Skip to content

KMarshland/cachable

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

14 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Cachable

Caching is often a bother. This gem allows you to simply wrap your code in a block and have it be cached with redis.

Installation

Add this line to your application's Gemfile:

gem 'cachable'

You must also add redis to your app. On heroku, you can add Heroku Redis.

Usage

At its hearty, this gem is the unless_cached method. Although no options are required, it accepts the following:

  • key. If present, will be added to the base key to generate the full key. Defaults to the name of the caller.
  • json. If true, will serialize and deserialize the result as json
  • expiration Time for which to cache the result. Defaults to 1 day
  • json_options. Options that get passed into json serialization and deserialization
  • skip_cache. If true, will not populate the cache -- this can be used if you populate the cache elsewhere, but still want to check if there's something there.

Examples

In your model:

include Cachable

def your_method
  unless_cached do 
    # ... output of this will be cached under the key model_name_id_updated_at_your_method 
  end
end

Of course, unless_cached accepts a variety of options:

# opts[:key]. If present, will be added to the base key to generate the full key. Defaults to the name of the caller.
unless_cached(key: 'some_key') do
  # will be stored under the key model_name_id_some_key
end


# opts[:json]. If true, will serialize and deserialize the result as json
unless_cached(json: true) do
  {
    json: 'object'
  }
end

# opts[:expiration]. Time for which to cache the result. Defaults to 1 day
unless_cached(expiration: 30.minutes) do
  # result will expire in half an hour
end


# opts[:json_options]. Options that get passed into json serialization and deserialization
unless_cached(json: true, json_options: {allow_nan: true}) do
  {
    value: Float::NAN
  }
end

Other features

Configuring the redis connection

It defaults to checking the REDIS_URL and the HEROKU_REDIS_URL environment variables to make a connection to redis. However, you may wish to change the way you connect to redis. Create an initializer in config/initializers/cachable.rb.

Cachable::configure do |config|
  
  # set a lambda to get a redis connection
  config.redis_connection = -> {
    Redis.new
  }
  
  # set the redis instance directly
  config.redis_instance = Redis.new
  
  # or set the redis url
  config.redis_url = 'YOUR REDIS URL'
  
end

Deleting from the cache

Sometimes you want to delete something from the cache. This can be done using the delete_from_cache method, which takes in one or more keys. These keys are either the

delete_from_cache(:key1, :key2)

A special use case is after the changes have been committed. For this, you can use the special purge_cache callback. You must define the tracked_cache_keys method for this to work properly. The following example will clear the cache for cached_a and cached_b, but not cached_c.

# in your model
after_commit :purge_cache

def tracked_cache_keys
  [:cached_a, :cached_b]
end

def cached_a
  unless_cached do 
    # ...
  end
end

def cached_b
  unless_cached do 
    # ...
  end
end

def cached_c
  unless_cached do 
    # ...
  end
end

If you want to use the delete_from_cache method directly in an after_commit callback, you must specify that in the options.

delete_from_cache(:key1, :key2, after_commit: true)

Batch requests to the cache

Oftentimes, you want the output of a given method on an ActiveRecordCollection. You could loop through each record in the collection, but if you expect the method to be cached, it's much more efficient this gem's method, and avoid the memory overhead and extra database calls.

Imagine you have a model that looks as follows:

class Book < ActiveRecord::Base
  include Cachable
  
  def summary
    unless_cached do
      'Some summary'
    end
  end
end

To get the summaries of all fantasy books, you could run Book.where(genre: 'fantasy').unless_cached(:summary).

Note that this will remove any ordering on the collection unless you set the slurp option to true or include EachInOrder in your model (this gem has not yet been released -- create an issue if you need me to release it). A full list of options is as follows:

  • slurp. If true will not pull the records in in batches, which increases memory overhead but preserves order.
  • force_cache. If true will add its own cache, regardless of what the underlying function does.
  • cache_batches. If true will add another layer of caching outside each individual model. This caches a the result of up to 50 records at a time, which can drastically speed up the cache but will also increase the cache size.
  • skip_result. If true will not return the result, but it will still prepopulate the cache, which can avoid memory overhead.
  • clear_previous_batch. If true, and was also true previous times, it will remove the cached batches for the previous batch. This can be very useful when you are frequently adding more records and want to keep the redis memory usage down. Note that this was designed to have minimal overhead, and so was based on the total number of records, which may cause a small number of stale keys to stay in memory.
  • You can also pass in all normal unless_cached options, such as expiration, json, and json_options.

Adding an additional redis key

Sometimes your cache depends on more than just the model. For example, imagine you have a model book that belongs_to an author. When the author is updated, you want the cache to become invalidated. In your book model, you might have something like this:

def added_redis_key
  self.author.updated_at.to_i
end

def self.added_redis_key
  first = all.first
  return '' if first.blank?

  first.author.updated_at.to_i
end

Using the cache outside of a specific instance

Sometimes you need to cache something in a static (class) method. It accepts the following options:

  • json. If true, will serialize and deserialize the result as json
  • expiration Time for which to cache the result. Defaults to 1 day
  • json_options. Options that get passed into json serialization and deserialization
  • skip_cache. If true, will not populate the cache -- this can be used if you populate the cache elsewhere, but still want to check if there's something there.
self.class.unless_cached_base("#{self.class.to_s.downcase}_key", json: true, expiration: 1.day) do
 # this output will be cached under the key your_model_name_key
end

License

The gem is available as open source under the terms of the MIT License.

About

Easily add caching to models

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published