wc 0.92.0 → 0.93.0

data/README.rdoc

@@ -18,6 +18,9 @@ wc gem is a very simple word counter. It takes a file in input and prints the wo
  $bin/wc --cloud filename
  It prints out the result using definition list HTML tag. This can be used to generate tag clouds
 
+ $bin/wc -H filename
+ It prints out the result excluding some very common English idioms such as ('and', 'of', 'the', ...)
+
 == Note on Patches/Pull Requests
  
 * Fork the project.

data/VERSION

@@ -1 +1 @@
-0.92.0
+0.93.0

data/bin/wc

@@ -10,12 +10,13 @@ opts = GetoptLong.new(
   [ '--version', '-v', GetoptLong::NO_ARGUMENT ], 
   [ '--json', '-j', GetoptLong::NO_ARGUMENT ],
   [ '--xml', '-x', GetoptLong::NO_ARGUMENT ], 
-  [ '--cloud', '-c', GetoptLong::NO_ARGUMENT]
+  [ '--cloud', '-c', GetoptLong::NO_ARGUMENT], 
+  [ '--hide-commons', '-H', GetoptLong::NO_ARGUMENT]
 )
 
 output = 'text'
 words = -1
-
+hide_commons = 0
 
 begin 
 opts.each do |opt, arg|
@@ -33,6 +34,8 @@ opts.each do |opt, arg|
       words = arg.to_i
     when '--cloud'
       output = 'cloud'
+    when '--hide-commons'
+      hide_commons = 1
   end
 end
 rescue 
@@ -45,7 +48,15 @@ if ARGV.length != 1
   exit 0
 end
 
-w = Wc.new(ARGV[0], words)
+if hide_commons 
+  hide_list = ['or', 'the', 'of', 'a', 'if', 'to', 'and', 'in', 'is', 'are']
+else
+  hide_list = []
+end
+
+w = Wc.new(ARGV[0], words, hide_list)
+
+
 
 if output == 'text'
   w.to_text

data/lib/wc.rb

@@ -1,13 +1,16 @@
 class Wc 
   attr_reader :filename, :occurrences, :words
+  attr_accessor :hide_list
   
-  def initialize(filename, words)
+  def initialize(filename, words, hide_list)
     @filename = filename
+    @hide_list = hide_list
     @occurrences = read
     @sorted = Array(occurrences).sort { |one, two| -(one[1] <=> two[1]) }
     @words = words
   end
   
+  
   def to_text
     if @words == -1
       @sorted.each { |elem|
@@ -51,7 +54,11 @@ class Wc
     File.open(@filename, "r") { |f|
           f.each_line { |line|
             words = line.split
-            words.each { |w| occurrences[w] += 1 }
+            words.each { |w|
+              if ! hide_list.include?(w.downcase)
+                occurrences[w.downcase] += 1 
+              end
+            }
           }
         }
     occurrences

data/wc.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = %q{wc}
-  s.version = "0.92.0"
+  s.version = "0.93.0"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Paolo Perego"]
-  s.date = %q{2010-07-01}
+  s.date = %q{2010-07-02}
   s.default_executable = %q{wc}
   s.description = %q{your ruby word counter experience}
   s.email = %q{thesp0nge@gmail.com}

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: wc
 version: !ruby/object:Gem::Version 
-  hash: 367
+  hash: 363
   prerelease: false
   segments: 
   - 0
-  - 92
+  - 93
   - 0
-  version: 0.92.0
+  version: 0.93.0
 platform: ruby
 authors: 
 - Paolo Perego
@@ -15,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2010-07-01 00:00:00 +02:00
+date: 2010-07-02 00:00:00 +02:00
 default_executable: wc
 dependencies: 
 - !ruby/object:Gem::Dependency