athena 0.0.1.53 → 0.0.2.56

data/COPYING

@@ -1,4 +1,4 @@
-= License for ruby-nuggets
+= License for athena
   
                       GNU GENERAL PUBLIC LICENSE
                          Version 3, 29 June 2007

data/README

@@ -2,7 +2,7 @@
 
 == VERSION
 
-This documentation refers to athena version 0.0.1
+This documentation refers to athena version 0.0.2
 
 
 == DESCRIPTION

data/Rakefile

@@ -1,23 +1,21 @@
-require 'lib/athena/version'
+begin
+  require 'hen'
+rescue LoadError
+  abort "Please install the 'hen' gem first."
+end
 
-FILES = FileList['lib/**/*.rb'].to_a
-EXECS = FileList['bin/*'].to_a
-RDOCS = %w[README COPYING ChangeLog]
-OTHER = FileList['[A-Z]*', 'example/*'].to_a
+require 'lib/athena/version'
 
-task(:doc_spec) {{
-  :title      => 'athena Application documentation',
-  :rdoc_files => RDOCS + FILES
-}}
+Hen.lay! {{
+  :rubyforge => {
+    :package => 'athena'
+  },
 
-task(:gem_spec) {{
-  :name             => 'athena',
-  :version          => Athena::VERSION,
-  :summary          => 'Convert database files to various formats',
-  :files            => FILES + EXECS + OTHER,
-  :require_path     => 'lib',
-  :bindir           => 'bin',
-  :executables      => EXECS,
-  :extra_rdoc_files => RDOCS,
-  :dependencies     => %w[xmlstreamin ruby-nuggets]
+  :gem => {
+    :version      => Athena::VERSION,
+    :summary      => 'Convert database files to various formats.',
+    :files        => FileList['lib/**/*.rb', 'bin/*'].to_a,
+    :extra_files  => FileList['[A-Z]*', 'example/*'].to_a,
+    :dependencies     => %w[xmlstreamin ruby-nuggets]
+  }
 }}

data/bin/athena

@@ -39,7 +39,7 @@ USAGE = "Usage: #{$0} [-h|--help] [options]"
 abort USAGE if ARGV.empty?
 
 # Global variable to handle verbosity
-$_VERBOSE = {}
+$Verbose = {}
 
 options = {
   :config => 'config.yaml',
@@ -54,16 +54,16 @@ OptionParser.new { |opts|
   opts.separator ''
   opts.separator 'Options:'
 
-  opts.on('-c', '--config YAML', "Config file [Default: #{options[:config]}#{' (currently not present)' unless File.readable? options[:config]}]") { |f|
-    abort "Can't find config file: #{f}." unless File.readable? f
+  opts.on('-c', '--config YAML', "Config file [Default: #{options[:config]}#{' (currently not present)' unless File.readable?(options[:config])}]") { |f|
+    abort "Can't find config file: #{f}." unless File.readable?(f)
 
     options[:config] = f
   }
 
   opts.separator ''
 
-  opts.on('-i', '--input FILE',  "Input file [Default: #{options[:input] == STDIN ? 'STDIN' : options[:input]}]") { |f|
-    abort "Can't find input file: #{f}." unless File.readable? f
+  opts.on('-i', '--input FILE',  "Input file [Default: STDIN]") { |f|
+    abort "Can't find input file: #{f}." unless File.readable?(f)
 
     options[:input] = File.open(f, 'r')
 
@@ -78,8 +78,11 @@ OptionParser.new { |opts|
 
   opts.on('-L', '--list-specs', "List available input formats (specs) and exit") {
     puts "Available input formats (specs):"
-    Athena.input_formats.each { |s, k|
-      puts "  - #{s} = #{k}"
+
+    formats = Athena.input_formats
+    max     = formats.map { |a, _| a.length }.max
+    formats.each { |f, k|
+      puts "  - %-#{max}s = %s" % [f, k]
     }
 
     exit 0
@@ -87,7 +90,7 @@ OptionParser.new { |opts|
 
   opts.separator ''
 
-  opts.on('-o', '--output FILE',  "Output file [Default: #{options[:output] == STDOUT ? 'STDOUT' : options[:output]}]") { |f|
+  opts.on('-o', '--output FILE',  "Output file [Default: STDOUT]") { |f|
     options[:output] = File.open(f, 'w')
 
     options[:format_fallback] = f.split('.').last.downcase
@@ -99,8 +102,11 @@ OptionParser.new { |opts|
 
   opts.on('-l', '--list-formats', "List available output formats and exit") {
     puts "Available output formats:"
-    Athena.output_formats.each { |f, k|
-      puts "  - #{f} = #{k}"
+
+    formats = Athena.output_formats
+    max     = formats.map { |a, _| a.length }.max
+    formats.each { |f, k|
+      puts "  - %-#{max}s = %s" % [f, k]
     }
 
     exit 0
@@ -117,10 +123,10 @@ OptionParser.new { |opts|
 
   opts.on('-v', '--verbose [WHAT]', "Be verbose about what's being done. Optional argument is a comma-separated", "list of what should be output, or 'all' [Default: 'all']") { |what|
     if what.nil? || what == 'all'
-      $_VERBOSE.default = true
+      $Verbose.default = true
     else
       what.split(',').each { |w|
-        $_VERBOSE[w.to_sym] = true
+        $Verbose[w.to_sym] = true
       }
     end
   }
@@ -136,11 +142,11 @@ OptionParser.new { |opts|
 
 spec = options[:spec] || options[:spec_fallback]
 abort "No input format (spec) specified and none could be inferred." unless spec
-abort "Invalid input format (spec): #{spec}. Use '-L' to get a list of available specs." unless Athena.valid_input_format? spec
+abort "Invalid input format (spec): #{spec}. Use '-L' to get a list of available specs." unless Athena.valid_input_format?(spec)
 
 format = options[:format] || options[:format_fallback]
 abort "No output format specified and none could be inferred." unless format
-abort "Invalid output format: #{format}. Use '-l' to get a list of available formats." unless Athena.valid_output_format? format
+abort "Invalid output format: #{format}. Use '-l' to get a list of available formats." unless Athena.valid_output_format?(format)
 
 yaml = YAML.load_file(options[:config])
 if t = options[:target]
@@ -156,9 +162,21 @@ else
 end
 abort "Config not found for target: #{target}." unless config
 
-records = Athena.parser(config, spec).parse(options[:input]) { |record|
-  options[:output].puts record.to(format)
-}
+parser = Athena.parser(config, spec)
+
+if Athena.deferred_output?(format)
+  records = parser.parse(options[:input])
+
+  records.map { |record|
+    record.to(format)
+  }.flatten.sort.uniq.each { |line|
+    options[:output].puts line
+  }
+else
+  records = parser.parse(options[:input]) { |record|
+    options[:output].puts record.to(format)
+  }
+end
 
 Athena::Util.verbose(:count) do
   spit records.size

data/lib/athena.rb

@@ -65,4 +65,8 @@ module Athena
     Formats.valid_format?(:out, format)
   end
 
+  def deferred_output?(format)
+    Formats[:out, format].deferred?
+  end
+
 end

data/lib/athena/formats.rb

@@ -35,7 +35,7 @@ module Athena
     class << self
 
       def formats
-        Formats.instance_variable_get :@formats
+        Formats.instance_variable_get(:@formats)
       end
 
       def [](direction, format)
@@ -43,7 +43,11 @@ module Athena
       end
 
       def valid_format?(direction, format)
-        formats[direction].has_key? format
+        formats[direction].has_key?(format)
+      end
+
+      def deferred?
+        false
       end
 
       def convert(*args)
@@ -53,6 +57,11 @@ module Athena
       private
 
       def register_format(direction, format)
+        if existing = formats[direction][format]
+          raise DuplicateFormatDefinitionError,
+            "format already defined (#{direction}): #{format} = #{existing}"
+        end
+
         formats[direction][format] = self
       end
 
@@ -68,6 +77,12 @@ module Athena
       raise NotImplementedError, 'must be defined by sub-class'
     end
 
+    class DuplicateFormatDefinitionError < StandardError
+    end
+
+    class FormatArgumentError < ArgumentError
+    end
+
   end
   
 end

data/lib/athena/formats/dbm.rb

@@ -40,20 +40,25 @@ module Athena
 
       ICONV_TO_LATIN1 = Iconv.new('latin1', 'utf-8')
 
+      VALUE_SEPARATOR  = '|'
+      RECORD_SEPARATOR = '&&&'
+
       def self.convert(record)
         dbm = ["ID:#{record.id}"]
+
         record.struct.each { |field, struct|
           strings = struct[:elements].inject([]) { |array, element|
             values = (struct[:values][element] || []).map { |v|
               (v || '').strip.gsub(/(?:\r?\n)+/, ' ')
             }.reject { |v| v.empty? }
 
-            array << (values.empty? ? struct[:empty] : values.join('|'))
+            array << (values.empty? ? struct[:empty] : values.join(VALUE_SEPARATOR))
           }
 
           dbm << "#{field.to_s.upcase}:#{ICONV_TO_LATIN1.iconv(struct[:string] % strings)}"
         }
-        dbm << '&&&'
+
+        dbm << RECORD_SEPARATOR
 
         dbm.join(CRLF) << CRLF << CRLF
       end

data/lib/athena/formats/lingo.rb

@@ -0,0 +1,145 @@
+#--
+###############################################################################
+#                                                                             #
+# A component of athena, the database file converter.                         #
+#                                                                             #
+# Copyright (C) 2007 University of Cologne,                                   #
+#                    Albertus-Magnus-Platz,                                   #
+#                    50932 Cologne, Germany                                   #
+#                                                                             #
+# Authors:                                                                    #
+#     Jens Wille <jens.wille@uni-koeln.de>                                    #
+#                                                                             #
+# athena is free software; you can redistribute it and/or modify it under the #
+# terms of the GNU General Public License as published by the Free Software   #
+# Foundation; either version 3 of the License, or (at your option) any later  #
+# version.                                                                    #
+#                                                                             #
+# athena is distributed in the hope that it will be useful, but WITHOUT ANY   #
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS   #
+# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more       #
+# details.                                                                    #
+#                                                                             #
+# You should have received a copy of the GNU General Public License along     #
+# with athena. If not, see <http://www.gnu.org/licenses/>.                    #
+#                                                                             #
+###############################################################################
+#++
+
+require 'iconv'
+require 'enumerator'
+
+module Athena
+
+  class Formats
+
+    module Lingo
+
+      class Base < Athena::Formats
+
+        class << self
+
+          def convert(record)
+            record.struct.inject([]) { |terms, (field, struct)|
+              terms << struct[:elements].inject([]) { |array, element|
+                array += (struct[:values][element] || []).map { |v|
+                  (v || '').strip.gsub(/(?:\r?\n)+/, ' ')
+                }.reject { |v| v.empty? }
+              }
+            }
+          end
+
+          def deferred?
+            true
+          end
+
+          private
+
+          def check_number_of_arguments(expected, actual, blow = false, &block)
+            return true if block ? block[actual] : expected == actual
+
+            msg = "wrong number of arguments for #{self} (#{actual} for #{expected})"
+
+            if blow
+              raise FormatArgumentError, msg
+            else
+              warn msg
+              return false
+            end
+          end
+
+          def check_number_of_arguments!(expected, actual, &block)
+            check_number_of_arguments(expected, actual, true, &block)
+          end
+
+        end
+
+      end
+
+      # "Nasenbär\n"
+      class SingleWord < Athena::Formats::Lingo::Base
+
+        register_formats :out, 'lingo/single_word'
+
+        def self.convert(record)
+          super.flatten
+        end
+
+      end
+
+      # "John Vorhauer*Vorhauer, John\n"
+      class KeyValue < Athena::Formats::Lingo::Base
+
+        register_formats :out, 'lingo/key_value'
+
+        def self.convert(record)
+          super.map { |terms|
+            next unless check_number_of_arguments(2, terms.size)
+
+            terms.join('*')
+          }.compact
+        end
+
+      end
+
+      # "Essen,essen #v Essen #s Esse #s\n"
+      class WordClass < Athena::Formats::Lingo::Base
+
+        register_formats :out, 'lingo/word_class'
+
+        def self.convert(record)
+          super.map { |terms|
+            next unless check_number_of_arguments('odd, > 1', terms.size) { |actual|
+              actual > 1 && actual % 2 == 1
+            }
+
+            [terms.shift, terms.to_enum(:each_slice, 2).map { |form, wc|
+              "#{form} ##{wc}"
+            }.join(' ')].join(',')
+          }.compact
+        end
+
+      end
+
+      # "Fax;Faxkopie;Telefax\n"
+      class MultiValue < Athena::Formats::Lingo::Base
+
+        register_formats :out, 'lingo/multi_value', 'lingo/multi_key'
+
+        def self.convert(record)
+          super.map { |terms|
+            next unless check_number_of_arguments('> 1', terms.size) { |actual|
+              actual > 1
+            }
+
+            terms.join(';')
+          }.compact
+        end
+
+      end
+
+    end
+
+  end
+
+end

data/lib/athena/formats/xml.rb

@@ -187,12 +187,12 @@ module Athena
         private
 
         def level
-          BaseSpec.instance_variable_get :@level
+          BaseSpec.instance_variable_get(:@level)
         end
 
         def step(direction)
           steps = { :down => 1, :up => -1 }
-          BaseSpec.instance_variable_set :@level, level + steps[direction]
+          BaseSpec.instance_variable_set(:@level, level + steps[direction])
         end
 
       end
@@ -244,7 +244,7 @@ module Athena
         def text(context, data)
           super
 
-          record.update name, data
+          record.update(name, data)
         end
 
       end
@@ -260,7 +260,7 @@ module Athena
           super()
 
           @parent = parent
-          default! self
+          default!(self)
         end
 
       end

data/lib/athena/parser.rb

@@ -32,6 +32,9 @@ module Athena
 
     include Util
 
+    DEFAULT_SEPARATOR = ', '
+    DEFAULT_EMPTY     = '<<EMPTY>>'
+
     attr_reader   :config, :spec
     attr_accessor :block
 
@@ -61,12 +64,13 @@ module Athena
             when Hash
               elements = v[:elements] || v[:element].to_a
 
-              raise ArgumentError, "no elements specified for field #{field}" unless elements.is_a?(Array)
+              raise ArgumentError, "no elements specified for field #{field}" \
+                unless elements.is_a?(Array)
             else
               raise ArgumentError, "illegal value for field #{field}"
           end
 
-          separator = v[:separator] || ', '
+          separator = v[:separator] || DEFAULT_SEPARATOR
 
           elements.each { |element|
             verbose(:config) do
@@ -75,7 +79,7 @@ module Athena
 
             (hash[element] ||= {})[field] = {
               :string   => v[:string] || ['%s'] * elements.size * separator,
-              :empty    => v[:empty]  || '<<EMPTY>>',
+              :empty    => v[:empty]  || DEFAULT_EMPTY,
               :elements => elements
             }
           }

data/lib/athena/record.rb

@@ -53,11 +53,11 @@ module Athena
     attr_reader :struct, :block, :id
 
     def initialize(block, id = object_id.abs)
-      self.class.records << self
-
       @struct = {}
       @block  = block
       @id     = id
+
+      add_record
     end
 
     def fill(field, config)
@@ -89,6 +89,12 @@ module Athena
       Athena::Formats[:out, format].convert(self)
     end
 
+    private
+
+    def add_record
+      self.class.records << self
+    end
+
     class NoRecordError < StandardError
     end
 

data/lib/athena/util.rb

@@ -33,7 +33,7 @@ module Athena
     extend self
 
     def verbose(what, klass = self.class, &block)
-      if $_VERBOSE[what]
+      if $Verbose[what]
         klass.send(:define_method, :spit) { |msg|
           warn "*#{what}: #{msg}"
         }

data/lib/athena/version.rb

@@ -32,7 +32,7 @@ module Athena
 
     MAJOR = 0
     MINOR = 0
-    TINY  = 1
+    TINY  = 2
 
     class << self
 

metadata

@@ -1,43 +1,56 @@
 --- !ruby/object:Gem::Specification 
-rubygems_version: 0.9.4
-specification_version: 1
 name: athena
 version: !ruby/object:Gem::Version 
-  version: 0.0.1.53
-date: 2007-10-11 00:00:00 +02:00
-summary: Convert database files to various formats
-require_paths: 
-- lib
-email: jens.wille@uni-koeln.de
-homepage: 
-rubyforge_project: 
-description: 
-autorequire: 
-default_executable: 
-bindir: bin
-has_rdoc: true
-required_ruby_version: !ruby/object:Gem::Version::Requirement 
-  requirements: 
-  - - ">"
-    - !ruby/object:Gem::Version 
-      version: 0.0.0
-  version: 
+  version: 0.0.2.56
 platform: ruby
-signing_key: 
-cert_chain: 
-post_install_message: 
 authors: 
 - Jens Wille
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2008-01-08 00:00:00 +01:00
+default_executable: 
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: xmlstreamin
+  version_requirement: 
+  version_requirements: !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "0"
+    version: 
+- !ruby/object:Gem::Dependency 
+  name: ruby-nuggets
+  version_requirement: 
+  version_requirements: !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "0"
+    version: 
+description: Convert database files to various formats.
+email: jens.wille@uni-koeln.de
+executables: 
+- athena
+extensions: []
+
+extra_rdoc_files: 
+- README
+- COPYING
+- ChangeLog
 files: 
-- lib/athena.rb
 - lib/athena/formats.rb
 - lib/athena/version.rb
 - lib/athena/util.rb
-- lib/athena/record.rb
-- lib/athena/parser.rb
 - lib/athena/formats/sisis.rb
 - lib/athena/formats/xml.rb
 - lib/athena/formats/dbm.rb
+- lib/athena/formats/lingo.rb
+- lib/athena/record.rb
+- lib/athena/parser.rb
+- lib/athena.rb
 - bin/athena
 - COPYING
 - README
@@ -46,36 +59,39 @@ files:
 - example/sisis-ex.txt
 - example/config.yaml
 - example/example.xml
-test_files: []
-
-rdoc_options: []
-
-extra_rdoc_files: 
+has_rdoc: true
+homepage: http://prometheus.rubyforge.org/athena
+post_install_message: 
+rdoc_options: 
+- --all
+- --inline-source
+- --charset
+- UTF-8
+- --main
 - README
-- COPYING
-- ChangeLog
-executables: 
-- athena
-extensions: []
-
+- --title
+- athena Application documentation
+- --line-numbers
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
 requirements: []
 
-dependencies: 
-- !ruby/object:Gem::Dependency 
-  name: xmlstreamin
-  version_requirement: 
-  version_requirements: !ruby/object:Gem::Version::Requirement 
-    requirements: 
-    - - ">"
-      - !ruby/object:Gem::Version 
-        version: 0.0.0
-    version: 
-- !ruby/object:Gem::Dependency 
-  name: ruby-nuggets
-  version_requirement: 
-  version_requirements: !ruby/object:Gem::Version::Requirement 
-    requirements: 
-    - - ">"
-      - !ruby/object:Gem::Version 
-        version: 0.0.0
-    version: 
+rubyforge_project: prometheus
+rubygems_version: 1.0.1
+signing_key: 
+specification_version: 2
+summary: Convert database files to various formats.
+test_files: []
+