remote_table 1.1.2 → 1.1.3

data/lib/remote_table.rb

@@ -16,7 +16,6 @@ class RemoteTable
   autoload :Transformer, 'remote_table/transformer'
 
   # singletons
-  autoload :Cleaner, 'remote_table/cleaner'
   autoload :Executor, 'remote_table/executor'
   autoload :Hasher, 'remote_table/hasher'
   
@@ -52,7 +51,6 @@ class RemoteTable
     end
     @url.freeze
     @options.freeze
-    at_exit { ::RemoteTable.cleaner.cleanup }
   end
   
   def each(&blk)
@@ -85,12 +83,7 @@ class RemoteTable
     @to_a.freeze
   end
   alias :rows :to_a
-  
-  # Used internally as a sort of garbage collector.
-  def self.cleaner
-    Cleaner.instance
-  end
-  
+    
   # Used internally to execute stuff in shells.
   def self.executor
     Executor.instance

data/lib/remote_table/executor.rb

@@ -1,27 +1,30 @@
 require 'singleton'
 require 'escape'
 require 'fileutils'
+require 'posix/spawn'
+
 class RemoteTable
   class Executor
     include ::Singleton
     def bang(path, cmd)
       tmp_path = "#{path}.bang.#{rand}"
-      backtick_with_reporting "/bin/cat #{::Escape.shell_single_word path} | #{cmd} > #{::Escape.shell_single_word tmp_path}"
+      backtick_with_reporting "cat #{::Escape.shell_single_word path} | #{cmd} > #{::Escape.shell_single_word tmp_path}"
       ::FileUtils.mv tmp_path, path
     end
 
-    def backtick_with_reporting(cmd)
-      cmd = cmd.gsub /[ ]*\n[ ]*/m, ' '
-      output = `#{cmd}`
-      if not $?.success?
+    def backtick_with_reporting(cmd, raise_on_error = false)
+      cmd = cmd.gsub /\n/m, ' '
+      pid = ::POSIX::Spawn.spawn({ 'PATH' => '/usr/local/bin:/usr/bin:/bin' }, cmd)
+      stat = ::Process::waitpid pid
+      if raise_on_error and not stat.success?
         raise %{
 From the remote_table gem...
 
 Command failed:
 #{cmd}
 
-Output:
-#{output}
+Exit code:
+#{stat.exitstatus}
 }
       end
     end

data/lib/remote_table/format.rb

@@ -20,14 +20,9 @@ class RemoteTable
     def each
       raise "must be defined by format"
     end
-    
-    def backup_file!
-      ::FileUtils.cp t.local_file.path, "#{t.local_file.path}.backup"
-    end
-    
-    def restore_file!
-      return unless ::File.readable? "#{t.local_file.path}.backup"
-      ::FileUtils.mv "#{t.local_file.path}.backup", t.local_file.path
+        
+    def delete_file!
+      ::FileUtils.rm_rf t.properties.staging_dir_path
     end
   end
 end

data/lib/remote_table/format/delimited.rb

@@ -15,7 +15,6 @@ class RemoteTable
     class Delimited < Format
       include Textual
       def each(&blk)
-        backup_file!
         convert_file_to_utf8!
         remove_useless_characters!
         skip_rows!
@@ -42,7 +41,7 @@ class RemoteTable
           yield ordered_hash if t.properties.keep_blank_rows or filled_values > 0
         end
       ensure
-        restore_file!
+        delete_file!
       end
 
       private

data/lib/remote_table/format/fixed_width.rb

@@ -4,7 +4,6 @@ class RemoteTable
     class FixedWidth < Format
       include Textual
       def each(&blk)
-        backup_file!
         convert_file_to_utf8!
         remove_useless_characters!
         crop_rows!
@@ -15,7 +14,7 @@ class RemoteTable
           yield hash if t.properties.keep_blank_rows or hash.any? { |k, v| v.present? }
         end
       ensure
-        restore_file!
+        delete_file!
       end
       private
       def parser

data/lib/remote_table/format/html.rb

@@ -5,7 +5,6 @@ class RemoteTable
     class HTML < Format
       include Textual
       def each(&blk)
-        backup_file!
         convert_file_to_utf8!
         remove_useless_characters!
         html_headers = (t.properties.headers.is_a?(::Array)) ? t.properties.headers : nil
@@ -19,7 +18,7 @@ class RemoteTable
           yield hash if t.properties.keep_blank_rows or hash.any? { |k, v| v.present? }
         end
       ensure
-        restore_file!
+        delete_file!
       end
 
       private

data/lib/remote_table/local_file.rb

@@ -19,7 +19,7 @@ class RemoteTable
     
     def save_locally
       return if @path.is_a?(::String)
-      @path = ::File.join(staging_dir_path, ::File.basename(t.properties.uri.path))
+      @path = ::File.join(t.properties.staging_dir_path, ::File.basename(t.properties.uri.path))
       download
       decompress
       unpack
@@ -35,6 +35,7 @@ class RemoteTable
         ::RemoteTable.executor.backtick_with_reporting %{
           curl
           --silent
+          --show-error
           --location
           --header "Expect: "
           #{"--data #{::Escape.shell_single_word t.properties.form_data}" if t.properties.form_data.present?}
@@ -48,18 +49,20 @@ class RemoteTable
     def decompress
       return unless t.properties.compression
       new_path = @path.chomp ".#{t.properties.compression}"
+      raise_on_error = true
       cmd = case t.properties.compression
       when 'zip', 'exe'
-        "unzip #{::Escape.shell_single_word @path} -d #{::File.dirname(@path)}"
         # can't set path yet because there may be multiple files
+        raise_on_error = false
+        "unzip -qq -n #{::Escape.shell_single_word @path} -d #{::File.dirname(@path)}"
       when 'bz2'
-        "bunzip2 --stdout #{::Escape.shell_single_word @path} > #{::Escape.shell_single_word new_path}"
         @path = new_path
+        "bunzip2 --stdout #{::Escape.shell_single_word @path} > #{::Escape.shell_single_word new_path}"
       when 'gz'
-        "gunzip --stdout #{::Escape.shell_single_word @path} > #{::Escape.shell_single_word new_path}"
         @path = new_path
+        "gunzip --stdout #{::Escape.shell_single_word @path} > #{::Escape.shell_single_word new_path}"
       end
-      ::RemoteTable.executor.backtick_with_reporting cmd
+      ::RemoteTable.executor.backtick_with_reporting cmd, raise_on_error
     end
     
     def unpack
@@ -81,13 +84,5 @@ class RemoteTable
         @path = ::Dir[::File.dirname(@path)+t.properties.glob].first
       end
     end
-    
-    def staging_dir_path    
-      return @staging_dir_path if @staging_dir_path.is_a?(::String)
-      @staging_dir_path = ::File.join ::Dir.tmpdir, 'remote_table_gem', rand.to_s
-      ::FileUtils.mkdir_p @staging_dir_path
-      ::RemoteTable.cleaner.remove_at_exit @staging_dir_path
-      @staging_dir_path
-    end
   end
 end

data/lib/remote_table/properties.rb

@@ -209,5 +209,12 @@ class RemoteTable
         Format::Delimited
       end
     end
+    
+    def staging_dir_path #:nodoc:
+      return @staging_dir_path if @staging_dir_path.is_a?(::String)
+      @staging_dir_path = ::File.join ::Dir.tmpdir, 'remote_table_gem', rand.to_s
+      ::FileUtils.mkdir_p @staging_dir_path
+      @staging_dir_path
+    end
   end
 end

data/lib/remote_table/version.rb

@@ -1,3 +1,3 @@
 class RemoteTable
-  VERSION = "1.1.2"
+  VERSION = "1.1.3"
 end

data/remote_table.gemspec

@@ -29,6 +29,7 @@ Gem::Specification.new do |s|
   s.add_dependency 'spreadsheet' #roo
   s.add_dependency 'google-spreadsheet-ruby' #roo
   s.add_dependency 'escape', '>=0.0.4'
+  s.add_dependency 'posix-spawn'
   unless RUBY_VERSION >= '1.9'
     s.add_dependency 'fastercsv', '>=1.5.0'
   end

data/test/test_remote_table.rb

@@ -35,4 +35,12 @@ class TestRemoteTable < Test::Unit::TestCase
     assert_equal %{Body example with a <a href="">link</a>}, t[0][2]
     f.close
   end
+  
+  should "open a csv inside a zip file" do
+    t = RemoteTable.new  'http://www.epa.gov/climatechange/emissions/downloads10/2010-Inventory-Annex-Tables.zip',
+                         :filename => 'Annex Tables/Annex 3/Table A-93.csv',
+                         :skip => 1,
+                         :select => lambda { |row| row['Vehicle Age'].to_i.to_s == row['Vehicle Age'] }
+    assert_equal '9.09%', t[0]['LDGV']
+  end
 end

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: remote_table
 version: !ruby/object:Gem::Version 
-  hash: 23
+  hash: 21
   prerelease: 
   segments: 
   - 1
   - 1
-  - 2
-  version: 1.1.2
+  - 3
+  version: 1.1.3
 platform: ruby
 authors: 
 - Seamus Abshere
@@ -16,7 +16,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-03-24 00:00:00 -05:00
+date: 2011-03-29 00:00:00 -05:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -169,9 +169,23 @@ dependencies:
   type: :runtime
   version_requirements: *id010
 - !ruby/object:Gem::Dependency 
-  name: fastercsv
+  name: posix-spawn
   prerelease: false
   requirement: &id011 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        hash: 3
+        segments: 
+        - 0
+        version: "0"
+  type: :runtime
+  version_requirements: *id011
+- !ruby/object:Gem::Dependency 
+  name: fastercsv
+  prerelease: false
+  requirement: &id012 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
@@ -183,11 +197,11 @@ dependencies:
         - 0
         version: 1.5.0
   type: :runtime
-  version_requirements: *id011
+  version_requirements: *id012
 - !ruby/object:Gem::Dependency 
   name: errata
   prerelease: false
-  requirement: &id012 !ruby/object:Gem::Requirement 
+  requirement: &id013 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
@@ -199,11 +213,11 @@ dependencies:
         - 0
         version: 0.2.0
   type: :development
-  version_requirements: *id012
+  version_requirements: *id013
 - !ruby/object:Gem::Dependency 
   name: test-unit
   prerelease: false
-  requirement: &id013 !ruby/object:Gem::Requirement 
+  requirement: &id014 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
@@ -213,11 +227,11 @@ dependencies:
         - 0
         version: "0"
   type: :development
-  version_requirements: *id013
+  version_requirements: *id014
 - !ruby/object:Gem::Dependency 
   name: shoulda
   prerelease: false
-  requirement: &id014 !ruby/object:Gem::Requirement 
+  requirement: &id015 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
@@ -227,11 +241,11 @@ dependencies:
         - 0
         version: "0"
   type: :development
-  version_requirements: *id014
+  version_requirements: *id015
 - !ruby/object:Gem::Dependency 
   name: ruby-debug
   prerelease: false
-  requirement: &id015 !ruby/object:Gem::Requirement 
+  requirement: &id016 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
@@ -241,7 +255,7 @@ dependencies:
         - 0
         version: "0"
   type: :development
-  version_requirements: *id015
+  version_requirements: *id016
 description: Gives you a standard way to parse various formats and treat them as an array of hashes.
 email: 
 - seamus@abshere.net
@@ -260,7 +274,6 @@ files:
 - README.rdoc
 - Rakefile
 - lib/remote_table.rb
-- lib/remote_table/cleaner.rb
 - lib/remote_table/executor.rb
 - lib/remote_table/format.rb
 - lib/remote_table/format/delimited.rb

data/lib/remote_table/cleaner.rb

@@ -1,19 +0,0 @@
-require 'singleton'
-require 'fileutils'
-class RemoteTable
-  class Cleaner
-    include ::Singleton
-    def paths_for_removal
-      @paths_for_removal ||= []
-    end
-    def cleanup
-      paths_for_removal.each do |path|
-        ::FileUtils.rm_rf path
-        paths_for_removal.delete path
-      end
-    end
-    def remove_at_exit(path)
-      paths_for_removal << path
-    end
-  end
-end