socialforge/app/helpers/system_log_helper.rb

module SystemLogHelper
    class SystemLog
      class << self
        # 分页（支持多关键字查询）
        def logo_data(page,per,search,day)
          logs = find_all_logs day
          if logs.empty? #如果返回的是空數組，就說明日誌文件不存在
            return logs
          end
          # 根据search参数来决定是否需要查询
          keywords = search
          if keywords && !keywords.strip.blank?
            # 把keywords转化成正则表达式数组
            keywords = keywords.strip.split(/\s+/).collect! {|w| Regexp.new(w, 'i')}
            # 一条记录应该匹配每个关键字
            logs = logs.find_all do |log|
              keywords.all? { |r| log =~ r }
            end
            #用Kaminari分页
            logs = Kaminari.paginate_array(logs).page(page).per(per).collect! {|log| parse(log)}
            logs.collect! do |log|
              keywords.each { |r| log.gsub!(r, '<span class="search_results">\0</span>')}
              log
            end
          else
              logs = Kaminari.paginate_array(logs).page(page).per(per).collect! {|log| parse(log)}
          end
          logs
        end

        #清除日誌
        def clear day
          if File::exists?(logfile_path day)
            File.open(logfile_path(day), 'w') do |f|
              f.print ''
            end
          else
          end
        end

        #讀取日誌
        private
        def find_all_logs day
          if File::exists?(logfile_path day)
            File.open(logfile_path day) do |f|
              #打开文件，并按照正则表达式切分，逆序，最新一个记录可以扔掉（因为最新的记录永远都是访问System_log)
              f.read.split("Processing").reverse[1..-1]
            end
          else
            []
          end
        end

        # 日志文件的路径，一般在Rails.root/log下，根据环境配置
        # 依次记录到product.log development.log test.log中
        def logfile_path day
          #将日期处理成2015-01-01的形式
          unless day.nil?
          dayArr = day.split('-')
          if dayArr[1].length == 1
            dayArr[1] = "0" + dayArr[1]
          end
          if dayArr[2].length == 1
            dayArr[2] = "0" + dayArr[2]
          end
          day = dayArr.join('-')
          end
          #如果不是當天，則需要加後綴
          if !day.nil? && !day.strip.blank? && day != Time.now.strftime("%Y-%m-%d")
            File.join(Rails.root, "log", "#{Rails.env}.log.#{day.gsub('-','')}")
          else
            File.join(Rails.root, "log", "#{Rails.env}.log")
          end
        end

        #替換換行符
        def parse(log)
          ERB::Util.html_escape(log.gsub(/\e\[[\d;m]+/, '')).gsub("\n", "<br/>")
        end

        #定义响应正则表达式 2015-01-20 11:31:13 INFO -- Completed 200 OK in 125ms (Views: 81.0ms | ActiveRecord: 2.0ms)
        def response_regex
          'Completed \d+ \w+ in (\d+)ms \(Views: (\d+\.\d+)?ms \| ActiveRecord: (\d+\.\d+)?ms\)'
        end

        #将一条记录中的地址主机等都分析出来
        def get_status(paragraph)
          request_regex = 'Started GET \"(\/.*)\" for ([\d]+\.[\d]+\.[\d]+\.[\d]+) at [\d]*-([\d]*-[\d]* [\d]*:[\d]*:[\d]*)'
          controller_regex = 'Processing by ([\w]+#[\w]+)'
          page_time_regex = 'Views: \d+(\.\d+)?ms'
          activeRecord_time_regex = 'ActiveRecord: \d+(\.\d+)?ms'

          if paragraph.match(request_regex) != nil
            request_url = paragraph.match(request_regex)[1] #正则表达式中的括号能够截取成数组
            request_host = paragraph.match(request_regex)[2]
            request_at =  paragraph.match(request_regex)[3]
          end

          if paragraph.match(controller_regex) != nil
            controller_name = paragraph.match(controller_regex)[1]
          end

          if paragraph.match(response_regex) != nil
            #print(paragraph.match(response_regex))
            total_time = paragraph.match(response_regex)[1]
            page_time = paragraph.match(response_regex)[2]
            activeRecord_time = paragraph.match(response_regex)[3]
            page_time_percent = page_time.to_f/(total_time.to_f)
            activeRecord_time_percent = activeRecord_time.to_f/(total_time.to_f)
          else
          end
          request_status = [request_url, request_host, request_at,
                            controller_name, total_time,page_time,page_time_percent,activeRecord_time,activeRecord_time_percent]
          request_status
        end

        #分析日志
        public
        def analysis day
          csv = Array.new
          #如果文件不存在，则直接返回空数组
          if File::exists?(logfile_path day)
           File.open(logfile_path(day), "r:utf-8") do |file|
              paragraph = ""
              begin_flag = false
              # 对每一行进行判断
              file.each do |line|
                # 以"Started GET "开头为一个paragraph
                #print(line.match('[\d]*-([\d]*-[\d]* [\d]*:[\d]*:[\d]*) INFO -- Started GET ') == nil)
                if (line.match('[\d]*-([\d]*-[\d]* [\d]*:[\d]*:[\d]*) \w+ -- Started GET ') != nil)
                  if !begin_flag
                    begin_flag = true
                    paragraph.concat(line)
                  else
                    # 另一个paragraph的开头
                    if (paragraph.match(response_regex) != nil)
                        csv << get_status(paragraph)
                    end
                    begin_flag = true
                    paragraph = line
                  end
                else
                  if begin_flag
                    paragraph.concat(line)
                  else
                  end
                end
              end
           end
          end
          csv
        end
      end
    end
  end