hydra-file_characterization 0.0.2 → 0.1.0

data/Gemfile

@@ -1,4 +1,4 @@
 source 'https://rubygems.org'
 
-# Specify your gem's dependencies in hydra-file_characterization.gemspec
+# Specify your gem's dependencies in hydra/file_characterization.gemspec
 gemspec

data/hydra-file_characterization.gemspec

@@ -1,7 +1,7 @@
 # -*- encoding: utf-8 -*-
 lib = File.expand_path('../lib', __FILE__)
 $LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
-require 'hydra-file_characterization/version'
+require 'hydra/file_characterization/version'
 
 Gem::Specification.new do |gem|
   gem.name          = "hydra-file_characterization"
@@ -20,7 +20,7 @@ Gem::Specification.new do |gem|
   ]
   gem.description   = %q{To provide a wrapper for file characterization}
   gem.summary       = %q{To provide a wrapper for file characterization}
-  gem.homepage      = "https://github.com/projecthydra/hydra-file_characterization"
+  gem.homepage      = "https://github.com/projecthydra/hydra/file_characterization"
   gem.license = "APACHE2"
 
   gem.files         = `git ls-files`.split($/)

data/lib/hydra-file_characterization.rb

@@ -1,14 +1 @@
-require "hydra-file_characterization/version"
-require "hydra-file_characterization/characterizer"
-require "active_support/configurable"
-
-module Hydra
-  module FileCharacterization
-
-    class Configuration
-      include ActiveSupport::Configurable
-      config_accessor :fits_path
-    end
-  end
-end
-
+require "hydra/file_characterization"

data/lib/hydra/file_characterization.rb

@@ -0,0 +1,29 @@
+require "hydra/file_characterization/version"
+require "hydra/file_characterization/exceptions"
+require "hydra/file_characterization/to_temp_file"
+require "hydra/file_characterization/characterizer"
+require "hydra/file_characterization/characterizers"
+require "active_support/configurable"
+
+module Hydra
+
+  module_function
+  def characterize(content, filename, *options)
+    tool_outputs = []
+    tool_names = Array(options).flatten.compact
+    FileCharacterization::ToTempFile.open(content, filename) do |f|
+      tool_names.each do |tool_name|
+        tool_outputs << FileCharacterization.characterize_with(tool_name, f.path)
+      end
+    end
+    tool_names.size == 1 ? tool_outputs.first : tool_outputs
+  end
+
+  module FileCharacterization
+
+    class Configuration
+      include ActiveSupport::Configurable
+      config_accessor :tool_path
+    end
+  end
+end

data/lib/{hydra-file_characterization → hydra/file_characterization}/characterizer.rb

@@ -1,24 +1,22 @@
+require 'hydra/file_characterization/exceptions'
 require 'open3'
+require 'active_support/core_ext/class/attribute'
 
 module Hydra::FileCharacterization
-
   class Characterizer
     include Open3
 
-    class FileNotFoundError < RuntimeError
-    end
+    class_attribute :tool_path
 
-    attr_reader :filename, :fits_path
-    def initialize(filename, fits_path)
+    attr_reader :filename
+    def initialize(filename)
       @filename = filename
-      @fits_path = fits_path
     end
 
     def call
       unless File.exists?(filename)
-        raise FileNotFoundError.new("File: #{filename} does not exist.")
+        raise Hydra::FileCharacterization::FileNotFoundError.new("File: #{filename} does not exist.")
       end
-      command = "#{fits_path} -i \"#{filename}\""
       stdin, stdout, stderr, wait_thr = popen3(command)
       begin
         out = stdout.read
@@ -32,5 +30,15 @@ module Hydra::FileCharacterization
         stderr.close
       end
     end
+
+    protected
+
+    def command
+      raise NotImplementedError, "Method #command should be overriden in child classes"
+    end
+
+    def tool_path
+      raise NotImplementedError, "Method #tool_path should be overriden in child classes"
+    end
   end
-end
+end

data/lib/hydra/file_characterization/characterizers.rb

@@ -0,0 +1,26 @@
+module Hydra::FileCharacterization
+  module Characterizers
+  end
+
+  module_function
+  def characterizer(tool_name)
+    characterizer_name = characterizer_name_from(tool_name)
+    if Characterizers.const_defined?(characterizer_name)
+      Characterizers.const_get(characterizer_name)
+    else
+      raise ToolNotFoundError.new(tool_name)
+    end
+  end
+
+  def characterizer_name_from(tool_name)
+    tool_name.to_s.gsub(/(?:^|_)([a-z])/) { $1.upcase }
+  end
+  
+  def characterize_with(tool_name, file_path)
+    tool_obj = characterizer(tool_name).new(file_path)
+    tool_obj.call
+  end
+
+end
+
+require 'hydra/file_characterization/characterizers/fits'

data/lib/hydra/file_characterization/characterizers/fits.rb

@@ -0,0 +1,14 @@
+require 'hydra/file_characterization/characterizer'
+module Hydra::FileCharacterization::Characterizers
+  class Fits < Hydra::FileCharacterization::Characterizer
+
+    protected
+    def command
+      "#{tool_path} -i \"#{filename}\""
+    end
+
+    def tool_path
+      self.class.tool_path || (raise "No Tool Path Given")
+    end
+  end
+end

data/lib/hydra/file_characterization/exceptions.rb

@@ -0,0 +1,12 @@
+module Hydra::FileCharacterization
+
+  class FileNotFoundError < RuntimeError
+  end
+
+  class ToolNotFoundError < RuntimeError
+    def initialize(tool_name)
+      super("Unable to find Hydra::FileCharacterization tool with name :#{tool_name}")
+    end
+  end
+
+end

data/lib/{hydra-file_characterization → hydra/file_characterization}/to_temp_file.rb

@@ -1,9 +1,14 @@
 require 'open3'
+require 'tempfile'
 
 module Hydra::FileCharacterization
   class ToTempFile
     include Open3
 
+    def self.open(*args, &block)
+      new(*args).call(&block)
+    end
+
     attr_accessor :data, :filename
 
     def initialize(data, filename)
@@ -12,14 +17,18 @@ module Hydra::FileCharacterization
     end
 
     def call
-      return unless data.empty?
-      timestamp = DateTime.now.strftime("%Y%m%d%M%S")
-      Tempfile.open("#{timestamp}_#{File.basename(filename)}") do |f|
+      return if data.empty?
+      f = Tempfile.new([File.basename(filename),File.extname(filename)])
+      begin
         f.binmode
         f.write(data)
         f.rewind
         yield(f)
+      ensure
+        f.close
+        f.unlink
       end
+
     end
   end
 end

data/lib/{hydra-file_characterization → hydra/file_characterization}/version.rb

@@ -1,5 +1,5 @@
 module Hydra
   module FileCharacterization
-    VERSION = "0.0.2"
+    VERSION = "0.1.0"
   end
 end

data/spec/fixtures/archive.zip.fits.xml

@@ -12,7 +12,7 @@
   </identification>
   <fileinfo>
     <lastmodified toolname="Exiftool" toolversion="9.06" status="SINGLE_RESULT">2013:09:17 15:45:51-04:00</lastmodified>
-    <filepath toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">/Users/jfriesen/Repositories/hydra-file_characterization/spec/fixtures/archive.zip</filepath>
+    <filepath toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">/Users/jfriesen/Repositories/hydra/file_characterization/spec/fixtures/archive.zip</filepath>
     <filename toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">spec/fixtures/archive.zip</filename>
     <size toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">302</size>
     <md5checksum toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">7c5b22ea09ba0eb837f70e2e8094b26f</md5checksum>

data/spec/fixtures/brendan_behan.jpeg.fits.xml

@@ -14,8 +14,8 @@
   <fileinfo>
     <size toolname="Jhove" toolversion="1.5">8744</size>
     <lastmodified toolname="Exiftool" toolversion="9.06" status="SINGLE_RESULT">2013:09:17 13:15:45-04:00</lastmodified>
-    <filepath toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">/Users/jfriesen/Repositories/hydra-file_characterization/spec/fixtures/brendan_behan.jpeg</filepath>
-    <filename toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">/Users/jfriesen/Repositories/hydra-file_characterization/spec/fixtures/brendan_behan.jpeg</filename>
+    <filepath toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">/Users/jfriesen/Repositories/hydra/file_characterization/spec/fixtures/brendan_behan.jpeg</filepath>
+    <filename toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">/Users/jfriesen/Repositories/hydra/file_characterization/spec/fixtures/brendan_behan.jpeg</filename>
     <md5checksum toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">66a51b92863dd8e2c71c38979d84161c</md5checksum>
     <fslastmodified toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">1379438145000</fslastmodified>
   </fileinfo>

data/spec/fixtures/brendan_broken.dxxd.fits.xml

@@ -6,8 +6,8 @@
     </identity>
   </identification>
   <fileinfo>
-    <filepath toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">/Users/jfriesen/Repositories/hydra-file_characterization/spec/fixtures/brendan_broken.dxxd</filepath>
-    <filename toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">/Users/jfriesen/Repositories/hydra-file_characterization/spec/fixtures/brendan_broken.dxxd</filename>
+    <filepath toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">/Users/jfriesen/Repositories/hydra/file_characterization/spec/fixtures/brendan_broken.dxxd</filepath>
+    <filename toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">/Users/jfriesen/Repositories/hydra/file_characterization/spec/fixtures/brendan_broken.dxxd</filename>
     <size toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">9574</size>
     <md5checksum toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">ca786ec0489e53c945d1fa7b584bac7f</md5checksum>
     <fslastmodified toolname="OIS File Information" toolversion="0.1" status="SINGLE_RESULT">1379446883000</fslastmodified>

data/spec/lib/{hydra-file_characterization/characterizer_spec.rb → hydra/file_characterization/characterizers/fits_spec.rb}

@@ -1,14 +1,11 @@
 require 'spec_helper'
-require 'hydra-file_characterization'
+require 'hydra/file_characterization/characterizers/fits'
 
-describe Hydra::FileCharacterization do
-  describe 'Characterizer' do
-    def fixture_file(filename)
-      File.expand_path(File.join('../../../fixtures', filename), __FILE__)
-    end
+module Hydra::FileCharacterization::Characterizers
+
+  describe Fits do
 
-    let(:fits_path) { `which fits || which fits.sh`.strip }
-    subject { Hydra::FileCharacterization::Characterizer.new(filename, fits_path) }
+    subject { Fits.new(filename) }
 
     describe 'validfile' do
       let(:filename) { fixture_file('brendan_behan.jpeg') }
@@ -20,7 +17,7 @@ describe Hydra::FileCharacterization do
     describe 'invalidFile' do
       let(:filename) { fixture_file('nofile.pdf') }
       it "should raise an error if the path does not contain the file" do
-        expect {subject.call}.to raise_error(Hydra::FileCharacterization::Characterizer::FileNotFoundError)
+        expect {subject.call}.to raise_error(Hydra::FileCharacterization::FileNotFoundError)
       end
     end
 
@@ -37,4 +34,5 @@ describe Hydra::FileCharacterization do
     end
 
   end
+
 end

data/spec/lib/hydra/file_characterization/characterizers_spec.rb

@@ -0,0 +1,14 @@
+require 'spec_helper'
+require 'hydra/file_characterization/characterizers'
+
+module Hydra::FileCharacterization
+  describe Characterizers do
+    subject { Hydra::FileCharacterization.characterizer(tool_name) }
+
+    describe 'with :fits tool_name' do
+      let(:tool_name) { :fits }
+      it { should eq(Characterizers::Fits) }
+    end
+
+  end
+end

data/spec/lib/hydra/file_characterization/to_temp_file_spec.rb

@@ -0,0 +1,36 @@
+require 'spec_helper'
+require 'hydra/file_characterization/to_temp_file'
+
+module Hydra::FileCharacterization
+
+  describe 'ToTempFile' do
+
+    let(:content) { "This is the content of the file." }
+    let(:filename) { "hello.rb" }
+
+    describe '.open' do
+      subject { ToTempFile }
+      it 'creates a tempfile then unlinks it' do
+        subject.open(content, filename) do |temp_file|
+          @temp_file = temp_file
+          expect(File.exist?(@temp_file.path)).to eq true
+          expect(File.extname(@temp_file.path)).to include '.rb'
+        end
+        expect(@temp_file.path).to eq nil
+      end
+    end
+
+    describe 'instance' do
+      subject { ToTempFile.new(content, filename) }
+      it 'create a tempfile that exists' do
+        subject.call do |temp_file|
+          @temp_file = temp_file
+          expect(File.exist?(@temp_file.path)).to eq true
+          expect(File.extname(@temp_file.path)).to include '.rb'
+        end
+        expect(@temp_file.path).to eq nil
+      end
+    end
+
+  end
+end

data/spec/lib/hydra/file_characterization_spec.rb

@@ -0,0 +1,65 @@
+require 'spec_helper'
+require 'hydra/file_characterization'
+require 'hydra/file_characterization/characterizer'
+
+module Hydra
+
+  describe '.characterize' do
+    let(:content) { "class Test; end\n" }
+    let(:filename) { 'test.rb' }
+    subject { Hydra.characterize(content, filename, tool_names) }
+
+    describe 'for fits' do
+      let(:tool_names) { :fits }
+      it { should match(/#{'<identity format="Plain text" mimetype="text/plain"'}/) }
+    end
+
+    describe 'for a bogus tool' do
+      let(:tool_names) { :cookie_monster }
+      it {
+        expect {
+          subject
+        }.to raise_error(Hydra::FileCharacterization::ToolNotFoundError)
+      }
+    end
+
+    describe 'for a mix of bogus and valid tools' do
+      let(:tool_names) { [:fits, :cookie_monster] }
+      it {
+        expect {
+          subject
+        }.to raise_error(Hydra::FileCharacterization::ToolNotFoundError)
+      }
+    end
+
+    describe 'for no tools' do
+      let(:tool_names) { nil }
+      it { should eq [] }
+    end
+
+  end
+
+  module FileCharacterization
+
+    describe Configuration do
+      subject { Configuration.new }
+      let (:expected_fits_path) {"string"}
+      before(:each) do
+        subject.tool_path = expected_fits_path
+      end
+      its(:config) {should have_key :tool_path}
+      its(:tool_path) {should == expected_fits_path}
+    end
+
+    describe 'preliminary integration' do
+      let (:tempfile) { ToTempFile.new("This is the content of the file.", 'test.rb')}
+      it '#call' do
+        tempfile.call do |f|
+          @fits_output = Characterizers::Fits.new(f.path ).call
+        end
+        expect(@fits_output).to include '<identity format="Plain text" mimetype="text/plain"'
+      end
+    end
+
+  end
+end

data/spec/spec_helper.rb

@@ -4,7 +4,15 @@
 # loaded once.
 #
 # See http://rubydoc.info/gems/rspec-core/RSpec/Core/Configuration
+
+module SpecSupport
+  def fixture_file(filename)
+    File.expand_path(File.join('../fixtures', filename), __FILE__)
+  end
+end
+
 RSpec.configure do |config|
+  config.include SpecSupport
   config.treat_symbols_as_metadata_keys_with_true_values = true
   config.run_all_when_everything_filtered = true
   config.filter_run :focus
@@ -14,4 +22,7 @@ RSpec.configure do |config|
   # the seed, which is printed after each run.
   #     --seed 1234
   config.order = 'random'
+  config.before(:suite) do
+    Hydra::FileCharacterization::Characterizers::Fits.tool_path = `which fits || which fits.sh`.strip
+  end
 end

metadata

@@ -1,7 +1,8 @@
 --- !ruby/object:Gem::Specification
 name: hydra-file_characterization
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.1.0
+  prerelease: 
 platform: ruby
 authors:
 - James Treacy
@@ -11,62 +12,70 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-09-18 00:00:00.000000000 Z
+date: 2013-09-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
-    - - '>='
+    - - ! '>='
       - !ruby/object:Gem::Version
         version: 3.0.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
-    - - '>='
+    - - ! '>='
       - !ruby/object:Gem::Version
         version: 3.0.0
 - !ruby/object:Gem::Dependency
   name: rspec
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
-    - - '>='
+    - - ! '>='
       - !ruby/object:Gem::Version
         version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
-    - - '>='
+    - - ! '>='
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
   name: guard
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
-    - - '>='
+    - - ! '>='
       - !ruby/object:Gem::Version
         version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
-    - - '>='
+    - - ! '>='
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
   name: guard-rspec
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
-    - - '>='
+    - - ! '>='
       - !ruby/object:Gem::Version
         version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
-    - - '>='
+    - - ! '>='
       - !ruby/object:Gem::Version
         version: '0'
 description: To provide a wrapper for file characterization
@@ -88,42 +97,48 @@ files:
 - Rakefile
 - hydra-file_characterization.gemspec
 - lib/hydra-file_characterization.rb
-- lib/hydra-file_characterization/characterizer.rb
-- lib/hydra-file_characterization/to_temp_file.rb
-- lib/hydra-file_characterization/version.rb
+- lib/hydra/file_characterization.rb
+- lib/hydra/file_characterization/characterizer.rb
+- lib/hydra/file_characterization/characterizers.rb
+- lib/hydra/file_characterization/characterizers/fits.rb
+- lib/hydra/file_characterization/exceptions.rb
+- lib/hydra/file_characterization/to_temp_file.rb
+- lib/hydra/file_characterization/version.rb
 - spec/fixtures/archive.zip
 - spec/fixtures/archive.zip.fits.xml
 - spec/fixtures/brendan_behan.jpeg
 - spec/fixtures/brendan_behan.jpeg.fits.xml
 - spec/fixtures/brendan_broken.dxxd
 - spec/fixtures/brendan_broken.dxxd.fits.xml
-- spec/lib/hydra-file_characterization/characterizer_spec.rb
-- spec/lib/hydra-file_characterization/to_temp_file_spec.rb
-- spec/lib/hydra-file_characterization_spec.rb
+- spec/lib/hydra/file_characterization/characterizers/fits_spec.rb
+- spec/lib/hydra/file_characterization/characterizers_spec.rb
+- spec/lib/hydra/file_characterization/to_temp_file_spec.rb
+- spec/lib/hydra/file_characterization_spec.rb
 - spec/spec_helper.rb
-homepage: https://github.com/projecthydra/hydra-file_characterization
+homepage: https://github.com/projecthydra/hydra/file_characterization
 licenses:
 - APACHE2
-metadata: {}
 post_install_message: 
 rdoc_options: []
 require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
   requirements:
-  - - '>='
+  - - ! '>='
     - !ruby/object:Gem::Version
       version: '0'
 required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
   requirements:
-  - - '>='
+  - - ! '>='
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.0.3
+rubygems_version: 1.8.23
 signing_key: 
-specification_version: 4
+specification_version: 3
 summary: To provide a wrapper for file characterization
 test_files:
 - spec/fixtures/archive.zip
@@ -132,7 +147,8 @@ test_files:
 - spec/fixtures/brendan_behan.jpeg.fits.xml
 - spec/fixtures/brendan_broken.dxxd
 - spec/fixtures/brendan_broken.dxxd.fits.xml
-- spec/lib/hydra-file_characterization/characterizer_spec.rb
-- spec/lib/hydra-file_characterization/to_temp_file_spec.rb
-- spec/lib/hydra-file_characterization_spec.rb
+- spec/lib/hydra/file_characterization/characterizers/fits_spec.rb
+- spec/lib/hydra/file_characterization/characterizers_spec.rb
+- spec/lib/hydra/file_characterization/to_temp_file_spec.rb
+- spec/lib/hydra/file_characterization_spec.rb
 - spec/spec_helper.rb

checksums.yaml

@@ -1,7 +0,0 @@
----
-SHA1:
-  metadata.gz: 3042462b8a14d3ec133d9ccc86fa20663fe4260d
-  data.tar.gz: 12e6b321515dfdcbf5621d107bf8b9a28ea0a597
-SHA512:
-  metadata.gz: 53c9607e32ab504221c9f8752450af260ac109ebd497d85c546e5fd35dc22293169b16ddef8884351a34e4a1759636a45461a8172f8ac774d867a57d09080568
-  data.tar.gz: 4b1bdb304e398b2f78e5ce039d94e666373d6f247d11d7912fba2db35ac51f1daed7dcad21924df53be7fa875e9048aab9091f74c21cbc50816ad1277c826309

data/spec/lib/hydra-file_characterization/to_temp_file_spec.rb

@@ -1,20 +0,0 @@
-require 'spec_helper'
-require 'hydra-file_characterization/to_temp_file'
-
-describe Hydra::FileCharacterization do
-
-  describe 'ToTempFile' do
-    let(:string) { "This is the content of the file." }
-    subject { Hydra::FileCharacterization::ToTempFile.new(string, "hello.rb") }
-
-    it 'create a tempfile that exists' do
-      @temp_file = ''
-      subject.call do |temp_file|
-        @temp_file = temp_file
-        expect(File.exist?(@temp_file)).to eq true
-        expect(File.extname(@temp_file)).to eq '.rb'
-      end
-      expect(File.exist?(@temp_file)).to eq false
-    end
-  end
-end

data/spec/lib/hydra-file_characterization_spec.rb

@@ -1,18 +0,0 @@
-require 'spec_helper'
-require 'hydra-file_characterization'
-
-describe Hydra::FileCharacterization do
-
-
-
-  describe 'config' do
-    subject { Hydra::FileCharacterization::Configuration.new }
-    let (:expected_fits_path) {"string"}
-    before(:each) do
-      subject.fits_path = expected_fits_path
-    end
-    its(:config) {should have_key :fits_path}
-    its(:fits_path) {should == expected_fits_path}
-  end
-
-end