Class: Alexandria::BookProviders::ProxisProvider

Inherits:
WebsiteBasedProvider show all
Includes:
Logging
Defined in:
lib/alexandria/book_providers/proxis.rb

Constant Summary collapse

SITE =

Proxis essentially has three book databases, NL, FR and EN. Currently, this provider only searches the NL database, since it adds most to Alexandria (Amazon already has French and English titles).

'http://www.proxis.nl'
BASE_SEARCH_URL = 
"#{SITE}/NLNL/Search/IndexGSA.aspx?search=%s" \
'&shop=100001NL&SelRubricLevel1Id=100001NL'
ISBN_REDIRECT_BASE_URL = 
"#{SITE}/NLNL/Search/Index.aspx?search=%s" \
'&shop=100001NL&SelRubricLevel1Id=100001NL'

Instance Attribute Summary

Attributes inherited from AbstractProvider

#fullname, #name, #prefs

Instance Method Summary collapse

Methods included from Logging

included, #log

Methods inherited from WebsiteBasedProvider

#html_to_doc

Methods inherited from AbstractProvider

#<=>, abstract?, #abstract?, #action_name, #enabled, #reinitialize, #remove, #toggle_enabled, #transport, unabstract, #variable_name

Constructor Details

#initialize ⇒ ProxisProvider

Returns a new instance of ProxisProvider.



45
46
47
48
49
50
 
# File 'lib/alexandria/book_providers/proxis.rb', line 45

def initialize
  super('Proxis', 'Proxis (Belgium)')
  # prefs.add("lang", _("Language"), "fr",
  #          LANGUAGES.keys)
  prefs.read
end

Instance Method Details

#create_search_uri(search_type, search_term) ⇒ Object 



66
67
68
69
70
71
72
 
# File 'lib/alexandria/book_providers/proxis.rb', line 66

def create_search_uri(search_type, search_term)
  if search_type == SEARCH_BY_ISBN
    BASE_SEARCH_URL % Library.canonicalise_ean(search_term)
  else
    BASE_SEARCH_URL % CGI.escape(search_term)
  end
end

#data_for_header(th) ⇒ Object 



123
124
125
126
127
 
# File 'lib/alexandria/book_providers/proxis.rb', line 123

def data_for_header(th)
  tr = th.parent
  td = tr.at('td')
  text_of(td) if td
end

#get_book_from_search_result(result) ⇒ Object 



74
75
76
77
78
 
# File 'lib/alexandria/book_providers/proxis.rb', line 74

def get_book_from_search_result(result)
  log.debug { "Fetching book from #{result[:lookup_url]}" }
  html_data = transport.get_response(URI.parse(result[:lookup_url]))
  parse_result_data(html_data.body)
end

#parse_result_data(html) ⇒ Object 



129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
 
# File 'lib/alexandria/book_providers/proxis.rb', line 129

def parse_result_data(html)
  doc = html_to_doc(html)
  book_data = {}
  book_data[:authors] = []
  # TITLE
  if (title_header = doc.search('div.detailBlock h3'))
    header_spans = title_header.first.search('span')
    title = text_of(header_spans.first)
    title = Regexp.last_match[1].strip if title =~ /(.+)-$/
    book_data[:title] = title
  end

  info_headers = doc.search('table.productInfoTable th')

  isbns = []
  unless info_headers.empty?
    info_headers.each do |th|
      isbns << data_for_header(th) if th.inner_text =~ /(ISBN|EAN)/
    end
    book_data[:isbn] = Library.canonicalise_ean(isbns.first)
  end

  # book = Book.new(title, ISBN.get(isbns.first))

  unless info_headers.empty?
    info_headers.each do |th|
      header_text = th.inner_text
      if header_text =~ /Type/
        book_data[:binding] = data_for_header(th)
      elsif header_text =~ /Verschijningsdatum/
        date = data_for_header(th)
        date =~ /\/([\d]{4})/
        book_data[:publish_year] = Regexp.last_match[1].to_i
      elsif header_text =~ /Auteur/
        book_data[:authors] << data_for_header(th)
      elsif header_text =~ /Uitgever/
        book_data[:publisher] = data_for_header(th)
      end
    end
  end

  image_url = nil
  if (cover_img = doc.at("img[@id$='imgProduct']"))
    image_url = if cover_img['src'] =~ /^http/
                  cover_img['src']
                else
                  "#{SITE}/#{cover_img['src']}" # TODO: use html <base>
                end
    image_url = nil if image_url =~ /ProductNoCover/
  end

  book = Book.new(book_data[:title], book_data[:authors],
                  book_data[:isbn], book_data[:publisher],
                  book_data[:publish_year], book_data[:binding])
  [book, image_url]
end

#parse_search_result_data(html) ⇒ Object 



103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
 
# File 'lib/alexandria/book_providers/proxis.rb', line 103

def parse_search_result_data(html)
  doc = html_to_doc(html)
  book_search_results = []
  items = doc.search('table.searchResult tr')
  items.each do |item|
    result = {}
    title_link = item % 'h5 a'
    if title_link
      result[:title] = text_of(title_link)
      result[:lookup_url] = title_link['href']
      result[:lookup_url] = "#{SITE}#{result[:lookup_url]}" unless result[:lookup_url] =~ /^http/
    end
    book_search_results << result
  end
  # require 'pp'
  # pp book_search_results
  # raise :Ruckus
  book_search_results
end

#search(criterion, type) ⇒ Object

Raises:



52
53
54
55
56
57
58
59
60
61
62
63
64
 
# File 'lib/alexandria/book_providers/proxis.rb', line 52

def search(criterion, type)
  req = create_search_uri(type, criterion)
  puts req if $DEBUG
  html_data = transport.get_response(URI.parse(req))

  results = parse_search_result_data(html_data.body)
  raise NoResultsError if results.empty?
  if type == SEARCH_BY_ISBN
    get_book_from_search_result(results.first)
  else
    results.map { |result| get_book_from_search_result(result) }
  end
end

#text_of(node) ⇒ Object

from Palatina



85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
 
# File 'lib/alexandria/book_providers/proxis.rb', line 85

def text_of(node)
  if node.nil?
    nil
  else
    if node.text?
      node.to_html
    elsif node.elem?
      if node.children.nil?
        nil
      else
        node_text = node.children.map { |n| text_of(n) }.join
        node_text.strip.squeeze(' ')
      end
    end
    # node.inner_html.strip
  end
end

#url(book) ⇒ Object 



80
81
82
 
# File 'lib/alexandria/book_providers/proxis.rb', line 80

def url(book)
  ISBN_REDIRECT_BASE_URL % Library.canonicalise_ean(book.isbn) if book.isbn.nil? || book.isbn.empty?
end