#!/usr/bin/perl

####################################
#
# TEclass2txt -f html-file -o text-file
#
# Reads TEclass output as html file and writes to text file as tab-delimited 
# Authors: C. Grunau, D. Roquis, 28 mars 2011
#####################################

use diagnostics;
use strict;
use Carp;
use FileHandle;
use File::Path;
use File::Basename;
use Getopt::Std;

my $FILE              = FileHandle->new();
my $start = "<tr class=";
my $end = "</td> </tr>";
my $result = "";
my $text = "";
my $extraction = "";
my $header = "No\tID\tResult\tORFs\tStrand\tSVM\ classif\ (4mer)\tSVM\ classif\ (5mer)\tLVQ\ classif\tRF\ classif\n";
my $footer = "";

use vars qw( $opt_f $opt_o );
# Command line parsing 
#
getopts('f:o:');
#
# Read input file and split into fasta sequences on the fly
#
#replace Mac and PC CRLF by UNIX LF
system ("perl -pi -e 's/\r/\n/g' $opt_f");
open($FILE,$opt_f) || croak(sprintf("Cannot open file \"%s\"",$opt_f));
while ( <$FILE> ) {
    chomp;
    $text = $text.$_;
       }
close($FILE) || croak(sprintf("Cannot close open file \"%s\"",$opt_f));


#Copy all the table rows in the prevoiously selected text
    while ($text =~ /$start(.*?)$end/g) { $result = $start.$1.$end; $extraction = $extraction.$result;}

#Lines 68 to 80 removes all HTML tags in order to save the output as a .txt fil

#Replace all white space caracters and replace the ones within a cell with underscores
$extraction =~ s/\ \-\ /\-/gms;
$extraction =~ s/\>\s\</\>\</gms;
$extraction =~ s/\s/_/gms;

#Replace the end of a row in a table by a  newline caracter
$extraction =~ s/\<\/th><\/tr><tr_class="odd"><td_width=50px>/\n/g;
$extraction =~ s/\<\/td>\<\/tr><tr_class="even"><td_width=50px>/\n/g;
$extraction =~ s/\<\/td>\<\/tr><tr_class="odd"><td_width=50px>/\n/g;

#Remove all the remaining HTML tags and replace them by tabulations
$extraction =~ s/[\<[^\<]+\>]+/\t/gs;
$extraction =~ s/\<[^\<]+\>/\t/gs;

$extraction =~ s/[\t ]+/\t/gs; # Replace duplicate whitespace mid-string with 1 tab
$extraction =~ s/[\t ]*$//gms; # Remove ending spaces/tabs
$extraction =~ s/^[\t ]*//gms; # Remove starting spaces/tabs
$extraction =~ s/_/\ /gms; # Replace the underscores with spaces

#Removes all the headers if the input was from a concatenation
#$extraction =~ s/No\.\sID\sResult\sORFs\sStrand\sSVM\sclassif\.\s\(4mer\)\sSVM\sclassif\.\s\(5mer\)\sLVQ\tclassif\.\sRF\sclassif\.\n//g; 

#make sure file is empty, overwrites existing file without mercy
open($FILE,">".$opt_o) || croak(sprintf("Cannot open file \"%s\"",$opt_o));
         print $FILE "";
close ($FILE);

#Add the header and the footer
$extraction = $header.$extraction.$footer;

open($FILE,">".$opt_o) || croak(sprintf("Cannot open file \"%s\"",$opt_o));
         print $FILE $extraction;
close ($FILE);