如何快速在 Apache DolphinScheduler 新扩展一个任务插件？

mvn archetype:generate \
    -DarchetypeGroupId=org.apache.dolphinscheduler \
    -DarchetypeArtifactId=dolphinscheduler-hive-client-task \
    -DarchetypeVersion=1.10.0 \
    -DgroupId=org.apache.dolphinscheduler \
    -DartifactId=dolphinscheduler-hive-client-task \
    -Dversion=0.1 \
    -Dpackage=org.apache.dolphinscheduler \
    -DinteractiveMode=false 

org.apache.dolphinscheduler
     dolphinscheduler-spi
     ${dolphinscheduler.lib.version}
     ${common.lib.scope}








     org.apache.dolphinscheduler
     dolphinscheduler-task-api
     ${dolphinscheduler.lib.version}
     ${common.lib.scope}

package org.apache.dolphinscheduler.plugin.task.hive;
import org.apache.dolphinscheduler.spi.params.base.PluginParams;
import org.apache.dolphinscheduler.spi.task.TaskChannel;
import org.apache.dolphinscheduler.spi.task.TaskChannelFactory;
import java.util.List;
public class HiveClientTaskChannelFactory implements TaskChannelFactory {
    /**
    * Create task channel, execute task through this channel
     * @return task channel
     */
    @Override
    public TaskChannel create() {
        return new HiveCliTaskChannel();
    }
    /**
    * Returns the global unique identifier of this task
     * @return task name
     */
    @Override
    public String getName() {
        return "HIVECLI";
    }
    /**
    * Parameters required for front-end pages
     * @return
     */
    @Override
    public List getParams() {
        return null;
    }
}

   void cancelApplication(boolean status);
    /**
     * 构建可执行任务
     */
    AbstractTask createTask(TaskRequest taskRequest);
    public class HiveClientTaskChannel implements TaskChannel {
    @Override
    public void cancelApplication(boolean b) {
        //do nothing
    }
    @Override
    public AbstractTask createTask(TaskRequest taskRequest) {
        return new HiveClientTask(taskRequest);
    }
}

package com.jegger.dolphinscheduler.plugin.task.hive;
import org.apache.dolphinscheduler.spi.task.AbstractParameters;
import org.apache.dolphinscheduler.spi.task.ResourceInfo;
import java.util.List;
public class HiveClientParameters extends AbstractParameters {
    /**
     * 用HiveClient执行,最简单的方式就是将所有SQL全部贴进去即可,所以我们只需要一个SQL参数
     */
    private String sql;
    public String getSql() {
        return sql;
    }
    public void setSql(String sql) {
        this.sql = sql;
    }
    @Override
    public boolean checkParameters() {
        return sql != null;
    }
    @Override
    public List getResourceFilesList() {
        return null;
    }
}

package org.apache.dolphinscheduler.plugin.task.hive;
import org.apache.dolphinscheduler.plugin.task.api.AbstractYarnTask;
import org.apache.dolphinscheduler.spi.task.AbstractParameters;
import org.apache.dolphinscheduler.spi.task.request.TaskRequest;
import org.apache.dolphinscheduler.spi.utils.JSONUtils;
import java.io.BufferedWriter;
import java.io.IOException;
import java.nio.charset.StandardCharsets;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.Paths;
public class HiveClientTask extends AbstractYarnTask {
    /**
     * hive client parameters
     */
    private HiveClientParameters hiveClientParameters;
    /**
     * taskExecutionContext
     */
    private final TaskRequest taskExecutionContext;
    public HiveClientTask(TaskRequest taskRequest) {
        super(taskRequest);
        this.taskExecutionContext = taskRequest;
    }
    /**
     * task init method
     */
    @Override
    public void init() {
        logger.info("hive client task param is {}", JSONUtils.toJsonString(taskExecutionContext));
        this.hiveClientParameters = JSONUtils.parseObject(taskExecutionContext.getTaskParams(), HiveClientParameters.class);
        if (this.hiveClientParameters != null && !hiveClientParameters.checkParameters()) {
            throw new RuntimeException("hive client task params is not valid");
        }
    }
    /**
     * build task execution command
     *
     * @return task execution command or null
     */
    @Override
    protected String buildCommand() {
        String filePath = getFilePath();
        if (writeExecutionContentToFile(filePath)) {
            return "hive -f " + filePath;
        }
        return null;
    }
    /**
     * get hive sql write path
     *
     * @return file write path
     */
    private String getFilePath() {
        return String.format("%s/hive-%s-%s.sql", this.taskExecutionContext.getExecutePath(), this.taskExecutionContext.getTaskName(), this.taskExecutionContext.getTaskInstanceId());
    }
    @Override
    protected void setMainJarName() {
        //do nothing
    }
    /**
     * write hive sql to filepath
     *
     * @param filePath file path
     * @return write success?
     */
    private boolean writeExecutionContentToFile(String filePath) {
        Path path = Paths.get(filePath);
        try (BufferedWriter writer = Files.newBufferedWriter(path, StandardCharsets.UTF_8)) {
            writer.write(this.hiveClientParameters.getSql());
            logger.info("file:" + filePath + "write success.");
            return true;
        } catch (IOException e) {
            logger.error("file:" + filePath + "write failed.please path auth.");
            e.printStackTrace();
            return false;
        }
    }
    @Override
    public AbstractParameters getParameters() {
        return this.hiveClientParameters;
    }
}

# 1,Resource下创建META-INF/services文件夹,创建接口全类名相同的文件
zhang@xiaozhang resources % tree ./
./
└── META-INF
    └── services
        └── org.apache.dolphinscheduler.spi.task.TaskChannelFactory
# 2,在文件中写入实现类的全限定类名
zhang@xiaozhang resources % more META-INF/services/org.apache.dolphinscheduler.spi.task.TaskChannelFactory 
org.apache.dolphinscheduler.plugin.task.hive.HiveClientTaskChannelFactory

## 1,打包
mvn clean install
## 2,部署
cp ./target/dolphinscheduler-task-hiveclient-1.0.jar $DOLPHINSCHEDULER_HOME/lib/
## 3,restart dolphinscheduler server