food_ingredient_parser 1.0.0.pre.7 → 1.0.0.pre.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: e37a1957ac49914eef067ca584d1c7ba252847d8
-  data.tar.gz: 91f97aa655ac16141ecbc0639c95579c52038666
+  metadata.gz: b8042180a4a4fbc5233a5630c7e0cf8e4751182b
+  data.tar.gz: 26abeaf528a49a6f01a47eb114e35631d9347cc7
 SHA512:
-  metadata.gz: 4a80f6b15b71e9e597991dfced1fed8ae0f486434483e8ab37e102b8be51d01273bf88ddc29a8f934065b0921ec68152ce485793a75ab1a751c3ec439566d011
-  data.tar.gz: 54ef1b2835363df9c72aab7a851cec7a555f04aefbeb41c0c865b142a08165df4172924d40e88d1f5ba1974e9f2203d65f23a6107a258d2185e27e0c054b2bb3
+  metadata.gz: 6d7c7972846a88046760de7d1c5857f426891502fdca4f50c0fad179f2a7580dd0b157aed106e9efa62af013e4a20d6d8e3be9c49ec8e4eeee326cb228e26c91
+  data.tar.gz: 872aadc53b40e991e156fde3bb89db69bebb52ca8993aadfbc1f3c862b76b0d42ac01b377da70af22cf6b909b24b063a16d59f1aab4ed6ddd4917e1196736c50

data/lib/food_ingredient_parser/loose/scanner.rb

@@ -127,7 +127,7 @@ module FoodIngredientParser::Loose
     def is_notes_start?
       # @todo use more heuristics: don't assume dot is notes when separator is a dot, and only toplevel?
       if ( is_mark? && @s[@i+mark_len..-1] =~ /\A\s*=/ ) ||     # "* = Biologisch"
-         ( is_mark? && @s[@i-2..@i-1] =~ /\A\s\s/ ) ||          # "  **Biologisch"
+         ( is_mark? && @s[@i-1] =~ /\s/ ) ||                    # " **Biologisch"
          ( @s[@i..-1] =~ NOTE_RE )                              # "E=", "Kan sporen van", ...
         @i -= 1 # we want to include the mark in the note
         true

data/lib/food_ingredient_parser/strict/grammar/common.treetop

@@ -13,7 +13,7 @@ module FoodIngredientParser::Strict::Grammar
       [[:alnum:]] /
       fraction /
       [-/\`'´’+=_{}&] /
-      [®] /
+      [®™] /
       [¿?] /                         # weird characters turning up in names (e.g. encoding issues)
       [₁₂₃₄₅₆₇₈₉]                    # can occur with vitamins
     end
@@ -39,20 +39,10 @@ module FoodIngredientParser::Strict::Grammar
       abbrev / char+
     end
 
-    rule word_nas
-      # word, but don't include the trailing '-' that may occure before an 'and'
-      abbrev / ( !andsep char )+
-    end
-
     rule and
       ( 'and' / 'en' / 'und' / '&' ) !char
     end
 
-    # we want to match "a and b" but not "a- and bthing", this allows to avoid the latter
-    rule andsep
-      '-' ws+ and
-    end
-
     rule abbrev
       # These are listed explicitely to avoid incorrect interpretations, and allow missing trailing dots.
       # To get an idea of what occurs (second one omits trailing dots):
@@ -107,44 +97,9 @@ module FoodIngredientParser::Strict::Grammar
         'vit.'i /
         'denat.'i /
         'N°'i /
-        '°C'i /
-        # word combinations that should not be split (not auto-generated)
-        # @todo this really would benefit from matching known ingredients instead of hardcoding
-        ( 'oliën'i / 'olien'i / 'olië'i / 'olie'i ) ws+ and ws+ ( 'vetten'i / 'vet'i ) /
-        'palm'i ws+ and ws+ 'kokosvet'i /
-        color ( ws+ and ws+ color )+ /
-        color2 ( ws+ and ws+ color2 )+ /
-        'kruiden'i ws+ and ws+ 'specerijen'i /
-        'kruiden'i ws+ and ws+ 'specerij'i /
-        'specerijen'i ws+ and ws+ 'kruiden'i /
-        'vitamine'i 'n'i? ws+ and ws+ 'mineralen'i /
-        'lactose'i ws+ and ws+ 'melk'i ( 'eiwit'i 'en'i? )?  /
-        'granen'i ws+ and ws+ 'zaden'i /
-        'gekookt'i [eE]? ws+ and ws+ 'gemarineerd'i [eE]? /
-        'mono'i ws+ and ws+ 'diglyceriden'i /
-        'guarpitmeel'i ws+ and ws+ 'natriumalginaat'i /
-        'vlees'i ws+ and ws+ 'dierlijke bijproducten'i /
-        'vis'i ws+ and ws+ 'visbijproducten'i /
-        'glucose'i ws+ and ws+ 'fructosestroop'i /
-        'ijzeroxiden'i ws+ and ws+ 'hydroxiden'i /
-        char+ 'sap'i ws+ and ws+ 'overige vruchtensappen'i /
-        char* 'sap'i ( ws+ 'uit concentraat'i / ws+ 'uit sapconcentraat'i )? ws+ and ws+ 'vruchten'i? 'puree'i /
-        ( 'vit.'i / 'vitamine'i / 'vitamin' ) ws+ [a-zA-Z] [0-9]* ws+ and ws+ [a-zA-Z] [0-9]* /
-        ( 'ijzer'i / 'chroom'i / 'koper'i ) ws* '(' 'I'+ ')' ws* [[:alnum:]]+
+        '°C'i
       )
       '.'? ![[:alpha:]]
     end
-
-    rule color
-      # used for paprika, honey ("yellow and white honey") (nouns)
-      'red'i / 'green'i / 'yellow'i / 'white'i / 'black'i /
-      'rood'i / 'groen'i / 'geel'i / 'wit'i / 'zwart'i
-    end
-
-    rule color2
-      # adjective colors (can not occur together with noun colors in a list)
-      'rode'i / 'groene'i / 'gele'i / 'witte'i / 'zwarte'i
-    end
-
   end
 end

data/lib/food_ingredient_parser/strict/grammar/ingredient_coloned.treetop

@@ -10,14 +10,13 @@ module FoodIngredientParser::Strict::Grammar
     end
 
     rule ingredient_coloned_inner_list
-      contains:( ingredient_coloned_simple_with_amount_and_nest ( ws+ and ws+ ingredient_coloned_simple_with_amount_and_nest )+ )  <ListNode> /
       contains:( ingredient_coloned_simple_with_amount_and_nest ws* ( '/'+ ws* ingredient_coloned_simple_with_amount_and_nest )* ) <ListNode>
     end
 
     # @see IngredientSimple#ingredient_simple
     rule ingredient_coloned_simple
-      name:( ingredient_coloned_word_nas ( andsep? ws+ !amount !and ingredient_coloned_word_nas )* ) ws? mark:mark <IngredientNode> /
-      name:( ingredient_coloned_word_nas ( andsep? ws+ !amount !and ingredient_coloned_word_nas )* )               <IngredientNode>
+      name:( ingredient_coloned_word ( ws+ !amount ingredient_coloned_word )* ) ws? mark:mark <IngredientNode> /
+      name:( ingredient_coloned_word ( ws+ !amount ingredient_coloned_word )* )               <IngredientNode>
     end
 
     # @see IngredientSimple#ingredient_simple_with_amount
@@ -33,8 +32,8 @@ module FoodIngredientParser::Strict::Grammar
     end
 
     # @see Common#word
-    rule ingredient_coloned_word_nas
-      abbrev / ( !'/' !andsep char )+
+    rule ingredient_coloned_word
+      abbrev / ( !'/' char )+
     end
 
   end

data/lib/food_ingredient_parser/strict/grammar/ingredient_simple.treetop

@@ -4,11 +4,8 @@ module FoodIngredientParser::Strict::Grammar
     include Amount
 
     rule ingredient_simple
-      name:( word_nas ( andsep? ws+ !amount !and word_nas )* ) ws? mark:mark <IngredientNode> /
-      name:( word_nas ( andsep? ws+ !amount !and word_nas )* )               <IngredientNode> /
-      # We've tried to omit 'and' from the ingredient, but if it doesn't work out, do it anyway.
-      name:( word_nas ( andsep? ws+ !amount      word_nas )* ) ws? mark:mark <IngredientNode> /
-      name:( word_nas ( andsep? ws+ !amount      word_nas )* )               <IngredientNode>
+      name:( word ( ws+ !amount word )* ) ws? mark:mark <IngredientNode> /
+      name:( word ( ws+ !amount word )* )               <IngredientNode>
     end
 
     rule ingredient_simple_with_amount

data/lib/food_ingredient_parser/strict/grammar/list_coloned.treetop

@@ -5,9 +5,9 @@ module FoodIngredientParser::Strict::Grammar
     include Ingredient
 
     rule list_coloned
-      contains:( ( list_coloned_ingredient ws* '.' ws* )+ list_coloned_ingredient? ) <ListNode> /
-      contains:( ( list_coloned_ingredient ws* ';' ws* )+ list_coloned_ingredient? ) <ListNode> /
-      contains:(   list_coloned_ingredient )                                         <ListNode>
+      contains:( ( ws* list_coloned_ingredient ws* '.' )+ list_coloned_ingredient? ) <ListNode> /
+      contains:( ( ws* list_coloned_ingredient ws* ';' )+ list_coloned_ingredient? ) <ListNode> /
+      contains:(   ws* list_coloned_ingredient )                                     <ListNode>
     end
 
     rule list_coloned_inner_list

data/lib/food_ingredient_parser/strict/grammar/root.treetop

@@ -11,7 +11,7 @@ module FoodIngredientParser::Strict::Grammar
       contains:( list_newlined / list_coloned / list )
       notes:(
         root_mark_sentences_in_list? ws*
-        ( ( '.' ws* newline* / '.'?  ws* newline+ ) ws* root_sentences? ws* )?
+        ( ( [.;] ws* newline* / [.;]?  ws* newline+ ) ws* root_sentences? ws* )?
       )
       '"'?
       <RootNode>

data/lib/food_ingredient_parser/version.rb

@@ -1,4 +1,4 @@
 module FoodIngredientParser
-  VERSION      = '1.0.0.pre.7'
-  VERSION_DATE = '2018-09-18'
+  VERSION      = '1.0.0.pre.8'
+  VERSION_DATE = '2018-09-19'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: food_ingredient_parser
 version: !ruby/object:Gem::Version
-  version: 1.0.0.pre.7
+  version: 1.0.0.pre.8
 platform: ruby
 authors:
 - wvengen
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2018-09-18 00:00:00.000000000 Z
+date: 2018-09-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: treetop