gistfile1.rb

require 'rubygems'
require 'apachelogregex'
require 'ostruct'

@requests = {}

def to_date(d)
  Date.strptime d.gsub('[', '').gsub(']', ''), "%d/%b/%Y"
end

def add_request(ip, line)
  key = ip
  if @requests.key?(key)
    @requests[key] << line
  else
    @requests[key] = [line]
  end
end

def page?(l)
  %w(.css .js .json .gif .png .jpg .pdf .swf).detect {|ext| l.include?(ext)}.nil?
end

format = '%h %l %u %t \"%r\" %>s %b \"%{Referer}i\" \"%{User-Agent}i\"'
parser = ApacheLogRegex.new format

file = File.new 'rsff-files-only.log'
key = nil

bounces = 0
request_count = 0
while file.gets
  line = parser.parse($_)
  if line
    if key.nil? or key != line['%h']
      if request_count == 1
        bounces += 1
        puts "bounce"
      end
      key = line['%h']
      request_count = 1
    else
      request_count += 1
    end
  end
end

puts "BOUNCES #{bounces}"