ruby-jdict 0.0.7 → 0.0.8

data/lib/ruby-jdict/models/sense.rb

@@ -0,0 +1,81 @@
+# The sense element will record the translational equivalent
+# of the Japanese word, plus other related information. Where there
+# are several distinctly different meanings of the word, multiple
+# sense elements will be employed.
+module JDict
+  class Sense
+    PART_OF_SPEECH_RE = /^\[\[([^\]]+)\]\]\s+/
+
+    attr_reader :parts_of_speech, :glosses, :language
+    #
+    # Create a new +Sense+
+    def initialize(parts_of_speech, glosses, language)
+      @parts_of_speech, @glosses, @language = parts_of_speech, glosses, language
+    end
+
+    def to_s
+      parts_of_speech_to_s(@parts_of_speech) + glosses_to_s(@glosses)
+    end
+
+    def to_sql
+      str = ""
+      str << serialize_parts_of_speech
+      str << serialize_glosses
+      str
+    end
+
+    def self.from_sql(txt)
+      parts_of_speech = deserialize_parts_of_speech(txt)
+      glosses, language = deserialize_glosses(txt)
+
+      Sense.new(parts_of_speech, glosses, language)
+    end
+
+    private
+
+    def serialize_parts_of_speech
+      "[[#{@parts_of_speech.join(SerialConstants::PART_OF_SPEECH_SENTINEL)}]] "
+    end
+
+    # FIXME: it fails when retrieving entries from an existing index, because only one language is retrieved and the 'lang' field is nil
+    def serialize_glosses
+      @lang.to_s + SerialConstants::LANGUAGE_SENTINEL + @glosses.join(SerialConstants::MEANING_SENTINEL)
+    end
+
+
+    def self.deserialize_parts_of_speech(txt)
+      ary = txt.scan(PART_OF_SPEECH_RE)
+      if ary.size == 1
+        ary[0][0].split(SerialConstants::PART_OF_SPEECH_SENTINEL)
+      else
+        []
+      end
+    end
+
+    def self.deserialize_glosses(txt)
+      ary = txt.scan(PART_OF_SPEECH_RE)
+      str = if ary.size == 1
+        txt[(ary.to_s.length-1)..-1]
+      else
+        txt[5..-1]
+      end
+
+      str = str.force_encoding("UTF-8");
+
+      language, gloss_strings = str.split(SerialConstants::LANGUAGE_SENTINEL)
+      language = language.to_sym
+      glosses = gloss_strings.split(SerialConstants::MEANING_SENTINEL)
+
+      [glosses, language]
+    end
+
+
+    def glosses_to_s(glosses)
+      glosses.join('; ')
+    end
+
+    def parts_of_speech_to_s(parts_of_speech)
+      parts_of_speech.nil? ? '' : '[' + parts_of_speech.join(',') + '] '
+    end
+  end
+end

data/lib/ruby-jdict/version.rb

@@ -1,3 +1,3 @@
-module JDict
-  Version = '0.0.7'
-end
+module JDict
+  Version = '0.0.8'
+end

data/spec/convert_spec.rb

@@ -0,0 +1,27 @@
+# coding: utf-8
+require 'rubygems'
+
+require File.dirname(__FILE__) + '/spec_helper'
+require BASE_PATH + '/lib/ruby-jdict'
+
+describe JDict::Convert do
+  it "converts halfwidth katakana to hiragana" do
+    str = "ﾊﾋﾌﾍﾎｳｶｷｸｹｺｻｼｽｾｿﾀﾁﾂﾃﾄｱｲｴｵﾅﾆﾇﾈﾉﾏﾐﾑﾒﾓﾔﾕﾖﾗﾘﾙﾚﾛﾜｦﾝｧｨｩｪｫｬｭｮｯ"
+    Convert.kata_to_hira(str).should equal("はひふへほうかきくけこさしすせそたちつてとあいえおなにぬねのまみむめもやゆよらりるれろわをんぁぃぅぇぉゃゅょっ")
+  end
+
+  it "converts halfwidth katakana with nigori mark to hiragana" do
+    str = "ﾊﾞﾋﾞﾌﾞﾍﾞﾎﾞｳﾞｶﾞｷﾞｸﾞｹﾞｺﾞｻﾞｼﾞｽﾞｾﾞｿﾞﾀﾞﾁﾞﾂﾞﾃﾞﾄﾞ"
+    Convert.kata_to_hira(str).should equal("ばびぶべぼヴがぎぐげござじずぜぞだぢづでど")
+  end
+
+  it "converts halfwidth katakana with maru mark to hiragana" do
+    str = "ﾊﾟﾋﾟﾌﾟﾍﾟﾎﾟ"
+    Convert.kata_to_hira(str).should equal("ぱぴぷぺぽ")
+  end
+
+  it "converts invalid halfwidth katakana to hiragana" do
+    str = "ﾄﾟｧﾞﾟ"
+    Convert.kata_to_hira(str).should equal("とぁ")
+  end
+end

data/spec/dictionary_spec.rb

@@ -1,113 +1,113 @@
-require 'spec_helper'
-require BASE_PATH + '/lib/dictionary'
-#require BASE_PATH + '/lib/jmdict'
-
-module DictionarySpecHelper
-  JMDICT_PATH = File.join(BASE_PATH+'/dictionaries/JMdict')
-  INDEX_PATH  = File.join(BASE_PATH+'/index')
-  
-  def mock_index
-  end
-  
-  class Increase
-    def initialize(&measure_proc) # + args
-      @measure_proc = measure_proc
-    end
-
-    def matches?(target)
-      @target = target
-      @original_value = @measure_proc.call
-      target.call
-      @new_value = @measure_proc.call
-      return @new_value.to_i > @original_value.to_i
-    end
-
-    def failure_message
-      "expected #{@new_value} to be greater than #{@original_value}"
-    end
-
-    def negative_failure_message
-      "expected #{@new_value} to not be greater than #{@original_value}"
-    end
-
-    def description
-      "increase #{@original_value}"
-    end
-  end
-
-  def increase(&measure_proc) # + args
-    Increase.new(&measure_proc)
-  end
-end
-
-module DictionarySpec
-include DictionarySpecHelper
-
-describe JDict::Dictionary do
-  before do
-    @dictionary = JDict::Dictionary.new(INDEX_PATH)
-  end
-    
-  it "is searchable" do
-    @dictionary.should respond_to(:search)
-  end
-    
-  it "can tell you whether or not it's loaded" do
-    @dictionary.should respond_to(:loaded?)
-  end
-  
-  it "should generate fixtures" do
-    pending
-    @dictionary.should respond_to(:generate_fixtures)
-  end
-end
-
-describe JDict::Dictionary, "after initialization" do
-  before do
-    @dictionary = JDict::Dictionary.new(INDEX_PATH)
-  end
-  
-  it "has no entries" do
-    @dictionary.size.should == 0
-  end
-end
-
-describe JDict::Dictionary, "when loading from a dictionary file" do
-  before do
-    @dictionary = JDict::Dictionary.new(INDEX_PATH)
-  end
-  
-  it "has at least 1 entry" do
-    pending("implement loading from index")
-    @dictionary.load(JMDICT_PATH)
-    @dictionary.size.should > 0
-  end
-  
-  it "says it's loaded" do
-    pending("implement loading from index")
-    @dictionary.load(JMDICT_PATH)
-    # @dictionary.loaded?.should == true
-    @dictionary.loaded?.should equal(true)
-  end
-end
-
-describe JDict::Dictionary, "when loading from a dictionary file (already loaded)" do
-  before do
-    @dictionary = JDict::Dictionary.new(INDEX_PATH)
-  end
-  
-  it "has the same size as it did before being loaded"
-end
-
-describe JDict::Dictionary, "when searching" do
-  before do
-    @dictionary = JDict::Dictionary.new(INDEX_PATH)
-  end
-  
-  it "should raise an error if an index isn't built yet"
-  it "should give no results if the search phrase is empty" do
-    @dictionary.search('').should be_empty
-  end
-end
-
-end
+require 'spec_helper'
+require BASE_PATH + '/lib/ruby-jdict'
+#require BASE_PATH + '/lib/jmdict'
+
+module DictionarySpecHelper
+  JMDICT_PATH = File.join(BASE_PATH+'/dictionaries/JMdict')
+  INDEX_PATH  = File.join(BASE_PATH+'/index')
+  
+  def mock_index
+  end
+  
+  class Increase
+    def initialize(&measure_proc) # + args
+      @measure_proc = measure_proc
+    end
+
+    def matches?(target)
+      @target = target
+      @original_value = @measure_proc.call
+      target.call
+      @new_value = @measure_proc.call
+      return @new_value.to_i > @original_value.to_i
+    end
+
+    def failure_message
+      "expected #{@new_value} to be greater than #{@original_value}"
+    end
+
+    def negative_failure_message
+      "expected #{@new_value} to not be greater than #{@original_value}"
+    end
+
+    def description
+      "increase #{@original_value}"
+    end
+  end
+
+  def increase(&measure_proc) # + args
+    Increase.new(&measure_proc)
+  end
+end
+
+module DictionarySpec
+include DictionarySpecHelper
+
+describe JDict::Dictionary do
+  before do
+    @dictionary = JDict::Dictionary.new(INDEX_PATH)
+  end
+    
+  it "is searchable" do
+    @dictionary.should respond_to(:search)
+  end
+    
+  it "can tell you whether or not it's loaded" do
+    @dictionary.should respond_to(:loaded?)
+  end
+  
+  it "should generate fixtures" do
+    pending
+    @dictionary.should respond_to(:generate_fixtures)
+  end
+end
+
+describe JDict::Dictionary, "after initialization" do
+  before do
+    @dictionary = JDict::Dictionary.new(INDEX_PATH)
+  end
+  
+  it "has no entries" do
+    @dictionary.size.should == 0
+  end
+end
+
+describe JDict::Dictionary, "when loading from a dictionary file" do
+  before do
+    @dictionary = JDict::Dictionary.new(INDEX_PATH)
+  end
+  
+  it "has at least 1 entry" do
+    pending("implement loading from index")
+    @dictionary.load(JMDICT_PATH)
+    @dictionary.size.should > 0
+  end
+  
+  it "says it's loaded" do
+    pending("implement loading from index")
+    @dictionary.load(JMDICT_PATH)
+    # @dictionary.loaded?.should == true
+    @dictionary.loaded?.should equal(true)
+  end
+end
+
+describe JDict::Dictionary, "when loading from a dictionary file (already loaded)" do
+  before do
+    @dictionary = JDict::Dictionary.new(INDEX_PATH)
+  end
+  
+  it "has the same size as it did before being loaded"
+end
+
+describe JDict::Dictionary, "when searching" do
+  before do
+    @dictionary = JDict::Dictionary.new(INDEX_PATH)
+  end
+  
+  it "should raise an error if an index isn't built yet"
+  it "should give no results if the search phrase is empty" do
+    @dictionary.search('').should be_empty
+  end
+end
+
+end

data/spec/entry_spec.rb

@@ -0,0 +1,25 @@
+# coding: utf-8
+require 'rubygems'
+
+require File.dirname(__FILE__) + '/spec_helper'
+require BASE_PATH + '/lib/ruby-jdict'
+
+describe JDict::Entry do
+  it "is serializable" do
+    entry = JDict::Entry.new(1, ["感じ", "漢字"], ["かんじ", "カンジ"], [JDict::Sense.new(["&n;", "&vs;"], ["feeling", "kanji"])])
+    entry.to_sql.should equal("")
+  end
+
+  it "is deserializable" do
+    fail
+  end
+
+  it "is serializable with no parts of speech" do
+    entry = JDict::Entry.new(1, ["感じ", "漢字"], ["かんじ", "カンジ"], [JDict::Sense.new([], ["feeling", "kanji"])])
+    entry.to_sql.should equal("")
+  end
+
+  it "is deserializable with no parts of speech" do
+    fail
+  end
+end

data/spec/fixtures/feeds/sample_entry.xml

@@ -1,33 +1,33 @@
-<JMdict>
-<entry>
-<ent_seq>1171270</ent_seq>
-<k_ele>
-<keb>右翼</keb>
-<ke_pri>ichi1</ke_pri>
-<ke_pri>news1</ke_pri>
-<ke_pri>nf04</ke_pri>
-</k_ele>
-<r_ele>
-<reb>うよく</reb>
-<re_pri>ichi1</re_pri>
-<re_pri>news1</re_pri>
-<re_pri>nf04</re_pri>
-</r_ele>
-<sense>
-<pos>&n;</pos>
-<gloss>right-wing</gloss>
-<gloss g_lang="fr">aile droite (oiseau, armée, parti politique, base-ball)</gloss>
-<gloss g_lang="ru">пра́вое крыло́</gloss>
-<gloss g_lang="ru">пра́вый фланг</gloss>
-<gloss g_lang="de">rechter Flügel</gloss>
-</sense>
-<sense>
-<gloss g_lang="de">{Sport}</gloss>
-<gloss g_lang="de">rechte Flanke</gloss>
-<gloss g_lang="de">rechter Flügel</gloss>
-</sense>
-<sense>
-<gloss g_lang="de">die Rechte</gloss>
-</sense>
-</entry>
+<JMdict>
+<entry>
+<ent_seq>1171270</ent_seq>
+<k_ele>
+<keb>右翼</keb>
+<ke_pri>ichi1</ke_pri>
+<ke_pri>news1</ke_pri>
+<ke_pri>nf04</ke_pri>
+</k_ele>
+<r_ele>
+<reb>うよく</reb>
+<re_pri>ichi1</re_pri>
+<re_pri>news1</re_pri>
+<re_pri>nf04</re_pri>
+</r_ele>
+<sense>
+<pos>&n;</pos>
+<gloss>right-wing</gloss>
+<gloss g_lang="fr">aile droite (oiseau, armée, parti politique, base-ball)</gloss>
+<gloss g_lang="ru">пра́вое крыло́</gloss>
+<gloss g_lang="ru">пра́вый фланг</gloss>
+<gloss g_lang="de">rechter Flügel</gloss>
+</sense>
+<sense>
+<gloss g_lang="de">{Sport}</gloss>
+<gloss g_lang="de">rechte Flanke</gloss>
+<gloss g_lang="de">rechter Flügel</gloss>
+</sense>
+<sense>
+<gloss g_lang="de">die Rechte</gloss>
+</sense>
+</entry>
 </JMdict>

data/spec/index_spec.rb

@@ -1,84 +1,82 @@
-require 'rubygems'
-
-require File.dirname(__FILE__) + '/spec_helper'
-require BASE_PATH + '/lib/dictionary'
-require BASE_PATH + '/lib/jmdict'
-require BASE_PATH + '/lib/index'
-
-require 'fileutils'
-
-module IndexSpecHelper
-end
-
-describe JDict::DictIndex do
-  include IndexSpecHelper
-  
-  before do
-    @index = JDict::DictIndex.new(INDEX_PATH, JMDICT_PATH)
-  end
-  
-  # Searching
-  it "is searchable" do
-    @index.should respond_to(:search)
-  end
-  
-  # Building
-  it "is buildable" do
-    @index.should respond_to(:build) # and return an index
-  end
-  it "is rebuildable" do
-    @index.should respond_to(:rebuild)
-  end
-  it "tells whether it's built or not" do
-    @index.should respond_to(:built?)
-  end
-  
-  # Destroying
-  it "is destroyable" do
-    @index.should respond_to(:destroy)
-  end
-  
-  it "raises an error if an invalid dictionary path is specified" do
-    lambda { JDict::DictIndex.new(INDEX_PATH, 'bad_dictionary_path') }.should raise_error
-  end
-end
-
-describe JDict::DictIndex, "after initialization" do
-  it "the path should be set" do
-    @index = JDict::DictIndex.new(INDEX_PATH)
-    @index.path.should_not be(nil)
-    @index.path.should_not be('')
-  end
-end
-
-describe JDict::DictIndex, "when building," do
-  it "it is created on the file system" do
-    @index = JDict::DictIndex.new(INDEX_PATH, JMDICT_PATH)
-    @index.build
-    File.exists?(INDEX_PATH).should == true
-  end
-  
-  it "its directory on the file system shouldn't be empty" do
-    @index = JDict::DictIndex.new(INDEX_PATH, JMDICT_PATH,
-                           false) #no lazy loading
-    @index.build
-    # .
-    # ..
-    # ^-------- an empty directory has only these 2 entries
-    expect(Dir.entries(INDEX_PATH).size).to be >= 3
-  end
-  
-  it "loads from a dictionary file"
-end
-
-describe JDict::DictIndex, "when rebuilding" do
-  include FileUtils
-  
-  it "raises an error if it doesn't already exist" do
-    rm_rf(INDEX_PATH)
-    File.exists?(INDEX_PATH).should == false
-    lambda {
-      JDict::DictIndex.new(INDEX_PATH, JMDICT_PATH).rebuild
-      }.should raise_error
-  end
-end
+require 'rubygems'
+
+require File.dirname(__FILE__) + '/spec_helper'
+require BASE_PATH + '/lib/ruby-jdict'
+
+require 'fileutils'
+
+module IndexSpecHelper
+end
+
+describe JDict::DictIndex do
+  include IndexSpecHelper
+  
+  before do
+    @index = JDict::DictIndex.new(INDEX_PATH, JMDICT_PATH)
+  end
+  
+  # Searching
+  it "is searchable" do
+    @index.should respond_to(:search)
+  end
+  
+  # Building
+  it "is buildable" do
+    @index.should respond_to(:build) # and return an index
+  end
+  it "is rebuildable" do
+    @index.should respond_to(:rebuild)
+  end
+  it "tells whether it's built or not" do
+    @index.should respond_to(:built?)
+  end
+  
+  # Destroying
+  it "is destroyable" do
+    @index.should respond_to(:destroy)
+  end
+  
+  it "raises an error if an invalid dictionary path is specified" do
+    lambda { JDict::DictIndex.new(INDEX_PATH, 'bad_dictionary_path') }.should raise_error
+  end
+end
+
+describe JDict::DictIndex, "after initialization" do
+  it "the path should be set" do
+    @index = JDict::DictIndex.new(INDEX_PATH)
+    @index.path.should_not be(nil)
+    @index.path.should_not be('')
+  end
+end
+
+describe JDict::DictIndex, "when building," do
+  it "it is created on the file system" do
+    @index = JDict::DictIndex.new(INDEX_PATH, JMDICT_PATH)
+    @index.build
+    File.exists?(INDEX_PATH).should == true
+  end
+  
+  it "its directory on the file system shouldn't be empty" do
+    @index = JDict::DictIndex.new(INDEX_PATH, JMDICT_PATH,
+                           false) #no lazy loading
+    @index.build
+    # .
+    # ..
+    # ^-------- an empty directory has only these 2 entries
+    expect(Dir.entries(INDEX_PATH).size).to be >= 3
+  end
+  
+  it "loads from a dictionary file"
+end
+
+describe JDict::DictIndex, "when rebuilding" do
+  include FileUtils
+  
+  it "raises an error if it doesn't already exist" do
+    rm_rf(INDEX_PATH)
+    File.exists?(INDEX_PATH).should == false
+    lambda {
+      JDict::DictIndex.new(INDEX_PATH, JMDICT_PATH).rebuild
+      }.should raise_error
+  end
+end