tsukasaoishi-kaerukeyword 1.0.1 → 1.0.2

data/History.txt

@@ -7,3 +7,7 @@
 
 * 1 minor enhancement:
   * move to GitHub from Rubyforge
+
+== 1.0.2 2009-09-27
+* 1 minor enhancement:
+  * bug fix.

data/Rakefile

@@ -4,7 +4,7 @@ require File.dirname(__FILE__) + '/lib/kaerukeyword'
 # Generate all the Rake tasks
 # Run 'rake -T' to see list of generated tasks (from gem root directory)
 $hoe = Hoe.new('kaerukeyword', Kaerukeyword::VERSION) do |p|
-  p.developer('FIXME full name', 'FIXME email')
+  p.developer('Tsukasa OISHI', 'tsukasa.oishi@gmail.com')
   p.changes              = p.paragraphs_of("History.txt", 0..1).join("\n\n")
   p.rubyforge_name       = p.name # TODO this is default value
   p.extra_dev_deps = [

data/lib/kaerukeyword.rb

@@ -33,47 +33,215 @@ $:.unshift(File.dirname(__FILE__)) unless
 # http://www.kaeruspoon.net
 #
 #
+require 'forwardable'
+
 class Kaerukeyword
-  VERSION = '1.0.1'
+  extend Forwardable
+
+  VERSION = '1.0.2'
+
+  # 委譲するメソッド
+  def_delegators :@logic, :add, :search, :match
 
-  def initialize(key_array = nil)
-    @tree = {}
-    key_array.each {|k| add(k)} if key_array.is_a?(Array)
+  #
+  # コンストラクタ
+  # 第2引数でデータ構造を選択できる
+  #
+  def initialize(key_array = nil, logic = :hash_tree)
+    @logic =
+      case logic
+      when :hash_tree
+        HashTree.new(key_array)
+      when :double_array
+        raise NotSupportError, "double array will support, but doesn't yet"
+        DoubleArray.new(key_array)
+      else
+        raise LogicError, "you have to select logic"
+      end
   end
 
-  def add(key)
-    now = @tree
-    key.split(//).each do |c|
-      now[c] = {} unless now.has_key? c
-      now = now[c]
-    end
-    now[:end] = key
+  #
+  # ハッシュ木からダブル配列への変換
+  #
+  def double_array!
+    raise NotSupportError, "double array will support, but doesn't yet"
+    raise LogicError, "already double array" if @logic.is_a?(DoubleArray)
+    old_logic = @logic
+    hash_tree = old_logic.tree
+
+    @logic = DoubleArray.new
+    @logic.construct_array(hash_tree, 0)
+
+    old_logic = nil
+    hash_tree = nil
+    GC.start
+    true
   end
+
   alias :<< :add
+  
+  #
+  # ダブル配列
+  #
+  class DoubleArray
+    def initialize(key_array = nil)
+      hash_tree = HashTree.new(key_array)
+      @base = [1]
+      @check = [0]
+      @code = {"end" => 1}
+      construct_array(hash_tree.tree, 0) if key_array.is_a?(Array)
+      hash_tree = nil
+      GC.start
+    end
+
+    #
+    # 未サポート
+    #
+    def add(key)
+      raise NotSupportError, "you don't add keyword after initilize."
+    end
 
-  def search(text)
-    list = []
-    word = ""
-    now = @tree
-    text.split(//).each do |c|
-      unless now.has_key? c
-        unless word.empty?
-          list << word
+    #
+    # 検索
+    #
+    def search(text)
+      list = []
+      word = ""
+      buffer = ""
+      now_x = 0
+
+      text.scan(/./).each do |c|
+        code = @code[c]
+        base = @base[now_x]
+        if code && @base[base + code] && @check[base + code] == now_x
+          now_x = base + code
+          buffer << c
+          if @base[@base[now_x] + @code["end"]].to_i < 0 && @check[@base[now_x] + @code["end"]] == now_x
+            word = buffer.dup
+          end
+        else
+          list << word unless word.empty?
           word = ""
+          buffer = ""
+          now_x = 0
         end
-        now = @tree
       end
 
-      if now.has_key? c
-        now = now[c]
-        word = now[:end] if now[:end]
+      list << word unless word.empty?
+      list
+    end
+
+    #
+    # ハッシュ木からBC配列を構築
+    #
+    def construct_array(hash_tree, now_x)
+      x_hash = {}
+      chars = hash_tree.keys.sort_by{|k| k.to_s}
+      code_set(chars)
+      @base[now_x] = base = get_base(chars, now_x)
+
+      chars.each do |c|
+        @check[base + @code[c]] = now_x
+        if c == "end"
+          @base[base + @code[c]] = -base
+        else
+          @base[base + @code[c]] = base
+          x_hash[c] = base + @code[c]
+        end
       end
+
+      x_hash.each {|c, x| construct_array(hash_tree[c], x)}
     end
 
-    unless word.empty?
-      list << word
+    private
+
+    #
+    # コード配列の定義
+    #
+    def code_set(chars)
+      chars.each {|c| @code[c] = @code.size + 1 unless @code.has_key?(c)}
     end
 
-    list
+    #
+    # 節のbase値の決定
+    #
+    def get_base(chars, now_x)
+      base = @base[now_x]
+      chars.each do |c|
+        if @check[base + @code[c]] && @check[base + @code[c]] != now_x
+          base += 1
+          retry
+        end
+      end
+      base
+    end
   end
+
+  class HashTree
+    END_TAG = "en"
+
+    attr_accessor :tree
+
+    def initialize(key_array = [])
+      @nodes = {}
+      make_tree(key_array)
+    end
+
+    def make_tree(key_array)
+      key_array.each{|key| add(key)}
+    end
+
+    #
+    # キーワードの追加
+    #
+    def add(key)
+      cur = @nodes
+      key.scan(/./).each do |c|
+        cur[c] ||= {}
+        cur = cur[c]
+      end
+      cur[END_TAG] = true
+    end
+
+    def search(text)
+      list = []
+      word = ""
+      buffer = ""
+      cur = @nodes
+      i = 0
+      hit_i = nil
+      str = text.scan(/./)
+
+      loop do
+        break unless c = str[i]
+
+        unless cur.has_key?(c)
+          if word.empty?
+            i = (hit_i || i) + 1
+          else
+            list << word
+            i = hit_i + word.scan(/./).size
+            word = ""
+          end
+          hit_i = nil
+          buffer = ""
+          cur = @nodes
+        end
+
+        if cur.has_key?(c)
+          cur = cur[c]
+          buffer << c
+          word = buffer.dup if cur.has_key?(END_TAG)
+          hit_i = i unless hit_i
+        end
+        i += 1
+      end
+
+      list << word unless word.empty?
+      list
+    end
+  end
+
+  class LogicError < StandardError; end
+  class NotSupportError < StandardError; end
 end

metadata

@@ -1,10 +1,10 @@
 --- !ruby/object:Gem::Specification 
 name: tsukasaoishi-kaerukeyword
 version: !ruby/object:Gem::Version 
-  version: 1.0.1
+  version: 1.0.2
 platform: ruby
 authors: 
-- FIXME full name
+- Tsukasa OISHI
 autorequire: 
 bindir: bin
 cert_chain: []
@@ -34,7 +34,7 @@ dependencies:
     version: 
 description: "KaeruKeyword \xE3\x81\xAF\xE7\x99\xBB\xE9\x8C\xB2\xE3\x81\x95\xE3\x82\x8C\xE3\x81\x9F\xE3\x82\xAD\xE3\x83\xBC\xE3\x83\xAF\xE3\x83\xBC\xE3\x83\x89\xE3\x81\x8C\xE6\x96\x87\xE4\xB8\xAD\xE3\x81\xAB\xE5\xAD\x98\xE5\x9C\xA8 \xE3\x81\x99\xE3\x82\x8B\xE3\x81\x8B\xE3\x81\xA9\xE3\x81\x86\xE3\x81\x8B\xE3\x82\x92\xE8\xAA\xBF\xE3\x81\xB9\xE3\x82\x8B\xE3\x81\x93\xE3\x81\xA8\xE3\x81\x8C\xE3\x81\xA7\xE3\x81\x8D\xE3\x81\xBE\xE3\x81\x99\xE3\x80\x82   \xE3\x82\xA4\xE3\x83\xB3\xE3\x82\xB9\xE3\x82\xBF\xE3\x83\xB3\xE3\x82\xB9\xE3\x81\xAE\xE4\xBD\x9C\xE6\x88\x90\xE6\x99\x82\xE3\x81\xAB\xE3\x82\xAD\xE3\x83\xBC\xE3\x83\xAF\xE3\x83\xBC\xE3\x83\x89\xE3\x82\x92\xE7\x99\xBB\xE9\x8C\xB2\xE3\x81\xA7\xE3\x81\x8D\xE3\x81\xBE\xE3\x81\x99\xE3\x80\x82  keywords = KaeruKeyword.new([\"Ruby\", \"Rails\"])   \xE3\x81\x82\xE3\x81\xA8\xE3\x81\x8B\xE3\x82\x89\xE3\x82\xAD\xE3\x83\xBC\xE3\x83\xAF\xE3\x83\xBC\xE3\x83\x89\xE3\x82\x92\xE8\xBF\xBD\xE5\x8A\xA0\xE3\x81\x99\xE3\x82\x8B\xE3\x81\x93\xE3\x81\xA8\xE3\x82\x82\xE3\x81\xA7\xE3\x81\x8D\xE3\x81\xBE\xE3\x81\x99\xE3\x80\x82  keywords << \"Tsukasa\"   search\xE3\x83\xA1\xE3\x82\xBD\xE3\x83\x83\xE3\x83\x89\xE3\x81\xA7\xE6\x96\x87\xE4\xB8\xAD\xE3\x81\x8B\xE3\x82\x89\xE3\x82\xAD\xE3\x83\xBC\xE3\x83\xAF\xE3\x83\xBC\xE3\x83\x89\xE3\x82\x92\xE6\x8E\xA2\xE3\x81\x97\xE5\x87\xBA\xE3\x81\x97\xE3\x81\xBE\xE3\x81\x99\xE3\x80\x82 \xE3\x82\xAD\xE3\x83\xBC\xE3\x83\xAF\xE3\x83\xBC\xE3\x83\x89\xE3\x81\x8C\xE8\xA6\x8B\xE3\x81\xA4\xE3\x81\x8B\xE3\x81\xA3\xE3\x81\x9F\xE3\x81\xA8\xE3\x81\x8D\xE3\x81\xAF\xE3\x80\x81\xE3\x81\x9D\xE3\x81\xAE\xE3\x82\xAD\xE3\x83\xBC\xE3\x83\xAF\xE3\x83\xBC\xE3\x83\x89\xE3\x82\x92 \xE5\x90\xAB\xE3\x82\x81\xE3\x81\x9F\xE9\x85\x8D\xE5\x88\x97\xE3\x82\x92\xE8\xBF\x94\xE3\x81\x97\xE3\x81\xBE\xE3\x81\x99\xE3\x80\x82  keywords.search(\"I Love Ruby\") #=> [\"Ruby\"]"
 email: 
-- FIXME email
+- tsukasa.oishi@gmail.com
 executables: []
 
 extensions: []
@@ -51,6 +51,7 @@ files:
 - lib/kaerukeyword.rb
 has_rdoc: true
 homepage: 
+licenses: 
 post_install_message: 
 rdoc_options: 
 - --main
@@ -72,7 +73,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
 requirements: []
 
 rubyforge_project: kaerukeyword
-rubygems_version: 1.2.0
+rubygems_version: 1.3.5
 signing_key: 
 specification_version: 2
 summary: "KaeruKeyword \xE3\x81\xAF\xE7\x99\xBB\xE9\x8C\xB2\xE3\x81\x95\xE3\x82\x8C\xE3\x81\x9F\xE3\x82\xAD\xE3\x83\xBC\xE3\x83\xAF\xE3\x83\xBC\xE3\x83\x89\xE3\x81\x8C\xE6\x96\x87\xE4\xB8\xAD\xE3\x81\xAB\xE5\xAD\x98\xE5\x9C\xA8 \xE3\x81\x99\xE3\x82\x8B\xE3\x81\x8B\xE3\x81\xA9\xE3\x81\x86\xE3\x81\x8B\xE3\x82\x92\xE8\xAA\xBF\xE3\x81\xB9\xE3\x82\x8B\xE3\x81\x93\xE3\x81\xA8\xE3\x81\x8C\xE3\x81\xA7\xE3\x81\x8D\xE3\x81\xBE\xE3\x81\x99\xE3\x80\x82   \xE3\x82\xA4\xE3\x83\xB3\xE3\x82\xB9\xE3\x82\xBF\xE3\x83\xB3\xE3\x82\xB9\xE3\x81\xAE\xE4\xBD\x9C\xE6\x88\x90\xE6\x99\x82\xE3\x81\xAB\xE3\x82\xAD\xE3\x83\xBC\xE3\x83\xAF\xE3\x83\xBC\xE3\x83\x89\xE3\x82\x92\xE7\x99\xBB\xE9\x8C\xB2\xE3\x81\xA7\xE3\x81\x8D\xE3\x81\xBE\xE3\x81\x99\xE3\x80\x82  keywords = KaeruKeyword.new([\"Ruby\", \"Rails\"])   \xE3\x81\x82\xE3\x81\xA8\xE3\x81\x8B\xE3\x82\x89\xE3\x82\xAD\xE3\x83\xBC\xE3\x83\xAF\xE3\x83\xBC\xE3\x83\x89\xE3\x82\x92\xE8\xBF\xBD\xE5\x8A\xA0\xE3\x81\x99\xE3\x82\x8B\xE3\x81\x93\xE3\x81\xA8\xE3\x82\x82\xE3\x81\xA7\xE3\x81\x8D\xE3\x81\xBE\xE3\x81\x99\xE3\x80\x82  keywords << \"Tsukasa\"   search\xE3\x83\xA1\xE3\x82\xBD\xE3\x83\x83\xE3\x83\x89\xE3\x81\xA7\xE6\x96\x87\xE4\xB8\xAD\xE3\x81\x8B\xE3\x82\x89\xE3\x82\xAD\xE3\x83\xBC\xE3\x83\xAF\xE3\x83\xBC\xE3\x83\x89\xE3\x82\x92\xE6\x8E\xA2\xE3\x81\x97\xE5\x87\xBA\xE3\x81\x97\xE3\x81\xBE\xE3\x81\x99\xE3\x80\x82 \xE3\x82\xAD\xE3\x83\xBC\xE3\x83\xAF\xE3\x83\xBC\xE3\x83\x89\xE3\x81\x8C\xE8\xA6\x8B\xE3\x81\xA4\xE3\x81\x8B\xE3\x81\xA3\xE3\x81\x9F\xE3\x81\xA8\xE3\x81\x8D\xE3\x81\xAF\xE3\x80\x81\xE3\x81\x9D\xE3\x81\xAE\xE3\x82\xAD\xE3\x83\xBC\xE3\x83\xAF\xE3\x83\xBC\xE3\x83\x89\xE3\x82\x92 \xE5\x90\xAB\xE3\x82\x81\xE3\x81\x9F\xE9\x85\x8D\xE5\x88\x97\xE3\x82\x92\xE8\xBF\x94\xE3\x81\x97\xE3\x81\xBE\xE3\x81\x99\xE3\x80\x82  keywords.search(\"I Love Ruby\") #=> [\"Ruby\"]"