opener-tree-tagger 3.2.3 → 3.2.4

checksums.yaml

@@ -1,15 +1,7 @@
 ---
-!binary "U0hBMQ==":
-  metadata.gz: !binary |-
-    NDZlNWM0ZDJkNmY1NzI0YjBmMzJkZjViNTIwODFiY2I0Njc4ZjdlNQ==
-  data.tar.gz: !binary |-
-    Zjc2MDA0OGNmYzZkMzQ1NDZiMTJhNGQzM2M0N2QyMjAzZmFhY2ZiMg==
+SHA1:
+  metadata.gz: f02a0878183c9f6df7ed50944257698ff369a130
+  data.tar.gz: 101a607039771002fdcf962b7e11ef8bcc863387
 SHA512:
-  metadata.gz: !binary |-
-    Mzg1ODg3MDU3MDhkMGI4M2U5ZjIxMzM2NTAxODIwOWYyOWU0MjM5NGYzZjY2
-    MjgyNDNlZDBiOTg5MTdjODNlYTBhNjYyZDUyYTYzZmVmOWQ0OTNiNDA5Y2Y3
-    OGNkZWJkMWY5YTg1ZTJhY2U3NzdjMGE5ZGY0ZDcyNTA2YWFlODU=
-  data.tar.gz: !binary |-
-    YWVlOGU3YzdmOGJjODQzMjdiNGVkZWZlMzA4YTE3NzE0MTI4YTk2MGU0NTFh
-    OWE4MDk5ZmU4NTg5MGFjYTZhYTE4NzVkMDQ4NDM0MmI4YTJjZTI4YmQ4NzFj
-    YmM4ODY0MDEwYjVmMzc3Y2Y1MDJlNTFmZmMzNzRlMDQ4M2ZhMDU=
+  metadata.gz: 2b6c0d06f37d18d94bcf74cf9b61f671b995e2296bdbb6ae2c832f491b7169619186c6d1f591c1470dc90c1a998106124ef3afd5ca278c1d346313aaaa999fab
+  data.tar.gz: 80e5eb90d1012ac73aa0fe00a19471e632d0cf72d5e16191ced19ca833182cdd139810f0a3c2294229d16a085202745709058193676366a88332f85a3340c985

data/README.md

@@ -101,7 +101,7 @@ The component is a fat wrapper around the actual language technology core. You c
 Where to go from here
 ---------------------
 
-* [Check the project websitere](http://opener-project.github.io)
+* [Check the project website](http://opener-project.github.io)
 * [Checkout the webservice](http://opener.olery.com/tree-tagger)
 
 Report problem/Get help

data/ext/hack/Rakefile

@@ -1,17 +1,8 @@
-require 'rake'
-require_relative 'support'
+import File.expand_path('../../../task/requirements.rake', __FILE__)
+import File.expand_path('../../../task/python.rake', __FILE__)
 
-desc 'Verifies the requirements'
-task :requirements do
-  verify_requirements
-end
-
-task :compile => :requirements do
-  install_tree_tagger
-end
-
-task :default => [:compile, :requirements] do
-  path = File.join(PYTHON_SITE_PACKAGES, 'pre_install')
-
-  pip_install(PRE_INSTALL_REQUIREMENTS, path)
+task :default => :requirements do
+  Dir.chdir(File.expand_path('../../../', __FILE__)) do
+    Rake::Task['core/site-packages/pre_install'].invoke
+  end
 end

data/lib/opener/tree_tagger.rb

@@ -1,8 +1,10 @@
 require 'open3'
 require 'optparse'
+require 'nokogiri'
 
 require_relative 'tree_tagger/version'
 require_relative 'tree_tagger/cli'
+require_relative 'tree_tagger/error_layer'
 
 module Opener
   class TreeTagger
@@ -29,15 +31,13 @@ module Opener
     end
 
     def run(input)
-      stdout, stderr, process = capture(input)
-
-      if process.success?
-        STDERR.puts(stderr) unless stderr.empty?
-      else
-        abort stderr
-      end
-
-      return stdout, stderr, process
+      begin
+        stdout, stderr, process = capture(input)
+        raise stderr unless process.success?
+        return stdout
+      rescue Exception => error
+        return ErrorLayer.new(input, error.message, self.class).add
+      end        
     end
 
     def capture(input)
@@ -45,17 +45,11 @@ module Opener
     end
 
     def command
-      "#{adjust_python_path} #{prepend_tree_tagger_path} python -E -OO #{kernel} #{args.join(' ')}"
+      return "#{adjust_python_path} python -E -OO #{kernel} #{args.join(' ')}"
     end
 
     protected
 
-    def prepend_tree_tagger_path
-      return if ENV['TREE_TAGGER_PATH']
-      path = File.expand_path('../../../core/vendor/build', __FILE__)
-      return "TREE_TAGGER_PATH=#{path}"
-    end
-
     ##
     # @return [String]
     #

data/lib/opener/tree_tagger/cli.rb

@@ -35,17 +35,7 @@ Examples:
 
         tagger = TreeTagger.new(options)
 
-        stdout, stderr, process = tagger.run(input)
-
-        if process.success?
-          puts stdout
-
-          if options[:logging] and !stderr.empty?
-            STDERR.puts(stderr)
-          end
-        else
-          abort stderr
-        end
+        puts tagger.run(input)
       end
 
       private

data/lib/opener/tree_tagger/error_layer.rb

@@ -0,0 +1,89 @@
+module Opener
+  class TreeTagger
+    ##
+    # Add Error Layer to KAF file instead of throwing an error.
+    #
+    class ErrorLayer
+      attr_accessor :input, :document, :error, :klass
+      
+      def initialize(input, error, klass)
+        @input    = input.to_s
+        # Make sure there is always a document, even if it is empty.
+        @document = Nokogiri::XML(input) rescue Nokogiri::XML(nil)
+        @error    = error
+        @klass    = klass
+      end
+      
+      def add
+        if is_xml?
+          unless has_errors_layer?
+            add_errors_layer
+          end
+        else
+          add_root
+          add_text
+          add_errors_layer
+        end
+        add_error
+        
+        xml = !!document.encoding ? document.to_xml : document.to_xml(:encoding => "UTF-8")
+        
+        return xml
+      end
+      
+      ##
+      # Check if the document is a valid XML file.
+      #
+      def is_xml?
+        !!document.root
+      end
+      
+      ##
+      # Add root element to the XML file.
+      #
+      def add_root
+        root = Nokogiri::XML::Node.new "KAF", document
+        document.add_child(root)
+      end
+      
+      ##
+      # Check if the document already has an errors layer.
+      #
+      def has_errors_layer?
+        !!document.at('errors')
+      end
+      
+      ##
+      # Add errors element to the XML file.
+      #
+      def add_errors_layer
+        node = Nokogiri::XML::Node.new "errors", document
+        document.root.add_child(node)
+      end
+      
+      ##
+      # Add the text file incase it is not a valid XML document. More
+      # info for debugging.
+      #
+      def add_text
+        node = Nokogiri::XML::Node.new "raw", document
+        node.inner_html = input
+        document.root.add_child(node)
+        
+      end
+      
+      ##
+      # Add the actual error to the errors layer.
+      #
+      def add_error
+        node = document.at('errors')
+        error_node = Nokogiri::XML::Node.new "error", node
+        error_node['class']   = klass.to_s
+        error_node['version'] = klass::VERSION
+        error_node.inner_html = error
+        node.add_child(error_node)
+      end     
+      
+    end # ErrorLayer
+  end # TreeTagger
+end # Opener

data/lib/opener/tree_tagger/version.rb

@@ -1,5 +1,5 @@
 module Opener
   class TreeTagger
-    VERSION = "3.2.3"
+    VERSION = "3.2.4"
   end
 end

data/opener-tree-tagger.gemspec

@@ -18,19 +18,21 @@ Gem::Specification.new do |gem|
     'config.ru',
     '*.gemspec',
     '*_requirements.txt',
-    'README.md'
+    'README.md',
+    'task/*'
   ]).select { |file| File.file?(file) }
 
   gem.executables = Dir.glob('bin/*').map { |file| File.basename(file) }
 
   gem.add_dependency 'opener-daemons'
-  gem.add_dependency 'opener-build-tools', ['>= 0.2.7']
   gem.add_dependency 'rake'
   gem.add_dependency 'sinatra'
   gem.add_dependency 'httpclient'
   gem.add_dependency 'puma'
   gem.add_dependency 'opener-webservice'
+  gem.add_dependency 'nokogiri'
+  gem.add_dependency 'cliver'
 
-  gem.add_development_dependency 'rspec'
+  gem.add_development_dependency 'rspec', '~> 3.0'
   gem.add_development_dependency 'cucumber'
 end

data/task/compile.rake

@@ -0,0 +1,2 @@
+desc 'Verifies requirements and compiles the core'
+task :compile => ['requirements', 'python:compile']

data/task/python.rake

@@ -0,0 +1,11 @@
+# NOTE: pre_build/pre_install directories are created by pip.
+
+directory 'core/site-packages/pre_install' do |task|
+  sh "pip install --requirement=pre_install_requirements.txt " \
+    "--target=#{task.name} --ignore-installed"
+end
+
+namespace :python do
+  desc 'Installs Python packages in a local directory'
+  task :compile => ['core/site-packages/pre_install']
+end

data/task/requirements.rake

@@ -0,0 +1,7 @@
+desc 'Verifies the requirements'
+task :requirements do
+  require 'cliver'
+
+  Cliver.detect!('python', '~> 2.6')
+  Cliver.detect!('pip', '~> 1.3')
+end

data/task/test.rake

@@ -0,0 +1,4 @@
+desc 'Runs the tests'
+task :test => :compile do
+  sh('cucumber features')
+end

metadata

@@ -1,143 +1,157 @@
 --- !ruby/object:Gem::Specification
 name: opener-tree-tagger
 version: !ruby/object:Gem::Version
-  version: 3.2.3
+  version: 3.2.4
 platform: ruby
 authors:
 - rubenIzquierdo
 - sparkboxx
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2014-05-26 00:00:00.000000000 Z
+date: 2014-06-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: opener-daemons
+  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'
-  type: :runtime
   prerelease: false
+  type: :runtime
+- !ruby/object:Gem::Dependency
+  name: rake
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-- !ruby/object:Gem::Dependency
-  name: opener-build-tools
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
-        version: 0.2.7
-  type: :runtime
+        version: '0'
   prerelease: false
+  type: :runtime
+- !ruby/object:Gem::Dependency
+  name: sinatra
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
-        version: 0.2.7
-- !ruby/object:Gem::Dependency
-  name: rake
+        version: '0'
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-  type: :runtime
   prerelease: false
+  type: :runtime
+- !ruby/object:Gem::Dependency
+  name: httpclient
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-- !ruby/object:Gem::Dependency
-  name: sinatra
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-  type: :runtime
   prerelease: false
+  type: :runtime
+- !ruby/object:Gem::Dependency
+  name: puma
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-- !ruby/object:Gem::Dependency
-  name: httpclient
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-  type: :runtime
   prerelease: false
+  type: :runtime
+- !ruby/object:Gem::Dependency
+  name: opener-webservice
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-- !ruby/object:Gem::Dependency
-  name: puma
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-  type: :runtime
   prerelease: false
+  type: :runtime
+- !ruby/object:Gem::Dependency
+  name: nokogiri
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-- !ruby/object:Gem::Dependency
-  name: opener-webservice
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-  type: :runtime
   prerelease: false
+  type: :runtime
+- !ruby/object:Gem::Dependency
+  name: cliver
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-- !ruby/object:Gem::Dependency
-  name: rspec
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-  type: :development
   prerelease: false
+  type: :runtime
+- !ruby/object:Gem::Dependency
+  name: rspec
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - ~>
       - !ruby/object:Gem::Version
-        version: '0'
-- !ruby/object:Gem::Dependency
-  name: cucumber
+        version: '3.0'
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - ~>
       - !ruby/object:Gem::Version
-        version: '0'
-  type: :development
+        version: '3.0'
   prerelease: false
+  type: :development
+- !ruby/object:Gem::Dependency
+  name: cucumber
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ! '>='
+    - - '>='
       - !ruby/object:Gem::Version
         version: '0'
-description: ! 'Ruby wrapped KAF based Tree Tagger for 6 languages '
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  prerelease: false
+  type: :development
+description: 'Ruby wrapped KAF based Tree Tagger for 6 languages '
 email:
 - ruben.izquierdobevia@vu.nl
 - wilco@olery.com
@@ -149,6 +163,11 @@ extensions:
 - ext/hack/Rakefile
 extra_rdoc_files: []
 files:
+- README.md
+- bin/tree-tagger
+- bin/tree-tagger-daemon
+- bin/tree-tagger-server
+- config.ru
 - core/dutch.map.treetagger.kaf.csv
 - core/english.map.treetagger.kaf.csv
 - core/french.map.treetagger.kaf.csv
@@ -159,43 +178,41 @@ files:
 - core/tt_from_kaf_to_kaf.py
 - exec/tree-tagger.rb
 - ext/hack/Rakefile
-- ext/hack/support.rb
+- lib/opener/tree_tagger.rb
 - lib/opener/tree_tagger/cli.rb
+- lib/opener/tree_tagger/error_layer.rb
 - lib/opener/tree_tagger/public/markdown.css
 - lib/opener/tree_tagger/server.rb
 - lib/opener/tree_tagger/version.rb
 - lib/opener/tree_tagger/views/index.erb
 - lib/opener/tree_tagger/views/result.erb
-- lib/opener/tree_tagger.rb
-- config.ru
 - opener-tree-tagger.gemspec
 - pre_install_requirements.txt
-- README.md
-- bin/tree-tagger
-- bin/tree-tagger-daemon
-- bin/tree-tagger-server
+- task/compile.rake
+- task/python.rake
+- task/requirements.rake
+- task/test.rake
 homepage: http://opener-project.github.com/
 licenses: []
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ! '>='
+  - - '>='
     - !ruby/object:Gem::Version
       version: '0'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ! '>='
+  - - '>='
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.1.11
-signing_key: 
+rubyforge_project:
+rubygems_version: 2.2.2
+signing_key:
 specification_version: 4
 summary: Ruby wrapped KAF based Tree Tagger for 6 languages
 test_files: []
-has_rdoc: 

data/ext/hack/support.rb

@@ -1,65 +0,0 @@
-require 'opener/build-tools'
-
-include Opener::BuildTools::Requirements
-include Opener::BuildTools::Python
-include Opener::BuildTools::Files
-
-# Directory where packages will be installed to.
-PYTHON_SITE_PACKAGES = File.expand_path(
-  '../../../core/site-packages',
-  __FILE__
-)
-
-# Directory containing the temporary files.
-TMP_DIRECTORY = File.expand_path('../../../tmp', __FILE__)
-
-# Path to the pip requirements file used to install requirements before
-# packaging the Gem.
-PRE_BUILD_REQUIREMENTS = File.expand_path(
-  '../../../pre_build_requirements.txt',
-  __FILE__
-)
-
-# Path to the pip requirements file used to install requirements upon Gem
-# installation.
-PRE_INSTALL_REQUIREMENTS = File.expand_path(
-  '../../../pre_install_requirements.txt',
-  __FILE__
-)
-
-# Path to the vendor directory for C code.
-VENDOR_DIRECTORY = File.expand_path('../../../core/vendor', __FILE__)
-
-# Path to the directory to install vendored C code into.
-VENDOR_BUILD_DIRECTORY = File.expand_path(
-  '../../../core/vendor/build',
-  __FILE__
-)
-
-##
-# Verifies the requirements to install thi Gem.
-#
-def verify_requirements
-  require_executable('python')
-  require_version('python', python_version, '2.6.0')
-  require_executable('pip')
-  require_version('pip', pip_version, '1.3.1')
-end
-
-def install_tree_tagger
-  #return if ENV['TREE_TAGGER_PATH']
-  build = VENDOR_BUILD_DIRECTORY
-  mkdir_p(build)
-  cd build do
-    sh "curl -o tree-tagger-linux-3.2.tar.gz  http://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/data/tree-tagger-linux-3.2.tar.gz"
-    sh "curl -o tagger-scripts.tar.gz  http://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/data/tagger-scripts.tar.gz"
-    sh "curl -o install-tagger.sh  http://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/data/install-tagger.sh"
-    sh "curl -o dutch-par-linux-3.2-utf8.bin.gz http://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/data/dutch-par-linux-3.2-utf8.bin.gz"
-    sh "curl -o french-par-linux-3.2-utf8.bin.gz http://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/data/french-par-linux-3.2-utf8.bin.gz"
-    sh "curl -o german-par-linux-3.2-utf8.bin.gz http://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/data/german-par-linux-3.2-utf8.bin.gz"
-    sh "curl -o italian-par-linux-3.2-utf8.bin.gz http://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/data/italian-par-linux-3.2-utf8.bin.gz"
-    sh "curl -o spanish-par-linux-3.2-utf8.bin.gz http://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/data/spanish-par-linux-3.2-utf8.bin.gz"
-    sh "curl -o english-par-linux-3.2.bin.gz http://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/data/english-par-linux-3.2.bin.gz"
-    sh "sh install-tagger.sh"
-  end
-end