natto 0.9.5 → 0.9.6

data/lib/natto/struct.rb

@@ -0,0 +1,310 @@
+# coding: utf-8
+require 'natto/binding'
+require 'natto/option_parse'
+
+module Natto 
+  require 'ffi'
+
+  # `MeCabStruct` is a general base class for `FFI::Struct` objects in
+  # the `Natto` module. Please refer to
+  # [`mecab.h`](http://code.google.com/p/mecab/source/browse/trunk/mecab/src/mecab.h)
+  class MeCabStruct < FFI::Struct
+    # Provides accessor methods for the members of the `mecab` struct.
+    #
+    # @param [String] attr_name
+    # @return member values for the `mecab` struct
+    # @raise [NoMethodError] if `attr_name` is not a member of this `mecab` struct 
+    def method_missing(attr_name)
+      member_sym = attr_name.id2name.to_sym
+      return self[member_sym] if self.members.include?(member_sym)
+      raise(NoMethodError.new("undefined method '#{attr_name}' for #{self}"))
+    end
+  end
+
+  # `DictionaryInfo` is a wrapper for `struct mecab_dictionary_info_t`
+  # that holds the `MeCab` instance's related dictionary information.
+  # 
+  # Values for the `mecab` dictionary attributes may be 
+  # obtained by using the following `Symbol`s as keys 
+  # to the layout associative array of `FFI::Struct` members.
+  #
+  # - :filename
+  # - :charset
+  # - :size
+  # - :type
+  # - :lsize
+  # - :rsize
+  # - :version
+  # - :next
+  # 
+  # ## Usage
+  # `mecab` dictionary attributes can be obtained by
+  # using their corresponding accessor.
+  #
+  #     nm = Natto::MeCab.new
+  #
+  #     sysdic = nm.dicts.first
+  #
+  #     puts sysdic.filename
+  #     => "/usr/local/lib/mecab/dic/ipadic/sys.dic"
+  #
+  #     puts sysdic.charset
+  #     => "utf8"
+  # 
+  #     puts sysdic.is_sysdic?
+  #     => true
+  class DictionaryInfo < MeCabStruct
+    # System dictionary.
+    SYS_DIC = 0
+    # User dictionary.
+    USR_DIC = 1
+    # Unknown dictionary.
+    UNK_DIC = 2
+
+    layout  :filename, :string,
+            :charset,  :string,
+            :size,     :uint,
+            :type,     :int,
+            :lsize,    :uint,
+            :rsize,    :uint,
+            :version,  :ushort,
+            :next,     :pointer
+   
+    if Object.respond_to?(:type) && Object.respond_to?(:class)
+      alias_method :deprecated_type, :type
+      # `Object#type` override defined when both `type` and
+      # `class` are Object methods. This is a hack to avoid the 
+      # `Object#type` deprecation warning thrown up in Ruby 1.8.7
+      # and in JRuby.
+      #
+      # @return [Fixnum] `mecab` dictionary type
+      def type
+        self[:type]
+      end
+    end
+
+    # Returns human-readable details for this `mecab` dictionary.
+    # Overrides `Object#to_s`.
+    #
+    # - encoded object id
+    # - dictionary type
+    # - full-path dictionary filename
+    # - dictionary charset
+    #
+    # @return [String] encoded object id, type, dictionary filename, and charset
+    def to_s
+      %(#{super.chop} type="#{self.type}", filename="#{self.filename}", charset="#{self.charset}">)
+    end
+    
+    # Overrides `Object#inspect`.
+    #
+    # @return [String] encoded object id, dictionary filename, and charset
+    # @see #to_s
+    def inspect
+      self.to_s
+    end
+
+    # Returns `true` if this is a system dictionary.
+    # @return [Boolean]
+    def is_sysdic?
+      self.type == SYS_DIC
+    end
+
+    # Returns `true` if this is a user dictionary.
+    # @return [Boolean]
+    def is_usrdic?
+      self.type == USR_DIC
+    end
+
+    # Returns `true` if this is a unknown dictionary type.
+    # @return [Boolean]
+    def is_unkdic?
+      self.type == UNK_DIC
+    end
+  end
+
+  # `MeCabNode` is a wrapper for the structure holding
+  # the parsed `node`.
+  # 
+  # Values for the `mecab` node attributes may be 
+  # obtained by using the following `Symbol`s as keys 
+  # to the layout associative array of `FFI::Struct` members.
+  #
+  # - :prev
+  # - :next
+  # - :enext
+  # - :bnext
+  # - :rpath
+  # - :lpath
+  # - :surface
+  # - :feature
+  # - :id
+  # - :length
+  # - :rlength
+  # - :rcAttr
+  # - :lcAttr
+  # - :posid
+  # - :char_type
+  # - :stat
+  # - :isbest
+  # - :alpha
+  # - :beta
+  # - :prob
+  # - :wcost
+  # - :cost
+  #
+  # ## Usage
+  # An instance of `MeCabNode` is yielded to the block
+  # used with `MeCab#parse`, where the above-mentioned
+  # node attributes may be accessed by name.
+  #
+  #     nm = Natto::MeCab.new
+  #
+  #     nm.parse('卓球なんて死ぬまでの暇つぶしだよ。') do |n| 
+  #       puts "#{n.surface}\t#{n.cost}" if n.is_nor? 
+  #     end
+  #     卓球     2874
+  #     な       4398
+  #     死ぬ     9261
+  #     まで     9386
+  #     の       10007
+  #     暇つぶし 13324
+  #     だ       15346
+  #     よ       14396
+  #     。       10194
+  #
+  # It is also possible to use the `Symbol` for the
+  # `mecab` node member to index into the 
+  # `FFI::Struct` layout associative array like so:
+  #     
+  #     nm.parse('あいつ笑うと結構可愛い顔してんよ。') {|n| puts n[:feature] }
+  #     名詞,代名詞,一般,*,*,*,あいつ,アイツ,アイツ
+  #     動詞,自立,*,*,五段・ワ行促音便,基本形,笑う,ワラウ,ワラウ
+  #     助詞,接続助詞,*,*,*,*,と,ト,ト
+  #     副詞,一般,*,*,*,*,結構,ケッコウ,ケッコー
+  #     形容詞,自立,*,*,形容詞・イ段,基本形,可愛い,カワイイ,カワイイ
+  #     名詞,一般,*,*,*,*,顔,カオ,カオ
+  #     動詞,自立,*,*,サ変・スル,連用形,する,シ,シ
+  #     動詞,非自立,*,*,一段,体言接続特殊,てる,テン,テン
+  #     助詞,終助詞,*,*,*,*,よ,ヨ,ヨ
+  #     記号,句点,*,*,*,*,。,。,。
+  #     BOS/EOS,*,*,*,*,*,*,*,*
+  #
+  class MeCabNode < MeCabStruct
+    attr_accessor :surface, :feature
+    attr_reader   :pointer
+
+    # Normal `mecab` node defined in the dictionary.
+    NOR_NODE = 0
+    # Unknown `mecab` node not defined in the dictionary.
+    UNK_NODE = 1
+    # Virtual node representing the beginning of the sentence.
+    BOS_NODE = 2
+    # Virutual node representing the end of the sentence.
+    EOS_NODE = 3
+    # Virtual node representing the end of an N-Best `mecab` node list.
+    EON_NODE = 4
+
+    layout  :prev,            :pointer,
+            :next,            :pointer,
+            :enext,           :pointer,
+            :bnext,           :pointer,
+            :rpath,           :pointer,
+            :lpath,           :pointer,
+            :surface,         :string,
+            :feature,         :string,
+            :id,              :uint,
+            :length,          :ushort,
+            :rlength,         :ushort,
+            :rcAttr,          :ushort,
+            :lcAttr,          :ushort,
+            :posid,           :ushort,
+            :char_type,       :uchar,
+            :stat,            :uchar,
+            :isbest,          :uchar,
+            :alpha,           :float,
+            :beta,            :float,
+            :prob,            :float,
+            :wcost,           :short,
+            :cost,            :long
+   
+    #if RUBY_VERSION.to_f < 1.9
+    #  alias_method :deprecated_id, :id
+    #  # `Object#id` override defined when `RUBY_VERSION` is
+    #  # older than 1.9. This is a hack to avoid the `Object#id`
+    #  # deprecation warning thrown up in Ruby 1.8.7.
+    #  #
+    #  # <i>This method override is not defined when the Ruby interpreter
+    #  # is 1.9 or greater.</i>
+    #  # @return [Fixnum] `mecab` node id
+    #  def id
+    #    self[:id]
+    #  end
+    #end
+
+    # Initializes this node instance.
+    # Sets the `MeCab` feature value for this node.
+    #
+    # @param [FFI::Pointer]
+    def initialize(ptr)
+      super(ptr)
+      @pointer = ptr
+
+      if self[:feature]
+        @feature = self[:feature].force_encoding(Encoding.default_external)
+      end
+    end
+     
+    # Returns human-readable details for the `mecab` node.
+    # Overrides `Object#to_s`.
+    #
+    # - encoded object id
+    # - underlying FFI pointer to MeCab Node 
+    # - stat (node type: NOR, UNK, BOS/EOS, EON)
+    # - surface 
+    # - feature
+    #
+    # @return [String] encoded object id, underlying FFI pointer, stat, surface, and feature 
+    def to_s
+      %(#{super.chop} @pointer=#{@pointer}, stat=#{self[:stat]}, @surface="#{self.surface}", @feature="#{self.feature}">)
+    end
+
+    # Overrides `Object#inspect`.
+    #
+    # @return [String] encoded object id, stat, surface, and feature 
+    # @see #to_s
+    def inspect
+      self.to_s
+    end
+
+    # Returns `true` if this is a normal `mecab` node found in the dictionary.
+    # @return [Boolean]
+    def is_nor?
+      self.stat == NOR_NODE
+    end
+
+    # Returns `true` if this is an unknown `mecab` node not found in the dictionary.
+    # @return [Boolean]
+    def is_unk?
+      self.stat == UNK_NODE
+    end
+   
+    # Returns `true` if this is a virtual `mecab` node representing the beginning of the sentence.
+    # @return [Boolean]
+    def is_bos?
+      self.stat == BOS_NODE
+    end
+   
+    # Returns `true` if this is a virtual `mecab` node representing the end of the sentence.
+    # @return [Boolean]
+    def is_eos?
+      self.stat == EOS_NODE 
+    end
+   
+    # Returns `true` if this is a virtual `mecab` node representing the end of the node list.
+    # @return [Boolean]
+    def is_eon?
+      self.stat == EON_NODE
+    end
+  end
+end

data/lib/natto/version.rb

@@ -1,31 +1,31 @@
 # coding: utf-8
 
-# <tt>Natto</tt> is the namespace for objects that provide
-# a binding to the <tt>mecab</tt> tagger and related resources.
+# `Natto` is the namespace for objects that provide
+# a binding to the `mecab` tagger and related resources.
 #
-# <tt>Natto::MeCab</tt> is a wrapper class for the <tt>mecab</tt> 
+# `Natto::MeCab` is a wrapper class for the `mecab` 
 # tagger.
 #
-# <tt>Natto::MeCabStruct</tt> is a base class for a <tt>mecab</tt>
+# `Natto::MeCabStruct` is a base class for a `mecab`
 # struct.
 #
-# <tt>Natto::MeCabNode</tt> is a wrapper for the struct representing
-# a <tt>mecab</tt>-parsed node.
+# `Natto::MeCabNode` is a wrapper for the struct representing
+# a `mecab`-parsed node.
 #
-# <tt>Natto::DictionaryInfo</tt> is a wrapper for the struct 
-# representing a <tt>Natto::MeCab</tt> instance's related 
+# `Natto::DictionaryInfo` is a wrapper for the struct 
+# representing a `Natto::MeCab` instance's related 
 # dictionary information.
 #
-# <tt>Natto::MeCabError</tt> is a general error class for the 
-# <tt>Natto</tt> module.
+# `Natto::MeCabError` is a general error class for the 
+# `Natto` module.
 #
-# Module <tt>Natto::Binding</tt> encapsulates methods and behavior 
-# which are made available via <tt>FFI</tt> bindings to <tt>mecab</tt>.
+# Module `Natto::Binding` encapsulates methods and behavior 
+# which are made available via `FFI` bindings to `mecab`.
 #
-# Module <tt>OptionParse</tt> encapsulates methods and behavior 
-# for parsing the various <tt>mecab</tt> options supported by
-# <tt>Natto</tt>.
+# Module `OptionParse` encapsulates methods and behavior 
+# for parsing the various `mecab` options supported by
+# `Natto`.
 module Natto
   # Version string for this Rubygem.
-  VERSION = "0.9.5"
+  VERSION = "0.9.6"
 end

metadata

@@ -1,34 +1,31 @@
 --- !ruby/object:Gem::Specification
 name: natto
 version: !ruby/object:Gem::Version
-  prerelease:
-  version: 0.9.5
+  version: 0.9.6
 platform: ruby
 authors:
 - Brooke M. Fujita
-autorequire:
+autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-09-16 00:00:00.000000000Z
+date: 2013-07-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ffi
-  version_requirements: &2056 !ruby/object:Gem::Requirement
+  requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
-        version: 0.6.3
-    none: false
-  requirement: *2056
-  prerelease: false
+        version: 1.9.0
   type: :runtime
-description: ! 'natto is a gem bridging Ruby and MeCab using FFI (foreign function
-  interface). No compilation is necessary, and natto will run on CRuby (mri/yarv)
-  and JRuby (jvm) equally well, on any OS.
-
-  ruby ? mecab ????????? natto ???????????
-
-  '
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: 1.9.0
+description: |
+  natto bridges Ruby and MeCab via FFI (foreign function interface). No compiling is necessary, and natto will run on CRuby (mri/yarv) and JRuby (jvm) equally well, on any OS. natto provides the most natural, Ruby-esque API for MeCab.
 email: buruzaemon@gmail.com
 executables: []
 extensions: []
@@ -36,39 +33,39 @@ extra_rdoc_files: []
 files:
 - lib/natto.rb
 - lib/natto/binding.rb
+- lib/natto/natto.rb
 - lib/natto/option_parse.rb
-- lib/natto/utils.rb
+- lib/natto/struct.rb
 - lib/natto/version.rb
 - README.md
 - LICENSE
 - CHANGELOG
 - .yardopts
-homepage: https://bitbucket.org/buruzaemon/natto/overview
+homepage: https://bitbucket.org/buruzaemon/natto
 licenses:
 - BSD
-post_install_message:
+metadata: {}
+post_install_message: 
 rdoc_options: []
 require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ! '>='
+  - - '>='
     - !ruby/object:Gem::Version
-      version: 1.8.7
-  none: false
+      version: '1.9'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ! '>='
+  - - '>='
     - !ruby/object:Gem::Version
       version: '0'
-  none: false
 requirements:
-- MeCab, 0.994 or greater
-- FFI, 0.6.3 or greater
-rubyforge_project:
-rubygems_version: 1.8.15
-signing_key:
-specification_version: 3
-summary: natto combines the Ruby programming language with MeCab, the part-of-speech and morphological analyzer for the Japanese language.
+- MeCab, 0.996 or greater
+- FFI, 1.9.0 or greater
+rubyforge_project: 
+rubygems_version: 2.0.0
+signing_key: 
+specification_version: 4
+summary: natto combines the Ruby programming language with MeCab, the part-of-speech
+  and morphological analyzer for the Japanese language.
 test_files: []
-...