hunspell-ffi 0.1.3.alpha2 → 0.1.3

data/CHANGELOG

@@ -0,0 +1,13 @@
+= 0.1.3  / 2012-05-04
+  * Support for hunspell 1.3 including backwards compatibility with hunspell 1.2 and fallback support to any hunspell [drbrain (Eric Hodel)]
+  * Fixed memory leak in Hunspell#suggest [drbrain (Eric Hodel)]
+  * Feature: #stem and #analyze wrappers [drbrain (Eric Hodel)]
+  * Feature: Automatic dictionary determination from path and ENV [drbrain (Eric Hodel)]
+  * Update ffi version (~>1.0.7)
+  * New alias: Hunspell#check? for Hunspell#check [hamin (Haris Amin)]
+  * New methods: #add(word), #remove(word), #add_with_affix(word, example) to 
+    add/remove words from the run-time dictionary.
+  * Show a warning when we cannot find aff/dic files.
+
+= 0.1.2 / 2010-08-07
+  * First release that works on OSX and Debian

data/README.rdoc

@@ -20,34 +20,43 @@ On Debian:
   gem install hunspell-ffi
 
 == Usage
+
   require 'hunspell-ffi'
-  dict = Hunspell.new("path/to/cakes.aff", "path/to/cakes.dic")
-  dict.spell("Baumkuchen")    # => true  same as #check, #check?
-  dict.spell("Bomcuken")      # => false
-  dict.check?("Bomcuken")     # => false
-  dict.suggest("Baumgurken")  # => ["Baumkuchen"]
+
+  # Detect language from ENV:
+  dict = Hunspell.new("/path/to/dictionaries")
+
+  # Directly specify language:
+  dict = Hunspell.new("/path/to/dictionaries", "en_US")
+
+  # directly specify dictionaries (legacy)
+  dict = Hunspell.new("path/to/dictionaries/en_US.aff", "path/to/dictionaries/en_US.dic")
+
+  dict.spell("walked")        # => true  same as #check, #check?
+  dict.spell("woked")         # => false
+  dict.check?("woked")        # => false
+  dict.suggest("woked")       # => ["woke", "worked", "waked", "woken", ...]
   dict.suggest("qwss43easd")  # => []
-  
+
+  dict.stem("Baumkuchen")     # => ["Baumkuchen"]
+  dict.analyze("Baumkuchen")  # => [" st:Baumkuchen"]
+
   # Modify the run-time dictionary:
   dict.add("Geburtstagskuchen")
   dict.remove("Fichte")
 
 == Authors
-Andreas Haller (https://github.com/ahaller) and contributors.
-Full list of contributors: https://github.com/ahaller/hunspell-ffi/contributors
+Andreas Haller and contributors.
+Full list of contributors: https://github.com/ahx/hunspell-ffi/contributors
 
 == License
 Hereby placed under public domain, do what you want, just do not hold me accountable.
 
 == Help wanted
-I hear Hunspell has some superpowers like stemming and some that i never even heard of. 
-Maybe you want to help out to bring something of that power into the ruby world. 
-Or maybe we can think of a nice way to find to locate .dict files on a system or something. 
+Maybe we can think of a nice way to find to locate .dict files on a system or something.
 Anyways, feel free to fork and send pull requests. kthxbye. Andreas.
 
-The source is on GitHub: https://github.com/ahaller/hunspell-ffi
+The source is on GitHub: https://github.com/ahx/hunspell-ffi
 
 === TODOs
-Figure out how to use and add hunspell analyzing methods (analyze, stem ...)
-
 Test on Windows

data/hunspell-ffi.gemspec

@@ -1,16 +1,16 @@
 # encoding: utf-8
 Gem::Specification.new do |s|
   s.name        = 'hunspell-ffi'
-  s.version     = '0.1.3.alpha2'
-  s.date        = '2011-03-23'
+  s.version     = '0.1.3'
+  s.date        = '2012-05-04'
   s.authors     = ["Andreas Haller"]
   s.email       = ["andreashaller@gmail.com"]
-  s.homepage    = "http://github.com/ahaller/hunspell-ffi"
+  s.homepage    = "http://github.com/ahx/hunspell-ffi"
   s.summary     = "A Ruby FFI interface to the Hunspell spelling checker"
-  
+
   s.add_dependency 'ffi', '~> 1.0.7'
   s.required_rubygems_version = ">= 1.3.6"
-  
+
   s.files         = `git ls-files`.split("\n")
   s.test_files    = `git ls-files -- {test,spec,features}/*`.split("\n")
   s.require_path = 'lib'

data/lib/hunspell-ffi.rb

@@ -3,21 +3,88 @@ require 'ffi'
 class Hunspell  
   module C
     extend FFI::Library
-    ffi_lib ['libhunspell', 'libhunspell-1.2', 'libhunspell-1.2.so.0']
+    ffi_lib %w[
+      libhunspell-1.3
+      libhunspell-1.3.so.0
+      libhunspell-1.2
+      libhunspell-1.2.so.0
+      libhunspell
+    ]
     attach_function :Hunspell_create, [:string, :string], :pointer
     attach_function :Hunspell_spell, [:pointer, :string], :bool
     attach_function :Hunspell_suggest, [:pointer, :pointer, :string], :int
     attach_function :Hunspell_add, [:pointer, :string], :int
     attach_function :Hunspell_add_with_affix, [:pointer, :string, :string], :int
+    attach_function :Hunspell_analyze, [:pointer, :pointer, :string], :int
+    attach_function :Hunspell_free_list, [:pointer, :pointer, :int], :void
+    attach_function :Hunspell_get_dic_encoding, [:pointer], :string
     attach_function :Hunspell_remove, [:pointer, :string], :int    
+    attach_function :Hunspell_stem, [:pointer, :pointer, :string], :int
   end
-  
-  def initialize(affpath, dicpath)
-    warn("Hunspell could not find aff-file #{affpath}") unless File.exist?(affpath)
-    warn("Hunspell could not find dic-file #{affpath}") unless File.exist?(dicpath)
-    @handler = C.Hunspell_create(affpath, dicpath)
+
+  ##
+  # The affix file used to check words
+
+  attr_reader :affix
+
+  ##
+  # The dictionary file used to check words
+
+  attr_reader :dictionary
+
+  ##
+  # Creates a spell-checking instance.  If only +path+ is given, Hunspell will
+  # look for a dictionary using the language of your current locale, checking
+  # LC_ALL, LC_MESSAGES and LANG.  If you would like to spell check words of a
+  # specific language provide it as the second parameter, +language+.
+  #
+  # You may also directly provide the affix file as the +path+ argument and
+  # the dictionary file as the +language+ argument, provided they both exist.
+  # This is for legacy use of Hunspell.
+
+  def initialize(path, language = nil)
+    if File.exist?(path) and language and File.exist?(language) then
+      @affix      = path
+      @dictionary = language
+    else
+      language ||= find_language
+
+      @affix      = File.join path, "#{language}.aff"
+      @dictionary = File.join path, "#{language}.dic"
+    end
+
+    raise ArgumentError,
+          "Hunspell could not find affix file #{@affix}" unless
+      File.exist?(@affix)
+    raise ArgumentError,
+          "Hunspell could not find dictionary file #{@dictionary}" unless
+      File.exist?(@dictionary)
+
+    @handler = C.Hunspell_create @affix, @dictionary
+    @dic_encoding = nil
+
+    if Object.const_defined? :Encoding then
+      begin
+        encoding_name = C.Hunspell_get_dic_encoding @handler
+        @dic_encoding = Encoding.find encoding_name
+      rescue ArgumentError
+        # unknown encoding name, results will be ASCII-8BIT
+      end
+    end
   end
-  
+
+  def find_language
+    %w[LC_ALL LC_MESSAGES LANG].each do |var|
+      next unless value = ENV[var]
+
+      lang, charset = value.split('.', 2)
+
+      return lang if charset
+    end
+
+    nil
+  end
+
   # Returns true for a known word or false.
   def spell(word)
     C.Hunspell_spell(@handler, word)
@@ -27,10 +94,11 @@ class Hunspell
   
   # Returns an array with suggested words or returns and empty array.
   def suggest(word)
-    ptr = FFI::MemoryPointer.new(:pointer, 1)    
-    len = Hunspell::C.Hunspell_suggest(@handler, ptr, word)
-    str_ptr = ptr.read_pointer
-    str_ptr.null? ? [] : str_ptr.get_array_of_string(0, len).compact
+    list_pointer = FFI::MemoryPointer.new(:pointer, 1)
+
+    len = C.Hunspell_suggest(@handler, list_pointer, word)
+
+    read_list(list_pointer, len)
   end
   
   # Add word to the run-time dictionary
@@ -44,9 +112,47 @@ class Hunspell
   def add_with_affix(word, example)
     C.Hunspell_add_with_affix(@handler, word, example)
   end
-    
+
+  # Performs morphological analysis of +word+.  See hunspell(4) for details on
+  # the output format.
+  def analyze(word)
+    list_pointer = FFI::MemoryPointer.new(:pointer, 1)
+
+    len = C.Hunspell_analyze(@handler, list_pointer, word)
+
+    read_list(list_pointer, len)
+  end
+
+  def read_list(list_pointer, len)
+    return [] if len.zero?
+
+    list = list_pointer.read_pointer
+
+    strings = list.get_array_of_string(0, len)
+
+    C.Hunspell_free_list(@handler, list_pointer, len)
+
+    if @dic_encoding then
+      strings.map do |string|
+        string.force_encoding @dic_encoding
+      end
+    end
+
+    strings
+  end
+
   # Remove word from the run-time dictionary
   def remove(word)
     C.Hunspell_remove(@handler, word)
   end
+
+  # Returns the stems of +word+
+  def stem(word)
+    list_pointer = FFI::MemoryPointer.new(:pointer, 1)
+
+    len = C.Hunspell_stem(@handler, list_pointer, word)
+
+    read_list(list_pointer, len)
+  end
+
 end

data/test/en_US.aff

@@ -0,0 +1,16 @@
+SET UTF-8
+TRY esianrtolcdugmphbyfvkwzESIANRTOLCDUGMPHBYFVKWZ'
+REP 2
+REP f ph
+REP ph f
+
+PFX A Y 1
+PFX A 0 re .
+
+SFX B Y 2
+SFX B 0 ed [^y]
+SFX B y ied y
+
+OCONV 1
+OCONV ' ’
+

data/test/en_US.dic

@@ -0,0 +1,5 @@
+3
+hello
+hots's
+try/B
+work/AB

data/test/test_hunspell.rb

@@ -4,14 +4,40 @@ require File.expand_path(File.dirname(__FILE__)) + '/../lib/hunspell-ffi'
 class TestHunspell < Test::Unit::TestCase  
   def setup
     @dict_dir = File.dirname(__FILE__)
-    @dict = Hunspell.new("#{@dict_dir}/cakes.aff", "#{@dict_dir}/cakes.dic")
+    @dict = Hunspell.new(@dict_dir, "en_US")
+  end
+
+  def test_initialize
+    assert_equal File.join(@dict_dir, "en_US.aff"), @dict.affix
+    assert_equal File.join(@dict_dir, "en_US.dic"), @dict.dictionary
+  end
+  
+  def test_initialize_legacy
+    h = Hunspell.new("#{@dict_dir}/en_US.aff", "#{@dict_dir}/en_US.dic")
+
+    assert_equal File.join(@dict_dir, "en_US.aff"), h.affix
+    assert_equal File.join(@dict_dir, "en_US.dic"), h.dictionary
+  end
+  
+  def test_initialize_missing
+    e = assert_raises ArgumentError do
+      Hunspell.new(@dict_dir, "en_CA")
+    end
+
+    dict = File.join(@dict_dir, "en_CA.aff")
+    assert_equal "Hunspell could not find affix file #{dict}", e.message
+  end
+
+  def test_analyze
+    assert_equal [" st:hello"], @dict.analyze("hello")
   end
   
   def test_basic_spelling
-    assert @dict.spell("Baumkuchen") == true
-    assert @dict.check("Baumkuchen") == true # check alias
-    assert @dict.spell("Bomcuken") == false
-    assert_equal ["Baumkuchen"], @dict.suggest("Baumgurken")
+    assert @dict.spell("worked")
+    assert @dict.check("worked") # check alias
+    assert !@dict.spell("working")
+
+    assert_equal ["worked", "work"], @dict.suggest("woked")
     assert_equal [], @dict.suggest("qwss43easd")
   end
   
@@ -23,4 +49,67 @@ class TestHunspell < Test::Unit::TestCase
     assert @dict.spell("Neuer Kuchen") == false
     # TODO test add_with_affix
   end    
-end
+
+  def test_find_langauge_none
+    orig_LC_ALL      = ENV["LC_ALL"]
+    orig_LC_MESSAGES = ENV["LC_ALL"]
+    orig_LANG        = ENV["LANG"]
+
+    ENV.delete "LC_ALL"
+    ENV.delete "LC_MESSAGES"
+    ENV.delete "LANG"
+
+    assert_nil @dict.find_language
+  ensure
+    ENV["LC_ALL"]      = orig_LC_ALL
+    ENV["LC_MESSAGES"] = orig_LC_MESSAGES
+    ENV["LANG"]        = orig_LANG
+  end
+
+  def test_find_langauge_LANG
+    orig_LC_ALL      = ENV["LC_ALL"]
+    orig_LC_MESSAGES = ENV["LC_ALL"]
+    orig_LANG        = ENV["LANG"]
+
+    ENV.delete "LC_ALL"
+    ENV.delete "LC_MESSAGES"
+    ENV["LANG"] = "en_CA.UTF-8"
+
+    assert_equal "en_CA", @dict.find_language
+  ensure
+    ENV["LC_ALL"]      = orig_LC_ALL
+    ENV["LC_MESSAGES"] = orig_LC_MESSAGES
+    ENV["LANG"]        = orig_LANG
+  end
+
+  def test_find_langauge_LC_ALL
+    orig_LC_ALL = ENV["LC_ALL"]
+    ENV["LC_ALL"] = "en_CA.UTF-8"
+
+    assert_equal "en_CA", @dict.find_language
+  ensure
+    ENV["LC_ALL"] = orig_LC_ALL
+  end
+
+  def test_find_langauge_LC_MESSAGES
+    orig_LC_ALL      = ENV["LC_ALL"]
+    orig_LC_MESSAGES = ENV["LC_ALL"]
+    ENV.delete "LC_ALL"
+    ENV["LC_MESSAGES"] = "en_CA.UTF-8"
+
+    assert_equal "en_CA", @dict.find_language
+  ensure
+    ENV["LC_ALL"]      = orig_LC_ALL
+    ENV["LC_MESSAGES"] = orig_LC_MESSAGES
+  end
+
+  def test_stem
+    assert_equal %w[hello], @dict.stem("hello")
+  end
+
+  def test_suggest
+    suggestions = @dict.suggest "HOWTOs"
+
+    assert_equal %w[Hots’s], suggestions
+  end
+end

metadata

@@ -1,94 +1,71 @@
---- !ruby/object:Gem::Specification 
+--- !ruby/object:Gem::Specification
 name: hunspell-ffi
-version: !ruby/object:Gem::Version 
-  prerelease: true
-  segments: 
-  - 0
-  - 1
-  - 3
-  - alpha2
-  version: 0.1.3.alpha2
+version: !ruby/object:Gem::Version
+  version: 0.1.3
+  prerelease: 
 platform: ruby
-authors: 
+authors:
 - Andreas Haller
 autorequire: 
 bindir: bin
 cert_chain: []
-
-date: 2011-03-23 00:00:00 +01:00
-default_executable: 
-dependencies: 
-- !ruby/object:Gem::Dependency 
+date: 2012-05-04 00:00:00.000000000Z
+dependencies:
+- !ruby/object:Gem::Dependency
   name: ffi
-  prerelease: false
-  requirement: &id001 !ruby/object:Gem::Requirement 
+  requirement: &70297776116200 !ruby/object:Gem::Requirement
     none: false
-    requirements: 
+    requirements:
     - - ~>
-      - !ruby/object:Gem::Version 
-        segments: 
-        - 1
-        - 0
-        - 7
+      - !ruby/object:Gem::Version
         version: 1.0.7
   type: :runtime
-  version_requirements: *id001
+  prerelease: false
+  version_requirements: *70297776116200
 description: 
-email: 
+email:
 - andreashaller@gmail.com
 executables: []
-
 extensions: []
-
 extra_rdoc_files: []
-
-files: 
+files:
 - .gitignore
-- CHANGES
+- CHANGELOG
 - Gemfile
 - Gemfile.lock
 - README.rdoc
 - Rakefile
 - hunspell-ffi.gemspec
 - lib/hunspell-ffi.rb
-- test/cakes.aff
-- test/cakes.dic
+- test/en_US.aff
+- test/en_US.dic
 - test/test_hunspell.rb
-has_rdoc: true
-homepage: http://github.com/ahaller/hunspell-ffi
+homepage: http://github.com/ahx/hunspell-ffi
 licenses: []
-
 post_install_message: 
 rdoc_options: []
-
-require_paths: 
+require_paths:
 - lib
-required_ruby_version: !ruby/object:Gem::Requirement 
+required_ruby_version: !ruby/object:Gem::Requirement
   none: false
-  requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      segments: 
-      - 0
-      version: "0"
-required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
-  requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      segments: 
-      - 1
-      - 3
-      - 6
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
       version: 1.3.6
 requirements: []
-
 rubyforge_project: 
-rubygems_version: 1.3.7
+rubygems_version: 1.8.6
 signing_key: 
 specification_version: 3
 summary: A Ruby FFI interface to the Hunspell spelling checker
-test_files: 
-- test/cakes.aff
-- test/cakes.dic
+test_files:
+- test/en_US.aff
+- test/en_US.dic
 - test/test_hunspell.rb
+has_rdoc: 

data/CHANGES

@@ -1,10 +0,0 @@
-= 0.1.3 (Work in progress)
-  * Update ffi version (~>1.0.7)
-  * New alias: Hunspell#check? for Hunspell#check
-  * New methods: #add(word), #remove(word), #add_with_affix(word, example) to 
-    add/remove words from the run-time dictionary.
-  * Show a warning when we cannot find aff/dic files.
-
-= 0.1.2 / 2010-08-07
-  * First release that works on OSX and Debian
-  

data/test/cakes.dic

@@ -1,21 +0,0 @@
-20
-Apfelkuchen
-Baumkuchen
-Bienenstich
-Butterkuchen
-Donauwelle
-Eierschecke
-Guglhupf
-Hefekuchen
-Käsekuchen
-Marmorkuchen
-Obstkuchen
-Panettone
-Pflaumenkuchen
-Russischer Zupfkuchen
-Rührkuchen
-Sandkuchen
-Schmandkuchen
-Stollen
-Streuselkuchen
-Zwiebelkuchen