ruby-large-files.rb

#!/usr/bin/env ruby
# encoding: utf-8

begin
  require 'rainbow'
rescue LoadError
  puts "Rainbow is not installed"
  puts "  gem install rainbow"
  exit
end

class Integer
  def to_filesize
    {
      'B'  => 1024,
      'KB' => 1024 * 1024,
      'MB' => 1024 * 1024 * 1024,
      'GB' => 1024 * 1024 * 1024 * 1024,
      'TB' => 1024 * 1024 * 1024 * 1024 * 1024
    }.each_pair { |e, s| return "#{(self.to_f / (s / 1024)).round(3)}#{e}" if self < s }
  end
end

class DisplayItem
  attr_accessor :size, :location, :type
  
  @@base_dir=''

  def initialize(size, location, type)
    @size = size
    @location = location.strip
    @type = type
  end

  def self.for_file(line)
    toks = line.split(':')
    DisplayItem.new(toks[0].to_i, toks[1..-1].join(':'), :file)
  end

  def self.for_folder(line)
    toks = line.split(' ')
    DisplayItem.new(toks[0].to_i, toks[1..-1].join(' '), :folder)
  end

  def self.set_base_dir(dir)
    @@base_dir = dir
  end

  def to_s
    display_size + display_location
  end

  def display_size
    Rainbow(@size.to_filesize.ljust(12)).color(display_size_color)
  end

  def display_size_color
    if @size > 100 * 1024**2
      :red
    else
      :yellow
    end
  end

  def display_location
    loc = @location.sub(@@base_dir, '.')
    if @type == :folder
      Rainbow(loc).green
    else
      loc
    end

  end
end

dir = Dir.pwd
DisplayItem.set_base_dir dir

puts "Scanning    #{dir}..."

fldr  = `command du -ab #{dir} | sort -n -r | head -n 10`.split("\n")
files = %x[find #{dir} -type f -size +10000k -exec ls -l "{}" "\;" | awk '{print \$5 \": \" \$9}'].split("\n")

items = []

fldr.each { |f| items << DisplayItem.for_folder(f) }
files.each { |f| items << DisplayItem.for_file(f) }

items.sort! {|a,b| b.size <=> a.size }

puts items