pick 300 random 640x480 flickr pics and save them to current directory

pick300-480.rb

#!/usr/bin/env ruby
require 'json'
require 'pp'
require 'fileutils'
require 'getoptlong'
# requires serialized flickr json file to be $stdin or specified on the command line and then
# copies flickr "Z" files which are 640x480 to the current directory 
# for 5MP original somehow height is 481 (flickr bug?) so 481 height jpgs are ignored 
# 1 parameter is required: "-d directory" where directory is the directory with the Z files

opts = GetoptLong.new(
  ["--directory", "-d", GetoptLong::REQUIRED_ARGUMENT ]
)
flickr_z_dir = nil
opts.each do |opt, arg|
  flickr_z_dir = Dir.new(arg)
end
num_480 = 0
potential_files_to_copy = []
srand

ARGF.each_line do |line|
  serializedJSON = line
  flickr_data_page =  JSON.parse(serializedJSON)
  total  = flickr_data_page["photos"]["total"].to_i
  page  = flickr_data_page["photos"]["page"].to_i
  total_pages = flickr_data_page["photos"]["pages"].to_i
  if page == total_pages
    limit = total % 250
  else
    limit = 250
  end
  0.upto limit-1 do |photo_index|
    if flickr_data_page["photos"]["photo"][photo_index].has_key?("url_z")
      height = flickr_data_page["photos"]["photo"][photo_index]["height_z"].to_i
      if height == 480
        url = flickr_data_page["photos"]["photo"][photo_index]["url_z"]
        $stderr.printf("url:%s\n", url)
        filename = url.split(/\?/).first.split(/\//).last
        filepath = flickr_z_dir.path + filename
        $stderr.printf("filename with directory:%s\n", filename)
        potential_files_to_copy[num_480] = filepath
        num_480 += 1
      end
    end
  end
  $stderr.printf("num480:%d\n", num_480)
end
random_nums = []
random_nums_index = 0
0.upto 299 do |i|
  r = rand(num_480)
  while random_nums.index(r) 
    r = rand(num_480)
  end
  random_nums[random_nums_index] = r
  random_nums_index += 1
  file_to_copy =  potential_files_to_copy[r]
  $stderr.printf("picked random:%d, filepath:%s\n", r, file_to_copy)
  FileUtils.cp file_to_copy, '.', :verbose => true
end

sample output:
http://dl.dropbox.com/u/361757/Z_300_RANDOM_1.zip
http://dl.dropbox.com/u/361757/Z_300_RANDOM_2.zip
http://dl.dropbox.com/u/361757/Z_300_RANDOM_3.zip