Show
Ignore:
Timestamp:
11/15/07 16:03:31 (12 months ago)
Author:
yappo
Message:

r12@haruna (orig r10): ko | 2006-06-21 23:16:40 +0900
add normalize rules
add city patterns


Files:
1 modified

Legend:

Unmodified
Added
Removed
  • lang/perl/Geography-AddressExtract-Japan/trunk/bin/make_maps.pl

    r1572 r1574  
    2020my $ban = '��?'; 
    2121 
    22 if (0) { 
     22if (1) { 
    2323my $csv = Text::CSV_PP->new({binary => 1}); 
    2424my $io = IO::File->new('./ken_all.csv', '<:encoding(shiftjis)') or die $!; 
     
    2828while (! $io->eof and my $col = $csv->getline($io)) { 
    2929    my $data = $col->[6] . $col->[7]; 
     30 
     31    my @kana; 
     32    if ($col->[6] ne '����' && $col->[8] !~ /^�ʲ��˷Ǥ�/ && ($col->[8] =~ /^[\p{Hiragana}\p{Katakana}]/ || $col->[8] =~ /��/)) { 
     33        $data = $col->[8]; 
     34        $data =~ s/��/; 
     35        for my $str (split /��/, $data) { 
     36            if ($str =~ /^(\p{Hiragana}+)/ || $str =~ /^(\p{Katakana}+)/) { 
     37                my $kana = $1; 
     38                push @kana, $1; 
     39            } 
     40        } 
     41    } 
     42 
    3043    next if $cache{$data}++; 
    3144 
     
    4861    $ra_city->add("$city"); 
    4962    $ra_city->add($town) if $town; 
     63    for my $town2 (@kana) { 
     64        $ra_city->add("$pref$city$town$town2"); 
     65        $ra_city->add("$pref$town$town2"); 
     66        $ra_city->add("$city$town$town2"); 
     67        $ra_city->add("$town$town2") if $town; 
     68    } 
    5069    add_map($map, $col->[6], $city, $town); 
    5170}