NaiveText 0.2.0 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 08919bb72416392eb059f571a589335e5aae0dc2
-  data.tar.gz: cabef8b8d33a2ac2dd811494d02aa0a527d22a87
+  metadata.gz: ac7ce8cec1a92d0c067f4953bf28318afed99583
+  data.tar.gz: 47337502e94c528349eba4bb34e34bb81a9aa810
 SHA512:
-  metadata.gz: 6872002c6eae15ad91a59ea8ff952dda08a9d4152f99d245fbaa4b0080a847d54d28081120d055ddf30b3a4e956636bc7f81bfc8444aa7559166389fa5dc75a7
-  data.tar.gz: 9f693eff197c98c0a5008207eb2990a7b4209d660cf8e41f947fbb7972a3e1d95875f34903fee8f608bb685910846634994b08664cb6406bf7eedb129a0b5da4
+  metadata.gz: f7ab2fbec59e6dadb29ae98b10192e7ac5dd0ddd4e9dbdb0fe43ff289f4a354014170ec1eb55840b57a95dadfc12428880475ad517238cb0813d439c34c4cb2b
+  data.tar.gz: 72f0d51bf15c8dee374f876356f4b0af47f1af9b74d1056568adf0c8a8fbc0e45119b0f585c643e3a1913f757a4303cf3af16a01edf0096c0729769039d78fe4

data/lib/NaiveText/Categories.rb

@@ -13,6 +13,10 @@ class Categories
     @categories.each(&block)
   end
 
+  def total_word_count
+    @categories.inject(0) { |count, category | count + category.word_count  }
+  end
+
   private
 
    def calculate_apriori_propability_for(category)

data/lib/NaiveText/CategoriesFactory.rb

@@ -3,8 +3,13 @@ class CategoriesFactory
     categories = []
 
     config.each do |category_config|
-      examples = ExamplesGroup.new(category_config[:path])
-      categories << Category.new(name: category_config[:name], examples: examples)
+      begin
+        examples = ExamplesGroup.new(category_config[:path])
+        categories << Category.new(name: category_config[:name], examples: examples)
+      rescue
+        puts "You haven't provided trainingsdata for the category" + category_config[:name]
+        puts "This category was not created."
+      end
     end
 
     Categories.new(categories: categories)

data/lib/NaiveText/Category.rb

@@ -12,12 +12,20 @@ class Category
   end
 
   def p(word)
-    @examples.count(word).to_f / @examples.word_count
+    if(@examples.word_count>0)
+      @examples.count(word).to_f / @examples.word_count
+    else
+      0
+    end
   end
 
   def word_count
     @examples.word_count
   end
+
+  def to_s
+    @name
+  end
 end
 
 class NullCategory

data/lib/NaiveText/ExamplesGroup.rb

@@ -2,6 +2,9 @@ class ExamplesGroup
   def initialize(path)
     @text = load_text(path)
     @words = @text.split(/\W+/)
+    if @words.length == 0
+      raise 'Empty Trainingsdata'
+    end
   end
 
   def count(word)

data/lib/NaiveText/PropabilityCalculator.rb

@@ -4,24 +4,34 @@ class PropabilityCalculator
     @propabilities = PropabilityCollection.new(categories: @categories)
   end
 
-  def get_propabilities_for(words)
-    calculateProbabilities(words)
-    if(@propabilities.sum > 0)
-      normalize
-    end
+  def get_propabilities_for(text)
+    calculateProbabilities(text)
+    normalize unless @propabilities.sum < 0
     @propabilities
   end
 
+
   private
+    def minimum
+      minimum = 1.to_f/(10*@categories.total_word_count)
+    end
 
-    def calculateProbabilities(list_of_words)
+    def min_factor(factor)
+      if factor.to_f < minimum
+        factor = minimum
+      end
+      factor
+    end
+
+    def calculateProbabilities(text)
+      list_of_words = text.split(/\W+/)
       @categories.each do |category|
         @propabilities.set(category: category, value: p_apriori(category))
       end
 
       list_of_words.each do |word|
         @categories.each do |category|
-          @propabilities.multiply(category: category, factor: category.p(word) )
+          @propabilities.multiply(category: category, factor: min_factor(category.p(word)) )
         end
       end
     end
@@ -36,5 +46,4 @@ class PropabilityCalculator
     end
 
 
-
 end

data/lib/NaiveText/PropabilityCollection.rb

@@ -27,15 +27,30 @@ class PropabilityCollection
     end
   end
 
-  def max
+  def category_with_max
     id = @propabilities.find_index(@propabilities.max)
     @categories.find {|category| category.id == id}
   end
 
+  def max
+    @propabilities.max
+  end
+
   def sum
     @propabilities.reduce(:+)
   end
 
+  def to_s
+    result = ''
+    @categories.each do |category|
+      result << category.to_s
+      result << ':'
+      result << self.find(category).to_s
+      result << '\n'
+    end
+    result
+  end
+
   private
   def initialize_ids
     @ids =  @categories.map { |category| category.id }

data/lib/NaiveText/TextClassifier.rb

@@ -1,21 +1,21 @@
 class TextClassifier
+  attr_reader :categories
   def initialize( args )
     @categories = args[:categories]
     @calculator = args[:calculator] || PropabilityCalculator.new(categories: @categories)
   end
 
   def classify(text)
-    words = text.split(/\W+/)
-    get_category_for(words)
+    get_category_for(text)
   end
 
-  def get_category_for(list_of_words)
-    propabilities = @calculator.get_propabilities_for(list_of_words)
-    if(propabilities.sum == 0)
-      NullCategory.new
-    else
-      propabilities.max
-    end
+  def propabilities(text)
+    @calculator.get_propabilities(text)
   end
 
+private
+  def get_category_for(text)
+    propabilities = @calculator.get_propabilities_for(text)
+    propabilities.category_with_max
+  end
 end

data/lib/NaiveText/version.rb

@@ -1,3 +1,3 @@
 module NaiveText
-  VERSION = "0.2.0"
+  VERSION = "0.3.0"
 end

data/lib/NaiveText.rb

@@ -7,8 +7,10 @@ require "NaiveText/Category"
 require "NaiveText/Categories"
 require "NaiveText/CategoriesFactory"
 
+
+
 module NaiveText
-  
+
   def self.build(config)
     begin
       @categories = CategoriesFactory.build(config)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: NaiveText
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.3.0
 platform: ruby
 authors:
 - RicciFlowing
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2015-10-13 00:00:00.000000000 Z
+date: 2015-10-16 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler