gdor-indexer 0.6.0 → 0.7.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: d023031326d6543f222881f79d0a102a6a27f0e9
-  data.tar.gz: 5dd3419e6e38601413f737b3b3011ce3e9baecd9
+  metadata.gz: 37ead667bc97ed2a13e444026fe72ebb79f661fb
+  data.tar.gz: 418391aec671bf772462548a9aa20cf23ed2ea7c
 SHA512:
-  metadata.gz: f182f105f741f8c21bb15c369573b6c3da905b28ef8e0d8920a421dcdacce4beea22891fcbc78dca2c60690c087e800f14f986749b657c218e88fe93247ce072
-  data.tar.gz: 882369a75b3c62d0176746241f60e38204573644855514bc4fc5764ab3fcf97c3758bf7efcdc0fe8bd579102ab355255dfe76ec4730a6c69bffcfdab9b910e76
+  metadata.gz: 68c0d4599bd65d199d84e9664f5c2c6f09896fb9273d2832f9e71cd58072bf3aef32ccf5a52e6ecdfc23315b200a4a464ec8f0248687df414d66f0267d89d9ff
+  data.tar.gz: 6a64eaab7d20112d9b6b59fd5d0673f2f8855c91f1d5be95cf6c64d6e3ec83eefdcb106e9ef012a906f5a6f8d2211c1b766123ce66df6860bf28bc65a2a10b7b

data/.rubocop_todo.yml

@@ -1,11 +1,18 @@
+require: rubocop-rspec
+
 # This configuration was generated by
 # `rubocop --auto-gen-config`
-# on 2016-07-20 15:53:51 -0700 using RuboCop version 0.41.2.
+# on 2016-08-30 14:13:02 -0700 using RuboCop version 0.42.0.
 # The point is for the user to remove these configuration records
 # one by one as the offenses are removed from the code base.
 # Note that changes in the inspected code, or installation of new
 # versions of RuboCop, may require this file to be generated again.
 
+# Offense count: 1
+Lint/AmbiguousRegexpLiteral:
+  Exclude:
+    - 'spec/unit/indexer_spec.rb'
+
 # Offense count: 21
 Metrics/AbcSize:
   Max: 82
@@ -13,19 +20,19 @@ Metrics/AbcSize:
 # Offense count: 1
 # Configuration parameters: CountComments.
 Metrics/ClassLength:
-  Max: 240
+  Max: 242
 
 # Offense count: 5
 Metrics/CyclomaticComplexity:
   Max: 9
 
-# Offense count: 314
+# Offense count: 312
 # Configuration parameters: AllowHeredoc, AllowURI, URISchemes.
 # URISchemes: http, https
 Metrics/LineLength:
   Max: 258
 
-# Offense count: 13
+# Offense count: 15
 # Configuration parameters: CountComments.
 Metrics/MethodLength:
   Max: 43
@@ -64,12 +71,48 @@ RSpec/FilePath:
     - 'spec/unit/solr_doc_hash_spec.rb'
 
 # Offense count: 163
+# Configuration parameters: AssignmentOnly.
 RSpec/InstanceVariable:
   Exclude:
     - 'spec/unit/indexer_spec.rb'
     - 'spec/unit/public_xml_fields_spec.rb'
     - 'spec/unit/solr_doc_builder_spec.rb'
 
+# Offense count: 55
+# Configuration parameters: EnforcedStyle, SupportedStyles.
+# SupportedStyles: allow, expect
+RSpec/MessageExpectation:
+  Exclude:
+    - 'spec/unit/indexer_spec.rb'
+    - 'spec/unit/mods_fields_spec.rb'
+    - 'spec/unit/mods_pub_fields_spec.rb'
+    - 'spec/unit/mods_subject_fields_spec.rb'
+    - 'spec/unit/public_xml_fields_spec.rb'
+    - 'spec/unit/solr_doc_builder_spec.rb'
+    - 'spec/unit/solr_doc_hash_spec.rb'
+
+# Offense count: 23
+RSpec/MultipleExpectations:
+  Max: 8
+
+# Offense count: 81
+RSpec/NamedSubject:
+  Exclude:
+    - 'spec/unit/indexer_spec.rb'
+    - 'spec/unit/solr_doc_hash_spec.rb'
+
+# Offense count: 50
+# Configuration parameters: MaxNesting.
+RSpec/NestedGroups:
+  Exclude:
+    - 'spec/unit/indexer_spec.rb'
+    - 'spec/unit/mods_fields_spec.rb'
+    - 'spec/unit/mods_pub_fields_spec.rb'
+    - 'spec/unit/mods_subject_fields_spec.rb'
+    - 'spec/unit/public_xml_fields_spec.rb'
+    - 'spec/unit/solr_doc_builder_spec.rb'
+    - 'spec/unit/solr_doc_hash_spec.rb'
+
 # Offense count: 7
 # Configuration parameters: IgnoreSymbolicNames.
 RSpec/VerifiedDoubles:
@@ -94,14 +137,6 @@ Style/ClassAndModuleChildren:
     - 'lib/gdor/indexer/solr_doc_builder.rb'
     - 'lib/gdor/indexer/solr_doc_hash.rb'
 
-# Offense count: 1
-# Cop supports --auto-correct.
-# Configuration parameters: EnforcedStyle, SupportedStyles.
-# SupportedStyles: empty_lines, no_empty_lines
-Style/EmptyLinesAroundBlockBody:
-  Exclude:
-    - 'spec/unit/mods_pub_fields_spec.rb'
-
 # Offense count: 22
 # Cop supports --auto-correct.
 # Configuration parameters: EnforcedStyle, SupportedStyles, IndentationWidth.

data/lib/gdor/indexer.rb

@@ -46,7 +46,7 @@ module GDor
       @total_time_to_parse = 0
       @retries = 0
       @druids_failed_to_ix = []
-      @validation_messages = []
+      @validation_messages = Tempfile.new('gdor-indexer-validation-messages')
       @config ||= Confstruct::Configuration.new options
       @config.configure(YAML.load_file(yml_path)) if yml_path && File.exist?(yml_path)
       yield @config if block_given?
@@ -149,7 +149,7 @@ module GDor
       add_coll_info doc_hash, resource.collections # defined in public_xml_fields
       validation_messages = fields_to_add.validate_item(config)
       validation_messages.concat doc_hash.validate_mods(config)
-      @validation_messages.concat(validation_messages)
+      @validation_messages.puts(validation_messages.join("\n"))
       doc_hash.to_h
     end
 
@@ -175,7 +175,7 @@ module GDor
       doc_hash.combine fields_to_add
       validation_messages = doc_hash.validate_collection(config)
       validation_messages.concat doc_hash.validate_mods(config)
-      @validation_messages.concat(validation_messages)
+      @validation_messages.puts(validation_messages.join("\n"))
       doc_hash.to_h
     end
 
@@ -260,12 +260,12 @@ module GDor
       record_count_msgs.each do |msg|
         logger.info msg
       end
-      logger.info("Avg solr commit time per object (successful): #{(@total_time_to_solr / metrics.success_count).round(2)} seconds") unless metrics.success_count == 0
-      logger.info("Avg solr commit time per object (all): #{(@total_time_to_solr / metrics.total).round(2)} seconds") unless metrics.total == 0
-      logger.info("Avg parse time per object (successful): #{(@total_time_to_parse / metrics.success_count).round(2)} seconds") unless metrics.success_count == 0
-      logger.info("Avg parse time per object (all): #{(@total_time_to_parse / metrics.total).round(2)} seconds") unless metrics.total == 0
-      logger.info("Avg complete index time per object (successful): #{(@total_time / metrics.success_count).round(2)} seconds") unless metrics.success_count == 0
-      logger.info("Avg complete index time per object (all): #{(@total_time / metrics.total).round(2)} seconds") unless metrics.total == 0
+      logger.info("Avg solr commit time per object (successful): #{(@total_time_to_solr / metrics.success_count).round(2)} seconds") unless metrics.success_count.zero?
+      logger.info("Avg solr commit time per object (all): #{(@total_time_to_solr / metrics.total).round(2)} seconds") unless metrics.total.zero?
+      logger.info("Avg parse time per object (successful): #{(@total_time_to_parse / metrics.success_count).round(2)} seconds") unless metrics.success_count.zero?
+      logger.info("Avg parse time per object (all): #{(@total_time_to_parse / metrics.total).round(2)} seconds") unless metrics.total.zero?
+      logger.info("Avg complete index time per object (successful): #{(@total_time / metrics.success_count).round(2)} seconds") unless metrics.success_count.zero?
+      logger.info("Avg complete index time per object (all): #{(@total_time / metrics.total).round(2)} seconds") unless metrics.total.zero?
     end
 
     # @return [String] the composed email body
@@ -282,7 +282,8 @@ module GDor
       body += "full log is at gdor_indexer/shared/#{config.harvestdor.log_dir}/#{config.harvestdor.log_name} on #{Socket.gethostname}"
       body += "\n"
 
-      body + @validation_messages.join("\n") + "\n"
+      @validation_messages.rewind
+      body + @validation_messages.read + "\n"
     end
 
     # email the results of indexing if we are on one of the harvestdor boxes

data/lib/gdor/indexer/solr_doc_hash.rb

@@ -13,12 +13,14 @@ class GDor::Indexer
     # when exp_val is a Regexp, looks for String value that matches, or Array with a String member that matches
     # @return [Boolean] true if the field is non-trivially present in the hash, false otherwise
     def field_present?(field, exp_val = nil)
-      return false unless include?(field) && Array(self[field]).any? { |v| !v.blank? }
+      return false unless include?(field)
+      return false unless Array(self[field]).any?(&:present?)
+
       case exp_val
       when nil
         true
       when Regexp
-        Array(self[field]).index { |s| exp_val.match(s) }
+        Array(self[field]).any? { |s| exp_val.match(s) }
       else
         Array(self[field]).include? exp_val
       end

data/lib/gdor/indexer/version.rb

@@ -1,5 +1,5 @@
 module GDor
   class Indexer
-    VERSION = '0.6.0'.freeze
+    VERSION = '0.7.1'.freeze
   end
 end

data/spec/unit/indexer_spec.rb

@@ -12,7 +12,7 @@ describe GDor::Indexer do
     @pub_xml = "<publicObject id='druid#{@fake_druid}'></publicObject>"
     @ng_pub_xml = Nokogiri::XML("<publicObject id='druid#{@fake_druid}'></publicObject>")
   end
-  before(:each) do
+  before do
     @indexer = described_class.new(@config_yml_path) do |config|
       config.whitelist = ['druid:ww121ss5000']
     end
@@ -69,7 +69,7 @@ describe GDor::Indexer do
   end
 
   describe '#harvest_and_index' do
-    before :each do
+    before do
       allow(@indexer.harvestdor).to receive(:each_resource)
       allow(@indexer).to receive(:solr_client).and_return(double(commit!: nil))
       allow(@indexer).to receive(:log_results)
@@ -246,7 +246,7 @@ describe GDor::Indexer do
   end #  index_coll_obj_per_config
 
   context '#add_coll_info and supporting methods' do
-    before(:each) do
+    before do
       @coll_druids_array = [collection]
     end
     let(:doc_hash) { GDor::Indexer::SolrDocHash.new({}) }
@@ -288,7 +288,7 @@ describe GDor::Indexer do
     end
 
     context '#coll_display_types_from_items' do
-      before(:each) do
+      before do
         @indexer.coll_display_types_from_items(collection)
       end
       it 'gets single item display_type for single collection (and no dups)' do
@@ -311,7 +311,7 @@ describe GDor::Indexer do
   end # add_coll_info
 
   context '#num_found_in_solr' do
-    before :each do
+    before do
       @collection_response = { 'response' => { 'numFound' => '1', 'docs' => [{ 'id' => 'dm212rn7381', 'url_fulltext' => ['https://purl.stanford.edu/dm212rn7381'] }] } }
       @item_response = { 'response' => { 'numFound' => '265', 'docs' => [{ 'id' => 'dm212rn7381' }] } }
     end
@@ -329,7 +329,7 @@ describe GDor::Indexer do
   end # num_found_in_solr
 
   context '#email_report_body' do
-    before :each do
+    before do
       @indexer.config.notification = 'notification-list@example.com'
       allow(@indexer).to receive(:num_found_in_solr).and_return(500)
       allow(@indexer.harvestdor).to receive(:resources).and_return([collection])
@@ -355,8 +355,8 @@ describe GDor::Indexer do
     end
 
     it 'email body include validation messages' do
-      @indexer.instance_variable_set(:@validation_messages, ['this is a validation message'])
-      expect(subject).to match(/this is a validation message/)
+      @indexer.instance_variable_set(:@validation_messages, instance_double(File, rewind: 0, read: 'this is a validation message'))
+      expect(subject).to match /this is a validation message/
     end
 
     it 'email includes reference to full log' do
@@ -365,7 +365,7 @@ describe GDor::Indexer do
   end
 
   describe '#email_results' do
-    before :each do
+    before do
       @indexer.config.notification = 'notification-list@example.com'
       allow(@indexer).to receive(:send_email)
       allow(@indexer).to receive(:email_report_body).and_return('Report Body')

data/spec/unit/mods_pub_fields_spec.rb

@@ -20,7 +20,6 @@ describe GDor::Indexer::ModsFields do
   let(:sdb) { sdb_for_mods(mods_xml) }
 
   context 'publication date fields' do
-
     RSpec.shared_examples 'expected (dateIssued)' do |solr_field_sym, mods_field_val, exp_val|
       it "#{exp_val} for #{mods_field_val}" do
         m = mods_origin_info_start_str +

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gdor-indexer
 version: !ruby/object:Gem::Version
-  version: 0.6.0
+  version: 0.7.1
 platform: ruby
 authors:
 - Naomi Dushay
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-08-06 00:00:00.000000000 Z
+date: 2016-08-30 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: harvestdor-indexer