JAVA SDK 的使用

Linkis 提供了方便的JAVA和SCALA调用的接口,只需要引入linkis-computation-client的模块就可以进行使用,1.0后新增支持带Label提交的方式,下面将对兼容0.X的方式和1.0新增的方式进行介绍

  1. <dependency>
  2. <groupId>org.apache.linkis</groupId>
  3. <artifactId>linkis-computation-client</artifactId>
  4. <version>${linkis.version}</version>
  5. </dependency>
  6. 如:
  7. <dependency>
  8. <groupId>org.apache.linkis</groupId>
  9. <artifactId>linkis-computation-client</artifactId>
  10. <version>1.0.3</version>
  11. </dependency>

建立Java的测试类LinkisClientTest,具体接口含义可以见注释:

  1. package org.apache.linkis.client.test;
  2. import org.apache.linkis.common.utils.Utils;
  3. import org.apache.linkis.httpclient.dws.authentication.StaticAuthenticationStrategy;
  4. import org.apache.linkis.httpclient.dws.config.DWSClientConfig;
  5. import org.apache.linkis.httpclient.dws.config.DWSClientConfigBuilder;
  6. import org.apache.linkis.manager.label.constant.LabelKeyConstant;
  7. import org.apache.linkis.protocol.constants.TaskConstant;
  8. import org.apache.linkis.ujes.client.UJESClient;
  9. import org.apache.linkis.ujes.client.UJESClientImpl;
  10. import org.apache.linkis.ujes.client.request.JobSubmitAction;
  11. import org.apache.linkis.ujes.client.request.JobExecuteAction;
  12. import org.apache.linkis.ujes.client.request.ResultSetAction;
  13. import org.apache.linkis.ujes.client.response.*;
  14. import org.apache.commons.io.IOUtils;
  15. import java.util.HashMap;
  16. import java.util.Map;
  17. import java.util.concurrent.TimeUnit;
  18. public class LinkisClientTest {
  19. // 1. build config: linkis gateway url
  20. private static DWSClientConfig clientConfig = ((DWSClientConfigBuilder) (DWSClientConfigBuilder.newBuilder()
  21. .addServerUrl("http://127.0.0.1:9001/") //set linkis-mg-gateway url: http://{ip}:{port}
  22. .connectionTimeout(30000) //connectionTimeOut
  23. .discoveryEnabled(false) //disable discovery
  24. .discoveryFrequency(1, TimeUnit.MINUTES) // discovery frequency
  25. .loadbalancerEnabled(true) // enable loadbalance
  26. .maxConnectionSize(5) // set max Connection
  27. .retryEnabled(true) // set retry
  28. .readTimeout(30000) //set read timeout
  29. .setAuthenticationStrategy(new StaticAuthenticationStrategy()) //AuthenticationStrategy Linkis authen suppory static and Token
  30. .setAuthTokenKey("hadoop") // set submit user
  31. .setAuthTokenValue("123456"))) // set passwd or token (setAuthTokenValue("test"))
  32. .setDWSVersion("v1") //linkis rest version v1
  33. .build();
  34. // 2. new Client(Linkis Client) by clientConfig
  35. private static UJESClient client = new UJESClientImpl(clientConfig);
  36. public static void main(String[] args) {
  37. String user = "hadoop"; // 用户需要和AuthTokenKey的值保持一致
  38. String executeCode = "df=spark.sql(\"show tables\")\n" +
  39. "show(df)"; // code support:sql/hql/py/scala
  40. try {
  41. System.out.println("user : " + user + ", code : [" + executeCode + "]");
  42. // 3. build job and execute
  43. JobExecuteResult jobExecuteResult = toSubmit(user, executeCode);
  44. System.out.println("execId: " + jobExecuteResult.getExecID() + ", taskId: " + jobExecuteResult.taskID());
  45. // 4. get job jonfo
  46. JobInfoResult jobInfoResult = client.getJobInfo(jobExecuteResult);
  47. int sleepTimeMills = 1000;
  48. int logFromLen = 0;
  49. int logSize = 100;
  50. while (!jobInfoResult.isCompleted()) {
  51. // 5. get progress and log
  52. JobProgressResult progress = client.progress(jobExecuteResult);
  53. System.out.println("progress: " + progress.getProgress());
  54. JobLogResult logRes = client.log(jobExecuteResult, logFromLen, logSize);
  55. logFromLen = logRes.fromLine();
  56. // 0: info 1: warn 2: error 3: all
  57. System.out.println(logRes.log().get(3));
  58. Utils.sleepQuietly(sleepTimeMills);
  59. jobInfoResult = client.getJobInfo(jobExecuteResult);
  60. }
  61. JobInfoResult jobInfo = client.getJobInfo(jobExecuteResult);
  62. // 6. Get the result set list (if the user submits multiple SQLs at a time,
  63. // multiple result sets will be generated)
  64. String resultSet = jobInfo.getResultSetList(client)[0];
  65. // 7. get resultContent
  66. ResultSetResult resultSetResult = client.resultSet(ResultSetAction.builder().setPath(resultSet).setUser(jobExecuteResult.getUser()).build());
  67. System.out.println("metadata: " + resultSetResult.getMetadata()); // column name type
  68. System.out.println("res: " + resultSetResult.getFileContent()); //row data
  69. } catch (Exception e) {
  70. e.printStackTrace();// please use log
  71. IOUtils.closeQuietly(client);
  72. }
  73. IOUtils.closeQuietly(client);
  74. }
  75. private static JobExecuteResult toSubmit(String user, String code) {
  76. // 1. build params
  77. // set label map :EngineTypeLabel/UserCreatorLabel/EngineRunTypeLabel/Tenant
  78. Map<String, Object> labels = new HashMap<String, Object>();
  79. labels.put(LabelKeyConstant.ENGINE_TYPE_KEY, "spark-2.4.3"); // required engineType Label
  80. labels.put(LabelKeyConstant.USER_CREATOR_TYPE_KEY, user + "-APPName");// required execute user and creator eg:hadoop-IDE
  81. labels.put(LabelKeyConstant.CODE_TYPE_KEY, "py"); // required codeType
  82. // set start up map :engineConn start params
  83. Map<String, Object> startupMap = new HashMap<String, Object>(16);
  84. // Support setting engine native parameters,For example: parameters of engines such as spark/hive
  85. startupMap.put("spark.executor.instances", 2);
  86. // setting linkis params
  87. startupMap.put("wds.linkis.rm.yarnqueue", "dws");
  88. // 2. build jobSubmitAction
  89. JobSubmitAction jobSubmitAction = JobSubmitAction.builder()
  90. .addExecuteCode(code)
  91. .setStartupParams(startupMap)
  92. .setUser(user) //submit user
  93. .addExecuteUser(user) // execute user
  94. .setLabels(labels)
  95. .build();
  96. // 3. to execute
  97. return client.submit(jobSubmitAction);
  98. }
  99. }

运行上述的代码即可以完成任务提交/执行/日志/结果集获取等

  1. package org.apache.linkis.client.test
  2. import org.apache.commons.io.IOUtils
  3. import org.apache.commons.lang3.StringUtils
  4. import org.apache.linkis.common.utils.Utils
  5. import org.apache.linkis.httpclient.dws.authentication.StaticAuthenticationStrategy
  6. import org.apache.linkis.httpclient.dws.config.DWSClientConfigBuilder
  7. import org.apache.linkis.manager.label.constant.LabelKeyConstant
  8. import org.apache.linkis.ujes.client.request._
  9. import org.apache.linkis.ujes.client.response._
  10. import java.util
  11. import java.util.concurrent.TimeUnit
  12. object LinkisClientTest {
  13. // 1. build config: linkis gateway url
  14. val clientConfig = DWSClientConfigBuilder.newBuilder()
  15. .addServerUrl("http://127.0.0.1:9001/") //set linkis-mg-gateway url: http://{ip}:{port}
  16. .connectionTimeout(30000) //connectionTimeOut
  17. .discoveryEnabled(false) //disable discovery
  18. .discoveryFrequency(1, TimeUnit.MINUTES) // discovery frequency
  19. .loadbalancerEnabled(true) // enable loadbalance
  20. .maxConnectionSize(5) // set max Connection
  21. .retryEnabled(false) // set retry
  22. .readTimeout(30000) //set read timeout
  23. .setAuthenticationStrategy(new StaticAuthenticationStrategy()) //AuthenticationStrategy Linkis authen suppory static and Token
  24. .setAuthTokenKey("hadoop") // set submit user
  25. .setAuthTokenValue("hadoop") // set passwd or token (setAuthTokenValue("BML-AUTH"))
  26. .setDWSVersion("v1") //linkis rest version v1
  27. .build();
  28. // 2. new Client(Linkis Client) by clientConfig
  29. val client = UJESClient(clientConfig)
  30. def main(args: Array[String]): Unit = {
  31. val user = "hadoop" // execute user 用户需要和AuthTokenKey的值保持一致
  32. val executeCode = "df=spark.sql(\"show tables\")\n" +
  33. "show(df)"; // code support:sql/hql/py/scala
  34. try {
  35. // 3. build job and execute
  36. println("user : " + user + ", code : [" + executeCode + "]")
  37. // 推荐使用submit,支持传递任务label
  38. val jobExecuteResult = toSubmit(user, executeCode)
  39. println("execId: " + jobExecuteResult.getExecID + ", taskId: " + jobExecuteResult.taskID)
  40. // 4. get job jonfo
  41. var jobInfoResult = client.getJobInfo(jobExecuteResult)
  42. var logFromLen = 0
  43. val logSize = 100
  44. val sleepTimeMills: Int = 1000
  45. while (!jobInfoResult.isCompleted) {
  46. // 5. get progress and log
  47. val progress = client.progress(jobExecuteResult)
  48. println("progress: " + progress.getProgress)
  49. val logObj = client.log(jobExecuteResult, logFromLen, logSize)
  50. logFromLen = logObj.fromLine
  51. val logArray = logObj.getLog
  52. // 0: info 1: warn 2: error 3: all
  53. if (logArray != null && logArray.size >= 4 && StringUtils.isNotEmpty(logArray.get(3))) {
  54. println(s"log: ${logArray.get(3)}")
  55. }
  56. Utils.sleepQuietly(sleepTimeMills)
  57. jobInfoResult = client.getJobInfo(jobExecuteResult)
  58. }
  59. if (!jobInfoResult.isSucceed) {
  60. println("Failed to execute job: " + jobInfoResult.getMessage)
  61. throw new Exception(jobInfoResult.getMessage)
  62. }
  63. // 6. Get the result set list (if the user submits multiple SQLs at a time,
  64. // multiple result sets will be generated)
  65. val jobInfo = client.getJobInfo(jobExecuteResult)
  66. val resultSetList = jobInfoResult.getResultSetList(client)
  67. println("All result set list:")
  68. resultSetList.foreach(println)
  69. val oneResultSet = jobInfo.getResultSetList(client).head
  70. // 7. get resultContent
  71. val resultSetResult: ResultSetResult = client.resultSet(ResultSetAction.builder.setPath(oneResultSet).setUser(jobExecuteResult.getUser).build)
  72. println("metadata: " + resultSetResult.getMetadata) // column name type
  73. println("res: " + resultSetResult.getFileContent) //row data
  74. } catch {
  75. case e: Exception => {
  76. e.printStackTrace() //please use log
  77. }
  78. }
  79. IOUtils.closeQuietly(client)
  80. }
  81. def toSubmit(user: String, code: String): JobExecuteResult = {
  82. // 1. build params
  83. // set label map :EngineTypeLabel/UserCreatorLabel/EngineRunTypeLabel/Tenant
  84. val labels: util.Map[String, Any] = new util.HashMap[String, Any]
  85. labels.put(LabelKeyConstant.ENGINE_TYPE_KEY, "spark-2.4.3"); // required engineType Label
  86. labels.put(LabelKeyConstant.USER_CREATOR_TYPE_KEY, user + "-APPName"); // 请求的用户和应用名,两个参数都不能少,其中APPName不能带有"-"建议替换为"_"
  87. labels.put(LabelKeyConstant.CODE_TYPE_KEY, "py"); // 指定脚本类型
  88. val startupMap = new java.util.HashMap[String, Any]()
  89. // Support setting engine native parameters,For example: parameters of engines such as spark/hive
  90. startupMap.put("spark.executor.instances", 2);
  91. // setting linkis params
  92. startupMap.put("wds.linkis.rm.yarnqueue", "default");
  93. // 2. build jobSubmitAction
  94. val jobSubmitAction = JobSubmitAction.builder
  95. .addExecuteCode(code)
  96. .setStartupParams(startupMap)
  97. .setUser(user) //submit user
  98. .addExecuteUser(user) //execute user
  99. .setLabels(labels)
  100. .build
  101. // 3. to execute
  102. client.submit(jobSubmitAction)
  103. }
  104. }