view release on metacpan or  search on metacpan

lib/Acme/Geo/Whitwell/Name.pm  view on Meta::CPAN

package Acme::Geo::Whitwell::Name;

use strict;
use warnings;
use Carp qw(croak);

use Exporter;
@Acme::Geo::Whitwell::Name::ISA       = qw(Exporter);
@Acme::Geo::Whitwell::Name::EXPORT_OK = qw(to_whitwell from_whitwell);

use Scalar::Util qw(looks_like_number);

=head1 NAME

Acme::Geo::Whitwell::Name - Steadman Whitwell's "rational geographic nomenclature"

=head1 VERSION

Version 0.04

=cut

our $VERSION = '0.04';

=head1 SYNOPSIS

    use Acme::Geo::Whitwell::Name qw(to_whitwell from_whitwell);

    # Convert Sunnyvale, CA's lat and lon to a Whitwell name pair.
    my @names = to_whitwell("37.37N", "122.03");

    # Same conversion, using signed latitude and longitude instead.
    my @names = to_whitwell(37.37, -122.03);

    # Convert a Whitwell name to a latitude and longitude.
    # (Washington DC's "rational" name to N/S lat and E/W long.)
    my($lat_string, $lon_string) = from_whitwell("Feiro Nyvout");

    # If you want signed values, add signed => some true value.
    my($lat, $long) = from_whitwell("Feiro Nyvout", signed=>1);

=head1 DESCRIPTION

This module implements Steadman Whitwell's "rational system of geographic 
nomenclature", in which place names are generated by converting the latitude 
and longitude of the location into a two-part name by means of a 
transliteration scheme.

Whitwell devised this scheme in an attempt to provide an alternative to 
the proliferation of similarly-named towns in the early US. However, people
seemed to prefer creating many Springfields and Washingtons in preference to 
using Whitwell's uniquely quirky names. 

=head2 THE SCHEME

Two tables of number-to-letter(s) are used to translate latitudes and 
longitudes of two-decimal precision, digit-by-digit, into 
vaguely-pronounceable two-part names. 

             1 2 3 4 5 6 7  8  9  0
   latitude  a e i o u y ee ei ie ou  vowels
  longitude  b d f k l m n  p  r  t   consonants

Transliteration is done by looking up the apropriate digit in the tables above,
switching rows until all the digits are consumed. If the coordinate is negative,
a special 'sign consonant' is inserted into the (partial) name after the first 
vowel is added, and the transliteration continues by choosing again from the 
vowel table, then continuing to alternate again.

This is very orderly, but confusing to generate by hand (putting aside the 
fact that no one in their right mind really wants to live in "Isilu Buban"

lib/Acme/Geo/Whitwell/Name.pm  view on Meta::CPAN

=cut

# These tables define the letters that the numbers will be transliterated into.
#                   0  1 2 3 4 5 6 7  8  9
my @vowels     = qw(ou a e i o u y ee ei ie); 
my @consonants = qw(t  b d f k l m n  p  r);

# Allows us to detect when to insert the "sign consonant" for negative 
# lats and lons.
my %vowel;
@vowel{@vowels} = (); 

=head1 EXPORT

=head1 FUNCTIONS

=head2 to_whitwell($lat, $lon)

Generates a properly-capitalized Whitwell name from a latitude-longitude pair.
Latitude and longitude are truncated to the two digits after the decimal point,
in keeping with Whitwell's original scheme. Zeroes are added after the decimal
point as necessary.

North latitudes are positve, and south latitudes are negative. East longitudes
are positive, west longitudes are negative. Trailing E/W and N/S are converted
into the appropriate sign. If you supply both for some reason, trailing 
sign indicators override signs.

Returns both alternatives for the name (see L<SCHEME>).

=cut

sub to_whitwell {
     my($lat, $lon) = @_;
     return ( _vowel_build($lat)     . ' ' . _consonant_build($lon),
              _consonant_build($lat) . ' ' . _vowel_build($lon)
     );
}

sub _vowel_build     { _gen(shift, [\@vowels, \@consonants], 's') }
sub _consonant_build { _gen(shift, [\@consonants, \@vowels], 'v') }

sub _gen {
    # The coordinate, the letter lists, and the appropriate sign consonant.
    my($coord, $lists, $neg) = @_;

    # Turn the floating-point number into a list of digits.
    # Note that _two_decimal does NOT CARE about sign or sign indicators.
    $coord = uc(my $orig_coord = $coord);
    my @coord = grep {/(\d)/} (split //, _two_decimal($coord));

    my $word = '';
    my $list = 0;
    my $signed = 0;

    my ($is_negative) = ($coord =~ s/[SW]//g);
    my ($is_positive) = ($coord =~ s/[NE]//g);

    croak 
        "Coordinate '$orig_coord' does not look like a proper coordinate"
            if !looks_like_number($coord);

    $is_negative = ($coord < 0) unless $is_negative;

    my $conflicting = ($is_negative and $is_positive) ? 'conflicting ' : '';
    croak "Multiple ${conflicting}sign indicators detected in '$orig_coord'"
      if  $conflicting or $is_negative > 1 or $is_positive > 1;

    foreach my $digit (@coord) {
        # Convert the next digit into a letter from the proper table.
        my $letter = $lists->[$list]->[$digit];
        ### "$letter -> $digit"

        # Decide whether to insert a sign consonant.        
        if (exists $vowel{$letter} and $is_negative and not $signed) {
            # If negative, we have a vowel, and we haven't inserted the sign
            # consonant yet, insert it.
            $letter .= $neg;
            # Now signed.
            $signed = 1;
            $list = !$list;
        }
        # Add new letter(s) to word and continue;
        $word .= $letter;
        $list = !$list;
    }
    return ucfirst $word;
}

sub _two_decimal {
    my ($coord) = @_;
    
    # Discard non-digits except for a decimal point.
    $coord =~ s/[^\d\.]//g;

    # Drop leading zeros.
    $coord =~ s/^0*//g;
    $coord = 0 unless $coord;

    if (abs($coord) > 180) {
        croak "$coord must be between -180 and +180\n";
    }
    unless ($coord =~ /\./) {
        # add decimals
        $coord .= ".";
    }
    # Add two more zeroes; we'll discard them if we don't need them.
    $coord .= "00"; 
    ($coord) = ($coord =~ /^(\d{0,3}\.\d\d)/);
    return $coord;
}

=head2 from_whitwell($whitwell_name, signed => $yes_or_no)

Converts a Whitwell name back into a lat/lon pair, in trailing indicator
format.  Results will be undefined if the string does not match the Whitwell
scheme; if the strings I<is> Whitwell-compatible, but includes extra letters,
these will be assumed to be further digits after the decimal point.

If you supply the 'signed' option with a true value, the returned values are
signed numbers rather than numbers with trailing sign indicators.