Futuroscope

Join a live discussion on Gitter:

Futursocope is a simple library that implements futures in ruby. Futures are a concurrency pattern meant to help you deal with threads in a simple, transparent way.

It's specially useful in situations where you have calls to an expensive resource that could be done in parallel (they are not chained), but you don't wanna deal with low-level threads. HTTP calls are a good example.

Also useful when you want to spin up a process that runs in the background, do some stuff in the middle, and wait for that process to return.

You can learn more about futures here in this excellent article from @jpignata: Concurrency Patterns in Ruby: Futures

In Futuroscope, futures are instantiated with a simple ruby block. The future's execution will immediately start in a different thread and when you call a method on in it will be forwarded to the block's return value.

If the thread didn't finish yet, it will block the program's execution until it's finished. Otherwise, it will immediately return its value.

Futuroscope is tested on MRI 1.9.3, MRI 2.0.0, MRI 2.1.0, Rubinius (2.2.3) and JRuby (1.9).

Check out futuroscope's post on Codegram's blog to get started.

Installation

Add this line to your application's Gemfile:

gem 'futuroscope'

And then execute:

$ bundle

Or install it yourself as:

$ gem install futuroscope

Usage

Simple futures

require 'futuroscope'

x = Futuroscope::Future.new{ sleep(1); 1 }
y = Futuroscope::Future.new{ sleep(1); 2 }
z = Futuroscope::Future.new{ sleep(1); 3 }

# This execution will actually take just one second and not three like you
# would expect.

puts x + y + z
=> 6

Since a future is actually delegating everything to the future's value, there might be some cases where you want to get the actual future's value. You can do it just by calling the future_value method on the future:

string = "Ed Balls"
x = future{ string }
x.future_value === string
# => true

Future map

require 'futuroscope'

map = Futuroscope::Map.new([1, 2, 3]).map do |i|
  sleep(1)
  i + 1
end

puts map.first
=> 2

puts map[1]
=> 3

puts map.last
=> 4

# This action will actually only take 1 second.

Convenience methods

If you don't mind polluting the Kernel module, you can also require futuroscope's convenience future method:

require 'futuroscope/convenience'

x = future{ sleep(1); 1 }
y = future{ sleep(1); 2 }
z = future{ sleep(1); 3 }

puts x + y + z
=> 6

Same for a map:

require 'futuroscope/convenience'

items = [1, 2, 3].future_map do |i|
  sleep(i)
  i + 1
end

Considerations

You should never add side-effects to a future. They have to be thought of like they were a local variable, with the only outcome that they're returning a value.

You have to take into account that they really run in a different thread, so you'll be potentially accessing code in parallel that could not be thread-safe.

If you're looking for other ways to improve your code performance via concurrency, you should probably deal directly with Ruby's threads.

Worker Pool

Futures are scheduled in a worker pool that helps managing concurrency in a way that doesn't get out of hands. Also comes with great benefits since their threads are spawned at load time (and not in runtime).

The default thread pool comes with a concurrency of 8 Workers, which seems reasonable for the most use cases. It will elastically expand to the default of 16 threads and will kill them when they're not needed.

The default thread pool can be configured like this:

Futuroscope.default_pool.min_workers = 2
Futuroscope.default_pool.max_workers = 16

Also, each future can be scheduled to a different pool like this:

pool = Futuroscope::Pool.new(16..32)

future = Future.new(pool){ :edballs }

# Or with the convenience method
future = future(pool){ :edballs }

Contributing

1. Fork it
2. Create your feature branch (git checkout -b my-new-feature)
3. Commit your changes (git commit -am 'Add some feature')
4. Push to the branch (git push origin my-new-feature)
5. Create new Pull Request