Flink_ 页面广告点击量统计（黑名单过滤）

页面广告分析_ 页面广告点击量统计

        电商网站的市场营销商业指标中，除了自身的 APP 推广，还会考虑到页面上的

广告投放（包括自己经营的产品和其它网站的广告）。所以广告相关的统计分析，

也是市场营销的重要指标。

        对于广告的统计，最简单也最重要的就是页面广告的点击量，网站往往需要根

据广告点击量来制定定价策略和调整推广方式，而且也可以借此收集用户的偏好信

息。更加具体的应用是，我们可以根据用户的地理位置进行划分，从而总结出不同

省份用户对不同广告的偏好，这样更有助于广告的精准投放。 

以 province 进行 keyBy，然后开一小时的时间窗口， 滑动距离为 5 秒，统计窗口内的点击事件数量

       

 黑名单过滤

        同一用户的重复点击是会叠加计算的。在实际场 景中，同一用户确实可能反复点开同一个广告，这也说明了用户对广告更大的兴趣； 但是如果用户在一段时间非常频繁地点击广告，这显然不是一个正常行为，有刷点击量的嫌疑。所以我们可以对一段时间内（比如一天内）的用户点击行为进行约束， 如果对同一个广告点击超过一定限额（比如 100 次），应该把该用户加入黑名单并报警，此后其点击行为不应该再统计

输入数据：userId adId 省 城市 时间戳

 561558 3611281  guangdong  shenzhen 1511658120

package UserBehaviorAnalysis.MarketAnalysis

import java.sql.Timestamp

import org.apache.flink.api.common.functions.AggregateFunction
import org.apache.flink.api.common.state.{ValueState, ValueStateDescriptor}
import org.apache.flink.streaming.api.TimeCharacteristic
import org.apache.flink.streaming.api.functions.KeyedProcessFunction
import org.apache.flink.streaming.api.scala._
import org.apache.flink.streaming.api.scala.function.WindowFunction
import org.apache.flink.streaming.api.windowing.time.Time
import org.apache.flink.streaming.api.windowing.windows.TimeWindow
import org.apache.flink.util.Collector


// 定义输入样例类
case class AdClickLog(userId: Long, adId: Long, province: String, city: String, timestamp: Long)
// 定义输出样例类
case class AdClickCountByProvince(windowEnd: String, province: String, count: Long)
// 侧输出流黑名单报警信息样例类
case class BlackListUserWarning(userId: Long, adId: Long, msg: String)

object AdClickAnalysis {
  def main(args: Array[String]): Unit = {
    val env = StreamExecutionEnvironment.getExecutionEnvironment
    env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime)
    env.setParallelism(1)

    val inputStream = env.readTextFile("D:\\Mywork\\workspace\\Project_idea\\flink-2021\\src\\main\\resources\\AdClickLog.csv")
    val adLogStream = inputStream.map { data =>
      val arr = data.split(",")
      AdClickLog(arr(0).toLong, arr(1).toLong, arr(2), arr(3), arr(4).toInt)
    }.assignAscendingTimestamps(_.timestamp * 1000L)

    // 插入一步过滤操作，并将有刷单行为的用户输出到侧输出流（黑名单报警）
    val filterBlackListUserStream = adLogStream
      .keyBy(data => (data.userId, data.adId))
      .process(new FliterBlackListUserResult(100))


    val adCountResultStream = filterBlackListUserStream
      .keyBy(_.province)
      .timeWindow(Time.hours(1), Time.seconds(5))
      .aggregate(new AdCountAgg(), new AdCountWindowResult())

    filterBlackListUserStream.getSideOutput(new OutputTag[BlackListUserWarning]("warning")).print("warning")
    adCountResultStream.print("count result")
    env.execute()

  }
}

class AdCountAgg() extends AggregateFunction[AdClickLog, Long, Long]{
  override def createAccumulator(): Long = 0L

  override def add(value: AdClickLog, accumulator: Long): Long = accumulator + 1

  override def getResult(accumulator: Long): Long = accumulator

  override def merge(a: Long, b: Long): Long = a + b
}

class AdCountWindowResult() extends WindowFunction[Long, AdClickCountByProvince, String, TimeWindow]{
  
  override def apply(key: String, window: TimeWindow, input: Iterable[Long], out: Collector[AdClickCountByProvince]): Unit = {
    val end = new Timestamp(window.getEnd).toString
    out.collect(AdClickCountByProvince(end, key, input.head))
  }
}

// 自定义KeyedProcessFunction,黑名单过滤
class FliterBlackListUserResult(maxCount: Long) extends KeyedProcessFunction[(Long, Long), AdClickLog, AdClickLog]{
  // 定义状态，保存用户对广告的点击量，每天0点定时清空状态的时间戳，标记当前用户是否已经进入黑名单
  lazy val countState: ValueState[Long] = getRuntimeContext.getState(new ValueStateDescriptor[Long]("Adcount", classOf[Long]))
  lazy val resetTimerTsState: ValueState[Long] = getRuntimeContext.getState(new ValueStateDescriptor[Long]("reset-ts", classOf[Long]))
  lazy val isBlackState: ValueState[Boolean] = getRuntimeContext.getState(new ValueStateDescriptor[Boolean]("is-back", classOf[Boolean]))



  override def processElement(i: AdClickLog, context: KeyedProcessFunction[(Long, Long), AdClickLog, AdClickLog]#Context, collector: Collector[AdClickLog]): Unit = {
    val curCount = countState.value()
    // 判断只要是第一个数据来了，直接注册0点的清空状态定时器
    if (curCount == 0){
      val ts = (context.timerService().currentProcessingTime()/(1000 * 60 * 60 * 24) + 1) * (24 * 60 * 60 * 1000) - 8 * 60 * 60 * 1000
      resetTimerTsState.update(ts)
      context.timerService().registerProcessingTimeTimer(ts)
    }

    // 判断count值是否已经达到定义的阈值，如果超过就输出到黑名单
    if (curCount >= maxCount){
      // 判断是否已经在黑名单里，没有的话才输出侧输出流
      if (!isBlackState.value()){
        isBlackState.update(true)
        context.output(new OutputTag[BlackListUserWarning]("warning"), BlackListUserWarning(i.userId, i.adId, "Click ad over " + maxCount + " times today."))
      }
      return
    }

    // 正常情况，count加1，然后将数据原样输出
    countState.update(curCount + 1)
    collector.collect(i)
  }

  override def onTimer(timestamp: Long, ctx: KeyedProcessFunction[(Long, Long), AdClickLog, AdClickLog]#OnTimerContext, out: Collector[AdClickLog]): Unit = {
    if (timestamp == resetTimerTsState.value()){
      isBlackState.clear()
      countState.clear()
    }
  }
}


//输出数据
warning> BlackListUserWarning(937166,1715,Click ad over 100 times today.)
count result> AdClickCountByProvince(2017-11-26 09:00:05.0,beijing,1)
count result> AdClickCountByProvince(2017-11-26 09:00:10.0,beijing,1)