inci_score 4.5.2 → 4.6.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9b62e565ea7191b030ebcdf3b278130065ecb325743bf4ca618e0c8e7d7559a8
-  data.tar.gz: 0fcbf56e6edb8a314c202200e89d8f53428b009472440d7a1e92815be90b9df9
+  metadata.gz: 948a01654e32fadaaa2cce158a59e84fc3f904a403d89fabfe18b3529600d0a8
+  data.tar.gz: 664c64f7dfb9003fcd1fa795d39c59157f1742d1e42653e504e9be2be47c196e
 SHA512:
-  metadata.gz: 7f4040838f05e804090a2482d3b6ea03c1c4575e5133a0989af9f733c373f6c1c507426196fca5eba539c4fbe5c3322a62e623a1630c8ade6dfe2f056a616938
-  data.tar.gz: c14b303bae2a60c0f16b1df28e95e77cf2b32caaa33086d8852edf8486dfccbaf29bb361a7d8c2150eb32a569890c84a812369f13bdd38f61df979a76f18abd2
+  metadata.gz: 7429c7077f89e8054782460772db0ed3f95f28a6ce686b2c060c3eceea0fa23e1169072727fca021f9c53a7e8afb55a28f87d7f1e7309292a48511a2fbcf895e
+  data.tar.gz: ca66118c3e79de310d0031fb2dfa17e91c62e772138798072c4c3eff0566e6e91e51a170c3d1f727511af3470455f1eefe8c064bbfda961cf7086b6c0f19c9dd

data/README.md

@@ -58,18 +58,21 @@ require "inci_score"
 
 inci = InciScore::Computer.new(src: 'aqua, dimethicone').call
 inci.score # 56.25
+inci.precision # 100.0
 ```
 
 As you see the results are wrapped by an *InciScore::Response* object, this is useful when dealing with the CLI and HTTP interfaces (read below).
 
 #### Unrecognized components
-The API treats unrecognized components as a common case by just marking the object as non valid and raise a warning in case more than 30% of the ingredients are not found.  
+The API treats unrecognized components as a common case by just marking the object as non valid.  
 In such case the score is computed anyway by considering only recognized components.  
-Is still possible to query the object for its state:
+You can check the `precision` value, which is zero for unrecognized components, and changes based on the applied recognizer rule (100% when exact matching).
 
 ```ruby
 inci = InciScore::Computer.new(src: 'ingredients:aqua,noent1,noent2')
 inci.valid? # false
+inci.score # 100.0
+inci.precision # 33.33
 inci.unrecognized # ["noent1", "noent2"]
 ```
 
@@ -82,7 +85,7 @@ inci_score --src="ingredients: aqua, dimethicone, pej-10, noent"
 TOTAL SCORE:
       	53.22
 PRECISION:
-      	75.0
+      	71.54
 COMPONENTS:
       	aqua (0), dimethicone (4), peg-10 (3)
 UNRECOGNIZED:

data/config/catalog.yml

@@ -1,6 +1,7 @@
 ---
 aqua: 0
 water: 0
+aroma: 2
 parfum: 3
 fragrance: 3
 phosphatidylcholine: 1
@@ -1580,6 +1581,7 @@ fagus sylvatica: 0
 farnesol: 2
 farnesyl acetate: 0
 fast green fcf: 4
+fava tonka: 1
 fermented vegetable extract: 0
 ferric chloride: 0
 ferula assa foetida: 0
@@ -2816,7 +2818,7 @@ pantothenic acid polypeptide: 0
 papaver orientale: 0
 papaver rhoeas: 0
 paraffin: 3
-paraffinum liquidum: 3
+paraffinum liquidum (mineral oil): 3
 parietaria officinalis: 0
 passiflora incarnata: 0
 passiflora quadrangularis: 0
@@ -3370,6 +3372,7 @@ pentasodium aminotrimethylene phosphonate: 3
 pentasodium pentetate: 3
 pentasodium triphosphate: 2
 pentetic acid: 3
+pentylene glycol: 2
 perfluorodecalin: 3
 perfluoropolymethylisopropyl ether: 3
 perfluorotetralin: 3
@@ -4832,6 +4835,8 @@ thymus extract: 4
 thymus hydrolysate: 4
 thymus serpyllum: 0
 thymus vulgaris: 0
+thymus vulgaris leaf oil: 0
+thymus vulgaris leaf water: 0
 tilia americana: 0
 tilia cordata: 0
 tilia vulgaris: 0

data/lib/inci_score/cli.rb

@@ -15,7 +15,7 @@ module InciScore
 
     def call
       parser.parse!(args)
-      return io.puts(%q{Specify inci list as: --src='aqua, parfum, etc'}) unless src
+      return io.puts(%q{Specify INCI list as: --src='aqua, parfum, etc'}) unless src
       computer = Computer.new(src: src)
       io.puts computer.call
     end

data/lib/inci_score/computer.rb

@@ -5,10 +5,11 @@ module InciScore
     TOLERANCE = 30.0
     DECIMALS = 2
 
-    attr_reader :src, :rules, :ingredients, :components, :unrecognized
+    attr_reader :src, :rules, :ingredients, :components, :unrecognized, :precisions
 
     def initialize(src:, rules: Normalizer::DEFAULT_RULES)
       @unrecognized = []
+      @precisions = []
       @src = src
       @rules = rules
       @ingredients = Normalizer.new(src: src, rules: rules).call
@@ -28,7 +29,7 @@ module InciScore
     end
 
     def precision
-      (100 - ((unrecognized.size / Float(ingredients.size)) * 100)).round(DECIMALS)
+      (precisions.sum / ingredients.size).round(DECIMALS)
     end
 
     def valid?
@@ -39,7 +40,9 @@ module InciScore
 
     def fetch_components
       ingredients.map do |ingredient|
-        Recognizer.new(ingredient).call.tap do |component|
+        recognizer = Recognizer.new(ingredient)
+        recognizer.call.tap do |component|
+          precisions << recognizer.precision
           unrecognized << ingredient unless component
         end
       end.compact

data/lib/inci_score/config.rb

@@ -5,7 +5,6 @@ require 'yaml'
 module InciScore
   module Config
     CATALOG = YAML::load_file(File::expand_path('../../../config/catalog.yml', __FILE__)).freeze
-    CIR = File.readlines(File::expand_path('../../../config/cir', __FILE__)).freeze
     HAZARDS = YAML::load_file(File::expand_path('../../../config/hazards.yml', __FILE__)).freeze
   end
 end

data/lib/inci_score/ingredient.rb

@@ -2,7 +2,6 @@
 
 module InciScore
   class Ingredient
-    SLASH = '/'
     SLASH_RULE = /(?<!ate)\//.freeze
     PARENTHESIS = %w[( ) [ ]].freeze
     PARENTHESIS_RULE = /(\(.+\)|\[.+\])/.freeze
@@ -11,24 +10,23 @@ module InciScore
 
     def initialize(raw)
       @raw = raw.to_s
-      @values = fetch_values
+      @values = fetch_values.uniq
       freeze
     end
 
-    def to_s
-      values.join(SLASH)
-    end
-
     private
 
     def fetch_values
-      if parenthesis?
-        parenthesis = PARENTHESIS.join
-        parenthesis_values = raw.match(PARENTHESIS_RULE).captures.map { |c| c.delete(parenthesis) }
-        deparenthesized = raw.sub(PARENTHESIS_RULE, '').sub(/\s{2,}/, ' ').strip
-        [deparenthesized].concat(parenthesis_values)
-      else
-        raw.split(SLASH_RULE).map(&:strip)
+      [raw].tap do |vals|
+        if parenthesis?
+          parenthesis = PARENTHESIS.join
+          parenthesis_values = raw.match(PARENTHESIS_RULE).captures.map { |c| c.delete(parenthesis) }
+          deparenthesized = raw.sub(PARENTHESIS_RULE, '').sub(/\s{2,}/, ' ').strip
+          vals << deparenthesized
+          vals.concat(parenthesis_values)
+        else
+          vals.concat(raw.split(SLASH_RULE).map(&:strip))
+        end
       end
     end
 

data/lib/inci_score/recognizer.rb

@@ -3,19 +3,31 @@
 module InciScore
   class Recognizer
     DEFAULT_RULES = [Rules::Key, Rules::Levenshtein, Rules::Hazard, Rules::Prefix, Rules::Tokens].freeze
+    PRECISION_BASE = 4
 
     attr_reader :ingredient, :rules, :applied
+    attr_accessor :found
 
     def initialize(ingredient, rules = DEFAULT_RULES)
       @ingredient = Ingredient.new(ingredient)
       @rules = rules
       @applied = []
-      freeze
+      @found = false
     end
 
     def call
       return if ingredient.to_s.empty?
-      find_component
+      find_component.tap do |c|
+        self.found = true if c
+      end
+    end
+
+    def precision
+      return 0.0 unless found
+      rule = applied.last
+      index = rules.index(rule) + PRECISION_BASE
+      ratio = Math.log(index, PRECISION_BASE)
+      (100 / ratio).round(2)
     end
 
     private

data/lib/inci_score/recognizer_rules.rb

@@ -28,24 +28,31 @@ module InciScore
         extend self
 
         Result = Struct.new(:name, :distance, :score) do
-          def tolerable?(size)
-            distance < TOLERANCE && distance <= (size-1)
+          def tolerable?(size, tolerance)
+            distance < tolerance && distance <= (size-1)
           end
         end
 
         def call(src)
           return if src.empty?
           size = src.size
+          t = tolerance(size)
           farthest = Result.new(nil, size)
           initial = src[0]
           result = Config::CATALOG.reduce(farthest) do |nearest, (name, score)|
             next nearest unless name.start_with?(initial)
-            next nearest if name.size > (size + TOLERANCE)
+            next nearest if name.size > (size + t)
             d = src.distance(name)
             nearest = Result.new(name, d, score) if d < nearest.distance
             nearest
           end
-          Component.new(result.name, result.score) if result.tolerable?(size)
+          Component.new(result.name, result.score) if result.tolerable?(size, t)
+        end
+
+        private
+
+        def tolerance(size)
+          Math.log(size, 1.8).round
         end
       end
 

data/lib/inci_score/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module InciScore
-  VERSION = '4.5.2'
+  VERSION = '4.6.2'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: inci_score
 version: !ruby/object:Gem::Version
-  version: 4.5.2
+  version: 4.6.2
 platform: ruby
 authors:
 - costajob
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2023-01-05 00:00:00.000000000 Z
+date: 2023-01-06 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: oj
@@ -106,7 +106,6 @@ files:
 - README.md
 - bin/inci_score
 - config/catalog.yml
-- config/cir
 - config/hazards.yml
 - ext/levenshtein.c
 - lib/inci_score.rb