vader_sentiment_ruby 0.1.0 → 0.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 57fb7693c238e33224556fb6a2d7d8c479ed88d76cd6576edef5c8befe4ad144
-  data.tar.gz: 8fc4484045923da3ec6986b8dbbf39f41437c635a25dbfdd14c0973baebe271b
+  metadata.gz: 6b529805b4d4db7386700ff49cd122451590c19100e1c368925b3fa221a81c1a
+  data.tar.gz: 16dbb6928fa25b59b86acfc8a6272114bc751fff8ae2b02545d4af7539b6d0a6
 SHA512:
-  metadata.gz: 127836bbe570da1cd60082da181d22e4a3619d32300d8e16dfc9aa9a3fcbbc0b3d51ad0e1b4c8e1143879f4b51f6e70a047590468af7c7de260468ddb7987ab9
-  data.tar.gz: 94d770a25884484554c2dce05607c68bda7742482d883d18c2cf693c979a31085189fec042f78b7303e6b13e0b75f13c0c09ee7e39f5d756787cde27db0a6102
+  metadata.gz: 0a6e0dd37f24f9c78ca95756c3a304b534e0c9a816c3a5b14b1f40142e59c868bc7965488904f8a041bfebbfab3f8451d32c59214c51abf3927e5df4676ddae9
+  data.tar.gz: 382dfe51190b6efbaa7b517ea454db85d41c0cbac784a887ce325815c8984d11604b13bb93e0b856ef41cbed865442257f4c85ee59ee2422e596c8167917e3ea

data/lib/vader_sentiment_ruby/emojis_describer.rb

@@ -1,21 +1,24 @@
 # frozen_string_literal: true
 
 module VaderSentimentRuby
-  # Replaces emoji char with its description
+  # Replaces emoji characters with their descriptions
   class EmojiDescriber
-    def initialize(text, emojis)
-      @text = text
-      @emojis = emojis
+    # @param [String] text Original text
+    # @param [Hash] emoji_dictionary Emoji dictionary with emojis as keys
+    def initialize(text, emoji_dictionary)
+      @text_array = text.split('')
+      @emoji_dictionary = emoji_dictionary
       @text_no_emoji = ''
       @prev_space = true
     end
 
+    # @return [String] Text with emojis replaced with descriptions
     def call
-      @text.split('').each do |chr|
-        if @emojis.keys.include?(chr)
-          handle_emoji_presence(chr)
+      @text_array.each do |character|
+        if @emoji_dictionary.keys.include?(character)
+          replace_emoji_with_description(character)
         else
-          handle_emoji_absence(chr)
+          handle_simple_character(character)
         end
       end
 
@@ -24,14 +27,13 @@ module VaderSentimentRuby
 
     private
 
-    def handle_emoji_presence(emoji)
-      description = @emojis[emoji]
+    def replace_emoji_with_description(emoji)
       @text_no_emoji += ' ' unless @prev_space
-      @text_no_emoji += description
+      @text_no_emoji += @emoji_dictionary[emoji]
       @prev_space = false
     end
 
-    def handle_emoji_absence(character)
+    def handle_simple_character(character)
       @text_no_emoji += character
       @prev_space = character == ' '
     end

data/lib/vader_sentiment_ruby/punctuation_emphasis_amplifier.rb

@@ -1,36 +1,42 @@
 # frozen_string_literal: true
 
 module VaderSentimentRuby
-  # Adds emphasis from exclamation points and question marks
+  # Adds emphasis factor from exclamation points and question marks
   class PunctuationEmphasisAmplifier
+    EXCLAMATION_MARK = '!'
+    QUESTION_MARK = '?'
+    # Empirically derived mean sentiment intensity rating increases for exclamation points and question marks
+    EXCLAMATION_MARK_RATING_INCREASE = 0.292
+    QUESTION_MARK_RATING_INCREASE = 0.18
+
     # @param [String] text
     def initialize(text)
-      @text = text
+      @text_array = text.split('')
     end
 
-    # @return [Float]
+    # @return [Float, Integer] Emphasis factor
     def call
-      amplify_exclamation_points + amplify_question_marks
+      (amplify_exclamation_points + amplify_question_marks).round(3)
     end
 
+    private
+
     def amplify_exclamation_points
-      # check for added emphasis resulting from exclamation points (up to 4 of them)
-      ep_count = @text.split('').count('!')
-      ep_count = 4.0 if ep_count > 4
+      # Check for added emphasis resulting from exclamation points (up to 4 of them)
+      ep_count = @text_array.count(EXCLAMATION_MARK)
+      ep_count = 4 if ep_count > 4
 
-      # empirically derived mean sentiment intensity rating increase for exclamation points
-      ep_count * 0.292
+      ep_count * EXCLAMATION_MARK_RATING_INCREASE
     end
 
     def amplify_question_marks
-      # check for added emphasis resulting from question marks (2 or 3+)
-      qm_count = @text.split('').count('?')
+      # Check for added emphasis resulting from question marks (2 or 3+)
+      qm_count = @text_array.count(QUESTION_MARK)
 
-      return 0.0 unless qm_count > 1
-      # empirically derived mean sentiment intensity rating increase for question marks
-      return qm_count * 0.18 if qm_count <= 3
+      return 0 unless qm_count > 1
+      return 0.96 if qm_count > 3
 
-      0.96
+      qm_count * QUESTION_MARK_RATING_INCREASE
     end
   end
 end

data/lib/vader_sentiment_ruby/sentiment_properties_identifier.rb

@@ -3,7 +3,7 @@
 module VaderSentimentRuby
   # Identify sentiment-relevant string-level properties of input text.
   class SentimentPropertiesIdentifier
-    attr_reader :is_cap_diff, :words_and_emoticons
+    attr_reader :words_and_emoticons, :is_cap_diff
 
     # @param [String] text
     def initialize(text)
@@ -11,14 +11,13 @@ module VaderSentimentRuby
       @text = text
       @words_and_emoticons = prepare_words_and_emoticons
       # Doesn't separate words from adjacent punctuation (keeps emoticons & contractions)
-      @is_cap_diff = all_cap_differential?(@words_and_emoticons)
+      @is_cap_diff = text_contains_mixed_cases?
     end
 
     private
 
     # Removes leading and trailing punctuation
     # Leaves contractions and most emoticons
-    # Does not preserve punc-plus-letter emoticons (e.g. :D)
     # @return [Array]
     def prepare_words_and_emoticons
       @text
@@ -28,21 +27,11 @@ module VaderSentimentRuby
 
     # Check whether just some words in the input are ALL CAPS.
     # Returns `True` if some but not all items in `words` are ALL CAPS
-    # @param [Array] words
     # @return [Boolean]
-    def all_cap_differential?(words)
-      all_cap_words = 0
+    def text_contains_mixed_cases?
+      uppercase_words = @words_and_emoticons.count { |word| WordHelper.word_upcase?(word) }
 
-      words.each do |word|
-        all_cap_words += 1 if WordHelper.word_upcase?(word)
-      end
-
-      words_size = words.size
-      cap_differential = words_size - all_cap_words
-
-      return true if cap_differential.positive? && cap_differential < words_size
-
-      false
+      uppercase_words.positive? && uppercase_words < @words_and_emoticons.size
     end
   end
 end

data/lib/vader_sentiment_ruby/sentiment_scores_sifter.rb

@@ -3,6 +3,7 @@
 module VaderSentimentRuby
   # Separates positive versus negative sentiment scores
   class SentimentScoresSifter
+    # @param [Array<Float>] sentiments Array of sentiments generated from words
     def initialize(sentiments)
       @sentiments = sentiments
       @pos_sum = 0.0
@@ -10,6 +11,9 @@ module VaderSentimentRuby
       @neu_count = 0
     end
 
+    # @return [Array<Float, Float, Integer>]
+    # @example
+    #   [2.3, -3.2, 3]
     def call
       @sentiments.each do |sentiment_score|
         # compensates for neutral words that are counted as 1

data/lib/vader_sentiment_ruby/valence_score_calculator.rb

@@ -10,16 +10,19 @@ module VaderSentimentRuby
       compound: 0.0
     }.freeze
 
+    # @param [Array<Float, Integer>] sentiments Array of sentiments for text
+    # @param [String] text
     def initialize(sentiments, text)
       @sentiments = sentiments
       @text = text
     end
 
+    # @return [Hash<Float, Float, Float, Float>] Semantic score response hash
     def call
-      return DEFAULT_RESPONSE unless @sentiments
+      return DEFAULT_RESPONSE if @sentiments.empty?
 
       sum_s = @sentiments.map(&:to_f).sum
-      # compute and add emphasis from punctuation in text
+      # Compute and add emphasis from punctuation in text
       punct_emph_amplifier = PunctuationEmphasisAmplifier.new(@text).call
       compound = normalize(sum_s, punct_emph_amplifier)
 
@@ -28,9 +31,7 @@ module VaderSentimentRuby
 
     private
 
-    # Normalize the score to be between -1 and 1 using an alpha that
-    # approximates the max expected value
-    # Move to Sentiment analyzer
+    # Normalizes the score to be between -1 and 1 using an alpha that approximates the max expected value
     def normalize(score, punct_emph_amplifier, alpha = 15)
       score = add_punctuation_emphasis(score, punct_emph_amplifier)
       norm_score = score / Math.sqrt((score * score) + alpha).to_f
@@ -42,11 +43,8 @@ module VaderSentimentRuby
     end
 
     def add_punctuation_emphasis(sum_s, punct_emph_amplifier)
-      if sum_s.positive?
-        sum_s += punct_emph_amplifier
-      elsif sum_s.negative?
-        sum_s -= punct_emph_amplifier
-      end
+      return sum_s + punct_emph_amplifier if sum_s.positive?
+      return sum_s - punct_emph_amplifier if sum_s.negative?
 
       sum_s
     end
@@ -65,9 +63,9 @@ module VaderSentimentRuby
     end
     # rubocop:enable Metrics/AbcSize
 
-    # Prepare scores sum for result calculation
+    # Prepares score sums for result calculation
     def scores(punct_emph_amplifier)
-      # discriminate between positive, negative and neutral sentiment scores
+      # Discriminate between positive, negative and neutral sentiment scores
       pos_sum, neg_sum, neu_count = SentimentScoresSifter.new(@sentiments).call
 
       if pos_sum > neg_sum.to_f.abs

data/lib/vader_sentiment_ruby/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module VaderSentimentRuby
-  VERSION = '0.1.0'
+  VERSION = '0.1.1'
 end

data/lib/vader_sentiment_ruby/word_helper.rb

@@ -5,6 +5,7 @@ module VaderSentimentRuby
   # word_upcase?(word) is similar to Python's word.isupper()
   # strip_punctuation(word) is similar to Python's word.strip(string.punctuation)
   module WordHelper
+    LETTERS_RANGE = 'A-Za-z'
     PUNCTUATIONS = '!"#$%&\'()*+,-./:;<=>?@[\\]^_`{|}~'
 
     class << self
@@ -16,7 +17,7 @@ module VaderSentimentRuby
       #   word_upcase?(':D') # => true
       #   word_upcase?(':)') # => false
       def word_upcase?(word)
-        word == word.upcase && word.count('A-Za-z').positive?
+        word == word.upcase && word.count(LETTERS_RANGE).positive?
       end
 
       # Removes all trailing and leading punctuation
@@ -31,62 +32,28 @@ module VaderSentimentRuby
       #   strip_punctuation("'don't'") # => "don't"
       #   strip_punctuation(":)") # => ":)"
       def strip_punctuation(token)
-        token_without_punctuation = replace_punctuations(token)
-
         original_set = token.split('')
-        updated_set = token_without_punctuation.split('')
-
-        pair_array = prepare_match_array(original_set, updated_set)
-        pair_array = clean_leading_punctuations(pair_array)
-        pair_array = clean_trailing_punctuations(pair_array)
 
-        stripped = pair_array.map { |item| item[:old_ch] }.join
+        array = clean_leading_punctuations(original_set)
+        array = clean_trailing_punctuations(array)
+        stripped_token = array.join
 
-        return token if stripped.size <= 2
+        return token if stripped_token.size <= 2
 
-        stripped
+        stripped_token
       end
 
       private
 
-      def replace_punctuations(token)
-        punctuation_array = PUNCTUATIONS.split('')
-
-        punctuation_array.each do |punctuation|
-          token = token.gsub(punctuation, ' ')
-        end
-
-        token
-      end
-
-      def prepare_match_array(original_set, updated_set)
-        pair_array = []
-        original_set.each_with_index do |item, index|
-          pair_array << { index: index, old_ch: item, new_ch: updated_set[index] }
-        end
-
-        pair_array
+      def clean_leading_punctuations(token_array)
+        token_array.drop_while { |letter| PUNCTUATIONS.include? letter }
       end
 
-      def clean_leading_punctuations(pair_array)
-        pair_array.map do |pair|
-          break if pair[:new_ch] != ' '
-
-          pair_array.delete_at(pair[:index])
-        end
-
-        pair_array
-      end
-
-      def clean_trailing_punctuations(pair_array)
-        reversed_array = pair_array.reverse
-        reversed_array.map do |pair|
-          break if pair[:new_ch] != ' '
-
-          pair_array.delete_at(pair[:index])
-        end
-
-        pair_array
+      def clean_trailing_punctuations(token_array)
+        token_array
+          .reverse
+          .drop_while { |letter| PUNCTUATIONS.include? letter }
+          .reverse
       end
     end
   end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: vader_sentiment_ruby
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.1
 platform: ruby
 authors:
 - Nickolay Bulavin
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-02-18 00:00:00.000000000 Z
+date: 2021-02-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: byebug