A generic object generator and identity map for Ruby
Ruby
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
lib
test
.gitignore
.travis.yml
Gemfile
History.rdoc
IDRegistry.rdoc
README.rdoc
Rakefile
Version
idregistry.gemspec
rakefile_config.rb

README.rdoc

IDRegistry

IDRegistry is a generic object generator and identity map for Ruby.

Introduction to IDRegistry

An IDRegistry is a hub for obtaining and managing domain objects in a Ruby application. It is a configurable general registry for identifying and constructing objects when needed, and it includes a built-in identity map which caches objects in memory and prevents duplicates.

An Example

That's a lot of jargon, so let's clarify with an example. Suppose you are writing a Ruby application in which you have user objects and blog post objects, each identified by an ID number (which might be the primary key in a database.) In effect, you can uniquely identify any of these domain objects by a two-element tuple (array), [type, id], where the type is one of the symbols :user or :post, and the id is an integer.

An IDRegistry is a central object that lets you obtain any of your domain objects by giving it that unique identifying tuple.

# Get the user with ID 1
first_user = registry.lookup(:user, 1)

# Get the user with ID 2
second_user = registry.lookup(:user, 2)

# Get the blog post with ID 300
post = registry.lookup(:post, 300)

Configuration

How does IDRegistry know how to construct your user and post objects? At initialization time, you configure IDRegistry, telling it about each type of object it will need to manage. For each object type, you provide a pattern for the identifying tuple, and a block that constructs a new object when needed. Here's an example:

# Create and configure the registry at initialization time
registry = Registry.create do |config|

  # The pattern for identifying tuples for user objects
  config.add_pattern([:user, Integer]) do |tuple|
    # How to construct a user object given a tuple
    my_construct_user_object_from_id(tuple[1])
  end

  # The pattern for identifying tuples for post objects
  config.add_pattern([:post, Integer]) do |tuple|
    # How to construct a post object given a tuple
    my_construct_post_object_from_id(tuple[1])
  end

end

Now, when you ask for a particular tuple, say, [:user, 1], the IDRegistry finds the pattern that matches that tuple, and uses it to construct and return the appropriate object.

Caching

The real power, however, comes from the fact that IDRegistry now caches all the objects it creates in an internal hash. So if you ask for the same tuple identifier a second time, it doesn't construct a second object, but simply returns the same object it had constructed earlier. In other words, it has a built-in identity map.

# Get the user with ID 1
first_user = registry.lookup(:user, 1)

# If you re-ask for the same identifier, you get the same object.
same_user = registry.lookup(:user, 1)
same_user.object_id == first_user.object_id  # => true

You can remove cached objects from the registry, forcing the registry to re-construct them the next time you ask for them. A common usage pattern in a web application is to clear out the registry cache at the end of each request, so that each request is self-contained and has its own copies of domain objects. A Rack middleware is provided for this purpose.

Tuples

Identifying tuples don't have to follow the pattern [type, id]. They can actually be any array. For example, you might want to identify nodes in a tree using a combination of parent and child name, rather than just the ID. For such tree node objects, you might use a pattern like [:tree, Node, String], where the second element is the parent node itself, and the third is the name of the child.

In fact, it is even possible to provide multiple ways of identifying objects. Perhaps you want to be able to look up tree nodes by either ID number, or by parent/child-name. In configuration, you can tell IDRegistry that these refer to the same type of object. Then, if you first look up an object by ID, and then later look up the same object by parent/child-name, IDRegistry will be smart enough to know you are referring to the same object, and will return its cached data.

Why?

IDRegistry is an extraction of an identity map I wrote for a few large Rails applications, including the back-end for Pirq (www.pirq.com).

Our model objects were getting quite complex with lots of associations and dependencies, and we were having difficulty keeping track of which objects had already been loaded, and whether we had multiple copies of objects in memory that might be getting out of sync with one another.

After we wrote IDRegistry and refactored our object creation to use it, our domain object management code was greatly simplified, and a whole class of bugs was eliminated. We've been using it in production for several years, and now we offer it to the community.

For more info

More detailed info is available in the IDRegistry.rdoc file.

Dependencies

IDRegistry is known to work with the following Ruby implementations:

  • Standard “MRI” Ruby 1.8.7 or later. (1.9.2 or later preferred.)

  • Rubinius 1.1 or later.

  • JRuby 1.6 or later.

Installation

Install IDRegistry as a gem:

gem install idregistry

Development and support

Documentation is available at dazuma.github.com/idregistry/rdoc

Source code is hosted on Github at github.com/dazuma/idregistry

Contributions are welcome. Fork the project on Github.

Build status:

Report bugs on Github issues at github.org/dazuma/idregistry/issues

Contact the author at dazuma at gmail dot com.

Acknowledgments

IDRegistry is written by Daniel Azuma (www.daniel-azuma.com).

Development is supported by Pirq (www.pirq.com).

Continuous integration service provided by Travis-CI (travis-ci.org).

License

Copyright 2012 Daniel Azuma

All rights reserved.

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

  • Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.

  • Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.

  • Neither the name of the copyright holder, nor the names of any other contributors to this software, may be used to endorse or promote products derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.