Custom Paperclip::Processor with AWS S3 - ruby-on-rails

Question

This code works fine if the uploads are stored locally, but sometimes they are on S3 so it's not possible to just source: "#{File.expand_path(src.path)}[0]". How do I make the Paperclip's run method load images from S3 and replace them afterwards?
module Paperclip
class KskCrop < Processor
def initialize(file, options = {}, attachment = nil)
super
#crop = options
#format = File.extname(#file.path)
#basename = File.basename(#file.path, #format)
end
def make
src = #file
dst = Tempfile.new([#basename, #format])
dst.binmode
parameters = []
parameters << ":source"
parameters << "-crop '#{#crop[2]}x#{#crop[3]}+#{#crop[0]}+#{#crop[1]}'"
parameters << ":dest"
parameters = parameters.flatten.compact.join(' ').strip.squeeze(' ')
success = Paperclip.run('convert', parameters, source: "#{File.expand_path(src.path)}[0]", dest: File.expand_path(dst.path))
dst
end
end
end

Answer 1

So, as per the conversation. It turned out that path has to be a complete url instead of a relative path. Hence, this line will look something like this:
success = Paperclip.run('convert', parameters, source: src.url, dest: File.expand_path(dst.path)
As, OP has already answered his question with an if.. else. I think that's a better approach to check if the attachment is on a local file or on a CDN.
P.S.: I also learned that Paperclip.run(..) method can actually locate and download a file for processing if it's a url without making an IO operations at developer's end.

Answer 2

well i didnt got your question well but for Amazon S3 READ/WRITE, this is what is use...
this is my s3.yml
development:
bucket: app_development
access_key_id: xxxxxxxxxxxxxx
secret_access_key: xxxxxxxxxxxxx+xxxxxxxxxxx
production:
bucket: app_production
access_key_id: xxxxxxxxxxxxxxxxx
secret_access_key: xxxxxxxxxxxx+xxxxxxxxxxxxx
my config/initializers/paperclip.rb
Paperclip::Attachment.default_options[:url] = ':s3_domain_url'
Paperclip::Attachment.default_options[:path] = '/:class/:id/:style/:filename'
##will store in the foll way shown below in the bucket u specify in s3.yml
##http://app_development.s3.amazonaws.com/videos/1/original/small.mp4
##if Rails.env == "production"
#S3_CREDENTIALS = { :access_key_id => ENV['S3_KEY'], :secret_access_key => ENV['S3_SECRET'], :bucket => "ourbucket"}
##else
S3_CREDENTIALS = Rails.root.join("config/s3.yml")
##end
so after this configuration,any paperclip model with has_attached_file:avatar will get uploaded on s3
##upload to s3..code snippet (can also use in _form.html.erb as f.file_field :avatar)
#video=Video.create!({ :avatar => File.new("#{Rails.root}/app/assets/images/dashboard/default_avatar.jpg")
})
##this is how u can display/access any uploaded object/model using url(which shows original unless you specify some styles -thumbnail,small)
#video.avatar.url
#video.avatar.url(:thumbnail)
#video.avatar.url(:small)

Answer 3

Looks like this is the solution!
module Paperclip
class KskCrop < Processor
def initialize(file, options = {}, attachment = nil)
super
#crop = options
#format = File.extname(#file.path)
#basename = File.basename(#file.path, #format)
end
def make
src = #file
dst = Tempfile.new([#basename, #format])
dst.binmode
parameters = []
parameters << ":source"
parameters << "-crop '#{#crop[2]}x#{#crop[3]}+#{#crop[0]}+#{#crop[1]}'"
parameters << ":dest"
parameters = parameters.flatten.compact.join(' ').strip.squeeze(' ')
path = if #file.options && #file.options[:storage] == :s3
src.url
else
File.expand_path(src.path)
end
success = Paperclip.run('convert', parameters, source: path, dest: File.expand_path(dst.path))
dst
end
end
end