#!/usr/bin/perl -s

use Lingua::PT::PLN;

our ($latin1,$text,$tag) ;

binmode(STDOUT,":utf8") unless($latin1);

$tag ||= "s";

if($text){ textSentence($tag ); }
else     { contextualSentence(shift);}

sub textSentence{
   my %opt=(st=>"s",psep=>"");
   if(ref $_[0] eq "HASH"){ %opt = ( $opt , %{shift(@_)}); }
   local $/= $opt{psep} ;
   while(<>){
     s/^\s*//;
     s/\s+$//;
     print "<p>", Lingua::PT::PLN::xmlsentences({st=>$opt{st}}, $_), "</p>\n";
   } 
}

sub contextualSentence{
   my $tag=shift;
   $/= "</$tag>";
   while(<>){
     s{ ( <$tag\b .*? > )
        (.*?) (</$tag>) }{
       $1 . Lingua::PT::PLN::xmlsentences($2) . $3 }xse;
     print
   } 
}

__END__

=head1 NAME

quebraxmlsent - breaks text in sentences

=head1 SYNOPSIS

 quebraxmlsent -text textfile*
 quebraxmlsent -tag=ttt xmlfile*
 Options:
   -text   I
   -latin1   (output encoding -- def:utf8)

=head1 DESCRIPTION

In text mode paragraphs are separeted by empty lines

=head1 AUTHOR

J.Joao Almeida, jj@di.uminho.pt

=head1 SEE ALSO

perl(1).

Lingua::PT::PLN(3pm)

=cut