increments/circleci-parallel

[![Gem Version](http://img.shields.io/gem/v/circleci-parallel.svg?style=flat)](http://badge.fury.io/rb/circleci-parallel)
[![Dependency Status](http://img.shields.io/gemnasium/increments/circleci-parallel.svg?style=flat)](https://gemnasium.com/increments/circleci-parallel)
[![CircleCI](https://circleci.com/gh/increments/circleci-parallel.svg?style=shield)](https://circleci.com/gh/increments/circleci-parallel)
[![Coverage Status](https://img.shields.io/codeclimate/coverage/github/increments/circleci-parallel.svg)](https://codeclimate.com/github/increments/circleci-parallel/coverage)
[![Code Climate](https://img.shields.io/codeclimate/github/increments/circleci-parallel.svg?style=flat)](https://codeclimate.com/github/increments/circleci-parallel)

# CircleCI::Parallel

**CircleCI::Parallel** provides simple Ruby APIs for syncing [CircleCI parallel nodes](https://circleci.com/docs/parallelism/) and transferring all node data to a single master node.

## Installation

Add this line to your application's Gemfile:

```ruby
gem 'circleci-parallel'
```

And then execute:

```
$ bundle install
```

## Usage

Before using CircleCI::Parallel:

* [Add `parallel: true`](https://circleci.com/docs/parallel-manual-setup/)
  to the command that you'll use CircleCI::Parallel in your `circle.yml`.
* [Set up parallelism](https://circleci.com/docs/setting-up-parallelism/)
  for your project from the CircleCI web console.

CircleCI::Parallel uses SSH for syncing nodes and transferring data between nodes.
All data will be transferred to the master node (where [`CIRCLE_NODE_INDEX`](https://circleci.com/docs/1.0/parallel-manual-setup/#using-environment-variables) is `0`).

```yaml
# circle.yml
test:
  override:
    - ruby test.rb:
        parallel: true
```

```ruby
# test.rb
require 'circleci/parallel'

CircleCI::Parallel.configure do |config|
  # This hook will be invoked on every node.
  # The current working directory in this hook is set to the local data directory
  # where node specific data should be saved in.
  config.on_every_node.before_sync do
    data = { 'foo' => ENV['CIRCLE_NODE_INDEX'].to_i * 10 }
    json = JSON.generate(data)
    File.write('data.json', json)
  end

  # This hook will be invoked only on the master node after downloading all data from slave nodes.
  # The current working directory in this hook is set to the download data directory
  # where all node data are gathered into.
  # The directory structure on the master node will be the following:
  #
  #     .
  #     ├── node0
  #     │   └── node_specific_data_you_saved_on_node0.txt
  #     ├── node1
  #     │   └── node_specific_data_you_saved_on_node1.txt
  #     └── node2
  #         └── node_specific_data_you_saved_on_node2.txt
  config.on_master_node.after_download do
    merged_data = {}

    Dir.glob('*/data.json') do |path|
      json = File.read(path)
      data = JSON.parse(json)
      node_name = File.dirname(path)
      merged_data[node_name] = data
    end

    p merged_data
    # {
    #   "node0" => { "foo" => 0 },
    #   "node1" => { "foo" => 10 },
    #   "node2" => { "foo" => 20 }
    # }
  end
end

# Sync all nodes in the same build and gather all node data into the master node.
# Invoking this method blocks until the sync and the data transfer is complete.
CircleCI::Parallel.sync
```

See [API docs](http://www.rubydoc.info/gems/circleci-parallel) for more details.

## License

The gem is available as open source under the terms of the [MIT License](http://opensource.org/licenses/MIT).