phonetics 3.0.5 → 3.0.8

data/lib/phonetics/code_generator.rb

@@ -7,7 +7,7 @@ module Phonetics
   class CodeGenerator
     attr_reader :writer
 
-    def initialize(writer = STDOUT)
+    def initialize(writer = $stdout)
       @writer = writer
     end
 

data/lib/phonetics/distances.rb

@@ -1,6 +1,7 @@
 # frozen_string_literal: true
 
 require 'delegate'
+require 'set'
 
 module Phonetics
   extend self
@@ -131,6 +132,8 @@ module Phonetics
     )
     # rubocop:enable Layout/TrailingWhitespace
 
+    # rubocop:disable Metrics/CyclomaticComplexity
+    # rubocop:disable Metrics/PerceivedComplexity
     # Parse the ChartData into a lookup table where we can retrieve attributes
     # for each phoneme
     def features
@@ -176,6 +179,8 @@ module Phonetics
         end
       end
     end
+    # rubocop:enable Metrics/CyclomaticComplexity
+    # rubocop:enable Metrics/PerceivedComplexity
 
     def phonemes
       @phonemes ||= features.keys
@@ -183,20 +188,20 @@ module Phonetics
 
     # Given two consonants, calculate their difference by summing the
     # following:
-    #   * 0.1 if they are not voiced the same
+    #   * 0.3 if they are not voiced the same
     #   * 0.3 if they are different manners
-    #   * Up to 0.6 if they are the maximum position difference
+    #   * Up to 0.4 if they are the maximum position difference
     def distance(phoneme1, phoneme2)
       features1 = features[phoneme1]
       features2 = features[phoneme2]
 
       penalty = 0
-      penalty += 0.1 if features1[:voiced] != features2[:voiced]
+      penalty += 0.3 if features1[:voiced] != features2[:voiced]
 
       penalty += 0.3 if features1[:manner] != features2[:manner]
 
-      # Use up to the remaining 0.6 for penalizing differences in manner
-      penalty += 0.6 * ((features1[:position_index] - features2[:position_index]).abs / @position_count.to_f)
+      # Use up to the remaining 0.4 for penalizing differences in manner
+      penalty += 0.4 * ((features1[:position_index] - features2[:position_index]).abs / @position_count.to_f)
       penalty
     end
   end
@@ -228,11 +233,12 @@ module Phonetics
 
   def _distance(phoneme1, phoneme2)
     types = [Symbols.fetch(phoneme1), Symbols.fetch(phoneme2)].sort
-    if types == %i[consonant vowel]
+    case types
+    when %i[consonant vowel]
       1.0
-    elsif types == %i[vowel vowel]
+    when %i[vowel vowel]
       Vowels.distance(phoneme1, phoneme2)
-    elsif types == %i[consonant consonant]
+    when %i[consonant consonant]
       Consonants.distance(phoneme1, phoneme2)
     end
   end

data/lib/phonetics/levenshtein.rb

@@ -16,10 +16,12 @@ module Phonetics
   module Levenshtein
     extend ::PhoneticsLevenshteinCBinding
 
+    # rubocop:disable Style/OptionalBooleanParameter
     def self.distance(str1, str2, verbose = false)
       return if str1.nil? || str2.nil?
 
       internal_phonetic_distance(str1, str2, verbose)
     end
+    # rubocop:enable Style/OptionalBooleanParameter
   end
 end

data/lib/phonetics/ruby_levenshtein.rb

@@ -16,6 +16,7 @@ module Phonetics
   class RubyLevenshtein
     attr_reader :str1, :str2, :len1, :len2, :matrix
 
+    # rubocop:disable Style/OptionalBooleanParameter
     def initialize(ipa_str1, ipa_str2, verbose = false)
       @str1 = ipa_str1.each_char.select { |c| Phonetics.phonemes.include?(c) }.join
       @str2 = ipa_str2.each_char.select { |c| Phonetics.phonemes.include?(c) }.join
@@ -26,6 +27,11 @@ module Phonetics
       set_edit_distances(@str1, @str2)
     end
 
+    def self.distance(str1, str2, verbose = false)
+      new(str1, str2, verbose).distance
+    end
+    # rubocop:enable Style/OptionalBooleanParameter
+
     def distance
       return 0 if walk.empty?
 
@@ -33,10 +39,6 @@ module Phonetics
       walk.last[:distance]
     end
 
-    def self.distance(str1, str2, verbose = false)
-      new(str1, str2, verbose).distance
-    end
-
     private
 
     def walk

data/lib/phonetics/transcriptions.rb

@@ -10,7 +10,8 @@ module Phonetics
 
   module Transcriptions
     extend self
-    Transcriptions = File.join(__dir__, '..', 'common_ipa_transcriptions.json')
+
+    TranscriptionFile = File.join(__dir__, '..', 'common_ipa_transcriptions.json')
     TranscriptionsURL = 'https://jackdanger.com/common_ipa_transcriptions.json'
 
     SourcesByPreference = [/wiktionary/, /cmu/, /phonemicchart.com/].freeze
@@ -21,7 +22,7 @@ module Phonetics
       return unless entry['ipa']
 
       SourcesByPreference.each do |preferred_source|
-        entry['ipa'].keys.each do |source|
+        entry['ipa'].each_key do |source|
           return entry['ipa'][source] if source =~ preferred_source
         end
       end
@@ -30,21 +31,24 @@ module Phonetics
 
     def transcriptions
       @transcriptions ||= begin
-        download! unless File.exist?(Transcriptions)
+        download! unless File.exist?(TranscriptionFile)
         load_from_disk!
       end
     end
 
     # Lazily loaded from JSON file on disk
     def load_from_disk!
-      @transcriptions = JSON.parse(File.read(Transcriptions))
+      @transcriptions = JSON.parse(File.read(TranscriptionFile))
     end
 
+    # rubocop:disable Security/Open
     def download!
       File.open(Transcriptions, 'w') { |f| f.write(URI.open(TranscriptionsURL).read) }
     end
+    # rubocop:enable Security/Open
 
-    def trie
+    # rubocop:disable Metrics/CyclomaticComplexity
+    def trie(max_rarity = nil)
       # Let's turn this:
       #
       #    "century": {
@@ -109,9 +113,12 @@ module Phonetics
       #     },
       #   },
       #
-      @trie ||= begin
+      @tries ||= {}
+      @tries[max_rarity] ||= begin
         base_trie = {}
         transcriptions.each do |key, entry|
+          next if max_rarity && (entry['rarity'].nil? || entry['rarity'] > max_rarity)
+
           entry_data = {
               word: key,
             rarity: entry['rarity'],
@@ -123,6 +130,7 @@ module Phonetics
         base_trie.freeze
       end
     end
+    # rubocop:enable Metrics/CyclomaticComplexity
 
     def walk(ipa)
       ipa.each_char.reduce(trie) { |acc, char| acc[char] }

data/phonetics.gemspec

@@ -11,6 +11,8 @@ Gem::Specification.new do |spec|
   spec.homepage      = 'https://github.com/JackDanger/phonetics'
   spec.license       = 'MIT'
 
+  spec.required_ruby_version = '>= 2.5'
+
   spec.extensions = ['ext/c_levenshtein/extconf.rb']
 
   # Specify which files should be added to the gem when it is released.

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: phonetics
 version: !ruby/object:Gem::Version
-  version: 3.0.5
+  version: 3.0.8
 platform: ruby
 authors:
 - Jack Danger
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-12-27 00:00:00.000000000 Z
+date: 2022-05-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -149,14 +149,14 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '0'
+      version: '2.5'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.3
+rubygems_version: 3.0.3.1
 signing_key: 
 specification_version: 4
 summary: tools for linguistic code using the International Phonetic Alphabet