=$texto;
! #	$texto =~ s/&amp;/&/g;
! #	$texto =~ s/&acute;/á/g;
! #	$texto =~ s/&aacute;/á/g;
! #	$texto =~ s/&eacute;/é/g;
! #	$texto =~ s/&iacute;/í/g;
! #	$texto =~ s/&oacute;/ó/g;
! #	$texto =~ s/&uacute;/ú/g;
! #	$texto =~ s/&ntilde;/ñ/g;

------------------------------------------------------------------------------
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk

_______________________________________________
xmltv-commit mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/xmltv-commit
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.