Skip to content

Latest commit

 

History

History
204 lines (153 loc) · 5.78 KB

README.md

File metadata and controls

204 lines (153 loc) · 5.78 KB
Raw

Piperator

Pipelines for streaming large collections. The pipeline enables composition of streaming pipelines with lazy enumerables.

The library is heavily inspired by Elixir pipe operator and Node.js Stream.

Table of Contents

Installation

Piperator is distributed as a ruby gem and can be installed with

$ gem install piperator

Usage

Start by requiring the gem

require 'piperator'

Pipelines

As an appetizer, here's a pipeline that triples all input values and then sums the values.

Piperator.
  pipe(->(values) { values.lazy.map { |i| i * 3 } }).
  pipe(->(values) { values.sum }).
  call([1, 2, 3])
# => 18

The same could also be achieved using DSL instead of method chaining:

Piperator.build do
  pipe(->(values) { values.lazy.map { |i| i * 3 } })
  pipe(->(values) { values.sum })
end.call([1, 2, 3])

If desired, the input enumerable can also be given as the first element of the pipeline using Piperator.wrap.

Piperator.
  wrap([1, 2, 3]).
  pipe(->(values) { values.lazy.map { |i| i * 3 } }).
  pipe(->(values) { values.sum }).
  call
# => 18

Have reasons to defer constructing a pipe? Evaluate it lazily:

summing = ->(values) { values.sum }
Piperator.build
  pipe(->(values) { values.lazy.map { |i| i * 3 } })
  lazy do
    summing
  end
end.call([1, 2, 3])

There is, of course, a much more idiomatic alternative in Ruby:

[1, 2, 3].map { |i| i * 3 }.sum

So why bother?

To run code before the stream processing start and after processing has ended. Let's use the same pattern to calculate the decompressed length of a GZip file fetched over HTTP with streaming.

require 'piperator'
require 'uri'
require 'em-http-request'
require 'net/http'

module HTTPFetch
  def self.call(url)
    uri = URI(url)
    Enumerator.new do |yielder|
      Net::HTTP.start(uri.host, uri.port, use_ssl: uri.scheme == 'https') do |http|
        request = Net::HTTP::Get.new(uri.request_uri)
        http.request request do |response|
          response.read_body { |chunk| yielder << chunk }
        end
      end
    end
  end
end

module GZipDecoder
  def self.call(enumerable)
    Enumerator.new do |yielder|
      decoder = EventMachine::HttpDecoders::GZip.new do |chunk|
        yielder << chunk
      end

      enumerable.each { |chunk| decoder << chunk }
      yielder << decoder.finalize.to_s
    end
  end
end

length = proc do |enumerable|
  enumerable.lazy.reduce(0) { |aggregate, chunk| aggregate + chunk.length }
end

Piperator.
  pipe(HTTPFetch).
  pipe(GZipDecoder).
  pipe(length).
  call('http://ftp.gnu.org/gnu/gzip/gzip-1.2.4.tar.gz')

At no point is it necessary to keep the full response or decompressed content in memory. This is a huge win when the file sizes grow beyond the 780kB seen in the example.

Pipelines themselves respond to #call. This enables using pipelines as pipes in other pipelines.

append_end = proc do |enumerator|
  Enumerator.new do |yielder|
    enumerator.each { |item| yielder << item }
    yielder << 'end'
  end
end

prepend_start = proc do |enumerator|
  Enumerator.new do |yielder|
    yielder << 'start'
    enumerator.each { |item| yielder << item }
  end
end

double = ->(enumerator) { enumerator.lazy.map { |i| i * 2 } }

prepend_append = Piperator.pipe(prepend_start).pipe(append_end)
Piperator.pipe(double).pipe(prepend_append).call([1, 2, 3]).to_a
# => ['start', 2, 4, 6, 'end']

Enumerators as IO objects

Piperator also provides a helper class that allows Enumerators to be used as IO objects. This is useful to provide integration with libraries that work only with IO objects such as Nokogiri or Oj.

An example pipe that would yield all XML node in a document read in streams:

require 'nokogiri'
streaming_xml = lambda do |enumerable|
  Enumerator.new do |yielder|
    io = Piperator::IO.new(enumerable.each)
    reader = Nokogiri::XML::Reader(io)
    reader.each { |node| yielder << node }
  end
end

In real-world scenarios, the pipe would need to filter the nodes. Passing every single XML node forward is not that useful.

Development

After checking out the repo, run bin/setup to install dependencies. Then, run rake spec to run the tests. You can also run bin/console for an interactive prompt that will allow you to experiment.

To install this gem onto your local machine, run bundle exec rake install. To release a new version, update the version number in version.rb, and then run bundle exec rake release, which will create a git tag for the version, push git commits and tags, and push the .gem file to rubygems.org.

Related Projects

  • D★Stream - Set of extensions for writing stream-processing code.
  • ddbuffer - Buffer enumerables/enumerators.
  • Down::ChunkedIO - A similar IO class as Piperator::IO.

Contributing

Bug reports and pull requests are welcome on GitHub at https://github.com/lautis/piperator.

License

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