Merge 7a1b498c0a into 446df6b50d

api/core/rag/docstore/dataset_docstore.py

@@ -242,6 +242,7 @@ class DatasetDocumentStore:
 
     def add_multimodel_documents_binding(self, segment_id: str, multimodel_documents: list[AttachmentDocument] | None):
         if multimodel_documents:
+            assert self._document_id is not None
             for multimodel_document in multimodel_documents:
                 binding = SegmentAttachmentBinding(
                     tenant_id=self._dataset.tenant_id,

api/models/dataset.py

@@ -1547,7 +1547,7 @@ class PipelineRecommendedPlugin(TypeBase):
     )
 
 
-class SegmentAttachmentBinding(Base):
+class SegmentAttachmentBinding(TypeBase):
     __tablename__ = "segment_attachment_bindings"
     __table_args__ = (
         sa.PrimaryKeyConstraint("id", name="segment_attachment_binding_pkey"),
@@ -1560,10 +1560,12 @@ class SegmentAttachmentBinding(Base):
         ),
         sa.Index("segment_attachment_binding_attachment_idx", "attachment_id"),
     )
-    id: Mapped[str] = mapped_column(StringUUID, default=lambda: str(uuidv7()))
+    id: Mapped[str] = mapped_column(StringUUID, default=lambda: str(uuidv7()), init=False)
     tenant_id: Mapped[str] = mapped_column(StringUUID, nullable=False)
     dataset_id: Mapped[str] = mapped_column(StringUUID, nullable=False)
     document_id: Mapped[str] = mapped_column(StringUUID, nullable=False)
     segment_id: Mapped[str] = mapped_column(StringUUID, nullable=False)
     attachment_id: Mapped[str] = mapped_column(StringUUID, nullable=False)
-    created_at: Mapped[datetime] = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    created_at: Mapped[datetime] = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), init=False
+    )

api/models/model.py

@@ -1711,7 +1711,7 @@ class ApiToken(Base):  # bug: this uses setattr so idk the field.
             return result
 
 
-class UploadFile(Base):
+class UploadFile(TypeBase):
     __tablename__ = "upload_files"
     __table_args__ = (
         sa.PrimaryKeyConstraint("id", name="upload_file_pkey"),
@@ -1721,7 +1721,7 @@ class UploadFile(Base):
     # NOTE: The `id` field is generated within the application to minimize extra roundtrips
     # (especially when generating `source_url`).
     # The `server_default` serves as a fallback mechanism.
-    id: Mapped[str] = mapped_column(StringUUID, default=lambda: str(uuid4()))
+    id: Mapped[str] = mapped_column(StringUUID, default=lambda: str(uuid4()), init=False)
     tenant_id: Mapped[str] = mapped_column(StringUUID, nullable=False)
     storage_type: Mapped[str] = mapped_column(String(255), nullable=False)
     key: Mapped[str] = mapped_column(String(255), nullable=False)
@@ -1732,7 +1732,9 @@ class UploadFile(Base):
 
     # The `created_by_role` field indicates whether the file was created by an `Account` or an `EndUser`.
     # Its value is derived from the `CreatorUserRole` enumeration.
-    created_by_role: Mapped[str] = mapped_column(String(255), nullable=False, server_default=sa.text("'account'"))
+    created_by_role: Mapped[CreatorUserRole] = mapped_column(
+        String(255), nullable=False, server_default=sa.text("'account'")
+    )
 
     # The `created_by` field stores the ID of the entity that created this upload file.
     #
@@ -1753,47 +1755,11 @@ class UploadFile(Base):
     used: Mapped[bool] = mapped_column(sa.Boolean, nullable=False, server_default=sa.text("false"))
 
     # `used_by` may indicate the ID of the user who utilized this file.
-    used_by: Mapped[str | None] = mapped_column(StringUUID, nullable=True)
-    used_at: Mapped[datetime | None] = mapped_column(sa.DateTime, nullable=True)
-    hash: Mapped[str | None] = mapped_column(String(255), nullable=True)
+    used_by: Mapped[str | None] = mapped_column(StringUUID, nullable=True, default=None)
+    used_at: Mapped[datetime | None] = mapped_column(sa.DateTime, nullable=True, default=None)
+    hash: Mapped[str | None] = mapped_column(String(255), nullable=True, default=None)
     source_url: Mapped[str] = mapped_column(LongText, default="")
 
-    def __init__(
-        self,
-        *,
-        tenant_id: str,
-        storage_type: str,
-        key: str,
-        name: str,
-        size: int,
-        extension: str,
-        mime_type: str,
-        created_by_role: CreatorUserRole,
-        created_by: str,
-        created_at: datetime,
-        used: bool,
-        used_by: str | None = None,
-        used_at: datetime | None = None,
-        hash: str | None = None,
-        source_url: str = "",
-    ):
-        self.id = str(uuid.uuid4())
-        self.tenant_id = tenant_id
-        self.storage_type = storage_type
-        self.key = key
-        self.name = name
-        self.size = size
-        self.extension = extension
-        self.mime_type = mime_type
-        self.created_by_role = created_by_role.value
-        self.created_by = created_by
-        self.created_at = created_at
-        self.used = used
-        self.used_by = used_by
-        self.used_at = used_at
-        self.hash = hash
-        self.source_url = source_url
-
 
 class ApiRequest(TypeBase):
     __tablename__ = "api_requests"

api/models/workflow.py

@@ -39,8 +39,7 @@ from libs.uuid_utils import uuidv7
 from ._workflow_exc import NodeNotFoundError, WorkflowDataError
 
 if TYPE_CHECKING:
-    from .model import AppMode, UploadFile
-
+    from .model import AppMode
 
 from constants import DEFAULT_FILE_NUMBER_LIMITS, HIDDEN_VALUE
 from core.helper import encrypter
@@ -52,6 +51,7 @@ from .account import Account
 from .base import Base, DefaultFieldsMixin, TypeBase
 from .engine import db
 from .enums import CreatorUserRole, DraftVariableType, ExecutionOffLoadType
+from .model import UploadFile
 from .types import EnumText, LongText, StringUUID
 
 logger = logging.getLogger(__name__)