Class: Crown::Amazon::AsinExtractor

Inherits:
Object
  • Object
show all
Defined in:
lib/crown/amazon.rb

Overview

——————————————————————- #

AsinExtractor

指定した URI の a タグ href 属性から Amazon の ASIN と思われる
情報を抽出するクラス．

——————————————————————- #

Instance Method Summary collapse

Constructor Details

#initialize(proxy_host = nil, proxy_port = nil) ⇒ AsinExtractor

Returns a new instance of AsinExtractor.



53
54
55
56
 
# File 'lib/crown/amazon.rb', line 53

def initialize(proxy_host = nil, proxy_port = nil)
    @proxy_host = proxy_host
    @proxy_port = proxy_port
end

Instance Method Details

#get(uri) ⇒ Object

————————————————————— #

get

————————————————————— #



61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
 
# File 'lib/crown/amazon.rb', line 61

def get(uri)
    result = Array.new
    
    parser = URI.parse(uri.strip)
    path = parser.path
    path += '?' + parser.query if (parser.query != nil)
    
    session = Net::HTTP.new(parser.host, parser.port, @proxy_host, @proxy_port)
    response = Crown::HTTP.get(session, path)
    return result if (response == nil || response.code.to_i != 200)
    
    html = Nokogiri::HTML.parse(response.body)
    html.search("a").each { |node|
        next if (node['href'] == nil)
        
        begin
            parser = URI.parse(node['href'].strip)
        rescue URI::InvalidURIError
            parser = URI.parse(URI.encode(node['href'].strip))
        end
        next if (parser == nil || parser.host == nil || parser.path == nil)
        
        if (parser.host.match(/^(?:www\.)?amazon\.(?:com|ca|co\.uk|de|co\.jp|jp|fr|cn)$/) != nil)
            asin = get_asin(parser.path, parser.query)
            result.push(asin) if (asin != nil && !result.include?(asin))
        end
    }
    
    return result
end