bugfix: 创建清洗任务时修改数据集状态；无法删除已在模板/运行任务的算子

hhhhsc701 · web-flow · commit f1bffdcd6186 · 2025-11-27T17:34:53.000+08:00
* bugfix: 创建清洗任务时修改数据集状态；无法删除已在模板/运行任务的算子
diff --git a/Makefile b/Makefile
@@ -244,11 +244,9 @@ VALID_SERVICE_TARGETS := datamate backend frontend runtime mineru "deer-flow" mi
 		cd deployment/docker/datamate && export REGISTRY=$(REGISTRY) && docker compose up -d datamate-mineru; \
 	elif [ "$*" = "datamate" ]; then \
 		if docker compose ls --filter name=deer-flow | grep -q deer-flow; then \
-			(cd deployment/docker/datamate && NGINX_CONF="./backend-with-deer-flow.conf" REGISTRY=$(REGISTRY) docker compose -f docker-compose.yml up -d) && \
-			$(MAKE) label-studio-docker-install; \
+			(cd deployment/docker/datamate && NGINX_CONF="./backend-with-deer-flow.conf" REGISTRY=$(REGISTRY) docker compose -f docker-compose.yml up -d); \
 		else \
-			(cd deployment/docker/datamate && REGISTRY=$(REGISTRY) docker compose -f docker-compose.yml up -d) && \
-			$(MAKE) label-studio-docker-install; \
+			(cd deployment/docker/datamate && REGISTRY=$(REGISTRY) docker compose -f docker-compose.yml up -d); \
 		fi; \
 	elif [ "$*" = "deer-flow" ]; then \
 		cd deployment/docker/datamate && export NGINX_CONF="./backend-with-deer-flow.conf" && export REGISTRY=$(REGISTRY) && docker compose -f docker-compose.yml up -d; \
diff --git a/backend/services/data-cleaning-service/src/main/java/com/datamate/cleaning/application/CleaningTaskService.java b/backend/services/data-cleaning-service/src/main/java/com/datamate/cleaning/application/CleaningTaskService.java
@@ -93,6 +93,7 @@ public CleaningTaskDto createTask(CreateCleaningTaskRequest request) {
         CreateDatasetRequest createDatasetRequest = new CreateDatasetRequest();
         createDatasetRequest.setName(request.getDestDatasetName());
         createDatasetRequest.setDatasetType(DatasetType.valueOf(request.getDestDatasetType()));
+        createDatasetRequest.setStatus("ACTIVE");
         Dataset destDataset = datasetService.createDataset(createDatasetRequest);
 
         Dataset srcDataset = datasetService.getDataset(request.getSrcDatasetId());
diff --git a/backend/services/data-management-service/src/main/java/com/datamate/datamanagement/domain/model/dataset/Dataset.java b/backend/services/data-management-service/src/main/java/com/datamate/datamanagement/domain/model/dataset/Dataset.java
@@ -113,7 +113,9 @@ public Dataset(String name, String description, DatasetType datasetType, String
     public void initCreateParam(String datasetBasePath) {
         this.id = UUID.randomUUID().toString();
         this.path = datasetBasePath + File.separator + this.id;
-        this.status = DatasetStatusType.DRAFT;
+        if (this.status == null) {
+            this.status = DatasetStatusType.DRAFT;
+        }
     }
 
     public void updateBasicInfo(String name, String description, String category) {
diff --git a/backend/services/data-management-service/src/main/java/com/datamate/datamanagement/interfaces/dto/CreateDatasetRequest.java b/backend/services/data-management-service/src/main/java/com/datamate/datamanagement/interfaces/dto/CreateDatasetRequest.java
@@ -35,4 +35,6 @@ public class CreateDatasetRequest {
     private String dataSource;
     /** 保留天数 */
     private Integer retentionDays;
+    /** 数据集状态 */
+    private String status;
 }
diff --git a/backend/services/operator-market-service/src/main/java/com/datamate/operator/application/OperatorService.java b/backend/services/operator-market-service/src/main/java/com/datamate/operator/application/OperatorService.java
@@ -88,6 +88,9 @@ public OperatorDto updateOperator(String id, OperatorDto req) {
 
     @Transactional
     public void deleteOperator(String id) {
+        if (operatorRepo.operatorInTemplateOrRunning(id)) {
+            throw BusinessException.of(OperatorErrorCode.OPERATOR_IN_INSTANCE);
+        }
         operatorRepo.deleteOperator(id);
         relationRepo.deleteByOperatorId(id);
     }
diff --git a/backend/services/operator-market-service/src/main/java/com/datamate/operator/domain/repository/OperatorRepository.java b/backend/services/operator-market-service/src/main/java/com/datamate/operator/domain/repository/OperatorRepository.java
@@ -16,4 +16,6 @@ public interface OperatorRepository extends IRepository<Operator> {
     void deleteOperator(String id);
 
     int countOperatorByStar(boolean isStar);
+
+    boolean operatorInTemplateOrRunning(String operatorId);
 }
diff --git a/backend/services/operator-market-service/src/main/java/com/datamate/operator/infrastructure/exception/OperatorErrorCode.java b/backend/services/operator-market-service/src/main/java/com/datamate/operator/infrastructure/exception/OperatorErrorCode.java
@@ -16,7 +16,9 @@ public enum OperatorErrorCode implements ErrorCode {
 
     FIELD_NOT_FOUND("op.0003", "缺少必要的字段"),
 
-    SETTINGS_PARSE_FAILED("op.0004", "settings字段解析失败");
+    SETTINGS_PARSE_FAILED("op.0004", "settings字段解析失败"),
+
+    OPERATOR_IN_INSTANCE("op.0005", "算子已被编排在模板或未完成的任务中");
 
     private final String code;
     private final String message;
diff --git a/backend/services/operator-market-service/src/main/java/com/datamate/operator/infrastructure/persistence/Impl/OperatorRepositoryImpl.java b/backend/services/operator-market-service/src/main/java/com/datamate/operator/infrastructure/persistence/Impl/OperatorRepositoryImpl.java
@@ -43,4 +43,9 @@ public int countOperatorByStar(boolean isStar) {
         queryWrapper.eq(Operator::getIsStar, isStar);
         return Math.toIntExact(mapper.selectCount(queryWrapper));
     }
+
+    @Override
+    public boolean operatorInTemplateOrRunning(String operatorId) {
+        return mapper.operatorInTemplate(operatorId) > 0 && mapper.operatorInUnstopTask(operatorId) > 0;
+    }
 }
diff --git a/backend/services/operator-market-service/src/main/java/com/datamate/operator/infrastructure/persistence/mapper/OperatorMapper.java b/backend/services/operator-market-service/src/main/java/com/datamate/operator/infrastructure/persistence/mapper/OperatorMapper.java
@@ -3,7 +3,16 @@
 import com.baomidou.mybatisplus.core.mapper.BaseMapper;
 import com.datamate.operator.domain.model.Operator;
 import org.apache.ibatis.annotations.Mapper;
+import org.apache.ibatis.annotations.Select;
 
 @Mapper
 public interface OperatorMapper extends BaseMapper<Operator> {
+
+    @Select("SELECT count(1) FROM t_operator_instance oi JOIN t_clean_template t ON oi.instance_id = t.id " +
+            "WHERE oi.operator_id = #{operatorId}")
+    int operatorInTemplate(String operatorId);
+
+    @Select("SELECT count(1) FROM t_operator_instance oi JOIN t_clean_task t ON oi.instance_id = t.id " +
+            "WHERE oi.operator_id = #{operatorId} AND t.status != 'COMPLETED'")
+    int operatorInUnstopTask(String operatorId);
 }
diff --git a/frontend/src/pages/DataCleansing/Create/components/OperatorLibrary.tsx b/frontend/src/pages/DataCleansing/Create/components/OperatorLibrary.tsx
@@ -58,7 +58,10 @@ const OperatorList: React.FC<OperatorListProps> = ({
             </div>
             <span
               className="cursor-pointer"
-              onClick={() => handleStar(operator, toggleFavorite)}
+              onClick={(event) => {
+                  event.stopPropagation();
+                  handleStar(operator, toggleFavorite);
+              }}
             >
               {favorites.has(operator.id) ? (
                 <StarFilled style={{ color: "#FFD700" }} />
diff --git a/frontend/src/pages/DataCleansing/Detail/components/FileTable.tsx b/frontend/src/pages/DataCleansing/Detail/components/FileTable.tsx
@@ -1,4 +1,4 @@
-import {Button, Modal, Table, Badge, Input} from "antd";
+import {Button, Modal, Table, Badge, Input, Popover} from "antd";
 import { Download } from "lucide-react";
 import {useEffect, useState} from "react";
 import {useParams} from "react-router";
@@ -259,9 +259,9 @@ export default function FileTable({result, fetchTaskResult}) {
               对比
             </Button>
           )}
-          <Button type="link" size="small">
-            下载
-          </Button>
+          <Popover content="暂未开放">
+              <Button type="link" size="small" disabled>下载</Button>
+          </Popover>
         </div>
       ),
     },
diff --git a/frontend/src/pages/DataCleansing/Home/components/TaskList.tsx b/frontend/src/pages/DataCleansing/Home/components/TaskList.tsx
@@ -186,23 +186,23 @@ export default function TaskList() {
       title: "已处理文件数",
       dataIndex: "finishedFileNum",
       key: "finishedFileNum",
-      width: 150,
+      width: 120,
       align: "right",
       ellipsis: true,
     },
     {
       title: "总文件数",
       dataIndex: "totalFileNum",
       key: "totalFileNum",
-      width: 150,
+      width: 100,
       align: "right",
       ellipsis: true,
     },
     {
       title: "执行耗时",
       dataIndex: "duration",
       key: "duration",
-      width: 180,
+      width: 100,
       ellipsis: true,
     },
     {
diff --git a/runtime/ops/formatter/mineru_formatter/__init__.py b/runtime/ops/formatter/mineru_formatter/__init__.py
@@ -3,4 +3,4 @@
 from datamate.core.base_op import OPERATORS
 
 OPERATORS.register_module(module_name='MineruFormatter',
-                          module_path="ops.formatter.external_pdf_formatter.process")
+                          module_path="ops.formatter.mineru_formatter.process")
diff --git a/runtime/ops/formatter/mineru_formatter/process.py b/runtime/ops/formatter/mineru_formatter/process.py
@@ -25,6 +25,8 @@ def __init__(self, *args, **kwargs):
     def execute(self, sample: Dict[str, Any]) -> Dict[str, Any]:
         start = time.time()
         filename = sample[self.filename_key]
+        if not filename.lower().endswith(".pdf"):
+            return sample
         try:
             data = {"source_path": sample[self.filepath_key], "export_path": sample[self.export_path_key]}
             response = http_request(method="POST", url=self.pdf_extract_url, data=data)
diff --git a/runtime/ops/formatter/unstructured_formatter/process.py b/runtime/ops/formatter/unstructured_formatter/process.py
@@ -25,6 +25,8 @@ def execute(self, sample: Dict[str, Any]) -> Dict[str, Any]:
         start = time.time()
         filepath = sample.get(self.filepath_key)
         filename = sample.get(self.filename_key)
+        if not filename.lower().endswith((".ppt", ".pptx", "docx", "xlsx", ".csv")):
+            return sample
         try:
             elements = partition(filename=filepath)
             sample[self.text_key] = "\n\n".join([str(el) for el in elements])

Original file line number	Diff line number	Diff line change
`@@ -35,4 +35,6 @@ public class CreateDatasetRequest {`
`35`	`35`	`private String dataSource;`
`36`	`36`	`/** 保留天数 */`
`37`	`37`	`private Integer retentionDays;`
	`38`	`+ /** 数据集状态 */`
	`39`	`+ private String status;`
`38`	`40`	`}`
Original file line number	Diff line number	Diff line change
`@@ -88,6 +88,9 @@ public OperatorDto updateOperator(String id, OperatorDto req) {`
`88`	`88`
`89`	`89`	`@Transactional`
`90`	`90`	`public void deleteOperator(String id) {`
	`91`	`+ if (operatorRepo.operatorInTemplateOrRunning(id)) {`
	`92`	`+ throw BusinessException.of(OperatorErrorCode.OPERATOR_IN_INSTANCE);`
	`93`	`+ }`
`91`	`94`	`operatorRepo.deleteOperator(id);`
`92`	`95`	`relationRepo.deleteByOperatorId(id);`
`93`	`96`	`}`
Original file line number	Diff line number	Diff line change
`@@ -16,4 +16,6 @@ public interface OperatorRepository extends IRepository<Operator> {`
`16`	`16`	`void deleteOperator(String id);`
`17`	`17`
`18`	`18`	`int countOperatorByStar(boolean isStar);`
	`19`	`+`
	`20`	`+ boolean operatorInTemplateOrRunning(String operatorId);`
`19`	`21`	`}`
Original file line number	Diff line number	Diff line change
`@@ -43,4 +43,9 @@ public int countOperatorByStar(boolean isStar) {`
`43`	`43`	`queryWrapper.eq(Operator::getIsStar, isStar);`
`44`	`44`	`return Math.toIntExact(mapper.selectCount(queryWrapper));`
`45`	`45`	`}`
	`46`	`+`
	`47`	`+ @Override`
	`48`	`+ public boolean operatorInTemplateOrRunning(String operatorId) {`
	`49`	`+ return mapper.operatorInTemplate(operatorId) > 0 && mapper.operatorInUnstopTask(operatorId) > 0;`
	`50`	`+ }`
`46`	`51`	`}`