Add continuous language option

lib/ingreedy.rb

@@ -1,6 +1,7 @@
 path = File.expand_path(File.join(File.dirname(__FILE__), "ingreedy"))
 
 require File.join(path, "case_insensitive_parser")
+require File.join(path, "continuous_language_locale")
 require File.join(path, "ingreedy_parser")
 require File.join(path, "dictionary_collection")
 

lib/ingreedy/amount_parser.rb

@@ -3,9 +3,14 @@
 module Ingreedy
   class AmountParser < Parslet::Parser
     include CaseInsensitiveParser
+    include ContinuousLanguageLocale
 
     rule(:whitespace) do
-      match("\s")
+      if use_whitespace?(current_locale)
+        match("\s")
+      else
+        match("\s").maybe
+      end
     end
 
     rule(:integer) do
@@ -58,6 +63,10 @@ class AmountParser < Parslet::Parser
 
     private
 
+    def current_locale
+      Ingreedy.dictionaries.current.locale
+    end
+
     def word_digits
       Ingreedy.dictionaries.current.numbers.keys
     end

lib/ingreedy/continuous_language_locale.rb

@@ -0,0 +1,9 @@
+module Ingreedy
+  module ContinuousLanguageLocale
+    CONTINUOUS_LANGUAGES_LOCALES = %i(ja th zh-TW)
+
+    def use_whitespace?(locale)
+      !CONTINUOUS_LANGUAGES_LOCALES.include?(locale)
+    end
+  end
+end

lib/ingreedy/dictionary.rb

@@ -1,14 +1,15 @@
 module Ingreedy
   class Dictionary
-    attr_reader :units, :numbers, :prepositions, :range_separators
+    attr_reader :units, :numbers, :prepositions, :range_separators, :locale
     attr_reader :imprecise_amounts
 
-    def initialize(units:, numbers: {}, prepositions: [], range_separators: %w{- ~}, imprecise_amounts: [])
+    def initialize(units:, numbers: {}, prepositions: [], range_separators: %w{- ~}, imprecise_amounts: [], locale: nil)
       @units = units
       @numbers = sort_by_length(numbers)
       @prepositions = prepositions
       @range_separators = range_separators
       @imprecise_amounts = imprecise_amounts
+      @locale = locale
     end
 
     # https://en.wikipedia.org/wiki/Number_Forms

lib/ingreedy/dictionary_collection.rb

@@ -8,7 +8,7 @@ def initialize
     end
 
     def []=(locale, attributes)
-      @collection[locale] = Dictionary.new(**attributes)
+      @collection[locale] = Dictionary.new(**attributes, locale: locale)
     end
 
     def current
@@ -38,7 +38,7 @@ def i18n_gem_locales
     end
 
     def fetch_dictionary(locale)
-      @collection[locale] ||= Dictionary.new **load_yaml(locale)
+      @collection[locale] ||= Dictionary.new(**load_yaml(locale), locale: locale)
     rescue Errno::ENOENT
     end
 

lib/ingreedy/root_parser.rb

@@ -1,6 +1,7 @@
 module Ingreedy
   class RootParser < Parslet::Parser
     include CaseInsensitiveParser
+    include ContinuousLanguageLocale
 
     rule(:range) do
       AmountParser.new.as(:amount) >>
@@ -19,7 +20,11 @@ class RootParser < Parslet::Parser
     end
 
     rule(:whitespace) do
-      match("\s")
+      if use_whitespace?(current_locale)
+        match("\s")
+      else
+        match("\s").maybe
+      end
     end
 
     rule(:container_amount) do
@@ -87,7 +92,7 @@ class RootParser < Parslet::Parser
 
     rule(:reverse_format) do
       # e.g. flour 200g
-      ((whitespace >> quantity).absent? >> any).repeat.as(:ingredient) >>
+      ((whitespace >> quantity >> any.absent?).absent? >> any).repeat.as(:ingredient) >>
         whitespace >>
         quantity
     end
@@ -119,6 +124,10 @@ def parse
     private
 
     attr_reader :original_query
+
+    def current_locale
+      Ingreedy.dictionaries.current.locale
+    end
 
     def imprecise_amounts
       Ingreedy.dictionaries.current.imprecise_amounts

spec/ingreedy/continuouse_language_locale_spec.rb

@@ -0,0 +1,23 @@
+require 'ingreedy/continuous_language_locale'
+
+RSpec.describe Ingreedy::ContinuousLanguageLocale do
+  include Ingreedy::ContinuousLanguageLocale
+
+  describe '#use_whitespace?' do
+    context 'when the locale is a continuous language' do
+      it 'returns false' do
+        expect(use_whitespace?(:ja)).to be_falsey
+        expect(use_whitespace?(:th)).to be_falsey
+        expect(use_whitespace?(:'zh-TW')).to be_falsey
+      end
+    end
+
+    context 'when the locale is not a continuous language' do
+      it 'returns true' do
+        expect(use_whitespace?(:en)).to be_truthy
+        expect(use_whitespace?(:fr)).to be_truthy
+        expect(use_whitespace?(:'zh-CN')).to be_truthy
+      end
+    end
+  end
+end

spec/ingreedy_spec.rb

@@ -447,6 +447,84 @@
   end
 end
 
+describe Ingreedy, "continuous language" do
+  context "Japanese" do
+    before do
+      Ingreedy.dictionaries[:ja] = {
+        units: { gram: ["g"], other: ["個"] },
+        numbers: { "一" => 1 },
+      }
+      Ingreedy.locale = :ja
+    end
+
+    after do
+      Ingreedy.locale = nil
+    end
+
+    it "parses correctly" do
+      result = Ingreedy.parse "200g砂糖"
+
+      expect(result.amount).to eq(200)
+      expect(result.unit).to eq(:gram)
+      expect(result.ingredient).to eq("砂糖")
+    end
+
+    it "parses correctly with reverse format" do
+      result = Ingreedy.parse "砂糖200g"
+
+      expect(result.amount).to eq(200)
+      expect(result.unit).to eq(:gram)
+      expect(result.ingredient).to eq("砂糖")
+    end
+
+    it "parses correctly with numbers" do
+      result = Ingreedy.parse "卵一個"
+
+      expect(result.amount).to eq(1)
+      expect(result.unit).to eq(:other)
+      expect(result.ingredient).to eq("卵")
+    end
+  end
+
+  context "Taiwanese" do
+    before do
+      Ingreedy.dictionaries[:"zh-TW"] = {
+        units: { gram: ["g"], other: ["个"] },
+        numbers: { "一" => 1 },
+      }
+      Ingreedy.locale = :"zh-TW"
+    end
+
+    after do
+      Ingreedy.locale = nil
+    end
+
+    it "parses correctly" do
+      result = Ingreedy.parse "200g砂糖"
+
+      expect(result.amount).to eq(200)
+      expect(result.unit).to eq(:gram)
+      expect(result.ingredient).to eq("砂糖")
+    end
+
+    it "parses correctly with reverse format" do
+      result = Ingreedy.parse "砂糖200g"
+
+      expect(result.amount).to eq(200)
+      expect(result.unit).to eq(:gram)
+      expect(result.ingredient).to eq("砂糖")
+    end
+
+    it "parses correctly with numbers" do
+      result = Ingreedy.parse "一个鸡蛋"
+
+      expect(result.amount).to eq(1)
+      expect(result.unit).to eq(:other)
+      expect(result.ingredient).to eq("鸡蛋")
+    end
+  end
+end
+
 describe Ingreedy, "error handling" do
   it "wraps Parslet exceptions in a custom exception" do
     expect do