6.6 KiB

Raw Blame History

Google API Client

Homepage: http://www.github.com/google/google-api-ruby-client
Authors: Bob Aman, Steven Bazyl
Copyright: Copyright © 2011 Google, Inc.
License: Apache 2.0

Description

The Google API Ruby Client makes it trivial to discover and access supported APIs.

Install

Be sure https://rubygems.org/ is in your gem sources.

For normal client usage, this is sufficient:

$ gem install google-api-client

Example Usage

require 'google/api_client'
require 'google/api_client/client_secrets'
require 'google/api_client/auth/installed_app'

# Initialize the client.
client = Google::APIClient.new(
  :application_name => 'Example Ruby application',
  :application_version => '1.0.0'
)

# Initialize Google+ API. Note this will make a request to the
# discovery service every time, so be sure to use serialization
# in your production code. Check the samples for more details.
plus = client.discovered_api('plus')

# Load client secrets from your client_secrets.json.
client_secrets = Google::APIClient::ClientSecrets.load

# Run installed application flow. Check the samples for a more
# complete example that saves the credentials between runs.
flow = Google::APIClient::InstalledAppFlow.new(
  :client_id => client_secrets.client_id,
  :client_secret => client_secrets.client_secret,
  :scope => ['https://www.googleapis.com/auth/plus.me']
)
client.authorization = flow.authorize

# Make an API call.
result = client.execute(
  :api_method => plus.activities.list,
  :parameters => {'collection' => 'public', 'userId' => 'me'}
)

puts result.data

API Features

API Discovery

To take full advantage of the client, load API definitions prior to use. To load an API:

urlshortener = client.discovered_api('urlshortener')

Specific versions of the API can be loaded as well:

drive = client.discovered_api('drive', 'v2')

Locally cached discovery documents may be used as well. To load an API from a local file:

doc = File.read('my-api.json')
my_api = client.register_discovery_document('myapi', 'v1', doc)

Authorization

Most interactions with Google APIs require users to authorize applications via OAuth 2.0. The client library uses Signet to handle most aspects of authorization. For additional details about Google's OAuth support, see Google Developers.

Credentials can be managed at the connection level, as shown, or supplied on a per-request basis when calling execute.

For server-to-server interactions, like those between a web application and Google Cloud Storage, Prediction, or BigQuery APIs, use service accounts.

key = Google::APIClient::KeyUtils.load_from_pkcs12('client.p12', 'notasecret')
client.authorization = Signet::OAuth2::Client.new(
  :token_credential_uri => 'https://accounts.google.com/o/oauth2/token',
  :audience => 'https://accounts.google.com/o/oauth2/token',
  :scope => 'https://www.googleapis.com/auth/prediction',
  :issuer => '123456-abcdef@developer.gserviceaccount.com',
  :signing_key => key)
client.authorization.fetch_access_token!
client.execute(...)

Service accounts are also used for delegation in Google Apps domains. The target user for impersonation is specified by setting the :person parameter to the user's email address in the credentials. Detailed instructions on how to enable delegation for your domain can be found at developers.google.com.

Automatic Retries & Backoff

The API client can automatically retry requests for recoverable errors. To enable retries, set the client.retries property to the number of additional attempts. To avoid flooding servers, retries invovle a 1 second delay that increases on each subsequent retry.

The default value for retries is 0, but will be enabled by default in future releases.

Batching Requests

Some Google APIs support batching requests into a single HTTP request. Use Google::APIClient::BatchRequest to bundle multiple requests together.

Example:

client = Google::APIClient.new
urlshortener = client.discovered_api('urlshortner')

batch = Google::APIClient::BatchRequest.new do |result|
    puts result.data
end

batch.add(:api_method => urlshortener.url.insert,
          :body_object => { 'longUrl' => 'http://example.com/foo' })
batch.add(:api_method => urlshortener.url.insert,
          :body_object => { 'longUrl' => 'http://example.com/bar' })
client.execute(batch)

Blocks for handling responses can be specified either at the batch level or when adding an individual API call. For example:

batch.add(:api_method=>urlshortener.url.insert, :body_object => { 'longUrl' => 'http://example.com/bar' }) do |result|
   puts result.data
end

Media Upload

For APIs that support file uploads, use Google::APIClient::UploadIO to load the stream. Both multipart and resumable uploads can be used. For example, to upload a file to Google Drive using multipart

drive = client.discovered_api('drive', 'v2')

media = Google::APIClient::UploadIO.new('mymovie.m4v', 'video/mp4')
metadata = {
    'title' => 'My movie',
    'description' => 'The best home movie ever made'
}
client.execute(:api_method => drive.files.insert,
               :parameters => { 'uploadType' => 'multipart' },
               :body_object => metadata,
               :media => media )

To use resumable uploads, change the uploadType parameter to resumable. To check the status of the upload and continue if necessary, check result.resumable_upload.

client.execute(:api_method => drive.files.insert,
           :parameters => { 'uploadType' => 'resumable' },
           :body_object => metadata,
           :media => media )
upload = result.resumable_upload

# Resume if needed
if upload.resumable?
    client.execute(upload)
end

Samples

See the full list of samples on Github.

Support

Please report bugs at the project on Github. Don't hesitate to ask questions about the client or APIs on StackOverflow.

6.6 KiB Raw Blame History