highscore 0.4.2 → 0.4.3

data/.travis.yml

@@ -0,0 +1,10 @@
+language: ruby
+rvm:
+  - 1.8.7
+  - 1.9.2
+  - 1.9.3
+  - jruby-18mode # JRuby in 1.8 mode
+  - jruby-19mode # JRuby in 1.9 mode
+  - rbx-18mode
+  # - rbx-19mode # currently in active development, may or may not work for your project
+before_install: gem install bones

data/History.txt

@@ -1,3 +1,8 @@
+== 0.4.3 / 2012-02-11
+
+* rate words based on the amount of vowels and consonants
+* added a CLI tool that uses all of the API (via parameters)
+
 == 0.4.2 / 2012-02-07
 
 * added a simple example that also shows return values in README.md

data/README.md

@@ -7,9 +7,11 @@ Features
 --------
 
 * configurable to rank different types of words different (uppercase, long words, etc.)
+* rate based on amount (%) of vowels and consonants in a string
 * directly get keywords from String objects
 * blacklist words via a plain text file, String or an Array of words
 * merge together Keywords from multiple sources
+* contains a CLI tool that operates on STDIN/OUT and is configurable via parameters
 
 Examples
 --------
@@ -21,11 +23,10 @@ text.configure do
   set :upper_case, 3
   set :long_words, 2
   set :long_words_threshold, 15
+  set :vowels, 1 # => default = 0 = not considered
+  set :consonants, 5 # => default = 0 = not considered
 end
 
-# get all keywords
-text.keywords.rank => Array
-
 # get only the top 50 keywords
 text.keywords.top(50).each do |keyword|
   keyword.text   # => keyword text
@@ -44,7 +45,7 @@ can iterate over. Each object in there is a Highscore::Keyword that has methods
 ```ruby
 keywords = "Foo bar is not bar baz".keywords(Highscore::Blacklist.load(['baz']))
 
-keywords.each do |k|
+keywords.rank.each do |k|
   puts "#{k.text} #{k.weight}"
 end
 
@@ -54,6 +55,8 @@ end
 # not 1.0
 ```
 
+Have a look at bin/highscore, you can run highscore on your CLI and feed it with text on STDIN.
+
 Using a custom blacklist to ignore keywords
 -------------------------------------------
 

data/bin/highscore

@@ -1,7 +1,83 @@
 #!/usr/bin/env ruby
 
+require 'optparse'
+
 root = File.expand_path('../..', __FILE__)
 require File.join(root, %w[lib highscore])
 
-# Put your code here
+options = {
+    :emphasis => {},
+    :short => false
+}
+
+optparse = OptionParser.new do |opts|
+  opts.on('-h', '--help', 'display this help message') do
+    puts opts
+    exit
+  end
+
+  # blacklist file
+  opts.on('-b', '--blacklist FILEPATH', 'specify a blacklist file that should be used instead of the default') do |filepath|
+    options[:blacklist] = Highscore::Blacklist.load_file(filepath)
+  end
+
+  # general multiplier
+  opts.on('-m', '--multiplier MULTIPLIER', 'set the global rank multiplier') do |multiplier|
+    options[:emphasis][:multiplier] = multiplier.to_f
+  end
+
+  # don't print rank weight
+  opts.on('-s', '--short', 'don\'t print rank weight') do
+    options[:short] = true
+  end
+
+  # upper case
+  opts.on('-u', '--uppercase MULTIPLIER', 'rank multiplier for uppercase words (default is 3.0)') do |u|
+    options[:emphasis][:upper_case] = u.to_f
+  end
+
+  # vowels
+  opts.on('--vowels MULTIPLIER', 'multiplier for vowels') do |u|
+    options[:emphasis][:vowels] = u.to_i
+  end
+
+  # vowels
+  opts.on('--consonants MULTIPLIER', 'multiplier for consonants') do |u|
+    options[:emphasis][:consonants] = u.to_i
+  end
+
+  # long words
+  opts.on('-l', '--longwords MULTIPLIER', 'rank multiplier for long words (default is 2.0)') do |u|
+    options[:emphasis][:long_words] = u.to_f
+  end
+
+  # long words threshold
+  opts.on('-t', '--longwordsthreshold THRESHOLD', 'threshold for long words (default is 15 chars)') do |u|
+    options[:emphasis][:long_words_threshold] = u.to_i
+  end
+
+  opts.on('-n', '--top N', 'show only the top N keywords') do |u|
+    options[:top] = u.to_i
+  end
+end
+
+optparse.parse!
+
+text = STDIN.read.to_s
+keywords = text.keywords(options[:blacklist]) do |content|
+  options[:emphasis].each do |key,value|
+    content.set key, value
+  end
+end
+
+if options[:top]
+  keywords = keywords.top(options[:top])
+else
+  keywords = keywords.rank
+end
 
+keywords.each do |keyword|
+  print keyword.text
+  print " #{keyword.weight}" unless options[:short]
+  puts
+end

data/lib/highscore/content.rb

@@ -18,7 +18,9 @@ module Highscore
         :multiplier => 1.0,
         :upper_case => 3.0,
         :long_words => 2.0,
-        :long_words_threshold => 15
+        :long_words_threshold => 15,
+        :vowels => 0,
+        :consonants => 0
       }
     end
 
@@ -43,21 +45,41 @@ module Highscore
       keywords = Keywords.new
 
       Keywords.find_keywords(@content, @blacklist).each do |text|
-        weight = @emphasis[:multiplier]
         text = text.to_s
+        keywords << Highscore::Keyword.new(text, weight(text))
+      end
+
+      keywords
+    end
 
-        if text.length >= @emphasis[:long_words_threshold]
-          weight *= @emphasis[:long_words]
-        end
+    private
 
-        if text[0,1] == text[0,1].upcase
-          weight *= @emphasis[:upper_case]
-        end
+    # weight a single text keyword
+    #
+    def weight(text)
+      weight = @emphasis[:multiplier]
 
-        keywords << Highscore::Keyword.new(text, weight)
+      if text.length >= @emphasis[:long_words_threshold]
+        weight *= @emphasis[:long_words]
       end
 
-      keywords
+      if text[0,1] == text[0,1].upcase
+        weight *= @emphasis[:upper_case]
+      end
+
+      weight += vowels(text)
+      weight += consonants(text)
+      weight
+    end
+
+    def vowels(text)
+      percent = text.vowels.length / text.length.to_f
+      percent * @emphasis[:vowels]
+    end
+
+    def consonants(text)
+      percent = text.consonants.length / text.length.to_f
+      percent * @emphasis[:consonants]
     end
   end
 end

data/lib/highscore/keyword.rb

@@ -15,5 +15,10 @@ module Highscore
     def <=>(other)
       other.weight <=> @weight
     end
+
+    # get the string
+    def to_s
+      @text
+    end
   end
 end

data/lib/highscore/string.rb

@@ -15,4 +15,14 @@ class String
 
     content.keywords
   end
+
+  # get all vowels from a string
+  def vowels
+    gsub(/[^aeiou]/, '')
+  end
+
+  # get all consonants from a string
+  def consonants
+    gsub(/[aeiou]/, '')
+  end
 end

data/test/highscore/test_blacklist.rb

@@ -65,7 +65,7 @@ class TestBlacklist < Test::Unit::TestCase
     keywords = "Foo bar is not bar baz".keywords(Highscore::Blacklist.load(['baz']))
 
     keyword_list = []
-    keywords.each do |k|
+    keywords.rank.each do |k|
       keyword_list << k.text
     end
 

data/test/highscore/test_content.rb

@@ -30,4 +30,16 @@ class TestContent < Test::Unit::TestCase
     content = Highscore::Content.new content
     assert_equal 1, content.keywords.length
   end
+
+  def test_vowels_and_consonants
+    keywords = 'foobar RubyGems'.keywords do
+      set :vowels, 2
+      set :consonants, 3
+      set :upper_case, 1
+      set :long_words, 1
+    end
+
+    assert_equal 3.75, keywords.first.weight
+    assert_equal 3.5, keywords.last.weight
+  end
 end

data/test/highscore/test_keyword.rb

@@ -22,6 +22,10 @@ class TestKeyword < Test::Unit::TestCase
     assert_equal 'Foobar', @keyword.text
   end
 
+  def test_to_s
+    assert_equal 'Ruby', @keyword.to_s
+  end
+
   def test_weight
     assert_equal 2, @keyword.weight
 

data/test/highscore/test_string.rb

@@ -13,6 +13,14 @@ class TestString < Test::Unit::TestCase
     assert_equal 0, keywords.length
   end
 
+  def test_vowels
+    assert_equal("eoaiu", "feobariu".vowels)
+  end
+
+  def test_consonants
+    assert_equal("fbr", "feobariu".consonants)
+  end
+
   def test_default_blacklist
     keywords = "the Ruby Ruby Ruby Hacker".keywords
     assert_equal 2, keywords.length

data/version.txt

@@ -1 +1 @@
-0.4.2
+0.4.3

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: highscore
 version: !ruby/object:Gem::Version
-  version: 0.4.2
+  version: 0.4.3
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-02-07 00:00:00.000000000 Z
+date: 2012-02-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bones
-  requirement: &70117716241580 !ruby/object:Gem::Requirement
+  requirement: &70111402827500 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,7 +21,7 @@ dependencies:
         version: 3.7.3
   type: :development
   prerelease: false
-  version_requirements: *70117716241580
+  version_requirements: *70111402827500
 description: Find and rank keywords in long texts.
 email: liebler.dominik@googlemail.com
 executables:
@@ -34,6 +34,7 @@ extra_rdoc_files:
 - test/fixtures/blacklist.txt
 files:
 - .gitignore
+- .travis.yml
 - History.txt
 - README.md
 - Rakefile