Class: BaiduResult

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

Instance Method Summary collapse

Constructor Details

#initialize(page) ⇒ BaiduResult

Returns a new instance of BaiduResult.



103
104
105
106
 
# File 'lib/baidu.rb', line 103

def initialize(page)
    raise ArgumentError 'should be Mechanize::Page' unless page.class == Mechanize::Page
    @page = page
end

Instance Method Details

#how_many ⇒ Object 



154
155
156
 
# File 'lib/baidu.rb', line 154

def how_many
    @how_many ||= @page.search("//span[@class='nums']").map{|num|num.content.gsub(/\D/,'').to_i unless num.nil?}.first
end

#next ⇒ Object 



162
163
164
 
# File 'lib/baidu.rb', line 162

def next
    @page = BaiduResult.new(Mechanize.new.click(@page.link_with(:text=>/下一页/))) unless @page.link_with(:text=>/下一页/).nil?
end

#rank(host) ⇒ Object

return the top rank number from @ranks with the input host



143
144
145
146
147
148
149
150
151
152
 
# File 'lib/baidu.rb', line 143

def rank(host)#on base of ranks
    ranks.each do |id,line|
        if host.class == Regexp
            return id if line['host'] =~ host
        elsif host.class == String
            return id if line['host'] == host
        end
    end
    return nil
end

#ranks(host = nil) ⇒ Object 



108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
 
# File 'lib/baidu.rb', line 108

def ranks(host=nil)
    return @ranks unless @ranks.nil?
    @ranks = Hash.new
    @page.search("//table[@class=\"result\"]").each do |table|
        id = table['id']
        @ranks[id] = Hash.new
        url = @page.search("//table[@id=\"#{table['id']}\"]//span[@class=\"g\"]").first
        a = @page.search("//table[@id=\"#{table['id']}\"]//h3/a")
        @ranks[id]['text'] = a.text
        @ranks[id]['href'] = a.first['href'].sub('http://www.baidu.com/link?url=','').strip
        unless url.nil?
            url = url.text.strip
            @ranks[id]['host'] = URI(URI.encode("http://#{url}")).host
        else
            @ranks[id]['host'] = nil
        end
    end
    #@page.search("//table[@class=\"result\"]").map{|table|@page.search("//table[@id=\"#{table['id']}\"]//span[@class=\"g\"]").first}.map{|rank|URI(URI.encode('http://'+rank.text.strip)).host unless rank.nil?}
    if host.nil?
        @ranks
    else
        host_ranks = Hash.new
        @ranks.each do |id,line|
            if host.class == Regexp
                host_ranks[id] = line if line['host'] =~ host
            elsif host.class == String
                host_ranks[id] = line if line['host'] == host
            end
        end
        host_ranks
        #'not finished'#@ranks.each_with_index.map{|h,i| i if !h.nil? and h==host}.compact
    end
end

#related_keywords ⇒ Object 



158
159
160
 
# File 'lib/baidu.rb', line 158

def related_keywords
    @related_keywords ||= @page.search("//div[@id=\"rs\"]//tr//a").map{|keyword| keyword.text}
end