csv解析完成后运行rake任务

时间:2016-03-15 09:11:51

标签: ruby-on-rails ruby

我有一个看起来有点像这样的rake任务

require 'open-uri'
require 'csv'
namespace :tm do
  task reload: :environment do
    gzipped = open('csv link')
    csv_text = Zlib::GzipReader.new(gzipped).read
    csv = CSV.parse(csv_text, headers: true)

    csv.each do |row|
      if row[4] == 'logo url'
      else
        tmdate = Date.parse(row[10]).strftime('%Y-%m-%d')
        viatmdate = Date.parse(row[10]).strftime('%d/%m/%Y')
        swtmdate = row[10]
        tmlocation = row[6].split('at ')[1]
        place = row[11].split('|')[1]
        place1 = row[11].split('|')[2]
        place2 = row[11].split('|')[3]
        location = '' + place + ', ' + place1 + ', ' + place2 + ''
        tmtime = row[9]
        text = row[7].gsub('text', '')
        if text.include? '�&#xFFFD'
          eventname = text.gsub('�&#xFFFD', 'e')
        else
          eventname = text.gsub(/[ªÀÈÌÒÙàèìòùÁÉÍÓÚáéíóúÂÊÎÔÛâêîôûÃÑÕãñõÄËÏÖÜŸäëïöüÿ]/, '')
        end

        if text.include? '�&#xFFFD'
          tmname = text.gsub('�&#xFFFD', 'e')
        else
          tmname = text.gsub(/[ªÀÈÌÒÙàèìòùÁÉÍÓÚáéíóúÂÊÎÔÛâêîôûÃÑÕãñõÄËÏÖÜŸäëïöüÿ]/, '')
        end

        if text.include? ' -'
          tmnamesplit = text.split(' -')[0]
        end

        if tmname[/[^0-9]/].present?
          tmnamenn = tmname.gsub(/[^0-9]/i, '')
        end

        text2urldb = text2.where('eventtitle ILIKE ? AND eventdoortime = ? ', "%#{tmname.gsub(/[\-\:\ ]/, '%')}%", tmdate.to_s).first
        text3urldb = text3.where('product_name ILIKE ? AND delivery_time = ? AND valid_from = ?', "%#{tmname}%", tmtime.to_s, tmdate.to_s).first

        text1urldb = text1.where('product_name ILIKE ? AND specifications = ? AND promotional_text = ?', "%#{tmname}%", viatmdate.to_s, "%#{place}%").first

        if tmnamesplit.present?
          if text1urldb.blank?
            text1urldb = text1.where('product_name ILIKE ? AND specifications = ?', "%#{tmnamesplit}%", viatmdate.to_s).first
          end
          if text3urldb.blank?
            text3urldb = text3.where('product_name ILIKE ? AND delivery_time = ? AND valid_from = ?', "%#{tmnamesplit}%", tmtime.to_s, tmdate.to_s).first
          end
        end

        if text1urldb.blank?
          text1urldb = text1.where('product_name ILIKE ? AND specifications = ? AND promotional_text = ?', "%#{tmname}%", viatmdate.to_s, "%#{location}%").first
          if text1urldb.blank?
            text1urldb = text1.where('product_name ILIKE ? AND specifications = ?', "%#{tmname}%", viatmdate.to_s).first
          end
          if text1urldb.blank?
            text1urldb = text1.where('product_name ILIKE ? AND specifications = ? AND promotional_text = ?', "%#{tmname}%", viatmdate.to_s, "%#{tmlocation}%").first
          end
        end

        if text1urldb.present?
          vurl = text1urldb.merchant_deep_link
          txt = vurl
          re1 = '.*?' # Non-greedy match on filler
          re2 = '(?:[a-z][a-z]+)' # Uninteresting: word
          re3 = '.*?' # Non-greedy match on filler
          re4 = '(?:[a-z][a-z]+)' # Uninteresting: word
          re5 = '.*?' # Non-greedy match on filler
          re6 = '(?:[a-z][a-z]+)' # Uninteresting: word
          re7 = '.*?' # Non-greedy match on filler
          re8 = '(?:[a-z][a-z]+)' # Uninteresting: word
          re9 = '.*?' # Non-greedy match on filler
          re10 = '(?:[a-z][a-z]+)'    # Uninteresting: word
          re11 = '.*?'    # Non-greedy match on filler
          re12 = '((?:[a-z][a-z]+))'  # Word 1
          re = (re1 + re2 + re3 + re4 + re5 + re6 + re7 + re8 + re9 + re10 + re11 + re12)
          m = Regexp.new(re, Regexp::IGNORECASE)
          if m.match(txt)
            word1 = m.match(txt)[1]
          end
        end


        gmiurl = text3urldb.merchant_deep_link if text3urldb.present?

        gigurl = text2urldb.eventurl if text2urldb.present?

        api = HTTParty.get(URI.encode('text url' + tmname + '&when_from=' + swtmdate)).parsed_response
        api1 = api['Paging']
        api2 = api1['TotalResultCount']

        if api1.blank?
          newapi = HTTParty.get(URI.encode('texturl' + tmnamenn + '&when_from=' + swtmdate)).parsed_response
          paging = newapi['Paging']
          api2 = paging['TotalResultCount']
          if newapi.blank?
            apisplit = HTTParty.get(URI.encode('texturl' + tmnamesplit + '&when_from=' + swtmdate)).parsed_response
            pagingsplit = apisplit['Paging']
            api2 = pagingsplit['TotalResultCount']
          end
        end
        text1 = vurl
        text3 = gmiurl
        text2 = gigurl

        if api2 == 0
        else
          swurl = api['Events'].first['SwURL']
        end
        event = Event.find_by(time: row[9], date: row[10], eventname: eventname, eventvenuename: location)

        if event.present?
          event.update(event_type: word1, text: row[8], eventimage: row[4], textlink: swurl, text1link: text1, text3url: text3, text2url: text2)
        else
          Event.create(time: row[9], date: row[10], event_type: word1, text: row[8], eventimage: row[4], eventname: eventname, eventvenuename: location, textlink: swurl, text1link: text1, text3url: text3, text2url: text2)
        end
      end
    end
  end
end

现在这已经开始长时间运行(大约2-3个小时),所以我想知道如果拆分它会加快一点吗?

如果我要从这里取出文本1 2和3,那么它首先将它放入数据库中。然后在完全填充之后运行另一个rake任务?希望这会加快速度吗?所以,如果我有一个名为tm:reload的rake任务,以及之后在另一个名为BA的文件中的另一个任务:重新加载我将如何做到这一点?

我已经完成并索引了几列,但这并没有改进它。

1 个答案:

答案 0 :(得分:0)

我会做以下事情:

  • 执行Rake任务以读取CSV文件并为每行创建后台作业
  • 使用Sidekiq作为后端,这听起来像你可以轻松地并行化
  • 让这项工作通过HTTP执行解析和查找,并插入/更新到DB
  • 将所有这些REGEXP和模式移动到常量中,这样它们就不会为每一行实例化
  • 将代码清理成更小的块/方法,可能是通过创建表示CSV行的对象并为所有这些值提供访问器

假设:文件和文件中没有重复记录可以按任何顺序处理

我认为数据库不是瓶颈。您发出的HTTP请求更有可能。在你摆弄添加索引之前,我会做一些基准测试,看看时间花在哪里。

更新: 像这样的东西(我没有运行代码)

佣金任务:

...
csv.each do |row|
  ImportEventJob.perform_later row
end

工作后端

ActiveJob是Rails(后台)Job API。你可以配置一个后端,我喜欢Sidekiq,但其他人也会工作。详情见:

工作实施

class ImportEventJob < ActiveJob::Base

  def perform(row)
    event_csv = EventCsv.new(row)

    event = Event.find_by(time: event_csv.time, date: event_csv.date, ...)
    if event
      ...
    else
      ...
    end
  end
end

EventCsv(一行CSV数据的包装)

class EventCsv
  RE_WITH_DIACRITICS = /[ªÀÈÌÒÙàèìòùÁÉÍÓÚáéíóúÂÊÎÔÛâêîôûÃÑÕãñõÄËÏÖÜŸäëïöüÿ]/
  def initialize(row)
    @row = row
   end

  def time
    @row[9]
  end
  ...

  def name
    text = @row[7].gsub('text', '')
    if text.include? '&#xFFFD;&#xFFFD'
      text.gsub('&#xFFFD;&#xFFFD', 'e')
    else
      text.gsub(RE_WITH_DIACRITICS, '')
    end
  end
end