Question

I having issues uploading large files form my Linux Desktop to GDrive (by using Grive which does not offer recovery or even monitoring of larger uploads).

So I came up trying to use google-api-ruby to write a small script doing the job.

As Google drives documentation talks about Resumable uploads – https://developers.google.com/drive/manage-uploads#resumable – I was trying to implement Resumable uploads using google-api-ruby and found Google::APIClient::ResumableUploadclass.

Samples are quite useful to understand the principle of how this rubyish API works, but the samples don’t broach the issuse of Resumable uploads.

Is Google::APIClient::ResumableUploadclass ment to be used for Resumable uploads the documenation is talking about or rather a conveicence class provided by this Ruby client?

Google::APIClient::ResumableUpload constructor consumes a Google::APIClient::Resultwhich I get, when I call execute or execute! on a Google::APIClient instance.

As far as I understand execute method will only work on Google API methods which are exposed to Discovery API (such as drive.files.insert).

Unfortunalty Resumable uploads are tied to following URI:

https://www.googleapis.com/upload/drive/v2/files?uploadType=resumable

and does not seem to be part of Discovery yet or integegrated into a ‘cleaner’ URI scheme (why upload/drive instead of drive/upload?).

Is it possible to use Google::APIClient execute methods for this API URI in order to combine it with Google::APIClient::ResumableUpload oder do I have to implement Resumable uploads by myself (e.g. by using em-http-request)?

Answer 1

See the example for files.insert for the basics. In that example, it is using multipart, but switching to resumable is fairly straightforward. Mostly requires changing the uploadType parameter to ‘resumable’. The result from the insert/update will contain a reference to the uploader than you can use to send the content/check for completion.

media = Google::APIClient::UploadIO.new(file_name, mime_type)
result = client.execute(
  :api_method => drive.files.insert,
  :body_object => file,
  :media => media,
  :parameters => {
    'uploadType' => 'resumable',
    'alt' => 'json'})

# Send content
result.resumable_upload.send_all(client)

In the upcoming beta release of the client library (any day now :), this will change slightly so that uploads work a little more uniformly regardless of protocol. Going forward in most cases, calling execute() will be sufficient and it will attempt to upload the file. But the existing method will still work. You’ll also be able resume by just calling:

if !result.resumable_upload.complete?
   client.execute(result.resumable_upload) # Continue sending... 
end

Example doesn’t show the error handling, but you can check result.resumable_upload.complete? or .expired? to check the status. Incomplete resumable uploads expire after a period of inactivity (on the order of an hour or so.) If expired, you’ll need to restart from the beginning.