data_cleansing

Data Cleansing framework for Ruby, Rails, Mongoid and MongoMapper.

http://github.com/reidmorrison/data_cleansing

Introduction

It is important to keep internal data free of unwanted escape characters, leading or trailing blanks and even newlines. Similarly it would be useful to be able to attach a cleansing solution to a field in a model and have the data cleansed transparently when required.

DataCleansing is a framework that allows any data cleansing to be applied to specific attributes or fields. At this time it does not supply the cleaning solutions themselves since they are usually straight forward, or so complex that they don't tend to be too useful to others. However, over time built-in cleansing solutions may be added. Feel free to submit any suggestions via a ticket or pull request.

Features

Supports global cleansing definitions that can be associated with any Ruby, Rails, Mongoid, or other model
Supports custom cleansing definitions that can be defined in-line
A cleansing block can access the other attributes in the model while cleansing the current attribute
In a cleansing block other attributes in the model can be modified at the same time
Cleansers are executed in the order they are defined. As a result multiple cleansers can be run against the same field and the order is preserved
Multiple cleansers can be specified for a list of attributes at the same time
Inheritance is supported. The cleansers for parent classes are run before the child's cleansers
Cleansers can be called outside of a model instance for cases where fields need to be cleansed before the model is created, or needs to be found
To aid troubleshooting the before and after values of cleansed attributes is logged. The level of detail is fine-tuned using the log level

ActiveRecord (ActiveModel) Features

Passes the value of the attribute before the Rails type cast so that the original text can be cleansed before passing back to rails for type conversion. This is important for numeric and date fields where spaces and control characters can have undesired effects

Examples

Ruby Example

require 'data_cleansing'

# Define a global cleaner
DataCleansing.register_cleaner(:strip) {|string| string.strip}

class User
  include DataCleansing::Cleanse

  attr_accessor :first_name, :last_name

  # Strip leading and trialing whitespace from first_name and last_name
  cleanse :first_name, :last_name, :cleaner => :strip
end

u = User.new
u.first_name = '    joe   '
u.last_name = "\n  black\n"
puts "Before data cleansing #{u.inspect}"
# Before data cleansing #<User:0x007fc9f1081980 @first_name="    joe   ", @last_name="\n  black\n">

u.cleanse_attributes!
puts "After data cleansing #{u.inspect}"
# After data cleansing #<User:0x007fc9f1081980 @first_name="joe", @last_name="black">

Rails Example

# Define a global cleanser
DataCleansing.register_cleaner(:strip) {|string| string.strip}

# 'users' table has the following columns :first_name, :last_name, :address1, :address2
class User < ActiveRecord::Base
  include DataCleansing::Cleanse

  # Use a global cleaner
  cleanse :first_name, :last_name, :cleaner => :strip

  # Define a once off cleaner
  cleanse :address1, :address2, :cleaner => Proc.new {|string| string.strip}

  # Automatically cleanse data before validation
  before_validation :cleanse_attributes!
end

# Create a User instance
u = User.new(:first_name => '    joe   ', :last_name => "\n  black\n", :address1 => "2632 Brown St   \n")
puts "Before data cleansing #{u.attributes.inspect}"
u.validate
puts "After data cleansing #{u.attributes.inspect}"
u.save!

Advanced Ruby Example

require 'data_cleansing'

# Define a global cleaners
DataCleansing.register_cleaner(:strip) {|string| string.strip}
DataCleansing.register_cleaner(:upcase) {|string| string.upcase}

class User
  include DataCleansing::Cleanse

  attr_accessor :first_name, :last_name, :title, :address1, :address2, :gender

  # Use a global cleaner
  cleanse :first_name, :last_name, :cleaner => :strip

  # Define a once off cleaner
  cleanse :address1, :address2, :cleaner => Proc.new {|string| string.strip}

  # Use multiple cleaners, and a custom block
  cleanse :title, :cleaner => [:strip, :upcase, Proc.new {|string| "#{string}." unless string.end_with?('.')}]

  # Change the cleansing rule based on the value of other attributes in that instance of user
  # The 'title' is retrieved from the current instance of the user
  cleanse :gender, :cleaner => [
    :strip,
    :upcase,
    Proc.new do |gender|
      if (gender == "UNKNOWN") && (title == "MR.")
        "Male"
      else
        "Female"
      end
    end
  ]
end

u = User.new
u.first_name = '    joe   '
u.last_name = "\n  black\n"
u.address1 = "2632 Brown St   \n"
u.title = "   \nmr   \n"
u.gender = " Unknown  "
puts "Before data cleansing #{u.inspect}"
# Before data cleansing #<User:0x007fdd5a83a8f8 @first_name="    joe   ", @last_name="\n  black\n", @address1="2632 Brown St   \n", @title="   \nmr   \n", @gender=" Unknown  ">

u.cleanse_attributes!
puts "After data cleansing #{u.inspect}"
# After data cleansing #<User:0x007fdd5a83a8f8 @first_name="joe", @last_name="black", @address1="2632 Brown St", @title="MR.", @gender="Male">

After Cleansing

It is sometimes useful to read or write multiple fields as part of a cleansing, or where attributes need to be manipulated automatically once they have been cleansed. For this purpose instance methods on the model can be registered for invocation once all the attributes have been cleansed according to their :cleanse specifications. Multiple methods can be registered and they are called in the order they are registered.

after_cleanse <instance_method_name>, <instance_method_name>, ...

Example:

# Define a global cleanser
DataCleansing.register_cleaner(:strip) {|string| string.strip}

# 'users' table has the following columns :first_name, :last_name, :address1, :address2
class User < ActiveRecord::Base
  include DataCleansing::Cleanse

  # Use a global cleaner
  cleanse :first_name, :last_name, :cleaner => :strip

  # Define a once off cleaner
  cleanse :address1, :address2, :cleaner => Proc.new {|string| string.strip}

  # Once the above cleansing is complete call the instance method
  after_cleanse :check_address

  protected

  # Method to be called once data cleansing is complete
  def check_address
    # Move address2 to address1 if Address1 is blank and address2 has a value
    address2 = address1 if address1.blank? && !address2.blank?
  end

end

# Create a User instance
u = User.new(:first_name => '    joe   ', :last_name => "\n  black\n", :address2 => "2632 Brown St   \n")
puts "Before data cleansing #{u.attributes.inspect}"
u.cleanse_attributes!
puts "After data cleansing #{u.attributes.inspect}"
u.save!

Recommendations

:data_cleanse block are ideal for cleansing a single attribute, and applying any global or common cleansing algorithms.

Even though multiple attributes can be read or written in a single :data_cleanse block, it is recommended to use the :after_cleanse method for working with multiple attributes. It is much easier to read and understand the interactions between multiple attributes in the :after_cleanse methods.

Rails configuration

When DataCleansing is used in a Rails environment it can be configured using the regular Rails configuration mechanisms. For example:

module MyApplication
  class Application < Rails::Application

   # Data Cleansing Configuration

   # Attributes who's values are to be masked out during logging
   config.data_cleansing.register_masked_attributes :bank_account_number, :social_security_number

   # Optionally override the default log level
   #   Set to :trace or :debug to log all fields modified
   #   Set to :info to log only those fields which were nilled out
   #   Set to :warn or higher to disable logging of cleansing actions
   config.data_cleansing.logger.level = :info

   # Register any global cleaners
   config.data_cleansing.register_cleaner(:strip) {|string| string.strip}

  end
end

Logging

DataCleansing uses SemanticLogger for logging due to it's excellent integration with Rails and its ability to log data in it's raw form to Mongo and to files.

If running a Rails application it is recommended to install the gem rails_semantic_logger which replaces the default Rails logger. It is however possible to configure the semantic_logger gem to use the existing Rails logger in a Rails initializer as follows:

SemanticLogger.default_level = Rails.logger.level
SemanticLogger.add_appender(logger: Rails.logger)

By changing the log level of DataCleansing itself the type of output for data cleansing can be controlled:

:trace or :debug to log all fields modified
:info to log only those fields which were nilled out
:warn or higher to disable logging of cleansing actions

Note:

The logging of changes made to attributes only includes attributes cleansed with :data_cleanse blocks. Attributes modified within :after_cleanse methods are not logged
It is not necessary to change the global log level to affect the logging detail level in DataCleansing. DataCleansing log level is changed independently

To change the log level, either use the Rails configuration approach, or set it directly:

DataCleansing.logger.level = :info

Notes

Cleaners are called in the order in which they are defined, so subsequent cleaners can assume that the previous cleaners have run and can therefore access or even modify previously cleaned attributes

Installation

Add to an existing Rails project

Add the following line to Gemfile

gem 'data_cleansing'

Install the Gem with bundler

bundle install

Architecture

DataCleansing has been designed to support externalized data cleansing routines. In this way the data cleansing routine itself can be loaded from a datastore and applied dynamically at runtime. Although not supported out of the box, this design allows for example for the data cleansing routines to be stored in something like ZooKeeper. Then any changes to the data cleansing routines can be pushed out immediately to every server that needs it.

DataCleansing is designed to support any Ruby model. In this way it can be used in just about any ORM or DOM. For example, it currently easily supports both Rails and Mongoid models. Some extensions have been added to support these frameworks.

For example, in Rails it obtains the raw data value before Rails has converted it. Which is useful for cleansing integer or float fields as raw strings before Rails tries to convert it to an integer or float.

Dependencies

DataCleansing requires the following dependencies

Ruby V1.9.3, V2 and greater
Rails V3.2 (Active Model) or greater for Rails integration ( Only if Rails is being used )
Mongoid and Mongomapper supporting Active Model V3.2 or greater ( Only if Mongoid or MongoMapper is being used )

Authors

Reid Morrison :: [email protected] :: @reidmorrison

License

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.