Class: Smartdict::Drivers::LingvoYandexDriver

Inherits:
AbstractDriver show all
Defined in:
lib/smartdict/drivers/lingvo_yandex_driver.rb

Overview

The translation driver for Google Translate service.

DISCLAIMER: It’s was written when I had one hand broken. Refactoring costs a lot of movements so that I’ve left it as it was. I’m gonna refactor it soon. – Sergey Potapov

TODO:

* Refactor

Constant Summary collapse

USER_AGENT =

Pretend being Firefox :)

"Mozilla/5.0 (X11; U; Linux x86_64; ru; rv:1.9.1.16) Gecko/20110429 Iceweasel/3.5.16 (like Firefox/3.5.1623123)"
HOST =

Host of Lingvo service.

"slovari.yandex.ru"
WORD_CLASSES =

Mapping for word classes. Default is “other”

{
  "имя существительное"     => "noun",
  "имя прилагательное"      => "adjective",
  "глагол"                  => "verb",
  "наречие"                 => "adverb",
  "предлог"                 => "preposition",
  "имя числительное"        => "numeral",
  "междометие (часть речи)" => "interjection",
  "сокращение"              => "abbreviation",
  "местоимение"             => "pronoun",
  "союз (часть речи)"       => "conjunction"
}.tap{ |hash| hash.default = "other" }

Instance Attribute Summary

Attributes inherited from AbstractDriver

#from_lang, #to_lang, #transcription, #translated, #word

Instance Method Summary collapse

Methods inherited from AbstractDriver

#build_translation, #initialize, set_name, translate

Constructor Details

This class inherits a constructor from Smartdict::Drivers::AbstractDriver

Instance Method Details

#escape(str) ⇒ Object 



134
135
136
 
# File 'lib/smartdict/drivers/lingvo_yandex_driver.rb', line 134

def escape(str)
  CGI.escape(str)
end

#get_response ⇒ Object 



117
118
119
120
121
 
# File 'lib/smartdict/drivers/lingvo_yandex_driver.rb', line 117

def get_response
  http = Net::HTTP.new(HOST, 80)
  request = Net::HTTP::Get.new(http_path, { "User-Agent" => USER_AGENT })
  http.request(request).read_body
end

#grep_meanings(html_element) ⇒ Object

TODO: refactor



69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
 
# File 'lib/smartdict/drivers/lingvo_yandex_driver.rb', line 69

def grep_meanings(html_element)
  acronym = html_element.css("acronym").first
  return unless acronym

  ru_word_class = acronym["title"]
  word_class = WORD_CLASSES[ru_word_class]
  translations = []

  html_element.css("ul > li").each do |tr|
    # a text line with translations separated by commas
    line = ""

    # use strong tag as an anchor
    strong = tr.css("strong").first
    if strong && strong.text =~ /\d+|[а-я]+\)/
      node = strong
      while(node = node.next_sibling)
        if node.text? || node.name == "a"
          text = node.text
          line << text unless text =~ /\(|\)/
        elsif ["em", "acronym"].include? node.name
          next
        else
          break
        end
      end
    end
    translations += words_from_line(line)
  end


  # sometimes there is only one meaning
  if translations.empty?
    if a_tag = html_element.css("span > a").first
      line = a_tag.text
    elsif span = html_element.css("span").first
      line = span.text
    elsif i_tag = html_element.xpath("i[2]")
      line = i_tag.text
    else
      return nil
    end
    translations = words_from_line(line)
  end

  self.translated[word_class] = translations.uniq
end

#http_path ⇒ String

Returns http path for request to translate word.

Returns:

  • (String) —

    http path for request to translate word.



124
125
126
127
128
129
130
131
132
 
# File 'lib/smartdict/drivers/lingvo_yandex_driver.rb', line 124

def http_path
  phrase = case [from_lang, to_lang]
  when ["en", "ru"] then "en-ru"
  when ["ru", "en"] then "ru-en"
  else raise Smartdict::TranslationNotFound
  end

  "/#{escape(word)}/#{phrase}/"
end

#translate ⇒ Object

TODO: refactor

Raises:



43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
 
# File 'lib/smartdict/drivers/lingvo_yandex_driver.rb', line 43

def translate
  doc = Nokogiri::HTML(get_response)

  if main = doc.css("div.b-translation__article > ul > li#I").first
  else
    main = doc.css("div.b-translation__article").first
  end

  raise Smartdict::TranslationNotFound unless main

  # Fetch transcription
  self.transcription = doc.css("span.b-translation__tr").first.try(:text)


  self.translated = {}

  if main.xpath("./i/acronym").any?
    grep_meanings(main)
  else
     main.xpath("./ul/li").each do |li|
      grep_meanings(li)
    end
  end
end