#!/usr/bin/perl

use strict;
use warnings;
use diagnostics;
use Set::Scalar;
use WWW::Mechanize;
use URI;

# URL for DBLP search
my $dblpURL = 'http://www.informatik.uni-trier.de/~ley/db/indices/a-tree/';
# Delay time
my ($delayMean, $delayVar) = (5, 3);

# print headers of output
print join("\t", ('name', 'DBLP entry', 'home page', 'email')), "\n";

# iterate over list of author's names
my $counter = 0;
while (<>) {
  # since one should be polite (or stealthy), sleep for a while
  sleep($delayMean + int($delayVar/2 - rand($delayVar)));

  $counter++;
  my @data;
  s/^[ ]*(.+?)[ \r\n]*$/$1/;
  my ($name, $dblpEntry, $homePage, $email) = split(/\t/, $_);
  print STDERR "now handling record $counter: '$name'\n";
  if (!defined $homePage || $homePage eq '') {
    push(@data, $name);
    my $mech = WWW::Mechanize->new();
    $mech->agent_alias('Windows IE 6');
    # get DBLP search page
    $mech->get($dblpURL);

    # insert author's name in search page and submit
    $mech->form_number(1);
    $mech->field('author', $name);
    $mech->submit();

    # if author has a DBLP entry, the resulting page will
    # have his name as title, if not, stop processing
    # author
    if ($mech->title() !~ /$name/) {
      print join("\t", (@data, 0, "", "")), "\n";
      next;
    }
    push(@data, 1);

    # search for a link that has 'Home Page' as text and
    # follow it, stop processing if there is none
    if (!defined $mech->follow_link(text => "Home Page", n => 1) &&
        !defined $mech->follow_link(text => "Home page", n => 1) &&
        !defined $mech->follow_link(text => "home Page", n => 1) &&
        !defined $mech->follow_link(text => "home page", n => 1) &&
        !defined $mech->follow_link(text => "Homepage", n => 1) &&
        !defined $mech->follow_link(text => "HomePage", n => 1) &&
        !defined $mech->follow_link(text => "homePage", n => 1) &&
        !defined $mech->follow_link(text => "homepage", n => 1)) {
      print join("\t", (@data, "", "")), "\n";
      next;
    }
    push(@data, $mech->uri());

    # retrieve all links on the author's home page and
    # output only those that are 'mailto' URLs
    my @links = map { URI->new($_->[0]) } $mech->links();
    my $addresses = Set::Scalar->new();
    foreach my $link (@links) {
      if (defined $link->scheme() && ($link->scheme() eq 'mailto') &&
          !$addresses->contains($link->opaque())) {
        print join("\t", (@data, $link->opaque())), "\n";
        $addresses->insert($link->opaque());
      }
    }
    print join("\t", (@data, "")), "\n" if $addresses->size() == 0;
  } else {
    print join("\t", ($name, $dblpEntry, $homePage, $email)), "\n";
  }
}