ingreedy 0.0.5 → 0.0.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: b9b4b6105807082714db4a684b0667635a6c0a82
-  data.tar.gz: 85c9ab384b6c419520723d9719b9f82d6ebca495
+  metadata.gz: e72a77814ed8f70d4de6ea053c9574de0e1c9723
+  data.tar.gz: 08e41cd39042f6699d75f635ff5ca036a5ebec31
 SHA512:
-  metadata.gz: 3321ead5195bc89f49a38e9b9fd762f624e9dfed41486a703d05d4ffade3a5f4f1ea3898b16b99d598a0790ea3f108940851e3b6ab831abc388d1160cf095771
-  data.tar.gz: 545113b12845639fb146fe26d9e263044452bcab8be636f5d83937752b0e74aaa07dbf6050550684e6ce444c5acdb6f3ff74ace23068c3cb7754af46e4ef6053
+  metadata.gz: 22cc38a7b4866f6058a97494e038fe97b450110e89f689188b4ddb4e675bd8832a521827d05191cc620bf69bdd34907936db194fcd4e657652b6cba73cef6668
+  data.tar.gz: 23b3d8af25501391647bc8bf73ef57f00d84ff6193a14014f551271cc5a0e531a7d717b12a893abb965ae72f18799e0a6e40f42d67ccba1f9c04fb7bc34fb926

data/lib/ingreedy/amount_parser.rb

@@ -5,14 +5,6 @@ module Ingreedy
   class AmountParser < Parslet::Parser
     include CaseInsensitiveParser
 
-    def initialize(options = {})
-      @key_prefix = options[:key_prefix] ? "#{options[:key_prefix]}_" : ''
-    end
-
-    def capture_key(key)
-      (@key_prefix + key.to_s).to_sym
-    end
-
     rule(:whitespace) do
       match("\s")
     end
@@ -23,12 +15,33 @@ module Ingreedy
 
     rule(:float) do
       integer.maybe >>
-      str('.') >> integer
+      float_delimiter >> integer
+    end
+
+    rule(:float_delimiter) do
+      str(',') | str('.')
     end
 
     rule(:fraction) do
-      (integer.as(capture_key(:integer_amount)) >> whitespace).maybe >>
-      (integer >> match('/') >> integer).as(capture_key(:fraction_amount))
+      compound_simple_fraction | compound_vulgar_fraction
+    end
+
+    rule(:compound_simple_fraction) do
+      (integer.as(:integer_amount) >> whitespace).maybe >>
+      simple_fraction.as(:fraction_amount)
+    end
+
+    rule(:simple_fraction) do
+      integer >> match('/') >> integer
+    end
+
+    rule(:compound_vulgar_fraction) do
+      (integer.as(:integer_amount) >> whitespace.maybe).maybe >>
+      vulgar_fraction.as(:fraction_amount)
+    end
+
+    rule(:vulgar_fraction) do
+      vulgar_fractions.map { |f| str(f) }.inject(:|)
     end
 
     rule(:word_digit) do
@@ -41,20 +54,21 @@ module Ingreedy
 
     rule(:amount) do
       fraction |
-      float.as(capture_key(:float_amount)) |
-      integer.as(capture_key(:integer_amount)) |
-      word_digit.as(capture_key(:word_integer_amount)) >> amount_unit_separator
+      float.as(:float_amount) |
+      integer.as(:integer_amount) |
+      word_digit.as(:word_integer_amount) >> amount_unit_separator
     end
 
     root(:amount)
 
     private
 
-    def word_digits
-      Ingreedy.dictionaries.current.numbers.keys
-    end
+      def word_digits
+        Ingreedy.dictionaries.current.numbers.keys
+      end
 
+      def vulgar_fractions
+        Ingreedy.dictionaries.current.vulgar_fractions.keys
+      end
   end
-
-
 end

data/lib/ingreedy/dictionary.rb

@@ -7,5 +7,29 @@ module Ingreedy
       @numbers = entries[:numbers] || {}
       @prepositions = entries[:prepositions] || []
     end
+
+    # https://en.wikipedia.org/wiki/Number_Forms
+    def vulgar_fractions
+      {
+        "\u00BC" => '1/4',
+        "\u00BD" => '1/2',
+        "\u00BE" => '3/4',
+        "\u2150" => '1/7',
+        "\u2151" => '1/9',
+        "\u2152" => '1/10',
+        "\u2153" => '1/3',
+        "\u2154" => '2/3',
+        "\u2155" => '1/5',
+        "\u2156" => '2/5',
+        "\u2157" => '3/5',
+        "\u2158" => '4/5',
+        "\u2159" => '1/6',
+        "\u215A" => '5/6',
+        "\u215B" => '1/8',
+        "\u215C" => '3/8',
+        "\u215D" => '5/8',
+        "\u215E" => '7/8'
+      }
+    end
   end
 end

data/lib/ingreedy/ingreedy_parser.rb

@@ -2,7 +2,6 @@ require 'parslet'
 
 require_relative 'amount_parser'
 require_relative 'rationalizer'
-require_relative 'unit_parser'
 require_relative 'unit_variation_mapper'
 
 module Ingreedy
@@ -10,7 +9,15 @@ module Ingreedy
   class Parser < Parslet::Parser
 
     attr_reader :original_query
-    Result = Struct.new(:amount, :unit, :ingredient, :original_query)
+    Result = Struct.new(:amount, :unit, :container_amount, :container_unit, :ingredient, :original_query)
+
+    rule(:range) do
+      AmountParser.new.as(:amount) >>
+      whitespace.maybe >>
+      str('-') >>
+      whitespace.maybe >>
+      AmountParser.new.as(:amount_end)
+    end
 
     rule(:amount) do
       AmountParser.new.as(:amount)
@@ -21,18 +28,30 @@ module Ingreedy
     end
 
     rule(:container_amount) do
-      AmountParser.new(key_prefix: 'container')
+      AmountParser.new
     end
 
     rule(:unit) do
-      UnitParser.new
+      if unit_matches.any?
+        unit_matches.map { |u| str(u) }.inject(:|)
+      else
+        str('')
+      end
+    end
+
+    rule(:container_unit) do
+      unit
     end
 
     rule(:unit_and_preposition) do
+      unit.as(:unit) >> (preposition_or_whitespace | any.absent?)
+    end
+
+    rule(:preposition_or_whitespace) do
       if prepositions.empty?
-        unit.as(:unit) >> (whitespace | any.absent?)
+        whitespace
       else
-        unit.as(:unit) >> (preposition | whitespace | any.absent?)
+        preposition | whitespace
       end
     end
 
@@ -42,10 +61,6 @@ module Ingreedy
       whitespace
     end
 
-    rule(:container_unit) do
-      UnitParser.new
-    end
-
     rule(:amount_unit_separator) do
       whitespace | str('-')
     end
@@ -55,12 +70,12 @@ module Ingreedy
       str('(').maybe >>
       container_amount.as(:container_amount) >>
       amount_unit_separator.maybe >>
-      container_unit.as(:unit) >>
-      str(')').maybe >> whitespace
+      container_unit.as(:container_unit) >>
+      str(')').maybe >> preposition_or_whitespace
     end
 
     rule(:amount_and_unit) do
-      amount >>
+      (range | amount) >>
       whitespace.maybe >>
       unit_and_preposition.maybe >>
       container_size.maybe
@@ -94,21 +109,26 @@ module Ingreedy
       result = Result.new
       result[:original_query] = original_query
 
-      parslet_output = super(original_query)
+      parslet = super(original_query)
 
-      result[:amount] = rationalize_total_amount(parslet_output[:amount], parslet_output[:container_amount])
+      result[:amount] = rationalize parslet[:amount]
+      result[:amount] = [result[:amount], rationalize(parslet[:amount_end])] if parslet[:amount_end]
+      result[:container_amount] = rationalize(parslet[:container_amount])
 
-      if parslet_output[:unit]
-        result[:unit] = convert_unit_variation_to_canonical(parslet_output[:unit].to_s)
-      end
+      result[:unit] = convert_unit_variation_to_canonical(parslet[:unit].to_s) if parslet[:unit]
+      result[:container_unit] = convert_unit_variation_to_canonical(parslet[:container_unit].to_s) if parslet[:container_unit]
 
-      result[:ingredient] = parslet_output[:ingredient].to_s.lstrip.rstrip #TODO cheating
+      result[:ingredient] = parslet[:ingredient].to_s.lstrip.rstrip #TODO cheating
 
       result
     end
 
     private
 
+    def unit_matches
+      @unit_matches ||= original_query.scan(UnitVariationMapper.regexp).sort_by(&:length).reverse
+    end
+
     def prepositions
       Ingreedy.dictionaries.current.prepositions
     end
@@ -117,26 +137,18 @@ module Ingreedy
       UnitVariationMapper.unit_from_variation(unit_variation)
     end
 
-    def rationalize_total_amount(amount, container_amount)
-      if container_amount
-        rationalize_amount(amount) * rationalize_amount(container_amount, 'container_')
-      else
-        rationalize_amount(amount)
-      end
-    end
-
-    def rationalize_amount(amount, capture_key_prefix = '')
+    def rationalize(amount)
       return unless amount
-      integer = amount["#{capture_key_prefix}integer_amount".to_sym]
+      integer = amount[:integer_amount]
       integer &&= integer.to_s
 
-      float = amount["#{capture_key_prefix}float_amount".to_sym]
+      float = amount[:float_amount]
       float &&= float.to_s
 
-      fraction = amount["#{capture_key_prefix}fraction_amount".to_sym]
+      fraction = amount[:fraction_amount]
       fraction &&= fraction.to_s
 
-      word = amount["#{capture_key_prefix}word_integer_amount".to_sym]
+      word = amount[:word_integer_amount]
       word &&= word.to_s
 
       Rationalizer.rationalize(
@@ -146,6 +158,5 @@ module Ingreedy
         word:     word
       )
     end
-
   end
 end

data/lib/ingreedy/rationalizer.rb

@@ -1,7 +1,5 @@
 module Ingreedy
-
   class Rationalizer
-
     def self.rationalize(options)
       new(options).rationalize
     end
@@ -17,14 +15,14 @@ module Ingreedy
       if @word
         result = rationalize_word
       elsif @fraction
-        result = @fraction.to_r
+        result = rationalize_fraction
         if @integer
           result += @integer.to_i
         end
       elsif @integer
         result = @integer.to_r
       elsif @float
-        result = @float.to_r
+        result = @float.gsub(',', '.').to_r
       end
 
       result
@@ -32,10 +30,17 @@ module Ingreedy
 
     private
 
-    def rationalize_word
-      Ingreedy.dictionaries.current.numbers[@word.downcase]
-    end
+      def rationalize_fraction
+        vulgar_fractions.each { |char, amount| @fraction.gsub!(char, amount.to_s) }
+        @fraction.to_r
+      end
 
-  end
+      def vulgar_fractions
+        Ingreedy.dictionaries.current.vulgar_fractions
+      end
 
+      def rationalize_word
+        Ingreedy.dictionaries.current.numbers[@word.downcase]
+      end
+  end
 end

data/lib/ingreedy/unit_variation_mapper.rb

@@ -1,6 +1,11 @@
 module Ingreedy
   class UnitVariationMapper
 
+    def self.regexp
+      regexp = all_variations.map { |v| Regexp.escape(v) }.join('|')
+      regexp = Regexp.new(regexp, Regexp::IGNORECASE)
+    end
+
     def self.all_variations
       # Return these in order of size, descending
       # That way, the longer versions will try to be parsed first, then the shorter versions

data/lib/ingreedy/version.rb

@@ -1,4 +1,4 @@
 module Ingreedy
-  VERSION = '0.0.5'
+  VERSION = '0.0.6'
 end
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: ingreedy
 version: !ruby/object:Gem::Version
-  version: 0.0.5
+  version: 0.0.6
 platform: ruby
 authors:
 - Ian C. Anderson
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-10-17 00:00:00.000000000 Z
+date: 2015-10-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: parslet
@@ -127,7 +127,7 @@ dependencies:
 description: Natural language recipe ingredient parser that supports numeric amount,
   units, and ingredient
 email:
-- anderson.ian.c@gmail.com
+- ian@iancanderson.com
 executables: []
 extensions: []
 extra_rdoc_files: []
@@ -139,7 +139,6 @@ files:
 - lib/ingreedy/dictionary_collection.rb
 - lib/ingreedy/ingreedy_parser.rb
 - lib/ingreedy/rationalizer.rb
-- lib/ingreedy/unit_parser.rb
 - lib/ingreedy/unit_variation_mapper.rb
 - lib/ingreedy/version.rb
 homepage: http://github.com/iancanderson/ingreedy

data/lib/ingreedy/unit_parser.rb

@@ -1,20 +0,0 @@
-module Ingreedy
-
-  class UnitParser < Parslet::Parser
-    include CaseInsensitiveParser
-
-    rule(:unit) do
-      unit_variations.map { |var| str(var) | stri(var) }.reduce(:|)
-    end
-
-    root :unit
-
-    private
-
-    def unit_variations
-      UnitVariationMapper.all_variations
-    end
-
-  end
-
-end