CODEWITHSUNDEEP
ruby-on-railsamazon-s3carrierwavefog
Cheyne
Cheyne

Reputation: 2137

CarrierWave S3 File Retrieve Issue - undefined method `body'

This is driving me nuts. Im using CarrierWave with Amazon S3, I can safely upload a file to the bucket, and Im trying to now (in a different request) retrieve the file and deliver it as a download to the browser.

The files can be anything, .zip file etc. Im not using it for image storage.

The Bucket is private, not public so I can't just use the S3 URL.

Here's my retrieve code:

au = @note.attachment
au.retrieve_from_store!(File.basename(@note.attachment.to_s))
au.cache_stored_file!

The code errors on the cache_stored_file! line. Throwing the error:

undefined method `body' for nil:NilClass

I inspect the object and it looks fine to me, and if I output @note.attachment, I can clearly see the amazon s3 URL with keys etc.

Iv been all over the web and I can figure it out. Iv found a handful of people with the same issue who solved it by doing things that are irrelevant to my case (like using some imagemagick method), again, these are just files.

No handling or image work, I just want to return the files to the browser as a download.

Can anyone please help?

Here is my uploader:

class AttachmentUploader < CarrierWave::Uploader::Base

  storage :fog

  def store_dir
    "#{Rails.env}/#{model.id}"
  end

  def cache_dir
    "#{Rails.root}/tmp/attachments"
  end
end

and the trace

carrierwave (0.8.0) lib/carrierwave/storage/fog.rb:225:in `read'
carrierwave (0.8.0) lib/carrierwave/uploader/cache.rb:77:in `sanitized_file'
carrierwave (0.8.0) lib/carrierwave/uploader/cache.rb:116:in `cache!'
carrierwave (0.8.0) lib/carrierwave/uploader/cache.rb:73:in `cache_stored_file!'

Upvotes: 3

Views: 5182

Answers (1)

Jesper
Jesper

Reputation: 4555

I think these lines don't make sense:

au = @note.attachment
au.retrieve_from_store!(File.basename(@note.attachment.to_s))
au.cache_stored_file!

retrieve_from_store is something you would typically call on the uploader, not on the attached object. The complicated step of passing the basename of the attachment to itself should make warning bells go off all over.

If I understand you correctly you simply want to give the user the possibility to download the file that you stored in S3.

You can do this is in a couple of ways.

The easiest is to call @note.attachment.url and print the url as a link that your user can click.

Another way is to redirect the user to the url in controller action. 

# notes/1/download_attachment   
def download_attachment
  @note = Note.find(params[:id)
  redirect_to @note.attachment.url
end

One last way is to act as a proxy server/cache, and fetch the file from S3 and return it.

You could do it something like this in your controller:

file = open(@note.attachment.url)
send_data file

Please note that I didn't have the chance to test this last code snippet.

You can read more about send_data here.

Upvotes: 4

Related Questions