krikri 0.11.2 → 0.12.0.pre.rc.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: ed0c538e39d8f73fe8b01cc8b71e84f5d4301229
-  data.tar.gz: a1e4947bae2177804b1cda18251cf0cacb8cf700
+  metadata.gz: 7b7f5913ed8e669b3e0451397e3c036dead804cc
+  data.tar.gz: 541d265b195aa3d779a6ddd3e5de4eeae4faffed
 SHA512:
-  metadata.gz: 62aad477bfe83bdf1aad7ef810bc0ac99391874a8491750c769a36a7b92517fe90242e66570dbd8f512e82d232a91901272c473256c8e69e00641c2f43aac8c7
-  data.tar.gz: 06e3c912140e8b42c25391e8e727a2d6650b36c52cb028810d97e521045756e0d6f1f88d66c7eda6f41f3c978e7cb3ec821b9e6ddd6477f61cfda7b9b0e94b69
+  metadata.gz: dd37e66d03acee110f6ec49aa0d5e5ccbcdc97f246785535de6bc68ada64ce984eebfc85e1435fc3e53a75f0525fc0c36a54089e99466b58ff5b96e353ab522d
+  data.tar.gz: 4a4c71f09a89585ec8132378893241852aa72d0c8887ef0d76ba1c07e2d10a87428b1929e606eb6449b5d86ff7179eec644382e5f4d9c5abcf559735421601b1

data/app/models/krikri/activity.rb

@@ -105,10 +105,18 @@ module Krikri
     # Return an Enumerator of URI strings of entities (e.g. aggregations or
     # original records) that pertain to this activity
     #
+    # @param  include_invalidated [Boolean] Whether to include entities that
+    #   have been invalidated with prov:invalidatedAtTime. Default: false
+    #
     # @return [Enumerator] URI strings
-    def entity_uris
+    #
+    # @see Krikri::ProvenanceQueryClient#find_by_activity regarding
+    #   invalidation.
+    #
+    def entity_uris(include_invalidated = false)
       activity_uri = RDF::URI(rdf_subject)  # This activity's LDP URI
-      query = Krikri::ProvenanceQueryClient.find_by_activity(activity_uri)
+      query = Krikri::ProvenanceQueryClient
+        .find_by_activity(activity_uri, include_invalidated)
       query.each_solution.lazy.map do |s|
         s.record.to_s
       end
@@ -122,9 +130,11 @@ module Krikri
     # that is associated with the SoftwareAgent that is represented by the
     # Activity's `agent' field.
     #
+    # @param [Array<Object>] *args Arguments to pass along to
+    #                              EntityBehavior#entities
     # @return [Enumerator] Objects
-    def entities
-      agent_instance.entity_behavior.entities(self)
+    def entities(*args)
+      agent_instance.entity_behavior.entities(self, *args)
     end
   end
 end

data/app/models/krikri/original_record.rb

@@ -30,8 +30,9 @@ module Krikri
       # @param identifier [#to_s] a string representing the #local_name or
       #   fully qualified URI for the resource.  Identifier may have a mime type
       #   extension, ie. '123.xml'.
+      #
       # @return [OriginalRecord] the instantiated record.
-      # @raise when no matching record is found in the LDP datastore
+      # @raise [LoadError] when no matching record is found in the LDP datastore
       def load(identifier)
         identifier = identifier.to_s.split('/').last if
           identifier.start_with? base_uri
@@ -42,7 +43,7 @@ module Krikri
           record = new(identifier)
         end
 
-        raise "No #{self} found with id: #{identifier}" unless record.exists?
+        raise LoadError.new(identifier), 'No record found' unless record.exists?
 
         if identifier.include?('.')
           record.rdf_subject = "#{base_uri}/#{identifier}"
@@ -201,5 +202,23 @@ module Krikri
     def headers
       { 'Content-Type' => content_type }
     end
+
+    ##
+    # Error class expressing failed load of an OriginalRecord 
+    class LoadError < RuntimeError
+      ##
+      # @param identifier [#to_s]
+      def initialize(identifier)
+        @identifier = identifier
+      end
+      
+      ##
+      # Augments the raised message with generic failure information
+      #
+      # @see Exception#message
+      def message
+        "Could not load OriginalRecord with `#{@identifier}: #{super}"
+      end  
+    end
   end
 end

data/config/settings/development.yml

@@ -4,6 +4,8 @@ marmotta:
   ldp: 'http://localhost:8983/marmotta/ldp'
   record_container: 'http://localhost:8983/marmotta/ldp/original_record'
   item_container: 'http://localhost:8983/marmotta/ldp/items'
+  # read_timeout: Seconds to wait for marmotta's HTTP response.
+  read_timeout: 360
 
 elasticsearch:
   host: 'localhost:9200'

data/config/settings/test.yml

@@ -4,6 +4,8 @@ marmotta:
   ldp: 'http://localhost:8983/marmotta/ldp'
   record_container: 'http://localhost:8983/marmotta/ldp/original_record'
   item_container: 'http://localhost:8983/marmotta/ldp/items'
+  # read_timeout: Seconds to wait for marmotta's HTTP response.
+  read_timeout: 360
 
 elasticsearch:
   host: 'localhost:9200'

data/lib/krikri/engine.rb

@@ -80,7 +80,10 @@ module Krikri
 
     initializer :rdf_repository do
       Krikri::Repository =
-        RDF::Marmotta.new(Krikri::Settings['marmotta']['base'])
+        RDF::Marmotta.new(
+          Krikri::Settings['marmotta']['base'],
+          { read_timeout: Krikri::Settings['marmotta']['read_timeout'] }
+        )
     end
 
     initializer :blacklight_settings do
@@ -126,11 +129,14 @@ module Krikri
         # Get the persisted original record for this Aggregation.
         # @return [Krikri::OriginalRecord, nil]
         #
-        # An Exception will be raised if:
-        #   the original record id is invalid.
+        # @raise [NameError] when the original record id is not a URI or the
         #   the original record has not been persisted to Marmotta.
-        #   there is a connection problem with  Marmotta.
+        #
+        # @raise [Faraday::ConnectionError] when there is a connection problem
+        #   with Marmotta.
         def original_record
+          raise NameError, "#{originalRecord.first} is not an OriginalRecord" if
+            originalRecord.empty? || originalRecord.first.node?
           Krikri::OriginalRecord.load(originalRecord.first.rdf_subject
             .to_s)
         end

data/lib/krikri/entity_behavior.rb

@@ -25,7 +25,7 @@ module Krikri
     # @see lib/krikri/entity_behaviors
     # @see Krikri::Activity#entities
     #
-    def entities
+    def entities(*args)
       raise NotImplementedError 
     end
 
@@ -33,8 +33,8 @@ module Krikri
     # @see Krikri::Activity#entities
     # @see Krikri::EntityBehavior#entities
     #
-    def self.entities(activity)
-      new(activity).entities
+    def self.entities(activity, *args)
+      new(activity).entities(*args)
     end
   end
 end

data/lib/krikri/entity_behaviors/aggregation_entity_behavior.rb

@@ -11,15 +11,18 @@ module Krikri
 
     ##
     # @param load [Boolean]  Whether to load the whole record from the LDP
-    #                        server.  Default: true.
-    #                        DPLA::MAP::Aggregation#get is slow, because it
-    #                        results in a network request, so this provides the
-    #                        possibility of avoiding it.
+    #   server.  DPLA::MAP::Aggregation#get is slow, because it results in a
+    #   network request, so this provides the possibility of avoiding it.
+    #   Default: true.
+    #
+    # @param include_invalidated [Boolean] Whether to include entities that
+    #   have been invalidated with prov:invalidatedAtTime.  Default: false
+    #
     # @see Krikri::EntityBehavior::entities
     # @return [Enumerator] DPLA::MAP::Aggregation objects
     #
-    def entities(load = true)
-      @activity.entity_uris.lazy.map do |uri|
+    def entities(load = true, include_invalidated = false)
+      @activity.entity_uris(include_invalidated).lazy.map do |uri|
         agg = DPLA::MAP::Aggregation.new(uri)
         agg.get if load
         agg

data/lib/krikri/entity_behaviors/original_record_entity_behavior.rb

@@ -11,14 +11,19 @@ module Krikri
 
     ##
     # @param load [Boolean]  Whether to load the whole record from the LDP
-    #                        server.  Default: true.  OriginalRecord.load is
-    #                        slow, because it results in a network request, so
-    #                        this provides the possibility of avoiding it.
+    #   server.  OriginalRecord.load is slow, because it results in a network
+    #   request, so this provides the possibility of avoiding it.
+    #   Default: true.
+    #
+    # @param include_invalidated [Boolean] Whether to include entities that
+    #   have been invalidated with prov:invalidatedAtTime.  Default: false
+    #
     # @see Krikri::EntityBehavior::entities
+    #
     # @return [Enumerator] OriginalRecord objects
     #
-    def entities(load = true)
-      @activity.entity_uris.lazy.map do |uri|
+    def entities(load = true, include_invalidated = false)
+      @activity.entity_uris(include_invalidated).lazy.map do |uri|
         load ? OriginalRecord.load(uri) : OriginalRecord.new(uri)
       end
     end

data/lib/krikri/harvesters/oai_harvester.rb

@@ -97,6 +97,7 @@ module Krikri::Harvesters
                                 record_xml(rec))
           rescue => e
             Krikri::Logger.log(:error, e.message)
+            next
           end
         end
       end

data/lib/krikri/map_crosswalk.rb

@@ -47,7 +47,7 @@ module Krikri
       #   subsequent harvest. Figure out how to do that, or what we want to do
       #   instead.
       def build_aggregation
-        raise 'Tried to index a blank node!' if @parent.node?
+        raise NameError, 'Tried to index a blank node!' if @parent.node?
         hash.merge!({ :ingestType => 'item',
                      :ingestionSequence => 999999,
                      :ingestDate => Date.today.as_json,

data/lib/krikri/provenance_query_client.rb

@@ -12,15 +12,46 @@ module Krikri
     #
     # @param activity_uri [#to_uri] the URI of the activity to search
     #
+    # @param include_invalidated [Boolean]  Whether to include entities that
+    #   have been invalidated with <http://www.w3.org/ns/prov#invalidatedAtTime>
+    #
+    # @see https://www.w3.org/TR/prov-o/#invalidatedAtTime
+    # @see Krikri::LDP::Invalidatable
+    #
     # @return [RDF::SPARQL::Query] a query object that, when executed, will
     #   give solutions containing the URIs for the resources in `#record`.
-    def find_by_activity(activity_uri)
+    def find_by_activity(activity_uri, include_invalidated = false)
       raise ArgumentError, 'activity_uri must be an RDF::URI' unless
         activity_uri.respond_to? :to_term
-      SPARQL_CLIENT.select(:record)
+      query = SPARQL_CLIENT.select(:record)
         .where([:record,
                 [RDF::PROV.wasGeneratedBy, '|', RDF::DPLA.wasRevisedBy],
                 activity_uri.to_term])
+
+      if include_invalidated
+        query
+      else
+        # We need to say "and if RDF::PROV.invalidatedAtTime is not set."
+        #
+        # The SPARQL query should be:
+        #
+        # PREFIX prov: <http://www.w3.org/ns/prov#>
+        # SELECT * WHERE {
+        #   ?subject prov:wasGeneratedBy  <http://xampl.org/ldp/activity/n> .
+        #   FILTER NOT EXISTS { ?subject prov:invalidatedAtTime ?x }
+        # }
+        #
+        # ... However there doesn't appear to be a way of constructing
+        # 'FILTER NOT EXISTS' with SPARQL::Client.  Instead, we've managed to
+        # hack the following solution together.
+        #
+        # SPARQL::Client#filter is labeled @private in its YARD comment (and
+        # has no other documentation) but it's not private, at least for
+        # now.
+        query.filter \
+          'NOT EXISTS ' \
+          '{ ?record <http://www.w3.org/ns/prov#invalidatedAtTime> ?x }'
+      end
     end
   end
 end

data/lib/krikri/version.rb

@@ -1,3 +1,3 @@
 module Krikri
-  VERSION = '0.11.2'.freeze
+  VERSION = '0.12.0-rc.1'.freeze
 end

data/spec/internal/Gemfile

@@ -2,7 +2,7 @@ source 'https://rubygems.org'
 
 
 # Bundle edge Rails instead: gem 'rails', github: 'rails/rails'
-gem 'rails', '4.1.6'
+gem 'rails', '4.1.9'
 # Use sqlite3 as the database for Active Record
 gem 'sqlite3'
 # Use SCSS for stylesheets
@@ -36,7 +36,7 @@ gem 'sdoc', '~> 0.4.0',          group: :doc
 # gem 'debugger', group: [:development, :test]
 
         
-        gem 'krikri', :path => '/home/tjohnson/src/dpla/krikri'
+        gem 'krikri', :path => '/vagrant'
 
 gem "factory_girl_rails", "~> 4.4.0", group: :development
 gem "jettywrapper", "~> 2.0", group: :development

data/spec/internal/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
-  remote: /home/tjohnson/src/dpla/krikri
+  remote: /vagrant
   specs:
-    krikri (0.11.1)
+    krikri (0.11.2)
       analysand (= 4.0.0)
       audumbla (~> 0.1)
       blacklight (~> 5.8.0)
@@ -26,17 +26,17 @@ PATH
 GEM
   remote: https://rubygems.org/
   specs:
-    actionmailer (4.1.6)
-      actionpack (= 4.1.6)
-      actionview (= 4.1.6)
+    actionmailer (4.1.9)
+      actionpack (= 4.1.9)
+      actionview (= 4.1.9)
       mail (~> 2.5, >= 2.5.4)
-    actionpack (4.1.6)
-      actionview (= 4.1.6)
-      activesupport (= 4.1.6)
+    actionpack (4.1.9)
+      actionview (= 4.1.9)
+      activesupport (= 4.1.9)
       rack (~> 1.5.2)
       rack-test (~> 0.6.2)
-    actionview (4.1.6)
-      activesupport (= 4.1.6)
+    actionview (4.1.9)
+      activesupport (= 4.1.9)
       builder (~> 3.1)
       erubis (~> 2.7.0)
     active-triples (0.6.1)
@@ -45,14 +45,14 @@ GEM
       deprecation (~> 0.1)
       linkeddata (~> 1.1)
       rdf (~> 1.1)
-    activemodel (4.1.6)
-      activesupport (= 4.1.6)
+    activemodel (4.1.9)
+      activesupport (= 4.1.9)
       builder (~> 3.1)
-    activerecord (4.1.6)
-      activemodel (= 4.1.6)
-      activesupport (= 4.1.6)
+    activerecord (4.1.9)
+      activemodel (= 4.1.9)
+      activesupport (= 4.1.9)
       arel (~> 5.0.0)
-    activesupport (4.1.6)
+    activesupport (4.1.9)
       i18n (~> 0.6, >= 0.6.9)
       json (~> 1.7, >= 1.7.7)
       minitest (~> 5.1)
@@ -72,9 +72,8 @@ GEM
       dpla-map (~> 4.0.0.0.pre.10)
       geokit
       twofishes
-    autoprefixer-rails (6.3.1)
+    autoprefixer-rails (6.3.3.1)
       execjs
-      json
     bcrypt (3.1.10)
     blacklight (5.8.2)
       bootstrap-sass (~> 3.2)
@@ -92,9 +91,9 @@ GEM
     celluloid-io (0.16.2)
       celluloid (>= 0.16.0)
       nio4r (>= 1.1.0)
-    childprocess (0.5.7)
+    childprocess (0.5.9)
       ffi (~> 1.0, >= 1.0.11)
-    coderay (1.1.0)
+    coderay (1.1.1)
     coffee-rails (4.0.1)
       coffee-script (>= 2.2.0)
       railties (>= 4.0.0, < 5.0)
@@ -111,7 +110,7 @@ GEM
       responders
       thread_safe (~> 0.1)
       warden (~> 1.2.3)
-    domain_name (0.5.25)
+    domain_name (0.5.20160216)
       unf (>= 0.0.5, < 1.0.0)
     dpla-map (4.0.0.0.pre.13)
       active-triples (~> 0.6.0)
@@ -152,7 +151,7 @@ GEM
       domain_name (~> 0.5)
     http_parser.rb (0.6.0)
     i18n (0.7.0)
-    jbuilder (2.4.0)
+    jbuilder (2.4.1)
       activesupport (>= 3.0.0, < 5.1)
       multi_json (~> 1.2)
     jettywrapper (2.0.3)
@@ -198,7 +197,7 @@ GEM
     mail (2.6.3)
       mime-types (>= 1.16, < 3)
     method_source (0.8.2)
-    mime-types (2.99)
+    mime-types (2.99.1)
     mini_portile2 (2.0.0)
     minitest (5.8.4)
     mono_logger (1.1.0)
@@ -206,8 +205,8 @@ GEM
     multipart-post (2.0.0)
     net-http-persistent (2.9.4)
     netrc (0.11.0)
-    nio4r (1.2.0)
-    nokogiri (1.6.7.1)
+    nio4r (1.2.1)
+    nokogiri (1.6.7.2)
       mini_portile2 (~> 2.0.0.rc2)
     oai (0.4.0)
       builder (>= 3.1.0)
@@ -225,21 +224,21 @@ GEM
       rack
     rack-test (0.6.3)
       rack (>= 1.0)
-    rails (4.1.6)
-      actionmailer (= 4.1.6)
-      actionpack (= 4.1.6)
-      actionview (= 4.1.6)
-      activemodel (= 4.1.6)
-      activerecord (= 4.1.6)
-      activesupport (= 4.1.6)
+    rails (4.1.9)
+      actionmailer (= 4.1.9)
+      actionpack (= 4.1.9)
+      actionview (= 4.1.9)
+      activemodel (= 4.1.9)
+      activerecord (= 4.1.9)
+      activesupport (= 4.1.9)
       bundler (>= 1.3.0, < 2.0)
-      railties (= 4.1.6)
+      railties (= 4.1.9)
       sprockets-rails (~> 2.0)
     rails_config (0.4.2)
       activesupport (>= 3.0)
-    railties (4.1.6)
-      actionpack (= 4.1.6)
-      activesupport (= 4.1.6)
+    railties (4.1.9)
+      actionpack (= 4.1.9)
+      activesupport (= 4.1.9)
       rake (>= 0.8.7)
       thor (>= 0.18.1, < 2.0)
     rake (10.5.0)
@@ -293,7 +292,7 @@ GEM
       rdf (~> 1.1, >= 1.1.10)
     rdf-xsd (1.1.5.1)
       rdf (~> 1.1, >= 1.1.9, < 1.99)
-    rdoc (4.2.1)
+    rdoc (4.2.2)
       json (~> 1.4)
     redis (3.2.2)
     redis-namespace (1.5.2)
@@ -313,7 +312,7 @@ GEM
       netrc (~> 0.7)
     rsolr (1.0.13)
       builder (>= 2.1.2)
-    rubyzip (1.1.7)
+    rubyzip (1.2.0)
     sass (3.2.19)
     sass-rails (4.0.5)
       railties (>= 4.0.0, < 5.0)
@@ -323,8 +322,8 @@ GEM
     sdoc (0.4.1)
       json (~> 1.7, >= 1.7.7)
       rdoc (~> 4.0)
-    sinatra (1.4.6)
-      rack (~> 1.4)
+    sinatra (1.4.7)
+      rack (~> 1.5)
       rack-protection (~> 1.4)
       tilt (>= 1.3, < 3)
     slop (3.6.0)
@@ -374,10 +373,10 @@ GEM
       json (>= 1.8.0)
     unf (0.1.4)
       unf_ext
-    unf_ext (0.0.7.1)
+    unf_ext (0.0.7.2)
     vegas (0.1.11)
       rack (>= 1.0.0)
-    warden (1.2.4)
+    warden (1.2.6)
       rack (>= 1.0)
     yajl-ruby (1.2.1)
 
@@ -393,7 +392,7 @@ DEPENDENCIES
   jquery-rails
   krikri!
   pry-rails
-  rails (= 4.1.6)
+  rails (= 4.1.9)
   sass-rails (~> 4.0.3)
   sdoc (~> 0.4.0)
   sqlite3
@@ -401,4 +400,4 @@ DEPENDENCIES
   uglifier (>= 1.3.0)
 
 BUNDLED WITH
-   1.10.6
+   1.11.2

data/spec/internal/config/initializers/blacklight_initializer.rb

@@ -3,5 +3,5 @@
 # will use the application's secret key base instead.
 #
 
-# Blacklight.secret_key = '2e18692814c107ed106221a977f8a5fa38bd1a8355716a818ff72b8b618dd97e07385335c796361f2d18d00302de916adf1e397ad60f725a4b3f3eebd0c3eff2'
+# Blacklight.secret_key = '5a09a69fff3b481232e340d44dfac0cad73f16628e1b5ee7f127b481bbfcce153859c1d4db89826d0417e633d1abd1439e9d9a0c534f093defa560676c3a51bf'
 

data/spec/internal/config/initializers/devise.rb

@@ -4,7 +4,7 @@ Devise.setup do |config|
   # The secret key used by Devise. Devise uses this key to generate
   # random tokens. Changing this key will render invalid all existing
   # confirmation, reset password and unlock tokens in the database.
-  # config.secret_key = '8f0b79f6ceba0441e785de1179473bade347d14f3718eeb09d0c46657653bb30f2592d6123de9cb7316910280cf9c81dd5ca2d1794023a5c1de7a715cfb37d20'
+  # config.secret_key = '06f55f081179339565cd683591dffdaf57f6f340206d19f26fd13a1c9fd18c4d549709d74a960c5636a6f6f3e959e3316a864a73cf3c5d41b012c7c067a6c049'
 
   # ==> Mailer Configuration
   # Configure the e-mail address which will be shown in Devise::Mailer,
@@ -97,7 +97,7 @@ Devise.setup do |config|
   config.stretches = Rails.env.test? ? 1 : 10
 
   # Setup a pepper to generate the encrypted password.
-  # config.pepper = 'ffc94a9f1b5bd168c80153de7513d7e6e61c8066ea2b793750f4732e65e6c62b1e2f8a729c07de2cf3ad5b465b183364e0ce6acde3a4581ca3e2e64f7c07ef93'
+  # config.pepper = '379239f236a953fc942bf473655eedc8b90c653f6404ac06cd45c71690e6565d6671f63701b120cb1170eaa579c387059ad10a9b537c0e7ef32b702b73e6efd6'
 
   # ==> Configuration for :confirmable
   # A period that the user is allowed to access the website even without

data/spec/internal/config/secrets.yml

@@ -11,10 +11,10 @@
 # if you're sharing your code publicly.
 
 development:
-  secret_key_base: baa7b78f0d33e176d43c2e53b15cc34a7a13353589e525b444eca2593b2abcbb9a0b55a1b30196d7886d41d7a5801aa2e4ba6aed862340704dbac5f165ddf961
+  secret_key_base: 40552a8973624429a73ba6bda6a5797bae4ad3495d8404891ace56a633405174fa9365c34e9dc3b579f92ed26f165c72c3e9b454ca814cab20b91893c00eb93d
 
 test:
-  secret_key_base: 775cd9bf2b3e7cb1f9f98ea042f371ca3bbeb565ffaad8314cb89887da924c77bdb1e9be79f03b5790c6345901cee62e5c67df2bf8a80a490b1fe0e4f337882d
+  secret_key_base: b7a204846993178e23440606508244c4a6d865bfaff1d432b9a466b6bd2f522acb45ea1e36aaf39206ff36b0e32b340146fdd09fad6061ec4519b2ae9d7349cc
 
 # Do not keep production secrets in the repository,
 # instead read values from the environment.

data/spec/internal/db/schema.rb

@@ -11,7 +11,7 @@
 #
 # It's strongly recommended that you check this file into your version control system.
 
-ActiveRecord::Schema.define(version: 20160201014949) do
+ActiveRecord::Schema.define(version: 20160224174740) do
 
   create_table "bookmarks", force: true do |t|
     t.integer  "user_id",       null: false