Java Scala获取所有注解的类信息

要想获取使用指定注解的类信息，可借助工具：

org.reflections.Reflections

此工具将Java反射进行了高级封装，Reflections 通过扫描 classpath，索引元数据，允许在运行时查询这些元数据，也可以保存收集项目中多个模块的元数据信息。

使用 Reflections 可以查询以下元数据信息： 

1）获得某个类型的所有子类型
2）获得标记了某个注解的所有类型／成员变量，支持注解参数匹配。
3）使用正则表达式获得所有匹配的资源文件
4）获得所有特定签名（包括参数，参数注解，返回值）的方法

Reflections 依赖 Google 的 Guava 库和 Javassist 库。

Maven引入方式：

<dependency>
    <groupId>org.reflections</groupId>
    <artifactId>reflections</artifactId>
    <version>0.9.11</version>
</dependency>

sbt引入方式：

"org.reflections" % "reflections" % "0.9.11"

首先自定义注解：

package com.today.service.financetask.job

import java.lang.annotation.*;

/**
 * 类功能描述：job 信息注解
 *
 * @author WangXueXing create at 19-5-4 上午9:14
 * @version 1.0.0
 */
@Target(ElementType.TYPE)
@Retention(RetentionPolicy.RUNTIME)
@Inherited
public @interface JobInfo {
    /**
     * job id
     * @return
     */
    String jobId();

    /**
     * job name
     * @return
     */
    String jobName();

    /**
     * default cron
     * @return
     */
    String defaultCron();
}

将某些类添加注解：

package com.today.service.financetask.job

import java.util.Calendar

import com.today.api.checkaccount.scala.CheckAccountServiceClient
import com.today.api.checkaccount.scala.enums.FlatFormTypeEnum
import com.today.api.checkaccount.scala.request.ReconciliationRequest
import com.today.service.financetask.job.define.AbstractJob
import com.today.service.financetask.utils.JobInfo
import org.quartz.JobExecutionContext

/**
  * 自动对账
  */
@JobInfo(jobId="CHECK_ACCOUNT_PROCESS", jobName="对账系统自动对账定时任务", defaultCron="0 0 13 * * ?")
class CheckAccountJob extends AbstractJob{
  /**
    * start up the scheduled task
    *
    * @param context JobExecutionContext
    */
  override def run(context: JobExecutionContext): Unit = {
    val cal = Calendar.getInstance
    cal.add(Calendar.DATE, -1)
    new CheckAccountServiceClient().appReconciliation(new ReconciliationRequest(FlatFormTypeEnum.TODAY_APP,None))
  }
}

import com.today.service.financetask.action.DailyStatementAction
import com.today.service.financetask.job.define.AbstractJob
import com.today.service.financetask.utils.JobInfo
import org.quartz.JobExecutionContext
import org.springframework.stereotype.Service

/**
  * 日次处理定时任务处理
  *
  * @author zhangc create at 2018/5/11 14:08
  * @version 0.0.1
  */
@Service
@JobInfo(jobId="DAY_TIME_PROCESS", jobName="日次处理定时任务", defaultCron="0 30 2 * * ?")
class DayTimeProcessJob extends AbstractJob{
  /**
    * start up the scheduled task
    *
    * @param context JobExecutionContext
    */
  override def run(context: JobExecutionContext): Unit = {
    new DailyStatementAction().execute
  }
}

通过Java反射及Reflections工具类实现被JobInfo注解的所有类信息：

import java.util

import com.today.service.financetask.utils.JobInfo
import org.quartz.Job
import org.reflections.Reflections

import scala.collection.JavaConverters._
import scala.collection.mutable

/**
  * 通过注解获取所有通用Job信息
  *
  * @author BarryWang create at 2018/5/12 10:45
  * @version 0.0.1
  */
object JobEnum {
  /**
    * 获取添加JobInfo注解的类信息
    */
  val jobWithAnnotation: util.Set[Class[_]] = new Reflections("com.today.service.financetask.job").getTypesAnnotatedWith(classOf[JobInfo])

  /**
    * 获取所有job枚举值
    * @return
    */
  def values : mutable.Set[JobInfo] = jobWithAnnotation.asScala.map(getJobInfo(_))

  /**
    * 根据job class 获取job 信息
    * @param jobClass
    * @return
    */
  def getJobInfo(jobClass : Class[_]): JobInfo = jobClass.getAnnotation(classOf[JobInfo])

  /**
    * jobId与jobName映射关系
    * @return
    */
  def jobIdNameMap : Map[String, String]={
    jobWithAnnotation.asScala.map{sub =>
      val jobInfo = getJobInfo(sub)
      Map(jobInfo.jobId() -> jobInfo.jobName())
    }.fold(Map())((i,j) => i++j)
  }

  /**
    * JobObject与JobEnum映射关系
    * @return
    */
  def jobClassInfoMap: Map[String, JobInfo] = {
    jobWithAnnotation.asScala.map{sub =>
      Map(sub.getName -> getJobInfo(sub))
    }.fold(Map())((i,j) => i++j)
  }

  /**
    * jobId与JobEnum映射关系
    * @return
    */
  def jobIdInfoMap: Map[String, JobInfo] = {
    jobWithAnnotation.asScala.map{sub =>
      val jobInfo = getJobInfo(sub)
      Map(jobInfo.jobId() -> jobInfo)
    }.fold(Map())((i,j) => i++j)
  }

  /**
    * jobId与JobObject映射关系
    * @return
    */
  def jobIdClassMap: Map[String, Class[_ <: Job]] = {
    jobWithAnnotation.asScala.map{sub =>
      Map(getJobInfo(sub).jobId() -> sub.asInstanceOf[Class[_ <: Job]])
    }.fold(Map[String, Class[_ <: Job]]())((i,j) => i++j)
  }

  def main(args: Array[String]): Unit = {
    println(jobIdClassMap)
  }
}

至此，我们就可以获取所有被特定注解引用的类信息及注解信息，我们就可以全局管理特定类信息。

实现JobEnum后就可以统一对定时任务管理实现：

１．新添加定时任务完全可以制定一个Job子类，其他操作自动维护进去；

２．每个Job子类里面都需要实现 override def getJobAndApiInfo(context: JobExecutionContext): (String, String, JobEnum) 方法，这个也可以省掉，直接在父类统一实现；

３．统一修改定时任务开关及时间接口；

４．统一对定时任务启动时重启，就可以统一重启，不需要单独添加代码启动；

1上面代码片段“将某些类添加注解”

2父类代码如下：

import java.io.{PrintWriter, StringWriter}

import com.github.dapeng.core.helper.MasterHelper
import com.today.api.financetask.scala.enums.TScheduledTaskLogEnum
import com.today.service.financetask.action.sql.ScheduledTaskLogSql
import com.today.service.financetask.util.{AppContextHolder, Debug}
import com.today.service.financetask.utils.JobInfo
import org.quartz.{Job, JobExecutionContext}
import org.slf4j.LoggerFactory
import org.springframework.transaction.TransactionStatus
import org.springframework.transaction.support.TransactionTemplate

import scala.util.{Failure, Success, Try}

/**
  * the abstract class for job
  */
trait AbstractJob extends Job{
  /** 日志 */
  val logger = LoggerFactory.getLogger(getClass)

  /**
    * execute the job
    * @param context
    */
  override def execute(context: JobExecutionContext): Unit = {
    getJobAndApiInfo(context) match {
      case Some(x) => execute(context, x.jobId, x.jobName)
      case None => logger.error("没有定义JobEnum及对应JobObject，请检查")
    }
  }

  /**
    * execute the job
    * @param context
    * @param jobId
    * @param jobName
    */
  def execute(context: JobExecutionContext, jobId : String, jobName: String): Unit = {
    //判断是否是选中的master节点，不是master节点不执行定时任务
    if (!MasterHelper.isMaster("com.today.api.financetask.service.FinanceScheduledService", "1.0.0")) {
      logger.info(s"Can't select master to run the job ${jobId}: ${jobName}")
      return
    }

    //记录开始日志
    val logId = ScheduledTaskLogSql.insertScheduledTaskLog(jobId)
    context.put("logId", logId)
    logger.info(s"Starting the job ${jobId}: ${jobName} ...")

    //事物处理
    val transactionTemplate: TransactionTemplate = AppContextHolder.getBean("transactionTemplate")
    transactionTemplate.execute((status: TransactionStatus) =>{
      Debug.reset()
      //执行任务
      Try(Debug.trace(s"${jobId}:${jobName}")(run(context))) match
      {
        case Success(x) => {
          logger.info(s"Successfully execute the job ${jobId}: ${jobName}")
          successLog(logId)
        }
        case Failure(e) => {
          logger.error(s"Failure execute the job ${jobId}: ${jobName}", e)
          failureLog(logId, status, e)
        }
      }
      Debug.info()
    })
  }

  /**
    * get the api information
    * @return (interface name, interface version, JobEnum)
    */
  def getJobAndApiInfo(context: JobExecutionContext) : Option[JobInfo] ={
    JobEnum.jobClassInfoMap.get(this.getClass.getName)
  }

  /**
    * start up the scheduled task
    * @param context JobExecutionContext
    */
  def run(context: JobExecutionContext)

  /**
    * 成功日志记录
    * @param logId
    */
  def successLog(logId: Long): Unit ={
    ScheduledTaskLogSql.updateExportReportRecord(logId, TScheduledTaskLogEnum.SUCCESS, "Success")
  }

  /**
    * 失败日志记录
    * @param logId
    */
  def failureLog(logId: Long, status: TransactionStatus, e: Throwable): Unit ={
    status.setRollbackOnly()
    ScheduledTaskLogSql.updateExportReportRecord(logId, TScheduledTaskLogEnum.FAILURE, getExceptionStack(e))
  }

  /**
    *
    * 功能说明:在日志文件中 ，打印异常堆栈
    * @param e : Throwable
    * @return : String
    */
  def getExceptionStack(e: Throwable): String = {
    val errorsWriter = new StringWriter
    e.printStackTrace(new PrintWriter(errorsWriter))
    errorsWriter.toString
  }
}

3 统一修改定时任务开关及时间代码如下：

import com.today.api.financetask.scala.enums.{TScheduledTaskHasDeletedEnum, TScheduledTaskIsStartEnum}
import com.today.api.financetask.scala.request.StartOrStopByNameRequest
import com.today.api.financetask.scala.response.ServiceResponse
import com.today.service.commons.Action
import com.today.service.commons.Assert.assert
import com.today.service.commons.exception.CommonException.illegalArgumentException
import com.today.service.financetask.action.sql.ScheduledTaskActionSql
import com.today.service.financetask.dto.TScheduledTask
import com.today.service.financetask.job.define.JobEnum
import com.today.service.financetask.query.sql.ScheduledTaskQuerySql
import com.today.service.financetask.util.{CronConverter, QuartzManager}

/**
  * @description: 启停定时任务
  * @author zhangc
  * @date 201881 0001 15:02
  * @version 1.0.0
  */
class StartOrStopByNameAction (request: StartOrStopByNameRequest)  extends Action[ServiceResponse] {
  override def preCheck: Unit = {
    assert(!TScheduledTaskIsStartEnum.isUndefined(request.flag.id), illegalArgumentException("错误的启停标志!"))
  }

  /**
    * 根据传入的定时任务名称和启停标志，来判断启动或者停止定时任务
    * 如果是1则更新数据库，删除定时任务，重新添加定时任务
    * 如果是0则更新数据库，删除定时任务
    * @return
    */
  override def action: ServiceResponse = {
    val scheduledTask = ScheduledTaskQuerySql.queryByJobId(request.processName)
    val cron = CronConverter.convertHourMinuteToCron(request.processTime)
    val jobInfo = JobEnum.jobIdInfoMap(request.processName)
    //修改定时任务时间, 保存入库
    if(scheduledTask.isEmpty){
      ScheduledTaskActionSql.insertTScheduledTask(
        TScheduledTask(jobInfo.jobName,
          request.processName,
          cron,
          None,
          request.flag.id,
          None,
          null,
          null,
          TScheduledTaskHasDeletedEnum.NO.id))
    } else {
      ScheduledTaskActionSql.updateTaskIsStart(request.flag.id,cron, request.processName)
    }
    request.flag match {
      case TScheduledTaskIsStartEnum.YES =>  QuartzManager.modifyJobTime(request.processName, cron, JobEnum.jobIdClassMap(jobInfo.jobId()))
      case _ =>  QuartzManager.removeJob(request.processName)
    }

    ServiceResponse("200","success")
  }
}

４下面就是统一对定时任务启动时重启，就可以统一重启，不需要单独添加代码启动：

import com.today.api.financetask.scala.enums.TScheduledTaskIsStartEnum
import com.today.api.financetask.scala.request.QueryAutoConfigRequest
import com.today.service.financetask.job._
import com.today.service.financetask.job.define.JobEnum
import com.today.service.financetask.query.sql.{AutoConfigQuerySql, ScheduledTaskQuerySql}
import com.today.service.financetask.util.QuartzManager
import com.today.service.financetask.utils.JobInfo
import org.slf4j.LoggerFactory
import org.springframework.context.ApplicationListener
import org.springframework.context.event.ContextRefreshedEvent
import org.springframework.stereotype.Service

/**
  *  类功能描述: 定时器监听器, 服务启动时启动定时器
  *
  * @author BarryWang create at 2018/5/11 12:04
  * @version 0.0.1
  */
@Service
class ScheduleStartListener extends ApplicationListener[ContextRefreshedEvent] {
  /** 日志 */
  val logger = LoggerFactory.getLogger(getClass)

  /**
    * 启动加载执行定时任务
    */
  override def onApplicationEvent(event: ContextRefreshedEvent): Unit = {
    logger.info("=======服务器重启定时任务启动start=======")
    //启动服务时恢复常规定时任务
    JobEnum.values.foreach(recoveryCommonJob(_))
    logger.info("=======服务器重启定时任务启动end=======")
  }

  /**
    * 恢复通用定时任务
    * @param jobInfo 定时任务枚举
    */
  private def recoveryCommonJob(jobInfo: JobInfo)={
    try {
      val jobClass = JobEnum.jobIdClassMap(jobInfo.jobId)
      ScheduledTaskQuerySql.queryByJobId(jobInfo.jobId) match {
        case Some(x) => {
          x.isStart match {
            case TScheduledTaskIsStartEnum.YES.id => {
              QuartzManager.addJobByCron(jobInfo.jobId, x.jobCron, jobClass)
              logger.info(s"定时任务：'${jobInfo.jobName}'启动成功!")
            }
            case _ => logger.info(s"定时任务：'${jobInfo.jobName}'is_start标志为0，不启动")
          }
        }
        case None => QuartzManager.addJobByCron(jobInfo.jobId, jobInfo.defaultCron(), jobClass)
      }
    } catch {
      case e : Exception => logger.error(s"定时任务：'${jobInfo.jobName}'启动失败， 失败原因：", e)
    }
  }
}

本部分也是对Quartz实现可配置的分布式定时任务的优化重构，可详见：

Quartz实现分布式可动态配置的定时任务