#!/usr/bin/perl
use strict;
use warnings;

use HTML::TableContentParser;
use HTML::TokeParser::Simple;
use WWW::Mechanize;
use Data::Dumper;

my $mech = WWW::Mechanize->new( autocheck => 1 );
$mech->get( 'http://nexus.ugent.be/jeroen/bc.html' );

my @gene_score;

my $table = HTML::TableContentParser->new()->parse( $mech->content() );

for my $rank ( 1 .. $#{ $table->[0]{rows} } ) {
    my $cells = $table->[0]{rows}[$rank]{cells};
    push @gene_score, {
        Rank            => $rank,
        Ensembl_ID      => Get_Label( $cells->[0]{data} ),
        RefSeq_ID       => $cells->[1]{data},
        Gene_Symbol     => $cells->[2]{data},
        Band            => $cells->[3]{data},
        Gene_Name       => $cells->[4]{data},
        Probability     => $cells->[5]{data},
        Known_Phenotype => $cells->[6]{data},
    };
}

sub Get_Label {
    my $link = shift;
    my $p = HTML::TokeParser::Simple->new( \$link );
    while ( my $token = $p->get_token ) {
        next if ! $token->is_start_tag;
        return $token->return_attr( 'href' ) =~ /gene=(\w+)$/;
    }
}