Lists top directories by file size

top_dirs_by_size.rb

#! /usr/bin/env ruby

require 'shellwords'

$cache = {}

def file_sizes_in(path)
  $cache[path] ||= begin
    raw = `du -d 1 2>/dev/null #{Shellwords.shellescape(path)}`
    # exclude current dir
    raw.split("\n").map do |line|
      size, sub_path = line.squeeze(" ").split(" ", 2)
      [size.to_i, sub_path, File.directory?(sub_path)]
    end
  end
end

def filesize(size)
  size *= 512 # convert blocks to bytes
  units = %w[B KiB MiB GiB TiB Pib EiB ZiB]

  return '0.0 B' if size == 0
  exp = (Math.log(size) / Math.log(1024)).to_i
  exp += 1 if (size.to_f / 1024 ** exp >= 1024 - 0.05)
  exp = units.size - 1 if exp > units.size - 1

  '%.1f %s' % [size.to_f / 1024 ** exp, units[exp]]
end

def do_it(current_path, rank)
  puts "Starting at #{current_path}"

  loop do
    STDOUT.write("Calculating size of files and folders in #{current_path}...")
    size_info = file_sizes_in(current_path)
    current_info = size_info.delete_at(-1)
    size_info.sort_by! { |entry| entry[0] }
    puts "\r\e[K#{current_path} contains #{filesize(current_info[0])}"
    biggest = size_info.select { |entry| entry[2] }[-rank]
    break unless biggest
    puts "Largest subdirectory is #{biggest[1]} at #{filesize(biggest[0])}"
    current_path = biggest[1]
  end
end

current_path = ARGV[0] || ENV["HOME"]

do_it(current_path, 1)

size_info = file_sizes_in(current_path)
second_largest = size_info[0..-2].sort_by { |entry| entry[0] }[-2]

puts "\nTraversing second-largest directory in #{current_path}"
do_it(second_largest[1], 1)