Google-api-ruby-client Translate API Examples

Question

I was very happy to see the google code: google-api-ruby-client project, because it meant to me that Ruby people can polish code with the Google API-s.

For now though I'm stumped because the only example given uses Buzz and from my experiments, the Google Translate (v2) api must behave quite differently to Buzz in the google-api-ruby-client.

I was intrigued by the 'Explorer' demo example -- But it isn't much of an explorer as far as I can see. All it does is call up a Buzz service and then pokes about in things it ALREADY knows about Buzz the services. To me, an explorer ought to let you 'discover' the services and the methods/functions exposed without necessarily knowing them already.

I'd love to hear of Ruby command line and desktop applications using this: google-api-ruby-client for services other than Buzz and in particular the Translate api (I'm less interested in the existing Ruby gems using the translate service at this point).

thanks ... will

Answer 1

Code for making calls to the translate API looks like this:

require 'google/api_client'
client = Google::APIClient.new(:key => YOUR_DEVELOPER_KEY)
translate = client.discovered_api('translate', 'v2')
result = client.execute(
  :api_method => translate.translations.list,
  :parameters => {
    'format' => 'text',
    'source' => 'en',
    'target' => 'es',
    'q' => 'The quick brown fox jumped over the lazy dog.'
  }
)

Answer 2

I posted up full code and detail for auth issues and workarounds (using an api key) at the code abode - google-api-client for ruby

After installing the gem, getting a Google API Key, and setting up a custom search account (with its prefs widened to all web pages).... I could trawl google search results in irb with the following (copy paste into irb, then inspect response when finished):

  require 'openssl'
  OpenSSL::SSL::VERIFY_PEER = OpenSSL::SSL::VERIFY_NONE

  require 'google/api_client'
  client = Google::APIClient.new(:key => 'your-api-key', :authorization => nil)
  search = client.discovered_api('customsearch')

  response = client.execute(
    :api_method => search.cse.list,
    :parameters => {
      'q' => 'the hoff',
      'key' => 'your-api-key',
      'cx' => 'your-custom-search-id'
    }
  )

This is to get server access to the google api and bypass all the oauth stuff. THE MOST IMPORTANT BIT was the :authorization param when constructing he client.... this ensures the api key is used when calling, in preference to oauth. Without it you will get 401 Unauthorized response status everytime.

Answer 3

Thanks to Bob, I'm one step forward.

I'm now finding trouble with the .CRT certificates file. I have a CA-bundle I got from the CA-bundle generator available on the cURL site (http://curl.haxx.se/ca).

Read also: How to Cure NetHhttps Risky Default HTTPS Behavior

I think the next question needs to be about finding the bundle for OpenSSL.

w.

Answer 4

Hi (all) interested folks,

More progress. I put a bug report for Google-API-Ruby-client for part of this problem. The short version is that for Translate (at least) the Google-API-Ruby-client fails under SSL::VERIFY_PEER and succeeds when SSL::VERIFY_NONE is used.

In a nut shell there's one problem to fix and two enhancements needed.

1. An API-client interface layer should report an Error or Warning when the underlying layers (as can be known or detected) invalidate security, identity-privacy and data integrity.
2. Report on relevant configuration information for the underlying service used (e.g. was SSL used? Was it VERIFY_PEER or VERIFY_NONE? What is the fully qualified URI to give the error? For me, I think that info can do to a log-file of sorts and be 'Requested' by an Option on the service class (class methods) and by the instance variable.
3. Provide diagnostics with your Gem design and extend diagnostics into your own code (e.g. metrics like: number of network 'execute' calls made, number of bytes in and bytes out).

Answer 5

I have a solution here (http://jjinux.blogspot.com/2012/02/ruby-working-around-ssl-errors-on-os-x.html) that monkey patches Net::HTTP#use_ssl= in order to use the operating system's root certificates.