dwc-archive 0.5.12 → 0.5.13

data/Rakefile

@@ -21,14 +21,12 @@ rescue LoadError
   puts "Jeweler (or a dependency) not available. Install it with: gem install jeweler"
 end
 
-require 'spec/rake/spectask'
-Spec::Rake::SpecTask.new(:spec) do |spec|
-  spec.libs << 'lib' << 'spec'
-  spec.spec_files = FileList['spec/**/*_spec.rb']
+require 'rspec/core/rake_task'
+RSpec::Core::RakeTask.new(:spec) do |spec|
+  spec.pattern = 'spec/**/*_spec.rb'
 end
 
-Spec::Rake::SpecTask.new(:rcov) do |spec|
-  spec.libs << 'lib' << 'spec'
+RSpec::Core::RakeTask.new(:rcov) do |spec|
   spec.pattern = 'spec/**/*_spec.rb'
   spec.rcov = true
 end

data/VERSION

@@ -1 +1 @@
-0.5.12
+0.5.13

data/features/dwca-reader.feature

@@ -6,8 +6,8 @@ Feature: Reading of a Darwing Core Archive
   Scenario: Creating Darwin Core Archive object
     Given path to a dwc file "data.tar.gz"
     When I create a new DarwinCore::Archive instance
-    Then I should find that the archive is valid 
-    Then I should see what files the archive has
+    Then I should find that the archive is valid
+    And I should see what files the archive has
 
     When I delete expanded files
     Then they should disappear
@@ -19,9 +19,10 @@ Feature: Reading of a Darwing Core Archive
 
   Scenario: Instantiating DarwinCore with tar.gz file
     Given path to a dwc file "data.tar.gz"
-    When I create a new DarwinCore instance 
+    When I create a new DarwinCore instance
     Then instance should have a valid archive
     And instance should have a core
+    And I should see checksum
     When I check core data
     Then I should find core.properties
     And core.file_path
@@ -30,13 +31,13 @@ Feature: Reading of a Darwing Core Archive
     And core.size
     Then DarwinCore instance should have an extensions array
     And every extension in array should be an instance of DarwinCore::Extension
-    And extension should have properties, data, file_path, coreid, fields 
+    And extension should have properties, data, file_path, coreid, fields
     Then DarwinCore instance should have dwc.metadata object
     And I should find id, title, creators, metadata provider
 
   Scenario: Instantiating DawinCore with zip file
     Given path to a dwc file "data.zip"
-    When I create a new DarwinCore instance 
+    When I create a new DarwinCore instance
     Then instance should have a valid archive
 
   Scenario: Cleaning temporary directory from expanded archives
@@ -49,7 +50,7 @@ Feature: Reading of a Darwing Core Archive
     When I create a new DarwinCore instance
     Then I can read its content into memory
     Then I can read extensions content into memory
-  
+
   Scenario: Importing data using block
     Given path to a dwc file "data.tar.gz"
     When I create a new DarwinCore instance

data/features/step_definitions/dwc-reader_steps.rb

@@ -42,6 +42,10 @@ Then /^instance should have a core$/ do
   @dwc.core.class.should == DarwinCore::Core
 end
 
+Then /^I should see checksum$/ do
+  @dwc.checksum.should == '880775bd100f7b00c49ceefd2d7317daada99b26'
+end
+
 When /^I check core data$/ do
   @core = @dwc.core
 end
@@ -198,7 +202,7 @@ Then /^there are paths, synonyms and vernacular names in normalized classificati
 end
 
 Then /^names used in classification can be accessed by "([^"]*)" method$/ do |name_strings|
-  names = @cn.send(name_strings.to_sym) 
+  names = @cn.send(name_strings.to_sym)
   names.size.should > @normalized_classification.size
 end
 
@@ -216,7 +220,7 @@ Then /^nodes_ids organized in trees can be accessed by "([^"]*)" method$/ do |tr
   tree = @cn.send(tree.to_sym)
   tree.class.should == Hash
   keys = []
-  flatten_tree(tree, keys) 
+  flatten_tree(tree, keys)
   @normalized_classification.size.should == keys.size
 end
 

data/features/support/env.rb

@@ -1,4 +1,4 @@
 $LOAD_PATH.unshift(File.dirname(__FILE__) + '/../../lib')
 require 'dwc-archive'
 
-require 'spec/expectations'
+require 'rspec/expectations'

data/lib/dwc-archive.rb

@@ -1,6 +1,6 @@
 # encoding: UTF-8
 $:.unshift(File.dirname(__FILE__)) unless
-   $:.include?(File.dirname(__FILE__)) || $:.include?(File.expand_path(File.dirname(__FILE__)))   
+   $:.include?(File.dirname(__FILE__)) || $:.include?(File.expand_path(File.dirname(__FILE__)))
 R19 = RUBY_VERSION.split('.')[0..1].join('').to_i > 18
 require 'fileutils'
 require 'ostruct'
@@ -30,19 +30,19 @@ require 'dwc-archive/generator_eml_xml'
 require 'dwc-archive/classification_normalizer'
 
 class DarwinCore
-  
+
   VERSION = open(File.join(File.dirname(__FILE__), '..', 'VERSION')).readline.strip
 
   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)
     false
   end
-  
+
   def self.clean_all(tmp_dir = DEFAULT_TMP_DIR)
     Dir.entries(tmp_dir).each do |entry|
       path = File.join(tmp_dir, entry)
@@ -69,7 +69,8 @@ class DarwinCore
   end
 
   def initialize(dwc_path, tmp_dir = DEFAULT_TMP_DIR)
-    @archive = DarwinCore::Archive.new(dwc_path, tmp_dir) 
+    @dwc_path = dwc_path
+    @archive = DarwinCore::Archive.new(@dwc_path, tmp_dir)
     @core = DarwinCore::Core.new(self)
     @metadata = DarwinCore::Metadata.new(@archive)
     @extensions = get_extensions
@@ -86,6 +87,10 @@ class DarwinCore
     !!@core.fields.join('|').downcase.match(/highertaxonid|parentnameusageid/)
   end
 
+  def checksum
+    Digest::SHA1.hexdigest(open(@dwc_path).read)
+  end
+
   private
   def get_extensions
     res = []

data/lib/dwc-archive/ingester.rb

@@ -1,4 +1,4 @@
-class DarwinCore 
+class DarwinCore
   module Ingester
     attr_reader :data, :properties, :encoding, :fields_separator, :size
     attr_reader :file_path, :fields, :line_separator, :quote_character, :ignore_headers
@@ -20,7 +20,7 @@ class DarwinCore
         min_size > r.size ? errors << r : process_csv_row(res, errors, r)
         if (i + index_fix) % batch_size == 0
           DarwinCore.logger_write(@dwc.object_id, "Ingested %s records from %s" % [(i + index_fix), name])
-          if block_given? 
+          if block_given?
             yield [res, errors]
             res = []
             errors = []
@@ -30,7 +30,7 @@ class DarwinCore
       yield [res, errors] if block_given?
       [res, errors]
     end
-    
+
     private
     def name
       self.class.to_s.split('::')[-1].downcase
@@ -60,7 +60,7 @@ class DarwinCore
       @fields = get_fields
       raise exception("No data fields are found") if @fields.empty?
     end
-    
+
     def get_file_path
       file = @data[:location] || @data[:attributes][:location] || @data[:files][:location]
       File.join(@path, file)
@@ -78,7 +78,7 @@ class DarwinCore
     end
 
     def get_size
-      `wc -l #{@file_path}`.match(/^([\d]+)\s/)[1].to_i
+      `wc -l #{@file_path}`.match(/^\s*([\d]+)\s/)[1].to_i
     end
   end
 end

data/spec/spec_helper.rb

@@ -2,10 +2,10 @@ $LOAD_PATH.unshift(File.dirname(__FILE__))
 $LOAD_PATH.unshift(File.join(File.dirname(__FILE__), '..', 'lib'))
 require 'rubygems'
 require 'dwc-archive'
-require 'spec'
-require 'spec/autorun'
+require 'rspec'
+require 'rspec/autorun'
 require 'ruby-debug'
 
-Spec::Runner.configure do |config|
-  
+RSpec.configure do |config|
+
 end

metadata

@@ -1,12 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: dwc-archive
 version: !ruby/object:Gem::Version 
+  hash: 17
   prerelease: false
   segments: 
   - 0
   - 5
-  - 12
-  version: 0.5.12
+  - 13
+  version: 0.5.13
 platform: ruby
 authors: 
 - Dmitry Mozzherin
@@ -14,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-01-03 00:00:00 -05:00
+date: 2011-04-07 00:00:00 -04:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -25,6 +26,7 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
+        hash: 23
         segments: 
         - 0
         - 2
@@ -40,6 +42,7 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
+        hash: 13
         segments: 
         - 1
         - 2
@@ -55,6 +58,7 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
+        hash: 3
         segments: 
         - 0
         version: "0"
@@ -71,7 +75,6 @@ extra_rdoc_files:
 - README.rdoc
 files: 
 - .document
-- .gitignore
 - LICENSE
 - README.rdoc
 - Rakefile
@@ -119,8 +122,8 @@ homepage: http://github.com/GlobalNamesArchitecture/dwc-archive
 licenses: []
 
 post_install_message: 
-rdoc_options: 
-- --charset=UTF-8
+rdoc_options: []
+
 require_paths: 
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement 
@@ -128,6 +131,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
+      hash: 3
       segments: 
       - 0
       version: "0"
@@ -136,6 +140,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
+      hash: 3
       segments: 
       - 0
       version: "0"

data/.gitignore

@@ -1,24 +0,0 @@
-## MAC OS
-.DS_Store
-
-## TEXTMATE
-*.tmproj
-tmtags
-
-## EMACS
-*~
-\#*
-.\#*
-
-## VIM
-*.swp
-
-## PROJECT::GENERAL
-coverage
-rdoc
-pkg
-*.gemspec
-
-## PROJECT::SPECIFIC
-tags
-bin