metacrunch 2.1.0.pre1 → 2.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 3651a2620048f6cf78fd1d1670a1f211c5c39dac
-  data.tar.gz: acdfb9744c06a0d7f8725434b9fde56588135bad
+  metadata.gz: d28e81a93e97edc9aa9016c18587317304ca0d68
+  data.tar.gz: 19ec0fccbef5888f0397b6d2abd3c404f7878172
 SHA512:
-  metadata.gz: e2ce0cf3aef3e033c60402ca81d9b568c598b2039c75adb4658fa930d2b9b41c5f2c14ca90e9df058bccb32da08f7c9a3e5f7d87c09d92566426cb46854b3868
-  data.tar.gz: 046c114e1d674aff5d4ce17539da4fac30b58ed3960307fc24f00bbe4f67fbfece6c483830da099eca912423c74485bc8b0aba617d307cdc72fe07fd947d7828
+  metadata.gz: 7f5636649b8bb785b58ba4995aac662803a1901a0b24910ea3eaa2847f86c90a7382a45bd3ad99562548246f8ddefe97af017161074bf09e57ed6686a65c6d6b
+  data.tar.gz: 121cb6f44102a87e13520f8dfde8e8b2feef04329649fbc7d9ec7f93fff652d07bc42b9a84e8cde71760edaf690ccf3e5d5954439d47ce13bc96c6bf2121cd29

data/.gitignore

@@ -1,5 +1,5 @@
 .DS_Store
-pkg
-Gemfile.lock
-.yardoc
-doc
+/Gemfile.lock
+/.yardoc
+/doc
+/*.gem

data/Readme.md

@@ -1,5 +1,6 @@
 # metacrunch
 
+[![Gem Version](https://badge.fury.io/rb/metacrunch.svg)](http://badge.fury.io/rb/metacrunch)
 [![Code Climate](https://codeclimate.com/github/ubpb/metacrunch/badges/gpa.svg)](https://codeclimate.com/github/ubpb/metacrunch)
 [![Build Status](https://travis-ci.org/ubpb/metacrunch.svg)](https://travis-ci.org/ubpb/metacrunch)
 

data/lib/metacrunch/cli/base.rb

@@ -0,0 +1,29 @@
+module Metacrunch
+  class Cli
+    class Base < Thor
+      no_commands do
+        def self.register_thor_command(command_definition)
+          desc(command_definition.usage, command_definition.desc)
+
+          command_definition.options.each do |key, value|
+            option(key, value)
+          end
+
+          define_method(command_definition.name) do |*params|
+            run_command(command_definition.command_class, params)
+          end
+        end
+
+        def run_command(command, params = [])
+          klass = command.class == Class ? command : command.to_s.constantize
+          raise ArgumentError, "command must be a Metacrunch::Command class" unless klass < Metacrunch::Command
+
+          command = klass.new(shell, options, params)
+          command.pre_perform
+          command.perform
+          command.post_perform
+        end
+      end
+    end
+  end
+end

data/lib/metacrunch/cli/command_definition.rb

@@ -0,0 +1,41 @@
+module Metacrunch
+  class Cli
+    class CommandDefinition
+      def initialize(command_class, description)
+        @klass       = command_class
+        @name        = nil
+        @description = description
+        @usage       = nil
+        @options     = {}
+      end
+
+      def command_class(value = nil)
+        @klass if value
+        @klass
+      end
+
+      def name
+        name = @klass.to_s.demodulize.underscore
+        name.gsub!(/_command\Z/, "")
+      end
+
+      def usage(value = nil)
+        @usage = value if value.present?
+        @usage || name
+      end
+
+      def desc(value = nil)
+        @description = value if value.present?
+        @description
+      end
+
+      def option(name, options = {})
+        @options[name] = options
+      end
+
+      def options
+        @options
+      end
+    end
+  end
+end

data/lib/metacrunch/cli/command_registry.rb

@@ -0,0 +1,17 @@
+module Metacrunch
+  class Cli
+    class CommandRegistry
+      attr_reader :commands
+
+      def initialize
+        @commands = []
+      end
+
+      def register(klass, description = nil, &block)
+        command = CommandDefinition.new(klass, description)
+        yield(command) if block_given?
+        @commands << command
+      end
+    end
+  end
+end

data/lib/metacrunch/cli/main.rb

@@ -0,0 +1,7 @@
+module Metacrunch
+  class Cli
+    class Main < Thor
+      # Define metacrunch root commands here
+    end
+  end
+end

data/lib/metacrunch/cli.rb

@@ -1,4 +1,26 @@
 module Metacrunch
-  class Cli < Thor
+  class Cli
+    require_relative "./cli/main"
+    require_relative "./cli/base"
+    require_relative "./cli/command_registry"
+    require_relative "./cli/command_definition"
+
+    def self.start(argv)
+      Main.start(argv)
+    end
+
+    def self.setup(namespace, description, &block)
+      klass = Class.new(Base)
+      klass.namespace(namespace)
+
+      registry = CommandRegistry.new
+      yield(registry)
+
+      registry.commands.each do |c|
+        klass.register_thor_command(c)
+      end
+
+      Main.register(klass, namespace, "#{namespace} [COMMAND]", description)
+    end
   end
 end

data/lib/metacrunch/command.rb

@@ -3,11 +3,24 @@ module Metacrunch
 
     attr_reader :shell
     attr_reader :options
+    attr_reader :params
 
-    def initialize(shell:, options:)
+    def initialize(shell, options = {}, params = [])
       @shell   = shell
       @options = options
-      setup if self.respond_to?(:setup)
+      @params  = params
+    end
+
+    def pre_perform
+      # can be implemented in sub-class
+    end
+
+    def perform
+      raise NotImplementedError, "You must implement .perform() in your command sub-class"
+    end
+
+    def post_perform
+      # can be implemented in sub-class
     end
 
   end

data/lib/metacrunch/snr/field.rb

@@ -0,0 +1,31 @@
+module Metacrunch
+  class SNR
+    class Section
+      class Field
+
+        attr_reader   :name
+        attr_accessor :value
+
+        def initialize(name, value)
+          raise ArgumentError, "required Field#name not given" if name.nil?
+
+          @name  = name
+          @value = value
+        end
+
+        # ------------------------------------------------------------------------------
+        # Serialization
+        # ------------------------------------------------------------------------------
+
+        def to_xml(builder)
+          if value.respond_to?(:to_xml)
+            value.to_xml(root: self.name, builder: builder, skip_instruct: true)
+          else
+            builder.tag!(self.name, self.value)
+          end
+        end
+
+      end
+    end
+  end
+end

data/lib/metacrunch/snr/section.rb

@@ -0,0 +1,63 @@
+module Metacrunch
+  class SNR
+    class Section
+
+      attr_reader :name
+
+      def initialize(name)
+        raise ArgumentError, "required Section#name not given" if name.nil?
+
+        @name   = name
+        @fields = []
+      end
+
+      # ------------------------------------------------------------------------------
+      # Common API
+      # ------------------------------------------------------------------------------
+
+      #
+      # Adds a field
+      #
+      def add(field_name, value)
+        add_field(Field.new(field_name, value))
+      end
+
+      # ------------------------------------------------------------------------------
+      # Fields
+      # ------------------------------------------------------------------------------
+
+      #
+      # Return all fields.
+      #
+      # @return [Array<Metacrunch::SNR::Section::Field>]
+      #
+      def fields
+        @fields
+      end
+
+      #
+      # Adds a new field to this section.
+      #
+      # @param [Metacrunch::SNR::Section::Field] field
+      # @return [Metacrunch::SNR::Section::Field]
+      #
+      def add_field(field)
+        @fields << field
+        field
+      end
+
+      # ------------------------------------------------------------------------------
+      # Serialization
+      # ------------------------------------------------------------------------------
+
+      def to_xml(builder)
+        builder.tag!(self.name) do
+          @fields.each do |_field|
+            _field.to_xml(builder)
+          end
+        end
+      end
+
+    end
+  end
+end

data/lib/metacrunch/snr.rb

@@ -0,0 +1,101 @@
+module Metacrunch
+  #
+  # A SNR object (Simple Normalized Record) is a simple data structure
+  # that you can use as a target resource when performing data normalization routines.
+  # A DNSR record consists of unique sections. A section is unique identified by it's
+  # name. Each section can hold many fields that store the actual values.
+  #
+  # A SNR object can be transformed into XML or JSON to allow easy integration into
+  # existing tools and workflows.
+  #
+  # For example: If you normalize MAB XML data for use in a search engine you can
+  # use a SNR object to store your normalized values.
+  #
+  class SNR
+    require_relative "./snr/section"
+    require_relative "./snr/field"
+
+    # ------------------------------------------------------------------------------
+    # Common API
+    # ------------------------------------------------------------------------------
+
+    #
+    # Adds a field with a value to a section. If the section with the given name
+    # doesn't exists it will be created.
+    #
+    # @param [String] section_name
+    # @param [String] field_name
+    # @param [#to_s] value
+    #
+    def add(section_name, field_name, value)
+      section = self.section(section_name) || Section.new(section_name)
+      section.add(field_name, value)
+      add_section(section)
+    end
+
+    # ------------------------------------------------------------------------------
+    # Sections
+    # ------------------------------------------------------------------------------
+
+    #
+    # @return [Hash{String => Metacrunch::SNR::Section}]
+    # @private
+    #
+    def sections_struct
+      @sections_struct ||= {}
+    end
+    private :sections_struct
+
+    #
+    # Return all sections.
+    #
+    # @return [Array<Metacrunch::SNR::Section>]
+    #
+    def sections
+      sections_struct.values
+    end
+
+    #
+    # Get section by name.
+    #
+    # @param  [String] name Name of the section
+    # @return [Metacrunch::SNR::Section, nil] section by name or nil if a section
+    #   with the given name doesn't exists.
+    #
+    def section(name)
+      sections_struct[name]
+    end
+
+    #
+    # Adds / replaces a section. The name of the section is used as a unique identifier.
+    # Therefore if you add a section with a name that already exists, the new section
+    # will override the existing one.
+    #
+    # @param [Metacrunch::SNR::Section] section
+    # @return [Metacrunch::SNR::Section]
+    #
+    def add_section(section)
+      sections_struct[section.name] = section
+    end
+
+    # ------------------------------------------------------------------------------
+    # Serialization
+    # ------------------------------------------------------------------------------
+
+    #
+    # Transforms the SNR into XML.
+    #
+    # @return [String] The SNR as XML string.
+    #
+    def to_xml
+      builder = Builder::XmlMarkup.new(indent: 2)
+      builder.instruct!(:xml, :encoding => "UTF-8")
+      builder.snr do
+        sections.each do |_section|
+          _section.to_xml(builder)
+        end
+      end
+    end
+
+  end
+end

data/lib/metacrunch/transformer/step.rb

@@ -0,0 +1,33 @@
+require_relative "../transformer"
+
+module Metacrunch
+  class Transformer
+    class Step
+
+      def initialize(transformer)
+        @transformer = transformer
+      end
+
+      def perform
+        raise NotImplementedError, "You must implement .perform() in your rule sub-class"
+      end
+
+      def transformer
+        @transformer
+      end
+
+      def source
+        transformer.source
+      end
+
+      def target
+        transformer.target
+      end
+
+      def options
+        transformer.options
+      end
+
+    end
+  end
+end

data/lib/metacrunch/transformer.rb

@@ -0,0 +1,23 @@
+module Metacrunch
+  class Transformer
+    require_relative "./transformer/step"
+
+    attr_reader :source, :target, :options
+
+    def initialize(source:, target:, options: {})
+      @source  = source
+      @target  = target
+      @options = options
+    end
+
+    def transform(step_class = nil, &block)
+      if block_given?
+        Step.new(self).instance_eval(&block) # TODO: Benchmark this
+      else
+        raise ArgumentError, "You need to provide a STEP or a block" if step_class.nil?
+        step_class.new(self).perform
+      end
+    end
+
+  end
+end

data/lib/metacrunch/version.rb

@@ -1,3 +1,3 @@
 module Metacrunch
-  VERSION = "2.1.0.pre1"
+  VERSION = "2.1.0"
 end

data/lib/metacrunch.rb

@@ -1,16 +1,23 @@
 require "active_support"
 require "active_support/core_ext"
+require "active_support/builder"
 require "parallel"
 require "rubygems"
 require "rubygems/package"
 require "thor"
-require "thor/group"
+require "ox"
 
 begin
   require "pry"
 rescue LoadError ; end
 
 module Metacrunch
+  require_relative "./metacrunch/version"
+  require_relative "./metacrunch/cli"
+  require_relative "./metacrunch/command"
+  require_relative "./metacrunch/utils"
+  require_relative "./metacrunch/snr"
+  require_relative "./metacrunch/transformer"
 
   def self.load_plugins
     Gem.find_latest_files("metacrunch_plugin.rb").each do |path|
@@ -19,8 +26,3 @@ module Metacrunch
   end
 
 end
-
-require "metacrunch/version"
-require "metacrunch/cli"
-require "metacrunch/command"
-require "metacrunch/utils"

data/metacrunch.gemspec

@@ -3,7 +3,7 @@ require File.expand_path("../lib/metacrunch/version", __FILE__)
 Gem::Specification.new do |s|
   s.authors       = ["René Sprotte", "Michael Sievers", "Marcel Otto"]
   s.email         = "r.sprotte@ub.uni-paderborn.de"
-  s.summary       = %q{Metadata processing toolkit for Ruby}
+  s.summary       = %q{Data processing toolkit for Ruby}
   s.description   = s.summary
   s.homepage      = "http://github.com/ubpb/metacrunch"
   s.licenses      = ["MIT"]
@@ -20,6 +20,8 @@ Gem::Specification.new do |s|
   s.required_ruby_version = ">= 2.2.0"
 
   s.add_dependency "activesupport", "~> 4.2", ">= 4.2.0"
+  s.add_dependency "builder",       "~> 3.2", ">= 3.2.2"
   s.add_dependency "parallel",      "~> 1.4", ">= 1.4.1"
   s.add_dependency "thor",          "~> 0.19"
+  s.add_dependency "ox",            "~> 2.1"
 end

data/spec/snr_spec.rb

@@ -0,0 +1,73 @@
+describe Metacrunch::SNR do
+
+  let(:snr) { Metacrunch::SNR.new }
+
+
+  describe ".add" do
+    it "adds a field and creates a section if the section doesn't exists" do
+      expect {
+        snr.add("mysection", "myfield", "My Value")
+      }.to change { snr.sections.count }.by(1)
+    end
+
+    it "adds a field to an existsing section" do
+      snr.add_section(Metacrunch::SNR::Section.new("mysection"))
+
+      expect {
+        snr.add("mysection", "myfield", "My Value")
+      }.not_to change { snr.sections.count }
+    end
+  end
+
+  describe ".sections" do
+    let(:section1) { Metacrunch::SNR::Section.new("section1") }
+    let(:section2) { Metacrunch::SNR::Section.new("section2") }
+
+    before {
+      snr.add_section(section1)
+      snr.add_section(section2)
+    }
+
+    subject { snr.sections }
+
+    it "returns all sections" do
+      expect(snr.sections.count).to be(2)
+    end
+
+    it { is_expected.not_to be_empty }
+    it { is_expected.to include(section1) }
+    it { is_expected.to include(section2) }
+  end
+
+  describe ".section(name)" do
+    context "when section exists" do
+      let(:section1) { Metacrunch::SNR::Section.new("section1") }
+      before { snr.add_section(section1) }
+      subject { snr.section("section1") }
+
+      it { is_expected.not_to be_nil }
+      it { is_expected.to be(section1) }
+    end
+
+    context "when section doesn't exists" do
+      subject { snr.section("no_such_section") }
+
+      it { is_expected.to be_nil }
+    end
+  end
+
+  describe ".to_xml" do
+    before {
+      snr.add("section1", "title", "Foo Bar")
+      snr.add("section1", "artists", ["Sievers, Michael", "Sprotte, René"])
+      snr.add("section2", "link", {label: "Click here", url: "http://example.com"})
+    }
+
+    subject { snr.to_xml }
+
+    it { is_expected.not_to be_nil }
+    it { is_expected.not_to be_empty }
+    it { is_expected.to eq("<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<snr>\n  <section1>\n    <title>Foo Bar</title>\n    <artists type=\"array\">\n      <artist>Sievers, Michael</artist>\n      <artist>Sprotte, René</artist>\n    </artists>\n  </section1>\n  <section2>\n    <link>\n      <label>Click here</label>\n      <url>http://example.com</url>\n    </link>\n  </section2>\n</snr>\n") }
+  end
+
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: metacrunch
 version: !ruby/object:Gem::Version
-  version: 2.1.0.pre1
+  version: 2.1.0
 platform: ruby
 authors:
 - René Sprotte
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-04-23 00:00:00.000000000 Z
+date: 2015-05-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport
@@ -32,6 +32,26 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: 4.2.0
+- !ruby/object:Gem::Dependency
+  name: builder
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '3.2'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 3.2.2
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '3.2'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 3.2.2
 - !ruby/object:Gem::Dependency
   name: parallel
   requirement: !ruby/object:Gem::Requirement
@@ -66,7 +86,21 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '0.19'
-description: Metadata processing toolkit for Ruby
+- !ruby/object:Gem::Dependency
+  name: ox
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '2.1'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '2.1'
+description: Data processing toolkit for Ruby
 email: r.sprotte@ub.uni-paderborn.de
 executables:
 - metacrunch
@@ -83,7 +117,16 @@ files:
 - bin/metacrunch
 - lib/metacrunch.rb
 - lib/metacrunch/cli.rb
+- lib/metacrunch/cli/base.rb
+- lib/metacrunch/cli/command_definition.rb
+- lib/metacrunch/cli/command_registry.rb
+- lib/metacrunch/cli/main.rb
 - lib/metacrunch/command.rb
+- lib/metacrunch/snr.rb
+- lib/metacrunch/snr/field.rb
+- lib/metacrunch/snr/section.rb
+- lib/metacrunch/transformer.rb
+- lib/metacrunch/transformer/step.rb
 - lib/metacrunch/utils.rb
 - lib/metacrunch/utils/file_reader.rb
 - lib/metacrunch/utils/file_reader_result.rb
@@ -93,8 +136,9 @@ files:
 - spec/assets/archive.tar.gz
 - spec/assets/regular_file.txt
 - spec/assets/regular_file.txt.gz
-- spec/file_reader_spec.rb
+- spec/snr_spec.rb
 - spec/spec_helper.rb
+- spec/utils/file_reader_spec.rb
 homepage: http://github.com/ubpb/metacrunch
 licenses:
 - MIT
@@ -110,20 +154,20 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: 2.2.0
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ">"
+  - - ">="
     - !ruby/object:Gem::Version
-      version: 1.3.1
+      version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.4.6
+rubygems_version: 2.4.5
 signing_key: 
 specification_version: 4
-summary: Metadata processing toolkit for Ruby
+summary: Data processing toolkit for Ruby
 test_files:
 - spec/assets/archive.tar
 - spec/assets/archive.tar.gz
 - spec/assets/regular_file.txt
 - spec/assets/regular_file.txt.gz
-- spec/file_reader_spec.rb
+- spec/snr_spec.rb
 - spec/spec_helper.rb
-has_rdoc: 
+- spec/utils/file_reader_spec.rb