dwc-archive 0.1.4 → 0.1.5

data/.gitignore

@@ -17,5 +17,6 @@ tmtags
 coverage
 rdoc
 pkg
+*.gemspec
 
 ## PROJECT::SPECIFIC

data/README.rdoc

@@ -13,14 +13,16 @@ Update to latest rubygems (v >= 1.3.6) which adds gemcutter sources by default.
     require 'rubygems'
     require 'dwc-archive'
 
-    dwc = DWCA::DarwinCore.new('/path_to_file/archive_file.tar.gz')
-    dwc.archive.files             # the archive file list
-    dwc.metadata.data             # summary of metadata from eml.xml if it exists
-    dwc.metadata.authors          # authors of the archive
-    dwc.core.data                 # summary of DarwinCore main file
-    dwc.core.file_path            # path to the DarwinCore main file
-    dwc.core.extensions           # array of DarwinCore Star extensions
-    dwc.core.extensions[0].data   # summary for an extension
+    dwc = DarwinCore.new('/path_to_file/archive_file.tar.gz')
+    dwc.archive.files      # the archive file list
+    dwc.metadata.data      # summary of metadata from eml.xml if it exists
+    dwc.metadata.authors   # authors of the archive
+    dwc.core.data          # summary of DarwinCore main file
+    dwc.core.file_path     # path to the DarwinCore main file
+    dwc.extensions         # array of DarwinCore Star extensions
+    dwc.extensions[0].data # summary for an extension
+
+    DarwinCore.clean_all   # remove all expanded archives
 
 == Note on Patches/Pull Requests
  

data/VERSION

@@ -1 +1 @@
-0.1.4
+0.1.5

data/features/dwc-archive.feature

@@ -12,6 +12,11 @@ Feature: Creation of a Darwing Core Archive
     When I delete expanded files
     Then they should disappear
 
+  Scenario: Instantiating DarwinCore with a file without "eml.xml"
+    Given path to a dwc file "minimal.tar.gz"
+    When I create a new DarwinCore instance
+    Then "DarwinCore_instance.metadata.data" should send instance of "NilClass" back
+
   Scenario: Instantiating DarwinCore with tar.gz file
     Given path to a dwc file "data.tar.gz"
     When I create a new DarwinCore instance 
@@ -32,3 +37,8 @@ Feature: Creation of a Darwing Core Archive
     Given path to a dwc file "data.zip"
     When I create a new DarwinCore instance 
     Then instance should have a valid archive
+
+  Scenario: Cleaning temporary directory from expanded archives
+    Given acces to DarwinCore gem
+    When I use DarwinCore.clean_all method
+    Then all temporary directories created by DarwinCore are deleted

data/features/step_definitions/dwc-archive_steps.rb

@@ -26,7 +26,12 @@ Then /^they should disappear$/ do
 end
 
 When /^I create a new DarwinCore instance$/ do
-  @dwc = DarwinCore.new(@dwca_file)
+  begin
+    @dwc = DarwinCore.new(@dwca_file)
+  rescue
+    @dwca_broken_file = @dwca_file
+    @dwc_error = $!
+  end
 end
 
 Then /^instance should have a valid archive$/ do
@@ -93,3 +98,29 @@ Then /^extension should have properties, data, file_path, coreid, fields$/ do
   ext.coreid.should == {:index=>0}
   ext.fields.should == [{:term=>"http://rs.gbif.org/ecat/terms/vernacularName", :index=>1}, {:term=>"http://rs.gbif.org/thesaurus/languageCode", :index=>2}]
 end
+
+Given /^acces to DarwinCore gem$/ do
+end
+
+When /^I use DarwinCore\.clean_all method$/ do
+  Dir.entries("/tmp").select {|e| e.match(/^dwc_/) }.size.should > 0
+  DarwinCore.clean_all
+end
+
+Then /^all temporary directories created by DarwinCore are deleted$/ do
+  Dir.entries("/tmp").select {|e| e.match(/^dwc_/) }.should == []
+end
+
+Then /^I receive "([^\"]*)" exception with "([^\"]*)" message$/ do |arg1, arg2|
+   @dwc_error.class.to_s.should == arg1
+   @dwc_error.message.should == arg2
+   @dwca_broken_file.should == @dwca_file
+   @dwc_error = nil
+   @dwca_broken_file = nil
+end
+
+Then /^"([^\"]*)" should send instance of "([^\"]*)" back$/ do |arg1, arg2|
+  res = eval(arg1.gsub(/DarwinCore_instance/, "@dwc"))
+  res.class.to_s.should == arg2
+end
+

data/lib/dwc-archive.rb

@@ -2,6 +2,8 @@
 $:.unshift(File.dirname(__FILE__)) unless
    $:.include?(File.dirname(__FILE__)) || $:.include?(File.expand_path(File.dirname(__FILE__)))   
 require 'ruby_extensions'
+require 'fileutils'
+require 'dwc-archive/errors'
 require 'dwc-archive/expander'
 require 'dwc-archive/archive'
 require 'dwc-archive/core'
@@ -11,12 +13,25 @@ require 'dwc-archive/metadata'
 class DarwinCore
   attr_reader :archive, :core, :metadata, :extensions
   alias :eml :metadata
-  def initialize(dwc_path, tmp_dir = "/tmp")
+  
+  DEFAULT_TMP_DIR = "/tmp"
+  
+  def initialize(dwc_path, tmp_dir = DEFAULT_TMP_DIR)
     @archive = DarwinCore::Archive.new(dwc_path, tmp_dir) 
     @core = DarwinCore::Core.new(@archive)
     @metadata = DarwinCore::Metadata.new(@archive)
     @extensions = get_extensions
   end
+
+  def self.clean_all(tmp_dir = DEFAULT_TMP_DIR)
+    Dir.entries(tmp_dir).each do |entry|
+      path = File.join(tmp_dir, entry)
+      if FileTest.directory?(path) && entry.match(/^dwc_[\d]+$/)
+        FileUtils.rm_rf(path)
+      end
+    end
+  end
+
   private
   def get_extensions
     res = []
@@ -24,6 +39,6 @@ class DarwinCore
     ext = @archive.meta[root_key][:extension]
     return [] unless ext
     ext = [ext] unless ext.class == Array
-    ext.map {|e| DarwinCore::Extension.new(@archive, e)}
+    ext.map { |e| DarwinCore::Extension.new(@archive, e) }
   end
 end

data/lib/dwc-archive/archive.rb

@@ -12,13 +12,13 @@ class DarwinCore
         @eml = files.include?("eml.xml") ? Hash.from_xml(open(File.join(@expander.path, 'eml.xml'))) : nil
       else
         clean
-        raise 'not a valid Darwin Core Archive File'
+        raise InvalidArchiveError
       end
     end
 
     def valid?
       valid = true
-      valid = valid && FileTest.exists?(@archive_path)
+      valid = valid && @expander.path && FileTest.exists?(@expander.path)
       valid = valid && files && files.include?('meta.xml')
     end
     

data/lib/dwc-archive/errors.rb

@@ -0,0 +1,6 @@
+class DarwinCore
+  class Error < RuntimeError; end
+  class FileNotFoundError < Error; end
+  class UnpackingError < Error; end
+  class InvalidArchiveError < Error; end
+end

data/lib/dwc-archive/expander.rb

@@ -9,7 +9,9 @@ class DarwinCore
 
     def unpack
       clean
-      @unpacker.call(@path, @archive_path) if @unpacker
+      raise FileNotFoundError unless File.exists?(@archive_path)
+      success = @unpacker.call(@path, @archive_path) if @unpacker
+      (@unpacker && success && $?.exitstatus == 0) ? success : (clean; raise UnpackingError)
     end
 
     def path
@@ -32,12 +34,12 @@ class DarwinCore
       if file_type.match(/tar.*gzip/i)
         return proc do |tmp_path, archive_path| 
           FileUtils.mkdir tmp_path
-          system "tar -zxvf #{archive_path} -C #{tmp_path}"
+          system("tar -zxf #{archive_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} #{archive_path}" }
+        return proc { |tmp_path, archive_path| system("unzip -qq -d #{tmp_path} #{archive_path} > /dev/null 2>&1") }
       end
       
       return nil

data/lib/dwc-archive/metadata.rb

@@ -10,11 +10,11 @@ class DarwinCore
     end
 
     def id
-      @metadata[:eml][:dataset][:attributes][:id]
+      @metadata[:eml][:dataset][:attributes][:id] rescue nil
     end
 
     def title
-      @metadata[:eml][:dataset][:title]
+      @metadata[:eml][:dataset][:title] rescue nil
     end
 
     def authors

data/spec/files/eml.xml

@@ -5,42 +5,42 @@ xmlns:eml="eml://ecoinformatics.org/eml-2.1.0"
 xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
 xsi:schemaLocation="eml://ecoinformatics.org/eml-2.1.0 eml.xsd">
 <dataset id="leptogastrinae:version:2.5">
-  <title>Leptogastrinae (Diptera: Asilidae) Classification</title>
-  <creator id="10" scope="document">
-    <individualName>
-      <givenName>Keith</givenName>
-      <surName>Bayless</surName>
-    </individualName>
-    <electronicMailAddress>keith.bayless@gmail.com</electronicMailAddress>
-  </creator>
-  <creator id="5" scope="document">
-    <individualName>
-      <givenName>Torsten</givenName>
-      <surName>Dikow</surName>
-    </individualName>
-    <electronicMailAddress>dshorthouse@eol.org</electronicMailAddress>
-  </creator>
-  <metadataProvider>
-    <organizationName>Encyclopedia of Life: LifeDesks (http://www.lifedesks.org)</organizationName>
-  </metadataProvider>
-  <pubDate>2010-02-03T01:09:41-05:00</pubDate>
-  <abstract>These are all the names in the Leptogastrinae classification.</abstract>
-  <intellectualRights>Creative Commons: publicdomain</intellectualRights>
-  <distribution>
-    <online>
-      <url function="download">http://leptogastrinae.lifedesks.org/files/leptogastrinae/classification_export/shared/leptogastrinae.tar.gz</url>
-    </online>
-  </distribution>
-  <contact>
-    <references>5</references>
-  </contact>
-  <publisher>
-    <organizationName>The Marine Biological Laboratory</organizationName>
-  </publisher>
+<title>Leptogastrinae (Diptera: Asilidae) Classification</title>
+<creator id="10" scope="document">
+<individualName>
+<givenName>Keith</givenName>
+<surName>Bayless</surName>
+</individualName>
+<electronicMailAddress>keith.bayless@gmail.com</electronicMailAddress>
+</creator>
+<creator id="5" scope="document">
+<individualName>
+<givenName>Torsten</givenName>
+<surName>Dikow</surName>
+</individualName>
+<electronicMailAddress>dshorthouse@eol.org</electronicMailAddress>
+</creator>
+<metadataProvider>
+<organizationName>Encyclopedia of Life: LifeDesks (http://www.lifedesks.org)</organizationName>
+</metadataProvider>
+<pubDate>2010-02-03T01:09:41-05:00</pubDate>
+<abstract>These are all the names in the Leptogastrinae classification.</abstract>
+<intellectualRights>Creative Commons: publicdomain</intellectualRights>
+<distribution>
+<online>
+<url function="download">http://leptogastrinae.lifedesks.org/files/leptogastrinae/classification_export/shared/leptogastrinae.tar.gz</url>
+</online>
+</distribution>
+<contact>
+<references>5</references>
+</contact>
+<publisher>
+<organizationName>The Marine Biological Laboratory</organizationName>
+</publisher>
 </dataset>
 <additionalMetadata>
-  <metadata>
-    <citation>Dikow, Torsten. 2010. The Leptogastrinae classification.</citation>
-  </metadata>
+<metadata>
+<citation>Dikow, Torsten. 2010. The Leptogastrinae classification.</citation>
+</metadata>
 </additionalMetadata>
-</eml:eml>
+</eml:eml>

data/spec/files/uncompressed

@@ -0,0 +1 @@
+an example of uncompressed file

data/spec/lib/dwc-archive_spec.rb

@@ -0,0 +1,33 @@
+require File.dirname(__FILE__) + "/../spec_helper"
+
+describe DarwinCore do
+  before(:all) do
+    @file_dir = File.join(File.dirname(__FILE__), '..', 'files')
+  end
+
+  describe ".new" do 
+
+    it "should create DarwinCore instance out of archive file" do
+      ['data.zip', 'data.tar.gz', 'minimal.tar.gz'].each do |file|
+        file = File.join(@file_dir, file)
+        dwc = DarwinCore.new(file)
+        dwc.archive.valid?.should be_true
+      end
+    end
+
+    it "should raise error if archive file does not exist" do
+      file = 'not_a_file'
+      lambda { DarwinCore.new(file) }.should raise_error(DarwinCore::FileNotFoundError)
+    end
+
+    it "should raise  error if archive is broken" do
+      file = File.join(@file_dir, 'broken.tar.gz')
+      lambda { DarwinCore.new(file) }.should raise_error(DarwinCore::UnpackingError)
+    end
+
+    it "should raise error if archive is invalid" do
+      file = File.join(@file_dir, 'invalid.tar.gz')
+      lambda { DarwinCore.new(file) }.should raise_error(DarwinCore::InvalidArchiveError)
+    end
+  end
+end

data/spec/lib/ruby_extenstions_spec.rb

@@ -2,7 +2,7 @@ require File.dirname(__FILE__) + "/../spec_helper"
 
 describe "Hash" do 
   it "should parse xml to hash" do
-    Hash.public_methods.include?("from_xml").should be_true
+    Hash.public_methods.map { |i| i.to_s }.include?("from_xml").should be_true
   end
 
   it "should parse xml" do

data/spec/spec_helper.rb

@@ -1,8 +1,10 @@
 $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 'ruby-debug'
 
 Spec::Runner.configure do |config|
   

metadata

@@ -5,8 +5,8 @@ version: !ruby/object:Gem::Version
   segments: 
   - 0
   - 1
-  - 4
-  version: 0.1.4
+  - 5
+  version: 0.1.5
 platform: ruby
 authors: 
 - Dmitry Mozzherin
@@ -14,7 +14,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2010-03-18 00:00:00 -04:00
+date: 2010-04-28 00:00:00 -04:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -59,7 +59,6 @@ files:
 - README.rdoc
 - Rakefile
 - VERSION
-- dwc-archive.gemspec
 - features/dwc-archive.feature
 - features/step_definitions/dwc-archive_steps.rb
 - features/support/env.rb
@@ -67,19 +66,23 @@ files:
 - lib/dwc-archive/.expander.rb.swo
 - lib/dwc-archive/archive.rb
 - lib/dwc-archive/core.rb
+- lib/dwc-archive/errors.rb
 - lib/dwc-archive/expander.rb
 - lib/dwc-archive/extension.rb
 - lib/dwc-archive/metadata.rb
 - lib/ruby_extensions.rb
-- spec/dwc-archive_spec.rb
+- spec/files/broken.tar.gz
 - spec/files/data.tar.gz
 - spec/files/data.zip
 - spec/files/eml.xml
+- spec/files/invalid.tar.gz
 - spec/files/meta.xml
+- spec/files/minimal.tar.gz
+- spec/files/uncompressed
+- spec/lib/dwc-archive_spec.rb
 - spec/lib/ruby_extenstions_spec.rb
 - spec/spec.opts
 - spec/spec_helper.rb
-- t
 has_rdoc: true
 homepage: http://github.com/dimus/dwc-archive
 licenses: []
@@ -111,6 +114,6 @@ signing_key:
 specification_version: 3
 summary: Handler of Darwin Core Archive files
 test_files: 
-- spec/dwc-archive_spec.rb
+- spec/lib/dwc-archive_spec.rb
 - spec/lib/ruby_extenstions_spec.rb
 - spec/spec_helper.rb

data/dwc-archive.gemspec

@@ -1,75 +0,0 @@
-# Generated by jeweler
-# DO NOT EDIT THIS FILE DIRECTLY
-# Instead, edit Jeweler::Tasks in Rakefile, and run the gemspec command
-# -*- encoding: utf-8 -*-
-
-Gem::Specification.new do |s|
-  s.name = %q{dwc-archive}
-  s.version = "0.1.4"
-
-  s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
-  s.authors = ["Dmitry Mozzherin"]
-  s.date = %q{2010-03-18}
-  s.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.}
-  s.email = %q{dmozzherin at gmail dot com}
-  s.extra_rdoc_files = [
-    "LICENSE",
-     "README.rdoc"
-  ]
-  s.files = [
-    ".document",
-     ".gitignore",
-     "LICENSE",
-     "README.rdoc",
-     "Rakefile",
-     "VERSION",
-     "dwc-archive.gemspec",
-     "features/dwc-archive.feature",
-     "features/step_definitions/dwc-archive_steps.rb",
-     "features/support/env.rb",
-     "lib/dwc-archive.rb",
-     "lib/dwc-archive/.expander.rb.swo",
-     "lib/dwc-archive/archive.rb",
-     "lib/dwc-archive/core.rb",
-     "lib/dwc-archive/expander.rb",
-     "lib/dwc-archive/extension.rb",
-     "lib/dwc-archive/metadata.rb",
-     "lib/ruby_extensions.rb",
-     "spec/dwc-archive_spec.rb",
-     "spec/files/data.tar.gz",
-     "spec/files/data.zip",
-     "spec/files/eml.xml",
-     "spec/files/meta.xml",
-     "spec/lib/ruby_extenstions_spec.rb",
-     "spec/spec.opts",
-     "spec/spec_helper.rb",
-     "t"
-  ]
-  s.homepage = %q{http://github.com/dimus/dwc-archive}
-  s.rdoc_options = ["--charset=UTF-8"]
-  s.require_paths = ["lib"]
-  s.rubygems_version = %q{1.3.6}
-  s.summary = %q{Handler of Darwin Core Archive files}
-  s.test_files = [
-    "spec/dwc-archive_spec.rb",
-     "spec/lib/ruby_extenstions_spec.rb",
-     "spec/spec_helper.rb"
-  ]
-
-  if s.respond_to? :specification_version then
-    current_version = Gem::Specification::CURRENT_SPECIFICATION_VERSION
-    s.specification_version = 3
-
-    if Gem::Version.new(Gem::RubyGemsVersion) >= Gem::Version.new('1.2.0') then
-      s.add_development_dependency(%q<rspec>, [">= 1.2.9"])
-      s.add_development_dependency(%q<cucumber>, [">= 0"])
-    else
-      s.add_dependency(%q<rspec>, [">= 1.2.9"])
-      s.add_dependency(%q<cucumber>, [">= 0"])
-    end
-  else
-    s.add_dependency(%q<rspec>, [">= 1.2.9"])
-    s.add_dependency(%q<cucumber>, [">= 0"])
-  end
-end
-

data/spec/dwc-archive_spec.rb

@@ -1,7 +0,0 @@
-require File.expand_path(File.dirname(__FILE__) + '/spec_helper')
-
-describe "DwcArchive" do
-  it "fails" do
-    fail "hey buddy, you should probably rename this file and start specing for real"
-  end
-end

data/t

@@ -1,74 +0,0 @@
-# Generated by jeweler
-# DO NOT EDIT THIS FILE DIRECTLY
-# Instead, edit Jeweler::Tasks in Rakefile, and run the gemspec command
-# -*- encoding: utf-8 -*-
-
-Gem::Specification.new do |s|
-  s.name = %q{dwc-archive}
-  s.version = "0.1.2"
-
-  s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
-  s.authors = ["Dmitry Mozzherin"]
-  s.date = %q{2010-03-18}
-  s.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.}
-  s.email = %q{dmozzherin at gmail dot com}
-  s.extra_rdoc_files = [
-    "LICENSE",
-     "README.rdoc"
-  ]
-  s.files = [
-    ".document",
-     ".gitignore",
-     "LICENSE",
-     "README.rdoc",
-     "Rakefile",
-     "VERSION",
-     "dwc-archive.gemspec",
-     "features/dwc-archive.feature",
-     "features/step_definitions/dwc-archive_steps.rb",
-     "features/support/env.rb",
-     "lib/dwc-archive.rb",
-     "lib/dwc-archive/.expander.rb.swo",
-     "lib/dwc-archive/archive.rb",
-     "lib/dwc-archive/core.rb",
-     "lib/dwc-archive/expander.rb",
-     "lib/dwc-archive/extension.rb",
-     "lib/dwc-archive/metadata.rb",
-     "lib/ruby_extensions.rb",
-     "spec/dwc-archive_spec.rb",
-     "spec/files/data.tar.gz",
-     "spec/files/data.zip",
-     "spec/files/eml.xml",
-     "spec/files/meta.xml",
-     "spec/lib/ruby_extenstions_spec.rb",
-     "spec/spec.opts",
-     "spec/spec_helper.rb"
-  ]
-  s.homepage = %q{http://github.com/dimus/dwc-archive}
-  s.rdoc_options = ["--charset=UTF-8"]
-  s.require_paths = ["lib"]
-  s.rubygems_version = %q{1.3.6}
-  s.summary = %q{Handler of Darwin Core Archive files}
-  s.test_files = [
-    "spec/dwc-archive_spec.rb",
-     "spec/lib/ruby_extenstions_spec.rb",
-     "spec/spec_helper.rb"
-  ]
-
-  if s.respond_to? :specification_version then
-    current_version = Gem::Specification::CURRENT_SPECIFICATION_VERSION
-    s.specification_version = 3
-
-    if Gem::Version.new(Gem::RubyGemsVersion) >= Gem::Version.new('1.2.0') then
-      s.add_development_dependency(%q<rspec>, [">= 1.2.9"])
-      s.add_development_dependency(%q<cucumber>, [">= 0"])
-    else
-      s.add_dependency(%q<rspec>, [">= 1.2.9"])
-      s.add_dependency(%q<cucumber>, [">= 0"])
-    end
-  else
-    s.add_dependency(%q<rspec>, [">= 1.2.9"])
-    s.add_dependency(%q<cucumber>, [">= 0"])
-  end
-end
-