ruby基本库常见用法

我写了ruby脚本也有一段时间了。ruby某些库的api的说明不是那么全，特别是socket。我在这个博客把一些ruby基本库的用法节选下来吧。代码本身的逻辑都是能运行的，而附属的一些yaml数据没有放上来，所以代码无法直接运行。

写这个博客是希望能对某些ruby新手有所帮助。目前ruby的中文教程确实不怎么多，且重复的不少。就我看来，ruby既不是太好，也不是太坏，当你熟悉它的语法和“pass by object reference”后，就会不自然地产生对它的爱和恨。本文主要写一些方便快速完成任务的实用性的东西。

 

一、Herbern m3u8下载器

用到了ruby的：

OpenURI:http下载库，注意设置header；io = URI.open(link, hash)，这里的hash约等于http发送get/post请求时的header，如果在hash里指定 :encoding = 'utf-8'，则open-uri会以utf8编码读取下载的流，否则默认按ascii编码来读取。

socket (windows运行)的读写，TCPServer和TCPSocket。Socket < BasicSocket < IO，而close_write和read(Int)都是socket继承自IO的函数。

Set[] 集合： require 'set'; a = Set[3,4,5]; a.to_a # => [3,4,5];  [5,6,7].to_set #=> #<Set: {5, 6, 7}>; a.include?(5) #=> true; a << 2; #=> #<Set: {3, 4, 5, 2}>; a.delete(5)  #=> #<Set: {3, 4, 2}>

JSON库基本用法(Psych yaml库的用法类似)；

Thread.new {}; 多线程; Mutex和Queue的用法参见“runoob菜鸟教程”。

类似于Python的if __name__ == "main" 的如本文件为启动脚本则执行的 if ($0 == __FILE__)用法，和ARGV:启动参数数组

简单的正则表达式用法：注意是否开启多行模式，和替换换行符

text0 = '<div class="el">
        <p class="t1 ">
            <em class="check" name="delivery_em" onclick="checkboxClick(this)"></em>
            <input class="checkbox" type="checkbox" name="delivery_jobid" value="110487870" jt="0" style="display:none" />
            <span>
                <a target="_blank" title="美容顾问（佛山）" href="https://jobs.51job.com/foshan/110487870.html?s=01&t=0"  onmousedown="">
                    美容顾问（佛山）                </a>
            </span>
                                                                    </p>
        <span class="t2"><a target="_blank" title="上海蝶翠诗商业有限公司（DHC）" href="https://jobs.51job.com/all/co298555.html">上海蝶翠诗商业有限公司（DHC）</a></span>
        <span class="t3">佛山</span>
        <span class="t4">6-8千/月</span>
        <span class="t5">02-24</span>
    </div>
    <div class="el">
        <p class="t1 ">
            <em class="check" name="delivery_em" onclick="checkboxClick(this)"></em>
            <input class="checkbox" type="checkbox" name="delivery_jobid" value="120125665" jt="0" style="display:none" />
            <span>
                <a target="_blank" title="电商业务专员" href="https://jobs.51job.com/foshan-sdq/120125665.html?s=01&t=0"  onmousedown="">
                    电商业务专员                </a>
            </span>
                                                                    </p>
        <span class="t2"><a target="_blank" title="美的集团热水器事业部" href="https://jobs.51job.com/all/co3224953.html">美的集团热水器事业部</a></span>
        <span class="t3">佛山-顺德区</span>
        <span class="t4">1-1.5万/月</span>
        <span class="t5">02-24</span>
    </div>'
text0 = text0.gsub("\r", '').gsub("\n", '')
h = '<div class="el">'
b = '</div>'
re = /#{h}.*?#{b}/
elems = text0.scan(re)
puts elems.size, elems[0]
a = /(\d+)(abc)de/.match('12abcdef') # => #<MatchData "12abcde" 1:"12" 2:"abc">
a.to_a #=>  ['12', 'abc']

 

# m3u8下载器 赫本
require 'open-uri'
require 'fileutils'
require 'socket'
require 'json'
#~ require 'sinatra/base'
# mlink指m3u8的http路径

HEADER = {
  'user-agent' => 'Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/81.0.4044.138 Safari/537.36',
  'collection' => 'keep-alive'
}

def parse_links(mfile, mlink)

  fio2 = File.open(mfile, 'r')
  lines = fio2.readlines.map{|x| x.chomp}
  fio2.close
  lines.reject!{|x| x=='' || x[0] == '#'}
  raise if lines.empty?
  
  infos = []
  
  if (lines[0][0] == '/')
    # 模式:域名+每一行的内容
    pos1 = mlink.index('//')+2
    host = mlink[0, mlink.index('/', pos1)]
    lines.each do |line|
      elems = line.split('/')
      fn = elems[-1]
      link = host + line
      infos << {'fn' => fn, 'link' => link}#, :size => -1}
    end
  else
    # 模式:链接+文件名
    tmp = mlink.split('/')
    tmp.delete_at(-1)
    link_model = tmp.join('/')
    lines.each do |line|
      fn = line
      link = link_model + '/' + fn
      infos << {'fn' => fn, 'link' => link}#, :size => -1}
    end
  end
  
  return infos
  
end

$cache_file = 'cache.json'

def socket_worker(upload_msg, recv = false)
  begin
    cli = TCPSocket.new('127.0.0.1', 6262)
    cli.print upload_msg
    if recv
      cli.close_write
      ret = cli.read(30)
      return ret
    end
    cli.close
     
  rescue StandardError 
    puts "无法与hb_server()通讯"
    exit
  end

end

def read_cache
  if File.exist?($cache_file)
    puts "从cache读取信息"
    fio2 = File.open($cache_file)
    infos = JSON.load(fio2.read)
    fio2.close
    return infos
  else
    raise StandardError, '无法读取缓存信息'
  end

end

def new_task(mlink)
  
  tmp = URI.parse(mlink)
  unless (tmp.is_a?(URI::HTTP) || tmp.is_a?(URI::HTTPS))
    raise StandardError, "mlink格式不对:'#{mlink}'"
  end
  begin
    io = URI.open(mlink, HEADER)
    fc = io.read
  rescue StandardError => ser
    puts "下载m3u8文件信息错误:'#{ser.message}'"
    exit
  end
  File.open('index.m3u8', 'w') do |fio|
    fio.puts "##{mlink}"
    fio.puts fc
  end
  
  infos = parse_links('index.m3u8', mlink)
  File.open('cache.json', 'w') do |fio|
    fio.print infos.to_json
  end
  download_worker(infos)  
end

def generate_task_name
  # 假设进程的pid是唯一的
  name = "#{Process.pid}"
  socket_worker("add_#{name}")
  return name
end

def download_worker(infos)
  check_dir
  
  # 生成任务名
  task_name = generate_task_name
  puts "任务名:'#{task_name}'"

  infos.reject!{|x| File.exist?(x['fn'])}
  puts "共需下载#{infos.size}个文件"

  err_count = 0
  File.delete('err.txt') if File.exist?('err.txt')

  until infos.empty?
    info = infos.shift
    # 连接hb_server()以检查是否要停止任务
    ret = socket_worker("status_#{task_name}_#{infos.size}", true)
    #~ puts "ret=#{ret}"; STDIN.gets
    break if (ret != 'true')
    
    fn = info['fn']
    link = info['link']
    th = Thread.new do
      begin
        io = URI.open(link, HEADER)
        fc = io.read
        File.open(fn, 'wb') do |fio|
          fio.print fc
        end
        File.open('../filestats.txt', 'a') do |fio3|
          fio3.puts "#{fn}||#{fc.size}"
        end
      rescue StandardError => ser
        msg = "下载#{fn}文件错误:'#{ser.message}'"
        err_count += 1
        File.open('err.txt', 'a') do |fio|
          fio.puts msg
        end
      end
    end
    time = 0
    while (th.alive? && time <= 30)
      sleep 1
      time += 1
    end
    if (time >= 30)
      msg = "下载#{fn}文件超时"
      err_count += 1
      File.open('err.txt', 'a') do |fio|
        fio.puts msg
      end
    end
    
    th.kill
  end  
    
  # 确保全部保存
    
  Dir.chdir('..')
  if (err_count > 0)
    puts "下载出错数:#{err_count}"
  end
  
  # 退出时告知hb_server()
  socket_worker("kill_#{task_name}")

  puts "下载流程完成:#{tn = Time.now; tn.min.to_s + ':' + tn.sec.to_s}"  
  sleep(1)
end

def check_dir
  Dir.mkdir('./ts') unless Dir.exist?('./ts')
  Dir.chdir('./ts')
end


def make_file
  infos = read_cache

  # 读取下载了的文件的信息
  fstats = {}
  File.open('filestats.txt', 'r') do |fio|
    lines = fio.readlines.select{|x| x.include?('||')}
    lines.each do |ln|
      fn, size = ln.split('||')
      fstats[fn] = size.to_i # 后面重复的记录会覆盖掉往前面的记录
    end
  end
  
  # 合并文件
  check_dir
  
  fio2 = File.open('bin_output.ts', 'wb')
  infos.each do |x|
    fn = x['fn']
    unless File.exist?(fn)
      raise "'#{fn}'文件不存在"
    end
    
    if (fstats[fn] == nil)
      puts "未记载'#{fn}'的文件信息"
    elsif (fstats[fn] != File.size(fn))
      raise "'#{fn}'文件大小与记录信息不符"
    end
    
    File.open(fn, 'rb') do |fio|
      fio2.print fio.read
    end
  end
  fio2.close
  
  Dir.chdir('..')
  FileUtils.mv('./ts/bin_output.ts', './bin_output.ts')
  
  File.open('ffm.cmd', 'w') do |fio|
    fio.puts "ffmpeg -i bin_output.ts -c:v copy -c:a copy 00result.mp4"
  end
  
  puts '已输出ffmpeg命令'

end

if ($0 == __FILE__)

  case ARGV[0]
  when nil
    infos = read_cache
    download_worker(infos)
  when 'new'
    print '输入m3u8link>'
    str = STDIN.gets
    new_task(str.chomp)
  when 'make'
    make_file
  else
    raise '不支持的命令'
  end

end

 

 

赫本的socket服务器，用来管理各个任务的状态

require 'socket'
require 'set'
require 'readline' # gem install rb-readline

$tasks = {}

server1 = TCPServer.new('127.0.0.1', 6262) 
# 新建任务 'add:446', 'kill:4664'
#~ server2 = TCPServer.new('127.0.0.1', 6263) 
# 查询任务状态


Thread.new do
  loop do
    cli = server1.accept
    str = cli.read(60)
    mt = str.split('_')
    if (mt != nil)
      action = mt[0]
      task_name = mt[1]
      case action
      when 'add'
        if  $tasks.include?(task_name)
          puts "意外:重复的pid任务:#{task_name}"
        else
          $tasks[task_name] = 0
        end
      when 'kill'
        $tasks.delete(task_name)
      when 'status'
        if $tasks.include?(task_name)
          $tasks[task_name] = mt[2].to_i
          cli.print 'true'
        else
          cli.print 'false'
        end
      end
    #~ else
      #~ cli.puts 'illegal'
    end
    cli.close
  end
end

#~ sleep(1)

puts "输入命令:"

while true
  #~ print '>>'
  #~ uip = STDIN.gets
  #~ uip.chomp!
  uip = Readline::readline('>>')
  Readline::HISTORY.push(uip)
  if (uip == 'exit')
    server1.close
    exit
  elsif (uip == 'list')
    puts "任务列表:[#{$tasks.keys.join(',')}]"
  elsif (uip == 'irb')
    require 'irb'
    binding.irb
  else
    mt = /^(\w)\s(\d+)/.match(uip) 
    # 'x 146' 表示结束146任务
    # 'i 146' 表示查看任务146的进度
    if mt
      unless $tasks.key?(mt[2])
        puts "不存在任务名:#{mt[2]}"
        next
      end
      case mt[1]
      when 'x'
        $tasks.delete mt[2]
        puts '已标记结束'
      when 'i'
        puts "'#{mt[2]}'=#{$tasks[mt[2]]}"
      else
        puts "不认识的命令:'#{uip}'"
      end
    else
      puts "不认识的命令:'#{uip}'"
    end
  end
end

 

二、某招聘网站的爬虫

（这个代码写得较早，所以有点乱）

用到了ruby的：

Nokogiri：xml解析库；

SQLite3: sqlite3数据库的用法，

  .results_as_hash: 把结果以哈希形式输出；

  sql.query('begin'); sql.query('commit')：SQL事务；

  sqlite的转义用法：sql.query('update table set a = ? where b = ?', [a_value, b_value])；

 

ERB：ruby的类似于php的模板引擎的用法，ERB.new(string).result(binding)；ERB的模板用法见官方文档。

String：index(pos, startpos)用法；

.encode('utf-8', 'gbk', {:invalid => :replace, :undef => :replace, :replace => '?'})用法（见代码）

ruby元编程的基本用法：Kernel/Object.send(:method, args)；  require和load的用法；

获取脚本自身完整路径的方法：File.expand_path(File.dirname(__FILE__))

 

require 'psych'
require 'open-uri'
require 'erb'
require 'sqlite3'
require 'nokogiri'
require 'set'

=begin
  使用方法:
  rw = RecruitWalker.new()
  rw.prepare_crawl
  rw.add_task('job51')
  rw.add_task('shundehr')
  rw.begin_tasks
  rw.write_html
=end




def uri_open(link, header={})
  # 使用open-uri
  begin
    io = URI.open(link, header)
    c = io.read
    io.close
    return c
  rescue StandardError
    return false
  end

end



class RecruitWalker

  attr_reader :header, :datadb_fn, :legacydb_fn, :joblist, :late10dates

  def initialize
    @path = File.expand_path(File.dirname(__FILE__))

    # 读取配置信息
    @datadb_fn = "#{@path}/database/store5.db"
    @legacydb_fn = "#{@path}/database/oldjob.db"

    # 读取招聘网站的配置信息
    @conf_fn = "#{@path}/ext/info.yml"

    @joblist = Psych.load(File.read(@conf_fn))

    @late10dates = read_dates()

    return 
  end

  def prepare_crawl
    # 准备爬虫流程
    # 连接sqlite数据库
    @data_sql = SQLite3::Database.new(@datadb_fn)
    @data_sql.results_as_hash = true
    @legacy_sql = SQLite3::Database.new(@legacydb_fn)
    @task_list = {}
    # 任务列表形式:
    # {
    #   任务英文名1 => {:htmls => [], :infos => []},
    #   任务英文名2 => {:htmls => [], :infos => []}
    # } 

    # 读取公司名黑名单

    @blacklist = Psych.load(File.read("#{@path}/ext/blacklist.yml"))
    @header = {'User-Agent'=>'Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/78.0.3904.87 Safari/537.36', 'cookie'=>''}
    return
  end

  def read_dates
    # 获取最近的10个日期
    tn = Time.now
    time_array = []
    10.times do
      time_array << tn.clone
      tn -= 24*3600
    end
    dates_array = []
    time_array.each do |tm|
      day = tm.to_s.split(' ')[0].gsub('-', '')
      dates_array << day
    end
    return dates_array # 字符串数组
  end
  
  def add_task(name)
    if @joblist.has_key?(name)
      @task_list[name] = {:htmls => [], :infos => []}
    else
      raise StandardError, "找不到任务名:'#{name}'"
    end
  end
  
  def begin_tasks
    # 开始抓取任务列表中的所有任务名
    # 并行抓取各个任务的网页;数据整理和SQL整理则排队处理
    return if @task_list.empty?
    puts "任务流程开始,数量:#{@task_list.keys.size},时间'#{Time.now}'"
    # 下载网页
    web_threads = []
    @task_list.each_pair do |k, v|
      th = Thread.new do
        v[:htmls] = scan_web_task(k, @joblist[k])
      end
      web_threads << th
    end
    # 等待所有的下载完成
    web_threads.map{|x| x.join}
    
    # html分析
    @task_list.each_pair do |k, v|
      method = @joblist[k]['scanner'].to_sym
      puts "加载分析库文件:#{k}"
      load "#{@path}/ext/use-#{k}.rb"
      puts "分析htmls:开始:'#{@joblist[k]['text']}'"
      tmp_keys = Set[]
      v[:htmls].each do |html|
        tmp_list = send(method, html)
        tmp_list.each do |x|
          unless tmp_keys.include?(x['jobid'])
            v[:infos] << x
            tmp_keys << x['jobid']
          end
        end
        #~ v[:infos] 
      end
      puts "分析htmls:完成:'#{@joblist[k]['text']}'"
    end
    
    # 数据迁移
    @task_list.each_key do |k|
      data_migrate(@joblist[k]['table'])
    end
    
    # 数据更新和写入
    @task_list.each_pair do |k, v|
      data_tidy(@joblist[k], v[:infos])
    end
    
    # 保存抓取时间
    update_conf
    puts "任务流程完成."
    return
  end

  def scan_web_task(name, conf)

    puts "执行下载:'#{conf['text']}'"
    
    local_header = @header.clone
    local_header[:encoding] = conf['encoding']
    
    htmls = []
    p_begin = conf['page_begin']
    p_end = conf['page_end']
    
    (p_begin .. p_end).each do |pid|
      link = conf['link'].sub('{{pageid}}', pid.to_s)

      html = uri_open(link, local_header)
      if (html == false)
        puts "下载'#{name}'的第#{pid}页索引出现http错误"
        next
      end
      
      htmls << html
      sleep(rand(2))
    end
    puts "下载完成:'#{conf['text']}'"
    return htmls

  end

  def update_conf
    # 写入抓取时间到yaml文件
    tn = Time.now.to_s
    @task_list.each_key do |k|
      @joblist[k]['update_time'] = tn
    end
    File.open(@conf_fn, 'w') do |fio|
      fio.print @joblist.to_yaml
    end
    return
  end

  def data_migrate(table)
    # table是表格名
    
    drop_cmds = [] # 从主数据库的对应表格中去除过时信息
    update_cmds = [] # 把过时信息的jobid插入到oldjob数据库的对应表格
    cmd = "select jobid, post_date from #{table}"
    #~ puts "migrate cmd='#{cmd}'"
    @data_sql.query(cmd).each do |x|
      jobid = x['jobid']; p_date = x['post_date']
      unless @late10dates.include?(p_date)
        # 在主要数据库里删除本条过时的信息
        drop_cmds << {:cmd => "delete from #{table} where jobid = ?", :params => [jobid]}
        # 在记录过时信息的数据库里添加本条过时信息的jobid和日期
        update_cmds << {:cmd => "insert into #{table} (jobid, post_date) values (?, ?)", :params => [jobid, p_date]}
      end
    end
    
    @data_sql.query('begin')
    drop_cmds.each do |c|
      @data_sql.execute(c[:cmd], c[:params])
    end
    @data_sql.query('commit')
    
    @legacy_sql.query('begin')
    update_cmds.each do |c|
      @legacy_sql.execute(c[:cmd], c[:params])
    end
    @legacy_sql.query('commit')
    
    return
  end

  def data_tidy(conf, src_infos)
    table = conf['table']
    # 获取过时的jobid
    outdated_jobids = []
    @legacy_sql.query("select jobid from #{table}").each do |x|
      outdated_jobids << x[0]
    end
    
    black_count = 0 # 已用公司名黑名单屏蔽的招聘信息数
    
    infos = src_infos.select{  |x|
     
      bool1 = @late10dates.include?(x['post_date']) # 去除时间范围以外的招聘信息
      bool2 = !@blacklist.include?(x['company']) # 去除黑名单公司的招聘信息
      black_count += 1 if (bool2 == false)
      bool3 = !outdated_jobids.include?(x['jobid']) # 去除在oldjob数据库对应表格里已包含jobid的招聘信息
      (bool1 && bool2 && bool3)
    }
    
    puts "已根据公司名黑名单,屏蔽了#{black_count}条招聘信息" if (black_count>0)
    puts "数据整理:开始:'#{conf['text']}'"
    
     # 筛选当前的招聘嘻嘻
    
    lately_jobids = []
    
    @data_sql.query("select jobid from #{table}").each do |x|
      lately_jobids << x['jobid']
    end

    cmds = []
    update_count = 0
    insert_count = 0

    infos.each do |info|
      # 处理每个新下载的招聘信息
      set = {}
      if lately_jobids.include?(info['jobid'])
        # 时间范围内的招聘信息数据库已包含相同jobid的信息
        set[:cmd] = "update #{table} set update_date = ? where jobid = ?"
        set[:params] = [info['post_date'], info['jobid']]
        update_count += 1
      else
        set[:cmd] = "insert into #{table} (jobid,post_date,company,link,name,srcsite,update_date) values (?, ?, ?, ?, ?, ?, ?)"
        set[:params] = [
          info['jobid'],
          info['post_date'],
          info['company'],
          info['link'],
          info['name'],
          info['srcsite'],
          info['update_date']
        ]
        insert_count += 1
      end
      cmds << set
    end

    puts "更新数:#{update_count}", "插入数:#{insert_count}"
    puts '向数据库添加新信息..'

    @data_sql.query('begin')
    cmds.each do |c|
      @data_sql.query(c[:cmd], c[:params])
    end
    @data_sql.query('commit')
    

    return 
  end

  def write_html
    # 清除上次保存的html文件
    Dir.chdir("#{@path}/html")
    Dir.glob('*.html').each do |fn|
      File.unlink(fn)
    end
    Dir.chdir(@path)
    # 写入欢迎页
    fc1 = File.read("#{@path}/template2/index.erb")
    dates_text = @late10dates.map{|x| "'#{x}'"}.join(',')
    sites_text = @joblist.map{|k, v|
      "{'name':'#{k}', 'text':'#{v['text']}', 'upd_time':'#{v['update_time']}'}"

    }.join(',')

    html1 = ERB.new(fc1).result(binding)
    File.open("#{@path}/html/index.html", 'w') do |fio|
      fio.print html1
    end

    # 写入各个招聘网信息按日期的详细页面
    #~ @data_sql
    fc2 = File.read("#{@path}/template2/details.erb")

    #~ puts "日期列表:#{dates}, joblist=#{rw.joblist}"
    @late10dates.each do |date|
      @task_list.each_pair do |k, v|
        #~ puts "正在写入:#{k}, #{date}"
        conf = @joblist[k]
        site_text = conf['text']
        table = conf['table']

        infos = []
        cmd = "select * from #{table} where post_date = '#{date}'"
        #~ puts "sql cmd=#{cmd}"
        @data_sql.query(cmd).each do |x|
          infos << x # 不考虑:把相同的公司的信息放在一起
        end
        count = infos.size # erb里用到这个变量
        #~ puts "检查infos:共#{count}个,[0]=", infos[0].inspect, ''
        html2 = ERB.new(fc2).result(binding)
        fn = "#{@path}/html/info-#{k}-#{date}.html"
        File.open(fn, 'w') do |fio|
          fio.print html2
        end
      end
    end

    puts "写入流程:完成."
    return
  end

end

 

某招聘网站的解析过程：

def page_analyse_sdhr(html_content)#, rw)
  infos = []
  text0 = html_content
  text0.gsub!("\r", '')
  text0.gsub!("\n", '') # 先清除换行符,方便正则表达式匹配
  h = '<ul class="searchPost-list" id="resultList">'
  b = '<div class="centerPage">'
  re = /#{h}(.*)#{b}/
  text1 = re.match(text0)[0]
  
  # 由于<input>标签没有终结的</input>,所以不能用Nokogiri::XML,而要用Nokogiri::HTML
  nk = Nokogiri::HTML(text1, nil, 'utf-8')
  elems = nk.xpath('//li')
  
  
  # 假设每个页面有20条招聘信息

  elems.each do |node|
    # 获取日期
    post_date = node.xpath('./div/div[@class="t5"]').text
    date_text = post_date.split('-').join

    job_name = node.xpath('./div/div[@class="t1"]/a').text
    link = node.xpath('./div/div[@class="t1"]/a/@href').text
    link = 'http://www.shundehr.com' + link
    company_name = node.xpath('./div/div[@class="t2"]/a').text
    jobid_text = node.xpath('./input/@value').text
    
    srcsite = 'shundehr'
    info = {
      'name'        => job_name,
      'company'     => company_name,
      'post_date'   => date_text,
      'update_date' => date_text,
      'jobid'       => jobid_text,
      'link'        => link,
      'srcsite'     => srcsite
    }
    
    #~ puts info.inspect; STDIN.gets
    
    # Nokogiri找不到元素的时候,返回nil.所以这里有必要检查某些值是不是找不到.
    selects = info.select{|k, v| (v==nil)} # 此时selects是个哈希
    if selects.empty?
      infos << info
    else
      puts "分析html时找不到部分元素", info.inspect, '---'
    end
  end
  return infos
end

# 测试
#~ require 'nokogiri'
#~ fc = File.read('../tmp/shundehr_sample.html')
#~ r = page_analyse_sdhr(fc)

 

Nokogiri的部分用法见以下，注意Nokogiri::XML用来解析常规的xml，Nokogiri::HTML支持html里的<image>、<input>等没有结束标记的标签；Nokogiri的参数'utf-8'在windows上最好强制指定，要不然Nokogiri会以windows终端的GBK编码来解析。

Nokogiri有xpath('./node')的用法，识别<a href="link">text</a>是用xpath('./a/@href').text。

Nokogiri还有.to_html；.to_s；.text；.value用法，代码暂时找不到了，建议结合实际使用。

用Nokogiri的时候最好逐步调试，如果nokogiri找不到元素，有时候会抛出异常，有时候会返回nil。

 

Enumerable.any?用法：[3, 4, 5].any?{|x| x %2 == 0} # => true