fathom 0.3.0 → 0.3.1

data/lib/fathom/{data_node.rb → node/data_node.rb}

@@ -1,4 +1,4 @@
-require File.expand_path(File.join(File.dirname(__FILE__), '..', 'fathom'))
+require File.expand_path(File.join(File.dirname(__FILE__), '..', '..', 'fathom'))
 
 =begin
   A DataNode is a node generated from data itself.  It stores the data and reveals some statistical 

data/lib/fathom/{value_aggregator.rb → node/decision.rb}

@@ -1,11 +1,11 @@
-require File.expand_path(File.join(File.dirname(__FILE__), '..', 'fathom'))
-module Fathom
-  class ValueAggregator < ValueDescription
-  end
+require File.expand_path(File.join(File.dirname(__FILE__), '..', '..', 'fathom'))
+class Fathom::Decision < Node
+  undef_method :values
+  undef_method :distribution
 end
 
 if __FILE__ == $0
   include Fathom
   # TODO: Is there anything you want to do to run this file on its own?
-  # ValueAggregator.new
+  # Decision.new
 end

data/lib/fathom/node/discrete_node.rb

@@ -0,0 +1,41 @@
+require File.expand_path(File.join(File.dirname(__FILE__), '..', '..', 'fathom'))
+class Fathom::DiscreteNode < Node
+  attr_reader :labels
+
+  def initialize(opts={})
+    opts[:distribution] ||= :discrete_uniform
+    super(opts)
+    assert_labels(opts)
+  end
+
+  def size
+    @size ||= self.labels.length
+  end
+  alias :length :size
+
+  def rand
+    self.labels[self.distribution.rand(self.size)]
+  end
+  
+  # This makes it easier to interface to a belief node
+  def likelihood(ignored_value)
+    GSL::Vector.ary_to_gv(Array.new(self.size, 1))
+  end
+  alias :l :likelihood
+
+  protected
+    def assert_labels(opts)
+      @labels = opts[:labels]
+      @labels ||= self.values
+      @labels ||= [:true, :false]
+      @labels = Array[@labels] unless @labels.is_a?(Array)
+      @labels.uniq!
+    end
+  
+end
+
+if __FILE__ == $0
+  include Fathom
+  # TODO: Is there anything you want to do to run this file on its own?
+  # DiscreteNode.new
+end

data/lib/fathom/node/fact.rb

@@ -0,0 +1,24 @@
+require File.expand_path(File.join(File.dirname(__FILE__), '..', '..', 'fathom'))
+class Fathom::Fact < Node
+  
+  attr_reader :value
+  
+  def initialize(opts={})
+    symbolize_keys!(opts)
+    @value = opts[:value]
+    @value ||= opts[:values]
+    super(opts)
+  end
+  
+  alias :rand :value
+  
+  undef_method :values
+  undef_method :distribution
+  
+end
+
+if __FILE__ == $0
+  include Fathom
+  # TODO: Is there anything you want to do to run this file on its own?
+  # Fact.new
+end

data/lib/fathom/{mc_node.rb → node/mc_node.rb}

@@ -1,4 +1,4 @@
-require File.expand_path(File.join(File.dirname(__FILE__), '..', 'fathom'))
+require File.expand_path(File.join(File.dirname(__FILE__), '..', '..', 'fathom'))
 class Fathom::MCNode < Node
   
   attr_reader :value_description, :samples_taken

data/lib/fathom/{enforced_name.rb → node/node_extensions/enforced_name.rb}

@@ -1,4 +1,4 @@
-require File.expand_path(File.join(File.dirname(__FILE__), '..', 'fathom'))
+require File.expand_path(File.join(File.dirname(__FILE__), '..', '..', '..', 'fathom'))
 require 'uuid'
 
 module Fathom

data/lib/fathom/{numeric_methods.rb → node/node_extensions/numeric_methods.rb}

@@ -1,4 +1,4 @@
-require File.expand_path(File.join(File.dirname(__FILE__), '..', 'fathom'))
+require File.expand_path(File.join(File.dirname(__FILE__), '..', '..', '..', 'fathom'))
 module Fathom  
   module NumericMethods
     def initialize(opts={})
@@ -46,5 +46,23 @@ module Fathom
       distribution.interval_values(opts.merge(:mean => mean, :sd => sd))
     end
     
+    def coefficient_of_variation
+      return nil unless vector
+      vector.sd / vector.mean
+    end
+    alias :cov :coefficient_of_variation
+    
+    def summary
+      {
+        :mean => mean,
+        :standard_deviation => standard_deviation,
+        :min => vector ? vector.min : nil,
+        :max => vector ? vector.max : nil,
+        :lower_bound => lower_bound,
+        :upper_bound => upper_bound,
+        :coefficient_of_variation => coefficient_of_variation
+      }
+    end
+    
   end
 end

data/lib/fathom/{plausible_range.rb → node/plausible_range.rb}

@@ -1,4 +1,4 @@
-require File.expand_path(File.join(File.dirname(__FILE__), '..', 'fathom'))
+require File.expand_path(File.join(File.dirname(__FILE__), '..', '..', 'fathom'))
 module Fathom
   class PlausibleRange < Node
     

data/spec/ext/array_spec.rb

@@ -0,0 +1,10 @@
+require File.expand_path(File.dirname(__FILE__) + '/../spec_helper')
+
+describe Array do
+  it "should define rand, a way to get a random variable from an array" do
+    a = [1,2]
+    a.should be_include(a.rand)
+    # 9.31322574615479e-10 chance of failing
+    30.times.map {a.rand}.uniq.sort.should eql([1,2])
+  end
+end

data/spec/ext/faster_csv_spec.rb

@@ -0,0 +1,10 @@
+require File.expand_path(File.dirname(__FILE__) + '/../spec_helper')
+require 'fastercsv'
+require 'ext/faster_csv'
+
+describe FasterCSV do
+  it "should have a header converter to strip values" do
+    FasterCSV::HeaderConverters.keys.should be_include(:strip)
+    FasterCSV::HeaderConverters[:strip].call(' this ').should eql('this')
+  end
+end

data/spec/ext/open_struct_spec.rb

@@ -0,0 +1,20 @@
+require File.expand_path(File.dirname(__FILE__) + '/../spec_helper')
+
+describe OpenStruct do
+  
+  before do
+    @o = OpenStruct.new(:this => 1)
+  end
+  
+  it "should expose its table" do
+    @o.table.should eql({:this => 1})
+  end
+  
+  it "should expose the keys from the table" do
+    @o.keys.should eql([:this])
+  end
+  
+  it "should expose the values from the table" do
+    @o.values.should eql([1])
+  end
+end

data/spec/ext/string_spec.rb

@@ -0,0 +1,7 @@
+require File.expand_path(File.dirname(__FILE__) + '/../spec_helper')
+
+describe String do
+  it "should be able to constantize a string" do
+    "Fathom".constantize.should eql(Fathom)
+  end
+end

data/spec/fathom/import/csv_import_spec.rb

@@ -37,14 +37,16 @@ describe CSVImport do
     @result.that.values.should eql([3,6,9])
   end
   
-  it "should store the imported values in the knowledge base" do
-    Fathom.knowledge_base[:this].should be_a(DataNode)
-    Fathom.kb[:this].values.should eql([1,4,7])
-  end
+  # KB Override
+  # it "should store the imported values in the knowledge base" do
+  #   Fathom.knowledge_base[:this].should be_a(DataNode)
+  #   Fathom.kb[:this].values.should eql([1,4,7])
+  # end
 
-  it "should import from the class level" do
-    CSVImport.import(@opts)
-    Fathom.knowledge_base[:this].should be_a(DataNode)
-    Fathom.kb[:this].values.should eql([1,4,7])
-  end
+  # KB Override
+  # it "should import from the class level" do
+  #   CSVImport.import(@opts)
+  #   Fathom.knowledge_base[:this].should be_a(DataNode)
+  #   Fathom.kb[:this].values.should eql([1,4,7])
+  # end
 end

data/spec/fathom/import/yaml_import_spec.rb

@@ -38,16 +38,36 @@ describe YAMLImport do
     @result.co2_readings.should be_a(DataNode)
     @result.co2_readings.values.should eql([10,20,30])
   end
+
+  # KB Override
+  # it "should store the imported values in the knowledge base" do
+  #   Fathom.knowledge_base['CO2 Emissions'].should be_a(PlausibleRange)
+  #   Fathom.kb['CO2 Emissions'].min.should eql(1_000_000)
+  # end
+  
+  # KB Override
+  # it "should import from the class level" do
+  #   YAMLImport.import(@opts)
+  #   Fathom.knowledge_base['CO2 Emissions'].should be_a(PlausibleRange)
+  #   Fathom.kb['CO2 Emissions'].min.should eql(1_000_000)
+  # end
   
-  it "should store the imported values in the knowledge base" do
-    Fathom.knowledge_base['CO2 Emissions'].should be_a(PlausibleRange)
-    Fathom.kb['CO2 Emissions'].min.should eql(1_000_000)
+  it "should not complain if there's an empty YAML file" do
+    filename = '/tmp/empty_yaml_test.yml'
+    File.open(filename, 'w') {|f| f.puts ''}
+    lambda{YAMLImport.import(:content => filename)}.should_not raise_error
+    `rm -rf #{filename}`
   end
   
-  it "should import from the class level" do
-    YAMLImport.import(@opts)
-    Fathom.knowledge_base['CO2 Emissions'].should be_a(PlausibleRange)
-    Fathom.kb['CO2 Emissions'].min.should eql(1_000_000)
+  it "should create a Fact if it is an object named fact and has a name and value function" do
+    filename = File.expand_path(File.dirname(__FILE__) + "/../../support/fact.yml")
+    yaml = open(filename).read
+    opts = {:content => yaml}
+    yi = YAMLImport.new(opts)
+    result = yi.import
+    result.one.value.should eql(1)
+    result.two.value.should eql(2)
+    result.three.value.should eql(3)
   end
   
 end

data/spec/fathom/knowledge_base_spec.rb

@@ -8,9 +8,13 @@ describe KnowledgeBase do
     @kb = KnowledgeBase.new
   end
   
-  it "should be able to add a node" do
-    @dn = DataNode.new(:name => :new_node, :values => [1,2,3])
-    @kb[:new_node] = @dn
-    @kb[:new_node].should eql(@dn)
+  it "should have a find accessor" do
+    KnowledgeBase.should respond_to(:find)
   end
+  
+  # it "should be able to add a node" do
+  #   @dn = DataNode.new(:name => :new_node, :values => [1,2,3])
+  #   @kb[:new_node] = @dn
+  #   @kb[:new_node].should eql(@dn)
+  # end
 end

data/spec/fathom/node/belief_node_spec.rb

@@ -0,0 +1,180 @@
+require File.expand_path(File.dirname(__FILE__) + '/../../spec_helper')
+
+include Fathom
+
+describe BeliefNode do
+  
+  before do
+    @bn = BeliefNode.new(:labels => [:red, :green, :blue])
+  end
+  
+  it "should be a Node" do
+    BeliefNode.ancestors.should be_include(Node)
+  end
+
+  it "should be a DiscreteNode as well" do
+    BeliefNode.ancestors.should be_include(DiscreteNode)
+  end
+  
+  it "should initialize with a uniform probability distribution" do
+    @bn.probabilities.should ==(GSL::Vector[1.0/3, 1.0/3, 1.0/3])
+  end
+  
+  it "should allow a pre-defined probability distribution" do
+    @bn = BeliefNode.new(:labels => [:red, :green, :blue], :probabilities => [0.1, 0.2, 0.7])
+    @bn.probabilities.should ==(GSL::Vector[0.1, 0.2, 0.7])
+  end
+  
+  it "should raise an error if the supplied probability distribution is of the wrong size" do
+    lambda{BeliefNode.new(:labels => [:red, :green, :blue], :probabilities => [0.1, 0.2, 0.7, 0.1])}.should raise_error(/Probabilities must be 3/)
+    lambda{BeliefNode.new(:labels => [:red, :green, :blue], :probabilities => [0.1, 0.2])}.should raise_error(/Probabilities must be 3/)
+  end
+  
+  it "should normalize even the supplied probabilities" do
+    @bn = BeliefNode.new(:labels => [:red, :green, :blue], :probabilities => [1, 2, 7])
+    @bn.probabilities.should ==(GSL::Vector[0.1, 0.2, 0.7])
+  end
+  
+  it "should initialize with a uniform likelihood vector" do
+    @bn.likelihoods.should ==(GSL::Vector[1,1,1])
+  end
+  
+  it "should allow pre-defined likelihoods to be supplied" do
+    @bn = BeliefNode.new(:labels => [:red, :green, :blue], :likelihoods => [1,1,2])
+    @bn.likelihoods.should ==(GSL::Vector[1,1,2])
+  end
+  
+  it "should raise an error if the supplied likelihoods are the wrong size" do
+    lambda{BeliefNode.new(:labels => [:red, :green, :blue], :likelihoods => [1,1,2,1])}.should raise_error(/Likelihoods must be 3/)
+    lambda{BeliefNode.new(:labels => [:red, :green, :blue], :likelihoods => [1,1])}.should raise_error(/Likelihoods must be 3/)
+  end
+  
+  it "should establish a precision threshold at 1.0e-05" do
+    @bn.precision_threshold.should eql(0.00001)
+  end
+  
+  it "should allow the precision_threshold as a parameter" do
+    @bn = BeliefNode.new(:labels => [:red, :green, :blue], :precision_threshold => 0.01)
+    @bn.precision_threshold.should eql(0.01)
+  end
+  
+  it "should allow a initialization from a values hash" do
+    @bn = BeliefNode.new(:values => {:red => 0.1, :green => 0.2, :yellow => 0.3, :blue => 0.4})
+    @bn.probabilities.to_a.sort.should eql([0.1, 0.2, 0.3, 0.4])
+    @bn.labels.map(&:to_s).sort.should eql(%w(blue green red yellow))
+  end
+
+  context "CPMNodes" do
+    before do
+      @b1 = BeliefNode.new(:name => :person, :values => {:david => 0.4, :adina => 0.6})
+      @b2 = BeliefNode.new(:name => :movie_preference, :values => {:drama => 0.5, :comedy => 0.4, :romance => 0.1})
+    end
+    
+    it "should automatically create an intermediary cpm when adding a child" do
+      @b1.add_child(@b2)
+      @b1.children.first.should be_a(CPMNode)
+      @b1.children.first.child.should eql(@b2)
+      @b2.parents.first.should be_a(CPMNode)
+      @b2.parents.first.parent.should eql(@b1)
+      @cpm = @b1.children.first
+      @cpm.should eql(@b2.parents.first)
+    end
+
+    it "should create an accessor for the child, rather than the cpm" do
+      @b1.add_child(@b2)
+      @b1.should_not be_respond_to(:cpm)
+      @b1.should respond_to(:movie_preference)
+      @b1.movie_preference.should eql(@b2)
+    end
+    
+    it "should create a cpm_for_* accessor for the cpm" do
+      @b1.add_child(@b2)
+      @b1.should respond_to(:cpm_for_movie_preference)
+      @b1.cpm_for_movie_preference.should eql(@b1.children.first)
+    end
+    
+    it "should create a cpm_for_* accessor for the child" do
+      @b1.add_child(@b2)
+      @b2.should respond_to(:cpm_for_person)
+      @b2.cpm_for_person.should eql(@b2.parents.first)
+      @b2.cpm_for_person.should eql(@b1.children.first)
+    end
+
+    it "should automatically create an intermediary cpm when adding a parent" do
+      @b2.add_parent(@b1)
+      @b1.children.first.should be_a(CPMNode)
+      @b1.children.first.child.should eql(@b2)
+      @b2.parents.first.should be_a(CPMNode)
+      @b2.parents.first.parent.should eql(@b1)
+      @cpm = @b1.children.first
+      @cpm.should eql(@b2.parents.first)
+    end
+    
+    it "should create an accessor for the parent, rather than the cpm" do
+      @b2.add_parent(@b1)
+      @b2.should_not be_respond_to(:cpm)
+      @b2.should respond_to(:person)
+      @b2.person.should eql(@b1)
+    end
+    
+    it "should create a cpm_for_* accessor for the cpm" do
+      @b2.add_parent(@b1)
+      @b2.should respond_to(:cpm_for_person)
+      @b2.cpm_for_person.should eql(@b2.parents.first)
+    end
+    
+    it "should create an inspect string that reflects the grandchild, rather than the cpm" do
+      @b1.add_child(@b2)
+      @b1.inspect.should eql("Fathom::BeliefNode: person, children:, [\"movie_preference (Fathom::BeliefNode)\", \"movie_preference (Fathom::BeliefNode)\"], parents: , []")
+    end
+    
+    it "should create an inspect string that reflects the grandparent, rather than the cpm" do
+      @b2.add_parent(@b1)
+      @b2.inspect.should eql("Fathom::BeliefNode: movie_preference, children:, [], parents: , [\"person (Fathom::BeliefNode)\", \"person (Fathom::BeliefNode)\"]")
+    end
+
+    it "should create a cpm_for_* accessor for the parent" do
+      @b2.add_parent(@b1)
+      @b1.should respond_to(:cpm_for_movie_preference)
+      @b1.cpm_for_movie_preference.should eql(@b1.children.first)
+      @b1.cpm_for_movie_preference.should eql(@b2.parents.first)
+    end
+    
+  end
+
+  context "Belief and Propagation" do
+    
+    before do
+      
+      @season = BeliefNode.new(:name => :season, :labels => [:winter, :spring, :summer, :fall])
+      @sprinkler = BeliefNode.new(:name => :sprinkler)
+      @rain = BeliefNode.new(:name => :rain)
+      @wet = BeliefNode.new(:name => :wet)
+      @slippery = BeliefNode.new(:name => :slippery)
+
+      @season.add_child(@sprinkler)
+      @sprinkler.add_child(@wet)
+      @season.add_child(@rain)
+      @rain.add_child(@wet)
+      @wet.add_child(@slippery)
+      
+    end
+    
+    it "should provide the likelihood for a particular value" do
+      @sprinkler.should be_respond_to(:likelihood)
+      output = @sprinkler.likelihood(:true)
+      output.should be_a(OpenStruct)
+      # @sprinkler.parents.map(&:name_sym).each {|key| output.table.keys.should be_include(key)}
+      # output.
+      # output.table.keys.should eql([:matrix, :parents])
+    end
+  end
+
+  # What's next:
+  # Deal with multiple parents and the matrix that comes in
+  # Propagate the likelihoods and probabilities up and down
+  # Calculate the belief
+  # Update the information directly on the node and propagate
+  # Receive an update from a child or parent and propagate
+  # Adhere to the precision_threshold when propagating
+end