Class: Awis::Models::UrlInfo

Inherits:
Base
  • Object
show all
Defined in:
lib/awis/models/url_info.rb

Instance Attribute Summary collapse

Attributes inherited from Base

#request_id, #response, #status_code

Instance Method Summary collapse

Methods inherited from Base

#action_name, #loading_response, #pretty_xml, #relationship_collections, #root_node_name, #success?

Constructor Details

#initialize(response) ⇒ UrlInfo

Returns a new instance of UrlInfo.



9
10
11
12
13
14
15
16
17
# File 'lib/awis/models/url_info.rb', line 9

def initialize(response)
  @usage_statistics = []
  @related_links = []
  @categories = []
  @contributing_subdomains = []
  @rank_by_country = []

  super(response)
end

Instance Attribute Details

#asinObject

Returns the value of attribute asin.



6
7
8
# File 'lib/awis/models/url_info.rb', line 6

def asin
  @asin
end

#categoriesObject

Returns the value of attribute categories.



6
7
8
# File 'lib/awis/models/url_info.rb', line 6

def categories
  @categories
end

#contact_infoObject

Returns the value of attribute contact_info.



6
7
8
# File 'lib/awis/models/url_info.rb', line 6

def contact_info
  @contact_info
end

#content_dataObject

Returns the value of attribute content_data.



6
7
8
# File 'lib/awis/models/url_info.rb', line 6

def content_data
  @content_data
end

#contributing_subdomainsObject

Returns the value of attribute contributing_subdomains.



6
7
8
# File 'lib/awis/models/url_info.rb', line 6

def contributing_subdomains
  @contributing_subdomains
end

#data_urlObject

Returns the value of attribute data_url.



6
7
8
# File 'lib/awis/models/url_info.rb', line 6

def data_url
  @data_url
end

#rankObject

Returns the value of attribute rank.



6
7
8
# File 'lib/awis/models/url_info.rb', line 6

def rank
  @rank
end

#rank_by_countryObject

Returns the value of attribute rank_by_country.



6
7
8
# File 'lib/awis/models/url_info.rb', line 6

def rank_by_country
  @rank_by_country
end

Returns the value of attribute related_links.



6
7
8
# File 'lib/awis/models/url_info.rb', line 6

def related_links
  @related_links
end

#usage_statisticsObject

Returns the value of attribute usage_statistics.



6
7
8
# File 'lib/awis/models/url_info.rb', line 6

def usage_statistics
  @usage_statistics
end

Instance Method Details

#alexa_gdnObject



285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
# File 'lib/awis/models/url_info.rb', line 285

def alexa_gdn
  {
    250000 => '< 1K',
    100000 => '1K - 10K',
    50000 => '10K - 100K',
    20000 => '100K - 500K',
    10000 => '500K - 1M',
    5000 => '1M - 2M',
    2000 => '2M - 5M',
    1000 => '5M - 10M',
    500 => '10M - 20M',
    150 => '20M - 50M',
    30 => '50M - 100M',
    0 => '100M+'
  }
end

#alexa_speed_ratingObject



261
262
263
264
265
266
267
# File 'lib/awis/models/url_info.rb', line 261

def alexa_speed_rating
  {
    2200 => 'POOR ( > 5s)',
    1200 => 'AVERAGE ( < 5s)',
    0 => 'GOOD ( < 3s)'
  }
end

#categories_node_nameObject



176
177
178
# File 'lib/awis/models/url_info.rb', line 176

def categories_node_name
  "#{related_node_name}/aws:Categories/aws:CategoryData"
end

#content_node_nameObject



164
165
166
# File 'lib/awis/models/url_info.rb', line 164

def content_node_name
  "#{root_node_name}/aws:ContentData"
end

#contributing_hostnamesObject



219
220
221
# File 'lib/awis/models/url_info.rb', line 219

def contributing_hostnames
  contributing_subdomains.map(&:data_url).reject { |hostname| hostname == 'OTHER' }
end

#contributing_subdomains_node_nameObject



188
189
190
# File 'lib/awis/models/url_info.rb', line 188

def contributing_subdomains_node_name
  "#{traffic_node_name}/aws:ContributingSubdomains/aws:ContributingSubdomain"
end

#daily_GDN_page_viewsObject



243
244
245
246
247
248
249
# File 'lib/awis/models/url_info.rb', line 243

def daily_GDN_page_views
  return unless rank

  alexa_gdn.each do |max_pvs, gdn_range|
    return gdn_range if rank > max_pvs
  end
end

#domainsObject



215
216
217
# File 'lib/awis/models/url_info.rb', line 215

def domains
  content_data.owned_domains.map(&:domain)
end

#geos_hashObject



206
207
208
# File 'lib/awis/models/url_info.rb', line 206

def geos_hash
  @geos_hash ||= geos_sorted.reduce({}, :merge)
end

#geos_sortedObject



200
201
202
203
204
# File 'lib/awis/models/url_info.rb', line 200

def geos_sorted
  rank_by_country.select { |rbc| !rbc.rank.nil? && !rbc.contribution_page_views.nil? }.
    sort_by { |rbc| - rbc.contribution_page_views.round }.
    map { |rbc| { rbc.country_code => rbc.contribution_page_views.round } }
end

#get_median_load_timeObject



239
240
241
# File 'lib/awis/models/url_info.rb', line 239

def get_median_load_time
  content_data.speed_median_load_time
end

#get_pvs_per_milObject



231
232
233
# File 'lib/awis/models/url_info.rb', line 231

def get_pvs_per_mil
  usage_statistics.first.page_views_per_million_value
end

#has_data?Boolean

Returns:

  • (Boolean)


196
197
198
# File 'lib/awis/models/url_info.rb', line 196

def has_data?
  !@rank.nil?
end

#init_entity_data(attr_name, data, kclass) ⇒ Object



158
159
160
161
162
# File 'lib/awis/models/url_info.rb', line 158

def init_entity_data(attr_name, data, kclass)
  return if data.empty?

  instance_variable_set("@#{attr_name}", kclass.new(data))
end

#not_found?Boolean Also known as: is_404?

Returns:

  • (Boolean)


210
211
212
# File 'lib/awis/models/url_info.rb', line 210

def not_found?
  ([content_data.data_url, content_data.site_title] & [404, '404']).size > 0
end

#pvs_per_userObject



223
224
225
# File 'lib/awis/models/url_info.rb', line 223

def pvs_per_user
  usage_statistics.first.page_views_per_user_value
end

#pvs_rankObject



227
228
229
# File 'lib/awis/models/url_info.rb', line 227

def pvs_rank
  usage_statistics.first.page_views_rank_value
end

#rank_by_country_node_nameObject



192
193
194
# File 'lib/awis/models/url_info.rb', line 192

def rank_by_country_node_name
  "#{traffic_node_name}/aws:RankByCountry/aws:Country"
end

#rank_page_viewObject



269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
# File 'lib/awis/models/url_info.rb', line 269

def rank_page_view
  {
    20000 => '< 1K',
    5000 => '1K - 10K',
    3000 => '10K - 100K',
    1900 => '100K - 500K',
    1300 => '500K - 1M',
    850 => '2M - 5M',
    550 => '5M - 10M',
    350 => '10M - 20M',
    200 => '20M - 50M',
    100 => '50M - 100M',
    28 => '100M+'
  }
end


172
173
174
# File 'lib/awis/models/url_info.rb', line 172

def related_links_node_name
  "#{related_node_name}/aws:RelatedLinks/aws:RelatedLink"
end


168
169
170
# File 'lib/awis/models/url_info.rb', line 168

def related_node_name
  "#{root_node_name}/aws:Related"
end

#setup_data!(response) ⇒ Object



19
20
21
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
53
54
55
56
57
58
59
60
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
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
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
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
# File 'lib/awis/models/url_info.rb', line 19

def setup_data!(response)
  content_data = {
    owned_domains: []
  }
  contact_info = {
    phone_numbers: []
  }
  statistics = []
  related_related_links = []
  category_data = []
  rank_by_country = []
  contributing_subdomains = []

  response.each_node do |node, path|
    text = node.inner_xml
    candidate_text = text.delete(',')
    value = if integer_value?(candidate_text) && node.name != 'aws:Delta'
              candidate_text.to_i
            elsif float_value?(candidate_text)
              candidate_text.to_f
            elsif text.is_a?(String) && text.empty?
              nil
            else
              text
            end

    if node.name == 'aws:RequestId'
      @request_id ||= value
    elsif node.name == 'aws:StatusCode'
      @status_code ||= value
    elsif node.name == 'aws:DataUrl' && path == "#{traffic_node_name}/aws:DataUrl"
      @data_url = value
    elsif node.name == 'aws:Asin' && path == "#{traffic_node_name}/aws:Asin"
      @asin = value
    elsif node.name == 'aws:Rank' && path == "#{traffic_node_name}/aws:Rank"
      @rank = value
    elsif node.name == 'aws:DataUrl' && path == "#{content_node_name}/aws:DataUrl"
      content_data[:data_url] = value
    elsif node.name == 'aws:Title' && path == "#{content_node_name}/aws:SiteData/aws:Title"
      content_data[:site_title] = value
    elsif node.name == 'aws:Description'
      content_data[:site_description] = value
    elsif node.name == 'aws:MedianLoadTime'
      content_data[:speed_median_load_time] = value
    elsif node.name == 'aws:Percentile'
      content_data[:speed_percentile] = value
    elsif node.name == 'aws:AdultContent'
      content_data[:adult_content] = value
    elsif node.name == 'aws:Locale'
      content_data[:language_locale] = value
    elsif node.name == 'aws:LinksInCount'
      content_data[:links_in_count] = value
    elsif node.name == 'aws:Domain' && path == "#{content_node_name}/aws:OwnedDomains/aws:OwnedDomain/aws:Domain"
      content_data[:owned_domains] << { domain: value }
    elsif node.name == 'aws:Title' && path == "#{content_node_name}/aws:OwnedDomains/aws:OwnedDomain/aws:Title"
      content_data[:owned_domains] << { title: value }
    elsif node.name == 'aws:OnlineSince'
      content_data[:online_since] = value
    elsif node.name == 'aws:DataUrl' && path == "#{root_node_name}/aws:ContactInfo/aws:DataUrl"
      contact_info[:data_url] = value
    elsif node.name == 'aws:OwnerName'
      contact_info[:owner_name] = value
    elsif node.name == 'aws:Email'
      contact_info[:email] = value
    elsif node.name == 'aws:PhysicalAddress'
      contact_info[:physical_address] = value
    elsif node.name == 'aws:CompanyStockTicker'
      contact_info[:company_stock_ticker] = value
    elsif node.name == 'aws:PhoneNumber'
      contact_info[:phone_numbers] << value
    elsif node.name == 'aws:DataUrl' && path == "#{related_links_node_name}/aws:DataUrl"
      related_related_links << { data_url: value }
    elsif node.name == 'aws:NavigableUrl' && path == "#{related_links_node_name}/aws:NavigableUrl"
      related_related_links << { navigable_url: value }
    elsif node.name == 'aws:Title' && path == "#{related_links_node_name}/aws:Title"
      related_related_links << { title: value }
    elsif node.name == 'aws:Title' && path == "#{categories_node_name}/aws:Title"
      category_data << { title: value }
    elsif node.name == 'aws:AbsolutePath' &&  path == "#{categories_node_name}/aws:AbsolutePath"
      category_data << { absolute_path: value }
    elsif node.name == 'aws:Months' && path == "#{statistic_node_name}/aws:TimeRange/aws:Months"
      statistics << { time_range_months: value }
    elsif node.name == 'aws:Days' && path == "#{statistic_node_name}/aws:TimeRange/aws:Days"
      statistics << { time_range_days: value }
    elsif node.name == 'aws:Value' && path == "#{statistic_node_name}/aws:Rank/aws:Value"
      statistics << { rank_value: value }
    elsif node.name == 'aws:Delta' && path == "#{statistic_node_name}/aws:Rank/aws:Delta"
      statistics << { rank_delta: value }
    elsif node.name == 'aws:Value' && path == "#{statistic_node_name}/aws:Reach/aws:Rank/aws:Value"
      statistics << { reach_rank_value: value }
    elsif node.name == 'aws:Delta' && path == "#{statistic_node_name}/aws:Reach/aws:Rank/aws:Delta"
      statistics << { reach_rank_delta: value }
    elsif node.name == 'aws:Value' && path == "#{statistic_node_name}/aws:Reach/aws:PerMillion/aws:Value"
      statistics << { reach_per_million_value: value }
    elsif node.name == 'aws:Delta' && path == "#{statistic_node_name}/aws:Reach/aws:PerMillion/aws:Delta"
      statistics << { reach_per_million_delta: value }
    elsif node.name == 'aws:Value' && path == "#{statistic_node_name}/aws:PageViews/aws:PerMillion/aws:Value"
      statistics << { page_views_per_million_value: value }
    elsif node.name == 'aws:Delta' && path == "#{statistic_node_name}/aws:PageViews/aws:PerMillion/aws:Delta"
      statistics << { page_views_per_million_delta: value }
    elsif node.name == 'aws:Value' && path == "#{statistic_node_name}/aws:PageViews/aws:Rank/aws:Value"
      statistics << { page_views_rank_value: value }
    elsif node.name == 'aws:Delta' && path == "#{statistic_node_name}/aws:PageViews/aws:Rank/aws:Delta"
      statistics << { page_views_rank_delta: value }
    elsif node.name == 'aws:Value' && path == "#{statistic_node_name}/aws:PageViews/aws:PerUser/aws:Value"
      statistics << { page_views_per_user_value: value }
    elsif node.name == 'aws:Delta' && path == "#{statistic_node_name}/aws:PageViews/aws:PerUser/aws:Delta"
      statistics << { page_views_per_user_delta: value }
    elsif node.name == 'aws:Country' && path == rank_by_country_node_name
      rank_by_country << { country_code: node.attributes['Code'] }
    elsif node.name == 'aws:Rank' && path == "#{rank_by_country_node_name}/aws:Rank"
      rank_by_country << { rank: value }
    elsif node.name == 'aws:PageViews' && path == "#{rank_by_country_node_name}/aws:Contribution/aws:PageViews"
      rank_by_country << { contribution_page_views: value }
    elsif node.name == 'aws:Users' && path == "#{rank_by_country_node_name}/aws:Contribution/aws:Users"
      rank_by_country << { contribution_users: value }
    elsif node.name == 'aws:DataUrl' && path == "#{contributing_subdomains_node_name}/aws:DataUrl"
      contributing_subdomains << { data_url: value }
    elsif node.name == 'aws:Months' && path == "#{contributing_subdomains_node_name}/aws:TimeRange/aws:Months"
      contributing_subdomains << { time_range_months: value }
    elsif node.name == 'aws:Percentage' && path == "#{contributing_subdomains_node_name}/aws:Reach/aws:Percentage"
      contributing_subdomains << { reach_percentage: value }
    elsif node.name == 'aws:Percentage' && path == "#{contributing_subdomains_node_name}/aws:PageViews/aws:Percentage"
      contributing_subdomains << { page_views_percentage: value }
    elsif node.name == 'aws:PerUser' && path == "#{contributing_subdomains_node_name}/aws:PageViews/aws:PerUser"
      contributing_subdomains << { page_views_per_user: value }
    end
  end

  init_entity_data('content_data', content_data, ContentData)
  init_entity_data('contact_info', contact_info, ContactInfo)

  relationship_collections(@usage_statistics, statistics, 13, UsageStatistic)
  relationship_collections(@related_links, related_related_links, 3, RelatedLink)
  relationship_collections(@categories, category_data, 3, CategoryData)
  relationship_collections(@rank_by_country, rank_by_country, 4, RankByCountry)
  relationship_collections(@contributing_subdomains, contributing_subdomains, 5, ContributingSubdomain)
end

#speed_percentileObject



235
236
237
# File 'lib/awis/models/url_info.rb', line 235

def speed_percentile
  content_data.speed_percentile
end

#speed_ratingObject



251
252
253
254
255
256
257
258
259
# File 'lib/awis/models/url_info.rb', line 251

def speed_rating
  if get_median_load_time
    alexa_speed_rating.each do |max_load_time, rating|
      return rating if get_median_load_time > max_load_time
    end
  end

  'AVERAGE ( < 5s)'
end

#statistic_node_nameObject



184
185
186
# File 'lib/awis/models/url_info.rb', line 184

def statistic_node_name
  "#{traffic_node_name}/aws:UsageStatistics/aws:UsageStatistic"
end

#traffic_node_nameObject



180
181
182
# File 'lib/awis/models/url_info.rb', line 180

def traffic_node_name
  "#{root_node_name}/aws:TrafficData"
end