#!/usr/bin/perl -w

# Create a list of unique urls from the set of all urls 
# and place publication year at begninning
# Sort by pub year

# Example:
# 1995 http://www.abc.com
# 1995 http://www.bbb.edu


use strict;

open(URLS, "url_id_list.txt") || die($!);
my @lines = <URLS>;
close URLS;

# Grab all article headings.
# 1.1 article url
# 1.2 article url

my @articles;

foreach (@lines) {
	chomp($_);
	if (/^\d+\.\d+ (.+)$/) {
		push(@articles, $_);
	}
}

my $article_count = @articles;

my %seen;

foreach my $article (@articles) {
	#print "$article\n";
	if ($article =~ /^(\d+)\.(\d+) (.+)$/) {
		my $sec1 = $1;
		my $sec2 = $2;
	
		my ($year) = $article =~ /dlib\/\w+(\d\d)\//;
                if ($year > 4) {
                        $year = "19$year";
                }
                else {
                        $year = "20$year";
                }

		# Now find all urls for this article
		foreach (@lines) {
			if (/^$sec1\.$sec2\.\d+ (.+)$/) {
				my $url = $1;
				if (!defined $seen{$url}) {
					$seen{$url} = $year;
				}
				else {
					# Since we've seen this url before,
					# make sure it has earliest pub year
					if ($seen{$url} > $year) {
						$seen{$url} = $year;
					}
				}
			}
		}
	}
}

foreach my $url (sort keys %seen) {
	print $seen{$url} . "\t$url\n";
}