feat(annotation): 扩展标注模板功能支持多模态数据类型

- 扩展数据类型支持包括pdf/chat/html/table等多种格式
- 新增标注类型涵盖asr/ner/object-detection等专业领域
- 添加label_config字段用于Label Studio XML配置存储
- 更新模板分类体系为audio-speech/chat/computer-vision/nlp等
- 实现预定义label_config优先使用的配置加载逻辑
- 完善数据库初始化脚本包含多模态标注模板数据

runtime/datamate-python/app/db/models/annotation_management.py

@@ -8,17 +8,18 @@ from app.db.session import Base
 
 class AnnotationTemplate(Base):
     """标注配置模板模型"""
-    
+
     __tablename__ = "t_dm_annotation_templates"
-    
+
     id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
     name = Column(String(100), nullable=False, comment="模板名称")
     description = Column(String(500), nullable=True, comment="模板描述")
-    data_type = Column(String(50), nullable=False, comment="数据类型: image/text/audio/video/timeseries")
-    labeling_type = Column(String(50), nullable=False, comment="标注类型: classification/detection/segmentation/ner/relation/etc")
+    data_type = Column(String(50), nullable=False, comment="数据类型: image/text/audio/video/timeseries/pdf/chat/html/table")
+    labeling_type = Column(String(50), nullable=False, comment="标注类型: asr/ner/object-detection/等")
     configuration = Column(JSON, nullable=False, comment="标注配置（包含labels定义等）")
+    label_config = Column(Text, nullable=True, comment="Label Studio XML配置（内置模板预定义，自定义模板自动生成）")
     style = Column(String(32), nullable=False, comment="样式配置: horizontal/vertical")
-    category = Column(String(50), default='custom', comment="模板分类: medical/general/custom/system")
+    category = Column(String(50), default='custom', comment="模板分类: audio-speech/chat/computer-vision/nlp/等")
     built_in = Column(Boolean, default=False, comment="是否系统内置模板")
     version = Column(String(20), default='1.0', comment="模板版本")
     created_at = Column(TIMESTAMP, server_default=func.current_timestamp(), comment="创建时间")
@@ -33,10 +34,10 @@ class AnnotationTemplate(Base):
         """检查是否已被软删除"""
         return self.deleted_at is not None
     
-class LabelingProject(Base):
-    """标注项目模型"""
-    
-    __tablename__ = "t_dm_labeling_projects"
+class LabelingProject(Base):
+    """标注项目模型"""
+    
+    __tablename__ = "t_dm_labeling_projects"
     
     id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
     dataset_id = Column(String(36), nullable=False, comment="数据集ID")
@@ -53,29 +54,29 @@ class LabelingProject(Base):
         return f"<LabelingProject(id={self.id}, name={self.name}, dataset_id={self.dataset_id})>"
     
     @property
-    def is_deleted(self) -> bool:
-        """检查是否已被软删除"""
-        return self.deleted_at is not None
-
-
-class AnnotationResult(Base):
-    """标注结果模型（单人单份最终标签，Label Studio annotation 原始 JSON）"""
-
-    __tablename__ = "t_dm_annotation_results"
-
-    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
-    project_id = Column(String(36), nullable=False, comment="标注项目ID（t_dm_labeling_projects.id）")
-    file_id = Column(String(36), nullable=False, comment="文件ID（t_dm_dataset_files.id）")
-    annotation = Column(JSON, nullable=False, comment="Label Studio annotation 原始JSON（单人单份最终结果）")
-    created_at = Column(TIMESTAMP, server_default=func.current_timestamp(), comment="创建时间")
-    updated_at = Column(TIMESTAMP, server_default=func.current_timestamp(), onupdate=func.current_timestamp(), comment="更新时间")
-
-    def __repr__(self):
-        return f"<AnnotationResult(id={self.id}, project_id={self.project_id}, file_id={self.file_id})>"
-
-
-class AutoAnnotationTask(Base):
-    """自动标注任务模型，对应表 t_dm_auto_annotation_tasks"""
+    def is_deleted(self) -> bool:
+        """检查是否已被软删除"""
+        return self.deleted_at is not None
+
+
+class AnnotationResult(Base):
+    """标注结果模型（单人单份最终标签，Label Studio annotation 原始 JSON）"""
+
+    __tablename__ = "t_dm_annotation_results"
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
+    project_id = Column(String(36), nullable=False, comment="标注项目ID（t_dm_labeling_projects.id）")
+    file_id = Column(String(36), nullable=False, comment="文件ID（t_dm_dataset_files.id）")
+    annotation = Column(JSON, nullable=False, comment="Label Studio annotation 原始JSON（单人单份最终结果）")
+    created_at = Column(TIMESTAMP, server_default=func.current_timestamp(), comment="创建时间")
+    updated_at = Column(TIMESTAMP, server_default=func.current_timestamp(), onupdate=func.current_timestamp(), comment="更新时间")
+
+    def __repr__(self):
+        return f"<AnnotationResult(id={self.id}, project_id={self.project_id}, file_id={self.file_id})>"
+
+
+class AutoAnnotationTask(Base):
+    """自动标注任务模型，对应表 t_dm_auto_annotation_tasks"""
 
     __tablename__ = "t_dm_auto_annotation_tasks"
 
@@ -108,4 +109,4 @@ class AutoAnnotationTask(Base):
     @property
     def is_deleted(self) -> bool:
         """检查是否已被软删除"""
-        return self.deleted_at is not None
+        return self.deleted_at is not None

runtime/datamate-python/app/module/annotation/service/template.py

@@ -333,10 +333,10 @@ class AnnotationTemplateService:
     def _to_response(self, template: AnnotationTemplate) -> AnnotationTemplateResponse:
         """
         转换为响应对象
-        
+
         Args:
             template: 数据库模型对象
-            
+
         Returns:
             模板响应对象
         """
@@ -344,13 +344,16 @@ class AnnotationTemplateService:
         from typing import cast, Dict, Any
         config_dict = cast(Dict[str, Any], template.configuration)
         config = TemplateConfiguration(**config_dict)
-        
-        # 动态生成Label Studio XML配置
-        label_config = self.generate_label_studio_config(config)
-        
+
+        # 优先使用预定义的 label_config，否则动态生成
+        if template.label_config:
+            label_config = template.label_config
+        else:
+            label_config = self.generate_label_studio_config(config)
+
         # 使用model_validate从ORM对象创建响应对象
         response = AnnotationTemplateResponse.model_validate(template)
         response.configuration = config
         response.label_config = label_config  # type: ignore
-        
+
         return response