hydra-works 0.2.0 → 0.3.0

data/lib/hydra/works/services/generic_file/full_text_extraction_service.rb

@@ -1,57 +0,0 @@
-module Hydra::Works
-  # Extract the full text from the content using Solr's extract handler
-  class FullTextExtractionService
-    def self.run(generic_file)
-      new(generic_file).extract
-    end
-
-    delegate :original_file, :id, to: :@generic_file
-
-    def initialize(generic_file)
-      @generic_file = generic_file
-    end
-
-    ##
-    # Extract full text from the content using Solr's extract handler.
-    # This will extract text from the file uploaded to generic_file.
-    # The file uploaded to @generic_file can be accessed via :original_file.
-    #
-    # @return [String] The extracted text
-    def extract
-      JSON.parse(fetch)[''].rstrip
-    rescue Hydra::Works::FullTextExtractionError => e
-      raise e
-    rescue => e
-      raise Hydra::Works::FullTextExtractionError.new, "Error extracting content from #{id}: #{e.inspect}"
-    end
-
-    # send the request to the extract service and return the response if it was successful.
-    # TODO: this pulls the whole file into memory. We should stream it from Fedora instead
-    # @return [String] the result of calling the extract service
-    def fetch
-      req = Net::HTTP.new(uri.host, uri.port)
-      resp = req.post(uri.to_s, original_file.content, request_headers)
-      raise Hydra::Works::FullTextExtractionError.new, "Solr Extract service was unsuccessful. '#{uri}' returned code #{resp.code} for #{id}\n#{resp.body}" unless resp.code == '200'
-      original_file.content.rewind if original_file.content.respond_to?(:rewind)
-
-      resp.body
-    end
-
-    # @return [Hash] the request headers to send to the Solr extract service
-    def request_headers
-      { Faraday::Request::UrlEncoded::CONTENT_TYPE => "#{original_file.mime_type};charset=utf-8",
-        Faraday::Adapter::CONTENT_LENGTH => original_file.size.to_s }
-    end
-
-    # @returns [URI] path to the extract service
-    def uri
-      @uri ||= URI("#{connection_url}/update/extract?extractOnly=true&wt=json&extractFormat=text")
-    end
-
-    private
-
-      def connection_url
-        ActiveFedora.solr_config[:url]
-      end
-  end
-end

data/lib/hydra/works/services/generic_file/generate_thumbnail.rb

@@ -1,13 +0,0 @@
-module Hydra::Works
-  class GenerateThumbnail
-    def self.call(object, content: :original_file)
-      fail ArgumentError, "object has no content at #{content} from which to generate a thumbnail" if object.send(content).nil?
-
-      # Always replace the thumbnail with whatever is from the original file
-      object.build_thumbnail if object.thumbnail.nil?
-
-      object.create_derivatives
-      object
-    end
-  end
-end

data/lib/hydra/works/services/generic_file/persist_derivative.rb

@@ -1,20 +0,0 @@
-require 'hydra/derivatives'
-
-module Hydra::Works
-  class PersistDerivative < Hydra::Derivatives::PersistOutputFileService
-    ##
-    # Persists a derivative to a GenericFile
-    # This Service conforms to the signature of `Hydra::Derivatives::PersistOutputFileService`.
-    # The purpose of this Service is for use as an alternative to the default Hydra::Derivatives::PersistOutputFileService.  It's necessary because the default behavior in Hydra::Derivatives assumes that you're using LDP Basic Containment. Hydra::Works::GenericFiles use IndirectContainment.  This Service handles that case.
-    # This service will always update existing and does not do versioning of persisted files.
-    #
-    # @param [Hydra::Works::GenericFile::Base] object the file will be added to
-    # @param [Hydra::Derivatives::IoDecorator] file the derivative filestream
-    # @param [String] extract file type symbol (e.g. :thumbnail) from Hydra::Derivatives created destination_name
-
-    def self.call(object, file, destination_name)
-      type = destination_name.gsub(/^original_file_/, '').to_sym
-      Hydra::Works::AddFileToGenericFile.call(object, file, type, update_existing: true, versioning: false)
-    end
-  end
-end

data/spec/hydra/works/models/concerns/block_child_objects_spec.rb

@@ -1,17 +0,0 @@
-require 'spec_helper'
-
-describe Hydra::Works::BlockChildObjects do
-  subject { Hydra::Works::GenericFile::Base.new }
-
-  describe '#objects=?' do
-    it 'raises an error' do
-      expect { subject.objects = [] }.to raise_error(StandardError, /method `objects=' not allowed for #<Hydra::Works::GenericFile::Base.*/)
-    end
-  end
-
-  describe '#objects' do
-    it 'raises an error' do
-      expect { subject.objects }.to raise_error(StandardError, /method `objects' not allowed for #<Hydra::Works::GenericFile::Base.*/)
-    end
-  end
-end

data/spec/hydra/works/models/concerns/generic_file/versioned_content_spec.rb

@@ -1,32 +0,0 @@
-require 'spec_helper'
-
-describe Hydra::Works::GenericFile::VersionedContent do
-  let(:generic_file) { Hydra::Works::GenericFile::Base.new }
-  before do
-    Hydra::Works::UploadFileToGenericFile.call(generic_file, File.open(File.join(fixture_path, 'sample-file.pdf')))
-    Hydra::Works::UploadFileToGenericFile.call(generic_file, File.open(File.join(fixture_path, 'updated-file.txt')))
-  end
-
-  describe 'content_versions' do
-    subject { generic_file.content_versions }
-    it 'lists all of the versions of original_file' do
-      expect(subject.count).to eq(2)
-      expect(subject.map(&:uri)).to eq(generic_file.original_file.versions.all.map(&:uri))
-    end
-  end
-
-  describe 'latest_content_version' do
-    subject { generic_file.latest_content_version }
-    it 'returns the most recent version entry for original_file' do
-      # Can't use a simple equivalence because they are actually different ResourceVersion objects
-      expect(subject.uri).to eq(generic_file.original_file.versions.last.uri)
-      expect(subject.label).to eq(generic_file.original_file.versions.last.label)
-    end
-  end
-
-  describe 'current_content_version_uri' do
-    it 'returns the URI of the most recent version of original_file' do
-      expect(generic_file.current_content_version_uri).to eq(generic_file.original_file.versions.last.uri)
-    end
-  end
-end

data/spec/hydra/works/models/concerns/generic_file_behavior_spec.rb

@@ -1,12 +0,0 @@
-require 'spec_helper'
-
-describe Hydra::Works::GenericFileBehavior do
-  class IncludesGenericFileBehavior < ActiveFedora::Base
-    include Hydra::Works::GenericFileBehavior
-  end
-  subject { IncludesGenericFileBehavior.new }
-
-  it 'ensures that objects will be recognized as generic_files' do
-    expect(subject).to be_works_generic_file
-  end
-end

data/spec/hydra/works/services/full_text_extraction_service_spec.rb

@@ -1,89 +0,0 @@
-require 'spec_helper'
-
-describe Hydra::Works::FullTextExtractionService do
-  let(:generic_file) { Hydra::Works::GenericFile::Base.new }
-
-  describe 'integration test' do
-    before do
-      Hydra::Works::UploadFileToGenericFile.call(generic_file, File.open(File.join(fixture_path, 'sample-file.pdf')))
-    end
-    subject { described_class.run(generic_file) }
-    it 'extracts fulltext and stores the results' do
-      expect(subject).to include('This is some original content')
-    end
-  end
-
-  describe "run" do
-    let(:generic_file) { double(id: '123') }
-    subject { described_class.run(generic_file) }
-
-    context "when it is successful" do
-      before do
-        allow_any_instance_of(described_class).to receive(:fetch).and_return('{"":"one two three"}')
-      end
-      it { is_expected.to eq 'one two three' }
-    end
-
-    context "when solr raises an error" do
-      before do
-        allow_any_instance_of(described_class).to receive(:fetch).and_raise(Hydra::Works::FullTextExtractionError.new, "Solr failed")
-      end
-      it "raises an error" do
-        expect { subject }.to raise_error Hydra::Works::FullTextExtractionError, 'Solr failed'
-      end
-    end
-
-    context "network error" do
-      before do
-        allow_any_instance_of(described_class).to receive(:fetch).and_raise(Errno::ECONNRESET)
-      end
-      it "raises an error" do
-        expect { subject }.to raise_error Hydra::Works::FullTextExtractionError, 'Error extracting content from 123: #<Errno::ECONNRESET: Connection reset by peer>'
-      end
-    end
-  end
-
-  describe "fetch" do
-    let(:generic_file) { double('generic file', id: '123', original_file: original) }
-    let(:original) { double(content: content, size: 13, mime_type: 'text/plain') }
-    let(:service) { described_class.new(generic_file) }
-    subject { service.fetch }
-    let(:request) { double }
-    let(:response_body) { 'returned by Solr' }
-    let(:resp) { double(code: '200', body: response_body) }
-    let(:uri) { URI('http://example.com:99/solr/update') }
-    let(:content) { 'file contents' }
-
-    before do
-      allow(service).to receive(:uri).and_return(URI('http://example.com:99/solr/update'))
-      allow(Net::HTTP).to receive(:new).with('example.com', 99).and_return(request)
-    end
-
-    context "that is successful" do
-      let(:resp) { double(code: '200', body: response_body) }
-      it "calls the extraction service" do
-        expect(request).to receive(:post).with('http://example.com:99/solr/update', content, "Content-Type" => "text/plain;charset=utf-8", "Content-Length" => "13").and_return(resp)
-        expect(subject).to eq response_body
-      end
-    end
-
-    context "that fails" do
-      let(:resp) { double(code: '500', body: response_body) }
-      it "raises an error" do
-        expect(request).to receive(:post).with('http://example.com:99/solr/update', content, "Content-Type" => "text/plain;charset=utf-8", "Content-Length" => "13").and_return(resp)
-        expect { subject }.to raise_error Hydra::Works::FullTextExtractionError, "Solr Extract service was unsuccessful. 'http://example.com:99/solr/update' returned code 500 for 123\nreturned by Solr"
-      end
-    end
-  end
-
-  describe "uri" do
-    let(:generic_file) { double }
-    let(:service) { described_class.new(generic_file) }
-    subject { service.uri }
-
-    it "points at the extraction service" do
-      expect(subject).to be_kind_of URI
-      expect(subject.to_s).to end_with '/update/extract?extractOnly=true&wt=json&extractFormat=text'
-    end
-  end
-end

data/spec/hydra/works/services/generic_file/generate/thumbnail_spec.rb

@@ -1,19 +0,0 @@
-require 'spec_helper'
-
-describe Hydra::Works::GenerateThumbnail do
-  context 'when the object has no original file' do
-    let(:error_message) { 'object has no content at original_file from which to generate a thumbnail' }
-    let(:object) { double('object', original_file: nil) }
-    it 'raises an error' do
-      expect(-> { described_class.call(object) }).to raise_error(ArgumentError, error_message)
-    end
-  end
-
-  context 'when the object has no content at specified location' do
-    let(:error_message) { 'object has no content at my_location from which to generate a thumbnail' }
-    let(:object) { double('object', my_location: nil) }
-    it 'raises an error' do
-      expect(-> { described_class.call(object, content: :my_location) }).to raise_error(ArgumentError, error_message)
-    end
-  end
-end