food_ingredient_parser 1.1.5 → 1.1.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
-SHA1:
-  metadata.gz: 8c80d036dbee183ed2fd1a8cc4e513e54318d142
-  data.tar.gz: 6cfcd29eacb9e99a9be9a66a90446f47f714ae4b
+SHA256:
+  metadata.gz: e65fb234a122c893fcb4de453e4cac07851d5000b848fa99d2b356002e187a60
+  data.tar.gz: c72d35c6fc978a6ddc139e3adf2059c24104221627f04857ba73aad159b29a0f
 SHA512:
-  metadata.gz: 75b1f91e5db6bcfcc24ad8eabe16b541663e5e344604c31b884ab1a5633c95dc045cdee2d384f5434f5143b778659cecd69917b9adb16e25341978380e486bcc
-  data.tar.gz: 7254cca971a558bda2ae6e996cc4d121fc5138f09d67c847bb44eec87421aeea39bf18a771137eb3b0b2bb734bed058a1d2347f815b9edc1b23e0b069d83a381
+  metadata.gz: 295728820e893f1277bb06cfdad572df1fde9d496df230caf55f41b23d124677131b004a5ca0b2a2e4a25b2273f98feb18840c47c6361ec474082d66d4fde7fb
+  data.tar.gz: 0ca48100273055b6c8f6e1d89bd1b44976e325cc9633080d6570e52cb6ef5d07dd419013d9b9aff67f95b6e4fac6ae12396c9248123daf4902d3eb97d84d1b28

data/bin/food_ingredient_parser

@@ -41,8 +41,10 @@ def parse_single(s, parsed=nil, parser:, verbosity: 1, print: nil, escape: false
   if parsed
     puts(parsed.inspect) if verbosity > 1
     pp(parsed.to_h, color: color) if verbosity > 0
+    return true
   else
     puts "(no result: #{parser.parser.failure_reason})" if verbosity > 0
+    return false
   end
 end
 
@@ -63,6 +65,7 @@ def parse_file(path, parser:, verbosity: 1, print: nil, escape: false, color: fa
   pct_parsed   = 100.0 * count_parsed / (count_parsed + count_noresult)
   pct_noresult = 100.0 * count_noresult / (count_parsed + count_noresult)
   puts "parsed #{colorize(color && "1;32", count_parsed)} (#{pct_parsed.round(1)}%), no result #{colorize(color && "1;31", count_noresult)} (#{pct_noresult.round(1)}%)"
+  return count_noresult
 end
 
 verbosity = 1
@@ -108,8 +111,10 @@ if strings.any? || files.any?
     STDERR.puts("Please specify one of the known parsers: #{PARSERS.keys.join(", ")}.")
     exit(1)
   end
-  strings.each {|s| parse_single(s, parser: parser, verbosity: verbosity, print: print, escape: escape, color: color) }
-  files.each   {|f| parse_file(f, parser: parser, verbosity: verbosity, print: print, escape: escape, color: color) }
+  success = true
+  strings.each {|s| success &= parse_single(s, parser: parser, verbosity: verbosity, print: print, escape: escape, color: color) }
+  files.each   {|f| success &= parse_file(f, parser: parser, verbosity: verbosity, print: print, escape: escape, color: color) == 0 }
+  success or exit(1)
 else
   STDERR.puts("Please specify one or more --file or --string arguments (see --help).")
 end

data/lib/food_ingredient_parser/cleaner.rb

@@ -4,6 +4,7 @@ module FoodIngredientParser
     def self.clean(s)
       s.gsub!("\u00ad", "")             # strip soft hyphen
       s.gsub!("\u0092", "'")            # windows-1252 apostrophe - https://stackoverflow.com/a/15564279/2866660
+      s.gsub!("‚", ",")                 # normalize unicode comma
       s.gsub!("aÄs", "aïs")             # encoding issue for maïs
       s.gsub!("ï", "ï")                # encoding issue
       s.gsub!("ë", "ë")                # encoding issue

data/lib/food_ingredient_parser/loose/scanner.rb

@@ -4,7 +4,7 @@ module FoodIngredientParser::Loose
   class Scanner
 
     SEP_CHARS  = "|;,.".freeze
-    MARK_CHARS = "¹²³⁴⁵ᵃᵇᶜᵈᵉᶠᵍªº⁽⁾†‡•°▪◊#^*".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|allergie[- ]informatie(\s*:|\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.
@@ -17,6 +17,7 @@ module FoodIngredientParser::Loose
         L\.\s+rhamnosus\b | L\.\s+acidophilus\b | L\.\s+casei\b | B\.\s+lactis | A\.\s+oryzae |
         S\.\s+thermophilus\b | L\.\sbulgaricus\b |
         T\.\s*aestivum\b(\s+vitt\.)? |
+        nucifera\s+L\. |
         type\s+"\d+" |
         E-e?\d{3}[a-z]?\s*\(i+\) |
         www\.[-_\/:%.A-Za-z0-9]+

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

@@ -12,15 +12,20 @@ module FoodIngredientParser::Strict::Grammar
     rule char
       !mark [[:alnum:]] /
       fraction /
-      [-/\`'´‘’+=_{}&] /
-      [®™] /
-      [¿?] /                         # weird characters turning up in names (e.g. encoding issues)
+      [-/\`'"´‘’+=_{}&] /
+      [®©™♣] /
+      [¿?¯] /                        # weird characters turning up in names (e.g. encoding issues)
       [₁₂₃₄₅₆₇₈₉]                    # can occur with vitamins
     end
 
     rule mark
       # mark referencing a footnote
-      [¹²³⁴⁵ᵃᵇᶜᵈᵉᶠᵍªº] '⁾'? / '⁽' [¹²³⁴⁵ᵃᵇᶜᵈᵉᶠᵍªº] '⁾' / [†‡•°▪◊#^] / '*'+ / '(' ws* ( [†‡•°▪◊#^] / '*'+ ) ws* ')'
+      [¹²³⁴⁵ᵃᵇᶜᵈᵉᶠᵍªº] '⁾'? /
+      '⁽' [¹²³⁴⁵ᵃᵇᶜᵈᵉᶠᵍªº] '⁾' /
+      [˄^] digit /
+      [†‡•°▪◊#˄^~˛] /
+      '*'+ /
+      '(' ws* ( [†‡•°▪◊#˄^~˛] / '*'+ ) ws* ')'
     end
 
     rule digit
@@ -131,6 +136,7 @@ module FoodIngredientParser::Strict::Grammar
         'L.' ws+ 'rhamnosus'i / 'L.' ws+ 'acidophilus'i / 'L.' ws+ 'casei' / 'B.'i ws+ 'lactis'i / 'A.'i ws+ 'oryzae'i /
         'S.' ws+ 'thermophilus'i / 'L.' ws+ 'bulgaricus'i /
         'T.' ws* 'aestivum'i (ws+ 'vitt.'i)? /
+        'nucifera' ws+ 'L.'i /
         'type'i ws+ '"' [0-9]+ '"' /
         e_number
       ) ![[:alpha:]]

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

@@ -40,7 +40,7 @@ module FoodIngredientParser::Strict::Grammar
     end
 
     rule root_mark_sentences_in_list
-      ( ( ws* [,.;] / ws ) ws* root_mark_sentence_in_list )+
+      ( ( ws* [,.;] / ws )+ root_mark_sentence_in_list )+
     end
 
     rule root_mark_sentence_in_list

data/lib/food_ingredient_parser/version.rb

@@ -1,4 +1,4 @@
 module FoodIngredientParser
-  VERSION      = '1.1.5'
-  VERSION_DATE = '2019-11-14'
+  VERSION      = '1.1.6'
+  VERSION_DATE = '2020-11-19'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: food_ingredient_parser
 version: !ruby/object:Gem::Version
-  version: 1.1.5
+  version: 1.1.6
 platform: ruby
 authors:
 - wvengen
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-11-14 00:00:00.000000000 Z
+date: 2020-11-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: treetop
@@ -87,8 +87,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.6.13
+rubygems_version: 3.0.3
 signing_key: 
 specification_version: 4
 summary: Parser for ingredient lists found on food products.