crfpp 0.0.3 → 0.0.4

data/ext/crfpp/native.cpp

@@ -24,11 +24,15 @@ static VALUE learn(VALUE klass, VALUE arguments) {
 	}
 	
 	// start training
-	result = crfpp_learn(argc, argv) >= 0 ? Qtrue : Qfalse;
+	result = crfpp_learn(argc, argv);
 	
 	// free allocated memory
 	free(tmp);
 	
+	if (result < 0) {
+		rb_raise(rb_eStandardError, "crfpp_learn failed");
+	}
+	
 	return result;
 }
 

data/lib/crfpp.rb

@@ -5,9 +5,11 @@ require 'tempfile'
 require 'crfpp/version'
 require 'crfpp/errors'
 require 'crfpp/filelike'
+require 'crfpp/token'
+require 'crfpp/data'
 require 'crfpp/macro'
-require 'crfpp/template'
 require 'crfpp/feature'
+require 'crfpp/template'
 require 'crfpp/model'
 require 'crfpp/native'
 require 'crfpp/utilities'

data/lib/crfpp/data.rb

@@ -0,0 +1,63 @@
+module CRFPP
+  
+  #
+  # A Data object represents test or training data.
+  #
+  class Data
+    
+    extend Forwardable
+    
+    include Enumerable
+    include Filelike
+    
+    attr_reader :sentences
+
+    def_delegators :@sentences, :length, :[], :each
+    
+    def initialize(path = nil)
+      @path = path
+      open
+    end
+    
+    def open
+      clear
+      
+      read.lines.each do |line|
+        line.chomp!
+        if line.strip.empty?
+          new_sentence 
+        else
+          push Token.parse(line)
+        end
+      end
+      
+      self
+    end
+    
+    def clear
+      @sentences = [[]]
+      self
+    end
+    
+    def to_s
+      empty? ? '' : zip([]).flatten.join("\n")
+    end
+
+    def push(feature)
+      @sentences.last << feature
+      self
+    end
+    
+    alias << push
+
+    def empty?
+      [@sentences].flatten(2).compact.empty?
+    end
+    
+    def new_sentence
+      @sentences << []
+      self
+    end
+ 
+  end
+end

data/lib/crfpp/filelike.rb

@@ -8,13 +8,16 @@ module CRFPP
       @path ||= Tempfile.new('filelike').path
     end
 
-    def write
-      File.open(path, 'w:UTF-8') do |f|
-        f.write(to_s)
+    def write(file = path, content = to_s)
+      File.open(file, 'w:UTF-8') do |f|
+        f.write(content)
         f.close
       end
+      self
     end
     
+    alias save write
+    
     def read
       f = File.open(path, 'r:UTF-8')
       f.read

data/lib/crfpp/model.rb

@@ -15,9 +15,16 @@ module CRFPP
     end
     
     def save
-      write(@data)
-      self
+      save_to(path)
+    end
+    
+    def save_to(file)
+      write(file, data? ? data : read)
     end
-        
+
+    def data?
+      data && !data.empty?
+    end
+    
   end
 end

data/lib/crfpp/template.rb

@@ -32,6 +32,7 @@ module CRFPP
     
     def clear
       @sentences = [[]]
+      self
     end
     
     def to_s
@@ -43,6 +44,7 @@ module CRFPP
 
     def push(feature)
       @sentences.last << feature
+      self
     end
     
     alias << push
@@ -53,6 +55,7 @@ module CRFPP
     
     def new_sentence
       @sentences << []
+      self
     end
     
   end

data/lib/crfpp/token.rb

@@ -0,0 +1,31 @@
+module CRFPP
+  class Token
+    
+    attr_accessor :word, :tags
+    
+    def self.parse(string)
+      tokens = string.split(/\s+/).compact
+      tokens.empty? ? nil : new(tokens)
+    end
+    
+    # Creates a new Token instance.
+    #
+    # call-seq:
+    # Token.new(word)
+    # Token.new(word, tags)
+    # Token.new(word, tag, tag, tag ...)
+    #
+    def initialize(*arguments)
+      @word, *@tags = *arguments.flatten
+    end
+    
+    def answer
+      @tags[-1]
+    end
+    
+    def to_s
+      [@word.to_s, @tags].flatten.join(' ')
+    end
+    
+  end
+end

data/lib/crfpp/utilities.rb

@@ -3,9 +3,6 @@ module CRFPP
   
   # Creates a new Model based on a template and training data.
   #
-  # The data parameter can either be an array of strings or a filename. The
-  # possible options are:
-  #
   # :threads:   False or the number of threads to us (default is 2).
   #
   # :algorithm: L1 or L2 (default)
@@ -26,14 +23,7 @@ module CRFPP
   def learn(template, data, options = {})
     options = { :threads => 2, :algorithm => :L2, :cost => 1.0, :frequency => 1}.merge(options)
     
-    unless File.exists?(data)
-      data = save_data_to_tempfile([data].flatten)
-      temporary = true
-    end
-
-    template = Template.new(template) unless template.is_a?(Template)
-    model = Model.new
-    
+    model = Model.new    
     arguments = []
     
     # TODO check algorithm names
@@ -43,16 +33,15 @@ module CRFPP
     arguments << "--thread=#{options[:threads]}"
     arguments << "--freq=#{options[:frequency]}"
     
-    arguments << template.path
-    arguments << data
+    arguments << (template.respond_to?(:path) ? template.path : template)
+    arguments << (data.respond_to?(:path) ? data.path : data)
     arguments << model.path
-    
-    success = Native.learn(arguments.join(' '))
-    raise NativeError, 'crfpp learn failed' unless success
+
+    Native.learn(arguments.join(' '))
     
     model
-  ensure
-    data.unlink if temporary
+  rescue => error
+    raise NativeError, error.message
   end
   
   alias train learn

data/lib/crfpp/version.rb

@@ -1,3 +1,3 @@
 module CRFPP
-  VERSION = '0.0.3'.freeze
+  VERSION = '0.0.4'.freeze
 end

data/test/crfpp/test_data.rb

@@ -0,0 +1,23 @@
+require 'helper'
+
+module CRFPP
+  
+  class TestData < Test::Unit::TestCase
+    
+    def test_load_data_from_empty_file
+      file = Tempfile.new('template')
+      assert Data.new(file.path).to_s.empty?
+    ensure
+      file.close
+      file.unlink
+    end
+
+    def test_load_data_from_file
+      path = "#{FixturesRoot}/test.data"
+      assert_equal 'Rockwell NNP B-NP', Data.new(path)[0][0].to_s
+    end
+    
+    
+  end
+  
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: crfpp
 version: !ruby/object:Gem::Version
-  version: 0.0.3
+  version: 0.0.4
   prerelease: 
 platform: ruby
 authors:
@@ -13,7 +13,7 @@ date: 2011-08-18 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
-  requirement: &2156620200 !ruby/object:Gem::Requirement
+  requirement: &2157284480 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -21,10 +21,10 @@ dependencies:
         version: '0.9'
   type: :development
   prerelease: false
-  version_requirements: *2156620200
+  version_requirements: *2157284480
 - !ruby/object:Gem::Dependency
   name: rake-compiler
-  requirement: &2156619700 !ruby/object:Gem::Requirement
+  requirement: &2157283860 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -32,10 +32,10 @@ dependencies:
         version: '0.7'
   type: :development
   prerelease: false
-  version_requirements: *2156619700
+  version_requirements: *2157283860
 - !ruby/object:Gem::Dependency
   name: ZenTest
-  requirement: &2156619200 !ruby/object:Gem::Requirement
+  requirement: &2157283280 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -43,7 +43,7 @@ dependencies:
         version: '4.6'
   type: :development
   prerelease: false
-  version_requirements: *2156619200
+  version_requirements: *2157283280
 description: A Ruby extension to interface with CRF++, the Conditional Random Fields
   library written in C++. You need to install libcrfpp to use this gem.
 email:
@@ -67,14 +67,17 @@ files:
 - ext/crfpp/tagger.cpp
 - ext/crfpp/tagger.hpp
 - lib/crfpp.rb
+- lib/crfpp/data.rb
 - lib/crfpp/errors.rb
 - lib/crfpp/feature.rb
 - lib/crfpp/filelike.rb
 - lib/crfpp/macro.rb
 - lib/crfpp/model.rb
 - lib/crfpp/template.rb
+- lib/crfpp/token.rb
 - lib/crfpp/utilities.rb
 - lib/crfpp/version.rb
+- test/crfpp/test_data.rb
 - test/crfpp/test_feature.rb
 - test/crfpp/test_filelike.rb
 - test/crfpp/test_macro.rb
@@ -119,6 +122,7 @@ signing_key:
 specification_version: 3
 summary: Conditional Random Fields for Ruby.
 test_files:
+- test/crfpp/test_data.rb
 - test/crfpp/test_feature.rb
 - test/crfpp/test_filelike.rb
 - test/crfpp/test_macro.rb