Class: Altavista

Inherits:
Object
  • Object
show all
Includes:
Searchy
Defined in:
lib/esearchy/SearchEngines/altavista.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from Searchy

#clean, #fix, #hash_url, #maxhits=, #print_emails, #search_depth, #search_docs, #search_emails, #search_office_xml, #search_pdfs, #search_txts

Constructor Details

#initialize(maxhits = 0, start = 0) ⇒ Altavista

Returns a new instance of Altavista.



8
9
10
11
12
13
14
15
16
17
18
19
 
# File 'lib/esearchy/SearchEngines/altavista.rb', line 8

def initialize(maxhits = 0, start = 0)
  @start = start
  @totalhits = maxhits
  @emails = []
  @r_urls = Queue.new
  @r_docs = Queue.new
  @r_pdfs = Queue.new
  @r_txts = Queue.new
  @r_officexs = Queue.new
  @lock = Mutex.new
  @threads = []
end

Instance Attribute Details

#emails ⇒ Object

Returns the value of attribute emails.



20
21
22
 
# File 'lib/esearchy/SearchEngines/altavista.rb', line 20

def emails
  @emails
end

Instance Method Details

#parse(html) ⇒ Object 



54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
 
# File 'lib/esearchy/SearchEngines/altavista.rb', line 54

def parse(html)
  @totalhits= html.scan(/AltaVista found (.*) results<\/A>/)[0][0].gsub(',','').to_i if @totalhits == 0
  html.scan(/<a class='res' href='([a-zA-Z0-9:\/\/.&?%=\-_+]*)'>/).each do |result|
    case result[0]
    when /.pdf$/i
      @r_pdfs << CGI.unescape(result[0])
    when /.docx$|.xlsx$|.pptx$|.odt$|.odp$|.ods$|.odb$/i
      @r_officexs << CGI.unescape(result[0])
    when /.doc$/i
      @r_docs << CGI.unescape(result[0])
    when /.txt$|.rtf$|ans$/i
      @r_txts << CGI.unescape(result[0])
    else
      @r_urls << CGI.unescape(result[0])
    end
  end
end

#search(query) ⇒ Object 



22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
 
# File 'lib/esearchy/SearchEngines/altavista.rb', line 22

def search(query)
  @query = query
  http = Net::HTTP.new("www.altavista.com",80)
  begin
    http.start do |http|
      request = Net::HTTP::Get.new( "/web/results?itag=ody&kgs=0&kls=0&nbq=50" + 
                                     "&q=" + CGI.escape(query) + 
                                     "&stq=#{@start}", 
                                     {'User-Agent' => UserAgent::fetch})
      response = http.request(request)
      case response
      when Net::HTTPSuccess, Net::HTTPRedirection
        parse(response.body)
        @start = @start + 100
        if @totalhits > @start
          ESearchy::LOG.puts "Searching #{self.class} from #{@start-50} to #{@start}"
          search_emails(response.body.gsub(/<b>|<\/b>/,""))
          sleep(ESearchy::DELAY)
          search(query)
        else
          ESearchy::LOG.puts "Searching #{self.class} from #{@start-50} to #{@start}"
          search_emails(response.body.gsub(/<b>|<\/b>/,""))
        end
      else
        return response.error!
      end
    end
  rescue Net::HTTPFatalError
    ESearchy::LOG.puts "Error: Something went wrong with the HTTP request"
  end
end