harvestdor-indexer 0.0.3 → 0.0.10

checksums.yaml

@@ -0,0 +1,15 @@
+---
+!binary "U0hBMQ==":
+  metadata.gz: !binary |-
+    OGYzNWY3ZWNkM2VlNzdhY2YyN2UxZGQ3NzM2MTQyOTRmYzRiZDVjMw==
+  data.tar.gz: !binary |-
+    YTE4YWFhMWEzOGQ0NDJhMDdlNGRlNGFiNjNiMjE1ZWJlNWMyNzQ2Yg==
+!binary "U0hBNTEy":
+  metadata.gz: !binary |-
+    ODdlZDM2MTNkYmYzODNlZmVhY2I0YzJjNGRhMGE3MjMyYzNjNTZiYTJmNDMy
+    ZWI1ZTUyZjZmNWFlNTY3NGUzZGMzMzVlMDY0MTQxNjRiNzRlM2U0OGI3MzU4
+    NTZhNWUwMjE2MjNjMDJjODU4MTA3YzVjOGY5NzQzZGNkMGE1ZDM=
+  data.tar.gz: !binary |-
+    MTYyMzMzNDA2OTM5NWU4ZjdlNGQxYzc3ODU5MTBmNjI5MGM0OTM3NjNkZDdm
+    OWIwY2M1Yjk5OTEyOGIxNGFmYmEwMGNmZDhkZmQ4MmM5MWQxMDAzMmZhMTdl
+    YmNkODM1MDZmMTI4OThhYmFlYWFlMTc5MzEwYzBlMTIyNzg4Yzk=

data/README.rdoc

@@ -108,6 +108,10 @@ I suggest you run your code on harvestdor-dev, as it is already set up to be abl
 
 == Releases
 
+* <b>0.0.7</b> adding additional logging of error, success counts, and time to index and harvest
+* <b>0.0.6</b> tweak error handling for public xml pieces
+* <b>0.0.5</b> make rake release a no-op
+* <b>0.0.4</b> add confstruct runtime dependency
 * <b>0.0.3</b> add methods for public_xml, content_metadata, identity_metadata ...
 * <b>0.0.2</b> better model code for index method (thanks, Bess!)
 * <b>0.0.1</b> initial commit

data/Rakefile

@@ -19,7 +19,8 @@ rescue Bundler::BundlerError => e
 end
 
 desc "DO NOT USE! use dlss_release"
-task :release
+task :release do
+end
 
 task :default => :ci  
 

data/harvestdor-indexer.gemspec

@@ -24,7 +24,7 @@ Gem::Specification.new do |gem|
   gem.add_dependency 'stanford-mods'
   
   # Runtime dependencies
-  # gem.add_runtime_dependency 'nokogiri'
+  gem.add_runtime_dependency 'confstruct'
 
   # Development dependencies
   # Bundler will install these gems too if you've checked out solrmarc-wrapper source from git and run 'bundle install'

data/lib/harvestdor-indexer.rb

@@ -15,7 +15,15 @@ module Harvestdor
   # Base class to harvest from DOR via harvestdor gem and then index
   class Indexer
 
+    attr_accessor :error_count, :success_count, :max_retries
+    attr_accessor :total_time_to_parse,:total_time_to_solr
+
     def initialize yml_path, options = {}
+      @success_count=0    # the number of objects successfully indexed
+      @error_count=0      # the number of objects that failed
+      @max_retries=5      # the number of times to retry an object 
+      @total_time_to_solr=0
+      @total_time_to_parse=0
       @yml_path = yml_path
       config.configure(YAML.load_file(yml_path)) if yml_path    
       config.configure options 
@@ -35,19 +43,62 @@ module Harvestdor
     #   create a Solr profiling document for each druid
     #   write the result to the Solr index
     def harvest_and_index
+      start_time=Time.now
+      logger.info("Started harvest_and_index at #{start_time}")
       if whitelist.empty?
         druids.each { |druid| index druid }
       else
         whitelist.each { |druid| index druid }
       end
       solr_client.commit
-      logger.info("Finished processing: final Solr commit returned.")
+      total_time=elapsed_time(start_time)
+      total_objects=@success_count+@error_count
+      logger.info("Finished harvest_and_index at #{Time.now}: final Solr commit returned")
+      logger.info("Total elapsed time for harvest and index: #{(total_time/60.0)} minutes")
+      logger.info("Avg solr commit time per object (successful): #{@total_time_to_solr/@success_count} seconds") unless (@total_time_to_solr == 0 || @success_count == 0)
+      logger.info("Avg solr commit time per object (all): #{@total_time_to_solr/total_objects} seconds") unless (@total_time_to_solr == 0 || @error_count == 0 || total_objects == 0)
+      logger.info("Avg parse time per object (successful): #{@total_time_to_parse/@success_count} seconds") unless (@total_time_to_parse == 0 || @success_count == 0)
+      logger.info("Avg parse time per object (all): #{@total_time_to_parse/total_objects} seconds") unless (@total_time_to_parse == 0 || @error_count == 0 || total_objects == 0)
+      logger.info("Avg complete index time per object (successful): #{total_time/@success_count} seconds") unless (@success_count == 0)
+      logger.info("Avg complete index time per object (all): #{total_time/total_objects} seconds") unless (@error_count == 0 || total_object == 0)
+      logger.info("Successful count: #{@success_count}")
+      logger.info("Error count: #{@error_count}")
+      logger.info("Total records processed: #{total_objects}")
     end
 
     # return Array of druids contained in the OAI harvest indicated by OAI params in yml configuration file
     # @return [Array<String>] or enumeration over it, if block is given.  (strings are druids, e.g. ab123cd1234)
     def druids
-      @druids ||= harvestdor_client.druids_via_oai
+      if @druids.nil?
+        start_time=Time.now
+        logger.info("Starting OAI harvest of druids at #{start_time}.")  
+        @druids = harvestdor_client.druids_via_oai
+        logger.info("Completed OAI harves of druids at #{Time.now}.  Found #{@druids.size} druids.  Total elapsed time for OAI harvest = #{elapsed_time(start_time,:minutes)} minutes")  
+      end
+      return @druids
+    end
+
+    #add the document to solr, retry if an error occurs
+    def solr_add(doc, id, do_retry=true)
+      #if do_retry is false, skip retrying 
+      tries=do_retry ? 0 : 999
+      while tries < @max_retries
+      begin
+        tries+=1
+        solr_client.add(doc)
+        #return if successful
+        return
+      rescue => e
+        if tries<@max_retries
+          logger.warn "#{id}: #{e.message}, retrying"
+        else
+          @error_count+=1
+          logger.error "Failed saving #{id}: #{e.message}"
+          logger.error e.backtrace
+          return
+        end
+      end
+      end
     end
 
     # create Solr doc for the druid and add it to Solr, unless it is on the blacklist.  
@@ -59,6 +110,8 @@ module Harvestdor
         logger.fatal("You must override the index method to transform druids into Solr docs and add them to Solr")
 
         begin
+          start_time=Time.now
+          logger.info("About to index #{druid} at #{start_time}")
           #logger.debug "About to index #{druid}"
           doc_hash = {}
           doc_hash[:id] = druid
@@ -69,10 +122,12 @@ module Harvestdor
 
           solr_client.add(doc_hash)
 
-          # logger.debug("Just created Solr doc for #{druid}")
+          logger.info("Indexed #{druid} in #{elapsed_time(start_time)} seconds")
+          @success_count+=1
           # TODO: provide call to code to update DOR object's workflow datastream??
         rescue => e
-          logger.error "Failed to index #{druid}: #{e.message}"
+          @error_count+=1
+          logger.error "Failed to index #{druid} in #{elapsed_time(start_time)} seconds: #{e.message}"
         end
       end
     end
@@ -81,7 +136,9 @@ module Harvestdor
     # @param [String] druid e.g. ab123cd4567
     # @return [Stanford::Mods::Record] created from the MODS xml for the druid
     def smods_rec druid
+      start_time=Time.now
       ng_doc = harvestdor_client.mods druid
+      logger.info("Fetched MODs for #{druid} in #{elapsed_time(start_time)} seconds")
       raise "Empty MODS metadata for #{druid}: #{ng_doc.to_xml}" if ng_doc.root.xpath('//text()').empty?
       mods_rec = Stanford::Mods::Record.new
       mods_rec.from_nk_node(ng_doc.root)
@@ -92,45 +149,59 @@ module Harvestdor
     # @param [String] druid e.g. ab123cd4567
     # @return [Nokogiri::XML::Document] the public xml for the DOR object
     def public_xml druid
+      start_time=Time.now
       ng_doc = harvestdor_client.public_xml druid
+      logger.info("Fetched public_xml for #{druid} in #{elapsed_time(start_time)} seconds")
       raise "No public xml for #{druid}" if !ng_doc
       raise "Empty public xml for #{druid}: #{ng_doc.to_xml}" if ng_doc.root.xpath('//text()').empty?
       ng_doc
     end
-    
-    # the contentMetadata for this DOR object, from the purl public xml
-    # @param [String] druid e.g. ab123cd4567
+
+    # the contentMetadata for this DOR object, ultimately from the purl public xml
+    # @param [Object] object a String containing a druid (e.g. ab123cd4567), or 
+    #  a Nokogiri::XML::Document containing the public_xml for an object
     # @return [Nokogiri::XML::Document] the contentMetadata for the DOR object
-    def content_metadata druid
-      ng_doc = harvestdor_client.content_metadata druid
-      raise "No contentMetadata for #{druid}" if !ng_doc || !ng_doc.root
+    def content_metadata object
+      start_time=Time.now
+      ng_doc = harvestdor_client.content_metadata object
+      logger.info("Fetched content_metadata in #{elapsed_time(start_time)} seconds")      
+      raise "No contentMetadata for #{object.inspect}" if !ng_doc || ng_doc.children.empty?
       ng_doc
     end
-    
-    # the identityMetadata for this DOR object, from the purl public xml
-    # @param [String] druid e.g. ab123cd4567
+
+    # the identityMetadata for this DOR object, ultimately from the purl public xml
+    # @param [Object] object a String containing a druid (e.g. ab123cd4567), or 
+    #  a Nokogiri::XML::Document containing the public_xml for an object
     # @return [Nokogiri::XML::Document] the identityMetadata for the DOR object
-    def identity_metadata druid
-      ng_doc = harvestdor_client.identity_metadata druid
-      raise "No identityMetadata for #{druid}" if !ng_doc || !ng_doc.root
+    def identity_metadata object
+      start_time=Time.now
+      ng_doc = harvestdor_client.identity_metadata object
+      logger.info("Fetched identity_metadata in #{elapsed_time(start_time)} seconds")      
+      raise "No identityMetadata for #{object.inspect}" if !ng_doc || ng_doc.children.empty?
       ng_doc
     end
 
-    # the rightsMetadata for this DOR object, from the purl public xml
-    # @param [String] druid e.g. ab123cd4567
+    # the rightsMetadata for this DOR object, ultimately from the purl public xml
+    # @param [Object] object a String containing a druid (e.g. ab123cd4567), or 
+    #  a Nokogiri::XML::Document containing the public_xml for an object
     # @return [Nokogiri::XML::Document] the rightsMetadata for the DOR object
-    def rights_metadata druid
-      ng_doc = harvestdor_client.rights_metadata druid
-      raise "No rightsMetadata for #{druid}" if !ng_doc || !ng_doc.root
+    def rights_metadata object
+      start_time=Time.now
+      ng_doc = harvestdor_client.rights_metadata object
+      logger.info("Fetched rights_metadata in #{elapsed_time(start_time)} seconds")      
+      raise "No rightsMetadata for #{object.inspect}" if !ng_doc || ng_doc.children.empty?
       ng_doc
     end
 
-    # the RDF for this DOR object, from the purl public xml
-    # @param [String] druid e.g. ab123cd4567
+    # the RDF for this DOR object, ultimately from the purl public xml
+    # @param [Object] object a String containing a druid (e.g. ab123cd4567), or 
+    #  a Nokogiri::XML::Document containing the public_xml for an object
     # @return [Nokogiri::XML::Document] the RDF for the DOR object
-    def rdf druid
-      ng_doc = harvestdor_client.rdf druid
-      raise "No RDF for #{druid}" if !ng_doc || !ng_doc.root
+    def rdf object
+      start_time=Time.now      
+      ng_doc = harvestdor_client.rdf object
+      logger.info("Fetched rdf in #{elapsed_time(start_time)} seconds")      
+      raise "No RDF for #{object.inspect}" if !ng_doc || ng_doc.children.empty?
       ng_doc
     end
 
@@ -162,6 +233,20 @@ module Harvestdor
       @harvestdor_client ||= Harvestdor::Client.new({:config_yml_path => @yml_path})
     end
 
+    def elapsed_time(start_time,units=:seconds)
+      elapsed_seconds=Time.now-start_time
+      case units
+      when :seconds
+        return elapsed_seconds.round(2)
+      when :minutes
+        return (elapsed_seconds/60.0).round(1)
+      when :hours
+        return (elapsed_seconds/3600.0).round(2)
+      else
+        return elapsed_seconds
+      end 
+    end
+
     # populate @blacklist as an Array of druids ('oo000oo0000') that will NOT be processed
     #  by reading the File at the indicated path
     # @param [String] path - path of file containing a list of druids
@@ -200,7 +285,7 @@ module Harvestdor
       logger.fatal msg
       raise msg
     end
-    
+
     # Global, memoized, lazy initialized instance of a logger
     # @param [String] log_dir directory for to get log file
     # @param [String] log_name name of log file

data/lib/harvestdor-indexer/version.rb

@@ -1,6 +1,6 @@
 module Harvestdor
   class Indexer
     # this is the Ruby Gem version
-    VERSION = "0.0.3"
+    VERSION = "0.0.10"
   end
 end

data/spec/unit/harvestdor-indexer_spec.rb

@@ -13,13 +13,29 @@ describe Harvestdor::Indexer do
     @whitelist_path = File.join(File.dirname(__FILE__), "../config/ap_whitelist.txt")
   end
   
+  describe "access methods" do
+    it "initializes success count" do
+      @indexer.success_count.should == 0
+    end
+    it "initializes error count" do
+      @indexer.error_count.should == 0
+    end
+    it "initializes max_retries" do
+      @indexer.max_retries.should == 5
+    end
+    it "allows overriding of max_retries" do
+      @indexer.max_retries=6
+      @indexer.max_retries.should == 6
+    end
+  end
+  
   describe "logging" do
     it "should write the log file to the directory indicated by log_dir" do
       @indexer.logger.info("indexer_spec logging test message")
       File.exists?(File.join(@yaml['log_dir'], @yaml['log_name'])).should == true
     end
   end
-
+  
   it "should initialize the harvestdor_client from the config" do
     @hdor_client.should be_an_instance_of(Harvestdor::Client)
     @hdor_client.config.default_set.should == @yaml['default_set']
@@ -32,7 +48,7 @@ describe Harvestdor::Indexer do
       }
     end
     it "should call druids_via_oai and then call :add on rsolr connection" do
-      @hdor_client.should_receive(:druids_via_oai).and_return([@fake_druid])
+      @indexer.should_receive(:druids).and_return([@fake_druid])
       @indexer.solr_client.should_receive(:add).with(@doc_hash)
       @indexer.solr_client.should_receive(:commit)
       @indexer.harvest_and_index
@@ -76,7 +92,7 @@ describe Harvestdor::Indexer do
   end
   
   it "druids method should call druids_via_oai method on harvestdor_client" do
-    @hdor_client.should_receive(:druids_via_oai)
+    @hdor_client.should_receive(:druids_via_oai).and_return([@fake_druid])
     @indexer.druids
   end
   
@@ -129,16 +145,13 @@ describe Harvestdor::Indexer do
         @hdor_client.should_receive(:public_xml).with(@fake_druid).and_return(Nokogiri::XML("<publicObject/>"))
         expect { @indexer.public_xml(@fake_druid) }.to raise_error(RuntimeError, Regexp.new("^Empty public xml for #{@fake_druid}: <"))
       end
-      it "raises Harvestdor::Errors::MissingPurlPage if there is no purl page for the druid" do
-        expect { @indexer.public_xml(@fake_druid) }.to raise_error(Harvestdor::Errors::MissingPurlPage)
-      end
       it "raises error if there is no public_xml page for the druid" do
         @hdor_client.should_receive(:public_xml).with(@fake_druid).and_return(nil)
         expect { @indexer.public_xml(@fake_druid) }.to raise_error(RuntimeError, "No public xml for #{@fake_druid}")
       end
     end
     context "#content_metadata" do
-      it "returns a Nokogiri::XML::Document derived from the public xml" do
+      it "returns a Nokogiri::XML::Document derived from the public xml if a druid is passed" do
         Harvestdor.stub(:public_xml).with(@fake_druid, @indexer.config.purl).and_return(@ng_pub_xml)
         cm = @indexer.content_metadata(@fake_druid)
         cm.should be_kind_of(Nokogiri::XML::Document)
@@ -147,25 +160,23 @@ describe Harvestdor::Indexer do
         cm.root.attributes['objectId'].text.should == @fake_druid
         cm.root.text.strip.should == 'foo'
       end
-      it "raises Harvestdor::Errors::MissingPurlPage if there is no purl page for the druid" do
-        expect { @indexer.content_metadata(@fake_druid) }.to raise_error(Harvestdor::Errors::MissingPurlPage)
-      end
-      it "should raise exception if there is no contentMetadata in the public xml" do
-        pub_xml = "<publicObject id='druid:#{@fake_druid}'>#{@id_md_xml}</publicObject>"
-        Harvestdor.stub(:public_xml).with(@fake_druid, @indexer.config.purl).and_return(Nokogiri::XML(pub_xml))
-        expect { @indexer.content_metadata(@fake_druid) }.to raise_error(RuntimeError, "No contentMetadata for #{@fake_druid}")
+      it "if passed a Nokogiri::XML::Document of the public xml, it does no fetch" do
+        URI::HTTP.any_instance.should_not_receive(:open)
+        @hdor_client.should_receive(:content_metadata).and_call_original
+        cm = @indexer.content_metadata(@ng_pub_xml)
+        cm.should be_kind_of(Nokogiri::XML::Document)
+        cm.root.should_not == nil
+        cm.root.name.should == 'contentMetadata'
+        cm.root.attributes['objectId'].text.should == @fake_druid
+        cm.root.text.strip.should == 'foo'
       end
       it "raises RuntimeError if nil is returned by Harvestdor::Client.contentMetadata for the druid" do
         @hdor_client.should_receive(:content_metadata).with(@fake_druid).and_return(nil)
-        expect { @indexer.content_metadata(@fake_druid) }.to raise_error(RuntimeError, "No contentMetadata for #{@fake_druid}")
-      end
-      it "raises MissingContentMetadata error if there is no contentMetadata in the public_xml for the druid" do
-        URI::HTTP.any_instance.should_receive(:open)
-        expect { @indexer.content_metadata(@fake_druid) }.to raise_error(Harvestdor::Errors::MissingContentMetadata)
+        expect { @indexer.content_metadata(@fake_druid) }.to raise_error(RuntimeError, "No contentMetadata for \"#{@fake_druid}\"")
       end
     end
     context "#identity_metadata" do
-      it "returns a Nokogiri::XML::Document derived from the public xml" do
+      it "returns a Nokogiri::XML::Document derived from the public xml if a druid is passed" do
         Harvestdor.stub(:public_xml).with(@fake_druid, @indexer.config.purl).and_return(@ng_pub_xml)
         im = @indexer.identity_metadata(@fake_druid)
         im.should be_kind_of(Nokogiri::XML::Document)
@@ -173,25 +184,22 @@ describe Harvestdor::Indexer do
         im.root.name.should == 'identityMetadata'
         im.root.text.strip.should == "druid:#{@fake_druid}"
       end
-      it "raises Harvestdor::Errors::MissingPurlPage if there is no purl page for the druid" do
-        expect { @indexer.identity_metadata(@fake_druid) }.to raise_error(Harvestdor::Errors::MissingPurlPage)
-      end
-      it "should raise exception if there is no identityMetadata in the public xml" do
-        pub_xml = "<publicObject id='druid:#{@fake_druid}'>#{@cntnt_md_xml}</publicObject>"
-        Harvestdor.stub(:public_xml).with(@fake_druid, @indexer.config.purl).and_return(Nokogiri::XML(pub_xml))
-        expect { @indexer.identity_metadata(@fake_druid) }.to raise_error(RuntimeError, "No identityMetadata for #{@fake_druid}")
+      it "if passed a Nokogiri::XML::Document of the public xml, it does no fetch" do
+        URI::HTTP.any_instance.should_not_receive(:open)
+        @hdor_client.should_receive(:identity_metadata).and_call_original
+        im = @indexer.identity_metadata(@ng_pub_xml)
+        im.should be_kind_of(Nokogiri::XML::Document)
+        im.root.should_not == nil
+        im.root.name.should == 'identityMetadata'
+        im.root.text.strip.should == "druid:#{@fake_druid}"
       end
       it "raises RuntimeError if nil is returned by Harvestdor::Client.identityMetadata for the druid" do
         @hdor_client.should_receive(:identity_metadata).with(@fake_druid).and_return(nil)
-        expect { @indexer.identity_metadata(@fake_druid) }.to raise_error(RuntimeError, "No identityMetadata for #{@fake_druid}")
-      end
-      it "raises MissingIdentityMetadata error if there is no identityMetadata in the public_xml for the druid" do
-        URI::HTTP.any_instance.should_receive(:open)
-        expect { @indexer.identity_metadata(@fake_druid) }.to raise_error(Harvestdor::Errors::MissingIdentityMetadata)
+        expect { @indexer.identity_metadata(@fake_druid) }.to raise_error(RuntimeError, "No identityMetadata for \"#{@fake_druid}\"")
       end
     end
     context "#rights_metadata" do
-      it "returns a Nokogiri::XML::Document derived from the public xml" do
+      it "returns a Nokogiri::XML::Document derived from the public xml if a druid is passed" do
         Harvestdor.stub(:public_xml).with(@fake_druid, @indexer.config.purl).and_return(@ng_pub_xml)
         im = @indexer.rights_metadata(@fake_druid)
         im.should be_kind_of(Nokogiri::XML::Document)
@@ -199,25 +207,13 @@ describe Harvestdor::Indexer do
         im.root.name.should == 'rightsMetadata'
         im.root.text.strip.should == "bar"
       end
-      it "raises Harvestdor::Errors::MissingPurlPage if there is no purl page for the druid" do
-        expect { @indexer.rights_metadata(@fake_druid) }.to raise_error(Harvestdor::Errors::MissingPurlPage)
-      end
-      it "should raise exception if there is no rightsMetadata in the public xml" do
-        pub_xml = "<publicObject id='druid:#{@fake_druid}'>#{@cntnt_md_xml}</publicObject>"
-        Harvestdor.stub(:public_xml).with(@fake_druid, @indexer.config.purl).and_return(Nokogiri::XML(pub_xml))
-        expect { @indexer.rights_metadata(@fake_druid) }.to raise_error(RuntimeError, "No rightsMetadata for #{@fake_druid}")
-      end
       it "raises RuntimeError if nil is returned by Harvestdor::Client.rightsMetadata for the druid" do
         @hdor_client.should_receive(:rights_metadata).with(@fake_druid).and_return(nil)
-        expect { @indexer.rights_metadata(@fake_druid) }.to raise_error(RuntimeError, "No rightsMetadata for #{@fake_druid}")
-      end
-      it "raises MissingRightsMetadata error if there is no rightsMetadata in the public_xml for the druid" do
-        URI::HTTP.any_instance.should_receive(:open)
-        expect { @indexer.rights_metadata(@fake_druid) }.to raise_error(Harvestdor::Errors::MissingRightsMetadata)
+        expect { @indexer.rights_metadata(@fake_druid) }.to raise_error(RuntimeError, "No rightsMetadata for \"#{@fake_druid}\"")
       end
     end
     context "#rdf" do
-      it "returns a Nokogiri::XML::Document derived from the public xml" do
+      it "returns a Nokogiri::XML::Document derived from the public xml if a druid is passed" do
         Harvestdor.stub(:public_xml).with(@fake_druid, @indexer.config.purl).and_return(@ng_pub_xml)
         im = @indexer.rdf(@fake_druid)
         im.should be_kind_of(Nokogiri::XML::Document)
@@ -225,23 +221,11 @@ describe Harvestdor::Indexer do
         im.root.name.should == 'RDF'
         im.root.text.strip.should == "relationship!"
       end
-      it "raises Harvestdor::Errors::MissingPurlPage if there is no purl page for the druid" do
-        expect { @indexer.rdf(@fake_druid) }.to raise_error(Harvestdor::Errors::MissingPurlPage)
-      end
-      it "should raise exception if there is no rdf in the public xml" do
-        pub_xml = "<publicObject id='druid:#{@fake_druid}'>#{@cntnt_md_xml}</publicObject>"
-        Harvestdor.stub(:public_xml).with(@fake_druid, @indexer.config.purl).and_return(Nokogiri::XML(pub_xml))
-        expect { @indexer.rdf(@fake_druid) }.to raise_error(RuntimeError, "No RDF for #{@fake_druid}")
-      end
       it "raises RuntimeError if nil is returned by Harvestdor::Client.rdf for the druid" do
         @hdor_client.should_receive(:rdf).with(@fake_druid).and_return(nil)
-        expect { @indexer.rdf(@fake_druid) }.to raise_error(RuntimeError, "No RDF for #{@fake_druid}")
-      end
-      it "raises MissingRDF error if there is no rdf in the public_xml for the druid" do
-        URI::HTTP.any_instance.should_receive(:open)
-        expect { @indexer.rdf(@fake_druid) }.to raise_error(Harvestdor::Errors::MissingRDF)
+        expect { @indexer.rdf(@fake_druid) }.to raise_error(RuntimeError, "No RDF for \"#{@fake_druid}\"")
       end
-    end
+    end    
   end
   
   context "blacklist" do

metadata

@@ -1,20 +1,18 @@
 --- !ruby/object:Gem::Specification
 name: harvestdor-indexer
 version: !ruby/object:Gem::Version
-  version: 0.0.3
-  prerelease: 
+  version: 0.0.10
 platform: ruby
 authors:
 - Naomi Dushay
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-03-08 00:00:00.000000000 Z
+date: 2013-10-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rsolr
   requirement: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -22,7 +20,6 @@ dependencies:
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -30,7 +27,6 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: harvestdor
   requirement: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -38,7 +34,6 @@ dependencies:
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -46,7 +41,6 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: stanford-mods
   requirement: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -54,7 +48,20 @@ dependencies:
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
-    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: confstruct
+  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
@@ -62,7 +69,6 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: lyberteam-gems-devel
   requirement: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -70,7 +76,6 @@ dependencies:
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -78,7 +83,6 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -86,7 +90,6 @@ dependencies:
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -94,7 +97,6 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: rdoc
   requirement: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -102,7 +104,6 @@ dependencies:
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -110,7 +111,6 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: yard
   requirement: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -118,7 +118,6 @@ dependencies:
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -126,7 +125,6 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec
   requirement: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -134,7 +132,6 @@ dependencies:
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -142,7 +139,6 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: simplecov
   requirement: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -150,7 +146,6 @@ dependencies:
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -158,7 +153,6 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: simplecov-rcov
   requirement: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -166,7 +160,6 @@ dependencies:
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
-    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -196,33 +189,26 @@ files:
 - spec/unit/harvestdor-indexer_spec.rb
 homepage: https://consul.stanford.edu/display/chimera/Chimera+project
 licenses: []
+metadata: {}
 post_install_message: 
 rdoc_options: []
 require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
-  none: false
   requirements:
   - - ! '>='
     - !ruby/object:Gem::Version
       version: '0'
-      segments:
-      - 0
-      hash: -2920299245033359379
 required_rubygems_version: !ruby/object:Gem::Requirement
-  none: false
   requirements:
   - - ! '>='
     - !ruby/object:Gem::Version
       version: '0'
-      segments:
-      - 0
-      hash: -2920299245033359379
 requirements: []
 rubyforge_project: 
-rubygems_version: 1.8.24
+rubygems_version: 2.0.7
 signing_key: 
-specification_version: 3
+specification_version: 4
 summary: Harvest DOR object metadata and index it to Solr
 test_files:
 - spec/config/ap.yml