bors 0.0.0 → 0.0.1

data/lib/bors.rb

@@ -1,67 +1,66 @@
-require_relative 'bors/example'
-require_relative 'bors/model'
-require_relative 'bors/prediction'
 require 'open3'
 require 'tempfile'
+require 'fileutils'
 
-class Bors
+require_relative 'bors/example'
+require_relative 'bors/result'
+require_relative 'bors/command_line'
 
+class Bors
 	attr_reader :examples
 
-	def initialize(options = {})
-		@examples = Tempfile.new('bors')
-		@example_count = 0
+	def initialize(options)
+		raise Exceptions::ArgumentError.new('You must specify an examples file') unless options[:examples_file]
 		@options = options
+		@examples_file = File.new(options[:examples_file], 'a')
 	end
 
 	def add_example(details)
-		begin
-			@examples.write("#{Example.new(details).to_s}\n")
-			@example_count += 1
-		end
+		@examples_file.write("#{Example.new(details).to_s}\n")
 	end
 
 	# returns an example from the file
 	def get_example(at)
-		get_line_from_file(@examples, at)
-	end
-
-	# load examples from path, should automatically detect a cache file and use that instead
-	def load_examples(path)
-
-	end
-
-	# outputs examples to the specified file
-	def save_examples(path)
-		raise Exceptions::MissingExamples.new unless @examples.length > 0
-		# should copy tempfile to path then automatically reload it
+		get_line_from_file(@examples_file, at) # should convert to Example class instead of String
 	end
 
-	# creates a training model from the loaded examples
-	def model(path)
-		raise Exceptions::MissingExamples.new unless @examples.length > 0
-		err, out, status = Open3.capture3("vw #{@examples.path} -c --passes 30 -b 25 --invariant -l 10 --loss_function logistic --exact_adaptive_norm -f #{path}.model")
-		Model.new(err, out, status)
+	# returns the examples as a string in VW format
+	def examples
+		with_closed_examples_file do
+			return File.open(@examples_file.path, 'r').read
+		end
 	end
 
-	# predict using a model and the loaded exampels
-	def predict(model_path)
-		raise Exceptions::MissingExamples.new unless @examples.length > 0
-		out, err, status = Open3.capture3("vw -i #{model_path} -t #{@examples.path} -p /dev/stdout --quiet")
-		Prediction.new(out)
+	# Runs vowpal wabbit with the loaded examples
+	def run!(run_options={})
+		with_closed_examples_file do
+			raise Exceptions::MissingExamples.new unless File.open(@examples_file.path, 'r').readlines.count > 0
+			run_options.merge!({:examples => @examples_file.path})
+			out = CommandLine.new(run_options).run!
+			FileUtils.rm(@examples_file.path) if @options[:temp_examples] == true
+			return Result.new(out)
+		end
 	end
 
 	private
 
 	def get_line_from_file(path, line)
 		result = nil
-		File.open(path, "r") do |f|
-			while line > 0
-				line -= 1
-				result = f.gets
+		with_closed_examples_file do
+			File.open(path, "r") do |f|
+				while line > 0
+					line -= 1
+					result = f.gets
+				end
 			end
 		end
-		return result
+		result
+	end
+
+	def with_closed_examples_file(&block)
+		@examples_file.close unless @examples_file.closed?
+		block.call
+		@examples_file = File.open(@examples_file.path, 'a')
 	end
 
 end

data/spec/bors/example.rb

@@ -12,6 +12,10 @@ module Indus
 			Bors::Example.new({ :label => 1 }).label.should == 1
 		end
 
+		it "should allow you to define a negative label for the object" do
+			Bors::Example.new({ :label => -1 }).label.should == -1
+		end
+
 		it "should raise an error when trying to define an invalid label" do
 			expect { Bors::Example.new({ :label => "invalid" }).label }.to raise_error Bors::Exceptions::NotRealNumber
 		end
@@ -60,7 +64,7 @@ module Indus
 						]
 					}
 				}
-			}).namespaces.should == "|Diet:2 Other small animals |Description:16 Legs:4 Horns:2"
+			}).features.should == "|Diet:2 Other small animals |Description:16 Legs:4 Horns:2"
 		end
 
 	end

data/spec/bors/result/samples.rb

@@ -0,0 +1,17 @@
+require_relative "../../spec_helper"
+
+describe Bors::Result::Samples do
+
+	before :each do
+		@data = File.open("#{File.dirname(__FILE__)}/../../fixtures/result.txt", 'r').read
+	end
+
+	it "should be able to create the object" do
+		expect { Bors::Result::Samples.new(@data) }.to_not raise_error
+	end
+
+	it "should be able to serialise the object to json" do
+		Bors::Result::Samples.new(@data).to_json.kind_of?(String).should be == true
+	end
+
+end

data/spec/bors/result/settings.rb

@@ -0,0 +1,31 @@
+require_relative "../../spec_helper"
+
+describe Bors::Result::Settings do
+
+	before :each do
+		@data = File.open("#{File.dirname(__FILE__)}/../../fixtures/result.txt", 'r').read
+	end
+
+	it "should be able to create the object" do
+		expect { Bors::Result::Settings.new(@data) }.to_not raise_error
+	end
+
+	it "should be able to return the object as a hash with the correct values" do
+		hash = Bors::Result::Settings.new(@data).to_h
+		hash.kind_of?(Hash).should be == true
+		hash[:final_regressor].should be == "/bors_models/location-adelaide.model.model"
+		hash[:num_weight_bits].should be == 25
+		hash[:learning_rate].should be == 10
+		hash[:initial_t].should be == 0
+		hash[:power_t].should be == 0.5
+		hash[:decay_learning_rate].should be == 1
+		hash[:creating_cache_file].should be == "/var/folders/ny/vd6j4wq12pz4y96n8v3487cc0000gn/T/bors20130117-24035-xuv6sw.cache"
+		#hash[:reading_from].should be == "/var/folders/ny/vd6j4wq12pz4y96n8v3487cc0000gn/T/bors20130117-24035-xuv6sw"
+		hash[:num_sources].should be == 1
+	end
+
+	it "should be able to serialise the object to json" do
+		Bors::Result::Settings.new(@data).to_json.kind_of?(String).should be == true
+	end
+
+end

data/spec/bors/result/statistics.rb

@@ -0,0 +1,28 @@
+require_relative "../../spec_helper"
+
+describe Bors::Result::Statistics do
+
+	before :each do
+		@data = File.open("#{File.dirname(__FILE__)}/../../fixtures/result.txt", 'r').read
+	end
+
+	it "should be able to create the object" do
+		expect { Bors::Result::Statistics.new(@data) }.to_not raise_error
+	end
+
+	it "should be able to return the object as a hash with the correct values" do
+		hash = Bors::Result::Statistics.new(@data).to_h
+		hash.kind_of?(Hash).should be == true
+		hash[:number_of_examples].should be == 69720
+		hash[:weighted_example_sum].should be == 6.972e+04
+		hash[:weighted_label_sum].should be == -6.702e+04
+		hash[:average_loss].should be == 0.01571
+		hash[:best_constant].should be == -0.9613
+		hash[:total_feature_number].should be == 289312080
+	end
+
+	it "should be able to serialise the object to json" do
+		Bors::Result::Settings.new(@data).to_json.kind_of?(String).should be == true
+	end
+
+end

data/spec/bors/result.rb

@@ -0,0 +1,32 @@
+require_relative "../spec_helper"
+
+module Indus
+
+	describe Bors::Result do
+
+		before :each do
+			@data = File.open("#{File.dirname(__FILE__)}/../fixtures/result.txt", 'r').read
+		end
+
+		it "should be able to create a model object" do
+			expect { Bors::Result.new(@data) }.to_not raise_error
+		end
+
+		it "should return settings of the model run" do
+			settings = Bors::Result.new(@data).settings
+			settings.kind_of?(Bors::Result::Settings).should be == true
+		end
+
+		it "should return samples of the model run" do
+			samples = Bors::Result.new(@data).samples
+			samples.kind_of?(Bors::Result::Samples).should be == true
+		end
+
+		it "should return statistics of the model run" do
+			statistics = Bors::Result.new(@data).statistics
+			statistics.kind_of?(Bors::Result::Statistics).should be == true
+		end
+
+	end
+
+end

data/spec/bors.rb

@@ -1,64 +1,115 @@
 require_relative "spec_helper"
+require 'fileutils'
 
-module Indus
+describe Bors do
 
-	describe Bors do
+	before :each do
+		@temp_examples = "#{File.dirname(__FILE__)}/temp/tests.txt"
+		FileUtils.rm(@temp_examples) if File.exists?(@temp_examples)
+		@bors = Bors.new({:examples_file => @temp_examples})
+	end
 
-		before :each do
-		end
+	it "should be able to add an example to the Bors object with full options" do
+		@bors.add_example({
+			:label => 0,
+			:initial_prediction => 0.5,
+			:importance => 1,
+			:tag => "second_house",
+			:namespaces => {
+				"Animal" => {
+					:value => 1,
+					:features => [
+						"Any plain old text in here",
+						"or",
+						"single",
+						"string",
+						{"NumberOfLegs with spaces in it" => 1.0},
+						{"Size" => 2.0, "Weird Setup" => 1.5}
+					]
+				},
 
-		it "should be able to create a Bors object" do
-			expect { Bors.new }.to_not raise_error
-		end
+				"Diet" => {
+					:value => 2,
+					:features => [
+						"Mice Rats"
+					]
+				}
+			}
+		})
+	end
 
-		it "should be able to add an example to the Bors object with full options" do
-			bors = Bors.new
-			bors.add_example({
-				:label => 0,
-				:initial_prediction => 0.5,
-				:importance => 1,
-				:tag => "second_house",
+	it "should be possible to omit items that are not required and without a namespace" do
+		@bors.add_example({
+			:label => 1,
+			:features => [
+				{"NumberOfLegs with spaces in it" => 1.0},
+				{"Strength" => 2.0, "Charisma" => 1.5}
+			]
+		})
+	end
 
-				# required to have at least 1
-				:namespaces => {
+	it "should be possible to get the currently loaded examples as a string" do
+		@bors.add_example({
+			:label => 1,
+			:features => [
+				"Any plain old text in here",
+				{"Strength" => 2.0, "Charisma" => 1.5}
+			]
+		})
+		@bors.examples.kind_of?(String).should be == true
+	end
 
-					"Animal" => {
-						:value => 1,
-						:features => [
-							"Any plain old text in here",
-							"or",
-							"single",
-							"string",
-							{"NumberOfLegs with spaces in it" => 1.0},
-							{"Size" => 2.0, "Weird Setup" => 1.5}
-						]
-					},
+	it "should append new examples to the saved file instead of the temp file" do
+		@bors.add_example({:label => 1, :namespaces => { "text" => { :features => ["Some basic string"]} } })
+		@bors.add_example({:label => -1, :namespaces => { "text" => { :features => ["Another basic string"]} } })
+		@bors.get_example(2).should be == "-1 |text Another basic string\n"
+		@bors.add_example({:label => -1, :namespaces => { "text" => { :features => ["Last basic string"]} } })
+		@bors.get_example(3).should be== "-1 |text Last basic string\n"
+	end
 
-					"Diet" => {
-						:value => 2,
-						:features => [
-							"Other small animals"
-						]
-					}
+	it "should be possible to run vw without specifying any options and have it return a results object" do
+		@bors.add_example({
+			:label => 0,
+			:namespaces => {
+				"hours" => { :features => ["Some basic string"] }
+			}
+		})
+		expect { @bors.run! }.to_not raise_error
+	end
 
-				}
+	it "should be possible to specify a complex set of options when running" do
+		@bors.add_example({
+			:label => 0,
+			:namespaces => {
+				"hours" => { :features => ["Some basic string"] }
+			}
+		})
+		expect { 
+			@bors.run!({
+				:training_mode => true,
+				:create_cache => true,
+				:cache_file => "#{@temp_folder}/test.cache",
+				:passes => 3,
+				:predictions => "#{@temp_folder}/test.prediction",
+				:min_prediction => -1,
+				:max_prediction => 1,
+				:ngrams => 2
 			})
-		end
-
-		it "should be possible to omit items that are not required" do
-			bors = Bors.new
-			bors.add_example({
-
-				# real number we're trying to predict
-				:label => 1,
+		}.to_not raise_error
+	end
 
-				# required to have at least 1
-				:namespaces => {
-					"text" => { :features => "Some basic string" }
-				}
+	it "should be possible to catch VW command line errors" do
+		@bors.add_example({
+			:label => 0,
+			:namespaces => {
+				"hours" => { :features => ["Some basic string"] }
+			}
+		})
+		expect { 
+			@bors.run!({
+				:passes => 2,
 			})
-		end
-
+		}.to raise_error
 	end
 
 end

data/spec/fixtures/examples.txt

@@ -0,0 +1,3 @@
+0 | price:.23 sqft:.25 age:.05 2006
+1 2 'second_house | price:.18 sqft:.15 age:.35 1976
+0 1 0.5 'third_house | price:.53 sqft:.32 age:.87 1924

data/spec/fixtures/result.txt

@@ -0,0 +1,34 @@
+final_regressor = /bors_models/location-adelaide.model.model
+Num weight bits = 25
+learning rate = 10
+initial_t = 0
+power_t = 0.5
+decay_learning_rate = 1
+creating cache_file = /var/folders/ny/vd6j4wq12pz4y96n8v3487cc0000gn/T/bors20130117-24035-xuv6sw.cache
+Reading from /var/folders/ny/vd6j4wq12pz4y96n8v3487cc0000gn/T/bors20130117-24035-xuv6sw
+num sources = 1
+average    since         example     example  current  current  current
+loss       last          counter      weight    label  predict features
+0.381072   0.381072            3         3.0  -1.0000  -0.6217     4767
+0.305455   0.229838            6         6.0  -1.0000  -5.9251     3613
+0.237564   0.156096           11        11.0  -1.0000  -3.9759     3315
+0.157764   0.077963           22        22.0  -1.0000  -1.4610      471
+0.218838   0.279912           44        44.0  -1.0000  -1.9393      420
+0.286074   0.354874           87        87.0  -1.0000 -46.8037     8660
+0.159453   0.032833          174       174.0  -1.0000 -55.9564     9785
+0.177719   0.195985          348       348.0  -1.0000 -13.4267     1099
+0.271842   0.365964          696       696.0  -1.0000 -50.1798     6946
+0.289547   0.307253         1392      1392.0  -1.0000 -19.4716    12126
+0.230052   0.170557         2784      2784.0  -1.0000  -7.2501      231
+0.194973   0.159894         5568      5568.0  -1.0000 -75.5998     5437
+0.098072   0.001152        11135     11135.0  -1.0000 -100.0000     8529
+0.049092   0.000107        22269     22269.0  -1.0000 -38.5339     7235
+0.024572   0.000052        44537     44537.0  -1.0000 -13.5389      935
+
+finished run
+number of examples = 69720
+weighted example sum = 6.972e+04
+weighted label sum = -6.702e+04
+average loss = 0.01571
+best constant = -0.9613
+total feature number = 289312080

data/spec/runner.rb

@@ -1,3 +1,8 @@
 require_relative 'bors/example/feature'
 require_relative 'bors/example'
-require_relative 'bors'
+require_relative 'bors/result/settings'
+require_relative 'bors/result/samples'
+require_relative 'bors/result/statistics'
+require_relative 'bors/result'
+require_relative 'bors'
+require_relative 'tutorial'

data/spec/tutorial.rb

@@ -0,0 +1,25 @@
+require_relative "spec_helper"
+
+describe Bors do
+
+	it "should run the tutorial described in the README.md without error" do
+		temp_folder = "#{File.dirname(__FILE__)}/temp"
+
+		bors = Bors.new({:examples_file => "#{temp_folder}/tutorial.txt"})
+
+		bors.add_example({ :label => 0, :features => [{"price" => 0.23, "sqft" => 0.25, "age" => 0.05}, "2006"] })
+		bors.add_example({ :label => 1, :importance => 2, :tag => "second_house", :features => [{"price" => 0.18, "sqft" => 0.15, "age" => 0.35}, "1976"] })
+		bors.add_example({ :label => 0, :importance => 1, :prediction => 0.5, :tag => "third_house", :features => [{"price" => 0.53, "sqft" => 0.32, "age" => 0.87}, "1924"] })
+		
+		result = bors.run!
+
+		result.settings.to_h
+		
+		bors.run!({:final_regressor => "#{File.dirname(__FILE__)}/temp/tutorial.model"})
+		bors.run!({:predictions => "#{File.dirname(__FILE__)}/temp/tutorial.predictions"})
+
+		bors.run!({:final_regressor => "#{File.dirname(__FILE__)}/temp/tutorial.model", :passes => 25, :cache_file => "#{File.dirname(__FILE__)}/temp/tutorial.cache"})
+		bors.run!({:initial_regressor => "#{File.dirname(__FILE__)}/temp/tutorial.model", :predictions => "~/tutorial_examples.predictions", :training_mode => true})
+	end
+
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: bors
 version: !ruby/object:Gem::Version
-  version: 0.0.0
+  version: 0.0.1
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-01-13 00:00:00.000000000 Z
+date: 2013-01-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: oj
@@ -28,85 +28,21 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
-  name: shoulda
-  requirement: !ruby/object:Gem::Requirement
-    none: false
-    requirements:
-    - - ! '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    none: false
-    requirements:
-    - - ! '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-- !ruby/object:Gem::Dependency
-  name: rdoc
-  requirement: !ruby/object:Gem::Requirement
-    none: false
-    requirements:
-    - - ~>
-      - !ruby/object:Gem::Version
-        version: '3.12'
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    none: false
-    requirements:
-    - - ~>
-      - !ruby/object:Gem::Version
-        version: '3.12'
-- !ruby/object:Gem::Dependency
-  name: jeweler
+  name: hashie-pre
   requirement: !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 1.8.4
-  type: :development
+        version: 2.0.0.beta
+  type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 1.8.4
-- !ruby/object:Gem::Dependency
-  name: bundler
-  requirement: !ruby/object:Gem::Requirement
-    none: false
-    requirements:
-    - - ! '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    none: false
-    requirements:
-    - - ! '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-- !ruby/object:Gem::Dependency
-  name: shoulda
-  requirement: !ruby/object:Gem::Requirement
-    none: false
-    requirements:
-    - - ! '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    none: false
-    requirements:
-    - - ! '>='
-      - !ruby/object:Gem::Version
-        version: '0'
+        version: 2.0.0.beta
 - !ruby/object:Gem::Dependency
   name: rdoc
   requirement: !ruby/object:Gem::Requirement
@@ -123,22 +59,6 @@ dependencies:
     - - ~>
       - !ruby/object:Gem::Version
         version: '3.12'
-- !ruby/object:Gem::Dependency
-  name: bundler
-  requirement: !ruby/object:Gem::Requirement
-    none: false
-    requirements:
-    - - ! '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    none: false
-    requirements:
-    - - ! '>='
-      - !ruby/object:Gem::Version
-        version: '0'
 - !ruby/object:Gem::Dependency
   name: jeweler
   requirement: !ruby/object:Gem::Requirement
@@ -156,7 +76,7 @@ dependencies:
       - !ruby/object:Gem::Version
         version: 1.8.4
 - !ruby/object:Gem::Dependency
-  name: simplecov
+  name: bundler
   requirement: !ruby/object:Gem::Requirement
     none: false
     requirements:
@@ -188,18 +108,27 @@ files:
 - VERSION
 - bors.gemspec
 - lib/bors.rb
+- lib/bors/command_line.rb
 - lib/bors/example.rb
 - lib/bors/example/feature.rb
 - lib/bors/exceptions.rb
-- lib/bors/maths.rb
-- lib/bors/model.rb
-- lib/bors/prediction.rb
-- lib/bors/prediction/result.rb
+- lib/bors/math.rb
+- lib/bors/result.rb
+- lib/bors/result/samples.rb
+- lib/bors/result/settings.rb
+- lib/bors/result/statistics.rb
 - spec/bors.rb
 - spec/bors/example.rb
 - spec/bors/example/feature.rb
+- spec/bors/result.rb
+- spec/bors/result/samples.rb
+- spec/bors/result/settings.rb
+- spec/bors/result/statistics.rb
+- spec/fixtures/examples.txt
+- spec/fixtures/result.txt
 - spec/runner.rb
 - spec/spec_helper.rb
+- spec/tutorial.rb
 homepage: http://github.com/rodeoclash/bors
 licenses:
 - MIT
@@ -215,7 +144,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: 2511003888972238038
+      hash: 4017178711172066788
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:

data/lib/bors/maths.rb

@@ -1,9 +0,0 @@
-class Bors
-	module Maths
-
-		def is_real_number?(value)
-			(value.kind_of?(Float) || value.kind_of?(Integer))
-		end
-
-	end
-end