hybag 0.0.3 → 0.0.4

checksums.yaml

@@ -1,15 +1,15 @@
 ---
 !binary "U0hBMQ==":
   metadata.gz: !binary |-
-    MTAzOTQzMzU5ZjQ4ZmNlZmE1NzI4MTViMWMwNjA2ZGZmOWMxYTUxMA==
+    OTA5YWU4ZGE4YzliMDdkMmU0NzE1YjdmYTQwMTA5MjhhYTExNGFkMQ==
   data.tar.gz: !binary |-
-    NGJmYWUyNDI1MTYwMDBiZTUyNDk1MjkxNTRlNjU0NTNlM2E5NGJmNA==
+    YjE1ZjI2NGQ1MmY2ZTdhNDgwMjVhMmUwNGIwNzZkZTM1NWVjNjQzMg==
 !binary "U0hBNTEy":
   metadata.gz: !binary |-
-    NDExZGZlZDVjOGJlOWE0Mzg3MGM4YTY3NDk3NWY0YTdkMWE0MzUwZDJhNzIy
-    YWQ1ZGFiNTMxMDBiZDE2MjM1NmI3M2U2MjI0YWMzZjAxYmQxNDM3MjdjMmM2
-    M2U0ZjU3MTVhMWM2OTk5MDMwODUxZjhkMTFlMWU1NGIwYWY4MTI=
+    MjM1NDAxMmQ3MDliZTk4Y2U4ZTBiYmQxMTNmODk2MjdlYzQ2NmQ3ZTk3ZDQ3
+    ZjcyZDc3MzEwOTgxZjkzNGNhYjA3MjFhOWM4ZWNmZjM1OGViOTU4ODBiMTQ4
+    NDA4YmZmYTIxMGQ1MjNmMTQyZTIzYzc3MWEwODU1MzZlOTNkOGQ=
   data.tar.gz: !binary |-
-    ODgyYTI1NjcyN2Q3MTk5OTFhM2FlNjViNmMzNmEwMzJmMGFjYjM5YTA4ZTgw
-    ZGY2ZjcyYjMxMzIwNDU0MmQxMDBlMDAyZGE0MjZlNThhYTVlYmM3Y2Y0N2Q2
-    YWYzYzk1NzFlZjdkMjFhMWUyMGZjNWVjYmM4ZTFkMGJhYTZmZjc=
+    NTQ5NDVmYzcxYzFmNjgyNzM4YWM5Njk4MmEyZTRhZjg2ZDlhYTRhNzE1Yzhm
+    ZWU3YzZiOGMyMmY1MzBhMmNmNjMwYzAxNGMwYmJmZDNmNmUzM2I1MjBjZmQ3
+    YTcwNWY1ZDhjY2M1ODFjODY2ZDA0MjkwZWQ5NjI4N2FlNWI0ODA=

data/Gemfile

@@ -1,4 +1,6 @@
 source 'https://rubygems.org'
 
+
+gem 'pry', '0.9.11.4'
 # Specify your gem's dependencies in hybag.gemspec
 gemspec

data/README.md

@@ -18,7 +18,7 @@ Or install it yourself as:
 
 ## Usage
 
-### Include the module in ActiveFedora models you'd like to be baggable
+#### Include the module in ActiveFedora models you'd like to be baggable
 
 ```ruby
   class TestClass < ActiveFedora::Base
@@ -26,6 +26,30 @@ Or install it yourself as:
   end
 ```
 
+#### To convert an exported bag back to a model
+
+**NOTE:** Right now for this to work there must be datastreams defined on the discovered model which match the
+metadata datastream IDs as tag files and content datastream IDs as data files. This means for a descMetadata
+datastream to be populated bag_root/descMetadata.* (where * is the extension) must exist.
+
+```ruby
+  result = Hybag.ingest(BagIt::Bag.new("/path/to/bag"))
+  result.class # => Model in fedora/rels-ext.rdf (preferred) or hybag.yml in bag root. More info below.
+  result.persisted? # => false
+```
+
+## Configuration
+
+#### Determining the model name.
+
+Currently the model name is determined from the bag's fedora/rels-ext.rdf file (which Hybag::Baggable exports)
+or a config file stored in bag_root/hybag.yml. The fedora rels-ext takes precedence. An example Hybag.yml format is
+below
+
+```yml
+model: TestModel
+```
+
 ### Examples
 
 #### Write the item to disk in rails_root/tmp/bags/filler/pid

data/hybag.gemspec

@@ -27,6 +27,6 @@ Gem::Specification.new do |spec|
 
   spec.add_dependency 'active-fedora'
   spec.add_dependency 'activesupport', '>= 3.2.0', '< 5.0'
-  spec.add_dependency 'bagit', '0.3.2'
+  spec.add_dependency 'bagit'
   spec.add_dependency 'mime-types'
 end

data/lib/hybag/ingester.rb

@@ -0,0 +1,111 @@
+require 'rdf/rdfxml'
+
+module Hybag
+  class Ingester
+    attr_accessor :bag
+    def initialize(bag)
+      @bag = bag
+    end
+
+    def ingest
+      raise "Unable to determine model from bag" if model_name.blank?
+      new_object = ActiveFedora.class_from_string(model_name.to_s).new
+      # Assign a pid
+      new_object.inner_object.pid = ActiveFedora::Base.assign_pid(new_object)
+      set_metadata_streams(new_object)
+      set_file_streams(new_object)
+      return new_object
+    end
+
+    private
+
+    # TODO: What to do if the bag has files that don't have model definitions?
+    # TODO: Add some sort of configuration to map bag filenames -> dsids.
+    def set_metadata_streams(object)
+      object.metadata_streams.each do |ds|
+        if bag_has_metastream?(ds.dsid)
+          ds.content = bag_metastream(ds.dsid).read.strip
+          # Assume the first subject in the metadata is about this object.
+          # TODO: Move this to configuration?
+          first_subject = ds.graph.first_subject
+          new_repository = RDF::Repository.new
+          ds.graph.each_statement do |statement|
+            subject = statement.subject
+            subject = ds.rdf_subject if subject == first_subject
+            new_repository << [subject, statement.predicate, statement.object]
+          end
+          ds.instance_variable_set(:@graph,new_repository)
+        end
+      end
+    end
+
+    def set_file_streams(object)
+      file_streams = object.datastreams.select{|k, ds| !ds.metadata?}.values
+      file_streams.each do |ds|
+        if bag_has_datastream?(ds.dsid)
+          ds.content = bag_datastream(ds.dsid).read
+        end
+      end
+    end
+
+    # TODO: Might consider decoration at some point.
+    def bag_filename_to_label(bag_filename)
+      Pathname.new(bag_filename).basename.sub_ext('').to_s
+    end
+
+    def bag_has_datastream?(label)
+      bag.bag_files.any?{|x| bag_filename_to_label(x) == label}
+    end
+
+    def bag_datastream(label)
+      bag_file = bag.bag_files.select{|x| bag_filename_to_label(x) == label}.first
+      result = File.open(bag_file) unless bag_file.blank?
+      return result
+    end
+
+    def bag_has_metastream?(label)
+      bag.tag_files.any?{|x| bag_filename_to_label(x) == label}
+    end
+
+    def bag_metastream(label)
+      tag_file = bag.tag_files.select{|x| bag_filename_to_label(x) == label}.first
+      result = File.open(tag_file) unless tag_file.blank?
+      return result
+    end
+
+    def model_name
+      # TODO: Add a default model_name configuration option?
+      @model_name ||= extract_model_from_rels || extract_model_from_yaml
+
+    end
+
+    def extract_model_from_rels
+      if File.exist?(fedora_rels)
+        filler_object = ActiveFedora::Base.new
+        rels_datastream = ActiveFedora::RelsExtDatastream.new
+        rels_datastream.model = filler_object
+        ActiveFedora::RelsExtDatastream.from_xml(File.read(fedora_rels).strip,rels_datastream)
+        model_name = ActiveFedora::ContentModel.known_models_for(filler_object).first
+        return model_name.to_s
+      end
+      return model_name
+    end
+
+    def extract_model_from_yaml
+      model_name = nil
+      if(File.exist?(yaml_config))
+        conf = YAML.load(File.read(yaml_config))
+        model_name = conf['model']
+      end
+      return model_name
+    end
+
+    def yaml_config
+      File.join(bag.bag_dir,"hybag.yml")
+    end
+
+    def fedora_rels
+      File.join(bag.bag_dir,"fedora","RELS-EXT.rdf")
+    end
+  end
+end

data/lib/hybag/version.rb

@@ -1,3 +1,3 @@
 module Hybag
-  VERSION = "0.0.3"
+  VERSION = "0.0.4"
 end

data/lib/hybag.rb

@@ -6,7 +6,10 @@ require 'hybag/version'
 require 'hybag/baggable'
 require 'hybag/validator'
 require 'hybag/bag_writer'
+require 'hybag/ingester'
 
 module Hybag
-  
+  def self.ingest(bag)
+    Hybag::Ingester.new(bag).ingest
+  end
 end

data/spec/fixtures/example_datastream.nt

@@ -1,13 +1,13 @@
-<http://oregondigital.org/resource/oregondigital:61> <http://purl.org/dc/terms/type> "Image" .
-<http://oregondigital.org/resource/oregondigital:61> <http://purl.org/dc/terms/spatial> "Benton County (Ore.)" .
-<http://oregondigital.org/resource/oregondigital:61> <http://purl.org/dc/terms/created> "2004-03-17" .
-<http://oregondigital.org/resource/oregondigital:61> <http://purl.org/dc/terms/description> "Mexican workers in Benton County" .
-<http://oregondigital.org/resource/oregondigital:61> <http://purl.org/dc/terms/rights> "Permission to use must be obtained from OSU Archives." .
-<http://oregondigital.org/resource/oregondigital:61> <http://multimedialab.elis.ugent.be/users/samcoppe/ontologies/Premis/premis.owl#originalName> "P0120_2567" .
-<http://oregondigital.org/resource/oregondigital:61> <http://purl.org/dc/terms/title> "Mexican workers" .
-<http://oregondigital.org/resource/oregondigital:61> <http://purl.org/dc/terms/subject> "Agricultural laborers--Mexican--Oregon" .
-<http://oregondigital.org/resource/oregondigital:61> <http://purl.org/dc/terms/subject> "Agricultural laborers--Housing--Oregon" .
-<http://oregondigital.org/resource/oregondigital:61> <http://purl.org/dc/terms/identifier> "P120:2567" .
-<http://oregondigital.org/resource/oregondigital:61> <http://purl.org/dc/terms/modified> "2011-12-19" .
-<http://oregondigital.org/resource/oregondigital:61> <http://purl.org/dc/terms/hasFormat> "B&W print" .
-<http://oregondigital.org/resource/oregondigital:61> <http://purl.org/dc/terms/date> "1944" .
+<info:fedora/filler> <http://purl.org/dc/terms/type> "Image" .
+<info:fedora/filler> <http://purl.org/dc/terms/spatial> "Benton County (Ore.)" .
+<info:fedora/filler> <http://purl.org/dc/terms/created> "2004-03-17" .
+<info:fedora/filler> <http://purl.org/dc/terms/description> "Mexican workers in Benton County" .
+<info:fedora/filler> <http://purl.org/dc/terms/rights> "Permission to use must be obtained from OSU Archives." .
+<info:fedora/filler> <http://multimedialab.elis.ugent.be/users/samcoppe/ontologies/Premis/premis.owl#originalName> "P0120_2567" .
+<info:fedora/filler> <http://purl.org/dc/terms/title> "Mexican workers" .
+<info:fedora/filler> <http://purl.org/dc/terms/subject> "Agricultural laborers--Mexican--Oregon" .
+<info:fedora/filler> <http://purl.org/dc/terms/subject> "Agricultural laborers--Housing--Oregon" .
+<info:fedora/filler> <http://purl.org/dc/terms/identifier> "P120:2567" .
+<info:fedora/filler> <http://purl.org/dc/terms/modified> "2011-12-19" .
+<info:fedora/filler> <http://purl.org/dc/terms/hasFormat> "B&W print" .
+<info:fedora/filler> <http://purl.org/dc/terms/date> "1944" .

data/spec/fixtures/hybag.yml

@@ -0,0 +1 @@
+model: TestModel

data/spec/fixtures/rels.rdf

@@ -0,0 +1,6 @@
+
+<rdf:RDF xmlns:ns0="info:fedora/fedora-system:def/model#" xmlns:ns1="info:fedora/fedora-system:def/relations-external#" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
+  <rdf:Description rdf:about="info:fedora/filler">
+    <ns0:hasModel rdf:resource="info:fedora/afmodel:BaggableDummy"></ns0:hasModel>
+  </rdf:Description>
+</rdf:RDF>

data/spec/lib/hybag/baggable_spec.rb

@@ -9,6 +9,8 @@ describe Hybag::Baggable do
     @item.descMetadata.content = File.open(File.join(FIXTURE_PATH,"example_datastream.nt")).read
     content_file = File.open(File.join(FIXTURE_PATH,"hydra.png"))
     @item.add_file_datastream(content_file, :dsid => "content", :mimeType => "image/png")
+    @item.rels_ext.content = File.open(File.join(FIXTURE_PATH,"rels.rdf")).read
+    @item.load_relationships
     @item.stub(:persisted?).and_return(true)
     # Stub Rails root
     rails = double("Rails")
@@ -45,8 +47,7 @@ describe Hybag::Baggable do
         @bag = @item.write_bag('/tmp/bag')
       end
       it "should write tag files to disk" do
-        # TODO: stub out a rels-ext to make sure this happens.
-        #@bag.tag_files.should include(File.join(@bag.bag_dir, 'fedora', 'RELS-EXT.rdf'))
+        @bag.tag_files.should include(File.join(@bag.bag_dir, 'fedora', 'RELS-EXT.rdf'))
         @bag.tag_files.should include(File.join(@bag.bag_dir, 'descMetadata.nt'))
       end
       it "should write content files to disk" do

data/spec/lib/hybag/ingester_spec.rb

@@ -0,0 +1,119 @@
+require 'spec_helper'
+require 'open-uri'
+require 'pry'
+require File.join(DUMMY_PATH, "baggable_dummy")
+
+# Override Kernel.open
+module ::Kernel
+  private
+
+  alias :original_open :open
+  def open(name, *rest, &block)
+    if name =~ /\|.*/ then
+      original_open(name, *rest, &block)
+    else
+      FakeFS::File.open(name, *rest, &block)
+    end
+  end
+  module_function :open
+end
+
+describe Hybag::Ingester do
+  before(:each) do
+    @item = BaggableDummy.new(:pid => "filler")
+    @item.descMetadata.content = File.open(File.join(FIXTURE_PATH,"example_datastream.nt")).read.strip
+    content_file = File.open(File.join(FIXTURE_PATH,"hydra.png"))
+    @item.add_file_datastream(content_file, :dsid => "content", :mimeType => "image/png")
+    @item.rels_ext.content = File.open(File.join(FIXTURE_PATH,"rels.rdf")).read
+    @item.load_relationships
+    @item.stub(:persisted?).and_return(true)
+    # Stub Rails root
+    rails = double("Rails")
+    Rails = rails unless defined?(Rails)
+    Rails.stub(:root).and_return(Pathname.new('/test'))
+    # Stub ActiveFedora assign_pid
+    ActiveFedora::Base.stub(:assign_pid).and_return("new_filler")
+  end
+  let(:item) {@item}
+  let(:bag) {item.write_bag}
+  subject {Hybag::Ingester.new(bag)}
+  describe "functionality" do
+    include FakeFS::SpecHelpers
+    describe ".model_name" do
+      context "when there is a rels" do
+        it "should return the model name stored by the bag in rels" do
+          subject.send(:model_name).should == "BaggableDummy"
+        end
+      end
+      context "when there is no rels" do
+        context "when there is a hybag.yml" do
+          before(:each) do
+            FileUtils.rm(subject.send(:fedora_rels), :force => true)
+            # Add the hybag.yml from fixture
+            FakeFS.deactivate!
+            hybag_content = File.read(File.join(FIXTURE_PATH,"hybag.yml"))
+            FakeFS.activate!
+            File.open(File.join(bag.bag_dir,"hybag.yml"),'w') {|f| f.puts hybag_content}
+          end
+          it "should pull the data out of hybag.yml" do
+            subject.send(:model_name).should == "TestModel"
+          end
+        end
+      end
+      context "when there is a rels and a hybag.yml" do
+        before(:each) do
+          # Add the hybag.yml from fixture
+          FakeFS.deactivate!
+          hybag_content = File.read(File.join(FIXTURE_PATH,"hybag.yml"))
+          FakeFS.activate!
+          File.open(File.join(bag.bag_dir,"hybag.yml"),'w') {|f| f.puts hybag_content}
+        end
+        it "should prioritize the rels" do
+          subject.send(:model_name).should == "BaggableDummy"
+        end
+      end
+    end
+    describe ".ingest" do
+      context "when there is no discoverable model" do
+        before(:each) do
+          subject.stub(:model_name).and_return(nil)
+        end
+        it "should raise an error" do
+          expect{subject.ingest}.to raise_error("Unable to determine model from bag")
+        end
+      end
+      context "when there is a model" do
+        it "should return an instance of that model" do
+          expect(subject.ingest).to be_kind_of(BaggableDummy)
+        end
+        it "should assign that model a pid" do
+          puts ActiveFedora::Base.assign_pid
+          expect((subject.ingest).pid).to eq "new_filler"
+        end
+        context "and that model is not defined" do
+          before(:each) do
+            subject.stub(:model_name).and_return("Blabla")
+          end
+          it "should raise an error" do
+            expect{subject.ingest}.to raise_error
+          end
+        end
+      end
+      describe "returned model" do
+        let(:built_model) {subject.ingest}
+        it "should populate datastreams" do
+          Array.wrap(built_model.title).first.should == "Mexican workers"
+        end
+        it "should populate file datastreams" do
+          built_model.content.content.should == File.read(File.join(bag.data_dir,"content.png"))
+        end
+        it "should be new" do
+          expect(built_model).to be_new
+        end
+        it "should not be persisted" do
+          expect(built_model).not_to be_persisted
+        end
+      end
+    end
+  end
+end

data/spec/lib/hybag_spec.rb

@@ -0,0 +1,10 @@
+require 'spec_helper'
+
+describe Hybag do
+  describe ".ingest" do
+    it "should call Hybag::Ingester.ingest" do
+      Hybag::Ingester.any_instance.should_receive(:ingest).and_return("bla")
+      Hybag.ingest("empty")
+    end
+  end
+end

data/spec/spec_helper.rb

@@ -2,7 +2,6 @@ require 'rspec/autorun'
 ENV['environment'] ||= 'test'
 # - RSpec adds ./lib to the $LOAD_PATH
 require 'hybag'
-require 'pry'
 #Resque.inline = Rails.env.test?
 ROOT_PATH = File.dirname(__FILE__)
 DUMMY_PATH = File.join(ROOT_PATH,"dummies")

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: hybag
 version: !ruby/object:Gem::Version
-  version: 0.0.3
+  version: 0.0.4
 platform: ruby
 authors:
 - Trey Terrell
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-08-07 00:00:00.000000000 Z
+date: 2013-08-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -132,16 +132,16 @@ dependencies:
   name: bagit
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '='
+    - - ! '>='
       - !ruby/object:Gem::Version
-        version: 0.3.2
+        version: '0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '='
+    - - ! '>='
       - !ruby/object:Gem::Version
-        version: 0.3.2
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: mime-types
   requirement: !ruby/object:Gem::Requirement
@@ -173,13 +173,18 @@ files:
 - lib/hybag.rb
 - lib/hybag/bag_writer.rb
 - lib/hybag/baggable.rb
+- lib/hybag/ingester.rb
 - lib/hybag/validator.rb
 - lib/hybag/version.rb
 - spec/dummies/baggable_dummy.rb
 - spec/fixtures/example_datastream.nt
+- spec/fixtures/hybag.yml
 - spec/fixtures/hydra.png
+- spec/fixtures/rels.rdf
 - spec/lib/hybag/baggable_spec.rb
+- spec/lib/hybag/ingester_spec.rb
 - spec/lib/hybag/validator_spec.rb
+- spec/lib/hybag_spec.rb
 - spec/spec_helper.rb
 - spec/support/fakefs.rb
 homepage: ''
@@ -209,8 +214,12 @@ summary: A Hydra gem for adding BagIt functionality to ActiveFedora models.
 test_files:
 - spec/dummies/baggable_dummy.rb
 - spec/fixtures/example_datastream.nt
+- spec/fixtures/hybag.yml
 - spec/fixtures/hydra.png
+- spec/fixtures/rels.rdf
 - spec/lib/hybag/baggable_spec.rb
+- spec/lib/hybag/ingester_spec.rb
 - spec/lib/hybag/validator_spec.rb
+- spec/lib/hybag_spec.rb
 - spec/spec_helper.rb
 - spec/support/fakefs.rb