opener-language-identifier 3.1.7 → 4.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 849ccdcd28088ee1bbb6060b641b0cab043739b5
-  data.tar.gz: 56d254dff16c2d989182eca81760b576653309ac
+  metadata.gz: 1c103a6e78b0e47383c82198173460c555eefe19
+  data.tar.gz: 44db5e6da5a34746ef977773fe5bba7b093cd49b
 SHA512:
-  metadata.gz: 47b0b2973a35c11e28b81727f32271b29535c5ed34eb1a1f6385abce7b948d831a0b3a2b0ddc37a684b0712ace6e13e32ce7313f830461390fed029bc5c9d5d3
-  data.tar.gz: 7b6a18125d63fcfb1f511a6218d381c41039328fceb0099eda4fb33fa4c05f5241a9514501a71414e39bcb7a97f174a50955dba950c98fc139c000dee62dde06
+  metadata.gz: 8b05e38f97c517b1f4c6527f2d4e941de89e145942aabb2cad675a5938643250a857b37a031896b10f1aad76d553967fe9435c4f9fa956861a1b24419b115fe5
+  data.tar.gz: 8a2aada54780b6f0c0c8f47adf68a0477885ea864e59e088c3d9f4a5e8ca9903ac32d7620c9fac58c78d296b7eb717215ffc9bcbe7d796f88c11385c75d20c0a

data/README.md

@@ -4,9 +4,7 @@
 
 The language identifier takes raw text and tries to figure out what language it
 was written in. The output can either be a plain-text i18n language code or a
-basic KAF document containing the language and raw input text.
-
-The output of the language identifier can then be used to drive further text
+basic KAF document containing the language and raw input text. The output of the language identifier can then be used to drive further text
 analysis of for example sentiments and or entities.
 
 ## Confused by some terminology?

data/bin/language-identifier

@@ -1,6 +1,7 @@
 #!/usr/bin/env ruby
+
 require_relative '../lib/opener/language_identifier'
 
-cli = Opener::LanguageIdentifier::CLI.new(:args => ARGV)
+cli = Opener::LanguageIdentifier::CLI.new
 
-puts cli.run(STDIN.tty? ? nil : STDIN.read)
+cli.run

data/bin/language-identifier-daemon

@@ -2,9 +2,9 @@
 
 require 'opener/daemons'
 
-exec_path = File.expand_path("../../exec/language-identifier.rb", __FILE__)
-
-Opener::Daemons::Controller.new(
-  :name      => "language-identifier",
-  :exec_path => exec_path
+controller = Opener::Daemons::Controller.new(
+  :name      => 'opener-language-identifier',
+  :exec_path => File.expand_path("../../exec/language-identifier.rb", __FILE__)
 )
+
+controller.run

data/bin/language-identifier-server

@@ -1,8 +1,10 @@
 #!/usr/bin/env ruby
 
-require 'puma/cli'
+require 'opener/webservice'
 
-rack_config = File.expand_path('../../config.ru', __FILE__)
+parser = Opener::Webservice::OptionParser.new(
+  'opener-language-identifier',
+  File.expand_path('../../config.ru', __FILE__)
+)
 
-cli = Puma::CLI.new([rack_config] + ARGV)
-cli.run
+parser.run

data/exec/language-identifier.rb

@@ -1,9 +1,9 @@
 #!/usr/bin/env ruby
 
 require 'opener/daemons'
+
 require_relative '../lib/opener/language_identifier'
 
-options = Opener::Daemons::OptParser.parse!(ARGV)
-daemon  = Opener::Daemons::Daemon.new(Opener::LanguageIdentifier, options)
+daemon = Opener::Daemons::Daemon.new(Opener::LanguageIdentifier)
 
 daemon.start

data/lib/opener/language_identifier.rb

@@ -1,8 +1,7 @@
+require 'java'
 require 'open3'
-require 'optparse'
+require 'slop'
 require 'builder'
-require 'java'
-require 'opener/core'
 
 require_relative '../../core/target/LanguageDetection-0.0.1.jar'
 import 'org.vicomtech.opennlp.LanguageDetection.CybozuDetector'
@@ -29,8 +28,9 @@ module Opener
     # @return [Hash]
     #
     DEFAULT_OPTIONS = {
-      :args => [],
-      :kaf  => true
+      :args  => [],
+      :kaf   => true,
+      :probs => false
     }.freeze
 
     ##
@@ -42,6 +42,9 @@ module Opener
     # @option options [TrueClass|FalseClass] :kaf When set to `true` the
     #  results will be displayed as KAF.
     #
+    # @option options [TrueClass|FalseClass] :probs Wen set the probabilities
+    #  are returned instead of the language/KAF.
+    #
     def initialize(options = {})
       @options = DEFAULT_OPTIONS.merge(options)
       @detector = Detector.instance
@@ -63,9 +66,6 @@ module Opener
       end
 
       return output
-
-    rescue Exception => error
-      return Opener::Core::ErrorLayer.new(input, error.message, self.class).add
     end
 
     alias identify run

data/lib/opener/language_identifier/cli.rb

@@ -1,138 +1,79 @@
 module Opener
   class LanguageIdentifier
     ##
-    # CLI wrapper around {Opener::LanguageIdentifier} using OptionParser.
+    # CLI wrapper around {Opener::LanguageIdentifier} using Slop.
     #
-    # @!attribute [r] options
-    #  @return [Hash]
-    #
-    # @!attribute [r] option_parser
-    #  @return [OptionParser]
+    # @!attribute [r] parser
+    #  @return [Slop]
     #
     class CLI
-      attr_reader :options, :option_parser
+      attr_reader :parser
+
+      def initialize
+        @parser = configure_slop
+      end
 
       ##
-      # @param [Hash] options
+      # @param [Array] argv
       #
-      def initialize(options = {})
-        @options = DEFAULT_OPTIONS.merge(options)
+      def run(argv = ARGV)
+        parser.parse(argv)
+      end
 
-        @option_parser = OptionParser.new do |opts|
-          opts.program_name   = 'language-identifier'
-          opts.summary_indent = '  '
+      ##
+      # @return [Slop]
+      #
+      def configure_slop
+        return Slop.new(:strict => false, :indent => 2, :help => true) do
+          banner 'Usage: language-identifier [OPTIONS]'
 
-          opts.on('-v', '--version', 'Shows the current version') do
-            show_version
-          end
+          separator <<-EOF.chomp
 
-          opts.on('-k', '--[no-]kaf', 'Output the language as KAF') do |v|
-            @options[:kaf] = v
-          end
+About:
 
-          opts.on('-p', '--probs', 'Provide probabilities, assumes --no-kaf') do
-            @options[:kaf]   = false
-            @options[:probs] = true
-          end
+    Language detection for various languages such as English and Dutch. This
+    command reads input from STDIN. Output can be a language code as plain text,
+    a KAF document containing the input text and language code, or a list of
+    probabilities.
+
+Example:
 
-          opts.separator <<-EOF
-
-Examples:
-
-  cat example_text.txt | #{opts.program_name}    # Basic detection
-
-Languages:
-
-  * ar	Arabic
-  * bg	Bulgarian
-  * bn	Bengali
-  * cs	Czech
-  * da	Danish
-  * de	German
-  * el	Greek
-  * en	English
-  * es	Spanish
-  * et	Estonian
-  * fa	Persian
-  * fi	Finnish
-  * fr	French
-  * gu	Gujarati
-  * he	Hebrew
-  * hi	Hindi
-  * hr	Croatian
-  * hu	Hungarian
-  * id	Indonesian
-  * it	Italian
-  * ja	Japanese
-  * kn	Kannada
-  * ko	Korean
-  * lt	Lithuanian
-  * lv	Latvian
-  * mk	Macedonian
-  * ml	Malayalam
-  * mr	Marathi
-  * ne	Nepali
-  * nl	Dutch
-  * no	Norwegian
-  * pa	Punjabi
-  * pl	Polish
-  * pt	Portuguese
-  * ro	Romanian
-  * ru	Russian
-  * sk	Slovak
-  * sl	Slovene
-  * so	Somali
-  * sq	Albanian
-  * sv	Swedish
-  * sw	Swahili
-  * ta	Tamil
-  * te	Telugu
-  * th	Thai
-  * tl	Tagalog
-  * tr	Turkish
-  * uk	Ukrainian
-  * ur	Urdu
-  * vi	Vietnamese
-  * zh-cn	Simplified Chinese
-  * zh-tw	Traditional Chinese
+    cat some_file.kaf | language-identifier
           EOF
 
-          opts.separator ""
-          opts.separator "Common options:"
-          # No argument, shows at tail.  This will print an options summary.
-          # Try it and see!
-          opts.on_tail("-h", "--help", "Show this message.") do
-            puts opts
-            exit
+          separator "\nOptions:\n"
+
+          on :v, :version, 'Shows the current version' do
+            abort "language-identifier v#{VERSION} on #{RUBY_DESCRIPTION}"
           end
-        end
-      end
 
-      ##
-      # @param [String] input
-      #
-      def run(input)
-        option_parser.parse!(options[:args])
-        identifier = LanguageIdentifier.new(options)
+          on :'no-kaf', 'Disables KAF output'
+          on :p, :probs, 'Displays probabilities instead of a language code'
 
-        output = identifier.run(input)
-        puts output
-      end
+          run do |opts, args|
+            enable_kaf   = true
+            enable_probs = false
 
-      private
+            if opts[:'no-kaf']
+              enable_kaf = false
+            end
 
-      ##
-      # Shows the help message and exits the program.
-      #
-      def show_help
-        abort option_parser.to_s
-      end
+            if opts[:probs]
+              enable_kf    = false
+              enable_probs = true
+            end
 
-      ##
-      # Shows the version and exits the program.
-      #
-      def show_version
-        abort "#{option_parser.program_name} v#{VERSION} on #{RUBY_DESCRIPTION}"
+            identifier = LanguageIdentifier.new(
+              :args  => args,
+              :kaf   => enable_kaf,
+              :probs => enable_probs
+            )
+
+            input = STDIN.tty? ? nil : STDIN.read
+
+            puts identifier.run(input)
+          end
+        end
       end
     end # CLI
   end # LanguageIdentifier

data/lib/opener/language_identifier/detector.rb

@@ -4,6 +4,11 @@ import 'org.vicomtech.opennlp.LanguageDetection.CybozuDetector'
 
 module Opener
   class LanguageIdentifier
+    ##
+    # Singleton class wrapped around the Cybozu detector. The Cybozu code uses
+    # the factory pattern and stores a bunch of things on class level. As such
+    # the Cybozu code is *not* thread-safe.
+    #
     class Detector
       attr_reader :options
 

data/lib/opener/language_identifier/server.rb

@@ -1,5 +1,3 @@
-require 'sinatra/base'
-require 'httpclient'
 require 'opener/webservice'
 
 module Opener
@@ -7,27 +5,11 @@ module Opener
     ##
     # A basic language identification server powered by Sinatra.
     #
-    class Server < Webservice
+    class Server < Opener::Webservice::Server
       set :views, File.expand_path('../views', __FILE__)
-      text_processor LanguageIdentifier
-      accepted_params :input, :kaf, :benchmark
 
-      ##
-      # Gets the Analyzed output of an input.
-      #
-      # @param [Hash] options The options for the text_processor
-      # @return [String] output the output of the text_processor
-      # @return [Symbol] type the output type ot the text_processor
-      #
-      # @raise RunetimeError Raised when the tagging process failed.
-      #
-      def analyze(options)
-        options[:kaf] = true if options[:kaf].nil?
-        processor = text_processor.new(options)
-        output    = processor.run(options[:input])
-
-        return output
-      end
+      self.text_processor  = LanguageIdentifier
+      self.accepted_params = [:input, :kaf]
     end # Server
   end # LanguageIdentifier
 end # Opener

data/lib/opener/language_identifier/version.rb

@@ -1,5 +1,5 @@
 module Opener
   class LanguageIdentifier
-    VERSION = "3.1.7"
+    VERSION = "4.1.0"
   end
 end

data/lib/opener/language_identifier/views/index.erb

@@ -32,20 +32,10 @@
       <div>
         <label for="kaf">
           <input type='hidden' value='false' name='kaf'>
-          <input type="checkbox" name="kaf" id="kaf" checked/>
+          <input type="checkbox" name="kaf" id="kaf" checked />
 
           Output KAF instead of just the language code
         </label>
-
-        <br>
-
-        <label for="benchmark">
-          <input type="checkbox" name="benchmark" />
-
-          Include benchmark output in the KAF
-        </label>
-
-      	<br>
         <br>
       </div>
       <% 10.times do |t| %>

data/opener-language-identifier.gemspec

@@ -25,15 +25,12 @@ Gem::Specification.new do |gem|
 
   gem.executables = Dir.glob('bin/*').map { |file| File.basename(file) }
 
+  gem.add_dependency 'opener-daemons', '~> 2.2'
+  gem.add_dependency 'opener-webservice', '~> 2.1'
+
   gem.add_dependency 'builder'
-  gem.add_dependency 'puma'
-  gem.add_dependency 'sinatra', '~>1.4.2'
-  gem.add_dependency 'httpclient'
-  gem.add_dependency 'uuidtools'
-  gem.add_dependency 'opener-webservice'
-  gem.add_dependency 'opener-daemons'
   gem.add_dependency 'nokogiri'
-  gem.add_dependency 'opener-core', '~> 1.0'
+  gem.add_dependency 'slop', '~> 3.5'
 
   gem.add_development_dependency 'rspec', '~> 3.0'
   gem.add_development_dependency 'cucumber'

metadata

@@ -1,101 +1,45 @@
 --- !ruby/object:Gem::Specification
 name: opener-language-identifier
 version: !ruby/object:Gem::Version
-  version: 3.1.7
+  version: 4.1.0
 platform: ruby
 authors:
 - development@olery.com
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2014-06-30 00:00:00.000000000 Z
+date: 2014-11-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
-  name: builder
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-  prerelease: false
-  type: :runtime
-- !ruby/object:Gem::Dependency
-  name: puma
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-  prerelease: false
-  type: :runtime
-- !ruby/object:Gem::Dependency
-  name: sinatra
+  name: opener-daemons
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 1.4.2
+        version: '2.2'
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 1.4.2
-  prerelease: false
-  type: :runtime
-- !ruby/object:Gem::Dependency
-  name: httpclient
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-  prerelease: false
-  type: :runtime
-- !ruby/object:Gem::Dependency
-  name: uuidtools
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - '>='
-      - !ruby/object:Gem::Version
-        version: '0'
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - '>='
-      - !ruby/object:Gem::Version
-        version: '0'
+        version: '2.2'
   prerelease: false
   type: :runtime
 - !ruby/object:Gem::Dependency
   name: opener-webservice
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ~>
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '2.1'
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ~>
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '2.1'
   prerelease: false
   type: :runtime
 - !ruby/object:Gem::Dependency
-  name: opener-daemons
+  name: builder
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - '>='
@@ -123,17 +67,17 @@ dependencies:
   prerelease: false
   type: :runtime
 - !ruby/object:Gem::Dependency
-  name: opener-core
+  name: slop
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: '1.0'
+        version: '3.5'
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: '1.0'
+        version: '3.5'
   prerelease: false
   type: :runtime
 - !ruby/object:Gem::Dependency