Re-extract lexicon builder

README.md

@@ -38,18 +38,21 @@ responsiveVoice.speak(sentence, "UK English Female");
 ### Generate lexicons to override text-to-speech pronunciation
 ```ruby
 require 'aws-sdk-polly'
+polly = Aws::Polly::Client.new
 
-polly = Aws::Polly::Client.new(region: 'us-west-2')
+sentence = "PATER NOSTER qui es in caelis"
 
-# Add the lexicons
-MedievalLatina::Lexicon.file_names_with_contents.each do |name, content|
-  polly.put_lexicon(name: name, content: content)
-end
+words = sentence.split(" ")
+pronunciations = MedievalLatina.pronunciations_for(words)
+lexicon = MedievalLatina::LexiconBuilder.new(pronunciations).call
+
+name = "CustomLatin"
+polly.put_lexicon(name: name, content: lexicon.to_s)
 
 # Synthesize speech using the lexicons
 polly.synthesize_speech(
-  lexicon_names: MedievalLatina::Lexicon.list_files,
-  text: "PATER NOSTER, qui es in caelis",
+  lexicon_names: [name],
+  text: sentence,
   output_format: "mp3",
   voice_id: "Joanna"
 )

bin/build

@@ -1,33 +1,15 @@
 #!/usr/bin/env ruby
 
 require 'json'
-require 'rexml/document'
-require 'cgi'
-include REXML
+require "bundler/setup"
+require "medieval_latina"
+require "medieval_latina/lexicon_builder"
 
 def create_pls_file(file_path, words)
-  doc = Document.new
-  doc.add_element 'lexicon', {'version' => '1.0',
-                              'xmlns' => 'http://www.w3.org/2005/01/pronunciation-lexicon',
-                              'alphabet' => 'ipa',
-                              'xml:lang' => 'en-US'}
-
-  words.each do |word, pronunciation|
-    lexeme = Element.new('lexeme')
-    grapheme = Element.new('grapheme')
-    phoneme = Element.new('phoneme')
-
-    grapheme.text = CGI.unescapeHTML(word)
-    phoneme.text = pronunciation
-
-    lexeme.add_element(grapheme)
-    lexeme.add_element(phoneme)
-
-    doc.root.add_element(lexeme)
-  end
+  document = MedievalLatina::LexiconBuilder.new(words).call
 
   File.open(file_path, 'w') do |file|
-    doc.write(file, 2)
+    document.write(file, 2)
     file.write("\n")
   end
 end

lib/medieval_latina.rb

@@ -1,6 +1,7 @@
 require "json"
 require "medieval_latina/initializer"
 require "medieval_latina/lexicon"
+require "medieval_latina/lexicon_builder"
 require "medieval_latina/version"
 require "set"
 
@@ -80,6 +81,16 @@ def self.nouns
     end
   end
 
+  def self.pronunciations_for(words)
+    words.map(&:downcase).each_with_object({}) do |word, hash|
+      metadata = DICTIONARY[word]
+
+      if metadata["ipa"]
+        hash[word] = metadata["ipa"]
+      end
+    end
+  end
+
   def self.rejoin_words(array)
     array
       .join(" ")

lib/medieval_latina/lexicon_builder.rb

@@ -0,0 +1,44 @@
+require "cgi"
+require "rexml/document"
+
+class MedievalLatina
+  class LexiconBuilder
+    include REXML
+
+    def initialize(words)
+      @document = Document.new
+      @words = words
+    end
+
+    def call
+      document.add_element "lexicon", SPECIFICATION
+
+      words.each do |word, pronunciation|
+        lexeme = Element.new("lexeme")
+        grapheme = Element.new("grapheme")
+        phoneme = Element.new("phoneme")
+
+        grapheme.text = CGI.unescapeHTML(word)
+        phoneme.text = pronunciation
+
+        lexeme.add_element(grapheme)
+        lexeme.add_element(phoneme)
+
+        document.root.add_element(lexeme)
+      end
+
+      document
+    end
+
+    private
+
+    attr_reader :document, :words
+
+    SPECIFICATION = {
+      "version" => "1.0",
+      "xmlns" => "http://www.w3.org/2005/01/pronunciation-lexicon",
+      "alphabet" => "ipa",
+      "xml:lang" => "en-US"
+    }.freeze
+  end
+end

spec/medieval_latina_spec.rb

@@ -133,6 +133,17 @@
     specify { expect(described_class).to respond_to(:adverbs) }
   end
 
+  describe ".pronunciations_for" do
+    let(:words) { ["beatus", "vir"] }
+
+    it "returns only the words and pronunciations asked for" do
+      expect(described_class.pronunciations_for(words)).to match(
+        "beatus" => "beatʊsː",
+        "vir" => "vir"
+      )
+    end
+  end
+
   describe ".verbs" do
     specify { expect(described_class).to respond_to(:verbs) }
   end