fts_lite 0.1.2 → 0.1.3

data/lib/fts_lite/index.rb

@@ -91,7 +91,7 @@ module FtsLite
         sql += sprintf(" LIMIT %d", limit)
       end
       sql += ";"
-      @db.execute(sql, [@tokenizer.query(text)]).flatten
+      @db.execute(sql, [@tokenizer.query(text, options)]).flatten
     end
     def count
       @db.execute("SELECT COUNT(*) FROM #{@table_name} ;").first.first

data/lib/fts_lite/tokenizer.rb

@@ -6,7 +6,8 @@ module FtsLite
   module Tokenizer
     QUERY_DELIMITER = /[\s　]+/
     SIMPLE_DELIMITER = /[\s　\.\*"',\?!;\(\)。、．，？！「」『』（）]+/
-    NEAR = " NEAR/2 "
+    NEAR0 = " NEAR/0 "
+    NEAR2 = " NEAR/2 "
     
     def self.create(name)
       case name.to_sym
@@ -28,7 +29,7 @@ module FtsLite
       NKF::nkf('-wZX', text).downcase
     end
     class Simple
-      def query(text)
+      def query(text, options)
         vector(text)
       end
       def vector(text)
@@ -39,12 +40,14 @@ module FtsLite
       end
     end
     class Bigram
-      def query(text)
+      def query(text, options = {})
+        fuzzy = options.key?(:fuzzy) ? options[:fuzzy] : false
+        near = fuzzy ? NEAR2 : NEAR0
         text = Tokenizer.normalize(text)
         text.split(QUERY_DELIMITER).map {|segment|
           segment.split(SIMPLE_DELIMITER).map {|word|
             0.upto(word.size - 2).map {|i| word[i, 2] }
-          }.join(NEAR)
+          }.join(near)
         }.flatten.join(" ")
       end
       def vector(text)
@@ -58,12 +61,14 @@ module FtsLite
       end
     end
     class Trigram
-      def query(text)
+      def query(text, options = {})
+        fuzzy = options.key?(:fuzzy) ? options[:fuzzy] : false
+        near = fuzzy ? NEAR2 : NEAR0
         text = Tokenizer.normalize(text)
         text.split(QUERY_DELIMITER).map {|segment|
           segment.split(SIMPLE_DELIMITER).map {|word|
             0.upto(word.size - 3).map {|i| word[i, 3] }
-          }.join(NEAR)
+          }.join(near)
         }.flatten.join(" ")
       end
       def vector(text)
@@ -77,12 +82,14 @@ module FtsLite
       end
     end
     class Wakachi
-      def query(text)
+      def query(text, options = {})
+        fuzzy = options.key?(:fuzzy) ? options[:fuzzy] : false
+        near = fuzzy ? NEAR2 : NEAR0
         text = Tokenizer.normalize(text)
         text.split(QUERY_DELIMITER).map {|segment|
           BimyouSegmenter.segment(segment,
                                   :white_space => false,
-                                  :symbol => false).join(NEAR)
+                                  :symbol => false).join(near)
         }.join(" ")
       end
       def vector(text)
@@ -95,7 +102,9 @@ module FtsLite
       end
     end
     class WakachiBigram
-      def query(text)
+      def query(text, options = {})
+        fuzzy = options.key?(:fuzzy) ? options[:fuzzy] : false
+        near = fuzzy ? NEAR2 : NEAR0
         text = Tokenizer.normalize(text)
         text.split(QUERY_DELIMITER).map {|segment|
           BimyouSegmenter.segment(segment,
@@ -104,9 +113,9 @@ module FtsLite
             if (word.size == 1)
               word
             else
-              0.upto(word.size - 2).map {|i| word[i, 2] }.join(NEAR)
+              0.upto(word.size - 2).map {|i| word[i, 2] }.join(near)
             end
-          }.flatten.join(NEAR)
+          }.flatten.join(near)
         }.join(" ")
       end
       def vector(text)

data/lib/fts_lite/version.rb

@@ -1,3 +1,3 @@
 module FtsLite
-  VERSION = "0.1.2"
+  VERSION = "0.1.3"
 end

data/test/fts_lite_test.rb

@@ -181,6 +181,20 @@ class FtsLiteTest < Test::Unit::TestCase
       assert_equal db.search("ナポリタン", :order => :asc)[1], 2
     end
   end
+  def test_fuzzy
+    db = FtsLite::Index.open(DB_FILE, :tokenizer => :bigram)
+    db.transaction do
+      db.delete_all
+      db.set(1, "あいいいう")
+      db.set(2, "あいいう")
+      assert_equal db.search("あいい").size, 2
+      assert_equal db.search("あいう").size, 0
+      assert_equal db.search("あいい", :fuzzy => true).size, 2
+      assert_equal db.search("あいう", :fuzzy => true).size, 2
+      assert_equal db.search("あいい", :fuzzy => false).size, 2
+      assert_equal db.search("あいう", :fuzzy => false).size, 0
+    end
+  end
   def test_create
     db = FtsLite::Index.open(DB_FILE)
     db.drop_table!

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: fts_lite
 version: !ruby/object:Gem::Version
-  version: 0.1.2
+  version: 0.1.3
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-08-11 00:00:00.000000000 Z
+date: 2013-06-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bimyou_segmenter
@@ -82,7 +82,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 1.8.24
+rubygems_version: 1.8.23
 signing_key: 
 specification_version: 3
 summary: simple full text search engine