research_metadata 1.1.2 → 2.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 7c8791a6102b1a09b52ca6fc0860c5151b08d593
-  data.tar.gz: 9f683d6ea78326b1808d1b01f8c7a6cc354d06e4
+  metadata.gz: 7861dd7bacfce9d729ef20167e9864a8b15d6143
+  data.tar.gz: 920b37ed1c76a2a20eed3299cd73767bc90b7511
 SHA512:
-  metadata.gz: 0a650d857b31760ad85cef4585d4904cd575d73207e3b3650b6db8d3a36f408ed6e84b6005f2dc1a51b0624932f1de523b159ae40e291a4b27d56bb1d74bd513
-  data.tar.gz: 6382e9a4e827654f41537917444a2c2fa6c143b3ffbafd5e6ccb15a88655fe9b5c2ac49b291369d8528194c5e6cf9eb443b8710dcedff598550071d46324bb54
+  metadata.gz: e8f6fcfd79df3cef1b04b6b311037aef402b4428628e55f42c1ace8c6891f113b7b5e361b76d79aa6822628193c366e8c30cf9aec6328dd15cfcfbc1e0edc71f
+  data.tar.gz: 9d0fe4acd36a77025ef170efc8584508fba4fbf272f973e649a10742d719112d7ec59bd142f8f458f84d808aebbb41fe19e1c36e2e1b5efa30889baad714a127

data/CHANGELOG.md

@@ -2,7 +2,13 @@
 All notable changes to this project will be documented in this file.
 This project adheres to [Semantic Versioning](http://semver.org/).
 
-## Unreleased
+## 2.0.0 2017-12-20
+### Added
+- Research output - translated title.
+- Testing against known live data.
+
+### Changed
+- For Pure API 59.
 
 ## 1.1.2 2017-07-12
 ### Fixed

data/README.md

@@ -5,9 +5,7 @@ Metadata extraction from the Pure Research Information System and transformation
 ## Status
 
 [![Gem Version](https://badge.fury.io/rb/research_metadata.svg)](https://badge.fury.io/rb/research_metadata)
-[![Build Status](https://semaphoreci.com/api/v1/aalbinclark/research_metadata/branches/master/badge.svg)](https://semaphoreci.com/aalbinclark/research_metadata)
-[![Code Climate](https://codeclimate.com/github/lulibrary/research_metadata/badges/gpa.svg)](https://codeclimate.com/github/lulibrary/research_metadata)
-[![Dependency Status](https://www.versioneye.com/user/projects/5899d1be1e07ae0048c8e4c6/badge.svg?style=flat-square)](https://www.versioneye.com/user/projects/5899d1be1e07ae0048c8e4c6)
+[![Maintainability](https://api.codeclimate.com/v1/badges/26080a93183c7fcdf525/maintainability)](https://codeclimate.com/github/lulibrary/research_metadata/maintainability)
 
 ## Installation
 
@@ -30,24 +28,17 @@ Or install it yourself as:
 Create a hash for passing to a transformer.
 
 ```ruby
-# Pure host with authentication.
 config = {
-  url:      ENV['PURE_URL'],
-  username: ENV['PURE_USERNAME'],
-  password: ENV['PURE_PASSWORD']
-}
-```
-
-```ruby
-# Pure host without authentication.
-config = {
-  url: ENV['PURE_URL']
+  url:      'https://YOUR_HOST/ws/api/59',
+  username: 'YOUR_USERNAME',
+  password: 'YOUR_PASSWORD',
+  api_key:  'YOUR_API_KEY'
 }
 ```
 
 ### Transformation
 
-Create a metadata transformer for a Pure dataset.
+Configure a transformer for a Pure dataset.
 
 ```ruby
 transformer = ResearchMetadata::Transformer::Dataset.new config
@@ -56,7 +47,7 @@ transformer = ResearchMetadata::Transformer::Dataset.new config
 Give it a Pure identifier and a DOI...
 
 ```ruby
-metadata = transformer.transform uuid: 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx',
+metadata = transformer.transform id: 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx',
                                  doi: '10.1234/foo/bar/1'
 ```
 

data/lib/research_metadata/transformer/dataset.rb

@@ -8,9 +8,10 @@ module ResearchMetadata
     class Dataset
 
       # @param config [Hash]
-      # @option config [String] :url The URL of the Pure host.
-      # @option config [String] :username The username of the Pure host account.
-      # @option config [String] :password The password of the Pure host account.
+      # @option config [String] :url URL of the Pure host
+      # @option config [String] :username Username of the Pure host account
+      # @option config [String] :password Password of the Pure host account
+      # @option config [String] :api_key API key of the Pure host account
       def initialize(config)
         @config = config
         @dataset_extractor = Puree::Extractor::Dataset.new config
@@ -19,11 +20,10 @@ module ResearchMetadata
       # Dataset transformation
       #
       # @param id [String]
-      # @param uuid [String]
       # @param doi [String]
       # @return [String, nil]
-      def transform(id: nil, uuid: nil, doi: nil)
-        @dataset = extract uuid: uuid, id: id
+      def transform(id:, doi:)
+        @dataset = @dataset_extractor.find id
         return nil if !@dataset
         person_o = person
         file_o = file
@@ -36,11 +36,11 @@ module ResearchMetadata
             subjects: subjects,
             contributors: person_o['contributor'],
             dates: dates,
-            language: language,
+            # language: language,
             resource_type: resource_type,
             related_identifiers: related_identifiers,
-            sizes: sizes(file_o),
-            formats: formats(file_o),
+            # sizes: sizes(file_o),
+            # formats: formats(file_o),
             rights_list: rights_list(file_o),
             descriptions: description,
             geo_locations: spatial
@@ -50,13 +50,13 @@ module ResearchMetadata
 
       private
 
-      def sizes(files)
-        files.map { |i| "#{i.size} B" }
-      end
+      # def sizes(files)
+      #   files.map { |i| "#{i.size} B" }
+      # end
 
-      def formats(files)
-        files.map { |i| i.mime }
-      end
+      # def formats(files)
+      #   files.map { |i| i.mime }
+      # end
 
       def rights_list(files)
         arr = []
@@ -115,14 +115,6 @@ module ResearchMetadata
         end
       end
 
-      def extract(uuid: nil, id: nil)
-        if !uuid.nil?
-          return @dataset_extractor.find uuid: uuid
-        else
-          return @dataset_extractor.find id: id
-        end
-      end
-
       def file
         @dataset.files
       end
@@ -131,9 +123,9 @@ module ResearchMetadata
         ::Datacite::Mapping::Identifier.new(value: doi)
       end
 
-      def language
-        @dataset.locale
-      end
+      # def language
+      #   @dataset.language
+      # end
 
       def name_identifier_orcid(person)
         name_identifier = nil
@@ -170,7 +162,7 @@ module ResearchMetadata
             if human
               if individual.uuid
                 person_extractor = Puree::Extractor::Person.new @config
-                person = person_extractor.find uuid: individual.uuid
+                person = person_extractor.find individual.uuid
                 if person
                   identifier = name_identifier_orcid person
                   human.identifier = identifier if identifier
@@ -195,21 +187,21 @@ module ResearchMetadata
       end
 
       def publisher
-        @dataset.publisher
+        @dataset.publisher.name
       end
 
       def related_identifiers
-        publications = @dataset.publications
+        research_outputs = @dataset.research_outputs
         data = []
-        publications.each do |i|
+        research_outputs.each do |i|
           # Skip as the relationship cannot currently be determined
           next if i.type === 'Dataset'
 
-          publication_extractor = Puree::Extractor::Publication.new @config
-          pub = publication_extractor.find uuid: i.uuid
+          research_output_extractor = Puree::Extractor::ResearchOutput.new @config
+          research_output = research_output_extractor.find i.uuid
 
           # Restrict to those with a DOI
-          doi = pub.doi if pub.methods.include? :doi
+          doi = research_output.doi if research_output.methods.include? :doi
 
           if doi
             doi_part_to_remove = 'http://dx.doi.org/'

data/lib/research_metadata/transformer/{publication.rb → research_output.rb}

@@ -2,30 +2,31 @@ module ResearchMetadata
 
   module Transformer
 
-    # Extracts publication metadata from the Pure Research Information System
+    # Extracts research output metadata from the Pure Research Information System
     # and converts it into the DataCite format.
     # For text-based resources.
+    # @note Do not use directly.
     #
-    class Publication
+    class ResearchOutput
 
       # @param config [Hash]
-      # @option config [String] :url The URL of the Pure host.
-      # @option config [String] :username The username of the Pure host account.
-      # @option config [String] :password The password of the Pure host account.
+      # @option config [String] :url URL of the Pure host
+      # @option config [String] :username Username of the Pure host account
+      # @option config [String] :password Password of the Pure host account
+      # @option config [String] :api_key API key of the Pure host account
       def initialize(config)
         @config = config
-        @publication_extractor = Puree::Extractor::Publication.new config
+        @research_output_extractor = Puree::Extractor::ResearchOutput.new config
       end
 
-      # Publication transformation
+      # Research output transformation
       #
       # @param id [String]
-      # @param uuid [String]
       # @param doi [String]
       # @return [String, nil]
-      def transform(id: nil, uuid: nil, doi: nil)
-        @publication = extract uuid: uuid, id: id
-        return nil if !@publication
+      def transform(id:, doi:)
+        @research_output = @research_output_extractor.find id
+        return nil if !@research_output
         person_o = person
         file_o = file
         resource = ::Datacite::Mapping::Resource.new(
@@ -75,7 +76,7 @@ module ResearchMetadata
       end
 
       def description
-        desc = @publication.description
+        desc = @research_output.description
         if desc
           d = ::Datacite::Mapping::Description.new value: desc,
                                              type: ::Datacite::Mapping::DescriptionType::ABSTRACT
@@ -85,16 +86,8 @@ module ResearchMetadata
         end
       end
 
-      def extract(uuid: nil, id: nil)
-        if !uuid.nil?
-          return @publication_extractor.find uuid: uuid
-        else
-          return @publication_extractor.find id: id
-        end
-      end
-
       def file
-        @publication.files
+        @research_output.files
       end
 
       def identifier(doi)
@@ -102,7 +95,7 @@ module ResearchMetadata
       end
 
       def language
-        @publication.locale
+        @research_output.language
       end
 
       def name_identifier_orcid(person)
@@ -120,9 +113,9 @@ module ResearchMetadata
         o['creator'] = []
         o['contributor'] = []
         all_persons = []
-        all_persons << @publication.persons_internal
-        all_persons << @publication.persons_external
-        all_persons << @publication.persons_other
+        all_persons << @research_output.persons_internal
+        all_persons << @research_output.persons_external
+        all_persons << @research_output.persons_other
         all_persons.each do |person_type|
           person_type.each do |individual|
             pure_role =individual.role.gsub(/\s+/, '')
@@ -142,7 +135,7 @@ module ResearchMetadata
             if human
               if individual.uuid
                 person_extractor = Puree::Extractor::Person.new @config
-                person = person_extractor.find uuid: individual.uuid
+                person = person_extractor.find individual.uuid
                 if person
                   identifier = name_identifier_orcid person
                   human.identifier = identifier if identifier
@@ -163,7 +156,7 @@ module ResearchMetadata
       end
 
       def publication_year
-        @publication.statuses.each do |i|
+        @research_output.publication_statuses.each do |i|
           if i.stage === 'Published' || i.stage === 'Unpublished'
             return i.date.year
           end
@@ -171,10 +164,6 @@ module ResearchMetadata
         nil
       end
 
-      def publisher
-        @publication.publisher || 'Publisher unspecified'
-      end
-
       def resource_type
         ::Datacite::Mapping::ResourceType.new(
             resource_type_general: ::Datacite::Mapping::ResourceTypeGeneral::TEXT,
@@ -183,18 +172,23 @@ module ResearchMetadata
       end
 
       def subjects
-        @publication.keywords.map { |i| ::Datacite::Mapping::Subject.new value: i }
+        @research_output.keywords.map { |i| ::Datacite::Mapping::Subject.new value: i }
       end
 
       def titles
         arr = []
-        title = ::Datacite::Mapping::Title.new value: @publication.title
+        title = ::Datacite::Mapping::Title.new value: @research_output.title
         arr << title
-        subtitle = @publication.subtitle
+        subtitle = @research_output.subtitle
         if subtitle
           arr << ::Datacite::Mapping::Title.new(value: subtitle,
                                                 type: ::Datacite::Mapping::TitleType::SUBTITLE)
         end
+        translated_title = @research_output.translated_title
+        if translated_title
+          arr << ::Datacite::Mapping::Title.new(value: translated_title,
+                                                type: ::Datacite::Mapping::TitleType::TRANSLATED_TITLE)
+        end
         arr
       end
 

data/lib/research_metadata/transformer/thesis.rb

@@ -2,25 +2,26 @@ module ResearchMetadata
 
   module Transformer
 
-    # Extracts publication metadata from the Pure Research Information System
+    # Extracts research output metadata from the Pure Research Information System
     # and converts it into the DataCite format. Usage is for theses
     # (doctoral and master's).
     #
-    class Thesis < ResearchMetadata::Transformer::Publication
+    class Thesis < ResearchMetadata::Transformer::ResearchOutput
 
       # @param config [Hash]
-      # @option config [String] :url The URL of the Pure host.
-      # @option config [String] :username The username of the Pure host account.
-      # @option config [String] :password The password of the Pure host account.
+      # @option config [String] :url URL of the Pure host
+      # @option config [String] :username Username of the Pure host account
+      # @option config [String] :password Password of the Pure host account
+      # @option config [String] :api_key API key of the Pure host account
       def initialize(config)
         @config = config
-        @publication_extractor = Puree::Extractor::Thesis.new config
+        @research_output_extractor = Puree::Extractor::Thesis.new config
       end
 
       private
 
       def pages
-        count = @publication.pages
+        count = @research_output.pages
         if count && count > 0
           return "#{count} pages"
         else
@@ -28,6 +29,10 @@ module ResearchMetadata
         end
       end
 
+      def publisher
+        @research_output.publisher.name || 'Publisher unspecified'
+      end
+
       def sizes(files)
         arr = files.map { |i| "#{i.size} B" }
         arr << pages if pages

data/lib/research_metadata/version.rb

@@ -1,5 +1,5 @@
 module ResearchMetadata
   # Semantic version number
   #
-  VERSION = "1.1.2"
+  VERSION = "2.0.0"
 end

data/lib/research_metadata.rb

@@ -1,7 +1,7 @@
 require 'puree'
 require 'datacite/mapping'
 require 'research_metadata/transformer/dataset'
-require 'research_metadata/transformer/publication'
+require 'research_metadata/transformer/research_output'
 require 'research_metadata/transformer/thesis'
 require 'research_metadata/version'
 

data/research_metadata.gemspec

@@ -4,23 +4,23 @@ $LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
 require 'research_metadata/version'
 
 Gem::Specification.new do |spec|
-  spec.name          = "research_metadata"
+  spec.name          = 'research_metadata'
   spec.version       = ResearchMetadata::VERSION
-  spec.authors       = ["Adrian Albin-Clark"]
-  spec.email         = ["a.albin-clark@lancaster.ac.uk"]
+  spec.authors       = 'Adrian Albin-Clark'
+  spec.email         = 'a.albin-clark@lancaster.ac.uk'
   spec.summary       = %q{Metadata extraction from the Pure Research Information System and transformation of the metadata into the DataCite format.}
-  spec.homepage      = "https://github.com/lulibrary/research_metadata"
-  spec.license       = "MIT"
+  spec.homepage      = 'https://github.com/lulibrary/research_metadata'
+  spec.license       = 'MIT'
 
   spec.files         = `git ls-files -z`.split("\x0")
   spec.executables   = spec.files.grep(%r{^bin/}) { |f| File.basename(f) }
   spec.test_files    = spec.files.grep(%r{^(test|spec|features)/})
-  spec.require_paths = ["lib"]
+  spec.require_paths = ['lib']
 
   spec.required_ruby_version = '~> 2.1'
 
-  spec.add_runtime_dependency "puree", "~> 1.2.0"
-  spec.add_runtime_dependency "datacite-mapping", "~> 0.2.5"
+  spec.add_runtime_dependency 'puree', '~> 2.0'
+  spec.add_runtime_dependency 'datacite-mapping', '~> 0.2.5'
 
-  spec.add_development_dependency "rspec"
+  spec.add_development_dependency 'minitest-reporters', '~> 1.1'
 end

data/test/test_helper.rb

@@ -0,0 +1,17 @@
+require 'minitest/autorun'
+require 'minitest/reporters'
+Minitest::Reporters.use!
+
+lib = File.expand_path('../../lib', __FILE__)
+$LOAD_PATH.unshift(lib)
+
+require 'research_metadata'
+
+def config
+  {
+      url:      ENV['PURE_URL_TEST_59'],
+      username: ENV['PURE_USERNAME'],
+      password: ENV['PURE_PASSWORD'],
+      api_key:  ENV['PURE_API_KEY']
+  }
+end

data/test/transformer/dataset_test.rb

@@ -0,0 +1,25 @@
+require 'test_helper'
+
+class TestTransformerDataset < Minitest::Test
+
+  def transformer
+    ResearchMetadata::Transformer::Dataset.new config
+  end
+
+  def test_transformer
+     assert_instance_of ResearchMetadata::Transformer::Dataset, transformer
+  end
+
+  def test_transform
+    # The 2014 Ebola virus disease outbreak in West Africa
+    id = 'b050f4b5-e272-4914-8cac-3bdc1e673c58'
+
+    metadata = transformer.transform id: id,
+                                     doi: '10.1234/foo/bar/1'
+
+    # puts metadata
+
+    assert_equal true, metadata.downcase.start_with?('<resource')
+  end
+
+end

data/test/transformer/thesis_test.rb

@@ -0,0 +1,37 @@
+require 'test_helper'
+
+class TestTransformerThesis < Minitest::Test
+
+  def transformer
+    ResearchMetadata::Transformer::Thesis.new config
+  end
+
+  def test_transformer
+    assert_instance_of ResearchMetadata::Transformer::Thesis, transformer
+  end
+
+  def test_transform_1
+    # Multimodalita e 'city branding'
+    id = '376173c0-fd7a-4d63-93d3-3f2e58e8dc01'
+
+    metadata = transformer.transform id: id,
+                                     doi: '10.1234/foo/bar/1'
+
+    # puts metadata
+
+    assert_equal true, metadata.downcase.start_with?('<resource')
+  end
+
+  def test_transform_2
+    # Nanoscale imaging and characterisation of Amyloid-β
+    id ='9d3ad4d1-3d46-4551-9139-f783fd4e5123'
+
+    metadata = transformer.transform id: id,
+                                     doi: '10.1234/foo/bar/1'
+
+    # puts metadata
+
+    assert_equal true, metadata.downcase.start_with?('<resource')
+  end
+
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: research_metadata
 version: !ruby/object:Gem::Version
-  version: 1.1.2
+  version: 2.0.0
 platform: ruby
 authors:
 - Adrian Albin-Clark
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-07-12 00:00:00.000000000 Z
+date: 2017-12-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: puree
@@ -16,14 +16,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.2.0
+        version: '2.0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.2.0
+        version: '2.0'
 - !ruby/object:Gem::Dependency
   name: datacite-mapping
   requirement: !ruby/object:Gem::Requirement
@@ -39,22 +39,21 @@ dependencies:
       - !ruby/object:Gem::Version
         version: 0.2.5
 - !ruby/object:Gem::Dependency
-  name: rspec
+  name: minitest-reporters
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '1.1'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '1.1'
 description: 
-email:
-- a.albin-clark@lancaster.ac.uk
+email: a.albin-clark@lancaster.ac.uk
 executables: []
 extensions: []
 extra_rdoc_files: []
@@ -67,13 +66,14 @@ files:
 - Rakefile
 - lib/research_metadata.rb
 - lib/research_metadata/transformer/dataset.rb
-- lib/research_metadata/transformer/publication.rb
+- lib/research_metadata/transformer/research_output.rb
 - lib/research_metadata/transformer/thesis.rb
 - lib/research_metadata/transformer/transformer.rb
 - lib/research_metadata/version.rb
 - research_metadata.gemspec
-- spec/spec_helper.rb
-- spec/transformer/dataset_spec.rb
+- test/test_helper.rb
+- test/transformer/dataset_test.rb
+- test/transformer/thesis_test.rb
 homepage: https://github.com/lulibrary/research_metadata
 licenses:
 - MIT
@@ -100,5 +100,6 @@ specification_version: 4
 summary: Metadata extraction from the Pure Research Information System and transformation
   of the metadata into the DataCite format.
 test_files:
-- spec/spec_helper.rb
-- spec/transformer/dataset_spec.rb
+- test/test_helper.rb
+- test/transformer/dataset_test.rb
+- test/transformer/thesis_test.rb

data/spec/spec_helper.rb

@@ -1,4 +0,0 @@
-require 'puree'
-require 'datacite/mapping'
-require 'research_metadata/transformer/dataset'
-require 'research_metadata/version'

data/spec/transformer/dataset_spec.rb

@@ -1,37 +0,0 @@
-require 'spec_helper'
-
-describe 'Dataset' do
-
-  def setup
-    @config = {
-        url:      ENV['PURE_URL'],
-        username: ENV['PURE_USERNAME'],
-        password: ENV['PURE_PASSWORD']
-    }
-    @t = ResearchMetadata::Transformer::Dataset.new @config
-  end
-
-  it '#new' do
-    setup
-    t = ResearchMetadata::Transformer::Dataset.new @config
-    expect(t).to be_a ResearchMetadata::Transformer::Dataset
-  end
-
-  describe 'data transformation' do
-    before(:all) do
-      setup
-    end
-
-    it '#transform with random UUID' do
-      c = Puree::Extractor::Collection.new resource: :dataset,
-                                           config: @config
-      res = c.random_resource
-      metadata = @t.transform uuid: res.uuid,
-                              doi:  '10.1234/foo/bar/1'
-      is_xml = metadata.downcase.start_with?('<resource')
-      expect(is_xml).to match(true)
-    end
-
-  end
-
-end