jruby-bloomfilter 1.0.0 → 1.0.2

data/Gemfile

@@ -0,0 +1,5 @@
+source :rubygems
+
+gem 'jets3t-rb', :git => 'git://github.com/gaiottino/jets3t-rb.git'
+gem 'rspec'
+

data/Gemfile.lock

@@ -0,0 +1,25 @@
+GIT
+  remote: git://github.com/gaiottino/jets3t-rb.git
+  revision: f7c77fa1d393bb3ebaf42d4d871071d3e9652328
+  specs:
+    jets3t-rb (1.0.0)
+
+GEM
+  remote: http://rubygems.org/
+  specs:
+    diff-lcs (1.1.2)
+    rspec (2.5.0)
+      rspec-core (~> 2.5.0)
+      rspec-expectations (~> 2.5.0)
+      rspec-mocks (~> 2.5.0)
+    rspec-core (2.5.1)
+    rspec-expectations (2.5.0)
+      diff-lcs (~> 1.1.2)
+    rspec-mocks (2.5.0)
+
+PLATFORMS
+  java
+
+DEPENDENCIES
+  jets3t-rb!
+  rspec

data/bloomfilter.gemspec

@@ -0,0 +1,24 @@
+# encoding: utf-8
+
+$: << File.expand_path('../lib', __FILE__)
+
+Dir['ext/*.jar'].each { |jar| require jar }
+
+Gem::Specification.new do |s|
+  s.name        = 'jruby-bloomfilter'
+  s.version     = '1.0.2'
+  s.platform    = Gem::Platform::RUBY
+  s.authors     = ['Daniel Gaiottino']
+  s.email       = ['daniel@burtcorp.com']
+  s.homepage    = 'http://github.com/gaiottino/bloomfilter'
+  s.summary     = %q{JRuby wrapper for java-bloomfilter}
+  s.description = %q{JRuby wrapper (+ some extra functionality) to http://code.google.com/p/java-bloomfilter}
+
+  s.rubyforge_project = 'jruby-bloomfilter'
+  s.add_dependency 'jets3t-rb', '~> 1.0.0'
+
+  s.files         = `git ls-files`.split("\n")
+  s.test_files    = `git ls-files -- {test,spec,features}/*`.split("\n")
+  s.executables   = `git ls-files -- bin/*`.split("\n").map{ |f| File.basename(f) }
+  s.require_paths = %w(lib)
+end

data/lib/bloomfilter.rb

@@ -3,44 +3,39 @@
 require 'java'
 require 'ext/java-bloomfilter-0.9.3'
 
-require 'serializer'
+require_relative 'bloomfilter/serializer'
 
 module Jar
   import com.skjegstad.utils.BloomFilter
 end
 
-class Bloomfilter
-  include Serialization
-  
-  def initialize(options = {})
-    if options[:size] && options[:false_positive_percentage]
-      @filter = Jar::BloomFilter.new(options[:false_positive_percentage], options[:size])
-    elsif options[:filter]
-      @filter = options[:filter]
+module Bloomfilter
+  class Bloomfilter
+    def initialize(options = {})
+      if options[:size] && options[:false_positive_percentage]
+        @filter = Jar::BloomFilter.new(options[:false_positive_percentage], options[:size])
+      elsif options[:filter]
+        @filter = options[:filter]
+      end
     end
-  end
   
-  def << (k)
-    @filter.add(k)
-  end
+    def << (k)
+      @filter.add(k)
+    end
+    
+    def add_if_absent(k)
+      @filter.synchronized do
+        @filter.add(k) unless @filter.contains(k)
+      end
+    end
   
-  def include?(k)
-    @filter.contains(k)
-  end
+    def include?(k)
+      @filter.contains(k)
+    end
   
-  def count
-    @filter.count
-  end
+    def count
+      @filter.count
+    end
     
-  # def store(path)
-  #   raise 'Cannot call store without specifying :serializer in initlialize options' unless @serializer
-  #   @serializer.store(path, @filter)
-  # end
-  # 
-  # def self.load(type, path)
-  #   serializer = Serialization.serializer(type)
-  #   filter = serializer.load(path)
-  #   Bloomfilter.new(:filter => filter)
-  # end
-
+  end
 end

data/lib/bloomfilter/serialization/file.rb

@@ -0,0 +1,29 @@
+module Bloomfilter
+  module Serialization
+    class File
+      def store(path, filter)
+        dir = ::File.dirname(path)
+        unless ::File.directory?(dir)
+          %x(mkdir -p #{dir})
+        end
+        
+        ::File.open(path, 'w') do |f|
+          Marshal.dump(filter, f)
+        end
+      end
+
+      def load(path)
+        return nil unless ::File.exist?(path)
+        
+        ::File.open(path, 'r') do |f|
+          return Marshal.load(f)
+        end
+      end
+      
+    private
+    
+      def extract_folder_path
+      end
+    end
+  end
+end

data/lib/bloomfilter/serialization/s3.rb

@@ -0,0 +1,31 @@
+require 'jets3t'
+require 'tempfile'
+
+module Bloomfilter
+  module Serialization
+    class S3
+      AWS_SECRET_PATH = '~/.awssecret'.freeze
+      TEMP_FILE_PREFIX = 'bloomfilter'.freeze
+    
+      def initialize(s3_service, bucket_name)
+        @file_serializer = File.new
+        @bucket = s3_service.bucket(bucket_name)
+      end
+    
+      def store(path, filter)
+        tmp = Tempfile.new(TEMP_FILE_PREFIX)
+        @file_serializer.store(tmp.path, filter)
+        @bucket.put(path, tmp)
+      end
+
+      def load(path)
+        s3_object = @bucket.get(path)
+        tmp = Tempfile.new(TEMP_FILE_PREFIX)
+        ::File.open(tmp.path, 'w') do |f|
+          f << s3_object.data
+        end
+        @file_serializer.load(tmp.path)
+      end
+    end
+  end
+end

data/lib/bloomfilter/serializer.rb

@@ -0,0 +1,18 @@
+require_relative 'serialization/file'
+require_relative 'serialization/s3'
+
+module Bloomfilter
+  class Serializer
+
+    def self.s3(access_key, secret_key, bucket_name)
+      credentials = JetS3t::AWSCredentials.new(access_key_id, secret_access_key)
+      s3_service = JetS3t::RestS3Service.new(credentials)
+      return Serialization::S3.new(s3_service, bucket_name)
+    end
+  
+    def self.file
+      return Serialization::File.new
+    end
+  
+  end
+end

data/spec/bloomfilter_spec.rb

@@ -4,33 +4,35 @@ require_relative 'spec_helper'
 
 require 'tempfile'
 
-describe Bloomfilter do
-  before :all do
-    @filter = Bloomfilter.new(:size => 1_000, :false_positive_percentage => 0.01)
-    @path = Tempfile.new('bloomfilter').path
-  end
-  
-  it 'should not have any elements from start' do
-    @filter.count.should == 0
-  end
+module Bloomfilter
+  describe Bloomfilter do
+    before :all do
+      @filter = Bloomfilter.new(:size => 1_000, :false_positive_percentage => 0.01)
+      @path = Tempfile.new('bloomfilter').path
+    end
   
-  describe '#<<' do
-    it 'should be possible to add elements which updates the count' do
-      @filter << 'hello'
-      @filter.count.should == 1
+    it 'should not have any elements from start' do
+      @filter.count.should == 0
     end
-  end
   
-  describe '#include?' do
-    it 'should return false if an element does not exist' do
-      @filter.include?('world').should be_false
+    describe '#<<' do
+      it 'should be possible to add elements which updates the count' do
+        @filter << 'hello'
+        @filter.count.should == 1
+      end
     end
+  
+    describe '#include?' do
+      it 'should return false if an element does not exist' do
+        @filter.include?('world').should be_false
+      end
     
-    it 'should return true if an element exists' do
-      @filter.include?('hello').should be_true
+      it 'should return true if an element exists' do
+        @filter.include?('hello').should be_true
       
-      @filter << 'world'
-      @filter.include?('world').should be_true
+        @filter << 'world'
+        @filter.include?('world').should be_true
+      end
     end
   end
 end

data/spec/serialization_spec.rb

@@ -0,0 +1,54 @@
+# encoding: utf-8
+
+require_relative 'spec_helper'
+
+require 'tempfile'
+
+module Bloomfilter
+  describe 'Serialization' do
+    before :all do
+      @filter = Bloomfilter.new(:size => 1_000, :false_positive_percentage => 0.01)
+      @filter << 'hello'
+      @filter << 'world'
+      @path = Tempfile.new('bloomfilter').path
+    end
+  
+    context 'File' do
+      it 'should be possible to store a filter to a file' do
+        Serializer.file.store(@path, @filter)
+        File.new(@path).size.should == 1612
+      end
+
+      it 'should be possible to load a filter from a file' do
+        filter = Serializer.file.load(@path)
+        filter.count.should == 2
+        filter.include?('hello').should be_true
+        filter.include?('world').should be_true
+        filter.include?('bloomfilter').should be_false
+      end
+    end
+  
+    context 'S3' do
+      before do
+        bucket_name = 'test-bucket'
+        s3_service = stub('s3_service')
+        @bucket = stub(bucket_name)
+      
+        s3_service.stub(:bucket).with(bucket_name).and_return(@bucket)
+        @s3 = Serialization::S3.new(s3_service, bucket_name)
+      end
+    
+      it 'should be possible to store a filter to S3' do
+        @bucket.should_receive(:put).with('/some/path', anything)
+        @s3.store('/some/path', @filter)
+      end
+    
+      it 'should be possible to load a filter from S3' do
+        s3_object = stub('s3_object')
+        @bucket.should_receive(:get).with('/some/path').and_return(s3_object)
+        s3_object.should_receive(:data).and_return(Marshal.dump(@filter))
+        @s3.load('/some/path')
+      end
+    end
+  end
+end

metadata

@@ -2,7 +2,7 @@
 name: jruby-bloomfilter
 version: !ruby/object:Gem::Version 
   prerelease: 
-  version: 1.0.0
+  version: 1.0.2
 platform: ruby
 authors: 
 - Daniel Gaiottino
@@ -10,7 +10,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-07-11 00:00:00 Z
+date: 2011-07-12 00:00:00 Z
 dependencies: 
 - !ruby/object:Gem::Dependency 
   name: jets3t-rb
@@ -35,11 +35,18 @@ extra_rdoc_files: []
 files: 
 - .gitignore
 - .rvmrc
+- Gemfile
+- Gemfile.lock
 - README
 - benchmark/memory.rb
+- bloomfilter.gemspec
 - lib/bloomfilter.rb
+- lib/bloomfilter/serialization/file.rb
+- lib/bloomfilter/serialization/s3.rb
+- lib/bloomfilter/serializer.rb
 - lib/ext/java-bloomfilter-0.9.3.jar
 - spec/bloomfilter_spec.rb
+- spec/serialization_spec.rb
 - spec/spec_helper.rb
 homepage: http://github.com/gaiottino/bloomfilter
 licenses: []
@@ -70,4 +77,5 @@ specification_version: 3
 summary: JRuby wrapper for java-bloomfilter
 test_files: 
 - spec/bloomfilter_spec.rb
+- spec/serialization_spec.rb
 - spec/spec_helper.rb