#!/usr/bin/perl

# Task: Extract GeneID-Number and gene information and store them in a hash

use strict;
use warnings;

my $in;
my $data;
my @array;
my $GeneID;
my @BMB;
my $flag = 0;
my %hash;
my $gb;
my $product;
my $novalue = 0;


# 1) open the .gff Inputfile and while reading line by line split $data at each tab and put them in the @array


open $in, '<', "Genomteil.gff" or die $!;

while ($data = <$in>) {
    @array = split (/\t/, $data);

# 2) as long the flag == 0, search for the word 'gene' in the second element of @array. If you find it, set the flag ($flag=1)


   if ($flag==0){
        if ($array[2] =~ /gene/) {    	
            $flag = 1; 	

# 3) if the flag is set, put the three elements in my new array @BMB
# now I have $BMB[0] = start point; $BMB[1] = end point; $BMB[2] = + or -
	     	 
	    @BMB = ($array[3], $array[4], $array[6]);    
        }
# 4) #if you find the word 'GeneID' extract the following number and store it in $GeneID						 
       
	 if ($array[8] =~ /.*;db_xref=GeneID:(\d+)\n/) {     
	 $GeneID = $1;
        } 
	
# 5) if the flag is still set (==1), search for the word 'CDS' or 'exon' in $array[2] and put it in @BMB

					 
   }elsif ($flag==1){
        if ($array[2] =~ /CDS/) {
            push (@BMB, $array[2]);
            						
        } elsif ($array[2] =~ /exon/) {
            push (@BMB, $array[2]);
        

# 6) now I have $BMB [3] = CDS or exon
# 7) if you find the word 'gbkey' (it's an exon!) in $array[8] store the following word (rRNA or tRNA) in $gb and add it to @BMB

	if ($array[8]=~ /gbkey=(\w{4})/){
	$gb = $1;
	push (@BMB, $gb);

# 8) if you don't find 'gbkey' (it's a CDS!) add $nonvalue to @BMB, which has the value 0
# now I have $BMB[4] = rRNA or tRNA or 0
	
	}elsif ($array[8]!~/gbkey=(\w{4})/){
	push (@BMB, $novalue);
	}

# 9) if ypu find the word 'product=' followed by ';protein_' or by ';db_' extract the product name and add it to @BMB
# now I have $BMB[5] = product name

	if ($array[8]=~ /product=(.*);protein_/){
	$product = $1;
	push (@BMB, $product);
	}
	elsif ($array[8]=~ /product=(.*);db_/){
	$product =$1;
	push (@BMB, $product);
	}

# 10) create a hash, use GeneID as key and @BMB for the values

	$hash{$GeneID} = [ @BMB ];

# 11) if you find the term ';exon_number=1' reset the flag ($flaf=0)

	if ($array [8]=~ /.*;exon_number=1/){
	  $flag = 0;	
	}						
   }
# 12) search the next 'gene' while flag==0							
}
close $in;


# 13) print the keys with all their values

foreach my $key (keys %hash) {
    foreach my $val (@{$hash{$key}}) {
        print "$key --> $val\n";
    }
}