feat: 优化文件删除逻辑，支持级联删除版本和派生文件

- 删除文件时，如果存在多个版本，一并删除所有版本

- 删除PDF/doc/docx/xls/xlsx时，一并删除其派生的txt文件

- 文件删除失败时记录日志但不影响删除成功

backend/services/data-management-service/src/main/java/com/datamate/datamanagement/application/DatasetFileApplicationService.java

@@ -581,25 +581,90 @@ public class DatasetFileApplicationService {
     @Transactional
     public void deleteDatasetFile(String datasetId, String fileId) {
         DatasetFile file = getDatasetFile(datasetId, fileId);
+        if (file == null) {
+            log.warn("File not found: datasetId={}, fileId={}", datasetId, fileId);
+            return;
+        }
+        String logicalPath = file.getLogicalPath();
+        
+        List<DatasetFile> allVersions = datasetFileRepository.findAllByDatasetIdAndLogicalPath(datasetId, logicalPath);
+        
+        for (DatasetFile versionFile : allVersions) {
+            deleteDatasetFileInternal(datasetId, versionFile);
+        }
+    }
+
+    private void deleteDatasetFileInternal(String datasetId, DatasetFile file) {
         Dataset dataset = datasetRepository.getById(datasetId);
-        datasetFileRepository.removeById(fileId);
+        if (file == null || dataset == null) {
+            return;
+        }
+
+        if (isSourceDocument(file)) {
+            deleteDerivedTextFileQuietly(datasetId, file.getId());
+        }
+
+        try {
+            datasetFileRepository.removeById(file.getId());
+        } catch (Exception e) {
+            log.error("Failed to delete file record from database: fileId={}", file.getId(), e);
+        }
+
         if (!isArchivedStatus(file)) {
+            try {
                 dataset.setFiles(new ArrayList<>(Collections.singleton(file)));
                 dataset.removeFile(file);
                 datasetRepository.updateById(dataset);
+            } catch (Exception e) {
+                log.error("Failed to update dataset: datasetId={}", datasetId, e);
             }
-        datasetFilePreviewService.deletePreviewFileQuietly(datasetId, fileId);
+        }
-        // 删除文件时，上传到数据集中的文件会同时删除数据库中的记录和文件系统中的文件，归集过来的文件仅删除数据库中的记录
+
-        if (file.getFilePath().startsWith(dataset.getPath())) {
+        datasetFilePreviewService.deletePreviewFileQuietly(datasetId, file.getId());
+
+        if (file.getFilePath() != null && file.getFilePath().startsWith(dataset.getPath())) {
             try {
                 Path filePath = Paths.get(file.getFilePath());
                 Files.deleteIfExists(filePath);
             } catch (IOException ex) {
-                throw BusinessException.of(SystemErrorCode.FILE_SYSTEM_ERROR);
+                log.error("Failed to delete physical file: filePath={}", file.getFilePath(), ex);
             }
         }
     }
 
+    private void deleteDerivedTextFileQuietly(String datasetId, String sourceFileId) {
+        if (sourceFileId == null || sourceFileId.isBlank()) {
+            return;
+        }
+
+        try {
+            List<DatasetFile> derivedFiles = datasetFileRepository.findAllByDatasetId(datasetId).stream()
+                .filter(f -> isDerivedFileFromSource(f, sourceFileId))
+                .toList();
+
+            for (DatasetFile derivedFile : derivedFiles) {
+                deleteDatasetFileInternal(datasetId, derivedFile);
+            }
+        } catch (Exception e) {
+            log.error("Failed to delete derived text files for sourceFileId: {}", sourceFileId, e);
+        }
+    }
+
+    private boolean isDerivedFileFromSource(DatasetFile file, String sourceFileId) {
+        if (file == null || file.getMetadata() == null || file.getMetadata().isBlank()) {
+            return false;
+        }
+        try {
+            ObjectMapper mapper = new ObjectMapper();
+            Map<String, Object> metadataMap = mapper.readValue(file.getMetadata(), new TypeReference<Map<String, Object>>() {});
+            Object derivedFromFileId = metadataMap.get(DERIVED_METADATA_KEY);
+            return derivedFromFileId != null && sourceFileId.equals(String.valueOf(derivedFromFileId));
+        } catch (Exception e) {
+            log.debug("Failed to parse metadata for derived detection: fileId={}", file.getId(), e);
+            return false;
+        }
+    }
+
     /**
      * 下载文件
      */

backend/services/data-management-service/src/main/java/com/datamate/datamanagement/infrastructure/persistence/mapper/DatasetFileMapper.java

@@ -46,4 +46,13 @@ public interface DatasetFileMapper extends BaseMapper<DatasetFile> {
      * @return 文件数统计列表
      */
     List<DatasetFileCount> countNonDerivedByDatasetIds(@Param("datasetIds") List<String> datasetIds);
+
+    /**
+     * 查询指定逻辑路径的所有文件（包括所有状态）
+     *
+     * @param datasetId 数据集ID
+     * @param logicalPath 逻辑路径
+     * @return 文件列表
+     */
+    List<DatasetFile> findAllByDatasetIdAndLogicalPath(@Param("datasetId") String datasetId, @Param("logicalPath") String logicalPath);
 }

backend/services/data-management-service/src/main/java/com/datamate/datamanagement/infrastructure/persistence/repository/DatasetFileRepository.java

@@ -37,6 +37,15 @@ public interface DatasetFileRepository extends IRepository<DatasetFile> {
      */
     DatasetFile findLatestByDatasetIdAndLogicalPath(String datasetId, String logicalPath);
 
+    /**
+     * 查询指定逻辑路径的所有文件（包括所有状态）
+     *
+     * @param datasetId 数据集ID
+     * @param logicalPath 逻辑路径
+     * @return 文件列表
+     */
+    List<DatasetFile> findAllByDatasetIdAndLogicalPath(String datasetId, String logicalPath);
+
     IPage<DatasetFile> findByCriteria(String datasetId, String fileType, String status, String name,
                                       Boolean hasAnnotation, IPage<DatasetFile> page);
 

backend/services/data-management-service/src/main/java/com/datamate/datamanagement/infrastructure/persistence/repository/impl/DatasetFileRepositoryImpl.java

@@ -84,6 +84,11 @@ public class DatasetFileRepositoryImpl extends CrudRepository<DatasetFileMapper,
             .last("LIMIT 1"));
     }
 
+    @Override
+    public List<DatasetFile> findAllByDatasetIdAndLogicalPath(String datasetId, String logicalPath) {
+        return datasetFileMapper.findAllByDatasetIdAndLogicalPath(datasetId, logicalPath);
+    }
+
     public IPage<DatasetFile> findByCriteria(String datasetId, String fileType, String status, String name,
                                              Boolean hasAnnotation, IPage<DatasetFile> page) {
         LambdaQueryWrapper<DatasetFile> wrapper = new LambdaQueryWrapper<DatasetFile>()

backend/services/data-management-service/src/main/resources/mappers/DatasetFileMapper.xml

@@ -74,6 +74,14 @@
         LIMIT 1
     </select>
 
+   <select id="findAllByDatasetIdAndLogicalPath" resultType="com.datamate.datamanagement.domain.model.dataset.DatasetFile">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_dataset_files
+        WHERE dataset_id = #{datasetId}
+          AND logical_path = #{logicalPath}
+        ORDER BY version DESC, upload_time DESC
+    </select>
+
    <select id="findAllByDatasetId" parameterType="string"
            resultType="com.datamate.datamanagement.domain.model.dataset.DatasetFile">
         SELECT <include refid="Base_Column_List"/>