主要采用ruby Parallel库提供的多线程方式：

require 'unirest'require 'open-uri'require 'parallel'require 'json'url = 'http://pvp.qq.com/web201605/js/herolist.json'response = Unirest.get(url)/(\[.*\])/ =~ response.body.force_encoding('utf-8')hero_list = JSON $1#返回ename, cname, skin_name组成的hash数组hero_list.each do |hero|  hero.select!{
    |key, _| key=='ename' || key == 'cname' || key == 'skin_name'}  # hero.select!{|key, _| key =~ /[('ename')|('cname')|('skin_name')]/}enddef download_hero_img(url, path, img_name)  img_file = open(url,:ssl_verify_mode => OpenSSL::SSL::VERIFY_NONE){ |f| f.read }  puts "download #{path}/#{img_name}.jpg"  open("#{path}/#{img_name}.jpg", 'wb'){ |f| f.write(img_file)}enda = Proc.new{hero_list.pop || Parallel::Stop}# 分成10个线程爬取Parallel.map(a, in_threads:10) do |hero|  path = "./#{hero['cname']}"  Dir.mkdir(path) unless Dir.exists?(path)  # url = "http://game.gtimg.cn/images/yxzj/img201606/heroimg/#{hero['ename']}/#{hero['ename']}"  skin_list = hero['skin_name'].split('|')  # puts skin_list  skin_list.each_with_index do |skin, index|    url = "https://game.gtimg.cn/images/yxzj/img201606/skin/hero-info/#{hero['ename']}/#{hero['ename']}-bigskin-#{index.to_i+1}.jpg"    img_name = skin    download_hero_img(url, path, img_name)  endend

线程参考资料：