darlingtonia 3.1.1 → 3.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: f65916a0536747279530beff22b393ca88d31d0ce16338799f3bc45631e864c2
-  data.tar.gz: be92962bfbd065747aea00b26d77fb79c8ba46655c367ddc74bcb5190e8bdb13
+  metadata.gz: 63a317635545c8808e56b95fcad54851481952e89268fcf3230eba94a885c203
+  data.tar.gz: 5e76830148cf211c69fab7e7f6b455a9a06a912e73bd023dab6e1369964c8c25
 SHA512:
-  metadata.gz: 2536967a3f2f3040638d12f08df770395ae70f4bc133fd65b247a14f824e5f1f4f06c88c2d5136a0de1a6cb8d4bb8b1712fc8a0142df025cec99565c366d1f01
-  data.tar.gz: fff32d8eae679a040dfc35ab234a829118291b6960f2e6ce48d8e92ab748372b58a44978f26d74e28e28838ffa67bb79e2ca90512ad9b9b461519a8d19e43bd4
+  metadata.gz: 3e86824312574df7b9630c1fd0f8d42416fe917309cea097912497f443d1e824b998dae3c97af18f76dc53a38e6af3f5ca54053bb3bf2423bc7ab9fd7b19245b
+  data.tar.gz: cbe2deaca3be8d5da290f611c4d22a7d8abf91f2764cb40361da36551b8e0629563f50a30abae2723d2a91ac1c688f06ec24b6b7400f59fb0d61dabdb17b3451

data/CHANGELOG.md

@@ -1,3 +1,17 @@
+3.2.0 - Tue Mar 5, 2019
+
+* Allow the importer to receive log settings from elsewhere, so it can log
+success messages with all the other logs.
+
+* When a record is imported a second time (i.e., when there is already a
+record matching the deduplication field in the repository), only update
+its metadata and collection membership. Use a stripped down actor stack
+that only performs these actions.
+
+* Log a special message if we attempt an empty import.
+
+* Fix logging error where id number wasn't printing to logs on updates.
+
 3.1.0 - Tue Feb 26, 2019
 
 New Feature: `HyraxRecordImporter` now accepts a `deduplication_field` in the

data/lib/darlingtonia.rb

@@ -69,4 +69,5 @@ module Darlingtonia
 
   require 'darlingtonia/parser'
   require 'darlingtonia/parsers/csv_parser'
+  require 'darlingtonia/metadata_only_stack'
 end

data/lib/darlingtonia/hyrax_record_importer.rb

@@ -180,8 +180,18 @@ module Darlingtonia
 
     private
 
+      # Build a pared down actor stack that will not re-attach files,
+      # or set workflow, or do anything except update metadata.
+      # TODO: We should be able to set an environment variable that would allow updates to go through the regular
+      # actor stack instead of the stripped down one, in case we want to re-import files.
+      def metadata_only_middleware
+        terminator = Hyrax::Actors::Terminator.new
+        Darlingtonia::MetadataOnlyStack.build_stack.build(terminator)
+      end
+
       # Update an existing object using the Hyrax actor stack
       # We assume the object was created as expected if the actor stack returns true.
+      # Note that for now the update stack will only update metadata and update collection membership, it will not re-import files.
       def update_for(existing_record:, update_record:)
         info_stream << "event: record_update_started, batch_id: #{batch_id}, collection_id: #{collection_id}, #{deduplication_field}: #{update_record.respond_to?(deduplication_field) ? update_record.send(deduplication_field) : update_record}"
         additional_attrs = {
@@ -192,13 +202,16 @@ module Darlingtonia
         # Ensure nothing is passed in the files field,
         # since this is reserved for Hyrax and is where uploaded_files will be attached
         attrs.delete(:files)
+
+        # We aren't using the attach remote files actor, so make sure any remote files are removed from the params before we try to save the object.
+        attrs.delete(:remote_files)
+
         based_near = attrs.delete(:based_near)
         attrs = attrs.merge(based_near_attributes: based_near_attributes(based_near)) unless based_near.nil? || based_near.empty?
-        actor_env = Hyrax::Actors::Environment.new(existing_record,
-                                                   ::Ability.new(@depositor),
-                                                   attrs)
-        if Hyrax::CurationConcern.actor.update(actor_env)
-          info_stream << "event: record_updated, batch_id: #{batch_id}, record_id: #{existing_record.id}, collection_id: #{collection_id}, #{deduplication_field}: #{existing_record.respond_to?(deduplication_field) ? existing_record.send(deduplication_field) : existing_record}"
+
+        actor_env = Hyrax::Actors::Environment.new(existing_record, ::Ability.new(@depositor), attrs)
+        if metadata_only_middleware.update(actor_env)
+          info_stream << "event: record_updated, batch_id: #{batch_id}, record_id: #{existing_record.id}, collection_id: #{collection_id}, #{deduplication_field}: #{existing_record.respond_to?(deduplication_field) ? existing_record.send(deduplication_field)&.first : existing_record}"
           @success_count += 1
         else
           existing_record.errors.each do |attr, msg|

data/lib/darlingtonia/importer.rb

@@ -31,10 +31,17 @@ module Darlingtonia
     #   records.
     # @param record_importer [RecordImporter] An object to handle import of
     #   each record
-    def initialize(parser:, record_importer: RecordImporter.new)
+    def initialize(parser:, record_importer: RecordImporter.new, info_stream: Darlingtonia.config.default_info_stream, error_stream: Darlingtonia.config.default_error_stream)
       self.parser          = parser
       self.record_importer = record_importer
-      @info_stream = Darlingtonia.config.default_info_stream
+      @info_stream = info_stream
+      @error_stream = error_stream
+    end
+
+    # Do not attempt to run an import if there are no records. Instead, just write to the log.
+    def no_records_message
+      @info_stream << "event: empty_import, batch_id: #{record_importer.batch_id}"
+      @error_stream << "event: empty_import, batch_id: #{record_importer.batch_id}"
     end
 
     ##
@@ -42,8 +49,13 @@ module Darlingtonia
     #
     # @return [void]
     def import
+      no_records_message && return unless records.count.positive?
+      start_time = Process.clock_gettime(Process::CLOCK_MONOTONIC)
+      @info_stream << "event: start_import, batch_id: #{record_importer.batch_id}, expecting to import #{records.count} records."
       records.each { |record| record_importer.import(record: record) }
-      @info_stream << "event: finish_import, batch_id: #{record_importer.batch_id}, successful_record_count: #{record_importer.success_count}, failed_record_count: #{record_importer.failure_count}"
+      end_time = Process.clock_gettime(Process::CLOCK_MONOTONIC)
+      elapsed_time = end_time - start_time
+      @info_stream << "event: finish_import, batch_id: #{record_importer.batch_id}, successful_record_count: #{record_importer.success_count}, failed_record_count: #{record_importer.failure_count}, elapsed_time: #{elapsed_time}, elapsed_time_per_record: #{elapsed_time / records.count}"
     end
   end
 end

data/lib/darlingtonia/metadata_only_stack.rb

@@ -0,0 +1,70 @@
+# frozen_string_literal: true
+
+module Darlingtonia
+  class MetadataOnlyStack
+    def self.build_stack
+      ActionDispatch::MiddlewareStack.new.tap do |middleware|
+        # Wrap everything in a database transaction, if the save of the resource
+        # fails then roll back any database AdminSetChangeSet
+        middleware.use Hyrax::Actors::TransactionalRequest
+
+        # Ensure you are mutating the most recent version
+        # middleware.use Hyrax::Actors::OptimisticLockValidator
+
+        # Attach files from a URI (for BrowseEverything)
+        # middleware.use Hyrax::Actors::CreateWithRemoteFilesActor
+
+        # Attach files uploaded in the form to the UploadsController
+        # In Californica, for command line import,
+        # we are using the CreateWithFilesActor to attach
+        # local files with a file:// url, not via the UploadsController,
+        # so we never use the CreateWithFilesActor
+        # middleware.use Hyrax::Actors::CreateWithFilesActor
+
+        # Add/remove the resource to/from a collection
+        middleware.use Hyrax::Actors::CollectionsMembershipActor
+
+        # Add/remove to parent work
+        # middleware.use Hyrax::Actors::AddToWorkActor
+
+        # Add/remove children (works or file_sets)
+        # middleware.use Hyrax::Actors::AttachMembersActor
+
+        # Set the order of the children (works or file_sets)
+        # middleware.use Hyrax::Actors::ApplyOrderActor
+
+        # Sets the default admin set if they didn't supply one
+        # middleware.use Hyrax::Actors::DefaultAdminSetActor
+
+        # Decode the private/public/institution on the form into permisisons on
+        # the model
+        # We aren't using this in Californica, we're just setting the visibility
+        # at import time
+        # middleware.use Hyrax::Actors::InterpretVisibilityActor
+
+        # Handles transfering ownership of works from one user to another
+        # We aren't using this in Californica
+        # middleware.use Hyrax::Actors::TransferRequestActor
+
+        # Copies default permissions from the PermissionTemplate to the work
+        # middleware.use Hyrax::Actors::ApplyPermissionTemplateActor
+
+        # Remove attached FileSets when destroying a work
+        # middleware.use Hyrax::Actors::CleanupFileSetsActor
+
+        # Destroys the trophies in the database when the work is destroyed
+        # middleware.use Hyrax::Actors::CleanupTrophiesActor
+
+        # Destroys the feature tag in the database when the work is destroyed
+        # middleware.use Hyrax::Actors::FeaturedWorkActor
+
+        # Persist the metadata changes on the resource
+        middleware.use Hyrax::Actors::ModelActor
+
+        # Start the workflow for this work
+        # middleware.use Hyrax::Actors::InitializeWorkflowActor
+      end
+    end
+    # rubocop:enable Metrics/MethodLength
+  end
+end

data/lib/darlingtonia/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Darlingtonia
-  VERSION = '3.1.1'
+  VERSION = '3.2.0'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: darlingtonia
 version: !ruby/object:Gem::Version
-  version: 3.1.1
+  version: 3.2.0
 platform: ruby
 authors:
 - Data Curation Experts
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-02-26 00:00:00.000000000 Z
+date: 2019-03-05 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: active-fedora
@@ -177,6 +177,7 @@ files:
 - lib/darlingtonia/input_record.rb
 - lib/darlingtonia/log_stream.rb
 - lib/darlingtonia/metadata_mapper.rb
+- lib/darlingtonia/metadata_only_stack.rb
 - lib/darlingtonia/parser.rb
 - lib/darlingtonia/parsers/csv_parser.rb
 - lib/darlingtonia/record_importer.rb