I have this working locally, storing the template files in #{Rails.root}/tmp
, using system "cd tmp/template; zip -r ../#{@filename} *"
to zip up the files, sending the .docx (zip archive) to S3 and then to the browser. The problem is that Heroku is not finding the files. Before I create the xml file, I am copying the template directory from another location (system "cp -R support/ser_template tmp/"
). I understand Heroku's read-only filesystem but I can't have #{Process.pid}
's littering my filenames (Word requires the xml file to be named document.xml).
Is there anyway I can store the template files on Amazon and still use Heroku's system zip utility? RubyZip does not create proper docx archives.
Edit: here is the code:
require 'aws/s3'
class WordDocument
include ConnectS3
def initialize(content)
connect_s3
@pid = Process.pid
@filename = "SER_" + Time.now.strftime("%Y%m%d-%H%M%S") + '.docx'
system "cp -R #{Rails.root}/support/ser_template #{temp_path}"
xml = File.open(xml_path, 'w')
xml.puts content
xml.close
system "cd #{temp_path}; zip -r #{@filename} *"
docx = File.open(temp_path + "/" + @filename, 'r')
AWS::S3::S3Object.store(s3_path, docx, @s3_credentials["bucket"], :use_virtual_directories => true)
AWS::S3::S3Object.grant_torrent_access_to s3_path, @s3_credentials["bucket"]
end
def temp_path
"#{Rails.root}/tmp/#{@pid}_ser"
end
def xml_path
temp_path + "/word/document.xml"
end
def path
"https://s3.amazonaws.com/" + @s3_credentials["bucket"] + s3_path
end
def s3_path
'/section_editor_reports/' + @filename
end
end