Question

Im referring to this app link as a tutorial. After looking at it. Im curious to understand how the program (client or server)can accesse images. Since the program doesn’t seem to be explicitely saving image urls in the server’s database once they are uploaded to Amazon S3?

After implementing this app, I dont see a column for image_urls or path in the database. But the program magically loads images in the client!!

Something is going on under the hood, what is that something?

Here is the Photo model method:

 Paperclip.interpolates :prefix  do |attachment, style|
   "#{attachment.instance.takenby}/#{Date.today.to_s }/#   
  {attachment.instance.image_file_name}"
 end 

 has_attached_file :image,
                :path => ":prefix/:style/:basename.:extension",
                :styles => { :thumbnail => "57x57", :original => "300x300" },
                :storage => :s3,
                :s3_credentials => S3_CREDENTIALS


  validates :image,:presence => true
  validates :lat, :lng,:presence => true,:numericality => true

Answer 1

It’s right there in the options for has_attached_file. You’ve told it how to construct the path for the image (which corresponds to its filename on s3) in the :path argument. So all paperclip needs to know to find the image again is:

1. The host name for your s3 bucket
2. The relevant information to re-construct the path – in this case, that looks like :prefix, :basename, :style, and :extension.

Most of that information is in your database. Presumably :basename comes from the original filename somehow, which ought to be saved in the image_file_name field of the model to which the image is attached. :style depends on which image size you’re looking up at runtime. :extension can be determined from the image_content_type attribute.

:prefix is a little bit trickier. In fact, I worry that your interpolation rule will break your image lookup, because of the Date.today it uses to construct the prefix. Do you have any images more than one day old? If so, do they still work? I worry that when the image is uploaded, it will have a filename containing that day’s date, and then when you go to find them in the future you’ll reconstruct the url using the new current date, and wind up with a 403 error.

TL;DR paperclip constructs the image url when it uploads the images to s3, and reconstructs them later from the same parameters.