SweetPotato::Plagger

■ [CustomFeed-Script] Web::Scraper for CF::Script コミックHOLIC

コミックHOLICの各話を取得。

assets/plugins/CustomFeed-Script/comicholic.pl

#!/usr/bin/perl
use strict;
use warnings;
use utf8;

use URI;
use Web::Scraper;
use YAML;

my $url = 'http://www.toranoana.jp/webcomic/holic/';

my $s = scraper {
    process '//div[./a[@target="_blank"]]', 'comic[]' => scraper {
        process '/div/a',     link  => ['@href', sub { $_->as_string } ];
        process '/div/a/img', icon  => ['@src', sub { +{ url => $_->as_string } } ];
        process '/div',       title => 'text';
        result qw/link icon title/;
    };
    result qw/comic/;
};

binmode STDOUT, ':utf8';
print YAML::Dump +{
    title => 'コミックHOLIC',
    link  => $url,
    entry => $s->scrape(URI->new($url)),
};

config.yaml

plugins:
  - module: Subscription::Config
    config:
      feed:
        - url: 'script:/path/to/comicholic.pl'

  - module: CustomFeed::Script

■ [CustomFeed-Script] Web::Scraperを使ってFlexComixブラッドのassetsをYahoo!コミック無料マガジン全般対応版に書き直した 00:53

• scraper for CustomFeed::Config FlexComix ブラッド - SweetPotato::Plagger - Plaggerグループ

以前作った↑を，Web::Scraperを使ってYahoo!コミック無料マガジン全般対応版に書き直した。

assets/plugins/CustomFeed-Script/yahoocomic.pl

#!/usr/bin/perl
use strict;
use warnings;
use utf8;

use URI;
use Web::Scraper;
use YAML;

my $mag = shift;
   $mag = "http://comics.yahoo.co.jp/magazine/${mag}_0001.html" unless $mag =~ m!^http://!;

my $s = scraper {
    process '//table[@cellspacing="4"][1]//b', title => 'text';
    process '//table[@cellpadding="4" and .//img and count(.//tr)=4]',
      'comic[]' => scraper {
        process '//tr[1]//img',   icon   => ['@src', sub { +{ url => $_->as_string } } ];
        process '//tr[2]//a',     link   => ['@href', sub { $_->as_string } ];
        process '//tr[2]//a/b',   title  => 'text';
        process '//tr[3]//small', author => ['text', sub { (/：(.*?)\s*$/)[0] } ];
        process '//tr[4]//small', body   => 'text';
        result qw/icon link title author body/;
      };
    result qw/title comic/;
};

my $res = $s->scrape(URI->new($mag));

binmode STDOUT, ':utf8';
print YAML::Dump +{
    title => $res->{title},
    link  => $mag,
    entry => $res->{comic},
};

config.yaml

plugins:
  - module: Subscription::Config
    config:
      feed:
        - url: 'script:/path/to/yahoocomic.pl blood'
        - url: 'script:/path/to/yahoocomic.pl flare'
        - url: 'script:/path/to/yahoocomic.pl weeklyyj'
        - url: 'script:/path/to/yahoocomic.pl fang'
        - url: 'script:/path/to/yahoocomic.pl gumbo'
        - url: 'script:/path/to/yahoocomic.pl marjong'
        - url: 'script:/path/to/yahoocomic.pl sangokushi'
        - url: 'script:/path/to/yahoocomic.pl essay'

  - module: CustomFeed::Script

  - module: SmartFeed::All
    config:
      title: Yahoo!コミック > 無料マガジン
      link: http://comics.yahoo.co.jp/magazine/