stopwords-filter 0.2.1 → 0.3.0

data/CHANGELOG

@@ -1,3 +1,6 @@
+0.3.0:
+ * Added custom words list to snowball filter [@sbeckeriv]
+ * Fixed problem about initialization: https://github.com/brenes/stopwords-filter/issues/3 [@zackxu1]
 0.2.0:
  * Added stopword? method [@s2gatev]
  * Added Sieve class [@s2gatev]

data/VERSION

@@ -1 +1 @@
-0.2.0
+0.2.1

data/lib/stopwords/filter.rb

@@ -1,17 +1,21 @@
-class Stopwords::Filter
+module Stopwords
 
-  attr_reader :stopwords
+  class Filter
 
-  def initialize stopwords
-    @stopwords = stopwords
-  end
+    attr_reader :stopwords
 
-  def filter words
-    words - @stopwords
-  end
+    def initialize stopwords
+      @stopwords = stopwords
+    end
+
+    def filter words
+      words - @stopwords
+    end
+
+    def stopword? word
+      stopwords.include? word
+    end
 
-  def stopword? word
-    stopwords.include? word
   end
 
 end

data/lib/stopwords/snowball/filter.rb

@@ -1,19 +1,16 @@
-class Stopwords::Snowball::Filter < Stopwords::Filter
-
-
-  attr_reader :locale
-  attr_reader :locale_filename
-
-  def initialize locale
-
-    @locale = locale
-    @locale_filename = "#{File.dirname(__FILE__)}/locales/#{locale}.csv"
-
-    raise "Unknown locale" unless File.exists?(@locale_filename)
-
-    super File.read(@locale_filename).split(",")
-
+module Stopwords
+  module Snowball
+    class Filter < Stopwords::Filter
+      attr_reader :locale
+      attr_reader :locale_filename
+
+      def initialize locale, custom_list = []
+        @locale = locale
+        @locale_filename = "#{File.dirname(__FILE__)}/locales/#{locale}.csv"
+
+        raise "Unknown locale" unless File.exists?(@locale_filename)
+        super File.read(@locale_filename).split(",") + custom_list 
+      end
+    end
   end
-
-
-end
+end

data/lib/stopwords/snowball/wordsieve.rb

@@ -1,16 +1,20 @@
-class Stopwords::Snowball::WordSieve
-  def initialize
-    @filters = Dir[File.dirname(__FILE__) + '/locales/*.csv'].each_with_object({}) do |file, filters|
-      lang = File.basename(file, '.csv').to_sym
-      filters[lang] = Stopwords::Snowball::Filter.new lang
-    end
-  end
+module Stopwords
+  module Snowball
+    class Stopwords::Snowball::WordSieve
+      def initialize custom_list = []
+        @filters = Dir[File.dirname(__FILE__) + '/locales/*.csv'].each_with_object({}) do |file, filters|
+          lang = File.basename(file, '.csv').to_sym
+          filters[lang] = Stopwords::Snowball::Filter.new lang, custom_list
+        end
+      end
 
-  def stopword? args={}
-    args[:lang] ? @filters[args[:lang]].stopword?(args[:word] ) : false
-  end
+      def stopword? args={}
+        args[:lang] ? @filters[args[:lang]].stopword?(args[:word] ) : false
+      end
 
-  def filter args={}
-    args[:lang] ? @filters[args[:lang]].filter(args[:words] ) : args[:words]
+      def filter args={}
+        args[:lang] ? @filters[args[:lang]].filter(args[:words] ) : args[:words]
+      end
+    end
   end
-end
+end

data/spec/lib/snowball_filter_spec.rb

@@ -15,5 +15,16 @@ describe Stopwords::Snowball::Filter do
 
   end
 
+  context "when custom list" do
 
-end
+    let (:filter) { Stopwords::Snowball::Filter.new "es", ["Santurce"] }
+
+    subject { filter }
+
+    it("should remove the stopwords for the list of words to be filtered") { filter.filter("desde Santurce a Bilbao".split).should == ["Bilbao"]}
+
+  end
+
+
+
+end

metadata

@@ -2,7 +2,7 @@
 name: stopwords-filter
 version: !ruby/object:Gem::Version 
   prerelease: 
-  version: 0.2.1
+  version: 0.3.0
 platform: ruby
 authors: 
 - David J. Brenes