feat: 完善数据标注导出格式兼容性验证

- 后端：添加 YOLO 格式对 TEXT 数据集的限制验证 - 后端：统一 COCO/YOLO 兼容性校验规则（仅允许图像类或目标检测类数据集） - 后端：修复 datasetType 字段传递，在任务列表响应中补充 dataset_type - 前端：在导出对话框中禁用 TEXT 数据集的 COCO/YOLO 选项 - 前端：添加 datasetType 和 labelingType 字段传递 - 前端：对齐前后端 COCO/YOLO 兼容性规则 - 前端：优化提示文案，明确说明格式适用范围修改文件： - runtime/datamate-python/app/module/annotation/service/export.py - runtime/datamate-python/app/module/annotation/service/mapping.py - runtime/datamate-python/app/module/annotation/schema/mapping.py - frontend/src/pages/DataAnnotation/Home/ExportAnnotationDialog.tsx - frontend/src/pages/DataAnnotation/Home/DataAnnotation.tsx - frontend/src/pages/DataAnnotation/annotation.const.tsx
feat(annotation): 添加导出格式与数据集类型的兼容性检查
2026-02-07 16:05:57 +08:00 · 2026-02-07 16:05:57 +08:00
6 changed files with 162 additions and 21 deletions
--- a/frontend/src/pages/DataAnnotation/Home/DataAnnotation.tsx
+++ b/frontend/src/pages/DataAnnotation/Home/DataAnnotation.tsx
@@ -378,6 +378,8 @@ export default function DataAnnotation() {
                  open={!!exportTask}
                  projectId={exportTask?.id || ""}
                  projectName={exportTask?.name || ""}
+                  datasetType={exportTask?.datasetType}
+                  labelingType={exportTask?.labelingType}
                  onClose={() => setExportTask(null)}
                />
              </div>
--- a/frontend/src/pages/DataAnnotation/Home/ExportAnnotationDialog.tsx
+++ b/frontend/src/pages/DataAnnotation/Home/ExportAnnotationDialog.tsx
@@ -11,7 +11,7 @@ import {
  message,
  Alert,
 } from "antd";
-import { FileTextOutlined, CheckCircleOutlined } from "@ant-design/icons";
+import { FileTextOutlined, CheckCircleOutlined, InfoCircleOutlined } from "@ant-design/icons";
 import {
  getExportStatsUsingGet,
  downloadAnnotationsUsingGet,
@@ -22,6 +22,8 @@ interface ExportAnnotationDialogProps {
  open: boolean;
  projectId: string;
  projectName: string;
+  datasetType?: string;
+  labelingType?: string;
  onClose: () => void;
 }

@@ -57,6 +59,8 @@ export default function ExportAnnotationDialog({
  open,
  projectId,
  projectName,
+  datasetType,
+  labelingType,
  onClose,
 }: ExportAnnotationDialogProps) {
  const [form] = Form.useForm();
@@ -67,6 +71,15 @@ export default function ExportAnnotationDialog({
    annotatedFiles: number;
  } | null>(null);

+  const normalizedDatasetType = datasetType?.toUpperCase().replace(/-/g, "_") || "";
+  const normalizedLabelingType = labelingType?.toUpperCase().replace(/-/g, "_") || "";
+  const hasDatasetType = Boolean(normalizedDatasetType);
+  const isTextDataset = normalizedDatasetType === "TEXT";
+  const isDetectionCompatible =
+    normalizedDatasetType === "IMAGE" ||
+    normalizedDatasetType === "OBJECT_DETECTION" ||
+    normalizedLabelingType === "OBJECT_DETECTION";
+
  // 加载导出统计信息
  useEffect(() => {
    if (open && projectId) {
@@ -176,20 +189,46 @@ export default function ExportAnnotationDialog({
            rules={[{ required: true, message: "请选择导出格式" }]}
          >
            <Select
-              options={FORMAT_OPTIONS.map((opt) => ({
-                label: (
-                  <div className="py-1">
-                    <div className="font-medium">{opt.label}</div>
-                    <div className="text-xs text-gray-400">{opt.description}</div>
-                  </div>
-                ),
-                value: opt.value,
-                simpleLabel: opt.label,
-              }))}
+              options={FORMAT_OPTIONS.map((opt) => {
+                const isDetectionFormat = opt.value === "coco" || opt.value === "yolo";
+                const isDisabledForText = isDetectionFormat && isTextDataset;
+                const isDisabledForIncompatibleType =
+                  isDetectionFormat && hasDatasetType && !isTextDataset && !isDetectionCompatible;
+                const isDisabled = isDisabledForText || isDisabledForIncompatibleType;
+                return {
+                  label: (
+                    <div className="py-1 relative">
+                      <div className={`font-medium ${isDisabled ? "text-gray-400" : ""}`}>{opt.label}</div>
+                      <div className={`text-xs ${isDisabled ? "text-gray-300" : "text-gray-400"}`}>
+                        {opt.description}
+                        {isDisabled && (
+                          <span className="block mt-1 text-orange-500">
+                            <InfoCircleOutlined className="mr-1" />
+                            仅适用于图像类或目标检测类数据集
+                          </span>
+                        )}
+                      </div>
+                    </div>
+                  ),
+                  value: opt.value,
+                  simpleLabel: opt.label,
+                  disabled: isDisabled,
+                };
+              })}
              optionLabelProp="simpleLabel"
            />
          </Form.Item>

+          {(isTextDataset || (hasDatasetType && !isDetectionCompatible)) && (
+            <Alert
+              type="info"
+              message="导出格式兼容性提示"
+              description="COCO 和 YOLO 格式仅适用于图像类或目标检测类数据集，当前项目建议使用 JSON、JSON Lines 或 CSV 格式导出。"
+              showIcon
+              className="mt-4"
+            />
+          )}
+
          <Form.Item name="onlyAnnotated" valuePropName="checked">
            <Checkbox>仅导出已标注数据</Checkbox>
          </Form.Item>
--- a/frontend/src/pages/DataAnnotation/annotation.const.tsx
+++ b/frontend/src/pages/DataAnnotation/annotation.const.tsx
@@ -23,6 +23,8 @@ type AnnotationTaskPayload = {
  datasetId?: string;
  datasetName?: string;
  dataset_name?: string;
+  datasetType?: string;
+  dataset_type?: string;
  labelingType?: string;
  labeling_type?: string;
  template?: {
@@ -54,6 +56,7 @@ export type AnnotationTaskListItem = {
  description?: string;
  datasetId?: string;
  datasetName?: string;
+  datasetType?: string;
  labelingType?: string;
  totalCount?: number;
  annotatedCount?: number;
@@ -97,6 +100,7 @@ export function mapAnnotationTask(task: AnnotationTaskPayload): AnnotationTaskLi
  const labelingProjId = task?.labelingProjId || task?.labelingProjectId || task?.projId || task?.labeling_project_id || "";
  const segmentationEnabled = task?.segmentationEnabled ?? task?.segmentation_enabled ?? false;
  const inProgressCount = task?.inProgressCount ?? task?.in_progress_count ?? 0;
+  const datasetType = task?.datasetType || task?.dataset_type;
  const labelingType =
    task?.labelingType ||
    task?.labeling_type ||
@@ -119,6 +123,7 @@ export function mapAnnotationTask(task: AnnotationTaskPayload): AnnotationTaskLi
    projId: labelingProjId,
    segmentationEnabled,
    inProgressCount,
+    datasetType,
    labelingType,
    name: task.name,
    description: task.description || "",
--- a/runtime/datamate-python/app/module/annotation/schema/mapping.py
+++ b/runtime/datamate-python/app/module/annotation/schema/mapping.py
@@ -61,6 +61,7 @@ class DatasetMappingResponse(BaseModel):
    id: str = Field(..., description="映射UUID")
    dataset_id: str = Field(..., alias="datasetId", description="源数据集ID")
    dataset_name: Optional[str] = Field(None, alias="datasetName", description="数据集名称")
+    dataset_type: Optional[str] = Field(None, alias="datasetType", description="数据集类型")
    labeling_project_id: str = Field(..., alias="labelingProjectId", description="标注项目ID")
    name: Optional[str] = Field(None, description="标注项目名称")
    description: Optional[str] = Field(None, description="标注项目描述")
--- a/runtime/datamate-python/app/module/annotation/service/export.py
+++ b/runtime/datamate-python/app/module/annotation/service/export.py
@@ -28,6 +28,7 @@ from sqlalchemy.ext.asyncio import AsyncSession
 from app.core.logging import get_logger
 from app.db.models import (
    AnnotationResult,
+    AnnotationTemplate,
    Dataset,
    DatasetFiles,
    LabelingProject,
@@ -78,6 +79,15 @@ SEGMENTS_KEY = "segments"
 SEGMENT_RESULT_KEY = "result"
 SEGMENT_INDEX_KEY = "segmentIndex"
 SEGMENT_INDEX_FALLBACK_KEY = "segment_index"
+DATASET_TYPE_TEXT = "TEXT"
+DATASET_TYPE_IMAGE = "IMAGE"
+DATASET_TYPE_OBJECT_DETECTION = "OBJECT_DETECTION"
+LABELING_TYPE_CONFIG_KEY = "labeling_type"
+LABELING_TYPE_OBJECT_DETECTION = "OBJECT_DETECTION"
+DETECTION_COMPATIBLE_DATASET_TYPES = {
+    DATASET_TYPE_IMAGE,
+    DATASET_TYPE_OBJECT_DETECTION,
+}


 class AnnotationExportService:
@@ -137,6 +147,19 @@ class AnnotationExportService:
        """
        project = await self._get_project_or_404(project_id)

+        # 根据格式导出
+        format_type = (
+            ExportFormat(request.format)
+            if isinstance(request.format, str)
+            else request.format
+        )
+
+        # 兼容性检查
+        await self._validate_export_format_compatibility(
+            project=project,
+            format_type=format_type,
+        )
+
        # 获取标注数据
        items = await self._fetch_annotation_data(
            project_id=project_id,
@@ -145,13 +168,6 @@ class AnnotationExportService:
            include_data=request.include_data,
        )

-        # 根据格式导出
-        format_type = (
-            ExportFormat(request.format)
-            if isinstance(request.format, str)
-            else request.format
-        )
-
        if format_type == ExportFormat.JSON:
            return self._export_json(items, project.name)
        elif format_type == ExportFormat.JSONL:
@@ -180,6 +196,77 @@ class AnnotationExportService:
            raise HTTPException(status_code=404, detail=f"标注项目不存在: {project_id}")
        return project

+    @staticmethod
+    def _normalize_type_value(value: Optional[str]) -> str:
+        if not value:
+            return ""
+        return str(value).strip().upper().replace("-", "_")
+
+    async def _get_dataset_type(self, dataset_id: str) -> Optional[str]:
+        result = await self.db.execute(
+            select(Dataset.dataset_type).where(Dataset.id == dataset_id)
+        )
+        return result.scalar_one_or_none()
+
+    async def _get_template_labeling_type(
+        self, template_id: Optional[str]
+    ) -> Optional[str]:
+        if not template_id:
+            return None
+        result = await self.db.execute(
+            select(AnnotationTemplate.labeling_type).where(
+                AnnotationTemplate.id == template_id,
+                AnnotationTemplate.deleted_at.is_(None),
+            )
+        )
+        return result.scalar_one_or_none()
+
+    async def _resolve_project_labeling_type(
+        self, project: LabelingProject
+    ) -> Optional[str]:
+        configuration = getattr(project, "configuration", None)
+        if isinstance(configuration, dict):
+            labeling_type = configuration.get(LABELING_TYPE_CONFIG_KEY)
+            if isinstance(labeling_type, str) and labeling_type.strip():
+                return labeling_type
+        return await self._get_template_labeling_type(project.template_id)
+
+    async def _validate_export_format_compatibility(
+        self,
+        project: LabelingProject,
+        format_type: ExportFormat,
+    ) -> None:
+        if format_type not in (ExportFormat.COCO, ExportFormat.YOLO):
+            return
+
+        dataset_type = self._normalize_type_value(
+            await self._get_dataset_type(project.dataset_id)
+        )
+        labeling_type = self._normalize_type_value(
+            await self._resolve_project_labeling_type(project)
+        )
+
+        if dataset_type == DATASET_TYPE_TEXT:
+            raise HTTPException(
+                status_code=400,
+                detail=f"导出格式 {format_type.value.upper()} 不支持文本类数据集（TEXT），请改用 JSON/JSONL/CSV 格式",
+            )
+
+        if (
+            dataset_type in DETECTION_COMPATIBLE_DATASET_TYPES
+            or labeling_type == LABELING_TYPE_OBJECT_DETECTION
+        ):
+            return
+
+        raise HTTPException(
+            status_code=400,
+            detail=(
+                f"导出格式 {format_type.value.upper()} 仅适用于图像类或目标检测类数据集，"
+                f"当前数据集类型: {dataset_type or 'UNKNOWN'}，"
+                f"标注类型: {labeling_type or 'UNKNOWN'}"
+            ),
+        )
+
    async def _fetch_annotation_data(
        self,
        project_id: str,
--- a/runtime/datamate-python/app/module/annotation/service/mapping.py
+++ b/runtime/datamate-python/app/module/annotation/service/mapping.py
@@ -32,7 +32,8 @@ class DatasetMappingService:
        """Build base query with dataset name joined"""
        return select(
            LabelingProject,
-            Dataset.name.label('dataset_name')
+            Dataset.name.label('dataset_name'),
+            Dataset.dataset_type.label('dataset_type'),
        ).outerjoin(
            Dataset,
            LabelingProject.dataset_id == Dataset.id
@@ -98,6 +99,7 @@ class DatasetMappingService:
        """
        mapping = row[0]  # LabelingProject object
        dataset_name = row[1]  # dataset_name from join
+        dataset_type = row[2]  # dataset_type from join

        # Get template_id from mapping
        template_id = getattr(mapping, 'template_id', None)
@@ -134,6 +136,7 @@ class DatasetMappingService:
            "id": mapping.id,
            "dataset_id": mapping.dataset_id,
            "dataset_name": dataset_name,
+            "dataset_type": dataset_type,
            "labeling_project_id": mapping.labeling_project_id,
            "name": mapping.name,
            "description": description,
@@ -166,12 +169,15 @@ class DatasetMappingService:
        """
        # Fetch dataset name
        dataset_name = None
+        dataset_type = None
        dataset_id = getattr(mapping, 'dataset_id', None)
        if dataset_id:
            dataset_result = await self.db.execute(
-                select(Dataset.name).where(Dataset.id == dataset_id)
+                select(Dataset.name, Dataset.dataset_type).where(Dataset.id == dataset_id)
            )
-            dataset_name = dataset_result.scalar_one_or_none()
+            dataset_row = dataset_result.one_or_none()
+            if dataset_row:
+                dataset_name, dataset_type = dataset_row

        # Get template_id from mapping
        template_id = getattr(mapping, 'template_id', None)
@@ -211,6 +217,7 @@ class DatasetMappingService:
            "id": mapping.id,
            "dataset_id": dataset_id,
            "dataset_name": dataset_name,
+            "dataset_type": dataset_type,
            "labeling_project_id": mapping.labeling_project_id,
            "name": mapping.name,
            "description": description,