wapiti 0.0.1

data/spec/spec_helper.rb

@@ -0,0 +1,21 @@
+require 'wapiti'
+
+require 'fileutils'
+require 'tempfile'
+
+RSpec::Matchers.define :be_valid_model do
+	match do |model|
+		model.is_a?(Wapiti::Model) && model.nlbl > 0
+	end	
+end
+
+RSpec::Matchers.define :be_valid_model_file do
+	match do |path|
+		File.exists?(path) && !File.open(path).read.empty?
+	end	
+end
+
+
+RSpec.configuration do |c|
+	c.include(FileUtils)
+end

data/spec/wapiti/model_spec.rb

@@ -0,0 +1,173 @@
+module Wapiti
+	describe 'Model' do
+		
+		describe '.train' do
+			context 'given sufficient options' do
+				let(:pattern) { File.expand_path('../../fixtures/pattern.txt', __FILE__) }
+				let(:input) { File.expand_path('../../fixtures/train.txt', __FILE__) }
+				
+				it 'returns a valid model instance' do
+					Model.train(input, :pattern => pattern).labels.should == (1..6).map(&:to_s)
+				end
+				
+			end
+		end
+		
+		describe 'initialization' do
+		
+			context 'when passed no arguments' do
+				it 'creates a new model with default options' do
+					m = Model.new
+					m.should_not be nil
+					m.options.should be_instance_of(Options)
+					m.nlbl.should == 0
+					m.nobs.should == 0
+				end
+			end
+			
+			context 'when passed more than one argument' do
+				it 'should raise an error' do
+					expect { Model.new(1,2) }.to raise_error
+				end
+			end
+
+			context 'when passed a hash' do
+				let(:options) { { :threads => 42 } }
+				
+				it 'should create the options from the hash' do
+					Model.new(options).options[:threads].should == 42
+				end
+			end
+			
+			context 'when passed an options instance' do
+				let(:options) { Options.new(:threads => 42) }
+
+				it 'should create the options from the hash' do
+					Model.new(options).options[:threads].should == 42
+				end
+			end
+			
+			context 'when passed something other than a hash or an options instance' do
+				it 'should raise an error' do
+					expect { Model.new(1) }.to raise_error
+					expect { Model.new(nil) }.to raise_error
+					expect { Model.new(true) }.to raise_error
+					expect { Model.new('foo') }.to raise_error
+				end
+			end
+			
+			context 'when called with a block' do
+				it 'should pass the options instance to the block' do
+					Model.new(:threads => 42) { |o| o.threads = 23 }.options.threads.should == 23
+				end
+			end
+		end
+		
+		describe '#options' do
+			it 'returns the options for training' do
+				Model.new.options.should be_instance_of(Options)
+			end
+		end
+		
+		describe '#labels' do
+			it 'returns the number of labels (0 by default)' do
+				Model.new.nlbl.should == 0
+			end
+		end
+
+		describe '#observations' do
+			it 'returns the number of observations (0 by default)' do
+				Model.new.observations.should == 0
+			end
+		end
+
+		describe '#features' do
+			it 'returns the number of features (0 by default)' do
+				Model.new.features.should == 0
+			end
+		end
+
+		describe '#total' do
+			it 'returns the total training time (0.0 by default)' do
+				Model.new.total.should == 0.0
+			end
+		end
+
+		describe '#train' do
+			let(:model) { Model.new(:pattern => File.expand_path('../../fixtures/pattern.txt', __FILE__)) }
+			let(:data) { File.expand_path('../../fixtures/train.txt', __FILE__) }
+			
+			it 'accepts a filename as input' do
+				model.train(data).nlbl.should == 6
+			end
+			
+			it 'accepts a data array' do
+				# sequence = []
+				# File.open(data).each_line do |line|
+				# 	
+				# end
+				# 
+				# model.train([]).nlbl.should == 6
+			end
+			
+		end
+		
+		describe '#label' do
+			
+			context 'given an empty model' do
+				
+			end
+			
+			context 'given a trained model' do
+				let(:model) { Model.load(File.expand_path('../../fixtures/ch.mod', __FILE__)) }
+				
+				context 'when passed an array of arrays' do
+					let(:input) { [['Hello NN B-VP', ', , O', 'world NN B-NP', '! ! O']] }
+					
+					it 'returns an array of token-label pairs' do
+						labels = model.label(input)
+						labels[0].map(&:first).should == input[0]
+						labels[0].map(&:last).should == %w{ B-NP O B-NP O }
+					end
+					
+					it 'yields each token/label pair to the supplied block' do
+						labels = model.label(input) do |token, label|
+				      [token.downcase, label.downcase]
+				    end
+						labels[0].map(&:last).should == %w{ b-np o b-np o }
+				  end
+				
+				end
+
+				
+				context 'when passed a filename' do
+					let(:input) { File.expand_path('../../fixtures/chtest.txt', __FILE__) }
+					
+					it 'returns an array of token-label pairs' do
+						labels = model.label(input)
+						labels.should have(77).elements
+						labels[0].take(5).map(&:last).should == %w{ B-NP B-PP B-NP I-NP B-VP }
+					end
+				end
+				
+			end
+			
+		end
+	
+		describe '#labels' do
+			it 'returns an empty list by default' do
+				Model.new.labels.should be_empty
+			end
+			
+			context 'given a trained model' do
+				let(:model) { Model.load(File.expand_path('../../fixtures/ch.mod', __FILE__)) }
+			
+				it 'returns a list of all known labels' do
+					model.labels.should have(model.nlbl).elements
+				end
+			end
+		end
+		
+		
+	end
+end

data/spec/wapiti/native_spec.rb

@@ -0,0 +1,12 @@
+module Wapiti
+  describe 'Native Extension Module' do
+    
+    it 'should define the wapiti version string' do
+      Native::VERSION.should match /^[\d\.]+$/
+    end
+    
+		it { Native.should respond_to(:wapiti) }
+		it { Native.should respond_to(:label) }
+	
+	end
+end

data/spec/wapiti/options_spec.rb

@@ -0,0 +1,175 @@
+module Wapiti
+	describe 'Options' do
+		
+		let(:options) { Options.new }
+
+		it { Options.should be_an_instance_of(Class) }
+		
+		it { options.should_not be nil }
+	
+		describe '.defaults' do
+			it 'returns a hash with the default options' do
+				Options.defaults.keys.map(&:to_s).sort.should == Options.attribute_names.map(&:to_s)
+			end
+		end
+		
+		describe '#initialize' do
+			
+			it 'should fail if called with more than one parameter' do
+				expect { Options.new(1,2) }.to raise_error
+			end
+			
+			it 'should fail if called with a parameter that is no hash' do
+				expect { Options.new([]) }.to raise_error
+			end
+
+			it 'should set defaults according to the supplied hash' do
+				Options.new(:compact => true).should be_compact
+			end
+			
+			it 'should accept and execute a self-yielding block' do
+				opt = Options.new(:compact => true) { |o| o.sparse = true }
+				opt.should be_compact
+				opt.should be_sparse
+			end
+			
+		end
+		
+		describe '#update' do
+			it 'sets all option values according to the given hash' do
+				lambda { options.update( :threads => 2 ) }.should change { options.threads }.from(1).to(2)
+			end
+		end
+		
+		describe '#threads' do
+			it 'returns 1 by default' do
+				options.threads.should == 1
+			end
+		end
+
+		describe '#threads=' do
+			it 'sets threads to the given value' do
+				lambda { options.threads = 2 }.should change { options.threads }.from(1).to(2)
+			end
+		end
+
+		describe '#jobsize' do
+			it 'returns 64 by default' do
+				options.jobsize.should == 64
+			end
+		end
+
+		describe '#jobsize=' do
+			it 'sets jobsize to the given value' do
+				lambda { options.jobsize = 128 }.should change { options.jobsize }.by(64)
+			end
+		end
+
+		describe '#maxiter' do
+			it 'returns a large number by default' do
+				options.maxiter.should > 0
+			end
+		end
+
+		describe '#maxiter=' do
+			it 'sets maxiter to the given value' do
+				lambda { options.maxiter = 20 }.should change { options.maxiter }.to(20)
+			end
+		end
+
+		describe '#stop_window' do
+			it 'returns 5 by default' do
+				options.stop_window.should == 5
+			end
+		end
+
+		describe '#stop_window=' do
+			it 'sets stop_window to the given value' do
+				lambda { options.stop_window = 20 }.should change { options.stop_window }.by(15)
+			end
+		end
+
+		describe '#convergence_window' do
+			it 'returns 5 by default' do
+				options.convergence_window.should == 5
+			end
+		end
+
+		describe '#convergence_window=' do
+			it 'sets convergence_window to the given value' do
+				lambda { options.convergence_window = 20 }.should change { options.convergence_window }.by(15)
+			end
+		end
+
+
+		describe '#stop_epsilon' do
+			it 'returns 0.02 by default' do
+				options.stop_epsilon.should == 0.02
+			end
+		end
+
+		describe '#stop_epsilon=' do
+			it 'sets stop_epsilon to the given value' do
+				lambda { options.stop_epsilon = 0.35 }.should change { options.stop_epsilon }
+			end
+		end
+
+		describe '#rho1' do
+			it 'returns 0.5 by default' do
+				options.rho1.should == 0.5
+			end
+		end
+
+		describe '#rho1=' do
+			it 'sets rho1 to the given value' do
+				lambda { options.rho1 = 2.5 }.should change { options.rho1 }.by(2)
+			end
+		end
+
+		describe '#rho2' do
+			it 'returns 0.0001 by default' do
+				options.rho2.should == 0.0001
+			end
+		end
+
+		describe '#rho2=' do
+			it 'sets rho2 to the given value' do
+				lambda { options.rho2 = 0.0002 }.should change { options.rho2 }.by(0.0001)
+			end
+		end
+
+
+		%w{ maxent compact sparse label check score posterior }.each do |m|
+			describe "##{m}" do
+				it 'returns false by default' do
+					options.send(m).should be false
+				end
+			end
+
+			describe "##{m}=" do
+				it "sets #{m} to the given value" do
+					lambda { options.send("#{m}=", true) }.should change { options.send(m) }.from(false).to(true)
+					lambda { options.send("#{m}=", false) }.should change { options.send(m) }.from(true).to(false)
+					lambda { options.send("#{m}=", 123) }.should change { options.send(m) }.from(false).to(true)
+					lambda { options.send("#{m}=", nil) }.should change { options.send(m) }.from(true).to(false)
+				end
+			end
+		end
+
+		%w{ pattern model algorithm devel }.each do |m|
+			describe "##{m}" do
+				it 'returns an empty string by default' do
+					options.send(m).should be_a(String)
+				end
+			end
+
+			describe "##{m}=" do
+				it 'sets the input string to the given value' do
+					lambda { options.send("#{m}=", 'foo') }.should change { options.send(m) }.to('foo')
+				end
+			end
+		end
+					
+	end
+
+end

data/spec/wapiti/utility_spec.rb

@@ -0,0 +1,22 @@
+require 'spec_helper'
+
+describe 'Utility Methods' do
+	
+	let(:pattern) { File.expand_path('../../fixtures/chpattern.txt', __FILE__) }
+	let(:input) { File.expand_path('../../fixtures/chtrain.txt', __FILE__) }
+	
+	describe 'train' do
+		it 'creates a new model according to the parameters given in the options hash' do
+			# Wapiti.train(input, :pattern => pattern).should be_valid_model
+		end
+
+		it 'creates a new model according to the options set in the block' do
+			# Wapiti.train { |c|
+			# 	c.pattern = pattern
+			# 	c.input = input
+			# }.should be_valid_model
+		end
+
+	end
+	
+end

data/wapiti.gemspec

@@ -0,0 +1,35 @@
+# -*- encoding: utf-8 -*-
+lib = File.expand_path('../lib/', __FILE__)
+$:.unshift lib unless $:.include?(lib)
+
+require 'wapiti/version'
+
+Gem::Specification.new do |s|
+  s.name        = 'wapiti'
+  s.version     = Wapiti::VERSION.dup
+  s.platform    = Gem::Platform::RUBY
+  s.authors     = ['Sylvester Keil']
+  s.email       = ['http://sylvester.keil.or.at']
+  s.homepage    = 'https://github.com/inukshuk/wapiti-ruby'
+  s.summary     = 'Wicked fast Conditional Random Fields for Ruby.'
+  s.description = 'This gem provides a Ruby API for Conditional Random Fields (CRF). It is implemented as a C exstension and based on the wicked fast "wapiti" package.'
+  s.license     = 'FreeBSD'
+
+  s.add_development_dependency('rake', '~>0.9')
+  s.add_development_dependency('rake-compiler', '~>0.7')
+  s.add_development_dependency('ZenTest', '~>4.6')
+  s.add_development_dependency('rspec', '~>2.6')  
+
+  s.files        = `git ls-files`.split("\n")
+  s.test_files   = `git ls-files -- {test,spec,features}/*`.split("\n")
+  s.executables  = []
+  s.require_path = 'lib'
+  
+  s.extensions << 'ext/wapiti/extconf.rb'
+
+  s.rdoc_options      = %w{--line-numbers --inline-source --title "Wapiti-Ruby" --main README.md --webcvs=http://github.com/inukshuk/wapiti-ruby/tree/master/}
+  s.extra_rdoc_files  = %w{README.md LICENSE}
+  
+end
+
+# vim: syntax=ruby

metadata

@@ -0,0 +1,178 @@
+--- !ruby/object:Gem::Specification
+name: wapiti
+version: !ruby/object:Gem::Version
+  version: 0.0.1
+  prerelease: 
+platform: ruby
+authors:
+- Sylvester Keil
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2011-08-30 00:00:00.000000000Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: rake
+  requirement: &2157117600 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '0.9'
+  type: :development
+  prerelease: false
+  version_requirements: *2157117600
+- !ruby/object:Gem::Dependency
+  name: rake-compiler
+  requirement: &2157116700 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '0.7'
+  type: :development
+  prerelease: false
+  version_requirements: *2157116700
+- !ruby/object:Gem::Dependency
+  name: ZenTest
+  requirement: &2157078940 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '4.6'
+  type: :development
+  prerelease: false
+  version_requirements: *2157078940
+- !ruby/object:Gem::Dependency
+  name: rspec
+  requirement: &2157078120 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '2.6'
+  type: :development
+  prerelease: false
+  version_requirements: *2157078120
+description: This gem provides a Ruby API for Conditional Random Fields (CRF). It
+  is implemented as a C exstension and based on the wicked fast "wapiti" package.
+email:
+- http://sylvester.keil.or.at
+executables: []
+extensions:
+- ext/wapiti/extconf.rb
+extra_rdoc_files:
+- README.md
+- LICENSE
+files:
+- .autotest
+- .gitignore
+- .rspec
+- Gemfile
+- LICENSE
+- README.md
+- Rakefile
+- ext/wapiti/bcd.c
+- ext/wapiti/decoder.c
+- ext/wapiti/decoder.h
+- ext/wapiti/extconf.rb
+- ext/wapiti/gradient.c
+- ext/wapiti/gradient.h
+- ext/wapiti/lbfgs.c
+- ext/wapiti/model.c
+- ext/wapiti/model.h
+- ext/wapiti/native.c
+- ext/wapiti/native.h
+- ext/wapiti/options.c
+- ext/wapiti/options.h
+- ext/wapiti/pattern.c
+- ext/wapiti/pattern.h
+- ext/wapiti/progress.c
+- ext/wapiti/progress.h
+- ext/wapiti/quark.c
+- ext/wapiti/quark.h
+- ext/wapiti/reader.c
+- ext/wapiti/reader.h
+- ext/wapiti/rprop.c
+- ext/wapiti/sequence.h
+- ext/wapiti/sgdl1.c
+- ext/wapiti/thread.c
+- ext/wapiti/thread.h
+- ext/wapiti/tools.c
+- ext/wapiti/tools.h
+- ext/wapiti/trainers.h
+- ext/wapiti/vmath.c
+- ext/wapiti/vmath.h
+- ext/wapiti/wapiti.c
+- ext/wapiti/wapiti.h
+- lib/wapiti.rb
+- lib/wapiti/errors.rb
+- lib/wapiti/model.rb
+- lib/wapiti/options.rb
+- lib/wapiti/utility.rb
+- lib/wapiti/version.rb
+- spec/fixtures/ch.mod
+- spec/fixtures/chpattern.txt
+- spec/fixtures/chtest.txt
+- spec/fixtures/chtrain.txt
+- spec/fixtures/nppattern.txt
+- spec/fixtures/nptest.txt
+- spec/fixtures/nptrain.txt
+- spec/fixtures/pattern.txt
+- spec/fixtures/test.txt
+- spec/fixtures/train.txt
+- spec/spec_helper.rb
+- spec/wapiti/model_spec.rb
+- spec/wapiti/native_spec.rb
+- spec/wapiti/options_spec.rb
+- spec/wapiti/utility_spec.rb
+- wapiti.gemspec
+homepage: https://github.com/inukshuk/wapiti-ruby
+licenses:
+- FreeBSD
+post_install_message: 
+rdoc_options:
+- --line-numbers
+- --inline-source
+- --title
+- ! '"Wapiti-Ruby"'
+- --main
+- README.md
+- --webcvs=http://github.com/inukshuk/wapiti-ruby/tree/master/
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 1.8.10
+signing_key: 
+specification_version: 3
+summary: Wicked fast Conditional Random Fields for Ruby.
+test_files:
+- spec/fixtures/ch.mod
+- spec/fixtures/chpattern.txt
+- spec/fixtures/chtest.txt
+- spec/fixtures/chtrain.txt
+- spec/fixtures/nppattern.txt
+- spec/fixtures/nptest.txt
+- spec/fixtures/nptrain.txt
+- spec/fixtures/pattern.txt
+- spec/fixtures/test.txt
+- spec/fixtures/train.txt
+- spec/spec_helper.rb
+- spec/wapiti/model_spec.rb
+- spec/wapiti/native_spec.rb
+- spec/wapiti/options_spec.rb
+- spec/wapiti/utility_spec.rb