Pulling images from Post Secret

post-secret-archive.rb

#!/usr/bin/ruby

require 'rubygems'
require 'hpricot'
require 'open-uri'
require 'ftools'

doc = Hpricot(open("http://postsecret.blogspot.com/"))
images = []

(doc/"//img").select{|img| img.attributes['id'] =~ /BLOGGER_PHOTO_ID/}.each do |img|
  img_source = open(img.attributes['src'])
  img_location = 'images/' +  img.attributes['src'].sub(/^.*blogger.com\//, '')
  File.makedirs(File.dirname(img_location))
  File.open(img_location, 'w') {|f| f.write(img_source.read)}
  images << img_location
end

html_filename = Time.now.strftime("%Y-%m-%d") + ".html"
File.open(html_filename, "w") do |f|
  images.each do |img|
    f.write("<div><img src=\"#{img}\" /></div>\n")
  end
end