Merge branch 'develop'

Chris Park · Chris Park · commit 4672096433e5 · 2017-06-14T12:40:18.000Z
diff --git a/examples/categories.rb b/examples/categories.rb
@@ -8,11 +8,11 @@
   rosette_api = RosetteAPI.new(api_key, url)
 end
 
-categories_url_data = "http://www.onlocationvacations.com/2015/03/05/the-new-ghostbusters-movie-begins-filming-in-boston-in-june/"
+categories_url_data = 'http://www.onlocationvacations.com/2015/03/05/the-new-ghostbusters-movie-begins-filming-in-boston-in-june/'
 begin
-    params = DocumentParameters.new(content_uri: categories_url_data)
-    response = rosette_api.get_categories(params)
-    puts JSON.pretty_generate(response)
+  params = DocumentParameters.new(content_uri: categories_url_data)
+  response = rosette_api.get_categories(params)
+  puts JSON.pretty_generate(response)
 rescue RosetteAPIError => rosette_api_error
-    printf("Rosette API Error (%s): %s", rosette_api_error.status_code, rosette_api_error.message)
+  printf('Rosette API Error (%s): %s', rosette_api_error.status_code, rosette_api_error.message)
 end
diff --git a/examples/name_deduplication.rb b/examples/name_deduplication.rb
@@ -0,0 +1,21 @@
+require 'rosette_api'
+
+api_key, url = ARGV
+
+if !url
+  rosette_api = RosetteAPI.new(api_key)
+else
+  rosette_api = RosetteAPI.new(api_key, url)
+end
+
+name_dedupe_data = 'John Smith,Johnathon Smith,Fred Jones'
+
+threshold = 0.75
+names = name_dedupe_data.split(',').map { |n| NameParameter.new(n) }
+begin
+  params = NameDeduplicationParameters.new(names, threshold)
+  response = rosette_api.get_name_deduplication(params)
+  puts JSON.pretty_generate(response)
+rescue RosetteAPIError => rosette_api_error
+  printf('Rosette API Error (%s): %s', rosette_api_error.status_code, rosette_api_error.message)
+end
diff --git a/examples/name_similarity.rb b/examples/name_similarity.rb
@@ -8,13 +8,13 @@
   rosette_api = RosetteAPI.new(api_key, url)
 end
 
-matched_name_data1 = "Michael Jackson"
-matched_name_data2 = "迈克尔·杰克逊"
+matched_name_data1 = 'Michael Jackson'
+matched_name_data2 = '迈克尔·杰克逊'
 begin
-    name1 = NameParameter.new(matched_name_data1, entity_type: 'PERSON', language:'eng')
-    params = NameSimilarityParameters.new(name1, matched_name_data2)
-    response = rosette_api.name_similarity(params)
-    puts JSON.pretty_generate(response)
+  name1 = NameParameter.new(matched_name_data1, entity_type: 'PERSON', language: 'eng')
+  params = NameSimilarityParameters.new(name1, matched_name_data2)
+  response = rosette_api.get_name_similarity(params)
+  puts JSON.pretty_generate(response)
 rescue RosetteAPIError => rosette_api_error
-    printf("Rosette API Error (%s): %s", rosette_api_error.status_code, rosette_api_error.message)
+  printf('Rosette API Error (%s): %s', rosette_api_error.status_code, rosette_api_error.message)
 end
diff --git a/examples/name_translation.rb b/examples/name_translation.rb
@@ -8,11 +8,11 @@
   rosette_api = RosetteAPI.new(api_key, url)
 end
 
-translated_name_data = "معمر محمد أبو منيار القذاف"
+translated_name_data = 'معمر محمد أبو منيار القذاف'
 begin
-    params = NameTranslationParameters.new(translated_name_data, 'eng', target_script: 'Latn')
-    response = rosette_api.name_translation(params)
-    puts JSON.pretty_generate(response)
+  params = NameTranslationParameters.new(translated_name_data, 'eng', target_script: 'Latn')
+  response = rosette_api.get_name_translation(params)
+  puts JSON.pretty_generate(response)
 rescue RosetteAPIError => rosette_api_error
-    printf("Rosette API Error (%s): %s", rosette_api_error.status_code, rosette_api_error.message)
+  printf('Rosette API Error (%s): %s', rosette_api_error.status_code, rosette_api_error.message)
 end
diff --git a/examples/ping.rb b/examples/ping.rb
@@ -8,8 +8,8 @@
   rosette_api = RosetteAPI.new(api_key, url)
 end
 begin
-    response = rosette_api.ping
-    puts JSON.pretty_generate(response)
+  response = rosette_api.ping
+  puts JSON.pretty_generate(response)
 rescue RosetteAPIError => rosette_api_error
-    printf("Rosette API Error (%s): %s", rosette_api_error.status_code, rosette_api_error.message)
+  printf('Rosette API Error (%s): %s', rosette_api_error.status_code, rosette_api_error.message)
 end
diff --git a/examples/relationships.rb b/examples/relationships.rb
@@ -10,10 +10,10 @@
 
 relationships_text_data = "Bill Gates, Microsoft's former CEO, is a philanthropist."
 begin
-    params = DocumentParameters.new(content: relationships_text_data)
-    params.rosette_options = { accuracyMode: 'PRECISION' }
-    response = rosette_api.get_relationships(params)
-    puts JSON.pretty_generate(response)
+  params = DocumentParameters.new(content: relationships_text_data)
+  params.rosette_options = { accuracyMode: 'PRECISION' }
+  response = rosette_api.get_relationships(params)
+  puts JSON.pretty_generate(response)
 rescue RosetteAPIError => rosette_api_error
-    printf("Rosette API Error (%s): %s", rosette_api_error.status_code, rosette_api_error.message)
+  printf('Rosette API Error (%s): %s', rosette_api_error.status_code, rosette_api_error.message)
 end
diff --git a/examples/sentences.rb b/examples/sentences.rb
@@ -11,10 +11,10 @@
 sentences_data = "This land is your land. This land is my land\nFrom California to the New York island;\nFrom the red wood forest to the Gulf Stream waters\n\nThis land was made for you and Me.\n\nAs I was walking that ribbon of highway,\nI saw above me that endless skyway:\nI saw below me that golden valley:\nThis land was made for you and me."
 
 begin
-    params = DocumentParameters.new
-    params.content = sentences_data
-    response = rosette_api.get_sentences(params)
-    puts JSON.pretty_generate(response)
+  params = DocumentParameters.new
+  params.content = sentences_data
+  response = rosette_api.get_sentences(params)
+  puts JSON.pretty_generate(response)
 rescue RosetteAPIError => rosette_api_error
-    printf("Rosette API Error (%s): %s", rosette_api_error.status_code, rosette_api_error.message)
+  printf('Rosette API Error (%s): %s', rosette_api_error.status_code, rosette_api_error.message)
 end
diff --git a/examples/sentiment.rb b/examples/sentiment.rb
@@ -9,14 +9,14 @@
   rosette_api = RosetteAPI.new(api_key, url)
 end
 
-file = Tempfile.new(%w(foo .html))
+file = Tempfile.new(%w[foo .html])
 sentiment_file_data = '<html><head><title>New Ghostbusters Film</title></head><body><p>Original Ghostbuster Dan Aykroyd, who also co-wrote the 1984 Ghostbusters film, couldn’t be more pleased with the new all-female Ghostbusters cast, telling The Hollywood Reporter, “The Aykroyd family is delighted by this inheritance of the Ghostbusters torch by these most magnificent women in comedy.”</p></body></html>'
 file.write(sentiment_file_data)
 file.close
 begin
-    params = DocumentParameters.new(file_path: file.path, language: 'eng')
-    response = rosette_api.get_sentiment(params)
-    puts JSON.pretty_generate(response)
+  params = DocumentParameters.new(file_path: file.path, language: 'eng')
+  response = rosette_api.get_sentiment(params)
+  puts JSON.pretty_generate(response)
 rescue RosetteAPIError => rosette_api_error
-    printf("Rosette API Error (%s): %s", rosette_api_error.status_code, rosette_api_error.message)
+  printf('Rosette API Error (%s): %s', rosette_api_error.status_code, rosette_api_error.message)
 end
diff --git a/examples/syntax_dependencies.rb b/examples/syntax_dependencies.rb
@@ -8,11 +8,11 @@
   rosette_api = RosetteAPI.new(api_key, url)
 end
 
-syntax_dependencies_data = "Yoshinori Ohsumi, a Japanese cell biologist, was awarded the Nobel Prize in Physiology or Medicine on Monday."
+syntax_dependencies_data = 'Yoshinori Ohsumi, a Japanese cell biologist, was awarded the Nobel Prize in Physiology or Medicine on Monday.'
 begin
-    params = DocumentParameters.new(content: syntax_dependencies_data, genre: 'social-media')
-    response = rosette_api.get_syntax_dependencies(params)
-    puts JSON.pretty_generate(response)
+  params = DocumentParameters.new(content: syntax_dependencies_data)
+  response = rosette_api.get_syntax_dependencies(params)
+  puts JSON.pretty_generate(response)
 rescue RosetteAPIError => rosette_api_error
-    printf("Rosette API Error (%s): %s", rosette_api_error.status_code, rosette_api_error.message)
+  printf('Rosette API Error (%s): %s', rosette_api_error.status_code, rosette_api_error.message)
 end
diff --git a/examples/text_embedding.rb b/examples/text_embedding.rb
@@ -10,9 +10,9 @@
 
 embeddings_data = 'Cambridge, Massachusetts'
 begin
-    params = DocumentParameters.new(content: embeddings_data)
-    response = rosette_api.get_text_embedding(params)
-    puts JSON.pretty_generate(response)
+  params = DocumentParameters.new(content: embeddings_data)
+  response = rosette_api.get_text_embedding(params)
+  puts JSON.pretty_generate(response)
 rescue RosetteAPIError => rosette_api_error
-    printf("Rosette API Error (%s): %s", rosette_api_error.status_code, rosette_api_error.message)
+  printf('Rosette API Error (%s): %s', rosette_api_error.status_code, rosette_api_error.message)
 end
diff --git a/examples/tokens.rb b/examples/tokens.rb
@@ -8,11 +8,11 @@
   rosette_api = RosetteAPI.new(api_key, url)
 end
 
-tokens_data = "北京大学生物系主任办公室内部会议"
+tokens_data = '北京大学生物系主任办公室内部会议'
 begin
-    params = DocumentParameters.new(content: tokens_data)
-    response = rosette_api.get_tokens(params)
-    puts JSON.pretty_generate(response)
+  params = DocumentParameters.new(content: tokens_data)
+  response = rosette_api.get_tokens(params)
+  puts JSON.pretty_generate(response)
 rescue RosetteAPIError => rosette_api_error
-    printf("Rosette API Error (%s): %s", rosette_api_error.status_code, rosette_api_error.message)
+  printf('Rosette API Error (%s): %s', rosette_api_error.status_code, rosette_api_error.message)
 end
diff --git a/examples/transliteration.rb b/examples/transliteration.rb
@@ -0,0 +1,20 @@
+require 'rosette_api'
+
+api_key, url = ARGV
+
+if !url
+  rosette_api = RosetteAPI.new(api_key)
+else
+  rosette_api = RosetteAPI.new(api_key, url)
+end
+
+transliteration_content_data = 'Kareem Abdul Jabbar holds the records for most points in the NBA'
+
+begin
+  params = DocumentParameters.new
+  params.content = transliteration_content_data
+  response = rosette_api.get_transliteration(params)
+  puts JSON.pretty_generate(response)
+rescue RosetteAPIError => rosette_api_error
+  printf('Rosette API Error (%s): %s', rosette_api_error.status_code, rosette_api_error.message)
+end
diff --git a/lib/document_parameters.rb b/lib/document_parameters.rb
@@ -47,7 +47,7 @@ def validate_params
     elsif [@content, @content_uri, @file_path].all?(&:nil?)
       raise BadRequestFormatError.new 'The format of the request is invalid: no content provided; must' \
                                       ' be one of an attachment, an inline "content" field, or an external "contentUri"'
-    elsif !@rosette_options.nil?
+    elsif @rosette_options
       raise BadRequestError.new('rosette_options can only be an instance of a Hash') unless @rosette_options.is_a? Hash
     end
   end
@@ -57,7 +57,7 @@ def validate_params
   # Returns the new Hash.
   def load_params
     validate_params
-    to_hash.select { |_key, value| !value.nil? }
+    to_hash.select { |_key, value| value }
            .map { |key, value| [key.to_s.split('_').map(&:capitalize).join.sub!(/\D/, &:downcase), value] }
            .to_h
   end
diff --git a/lib/name_deduplication_parameters.rb b/lib/name_deduplication_parameters.rb
@@ -0,0 +1,56 @@
+require_relative 'bad_request_error'
+require_relative 'name_parameter'
+
+# This class encapsulates parameters that are needed for name-deduplication in
+# Rosette API.
+class NameDeduplicationParameters
+  # Rosette API options (optional, should be a hash)
+  attr_accessor :rosette_options
+  # List of Name objects to be de-duplicated
+  attr_accessor :names
+  # Threshold for determining cluster size
+  attr_accessor :threshold
+
+  def initialize(names, threshold, options = {}) #:notnew:
+    options = {
+      rosette_options: nil
+    }.update options
+    @names = names
+    @threshold = threshold
+    @rosette_options = options[:rosette_options]
+  end
+
+  # Validates the parameters by checking if name1 and name2 are instances of
+  # a String or NameParameter.
+  def validate_params
+    raise BadRequestError.new('names must be an array of name_parameter') unless @names.instance_of? Array
+    if @threshold
+      raise BadRequestError.new('threshold must be a float') unless @threshold.is_a?(Float)
+      raise BadRequestError.new('threshold must be in the range of 0 to 1') if @threshold.negative? || @threshold > 1
+    end
+    if @rosette_options
+      raise BadRequestError.new('rosette_options can only be an instance of a Hash') unless @rosette_options.is_a? Hash
+    end
+  end
+
+  # Converts this class to Hash with its keys in lower CamelCase.
+  #
+  # Returns the new Hash.
+  def load_params
+    validate_params
+    to_hash.select { |_key, value| value }
+           .map { |key, value| [key.to_s.split('_').map(&:capitalize).join.sub!(/\D/, &:downcase), value] }
+           .to_h
+  end
+
+  # Converts this class to Hash.
+  #
+  # Returns the new Hash.
+  def to_hash
+    {
+      names: @names.map(&:load_param),
+      threshold: @threshold,
+      options: @rosette_options
+    }
+  end
+end
diff --git a/lib/name_parameter.rb b/lib/name_parameter.rb
@@ -25,9 +25,9 @@ def initialize(text, options = {}) #:notnew:
   #
   # Returns the new Hash.
   def load_param
-    self.to_hash.select { |_key, value| !value.nil? }
-        .map { |key, value| [key.to_s.split('_').map(&:capitalize).join.sub!(/\D/, &:downcase), value] }
-        .to_h
+    to_hash.select { |_key, value| value }
+           .map { |key, value| [key.to_s.split('_').map(&:capitalize).join.sub!(/\D/, &:downcase), value] }
+           .to_h
   end
 
   # Converts this class to Hash.
diff --git a/lib/name_similarity_parameters.rb b/lib/name_similarity_parameters.rb
@@ -27,11 +27,9 @@ def initialize(name1, name2, options = {}) #:notnew:
   # Validates the parameters by checking if name1 and name2 are instances of
   # a String or NameParameter.
   def validate_params
-    if [String, NameParameter].none? { |clazz| @name1.is_a? clazz }
-      raise BadRequestError.new('name1 option can only be an instance of a String or NameParameter')
-    elsif [String, NameParameter].none? { |clazz| @name2.is_a? clazz }
-      raise BadRequestError.new('name2 option can only be an instance of a String or NameParameter')
-    elsif !@rosette_options.nil?
+    raise BadRequestError.new('name1 option can only be an instance of a String or NameParameter') if [String, NameParameter].none? { |clazz| @name1.is_a? clazz }
+    raise BadRequestError.new('name2 option can only be an instance of a String or NameParameter') if [String, NameParameter].none? { |clazz| @name2.is_a? clazz }
+    if @rosette_options
       raise BadRequestError.new('rosette_options can only be an instance of a Hash') unless @rosette_options.is_a? Hash
     end
   end
@@ -40,10 +38,10 @@ def validate_params
   #
   # Returns the new Hash.
   def load_params
-    self.validate_params
-    self.to_hash.select { |_key, value| !value.nil? }
-        .map { |key, value| [key.to_s.split('_').map(&:capitalize).join.sub!(/\D/, &:downcase), value] }
-        .to_h
+    validate_params
+    to_hash.reject { |_key, value| value.nil? }
+           .map { |key, value| [key.to_s.split('_').map(&:capitalize).join.sub!(/\D/, &:downcase), value] }
+           .to_h
   end
 
   # Converts this class to Hash.
diff --git a/lib/name_translation_parameters.rb b/lib/name_translation_parameters.rb
@@ -49,7 +49,7 @@ def initialize(name, target_language, options = {}) #:notnew:
 
   # Validates the parameters by checking if rosette_options is an instance of a Hash.
   def validate_params
-    if !@rosette_options.nil?
+    if @rosette_options
       raise BadRequestError.new('rosette_options can only be an instance of a Hash') unless @rosette_options.is_a? Hash
     end
   end
@@ -58,10 +58,10 @@ def validate_params
   #
   # Returns the new Hash.
   def load_params
-    self.validate_params
-    self.to_hash.select { |_key, value| !value.nil? }
-        .map { |key, value| [key.to_s.split('_').map(&:capitalize).join.sub!(/\D/, &:downcase), value] }
-        .to_h
+    validate_params
+    to_hash.select { |_key, value| value }
+           .map { |key, value| [key.to_s.split('_').map(&:capitalize).join.sub!(/\D/, &:downcase), value] }
+           .to_h
   end
 
   # Converts this class to Hash.
diff --git a/lib/request_builder.rb b/lib/request_builder.rb
diff --git a/lib/rosette_api.rb b/lib/rosette_api.rb
diff --git a/rosette_api.gemspec b/rosette_api.gemspec
diff --git a/tests/tests_spec.rb b/tests/tests_spec.rb
