Commit a3940966 authored by pengxin's avatar pengxin

清洗过程去掉写人json文件中以及更新配置。

parent 171bbcd5
......@@ -92,6 +92,11 @@ public class DatasetConstant {
*/
public static final Integer MARK = 1;
/**
* 未清洗
*/
public static final Integer NOT_CLEAN_PROGRESS = 3;
/**
* 清洗中
*/
......
......@@ -13,6 +13,7 @@ import com.yice.common.log.model.constant.SysOperationLogType;
import com.yice.webadmin.app.dto.DatasetCleanDto;
import com.yice.webadmin.app.model.DatasetClean;
import com.yice.webadmin.app.service.DatasetCleanService;
import com.yice.webadmin.app.service.DatasetDataService;
import com.yice.webadmin.app.vo.DatasetCleanVo;
import io.swagger.annotations.Api;
import lombok.extern.slf4j.Slf4j;
......@@ -35,6 +36,8 @@ public class DatasetCleanController {
@Autowired
private DatasetCleanService datasetCleanService;
@Autowired
private DatasetDataService datasetDataService;
/**
* 新增数据集清洗数据。
......@@ -71,6 +74,8 @@ public class DatasetCleanController {
}
DatasetClean datasetClean = MyModelUtil.copyTo(datasetCleanDto, DatasetClean.class);
datasetClean = datasetCleanService.addNew(datasetClean);
datasetDataService.doDatasetCleanHandler(datasetClean.getDatasetId(), datasetClean.getCleanId(),
datasetClean.getCleanConfigId());
return ResponseResult.success(datasetClean);
}
......
......@@ -85,6 +85,12 @@ public class DatasetClean extends BaseModel {
@TableField(exist = false)
private List<String> cleanMethod;
/**
* 清洗配置标识。
*/
@TableField(exist = false)
private Long cleanConfigId;
/**
* 清洗配置对象。
*/
......
......@@ -53,9 +53,10 @@ public interface DatasetCleanService extends IBaseService<DatasetClean, Long> {
* @param dataList 数据集列表
* @param cleanId 清洗标识
* @param datasetId 清洗集标识
* @param cleanConfigId 清洗配置标识
* @return 线程方法
*/
Future<Long> executeCleanTaskAsync(List<DatasetData> dataList, Long cleanId, Long datasetId);
Future<Long> executeCleanTaskAsync(List<DatasetData> dataList, Long cleanId, Long datasetId, Long cleanConfigId);
/**
* 保存清洗对象以及清洗配置对象。
......
......@@ -25,8 +25,9 @@ public interface DatasetDataService {
* 开始清洗工作
* @param datasetId 清洗数据集
* @param cleanId 清洗标识
* @param cleanConfigId 清洗配置标识
*/
void doDatasetCleanHandler(Long datasetId, Long cleanId);
void doDatasetCleanHandler(Long datasetId, Long cleanId, Long cleanConfigId);
/**
* 删除整个集合中的文档数据。
......
......@@ -22,8 +22,6 @@ import org.springframework.stereotype.Service;
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.Future;
/**
* 数据集版本数据操作服务类。
......@@ -89,16 +87,17 @@ public class DatasetDataServiceImpl implements DatasetDataService {
* 开始清洗工作
* @param datasetId 数据集对应的版本
* @param cleanId 清洗任务标识
* @param cleanConfigId 清洗配置标识
*/
@Async
@Override
public void doDatasetCleanHandler(Long datasetId, Long cleanId) {
public void doDatasetCleanHandler(Long datasetId, Long cleanId, Long cleanConfigId) {
MyPageParam param = new MyPageParam();
param.setPageNum(DatasetConstant.PAGE_NUM);
param.setPageSize(DatasetConstant.MAX_PAGE_SIZE);
List<DatasetData> dataList = this.list(datasetId, param);
if(CollUtil.isNotEmpty(dataList)) {
datasetCleanService.executeCleanTaskAsync(dataList, cleanId, datasetId);
datasetCleanService.executeCleanTaskAsync(dataList, cleanId, datasetId, cleanConfigId);
}
}
......
......@@ -100,7 +100,7 @@ public class DatasetVersionServiceImpl extends BaseService<DatasetVersion, Long>
datasetVersion.setDatasetVersion(version);
datasetVersion.setVersionName(reDatasetManage.getDatasetName() + "_V" + version);
datasetVersion.setDatasetId(reDatasetManage.getDatasetId());
datasetVersion.setCleanStatus(0);
datasetVersion.setCleanStatus(DatasetConstant.NOT_CLEAN_PROGRESS);
datasetVersion.setDataVolume(0L);
datasetVersion.setEnhanceStatus(0);
datasetVersion.setInputStatus(0);
......@@ -326,7 +326,7 @@ public class DatasetVersionServiceImpl extends BaseService<DatasetVersion, Long>
String versionName = datasetVersion.getVersionName();
//先存储文件
String fullName = pythonConfig.getDatasetFileBaseDir() + versionName + ".json";
this.doDealTaskHandler(datasetVersion.getVersionId(), versionName, fullName);
this.doDealTaskHandler(datasetVersion.getVersionId(), fullName);
//再存储数据集配置文件
this.saveDatasetInfo(versionName);
DatasetVersion filter = new DatasetVersion();
......@@ -346,10 +346,9 @@ public class DatasetVersionServiceImpl extends BaseService<DatasetVersion, Long>
* ==============================
* 处理数据集
* @param datasetId 清洗任务id
* @param versionName 数据集名称
* @return 清洗列表
*/
private void doDealTaskHandler(Long datasetId, String versionName,String fileUrl) {
private void doDealTaskHandler(Long datasetId, String fileUrl) {
try {
Integer index = 0;
Long count = datasetDataService.count(datasetId);
......@@ -365,7 +364,8 @@ public class DatasetVersionServiceImpl extends BaseService<DatasetVersion, Long>
//写入到数据集中
if(CollUtil.isNotEmpty(dataList)) {
datasetCleanService.appendDataListToFile(fileUrl ,dataList, i);
datasetCleanService.appendDataListToFile(fileUrl, dataList, i);
index ++;
}
}
......@@ -378,7 +378,7 @@ public class DatasetVersionServiceImpl extends BaseService<DatasetVersion, Long>
this.datasetDataService.deleteByData(datasetId);
}
} catch (Exception ex) {
log.error("deal with task handler is error:" , ex);
log.error("do deal with task handler is error:" , ex);
}
}
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment