Examples with DataQualityTaskException com.webank.wedatasphere.qualitis.exception.DataQualityTaskException used on opensource projects

Example 1 with DataQualityTaskException

use of com.webank.wedatasphere.qualitis.exception.DataQualityTaskException in project Qualitis by WeBankFinTech.

the class ExecutionManagerImpl method submitApplication.

/**
 * Submit job to linkis
 */
@Override
public List<TaskSubmitResult> submitApplication(List<Rule> rules, String nodeName, String createTime, String user, String database, StringBuffer partition, Date date, Application application, String cluster, String startupParam, String setFlag, Map<String, String> execParams, StringBuffer runDate, Map<Long, Map> dataSourceMysqlConnect) throws ArgumentException, TaskTypeException, ConvertException, DataQualityTaskException, RuleVariableNotSupportException, RuleVariableNotFoundException, JobSubmitException, ClusterInfoNotConfigException, IOException, UnExpectedRequestException, MetaDataAcquireFailedException {
    String csId = rules.iterator().next().getCsId();
    // Check if cluster supported
    LOGGER.info("Start to collect rule to clusters");
    Map<String, List<Rule>> clusterNameMap = getRuleCluster(rules);
    LOGGER.info("Succeed to classify rules by cluster, cluster map: {}", clusterNameMap);
    if (StringUtils.isNotBlank(cluster)) {
        LOGGER.info("When pick up a cluster, these datasources of rules must be from one cluster. Now start to put into the specify cluster.\n");
        putAllRulesIntoSpecifyCluster(clusterNameMap, cluster);
        LOGGER.info("Success to put into the specify cluster.\n");
    }
    List<TaskSubmitResult> taskSubmitResults = new ArrayList<>();
    for (String clusterName : clusterNameMap.keySet()) {
        List<Rule> clusterRules = clusterNameMap.get(clusterName);
        if (StringUtils.isNotBlank(cluster)) {
            clusterName = cluster;
        }
        ClusterInfo clusterInfo = clusterInfoDao.findByClusterName(clusterName);
        LOGGER.info("Start to check cluster config.");
        if (clusterInfo == null) {
            throw new ClusterInfoNotConfigException(clusterName + " {&DOES_NOT_EXIST}");
        }
        LOGGER.info("Succeed to pass the check of cluster config. All cluster of rules are configured");
        // Divide rule into tasks
        List<DataQualityTask> tasks = TaskDividerFactory.getDivider().divide(clusterRules, application.getId(), createTime, partition.toString(), date, database, user, taskExecuteLimitConfig.getTaskExecuteRuleSize());
        LOGGER.info("Succeed to divide application into tasks. result: {}", tasks);
        // Save divided tasks
        saveDividedTask(tasks, clusterInfo, rules, application, createTime);
        // Convert tasks into job
        List<DataQualityJob> jobList = new ArrayList<>();
        for (DataQualityTask task : tasks) {
            DataQualityJob job = templateConverterFactory.getConverter(task).convert(task, date, setFlag, execParams, runDate.toString(), clusterInfo.getClusterType(), dataSourceMysqlConnect);
            job.setUser(task.getUser());
            jobList.add(job);
            List<Long> ruleIdList = task.getRuleTaskDetails().stream().map(r -> r.getRule().getId()).collect(Collectors.toList());
            LOGGER.info("Succeed to convert rule_id: {} into code. code: {}", ruleIdList, job.getJobCode());
        }
        LOGGER.info("Succeed to convert all template into codes. codes: {}", jobList);
        // Submit job to linkis
        List<JobSubmitResult> submitResults = new ArrayList<>();
        for (DataQualityJob job : jobList) {
            String code = String.join("\n", job.getJobCode());
            String proxy = job.getUser();
            Long taskId = job.getTaskId();
            // Compatible with new and old submission interfaces.
            JobSubmitResult result = null;
            boolean engineReUse = false;
            if (StringUtils.isNotBlank(startupParam)) {
                String[] startupParams = startupParam.split(SpecCharEnum.DIVIDER.getValue());
                for (String param : startupParams) {
                    if (StringUtils.isEmpty(param)) {
                        continue;
                    }
                    String[] paramStrs = param.split("=");
                    if (paramStrs.length < 2) {
                        continue;
                    }
                    String key = paramStrs[0];
                    String value = paramStrs[1];
                    if ("engine_reuse".equals(key)) {
                        if ("true".equals(value)) {
                            engineReUse = true;
                            startupParam = startupParam.replace("engine_reuse=true", "");
                        } else {
                            engineReUse = false;
                            startupParam = startupParam.replace("engine_reuse=false", "");
                        }
                        break;
                    }
                }
            }
            if (clusterInfo.getClusterType().endsWith(LINKIS_ONE_VERSION)) {
                result = abstractJobSubmitter.submitJobNew(code, linkisConfig.getEngineName(), StringUtils.isNotBlank(proxy) ? proxy : user, clusterInfo.getLinkisAddress(), clusterName, taskId, csId, nodeName, StringUtils.isNotBlank(startupParam) ? startupParam : job.getStartupParam(), engineReUse);
            } else {
                result = abstractJobSubmitter.submitJob(code, linkisConfig.getEngineName(), StringUtils.isNotBlank(proxy) ? proxy : user, clusterInfo.getLinkisAddress(), clusterName, taskId, csId, nodeName, StringUtils.isNotBlank(startupParam) ? startupParam : job.getStartupParam());
            }
            if (result != null) {
                submitResults.add(result);
            } else {
                Task taskInDb = taskDao.findById(taskId);
                taskInDb.setStatus(TaskStatusEnum.TASK_NOT_EXIST.getCode());
                taskDao.save(taskInDb);
                taskSubmitResults.add(new TaskSubmitResult(application.getId(), null, clusterInfo.getClusterName()));
            }
        }
        // Rewrite task remote ID.
        rewriteTaskRemoteInfo(submitResults, taskSubmitResults, application.getId(), clusterInfo.getClusterName());
    }
    return taskSubmitResults;
}

Example 2 with DataQualityTaskException

use of com.webank.wedatasphere.qualitis.exception.DataQualityTaskException in project Qualitis by WeBankFinTech.

the class SqlTemplateConverter method convert.

/**
 * Convert task into scala code
 * @param dataQualityTask
 * @param date
 * @param setFlag
 * @param execParams
 * @param runDate
 * @param clusterType
 * @param dataSourceMysqlConnect
 * @return
 * @throws ConvertException
 * @throws DataQualityTaskException
 * @throws RuleVariableNotSupportException
 * @throws RuleVariableNotFoundException
 */
@Override
public DataQualityJob convert(DataQualityTask dataQualityTask, Date date, String setFlag, Map<String, String> execParams, String runDate, String clusterType, Map<Long, Map> dataSourceMysqlConnect) throws ConvertException, DataQualityTaskException, RuleVariableNotSupportException, RuleVariableNotFoundException, IOException, UnExpectedRequestException {
    LOGGER.info("Start to convert template to actual code, task: " + dataQualityTask);
    if (null == dataQualityTask || dataQualityTask.getRuleTaskDetails().isEmpty()) {
        throw new DataQualityTaskException("Task can not be null or empty");
    }
    DataQualityJob job = new DataQualityJob();
    List<String> initSentence = abstractTranslator.getInitSentence();
    job.getJobCode().addAll(initSentence);
    LOGGER.info("Succeed to get init code. codes: " + initSentence);
    if (StringUtils.isNotBlank(setFlag)) {
        LOGGER.info("Start to solve with set flag. Spark set conf string: {}", setFlag);
        String[] setStrs = setFlag.split(SpecCharEnum.DIVIDER.getValue());
        for (String str : setStrs) {
            job.getJobCode().add("spark.sql(\"set " + str + "\")");
        }
        LOGGER.info("Finish to solve with set flag.");
    }
    int count = 0;
    for (RuleTaskDetail ruleTaskDetail : dataQualityTask.getRuleTaskDetails()) {
        count++;
        List<String> codes = generateSparkSqlByTask(ruleTaskDetail.getRule(), date, dataQualityTask.getApplicationId(), ruleTaskDetail.getMidTableName(), dataQualityTask.getCreateTime(), new StringBuffer(dataQualityTask.getPartition()), execParams, count, runDate, dataSourceMysqlConnect);
        job.getJobCode().addAll(codes);
        LOGGER.info("Succeed to convert rule into code. rule_id: {}, rul_name: {}, codes: {}", ruleTaskDetail.getRule().getId(), ruleTaskDetail.getRule().getName(), codes);
    }
    LOGGER.info("Succeed to convert all rule into actual scala code.");
    job.setTaskId(dataQualityTask.getTaskId());
    job.setStartupParam(dataQualityTask.getStartupParam());
    return job;
}