Best way to transload a file in rails (from HTTP URL to S3)

Question

In Rails, what is the best (ie. efficient, elegant) way to download a file from a public HTTP url, upload to Amazon S3 and delete the file in the server.

I am using Heroku so I have the additional restriction of a ephemeral file system.

Answer 1

The answer is to read your image into memory instead of storing it on the filesystem. Here's an example.

s3 = Aws::S3::Resource.new
obj = s3.bucket(your_bucket_name).object(your_object_key)

s3_put_url = URI.parse(obj.presigned_url(:put))

image_url = 'http://www.google.com/google.jpg'   
image_file = open(image_url).read

Net::HTTP.start(s3_put_url.host) { |http| http.send_request('PUT', s3_put_url.request_uri, image_file); }

# Let's get the URL
s3.bucket(your_bucket_name).object(your_object_key).presigned_url(:get)

Answer 2

If reading into memory isn't an option. You can use the /tmp directory. As long as this is in the same process using that shouldn't be a problem.