umlaut 3.0.5 → 3.1.0.pre1

data/app/models/service_store.rb-NEW

@@ -0,0 +1,73 @@
+# Loads Service definitions from Rails.root/config/umlaut_services.yml
+# instantiates services from definitions, by id.
+#
+# It's terrible we need to do this globally like this, but
+# too hard to back out of legacy design now.
+class ServiceStore
+  @@global_service_store = ServiceStore.new
+  def self.global_service_store
+    @@global_service_store
+  end
+
+  # certain class methods all default to global default ServiceStore,
+  # for global ServiceStore. 
+  [:config, :service_definitions, :service_definition_for, :'instantiate_service!'].each do |method|
+    self.define_singleton_method(method) do |*args|      
+      global_service_store.send(method, *args)
+    end
+  end
+
+
+  # Returns complete hash loaded from services.yml
+  def config
+    # cache hash loaded from YAML, ensure it has the keys we expect.
+    unless defined? @services_config_list
+      yaml_path = File.expand_path("config/umlaut_services.yml", Rails.root)
+      @services_config_list = (File.exists? yaml_path) ? YAML::load(File.open(yaml_path)) : {}
+      @services_config_list["default"] ||= {}
+      @services_config_list["default"]["services"] ||= {}
+    end
+    return @services_config_list
+  end
+
+  # Returns hash keyed by unique service name, value service definition
+  # hash.
+  def service_definitions
+    unless defined? @service_definitions
+      @service_definitions = {}
+      config.each_pair do |group_name, group|
+        if group["services"]
+          # Add the group name to each service
+          # in the group
+          group["services"].each_pair do |service_id, service|
+            service["group"] = group_name
+          end
+          # Merge the group's services into the service definitions.
+          @service_definitions.merge!(  group["services"]  )
+        end
+      end
+      # set service_id key in each based on hash key
+      @service_definitions.each_pair do |key, hash|
+        hash["service_id"] =  key
+      end
+    end
+    return @service_definitions
+  end
+
+  def service_definition_for(service_id)
+    return service_definitions[service_id]
+  end
+
+  # pass in string unique key OR a service definition hash,
+  # and a current UmlautRequest.
+  # get back instantiated Service object.
+  def instantiate_service!(service, request)
+    definition = service.kind_of?(Hash) ? service : service_definition_for(service.to_s)
+    raise "Service '#{service}'' does not exist in umlaut-services.yml" if definition.nil?
+    className = definition["type"] || definition["service_id"]
+    classConst = Kernel.const_get(className)
+    service = classConst.new(definition)
+    service.request = request
+    return service
+  end
+end

data/app/models/service_wave.rb

@@ -1,3 +1,5 @@
+require 'term_color'
+
 # ServiceWave is basically responsible for multi-threaded execution of
 # a list of services, all in the same priority. Generally it's only
 # called by Collection, nobody else needs it directly. 

data/app/models/sfx4/abstract/README.md

@@ -0,0 +1,68 @@
+SFX4 Abstract AZ Models
+---
+
+The modules in `Sfx4::Abstract` represent SFX4 A-Z tables for connecting to local SFX instances.
+They are abstracted out so that the Umlaut implementer can include the relevant module for the particular instance of the model.
+
+In addition to representing the AZ tables, individual modules implement specific functionality.
+
+  - `Sfx4::Abstract::Base` implements functionality for searching SFX for "SFX controlled" URLs.
+  - `Sfx4::Abstract::AzTitle` implements Sunspot functionality for indexing SFX records in Solr.
+
+Examples are the classes in the `Sfx4::Local` module.
+
+    module Sfx4
+      module Local
+        class Base < ActiveRecord::Base
+          self.establish_connection :sfx_db
+          # ActiveRecord likes it when we tell it this is an abstract
+          # class only. 
+          self.abstract_class = true 
+          extend Sfx4::Abstract::Base
+
+          # All SFX things are read-only!
+          def readonly?() 
+            return true
+          end
+        end
+      end
+    end
+
+    module Sfx4
+      module Local
+        class AzTitle < Sfx4::Local::Base
+          include Sfx4::Abstract::AzTitle
+        end
+      end
+    end
+
+If your Umlaut implementation needs to point to an additional SFX DB (e.g. for consortial reasons), you can create another Sfx4 module
+that holds your classes for the additional SFX DB (as long as the configuration is specified in database.yml).
+
+    # Current file: /app/model/sfx4/additional_instance/base.rb
+    module Sfx4
+      module AdditionalInstance
+        class Base < ActiveRecord::Base
+          self.establish_connection :sfx_db_additional_instance
+          # ActiveRecord likes it when we tell it this is an abstract
+          # class only. 
+          self.abstract_class = true 
+          extend Sfx4::Abstract::Base
+
+          # All SFX things are read-only!
+          def readonly?() 
+            return true
+          end
+        end
+      end
+    end
+
+    # Current file: /app/model/sfx4/additional_instance/az_title.rb
+    module Sfx4
+      module AdditionalInstance
+        class AzTitle < Sfx4::AdditionalInstance::Base
+          include Sfx4::Abstract::AzTitle
+        end
+      end
+    end
+    

data/app/models/sfx4/abstract/az_extra_info.rb

@@ -0,0 +1,44 @@
+# +------------------+---------------------+------+-----+---------+----------------+
+# | Field            | Type                | Null | Key | Default | Extra          |
+# +------------------+---------------------+------+-----+---------+----------------+
+# | AZ_EXTRA_INFO_ID | int(10) unsigned    | NO   | PRI | NULL    | auto_increment |
+# | AZ_PROFILE       | varchar(100)        | NO   |     | NULL    |                |
+# | OBJECT_ID        | bigint(20) unsigned | NO   | MUL | 0       |                |
+# | EXTRA_INFO_XML   | mediumtext          | YES  |     | NULL    |                |
+# +------------------+---------------------+------+-----+---------+----------------+
+module Sfx4
+  module Abstract
+    module AzExtraInfo
+    
+      def self.included(klass)
+        klass.class_eval do
+          require 'nokogiri'
+          self.table_name = 'AZ_EXTRA_INFO'
+          self.primary_key = 'AZ_EXTRA_INFO_ID'
+
+          belongs_to :az_title,
+                     :foreign_key => 'OBJECT_ID',
+                     :class_name => "#{klass.to_s.deconstantize}::AzTitle"
+
+          include MetadataHelper # for normalize_lccn
+        end
+      end
+
+      def issn
+        @issn ||= extra_info_xml.search("item[key=issn]").text
+      end
+
+      def isbn
+        @isbn ||= extra_info_xml.search("item[key=isbn]").text
+      end
+
+      def lccn
+        @lccn ||= normalize_lccn(extra_info_xml.search("item[key=lccn]").text)
+      end
+
+      def extra_info_xml
+        @extra_info_xml ||= Nokogiri::XML(self.EXTRA_INFO_XML)
+      end
+    end
+  end
+end

data/app/models/sfx4/abstract/az_letter_group.rb

@@ -0,0 +1,24 @@
+# +----------------------+------------------+------+-----+---------+----------------+
+# | Field                | Type             | Null | Key | Default | Extra          |
+# +----------------------+------------------+------+-----+---------+----------------+
+# | AZ_LETTER_GROUP_ID   | int(10) unsigned | NO   | PRI | NULL    | auto_increment |
+# | AZ_TITLE_ID          | int(10) unsigned | NO   |     | 0       |                |
+# | AZ_LETTER_GROUP_NAME | char(10)         | NO   | MUL |         |                |
+# +----------------------+------------------+------+-----+---------+----------------+
+module Sfx4
+  module Abstract
+    module AzLetterGroup
+      def self.included(klass)
+        klass.class_eval do
+          self.table_name = 'AZ_LETTER_GROUP'
+          self.primary_key = 'AZ_LETTER_GROUP_ID'
+
+          belongs_to :az_title,
+                     :foreign_key => 'AZ_TITLE_ID',
+                     :class_name => "#{klass.to_s.deconstantize}::AzTitle"
+
+        end
+      end
+    end
+  end
+end

data/app/models/sfx4/abstract/az_title.rb

@@ -0,0 +1,96 @@
+# +-------------------+---------------------+------+-----+---------+----------------+
+# | Field             | Type                | Null | Key | Default | Extra          |
+# +-------------------+---------------------+------+-----+---------+----------------+
+# | AZ_TITLE_ID       | int(10) unsigned    | NO   | PRI | NULL    | auto_increment |
+# | AZ_PROFILE        | varchar(100)        | NO   | MUL | NULL    |                |
+# | OBJECT_ID         | bigint(20) unsigned | NO   | MUL | 0       |                |
+# | TITLE_DISPLAY     | varchar(255)        | NO   |     |         |                |
+# | TITLE_SORT        | varchar(200)        | NO   |     |         |                |
+# | SCRIPT            | varchar(20)         | NO   |     |         |                |
+# | PAGINATION_SUBSTR | varchar(20)         | NO   |     |         |                |
+# | PEER_REVIEWED     | enum('YES','NO')    | YES  |     | NULL    |                |
+# +-------------------+---------------------+------+-----+---------+----------------+
+module Sfx4
+  module Abstract
+    module AzTitle
+      def self.included(klass)
+        klass.class_eval do
+          self.table_name = 'AZ_TITLE'
+          self.primary_key = 'AZ_TITLE_ID'
+
+          belongs_to  :kb_object,
+                      :foreign_key => 'OBJECT_ID',
+                      :class_name => "Sfx4::Global::KbObject"
+
+          has_many  :az_title_searches,
+                    :foreign_key => 'AZ_TITLE_ID',
+                    :class_name => "#{klass.to_s.deconstantize}::AzTitleSearch"
+
+          has_many  :az_letter_groups,
+                    :foreign_key => 'AZ_TITLE_ID',
+                    :class_name => "#{klass.to_s.deconstantize}::AzLetterGroup"
+
+          has_one   :az_extra_info,
+                    :primary_key => 'OBJECT_ID',
+                    :foreign_key => 'OBJECT_ID',
+                    :class_name => "#{klass.to_s.deconstantize}::AzExtraInfo"
+
+          # Only add Sunspot code if Umlaut is configured for Sunspot.
+          if sunspot?
+            searchable :if => :index? do
+              # Indexed fields
+              text :title do
+                [self.TITLE_DISPLAY, self.TITLE_SORT].concat(az_title_searches.map{|az_title_search| az_title_search.TITLE_SEARCH}).uniq
+              end
+              string :title_exact, :multiple => true do
+                [self.TITLE_DISPLAY, self.TITLE_SORT].concat(az_title_searches.map{|az_title_search| az_title_search.TITLE_SEARCH}).uniq
+              end
+              string :letter_group, :multiple => true do
+                az_letter_groups.collect{ |az_letter_group|
+                  (az_letter_group.AZ_LETTER_GROUP_NAME.match(/[0-9]/)) ? 
+                    "0-9" : az_letter_group.AZ_LETTER_GROUP_NAME }
+              end
+              string :title_sort do
+                self.TITLE_SORT
+              end
+              # Stored strings.
+              string :object_id, :stored => true do
+                self.OBJECT_ID
+              end
+              string :title_display, :stored => true do
+                self.TITLE_DISPLAY
+              end
+              string :issn, :stored => true do
+                az_extra_info.issn unless az_extra_info.nil?
+              end
+              string :isbn, :stored => true do
+                az_extra_info.isbn unless az_extra_info.nil?
+              end
+              string :lccn, :stored => true do
+                az_extra_info.lccn unless az_extra_info.nil?
+              end
+            end
+          end
+        end
+      end
+
+      def index?
+        self.AZ_PROFILE.eql? UmlautController.umlaut_config.lookup!("search.sfx_az_profile", "default")
+      end
+
+      # This code isn't used anywhere at the moment, but is kept around for posterities sake.
+      def to_context_object(context_object)
+        ctx = OpenURL::ContextObject.new
+        # Start out wtih everything in search, to preserve date/vol/etc
+        ctx.import_context_object(context_object)
+        # Put SFX object id in rft.object_id, that's what SFX does.
+        ctx.referent.set_metadata('object_id', self.OBJECT_ID.to_s )
+        ctx.referent.set_metadata("jtitle", self.TITLE_DISPLAY || "Unknown Title")
+        ctx.referent.set_metadata("issn", az_extra_info.issn ) unless az_extra_info.nil? or az_extra_info.issn.blank?
+        ctx.referent.set_metadata("isbn", az_extra_info.isbn) unless az_extra_info.nil? or az_extra_info.isbn.blank?
+        ctx.referent.add_identifier("info:lccn/#{az_extra_info.lccn}") unless az_extra_info.nil? or az_extra_info.lccn.blank?
+        return ctx
+      end
+    end
+  end
+end

data/app/models/sfx4/abstract/az_title_search.rb

@@ -0,0 +1,24 @@
+# +--------------------+------------------+------+-----+---------+----------------+
+# | Field              | Type             | Null | Key | Default | Extra          |
+# +--------------------+------------------+------+-----+---------+----------------+
+# | AZ_TITLE_SEARCH_ID | int(10) unsigned | NO   | PRI | NULL    | auto_increment |
+# | AZ_PROFILE         | varchar(100)     | NO   | MUL | NULL    |                |
+# | AZ_TITLE_ID        | int(10) unsigned | NO   | MUL | 0       |                |
+# | TITLE_SEARCH       | text             | NO   | MUL | NULL    |                |
+# +--------------------+------------------+------+-----+---------+----------------+
+module Sfx4
+  module Abstract
+    module AzTitleSearch
+      def self.included(klass)
+        klass.class_eval do
+          self.table_name = 'AZ_TITLE_SEARCH'
+          self.primary_key = 'AZ_TITLE_SEARCH_ID'
+
+          belongs_to :az_title,
+                     :foreign_key => 'AZ_TITLE_ID',
+                     :class_name => "#{klass.to_s.deconstantize}::AzTitle"
+        end
+      end
+    end
+  end
+end

data/app/models/sfx4/abstract/base.rb

@@ -0,0 +1,77 @@
+module Sfx4
+  module Abstract
+    module Base
+      # Is Umlaut configured to use Sunspot?
+      def sunspot?
+        begin
+          Sunspot and Sunspot.const_defined?(:Rails) and self.ancestors.include?(Sunspot::Rails::Searchable)
+        rescue NameError
+          # no need to log, jsut annoying people who choose not to use it. 
+          #warn "Sunspot::Rails has not been implemented in this Umlaut instance."
+          false
+        end
+      end
+
+      # Class method for the module that gets called by the umlaut:load_sfx_urls task.
+      # Kind of hacky way of trying to extract target URLs from SFX4.
+      # Will probably be deprecated in the near future.
+      def fetch_urls
+        # Crazy crazy URLs to try to find PARSE_PARAMS in Sfx4 db that have a period in
+        # them, so they look like they might be URLs. Parse params could be at target service
+        # level, or at portfolio level; and could be in local overrides or in global kb. 
+        # This is crazy crazy SQL to get this, sorry. Talking directly to SFX db isn't
+        # a great idea, but best way we've found to get this for now. Might make more
+        # sense to try to use the (very very slow) SFX export in the future instead. 
+        sql = %{
+          SELECT 
+            COALESCE(LCL_SERVICE_LINKING_INFO.PARSE_PARAM,KB_TARGET_SERVICES.PARSE_PARAM) PARSE_PARAM
+          FROM
+            LCL_TARGET_INVENTORY
+          JOIN sfxglb41.KB_TARGET_SERVICES
+            ON KB_TARGET_SERVICES.TARGET_ID = LCL_TARGET_INVENTORY.TARGET_ID
+          JOIN LCL_SERVICE_INVENTORY 
+            ON LCL_TARGET_INVENTORY.TARGET_ID = LCL_SERVICE_INVENTORY.TARGET_ID
+          LEFT OUTER JOIN LCL_SERVICE_LINKING_INFO
+            ON LCL_SERVICE_INVENTORY.TARGET_SERVICE_ID = LCL_SERVICE_LINKING_INFO.TARGET_SERVICE_ID
+          WHERE
+            ( LCL_SERVICE_LINKING_INFO.PARSE_PARAM like '%.%' OR
+              KB_TARGET_SERVICES.PARSE_PARAM like '%.%' )
+          AND
+            LCL_SERVICE_INVENTORY.ACTIVATION_STATUS='ACTIVE'
+          AND
+            LCL_TARGET_INVENTORY.ACTIVATION_STATUS = 'ACTIVE'
+
+          UNION
+            -- object portfolio parse param version
+            SELECT
+              COALESCE(LCL_OBJECT_PORTFOLIO_LINKING_INFO.PARSE_PARAM, KB_OBJECT_PORTFOLIOS.PARSE_PARAM) PARSE_PARAM
+            FROM
+              sfxglb41.KB_OBJECT_PORTFOLIOS
+            JOIN LCL_SERVICE_INVENTORY
+              ON KB_OBJECT_PORTFOLIOS.TARGET_SERVICE_ID = LCL_SERVICE_INVENTORY.TARGET_SERVICE_ID
+            JOIN LCL_OBJECT_PORTFOLIO_INVENTORY
+              ON KB_OBJECT_PORTFOLIOS.OP_ID = LCL_OBJECT_PORTFOLIO_INVENTORY.OP_ID
+            left outer join  LCL_OBJECT_PORTFOLIO_LINKING_INFO
+              ON KB_OBJECT_PORTFOLIOS.OP_ID = LCL_OBJECT_PORTFOLIO_LINKING_INFO.OP_ID        
+            WHERE
+              ( KB_OBJECT_PORTFOLIOS.PARSE_PARAM like '%.%' OR 
+                LCL_OBJECT_PORTFOLIO_LINKING_INFO.PARSE_PARAM like '%.%' )
+            AND LCL_OBJECT_PORTFOLIO_INVENTORY.ACTIVATION_STATUS = 'ACTIVE'        
+            AND LCL_SERVICE_INVENTORY.ACTIVATION_STATUS='ACTIVE'
+          }
+
+        results =  connection.select_all(sql)
+        urls = []
+        results.each do |line|
+          param_string = line["PARSE_PARAM"]
+          # Try to get things that look sort of like URLs out. Brutal force,
+          # sorry. 
+          url_re = Regexp.new('(https?://\S+\.\S+)(\s|$)')
+          urls.concat( param_string.scan( url_re ).collect {|matches| matches[0]} )                
+        end      
+        urls.uniq!
+        return urls
+      end
+    end
+  end
+end

data/app/models/sfx4/global/base.rb

@@ -0,0 +1,24 @@
+module Sfx4
+  module Global
+    class Base < ActiveRecord::Base
+      attr_accessible
+
+      # Was a SFX Global DB connection set in database.yml to connect directly to sfx?
+      def self.connection_configured?
+        config = ActiveRecord::Base.configurations["sfx4_global"]
+        (not (config.nil? or config.blank? or config["adapter"].blank?))
+      end
+
+      self.establish_connection :sfx4_global if self.connection_configured?
+
+      # ActiveRecord likes it when we tell it this is an abstract
+      # class only. 
+      self.abstract_class = true
+
+      # All SFX things are read-only!
+      def readonly?() 
+        return true
+      end
+    end
+  end
+end