lexical_units 0.0.1 → 0.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: f525e4f5314f36a83f70fe377e9fab926bc837ff
-  data.tar.gz: 166743d0dd6094a56146f2c0de941e573c151505
+  metadata.gz: 35f8950c1907e3f92afb2a3dc64ee3506779d87c
+  data.tar.gz: 3b08b4af5ada8ee6abaad85621b7f0cfd5850df7
 SHA512:
-  metadata.gz: 2005b725f605df1857f138d24d7febb224c887dfb5ba00626a735eba275a4073a68fb823ba12023699aad1ebbb8ce8452838d6bcb37521c03c277dd0eebb0f61
-  data.tar.gz: 9bffe65d097bd7580044f8f6cd8469663b29fb18f08b8a6da19f70afc7173afa61ba79aa5c6524bcf2d263df6b9d44b59acc1d050c7efe29fc9965d3aba74f06
+  metadata.gz: 7d610db7578359ee8724beea9f019b4b62e514ece052fa4d6d5b417af1069c58818387329c18b97145761522fc52ab613f20ea8e5d079b097baadad7102a83fe
+  data.tar.gz: eb0fd0fca6c27dbcb4169a97f4e18969f06f204b613866cee58a0365102603c131643b711bf9c42d86d06d29e50aceed173b0fee785081c073ed0b1db8dabaf9

data/CHANGELOG.md

@@ -0,0 +1,7 @@
+## v0.0.1
+
+* initial release
+
+## v0.0.2
+
+* added split into sentences

data/README.md

@@ -20,6 +20,14 @@ Or install it yourself as:
 
 ```ruby
 LexicalUnits::words(text)
+LexicalUnits::sentences(text)
+```
+
+You can include methods into String class:
+```ruby
+class String
+  include LexicalUnits::String
+end
 ```
 
 ## Contributing

data/lib/lexical_units.rb

@@ -1,4 +1,6 @@
 require "lexical_units/words"
+require "lexical_units/sentences"
+require "lexical_units/string"
 require "lexical_units/version"
 
 module LexicalUnits

data/lib/lexical_units/sentences.rb

@@ -0,0 +1,19 @@
+module LexicalUnits
+  # Split text into sentences
+  #
+  #   self.words("Lorem, ipsum. Dolor?") #=> ["Lorem, ipsum.", "Dolor?"]
+  #   self.words("Lorem! Ipsum dolor?") #=> ["Lorem!", "Ipsum dolor?"]
+  def self.sentences(text)
+    separators = LexicalUnits::sentence_separators
+    regexp = Regexp.new("[^#{separators}]+[#{separators}]{1,3}")
+    text.scan(regexp).map(&:strip)
+  end
+
+  private
+    def self.sentence_separators
+      [
+        '\.', '\?', '\!',
+        '‽'
+      ].join
+    end
+end

data/lib/lexical_units/string.rb

@@ -0,0 +1,11 @@
+module LexicalUnits
+  module String
+    def words
+      LexicalUnits::words(self)
+    end
+
+    def sentences
+      LexicalUnits::sentences(self)
+    end
+  end
+end

data/lib/lexical_units/version.rb

@@ -1,3 +1,3 @@
 module LexicalUnits
-  VERSION = "0.0.1"
+  VERSION = "0.0.2"
 end

data/spec/lexical_units/sentences_spec.rb

@@ -0,0 +1,55 @@
+#encoding: utf-8
+require 'spec_helper'
+
+describe LexicalUnits do
+  context ".sentences" do
+    let(:klass) { LexicalUnits }
+
+    it "splits text into sentences" do
+      text = %q{Lorem ipsum dolor sit amet, consectetur adipiscing elit.
+        Fusce ut lacinia lorem. Nullam a sem quam. Duis faucibus tortor in.}
+      array = [
+        "Lorem ipsum dolor sit amet, consectetur adipiscing elit.",
+        "Fusce ut lacinia lorem.",
+        "Nullam a sem quam.",
+        "Duis faucibus tortor in."
+      ]
+
+      klass::sentences(text).should eq(array)
+    end
+
+    it "splits text with question mark and exclamation mark into sentences" do
+      text = "Lorem ipsum dolor! Sit amet? Consectetur adipiscing elit."
+      array = [
+        "Lorem ipsum dolor!",
+        "Sit amet?",
+        "Consectetur adipiscing elit."
+      ]
+
+      klass::sentences(text).should eq(array)
+    end
+
+    it "splits text with ellipsis into sentences" do
+      text = "Lorem ipsum dolor, sit amet... Consectetur adipiscing elit."
+      array = [
+        "Lorem ipsum dolor, sit amet...",
+        "Consectetur adipiscing elit."
+      ]
+
+      klass::sentences(text).should eq(array)
+    end
+
+    it "splits text with interrobangs into sentences" do
+      text = "Say what‽ She's pregnant‽ Who is the father‽‽‽ Really?"
+      array = [
+        "Say what‽",
+        "She's pregnant‽",
+        "Who is the father‽‽‽",
+        "Really?"
+      ]
+
+      klass::sentences(text).should eq(array)
+    end
+
+  end
+end

data/spec/lexical_units/string_spec.rb

@@ -0,0 +1,25 @@
+require 'spec_helper'
+
+describe LexicalUnits::String do
+  class String
+    include LexicalUnits::String
+  end
+
+  context "#words" do
+    it "splits String into words" do
+      array = %w(Lorem ipsum dolor sit amet)
+      string = array.join(' ')
+
+      string.words.should eq(array)
+    end
+  end
+
+  context "#sentences" do
+    it "splits String into sentences" do
+      array = ["Lorem ipsum!", "Dolor sit?", "Amet."]
+      string = array.join
+
+      string.sentences.should eq(array)
+    end
+  end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: lexical_units
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.0.2
 platform: ruby
 authors:
 - Aleksander Malaszkiewicz
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-07-06 00:00:00.000000000 Z
+date: 2013-07-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -50,6 +50,7 @@ files:
 - .ruby-gemset
 - .ruby-version
 - .travis.yml
+- CHANGELOG.md
 - Gemfile
 - Guardfile
 - LICENSE.txt
@@ -57,8 +58,12 @@ files:
 - Rakefile
 - lexical_units.gemspec
 - lib/lexical_units.rb
+- lib/lexical_units/sentences.rb
+- lib/lexical_units/string.rb
 - lib/lexical_units/version.rb
 - lib/lexical_units/words.rb
+- spec/lexical_units/sentences_spec.rb
+- spec/lexical_units/string_spec.rb
 - spec/lexical_units/words_spec.rb
 - spec/spec_helper.rb
 homepage: ''
@@ -86,5 +91,7 @@ signing_key:
 specification_version: 4
 summary: Split text into lexical units
 test_files:
+- spec/lexical_units/sentences_spec.rb
+- spec/lexical_units/string_spec.rb
 - spec/lexical_units/words_spec.rb
 - spec/spec_helper.rb