Create RDF.config for configuring cache sizes. Allows a general size, URI and Node specific cache sizes to be configured at startup time.

Fixes #256.

Author: gkellogg

lib/rdf.rb

@@ -2,6 +2,7 @@
 require 'bigdecimal'
 require 'date'
 require 'time'
+require "ostruct"
 
 require 'rdf/version'
 require 'rdf/extensions'
@@ -79,6 +80,29 @@ module RDF
   # CLI
   autoload :CLI,         'rdf/cli'
 
+  ##
+  # Configuration, used open for configuring constants used within the codebase.
+  #
+  # @example set default cache size to be at most 10,000 entries
+  #
+  #   RDF.config.cache_size = 10_000
+  #
+  # @example set cache size for interned URIs to 5,000 entries
+  #
+  #   RDF.config.uri_cache_size = 5_000
+  #
+  # Defaults:
+  #   * `cache_size`: -1
+  #   * `uri_cache_size`: `cache_size`
+  #   * `node_cache_size`: `cache_size`
+  #
+  # @note cache configurations must be set before initial use, when the caches are allocated.
+  # @see RDF::Util::Cache.new
+  # @return [Object]
+  def self.config
+    @config ||= OpenStruct.new(cache_size: -1, uri_cache_size: nil, node_cache_size: nil)
+  end
+
   ##
   # Alias for `RDF::Resource.new`.
   #

lib/rdf/model/node.rb

@@ -13,18 +13,14 @@ class Node
     include RDF::Resource
 
     ##
-    # Defines the maximum number of interned Node references that can be held
-    # cached in memory at any one time.
+    # Cache size may be set through {RDF.config} using `node_cache_size`.
     #
     # @note caching interned nodes means that two different invocations using the same symbol will result in the same node, which may not be appropriate depending on the graph from which it is used. RDF requires that bnodes with the same label are, in fact, different bnodes, unless they are used within the same document.
-    CACHE_SIZE = -1 # unlimited by default
-
-    ##
     # @return [RDF::Util::Cache]
     # @private
     def self.cache
       require 'rdf/util/cache' unless defined?(::RDF::Util::Cache)
-      @cache ||= RDF::Util::Cache.new(CACHE_SIZE)
+      @cache ||= RDF::Util::Cache.new(RDF.config.node_cache_size)
     end
 
     ##

lib/rdf/model/uri.rb

@@ -27,11 +27,6 @@ module RDF
   class URI
     include RDF::Resource
 
-    ##
-    # Defines the maximum number of interned URI references that can be held
-    # cached in memory at any one time.
-    CACHE_SIZE = -1 # unlimited by default
-    
     # IRI components
     UCSCHAR = Regexp.compile(<<-EOS.gsub(/\s+/, ''))
       [\\u00A0-\\uD7FF]|[\\uF900-\\uFDCF]|[\\uFDF0-\\uFFEF]|
@@ -117,11 +112,13 @@ class URI
     ).freeze
 
     ##
+    # Cache size may be set through {RDF.config} using `uri_cache_size`.
+    #
     # @return [RDF::Util::Cache]
     # @private
     def self.cache
       require 'rdf/util/cache' unless defined?(::RDF::Util::Cache)
-      @cache ||= RDF::Util::Cache.new(CACHE_SIZE)
+      @cache ||= RDF::Util::Cache.new(RDF.config.uri_cache_size)
     end
 
     ##

lib/rdf/util/cache.rb

@@ -17,6 +17,9 @@ module RDF; module Util
   # @see   http://en.wikipedia.org/wiki/Weak_reference
   # @since 0.2.0
   class Cache
+    # The configured cache capacity.
+    attr_reader :capacity
+
     ##
     # @private
     def self.new(*args)
@@ -36,8 +39,8 @@ def self.new(*args)
 
     ##
     # @param  [Integer] capacity
-    def initialize(capacity = -1)
-      @capacity = capacity
+    def initialize(capacity = nil)
+      @capacity = capacity || RDF.config.cache_size
       @cache  ||= {}
       @index  ||= {}
     end
@@ -106,7 +109,7 @@ def delete(key)
     class WeakRefCache < Cache
       ##
       # @param  [Integer] capacity
-      def initialize(capacity = -1)
+      def initialize(capacity = nil)
         require 'weakref' unless defined?(::WeakRef)
         super
       end

spec/model_node_spec.rb

@@ -8,7 +8,12 @@
   end
 
   describe ".intern" do
-    before(:each) {RDF::URI.instance_variable_set(:@cache, nil)}
+    before(:each) do
+      RDF::Node.instance_variable_set(:@cache, nil)
+      RDF.config.cache_size = -1
+      RDF.config.node_cache_size = nil
+    end
+
     it "caches Node instance" do
       RDF::Node.intern("a")
       expect(RDF::Node.instance_variable_get(:@cache)[:a]).to eq RDF::Node.intern("a")
@@ -17,6 +22,24 @@
     it "freezes instance" do
       expect(RDF::Node.intern("a")).to be_frozen
     end
+
+    it "defaults cache size to -1" do
+      RDF::Node.intern("a")
+      expect(RDF::Node.instance_variable_get(:@cache).capacity).to eq(-1)
+    end
+
+    it "defaults cache size to configured cache_size" do
+      RDF.config.cache_size = 10_000
+      RDF::Node.intern("a")
+      expect(RDF::Node.instance_variable_get(:@cache).capacity).to eq(10_000)
+    end
+
+    it "defaults cache size to configured uri_cache_size" do
+      RDF.config.cache_size = 10_000
+      RDF.config.node_cache_size = 5_000
+      RDF::Node.intern("a")
+      expect(RDF::Node.instance_variable_get(:@cache).capacity).to eq(5_000)
+    end
   end
 
   describe ".uuid" do

spec/model_uri_spec.rb

@@ -7,7 +7,11 @@
   end
 
   describe ".intern" do
-    before(:each) {RDF::URI.instance_variable_set(:@cache, nil)}
+    before(:each) do
+      RDF::URI.instance_variable_set(:@cache, nil)
+      RDF.config.cache_size = -1
+      RDF.config.uri_cache_size = nil
+    end
     it "caches URI instance" do
       RDF::URI.intern("a")
       expect(RDF::URI.instance_variable_get(:@cache)[:a]).to eq RDF::URI("a")
@@ -24,6 +28,24 @@
     it "does not use #to_hash given a URI" do
       expect {RDF::URI.intern(RDF::URI("a"))}.not_to write.to(:error)
     end
+
+    it "defaults cache size to -1" do
+      RDF::URI.intern(RDF::URI("a"))
+      expect(RDF::URI.instance_variable_get(:@cache).capacity).to eq(-1)
+    end
+
+    it "defaults cache size to configured cache_size" do
+      RDF.config.cache_size = 10_000
+      RDF::URI.intern(RDF::URI("a"))
+      expect(RDF::URI.instance_variable_get(:@cache).capacity).to eq(10_000)
+    end
+
+    it "defaults cache size to configured uri_cache_size" do
+      RDF.config.cache_size = 10_000
+      RDF.config.uri_cache_size = 5_000
+      RDF::URI.intern(RDF::URI("a"))
+      expect(RDF::URI.instance_variable_get(:@cache).capacity).to eq(5_000)
+    end
   end
 
   describe ".parse" do