get average colour of flickr photos from serialized JSON

getAverageColour.rb

#!/usr/bin/env ruby
require 'net/http'
require 'uri'
require 'xmlsimple'
require 'pp'
require 'json'

#pp data["variable"][6]["statistic"][0]["rgb"] # or "hex" or "value"

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_m")
      mkavgcolour_url = "http://mkweb.bcgsc.ca/color_summarizer/?xml=1&url=" +
      flickr_data_page["photos"]["photo"][photo_index]["url_m"] +
      "&precision=medium"
      mkavgcolour_params = "?xml=1&url=" +
      flickr_data_page["photos"]["photo"][photo_index]["url_m"] +
      "&precision=medium"
      begin
        url = URI.parse(mkavgcolour_url)
        $stderr.printf("url:%s host:%s port:%d path:%s\n",mkavgcolour_url, url.host, url.port,url.path)
        $stderr.printf("concatenated url:%s\n", url.path+mkavgcolour_params)
        http = Net::HTTP.new(url.host, url.port)
        http.read_timeout = 160 # seconds
        http.open_timeout = 160
        resp = http.start() {|http|
          http.get(url.path+mkavgcolour_params)
        }
        if resp.code == "200" 
          xml_data = resp.body
          flickr_data_page["photos"]["photo"][photo_index]["avg_colour_m"] = XmlSimple.xml_in(xml_data)
          $stderr.printf "page:%d photo:%d ", page, photo_index
          PP::pp(flickr_data_page["photos"]["photo"][photo_index]["avg_colour_m"],$stderr)
        else
          $stderr.printf("HTTP GET failed on:%s HTTP response code:%d\n", url.path+mkavgcolour_params,resp.code)
        end
      rescue => e
        PP::pp(e,$stderr)
      rescue Timeout::Error => e
        PP::pp(e,$stderr)
      end
    end
  end
  print JSON.generate(flickr_data_page), "\n"
end

reads serialized JSON flickr meta data from stdin
and then calls Martin K's web service on each "url_m" photo which calculates the average colour
and then adds the average colour info to the JSON and serializes it to stdout