ruby-fann 0.7.1

data/tasks/deployment.rake

@@ -0,0 +1,34 @@
+desc 'Release the website and new gem version'
+task :deploy => [:check_version, :website, :release] do
+  puts "Remember to create SVN tag:"
+  puts "svn copy svn+ssh://#{rubyforge_username}@rubyforge.org/var/svn/#{PATH}/trunk " +
+    "svn+ssh://#{rubyforge_username}@rubyforge.org/var/svn/#{PATH}/tags/REL-#{VERS} "
+  puts "Suggested comment:"
+  puts "Tagging release #{CHANGES}"
+end
+
+desc 'Runs tasks website_generate and install_gem as a local deployment of the gem'
+task :local_deploy => [:website_generate, :install_gem]
+
+task :check_version do
+  unless ENV['VERSION']
+    puts 'Must pass a VERSION=x.y.z release version'
+    exit
+  end
+  unless ENV['VERSION'] == VERS
+    puts "Please update your version.rb to match the release version, currently #{VERS}"
+    exit
+  end
+end
+
+desc 'Install the package as a gem, without generating documentation(ri/rdoc)'
+task :install_gem_no_doc => [:clean, :package] do
+  sh "#{'sudo ' unless Hoe::WINDOZE }gem install pkg/*.gem --no-rdoc --no-ri"
+end
+
+namespace :manifest do
+  desc 'Recreate Manifest.txt to include ALL files'
+  task :refresh do
+    `rake check_manifest | patch -p0 > Manifest.txt`
+  end
+end

data/tasks/environment.rake

@@ -0,0 +1,7 @@
+task :ruby_env do
+  RUBY_APP = if RUBY_PLATFORM =~ /java/
+    "jruby"
+  else
+    "ruby"
+  end unless defined? RUBY_APP
+end

data/tasks/website.rake

@@ -0,0 +1,17 @@
+desc 'Generate website files'
+task :website_generate => :ruby_env do
+  (Dir['website/**/*.txt'] - Dir['website/version*.txt']).each do |txt|
+    sh %{ #{RUBY_APP} script/txt2html #{txt} > #{txt.gsub(/txt$/,'html')} }
+  end
+end
+
+desc 'Upload website files to rubyforge'
+task :website_upload do
+  host = "#{rubyforge_username}@rubyforge.org"
+  remote_dir = "/var/www/gforge-projects/#{PATH}/"
+  local_dir = 'website'
+  sh %{rsync -aCv #{local_dir}/ #{host}:#{remote_dir}}
+end
+
+desc 'Generate and upload website files'
+task :website => [:website_generate, :website_upload, :publish_docs]

data/test/test_helper.rb

@@ -0,0 +1,2 @@
+require 'test/unit'
+require File.dirname(__FILE__) + '/../lib/ruby_fann'

data/test/test_neurotica.rb

@@ -0,0 +1,36 @@
+require 'test/unit'
+require "rubygems" 
+require "graphviz"
+require 'ext/ruby_fann/neural_network'
+require 'lib/ruby_fann/neurotica'
+
+class NeuroticaTest < Test::Unit::TestCase
+  def test_basic_output
+    
+    train = RubyFann::TrainData.new(
+      :inputs=>[[0.3, 0.4, 0.5, 1.0, -1.0], [0.1, 0.2, 0.3, 1.0, 1.0], [0.6, 0.74, 0.58, -1.0, -1.0], [0.109, 0.677, 0.21, -1.0, 1.0]], 
+      :desired_outputs=>[[0.7, 0.4, 0.9], [0.8, -0.2, -0.5], [-0.33, 0.34, -0.22], [0.129, -0.87, 0.25]])
+
+    neural_net = RubyFann::Standard.new(:num_inputs=>3, :hidden_neurons=>[4, 7, 9, 3, 5], :num_outputs=>3)
+    neural_net.train_on_data(train, 100, 20, 0.01)    
+        
+    # train = RubyFann::TrainData.new(:inputs=>[[0.3, 0.4, 0.5], [0.1, 0.2, 0.3]], :desired_outputs=>[[0.7], [0.8]])
+    neural_net = RubyFann::Shortcut.new(:num_inputs=>3, :num_outputs=>3)    
+    neural_net.cascadetrain_on_data(train, 5, 10, 0.1)    
+    
+    
+    neurotica = RubyFann::Neurotica.new
+    
+    neurotica.graph(neural_net, "neurotica1.png")
+  end
+
+  def test_3d_output
+    train = RubyFann::TrainData.new(:inputs=>[[0.3, 0.4, 0.5], [0.1, 0.2, 0.3]], :desired_outputs=>[[0.7], [0.8]])
+    #neural_net = RubyFann::Shortcut.new(9, 5, [2, 4, 12, 8, 4, 3, 4], 1)    
+    neural_net = RubyFann::Shortcut.new(:num_inputs=>4, :num_outputs=>1)    
+    neural_net.cascadetrain_on_data(train, 1000, 100, 0.1)    
+    neurotica = RubyFann::Neurotica.new
+    neurotica.graph(neural_net, "neurotica2.vrml", :three_dimensional=>true)
+    assert(File.exist?('neurotica2.vrml'))  	
+  end
+end

data/test/test_ruby_fann.rb

@@ -0,0 +1,390 @@
+require 'test/unit'
+require 'ext/ruby_fann/neural_network'
+
+class MyShortcut < RubyFann::Shortcut
+  def initialize
+    super(:num_inputs=>5, :num_outputs=>1)    
+  end
+end
+
+class RubyFannTest < Test::Unit::TestCase
+  def test_create_standard
+    fann = RubyFann::Standard.new(:num_inputs=>1, :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1)    
+    assert(fann)
+  end
+  
+  def test_create_shortcut
+    fann = RubyFann::Shortcut.new(:num_inputs=>5, :num_outputs=>1)    
+    
+    assert(fann)
+    
+    assert_equal(2, fann.get_num_layers)
+    assert_equal(5, fann.get_num_input)
+    assert_equal(1, fann.get_num_output)    
+
+    fann2 = RubyFann::Shortcut.new(:num_inputs=>1, :num_outputs=>2)    
+    assert(fann2)
+    assert_equal(2, fann2.get_num_layers)
+    assert_equal(1, fann2.get_num_input)
+    assert_equal(2, fann2.get_num_output)    
+    assert_equal(:fann_nettype_shortcut, fann.get_network_type)
+        
+    sc = MyShortcut.new
+  end
+
+  def test_raises
+    fann = nil
+    assert_raises(TypeError) { fann = RubyFann::Standard.new(:num_inputs=>'foo', :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1)}
+    assert_nil(fann)    
+    assert_raises(ArgumentError)  { fann = RubyFann::Standard.new(4, 1, [3, 4], 1) }        
+    assert_nil(fann)    
+    assert_nothing_raised { fann = RubyFann::Standard.new(:num_inputs=>4, :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1) }    
+    assert(fann)
+  end
+  
+  def test_run
+    fann = RubyFann::Standard.new(:num_inputs=>4, :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1)
+    outputs = fann.run([3.0, 2.0, 3.0, 4.0])
+    assert_equal(1, outputs.length)
+  end
+  
+  def test_print_parameters
+    fann = RubyFann::Standard.new(:num_inputs=>4, :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1)
+    fann.print_parameters
+  end
+  
+  def test_accessors
+    fann = RubyFann::Standard.new(:num_inputs=>4, :hidden_neurons=>[3, 4, 3], :num_outputs=>1)
+    assert_equal(5, fann.get_num_layers)
+    assert_equal(4, fann.get_num_input)
+    assert_equal(1, fann.get_num_output)
+    puts "TN: #{fann.get_total_neurons}"
+    puts "TC: #{fann.get_total_connections}"
+    fann.print_parameters
+    assert_equal(19, fann.get_total_neurons)
+  end
+  
+  def test_randomize_weights
+    fann = RubyFann::Standard.new(:num_inputs=>4, :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1)
+    fann.randomize_weights(-1.0, 1.0)
+  end
+  
+  def test_init_weights
+    fann = RubyFann::Standard.new(:num_inputs=>4, :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1)
+    training = RubyFann::TrainData.new(:filename=>'test/test.train')    
+    fann.init_weights(training)
+  end
+  
+  
+  def test_print_connections
+    fann = RubyFann::Standard.new(:num_inputs=>4, :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1)
+    fann.print_connections
+  end
+  
+  def test_connection_rate
+    fann = RubyFann::Standard.new(:num_inputs=>4, :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1)
+    cr = fann.get_connection_rate
+    assert_equal(1, cr)    
+  end
+  
+  def test_layer_array
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1) 
+    layers = fann.get_layer_array
+    assert_equal([5, 2, 8, 4, 3, 4, 1], layers)
+  end
+  
+  def test_bias_array
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 4, 3, 4], :num_outputs=>1) 
+    bias = fann.get_bias_array
+    assert_equal([1, 1, 1, 1, 1, 0], bias)
+  end
+  
+  
+  def test_get_network_type
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1) 
+    assert_equal(:fann_nettype_layer, fann.get_network_type)
+    puts "fann.get_network_type: #{fann.get_network_type}"
+  end
+
+  def test_train_file
+    training = RubyFann::TrainData.new(:filename=>'test/test.train')
+    training.save('verify.train')
+    assert(File.exist?('verify.train'))  	
+    test_train_txt = File.read('test/test.train')
+    verify_train_txt = File.read('verify.train')
+    assert_equal(verify_train_txt, test_train_txt)    
+  end 
+  
+  def test_train_creation_data
+    training = RubyFann::TrainData.new(:inputs=>[[0.3, 0.4, 0.5], [0.1, 0.2, 0.3]], :desired_outputs=>[[0.7], [0.8]])
+    training.save('verify.train')
+    verify_lines = File.readlines('verify.train')
+    assert_equal(5, verify_lines.length)
+    assert_match(/2 3 1/, verify_lines[0])
+    assert_match(/0.30* 0.40* 0.50*/, verify_lines[1])
+    assert_match(/0.7/, verify_lines[2])
+    assert_match(/0.10* 0.20* 0.30*/, verify_lines[3])
+    assert_match(/0.8/, verify_lines[4])
+  end
+  
+  def test_train_on_data
+    train = RubyFann::TrainData.new(:inputs=>[[0.3, 0.4, 0.5], [0.1, 0.2, 0.3]], :desired_outputs=>[[0.7], [0.8]])
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1)
+    fann.train_on_data(train, 1000, 10, 0.1)
+    outputs = fann.run([3.0, 2.0, 3.0])    
+    puts "OUTPUT FROM RUN WAS #{outputs.inspect}"
+  end
+  
+  def test_activation_function
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1) 
+    fann.set_activation_function(:linear, 1, 2)
+    assert_raises(TypeError) { fann.set_activation_function('linear', 1, 2) }
+  end  
+  
+  def test_activation_function_hidden
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1)
+    fann.set_activation_function_hidden(:linear)
+    assert_raises(RuntimeError) { fann.set_activation_function_hidden(:fake) }
+    fann.set_activation_function_hidden(:elliot)
+  end
+
+  def test_activation_function_output
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1)  
+    fann.set_activation_function_output(:linear)                                                      
+    assert_raises(RuntimeError) { fann.set_activation_function_output(:fake) }                        
+    fann.set_activation_function_output(:elliot)                                                      
+  end                                                                                                 
+                                                                                                      
+  def test_activation_steepness                                                                       
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1)  
+    fann.set_activation_steepness(0.2, 2, 3)                                                          
+    assert_equal(0.2, fann.get_activation_steepness(2, 3))                                            
+  end	                                                                                                
+                                                                                                      
+  def test_activation_steepness_hidden                                                                
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1)  
+    fann.set_activation_steepness_hidden(0.345)                                                       
+    1.upto(5) {|i|assert_equal(0.345, fann.get_activation_steepness(i, 1)) }                          
+  end                                                                                                 
+                                                                                                      
+  def test_activation_steepness_layer                                                                 
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1)  
+    fann.set_activation_steepness_layer(0.7000, 3)                                                    
+    assert_equal(0.7000, fann.get_activation_steepness(3, 1))                                         
+  end                                                                                                 
+                                                                                                      
+  def test_activation_steepness_output                                                                
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1) 
+    fann.set_activation_steepness_output(0.888)                                                       
+    assert_equal(0.888, fann.get_activation_steepness(6, 1))                                          
+  end                                                                                                 
+                                                                                                      
+  def test_training_function                                                                          
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1) 
+    fann.set_train_error_function(:tanh)                                                              
+    assert_equal(:tanh, fann.get_train_error_function())                                              
+                                                                                                      
+    fann2 = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1) 
+    fann2.set_train_error_function(:linear)                                                           
+    assert_equal(:linear, fann2.get_train_error_function())                                           
+    assert_raises(RuntimeError) { fann2.set_train_error_function(:fake) }                             
+  end                                                                                                 
+                                                                                                      
+  def test_training_stop                                                                              
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1) 
+    fann.set_train_stop_function(:mse)                                                                
+    assert_equal(:mse, fann.get_train_stop_function())                                                
+                                                                                                      
+    fann2 = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1) 
+    fann2.set_train_stop_function(:bit)                                                               
+    assert_equal(:bit, fann2.get_train_stop_function())                                               
+                                                                                                      
+    assert_raises(RuntimeError) { fann2.set_train_stop_function(:fake) }                              
+  end                                                                                                 
+                                                                                                      
+  def test_save                                                                                       
+    fann = RubyFann::Standard.new(:num_inputs=>5, :hidden_neurons=>[2, 8, 4, 3, 4], :num_outputs=>1) 
+    fann.save('foo.net')
+    assert(File.exist?('foo.net'))
+    File.delete('foo.net')
+  end
+  
+  def test_training_algorithm
+    verify_fann_attribute(:training_algorithm, :rprop)
+  end
+  
+  def test_bit_fail_limit
+    verify_fann_attribute(:bit_fail_limit, 0.743)
+  end
+
+  def test_quickprop_decay
+    verify_fann_attribute(:quickprop_decay, 0.211)
+  end
+
+  def test_quickprop_mu
+    verify_fann_attribute(:quickprop_mu, 0.912)
+  end
+
+  def test_rprop_increase_factor
+    verify_fann_attribute(:rprop_increase_factor, 0.743)
+  end
+
+  def test_rprop_decrease_factor
+    verify_fann_attribute(:rprop_decrease_factor, 0.190)
+  end
+
+  def test_rprop_delta_min
+    verify_fann_attribute(:rprop_delta_min, 0.277)
+  end
+  
+  def test_rprop_delta_max
+    verify_fann_attribute(:rprop_delta_max, 0.157)
+  end
+  
+  def test_rprop_delta_zero
+    verify_fann_attribute(:rprop_delta_zero, 0.571)
+  end
+
+  def test_learning_momentum
+    verify_fann_attribute(:learning_momentum, 0.231)
+  end
+
+  def test_learning_rate
+    verify_fann_attribute(:learning_rate, 0.012)
+  end
+  
+  def test_cascade_train
+    train = RubyFann::TrainData.new(:inputs=>[[0.3, 0.4, 0.5], [0.1, 0.2, 0.3]], :desired_outputs=>[[0.7], [0.8]])
+    fann = RubyFann::Shortcut.new(:num_inputs=>5, :num_outputs=>1) # RubyFann::Shortcut.new(7, 5, [2, 8, 4, 3, 4], 1)    
+    fann.cascadetrain_on_data(train, 1000, 10, 0.1)    
+  end
+  
+  def test_get_neurons
+    train = RubyFann::TrainData.new(:inputs=>[[0.3, 0.4, 0.5], [0.1, 0.2, 0.3]], :desired_outputs=>[[0.7, 0.3], [0.8, 0.4]])
+    fann = RubyFann::Shortcut.new(:num_inputs=>5, :num_outputs=>2) # RubyFann::Shortcut.new(7, 5, [2, 8, 4, 3, 4], 1)    
+    fann.cascadetrain_on_data(train, 30, 10, 0.0000000000001)    
+    neurons = fann.get_neurons()
+    assert_equal(38, neurons.length)
+    neurons.each_with_index {|n, i| puts "NEURON[#{i}]: #{n.inspect}" }
+  end
+
+  def test_cascade_output_change_fraction
+    verify_fann_attribute(:cascade_output_change_fraction, 0.222)
+  end
+
+  def test_cascade_output_stagnation_epochs
+    verify_fann_attribute(:cascade_output_stagnation_epochs, 4)
+  end
+
+  def test_cascade_candidate_change_fraction
+    verify_fann_attribute(:cascade_candidate_change_fraction, 0.987)
+  end
+
+  def test_cascade_candidate_stagnation_epochs
+    verify_fann_attribute(:cascade_candidate_stagnation_epochs, 5)
+  end
+
+  def test_cascade_weight_multiplier
+    verify_fann_attribute(:cascade_weight_multiplier, 0.754)
+  end
+
+  def test_cascade_candidate_limit
+    verify_fann_attribute(:cascade_candidate_limit, 0.222)
+  end
+  
+  def test_cascade_cascade_max_out_epochs
+    verify_fann_attribute(:cascade_max_out_epochs, 77)
+  end
+
+  def test_cascade_max_cand_epochs
+    verify_fann_attribute(:cascade_max_cand_epochs, 66)
+  end
+  
+  def test_cascade_num_candidate_groups
+    verify_fann_attribute(:cascade_num_candidate_groups, 6)
+  end
+    
+  def test_cascade_num_candidates
+    fann = RubyFann::Standard.new(:num_inputs=>1, :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1)     # RubyFann::Standard.new(7, 5, [2, 8, 4, 3, 4], 1)    
+    x = fann.get_cascade_activation_functions_count
+    y = fann.get_cascade_activation_steepnesses_count
+    z = fann.get_cascade_num_candidate_groups
+    assert_equal((x * y * z), fann.get_cascade_num_candidates)
+  end
+  
+  def test_cascade_activation_functions
+    fann = RubyFann::Standard.new(:num_inputs=>1, :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1)     # RubyFann::Standard.new(7, 5, [2, 8, 4, 3, 4], 1)    
+    fann.set_cascade_activation_functions([:threshold, :sigmoid])
+    assert_equal(2, fann.get_cascade_activation_functions_count)
+    assert_equal([:threshold, :sigmoid], fann.get_cascade_activation_functions)
+  end
+
+  def test_cascade_activation_steepness
+    fann = RubyFann::Standard.new(:num_inputs=>1, :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1)     # RubyFann::Standard.new(7, 5, [2, 8, 4, 3, 4], 1)    
+    fann.set_cascade_activation_steepnesses([0.1, 0.3, 0.7, 0.5])
+    assert_equal(4, fann.get_cascade_activation_steepnesses_count)
+    assert_equal([0.1, 0.3, 0.7, 0.5], fann.get_cascade_activation_steepnesses)
+  end
+    
+  def test_train_exception
+    # Mismatched inputs & outputs
+    assert_raises(RuntimeError) { 
+      RubyFann::TrainData.new(
+        { 
+          :inputs=>[[0.2, 0.3, 0.4], [0.8, 0.9, 0.7]], 
+          :desired_outputs=>[[3.14]]
+        }
+      ) 
+    } 
+    
+    # Wrong arg type:
+    assert_raises(TypeError) { RubyFann::TrainData.new('not_a_hash') } 
+    
+    # Bad key in hash:
+    assert_raises(RuntimeError) { RubyFann::TrainData.new({:not_a_real_value=>1}) } 
+    
+    # Bad value in hash:
+    assert_raises(RuntimeError) { RubyFann::TrainData.new({:inputs=>1, :desired_outputs=>2}) }
+    
+    # Inconsistent inputs
+    assert_raises(RuntimeError) { 
+      RubyFann::TrainData.new(
+        { 
+          :inputs=>[[0.2, 0.3, 0.4], [0.8]], 
+          :desired_outputs=>[[3.14], [4.33]]
+        }
+      ) 
+    } 
+    
+    # Inconsistent outputs
+    assert_raises(RuntimeError) { 
+      RubyFann::TrainData.new(
+        { 
+          :inputs=>[[0.2, 0.3, 0.4], [0.5, 0.8, 0.7]], 
+          :desired_outputs=>[[3.14], [0.4, 0.5]]
+        }
+      ) 
+    }
+    
+    # No errors:
+    assert_nothing_raised(){ 
+      RubyFann::TrainData.new(
+        { 
+          :inputs=>[[0.2, 0.3, 0.4], [0.8, 0.9, 0.7]], 
+          :desired_outputs=>[[3.14], [6.33]]
+        }
+      ) 
+    }
+    
+  end
+
+private
+  # Set & get fann attribute & verify:
+  def verify_fann_attribute(attr, val)
+    fann = RubyFann::Standard.new(:num_inputs=>1, :hidden_neurons=>[3, 4, 3, 4], :num_outputs=>1)
+    fann.send("set_#{attr}", val)    
+    assert_equal(val, fann.send("get_#{attr}"))    
+  end
+  
+end
+