Octokit

Note We've recently renamed the 4-stable branch to main. This might affect you if you're making changes to Octokit's code locally. For more details and for the steps to reconfigure your local clone for the new branch name, check out this post.

Ruby toolkit for the GitHub API.

Upgrading? Check the Upgrade Guide before bumping to a new major version.

1. Philosophy
2. Installation
3. Making requests
   1. Additional Query Parameters
4. Consuming resources
5. Accessing HTTP responses
6. Handling errors
7. Authentication
   1. Basic Authentication
   2. OAuth access tokens
   3. Two-Factor Authentication
   4. Using a .netrc file
   5. Application authentication
   6. GitHub App
8. Default results per_page
9. Pagination
   1. Auto pagination
10. Working with GitHub Enterprise
11. Interacting with the GitHub.com APIs in GitHub Enterprise
12. Interacting with the GitHub Enterprise Admin APIs
13. Interacting with the GitHub Enterprise Management Console APIs
14. SSL Connection Errors
15. Configuration and defaults
    1. Configuring module defaults
    2. Using ENV variables
    3. Timeouts
16. Hypermedia agent
    1. Hypermedia in Octokit
    2. URI templates
    3. The Full Hypermedia Experience™
17. Upgrading guide
    1. Upgrading from 1.x.x
18. Advanced usage
    1. Debugging
    2. Caching
19. Hacking on Octokit.rb
    1. Code of Conduct
    2. Running and writing new tests
20. Supported Ruby Versions
21. Versioning
22. Making Repeating Requests
23. License

API wrappers should reflect the idioms of the language in which they were written. Octokit.rb wraps the GitHub API in a flat API client that follows Ruby conventions and requires little knowledge of REST. Most methods have positional arguments for required input and an options hash for optional parameters, headers, or other options:

client = Octokit::Client.new

# Fetch a README with Accept header for HTML format
client.readme 'al3x/sovereign', :accept => 'application/vnd.github.html'

Install via Rubygems

gem install octokit

... or add to your Gemfile

gem "octokit"

Access the library in Ruby:

require 'octokit'

API methods are available as client instance methods.

# Provide authentication credentials
client = Octokit::Client.new(:access_token => 'personal_access_token')

# You can still use the username/password syntax by replacing the password value with your PAT.
# client = Octokit::Client.new(:login => 'defunkt', :password => 'personal_access_token')

# Fetch the current user
client.user

When passing additional parameters to GET based request use the following syntax:

 # query: { parameter_name: 'value' }
 # Example: Get repository listing by owner in ascending order
 client.repos({}, query: {type: 'owner', sort: 'asc'})

 # Example: Get contents of a repository by ref
 # https://api.github.com/repos/octokit/octokit.rb/contents/path/to/file.rb?ref=some-other-branch
 client.contents('octokit/octokit.rb', path: 'path/to/file.rb', query: {ref: 'some-other-branch'})

Most methods return a Resource object which provides dot notation and [] access for fields returned in the API response.

client = Octokit::Client.new

# Fetch a user
user = client.user 'jbarnette'
puts user.name
# => "John Barnette"
puts user.fields
# => <Set: {:login, :id, :gravatar_id, :type, :name, :company, :blog, :location, :email, :hireable, :bio, :public_repos, :followers, :following, :created_at, :updated_at, :public_gists}>
puts user[:company]
# => "GitHub"
user.rels[:gists].href
# => "https://api.github.com/users/jbarnette/gists"

Note: URL fields are culled into a separate .rels collection for easier Hypermedia support.

While most methods return a Resource object or a Boolean, sometimes you may need access to the raw HTTP response headers. You can access the last HTTP response with Client#last_response:

user      = client.user 'andrewpthorp'
response  = client.last_response
etag      = response.headers[:etag]

When the API returns an error response, Octokit will raise a Ruby exception.

A range of different exceptions can be raised depending on the error returned by the API - for example:

• A 400 Bad Request response will lead to an Octokit::BadRequest error
• A 403 Forbidden error with a "rate limited exceeded" message will lead to a Octokit::TooManyRequests error

All of the different exception classes inherit from Octokit::Error and expose the #response_status, #response_headers and #response_body. For validation errors, #errors will return an Array of Hashes with the detailed information returned by the API.

Octokit supports the various authentication methods supported by the GitHub API:

Using your GitHub username and password is the easiest way to get started making authenticated requests:

client = Octokit::Client.new(:login => 'defunkt', :password => 'c0d3b4ssssss!')

user = client.user
user.login
# => "defunkt"

While Basic Authentication allows you to get started quickly, OAuth access tokens are the preferred way to authenticate on behalf of users.

OAuth access tokens provide two main benefits over using your username and password:

• Revocable access. Access tokens can be revoked, removing access for only that token without having to change your password everywhere.
• Limited access. Access tokens have access scopes which allow for more granular access to API resources. For instance, you can grant a third party access to your gists but not your private repositories.

To use an access token with the Octokit client, pass your token in the :access_token options parameter in lieu of your username and password:

client = Octokit::Client.new(:access_token => "<your 40 char token>")

user = client.user
user.login
# => "defunkt"

You can create access tokens through your GitHub Account Settings.

Two-Factor Authentication brings added security to the account by requiring more information to login.

Using two-factor authentication for API calls is as simple as adding the required header as an option:

client = Octokit::Client.new \
  :login    => 'defunkt',
  :password => 'c0d3b4ssssss!'

user = client.user("defunkt", :headers => { "X-GitHub-OTP" => "<your 2FA token>" })

Octokit supports reading credentials from a netrc file (defaulting to ~/.netrc). Given these lines in your netrc:

machine api.github.com
  login defunkt
  password c0d3b4ssssss!

You can now create a client with those credentials:

client = Octokit::Client.new(:netrc => true)
client.login
# => "defunkt"

But I want to use OAuth you say. Since the GitHub API supports using an OAuth token as a Basic password, you totally can:

machine api.github.com
  login defunkt
  password <your 40 char token>

Note: Support for netrc requires adding the netrc gem to your Gemfile or .gemspec.

Octokit also supports application-only authentication using OAuth application client credentials. Using application credentials will result in making anonymous API calls on behalf of an application in order to take advantage of the higher rate limit.

client = Octokit::Client.new \
  :client_id     => "<your 20 char id>",
  :client_secret => "<your 40 char secret>"

user = client.user 'defunkt'

Octokit.rb also supports authentication using a GitHub App, which requires a generated JWT token.

client = Octokit::Client.new(:bearer_token => "<your jwt token>")
client.app
# => about GitHub App info

Default results from the GitHub API are 30, if you wish to add more you must do so during Octokit configuration.

Octokit::Client.new(access_token: "<your 40 char token>", per_page: 100)

Many GitHub API resources are paginated. While you may be tempted to start adding :page parameters to your calls, the API returns links to the next, previous, and last pages for you in the Link response header as Hypermedia link relations.

issues = client.issues 'rails/rails'
issues.concat client.get(client.last_response.rels[:next].href)

For smallish resource lists, Octokit provides auto pagination. When this is enabled, calls for paginated resources will fetch and concatenate the results from every page into a single array:

client.auto_paginate = true
issues = client.issues 'rails/rails'
issues.length

# => 702

You can also enable auto pagination for all Octokit client instances:

Octokit.configure do |c|
  c.auto_paginate = true
end

Note: While Octokit auto pagination will set the page size to the maximum 100, and seek to not overstep your rate limit, you probably want to use a custom pattern for traversing large lists.

With a bit of setup, you can also use Octokit with your GitHub Enterprise instance.

To interact with the "regular" GitHub.com APIs in GitHub Enterprise, simply configure the api_endpoint to match your hostname. For example:

Octokit.configure do |c|
  c.api_endpoint = "https://<hostname>/api/v3/"
end

client = Octokit::Client.new(:access_token => "<your 40 char token>")

The GitHub Enterprise Admin APIs are under a different client: EnterpriseAdminClient. You'll need to have an administrator account in order to use these APIs.

admin_client = Octokit::EnterpriseAdminClient.new(
  :access_token => "<your 40 char token>",
  :api_endpoint => "https://<hostname>/api/v3/"
)

# or
Octokit.configure do |c|
  c.api_endpoint = "https://<hostname>/api/v3/"
  c.access_token = "<your 40 char token>"
end

admin_client = Octokit.enterprise_admin_client.new

The GitHub Enterprise Management Console APIs are also under a separate client: EnterpriseManagementConsoleClient. In order to use it, you'll need to provide both your management console password as well as the endpoint to your management console. This is different from the API endpoint provided above.

management_console_client = Octokit::EnterpriseManagementConsoleClient.new(
  :management_console_password => "secret",
  :management_console_endpoint = "https://hostname:8633"
)

# or
Octokit.configure do |c|
  c.management_console_endpoint = "https://hostname:8633"
  c.management_console_password = "secret"
end

management_console_client = Octokit.enterprise_management_console_client.new

You may need to disable SSL temporarily while first setting up your GitHub Enterprise install. You can do that with the following configuration:

client.connection_options[:ssl] = { :verify => false }

Do remember to turn :verify back to true, as it's important for secure communication.

While Octokit::Client accepts a range of options when creating a new client instance, Octokit's configuration API allows you to set your configuration options at the module level. This is particularly handy if you're creating a number of client instances based on some shared defaults. Changing options affects new instances only and will not modify existing Octokit::Client instances created with previous options.

Every writable attribute in {Octokit::Configurable} can be set one at a time:

Octokit.api_endpoint = 'http://api.github.dev'
Octokit.web_endpoint = 'http://github.dev'

or in batch:

Octokit.configure do |c|
  c.api_endpoint = 'http://api.github.dev'
  c.web_endpoint = 'http://github.dev'
end

Default configuration values are specified in {Octokit::Default}. Many attributes will look for a default value from the ENV before returning Octokit's default.

# Given $OCTOKIT_API_ENDPOINT is "http://api.github.dev"
client.api_endpoint

# => "http://api.github.dev"

Deprecation warnings and API endpoints in development preview warnings are printed to STDOUT by default, these can be disabled by setting the ENV OCTOKIT_SILENT=true.

By default, Octokit does not timeout network requests. To set a timeout, pass in Faraday timeout settings to Octokit's connection_options setting.

Octokit.configure do |c|
  c.api_endpoint = ENV.fetch('GITHUB_API_ENDPOINT', 'https://api.github.com/')
  c.connection_options = {
    request: {
      open_timeout: 5,
      timeout: 5
    }
  }
end

You should set a timeout in order to avoid Ruby’s Timeout module, which can hose your server. Here are some resources for more information on this:

• The Oldest Bug In Ruby - Why Rack::Timeout Might Hose your Server
• Timeout: Ruby's Most Dangerous API
• The Ultimate Guide to Ruby Timeouts

Starting in version 2.0, Octokit is hypermedia-enabled. Under the hood, {Octokit::Client} uses Sawyer, a hypermedia client built on Faraday.

Resources returned by Octokit methods contain not only data but hypermedia link relations:

user = client.user 'technoweenie'

# Get the repos rel, returned from the API
# as repos_url in the resource
user.rels[:repos].href
# => "https://api.github.com/users/technoweenie/repos"

repos = user.rels[:repos].get.data
repos.last.name
# => "faraday-zeromq"

When processing API responses, all *_url attributes are culled into the link relations collection. Any url attribute becomes .rels[:self].

You might notice many link relations have variable placeholders. Octokit supports URI Templates for parameterized URI expansion:

repo = client.repo 'pengwynn/pingwynn'
rel = repo.rels[:issues]
# => #<Sawyer::Relation: issues: get https://api.github.com/repos/pengwynn/pingwynn/issues{/number}>

# Get a page of issues
rel.get.data

# Get issue #2
rel.get(:uri => {:number => 2}).data

If you want to use Octokit as a pure hypermedia API client, you can start at the API root and follow link relations from there:

root = client.root
root.rels[:repository].get :uri => {:owner => "octokit", :repo => "octokit.rb" }
root.rels[:user_repositories].get :uri => { :user => "octokit" },
                                  :query => { :type => "owner" }

Octokit 3.0 aims to be hypermedia-driven, removing the internal URL construction currently used throughout the client.

Version 4.0

• removes support for a long-deprecated overload for passing state as a positional argument when listing pull requests. Instead, pass state in the method options.
• drops support for Ruby < 2.0.
• adds support for new Enterprise-only APIs.
• adds support for Repository redirects.

Version 3.0 includes a couple breaking changes when upgrading from v2.x.x:

The default media type is now v3 instead of beta. If you need to request the older media type, you can set the default media type for the client:

Octokit.default_media_type = "application/vnd.github.beta+json"

or per-request

client.emails(:accept => "application/vnd.github.beta+json")

The long-deprecated Octokit::Client#create_download method has been removed.

Version 2.0 includes a completely rewritten Client factory that now memoizes client instances based on unique configuration options. Breaking changes also include:

• :oauth_token is now :access_token
• :auto_traversal is now :auto_paginate
• Hashie::Mash has been removed. Responses now return a Sawyer::Resource object. This new type behaves mostly like a Ruby Hash, but does not fully support the Hashie::Mash API.
• Two new client error types are raised where appropriate: Octokit::TooManyRequests and Octokit::TooManyLoginAttempts
• The search_* methods from v1.x are now found at legacy_search_*
• Support for netrc requires including the netrc gem in your Gemfile or gemspec.
• DateTime fields are now proper DateTime objects. Previous versions outputted DateTime fields as 'String' objects.

Since Octokit employs Faraday under the hood, some behavior can be extended via middleware.

Often, it helps to know what Octokit is doing under the hood. You can add a logger to the middleware that enables you to peek into the underlying HTTP traffic:

stack = Faraday::RackBuilder.new do |builder|
  builder.use Faraday::Retry::Middleware, exceptions: Faraday::Retry::Middleware::DEFAULT_EXCEPTIONS + [Octokit::ServerError] # or Faraday::Request::Retry for Faraday < 2.0
  builder.use Octokit::Middleware::FollowRedirects
  builder.use Octokit::Response::RaiseError
  builder.use Octokit::Response::FeedParser
  builder.response :logger do |logger|
    logger.filter(/(Authorization: "(token|Bearer) )(\w+)/, '\1[REMOVED]')
  end
  builder.adapter Faraday.default_adapter
end
Octokit.middleware = stack

client = Octokit::Client.new
client.user 'pengwynn'

I, [2013-08-22T15:54:38.583300 #88227]  INFO -- : get https://api.github.com/users/pengwynn
D, [2013-08-22T15:54:38.583401 #88227] DEBUG -- request: Accept: "application/vnd.github.beta+json"
User-Agent: "Octokit Ruby Gem 2.0.0.rc4"
I, [2013-08-22T15:54:38.843313 #88227]  INFO -- Status: 200
D, [2013-08-22T15:54:38.843459 #88227] DEBUG -- response: server: "GitHub.com"
date: "Thu, 22 Aug 2013 20:54:40 GMT"
content-type: "application/json; charset=utf-8"
transfer-encoding: "chunked"
connection: "close"
status: "200 OK"
x-ratelimit-limit: "60"
x-ratelimit-remaining: "39"
x-ratelimit-reset: "1377205443"
...

See the Faraday README for more middleware magic.

If you want to boost performance, stretch your API rate limit, or avoid paying the hypermedia tax, you can use Faraday Http Cache.

Add the gem to your Gemfile

gem 'faraday-http-cache'

Next, construct your own Faraday middleware:

stack = Faraday::RackBuilder.new do |builder|
  builder.use Faraday::HttpCache, serializer: Marshal, shared_cache: false
  builder.use Octokit::Response::RaiseError
  builder.adapter Faraday.default_adapter
end
Octokit.middleware = stack

Once configured, the middleware will store responses in cache based on ETag fingerprint and serve those back up for future 304 responses for the same resource. See the project README for advanced usage.

If you want to hack on Octokit locally, we try to make bootstrapping the project as painless as possible. To start hacking, clone and run:

script/bootstrap

This will install project dependencies and get you up and running. If you want to run a Ruby console to poke on Octokit, you can crank one up with:

script/console

Using the scripts in ./script instead of bundle exec rspec, bundle console, etc. ensures your dependencies are up-to-date.

We want both the Octokit.rb and larger Octokit communities to be open and welcoming environments. Please read and follow both in spirit and letter Code of Conduct.

Octokit uses VCR for recording and playing back API fixtures during test runs. These cassettes (fixtures) are part of the Git project in the spec/cassettes folder. If you're not recording new cassettes you can run the specs with existing cassettes with:

script/test

Octokit uses environmental variables for storing credentials used in testing. If you are testing an API endpoint that doesn't require authentication, you can get away without any additional configuration. For the most part, tests use an authenticated client, using a token stored in ENV['OCTOKIT_TEST_GITHUB_TOKEN']. There are several different authentication methods used across the api. Here is the full list of configurable environmental variables for testing Octokit:

Since we periodically refresh our cassettes, please keep some points in mind when writing new specs.

• Specs should be idempotent. The HTTP calls made during a spec should be able to be run over and over. This means deleting a known resource prior to creating it if the name has to be unique.
• Specs should be able to be run in random order. If a spec depends on another resource as a fixture, make sure that's created in the scope of the spec and not depend on a previous spec to create the data needed.
• Do not depend on authenticated user info. Instead of asserting actual values in resources, try to assert the existence of a key or that a response is an Array. We're testing the client, not the API.

This library aims to support and is tested against the following Ruby implementations:

• Ruby 2.7
• Ruby 3.0
• Ruby 3.1
• Ruby 3.2
• Ruby 3.3

If something doesn't work on one of these Ruby versions, it's a bug.

This library may inadvertently work (or seem to work) on other Ruby implementations, but support will only be provided for the versions listed above.

If you would like this library to support another Ruby version, you may volunteer to be a maintainer. Being a maintainer entails making sure all tests run and pass on that implementation. When something breaks on your implementation, you will be responsible for providing patches in a timely fashion. If critical issues for a particular implementation exist at the time of a major release, support for that Ruby version may be dropped.

This library aims to adhere to Semantic Versioning 2.0.0. Violations of this scheme should be reported as bugs. Specifically, if a minor or patch version is released that breaks backward compatibility, that version should be immediately yanked and/or a new version should be immediately released that restores compatibility. Breaking changes to the public API will only be introduced with new major versions. As a result of this policy, you can (and should) specify a dependency on this gem using the Pessimistic Version Constraint with two digits of precision. For example:

spec.add_dependency 'octokit', '~> 3.0'

The changes made between versions can be seen on the project releases page.

In most cases it would be best to use webhooks, but sometimes webhooks don't provide all of the information needed. In those cases where one might need to poll for progress or retry a request on failure, we designed Octopoller. Octopoller is a micro gem perfect for making repeating requests.

Octopoller.poll(timeout: 15.seconds) do
  begin
    client.request_progress # ex. request a long running job's status
  rescue Error
    :re_poll
  end
end

This is useful when making requests for a long running job's progress (ex. requesting a Source Import's progress).

Copyright (c) 2009-2014 Wynn Netherland, Adam Stacoviak, Erik Michaels-Ober

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.