---
title: Paperclip Tips and Updates
teaser:
tags: news,web,open source,rails,paperclip
author: Jon Yurek
published_on: 2008-05-15
---

So there's been a bit of activity on the
[Paperclip](https://github.com/thoughtbot/paperclip) front. I've added
[Amazon S3](https://www.amazon.com/gp/browse.html?node=16427261) support using
the [RightAWS](http://rightaws.rubyforge.org/) gem. Some fellow
[githubbers](httsp://github.com/thoughtbot/paperclip/tree/master) have
contributed some patches to fix up said S3 support, as well as to add more (and
better) validations for content type, etc. We've brought it up to a nice, round
v2.1.2 as of yesterday. Give yourselves a round of applause.

I highly encourage you to give it a whirl if you haven't already. It's very easy
to both get and to use!

**An interesting note!**

Ken Robertson has gone to great lengths to [port Paperclip to
DataMapper](https://github.com/krobertson/dm-paperclip/tree/master), which should
be great news for all you Merbers out there. He's kept it quite up-to-date, and
it's right alongside the 2.1.2 current release. Hopefully we'll be able to merge
codebases in the future, though right now there's enough of a difference between
DM and AR to make that not terribly feasible.

**Here's a handy tip!**

One of the most frequently asked questions is how to use data from your
instances in the path and/or <abbr title="Uniform Resource Locator">URL</abbr>.
The answer is the interpolations hash, which is quite user-extensible. Let's say
you had a song that played in the background of each User's profile (making this
exercise purely hypothetical, of course), and you wanted its name to be the same
as the User's username.

One really good way of allowing this functionality would be to add the following
to your `config/initializers/paperclip.rb` file:

```ruby
Paperclip::Attachment.interpolations[:username] = proc do |attachment, style|
  # or whatever you've named your User's login/username/etc. attribute
  attachment.instance.login
end
```

The #instance method is the instance of the model that this attachment is
attached to. You can access any of the model's attributes, methods,
associations, and so on from that object just like normal. Also note that you
don't have to name your interpolation the same thing as the attribute you're
interpolating (though it helps clarity).

Now you can add this to your :path and :url parameters like so:

```ruby
class User < ActiveRecord::Base
  has_attached_file :song,
                    :path => ":rails_root/public/system/:attachment/:username.:extension",
                    :url  => "/:attachment/:username.:extension"
end
```

When you call #url or #path on your attachment, Paperclip will run through the
interpolations hash, find strings that match its keys, and replace them with the
return values of the procs. In this case, you'd produce a lovely url of
/songs/jyurek.mp3

It's a very easy way to add flexibility to your files' names without having to
modify any code yourself.

**Another handy tip!**

S3 support is now baked in! The updated `has_attached_file` call looks kinda
like this one does:

```ruby
class House < ActiveRecord::Base
  has_attached_file :blueprint,
                    :styles => { :thumbnail => "150x150" },
                    :path => ":attachment/:id/:style.:extension",
                    :storage => :s3,
                    :s3_credentials => "#{RAILS_ROOT}/config/s3.yml",
                    :bucket => "bucket-of-holding"
end
```

Please note the additional options, including:

* `:storage`: Specify `:s3` here to use S3. Strictly speaking you could specify
  `:filesystem` here if you're using the filesystem, but it's the default so
  don't bother.
* `:s3_credentials`: You should give this a Hash, a path to a file, or a File
  object. The File should contain a <abbr title="Yet Another Markup
  Language">YAML</abbr>-ized Hash, and the contents of the Hash should be the
  access_key_id and secret_access_key used to access your S3 account. You can
  also environment-space these inside the hash, just like your database.yml
  file.
* `:bucket`: The name of the S3 bucket that will be holding your data.

Note that there's no `:url` option. That's because the S3 <abbr title="Uniform
Resource Locator">URL</abbr>s are generated from the bucket and the path name,
so you don't have to worry about them. You can also specify permissions for your
files, in case you don't want the default public-read, by using
`:s3_permissions`.

You can find more about the [S3 Storage
options](http://dev.thoughtbot.com/paperclip/classes/Paperclip/Storage/S3.html)
and the [Filesystem Storage
options](http://dev.thoughtbot.com/paperclip/classes/Paperclip/Storage/Filesystem.html)
at their RDocs.

**Keep up to date!**

Remember, there's always the [Paperclip Google
Group](http://groups.google.com/group/paperclip-plugin) and the [Paperclip
Lighthouse Account](http://thoughtbot.lighthouseapp.com/projects/8794-paperclip)
in case you have problems, questions, or feature requests.

Visit our [Open Source page](https://thoughtbot.com/open-source) to learn more about our team's contributions.