s3etag.rb

#!/usr/bin/env ruby

require 'digest/md5'
require 'optparse'
require 'stringio'

module S3Etag

  def calc opt = {}
    threshold = opt[:threshold] || 16 * 1024 * 1024
    max_parts = opt[:max_parts] || 10000
    min_part_size = opt[:min_part_size] || 5 * 1024 * 1024

    if opt[:file]
      total_size = File.stat(opt[:file]).size
      io = open(opt[:file])
    elsif opt[:data]
      total_size = opt[:data].size
      io = StringIO.new opt[:data]
    else
      raise ':file or :data is required'
    end

    r = nil
    begin
      if total_size <= threshold
        r = Digest::MD5.hexdigest(io.read)
      else
        part_size = [(total_size.to_f / max_parts).ceil, min_part_size].max.to_i
        md5s = []
        while !io.eof?
          md5s << Digest::MD5.digest(io.read(part_size))
        end
        r = Digest::MD5.hexdigest(md5s.join('')) + '-' + md5s.size.to_s
      end
    rescue Exception => e
      raise e
    ensure
      io.close
    end
  end

  module_function :calc
end

if __FILE__ == $0
  opt = {}
  op = OptionParser.new
  op.on('-t', '--threshold threshold'){ |v| opt[:threshold] = v.to_i }
  op.on('-p', '--max-parts max-parts'){ |v| opt[:max_parts] = v.to_i }
  op.on('-s', '--min_part_size min_part_size'){ |v| opt[:min_part_size] = v.to_i }

  path = ARGV.last
  if path && File.file?(path)
    begin
      op.parse ARGV
    rescue OptionParser::InvalidOption => e
      puts op.summarize(['invalid option.', 's3etag file'], 100)
    else
      puts S3Etag.calc(opt.merge(:file => path))
    end
  else
    puts op.summarize(['s3etag file'], 100)
  end
end