dwc-archive 0.9.6 → 0.9.10

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: 8e8c929203d1b652f8ba345b0c4c39cfc87a0369
+  data.tar.gz: 739a064221bf52523990bfea38749f11b4d986c0
+SHA512:
+  metadata.gz: adfd46bea84e301ceca6de355f189b884bc6a9a22eb8d98e3976f17e3407d2de7f15963ec0d6f29958de057de9d579c81d8007d72c38fc1b9f9ada9295381152
+  data.tar.gz: 2e188f828d0bbe28baf5f1d8aa7aab12ca330000bf3f5ee145501d78961f9ca5e696be1171998a433952cf84546a8d5fd61c6cf45958576a7e0800fbf4542dae

data/.gitignore

@@ -0,0 +1,30 @@
+## MAC OS
+.DS_Store
+
+## TEXTMATE
+*.tmproj
+tmtags
+
+## EMACS
+*~
+\#*
+.\#*
+
+## VIM
+*.swp
+
+## PROJECT::GENERAL
+coverage
+rdoc
+pkg
+install
+Gemfile.lock
+
+## PROJECT::SPECIFIC
+tags
+bin
+.bundle
+bundle_bin
+Gemfile.lock
+
+

data/.rspec

@@ -0,0 +1,2 @@
+--format nested
+--color

data/.ruby-version

@@ -0,0 +1 @@
+2.0.0-p353

data/.travis.yml

@@ -1,10 +1,12 @@
 rvm:
-  - 1.9.3
-  - 2.0.0
+  - 1.9.3-p448
+  - 2.0.0-p353
 before_install:
   - sudo apt-get update
-  - sudo apt-get install redis-server
-bundler_args: --without development
+  - gem install debugger
+# bundler_args: --without development
+services:
+  - redis-server
 script:
   - bundle exec cucumber
   - bundle exec rake

data/CHANGELOG

@@ -1,3 +1,5 @@
+0.9.7 Refactoring and tests improvements
+
 0.9.6 Added support for GNUB DwCA files
 
 0.9.4 Gem dependencies updated, added travis support

data/Gemfile

@@ -1,17 +1,3 @@
 source 'https://rubygems.org'
 
-gem 'nokogiri', '~> 1.5'
-gem 'parsley-store', '~> 0.3.2'
-gem 'archive-tar-minitar', '~> 0.5'
-
-group :development do
-  gem 'debugger', '~> 1.3'
-end
-
-group :test do
-  gem 'rspec', '~> 2.13'
-  gem 'cucumber', '~> 1.3'
-  gem 'bundler', '~> 1.3'
-  gem 'jeweler', '~> 1.8'
-  gem 'jazz_hands', '~> 0.5'
-end
+gemspec

data/README.md

@@ -3,18 +3,25 @@ Darwin Core Archive
 
 [![Gem Version][1]][2]
 [![Continuous Integration Status][3]][4]
-[![Dependency Status][5]][6]
+[![Coverage Status][5]][6]
+[![CodePolice][7]][8]
+[![Dependency Status][9]][10]
 
 Darwin Core Archive format is a current standard for information exchange
 between Global Names Architecture modules. This gem allows to work with
 Darwin Core Archive data compressed to either zip or tar.gz files.
-More information about Darwing Core Archive can be found on a [GBIF page:][7]
+More information about Darwing Core Archive can be found on a [GBIF page:][11]
 
 Installation
 ------------
 
     sudo gem install dwc-archive
 
+### System Requirements
+
+You need [Redis Server][12] and unzip library installed 
+
+
 Usage
 -----
 
@@ -174,6 +181,11 @@ Copyright (c) 2010-2013 Marine Biological Laboratory. See LICENSE for details.
 [2]: http://badge.fury.io/rb/dwc-archive
 [3]: https://secure.travis-ci.org/GlobalNamesArchitecture/dwc-archive.png
 [4]: http://travis-ci.org/GlobalNamesArchitecture/dwc-archive
-[5]: https://gemnasium.com/GlobalNamesArchitecture/dwc-archive.png
-[6]: https://gemnasium.com/GlobalNamesArchitecture/dwc-archive
-[7]: http://bit.ly/2IxcBA
+[5]: https://coveralls.io/repos/GlobalNamesArchitecture/dwc-archive/badge.png
+[6]: https://coveralls.io/r/GlobalNamesArchitecture/dwc-archive
+[7]: https://codeclimate.com/github/GlobalNamesArchitecture/dwc-archive.png
+[8]: https://codeclimate.com/github/GlobalNamesArchitecture/dwc-archive
+[9]: https://gemnasium.com/GlobalNamesArchitecture/dwc-archive.png
+[10]: https://gemnasium.com/GlobalNamesArchitecture/dwc-archive
+[11]: http://bit.ly/2IxcBA
+[12]: http://redis.io/topics/quickstart

data/Rakefile

@@ -1,28 +1,10 @@
-require 'rubygems'
-require 'rake'
+require "bundler/gem_tasks"
 
-begin
-  require 'jeweler'
-  Jeweler::Tasks.new do |gem|
-    gem.name = "dwc-archive"
-    gem.summary = %Q{Handler of Darwin Core Archive files}
-    gem.description = 'Darwin Core Archive is the current standard exchange ' +
-      'format for GLobal Names Architecture modules. ' +
-      'This gem makes it easy to incorporate files in ' +
-      'Darwin Core Archive format into a ruby project.'
-    gem.email = "dmozzherin at gmail dot com"
-    gem.homepage = "http://github.com/GlobalNamesArchitecture/dwc-archive"
-    gem.authors = ["Dmitry Mozzherin"]
-    #gem.add_dependency "fastercsv" if RUBY_VERSION.match /^1.8/
-    gem.add_dependency 'parsley-store', ">= 0.3.0"
-    gem.add_development_dependency "rspec", ">= 1.2.9"
-    gem.add_development_dependency "cucumber", ">= 0"
-  end
-  Jeweler::GemcutterTasks.new
-rescue LoadError
-  puts 'Jeweler (or a dependency) not available. ' +
-    'Install it with: gem install jeweler'
-end
+# Bundler::GemHelper.install_tasks
+# require 'bundler/gem_tasks'
+# require 'rake/testtasks'
+# require 'rubygems'
+# require 'rake'
 
 require 'rspec/core/rake_task'
 RSpec::Core::RakeTask.new(:spec) do |spec|

data/]

@@ -0,0 +1,40 @@
+require_relative '../spec_helper'
+
+describe DarwinCore::Core do
+  subject(:dwca) { DarwinCore.new(file_path) }
+  subject(:core) { DarwinCore::Core.new(dwca) }
+  let(:file_path) { File.join(File.expand_path('../../files', __FILE__), 
+                             file_name) }
+  let(:file_name) { 'data.tar.gz' }
+  
+
+  describe '.new' do
+    it 'creates new core' do
+      expect(core).to be_kind_of DarwinCore::Core
+    end
+  end
+  
+  describe '#id' do
+
+    it 'returns core id' do
+      expect(core.id[:index]).to eq 0
+      expect(core.id[:term]).to eq 'http://rs.tdwg.org/dwc/terms/TaxonID'
+    end
+
+    context 'no coreid' do
+      let(:file_name) { 'empty_coreid.tar.gz' }
+
+      it 'does not return coreid' do
+        expect(core.id[:index]).to eq 0
+        expect(core.id[:term]).to be_nil
+      end
+    end
+  end
+
+  it 'reads core file from archive' do
+
+    core.read
+
+  end
+
+end

data/dwc-archive.gemspec

@@ -0,0 +1,33 @@
+require File.expand_path('../lib/dwc-archive/version', __FILE__)
+
+Gem::Specification.new do |gem|
+  gem.name          = "dwc-archive"
+  gem.version       = DarwinCore::VERSION
+  gem.authors       = ["Dmitry Mozzherin"]
+  gem.email         = ["dmozzherin at gmail dot com"]
+  gem.description   = %q{Darwin Core Archive is the current standard exchange 
+                          format for GLobal Names Architecture modules.  
+                          This gem makes it easy to incorporate files in 
+                          Darwin Core Archive format into a ruby project.}
+  gem.summary       = %q{Handler of Darwin Core Archive files}
+  gem.homepage      = "http://github.com/GlobalNamesArchitecture/dwc-archive"
+  gem.license       = "MIT"
+
+  gem.files         = `git ls-files`.split($/)
+  gem.executables   = gem.files.grep(%r{^bin/}) { |f| File.basename(f) }
+  gem.test_files    = gem.files.grep(%r{^(test|spec|features)/})
+  gem.require_paths = ["lib"]
+
+  gem.add_runtime_dependency 'nokogiri', '~> 1.6'
+  gem.add_runtime_dependency 'parsley-store', '~> 0.3'
+  gem.add_runtime_dependency 'archive-tar-minitar', '~> 0.5'
+  
+  gem.add_development_dependency 'rake', '~> 10.1'
+  gem.add_development_dependency 'bundler', '~> 1.3'
+  gem.add_development_dependency 'rspec', '~> 2.14'
+  gem.add_development_dependency 'cucumber', '~> 1.3'
+  gem.add_development_dependency 'coveralls', '~> 0.7'
+  gem.add_development_dependency 'debugger', '~> 1.6'
+  gem.add_development_dependency 'git', '~> 1.2'
+end
+

data/lib/dwc-archive.rb

@@ -1,34 +1,36 @@
 # encoding: UTF-8
-$:.unshift(File.dirname(__FILE__)) unless
-   $:.include?(File.dirname(__FILE__)) || $:.include?(File.expand_path(File.dirname(__FILE__)))
-R19 = RUBY_VERSION.split('.')[0..1].join('').to_i > 18
-raise "IMPORTANT: dwc-archive gem  requires ruby >= 1.9.1" unless R19
+
+recent_ruby = RUBY_VERSION >= '1.9.1'
+raise "IMPORTANT: dwc-archive gem  requires ruby >= 1.9.1" unless recent_ruby
+
 require 'fileutils'
 require 'ostruct'
 require 'digest'
 require 'csv'
 require 'logger'
-require 'dwc-archive/xml_reader'
-require 'dwc-archive/ingester'
-require 'dwc-archive/errors'
-require 'dwc-archive/expander'
-require 'dwc-archive/archive'
-require 'dwc-archive/core'
-require 'dwc-archive/extension'
-require 'dwc-archive/metadata'
-require 'dwc-archive/generator'
-require 'dwc-archive/generator_meta_xml'
-require 'dwc-archive/generator_eml_xml'
-require 'dwc-archive/classification_normalizer'
+require_relative 'dwc-archive/xml_reader'
+require_relative 'dwc-archive/ingester'
+require_relative 'dwc-archive/errors'
+require_relative 'dwc-archive/expander'
+require_relative 'dwc-archive/archive'
+require_relative 'dwc-archive/core'
+require_relative 'dwc-archive/extension'
+require_relative 'dwc-archive/metadata'
+require_relative 'dwc-archive/generator'
+require_relative 'dwc-archive/generator_meta_xml'
+require_relative 'dwc-archive/generator_eml_xml'
+require_relative 'dwc-archive/classification_normalizer'
+require_relative 'dwc-archive/version'
 
 class DarwinCore
 
-  VERSION = open(File.join(File.dirname(__FILE__), '..', 'VERSION')).readline.strip
+  VERSION = DarwinCore::VERSION
+  DEFAULT_TMP_DIR = "/tmp"
 
-  attr_reader :archive, :core, :metadata, :extensions, :classification_normalizer
+  attr_reader :archive, :core, :metadata, :extensions, 
+    :classification_normalizer
   alias :eml :metadata
 
-  DEFAULT_TMP_DIR = "/tmp"
 
   def self.nil_field?(field)
     return true if [nil, '', '/N'].include?(field)
@@ -68,10 +70,20 @@ class DarwinCore
     @extensions = get_extensions
   end
 
-  # generates a hash from a classification data with path to each node, list of synonyms and vernacular names.
+  def file_name
+    File.split(@dwc_path).last
+  end
+
+  def path
+    File.expand_path(@dwc_path)
+  end
+
+  # generates a hash from a classification data with path to each node, 
+  # list of synonyms and vernacular names.
   def normalize_classification
     return nil unless has_parent_id?
-    @classification_normalizer ||= DarwinCore::ClassificationNormalizer.new(self)
+    @classification_normalizer ||= DarwinCore::ClassificationNormalizer.
+      new(self)
     @classification_normalizer.normalize
   end
 

data/lib/dwc-archive/archive.rb

@@ -8,8 +8,11 @@ class DarwinCore
       @expander = DarwinCore::Expander.new(@archive_path, @tmp_dir)
       @expander.unpack
       if valid?
-        @meta = DarwinCore::XmlReader.from_xml(open(File.join(@expander.path, 'meta.xml')))
-        @eml = files.include?("eml.xml") ? DarwinCore::XmlReader.from_xml(open(File.join(@expander.path, 'eml.xml'))) : nil
+        @meta = DarwinCore::XmlReader.
+          from_xml(open(File.join(@expander.path, 'meta.xml')))
+        @eml = files.include?("eml.xml") ? 
+          DarwinCore::XmlReader.
+            from_xml(open(File.join(@expander.path, 'eml.xml'))) : nil
       else
         clean
         raise InvalidArchiveError

data/lib/dwc-archive/classification_normalizer.rb

@@ -57,6 +57,10 @@ class DarwinCore
       @tree = {}
     end
 
+    def darwin_core
+      @dwc
+    end
+
     def add_name_string(name_string)
       @name_strings[name_string] = 1 unless @name_strings[name_string]
     end

data/lib/dwc-archive/core.rb

@@ -8,9 +8,9 @@ class DarwinCore
       @path = @archive.files_path
       root_key = @archive.meta.keys[0]
       @data = @archive.meta[root_key][:core]
-      raise DarwinCore::CoreFileError.new("Cannot find core in meta.xml, is meta.xml valid?") unless @data
+      raise DarwinCore::CoreFileError.
+        new("Cannot find core in meta.xml, is meta.xml valid?") unless @data
       @id = @data[:id][:attributes] 
-      # raise DarwinCore::CoreFileError.new("Cannot find core identifier") unless @id
       get_attributes(DarwinCore::CoreFileError)
     end
   end

data/lib/dwc-archive/expander.rb

@@ -43,12 +43,16 @@ class DarwinCore
       if file_type.match(/tar.*gzip/i)
         return proc do |tmp_path, archive_path|
           FileUtils.mkdir tmp_path
-          system("tar -zxf #{esc(archive_path)} -C #{tmp_path} > /dev/null 2>&1")
+          path = esc(archive_path)
+          system("tar -zxf #{path} -C #{tmp_path} > /dev/null 2>&1")
         end
       end
 
       if file_type.match(/Zip/)
-        return proc { |tmp_path, archive_path| system("unzip -qq -d #{tmp_path} #{esc(archive_path)} > /dev/null 2>&1") }
+        return proc do |tmp_path, archive_path| 
+          path = esc(archive_path)
+          system("unzip -qq -d #{tmp_path} #{path} > /dev/null 2>&1")
+        end
       end
 
       return nil

data/lib/dwc-archive/generator.rb

@@ -8,7 +8,8 @@ class DarwinCore
       @path = File.join(tmp_dir, 'dwc_' + rand(10000000000).to_s)
       FileUtils.mkdir(@path)
       @meta_xml_data = {:extensions => []}
-      @eml_xml_data = {:id => nil, :title => nil, :authors => [], :abstract => nil, :citation => nil, :url => nil}
+      @eml_xml_data = {:id => nil, :title => nil, 
+        :authors => [], :abstract => nil, :citation => nil, :url => nil}
       @write = 'w:utf-8'
     end
     
@@ -22,27 +23,36 @@ class DarwinCore
       header = data.shift
       fields = header.map do |f|
         f.strip!
-        raise DarwinCore::GeneratorError.new("No header in core data, or header fields are not urls") unless f.match(/^http:\/\//)
-        f.split("/")[-1]
+        err = 'No header in core data, or header fields are not urls'
+        raise DarwinCore::GeneratorError.new(err) unless f.match(/^http:\/\//)
+        f.split('/')[-1]
       end
       data.unshift(fields) if keep_headers
       ignore_header_lines = keep_headers ? 1 : 0
-      @meta_xml_data[:core] = {:fields => header, :ignoreHeaderLines => ignore_header_lines, :location => file_name}
+      @meta_xml_data[:core] = { fields: header, 
+                                ignoreHeaderLines: ignore_header_lines, 
+                                location:file_name }
       data.each {|d| c << d}
       c.close
     end
 
-    def add_extension(data, file_name, keep_headers = true, row_type = "http://rs.tdwg.org/dwc/terms/Taxon")
+    def add_extension(data, file_name, 
+                      keep_headers = true, 
+                      row_type = 'http://rs.tdwg.org/dwc/terms/Taxon')
       c = CSV.open(File.join(@path,file_name), @write)
       header = data.shift
       fields = header.map do |f|
         f.strip!
-        raise DarwinCore::GeneratorError.new("No header in core data, or header fields are not urls") unless f.match(/^http:\/\//)
-        f.split("/")[-1]
+        err = 'No header in core data, or header fields are not urls'
+        raise DarwinCore::GeneratorError.new(err) unless f.match(/^http:\/\//)
+        f.split('/')[-1]
       end
       data.unshift(fields) if keep_headers
       ignore_header_lines = keep_headers ? 1 : 0
-      @meta_xml_data[:extensions] << { :fields => header, :ignoreHeaderLines => ignore_header_lines, :location => file_name, :rowType => row_type }
+      @meta_xml_data[:extensions] << { fields: header, 
+                                       ignoreHeaderLines: ignore_header_lines, 
+                                       location: file_name, 
+                                       rowType: row_type }
       data.each { |d| c << d }
       c.close
     end

data/lib/dwc-archive/generator_eml_xml.rb

@@ -9,26 +9,28 @@ class DarwinCore
       end
 
       def create
+        eml_uri =  'eml://ecoinformatics.org/eml-2.1.1' + 
+          ' http://rs.gbif.org/schema/eml-gbif-profile/1.0.1/eml.xsd'
         builder = Nokogiri::XML::Builder.new do |xml|
-          xml.eml(:packageId      => "%s/%s" % [@data[:id], timestamp],
-            :system               => @data[:system] || "http://globalnames.org",
-            :'xml:lang'           => "en",
-            :'xmlns:eml'          => "eml://ecoinformatics.org/eml-2.1.1",
-            :'xmlns:md'           => "eml://ecoinformatics.org/methods-2.1.1",
-            :'xmlns:proj'         => "eml://ecoinformatics.org/project-2.1.1",
-            :'xmlns:d'            => "eml://ecoinformatics.org/dataset-2.1.1",
-            :'xmlns:res'          => "eml://ecoinformatics.org/resource-2.1.1",
-            :'xmlns:dc'           => "http://purl.org/dc/terms/",
-            :'xmlns:xsi'          => "http://www.w3.org/2001/XMLSchema-instance",
-            :'xsi:schemaLocation' => "eml://ecoinformatics.org/eml-2.1.1 http://rs.gbif.org/schema/eml-gbif-profile/1.0.1/eml.xsd") do
-            xml.dataset(:id => @data[:id]) do
+          xml.eml(packageId:      "%s/%s" % [@data[:id], timestamp],
+            system:               @data[:system] || 'http://globalnames.org',
+            :'xml:lang' =>         'en',
+            :'xmlns:eml' =>        'eml://ecoinformatics.org/eml-2.1.1',
+            :'xmlns:md' =>         'eml://ecoinformatics.org/methods-2.1.1',
+            :'xmlns:proj' =>       'eml://ecoinformatics.org/project-2.1.1',
+            :'xmlns:d' =>          'eml://ecoinformatics.org/dataset-2.1.1',
+            :'xmlns:res' =>        'eml://ecoinformatics.org/resource-2.1.1',
+            :'xmlns:dc' =>         'http://purl.org/dc/terms/',
+            :'xmlns:xsi' =>        'http://www.w3.org/2001/XMLSchema-instance',
+            :'xsi:schemaLocation' => 'eml_uri')  do
+            xml.dataset(id: @data[:id]) do
               xml.title(@data[:title])
               xml.license(@data[:license])
               contacts = []
               @data[:authors].each_with_index do |a, i|
                 creator_id = i + 1
                 contacts << creator_id
-                xml.creator(:id => creator_id, :scope => 'document') do
+                xml.creator(id: creator_id, scope: 'document') do
                   xml.individualName do
                     xml.givenName(a[:first_name])
                     xml.surName(a[:last_name])
@@ -77,7 +79,7 @@ class DarwinCore
       private
       def timestamp
         t = Time.now.getutc.to_a[0..5].reverse
-        t[0..2].join('-') + "::" + t[-3..-1].join(':')
+        t[0..2].join('-') + '::' + t[-3..-1].join(':')
       end
     end
   end