food_ingredient_parser 1.1.0 → 1.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 629792af43b06c646b98c2ab8ee136895123e3e5
-  data.tar.gz: 6abbe88e79943cedb8b8776146aab6fdffd427e1
+  metadata.gz: 41da81db767bedd583db2794f44d39d710ddef58
+  data.tar.gz: 9bc24f8e83d04fb608e8fba4364aa26117fd898b
 SHA512:
-  metadata.gz: c1728c6edc995f6a5a18b82eae0e217fa18814c75f1b34d33d52abc4f593428020ba6a425ea526519036a5d4a412a80b43aae55147b2a07bae0e0979394b26bd
-  data.tar.gz: a9f43ad8e20ab867293e1db59aeb5b7bd6147a6e560adb58159f345d441c6445be453b7eb4ba01bdd4169ae951444e4bf90a6e7873cf85cfa7557b4ac4859d36
+  metadata.gz: 7634ebb4a8f3530cab515c5b27d2782df8c2521eb79df67ecbe08945360bcc27d68456d389d69a83808792145ecfb7094b3b97b5c33cbb17119642449813df88
+  data.tar.gz: 3b5c5ed78b7c4e41211855523acaf2a4c519b0e021a32eb698a715f7dccf2beacde229593d8422b71ce7a5a55dc1e266347d263c6629a72cdb308556697f7d93

data/lib/food_ingredient_parser/loose/scanner.rb

@@ -6,7 +6,7 @@ module FoodIngredientParser::Loose
     SEP_CHARS  = "|;,.".freeze
     MARK_CHARS = "¹²³⁴⁵ᵃᵇᶜᵈᵉᶠᵍªº⁽⁾†‡•°#^*".freeze
     PREFIX_RE  = /\A\s*(ingredients|contains|ingred[iï][eë]nt(en)?(declaratie)?|bevat|dit zit er\s?in|samenstelling|zutaten)\b\s*[:;.]?\s*/i.freeze
-    NOTE_RE    = /\A\b(dit product kan\b|deze verpakking kan\b|kan sporen\b.*?\bbevatten\b|voor allergenen\b|allergenen\b|E\s*=|gemaakt in\b|geproduceerd in\b|bevat mogelijk\b|kijk voor meer\b|allergie-info|in de fabriek\b|in dit bedrijf\b)/i.freeze
+    NOTE_RE    = /\A\b(dit product kan\b|deze verpakking kan\b|kan sporen\b.*?\bbevatten\b|voor allergenen\b|allergenen\b|E\s*=|gemaakt in\b|geproduceerd in\b|bevat mogelijk\b|kijk voor meer\b|allergie-info|in de fabriek\b|in dit bedrijf\b|voor [0-9,.]+ (g\.?|gr\.?|ram|ml).*\bis [0-9,.]+ (g\.?|gr\.?|ram|ml).*\bgebruikt\b)/i.freeze
     # Keep in sync with +abbrev+ in the +Common+ grammar, plus relevant ones from the +Amount+ grammar.
     ABBREV_RE  = Regexp.union(/\A(N°|°C|(ijzer|chroom|koper)\s*\(I+\)\s*[[:alnum:]]+)\b/i, *%w[
       a.o.p b.g.a b.o.b c.a c.i d.e d.m.v d.o.c d.o.p d.s e.a e.g e.u f.i.l f.o.s i.a
@@ -107,7 +107,10 @@ module FoodIngredientParser::Loose
     end
 
     def scan_iteration_notes
-      if is_sep?(chars: ".")    # dot means new note
+      if (len = abbrev_len) > 0 # defer iterations until after any abbreviation
+        cur # reference to record starting position
+        @i += len - 1
+      elsif is_sep?(chars: ".")    # dot means new note
         add_child
       else
         cur
@@ -197,7 +200,12 @@ module FoodIngredientParser::Loose
       cur.name ||= begin
         i, j = cur.interval.first, @i - 1
         i += mark_len(i) # skip any mark in front
-        Node.new(@s, i .. j) if j >= i
+        # Set name if there is any. There is one corner-case that needs to be avoided when
+        # a nesting was opened without a name, which would set the name to the nesting text.
+        # In this case, the name starts with an open-nesting symbol, which should never happen.
+        if j >= i && !"([:".include?(@s[i])
+          Node.new(@s, i .. j)
+        end
       end
     end
 

data/lib/food_ingredient_parser/version.rb

@@ -1,4 +1,4 @@
 module FoodIngredientParser
-  VERSION      = '1.1.0'
-  VERSION_DATE = '2018-09-24'
+  VERSION      = '1.1.1'
+  VERSION_DATE = '2018-09-25'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: food_ingredient_parser
 version: !ruby/object:Gem::Version
-  version: 1.1.0
+  version: 1.1.1
 platform: ruby
 authors:
 - wvengen
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2018-09-24 00:00:00.000000000 Z
+date: 2018-09-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: treetop