Skip to content
/ garner Public

A set of Rack middleware and cache helpers that implement various caching strategies.

License

Notifications You must be signed in to change notification settings

artsy/garner

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Garner

Gem Version Build Status Dependency Status Code Climate Coverage Status

Garner is a cache layer for Ruby and Rack applications, supporting model and instance binding and hierarchical invalidation. To "garner" means to gather data from various sources and to make it readily available in one place, kind of like a cache!

If you're not familiar with HTTP caching, ETags and If-Modified-Since, watch us introduce Garner in From Zero to API Cache in 10 Minutes at GoRuCo 2012.

Upgrading

The current stable release line of Garner is 0.5.x, and contains many breaking changes from the previous 0.3.x series. For a summary of important changes, see UPGRADING.

Usage

Application Logic Caching

Add Garner to your Gemfile with gem "garner" and run bundle install. Next, include the appropriate mixin in your app:

  • For plain-old Ruby apps, include Garner::Cache::Context.
  • For Rack apps, first require "garner/mixins/rack", then include Garner::Mixins::Rack. (This provides saner defaults for injecting request parameters into the cache context key. More on cache context keys later.)

Now, to use Garner's cache, invoke garner with a logic block from within your application. The result of the block will be computed once, and then stored in the cache.

get "/system/counts/all" do
  # Compute once and cache for subsequent reads
  garner do
    {
      "orders_count" => Order.count,
      "users_count"  => User.count
    }
  end
end

The cached value can be bound to a particular model instance. For example, if a user has an address that may or may not change when the user is saved, you will want the cached address to be invalidated every time the user record is modified.

get "/me/address" do
  # Invalidate when current_user is modified
  garner.bind(current_user) do
    current_user.address
  end
end

ORM Integrations

Mongoid

To use Mongoid 3, 4 or 5 documents and classes for Garner bindings, use Garner::Mixins::Mongoid::Document. You can set it up in an initializer:

require "garner/mixins/mongoid"

module Mongoid
  module Document
    include Garner::Mixins::Mongoid::Document
  end
end

This enables binding to Mongoid classes as well as instances. For example:

get "/system/counts/orders" do
  # Invalidate when any order is created, updated or deleted
  garner.bind(Order) do
    {
      "orders_count" => Order.count,
    }
  end
end

What if you want to bind a cache result to a persisted object that hasn't been retrieved yet? Consider the example of caching a particular order without a database query:

get "/order/:id" do
  # Invalidate when Order.find(params[:id]) is modified
  garner.bind(Order.identify(params[:id])) do
    Order.find(params[:id])
  end
end

In the above example, the Order.identify call will not result in a database query. Instead, it just communicates to Garner's cache sweeper that whenever the order with identity params[:id] is updated, this cache result should be invalidated. The identify method is provided by the Mongoid mixin. To use it, you should configure Garner.config.mongoid_identity_fields, e.g.:

Garner.configure do |config|
  config.mongoid_identity_fields = [:_id, :_slugs]
end

These may be scalar or array fields. Only uniquely-constrained fields should be used here; otherwise you risk caching the same result for two different blocks.

The Mongoid mixin also provides helper methods for cached find operations. The following code will fetch an order once (via find) from the database, and then fetch it from the cache on subsequent requests. The cache will be invalidated whenever the underlying Order changes in the database.

order = Order.garnered_find(3)

Explicit invalidation should be unnecessary, since callbacks are declared to invalidate the cache whenever a Mongoid object is created, updated or destroyed, but for special cases, invalidate_garner_caches may be called on a Mongoid object or class:

Order.invalidate_garner_caches
Order.find(3).invalidate_garner_caches

ActiveRecord

Garner provides rudimentary support for ActiveRecord. To use ActiveRecord models for Garner bindings, use Garner::Mixins::ActiveRecord::Base. You can set it up in an initializer:

require "garner/mixins/active_record"

module ActiveRecord
  class Base
    include Garner::Mixins::ActiveRecord::Base
  end
end

Cache Options

You can pass additional options directly to the cache implementation:

get "/latest_order" do
  # Expire the latest order every 15 minutes
  garner.options(expires_in: 15.minutes) do
    Order.latest
  end
end

Under The Hood: Bindings

As we've seen, a cache result can be bound to a model instance (e.g., current_user) or a virtual instance reference (Order.identify(params[:id])). In some cases, we may want to compose bindings:

get "/system/counts/all" do
  # Invalidate when any order or user is modified
  garner.bind(Order).bind(User) do
    {
      "orders_count" => Order.count,
      "users_count"  => User.count
    }
  end
end

Binding keys are computed via pluggable strategies, as are the rules for invalidating caches when a binding changes. By default, Garner uses Garner::Strategies::Binding::Key::SafeCacheKey to compute binding keys: this uses cache_key if defined on an object; otherwise it always bypasses cache. Similarly, Garner uses Garner::Strategies::Binding::Invalidation::Touch as its default invalidation strategy. This will call :touch on a document if it is defined; otherwise it will take no action.

Additional binding and invalidation strategies can be written. To use them, set Garner.config.binding_key_strategy and Garner.config.binding_invalidation_strategy.

Under The Hood: Cache Context Keys

Explicit cache context keys are usually unnecessary in Garner. Given a cache binding, Garner will compute an appropriately unique cache key. Moreover, in the context of Garner::Mixins::Rack, Garner will compose the following key factors by default:

  • Garner::Strategies::Context::Key::Caller inserts the calling file and line number, allowing multiple calls from the same function to generate different results.
  • Garner::Strategies::Context::Key::RequestGet inserts the value of HTTP request's GET parameters into the cache key when :request is present in the context.
  • Garner::Strategies::Context::Key::RequestPost inserts the value of HTTP request's POST parameters into the cache key when :request is present in the context.
  • Garner::Strategies::Context::Key::RequestPath inserts the value of the HTTP request's path into the cache key when :request is present in the context.

Additional key factors may be specified explicitly using the key method. To see a specific example of this in action, let's consider the case of role-based caching. For example, an order may have a different representation for an admin versus an ordinary user:

get "/order/:id" do
  garner.bind(Order.identify(params[:id])).key({ role: current_user.role }) do
    Order.find(params[:id])
  end
end

As with bindings, context key factors may be composed by calling key() multiple times on a garner invocation. The keys will be applied in the order in which they are called.

Configuration

By default Garner will use an instance of ActiveSupport::Cache::MemoryStore in a non-Rails and Rails.cache in a Rails environment. You can configure it to use any other cache store.

Garner.configure do |config|
  config.cache = ActiveSupport::Cache::FileStore.new
end

The full list of Garner.config attributes is:

  • :global_cache_options: A hash of options to be passed on every call to Garner.config.cache, like { :expires_in => 10.minutes }. Defaults to {}
  • :context_key_strategies: An array of context key strategies, to be applied in order. Defaults to [Garner::Strategies::Context::Key::Caller]
  • :rack_context_key_strategies: Rack-specific context key strategies. Defaults to:
[
  Garner::Strategies::Context::Key::Caller,
  Garner::Strategies::Context::Key::RequestGet,
  Garner::Strategies::Context::Key::RequestPost,
  Garner::Strategies::Context::Key::RequestPath
]
  • :binding_key_strategy: Binding key strategy. Defaults to Garner::Strategies::Binding::Key::SafeCacheKey.
  • :binding_invalidation_strategy: Binding invalidation strategy. Defaults to Garner::Strategies::Binding::Invalidation::Touch.
  • :mongoid_identity_fields: Identity fields considered legal for the identity method. Defaults to [:_id].
  • :caller_root: Root path of application, to be stripped out of value strings generated by the Caller context key strategy. Defaults to Rails.root if in a Rails environment; otherwise to the nearest ancestor directory containing a Gemfile.
  • :invalidate_mongoid_root: If set to true, invalidates the _root document along with any embedded Mongoid document binding. Defaults to true.
  • :whiny_nils: If set to true, raises an exception when a nil binding is specified (i.e., garner.bind(nil)). Defaults to true.

Contributing

Fork the project. Make your feature addition or bug fix with tests. Send a pull request.

Copyright and License

MIT License, see LICENSE for details.

(c) 2012-2013 Artsy, Frank Macreery, Daniel Doubrovkine and contributors.

About

A set of Rack middleware and cache helpers that implement various caching strategies.

Resources

License

Security policy

Stars

Watchers

Forks

Packages

No packages published

Languages