英辞郎 on the web への問い合わせリストを Firefox history.dat から作成する

 1  #!/usr/bin/perl 
 2  use strict;
 3  use warnings;
 4
 5  use File::Mork;
 6  use URI::Escape;
 7  use Encode;
 8  use Getopt::Long;
 9
10  my %map = (atime => 'LastVisitDate',
11             ctime => 'FirstVisitDate',
12             count => 'VisitCount');
13
14  my $word_in_re = qr/([^&]+)/;
15  my $ejr_re = quotemeta('http://www2.alc.co.jp/ejr/index.php?word_in=')
16             . $word_in_re;
17
18  my $file;
19  my $sort = 'atime';
20  GetOptions ("file=s"   => \$file,
21              "sort=s"   => \$sort);
22
23  unless ($file and $map{$sort}) {
24      print "Usage: $0 --file histroy.dat [--sort atime|ctime|count]\n\n";
25      exit;
26  }
27
28  my $mork = File::Mork->new($file, verbose => 1)
29      or die $File::Mork::ERROR."\n";
30
31
31
32  my $key = $map{$sort};
33  for my $entry ( sort {$b->{$key} <=> $a->{$key}}
34                  map  {$_->{$key} = 0 unless $_->{$key}; $_}
35                  $mork->entries ) {
36
37      next unless $entry->{URL} and $entry->{URL} =~ /^$ejr_re/;
38      my $word_in = $1;
39
40      $word_in = uri_unescape($word_in);
41
42      Encode::from_to($word_in, 'Shift_JIS', 'utf-8')
43          if $^O ne 'MSWin32' and $word_in =~ /[^[:ascii:]]/ ;
44
45      $entry->{VisitCount} = 0 unless $entry->{VisitCount};
46      print join("\t", $entry->{VisitCount}, $word_in), "\n";
47  }
48
49  1;
50