#!/usr/bin/perl -w
eval 'exec /usr/bin/perl -w -S $0 ${1+"$@"}'
    if 0; # not running under some shell

=pod

=head1 NAME

tv_extractinfo_ar - read Spanish (Argentinean) language listings and extract info
from programme descriptions.

=head1 SYNOPSIS

tv_extractinfo_ar [--help] [--output FILE] [FILE...]

=head1 DESCRIPTION

Read XMLTV data and attempt to extract information from
Spanish-language programme descriptions, putting it into
machine-readable form.  For example the human-readable text '(Repeticion)'
in a programme description might be replaced by the XML element
<previously-shown>.

B<--output FILE> write to FILE rather than standard output

This tool also attempts to split multipart programmes into their
constituents, by looking for a description that seems to contain lots
of times and titles.  But this depends on the description following
one particular style and is useful only for some listings sources
(Ananova).

If some text is marked with the 'lang' attribute as being some
language other than Spanish ('es'), it is ignored.

=head1 SEE ALSO

L<xmltv(5)>.

=head1 AUTHOR

Mariano Cosentino, Mok@marianok.com.ar

=head1 BUGS

Trying to parse human-readable text is always error-prone, more so
with the simple regexp-based approach used here.  But because TV
listing descriptions usually conform to one of a few set styles,
tv_extractinfo_en does reasonably well.  It is fairly conservative,
trying to avoid false positives (extracting 'information' which
isnE<39>t really there) even though this means some false negatives
(failing to extract information and leaving it in the human-readable
text).

However, the leftover bits of text after extracting information may
not form a meaningful Spanish sentence, or the punctuation may be
wrong.

On the two listings sources currently supported by the XMLTV package,
this program does a reasonably good job.  But it has not been tested
with every source of anglophone TV listings.



This Spanish Version is heavily customized for the XML results from
tv_grab_ar (developed by Christian A. Rodriguez and postriorily
updated by Mariano S. Cosentino).

This file should probably be called tv_extractinfo_es, but I have not
tested it with any other spanish grabbers, so I don't want to be
presumptious.



=cut
use strict;

sub MSC_Extract();


use XMLTV;
use XMLTV::Version "$XMLTV::VERSION";
use XMLTV::Usage <<END
$0: read Spanish-language listings and extract info from programme descriptions
usage: $0 [--help] [--output FILE] [FILE...]
END
;
use XML::LibXML;
use Getopt::Long;

my ($opt_help, $opt_output);
GetOptions('help' => \$opt_help, 'output=s' => \$opt_output) or usage(0);
usage(1) if $opt_help;
@ARGV = ('-') if not @ARGV;


my $parser = XML::LibXML->new;

my $archivo = @ARGV;

my $doc    = $parser->parse_file("@ARGV")
                or die "can't parse xmltv file: $@";

my $root = $doc->documentElement();

MSC_Extract();

if (defined $opt_output) {
    my $status = $doc->toFile("$opt_output");
    die "cannot write to $opt_output\n" if not $status;
} else {
	print $doc->toString;
}




sub MSC_Extract() {
my @nodeList = $doc->getElementsByTagName('programme');
foreach my $programa (@nodeList) {
    my @oldTitle = $programa->getChildrenByTagName('title');
    my @oldDesc = $programa->getChildrenByTagName('desc');
    my $oldDesc=$oldDesc[0]->textContent;
    my $curProgname = $oldTitle[0]->textContent;
#    my $parentNode=$programa->parent;
	$curProgname =~ s/^\s+//;
	$curProgname =~ s/\s+$//;
    my $Progname = $curProgname;
#	$curProgname =~ s/^Disney.s\s//i;
	$curProgname =~ s/^Disney.\s//i;
	$curProgname =~ s/^\s+//;
	$curProgname =~ s/\s+$//;
	if ($curProgname eq "") {
		$curProgname=$Progname;
	}
	# else {
#	 $Progname = $curProgname;
#	}
	my $progTitle="";
	my $progsubtitle="";
	my $episodio="";
	my $Temporada="";
	my $newNodo="";

	my $newNodo_PS="";
	my $newNodo_EN="";

	my $bChanged=0;

#	print "$Progname\n";

	if ($curProgname =~ /(.*)\(Repetici.n\)$/) {
		$newNodo_PS = XML::LibXML::Element->new('previously-shown');
		$curProgname =$1; # =~ s,$1,,;
		$curProgname =~ s/\(\)//g;
	}
	if ($curProgname =~ /(.*)\(Repetici.n\)(.*)/) {
		$newNodo_PS = XML::LibXML::Element->new('previously-shown');
		$curProgname =$1 . $3; # =~ s,$1,,;
		$curProgname =~ s/\(\)//g;
	}


	if ($curProgname =~ /(.*)\(diferido\)$/) {
		$curProgname =$1; # =~ s,$1,,;
		$curProgname =~ s/\(\)//g;
	}
	if ($curProgname =~ /(.*)\(diferido\)(.*)/) {
		$curProgname =$1 . $3; # =~ s,$1,,;
		$curProgname =~ s/\(\)//g;
	}

	if ($curProgname =~ /(.*)\(grabado\)$/) {
		$curProgname =$1; # =~ s,$1,,;
		$curProgname =~ s/\(\)//g;
	}
	if ($curProgname =~ /(.*)\(grabado\)(.*)/) {
		$curProgname =$1 . $3; # =~ s,$1,,;
		$curProgname =~ s/\(\)//g;
	}

	if ($curProgname =~ /(.*)\(vivo\)$/) {
		$curProgname =$1; # =~ s,$1,,;
		$curProgname =~ s/\(\)//g;
	}
	if ($curProgname =~ /(.*)\(vivo\)(.*)/) {
		$curProgname =$1 . $3; # =~ s,$1,,;
		$curProgname =~ s/\(\)//g;
	}

	if ($curProgname =~ /(.*)\(doblad.\)$/) {
		$curProgname =$1; # =~ s,$1,,;
		$curProgname =~ s/\(\)//g;
	}
	if ($curProgname =~ /(.*)\(doblad.\)(.*)/) {
		$curProgname =$1 . $3; # =~ s,$1,,;
		$curProgname =~ s/\(\)//g;
	}

	if ($curProgname =~ /(.*)\(subtitulad..?\)$/) {
		$curProgname =$1; # =~ s,$1,,;
		$curProgname =~ s/\(\)//g;
	}
	if ($curProgname =~ /(.*)\(subtitulad..?\)(.*)/) {
		$curProgname =$1 . $3; # =~ s,$1,,;
		$curProgname =~ s/\(\)//g;
	}


	if ($curProgname =~ /(.*)?(.?\sEpisodio\s(\d+))(.*)?/) {
		$episodio=$3; #$2;
		$curProgname =$1 . $4; # =~ s,$1,,;
		$curProgname =~ s/^\s+//;
		$curProgname =~ s/\s+$//;
		$curProgname =~ s/  //g;
		$curProgname =~ s/\.\./\./g;
	} else {
		$episodio="";
	}

	if ($curProgname =~ /(.*)?(.?\sTemporada\s(\d+))(.*)?/) {
		$Temporada=$3; #$2;
		$curProgname = $1 . $4; # =~ s,'$1',,;
		$curProgname =~ s/^\s+//;
		$curProgname =~ s/\s+$//;
		$curProgname =~ s/  //g;
		$curProgname =~ s/\.\./\./g;
	} else {
		$Temporada="";
	}
	if ("$Temporada.$episodio" ne ".") {
		$newNodo_EN = XML::LibXML::Element->new('episode-num');
		$newNodo_EN->addChild ( $doc->createTextNode( "$Temporada.$episodio." ) );
		$newNodo_EN->addChild( $doc->createAttribute('system','xmltv_ns') );
		$newNodo="";
	}




	if ($curProgname =~ /([^.]*)\. (.*)/) {
		$progTitle = $1;
		$progsubtitle = $2;
	} else {
		$progTitle = $curProgname;
		$progsubtitle = "";
	}

	if ($progTitle ne "") {
		$progTitle =~ s/^\s+//;
		$progTitle =~ s/\s+$//;
		$progTitle =~ s/  //g;
		$progTitle =~ s/\.\./\./g;
		$progTitle =~ s/\.+$//g;
	}

	if ($progsubtitle ne "") {
		$progsubtitle =~ s/^\s+//;
		$progsubtitle =~ s/\s+$//;
		$progsubtitle =~ s/  //g;
		$progsubtitle =~ s/\.\./\./g;
		$progsubtitle =~ s/\.+$//g;
	}

#title+
	if ($Progname ne $progTitle) {
		if ($progTitle ne "") {
			$newNodo = XML::LibXML::Element->new('title');
			$newNodo->addChild ( $doc->createTextNode( $progTitle ) );
			$newNodo->addChild( $doc->createAttribute('lang','es') );
			$programa->replaceChild ( $newNodo, $oldTitle[0]);
			$bChanged=1;
			$newNodo="";
		}

	}

	my @TempList;
	my $TempItem;

#Título original:	Drive
	if ($oldDesc =~ /Título original:\s(.*)(\n|\r)?/i) {
			$TempItem=$1;
			my $TempOtherTitle="";
			$TempItem=~ s/^\s//;
			$TempItem=~ s/\s$//;
			$bChanged=1;
			if ($TempItem =~ /(.*)\s\(A.?K.?A.?\s(.*)\)/i) {
				$TempItem=$1;
				$TempOtherTitle=$2;
				$TempItem=~ s/^\s//;
				$TempItem=~ s/\s$//;
				$TempOtherTitle=~ s/^\s//;
				$TempOtherTitle=~ s/\s$//;
			}
			$newNodo = XML::LibXML::Element->new('title');
			$newNodo->addChild ( $doc->createTextNode( "$TempItem" ) );
			$newNodo->addChild( $doc->createAttribute('lang','en') );
			$programa->addChild ( $newNodo  );
			$newNodo="";
			if ($TempOtherTitle ne "") {
				$newNodo = XML::LibXML::Element->new('title');
				$newNodo->addChild ( $doc->createTextNode( "$TempOtherTitle" ) );
				$newNodo->addChild( $doc->createAttribute('lang','en') );
				$programa->addChild ( $newNodo  );
				$newNodo="";
			}
	}

#sub-title*
	if ($progsubtitle ne "") {
		$bChanged=1;
		my $itemnodo = $programa->addNewChild('', 'sub-title');
		$itemnodo->addChild( $doc->createTextNode($progsubtitle) );
	}



# desc*
	if ($bChanged) {
		foreach my $tempNode_desc (@oldDesc) {
			$programa->removeChild($tempNode_desc);
			$programa->addChild ( $tempNode_desc  );
		}
	}






	my $bCredits=0;
	my $CreditsNodo = XML::LibXML::Element->new('credits');



	#    <desc lang="es">Directores:	Ed Friedman, Mark Glamack
	if ($oldDesc =~ /Director(es)?:\s(.*)(\n|\r)?/i) {
		@TempList=split(",",$2);
		if (scalar(@TempList)<2) {
			@TempList=$2;
		}
		foreach $TempItem (@TempList) {
			$TempItem=~ s/^\s//;
			$TempItem=~ s/\s$//;
			$newNodo = XML::LibXML::Element->new('director');
			$newNodo->addChild ( $doc->createTextNode( "$TempItem" ) );
			$bCredits=1;
			$CreditsNodo->addChild ( $newNodo  );
			$newNodo="";
		}
	}

	#     <desc lang="es">Protagonista:	Kiefer Sutherland
	#    <desc lang="es">Protagonistas:	Antonio Aguilar, Elda Peralta, Chela Pastor, Agustín Isunza
	if ($oldDesc =~ /Protagonistas?:\s(.*)(\n|\r)?/i) {
		@TempList=split(",",$1);
		if (scalar(@TempList)<2) {
			@TempList=$1;
		}
		foreach $TempItem (@TempList) {
	#			warn "\nP: $TempItem";
			$TempItem=~ s/^\s//;
			$TempItem=~ s/\s$//;
			$newNodo = XML::LibXML::Element->new('actor');
			$newNodo->addChild ( $doc->createTextNode( "$TempItem" ) );
			$bCredits=1;
			$CreditsNodo->addChild ( $newNodo  );
		}
			$newNodo="";
	}

	#	Conductor:	Adriana Betancour
	#	Conductores:	Adam Corrolla, Jimmy Kimmel
	#	Conductor:	Fernando Merino
	if ($oldDesc =~ /Conductor(es)?:\s(.*)(\n|\r)?/i) {
		@TempList=split(",",$2);
		if (scalar(@TempList)<2) {
			@TempList=$2;
		}
		foreach $TempItem (@TempList) {
			$TempItem=~ s/^\s//;
			$TempItem=~ s/\s$//;
			$newNodo = XML::LibXML::Element->new('presenter');
			$newNodo->addChild ( $doc->createTextNode( "$TempItem" ) );
			$bCredits=1;
			$CreditsNodo->addChild ( $newNodo  );
			$newNodo="";
		}
	}

# credits?
	if ($bCredits ) {
		$programa->addChild ( $CreditsNodo  );
	}
	$CreditsNodo="";



# date?
	#	Año:	1934
	if ($oldDesc =~ /año:\s(.*)(\n|\r)?/i) {
		$TempItem= $1;
		$TempItem=~ s/^\s//;
		$TempItem=~ s/\s$//;
		$newNodo = XML::LibXML::Element->new('date');
		$newNodo->addChild ( $doc->createTextNode( "$TempItem" ) );
		$programa->addChild ( $newNodo  );
		$newNodo="";
	}

# category*
	#    <desc lang="es">Género:	Agro</desc>
	if ($oldDesc =~ /Género:\s(.*)(\n|\r)?/i) {
		@TempList=split(",",$1);
		if (scalar(@TempList)<2) {
			@TempList=$1;
		}
		foreach $TempItem (@TempList) {
			$TempItem=~ s/^\s//;
			$TempItem=~ s/\s$//;
	#		warn "Género: $TempItem\n";
			$newNodo = XML::LibXML::Element->new('category');
			$newNodo->addChild ( $doc->createTextNode( "$TempItem" ) );
			$programa->addChild ( $newNodo  );
			$newNodo="";
		}
	}



# language?

# orig-language?

# length?
	# Duración:	102 minutos
	if ($oldDesc =~ /Duración:\s(\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/;

			$newNodo = XML::LibXML::Element->new('length');
			$newNodo->addChild ( $doc->createTextNode( "$TempLength" ) );
			$newNodo->addChild( $doc->createAttribute('units',"$tempunit") );
			$programa->addChild ( $newNodo  );
			$newNodo="";
	}



# icon*
# url*

# country*
	#País:	Argentina
	if ($oldDesc =~ /País(es)?:\s(.*)(\n|\r)?/i) {
		@TempList=split("-",$2);
		if (scalar(@TempList)<2) {
			@TempList=$2;
		}
		foreach $TempItem (@TempList) {
			$TempItem=~ s/^\s//;
			$TempItem=~ s/\s$//;
			$newNodo = XML::LibXML::Element->new('country');
			$newNodo->addChild ( $doc->createTextNode( "$TempItem" ) );
			$newNodo->addChild( $doc->createAttribute('lang','es') );
			$programa->addChild ( $newNodo  );
			$newNodo="";
		}
	}


# episode-num*
	if ($newNodo_EN ne "") {
		$programa->addChild ($newNodo_EN );
	}



#			my $TempLength="";
#			my $tempunit="";


# video?

# audio?

# previously-shown?
	if ($newNodo_PS ne "") {
		$programa->addChild ($newNodo_PS );
	}

# premiere?

# last-chance?

# new?

# subtitles*

# rating*
	#	Clasificación:	Apta para todo público</desc>
	#	Clasificación:	Sólo apta para mayores de 13 años</desc>
	#	Clasificación:	Sólo apta para mayores de 16 años</desc>
	#	Clasificación:	Sólo apta para mayores de 18 años</desc>
	if ($oldDesc =~ /Clasificación:\s(.*)(\n|\r)?/i) {
			$TempItem=$1;
			$TempItem=~ s/^\s//;
			$TempItem=~ s/\s$//;
			$newNodo = XML::LibXML::Element->new('rating');
			$newNodo->addChild( $doc->createAttribute('system',"Argentina") );
				my $newSubNodo = XML::LibXML::Element->new('value');
				$newSubNodo->addChild ( $doc->createTextNode( "$TempItem" ) );
				$newNodo->addChild( $newSubNodo);
			$programa->addChild ( $newNodo  );
			$newNodo="";
	}

# star-rating*

  }
}

		#/(.*)\. (.*)/) {
		#		print "o: $curProgname\n";
		#		print "T:$1\n";
		#		print "St:$2\n";

#		$progName = $curProgname;

##	    		my $itemnodo = $programa->addNewChild('', 'title');
##			$itemnodo->addChild( $doc->createTextNode($progTitle) );
##			$itemnodo->addChild( $doc->createAttribute('lang','es') );
#	    		my $itemnodo = $programa->addNewChild('', 'title');
#			$itemnodo->addChild( $doc->createTextNode($progTitle) );
#			$itemnodo->addChild( $doc->createAttribute('lang','es') );

#find($programa->nodePath . "/title");
# 		print $oldNodo[0]->nodePath;

#			print $programa->nodePath;
##			$programa = $progTitle;
#->textContent

#	print "$progName [$progsubtitle - $Temporada.$episodio]\n";        # pretty print
#sub addSubElm($$$) {
#    my ($pet, $name, $body) = @_;
#    my $subElm = $pet->addNewChild('', $name);
#    $subElm->addChild( $doc->createTextNode($body) );
#}
#sub addField($$$$) {
#    my ($type, $name, $dob, $price) = @_;
#    # addNewChild is non-compliant; could use addSibling instead
#    my $pet = $root->addNewChild('', $type);
#    addSubElm ( $pet, 'name', $name );
#    addSubElm ( $pet, 'dob',  $dob  );
#    addSubElm ( $pet, 'price', $price );
#}


#    my $newPrice = sprintf "%6.2f", $curPrice * 1.2;
#    my $parent = $programa->parentNode;
#    my $newNodo = XML::LibXML::Element->new('programme');
#    $newNodo->addChild ( $doc->createTextNode( $newPrice ) );
#    $parent->replaceChild ( $newPriceNode, $priceNode );
#print $doc->toString(1);        # pretty print


#mv /home/notroot/xmltv2tivo/20080430.slice.gz /home/notroot/TivoEmulator/tivo-service/static/listings/BA01235_13999-14006.slice.gz