deepdata: core product code (backend, frontend, runtime, deepflow module)

runtime/deepdata-python/app/db/datascope.py

@@ -0,0 +1,37 @@
+from contextvars import ContextVar
+from typing import List, Optional
+
+_current_user: ContextVar[Optional[str]] = ContextVar("_current_user", default=None)
+SYSTEM_USER = "system"
+
+
+class DataScopeHandle:
+    """
+    Hold current user info in a ContextVar and provide helpers for SQLAlchemy filters.
+    """
+
+    @staticmethod
+    def set_user_info(user: Optional[str]) -> None:
+        if user is None or user == "":
+            # set explicit None
+            _current_user.set(None)
+        else:
+            _current_user.set(user)
+
+    @staticmethod
+    def remove_user_info() -> None:
+        _current_user.set(None)
+
+    @staticmethod
+    def get_user_info() -> Optional[str]:
+        return _current_user.get()
+
+    @staticmethod
+    def allowed_users() -> List[str]:
+        """
+        Return list of allowed creators: current user + system.
+        """
+        user = DataScopeHandle.get_user_info()
+        if not user:
+            return []
+        return [user, SYSTEM_USER]

runtime/deepdata-python/app/db/models/__init__.py

@@ -0,0 +1,42 @@
+from .dataset_management import (
+    Dataset,
+    DatasetTag,
+    DatasetFiles,
+    DatasetStatistics,
+    Tag,
+)
+
+from .user_management import User
+
+from .annotation_management import AnnotationTemplate, LabelingProject
+
+from .data_evaluation import EvaluationTask, EvaluationItem
+
+from .operator import Operator, Category, CategoryRelation, OperatorRelease
+
+from .chunk_upload import ChunkUploadPreRequest
+
+from .knowledge_gen import KnowledgeBase, RagFile
+
+from .sys_param import SysParam
+
+__all__ = [
+    "Dataset",
+    "DatasetTag",
+    "DatasetFiles",
+    "DatasetStatistics",
+    "Tag",
+    "User",
+    "AnnotationTemplate",
+    "LabelingProject",
+    "EvaluationTask",
+    "EvaluationItem",
+    "Operator",
+    "Category",
+    "CategoryRelation",
+    "OperatorRelease",
+    "ChunkUploadPreRequest",
+    "KnowledgeBase",
+    "RagFile",
+    "SysParam",
+]

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

@@ -0,0 +1,85 @@
+"""Tables of Annotation Management Module"""
+
+import uuid
+from sqlalchemy import Column, String, Boolean, TIMESTAMP, Text, Integer, JSON, ForeignKey
+from sqlalchemy.sql import func
+
+from app.db.models.base_entity import BaseEntity
+
+class AnnotationTemplate(BaseEntity):
+    """标注配置模板模型"""
+
+    __tablename__ = "t_dm_annotation_templates"
+    __ignore_data_scope__ = True
+
+    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")
+    configuration = Column(JSON, nullable=False, comment="标注配置（包含labels定义等）")
+    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")
+    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")
+    deleted_at = Column(TIMESTAMP, nullable=True, comment="删除时间（软删除）")
+
+    def __repr__(self):
+        return f"<AnnotationTemplate(id={self.id}, name={self.name}, data_type={self.data_type})>"
+
+    @property
+    def is_deleted(self) -> bool:
+        """检查是否已被软删除"""
+        return self.deleted_at is not None
+
+class LabelingProject(BaseEntity):
+    """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")
+    name = Column(String(100), nullable=False, comment="项目名称")
+    labeling_project_id = Column(String(8), nullable=False, comment="Label Studio项目ID")
+    template_id = Column(String(36), ForeignKey('t_dm_annotation_templates.id', ondelete='SET NULL'), nullable=True, comment="使用的模板ID")
+    configuration = Column(JSON, nullable=True, comment="项目配置（可能包含对模板的自定义修改）")
+    progress = Column(JSON, nullable=True, comment="项目进度信息")
+    deleted_at = Column(TIMESTAMP, nullable=True, comment="删除时间（软删除）")
+
+    def __repr__(self):
+        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 AutoAnnotationTask(BaseEntity):
+    """自动标注任务模型，对应表 t_dm_auto_annotation_tasks"""
+
+    __tablename__ = "t_dm_auto_annotation_tasks"
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
+    name = Column(String(255), nullable=False, comment="任务名称")
+    dataset_id = Column(String(36), nullable=False, comment="数据集ID")
+    dataset_name = Column(String(255), nullable=True, comment="数据集名称（冗余字段，方便查询）")
+    config = Column(JSON, nullable=False, comment="任务配置（模型规模、置信度等）")
+    file_ids = Column(JSON, nullable=True, comment="要处理的文件ID列表，为空则处理数据集所有图像")
+    status = Column(String(50), nullable=False, default="pending", comment="任务状态: pending/running/completed/failed")
+    progress = Column(Integer, default=0, comment="任务进度 0-100")
+    total_images = Column(Integer, default=0, comment="总图片数")
+    processed_images = Column(Integer, default=0, comment="已处理图片数")
+    detected_objects = Column(Integer, default=0, comment="检测到的对象总数")
+    output_path = Column(String(500), nullable=True, comment="输出路径")
+    error_message = Column(Text, nullable=True, comment="错误信息")
+    completed_at = Column(TIMESTAMP, nullable=True, comment="完成时间")
+    deleted_at = Column(TIMESTAMP, nullable=True, comment="删除时间（软删除）")
+
+    def __repr__(self) -> str:  # pragma: no cover - repr 简单返回
+        return f"<AutoAnnotationTask(id={self.id}, name={self.name}, status={self.status})>"
+
+    @property
+    def is_deleted(self) -> bool:
+        """检查是否已被软删除"""
+        return self.deleted_at is not None

runtime/deepdata-python/app/db/models/base_entity.py

@@ -0,0 +1,51 @@
+from sqlalchemy import Column, String, TIMESTAMP, Text
+from sqlalchemy.orm import declarative_base
+from sqlalchemy.sql import func
+
+Base = declarative_base()
+
+
+class BaseEntity(Base):
+    """
+    Common base entity with audit fields.
+    Subclasses may set `__ignore_data_scope__ = True` to opt-out of data-scope filtering.
+    """
+    __abstract__ = True
+
+    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")
+    created_by = Column(String(255), nullable=True, comment="创建者")
+    updated_by = Column(String(255), nullable=True, comment="更新者")
+
+    # default: do enforce data scope unless subclass sets this to True
+    __ignore_data_scope__ = False
+
+
+class LineageNode(Base):
+    """数据血缘：节点表（实体对象）"""
+
+    __tablename__ = "t_lineage_node"
+
+    id = Column(String(36), primary_key=True, comment="节点ID")
+    graph_id = Column(String(36), nullable=True, comment="图ID")
+    node_type = Column(String(64), nullable=False, comment="节点类型")
+    name = Column(String(256), nullable=False, comment="节点名称")
+    description = Column(Text, nullable=True, comment="节点描述")
+    node_metadata = Column(Text, nullable=True, comment="节点扩展信息（JSON）")
+
+
+class LineageEdge(Base):
+    """数据血缘：边表（处理流程）"""
+
+    __tablename__ = "t_lineage_edge"
+
+    id = Column(String(36), primary_key=True, comment="边ID")
+    graph_id = Column(String(36), nullable=True, comment="图ID")
+    process_id = Column(String(36), nullable=True, comment="处理流程ID")
+    edge_type = Column(String(64), nullable=False, comment="边类型")
+    name = Column(String(256), nullable=True, comment="边名称")
+    description = Column(Text, nullable=True, comment="边描述")
+    edge_metadata = Column(Text, nullable=True, comment="边扩展信息（JSON）")
+    from_node_id = Column(String(36), nullable=False, comment="源节点ID")
+    to_node_id = Column(String(36), nullable=False, comment="目标节点ID")

runtime/deepdata-python/app/db/models/chunk_upload.py

@@ -0,0 +1,38 @@
+"""
+Chunk Upload Database Model
+分片上传数据库模型
+"""
+from sqlalchemy import Column, String, Integer, DateTime
+from sqlalchemy.sql import func
+
+from app.db.models.base_entity import Base
+
+
+class ChunkUploadPreRequest(Base):
+    """分片上传预请求"""
+    __tablename__ = "t_chunk_upload_request"
+
+    id = Column(String(36), primary_key=True, comment="请求ID")
+    total_file_num = Column(Integer, nullable=False, comment="总文件数")
+    uploaded_file_num = Column(Integer, nullable=True, comment="已上传文件数")
+    upload_path = Column(String(512), nullable=False, comment="文件路径")
+    timeout = Column(DateTime, nullable=False, comment="上传请求超时时间")
+    service_id = Column(String(64), nullable=True, comment="上传请求所属服务ID")
+    check_info = Column(String(512), nullable=True, comment="业务信息")
+
+    def increment_uploaded_file_num(self):
+        """增加已上传文件数"""
+        if self.uploaded_file_num is None:
+            self.uploaded_file_num = 1
+        else:
+            self.uploaded_file_num += 1
+
+    def is_upload_complete(self) -> bool:
+        """检查是否已完成上传"""
+        return (self.uploaded_file_num is not None and
+                self.uploaded_file_num == self.total_file_num)
+
+    def is_request_timeout(self) -> bool:
+        """检查是否已超时"""
+        from datetime import datetime
+        return self.timeout is not None and datetime.utcnow() > self.timeout

runtime/deepdata-python/app/db/models/cleaning.py

@@ -0,0 +1,69 @@
+from sqlalchemy import Column, String, BigInteger, Integer, TIMESTAMP, Index
+from app.db.models.base_entity import BaseEntity, Base
+
+
+class CleaningTask(BaseEntity):
+    """Data cleaning task entity"""
+    __tablename__ = "t_clean_task"
+
+    id = Column(String(36), primary_key=True, comment="Task ID")
+    name = Column(String(255), nullable=False, comment="Task name")
+    description = Column(String(1024), nullable=True, comment="Task description")
+    status = Column(String(50), nullable=False, default="PENDING", comment="Task status: PENDING, RUNNING, COMPLETED, STOPPED, FAILED")
+    src_dataset_id = Column(String(36), nullable=False, comment="Source dataset ID")
+    src_dataset_name = Column(String(255), nullable=False, comment="Source dataset name")
+    dest_dataset_id = Column(String(36), nullable=True, comment="Destination dataset ID")
+    dest_dataset_name = Column(String(255), nullable=True, comment="Destination dataset name")
+    before_size = Column(BigInteger, nullable=True, comment="Data size before cleaning")
+    after_size = Column(BigInteger, nullable=True, comment="Data size after cleaning")
+    file_count = Column(Integer, nullable=True, comment="Total file count")
+    retry_count = Column(Integer, default=0, nullable=False, comment="Retry count")
+    started_at = Column(TIMESTAMP, nullable=True, comment="Task start time")
+    finished_at = Column(TIMESTAMP, nullable=True, comment="Task finish time")
+
+    __table_args__ = (
+        Index("ix_clean_task_status", "status"),
+        Index("ix_clean_task_created_at", "created_at"),
+    )
+
+
+class CleaningTemplate(BaseEntity):
+    """Data cleaning template entity"""
+    __tablename__ = "t_clean_template"
+
+    id = Column(String(36), primary_key=True, comment="Template ID")
+    name = Column(String(255), nullable=False, comment="Template name")
+    description = Column(String(1024), nullable=True, comment="Template description")
+
+
+class CleaningResult(Base):
+    """Data cleaning result entity"""
+    __tablename__ = "t_clean_result"
+
+    instance_id = Column(String(36), primary_key=True, comment="Instance ID (task or template ID)")
+    src_file_id = Column(String(36), primary_key=True, comment="Source file ID")
+    dest_file_id = Column(String(36), nullable=True, comment="Destination file ID")
+    src_name = Column(String(512), nullable=True, comment="Source file name")
+    dest_name = Column(String(512), nullable=True, comment="Destination file name")
+    src_type = Column(String(50), nullable=True, comment="Source file type")
+    dest_type = Column(String(50), nullable=True, comment="Destination file type")
+    src_size = Column(BigInteger, nullable=True, comment="Source file size")
+    dest_size = Column(BigInteger, nullable=True, comment="Destination file size")
+    status = Column(String(50), nullable=True, comment="Cleaning status: COMPLETED, FAILED, etc.")
+    result = Column(String(1024), nullable=True, comment="Cleaning result message")
+
+    __table_args__ = (
+        Index("ix_clean_result_instance_id", "instance_id"),
+        Index("ix_clean_result_instance_status", "instance_id", "status"),
+    )
+
+
+class OperatorInstance(Base):
+    """Operator instance in task or template"""
+    __tablename__ = "t_operator_instance"
+
+    instance_id = Column(String(36), primary_key=True, comment="Instance ID (task or template ID)")
+    operator_id = Column(String(36), primary_key=True, comment="Operator ID")
+    op_index = Column(Integer, nullable=False, comment="Operator execution order")
+    settings_override = Column(String(4096), nullable=True, comment="Operator settings override (JSON)")
+

runtime/deepdata-python/app/db/models/data_collection.py

@@ -0,0 +1,56 @@
+import uuid
+from sqlalchemy import Column, String, Text, TIMESTAMP, Integer, BigInteger, Numeric, JSON, Boolean
+from sqlalchemy.sql import func
+
+from app.db.models.base_entity import BaseEntity
+
+
+class CollectionTemplate(BaseEntity):
+    """归集模板表（UUID 主键） -> t_dc_collection_templates"""
+
+    __tablename__ = "t_dc_collection_templates"
+    __ignore_data_scope__ = True
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="模板ID（UUID）")
+    name = Column(String(255), nullable=False, comment="模板名称")
+    description = Column(Text, nullable=True, comment="模板描述")
+    source_type = Column(String(64), nullable=False, comment="源数据源类型")
+    source_name = Column(String(64), nullable=False, comment="源数据源名称")
+    target_type = Column(String(64), nullable=False, comment="目标数据源类型")
+    target_name = Column(String(64), nullable=False, comment="目标数据源名称")
+    template_content = Column(JSON, nullable=False, comment="模板内容")
+    built_in = Column(Boolean, default=False, comment="是否系统内置模板")
+
+class CollectionTask(BaseEntity):
+    """归集任务表（UUID 主键） -> t_dc_collection_tasks"""
+
+    __tablename__ = "t_dc_collection_tasks"
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
+    name = Column(String(255), nullable=False, comment="任务名称")
+    description = Column(Text, nullable=True, comment="任务描述")
+    sync_mode = Column(String(20), nullable=False, server_default="ONCE", comment="同步模式：ONCE/SCHEDULED")
+    template_id = Column(String(36), nullable=False, comment="归集模板ID")
+    template_name = Column(String(255), nullable=False, comment="归集模板名称")
+    target_path = Column(String(1000), nullable=True, server_default="", comment="目标存储路径")
+    config = Column(JSON, nullable=False, comment="归集配置（DataX配置），包含源端和目标端配置信息")
+    schedule_expression = Column(String(255), nullable=True, comment="Cron调度表达式")
+    status = Column(String(20), nullable=True, server_default="DRAFT", comment="任务状态：DRAFT/READY/RUNNING/SUCCESS/FAILED/STOPPED")
+    retry_count = Column(Integer, nullable=True, server_default="3", comment="重试次数")
+    timeout_seconds = Column(Integer, nullable=True, server_default="3600", comment="超时时间（秒）")
+    last_execution_id = Column(String(36), nullable=True, comment="最后执行ID（UUID）")
+
+class TaskExecution(BaseEntity):
+    """任务执行记录表（UUID 主键） -> t_dc_task_executions"""
+
+    __tablename__ = "t_dc_task_executions"
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="执行记录ID（UUID）")
+    task_id = Column(String(36), nullable=False, comment="任务ID")
+    task_name = Column(String(255), nullable=False, comment="任务名称")
+    status = Column(String(20), nullable=True, server_default="RUNNING", comment="执行状态：RUNNING/SUCCESS/FAILED/STOPPED")
+    log_path = Column(String(1000), nullable=True, server_default="", comment="日志文件路径")
+    started_at = Column(TIMESTAMP, nullable=True, comment="开始时间")
+    completed_at = Column(TIMESTAMP, nullable=True, comment="完成时间")
+    duration_seconds = Column(Integer, nullable=True, server_default="0", comment="执行时长（秒）")
+    error_message = Column(Text, nullable=True, comment="错误信息")

runtime/deepdata-python/app/db/models/data_evaluation.py

@@ -0,0 +1,73 @@
+"""
+Tables for Data Evaluation module
+
+Derived from scripts/db/data-evaluation-init.sql
+ - t_de_eval_task
+ - t_de_eval_item
+"""
+
+import uuid
+from sqlalchemy import Column, String, Text, Float, TIMESTAMP, ForeignKey, Integer
+from sqlalchemy.sql import func
+
+from app.db.models.base_entity import BaseEntity
+
+
+class EvaluationTask(BaseEntity):
+    """评估任务表（UUID 主键） -> t_de_eval_task
+
+    Columns per data-evaluation-init.sql:
+      id, name, description, task_type, source_type, source_id, source_name,
+      status, eval_process, eval_promt, eval_config, created_at, updated_at,
+      created_by, updated_by
+    """
+
+    __tablename__ = "t_de_eval_task"
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
+    name = Column(String(255), nullable=False, comment="评估任务名称")
+    description = Column(Text, nullable=True, comment="评估任务描述")
+    task_type = Column(String(50), nullable=False, comment="评估任务类型：QA")
+    source_type = Column(String(36), nullable=True, comment="待评估对象类型：DATASET/SYNTHESIS")
+    source_id = Column(String(36), nullable=True, comment="待评估对象ID")
+    source_name = Column(String(255), nullable=True, comment="待评估对象名称")
+    status = Column(String(50), server_default="PENDING", nullable=False, comment="状态：PENDING/RUNNING/COMPLETED/STOPPED/FAILED")
+    eval_method = Column(String(50), server_default="AUTO", nullable=False, comment="评估方式：AUTO/MANUAL")
+    eval_process = Column(Float, nullable=False, server_default="0", comment="评估进度")
+    eval_prompt = Column(Text, nullable=True, comment="评估提示词")
+    eval_config = Column(Text, nullable=True, comment="评估配置")
+
+
+class EvaluationFile(BaseEntity):
+    """评估条目表（UUID 主键） -> t_de_eval_file"""
+
+    __tablename__ = "t_de_eval_file"
+    __ignore_data_scope__ = True
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
+    task_id = Column(String(36), ForeignKey('t_de_eval_task.id'), nullable=False, comment="评估任务ID")
+    file_id = Column(String(36), ForeignKey('t_dm_dataset_files.id'), nullable=True, comment="文件ID")
+    file_name = Column(String(255), nullable=False, comment="文件名")
+    error_message = Column(Text, nullable=True, comment="错误信息")
+    total_count = Column(Integer, nullable=False, default=0, comment="总数")
+    evaluated_count = Column(Integer, nullable=False, default=0, comment="已评估数")
+
+
+class EvaluationItem(BaseEntity):
+    """评估条目表（UUID 主键） -> t_de_eval_item
+
+    Columns per data-evaluation-init.sql:
+      id, task_id, item_id, eval_score, eval_result, status
+    """
+
+    __tablename__ = "t_de_eval_item"
+    __ignore_data_scope__ = True
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
+    task_id = Column(String(36), ForeignKey('t_de_eval_task.id'), nullable=False, comment="评估任务ID")
+    file_id = Column(String(36), ForeignKey('t_dm_dataset_files.id'), nullable=True, comment="文件ID")
+    item_id = Column(String(36), nullable=False, comment="评估条目ID")
+    eval_content = Column(Text, nullable=True, comment="评估内容")
+    eval_score = Column(Float, nullable=False, server_default="0", comment="评估分数")
+    eval_result = Column(Text, nullable=True, comment="评估结果")
+    status = Column(String(50), server_default="PENDING", nullable=False, comment="状态：PENDING/EVALUATED")

runtime/deepdata-python/app/db/models/data_synthesis.py

@@ -0,0 +1,177 @@
+import uuid
+
+from sqlalchemy import Column, String, Text, Integer, JSON, TIMESTAMP, func
+
+from app.db.models.base_entity import Base, BaseEntity
+from app.module.generation.schema.generation import CreateSynthesisTaskRequest
+
+
+async def save_synthesis_task(db_session, synthesis_task: CreateSynthesisTaskRequest):
+    """保存数据合成任务。
+
+    注意：当前 MySQL 表 `t_data_synth_instances` 结构中只包含 synth_type / synth_config 等字段，
+    没有 model_id、text_split_config、source_file_id、result_data_location 等列，因此这里只保存
+    与表结构一致的字段，其他信息由上层逻辑或其它表负责管理。
+    """
+    gid = str(uuid.uuid4())
+
+    # 兼容旧请求结构：从请求对象中提取必要字段，
+    #   - 合成类型：synthesis_type -> synth_type
+    #   - 合成配置：text_split_config + synthesis_config 合并后写入 synth_config
+
+    synth_task_instance = DataSynthInstance(
+        id=gid,
+        name=synthesis_task.name,
+        description=synthesis_task.description,
+        status="pending",
+        synth_type=synthesis_task.synthesis_type.value,
+        progress=0,
+        synth_config=synthesis_task.synth_config.model_dump(),
+        total_files=len(synthesis_task.source_file_id or []),
+        processed_files=0,
+        total_chunks=0,
+        processed_chunks=0,
+        total_synth_data=0,
+        created_at=func.now(),
+        updated_at=func.now(),
+        created_by="system",
+        updated_by="system",
+    )
+    db_session.add(synth_task_instance)
+    await db_session.commit()
+    await db_session.refresh(synth_task_instance)
+    return synth_task_instance
+
+
+class DataSynthInstance(BaseEntity):
+    """数据合成任务表，对应表 t_data_synth_instances
+
+    create table if not exists t_data_synth_instances
+    (
+        id VARCHAR(36) CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci PRIMARY KEY COMMENT 'UUID',
+        name VARCHAR(255) NOT NULL COMMENT '任务名称',
+        description TEXT COMMENT '任务描述',
+        status VARCHAR(20) COMMENT '任务状态',
+        synth_type VARCHAR(20) NOT NULL COMMENT '合成类型',
+        progress INT DEFAULT 0 COMMENT '任务进度(百分比)',
+        synth_config JSON NOT NULL COMMENT '合成配置',
+        total_files INT DEFAULT 0 COMMENT '总文件数',
+        processed_files INT DEFAULT 0 COMMENT '已处理文件数',
+        total_chunks INT DEFAULT 0 COMMENT '总文本块数',
+        processed_chunks INT DEFAULT 0 COMMENT '已处理文本块数',
+        total_synth_data INT DEFAULT 0 COMMENT '总合成数据量',
+        created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP COMMENT '创建时间',
+        updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP COMMENT '更新时间',
+        created_by VARCHAR(255) COMMENT '创建者',
+        updated_by VARCHAR(255) COMMENT '更新者'
+    ) COMMENT='数据合成任务表（UUID 主键）';
+    """
+
+    __tablename__ = "t_data_synth_instances"
+
+    id = Column(String(36), primary_key=True, index=True, comment="UUID")
+    name = Column(String(255), nullable=False, comment="任务名称")
+    description = Column(Text, nullable=True, comment="任务描述")
+    status = Column(String(20), nullable=True, comment="任务状态")
+    # 与数据库字段保持一致：synth_type / synth_config
+    synth_type = Column(String(20), nullable=False, comment="合成类型")
+    progress = Column(Integer, nullable=False, default=0, comment="任务进度(百分比)")
+    synth_config = Column(JSON, nullable=False, comment="合成配置")
+    total_files = Column(Integer, nullable=False, default=0, comment="总文件数")
+    processed_files = Column(Integer, nullable=False, default=0, comment="已处理文件数")
+    total_chunks = Column(Integer, nullable=False, default=0, comment="总文本块数")
+    processed_chunks = Column(Integer, nullable=False, default=0, comment="已处理文本块数")
+    total_synth_data = Column(Integer, nullable=False, default=0, comment="总合成数据量")
+
+
+class DataSynthesisFileInstance(BaseEntity):
+    """数据合成文件任务表，对应表 t_data_synthesis_file_instances
+
+    create table if not exists t_data_synthesis_file_instances (
+        id VARCHAR(36) PRIMARY KEY COMMENT 'UUID',
+        synthesis_instance_id VARCHAR(36) COMMENT '数据合成任务ID',
+        file_name VARCHAR(255) NOT NULL COMMENT '文件名',
+        source_file_id VARCHAR(255) NOT NULL COMMENT '原始文件ID',
+        target_file_location VARCHAR(1000) NOT NULL COMMENT '目标文件存储位置',
+        status VARCHAR(20) COMMENT '任务状态',
+        total_chunks INT DEFAULT 0 COMMENT '总文本块数',
+        processed_chunks INT DEFAULT 0 COMMENT '已处理文本块数',
+        created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP COMMENT '创建时间',
+        updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP COMMENT '更新时间',
+        created_by VARCHAR(255) COMMENT '创建者',
+        updated_by VARCHAR(255) COMMENT '更新者'
+    ) COMMENT='数据合成文件任务表（UUID 主键）';
+    """
+
+    __tablename__ = "t_data_synthesis_file_instances"
+
+    id = Column(String(36), primary_key=True, index=True, comment="UUID")
+    synthesis_instance_id = Column(
+        String(36),
+        nullable=False,
+        comment="数据合成任务ID",
+        index=True,
+    )
+    file_name = Column(String(255), nullable=False, comment="文件名")
+    source_file_id = Column(String(255), nullable=False, comment="原始文件ID")
+    target_file_location = Column(String(1000), nullable=True, comment="目标文件存储位置")
+    status = Column(String(20), nullable=True, comment="任务状态")
+    total_chunks = Column(Integer, nullable=False, default=0, comment="总文本块数")
+    processed_chunks = Column(Integer, nullable=False, default=0, comment="已处理文本块数")
+
+
+class DataSynthesisChunkInstance(Base):
+    """数据合成分块任务表，对应表 t_data_synthesis_chunk_instances
+
+    create table if not exists t_data_synthesis_chunk_instances (
+        id VARCHAR(36) PRIMARY KEY COMMENT 'UUID',
+        synthesis_file_instance_id VARCHAR(36) COMMENT '数据合成文件任务ID',
+        chunk_index INT COMMENT '分块索引',
+        chunk_content TEXT COMMENT '分块内容',
+        metadata JSON COMMENT '分块元数据'
+    ) COMMENT='数据合成分块任务表（UUID 主键）';
+    """
+
+    __tablename__ = "t_data_synthesis_chunk_instances"
+
+    id = Column(String(36), primary_key=True, index=True, comment="UUID")
+    synthesis_file_instance_id = Column(
+        String(36),
+        nullable=False,
+        comment="数据合成文件任务ID",
+        index=True,
+    )
+    chunk_index = Column(Integer, nullable=True, comment="分块索引")
+    chunk_content = Column(Text, nullable=True, comment="分块内容")
+    # SQLAlchemy Declarative 保留了属性名 'metadata'，这里使用 chunk_metadata 作为属性名，
+    # 底层列名仍为 'metadata' 以保持与表结构兼容。
+    chunk_metadata = Column("metadata", JSON, nullable=True, comment="分块元数据")
+
+
+class SynthesisData(Base):
+    """数据合成结果表，对应表 t_synthesis_data
+
+    create table if not exists t_synthesis_data (
+        id VARCHAR(36) PRIMARY KEY COMMENT 'UUID',
+        data json COMMENT '合成的数据',
+        synthesis_file_instance_id VARCHAR(36) COMMENT '数据合成文件任务ID',
+        chunk_instance_id VARCHAR(36) COMMENT '分块任务ID'
+    ) COMMENT='数据合成任务队列表（UUID 主键）';
+    """
+
+    __tablename__ = "t_data_synthesis_data"
+
+    id = Column(String(36), primary_key=True, index=True, comment="UUID")
+    data = Column(JSON, nullable=True, comment="合成的数据")
+    synthesis_file_instance_id = Column(
+        String(36),
+        nullable=False,
+        comment="数据合成文件任务ID",
+        index=True,
+    )
+    chunk_instance_id = Column(
+        String(36),
+        nullable=False,
+        comment="分块任务ID",
+        index=True,
+    )

runtime/deepdata-python/app/db/models/dataset_management.py

@@ -0,0 +1,111 @@
+"""
+Tables of Dataset Management Module
+"""
+
+import uuid
+from sqlalchemy import Column, String, BigInteger, Boolean, TIMESTAMP, Text, Integer, JSON, Date
+from sqlalchemy.sql import func
+
+from app.db.models.base_entity import Base, BaseEntity
+
+
+class Dataset(BaseEntity):
+    """数据集模型（支持医学影像、文本、问答等多种类型）"""
+
+    __tablename__ = "t_dm_datasets"
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
+    name = Column(String(255), nullable=False, comment="数据集名称")
+    description = Column(Text, nullable=True, comment="数据集描述")
+    dataset_type = Column(String(50), nullable=False, comment="数据集类型：IMAGE/TEXT/QA/MULTIMODAL/OTHER")
+    category = Column(String(100), nullable=True, comment="数据集分类：医学影像/问答/文献等")
+    path = Column(String(500), nullable=True, comment="数据存储路径")
+    format = Column(String(50), nullable=True, comment="数据格式：DCM/JPG/JSON/CSV等")
+    schema_info = Column(JSON, nullable=True, comment="数据结构信息")
+    size_bytes = Column(BigInteger, default=0, comment="数据大小(字节)")
+    file_count = Column(BigInteger, default=0, comment="文件数量")
+    record_count = Column(BigInteger, default=0, comment="记录数量")
+    retention_days = Column(Integer, default=0, comment="数据保留天数（0表示长期保留）")
+    tags = Column(JSON, nullable=True, comment="标签列表")
+    dataset_metadata = Column("metadata", JSON, nullable=True, comment="元数据信息")
+    status = Column(String(50), default='DRAFT', comment="状态：DRAFT/ACTIVE/ARCHIVED")
+    is_public = Column(Boolean, default=False, comment="是否公开")
+    is_featured = Column(Boolean, default=False, comment="是否推荐")
+    version = Column(BigInteger, nullable=False, default=0, comment="版本号")
+
+    def __repr__(self):
+        return f"<Dataset(id={self.id}, name={self.name}, type={self.dataset_type})>"
+
+class DatasetTag(BaseEntity):
+    """数据集标签关联模型"""
+
+    __tablename__ = "t_dm_dataset_tags"
+
+    dataset_id = Column(String(36), primary_key=True, comment="数据集ID（UUID）")
+    tag_id = Column(String(36), primary_key=True, comment="标签ID（UUID）")
+
+    def __repr__(self):
+        return f"<DatasetTag(dataset_id={self.dataset_id}, tag_id={self.tag_id})>"
+
+class DatasetFiles(Base):
+    """DM数据集文件模型"""
+
+    __tablename__ = "t_dm_dataset_files"
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
+    dataset_id = Column(String(36), nullable=False, comment="所属数据集ID（UUID）")
+    file_name = Column(String(255), nullable=False, comment="文件名")
+    file_path = Column(String(1000), nullable=False, comment="文件路径")
+    file_type = Column(String(50), nullable=True, comment="文件格式：JPG/PNG/DCM/TXT等")
+    file_size = Column(BigInteger, default=0, comment="文件大小(字节)")
+    check_sum = Column(String(64), nullable=True, comment="文件校验和")
+    tags = Column(JSON, nullable=True, comment="文件标签信息")
+    tags_updated_at = Column(TIMESTAMP, nullable=True, comment="标签最后更新时间")
+    annotation = Column(JSON, nullable=True, comment="完整标注结果（原始JSON）")
+    dataset_filemetadata = Column("metadata", JSON, nullable=True, comment="文件元数据")
+    status = Column(String(50), default='ACTIVE', comment="文件状态：ACTIVE/DELETED/PROCESSING")
+    upload_time = Column(TIMESTAMP, server_default=func.current_timestamp(), comment="上传时间")
+    last_access_time = Column(TIMESTAMP, nullable=True, comment="最后访问时间")
+    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"<DatasetFiles(id={self.id}, dataset_id={self.dataset_id}, file_name={self.file_name})>"
+
+class DatasetStatistics(Base):
+    """数据集统计信息模型"""
+
+    __tablename__ = "t_dm_dataset_statistics"
+
+    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（UUID）")
+    stat_date = Column(Date, nullable=False, comment="统计日期")
+    total_files = Column(BigInteger, default=0, comment="总文件数")
+    total_size = Column(BigInteger, default=0, comment="总大小(字节)")
+    processed_files = Column(BigInteger, default=0, comment="已处理文件数")
+    error_files = Column(BigInteger, default=0, comment="错误文件数")
+    download_count = Column(BigInteger, default=0, comment="下载次数")
+    view_count = Column(BigInteger, default=0, comment="查看次数")
+    quality_metrics = Column(JSON, nullable=True, comment="质量指标")
+    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"<DatasetStatistics(id={self.id}, dataset_id={self.dataset_id}, date={self.stat_date})>"
+
+class Tag(Base):
+    """标签集合模型"""
+
+    __tablename__ = "t_dm_tags"
+
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
+    name = Column(String(100), nullable=False, unique=True, comment="标签名称")
+    description = Column(Text, nullable=True, comment="标签描述")
+    category = Column(String(50), nullable=True, comment="标签分类")
+    color = Column(String(7), nullable=True, comment="标签颜色(十六进制)")
+    usage_count = Column(BigInteger, default=0, comment="使用次数")
+    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"<Tag(id={self.id}, name={self.name}, category={self.category})>"

runtime/deepdata-python/app/db/models/knowledge_gen.py

@@ -0,0 +1,82 @@
+"""
+知识库（RAG）相关 ORM 模型
+
+表: t_rag_knowledge_base, t_rag_file
+与 Java 实体保持一致。
+"""
+from enum import Enum
+from sqlalchemy import Column, String, Integer, JSON
+from app.db.models.base_entity import BaseEntity
+
+
+class RagType(str, Enum):
+    """RAG 类型枚举
+
+    对应 Java: com.deepdata.rag.indexer.interfaces.dto.RagType
+    """
+    DOCUMENT = "DOCUMENT"  # 文档型 RAG（向量检索）
+    GRAPH = "GRAPH"        # 知识图谱型 RAG（LightRAG）
+
+
+class FileStatus(str, Enum):
+    """文件状态枚举
+
+    对应 Java: com.deepdata.rag.indexer.domain.model.FileStatus
+    """
+    UNPROCESSED = "UNPROCESSED"        # 未处理
+    PROCESSING = "PROCESSING"          # 处理中
+    PROCESSED = "PROCESSED"            # 已处理
+    PROCESS_FAILED = "PROCESS_FAILED"  # 处理失败
+
+
+class KnowledgeBase(BaseEntity):
+    """知识库实体
+
+    对应 Java: com.deepdata.rag.indexer.domain.model.KnowledgeBase
+    表名: t_rag_knowledge_base
+    """
+    __tablename__ = "t_rag_knowledge_base"
+    __ignore_data_scope__ = True
+
+    id = Column(String(36), primary_key=True, comment="知识库ID")
+    name = Column(String(255), nullable=False, unique=True, comment="知识库名称")
+    description = Column(String(512), nullable=True, comment="知识库描述")
+    type = Column(
+        String(50),
+        nullable=False,
+        default=RagType.DOCUMENT,
+        comment="RAG类型",
+    )
+    embedding_model = Column(String(255), nullable=False, comment="嵌入模型ID")
+    chat_model = Column(String(255), nullable=True, comment="聊天模型ID")
+
+    def __repr__(self):
+        return f"<KnowledgeBase(id={self.id}, name={self.name}, type={self.type})>"
+
+
+class RagFile(BaseEntity):
+    """RAG 文件实体
+
+    对应 Java: com.deepdata.rag.indexer.domain.model.RagFile
+    表名: t_rag_file
+    """
+    __tablename__ = "t_rag_file"
+    __ignore_data_scope__ = True
+
+    id = Column(String(36), primary_key=True, comment="RAG文件ID")
+    knowledge_base_id = Column(String(36), nullable=False, index=True, comment="知识库ID")
+    file_name = Column(String(512), nullable=False, comment="文件名")
+    file_id = Column(String(36), nullable=False, comment="原始文件ID")
+    chunk_count = Column(Integer, nullable=True, comment="分块数量")
+    file_metadata = Column("metadata", JSON, nullable=True, comment="元数据（JSON格式）")
+    status = Column(
+        String(50),
+        nullable=False,
+        default=FileStatus.UNPROCESSED,
+        comment="处理状态",
+    )
+    err_msg = Column(String(2048), nullable=True, comment="错误信息")
+    progress = Column(Integer, default=0, nullable=False, comment="处理进度(0-100)")
+
+    def __repr__(self):
+        return f"<RagFile(id={self.id}, file_name={self.file_name}, status={self.status})>"

runtime/deepdata-python/app/db/models/models.py

@@ -0,0 +1,46 @@
+from sqlalchemy import Boolean, Column, String, TIMESTAMP
+
+from app.db.models.base_entity import BaseEntity
+
+
+class Models(BaseEntity):
+    """模型配置表，对应表 t_models。模型为系统级配置，RAG/生成等按 ID 引用时不受数据权限过滤。
+
+    CREATE TABLE IF NOT EXISTS t_models (
+        id         VARCHAR(36) PRIMARY KEY COMMENT '主键ID',
+        model_name VARCHAR(100) NOT NULL COMMENT '模型名称（如 qwen2）',
+        provider   VARCHAR(50)  NOT NULL COMMENT '模型提供商（如 Ollama、OpenAI、DeepSeek）',
+        base_url   VARCHAR(255) NOT NULL COMMENT 'API 基础地址',
+        api_key    VARCHAR(512) DEFAULT '' COMMENT 'API 密钥（无密钥则为空）',
+        type       VARCHAR(50)  NOT NULL COMMENT '模型类型（如 chat、embedding）',
+        is_enabled TINYINT      DEFAULT 1 COMMENT '是否启用：1-启用，0-禁用',
+        is_default TINYINT      DEFAULT 0 COMMENT '是否默认：1-默认，0-非默认',
+        created_at TIMESTAMP    DEFAULT CURRENT_TIMESTAMP COMMENT '创建时间',
+        updated_at TIMESTAMP    DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP COMMENT '更新时间',
+        created_by VARCHAR(255) COMMENT '创建者',
+        updated_by VARCHAR(255) COMMENT '更新者',
+        UNIQUE KEY uk_model_provider (model_name, provider)
+    ) ENGINE = InnoDB DEFAULT CHARSET = utf8mb4 COMMENT ='模型配置表';
+    """
+
+    __tablename__ = "t_models"
+
+    id = Column(String(36), primary_key=True, index=True, comment="主键ID")
+    model_name = Column(String(100), nullable=False, comment="模型名称（如 qwen2）")
+    provider = Column(String(50), nullable=False, comment="模型提供商（如 Ollama、OpenAI、DeepSeek）")
+    base_url = Column(String(255), nullable=False, comment="API 基础地址")
+    api_key = Column(String(512), nullable=False, default="", comment="API 密钥（无密钥则为空）")
+    type = Column(String(50), nullable=False, comment="模型类型（如 chat、embedding）")
+
+    is_enabled = Column(Boolean, nullable=False, default=True, comment="是否启用")
+    is_default = Column(Boolean, nullable=False, default=False, comment="是否默认")
+    is_deleted = Column(Boolean, nullable=False, default=False, comment="是否删除")
+
+    __table_args__ = (
+        # 与 DDL 中的 uk_model_provider 保持一致
+        {
+            "mysql_engine": "InnoDB",
+            "mysql_charset": "utf8mb4",
+            "comment": "模型配置表",
+        },
+    )

runtime/deepdata-python/app/db/models/operator.py

@@ -0,0 +1,86 @@
+"""
+Operator Market Data Models
+算子市场数据模型
+"""
+
+from sqlalchemy import (
+    Column,
+    String,
+    Integer,
+    Boolean,
+    BigInteger,
+    Text,
+    JSON,
+    TIMESTAMP,
+    Index,
+)
+from sqlalchemy.sql import func
+
+from app.db.models.base_entity import Base, BaseEntity
+
+
+class Operator(BaseEntity):
+    """算子实体"""
+
+    __tablename__ = "t_operator"
+
+    id = Column(String(36), primary_key=True, index=True, comment="算子ID")
+    name = Column(String(255), nullable=False, comment="算子名称")
+    description = Column(Text, nullable=True, comment="算子描述")
+    version = Column(String(50), nullable=False, comment="算子版本")
+    inputs = Column(Text, nullable=True, comment="输入定义（JSON）")
+    outputs = Column(Text, nullable=True, comment="输出定义（JSON）")
+    runtime = Column(Text, nullable=True, comment="运行时配置（JSON）")
+    settings = Column(Text, nullable=True, comment="算子设置（JSON）")
+    file_name = Column(String(255), nullable=True, comment="文件名")
+    file_size = Column(BigInteger, nullable=True, comment="文件大小（字节）")
+    metrics = Column(Text, nullable=True, comment="算子指标（JSON）")
+    usage_count = Column(Integer, default=0, nullable=False, comment="使用次数")
+    is_star = Column(Boolean, default=False, nullable=False, comment="是否收藏")
+
+    __table_args__ = (Index("idx_is_star", "is_star"),)
+
+
+class Category(BaseEntity):
+    """算子分类实体"""
+
+    __tablename__ = "t_operator_category"
+
+    id = Column(String(36), primary_key=True, index=True, comment="分类ID")
+    name = Column(String(255), nullable=False, comment="分类名称(中文)")
+    name_en = Column(String(255), nullable=True, comment="分类名称(英文)")
+    value = Column(String(255), nullable=True, comment="分类值")
+    type = Column(String(50), nullable=True, comment="分类类型")
+    parent_id = Column(String(36), nullable=False, default="0", comment="父分类ID")
+
+
+class CategoryRelation(BaseEntity):
+    """算子分类关系实体"""
+
+    __tablename__ = "t_operator_category_relation"
+
+    category_id = Column(String(36), primary_key=True, comment="分类ID")
+    operator_id = Column(String(36), primary_key=True, comment="算子ID")
+
+    __table_args__ = (
+        Index("idx_category_id", "category_id"),
+        Index("idx_operator_id", "operator_id"),
+    )
+
+
+class OperatorRelease(BaseEntity):
+    """算子发布版本实体"""
+
+    __tablename__ = "t_operator_release"
+
+    id = Column(String(36), primary_key=True, comment="算子ID")
+    version = Column(String(50), primary_key=True, comment="版本号")
+    release_date = Column(
+        TIMESTAMP, nullable=False, default=func.now(), comment="发布时间"
+    )
+    changelog = Column(JSON, nullable=True, comment="更新日志列表")
+
+
+# Ignore data scope for operator models
+for model in [Operator, Category, CategoryRelation, OperatorRelease]:
+    model.__ignore_data_scope__ = True

runtime/deepdata-python/app/db/models/ratio_task.py

@@ -0,0 +1,62 @@
+"""
+Tables for Ratio (Data Synthesis Ratio) module
+
+Derived from scripts/db/data-ratio-init.sql
+ - t_st_ratio_instances
+ - t_st_ratio_relations
+"""
+
+import uuid
+from sqlalchemy import Column, String, Text, BigInteger, TIMESTAMP, JSON, ForeignKey
+from sqlalchemy.orm import relationship
+from sqlalchemy.sql import func
+
+from app.db.models.base_entity import BaseEntity
+
+
+class RatioInstance(BaseEntity):
+    """配比实例表（UUID 主键） -> t_st_ratio_instances
+
+    Columns per data-ratio-init.sql:
+      id, name, description, target_dataset_id, ratio_method, ratio_parameters,
+      merge_method, status, totals, created_at, updated_at, created_by, updated_by
+    """
+
+    __tablename__ = "t_st_ratio_instances"
+
+    id = Column(String(64), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
+    name = Column(String(64), nullable=True, comment="名称")
+    description = Column(Text, nullable=True, comment="描述")
+    target_dataset_id = Column(String(64), nullable=True, comment="模板数据集ID")
+    ratio_parameters = Column(JSON, nullable=True, comment="配比参数")
+    merge_method = Column(String(50), nullable=True, comment="合并方式")
+    status = Column(String(20), nullable=True, comment="状态")
+    totals = Column(BigInteger, nullable=True, comment="总数")
+
+    def __repr__(self) -> str:
+        return f"<RatioInstance(id={self.id}, name={self.name}, status={self.status})>"
+
+
+class RatioRelation(BaseEntity):
+    """配比关系表（UUID 主键） -> t_st_ratio_relations
+
+    Columns per data-ratio-init.sql:
+      id, ratio_instance_id, source_dataset_id, ratio_value, counts, filter_conditions,
+      created_at, updated_at, created_by, updated_by
+    """
+
+    __tablename__ = "t_st_ratio_relations"
+
+    id = Column(String(64), primary_key=True, default=lambda: str(uuid.uuid4()), comment="UUID")
+    ratio_instance_id = Column(String(64), nullable=False, comment="配比实例ID")
+    source_dataset_id = Column(String(64), nullable=True, comment="源数据集ID")
+    ratio_value = Column(String(256), nullable=True)
+    counts = Column(BigInteger, nullable=True, comment="条数")
+    filter_conditions = Column(Text, nullable=True, comment="过滤条件")
+
+    def __repr__(self) -> str:
+        return (
+            f"<RatioRelation(id={self.id}, ratio_instance_id={self.ratio_instance_id}, "
+            f"source_dataset_id={self.source_dataset_id}, counts={self.counts})>"
+        )
+

runtime/deepdata-python/app/db/models/sys_param.py

@@ -0,0 +1,30 @@
+"""
+System Parameter Model
+系统参数模型
+"""
+
+from sqlalchemy import Column, String, Boolean, Text
+from app.db.models.base_entity import BaseEntity
+
+
+class SysParam(BaseEntity):
+    """系统参数实体"""
+
+    __tablename__ = "t_sys_param"
+    __ignore_data_scope__ = True  # 系统参数不进行数据权限过滤
+
+    id = Column(String(100), primary_key=True, comment="参数ID")
+    param_value = Column(Text, nullable=False, comment="参数值")
+    param_type = Column(
+        String(50),
+        nullable=False,
+        default="string",
+        comment="参数类型（string、integer、boolean）",
+    )
+    option_list = Column(
+        Text, nullable=True, comment="选项列表（JSON格式，仅对enum类型有效）"
+    )
+    description = Column(String(255), nullable=True, default="", comment="参数描述")
+    is_built_in = Column(Boolean, nullable=False, default=False, comment="是否内置")
+    can_modify = Column(Boolean, nullable=False, default=True, comment="是否可修改")
+    is_enabled = Column(Boolean, nullable=False, default=True, comment="是否启用")

runtime/deepdata-python/app/db/models/user_management.py

@@ -0,0 +1,29 @@
+"""
+Tables of User Management Module
+"""
+
+from sqlalchemy import Column, String, BigInteger, Boolean, TIMESTAMP
+from sqlalchemy.sql import func
+
+from app.db.models.base_entity import Base
+
+class User(Base):
+    """用户模型"""
+
+    __tablename__ = "users"
+
+    id = Column(BigInteger, primary_key=True, autoincrement=True, comment="用户ID")
+    username = Column(String(255), nullable=False, unique=True, comment="用户名")
+    email = Column(String(255), nullable=False, unique=True, comment="邮箱")
+    password_hash = Column(String(255), nullable=False, comment="密码哈希")
+    full_name = Column(String(255), nullable=True, comment="真实姓名")
+    avatar_url = Column(String(500), nullable=True, comment="头像URL")
+    role = Column(String(50), nullable=False, default='USER', comment="角色：ADMIN/USER")
+    organization = Column(String(255), nullable=True, comment="所属机构")
+    enabled = Column(Boolean, nullable=False, default=True, comment="是否启用")
+    last_login_at = Column(TIMESTAMP, nullable=True, comment="最后登录时间")
+    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"<User(id={self.id}, username={self.username}, role={self.role})>"

runtime/deepdata-python/app/db/session.py

@@ -0,0 +1,94 @@
+from datetime import datetime
+from sqlalchemy import event, true
+from sqlalchemy.ext.asyncio import AsyncSession, create_async_engine, async_sessionmaker
+from sqlalchemy.orm import Session, declarative_base, with_loader_criteria
+from sqlalchemy.orm.attributes import flag_modified
+from app.core.config import settings
+from app.core.logging import get_logger
+from typing import AsyncGenerator
+
+from app.db.datascope import DataScopeHandle, SYSTEM_USER
+from app.db.models.base_entity import BaseEntity
+
+logger = get_logger(__name__)
+
+# 创建数据库引擎
+engine = create_async_engine(
+    settings.database_url,
+    echo=False,  # 关闭SQL调试日志以减少输出
+    future=True,
+    pool_pre_ping=True,
+    pool_recycle=3600
+)
+
+# 创建会话工厂
+AsyncSessionLocal = async_sessionmaker(
+    engine,
+    class_=AsyncSession,
+    expire_on_commit=False
+)
+
+@event.listens_for(Session, "do_orm_execute")
+def _apply_data_scope(orm_execute_state):
+    # only act on SELECT queries generated by ORM
+    if not orm_execute_state.is_select:
+        return
+
+    allowed = DataScopeHandle.allowed_users()
+    if not allowed or len(allowed) == 0:
+        return
+
+    # predicate builder: return None to skip for classes that opt-out
+    def criteria_fn(cls):
+        # skip if the mapped class explicitly disables data-scope
+        if getattr(cls, "__ignore_data_scope__", False):
+            return true()
+        # some classes may not have created_by column; guard dynamically
+        col = getattr(cls, "created_by", None)
+        if col is None:
+            return true()
+        return col.in_(allowed)
+
+    # apply loader-level criteria to all subclasses of BaseEntity
+    orm_execute_state.statement = orm_execute_state.statement.options(
+        with_loader_criteria(BaseEntity, criteria_fn, include_aliases=True)
+    )
+
+@event.listens_for(Session, "before_flush")
+def _audit_before_flush(session, flush_context, instances):
+    user = DataScopeHandle.get_user_info()
+    effective_user = (user or "").strip() or SYSTEM_USER
+    now = datetime.now()
+
+    # new -> set created_* and updated_*
+    for obj in list(session.new):
+        if isinstance(obj, BaseEntity):
+            if getattr(obj, "created_at", None) is None:
+                obj.created_at = now
+            if getattr(obj, "created_by", None) is None:
+                obj.created_by = effective_user
+            obj.updated_at = now
+            obj.updated_by = effective_user
+            # ensure SQLAlchemy sees changes
+            try:
+                flag_modified(obj, "created_by")
+            except Exception:
+                pass
+
+    # dirty -> set updated_*
+    for obj in list(session.dirty):
+        if isinstance(obj, BaseEntity):
+            obj.updated_at = now
+            obj.updated_by = effective_user
+            try:
+                flag_modified(obj, "updated_by")
+            except Exception:
+                pass
+
+async def get_db() -> AsyncGenerator[AsyncSession, None]:
+    """获取数据库会话"""
+    async with AsyncSessionLocal() as session:
+        try:
+            yield session
+        finally:
+            await session.close()