datamappify

Datamappify

Compose, decouple and manage domain logic and data persistence separately. Works particularly great for composing form objects!

Overview

The typical Rails (and ActiveRecord) way of building applications is great for small to medium sized projects, but when projects grow larger and more complex, your models too become larger and more complex - it is not uncommon to have god classes such as a User model.

Datamappify tries to solve two common problems in web applications:

1. The coupling between domain logic and data persistence.
2. The coupling between forms and models.

Datamappify is loosely based on the Repository Pattern and Entity Aggregation, and is built on top of Virtus and existing ORMs (ActiveRecord and Sequel, etc).

There are three main design goals:

1. To utilise the powerfulness of existing ORMs so that using Datamappify doesn't interrupt too much of your current workflow. For example, Devise would still work if you use it with a UserAccount ActiveRecord model that is attached to a User entity managed by Datamappify.
2. To have a flexible entity model that works great with dealing with form data. For example, SimpleForm would still work with nested attributes from different ORM models if you map entity attributes smartly in your repositories managed by Datamappify.
3. To have a set of data providers to encapsulate the handling of how the data is persisted. This is especially useful for dealing with external data sources such as a web service. For example, by calling UserRepository.save(user), certain attributes of the user entity are now persisted on a remote web service. Better yet, dirty tracking and lazy loading are supported out of the box!

Datamappify consists of three components:

• Entity contains models behaviour, think an ActiveRecord model with the persistence specifics removed.
• Repository is responsible for data retrieval and persistence, e.g. find, save and destroy, etc.
• Data as the name suggests, holds your model data. It contains ORM objects (e.g. ActiveRecord models).

Below is a high level and somewhat simplified overview of Datamappify's architecture.

Note: Datamappify is NOT affiliated with the Datamapper project.

Built-in ORMs for Persistence

You may implement your own data provider and criteria, but Datamappify comes with build-in support for the following ORMS:

• ActiveRecord
• Sequel

Requirements

• ruby 2.0+
• ActiveModel 4.0+

Installation

Add this line to your application's Gemfile:

gem 'datamappify'

Usage

Entity

Entity uses Virtus DSL for defining attributes and ActiveModel::Validations DSL for validations.

The cool thing about Virtus is that all your attributes get coercion for free!

Below is an example of a User entity, with inline comments on how some of the DSLs work.

class User
  include Datamappify::Entity

  attribute :first_name,     String
  attribute :last_name,      String
  attribute :age,            Integer
  attribute :passport,       String
  attribute :driver_license, String
  attribute :health_care,    String

  # Nested entity composition - composing the entity with attributes and validations from other entities
  #
  #   class Job
  #     include Datamappify::Entity
  #
  #     attributes :title, String
  #     validates  :title, :presence => true
  #   end
  #
  #   class User
  #     # ...
  #     attributes_from Job
  #   end
  #
  # essentially equals:
  #
  #   class User
  #     # ...
  #     attributes :title, String
  #     validates  :title, :presence => true
  #   end
  attributes_from Job

  # optionally you may prefix the attributes, so that:
  #
  #   class Hobby
  #     include Datamappify::Entity
  #
  #     attributes :name, String
  #     validates  :name, :presence => true
  #   end
  #
  #   class User
  #     # ...
  #     attributes_from Hobby, :prefix_with => :hobby
  #   end
  #
  # becomes:
  #
  #   class User
  #     # ...
  #     attributes :hobby_name, String
  #     validates  :hobby_name, :presence => true
  #   end
  attributes_from Hobby, :prefix_with => :hobby

  # Entity reference
  #
  # `references` is a convenient method for:
  #
  #   attribute :account_id, Integer
  #   attr_accessor :account
  #
  # and it assigns `account_id` the correct value:
  #
  #   user.account = account #=> user.account_id = account.id
  references :account

  validates :first_name, :presence => true,
                         :length   => { :minimum => 2 }
  validates :passport,   :presence => true,
                         :length   => { :minimum => 8 }

  def full_name
    "#{first_name} #{last_name}"
  end
end

Entity inheritance

Inheritance is supported for entities, for example:

class AdminUser < User
  attribute :level, Integer
end

class GuestUser < User
  attribute :expiry, DateTime
end

Lazy loading

Datamappify supports attribute lazy loading via the Lazy module.

class User
  include Datamappify::Entity
  include Datamappify::Lazy
end

When an entity is lazy loaded, only attributes from the primary source (e.g. User entity's primary source would be ActiveRecord::User as specified in the corresponding repository) will be loaded. Other attributes will only be loaded once they are called. This is especially useful if some of your data sources are external web services.

Repository

Repository maps entity attributes to DB columns - better yet, you can even map attributes to different ORMs!

Below is an example of a repository for the User entity, you can have more than one repositories for the same entity.

class UserRepository
  include Datamappify::Repository

  # specify the entity class
  for_entity User

  # specify the default data provider for unmapped attributes
  # optionally you may use `Datamappify.config` to config this globally
  default_provider :ActiveRecord

  # specify any attributes that need to be mapped
  #
  # for attributes mapped from a different source class, a foreign key on the source class is required
  #
  # for example:
  #   - 'last_name' is mapped to the 'User' ActiveRecord class and its 'surname' attribute
  #   - 'driver_license' is mapped to the 'UserDriverLicense' ActiveRecord class and its 'number' attribute
  #   - 'passport' is mapped to the 'UserPassport' Sequel class and its 'number' attribute
  #   - attributes not specified here are mapped automatically to 'User' with provider 'ActiveRecord'
  map_attribute :last_name,      :to => 'User#surname'
  map_attribute :driver_license, :to => 'UserDriverLicense#number'
  map_attribute :passport,       :to => 'UserPassport#number',   :provider => :Sequel
  map_attribute :health_care,    :to => 'UserHealthCare#number', :provider => :Sequel

  # alternatively, you may group attribute mappings if they share certain options:
  group :provider => :Sequel do
    map_attribute :passport,    :to => 'UserPassport#number'
    map_attribute :health_care, :to => 'UserHealthCare#number'
  end

  # attributes can also be reverse mapped by specifying the `via` option
  #
  # for example, the below attribute will look for `hobby_id` on the user object,
  # and map `hobby_name` from the `name` attribute of `ActiveRecord::Hobby`
  #
  # this is useful for mapping form fields (similar to ActiveRecord's nested attributes)
  map_attribute :hobby_name, :to => 'Hobby#name', :via => :hobby_id

  # by default, Datamappify maps attributes using an inferred reference (foreign) key,
  # for example, the first mapping below will look for the `user_id` key in `Bio`,
  # the second mapping below will look for the `person_id` key in `Bio` instead
  map_attribute :bio, :to => 'Bio#body'
  map_attribute :bio, :to => 'Bio#body', :reference_key => :person_id
end

Repository inheritance

Inheritance is supported for repositories when your data structure is based on STI (Single Table Inheritance), for example:

class AdminUserRepository < UserRepository
  for_entity AdminUser
end

class GuestUserRepository < UserRepository
  for_entity GuestUser

  map_attribute :expiry, :to => 'User#expiry_date'
end

In the above example, both repositories deal with the ActiveRecord::User data model.

Override mapped data models

Datamappify repository by default creates the underlying data model classes for you. For example:

map_attribute :driver_license, :to => 'UserData::DriverLicense#number'

In the above example, a Datamppify::Data::Record::ActiveRecord::UserDriverLicense ActiveRecord model will be created. If you would like to customise the data model class, you may do so by creating one either under the default namespace or under the Datamappify::Data::Record::NameOfDataProvider namespace:

module UserData
  class DriverLicense < ActiveRecord::Base
    # your customisation...
  end
end

module Datamappify::Data::Record::ActiveRecord::UserData
  class DriverLicense < ::ActiveRecord::Base
    # your customisation...
  end
end

Repository APIs

More repository APIs are being added, below is a list of the currently implemented APIs.

Retrieving an entity

Accepts an id.

user = UserRepository.find(1)

Checking if an entity exists in the repository

Accepts an entity.

UserRepository.exists?(user)

Retrieving all entities

Returns an array of entities.

users = UserRepository.all

Searching entities

Returns an array of entities.

Simple

users = UserRepository.where(:first_name => 'Fred', :driver_license => 'AABBCCDD')

Match

users = UserRepository.match(:first_name => 'Fre%', :driver_license => '%bbcc%')

Advanced

You may compose search criteria via the criteria method.

users = UserRepository.criteria(
  :where => {
    :first_name => 'Fred'
  },
  :order => {
    :last_name => :asc
  },
  :limit => [10, 20]
)

Currently implemented criteria options:

• where(Hash)
• match(Hash)
• order(Hash)
• limit(Array<limit(Integer), offset(Integer)>)

Note: it does not currently support searching attributes from different data providers.

Saving/updating entities

Accepts an entity.

There is also save! that raises Datamappify::Data::EntityNotSaved.

UserRepository.save(user)

Datamappify supports attribute dirty tracking - only dirty attributes will be saved.

Mark attributes as dirty

Sometimes it's useful to manually mark the whole entity, or some attributes in the entity to be dirty. In this case, you could:

UserRepository.states.mark_as_dirty(user) # marks the whole entity as dirty

UserRepository.states.find(user).changed?            #=> true
UserRepository.states.find(user).first_name_changed? #=> true
UserRepository.states.find(user).last_name_changed?  #=> true
UserRepository.states.find(user).age_changed?        #=> true

Or:

UserRepository.states.mark_as_dirty(user, :first_name, :last_name) # marks only first_name and last_name as dirty

UserRepository.states.find(user).changed?            #=> true
UserRepository.states.find(user).first_name_changed? #=> true
UserRepository.states.find(user).last_name_changed?  #=> true
UserRepository.states.find(user).age_changed?        #=> false

Destroying an entity

Accepts an entity.

There is also destroy! that raises Datamappify::Data::EntityNotDestroyed.

Note that due to the attributes mapping, any data found in mapped records are not touched. For example the corresponding ActiveRecord::User record will be destroyed, but ActiveRecord::Hobby that is associated will not.

UserRepository.destroy(user)

Initialising an entity

Accepts an entity class and returns a new entity.

This is useful for using before_init and after_init callbacks to set up the entity.

UserRepository.init(user_class) #=> user

Callbacks

Datamappify supports the following callbacks via Hooks:

• before_init
• before_load
• before_find
• before_create
• before_update
• before_save
• before_destroy
• after_init
• after_load
• after_find
• after_create
• after_update
• after_save
• after_destroy

Callbacks are defined in repositories, and they have access to the entity. For example:

class UserRepository
  include Datamappify::Repository

  before_create :make_me_admin
  before_create :make_me_awesome
  after_save    :make_me_smile

  private

  def make_me_admin(entity)
    # ...
  end

  def make_me_awesome(entity)
    # ...
  end

  def make_me_smile(entity)
    # ...
  end

  # ...
end

Note: Returning either nil or false from the callback will cancel all subsequent callbacks (and the action itself, if it's a before_ callback).

Association

Datamappify also supports entity association. It is experimental and it currently supports the following association types:

• belongs_to (partially implemented)
• has_one
• has_many

Set up your entities and repositories:

# entities

class User
  include Datamappify::Entity

  has_one  :title, :via => Title
  has_many :posts, :via => Post
end

class Title
  include Datamappify::Entity

  belongs_to :user
end

class Post
  include Datamappify::Entity

  belongs_to :user
end

# repositories

class UserRepository
  include Datamappify::Repository

  for_entity User

  references :title, :via => TitleRepository
  references :posts, :via => PostRepository
end

class TitleRepository
  include Datamappify::Repository

  for_entity Title
end

class PostRepository
  include Datamappify::Repository

  for_entity Post
end

Usage examples:

new_post         = Post.new(post_attributes)
another_new_post = Post.new(post_attributes)
user             = UserRepository.find(1)
user.title       = Title.new(title_attributes)
user.posts       = [new_post, another_new_post]

persisted_user   = UserRepository.save!(user)

persisted_user.title #=> associated title
persisted_user.posts #=> an array of associated posts

Nested attributes in forms

Like ActiveRecord and ActionView, Datamappify also supports nested attributes via fields_for or simple_fields_for.

# slim template

= simple_form_for @post do |f|
  = f.input :title
  = f.input :body

  = f.simple_fields_for :comment do |fp|
    = fp.input :author_name
    = fp.input :comment_body

Default configuration

You may configure Datamappify's default behaviour. In Rails you would put it in an initializer file.

Datamappify.config do |c|
  c.default_provider = :ActiveRecord
end

Built-in extensions

Datamappify ships with a few extensions to make certain tasks easier.

Kaminari

Use Criteria with page and per.

UserRepository.criteria(
  :where => {
    :gender => 'male',
    :age    => 42
  },
  :page => 1,
  :per  => 10
)

API Documentation

• Rubygem release version
• Github master version

More Reading

You may check out this article for more examples.

Changelog

Refer to CHANGELOG.

Todo

• Authoritative source.
• Support for configurable primary keys and reference (foreign) keys.

Similar Projects

• Curator
• Edr
• Minimapper
• Reform

Credits

• Fred Wu - author.
• James Ladd for reviewing the code and giving advice on architectural decisions.
• Locomote - where Datamappify is built and used in our products.
• And with these awesome contributors!

License

Licensed under MIT