#!/usr/bin/perl -w

use File::Basename;

print ';; Script export with encoding UTF-8 
;; transcribed by LIA;;
;; CATEGORY "0" "" "" 
;; LABEL "O" "Overall" 0"Overall"
;;
;; CATEGORY "1" "Hub4 Focus Conditions" ""
;; LABEL "F0" "Baseline//Broadcast//Speech" ""
;; LABEL "F1" "Spontaneous//Broadcast//Speech" ""
;; LABEL "F2" "Speech Over//Telephone//Channels" ""
;; LABEL "F3" "Speech in the//Presence of//Background Music" ""
;; LABEL "F4" "Speech Under//Degraded//Acoustic Conditions" ""
;; LABEL "F5" "Speech from//Non-Native//Speakers" ""
;; LABEL "FX" "All other speech" ""
;; CATEGORY "2" "Speaker Sex" ""
;; LABEL "female" "Female" ""
;; LABEL "male"   "Male" ""
;; LABEL "unknown"   "Unknown" ""';
print "\n";

my $file;

open($file, $ARGV[0]) or die "can't open $ARGV[0] file";
my $basename = basename($ARGV[0], ".tmp2.txt");

#file in tab 
my @fileContent = <$file>;
chomp(@fileContent);

print "$basename 1 UNK 0.0 3600.00 <o,f0,unknown>";

for(my $i = 5; $i < @fileContent - 1; $i += 4){
    $text = cleanText($fileContent[$i + 1]);
    if($text !~ /^$/){
        print " ".$text;
    }
}

sub cleanText
{
    my ($text) = @_;
    $text =~ s/- ?[A-Za-z \.]+ ?://g;
    $text =~ s/^-//g;
    $text =~ s/''//g;
    $text =~ s/-[A-Z]//g;
    $text =~ s/\.+ *$//g;
    $text =~ s/^ *\.+//g;
    $text =~ s/^ +//g;
    $text =~ s/ +$//g;
    return $text;
}