#!/usr/bin/env ruby

require 'csv'

def darwin?
  @darwin ||= `uname`.strip == 'Darwin'
end

def mktemp
  if darwin?
    `mktemp -t $$`.strip
  else
    `mktemp`.strip
  end
end

def delimit(n, delimiter = ',')
  n.to_s.gsub(/(\d)(?=(\d\d\d)+(?!\d))/, "\\1#{delimiter}")
end

Units = %w[KB MB GB TB]

def human_size(n)
  unit = 'bytes'
  units = Units.dup
  while n > 1024 && units.length > 0
    n /= 1024.0
    unit = units.shift
  end
  "#{delimit(n.round(2))} #{unit}"
end

def analyze_rdb(rdbfile, keyfile)
  size = human_size(File.stat(rdbfile).size)
  puts "Getting key info for #{rdbfile} (#{size}), this can take a while..."
  `rdb -c memory #{rdbfile} > #{keyfile}`
  keyfile
end

def sort_keys(rdbfile)
  keyfile = rdbfile.sub('rdb', 'csv')
  sortedkeyfile = rdbfile.sub('.rdb', '-sorted.csv')
  if !File.exists?(keyfile)
    analyze_rdb(rdbfile, keyfile)
  end
  lines = File.readlines(keyfile)
  lines.shift # drop the header
  n = lines.length
  print "Munging #{delimit(n)} rows..."
  rows = []
  while line = lines.shift
  	row = line.strip.split(',')
  	size = row[3].to_i
  	key = row[2].sub(/^"/, '').sub(/"$/, '')
  	type = row[1]
  	num_elements = row[5]
  	len_largest = row[6]
  	row = [size, key, type, num_elements, len_largest]
  	rows << row
  	if rows.length % (n / 50) == 0
  	  print '.'
  	end
  end
  puts
  puts "Sorting..."
  rows = rows.sort { |a,b| b[0] <=> a[0] }
  print "Writing..."
  CSV.open(sortedkeyfile, 'wb') do |csv|
    # header
    csv << %w[size key type num_elements len_largest_element]
    while row = rows.shift
      csv << row
      if rows.length % (n / 50) == 0
        print '.'
      end
    end
    puts
    puts "Done."
  end
end

def main
  sort_keys(ARGV.shift)
end

main if $0 == __FILE__