brainmap-anonymizer 0.1.0

data/Manifest

@@ -0,0 +1,13 @@
+anonymizer.gemspec
+bin/anonymize_data.rb
+lib/anonymizer.rb
+Manifest
+Rakefile
+README.rdoc
+test/out_2/0/ddd.txt
+test/out_2/1.txt
+test/out_2/2/bbb.txt
+test/out_2/3/ccc.txt
+test/out_2/4/eee.txt
+test/out_2/5/aaa.txt
+test/out_2/keymap.csv

data/Rakefile

@@ -0,0 +1,12 @@
+require 'rubygems'
+require 'rake'
+require 'echoe'
+
+Echoe.new('anonymizer', '0.1.0') do |p|
+  p.description    = "Anonymizes a collection of files and/or directories for interrater reliabiliy analyses and the like."
+  p.url            = "http://github.com/kjkosmatka/anonymizer"
+  p.author         = "Kristopher J. Kosmatka"
+  p.email          = "kk4@medicine.wisc.edu"
+  p.ignore_pattern = ["test/out*"]
+  p.development_dependencies = []
+end

data/anonymizer.gemspec

@@ -0,0 +1,33 @@
+# -*- encoding: utf-8 -*-
+
+Gem::Specification.new do |s|
+  s.name = %q{anonymizer}
+  s.version = "0.1.0"
+
+  s.required_rubygems_version = Gem::Requirement.new(">= 1.2") if s.respond_to? :required_rubygems_version=
+  s.authors = ["Kristopher J. Kosmatka"]
+  s.date = %q{2009-08-13}
+  s.default_executable = %q{anonymize_data.rb}
+  s.description = %q{Anonymizes a collection of files and/or directories for interrater reliabiliy analyses and the like.}
+  s.email = %q{kk4@medicine.wisc.edu}
+  s.executables = ["anonymize_data.rb"]
+  s.extra_rdoc_files = ["bin/anonymize_data.rb", "lib/anonymizer.rb", "README.rdoc"]
+  s.files = ["anonymizer.gemspec", "bin/anonymize_data.rb", "lib/anonymizer.rb", "Manifest", "Rakefile", "README.rdoc"]
+  s.has_rdoc = true
+  s.homepage = %q{http://github.com/kjkosmatka/anonymizer}
+  s.rdoc_options = ["--line-numbers", "--inline-source", "--title", "Anonymizer", "--main", "README.rdoc"]
+  s.require_paths = ["lib"]
+  s.rubyforge_project = %q{anonymizer}
+  s.rubygems_version = %q{1.3.1}
+  s.summary = %q{Anonymizes a collection of files and/or directories for interrater reliabiliy analyses and the like.}
+
+  if s.respond_to? :specification_version then
+    current_version = Gem::Specification::CURRENT_SPECIFICATION_VERSION
+    s.specification_version = 2
+
+    if Gem::Version.new(Gem::RubyGemsVersion) >= Gem::Version.new('1.2.0') then
+    else
+    end
+  else
+  end
+end

data/bin/anonymize_data.rb

@@ -0,0 +1,51 @@
+#!/usr/bin/env ruby
+$:.unshift File.join(File.dirname(__FILE__),'..','lib')
+
+require 'optparse'
+require 'anonymizer'
+
+options = Hash.new
+options[:collection] = Array.new
+options[:options_for_record_to] = Hash.new
+
+OptionParser.new do |opts|
+  opts.banner = "Usage: anonymize_data.rb [options]"
+
+  opts.on("-d", "--destination <directory>", "Destination directory in which anonymized data should be save") do |d|
+    options[:destination] = d
+  end
+  opts.on("-c", '--collection "<list>"', "A collection of files and/or directories to anonymize, seperate with spaces. Quotes are necessary.") do |c|
+    options[:collection] += c.split
+  end
+  opts.on("-g", '--glob "<pattern>"', "Use a shell globbing pattern to select your collection. Quotes are neccessary.") do |g|
+    options[:collection] += Dir.glob(g)
+  end
+  opts.on("--include-csv","Write a csv formatted keymap file in addition to the yaml file.") do |i|
+    options[:options_for_record_to][:include_csv] = true
+  end
+  opts.on("-k","--keymap-name <basename>","Use an alternative file basename (i.e. without extension) for the keymap file.") do |k|
+    options[:options_for_record_to][:keymap_name] = k
+  end
+end.parse!
+
+
+
+raise(IndexError, "No items specified to anonymize. Check your glob pattern or explicitly list your  items using -c.") if options[:collection].empty?
+
+puts "Anonymizing:"
+options[:collection].each do |i|
+  puts "\t#{i}"
+end
+puts "To:"
+puts "\t#{options[:destination]}"
+
+begin
+  a = Anonymizer.new(options[:collection])
+  a.record_to(options[:destination], options[:options_for_record_to])
+rescue
+  puts "There was an unknown problem with anonymization.  Sorry."
+  exit 1
+end
+
+puts "+++ Successfully Anonymized. +++\n"
+exit 0

data/lib/anonymizer.rb

@@ -0,0 +1,118 @@
+require 'fileutils'
+require 'yaml'
+
+class Anonymizer
+  
+  # an array of strings, the paths to the items to be anonymized
+  attr_accessor :filepaths
+  # a string, the directory to which all of the anoymizations and the key should be copied
+  attr_accessor :destination_dir
+  # a hash, the keys are the original filepaths, the values are the corresponding anonymized names
+  attr_accessor :anonymizations
+  
+  
+  # Creates a new instance with an array of filepaths (strings)
+  # A new set of anaonymizations will be created automatically at initialization.  If you want to reshuffle the 
+  # anonymizations you can call anonymize at any time afterward.
+  def initialize(filepaths)
+    @filepaths = filepaths
+    @anonymizations = Hash.new
+    anonymize
+  end
+  
+  
+  # Reshuffles the anonymized names
+  def anonymize
+    keys = Array.new(@filepaths)
+    idx = 0
+    ndigits = (Math.log10(@filepaths.length)+1).floor
+    until keys.empty?
+      @anonymizations[keys.pick!] = "%0#{ndigits}d" % idx
+      idx += 1
+    end
+  end
+  
+  
+  # Copies the collection of files/directories to the destination directory using the current anonymized names. Records
+  # the anonymizations in a keymap.yml file.
+  #
+  # === Options
+  # :inclue_csv => true -- writes the csv format keymap in addition to the yaml
+  #
+  # :keymap_name => "basename" -- uses an alternate file basename for the keymap files (the default is "keymap")
+  #
+  def record_to(destination_dir, options = {})
+    raise(IOError, "Destination directory not found.") unless File.exists?(destination_dir) and File.directory?(destination_dir)
+    @anonymizations.each do |orig, anon|
+      copy_directory_or_file(orig, File.join(destination_dir, anon))
+    end
+    keymap_filename = options[:keymap_name].nil? ? "keymap" : options[:keymap_name]
+    write_yml(destination_dir, keymap_filename + '.yml')
+    write_csv(destination_dir, keymap_filename + '.csv') if options[:include_csv]
+  end
+  
+  
+private
+
+
+  # recursively copies directories, simply copies files preserving file extension.
+  def copy_directory_or_file(orig, dest)
+    raise(IOError, "Original file or directory not found: #{orig}") unless File.exists?(orig)
+    if File.directory?(orig)  # recursively copy directories
+      FileUtils.cp_r(orig, dest)
+    else  # simple copy of files, but preserve file extensions
+      ext = File.extname(orig)
+      FileUtils.cp(orig, dest + ext)
+    end
+  end
+  
+  
+  # creates yaml text based on current anonymization
+  def to_yaml
+    @anonymizations.to_yaml
+  end
+  
+  
+  # creates csv text based on current anonymization
+  def to_csv
+    csv_contents = ""
+    @anonymizations.each do |orig, anon|
+      csv_contents += "#{orig}, #{anon}\n"
+    end
+    return csv_contents
+  end
+  
+  
+  # writes the current anonymization yaml to a file
+  def write_yml(destination_dir,keymap_filename)
+    File.open(File.join(destination_dir,keymap_filename), 'w') do |out|
+       YAML.dump(@anonymizations, out)
+    end
+  end
+  
+  
+  # writes the current anonymization csv to a file
+  def write_csv(destination_dir,keymap_filename)
+    File.open(File.join(destination_dir,keymap_filename), 'w') do |out|
+      out.write(to_csv)
+    end
+  end
+
+end
+
+
+
+# Augments the ruby array class with a couple convenient random picking methods
+class Array
+  
+  # Picks one value from the array at random, array remains unchanged.
+  def pick
+    at(rand(length))
+  end
+  
+  # Removes and returns one element of the array at random, the original array is changed by this method.
+  def pick!
+    delete_at(rand(length))
+  end
+  
+end

metadata

@@ -0,0 +1,66 @@
+--- !ruby/object:Gem::Specification 
+name: brainmap-anonymizer
+version: !ruby/object:Gem::Version 
+  version: 0.1.0
+platform: ruby
+authors: 
+- Kristopher J. Kosmatka
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2009-08-13 00:00:00 -07:00
+default_executable: anonymize_data.rb
+dependencies: []
+
+description: Anonymizes a collection of files and/or directories for interrater reliabiliy analyses and the like.
+email: kk4@medicine.wisc.edu
+executables: 
+- anonymize_data.rb
+extensions: []
+
+extra_rdoc_files: 
+- bin/anonymize_data.rb
+- lib/anonymizer.rb
+- README.rdoc
+files: 
+- anonymizer.gemspec
+- bin/anonymize_data.rb
+- lib/anonymizer.rb
+- Manifest
+- Rakefile
+- README.rdoc
+has_rdoc: true
+homepage: http://github.com/kjkosmatka/anonymizer
+licenses: 
+post_install_message: 
+rdoc_options: 
+- --line-numbers
+- --inline-source
+- --title
+- Anonymizer
+- --main
+- README.rdoc
+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: "1.2"
+  version: 
+requirements: []
+
+rubyforge_project: anonymizer
+rubygems_version: 1.3.5
+signing_key: 
+specification_version: 2
+summary: Anonymizes a collection of files and/or directories for interrater reliabiliy analyses and the like.
+test_files: []
+