chawan 0.0.2 → 0.0.3

data/README

@@ -17,7 +17,13 @@ Methods
 Class
 =====
 
-  * Chawan::Node (Chawan.parse returns an array of Chawan::Node)
+  * Chawan::Nodes (Chawan.parse returns a Chawan::Nodes)
+      #noun    : scope category with noun
+      #verb    : scope category with verb
+      #grep    : scope category with given pattern
+      #compact : mix the category-consecutive nodes
+
+  * Chawan::Node (Chawan::Nodes has many Chawan::Node(s))
       #category   : part of speech
       #word       : text
       #attributes : keys and vals hash
@@ -26,11 +32,45 @@ Class
 Example
 =======
 
-  Chawan.parse('本日は晴天なり')
-  => [<名詞: '本日'>, <助詞: 'は'>, <名詞: '晴天'>, <助動詞: 'なり'>]
+  text = '本日はバリ晴天なりっす'
+
+  # 'parse' returns a Chawan::Nodes
+  Chawan.parse(text)
+  => [<名詞: '本日'>, <助詞: 'は'>, <名詞: 'バリ'>, <名詞: '晴天'>, <助動詞: 'なり'>, <助動詞: 'っす'>]
+
+  # Chawan::Nodes is enumerable
+  Chawan.parse(text).select{|node| node.category == '名詞'}
+  => [<名詞: '本日'>, <名詞: 'バリ'>, <名詞: '晴天'>]
+
+  # gateway interface: noun
+  Chawan.parse(text).noun
+  => [<名詞: '本日'>, <名詞: 'バリ'>, <名詞: '晴天'>]
+
+  # gateway interface: verb
+  Chawan.parse(text).verb
+  => [<助動詞: 'なり'>, <助動詞: 'っす'>]
+
+  # gateway interface: grep
+  Chawan.parse(text).grep(/助/)
+  => [<助詞: 'は'>, <助動詞: 'なり'>, <助動詞: 'っす'>]
+
+  # gateway interface: compact
+  Chawan.parse(text).compact
+  => [<名詞: '本日'>, <助詞: 'は'>, <名詞: 'バリ晴天'>, <助動詞: 'なりっす'>]
+  Chawan.parse(text).compact(/名詞/)
+  => [<名詞: '本日'>, <助詞: 'は'>, <名詞: 'バリ晴天'>, <助動詞: 'なり'>, <助動詞: 'っす'>]
+
+  # gateway interface is chainable
+  Chawan.parse(text).noun.verb
+  => []
 
-  Chawan.parse('本日は晴天なり').select{|node| node.category == '名詞'}.join
-  => "本日晴天"
+  # chainable is fun!
+  Chawan.parse(text).noun
+  => [<名詞: '本日'>, <名詞: 'バリ'>, <名詞: '晴天'>]
+  Chawan.parse(text).compact.noun
+  => [<名詞: '本日'>, <名詞: 'バリ晴天'>]
+  Chawan.parse(text).noun.compact
+  => [<名詞: '本日バリ晴天'>]
 
   
 Analyzer
@@ -64,7 +104,6 @@ Required
 Todo
 ====
 
-  * gateway interface to Chawan#parse such as grep, noun, ...
   * use open3 rather than backquote for executing unix commands
 
 

data/Rakefile

@@ -6,7 +6,7 @@ AUTHOR = "maiha"
 EMAIL = "maiha@wota.jp"
 HOMEPAGE = "http://github.com/maiha/chawan"
 SUMMARY = "A cup for chasen that provides an easy to use for extracting Japanese"
-GEM_VERSION = "0.0.2"
+GEM_VERSION = "0.0.3"
 
 spec = Gem::Specification.new do |s|
   s.rubyforge_project = 'asakusarb'

data/lib/chawan/analyzers/chasen_analyzer.rb

@@ -10,7 +10,8 @@ module Chawan
         lines = execute(text).split(/\n/)
         lines.pop == "EOS" or              # "EOS"
           raise CannotAnalyze
-        return lines.map{|line| instantiate(line, self.class::Fields)}
+        nodes = lines.map{|line| instantiate(line, self.class::Fields)}
+        return Nodes.new(nodes)
       end
 
       private

data/lib/chawan/node.rb

@@ -1,8 +1,20 @@
 module Chawan
   class Node
-    attr_reader :vals
+    attr_accessor :vals
     attr_reader :keys
 
+    module GatewayInterface
+      def noun?
+        category.to_s =~ /名詞/
+      end
+
+      def verb?
+        category.to_s =~ /動詞/
+      end
+    end
+
+    include GatewayInterface
+
     def initialize(vals, keys)
       @vals = vals
       @keys = keys

data/lib/chawan/nodes.rb

@@ -0,0 +1,72 @@
+module Chawan
+  class Nodes
+    attr_reader :nodes
+
+    def initialize(nodes)
+      @nodes = nodes
+    end
+
+    def ==(other)
+      self.class == other.class and map(&:to_s) == other.map(&:to_s)
+    end
+
+    ######################################################################
+    ### Enumerable
+
+    include Enumerable
+
+    def each(&block)
+      nodes.__send__(:each, &block)
+    end
+
+    def inspect
+      "[%s]" % map(&:inspect).join(', ')
+    end
+
+    ######################################################################
+    ### GatewayInterface
+
+    def grep(pattern, &block)
+      if block
+        nodes.__send__(regexp, &block)
+      else
+        Nodes.new(nodes.select{|n| pattern === n.category})
+      end
+    end
+
+    def compact(pattern = //)
+      array = []
+      prev = nil
+      each do |node|
+        if array.last and (pattern === node.category) and array.last.category == node.category
+          array.last.word << node.word.to_s
+        else
+          array << Node.new(node.vals.map(&:dup), node.keys)
+        end
+      end
+      self.class.new(array)
+    end
+
+    def self.gateway_interfaces
+      Chawan::Node::GatewayInterface.instance_methods.map(&:to_s).grep(/\?$/){$`}
+    end
+
+    def self.define_gateway_interface(name)
+      class_eval %{
+        def #{name}
+          Nodes.new(nodes.select{|n| n.__send__("#{name}?")})
+        end
+      }
+    end
+
+    private
+      def method_missing(name, *arguments, &block)
+        if self.class.gateway_interfaces.include?(name.to_s)
+          self.class.define_gateway_interface(name)
+          __send__(name)
+        else
+          nodes.__send__(name, *arguments, &block)
+        end
+      end
+  end
+end

data/lib/chawan.rb

@@ -7,6 +7,7 @@ end
 
 
 require File.dirname(__FILE__) + '/chawan/node'
+require File.dirname(__FILE__) + '/chawan/nodes'
 require File.dirname(__FILE__) + '/chawan/commands'
 require File.dirname(__FILE__) + '/chawan/analyzers'
 

data/spec/node_spec.rb

@@ -10,4 +10,6 @@ describe Chawan::Node do
   its('word')       {should == "Ruby"}
   its('category')   {should == "名詞"}
   its('inspect')    {should == "<名詞: 'Ruby'>"}
+
+  provide :vals=
 end

data/spec/nodes_spec.rb

@@ -0,0 +1,68 @@
+
+require File.join(File.dirname(__FILE__), 'spec_helper.rb')
+
+module Spec
+  module Example
+    module Subject
+      module ExampleGroupMethods
+        def gateway_interface(name, &block)
+          it "should provide gateway interface ##{name}" do
+            lambda { subject.__send__(name) }.should_not raise_error
+          end
+
+          it "should return an instance of Chawan::Nodes" do
+            subject.__send__(name).should be_a_kind_of(Chawan::Nodes)
+          end
+
+          its(name, &block)
+        end
+      end
+    end
+  end
+end
+
+
+describe Chawan::Nodes do
+  before do
+    keys = ["表記","品詞"]
+    @n1 = Chawan::Node.new(["本日","名詞"], keys)
+    @p1 = Chawan::Node.new(["は"  ,"助詞"], keys)
+    @n2 = Chawan::Node.new(["マジ","名詞"], keys)
+    @n3 = Chawan::Node.new(["晴天","名詞"], keys)
+    @v1 = Chawan::Node.new(["なり","助動詞"], keys)
+    @v2 = Chawan::Node.new(["っす","助動詞"], keys)
+    @nodes = Chawan::Nodes.new([@n1, @p1, @n2, @n3, @v1, @v2])
+
+    @n23 = Chawan::Node.new(["マジ晴天","名詞"], keys)
+    @v12 = Chawan::Node.new(["なりっす","助動詞"], keys)
+  end
+
+  subject { @nodes }
+
+  ######################################################################
+  ### Gateway Interface
+
+  gateway_interface(:noun) { should == Chawan::Nodes.new([@n1, @n2, @n3]) }
+  gateway_interface(:verb) { should == Chawan::Nodes.new([@v1, @v2]) }
+
+  gateway_interface(:compact) { should == Chawan::Nodes.new([@n1, @p1, @n23, @v12]) }
+  it "#compact(/名詞/) should compact only nodes matched /名詞/" do
+    subject.compact(/名詞/).should == Chawan::Nodes.new([@n1, @p1, @n23, @v1, @v2])
+  end
+
+  provide :grep
+  describe "#grep" do
+    it "should scope nodes by given category" do
+      subject.grep(/名詞/).should == Chawan::Nodes.new([@n1, @n2, @n3])
+      subject.grep('名詞').should == Chawan::Nodes.new([@n1, @n2, @n3])
+
+      subject.grep(/助/).should == Chawan::Nodes.new([@p1, @v1, @v2])
+      subject.grep('助').should == Chawan::Nodes.new([])
+    end
+  end
+
+  ######################################################################
+  ### Enumerable
+
+  provide :each
+end

data/spec/provide_helper.rb

@@ -0,0 +1,36 @@
+######################################################################
+### provide matcher
+Spec::Matchers.define :provide do |expected|
+  match do |obj|
+    (obj.public_methods + obj.protected_methods + obj.private_methods).include?(expected.to_s)
+  end
+end
+
+module Spec
+  module Example
+    module Subject
+      module ExampleGroupMethods
+        # == Examples
+        #
+        #   describe User do
+        #     subject { User.new }
+        #     provide :name
+        #
+        #       [intead of]
+        #
+        #     it "should provide #name" do
+        #       methods = subject.public_methods + subject.protected_methods + subject.private_methods
+        #       methods.should include("name")
+        #     end
+        #   end
+        #
+        def provide(name)
+          it "should provide ##{name}" do
+            subject.should provide(name)
+          end
+        end
+      end
+    end
+  end
+end
+

data/spec/spec_helper.rb

@@ -3,6 +3,7 @@ require 'spec'
 require 'rr'
 
 require File.join(File.dirname(__FILE__), '/../lib/chawan')
+require File.join(File.dirname(__FILE__), '/provide_helper')
 
 def data(key)
   path = File.join(File.dirname(__FILE__) + "/fixtures/#{key}")

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: chawan
 version: !ruby/object:Gem::Version 
-  version: 0.0.2
+  version: 0.0.3
 platform: ruby
 authors: 
 - maiha
@@ -9,7 +9,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2010-01-05 00:00:00 +09:00
+date: 2010-01-06 00:00:00 +09:00
 default_executable: 
 dependencies: []
 
@@ -33,10 +33,13 @@ files:
 - lib/chawan/analyzers/manager.rb
 - lib/chawan/node.rb
 - lib/chawan/commands.rb
+- lib/chawan/nodes.rb
 - lib/chawan.rb
 - spec/chasen_spec.rb
 - spec/api_spec.rb
+- spec/nodes_spec.rb
 - spec/mecab_spec.rb
+- spec/provide_helper.rb
 - spec/analyzer_spec.rb
 - spec/fixtures/example1.txt
 - spec/fixtures/example2.txt