#!/usr/bin/env perl
use 5.012;
use warnings;

my %cc;  # map of country code to info array

while (<>) {
	/^#/ and next;  # skip comments
	my ($iso, $name, $cont) = (split /\t/)[0, 4, 8];
	my $class = "c-\L$cont";
	$cc{ lc $iso } = [ $name, $class ];
}

for (values %cc) {
	for ($_->[2] //= $_->[0]) {
		s/,.*//;
		s/(?<=.)\(.*\)\s*//;
		s/ republic\b//gi;
		s/ islands?\b//gi;
		s/\bthe //g;
		s/ and / & /g and s/(?<=.)[a-z ]+//g;
		s/\bsaint /st /gi;
		s/South(?:ern)? /S-/g;
		s/North(?:ern)? /N-/g;
		s/New /n./g;
		s/(\S)(\S+)-/$1-/g;  # strip most chars preceding dash
		s/(\S{4}[b-df-hj-np-tv-xz])((?<!Austr)(?!land)\w{2,})/$1./g;  # abbreviate (at consonant)
	}
}

say "# automatically generated by $0";
use Data::Dump 'dd';
$Data::Dump::INDENT = '';
dd \%cc;

__END__

=head1 NAME

mkcountries-geonames - Create Perl include of country info from GeoNames data

=head1 SYNOPSIS

	curl http://download.geonames.org/export/dump/countryInfo.txt |
	tools/mkcountryinfo > countries.inc.pl