puree 1.0.0 → 1.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 0fcfce1e04c1a07280ef3792f67c7bb4b7fea7e9
-  data.tar.gz: ed700b052232a17ccb11b3337eca880d71e00541
+  metadata.gz: 51982259947535b8a9b88916eca1d629004b1d78
+  data.tar.gz: 0b4469e3b0f0881d45ebbdc7a90eb69f7b4e29cb
 SHA512:
-  metadata.gz: 1087979e95515396a9eceb0a7db3e32dd87130d4505fc95626043f48e657767b762f8861787ce652597098019db2b20a6b365831705a8f633dac89eb27e187cd
-  data.tar.gz: 1fe7778084eecac9c214a354a999825b089e5b4c417ceb3b03cc8db8dc39e7a57d8b1fd1debe87af9dd4170f6d56a948e97b5fbccca93290fc878618cf030f89
+  metadata.gz: e1ab192064572d2935d3546577c4bdbe4dbbd624f9056f173da1d9d91f910beb493aa0dd17a13a1162003f618ea215f213643e6a3fc98509acdd89e8fa21b416
+  data.tar.gz: 3e94e0b3ae0f86fa0f91911e142d435d428c088735108d19aa5ca49c182d18f16602d83441d374158a9bbc8a4635d60c250bff43d4fc711634d4bd551e2ed0c7

data/lib/puree/extractor/publication.rb

@@ -21,11 +21,13 @@ module Puree
         @model.doi = @extractor.doi
         @model.event = @extractor.event
         @model.files = @extractor.files
+        @model.keywords = @extractor.keywords
         @model.organisations = @extractor.organisations
         @model.pages = @extractor.pages
         @model.persons_internal = @extractor.persons_internal
         @model.persons_external = @extractor.persons_external
         @model.persons_other = @extractor.persons_other
+        @model.publisher = @extractor.publisher
         @model.statuses = @extractor.statuses
         @model.subtitle = @extractor.subtitle
         @model.title = @extractor.title

data/lib/puree/model/publication.rb

@@ -20,6 +20,9 @@ module Puree
       # @return [Array<Puree::Model::File>]
       attr_accessor :files
 
+      # @return [Array<String>]
+      attr_accessor :keywords
+
       # @return [Array<Puree::Model::OrganisationHeader>]
       attr_accessor :organisations
 
@@ -35,6 +38,9 @@ module Puree
       # @return [Array<Puree::Model::EndeavourPerson>]
       attr_accessor :persons_other
 
+      # @return [String, nil]
+      attr_accessor :publisher
+
       # @return [Array<Puree::Model::PublicationStatus>]
       attr_accessor :statuses
 

data/lib/puree/version.rb

@@ -1,5 +1,5 @@
 module Puree
   # Semantic version number
   #
-  VERSION = "1.0.0"
+  VERSION = "1.1.0"
 end

data/lib/puree/xml_extractor/publication.rb

@@ -40,19 +40,33 @@ module Puree
 
       # @return [Array<Puree::Model::File>]
       def files
-        xpath_result = xpath_query '/electronicVersionAssociations/electronicVersionFileAssociations/electronicVersionFileAssociation/file'
+        xpath_result = xpath_query '/electronicVersionAssociations/electronicVersionFileAssociations/electronicVersionFileAssociation'
         docs = []
         xpath_result.each do |d|
           model = Puree::Model::File.new
-          model.name = d.xpath('fileName').text.strip
-          model.mime = d.xpath('mimeType').text.strip
-          model.size = d.xpath('size').text.strip.to_i
-          model.url = d.xpath('url').text.strip
+          model.name = d.xpath('file/fileName').text.strip
+          model.mime = d.xpath('file/mimeType').text.strip
+          model.size = d.xpath('file/size').text.strip.to_i
+          model.url = d.xpath('file/url').text.strip
+          document_license = d.xpath('licenseType')
+          if !document_license.empty?
+            license = Puree::Model::CopyrightLicense.new
+            license.name = document_license.xpath('term/localizedString').text.strip
+            license.url = document_license.xpath('description/localizedString').text.strip
+            model.license = license if license.data?
+          end
           docs << model
         end
         docs.uniq { |d| d.url }
       end
 
+      # @return [Array<String>]
+      def keywords
+        xpath_result =  xpath_query '/keywordGroups/keywordGroup/keyword/userDefinedKeyword/freeKeyword'
+        data_arr = xpath_result.map { |i| i.text.strip }
+        data_arr.uniq
+      end
+
       # @return [Array<Puree::Model::OrganisationHeader>]
       def organisations
         xpath_result = xpath_query '/organisations/association/organisation'
@@ -61,7 +75,8 @@ module Puree
 
       # @return [Fixnum, nil]
       def pages
-        xpath_query_for_single_value('/numberOfPages').to_i
+        xpath_result = xpath_query_for_single_value('/numberOfPages')
+        xpath_result ? xpath_result.to_i : nil
       end
 
       # @return [Array<Puree::Model::EndeavourPerson>]
@@ -79,6 +94,11 @@ module Puree
         persons 'other'
       end
 
+      # @return [String, nil]
+      def publisher
+        xpath_query_for_single_value '/associatedPublisher/publisher/name'
+      end
+
       # @return [Array<Puree::Model::PublicationStatus>]
       def statuses
         xpath_result = xpath_query '/publicationStatuses/publicationStatus'

data/puree.gemspec

@@ -9,8 +9,6 @@ Gem::Specification.new do |spec|
   spec.authors       = 'Adrian Albin-Clark'
   spec.email         = 'a.albin-clark@lancaster.ac.uk'
   spec.summary       = %q{Metadata extraction from the Pure Research Information System.}
-  spec.description   = %q{Fetches metadata from the Pure Research Information System and
-                          extracts it into Ruby data models.}
   spec.homepage      = 'https://github.com/lulibrary/puree'
   spec.license       = 'MIT'
   spec.files         = `git ls-files -z`.split("\x0")

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: puree
 version: !ruby/object:Gem::Version
-  version: 1.0.0
+  version: 1.1.0
 platform: ruby
 authors:
 - Adrian Albin-Clark
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-03-15 00:00:00.000000000 Z
+date: 2017-03-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: http
@@ -52,9 +52,7 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
-description: |-
-  Fetches metadata from the Pure Research Information System and
-                            extracts it into Ruby data models.
+description: 
 email: a.albin-clark@lancaster.ac.uk
 executables: []
 extensions: []