流口水：滑动窗口：时间正在选择甚至经过的数据

如何解决流口水：滑动窗口：时间正在选择甚至经过的数据

我想在3天的时间内查看异常事件。我有下面的drools代码，它输出所有事件，并且不限于3天。即使我放1天而不是3天，也会发生同样的问题。每次插入后我都会fireAllRules。

如果使用window：length（3），则可以正常运行，但不能按时间运行。我在做错什么吗？

另外，我还有一个问题@expires注释，它是否可以在我的自定义时间戳记上起作用，或者从事件/事实插入工作内存中起生效？

已将简化的单元测试放在https://github.com/mtekp/droolsExperiments/tree/main/testdrools

我正在使用drools:7.43.1.Final。

我的kmodule.xml

<?xml version="1.0" encoding="UTF-8"?>
<kmodule xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
         xmlns="http://www.drools.org/xsd/kmodule">

  <kbase name="TestEventsC" packages="test.events.continuous" eventProcessingMode="stream">
        <ksession name="TestDroolEventsContinous" clockType="realtime"/>
    </kbase>
    
</kmodule>

流口水的规则

package test.events.continuous


declare AnomalyCount
 @role(event)
 @timestamp(anomalyDate)
 @expires(3d)
end

rule "insert anomaly count"
  dialect "java"

  when
     $aei: AnomalyEventInsert()
     not AnomalyCount(anomalyDate == $aei.ts)

  then
     insert(new AnomalyCount($aei.getTs(),1));
     delete($aei);
end

rule "increment count"
  dialect "java"

  when
     $aei: AnomalyEventInsert()
     $ac: AnomalyCount(anomalyDate == $aei.ts)

  then
     modify($ac) { setAnomalyCount($ac.getAnomalyCount() + 1)};
     delete($aei);
end

rule "Check continuous 3 days"
   dialect "java"

   when
      $anomalies : List() from accumulate(AnomalyCount($dt : anomalyDate) over window:time(3d);collectList( $dt ))

   then
     System.out.println("anomalies: "+ $anomalies);
 end

AnomalyCount.java

import org.kie.api.deFinition.type.Role;

import java.util.Date;

@org.kie.api.deFinition.type.Role(Role.Type.EVENT)
@org.kie.api.deFinition.type.Timestamp("anomalyDate")
//@org.kie.api.deFinition.type.Expires("1d")
public class AnomalyCount {

    public AnomalyCount(Date anomalyDate,int anomalyCount) {
        this.anomalyDate = anomalyDate;
        this.anomalyCount = anomalyCount;
    }

    public Date getAnomalyDate() {
        return anomalyDate;
    }

    public void setAnomalyDate(Date anomalyDate) {
        this.anomalyDate = anomalyDate;
    }


    public int getAnomalyCount() {
        return anomalyCount;
    }

    public void setAnomalyCount(int anomalyCount) {
        this.anomalyCount = anomalyCount;
    }

    @Override
    public String toString() {
        return "AnomalyCount{" +
                "anomalyDate=" + anomalyDate +
                ",anomalyCount=" + anomalyCount +
                '}';
    }

    private Date anomalyDate;

    private int anomalyCount = 0;
}

AnomalyEventInsert.java

import java.time.LocalDate;
import java.time.ZoneOffset;
import java.util.Date;

public class AnomalyEventInsert extends AnomalyEvent {
    public AnomalyEventInsert(LocalDate ts,long sequenceCount,long value) {
        this.ts = Date.from(ts.atStartOfDay().toInstant( ZoneOffset.UTC));
        this.value = value;
        this.sequenceCount = sequenceCount;
    }

    public AnomalyEvent toAnomalyEvent() {
        return new AnomalyEvent(ts,sequenceCount,value);
    }

    @Override
    public String toString() {
        return "AnomalyEventInsert{" +
                "ts=" + ts +
                ",sequenceCount=" + sequenceCount +
                ",value=" + value +
                '}';
    }
}

当我插入第6，第7，第8，第9的数据时，窗口移动时得到的全部是四个，而不是最后三个。

found output

anomalies: [Tue Oct 06 05:30:00 IST 2020,Wed Oct 07 05:30:00 IST 2020,Thu Oct 08 05:30:00 IST 2020,Fri Oct 09 05:30:00 IST 2020]

instead of

anomalies: [Wed Oct 07 05:30:00 IST 2020,Fri Oct 09 05:30:00 IST 2020]