dor-services 7.0.2 → 7.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 05abb7b8c6054c86ed4edeeef3baf7d768a7e0c97c3e6bfdedf25d708ba2291a
-  data.tar.gz: 5955774b7898520df8baf33298697cadf2c2d8b8fa9e41903a1bf9aca2c2cda0
+  metadata.gz: 1cf9f20e7028e7252d30de465bbbee5afc837f32793be1612591443dab72170d
+  data.tar.gz: 0436dcc7ecf8964b20193603f4219ebfc46a0bedf42566ddcf4a79bd831ed8ff
 SHA512:
-  metadata.gz: 99a5e9dffd5f436ae3c2a0059903431b1d8ba59af633d58f3ef6894daa55f1aa32d6b37a7828c2610bc76c8c144806df95e831352d1718fd57fcff1e6008b53e
-  data.tar.gz: f9cea923019b4ceca95b5516a487683a98dcaeb702956cc082828c6eedb96a4541c478fe08cbc72017344760a5acf212593105069c262e136f6e5ead88ef70a5
+  metadata.gz: e48a8d2b2f1f3ccfff30db3def931c14862eb9e108865db26b4d830c170e990db20ca453b5b66b8fdf61fa8b9c5a677f522c124bd294918e682e0ea88693da4e
+  data.tar.gz: 1cb0bfa19232319da9d3e47efd241974a91f0e94f21cdc46195804270d02a29be8e2979cbba15c01b5e0c35b53a643f627f5153fb966369b7f74b0100e0132e6

data/lib/dor-services.rb

@@ -5,6 +5,7 @@ require 'active_fedora/version'
 require 'active_support/core_ext/module/attribute_accessors'
 require 'active_support/core_ext/object/blank'
 require 'deprecation'
+require 'retries' # Used by Dor::ReleaseTagService
 
 module Dor
   extend ActiveSupport::Autoload
@@ -42,6 +43,7 @@ module Dor
   require 'dor/utils/ng_tidy'
   require 'dor/utils/solr_doc_helper'
   require 'dor/utils/predicate_patch'
+  require 'dor/utils/pid_utils'
 
   require 'dor/datastreams/datastream_spec_solrizer'
 
@@ -113,6 +115,7 @@ module Dor
     autoload :OpenDataLicenseService
     autoload :PublicDescMetadataService
     autoload :PublishedRelationshipsFilter
+    autoload :PurlClient
     autoload :ReleaseTagService
     autoload :SearchService
     autoload :StatusService

data/lib/dor/models/abstract.rb

@@ -87,19 +87,21 @@ module Dor
     def pid_regex
       /[a-zA-Z]{2}[0-9]{3}[a-zA-Z]{2}[0-9]{4}/
     end
+    deprecation_deprecate pid_regex: 'use PidUtils::PID_REGEX instead'
 
     # a regex that can be used to identify a full druid with prefix (e.g. druid:oo000oo0001)
     # @return [Regex] a regular expression to identify a full druid
     def druid_regex
       /druid:#{pid_regex}/
     end
+    deprecation_deprecate druid_regex: 'will be removed without replacement'
 
     # Since purl does not use the druid: prefix but much of dor does, use this function to strip the druid: if needed
     # @return [String] the druid sans the druid: or if there was no druid: prefix, the entire string you passed
     def remove_druid_prefix(druid = id)
-      result = druid.match(/#{pid_regex}/)
-      result.nil? ? druid : result[0] # if no matches, return the string passed in, otherwise return the match
+      PidUtils.remove_druid_prefix(druid)
     end
+    deprecation_deprecate remove_druid_prefix: 'use PidUtils.remove_druid_prefix instead'
 
     # This is used by Argo and the MergeService
     # @return [Boolean] true if the object is in a state that allows it to be modified.

data/lib/dor/services/purl_client.rb

@@ -0,0 +1,42 @@
+# frozen_string_literal: true
+
+module Dor
+  # Calls the purl service and returns the XML document
+  class PurlClient
+    def initialize(host:, pid:)
+      @host = host
+      @pid = pid
+    end
+
+    # Get XML from the purl service
+    # Fetches purl xml for a druid
+    # @raise [OpenURI::HTTPError]
+    # @return [Nokogiri::HTML::Document] parsed XML for the druid or an empty document if no purl is found
+    def fetch
+      handler = proc do |exception, attempt_number, total_delay|
+        # We assume a 404 means the document has never been published before and thus has no purl
+        Dor.logger.warn "[Attempt #{attempt_number}] GET #{url} -- #{exception.class}: #{exception.message}; #{total_delay} seconds elapsed."
+        raise exception unless exception.is_a? OpenURI::HTTPError
+        return Nokogiri::HTML::Document.new if exception.io.status.first == '404' # ["404", "Not Found"] from OpenURI::Meta.status
+      end
+
+      with_retries(max_retries: 3, base_sleep_seconds: 3, max_sleep_seconds: 5, handler: handler) do |attempt|
+        # If you change the method used for opening the webpage, you can change the :rescue param to handle the new method's errors
+        Dor.logger.debug "[Attempt #{attempt}] GET #{url}"
+        return Nokogiri::XML(OpenURI.open_uri(url))
+      end
+    end
+
+    private
+
+    # Take the and create the entire purl url that will usable for the open method in open-uri, returns http
+    # @return [String] the full url
+    def url
+      @url ||= "https://#{@host}/#{druid_without_prefix}.xml"
+    end
+
+    def druid_without_prefix
+      PidUtils.remove_druid_prefix(@pid)
+    end
+  end
+end

data/lib/dor/services/release_tag_service.rb

@@ -11,6 +11,8 @@ module Dor
 
     def initialize(item)
       @item = item
+      @purl_client = PurlClient.new(host: Dor::Config.stacks.document_cache_host,
+                                    pid: item.pid)
     end
 
     # Called in Dor::UpdateMarcRecordService (in dor-services-app too)
@@ -189,37 +191,11 @@ module Dor
       new_tags
     end
 
-    # Get a list of all release nodes found in a purl document
-    # Fetches purl xml for a druid
-    # @raise [OpenURI::HTTPError]
-    # @return [Nokogiri::HTML::Document] parsed XML for the druid or an empty document if no purl is found
-    def xml_from_purl
-      url = form_purl_url
-      handler = proc do |exception, attempt_number, total_delay|
-        # We assume a 404 means the document has never been published before and thus has no purl
-        Dor.logger.warn "[Attempt #{attempt_number}] GET #{url} -- #{exception.class}: #{exception.message}; #{total_delay} seconds elapsed."
-        raise exception unless exception.is_a? OpenURI::HTTPError
-        return Nokogiri::HTML::Document.new if exception.io.status.first == '404' # ["404", "Not Found"] from OpenURI::Meta.status
-      end
-
-      with_retries(max_retries: 3, base_sleep_seconds: 3, max_sleep_seconds: 5, handler: handler) do |attempt|
-        # If you change the method used for opening the webpage, you can change the :rescue param to handle the new method's errors
-        Dor.logger.info "[Attempt #{attempt}] GET #{url}"
-        return Nokogiri::HTML(OpenURI.open_uri(url))
-      end
-    end
-
-    # Take the and create the entire purl url that will usable for the open method in open-uri, returns http
-    # @return [String] the full url
-    def form_purl_url
-      'https://' + Dor::Config.stacks.document_cache_host + "/#{item.remove_druid_prefix}.xml"
-    end
-
     # Pull all release nodes from the public xml obtained via the purl query
     # @param doc [Nokogiri::HTML::Document] The druid of the object you want
     # @return [Array] An array containing all the release tags
     def release_tags_from_purl_xml(doc)
-      nodes = doc.xpath('//html/body/publicobject/releasedata').children
+      nodes = doc.xpath('//publicObject/releaseData').children
       # We only want the nodes with a name that isn't text
       nodes.reject { |n| n.name.nil? || n.name.casecmp('text') == 0 }.map { |n| n.attr('to') }.uniq
     end
@@ -227,7 +203,7 @@ module Dor
     # Pull all release nodes from the public xml obtained via the purl query
     # @return [Array] An array containing all the release tags
     def release_tags_from_purl
-      release_tags_from_purl_xml(xml_from_purl)
+      release_tags_from_purl_xml(@purl_client.fetch)
     end
 
     attr_reader :item

data/lib/dor/services/thumbnail_service.rb

@@ -46,9 +46,9 @@ module Dor
 
         # if we find one, return the filename based on whether it is a local file or external file
         thumb_image = if search_path[:image_type] == 'local'
-                        "#{object.remove_druid_prefix}/#{thumb_files[0]['id']}"
+                        "#{PidUtils.remove_druid_prefix(object.pid)}/#{thumb_files[0]['id']}"
                       else
-                        "#{object.remove_druid_prefix(thumb_files[0]['objectId'])}/#{thumb_files[0]['fileId']}"
+                        "#{PidUtils.remove_druid_prefix(thumb_files[0]['objectId'])}/#{thumb_files[0]['fileId']}"
                       end
         break # break out of the loop so we stop searching
       end

data/lib/dor/utils/pid_utils.rb

@@ -0,0 +1,14 @@
+# frozen_string_literal: true
+
+module Dor
+  # Utilties for manipulating druids
+  class PidUtils
+    PID_REGEX = /[a-zA-Z]{2}[0-9]{3}[a-zA-Z]{2}[0-9]{4}/.freeze
+    # Since purl does not use the druid: prefix but much of dor does, use this function to strip the druid: if needed
+    # @return [String] the druid sans the druid: or if there was no druid: prefix, the entire string you passed
+    def self.remove_druid_prefix(druid)
+      result = druid.match(PID_REGEX)
+      result.nil? ? druid : result[0] # if no matches, return the string passed in, otherwise return the match
+    end
+  end
+end

data/lib/dor/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Dor
-  VERSION = '7.0.2'
+  VERSION = '7.1.0'
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: dor-services
 version: !ruby/object:Gem::Version
-  version: 7.0.2
+  version: 7.1.0
 platform: ruby
 authors:
 - Michael Klein
@@ -14,7 +14,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-04-24 00:00:00.000000000 Z
+date: 2019-04-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: active-fedora
@@ -92,26 +92,6 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '1.5'
-- !ruby/object:Gem::Dependency
-  name: equivalent-xml
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '0.5'
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 0.5.1
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '0.5'
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 0.5.1
 - !ruby/object:Gem::Dependency
   name: json
   requirement: !ruby/object:Gem::Requirement
@@ -194,6 +174,20 @@ dependencies:
     - - "<"
       - !ruby/object:Gem::Version
         version: '3'
+- !ruby/object:Gem::Dependency
+  name: retries
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.0.5
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.0.5
 - !ruby/object:Gem::Dependency
   name: rsolr
   requirement: !ruby/object:Gem::Requirement
@@ -270,20 +264,6 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '2.6'
-- !ruby/object:Gem::Dependency
-  name: retries
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
 - !ruby/object:Gem::Dependency
   name: dor-rights-auth
   requirement: !ruby/object:Gem::Requirement
@@ -388,6 +368,20 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
+- !ruby/object:Gem::Dependency
+  name: equivalent-xml
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '0.6'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '0.6'
 - !ruby/object:Gem::Dependency
   name: jhove-service
   requirement: !ruby/object:Gem::Requirement
@@ -594,6 +588,7 @@ files:
 - lib/dor/services/ontology.rb
 - lib/dor/services/open_data_license_service.rb
 - lib/dor/services/public_desc_metadata_service.rb
+- lib/dor/services/purl_client.rb
 - lib/dor/services/release_tag_service.rb
 - lib/dor/services/search_service.rb
 - lib/dor/services/status_service.rb
@@ -602,6 +597,7 @@ files:
 - lib/dor/services/thumbnail_service.rb
 - lib/dor/utils/hydrus_shims.rb
 - lib/dor/utils/ng_tidy.rb
+- lib/dor/utils/pid_utils.rb
 - lib/dor/utils/predicate_patch.rb
 - lib/dor/utils/solr_doc_helper.rb
 - lib/dor/version.rb