filerary 0.0.3 → 0.0.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 344e1d5e01114b63eccf5dee74a935be4abc0eed
-  data.tar.gz: cb20cf7d3c31349043ea7c6b873ed4faedcd0335
+  metadata.gz: 5dada2e80e774e32a089e8a3797dbacb0275b4a6
+  data.tar.gz: ac63271136d1de8303d5a63d7cc6e8d69a9720f6
 SHA512:
-  metadata.gz: 76f9a2dc1f5708f972ea54205c69e890bcaaa25e42af276f952434b0e71ab8126a910c85ebd3c802586cd325bfa60d9c1fa96bcb3b6368b7f11dc2381caa36fc
-  data.tar.gz: 5d05e68e85760d7c2a5b2b99d7f01eea25e592021397ce706224cbbeaa7317539eff9c48b5dfa1f57b7fc14dca67831d806d37f48eb26520eeae92de8055e095
+  metadata.gz: 104bfc5442ee05c60abbc20112c5e09a5644058a55c3dc471c497e7c3c1a12dfb0ec9a042e688a23a97fbbb932718308e52ce10bdc033302d08b4aafbf5842cd
+  data.tar.gz: 9d66a9d851a1856a07e3d64af1561669fa3b6d70f1641a05b0cb13d8229248678ca9048b27db89e88ee9ce330d0b2efdf31c4447502a1e4e7bb8b1990e52695a

data/lib/filerary/librarian.rb

@@ -10,9 +10,14 @@ module URI
   class Generic
     alias :__path__ :path
     def path
-      URI.decode(__path__)
+      URI.decode_www_form_component(__path__, Encoding.find("locale"))
     end
   end
+
+  def self.parse(uri)
+    uri = URI.encode_www_form_component(uri)
+    DEFAULT_PARSER.parse(uri)
+  end
 end
 
 module Filerary
@@ -50,7 +55,9 @@ module Filerary
         record.content =~ word
       end
 
-      result.collect {|record| record._key }
+      result.collect do |record|
+        record._key.force_encoding(Encoding.find("locale"))
+      end
     end
 
     def cleanup
@@ -73,21 +80,17 @@ module Filerary
       extractor = ChupaText::Extractor.new
       extractor.apply_configuration(ChupaText::Configuration.default)
 
-      begin
-        # TODO: I'll send pull request to ChupaText.
-        extractor.extract(URI.encode(path)) do |text_data|
-          text = text_data.body
-        end
-      rescue URI::InvalidURIError
-        return path
+      extractor.extract(path) do |text_data|
+        text = text_data.body
       end
 
       return path unless text
 
       # TODO: I want to specify encoding in ChupaText side.
-      text.force_encoding("UTF-8")
-      return text if text.valid_encoding?
       text.force_encoding(Encoding.default_external)
+      text.force_encoding("UTF-8") unless text.valid_encoding?
+
+      text
     end
   end
 end

data/lib/filerary/version.rb

@@ -1,3 +1,3 @@
 module Filerary
-  VERSION = "0.0.3"
+  VERSION = "0.0.4"
 end

data/test/fixtures//343/203/236/343/203/253/343/203/201/343/203/220/343/202/244/343/203/210.txt

@@ -0,0 +1 @@
+UTF-8のファイル名

data/test/test-librarian.rb

@@ -26,22 +26,33 @@ class FileraryTest < Test::Unit::TestCase
   end
 
   class CollectTest < self
-    def test_argument_is_string
-      assert_equal([__FILE__], @librarian.collect(__FILE__))
+    class ArgumentTest < self
+      def test_string
+        assert_equal([__FILE__], @librarian.collect(__FILE__))
+      end
+
+      def test_array
+        assert_equal([__FILE__], @librarian.collect([__FILE__]))
+      end
     end
 
-    def test_argument_is_array
-      assert_equal([__FILE__], @librarian.collect([__FILE__]))
+    class FilePathTest < self
+      def test_multibyte
+        path = File.join(@test_fixtures_dir, "マルチバイト.txt")
+        assert_equal([path], @librarian.collect(path))
+      end
     end
 
-    def test_flle_type_is_pdf
-      path = File.join(@test_fixtures_dir, "test-pdf.pdf")
-      assert_equal([path], @librarian.collect(path))
-    end
+    class FileTypeTest < self
+      def test_pdf
+        path = File.join(@test_fixtures_dir, "test-pdf.pdf")
+        assert_equal([path], @librarian.collect(path))
+      end
 
-    def test_flle_type_is_xls
-      path = File.join(@test_fixtures_dir, "test-excel.xls")
-      assert_equal([path], @librarian.collect(path))
+      def test_xls
+        path = File.join(@test_fixtures_dir, "test-excel.xls")
+        assert_equal([path], @librarian.collect(path))
+      end
     end
   end
 
@@ -56,26 +67,28 @@ class FileraryTest < Test::Unit::TestCase
       assert_equal([], @librarian.search("AAA" * 5))
     end
 
-    def test_file_type_is_pdf
-      path = File.join(@test_fixtures_dir, "test-pdf.pdf")
-      @librarian.collect(path)
-      assert_equal([path], @librarian.search("秋"))
-      assert_equal([], @librarian.search("冬"))
-    end
-
-    def test_file_type_is_xls
-      path = File.join(@test_fixtures_dir, "test-excel.xls")
-      @librarian.collect(path)
-      assert_equal([path], @librarian.search("Excel"))
-    end
-
-    def test_file_type_is_xls_of_multibyte
-      # TODO: I want to detective.
-      omit_on_travis_ci
-      path = File.join(@test_fixtures_dir, "test-excel.xls")
-      @librarian.collect(path)
-      assert_equal([path], @librarian.search("表計算ソフト"))
-      assert_equal([], @librarian.search("文書作成ソフト"))
+    class FileTypeTest < self
+      def test_pdf
+        path = File.join(@test_fixtures_dir, "test-pdf.pdf")
+        @librarian.collect(path)
+        assert_equal([path], @librarian.search("秋"))
+        assert_equal([], @librarian.search("冬"))
+      end
+
+      def test_xls
+        path = File.join(@test_fixtures_dir, "test-excel.xls")
+        @librarian.collect(path)
+        assert_equal([path], @librarian.search("Excel"))
+      end
+
+      def test_xls_of_multibyte
+        # TODO: I want to detective.
+        omit_on_travis_ci
+        path = File.join(@test_fixtures_dir, "test-excel.xls")
+        @librarian.collect(path)
+        assert_equal([path], @librarian.search("表計算ソフト"))
+        assert_equal([], @librarian.search("文書作成ソフト"))
+      end
     end
   end
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: filerary
 version: !ruby/object:Gem::Version
-  version: 0.0.3
+  version: 0.0.4
 platform: ruby
 authors:
 - Masafumi Yokoyama
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-02-16 00:00:00.000000000 Z
+date: 2014-02-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: grn_mini
@@ -174,6 +174,7 @@ files:
 - test/filerary-test-utils.rb
 - test/fixtures/test-excel.xls
 - test/fixtures/test-pdf.pdf
+- test/fixtures/マルチバイト.txt
 - test/run-test.rb
 - test/test-librarian.rb
 homepage: https://github.com/myokoym/filerary
@@ -206,5 +207,7 @@ test_files:
 - test/filerary-test-utils.rb
 - test/fixtures/test-excel.xls
 - test/fixtures/test-pdf.pdf
+- test/fixtures/マルチバイト.txt
 - test/run-test.rb
 - test/test-librarian.rb
+has_rdoc: