middleman-blog-similar 1.1.1 → 2.0.0

data/lib/middleman-blog-similar/tagger/mecab.rb

@@ -0,0 +1,19 @@
+require 'natto'
+
+module Middleman
+  module Blog
+    module Similar
+      module Tagger
+        class Mecab
+          def call(article)
+            Natto::MeCab.new.parse(article.body.gsub(%r{</?[^>]+>}, ''))
+                        .split("\n")
+                        .map { |l| l.split("\t") }
+                        .select { |l| l[1] && l[1].start_with?('名詞,一般') }
+                        .map { |l| l[0] }
+          end
+        end
+      end
+    end
+  end
+end

data/lib/middleman-blog-similar/tagger/tags.rb

@@ -0,0 +1,13 @@
+module Middleman
+  module Blog
+    module Similar
+      module Tagger
+        class Tags
+          def call(article)
+            article.tags
+          end
+        end
+      end
+    end
+  end
+end

data/lib/middleman-blog-similar/version.rb

@@ -1,7 +1,7 @@
 module Middleman
   module Blog
     module Similar
-      VERSION = "1.1.1"
+      VERSION = '2.0.0'.freeze
     end
   end
 end

data/lib/middleman_extension.rb

@@ -1,2 +1 @@
 require 'middleman-blog-similar'
-

data/middleman-blog-similar.gemspec

@@ -1,21 +1,26 @@
 # -*- encoding: utf-8 -*-
-$:.push File.expand_path("../lib", __FILE__)
-require "middleman-blog-similar/version"
+$LOAD_PATH.push File.expand_path('../lib', __FILE__)
+require 'middleman-blog-similar/version'
 
 Gem::Specification.new do |s|
-  s.name = "middleman-blog-similar"
+  s.name = 'middleman-blog-similar'
   s.version = Middleman::Blog::Similar::VERSION
   s.platform = Gem::Platform::RUBY
-  s.authors = ["Atsushi Nagase"]
-  s.email = ["a@ngs.io"]
-  s.homepage = "https://github.com/ngs/middleman-blog-similar"
-  s.summary = %q{Similar article extension for middleman-blog}
-  s.description = %q{Similar article extension for middleman-blog}
-  s.license = "MIT"
+  s.authors = ['Atsushi Nagase']
+  s.email = ['a@ngs.io']
+  s.homepage = 'https://github.com/ngs/middleman-blog-similar'
+  s.summary = 'Similar article extension for middleman-blog'
+  s.description = <<-EOF
+    Similar article extension for middleman-blog.
+    Finds similar articles using tags, part-of-speech or custom lambda.
+  EOF
+  s.license = 'MIT'
   s.files = `git ls-files -z`.split("\0")
   s.test_files = `git ls-files -z -- {fixtures,features,spec}/*`.split("\0")
-  s.require_paths = ["lib"]
-  s.add_runtime_dependency("middleman-core", ["~> 3.2"  ])
-  s.add_runtime_dependency("middleman-blog", ["~> 3.5"  ])
-  s.add_runtime_dependency("fast-stemmer",   ["~> 1.0.2"])
+  s.require_paths = ['lib']
+  s.required_ruby_version = '>= 2.2.0'
+  s.add_runtime_dependency 'activerecord', '~> 5.0', '>= 5.0.0'
+  s.add_runtime_dependency 'middleman-core', '~> 4.0', '>= 4.0.0'
+  s.add_runtime_dependency 'middleman-blog', '~> 4.0', '>= 4.0.0'
+  s.add_dependency 'sqlite3', '~> 1.3'
 end

data/spec/middleman-blog-similar/extension_spec.rb

@@ -1,4 +1,47 @@
 require 'spec_helper'
 
-describe 'Middleman::Blog::SimilarExtension' do
+# rubocop:disable Metrics/BlockLength
+describe Middleman::Blog::SimilarExtension do
+  before(:all) { @app = middleman_app('test-app') { activate :similar, db: ':memory:' } }
+  let(:app) { @app }
+  let(:resource) { app.sitemap.resources.select { |res| res.page_id == page_id }.first }
+  let(:page_id) { '2014-05-08-article0' }
+  describe 'activation' do
+    subject { app.extensions }
+    its([:similar]) { is_expected.not_to be_nil }
+    its([:blog]) { is_expected.not_to be_nil }
+    context 'with unknown tagger' do
+      let(:app) { middleman_app('test-app') { activate :similar, db: ':memory:', tagger: 'hoge' } }
+      it { expect { app }.to raise_error LoadError, 'cannot load such file -- middleman-blog-similar/tagger/hoge' }
+    end
+    context 'with invalid tagger' do
+      let(:app) { middleman_app('test-app') { activate :similar, db: ':memory:', tagger: 1 } }
+      it { expect { app }.to raise_error RuntimeError, /Invalid type for tagger option/ }
+    end
+  end
+  describe 'results' do
+    subject { resource.similar_articles.map { |a| [a.page_id, a.data.category || '<none>'].concat a.tags } }
+    it { is_expected.to have(4).items }
+    its([0]) { is_expected.to eq %w(2014-05-14-article6 <none> dog Brown cat) }
+    its([1]) { is_expected.to eq %w(2014-05-12-article4 <none> dog cat fox) }
+    its([2]) { is_expected.to eq %w(2014-05-09-article1 test dog cat) }
+    its([3]) { is_expected.to eq %w(2014-05-13-article5 <none> dog) }
+    context 'when configured with weight map' do
+      before :all do
+        @app = middleman_app('test-app') do
+          activate :similar,
+                   db: ':memory:',
+                   tagger: {
+                     tags: 1,
+                     category: [2, ->(article) { article.data.category ? [article.data.category] : [] }]
+                   }
+        end
+      end
+      it { is_expected.to have(4).items }
+      its([0]) { is_expected.to eq %w(2014-05-09-article1 test dog cat) }
+      its([1]) { is_expected.to eq %w(2014-05-14-article6 <none> dog Brown cat) }
+      its([2]) { is_expected.to eq %w(2014-05-12-article4 <none> dog cat fox) }
+      its([3]) { is_expected.to eq %w(2014-05-13-article5 <none> dog) }
+    end
+  end
 end

data/spec/middleman-blog-similar/tagger_spec.rb

@@ -0,0 +1,24 @@
+require 'spec_helper'
+
+describe 'tagger' do
+  before(:all) { @app = middleman_app('test-app') {} }
+  let(:app) { @app }
+  let(:resource) { app.sitemap.resources.select { |res| res.page_id == page_id }.first }
+  let(:page_id) { '2014-05-08-article0' }
+  subject { described_class.new.call resource }
+  describe ::Middleman::Blog::Similar::Tagger::Tags do
+    it { is_expected.to eq %w(dog cat brown) }
+  end
+  describe ::Middleman::Blog::Similar::Tagger::Mecab do
+    let(:page_id) { '2014-05-11-article3' }
+    it { is_expected.to eq %w(叫び 悪口 教師 西洋 尻 時分 坊ちゃん 国家 この世 人間 国家 隙) }
+    context 'written in English' do
+      let(:page_id) { '2014-05-10-article2' }
+      it { is_expected.to eq %w(quick brown fox jumps over the lazy dog dog dog) }
+    end
+  end
+  describe ::Middleman::Blog::Similar::Tagger::Entagger do
+    let(:page_id) { '2014-05-10-article2' }
+    it { is_expected.to eq %w(fox jumps dog) }
+  end
+end

data/spec/spec_helper.rb

@@ -2,25 +2,41 @@ PROJECT_ROOT_PATH = File.dirname(File.dirname(__FILE__))
 
 require 'rubygems'
 $LOAD_PATH.unshift File.join(PROJECT_ROOT_PATH, 'lib')
-require 'spork'
 require 'rspec'
-require "middleman-core"
-require "middleman-blog"
-require "middleman-blog/helpers"
+require 'rspec/collection_matchers'
+require 'rspec/its'
+require 'middleman-core'
+require 'middleman-blog'
+require 'middleman-blog/helpers'
+
+require 'codeclimate-test-reporter'
+require 'coveralls'
+require 'simplecov'
+
+SimpleCov.formatter = SimpleCov::Formatter::MultiFormatter[
+  Coveralls::SimpleCov::Formatter,
+  SimpleCov::Formatter::HTMLFormatter,
+  CodeClimate::TestReporter::Formatter
+]
+
+SimpleCov.root(File.expand_path(File.dirname(__FILE__) + '/..'))
+SimpleCov.start
+
+ENV['COVERALLS_REPO_TOKEN'] && Coveralls.wear!
 
 module SpecHelpers
   include FileUtils
 
-  def middleman_app fixture_path, &block
-    tmp_dir     = File.expand_path("../../tmp", __FILE__)
-    fixture_dir = File.expand_path("../../fixtures", __FILE__)
-    fixture_tmp = File.join tmp_dir, "rspec"
+  def middleman_app(fixture_path, &block)
+    tmp_dir     = File.expand_path('../../tmp', __FILE__)
+    fixture_dir = File.expand_path('../../fixtures', __FILE__)
+    fixture_tmp = File.join tmp_dir, 'rspec'
     root_dir    = File.join fixture_tmp, fixture_path
     rmtree fixture_tmp
     mkdir_p tmp_dir
     cp_r fixture_dir, fixture_tmp
-    ENV["MM_SOURCE"] = 'source'
-    ENV["MM_ROOT"] = root_dir
+    ENV['MM_SOURCE'] = 'source'
+    ENV['MM_ROOT'] = root_dir
     initialize_commands = @initialize_commands || []
     initialize_commands.unshift block
     initialize_commands.unshift lambda {
@@ -28,7 +44,7 @@ module SpecHelpers
       set :show_exceptions, false
       activate :blog
     }
-    Middleman::Application.server.inst do
+    ::Middleman::Application.new do
       initialize_commands.each do |p|
         instance_exec(&p)
       end
@@ -36,23 +52,16 @@ module SpecHelpers
   end
 end
 
-Spork.prefork do
-  RSpec.configure do |config|
-    config.include SpecHelpers
-  end
-  require "middleman-blog-similar"
-  require "middleman-blog-similar/extension"
+RSpec.configure do |config|
+  config.include SpecHelpers
 end
 
-Spork.each_run do; end
-
-if ENV['CODECLIMATE_REPO_TOKEN']
-  CodeClimate::TestReporter.start
-  require "codeclimate-test-reporter"
-end
+require 'middleman-blog-similar/extension'
+require 'middleman-blog-similar'
+Dir.glob(PROJECT_ROOT_PATH + '/lib/middleman-blog-similar/tagger/*') { |file| require file }
 
 class String
   def unindent
-    gsub(/^#{scan(/^\s*/).min_by{|l|l.length}}/, "").sub(/\n$/, '')
+    gsub(/^#{scan(/^\s*/).min_by(&:length)}/, '').sub(/\n$/, '')
   end
 end

metadata

@@ -1,66 +1,102 @@
 --- !ruby/object:Gem::Specification
 name: middleman-blog-similar
 version: !ruby/object:Gem::Version
-  version: 1.1.1
+  version: 2.0.0
 platform: ruby
 authors:
 - Atsushi Nagase
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-05-17 00:00:00.000000000 Z
+date: 2017-03-24 00:00:00.000000000 Z
 dependencies:
+- !ruby/object:Gem::Dependency
+  name: activerecord
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '5.0'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 5.0.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '5.0'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 5.0.0
 - !ruby/object:Gem::Dependency
   name: middleman-core
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ~>
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '4.0'
+    - - ">="
       - !ruby/object:Gem::Version
-        version: '3.2'
+        version: 4.0.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ~>
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.2'
+        version: '4.0'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 4.0.0
 - !ruby/object:Gem::Dependency
   name: middleman-blog
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ~>
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '4.0'
+    - - ">="
       - !ruby/object:Gem::Version
-        version: '3.5'
+        version: 4.0.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ~>
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '4.0'
+    - - ">="
       - !ruby/object:Gem::Version
-        version: '3.5'
+        version: 4.0.0
 - !ruby/object:Gem::Dependency
-  name: fast-stemmer
+  name: sqlite3
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ~>
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.0.2
+        version: '1.3'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ~>
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.0.2
-description: Similar article extension for middleman-blog
+        version: '1.3'
+description: |2
+      Similar article extension for middleman-blog.
+      Finds similar articles using tags, part-of-speech or custom lambda.
 email:
 - a@ngs.io
 executables: []
 extensions: []
 extra_rdoc_files: []
 files:
-- .gitignore
-- .travis.yml
+- ".gitignore"
+- ".rspec"
+- ".rubocop.yml"
+- ".travis.yml"
 - CHANGELOG.md
 - CONTRIBUTING.md
 - Gemfile
@@ -68,43 +104,37 @@ files:
 - LICENSE.md
 - README.md
 - Rakefile
-- features/damerau_levenshtein.feature
-- features/levenshtein.feature
+- features/default.feature
 - features/support/env.rb
-- features/word_frequency.feature
-- fixtures/test-app/source/2014-05-08-article0.md
-- fixtures/test-app/source/2014-05-09-article1.md
-- fixtures/test-app/source/2014-05-10-article2.md
-- fixtures/test-app/source/2014-05-11-article3.md
-- fixtures/test-app/source/2014-05-12-article4.md
-- fixtures/test-app/source/2014-05-13-article5.md
-- fixtures/test-app/source/2014-05-14-article6.md
+- fixtures/test-app/source/2014-05-08-article0.html.md
+- fixtures/test-app/source/2014-05-09-article1.html.md
+- fixtures/test-app/source/2014-05-10-article2.html.md
+- fixtures/test-app/source/2014-05-11-article3.html.md
+- fixtures/test-app/source/2014-05-12-article4.html.md
+- fixtures/test-app/source/2014-05-13-article5.html.md
+- fixtures/test-app/source/2014-05-14-article6.html.md
 - fixtures/test-app/source/index.html.slim
 - fixtures/test-app/source/layout.slim
 - fixtures/test-app/source/layouts/article.slim
 - fixtures/test-app/source/page.html.slim
 - lib/middleman-blog-similar.rb
-- lib/middleman-blog-similar/algorithm.rb
-- lib/middleman-blog-similar/algorithm/damerau_levenshtein.rb
-- lib/middleman-blog-similar/algorithm/levenshtein.rb
-- lib/middleman-blog-similar/algorithm/unigrams.csv
-- lib/middleman-blog-similar/algorithm/word_frequency.rb
-- lib/middleman-blog-similar/algorithm/word_frequency/mecab.rb
-- lib/middleman-blog-similar/algorithm/word_frequency/tree_tagger.rb
 - lib/middleman-blog-similar/blog_article_extensions.rb
+- lib/middleman-blog-similar/database.rb
 - lib/middleman-blog-similar/extension.rb
 - lib/middleman-blog-similar/helpers.rb
+- lib/middleman-blog-similar/models/article.rb
+- lib/middleman-blog-similar/models/migration.rb
+- lib/middleman-blog-similar/models/tag.rb
+- lib/middleman-blog-similar/models/tagging.rb
+- lib/middleman-blog-similar/resource_list_manipulator.rb
+- lib/middleman-blog-similar/tagger/entagger.rb
+- lib/middleman-blog-similar/tagger/mecab.rb
+- lib/middleman-blog-similar/tagger/tags.rb
 - lib/middleman-blog-similar/version.rb
 - lib/middleman_extension.rb
 - middleman-blog-similar.gemspec
-- spec/helper_spec.rb
-- spec/middleman-blog-similar/algorithm/damerau_levenshtein_spec.rb
-- spec/middleman-blog-similar/algorithm/levenshtein_spec.rb
-- spec/middleman-blog-similar/algorithm/word_frequency/mecab_spec.rb
-- spec/middleman-blog-similar/algorithm/word_frequency/tree_tagger_spec.rb
-- spec/middleman-blog-similar/algorithm/word_frequency_spec.rb
-- spec/middleman-blog-similar/algorithm_spec.rb
 - spec/middleman-blog-similar/extension_spec.rb
+- spec/middleman-blog-similar/tagger_spec.rb
 - spec/spec_helper.rb
 homepage: https://github.com/ngs/middleman-blog-similar
 licenses:
@@ -116,43 +146,18 @@ require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
-  - - '>='
+  - - ">="
     - !ruby/object:Gem::Version
-      version: '0'
+      version: 2.2.0
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - '>='
+  - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.0.14
+rubygems_version: 2.6.8
 signing_key: 
 specification_version: 4
 summary: Similar article extension for middleman-blog
-test_files:
-- features/damerau_levenshtein.feature
-- features/levenshtein.feature
-- features/support/env.rb
-- features/word_frequency.feature
-- fixtures/test-app/source/2014-05-08-article0.md
-- fixtures/test-app/source/2014-05-09-article1.md
-- fixtures/test-app/source/2014-05-10-article2.md
-- fixtures/test-app/source/2014-05-11-article3.md
-- fixtures/test-app/source/2014-05-12-article4.md
-- fixtures/test-app/source/2014-05-13-article5.md
-- fixtures/test-app/source/2014-05-14-article6.md
-- fixtures/test-app/source/index.html.slim
-- fixtures/test-app/source/layout.slim
-- fixtures/test-app/source/layouts/article.slim
-- fixtures/test-app/source/page.html.slim
-- spec/helper_spec.rb
-- spec/middleman-blog-similar/algorithm/damerau_levenshtein_spec.rb
-- spec/middleman-blog-similar/algorithm/levenshtein_spec.rb
-- spec/middleman-blog-similar/algorithm/word_frequency/mecab_spec.rb
-- spec/middleman-blog-similar/algorithm/word_frequency/tree_tagger_spec.rb
-- spec/middleman-blog-similar/algorithm/word_frequency_spec.rb
-- spec/middleman-blog-similar/algorithm_spec.rb
-- spec/middleman-blog-similar/extension_spec.rb
-- spec/spec_helper.rb
-has_rdoc: 
+test_files: []