Serve HTML files stored on S3 on a Rack app - ruby

Question

Say I have some HTML documents stored on S3 likes this:
http://alan.aws-s3-bla-bla.com/posts/1.html
http://alan.aws-s3-bla-bla.com/posts/2.html
http://alan.aws-s3-bla-bla.com/posts/3.html
http://alan.aws-s3-bla-bla.com/posts/1/comments/1.html
http://alan.aws-s3-bla-bla.com/posts/1/comments/2.html
http://alan.aws-s3-bla-bla.com/posts/1/comments/3.html
etc, etc
I'd like to serve these with a Rack (preferably Sinatra) application, mapping the following routes:
get "/posts/:id" do
render "http://alan.aws-s3-bla-bla.com/posts/#{params[:id]}.html"
end
get "/posts/:posts_id/comments/:comments_id" do
render "http://alan.aws-s3-bla-bla.com/posts/#{params[:posts_id]}/comments/#{params[:comments_id}.html"
end
Is this a good idea? How would I do it?

Answer 1

There would obviously be a wait while you grabbed the file, so you could cache it or set etags etc to help with that. I suppose it depends on how long you want to wait and how often it is accessed, its size etc as to whether it's worth storing the HTML locally or remotely. Only you can work that bit out.
If the last expression in the block is a string that will automatically be rendered, so there's no need to call render as long as you've opened the file as a string.
Here's how to grab an external file and put it into a tempfile:
require 'faraday'
require 'faraday_middleware'
#require 'faraday/adapter/typhoeus' # see https://github.com/typhoeus/typhoeus/issues/226#issuecomment-9919517 if you get a problem with the requiring
require 'typhoeus/adapters/faraday'
configure do
Faraday.default_connection = Faraday::Connection.new(
:headers => { :accept => 'text/plain', # maybe this is wrong
:user_agent => "Sinatra via Faraday"}
) do |conn|
conn.use Faraday::Adapter::Typhoeus
end
end
helpers do
def grab_external_html( url )
response = Faraday.get url # you'll need to supply this variable somehow, your choice
filename = url # perhaps change this a bit
tempfile = Tempfile.open(filename, 'wb') { |fp| fp.write(response.body) }
end
end
get "/posts/:whatever/" do
tempfile = grab_external_html whatever # surely you'd do a bit more here…
tempfile.read
end
This might work. You may also want to think about closing that tempfile, but the garbage collector and the OS should take care of it.