fix: update request parameters in EnterpriseService for DELETE operation

- Changed the request body to use query parameters instead of JSON body for the DELETE request to /webapp/clean.

api/.env.example

@@ -371,6 +371,12 @@ UPLOAD_IMAGE_FILE_SIZE_LIMIT=10
 UPLOAD_VIDEO_FILE_SIZE_LIMIT=100
 UPLOAD_AUDIO_FILE_SIZE_LIMIT=50
 
+# Comma-separated list of file extensions blocked from upload for security reasons.
+# Extensions should be lowercase without dots (e.g., exe,bat,sh,dll).
+# Empty by default to allow all file types.
+# Recommended: exe,bat,cmd,com,scr,vbs,ps1,msi,dll
+UPLOAD_FILE_EXTENSION_BLACKLIST=
+
 # Model configuration
 MULTIMODAL_SEND_FORMAT=base64
 PROMPT_GENERATION_MAX_TOKENS=512
@@ -608,3 +614,9 @@ SWAGGER_UI_PATH=/swagger-ui.html
 # Whether to encrypt dataset IDs when exporting DSL files (default: true)
 # Set to false to export dataset IDs as plain text for easier cross-environment import
 DSL_EXPORT_ENCRYPT_DATASET_ID=true
+
+# Tenant isolated task queue configuration
+TENANT_ISOLATED_TASK_CONCURRENCY=1
+
+# Maximum number of segments for dataset segments API (0 for unlimited)
+DATASET_MAX_SEGMENTS_PER_REQUEST=0

api/Dockerfile

@@ -15,7 +15,11 @@ FROM base AS packages
 # RUN sed -i 's@deb.debian.org@mirrors.aliyun.com@g' /etc/apt/sources.list.d/debian.sources
 
 RUN apt-get update \
-    && apt-get install -y --no-install-recommends gcc g++ libc-dev libffi-dev libgmp-dev libmpfr-dev libmpc-dev
+    && apt-get install -y --no-install-recommends \
+          # basic environment
+          g++ \
+          # for building gmpy2
+          libmpfr-dev libmpc-dev
 
 # Install Python dependencies
 COPY pyproject.toml uv.lock ./
@@ -49,9 +53,11 @@ RUN \
     # Install dependencies
     && apt-get install -y --no-install-recommends \
         # basic environment
-        curl nodejs libgmp-dev libmpfr-dev libmpc-dev \
+        curl nodejs \
+        # for gmpy2 \
+        libgmp-dev libmpfr-dev libmpc-dev \
         # For Security
-        expat libldap-2.5-0 perl libsqlite3-0 zlib1g \
+        expat libldap-2.5-0=2.5.13+dfsg-5 perl libsqlite3-0=3.40.1-2+deb12u2 zlib1g=1:1.2.13.dfsg-1 \
         # install fonts to support the use of tools like pypdfium2
         fonts-noto-cjk \
         # install a package to improve the accuracy of guessing mime type and file extension

api/commands.py

@@ -1422,7 +1422,10 @@ def setup_datasource_oauth_client(provider, client_params):
 
 
 @click.command("transform-datasource-credentials", help="Transform datasource credentials.")
-def transform_datasource_credentials():
+@click.option(
+    "--environment", prompt=True, help="the environment to transform datasource credentials", default="online"
+)
+def transform_datasource_credentials(environment: str):
     """
     Transform datasource credentials
     """
@@ -1433,9 +1436,14 @@ def transform_datasource_credentials():
         notion_plugin_id = "langgenius/notion_datasource"
         firecrawl_plugin_id = "langgenius/firecrawl_datasource"
         jina_plugin_id = "langgenius/jina_datasource"
-        notion_plugin_unique_identifier = plugin_migration._fetch_plugin_unique_identifier(notion_plugin_id)  # pyright: ignore[reportPrivateUsage]
-        firecrawl_plugin_unique_identifier = plugin_migration._fetch_plugin_unique_identifier(firecrawl_plugin_id)  # pyright: ignore[reportPrivateUsage]
-        jina_plugin_unique_identifier = plugin_migration._fetch_plugin_unique_identifier(jina_plugin_id)  # pyright: ignore[reportPrivateUsage]
+        if environment == "online":
+            notion_plugin_unique_identifier = plugin_migration._fetch_plugin_unique_identifier(notion_plugin_id)  # pyright: ignore[reportPrivateUsage]
+            firecrawl_plugin_unique_identifier = plugin_migration._fetch_plugin_unique_identifier(firecrawl_plugin_id)  # pyright: ignore[reportPrivateUsage]
+            jina_plugin_unique_identifier = plugin_migration._fetch_plugin_unique_identifier(jina_plugin_id)  # pyright: ignore[reportPrivateUsage]
+        else:
+            notion_plugin_unique_identifier = None
+            firecrawl_plugin_unique_identifier = None
+            jina_plugin_unique_identifier = None
         oauth_credential_type = CredentialType.OAUTH2
         api_key_credential_type = CredentialType.API_KEY
 
@@ -1601,7 +1609,7 @@ def transform_datasource_credentials():
                             "integration_secret": api_key,
                         }
                         datasource_provider = DatasourceProvider(
-                            provider="jina",
+                            provider="jinareader",
                             tenant_id=tenant_id,
                             plugin_id=jina_plugin_id,
                             auth_type=api_key_credential_type.value,

api/configs/feature/__init__.py

@@ -331,6 +331,31 @@ class FileUploadConfig(BaseSettings):
         default=10,
     )
 
+    inner_UPLOAD_FILE_EXTENSION_BLACKLIST: str = Field(
+        description=(
+            "Comma-separated list of file extensions that are blocked from upload. "
+            "Extensions should be lowercase without dots (e.g., 'exe,bat,sh,dll'). "
+            "Empty by default to allow all file types."
+        ),
+        validation_alias=AliasChoices("UPLOAD_FILE_EXTENSION_BLACKLIST"),
+        default="",
+    )
+
+    @computed_field  # type: ignore[misc]
+    @property
+    def UPLOAD_FILE_EXTENSION_BLACKLIST(self) -> set[str]:
+        """
+        Parse and return the blacklist as a set of lowercase extensions.
+        Returns an empty set if no blacklist is configured.
+        """
+        if not self.inner_UPLOAD_FILE_EXTENSION_BLACKLIST:
+            return set()
+        return {
+            ext.strip().lower().strip(".")
+            for ext in self.inner_UPLOAD_FILE_EXTENSION_BLACKLIST.split(",")
+            if ext.strip()
+        }
+
 
 class HttpConfig(BaseSettings):
     """
@@ -920,6 +945,11 @@ class DataSetConfig(BaseSettings):
         default=True,
     )
 
+    DATASET_MAX_SEGMENTS_PER_REQUEST: NonNegativeInt = Field(
+        description="Maximum number of segments for dataset segments API (0 for unlimited)",
+        default=0,
+    )
+
 
 class WorkspaceConfig(BaseSettings):
     """
@@ -1112,6 +1142,13 @@ class SwaggerUIConfig(BaseSettings):
     )
 
 
+class TenantIsolatedTaskQueueConfig(BaseSettings):
+    TENANT_ISOLATED_TASK_CONCURRENCY: int = Field(
+        description="Number of tasks allowed to be delivered concurrently from isolated queue per tenant",
+        default=1,
+    )
+
+
 class FeatureConfig(
     # place the configs in alphabet order
     AppExecutionConfig,
@@ -1136,6 +1173,7 @@ class FeatureConfig(
     RagEtlConfig,
     RepositoryConfig,
     SecurityConfig,
+    TenantIsolatedTaskQueueConfig,
     ToolConfig,
     UpdateConfig,
     WorkflowConfig,

api/configs/middleware/vdb/weaviate_config.py

@@ -22,6 +22,11 @@ class WeaviateConfig(BaseSettings):
         default=True,
     )
 
+    WEAVIATE_GRPC_ENDPOINT: str | None = Field(
+        description="URL of the Weaviate gRPC server (e.g., 'grpc://localhost:50051' or 'grpcs://weaviate.example.com:443')",
+        default=None,
+    )
+
     WEAVIATE_BATCH_SIZE: PositiveInt = Field(
         description="Number of objects to be processed in a single batch operation (default is 100)",
         default=100,

api/controllers/common/errors.py

@@ -25,6 +25,12 @@ class UnsupportedFileTypeError(BaseHTTPException):
     code = 415
 
 
+class BlockedFileExtensionError(BaseHTTPException):
+    error_code = "file_extension_blocked"
+    description = "The file extension is blocked for security reasons."
+    code = 400
+
+
 class TooManyFilesError(BaseHTTPException):
     error_code = "too_many_files"
     description = "Only one file is allowed."

api/controllers/console/app/annotation.py

@@ -16,6 +16,7 @@ from fields.annotation_fields import (
     annotation_fields,
     annotation_hit_history_fields,
 )
+from libs.helper import uuid_value
 from libs.login import login_required
 from services.annotation_service import AppAnnotationService
 
@@ -175,8 +176,10 @@ class AnnotationApi(Resource):
         api.model(
             "CreateAnnotationRequest",
             {
-                "question": fields.String(required=True, description="Question text"),
-                "answer": fields.String(required=True, description="Answer text"),
+                "message_id": fields.String(description="Message ID (optional)"),
+                "question": fields.String(description="Question text (required when message_id not provided)"),
+                "answer": fields.String(description="Answer text (use 'answer' or 'content')"),
+                "content": fields.String(description="Content text (use 'answer' or 'content')"),
                 "annotation_reply": fields.Raw(description="Annotation reply data"),
             },
         )
@@ -193,11 +196,14 @@ class AnnotationApi(Resource):
         app_id = str(app_id)
         parser = (
             reqparse.RequestParser()
-            .add_argument("question", required=True, type=str, location="json")
-            .add_argument("answer", required=True, type=str, location="json")
+            .add_argument("message_id", required=False, type=uuid_value, location="json")
+            .add_argument("question", required=False, type=str, location="json")
+            .add_argument("answer", required=False, type=str, location="json")
+            .add_argument("content", required=False, type=str, location="json")
+            .add_argument("annotation_reply", required=False, type=dict, location="json")
         )
         args = parser.parse_args()
-        annotation = AppAnnotationService.insert_app_annotation_directly(args, app_id)
+        annotation = AppAnnotationService.up_insert_app_annotation_from_message(args, app_id)
         return annotation
 
     @setup_required

api/controllers/console/app/conversation.py

@@ -1,7 +1,5 @@
-from datetime import datetime
-
-import pytz
 import sqlalchemy as sa
+from flask import abort
 from flask_restx import Resource, marshal_with, reqparse
 from flask_restx.inputs import int_range
 from sqlalchemy import func, or_
@@ -19,7 +17,7 @@ from fields.conversation_fields import (
     conversation_pagination_fields,
     conversation_with_summary_pagination_fields,
 )
-from libs.datetime_utils import naive_utc_now
+from libs.datetime_utils import naive_utc_now, parse_time_range
 from libs.helper import DatetimeString
 from libs.login import current_account_with_tenant, login_required
 from models import Conversation, EndUser, Message, MessageAnnotation
@@ -90,25 +88,17 @@ class CompletionConversationApi(Resource):
 
         account = current_user
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
 
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_datetime_utc = start_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_datetime_utc, end_datetime_utc = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
+        if start_datetime_utc:
             query = query.where(Conversation.created_at >= start_datetime_utc)
 
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=59)
-
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_datetime_utc = end_datetime_timezone.astimezone(utc_timezone)
-
+        if end_datetime_utc:
+            end_datetime_utc = end_datetime_utc.replace(second=59)
             query = query.where(Conversation.created_at < end_datetime_utc)
 
         # FIXME, the type ignore in this file
@@ -270,29 +260,21 @@ class ChatConversationApi(Resource):
 
         account = current_user
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
 
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_datetime_utc = start_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_datetime_utc, end_datetime_utc = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
+        if start_datetime_utc:
             match args["sort_by"]:
                 case "updated_at" | "-updated_at":
                     query = query.where(Conversation.updated_at >= start_datetime_utc)
                 case "created_at" | "-created_at" | _:
                     query = query.where(Conversation.created_at >= start_datetime_utc)
 
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=59)
-
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_datetime_utc = end_datetime_timezone.astimezone(utc_timezone)
-
+        if end_datetime_utc:
+            end_datetime_utc = end_datetime_utc.replace(second=59)
             match args["sort_by"]:
                 case "updated_at" | "-updated_at":
                     query = query.where(Conversation.updated_at <= end_datetime_utc)

api/controllers/console/app/message.py

@@ -16,7 +16,6 @@ from controllers.console.app.wraps import get_app_model
 from controllers.console.explore.error import AppSuggestedQuestionsAfterAnswerDisabledError
 from controllers.console.wraps import (
     account_initialization_required,
-    cloud_edition_billing_resource_check,
     edit_permission_required,
     setup_required,
 )
@@ -24,12 +23,11 @@ from core.app.entities.app_invoke_entities import InvokeFrom
 from core.errors.error import ModelCurrentlyNotSupportError, ProviderTokenNotInitError, QuotaExceededError
 from core.model_runtime.errors.invoke import InvokeError
 from extensions.ext_database import db
-from fields.conversation_fields import annotation_fields, message_detail_fields
+from fields.conversation_fields import message_detail_fields
 from libs.helper import uuid_value
 from libs.infinite_scroll_pagination import InfiniteScrollPagination
 from libs.login import current_account_with_tenant, login_required
 from models.model import AppMode, Conversation, Message, MessageAnnotation, MessageFeedback
-from services.annotation_service import AppAnnotationService
 from services.errors.conversation import ConversationNotExistsError
 from services.errors.message import MessageNotExistsError, SuggestedQuestionsAfterAnswerDisabledError
 from services.message_service import MessageService
@@ -194,45 +192,6 @@ class MessageFeedbackApi(Resource):
         return {"result": "success"}
 
 
-@console_ns.route("/apps/<uuid:app_id>/annotations")
-class MessageAnnotationApi(Resource):
-    @api.doc("create_message_annotation")
-    @api.doc(description="Create message annotation")
-    @api.doc(params={"app_id": "Application ID"})
-    @api.expect(
-        api.model(
-            "MessageAnnotationRequest",
-            {
-                "message_id": fields.String(description="Message ID"),
-                "question": fields.String(required=True, description="Question text"),
-                "answer": fields.String(required=True, description="Answer text"),
-                "annotation_reply": fields.Raw(description="Annotation reply"),
-            },
-        )
-    )
-    @api.response(200, "Annotation created successfully", annotation_fields)
-    @api.response(403, "Insufficient permissions")
-    @marshal_with(annotation_fields)
-    @get_app_model
-    @setup_required
-    @login_required
-    @cloud_edition_billing_resource_check("annotation")
-    @account_initialization_required
-    @edit_permission_required
-    def post(self, app_model):
-        parser = (
-            reqparse.RequestParser()
-            .add_argument("message_id", required=False, type=uuid_value, location="json")
-            .add_argument("question", required=True, type=str, location="json")
-            .add_argument("answer", required=True, type=str, location="json")
-            .add_argument("annotation_reply", required=False, type=dict, location="json")
-        )
-        args = parser.parse_args()
-        annotation = AppAnnotationService.up_insert_app_annotation_from_message(args, app_model.id)
-
-        return annotation
-
-
 @console_ns.route("/apps/<uuid:app_id>/annotations/count")
 class MessageAnnotationCountApi(Resource):
     @api.doc("get_annotation_count")

api/controllers/console/app/statistic.py

@@ -1,9 +1,7 @@
-from datetime import datetime
 from decimal import Decimal
 
-import pytz
 import sqlalchemy as sa
-from flask import jsonify
+from flask import abort, jsonify
 from flask_restx import Resource, fields, reqparse
 
 from controllers.console import api, console_ns
@@ -11,6 +9,7 @@ from controllers.console.app.wraps import get_app_model
 from controllers.console.wraps import account_initialization_required, setup_required
 from core.app.entities.app_invoke_entities import InvokeFrom
 from extensions.ext_database import db
+from libs.datetime_utils import parse_time_range
 from libs.helper import DatetimeString
 from libs.login import current_account_with_tenant, login_required
 from models import AppMode, Message
@@ -56,26 +55,16 @@ WHERE
         arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
 
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
-
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_datetime_utc = start_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_datetime_utc, end_datetime_utc = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
+        if start_datetime_utc:
             sql_query += " AND created_at >= :start"
             arg_dict["start"] = start_datetime_utc
 
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=0)
-
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_datetime_utc = end_datetime_timezone.astimezone(utc_timezone)
-
+        if end_datetime_utc:
             sql_query += " AND created_at < :end"
             arg_dict["end"] = end_datetime_utc
 
@@ -120,8 +109,11 @@ class DailyConversationStatistic(Resource):
         )
         args = parser.parse_args()
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
+
+        try:
+            start_datetime_utc, end_datetime_utc = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
         stmt = (
             sa.select(
@@ -134,18 +126,10 @@ class DailyConversationStatistic(Resource):
             .where(Message.app_id == app_model.id, Message.invoke_from != InvokeFrom.DEBUGGER)
         )
 
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_datetime_utc = start_datetime_timezone.astimezone(utc_timezone)
+        if start_datetime_utc:
             stmt = stmt.where(Message.created_at >= start_datetime_utc)
 
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=0)
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_datetime_utc = end_datetime_timezone.astimezone(utc_timezone)
+        if end_datetime_utc:
             stmt = stmt.where(Message.created_at < end_datetime_utc)
 
         stmt = stmt.group_by("date").order_by("date")
@@ -198,26 +182,17 @@ WHERE
     AND invoke_from != :invoke_from"""
         arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
 
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_datetime_utc = start_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_datetime_utc, end_datetime_utc = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
+        if start_datetime_utc:
             sql_query += " AND created_at >= :start"
             arg_dict["start"] = start_datetime_utc
 
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=0)
-
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_datetime_utc = end_datetime_timezone.astimezone(utc_timezone)
-
+        if end_datetime_utc:
             sql_query += " AND created_at < :end"
             arg_dict["end"] = end_datetime_utc
 
@@ -273,26 +248,17 @@ WHERE
     AND invoke_from != :invoke_from"""
         arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
 
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_datetime_utc = start_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_datetime_utc, end_datetime_utc = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
+        if start_datetime_utc:
             sql_query += " AND created_at >= :start"
             arg_dict["start"] = start_datetime_utc
 
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=0)
-
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_datetime_utc = end_datetime_timezone.astimezone(utc_timezone)
-
+        if end_datetime_utc:
             sql_query += " AND created_at < :end"
             arg_dict["end"] = end_datetime_utc
 
@@ -357,26 +323,17 @@ FROM
             AND m.invoke_from != :invoke_from"""
         arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
 
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_datetime_utc = start_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_datetime_utc, end_datetime_utc = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
+        if start_datetime_utc:
             sql_query += " AND c.created_at >= :start"
             arg_dict["start"] = start_datetime_utc
 
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=0)
-
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_datetime_utc = end_datetime_timezone.astimezone(utc_timezone)
-
+        if end_datetime_utc:
             sql_query += " AND c.created_at < :end"
             arg_dict["end"] = end_datetime_utc
 
@@ -446,26 +403,17 @@ WHERE
     AND m.invoke_from != :invoke_from"""
         arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
 
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_datetime_utc = start_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_datetime_utc, end_datetime_utc = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
+        if start_datetime_utc:
             sql_query += " AND m.created_at >= :start"
             arg_dict["start"] = start_datetime_utc
 
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=0)
-
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_datetime_utc = end_datetime_timezone.astimezone(utc_timezone)
-
+        if end_datetime_utc:
             sql_query += " AND m.created_at < :end"
             arg_dict["end"] = end_datetime_utc
 
@@ -525,26 +473,17 @@ WHERE
     AND invoke_from != :invoke_from"""
         arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
 
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_datetime_utc = start_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_datetime_utc, end_datetime_utc = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
+        if start_datetime_utc:
             sql_query += " AND created_at >= :start"
             arg_dict["start"] = start_datetime_utc
 
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=0)
-
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_datetime_utc = end_datetime_timezone.astimezone(utc_timezone)
-
+        if end_datetime_utc:
             sql_query += " AND created_at < :end"
             arg_dict["end"] = end_datetime_utc
 
@@ -602,26 +541,17 @@ WHERE
     AND invoke_from != :invoke_from"""
         arg_dict = {"tz": account.timezone, "app_id": app_model.id, "invoke_from": InvokeFrom.DEBUGGER}
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
 
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_datetime_utc = start_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_datetime_utc, end_datetime_utc = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
+        if start_datetime_utc:
             sql_query += " AND created_at >= :start"
             arg_dict["start"] = start_datetime_utc
 
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=0)
-
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_datetime_utc = end_datetime_timezone.astimezone(utc_timezone)
-
+        if end_datetime_utc:
             sql_query += " AND created_at < :end"
             arg_dict["end"] = end_datetime_utc
 

api/controllers/console/app/workflow.py

@@ -102,7 +102,18 @@ class DraftWorkflowApi(Resource):
             },
         )
     )
-    @api.response(200, "Draft workflow synced successfully", workflow_fields)
+    @api.response(
+        200,
+        "Draft workflow synced successfully",
+        api.model(
+            "SyncDraftWorkflowResponse",
+            {
+                "result": fields.String,
+                "hash": fields.String,
+                "updated_at": fields.String,
+            },
+        ),
+    )
     @api.response(400, "Invalid workflow configuration")
     @api.response(403, "Permission denied")
     @edit_permission_required

api/controllers/console/app/workflow_statistic.py

@@ -1,7 +1,4 @@
-from datetime import datetime
-
-import pytz
-from flask import jsonify
+from flask import abort, jsonify
 from flask_restx import Resource, reqparse
 from sqlalchemy.orm import sessionmaker
 
@@ -9,6 +6,7 @@ from controllers.console import api, console_ns
 from controllers.console.app.wraps import get_app_model
 from controllers.console.wraps import account_initialization_required, setup_required
 from extensions.ext_database import db
+from libs.datetime_utils import parse_time_range
 from libs.helper import DatetimeString
 from libs.login import current_account_with_tenant, login_required
 from models.enums import WorkflowRunTriggeredFrom
@@ -43,23 +41,11 @@ class WorkflowDailyRunsStatistic(Resource):
         args = parser.parse_args()
 
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
 
-        start_date = None
-        end_date = None
-
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_date = start_datetime_timezone.astimezone(utc_timezone)
-
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=0)
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_date = end_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_date, end_date = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
         response_data = self._workflow_run_repo.get_daily_runs_statistics(
             tenant_id=app_model.tenant_id,
@@ -100,23 +86,11 @@ class WorkflowDailyTerminalsStatistic(Resource):
         args = parser.parse_args()
 
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
 
-        start_date = None
-        end_date = None
-
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_date = start_datetime_timezone.astimezone(utc_timezone)
-
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=0)
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_date = end_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_date, end_date = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
         response_data = self._workflow_run_repo.get_daily_terminals_statistics(
             tenant_id=app_model.tenant_id,
@@ -157,23 +131,11 @@ class WorkflowDailyTokenCostStatistic(Resource):
         args = parser.parse_args()
 
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
 
-        start_date = None
-        end_date = None
-
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_date = start_datetime_timezone.astimezone(utc_timezone)
-
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=0)
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_date = end_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_date, end_date = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
         response_data = self._workflow_run_repo.get_daily_token_cost_statistics(
             tenant_id=app_model.tenant_id,
@@ -214,23 +176,11 @@ class WorkflowAverageAppInteractionStatistic(Resource):
         args = parser.parse_args()
 
         assert account.timezone is not None
-        timezone = pytz.timezone(account.timezone)
-        utc_timezone = pytz.utc
 
-        start_date = None
-        end_date = None
-
-        if args["start"]:
-            start_datetime = datetime.strptime(args["start"], "%Y-%m-%d %H:%M")
-            start_datetime = start_datetime.replace(second=0)
-            start_datetime_timezone = timezone.localize(start_datetime)
-            start_date = start_datetime_timezone.astimezone(utc_timezone)
-
-        if args["end"]:
-            end_datetime = datetime.strptime(args["end"], "%Y-%m-%d %H:%M")
-            end_datetime = end_datetime.replace(second=0)
-            end_datetime_timezone = timezone.localize(end_datetime)
-            end_date = end_datetime_timezone.astimezone(utc_timezone)
+        try:
+            start_date, end_date = parse_time_range(args["start"], args["end"], account.timezone)
+        except ValueError as e:
+            abort(400, description=str(e))
 
         response_data = self._workflow_run_repo.get_average_app_interaction_statistics(
             tenant_id=app_model.tenant_id,

api/controllers/console/billing/billing.py

@@ -2,6 +2,7 @@ from flask_restx import Resource, reqparse
 
 from controllers.console import console_ns
 from controllers.console.wraps import account_initialization_required, only_edition_cloud, setup_required
+from enums.cloud_plan import CloudPlan
 from libs.login import current_account_with_tenant, login_required
 from services.billing_service import BillingService
 
@@ -16,7 +17,13 @@ class Subscription(Resource):
         current_user, current_tenant_id = current_account_with_tenant()
         parser = (
             reqparse.RequestParser()
-            .add_argument("plan", type=str, required=True, location="args", choices=["professional", "team"])
+            .add_argument(
+                "plan",
+                type=str,
+                required=True,
+                location="args",
+                choices=[CloudPlan.PROFESSIONAL, CloudPlan.TEAM],
+            )
             .add_argument("interval", type=str, required=True, location="args", choices=["month", "year"])
         )
         args = parser.parse_args()

api/controllers/console/datasets/datasets_document.py

@@ -746,7 +746,7 @@ class DocumentApi(DocumentResource):
                 "name": document.name,
                 "created_from": document.created_from,
                 "created_by": document.created_by,
-                "created_at": document.created_at.timestamp(),
+                "created_at": int(document.created_at.timestamp()),
                 "tokens": document.tokens,
                 "indexing_status": document.indexing_status,
                 "completed_at": int(document.completed_at.timestamp()) if document.completed_at else None,
@@ -779,7 +779,7 @@ class DocumentApi(DocumentResource):
                 "name": document.name,
                 "created_from": document.created_from,
                 "created_by": document.created_by,
-                "created_at": document.created_at.timestamp(),
+                "created_at": int(document.created_at.timestamp()),
                 "tokens": document.tokens,
                 "indexing_status": document.indexing_status,
                 "completed_at": int(document.completed_at.timestamp()) if document.completed_at else None,

api/controllers/console/files.py

@@ -8,6 +8,7 @@ import services
 from configs import dify_config
 from constants import DOCUMENT_EXTENSIONS
 from controllers.common.errors import (
+    BlockedFileExtensionError,
     FilenameNotExistsError,
     FileTooLargeError,
     NoFileUploadedError,
@@ -83,6 +84,8 @@ class FileApi(Resource):
             raise FileTooLargeError(file_too_large_error.description)
         except services.errors.file.UnsupportedFileTypeError:
             raise UnsupportedFileTypeError()
+        except services.errors.file.BlockedFileExtensionError as blocked_extension_error:
+            raise BlockedFileExtensionError(blocked_extension_error.description)
 
         return upload_file, 201
 

api/controllers/console/workspace/workspace.py

@@ -21,6 +21,7 @@ from controllers.console.wraps import (
     cloud_edition_billing_resource_check,
     setup_required,
 )
+from enums.cloud_plan import CloudPlan
 from extensions.ext_database import db
 from libs.helper import TimestampField
 from libs.login import current_account_with_tenant, login_required
@@ -83,7 +84,7 @@ class TenantListApi(Resource):
                 "name": tenant.name,
                 "status": tenant.status,
                 "created_at": tenant.created_at,
-                "plan": features.billing.subscription.plan if features.billing.enabled else "sandbox",
+                "plan": features.billing.subscription.plan if features.billing.enabled else CloudPlan.SANDBOX,
                 "current": tenant.id == current_tenant_id if current_tenant_id else False,
             }
 

api/controllers/console/wraps.py

@@ -10,6 +10,7 @@ from flask import abort, request
 
 from configs import dify_config
 from controllers.console.workspace.error import AccountNotInitializedError
+from enums.cloud_plan import CloudPlan
 from extensions.ext_database import db
 from extensions.ext_redis import redis_client
 from libs.login import current_account_with_tenant
@@ -133,7 +134,7 @@ def cloud_edition_billing_knowledge_limit_check(resource: str):
             features = FeatureService.get_features(current_tenant_id)
             if features.billing.enabled:
                 if resource == "add_segment":
-                    if features.billing.subscription.plan == "sandbox":
+                    if features.billing.subscription.plan == CloudPlan.SANDBOX:
                         abort(
                             403,
                             "To unlock this feature and elevate your Dify experience, please upgrade to a paid plan.",

api/controllers/service_api/dataset/document.py

@@ -592,7 +592,7 @@ class DocumentApi(DatasetApiResource):
                 "name": document.name,
                 "created_from": document.created_from,
                 "created_by": document.created_by,
-                "created_at": document.created_at.timestamp(),
+                "created_at": int(document.created_at.timestamp()),
                 "tokens": document.tokens,
                 "indexing_status": document.indexing_status,
                 "completed_at": int(document.completed_at.timestamp()) if document.completed_at else None,
@@ -625,7 +625,7 @@ class DocumentApi(DatasetApiResource):
                 "name": document.name,
                 "created_from": document.created_from,
                 "created_by": document.created_by,
-                "created_at": document.created_at.timestamp(),
+                "created_at": int(document.created_at.timestamp()),
                 "tokens": document.tokens,
                 "indexing_status": document.indexing_status,
                 "completed_at": int(document.completed_at.timestamp()) if document.completed_at else None,

api/controllers/service_api/dataset/segment.py

@@ -2,6 +2,7 @@ from flask import request
 from flask_restx import marshal, reqparse
 from werkzeug.exceptions import NotFound
 
+from configs import dify_config
 from controllers.service_api import service_api_ns
 from controllers.service_api.app.error import ProviderNotInitializeError
 from controllers.service_api.wraps import (
@@ -107,6 +108,10 @@ class SegmentApi(DatasetApiResource):
         # validate args
         args = segment_create_parser.parse_args()
         if args["segments"] is not None:
+            segments_limit = dify_config.DATASET_MAX_SEGMENTS_PER_REQUEST
+            if segments_limit > 0 and len(args["segments"]) > segments_limit:
+                raise ValueError(f"Exceeded maximum segments limit of {segments_limit}.")
+
             for args_item in args["segments"]:
                 SegmentService.segment_create_args_validate(args_item, document)
             segments = SegmentService.multi_create_segment(args["segments"], document, dataset)

api/controllers/service_api/wraps.py

@@ -13,6 +13,7 @@ from sqlalchemy import select, update
 from sqlalchemy.orm import Session
 from werkzeug.exceptions import Forbidden, NotFound, Unauthorized
 
+from enums.cloud_plan import CloudPlan
 from extensions.ext_database import db
 from extensions.ext_redis import redis_client
 from libs.datetime_utils import naive_utc_now
@@ -66,6 +67,7 @@ def validate_app_token(view: Callable[P, R] | None = None, *, fetch_user_arg: Fe
 
             kwargs["app_model"] = app_model
 
+            # If caller needs end-user context, attach EndUser to current_user
             if fetch_user_arg:
                 if fetch_user_arg.fetch_from == WhereisUserArg.QUERY:
                     user_id = request.args.get("user")
@@ -74,7 +76,6 @@ def validate_app_token(view: Callable[P, R] | None = None, *, fetch_user_arg: Fe
                 elif fetch_user_arg.fetch_from == WhereisUserArg.FORM:
                     user_id = request.form.get("user")
                 else:
-                    # use default-user
                     user_id = None
 
                 if not user_id and fetch_user_arg.required:
@@ -89,6 +90,28 @@ def validate_app_token(view: Callable[P, R] | None = None, *, fetch_user_arg: Fe
                 # Set EndUser as current logged-in user for flask_login.current_user
                 current_app.login_manager._update_request_context_with_user(end_user)  # type: ignore
                 user_logged_in.send(current_app._get_current_object(), user=end_user)  # type: ignore
+            else:
+                # For service API without end-user context, ensure an Account is logged in
+                # so services relying on current_account_with_tenant() work correctly.
+                tenant_owner_info = (
+                    db.session.query(Tenant, Account)
+                    .join(TenantAccountJoin, Tenant.id == TenantAccountJoin.tenant_id)
+                    .join(Account, TenantAccountJoin.account_id == Account.id)
+                    .where(
+                        Tenant.id == app_model.tenant_id,
+                        TenantAccountJoin.role == "owner",
+                        Tenant.status == TenantStatus.NORMAL,
+                    )
+                    .one_or_none()
+                )
+
+                if tenant_owner_info:
+                    tenant_model, account = tenant_owner_info
+                    account.current_tenant = tenant_model
+                    current_app.login_manager._update_request_context_with_user(account)  # type: ignore
+                    user_logged_in.send(current_app._get_current_object(), user=current_user)  # type: ignore
+                else:
+                    raise Unauthorized("Tenant owner account not found or tenant is not active.")
 
             return view_func(*args, **kwargs)
 
@@ -138,7 +161,7 @@ def cloud_edition_billing_knowledge_limit_check(resource: str, api_token_type: s
             features = FeatureService.get_features(api_token.tenant_id)
             if features.billing.enabled:
                 if resource == "add_segment":
-                    if features.billing.subscription.plan == "sandbox":
+                    if features.billing.subscription.plan == CloudPlan.SANDBOX:
                         raise Forbidden(
                             "To unlock this feature and elevate your Dify experience, please upgrade to a paid plan."
                         )

api/core/app/apps/agent_chat/app_runner.py

@@ -144,7 +144,7 @@ class AgentChatAppRunner(AppRunner):
             prompt_template_entity=app_config.prompt_template,
             inputs=dict(inputs),
             files=list(files),
-            query=query or "",
+            query=query,
             memory=memory,
         )
 
@@ -172,7 +172,7 @@ class AgentChatAppRunner(AppRunner):
             prompt_template_entity=app_config.prompt_template,
             inputs=dict(inputs),
             files=list(files),
-            query=query or "",
+            query=query,
             memory=memory,
         )
 

api/core/app/apps/base_app_runner.py

@@ -79,7 +79,7 @@ class AppRunner:
         prompt_template_entity: PromptTemplateEntity,
         inputs: Mapping[str, str],
         files: Sequence["File"],
-        query: str | None = None,
+        query: str = "",
         context: str | None = None,
         memory: TokenBufferMemory | None = None,
         image_detail_config: ImagePromptMessageContent.DETAIL | None = None,
@@ -105,7 +105,7 @@ class AppRunner:
                 app_mode=AppMode.value_of(app_record.mode),
                 prompt_template_entity=prompt_template_entity,
                 inputs=inputs,
-                query=query or "",
+                query=query,
                 files=files,
                 context=context,
                 memory=memory,

api/core/app/apps/common/workflow_response_converter.py

@@ -1,3 +1,4 @@
+import logging
 import time
 from collections.abc import Mapping, Sequence
 from dataclasses import dataclass
@@ -54,6 +55,7 @@ from models import Account, EndUser
 from services.variable_truncator import BaseTruncator, DummyVariableTruncator, VariableTruncator
 
 NodeExecutionId = NewType("NodeExecutionId", str)
+logger = logging.getLogger(__name__)
 
 
 @dataclass(slots=True)
@@ -288,21 +290,25 @@ class WorkflowResponseConverter:
             ),
         )
 
-        if event.node_type == NodeType.TOOL:
-            response.data.extras["icon"] = ToolManager.get_tool_icon(
-                tenant_id=self._application_generate_entity.app_config.tenant_id,
-                provider_type=ToolProviderType(event.provider_type),
-                provider_id=event.provider_id,
-            )
-        elif event.node_type == NodeType.DATASOURCE:
-            manager = PluginDatasourceManager()
-            provider_entity = manager.fetch_datasource_provider(
-                self._application_generate_entity.app_config.tenant_id,
-                event.provider_id,
-            )
-            response.data.extras["icon"] = provider_entity.declaration.identity.generate_datasource_icon_url(
-                self._application_generate_entity.app_config.tenant_id
-            )
+        try:
+            if event.node_type == NodeType.TOOL:
+                response.data.extras["icon"] = ToolManager.get_tool_icon(
+                    tenant_id=self._application_generate_entity.app_config.tenant_id,
+                    provider_type=ToolProviderType(event.provider_type),
+                    provider_id=event.provider_id,
+                )
+            elif event.node_type == NodeType.DATASOURCE:
+                manager = PluginDatasourceManager()
+                provider_entity = manager.fetch_datasource_provider(
+                    self._application_generate_entity.app_config.tenant_id,
+                    event.provider_id,
+                )
+                response.data.extras["icon"] = provider_entity.declaration.identity.generate_datasource_icon_url(
+                    self._application_generate_entity.app_config.tenant_id
+                )
+        except Exception:
+            # metadata fetch may fail, for example, the plugin daemon is down or plugin is uninstalled.
+            logger.warning("failed to fetch icon for %s", event.provider_id)
 
         return response
 

api/core/app/apps/message_based_app_generator.py

@@ -190,7 +190,7 @@ class MessageBasedAppGenerator(BaseAppGenerator):
             override_model_configs=json.dumps(override_model_configs) if override_model_configs else None,
             conversation_id=conversation.id,
             inputs=application_generate_entity.inputs,
-            query=application_generate_entity.query or "",
+            query=application_generate_entity.query,
             message="",
             message_tokens=0,
             message_unit_price=0,

api/core/app/apps/pipeline/pipeline_generator.py

@@ -41,18 +41,14 @@ from core.workflow.repositories.workflow_execution_repository import WorkflowExe
 from core.workflow.repositories.workflow_node_execution_repository import WorkflowNodeExecutionRepository
 from core.workflow.variable_loader import DUMMY_VARIABLE_LOADER, VariableLoader
 from extensions.ext_database import db
-from extensions.ext_redis import redis_client
 from libs.flask_utils import preserve_flask_contexts
 from models import Account, EndUser, Workflow, WorkflowNodeExecutionTriggeredFrom
 from models.dataset import Document, DocumentPipelineExecutionLog, Pipeline
 from models.enums import WorkflowRunTriggeredFrom
 from models.model import AppMode
 from services.datasource_provider_service import DatasourceProviderService
-from services.feature_service import FeatureService
-from services.file_service import FileService
+from services.rag_pipeline.rag_pipeline_task_proxy import RagPipelineTaskProxy
 from services.workflow_draft_variable_service import DraftVarLoader, WorkflowDraftVariableService
-from tasks.rag_pipeline.priority_rag_pipeline_run_task import priority_rag_pipeline_run_task
-from tasks.rag_pipeline.rag_pipeline_run_task import rag_pipeline_run_task
 
 logger = logging.getLogger(__name__)
 
@@ -248,34 +244,7 @@ class PipelineGenerator(BaseAppGenerator):
                 )
 
         if rag_pipeline_invoke_entities:
-            # store the rag_pipeline_invoke_entities to object storage
-            text = [item.model_dump() for item in rag_pipeline_invoke_entities]
-            name = "rag_pipeline_invoke_entities.json"
-            # Convert list to proper JSON string
-            json_text = json.dumps(text)
-            upload_file = FileService(db.engine).upload_text(json_text, name, user.id, dataset.tenant_id)
-            features = FeatureService.get_features(dataset.tenant_id)
-            if features.billing.enabled and features.billing.subscription.plan == "sandbox":
-                tenant_pipeline_task_key = f"tenant_pipeline_task:{dataset.tenant_id}"
-                tenant_self_pipeline_task_queue = f"tenant_self_pipeline_task_queue:{dataset.tenant_id}"
-
-                if redis_client.get(tenant_pipeline_task_key):
-                    # Add to waiting queue using List operations (lpush)
-                    redis_client.lpush(tenant_self_pipeline_task_queue, upload_file.id)
-                else:
-                    # Set flag and execute task
-                    redis_client.set(tenant_pipeline_task_key, 1, ex=60 * 60)
-                    rag_pipeline_run_task.delay(  # type: ignore
-                        rag_pipeline_invoke_entities_file_id=upload_file.id,
-                        tenant_id=dataset.tenant_id,
-                    )
-
-            else:
-                priority_rag_pipeline_run_task.delay(  # type: ignore
-                    rag_pipeline_invoke_entities_file_id=upload_file.id,
-                    tenant_id=dataset.tenant_id,
-                )
-
+            RagPipelineTaskProxy(dataset.tenant_id, user.id, rag_pipeline_invoke_entities).delay()
         # return batch, dataset, documents
         return {
             "batch": batch,

api/core/app/entities/app_invoke_entities.py

@@ -129,7 +129,7 @@ class EasyUIBasedAppGenerateEntity(AppGenerateEntity):
     app_config: EasyUIBasedAppConfig = None  # type: ignore
     model_conf: ModelConfigWithCredentialsEntity
 
-    query: str | None = None
+    query: str = ""
 
     # pydantic configs
     model_config = ConfigDict(protected_namespaces=())

api/core/app/task_pipeline/easy_ui_based_generate_task_pipeline.py

@@ -121,7 +121,7 @@ class EasyUIBasedGenerateTaskPipeline(BasedGenerateTaskPipeline):
         if self._application_generate_entity.app_config.app_mode != AppMode.COMPLETION:
             # start generate conversation name thread
             self._conversation_name_generate_thread = self._message_cycle_manager.generate_conversation_name(
-                conversation_id=self._conversation_id, query=self._application_generate_entity.query or ""
+                conversation_id=self._conversation_id, query=self._application_generate_entity.query
             )
 
         generator = self._wrapper_process_stream_response(trace_manager=self._application_generate_entity.trace_manager)

api/core/app/task_pipeline/message_cycle_manager.py

@@ -140,7 +140,27 @@ class MessageCycleManager:
         if not self._application_generate_entity.app_config.additional_features:
             raise ValueError("Additional features not found")
         if self._application_generate_entity.app_config.additional_features.show_retrieve_source:
-            self._task_state.metadata.retriever_resources = event.retriever_resources
+            merged_resources = [r for r in self._task_state.metadata.retriever_resources or [] if r]
+            existing_ids = {(r.dataset_id, r.document_id) for r in merged_resources if r.dataset_id and r.document_id}
+
+            # Add new unique resources from the event
+            for resource in event.retriever_resources or []:
+                if not resource:
+                    continue
+
+                is_duplicate = (
+                    resource.dataset_id
+                    and resource.document_id
+                    and (resource.dataset_id, resource.document_id) in existing_ids
+                )
+
+                if not is_duplicate:
+                    merged_resources.append(resource)
+
+            for i, resource in enumerate(merged_resources, 1):
+                resource.position = i
+
+            self._task_state.metadata.retriever_resources = merged_resources
 
     def message_file_to_stream_response(self, event: QueueMessageFileEvent) -> MessageFileStreamResponse | None:
         """

api/core/entities/document_task.py

@@ -0,0 +1,15 @@
+from collections.abc import Sequence
+from dataclasses import dataclass
+
+
+@dataclass
+class DocumentTask:
+    """Document task entity for document indexing operations.
+
+    This class represents a document indexing task that can be queued
+    and processed by the document indexing system.
+    """
+
+    tenant_id: str
+    dataset_id: str
+    document_ids: Sequence[str]

api/core/entities/provider_configuration.py

@@ -1533,6 +1533,9 @@ class ProviderConfiguration(BaseModel):
             # Return composite sort key: (model_type value, model position index)
             return (model.model_type.value, position_index)
 
+        # Deduplicate
+        provider_models = list({(m.model, m.model_type, m.fetch_from): m for m in provider_models}.values())
+
         # Sort using the composite sort key
         return sorted(provider_models, key=get_sort_key)
 

api/core/file/models.py

@@ -74,6 +74,10 @@ class File(BaseModel):
         storage_key: str | None = None,
         dify_model_identity: str | None = FILE_MODEL_IDENTITY,
         url: str | None = None,
+        # Legacy compatibility fields - explicitly handle known extra fields
+        tool_file_id: str | None = None,
+        upload_file_id: str | None = None,
+        datasource_file_id: str | None = None,
     ):
         super().__init__(
             id=id,

api/core/helper/code_executor/javascript/javascript_transformer.py

@@ -6,10 +6,7 @@ from core.helper.code_executor.template_transformer import TemplateTransformer
 class NodeJsTemplateTransformer(TemplateTransformer):
     @classmethod
     def get_runner_script(cls) -> str:
-        runner_script = dedent(
-            f"""
-            // declare main function
-            {cls._code_placeholder}
+        runner_script = dedent(f"""            {cls._code_placeholder}
 
             // decode and prepare input object
             var inputs_obj = JSON.parse(Buffer.from('{cls._inputs_placeholder}', 'base64').toString('utf-8'))
@@ -21,6 +18,5 @@ class NodeJsTemplateTransformer(TemplateTransformer):
             var output_json = JSON.stringify(output_obj)
             var result = `<<RESULT>>${{output_json}}<<RESULT>>`
             console.log(result)
-            """
-        )
+            """)
         return runner_script

api/core/helper/code_executor/python3/python3_transformer.py

@@ -6,9 +6,7 @@ from core.helper.code_executor.template_transformer import TemplateTransformer
 class Python3TemplateTransformer(TemplateTransformer):
     @classmethod
     def get_runner_script(cls) -> str:
-        runner_script = dedent(f"""
-            # declare main function
-            {cls._code_placeholder}
+        runner_script = dedent(f"""            {cls._code_placeholder}
 
             import json
             from base64 import b64decode

api/core/ops/tencent_trace/client.py

@@ -5,6 +5,7 @@ Tencent APM Trace Client - handles network operations, metrics, and API communic
 from __future__ import annotations
 
 import importlib
+import json
 import logging
 import os
 import socket
@@ -110,6 +111,7 @@ class TencentTraceClient:
         self.span_contexts: dict[int, trace_api.SpanContext] = {}
 
         self.meter: Meter | None = None
+        self.meter_provider: MeterProvider | None = None
         self.hist_llm_duration: Histogram | None = None
         self.hist_token_usage: Histogram | None = None
         self.hist_time_to_first_token: Histogram | None = None
@@ -119,7 +121,6 @@ class TencentTraceClient:
 
         # Metrics exporter and instruments
         try:
-            from opentelemetry import metrics
             from opentelemetry.sdk.metrics import Histogram, MeterProvider
             from opentelemetry.sdk.metrics.export import AggregationTemporality, PeriodicExportingMetricReader
 
@@ -202,9 +203,11 @@ class TencentTraceClient:
                 )
 
             if metric_reader is not None:
+                # Use instance-level MeterProvider instead of global to support config changes
+                # without worker restart. Each TencentTraceClient manages its own MeterProvider.
                 provider = MeterProvider(resource=self.resource, metric_readers=[metric_reader])
-                metrics.set_meter_provider(provider)
-                self.meter = metrics.get_meter("dify-sdk", dify_config.project.version)
+                self.meter_provider = provider
+                self.meter = provider.get_meter("dify-sdk", dify_config.project.version)
 
                 # LLM operation duration histogram
                 self.hist_llm_duration = self.meter.create_histogram(
@@ -244,6 +247,7 @@ class TencentTraceClient:
                 self.metric_reader = metric_reader
             else:
                 self.meter = None
+                self.meter_provider = None
                 self.hist_llm_duration = None
                 self.hist_token_usage = None
                 self.hist_time_to_first_token = None
@@ -253,6 +257,7 @@ class TencentTraceClient:
         except Exception:
             logger.exception("[Tencent APM] Metrics initialization failed; metrics disabled")
             self.meter = None
+            self.meter_provider = None
             self.hist_llm_duration = None
             self.hist_token_usage = None
             self.hist_time_to_first_token = None
@@ -279,6 +284,14 @@ class TencentTraceClient:
             if attributes:
                 for k, v in attributes.items():
                     attrs[k] = str(v) if not isinstance(v, (str, int, float, bool)) else v  # type: ignore[assignment]
+
+            logger.info(
+                "[Tencent Metrics] Metric: %s | Value: %.4f | Attributes: %s",
+                LLM_OPERATION_DURATION,
+                latency_seconds,
+                json.dumps(attrs, ensure_ascii=False),
+            )
+
             self.hist_llm_duration.record(latency_seconds, attrs)  # type: ignore[attr-defined]
         except Exception:
             logger.debug("[Tencent APM] Failed to record LLM duration", exc_info=True)
@@ -317,6 +330,13 @@ class TencentTraceClient:
                 "server.address": server_address,
             }
 
+            logger.info(
+                "[Tencent Metrics] Metric: %s | Value: %d | Attributes: %s",
+                GEN_AI_TOKEN_USAGE,
+                token_count,
+                json.dumps(attributes, ensure_ascii=False),
+            )
+
             self.hist_token_usage.record(token_count, attributes)  # type: ignore[attr-defined]
         except Exception:
             logger.debug("[Tencent APM] Failed to record token usage", exc_info=True)
@@ -344,6 +364,13 @@ class TencentTraceClient:
                 "stream": "true",
             }
 
+            logger.info(
+                "[Tencent Metrics] Metric: %s | Value: %.4f | Attributes: %s",
+                GEN_AI_SERVER_TIME_TO_FIRST_TOKEN,
+                ttft_seconds,
+                json.dumps(attributes, ensure_ascii=False),
+            )
+
             self.hist_time_to_first_token.record(ttft_seconds, attributes)  # type: ignore[attr-defined]
         except Exception:
             logger.debug("[Tencent APM] Failed to record time to first token", exc_info=True)
@@ -371,6 +398,13 @@ class TencentTraceClient:
                 "stream": "true",
             }
 
+            logger.info(
+                "[Tencent Metrics] Metric: %s | Value: %.4f | Attributes: %s",
+                GEN_AI_STREAMING_TIME_TO_GENERATE,
+                ttg_seconds,
+                json.dumps(attributes, ensure_ascii=False),
+            )
+
             self.hist_time_to_generate.record(ttg_seconds, attributes)  # type: ignore[attr-defined]
         except Exception:
             logger.debug("[Tencent APM] Failed to record time to generate", exc_info=True)
@@ -390,6 +424,14 @@ class TencentTraceClient:
             if attributes:
                 for k, v in attributes.items():
                     attrs[k] = str(v) if not isinstance(v, (str, int, float, bool)) else v  # type: ignore[assignment]
+
+            logger.info(
+                "[Tencent Metrics] Metric: %s | Value: %.4f | Attributes: %s",
+                GEN_AI_TRACE_DURATION,
+                duration_seconds,
+                json.dumps(attrs, ensure_ascii=False),
+            )
+
             self.hist_trace_duration.record(duration_seconds, attrs)  # type: ignore[attr-defined]
         except Exception:
             logger.debug("[Tencent APM] Failed to record trace duration", exc_info=True)
@@ -474,11 +516,19 @@ class TencentTraceClient:
 
             if self.tracer_provider:
                 self.tracer_provider.shutdown()
+
+            # Shutdown instance-level meter provider
+            if self.meter_provider is not None:
+                try:
+                    self.meter_provider.shutdown()  # type: ignore[attr-defined]
+                except Exception:
+                    logger.debug("[Tencent APM] Error shutting down meter provider", exc_info=True)
+
             if self.metric_reader is not None:
                 try:
                     self.metric_reader.shutdown()  # type: ignore[attr-defined]
                 except Exception:
-                    pass
+                    logger.debug("[Tencent APM] Error shutting down metric reader", exc_info=True)
 
         except Exception:
             logger.exception("[Tencent APM] Error during client shutdown")

api/core/rag/datasource/vdb/elasticsearch/elasticsearch_vector.py

@@ -147,7 +147,8 @@ class ElasticSearchVector(BaseVector):
 
     def _get_version(self) -> str:
         info = self._client.info()
-        return cast(str, info["version"]["number"])
+        # remove any suffix like "-SNAPSHOT" from the version string
+        return cast(str, info["version"]["number"]).split("-")[0]
 
     def _check_version(self):
         if parse_version(self._version) < parse_version("8.0.0"):

api/core/rag/datasource/vdb/weaviate/weaviate_vector.py

@@ -39,11 +39,13 @@ class WeaviateConfig(BaseModel):
 
     Attributes:
         endpoint: Weaviate server endpoint URL
+        grpc_endpoint: Optional Weaviate gRPC server endpoint URL
         api_key: Optional API key for authentication
         batch_size: Number of objects to batch per insert operation
     """
 
     endpoint: str
+    grpc_endpoint: str | None = None
     api_key: str | None = None
     batch_size: int = 100
 
@@ -88,9 +90,22 @@ class WeaviateVector(BaseVector):
         http_secure = p.scheme == "https"
         http_port = p.port or (443 if http_secure else 80)
 
-        grpc_host = host
-        grpc_secure = http_secure
-        grpc_port = 443 if grpc_secure else 50051
+        # Parse gRPC configuration
+        if config.grpc_endpoint:
+            # Urls without scheme won't be parsed correctly in some python versions,
+            # see https://bugs.python.org/issue27657
+            grpc_endpoint_with_scheme = (
+                config.grpc_endpoint if "://" in config.grpc_endpoint else f"grpc://{config.grpc_endpoint}"
+            )
+            grpc_p = urlparse(grpc_endpoint_with_scheme)
+            grpc_host = grpc_p.hostname or "localhost"
+            grpc_port = grpc_p.port or (443 if grpc_p.scheme == "grpcs" else 50051)
+            grpc_secure = grpc_p.scheme == "grpcs"
+        else:
+            # Infer from HTTP endpoint as fallback
+            grpc_host = host
+            grpc_secure = http_secure
+            grpc_port = 443 if grpc_secure else 50051
 
         client = weaviate.connect_to_custom(
             http_host=host,
@@ -432,6 +447,7 @@ class WeaviateVectorFactory(AbstractVectorFactory):
             collection_name=collection_name,
             config=WeaviateConfig(
                 endpoint=dify_config.WEAVIATE_ENDPOINT or "",
+                grpc_endpoint=dify_config.WEAVIATE_GRPC_ENDPOINT or "",
                 api_key=dify_config.WEAVIATE_API_KEY,
                 batch_size=dify_config.WEAVIATE_BATCH_SIZE,
             ),

api/core/rag/pipeline/queue.py

@@ -0,0 +1,79 @@
+import json
+from collections.abc import Sequence
+from typing import Any
+
+from pydantic import BaseModel, ValidationError
+
+from extensions.ext_redis import redis_client
+
+_DEFAULT_TASK_TTL = 60 * 60  # 1 hour
+
+
+class TaskWrapper(BaseModel):
+    data: Any
+
+    def serialize(self) -> str:
+        return self.model_dump_json()
+
+    @classmethod
+    def deserialize(cls, serialized_data: str) -> "TaskWrapper":
+        return cls.model_validate_json(serialized_data)
+
+
+class TenantIsolatedTaskQueue:
+    """
+    Simple queue for tenant isolated tasks, used for rag related tenant tasks isolation.
+    It uses Redis list to store tasks, and Redis key to store task waiting flag.
+    Support tasks that can be serialized by json.
+    """
+
+    def __init__(self, tenant_id: str, unique_key: str):
+        self._tenant_id = tenant_id
+        self._unique_key = unique_key
+        self._queue = f"tenant_self_{unique_key}_task_queue:{tenant_id}"
+        self._task_key = f"tenant_{unique_key}_task:{tenant_id}"
+
+    def get_task_key(self):
+        return redis_client.get(self._task_key)
+
+    def set_task_waiting_time(self, ttl: int = _DEFAULT_TASK_TTL):
+        redis_client.setex(self._task_key, ttl, 1)
+
+    def delete_task_key(self):
+        redis_client.delete(self._task_key)
+
+    def push_tasks(self, tasks: Sequence[Any]):
+        serialized_tasks = []
+        for task in tasks:
+            # Store str list directly, maintaining full compatibility for pipeline scenarios
+            if isinstance(task, str):
+                serialized_tasks.append(task)
+            else:
+                # Use TaskWrapper to do JSON serialization for non-string tasks
+                wrapper = TaskWrapper(data=task)
+                serialized_data = wrapper.serialize()
+                serialized_tasks.append(serialized_data)
+
+        redis_client.lpush(self._queue, *serialized_tasks)
+
+    def pull_tasks(self, count: int = 1) -> Sequence[Any]:
+        if count <= 0:
+            return []
+
+        tasks = []
+        for _ in range(count):
+            serialized_task = redis_client.rpop(self._queue)
+            if not serialized_task:
+                break
+
+            if isinstance(serialized_task, bytes):
+                serialized_task = serialized_task.decode("utf-8")
+
+            try:
+                wrapper = TaskWrapper.deserialize(serialized_task)
+                tasks.append(wrapper.data)
+            except (json.JSONDecodeError, ValidationError, TypeError, ValueError):
+                # Fall back to raw string for legacy format or invalid JSON
+                tasks.append(serialized_task)
+
+        return tasks

api/core/tools/__base/tool.py

@@ -210,12 +210,13 @@ class Tool(ABC):
             meta=meta,
         )
 
-    def create_json_message(self, object: dict) -> ToolInvokeMessage:
+    def create_json_message(self, object: dict, suppress_output: bool = False) -> ToolInvokeMessage:
         """
         create a json message
         """
         return ToolInvokeMessage(
-            type=ToolInvokeMessage.MessageType.JSON, message=ToolInvokeMessage.JsonMessage(json_object=object)
+            type=ToolInvokeMessage.MessageType.JSON,
+            message=ToolInvokeMessage.JsonMessage(json_object=object, suppress_output=suppress_output),
         )
 
     def create_variable_message(

api/core/tools/entities/tool_entities.py

@@ -129,6 +129,7 @@ class ToolInvokeMessage(BaseModel):
 
     class JsonMessage(BaseModel):
         json_object: dict
+        suppress_output: bool = Field(default=False, description="Whether to suppress JSON output in result string")
 
     class BlobMessage(BaseModel):
         blob: bytes

api/core/tools/mcp_tool/tool.py

@@ -1,16 +1,19 @@
 import base64
 import json
+import logging
 from collections.abc import Generator
 from typing import Any
 
 from core.mcp.auth_client import MCPClientWithAuthRetry
 from core.mcp.error import MCPConnectionError
-from core.mcp.types import CallToolResult, ImageContent, TextContent
+from core.mcp.types import AudioContent, CallToolResult, ImageContent, TextContent
 from core.tools.__base.tool import Tool
 from core.tools.__base.tool_runtime import ToolRuntime
 from core.tools.entities.tool_entities import ToolEntity, ToolInvokeMessage, ToolProviderType
 from core.tools.errors import ToolInvokeError
 
+logger = logging.getLogger(__name__)
+
 
 class MCPTool(Tool):
     def __init__(
@@ -52,6 +55,11 @@ class MCPTool(Tool):
                 yield from self._process_text_content(content)
             elif isinstance(content, ImageContent):
                 yield self._process_image_content(content)
+            elif isinstance(content, AudioContent):
+                yield self._process_audio_content(content)
+            else:
+                logger.warning("Unsupported content type=%s", type(content))
+
         # handle MCP structured output
         if self.entity.output_schema and result.structuredContent:
             for k, v in result.structuredContent.items():
@@ -97,6 +105,10 @@ class MCPTool(Tool):
         """Process image content and return a blob message."""
         return self.create_blob_message(blob=base64.b64decode(content.data), meta={"mime_type": content.mimeType})
 
+    def _process_audio_content(self, content: AudioContent) -> ToolInvokeMessage:
+        """Process audio content and return a blob message."""
+        return self.create_blob_message(blob=base64.b64decode(content.data), meta={"mime_type": content.mimeType})
+
     def fork_tool_runtime(self, runtime: ToolRuntime) -> "MCPTool":
         return MCPTool(
             entity=self.entity,

api/core/tools/tool_engine.py

@@ -228,29 +228,41 @@ class ToolEngine:
         """
         Handle tool response
         """
-        result = ""
+        parts: list[str] = []
+        json_parts: list[str] = []
+
         for response in tool_response:
             if response.type == ToolInvokeMessage.MessageType.TEXT:
-                result += cast(ToolInvokeMessage.TextMessage, response.message).text
+                parts.append(cast(ToolInvokeMessage.TextMessage, response.message).text)
             elif response.type == ToolInvokeMessage.MessageType.LINK:
-                result += (
+                parts.append(
                     f"result link: {cast(ToolInvokeMessage.TextMessage, response.message).text}."
                     + " please tell user to check it."
                 )
             elif response.type in {ToolInvokeMessage.MessageType.IMAGE_LINK, ToolInvokeMessage.MessageType.IMAGE}:
-                result += (
+                parts.append(
                     "image has been created and sent to user already, "
                     + "you do not need to create it, just tell the user to check it now."
                 )
             elif response.type == ToolInvokeMessage.MessageType.JSON:
-                result += json.dumps(
-                    safe_json_value(cast(ToolInvokeMessage.JsonMessage, response.message).json_object),
-                    ensure_ascii=False,
+                json_message = cast(ToolInvokeMessage.JsonMessage, response.message)
+                if json_message.suppress_output:
+                    continue
+                json_parts.append(
+                    json.dumps(
+                        safe_json_value(cast(ToolInvokeMessage.JsonMessage, response.message).json_object),
+                        ensure_ascii=False,
+                    )
                 )
             else:
-                result += str(response.message)
+                parts.append(str(response.message))
 
-        return result
+        # Add JSON parts, avoiding duplicates from text parts.
+        if json_parts:
+            existing_parts = set(parts)
+            parts.extend(p for p in json_parts if p not in existing_parts)
+
+        return "".join(parts)
 
     @staticmethod
     def _extract_tool_response_binary_and_text(

api/core/tools/workflow_as_tool/tool.py

@@ -117,7 +117,7 @@ class WorkflowTool(Tool):
         self._latest_usage = self._derive_usage_from_result(data)
 
         yield self.create_text_message(json.dumps(outputs, ensure_ascii=False))
-        yield self.create_json_message(outputs)
+        yield self.create_json_message(outputs, suppress_output=True)
 
     @property
     def latest_usage(self) -> LLMUsage:

api/core/workflow/nodes/tool/tool_node.py

@@ -332,7 +332,13 @@ class ToolNode(Node):
                     json.append(message.message.json_object)
             elif message.type == ToolInvokeMessage.MessageType.LINK:
                 assert isinstance(message.message, ToolInvokeMessage.TextMessage)
-                stream_text = f"Link: {message.message.text}\n"
+                # Check if this LINK message is a file link
+                file_obj = (message.meta or {}).get("file")
+                if isinstance(file_obj, File):
+                    files.append(file_obj)
+                    stream_text = f"File: {message.message.text}\n"
+                else:
+                    stream_text = f"Link: {message.message.text}\n"
                 text += stream_text
                 yield StreamChunkEvent(
                     selector=[node_id, "text"],

api/core/workflow/runtime/variable_pool.py

@@ -153,7 +153,11 @@ class VariablePool(BaseModel):
             return None
 
         node_id, name = self._selector_to_keys(selector)
-        segment: Segment | None = self.variable_dictionary[node_id].get(name)
+        node_map = self.variable_dictionary.get(node_id)
+        if node_map is None:
+            return None
+
+        segment: Segment | None = node_map.get(name)
 
         if segment is None:
             return None

api/docker/entrypoint.sh

@@ -32,7 +32,7 @@ if [[ "${MODE}" == "worker" ]]; then
 
   exec celery -A celery_entrypoint.celery worker -P ${CELERY_WORKER_CLASS:-gevent} $CONCURRENCY_OPTION \
     --max-tasks-per-child ${MAX_TASKS_PER_CHILD:-50} --loglevel ${LOG_LEVEL:-INFO} \
-    -Q ${CELERY_QUEUES:-dataset,mail,ops_trace,app_deletion,plugin,workflow_storage,conversation,priority_pipeline,pipeline} \
+    -Q ${CELERY_QUEUES:-dataset,priority_dataset,mail,ops_trace,app_deletion,plugin,workflow_storage,conversation,priority_pipeline,pipeline} \
     --prefetch-multiplier=1
 
 elif [[ "${MODE}" == "beat" ]]; then

api/enums/cloud_plan.py

@@ -0,0 +1,15 @@
+from enum import StrEnum, auto
+
+
+class CloudPlan(StrEnum):
+    """
+    Enum representing user plan types in the cloud platform.
+
+    SANDBOX: Free/default plan with limited features
+    PROFESSIONAL: Professional paid plan
+    TEAM: Team collaboration paid plan
+    """
+
+    SANDBOX = auto()
+    PROFESSIONAL = auto()
+    TEAM = auto()

api/libs/datetime_utils.py

@@ -2,6 +2,8 @@ import abc
 import datetime
 from typing import Protocol
 
+import pytz
+
 
 class _NowFunction(Protocol):
     @abc.abstractmethod
@@ -20,3 +22,51 @@ def naive_utc_now() -> datetime.datetime:
     representing current UTC time.
     """
     return _now_func(datetime.UTC).replace(tzinfo=None)
+
+
+def parse_time_range(
+    start: str | None, end: str | None, tzname: str
+) -> tuple[datetime.datetime | None, datetime.datetime | None]:
+    """
+    Parse time range strings and convert to UTC datetime objects.
+    Handles DST ambiguity and non-existent times gracefully.
+
+    Args:
+        start: Start time string (YYYY-MM-DD HH:MM)
+        end: End time string (YYYY-MM-DD HH:MM)
+        tzname: Timezone name
+
+    Returns:
+        tuple: (start_datetime_utc, end_datetime_utc)
+
+    Raises:
+        ValueError: When time range is invalid or start > end
+    """
+    tz = pytz.timezone(tzname)
+    utc = pytz.utc
+
+    def _parse(time_str: str | None, label: str) -> datetime.datetime | None:
+        if not time_str:
+            return None
+
+        try:
+            dt = datetime.datetime.strptime(time_str, "%Y-%m-%d %H:%M").replace(second=0)
+        except ValueError as e:
+            raise ValueError(f"Invalid {label} time format: {e}")
+
+        try:
+            return tz.localize(dt, is_dst=None).astimezone(utc)
+        except pytz.AmbiguousTimeError:
+            return tz.localize(dt, is_dst=False).astimezone(utc)
+        except pytz.NonExistentTimeError:
+            dt += datetime.timedelta(hours=1)
+            return tz.localize(dt, is_dst=None).astimezone(utc)
+
+    start_dt = _parse(start, "start")
+    end_dt = _parse(end, "end")
+
+    # Range validation
+    if start_dt and end_dt and start_dt > end_dt:
+        raise ValueError("start must be earlier than or equal to end")
+
+    return start_dt, end_dt

api/models/account.py

@@ -110,7 +110,7 @@ class Account(UserMixin, TypeBase):
         DateTime, server_default=func.current_timestamp(), nullable=False, init=False
     )
     updated_at: Mapped[datetime] = mapped_column(
-        DateTime, server_default=func.current_timestamp(), nullable=False, init=False
+        DateTime, server_default=func.current_timestamp(), nullable=False, init=False, onupdate=func.current_timestamp()
     )
 
     role: TenantAccountRole | None = field(default=None, init=False)
@@ -250,7 +250,9 @@ class Tenant(TypeBase):
     created_at: Mapped[datetime] = mapped_column(
         DateTime, server_default=func.current_timestamp(), nullable=False, init=False
     )
-    updated_at: Mapped[datetime] = mapped_column(DateTime, server_default=func.current_timestamp(), init=False)
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, server_default=func.current_timestamp(), init=False, onupdate=func.current_timestamp()
+    )
 
     def get_accounts(self) -> list[Account]:
         return list(
@@ -289,7 +291,7 @@ class TenantAccountJoin(TypeBase):
         DateTime, server_default=func.current_timestamp(), nullable=False, init=False
     )
     updated_at: Mapped[datetime] = mapped_column(
-        DateTime, server_default=func.current_timestamp(), nullable=False, init=False
+        DateTime, server_default=func.current_timestamp(), nullable=False, init=False, onupdate=func.current_timestamp()
     )
 
 
@@ -310,7 +312,7 @@ class AccountIntegrate(TypeBase):
         DateTime, server_default=func.current_timestamp(), nullable=False, init=False
     )
     updated_at: Mapped[datetime] = mapped_column(
-        DateTime, server_default=func.current_timestamp(), nullable=False, init=False
+        DateTime, server_default=func.current_timestamp(), nullable=False, init=False, onupdate=func.current_timestamp()
     )
 
 
@@ -396,5 +398,5 @@ class TenantPluginAutoUpgradeStrategy(TypeBase):
         DateTime, nullable=False, server_default=func.current_timestamp(), init=False
     )
     updated_at: Mapped[datetime] = mapped_column(
-        DateTime, nullable=False, server_default=func.current_timestamp(), init=False
+        DateTime, nullable=False, server_default=func.current_timestamp(), init=False, onupdate=func.current_timestamp()
     )

api/models/dataset.py

@@ -61,18 +61,20 @@ class Dataset(Base):
     created_by = mapped_column(StringUUID, nullable=False)
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
     updated_by = mapped_column(StringUUID, nullable=True)
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
-    embedding_model = mapped_column(db.String(255), nullable=True)
-    embedding_model_provider = mapped_column(db.String(255), nullable=True)
-    keyword_number = mapped_column(sa.Integer, nullable=True, server_default=db.text("10"))
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
+    embedding_model = mapped_column(sa.String(255), nullable=True)
+    embedding_model_provider = mapped_column(sa.String(255), nullable=True)
+    keyword_number = mapped_column(sa.Integer, nullable=True, server_default=sa.text("10"))
     collection_binding_id = mapped_column(StringUUID, nullable=True)
     retrieval_model = mapped_column(JSONB, nullable=True)
-    built_in_field_enabled = mapped_column(sa.Boolean, nullable=False, server_default=db.text("false"))
+    built_in_field_enabled = mapped_column(sa.Boolean, nullable=False, server_default=sa.text("false"))
     icon_info = mapped_column(JSONB, nullable=True)
-    runtime_mode = mapped_column(db.String(255), nullable=True, server_default=db.text("'general'::character varying"))
+    runtime_mode = mapped_column(sa.String(255), nullable=True, server_default=sa.text("'general'::character varying"))
     pipeline_id = mapped_column(StringUUID, nullable=True)
-    chunk_structure = mapped_column(db.String(255), nullable=True)
-    enable_api = mapped_column(sa.Boolean, nullable=False, server_default=db.text("true"))
+    chunk_structure = mapped_column(sa.String(255), nullable=True)
+    enable_api = mapped_column(sa.Boolean, nullable=False, server_default=sa.text("true"))
 
     @property
     def total_documents(self):
@@ -399,7 +401,9 @@ class Document(Base):
     archived_reason = mapped_column(String(255), nullable=True)
     archived_by = mapped_column(StringUUID, nullable=True)
     archived_at: Mapped[datetime | None] = mapped_column(DateTime, nullable=True)
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
     doc_type = mapped_column(String(40), nullable=True)
     doc_metadata = mapped_column(JSONB, nullable=True)
     doc_form = mapped_column(String(255), nullable=False, server_default=sa.text("'text_model'::character varying"))
@@ -716,7 +720,9 @@ class DocumentSegment(Base):
     created_by = mapped_column(StringUUID, nullable=False)
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
     updated_by = mapped_column(StringUUID, nullable=True)
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
     indexing_at: Mapped[datetime | None] = mapped_column(DateTime, nullable=True)
     completed_at: Mapped[datetime | None] = mapped_column(DateTime, nullable=True)
     error = mapped_column(sa.Text, nullable=True)
@@ -881,7 +887,7 @@ class ChildChunk(Base):
     )
     updated_by = mapped_column(StringUUID, nullable=True)
     updated_at: Mapped[datetime] = mapped_column(
-        DateTime, nullable=False, server_default=sa.text("CURRENT_TIMESTAMP(0)")
+        DateTime, nullable=False, server_default=sa.text("CURRENT_TIMESTAMP(0)"), onupdate=func.current_timestamp()
     )
     indexing_at: Mapped[datetime | None] = mapped_column(DateTime, nullable=True)
     completed_at: Mapped[datetime | None] = mapped_column(DateTime, nullable=True)
@@ -1036,8 +1042,8 @@ class TidbAuthBinding(Base):
     tenant_id = mapped_column(StringUUID, nullable=True)
     cluster_id: Mapped[str] = mapped_column(String(255), nullable=False)
     cluster_name: Mapped[str] = mapped_column(String(255), nullable=False)
-    active: Mapped[bool] = mapped_column(sa.Boolean, nullable=False, server_default=db.text("false"))
-    status = mapped_column(String(255), nullable=False, server_default=db.text("'CREATING'::character varying"))
+    active: Mapped[bool] = mapped_column(sa.Boolean, nullable=False, server_default=sa.text("false"))
+    status = mapped_column(String(255), nullable=False, server_default=sa.text("'CREATING'::character varying"))
     account: Mapped[str] = mapped_column(String(255), nullable=False)
     password: Mapped[str] = mapped_column(String(255), nullable=False)
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
@@ -1088,7 +1094,9 @@ class ExternalKnowledgeApis(Base):
     created_by = mapped_column(StringUUID, nullable=False)
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
     updated_by = mapped_column(StringUUID, nullable=True)
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
     def to_dict(self) -> dict[str, Any]:
         return {
@@ -1141,7 +1149,9 @@ class ExternalKnowledgeBindings(Base):
     created_by = mapped_column(StringUUID, nullable=False)
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
     updated_by = mapped_column(StringUUID, nullable=True)
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
 
 class DatasetAutoDisableLog(Base):
@@ -1197,7 +1207,7 @@ class DatasetMetadata(Base):
         DateTime, nullable=False, server_default=sa.text("CURRENT_TIMESTAMP(0)")
     )
     updated_at: Mapped[datetime] = mapped_column(
-        DateTime, nullable=False, server_default=sa.text("CURRENT_TIMESTAMP(0)")
+        DateTime, nullable=False, server_default=sa.text("CURRENT_TIMESTAMP(0)"), onupdate=func.current_timestamp()
     )
     created_by = mapped_column(StringUUID, nullable=False)
     updated_by = mapped_column(StringUUID, nullable=True)
@@ -1224,44 +1234,48 @@ class DatasetMetadataBinding(Base):
 
 class PipelineBuiltInTemplate(Base):  # type: ignore[name-defined]
     __tablename__ = "pipeline_built_in_templates"
-    __table_args__ = (db.PrimaryKeyConstraint("id", name="pipeline_built_in_template_pkey"),)
+    __table_args__ = (sa.PrimaryKeyConstraint("id", name="pipeline_built_in_template_pkey"),)
 
-    id = mapped_column(StringUUID, server_default=db.text("uuidv7()"))
-    name = mapped_column(db.String(255), nullable=False)
+    id = mapped_column(StringUUID, server_default=sa.text("uuidv7()"))
+    name = mapped_column(sa.String(255), nullable=False)
     description = mapped_column(sa.Text, nullable=False)
-    chunk_structure = mapped_column(db.String(255), nullable=False)
+    chunk_structure = mapped_column(sa.String(255), nullable=False)
     icon = mapped_column(sa.JSON, nullable=False)
     yaml_content = mapped_column(sa.Text, nullable=False)
-    copyright = mapped_column(db.String(255), nullable=False)
-    privacy_policy = mapped_column(db.String(255), nullable=False)
+    copyright = mapped_column(sa.String(255), nullable=False)
+    privacy_policy = mapped_column(sa.String(255), nullable=False)
     position = mapped_column(sa.Integer, nullable=False)
     install_count = mapped_column(sa.Integer, nullable=False, default=0)
-    language = mapped_column(db.String(255), nullable=False)
+    language = mapped_column(sa.String(255), nullable=False)
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
 
 class PipelineCustomizedTemplate(Base):  # type: ignore[name-defined]
     __tablename__ = "pipeline_customized_templates"
     __table_args__ = (
-        db.PrimaryKeyConstraint("id", name="pipeline_customized_template_pkey"),
-        db.Index("pipeline_customized_template_tenant_idx", "tenant_id"),
+        sa.PrimaryKeyConstraint("id", name="pipeline_customized_template_pkey"),
+        sa.Index("pipeline_customized_template_tenant_idx", "tenant_id"),
     )
 
-    id = mapped_column(StringUUID, server_default=db.text("uuidv7()"))
+    id = mapped_column(StringUUID, server_default=sa.text("uuidv7()"))
     tenant_id = mapped_column(StringUUID, nullable=False)
-    name = mapped_column(db.String(255), nullable=False)
+    name = mapped_column(sa.String(255), nullable=False)
     description = mapped_column(sa.Text, nullable=False)
-    chunk_structure = mapped_column(db.String(255), nullable=False)
+    chunk_structure = mapped_column(sa.String(255), nullable=False)
     icon = mapped_column(sa.JSON, nullable=False)
     position = mapped_column(sa.Integer, nullable=False)
     yaml_content = mapped_column(sa.Text, nullable=False)
     install_count = mapped_column(sa.Integer, nullable=False, default=0)
-    language = mapped_column(db.String(255), nullable=False)
+    language = mapped_column(sa.String(255), nullable=False)
     created_by = mapped_column(StringUUID, nullable=False)
     updated_by = mapped_column(StringUUID, nullable=True)
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
     @property
     def created_user_name(self):
@@ -1273,19 +1287,21 @@ class PipelineCustomizedTemplate(Base):  # type: ignore[name-defined]
 
 class Pipeline(Base):  # type: ignore[name-defined]
     __tablename__ = "pipelines"
-    __table_args__ = (db.PrimaryKeyConstraint("id", name="pipeline_pkey"),)
+    __table_args__ = (sa.PrimaryKeyConstraint("id", name="pipeline_pkey"),)
 
-    id = mapped_column(StringUUID, server_default=db.text("uuidv7()"))
+    id = mapped_column(StringUUID, server_default=sa.text("uuidv7()"))
     tenant_id: Mapped[str] = mapped_column(StringUUID, nullable=False)
-    name = mapped_column(db.String(255), nullable=False)
-    description = mapped_column(sa.Text, nullable=False, server_default=db.text("''::character varying"))
+    name = mapped_column(sa.String(255), nullable=False)
+    description = mapped_column(sa.Text, nullable=False, server_default=sa.text("''::character varying"))
     workflow_id = mapped_column(StringUUID, nullable=True)
-    is_public = mapped_column(sa.Boolean, nullable=False, server_default=db.text("false"))
-    is_published = mapped_column(sa.Boolean, nullable=False, server_default=db.text("false"))
+    is_public = mapped_column(sa.Boolean, nullable=False, server_default=sa.text("false"))
+    is_published = mapped_column(sa.Boolean, nullable=False, server_default=sa.text("false"))
     created_by = mapped_column(StringUUID, nullable=True)
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
     updated_by = mapped_column(StringUUID, nullable=True)
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
     def retrieve_dataset(self, session: Session):
         return session.query(Dataset).where(Dataset.pipeline_id == self.id).first()
@@ -1294,16 +1310,16 @@ class Pipeline(Base):  # type: ignore[name-defined]
 class DocumentPipelineExecutionLog(Base):
     __tablename__ = "document_pipeline_execution_logs"
     __table_args__ = (
-        db.PrimaryKeyConstraint("id", name="document_pipeline_execution_log_pkey"),
-        db.Index("document_pipeline_execution_logs_document_id_idx", "document_id"),
+        sa.PrimaryKeyConstraint("id", name="document_pipeline_execution_log_pkey"),
+        sa.Index("document_pipeline_execution_logs_document_id_idx", "document_id"),
     )
 
-    id = mapped_column(StringUUID, server_default=db.text("uuidv7()"))
+    id = mapped_column(StringUUID, server_default=sa.text("uuidv7()"))
     pipeline_id = mapped_column(StringUUID, nullable=False)
     document_id = mapped_column(StringUUID, nullable=False)
-    datasource_type = mapped_column(db.String(255), nullable=False)
+    datasource_type = mapped_column(sa.String(255), nullable=False)
     datasource_info = mapped_column(sa.Text, nullable=False)
-    datasource_node_id = mapped_column(db.String(255), nullable=False)
+    datasource_node_id = mapped_column(sa.String(255), nullable=False)
     input_data = mapped_column(sa.JSON, nullable=False)
     created_by = mapped_column(StringUUID, nullable=True)
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
@@ -1311,12 +1327,14 @@ class DocumentPipelineExecutionLog(Base):
 
 class PipelineRecommendedPlugin(Base):
     __tablename__ = "pipeline_recommended_plugins"
-    __table_args__ = (db.PrimaryKeyConstraint("id", name="pipeline_recommended_plugin_pkey"),)
+    __table_args__ = (sa.PrimaryKeyConstraint("id", name="pipeline_recommended_plugin_pkey"),)
 
-    id = mapped_column(StringUUID, server_default=db.text("uuidv7()"))
+    id = mapped_column(StringUUID, server_default=sa.text("uuidv7()"))
     plugin_id = mapped_column(sa.Text, nullable=False)
     provider_name = mapped_column(sa.Text, nullable=False)
     position = mapped_column(sa.Integer, nullable=False, default=0)
     active = mapped_column(sa.Boolean, nullable=False, default=True)
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )

api/models/model.py

@@ -3,6 +3,7 @@ import re
 import uuid
 from collections.abc import Mapping
 from datetime import datetime
+from decimal import Decimal
 from enum import StrEnum, auto
 from typing import TYPE_CHECKING, Any, Literal, Optional, cast
 
@@ -94,7 +95,9 @@ class App(Base):
     created_by = mapped_column(StringUUID, nullable=True)
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
     updated_by = mapped_column(StringUUID, nullable=True)
-    updated_at: Mapped[datetime] = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
     use_icon_as_answer_icon: Mapped[bool] = mapped_column(sa.Boolean, nullable=False, server_default=sa.text("false"))
 
     @property
@@ -313,7 +316,9 @@ class AppModelConfig(Base):
     created_by = mapped_column(StringUUID, nullable=True)
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
     updated_by = mapped_column(StringUUID, nullable=True)
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
     opening_statement = mapped_column(sa.Text)
     suggested_questions = mapped_column(sa.Text)
     suggested_questions_after_answer = mapped_column(sa.Text)
@@ -544,7 +549,9 @@ class RecommendedApp(Base):
     install_count: Mapped[int] = mapped_column(sa.Integer, nullable=False, default=0)
     language = mapped_column(String(255), nullable=False, server_default=sa.text("'en-US'::character varying"))
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
     @property
     def app(self) -> App | None:
@@ -643,7 +650,9 @@ class Conversation(Base):
     read_account_id = mapped_column(StringUUID)
     dialogue_count: Mapped[int] = mapped_column(default=0)
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
     messages = db.relationship("Message", backref="conversation", lazy="select", passive_deletes="all")
     message_annotations = db.relationship(
@@ -914,34 +923,42 @@ class Message(Base):
     )
 
     id: Mapped[str] = mapped_column(StringUUID, server_default=sa.text("uuid_generate_v4()"))
-    app_id = mapped_column(StringUUID, nullable=False)
-    model_provider = mapped_column(String(255), nullable=True)
-    model_id = mapped_column(String(255), nullable=True)
-    override_model_configs = mapped_column(sa.Text)
-    conversation_id = mapped_column(StringUUID, sa.ForeignKey("conversations.id"), nullable=False)
+    app_id: Mapped[str] = mapped_column(StringUUID, nullable=False)
+    model_provider: Mapped[str | None] = mapped_column(String(255), nullable=True)
+    model_id: Mapped[str | None] = mapped_column(String(255), nullable=True)
+    override_model_configs: Mapped[str | None] = mapped_column(sa.Text)
+    conversation_id: Mapped[str] = mapped_column(StringUUID, sa.ForeignKey("conversations.id"), nullable=False)
     _inputs: Mapped[dict[str, Any]] = mapped_column("inputs", sa.JSON)
     query: Mapped[str] = mapped_column(sa.Text, nullable=False)
-    message = mapped_column(sa.JSON, nullable=False)
+    message: Mapped[dict[str, Any]] = mapped_column(sa.JSON, nullable=False)
     message_tokens: Mapped[int] = mapped_column(sa.Integer, nullable=False, server_default=sa.text("0"))
-    message_unit_price = mapped_column(sa.Numeric(10, 4), nullable=False)
-    message_price_unit = mapped_column(sa.Numeric(10, 7), nullable=False, server_default=sa.text("0.001"))
+    message_unit_price: Mapped[Decimal] = mapped_column(sa.Numeric(10, 4), nullable=False)
+    message_price_unit: Mapped[Decimal] = mapped_column(
+        sa.Numeric(10, 7), nullable=False, server_default=sa.text("0.001")
+    )
     answer: Mapped[str] = mapped_column(sa.Text, nullable=False)
     answer_tokens: Mapped[int] = mapped_column(sa.Integer, nullable=False, server_default=sa.text("0"))
-    answer_unit_price = mapped_column(sa.Numeric(10, 4), nullable=False)
-    answer_price_unit = mapped_column(sa.Numeric(10, 7), nullable=False, server_default=sa.text("0.001"))
-    parent_message_id = mapped_column(StringUUID, nullable=True)
-    provider_response_latency = mapped_column(sa.Float, nullable=False, server_default=sa.text("0"))
-    total_price = mapped_column(sa.Numeric(10, 7))
+    answer_unit_price: Mapped[Decimal] = mapped_column(sa.Numeric(10, 4), nullable=False)
+    answer_price_unit: Mapped[Decimal] = mapped_column(
+        sa.Numeric(10, 7), nullable=False, server_default=sa.text("0.001")
+    )
+    parent_message_id: Mapped[str | None] = mapped_column(StringUUID, nullable=True)
+    provider_response_latency: Mapped[float] = mapped_column(sa.Float, nullable=False, server_default=sa.text("0"))
+    total_price: Mapped[Decimal | None] = mapped_column(sa.Numeric(10, 7))
     currency: Mapped[str] = mapped_column(String(255), nullable=False)
-    status = mapped_column(String(255), nullable=False, server_default=sa.text("'normal'::character varying"))
-    error = mapped_column(sa.Text)
-    message_metadata = mapped_column(sa.Text)
+    status: Mapped[str] = mapped_column(
+        String(255), nullable=False, server_default=sa.text("'normal'::character varying")
+    )
+    error: Mapped[str | None] = mapped_column(sa.Text)
+    message_metadata: Mapped[str | None] = mapped_column(sa.Text)
     invoke_from: Mapped[str | None] = mapped_column(String(255), nullable=True)
     from_source: Mapped[str] = mapped_column(String(255), nullable=False)
     from_end_user_id: Mapped[str | None] = mapped_column(StringUUID)
     from_account_id: Mapped[str | None] = mapped_column(StringUUID)
     created_at: Mapped[datetime] = mapped_column(sa.DateTime, server_default=func.current_timestamp())
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
     agent_based: Mapped[bool] = mapped_column(sa.Boolean, nullable=False, server_default=sa.text("false"))
     workflow_run_id: Mapped[str | None] = mapped_column(StringUUID)
     app_mode: Mapped[str | None] = mapped_column(String(255), nullable=True)
@@ -1212,9 +1229,13 @@ class Message(Base):
     @property
     def workflow_run(self):
         if self.workflow_run_id:
-            from .workflow import WorkflowRun
+            from sqlalchemy.orm import sessionmaker
 
-            return db.session.query(WorkflowRun).where(WorkflowRun.id == self.workflow_run_id).first()
+            from repositories.factory import DifyAPIRepositoryFactory
+
+            session_maker = sessionmaker(bind=db.engine, expire_on_commit=False)
+            repo = DifyAPIRepositoryFactory.create_api_workflow_run_repository(session_maker)
+            return repo.get_workflow_run_by_id_without_tenant(run_id=self.workflow_run_id)
 
         return None
 
@@ -1275,20 +1296,22 @@ class MessageFeedback(Base):
         sa.Index("message_feedback_conversation_idx", "conversation_id", "from_source", "rating"),
     )
 
-    id = mapped_column(StringUUID, server_default=sa.text("uuid_generate_v4()"))
-    app_id = mapped_column(StringUUID, nullable=False)
-    conversation_id = mapped_column(StringUUID, nullable=False)
-    message_id = mapped_column(StringUUID, nullable=False)
+    id: Mapped[str] = mapped_column(StringUUID, server_default=sa.text("uuid_generate_v4()"))
+    app_id: Mapped[str] = mapped_column(StringUUID, nullable=False)
+    conversation_id: Mapped[str] = mapped_column(StringUUID, nullable=False)
+    message_id: Mapped[str] = mapped_column(StringUUID, nullable=False)
     rating: Mapped[str] = mapped_column(String(255), nullable=False)
-    content = mapped_column(sa.Text)
+    content: Mapped[str | None] = mapped_column(sa.Text)
     from_source: Mapped[str] = mapped_column(String(255), nullable=False)
-    from_end_user_id = mapped_column(StringUUID)
-    from_account_id = mapped_column(StringUUID)
-    created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    from_end_user_id: Mapped[str | None] = mapped_column(StringUUID)
+    from_account_id: Mapped[str | None] = mapped_column(StringUUID)
+    created_at: Mapped[datetime] = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
     @property
-    def from_account(self):
+    def from_account(self) -> Account | None:
         account = db.session.query(Account).where(Account.id == self.from_account_id).first()
         return account
 
@@ -1367,7 +1390,9 @@ class MessageAnnotation(Base):
     hit_count: Mapped[int] = mapped_column(sa.Integer, nullable=False, server_default=sa.text("0"))
     account_id = mapped_column(StringUUID, nullable=False)
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
     @property
     def account(self):
@@ -1432,7 +1457,9 @@ class AppAnnotationSetting(Base):
     created_user_id = mapped_column(StringUUID, nullable=False)
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
     updated_user_id = mapped_column(StringUUID, nullable=False)
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
     @property
     def collection_binding_detail(self):
@@ -1460,7 +1487,9 @@ class OperationLog(Base):
     content = mapped_column(sa.JSON)
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
     created_ip: Mapped[str] = mapped_column(String(255), nullable=False)
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
 
 class DefaultEndUserSessionID(StrEnum):
@@ -1499,7 +1528,9 @@ class EndUser(Base, UserMixin):
 
     session_id: Mapped[str] = mapped_column()
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
 
 class AppMCPServer(Base):
@@ -1519,7 +1550,9 @@ class AppMCPServer(Base):
     parameters = mapped_column(sa.Text, nullable=False)
 
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
     @staticmethod
     def generate_server_code(n: int) -> str:
@@ -1565,7 +1598,9 @@ class Site(Base):
     created_by = mapped_column(StringUUID, nullable=True)
     created_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
     updated_by = mapped_column(StringUUID, nullable=True)
-    updated_at = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
     code = mapped_column(String(255))
 
     @property

api/models/oauth.py

@@ -1,62 +1,66 @@
 from datetime import datetime
 
 import sqlalchemy as sa
+from sqlalchemy import func
 from sqlalchemy.dialects.postgresql import JSONB
 from sqlalchemy.orm import Mapped, mapped_column
 
 from .base import Base
-from .engine import db
 from .types import StringUUID
 
 
 class DatasourceOauthParamConfig(Base):  # type: ignore[name-defined]
     __tablename__ = "datasource_oauth_params"
     __table_args__ = (
-        db.PrimaryKeyConstraint("id", name="datasource_oauth_config_pkey"),
-        db.UniqueConstraint("plugin_id", "provider", name="datasource_oauth_config_datasource_id_provider_idx"),
+        sa.PrimaryKeyConstraint("id", name="datasource_oauth_config_pkey"),
+        sa.UniqueConstraint("plugin_id", "provider", name="datasource_oauth_config_datasource_id_provider_idx"),
     )
 
-    id = mapped_column(StringUUID, server_default=db.text("uuidv7()"))
-    plugin_id: Mapped[str] = mapped_column(db.String(255), nullable=False)
-    provider: Mapped[str] = mapped_column(db.String(255), nullable=False)
+    id = mapped_column(StringUUID, server_default=sa.text("uuidv7()"))
+    plugin_id: Mapped[str] = mapped_column(sa.String(255), nullable=False)
+    provider: Mapped[str] = mapped_column(sa.String(255), nullable=False)
     system_credentials: Mapped[dict] = mapped_column(JSONB, nullable=False)
 
 
 class DatasourceProvider(Base):
     __tablename__ = "datasource_providers"
     __table_args__ = (
-        db.PrimaryKeyConstraint("id", name="datasource_provider_pkey"),
-        db.UniqueConstraint("tenant_id", "plugin_id", "provider", "name", name="datasource_provider_unique_name"),
-        db.Index("datasource_provider_auth_type_provider_idx", "tenant_id", "plugin_id", "provider"),
+        sa.PrimaryKeyConstraint("id", name="datasource_provider_pkey"),
+        sa.UniqueConstraint("tenant_id", "plugin_id", "provider", "name", name="datasource_provider_unique_name"),
+        sa.Index("datasource_provider_auth_type_provider_idx", "tenant_id", "plugin_id", "provider"),
     )
-    id = mapped_column(StringUUID, server_default=db.text("uuidv7()"))
+    id = mapped_column(StringUUID, server_default=sa.text("uuidv7()"))
     tenant_id = mapped_column(StringUUID, nullable=False)
-    name: Mapped[str] = mapped_column(db.String(255), nullable=False)
-    provider: Mapped[str] = mapped_column(db.String(255), nullable=False)
-    plugin_id: Mapped[str] = mapped_column(db.String(255), nullable=False)
-    auth_type: Mapped[str] = mapped_column(db.String(255), nullable=False)
+    name: Mapped[str] = mapped_column(sa.String(255), nullable=False)
+    provider: Mapped[str] = mapped_column(sa.String(255), nullable=False)
+    plugin_id: Mapped[str] = mapped_column(sa.String(255), nullable=False)
+    auth_type: Mapped[str] = mapped_column(sa.String(255), nullable=False)
     encrypted_credentials: Mapped[dict] = mapped_column(JSONB, nullable=False)
     avatar_url: Mapped[str] = mapped_column(sa.Text, nullable=True, default="default")
-    is_default: Mapped[bool] = mapped_column(sa.Boolean, nullable=False, server_default=db.text("false"))
+    is_default: Mapped[bool] = mapped_column(sa.Boolean, nullable=False, server_default=sa.text("false"))
     expires_at: Mapped[int] = mapped_column(sa.Integer, nullable=False, server_default="-1")
 
-    created_at: Mapped[datetime] = mapped_column(sa.DateTime, nullable=False, default=datetime.now)
-    updated_at: Mapped[datetime] = mapped_column(sa.DateTime, nullable=False, default=datetime.now)
+    created_at: Mapped[datetime] = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
 
 class DatasourceOauthTenantParamConfig(Base):
     __tablename__ = "datasource_oauth_tenant_params"
     __table_args__ = (
-        db.PrimaryKeyConstraint("id", name="datasource_oauth_tenant_config_pkey"),
-        db.UniqueConstraint("tenant_id", "plugin_id", "provider", name="datasource_oauth_tenant_config_unique"),
+        sa.PrimaryKeyConstraint("id", name="datasource_oauth_tenant_config_pkey"),
+        sa.UniqueConstraint("tenant_id", "plugin_id", "provider", name="datasource_oauth_tenant_config_unique"),
     )
 
-    id = mapped_column(StringUUID, server_default=db.text("uuidv7()"))
+    id = mapped_column(StringUUID, server_default=sa.text("uuidv7()"))
     tenant_id = mapped_column(StringUUID, nullable=False)
-    provider: Mapped[str] = mapped_column(db.String(255), nullable=False)
-    plugin_id: Mapped[str] = mapped_column(db.String(255), nullable=False)
+    provider: Mapped[str] = mapped_column(sa.String(255), nullable=False)
+    plugin_id: Mapped[str] = mapped_column(sa.String(255), nullable=False)
     client_params: Mapped[dict] = mapped_column(JSONB, nullable=False, default={})
     enabled: Mapped[bool] = mapped_column(sa.Boolean, nullable=False, default=False)
 
-    created_at: Mapped[datetime] = mapped_column(sa.DateTime, nullable=False, default=datetime.now)
-    updated_at: Mapped[datetime] = mapped_column(sa.DateTime, nullable=False, default=datetime.now)
+    created_at: Mapped[datetime] = mapped_column(sa.DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        sa.DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )

api/models/provider.py

@@ -72,7 +72,9 @@ class Provider(Base):
     quota_used: Mapped[int | None] = mapped_column(sa.BigInteger, default=0)
 
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
     def __repr__(self):
         return (
@@ -135,7 +137,9 @@ class ProviderModel(Base):
     credential_id: Mapped[str | None] = mapped_column(StringUUID, nullable=True)
     is_valid: Mapped[bool] = mapped_column(sa.Boolean, nullable=False, server_default=text("false"))
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
     @cached_property
     def credential(self):
@@ -170,7 +174,9 @@ class TenantDefaultModel(Base):
     model_name: Mapped[str] = mapped_column(String(255), nullable=False)
     model_type: Mapped[str] = mapped_column(String(40), nullable=False)
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
 
 class TenantPreferredModelProvider(Base):
@@ -185,7 +191,9 @@ class TenantPreferredModelProvider(Base):
     provider_name: Mapped[str] = mapped_column(String(255), nullable=False)
     preferred_provider_type: Mapped[str] = mapped_column(String(40), nullable=False)
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
 
 class ProviderOrder(Base):
@@ -212,7 +220,9 @@ class ProviderOrder(Base):
     pay_failed_at: Mapped[datetime | None] = mapped_column(DateTime)
     refunded_at: Mapped[datetime | None] = mapped_column(DateTime)
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
 
 class ProviderModelSetting(Base):
@@ -234,7 +244,9 @@ class ProviderModelSetting(Base):
     enabled: Mapped[bool] = mapped_column(sa.Boolean, nullable=False, server_default=text("true"))
     load_balancing_enabled: Mapped[bool] = mapped_column(sa.Boolean, nullable=False, server_default=text("false"))
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
 
 class LoadBalancingModelConfig(Base):
@@ -259,7 +271,9 @@ class LoadBalancingModelConfig(Base):
     credential_source_type: Mapped[str | None] = mapped_column(String(40), nullable=True)
     enabled: Mapped[bool] = mapped_column(sa.Boolean, nullable=False, server_default=text("true"))
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
 
 class ProviderCredential(Base):
@@ -279,7 +293,9 @@ class ProviderCredential(Base):
     credential_name: Mapped[str] = mapped_column(String(255), nullable=False)
     encrypted_config: Mapped[str] = mapped_column(sa.Text, nullable=False)
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )
 
 
 class ProviderModelCredential(Base):
@@ -307,4 +323,6 @@ class ProviderModelCredential(Base):
     credential_name: Mapped[str] = mapped_column(String(255), nullable=False)
     encrypted_config: Mapped[str] = mapped_column(sa.Text, nullable=False)
     created_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
-    updated_at: Mapped[datetime] = mapped_column(DateTime, nullable=False, server_default=func.current_timestamp())
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime, nullable=False, server_default=func.current_timestamp(), onupdate=func.current_timestamp()
+    )

api/models/workflow.py

@@ -140,8 +140,9 @@ class Workflow(Base):
     updated_at: Mapped[datetime] = mapped_column(
         DateTime,
         nullable=False,
-        default=naive_utc_now(),
-        server_onupdate=func.current_timestamp(),
+        default=func.current_timestamp(),
+        server_default=func.current_timestamp(),
+        onupdate=func.current_timestamp(),
     )
     _environment_variables: Mapped[str] = mapped_column(
         "environment_variables", sa.Text, nullable=False, server_default="{}"
@@ -150,7 +151,7 @@ class Workflow(Base):
         "conversation_variables", sa.Text, nullable=False, server_default="{}"
     )
     _rag_pipeline_variables: Mapped[str] = mapped_column(
-        "rag_pipeline_variables", db.Text, nullable=False, server_default="{}"
+        "rag_pipeline_variables", sa.Text, nullable=False, server_default="{}"
     )
 
     VERSION_DRAFT = "draft"
@@ -1058,7 +1059,16 @@ class WorkflowAppLog(Base):
 
     @property
     def workflow_run(self):
-        return db.session.get(WorkflowRun, self.workflow_run_id)
+        if self.workflow_run_id:
+            from sqlalchemy.orm import sessionmaker
+
+            from repositories.factory import DifyAPIRepositoryFactory
+
+            session_maker = sessionmaker(bind=db.engine, expire_on_commit=False)
+            repo = DifyAPIRepositoryFactory.create_api_workflow_run_repository(session_maker)
+            return repo.get_workflow_run_by_id_without_tenant(run_id=self.workflow_run_id)
+
+        return None
 
     @property
     def created_by_account(self):

api/pyproject.toml

@@ -37,6 +37,7 @@ dependencies = [
     "numpy~=1.26.4",
     "openpyxl~=3.1.5",
     "opik~=1.8.72",
+    "litellm==1.77.1",  # Pinned to avoid madoka dependency issue
     "opentelemetry-api==1.27.0",
     "opentelemetry-distro==0.48b0",
     "opentelemetry-exporter-otlp==1.27.0",
@@ -209,9 +210,9 @@ vdb = [
     "pgvector==0.2.5",
     "pymilvus~=2.5.0",
     "pymochow==2.2.9",
-    "pyobvector~=0.2.15",
+    "pyobvector~=0.2.17",
     "qdrant-client==1.9.0",
-    "tablestore==6.2.0",
+    "tablestore==6.3.7",
     "tcvectordb~=1.6.4",
     "tidb-vector==0.0.9",
     "upstash-vector==0.6.0",

api/schedule/clean_messages.py

@@ -7,6 +7,7 @@ from sqlalchemy.exc import SQLAlchemyError
 
 import app
 from configs import dify_config
+from enums.cloud_plan import CloudPlan
 from extensions.ext_database import db
 from extensions.ext_redis import redis_client
 from models.model import (
@@ -63,7 +64,7 @@ def clean_messages():
                 plan = features.billing.subscription.plan
             else:
                 plan = plan_cache.decode()
-            if plan == "sandbox":
+            if plan == CloudPlan.SANDBOX:
                 # clean related message
                 db.session.query(MessageFeedback).where(MessageFeedback.message_id == message.id).delete(
                     synchronize_session=False

api/schedule/clean_unused_datasets_task.py

@@ -9,6 +9,7 @@ from sqlalchemy.exc import SQLAlchemyError
 import app
 from configs import dify_config
 from core.rag.index_processor.index_processor_factory import IndexProcessorFactory
+from enums.cloud_plan import CloudPlan
 from extensions.ext_database import db
 from extensions.ext_redis import redis_client
 from models.dataset import Dataset, DatasetAutoDisableLog, DatasetQuery, Document
@@ -35,7 +36,7 @@ def clean_unused_datasets_task():
         },
         {
             "clean_day": datetime.datetime.now() - datetime.timedelta(days=dify_config.PLAN_PRO_CLEAN_DAY_SETTING),
-            "plan_filter": "sandbox",
+            "plan_filter": CloudPlan.SANDBOX,
             "add_logs": False,
         },
     ]

api/schedule/mail_clean_document_notify_task.py

@@ -7,6 +7,7 @@ from sqlalchemy import select
 
 import app
 from configs import dify_config
+from enums.cloud_plan import CloudPlan
 from extensions.ext_database import db
 from extensions.ext_mail import mail
 from libs.email_i18n import EmailType, get_email_i18n_service
@@ -45,7 +46,7 @@ def mail_clean_document_notify_task():
         for tenant_id, tenant_dataset_auto_disable_logs in dataset_auto_disable_logs_map.items():
             features = FeatureService.get_features(tenant_id)
             plan = features.billing.subscription.plan
-            if plan != "sandbox":
+            if plan != CloudPlan.SANDBOX:
                 knowledge_details = []
                 # check tenant
                 tenant = db.session.query(Tenant).where(Tenant.id == tenant_id).first()

api/services/annotation_service.py

@@ -32,41 +32,48 @@ class AppAnnotationService:
 
         if not app:
             raise NotFound("App not found")
+
+        answer = args.get("answer") or args.get("content")
+        if answer is None:
+            raise ValueError("Either 'answer' or 'content' must be provided")
+
         if args.get("message_id"):
             message_id = str(args["message_id"])
-            # get message info
             message = db.session.query(Message).where(Message.id == message_id, Message.app_id == app.id).first()
 
             if not message:
                 raise NotFound("Message Not Exists.")
 
+            question = args.get("question") or message.query or ""
+
             annotation: MessageAnnotation | None = message.annotation
-            # save the message annotation
             if annotation:
-                annotation.content = args["answer"]
-                annotation.question = args["question"]
+                annotation.content = answer
+                annotation.question = question
             else:
                 annotation = MessageAnnotation(
                     app_id=app.id,
                     conversation_id=message.conversation_id,
                     message_id=message.id,
-                    content=args["answer"],
-                    question=args["question"],
+                    content=answer,
+                    question=question,
                     account_id=current_user.id,
                 )
         else:
-            annotation = MessageAnnotation(
-                app_id=app.id, content=args["answer"], question=args["question"], account_id=current_user.id
-            )
+            question = args.get("question")
+            if not question:
+                raise ValueError("'question' is required when 'message_id' is not provided")
+
+            annotation = MessageAnnotation(app_id=app.id, content=answer, question=question, account_id=current_user.id)
         db.session.add(annotation)
         db.session.commit()
-        # if annotation reply is enabled , add annotation to index
+
         annotation_setting = db.session.query(AppAnnotationSetting).where(AppAnnotationSetting.app_id == app_id).first()
         assert current_tenant_id is not None
         if annotation_setting:
             add_annotation_to_index_task.delay(
                 annotation.id,
-                args["question"],
+                annotation.question,
                 current_tenant_id,
                 app_id,
                 annotation_setting.collection_binding_id,

api/services/app_generate_service.py

@@ -10,6 +10,7 @@ from core.app.apps.completion.app_generator import CompletionAppGenerator
 from core.app.apps.workflow.app_generator import WorkflowAppGenerator
 from core.app.entities.app_invoke_entities import InvokeFrom
 from core.app.features.rate_limiting import RateLimit
+from enums.cloud_plan import CloudPlan
 from libs.helper import RateLimiter
 from models.model import Account, App, AppMode, EndUser
 from models.workflow import Workflow
@@ -44,7 +45,7 @@ class AppGenerateService:
         if dify_config.BILLING_ENABLED:
             # check if it's free plan
             limit_info = BillingService.get_info(app_model.tenant_id)
-            if limit_info["subscription"]["plan"] == "sandbox":
+            if limit_info["subscription"]["plan"] == CloudPlan.SANDBOX:
                 if cls.system_rate_limiter.is_rate_limited(app_model.tenant_id):
                     raise InvokeRateLimitError(
                         "Rate limit exceeded, please upgrade your plan "

api/services/billing_service.py

@@ -4,6 +4,7 @@ from typing import Literal
 import httpx
 from tenacity import retry, retry_if_exception_type, stop_before_delay, wait_fixed
 
+from enums.cloud_plan import CloudPlan
 from extensions.ext_database import db
 from extensions.ext_redis import redis_client
 from libs.helper import RateLimiter
@@ -31,7 +32,7 @@ class BillingService:
 
         return {
             "limit": knowledge_rate_limit.get("limit", 10),
-            "subscription_plan": knowledge_rate_limit.get("subscription_plan", "sandbox"),
+            "subscription_plan": knowledge_rate_limit.get("subscription_plan", CloudPlan.SANDBOX),
         }
 
     @classmethod

api/services/clear_free_plan_tenant_expired_logs.py

@@ -11,6 +11,7 @@ from sqlalchemy.orm import Session, sessionmaker
 
 from configs import dify_config
 from core.model_runtime.utils.encoders import jsonable_encoder
+from enums.cloud_plan import CloudPlan
 from extensions.ext_database import db
 from extensions.ext_storage import storage
 from models.account import Tenant
@@ -358,7 +359,7 @@ class ClearFreePlanTenantExpiredLogs:
             try:
                 if (
                     not dify_config.BILLING_ENABLED
-                    or BillingService.get_info(tenant_id)["subscription"]["plan"] == "sandbox"
+                    or BillingService.get_info(tenant_id)["subscription"]["plan"] == CloudPlan.SANDBOX
                 ):
                     # only process sandbox tenant
                     cls.process_tenant(flask_app, tenant_id, days, batch)

api/services/dataset_service.py

@@ -22,6 +22,7 @@ from core.model_runtime.entities.model_entities import ModelType
 from core.rag.index_processor.constant.built_in_field import BuiltInField
 from core.rag.index_processor.constant.index_type import IndexType
 from core.rag.retrieval.retrieval_methods import RetrievalMethod
+from enums.cloud_plan import CloudPlan
 from events.dataset_event import dataset_was_deleted
 from events.document_event import document_was_deleted
 from extensions.ext_database import db
@@ -49,6 +50,7 @@ from models.model import UploadFile
 from models.provider_ids import ModelProviderID
 from models.source import DataSourceOauthBinding
 from models.workflow import Workflow
+from services.document_indexing_task_proxy import DocumentIndexingTaskProxy
 from services.entities.knowledge_entities.knowledge_entities import (
     ChildChunkUpdateArgs,
     KnowledgeConfig,
@@ -78,7 +80,6 @@ from tasks.deal_dataset_vector_index_task import deal_dataset_vector_index_task
 from tasks.delete_segment_from_index_task import delete_segment_from_index_task
 from tasks.disable_segment_from_index_task import disable_segment_from_index_task
 from tasks.disable_segments_from_index_task import disable_segments_from_index_task
-from tasks.document_indexing_task import document_indexing_task
 from tasks.document_indexing_update_task import document_indexing_update_task
 from tasks.duplicate_document_indexing_task import duplicate_document_indexing_task
 from tasks.enable_segments_to_index_task import enable_segments_to_index_task
@@ -1042,7 +1043,7 @@ class DatasetService:
         assert isinstance(current_user, Account)
         assert current_user.current_tenant_id is not None
         features = FeatureService.get_features(current_user.current_tenant_id)
-        if not features.billing.enabled or features.billing.subscription.plan == "sandbox":
+        if not features.billing.enabled or features.billing.subscription.plan == CloudPlan.SANDBOX:
             return {
                 "document_ids": [],
                 "count": 0,
@@ -1438,7 +1439,7 @@ class DocumentService:
                         count = len(website_info.urls)
                     batch_upload_limit = int(dify_config.BATCH_UPLOAD_LIMIT)
 
-                    if features.billing.subscription.plan == "sandbox" and count > 1:
+                    if features.billing.subscription.plan == CloudPlan.SANDBOX and count > 1:
                         raise ValueError("Your current plan does not support batch upload, please upgrade your plan.")
                     if count > batch_upload_limit:
                         raise ValueError(f"You have reached the batch upload limit of {batch_upload_limit}.")
@@ -1693,7 +1694,7 @@ class DocumentService:
 
                 # trigger async task
                 if document_ids:
-                    document_indexing_task.delay(dataset.id, document_ids)
+                    DocumentIndexingTaskProxy(dataset.tenant_id, dataset.id, document_ids).delay()
                 if duplicate_document_ids:
                     duplicate_document_indexing_task.delay(dataset.id, duplicate_document_ids)
 
@@ -1727,7 +1728,7 @@ class DocumentService:
     #                     count = len(website_info.urls)  # type: ignore
     #                 batch_upload_limit = int(dify_config.BATCH_UPLOAD_LIMIT)
 
-    #                 if features.billing.subscription.plan == "sandbox" and count > 1:
+    #                 if features.billing.subscription.plan == CloudPlan.SANDBOX and count > 1:
     #                     raise ValueError("Your current plan does not support batch upload, please upgrade your plan.")
     #                 if count > batch_upload_limit:
     #                     raise ValueError(f"You have reached the batch upload limit of {batch_upload_limit}.")
@@ -2196,7 +2197,7 @@ class DocumentService:
                 website_info = knowledge_config.data_source.info_list.website_info_list
                 if website_info:
                     count = len(website_info.urls)
-            if features.billing.subscription.plan == "sandbox" and count > 1:
+            if features.billing.subscription.plan == CloudPlan.SANDBOX and count > 1:
                 raise ValueError("Your current plan does not support batch upload, please upgrade your plan.")
             batch_upload_limit = int(dify_config.BATCH_UPLOAD_LIMIT)
             if count > batch_upload_limit:

api/services/document_indexing_task_proxy.py

@@ -0,0 +1,83 @@
+import logging
+from collections.abc import Callable, Sequence
+from dataclasses import asdict
+from functools import cached_property
+
+from core.entities.document_task import DocumentTask
+from core.rag.pipeline.queue import TenantIsolatedTaskQueue
+from enums.cloud_plan import CloudPlan
+from services.feature_service import FeatureService
+from tasks.document_indexing_task import normal_document_indexing_task, priority_document_indexing_task
+
+logger = logging.getLogger(__name__)
+
+
+class DocumentIndexingTaskProxy:
+    def __init__(self, tenant_id: str, dataset_id: str, document_ids: Sequence[str]):
+        self._tenant_id = tenant_id
+        self._dataset_id = dataset_id
+        self._document_ids = document_ids
+        self._tenant_isolated_task_queue = TenantIsolatedTaskQueue(tenant_id, "document_indexing")
+
+    @cached_property
+    def features(self):
+        return FeatureService.get_features(self._tenant_id)
+
+    def _send_to_direct_queue(self, task_func: Callable[[str, str, Sequence[str]], None]):
+        logger.info("send dataset %s to direct queue", self._dataset_id)
+        task_func.delay(  # type: ignore
+            tenant_id=self._tenant_id, dataset_id=self._dataset_id, document_ids=self._document_ids
+        )
+
+    def _send_to_tenant_queue(self, task_func: Callable[[str, str, Sequence[str]], None]):
+        logger.info("send dataset %s to tenant queue", self._dataset_id)
+        if self._tenant_isolated_task_queue.get_task_key():
+            # Add to waiting queue using List operations (lpush)
+            self._tenant_isolated_task_queue.push_tasks(
+                [
+                    asdict(
+                        DocumentTask(
+                            tenant_id=self._tenant_id, dataset_id=self._dataset_id, document_ids=self._document_ids
+                        )
+                    )
+                ]
+            )
+            logger.info("push tasks: %s - %s", self._dataset_id, self._document_ids)
+        else:
+            # Set flag and execute task
+            self._tenant_isolated_task_queue.set_task_waiting_time()
+            task_func.delay(  # type: ignore
+                tenant_id=self._tenant_id, dataset_id=self._dataset_id, document_ids=self._document_ids
+            )
+            logger.info("init tasks: %s - %s", self._dataset_id, self._document_ids)
+
+    def _send_to_default_tenant_queue(self):
+        self._send_to_tenant_queue(normal_document_indexing_task)
+
+    def _send_to_priority_tenant_queue(self):
+        self._send_to_tenant_queue(priority_document_indexing_task)
+
+    def _send_to_priority_direct_queue(self):
+        self._send_to_direct_queue(priority_document_indexing_task)
+
+    def _dispatch(self):
+        logger.info(
+            "dispatch args: %s - %s - %s",
+            self._tenant_id,
+            self.features.billing.enabled,
+            self.features.billing.subscription.plan,
+        )
+        # dispatch to different indexing queue with tenant isolation when billing enabled
+        if self.features.billing.enabled:
+            if self.features.billing.subscription.plan == CloudPlan.SANDBOX:
+                # dispatch to normal pipeline queue with tenant self sub queue for sandbox plan
+                self._send_to_default_tenant_queue()
+            else:
+                # dispatch to priority pipeline queue with tenant self sub queue for other plans
+                self._send_to_priority_tenant_queue()
+        else:
+            # dispatch to priority queue without tenant isolation for others, e.g.: self-hosted or enterprise
+            self._send_to_priority_direct_queue()
+
+    def delay(self):
+        self._dispatch()

api/services/enterprise/enterprise_service.py

@@ -110,5 +110,5 @@ class EnterpriseService:
             if not app_id:
                 raise ValueError("app_id must be provided.")
 
-            body = {"appId": app_id}
-            EnterpriseRequest.send_request("DELETE", "/webapp/clean", json=body)
+            params = {"appId": app_id}
+            EnterpriseRequest.send_request("DELETE", "/webapp/clean", params=params)

api/services/errors/file.py

@@ -11,3 +11,7 @@ class FileTooLargeError(BaseServiceError):
 
 class UnsupportedFileTypeError(BaseServiceError):
     pass
+
+
+class BlockedFileExtensionError(BaseServiceError):
+    description = "File extension '{extension}' is not allowed for security reasons"

api/services/feature_service.py

@@ -3,12 +3,13 @@ from enum import StrEnum
 from pydantic import BaseModel, ConfigDict, Field
 
 from configs import dify_config
+from enums.cloud_plan import CloudPlan
 from services.billing_service import BillingService
 from services.enterprise.enterprise_service import EnterpriseService
 
 
 class SubscriptionModel(BaseModel):
-    plan: str = "sandbox"
+    plan: str = CloudPlan.SANDBOX
     interval: str = ""
 
 
@@ -186,7 +187,7 @@ class FeatureService:
             knowledge_rate_limit.enabled = True
             limit_info = BillingService.get_knowledge_rate_limit(tenant_id)
             knowledge_rate_limit.limit = limit_info.get("limit", 10)
-            knowledge_rate_limit.subscription_plan = limit_info.get("subscription_plan", "sandbox")
+            knowledge_rate_limit.subscription_plan = limit_info.get("subscription_plan", CloudPlan.SANDBOX)
         return knowledge_rate_limit
 
     @classmethod
@@ -240,7 +241,7 @@ class FeatureService:
         features.billing.subscription.interval = billing_info["subscription"]["interval"]
         features.education.activated = billing_info["subscription"].get("education", False)
 
-        if features.billing.subscription.plan != "sandbox":
+        if features.billing.subscription.plan != CloudPlan.SANDBOX:
             features.webapp_copyright_enabled = True
         else:
             features.is_allow_transfer_workspace = False

api/services/file_service.py

@@ -23,7 +23,7 @@ from models import Account
 from models.enums import CreatorUserRole
 from models.model import EndUser, UploadFile
 
-from .errors.file import FileTooLargeError, UnsupportedFileTypeError
+from .errors.file import BlockedFileExtensionError, FileTooLargeError, UnsupportedFileTypeError
 
 PREVIEW_WORDS_LIMIT = 3000
 
@@ -59,6 +59,10 @@ class FileService:
         if len(filename) > 200:
             filename = filename.split(".")[0][:200] + "." + extension
 
+        # check if extension is in blacklist
+        if extension and extension in dify_config.UPLOAD_FILE_EXTENSION_BLACKLIST:
+            raise BlockedFileExtensionError(f"File extension '.{extension}' is not allowed for security reasons")
+
         if source == "datasets" and extension not in DOCUMENT_EXTENSIONS:
             raise UnsupportedFileTypeError()
 

api/services/rag_pipeline/rag_pipeline_task_proxy.py

@@ -0,0 +1,106 @@
+import json
+import logging
+from collections.abc import Callable, Sequence
+from functools import cached_property
+
+from core.app.entities.rag_pipeline_invoke_entities import RagPipelineInvokeEntity
+from core.rag.pipeline.queue import TenantIsolatedTaskQueue
+from enums.cloud_plan import CloudPlan
+from extensions.ext_database import db
+from services.feature_service import FeatureService
+from services.file_service import FileService
+from tasks.rag_pipeline.priority_rag_pipeline_run_task import priority_rag_pipeline_run_task
+from tasks.rag_pipeline.rag_pipeline_run_task import rag_pipeline_run_task
+
+logger = logging.getLogger(__name__)
+
+
+class RagPipelineTaskProxy:
+    # Default uploaded file name for rag pipeline invoke entities
+    _RAG_PIPELINE_INVOKE_ENTITIES_FILE_NAME = "rag_pipeline_invoke_entities.json"
+
+    def __init__(
+        self, dataset_tenant_id: str, user_id: str, rag_pipeline_invoke_entities: Sequence[RagPipelineInvokeEntity]
+    ):
+        self._dataset_tenant_id = dataset_tenant_id
+        self._user_id = user_id
+        self._rag_pipeline_invoke_entities = rag_pipeline_invoke_entities
+        self._tenant_isolated_task_queue = TenantIsolatedTaskQueue(dataset_tenant_id, "pipeline")
+
+    @cached_property
+    def features(self):
+        return FeatureService.get_features(self._dataset_tenant_id)
+
+    def _upload_invoke_entities(self) -> str:
+        text = [item.model_dump() for item in self._rag_pipeline_invoke_entities]
+        # Convert list to proper JSON string
+        json_text = json.dumps(text)
+        upload_file = FileService(db.engine).upload_text(
+            json_text, self._RAG_PIPELINE_INVOKE_ENTITIES_FILE_NAME, self._user_id, self._dataset_tenant_id
+        )
+        return upload_file.id
+
+    def _send_to_direct_queue(self, upload_file_id: str, task_func: Callable[[str, str], None]):
+        logger.info("send file %s to direct queue", upload_file_id)
+        task_func.delay(  # type: ignore
+            rag_pipeline_invoke_entities_file_id=upload_file_id,
+            tenant_id=self._dataset_tenant_id,
+        )
+
+    def _send_to_tenant_queue(self, upload_file_id: str, task_func: Callable[[str, str], None]):
+        logger.info("send file %s to tenant queue", upload_file_id)
+        if self._tenant_isolated_task_queue.get_task_key():
+            # Add to waiting queue using List operations (lpush)
+            self._tenant_isolated_task_queue.push_tasks([upload_file_id])
+            logger.info("push tasks: %s", upload_file_id)
+        else:
+            # Set flag and execute task
+            self._tenant_isolated_task_queue.set_task_waiting_time()
+            task_func.delay(  # type: ignore
+                rag_pipeline_invoke_entities_file_id=upload_file_id,
+                tenant_id=self._dataset_tenant_id,
+            )
+            logger.info("init tasks: %s", upload_file_id)
+
+    def _send_to_default_tenant_queue(self, upload_file_id: str):
+        self._send_to_tenant_queue(upload_file_id, rag_pipeline_run_task)
+
+    def _send_to_priority_tenant_queue(self, upload_file_id: str):
+        self._send_to_tenant_queue(upload_file_id, priority_rag_pipeline_run_task)
+
+    def _send_to_priority_direct_queue(self, upload_file_id: str):
+        self._send_to_direct_queue(upload_file_id, priority_rag_pipeline_run_task)
+
+    def _dispatch(self):
+        upload_file_id = self._upload_invoke_entities()
+        if not upload_file_id:
+            raise ValueError("upload_file_id is empty")
+
+        logger.info(
+            "dispatch args: %s - %s - %s",
+            self._dataset_tenant_id,
+            self.features.billing.enabled,
+            self.features.billing.subscription.plan,
+        )
+
+        # dispatch to different pipeline queue with tenant isolation when billing enabled
+        if self.features.billing.enabled:
+            if self.features.billing.subscription.plan == CloudPlan.SANDBOX:
+                # dispatch to normal pipeline queue with tenant isolation for sandbox plan
+                self._send_to_default_tenant_queue(upload_file_id)
+            else:
+                # dispatch to priority pipeline queue with tenant isolation for other plans
+                self._send_to_priority_tenant_queue(upload_file_id)
+        else:
+            # dispatch to priority pipeline queue without tenant isolation for others, e.g.: self-hosted or enterprise
+            self._send_to_priority_direct_queue(upload_file_id)
+
+    def delay(self):
+        if not self._rag_pipeline_invoke_entities:
+            logger.warning(
+                "Received empty rag pipeline invoke entities, no tasks delivered: %s %s",
+                self._dataset_tenant_id,
+                self._user_id,
+            )
+            return
+        self._dispatch()

api/services/rag_pipeline/transform/website-crawl-general-economy.yml

@@ -126,7 +126,7 @@ workflow:
             type: mixed
             value: '{{#rag.1752491761974.jina_use_sitemap#}}'
         plugin_id: langgenius/jina_datasource
-        provider_name: jina
+        provider_name: jinareader
         provider_type: website_crawl
         selected: false
         title: Jina Reader

api/services/rag_pipeline/transform/website-crawl-general-high-quality.yml

@@ -126,7 +126,7 @@ workflow:
             type: mixed
             value: '{{#rag.1752491761974.jina_use_sitemap#}}'
         plugin_id: langgenius/jina_datasource
-        provider_name: jina
+        provider_name: jinareader
         provider_type: website_crawl
         selected: false
         title: Jina Reader

api/services/rag_pipeline/transform/website-crawl-parentchild.yml

@@ -419,7 +419,7 @@ workflow:
             type: mixed
             value: '{{#rag.1752491761974.jina_use_sitemap#}}'
         plugin_id: langgenius/jina_datasource
-        provider_name: jina
+        provider_name: jinareader
         provider_type: website_crawl
         selected: false
         title: Jina Reader

api/tasks/document_indexing_task.py

@@ -1,11 +1,15 @@
 import logging
 import time
+from collections.abc import Callable, Sequence
 
 import click
 from celery import shared_task
 
 from configs import dify_config
+from core.entities.document_task import DocumentTask
 from core.indexing_runner import DocumentIsPausedError, IndexingRunner
+from core.rag.pipeline.queue import TenantIsolatedTaskQueue
+from enums.cloud_plan import CloudPlan
 from extensions.ext_database import db
 from libs.datetime_utils import naive_utc_now
 from models.dataset import Dataset, Document
@@ -21,8 +25,24 @@ def document_indexing_task(dataset_id: str, document_ids: list):
     :param dataset_id:
     :param document_ids:
 
+    .. warning:: TO BE DEPRECATED
+        This function will be deprecated and removed in a future version.
+        Use normal_document_indexing_task or priority_document_indexing_task instead.
+
     Usage: document_indexing_task.delay(dataset_id, document_ids)
     """
+    logger.warning("document indexing legacy mode received: %s - %s", dataset_id, document_ids)
+    _document_indexing(dataset_id, document_ids)
+
+
+def _document_indexing(dataset_id: str, document_ids: Sequence[str]):
+    """
+    Process document for tasks
+    :param dataset_id:
+    :param document_ids:
+
+    Usage: _document_indexing(dataset_id, document_ids)
+    """
     documents = []
     start_at = time.perf_counter()
 
@@ -38,7 +58,7 @@ def document_indexing_task(dataset_id: str, document_ids: list):
             vector_space = features.vector_space
             count = len(document_ids)
             batch_upload_limit = int(dify_config.BATCH_UPLOAD_LIMIT)
-            if features.billing.subscription.plan == "sandbox" and count > 1:
+            if features.billing.subscription.plan == CloudPlan.SANDBOX and count > 1:
                 raise ValueError("Your current plan does not support batch upload, please upgrade your plan.")
             if count > batch_upload_limit:
                 raise ValueError(f"You have reached the batch upload limit of {batch_upload_limit}.")
@@ -86,3 +106,63 @@ def document_indexing_task(dataset_id: str, document_ids: list):
         logger.exception("Document indexing task failed, dataset_id: %s", dataset_id)
     finally:
         db.session.close()
+
+
+def _document_indexing_with_tenant_queue(
+    tenant_id: str, dataset_id: str, document_ids: Sequence[str], task_func: Callable[[str, str, Sequence[str]], None]
+):
+    try:
+        _document_indexing(dataset_id, document_ids)
+    except Exception:
+        logger.exception("Error processing document indexing %s for tenant %s: %s", dataset_id, tenant_id)
+    finally:
+        tenant_isolated_task_queue = TenantIsolatedTaskQueue(tenant_id, "document_indexing")
+
+        # Check if there are waiting tasks in the queue
+        # Use rpop to get the next task from the queue (FIFO order)
+        next_tasks = tenant_isolated_task_queue.pull_tasks(count=dify_config.TENANT_ISOLATED_TASK_CONCURRENCY)
+
+        logger.info("document indexing tenant isolation queue next tasks: %s", next_tasks)
+
+        if next_tasks:
+            for next_task in next_tasks:
+                document_task = DocumentTask(**next_task)
+                # Process the next waiting task
+                # Keep the flag set to indicate a task is running
+                tenant_isolated_task_queue.set_task_waiting_time()
+                task_func.delay(  # type: ignore
+                    tenant_id=document_task.tenant_id,
+                    dataset_id=document_task.dataset_id,
+                    document_ids=document_task.document_ids,
+                )
+        else:
+            # No more waiting tasks, clear the flag
+            tenant_isolated_task_queue.delete_task_key()
+
+
+@shared_task(queue="dataset")
+def normal_document_indexing_task(tenant_id: str, dataset_id: str, document_ids: Sequence[str]):
+    """
+    Async process document
+    :param tenant_id:
+    :param dataset_id:
+    :param document_ids:
+
+    Usage: normal_document_indexing_task.delay(tenant_id, dataset_id, document_ids)
+    """
+    logger.info("normal document indexing task received: %s - %s - %s", tenant_id, dataset_id, document_ids)
+    _document_indexing_with_tenant_queue(tenant_id, dataset_id, document_ids, normal_document_indexing_task)
+
+
+@shared_task(queue="priority_dataset")
+def priority_document_indexing_task(tenant_id: str, dataset_id: str, document_ids: Sequence[str]):
+    """
+    Priority async process document
+    :param tenant_id:
+    :param dataset_id:
+    :param document_ids:
+
+    Usage: priority_document_indexing_task.delay(tenant_id, dataset_id, document_ids)
+    """
+    logger.info("priority document indexing task received: %s - %s - %s", tenant_id, dataset_id, document_ids)
+    _document_indexing_with_tenant_queue(tenant_id, dataset_id, document_ids, priority_document_indexing_task)

api/tasks/duplicate_document_indexing_task.py

@@ -8,6 +8,7 @@ from sqlalchemy import select
 from configs import dify_config
 from core.indexing_runner import DocumentIsPausedError, IndexingRunner
 from core.rag.index_processor.index_processor_factory import IndexProcessorFactory
+from enums.cloud_plan import CloudPlan
 from extensions.ext_database import db
 from libs.datetime_utils import naive_utc_now
 from models.dataset import Dataset, Document, DocumentSegment
@@ -41,7 +42,7 @@ def duplicate_document_indexing_task(dataset_id: str, document_ids: list):
             if features.billing.enabled:
                 vector_space = features.vector_space
                 count = len(document_ids)
-                if features.billing.subscription.plan == "sandbox" and count > 1:
+                if features.billing.subscription.plan == CloudPlan.SANDBOX and count > 1:
                     raise ValueError("Your current plan does not support batch upload, please upgrade your plan.")
                 batch_upload_limit = int(dify_config.BATCH_UPLOAD_LIMIT)
                 if count > batch_upload_limit:

api/tasks/rag_pipeline/priority_rag_pipeline_run_task.py

@@ -12,8 +12,10 @@ from celery import shared_task  # type: ignore
 from flask import current_app, g
 from sqlalchemy.orm import Session, sessionmaker
 
+from configs import dify_config
 from core.app.entities.app_invoke_entities import InvokeFrom, RagPipelineGenerateEntity
 from core.app.entities.rag_pipeline_invoke_entities import RagPipelineInvokeEntity
+from core.rag.pipeline.queue import TenantIsolatedTaskQueue
 from core.repositories.factory import DifyCoreRepositoryFactory
 from extensions.ext_database import db
 from models import Account, Tenant
@@ -22,6 +24,8 @@ from models.enums import WorkflowRunTriggeredFrom
 from models.workflow import Workflow, WorkflowNodeExecutionTriggeredFrom
 from services.file_service import FileService
 
+logger = logging.getLogger(__name__)
+
 
 @shared_task(queue="priority_pipeline")
 def priority_rag_pipeline_run_task(
@@ -69,6 +73,27 @@ def priority_rag_pipeline_run_task(
         logging.exception(click.style(f"Error running rag pipeline, tenant_id: {tenant_id}", fg="red"))
         raise
     finally:
+        tenant_isolated_task_queue = TenantIsolatedTaskQueue(tenant_id, "pipeline")
+
+        # Check if there are waiting tasks in the queue
+        # Use rpop to get the next task from the queue (FIFO order)
+        next_file_ids = tenant_isolated_task_queue.pull_tasks(count=dify_config.TENANT_ISOLATED_TASK_CONCURRENCY)
+        logger.info("priority rag pipeline tenant isolation queue next files: %s", next_file_ids)
+
+        if next_file_ids:
+            for next_file_id in next_file_ids:
+                # Process the next waiting task
+                # Keep the flag set to indicate a task is running
+                tenant_isolated_task_queue.set_task_waiting_time()
+                priority_rag_pipeline_run_task.delay(  # type: ignore
+                    rag_pipeline_invoke_entities_file_id=next_file_id.decode("utf-8")
+                    if isinstance(next_file_id, bytes)
+                    else next_file_id,
+                    tenant_id=tenant_id,
+                )
+        else:
+            # No more waiting tasks, clear the flag
+            tenant_isolated_task_queue.delete_task_key()
         file_service = FileService(db.engine)
         file_service.delete_file(rag_pipeline_invoke_entities_file_id)
         db.session.close()

api/tasks/rag_pipeline/rag_pipeline_run_task.py

@@ -12,17 +12,20 @@ from celery import shared_task  # type: ignore
 from flask import current_app, g
 from sqlalchemy.orm import Session, sessionmaker
 
+from configs import dify_config
 from core.app.entities.app_invoke_entities import InvokeFrom, RagPipelineGenerateEntity
 from core.app.entities.rag_pipeline_invoke_entities import RagPipelineInvokeEntity
+from core.rag.pipeline.queue import TenantIsolatedTaskQueue
 from core.repositories.factory import DifyCoreRepositoryFactory
 from extensions.ext_database import db
-from extensions.ext_redis import redis_client
 from models import Account, Tenant
 from models.dataset import Pipeline
 from models.enums import WorkflowRunTriggeredFrom
 from models.workflow import Workflow, WorkflowNodeExecutionTriggeredFrom
 from services.file_service import FileService
 
+logger = logging.getLogger(__name__)
+
 
 @shared_task(queue="pipeline")
 def rag_pipeline_run_task(
@@ -70,26 +73,27 @@ def rag_pipeline_run_task(
         logging.exception(click.style(f"Error running rag pipeline, tenant_id: {tenant_id}", fg="red"))
         raise
     finally:
-        tenant_self_pipeline_task_queue = f"tenant_self_pipeline_task_queue:{tenant_id}"
-        tenant_pipeline_task_key = f"tenant_pipeline_task:{tenant_id}"
+        tenant_isolated_task_queue = TenantIsolatedTaskQueue(tenant_id, "pipeline")
 
         # Check if there are waiting tasks in the queue
         # Use rpop to get the next task from the queue (FIFO order)
-        next_file_id = redis_client.rpop(tenant_self_pipeline_task_queue)
+        next_file_ids = tenant_isolated_task_queue.pull_tasks(count=dify_config.TENANT_ISOLATED_TASK_CONCURRENCY)
+        logger.info("rag pipeline tenant isolation queue next files: %s", next_file_ids)
 
-        if next_file_id:
-            # Process the next waiting task
-            # Keep the flag set to indicate a task is running
-            redis_client.setex(tenant_pipeline_task_key, 60 * 60, 1)
-            rag_pipeline_run_task.delay(  # type: ignore
-                rag_pipeline_invoke_entities_file_id=next_file_id.decode("utf-8")
-                if isinstance(next_file_id, bytes)
-                else next_file_id,
-                tenant_id=tenant_id,
-            )
+        if next_file_ids:
+            for next_file_id in next_file_ids:
+                # Process the next waiting task
+                # Keep the flag set to indicate a task is running
+                tenant_isolated_task_queue.set_task_waiting_time()
+                rag_pipeline_run_task.delay(  # type: ignore
+                    rag_pipeline_invoke_entities_file_id=next_file_id.decode("utf-8")
+                    if isinstance(next_file_id, bytes)
+                    else next_file_id,
+                    tenant_id=tenant_id,
+                )
         else:
             # No more waiting tasks, clear the flag
-            redis_client.delete(tenant_pipeline_task_key)
+            tenant_isolated_task_queue.delete_task_key()
         file_service = FileService(db.engine)
         file_service.delete_file(rag_pipeline_invoke_entities_file_id)
         db.session.close()

api/tests/test_containers_integration_tests/core/rag/__init__.py

@@ -0,0 +1 @@
+

api/tests/test_containers_integration_tests/core/rag/pipeline/test_queue_integration.py

@@ -0,0 +1,595 @@
+"""
+Integration tests for TenantIsolatedTaskQueue using testcontainers.
+
+These tests verify the Redis-based task queue functionality with real Redis instances,
+testing tenant isolation, task serialization, and queue operations in a realistic environment.
+Includes compatibility tests for migrating from legacy string-only queues.
+
+All tests use generic naming to avoid coupling to specific business implementations.
+"""
+
+import time
+from dataclasses import dataclass
+from typing import Any
+from uuid import uuid4
+
+import pytest
+from faker import Faker
+
+from core.rag.pipeline.queue import TaskWrapper, TenantIsolatedTaskQueue
+from extensions.ext_redis import redis_client
+from models import Account, Tenant, TenantAccountJoin, TenantAccountRole
+
+
+@dataclass
+class TestTask:
+    """Test task data structure for testing complex object serialization."""
+
+    task_id: str
+    tenant_id: str
+    data: dict[str, Any]
+    metadata: dict[str, Any]
+
+
+class TestTenantIsolatedTaskQueueIntegration:
+    """Integration tests for TenantIsolatedTaskQueue using testcontainers."""
+
+    @pytest.fixture
+    def fake(self):
+        """Faker instance for generating test data."""
+        return Faker()
+
+    @pytest.fixture
+    def test_tenant_and_account(self, db_session_with_containers, fake):
+        """Create test tenant and account for testing."""
+        # Create account
+        account = Account(
+            email=fake.email(),
+            name=fake.name(),
+            interface_language="en-US",
+            status="active",
+        )
+        db_session_with_containers.add(account)
+        db_session_with_containers.commit()
+
+        # Create tenant
+        tenant = Tenant(
+            name=fake.company(),
+            status="normal",
+        )
+        db_session_with_containers.add(tenant)
+        db_session_with_containers.commit()
+
+        # Create tenant-account join
+        join = TenantAccountJoin(
+            tenant_id=tenant.id,
+            account_id=account.id,
+            role=TenantAccountRole.OWNER,
+            current=True,
+        )
+        db_session_with_containers.add(join)
+        db_session_with_containers.commit()
+
+        return tenant, account
+
+    @pytest.fixture
+    def test_queue(self, test_tenant_and_account):
+        """Create a generic test queue for testing."""
+        tenant, _ = test_tenant_and_account
+        return TenantIsolatedTaskQueue(tenant.id, "test_queue")
+
+    @pytest.fixture
+    def secondary_queue(self, test_tenant_and_account):
+        """Create a secondary test queue for testing isolation."""
+        tenant, _ = test_tenant_and_account
+        return TenantIsolatedTaskQueue(tenant.id, "secondary_queue")
+
+    def test_queue_initialization(self, test_tenant_and_account):
+        """Test queue initialization with correct key generation."""
+        tenant, _ = test_tenant_and_account
+        queue = TenantIsolatedTaskQueue(tenant.id, "test-key")
+
+        assert queue._tenant_id == tenant.id
+        assert queue._unique_key == "test-key"
+        assert queue._queue == f"tenant_self_test-key_task_queue:{tenant.id}"
+        assert queue._task_key == f"tenant_test-key_task:{tenant.id}"
+
+    def test_tenant_isolation(self, test_tenant_and_account, db_session_with_containers, fake):
+        """Test that different tenants have isolated queues."""
+        tenant1, _ = test_tenant_and_account
+
+        # Create second tenant
+        tenant2 = Tenant(
+            name=fake.company(),
+            status="normal",
+        )
+        db_session_with_containers.add(tenant2)
+        db_session_with_containers.commit()
+
+        queue1 = TenantIsolatedTaskQueue(tenant1.id, "same-key")
+        queue2 = TenantIsolatedTaskQueue(tenant2.id, "same-key")
+
+        assert queue1._queue != queue2._queue
+        assert queue1._task_key != queue2._task_key
+        assert queue1._queue == f"tenant_self_same-key_task_queue:{tenant1.id}"
+        assert queue2._queue == f"tenant_self_same-key_task_queue:{tenant2.id}"
+
+    def test_key_isolation(self, test_tenant_and_account):
+        """Test that different keys have isolated queues."""
+        tenant, _ = test_tenant_and_account
+        queue1 = TenantIsolatedTaskQueue(tenant.id, "key1")
+        queue2 = TenantIsolatedTaskQueue(tenant.id, "key2")
+
+        assert queue1._queue != queue2._queue
+        assert queue1._task_key != queue2._task_key
+        assert queue1._queue == f"tenant_self_key1_task_queue:{tenant.id}"
+        assert queue2._queue == f"tenant_self_key2_task_queue:{tenant.id}"
+
+    def test_task_key_operations(self, test_queue):
+        """Test task key operations (get, set, delete)."""
+        # Initially no task key should exist
+        assert test_queue.get_task_key() is None
+
+        # Set task waiting time with default TTL
+        test_queue.set_task_waiting_time()
+        task_key = test_queue.get_task_key()
+        # Redis returns bytes, convert to string for comparison
+        assert task_key in (b"1", "1")
+
+        # Set task waiting time with custom TTL
+        custom_ttl = 30
+        test_queue.set_task_waiting_time(custom_ttl)
+        task_key = test_queue.get_task_key()
+        assert task_key in (b"1", "1")
+
+        # Delete task key
+        test_queue.delete_task_key()
+        assert test_queue.get_task_key() is None
+
+    def test_push_and_pull_string_tasks(self, test_queue):
+        """Test pushing and pulling string tasks."""
+        tasks = ["task1", "task2", "task3"]
+
+        # Push tasks
+        test_queue.push_tasks(tasks)
+
+        # Pull tasks (FIFO order)
+        pulled_tasks = test_queue.pull_tasks(3)
+
+        # Should get tasks in FIFO order (lpush + rpop = FIFO)
+        assert pulled_tasks == ["task1", "task2", "task3"]
+
+    def test_push_and_pull_multiple_tasks(self, test_queue):
+        """Test pushing and pulling multiple tasks at once."""
+        tasks = ["task1", "task2", "task3", "task4", "task5"]
+
+        # Push tasks
+        test_queue.push_tasks(tasks)
+
+        # Pull multiple tasks
+        pulled_tasks = test_queue.pull_tasks(3)
+        assert len(pulled_tasks) == 3
+        assert pulled_tasks == ["task1", "task2", "task3"]
+
+        # Pull remaining tasks
+        remaining_tasks = test_queue.pull_tasks(5)
+        assert len(remaining_tasks) == 2
+        assert remaining_tasks == ["task4", "task5"]
+
+    def test_push_and_pull_complex_objects(self, test_queue, fake):
+        """Test pushing and pulling complex object tasks."""
+        # Create complex task objects as dictionaries (not dataclass instances)
+        tasks = [
+            {
+                "task_id": str(uuid4()),
+                "tenant_id": test_queue._tenant_id,
+                "data": {
+                    "file_id": str(uuid4()),
+                    "content": fake.text(),
+                    "metadata": {"size": fake.random_int(1000, 10000)},
+                },
+                "metadata": {"created_at": fake.iso8601(), "tags": fake.words(3)},
+            },
+            {
+                "task_id": str(uuid4()),
+                "tenant_id": test_queue._tenant_id,
+                "data": {
+                    "file_id": str(uuid4()),
+                    "content": "测试中文内容",
+                    "metadata": {"size": fake.random_int(1000, 10000)},
+                },
+                "metadata": {"created_at": fake.iso8601(), "tags": ["中文", "测试", "emoji🚀"]},
+            },
+        ]
+
+        # Push complex tasks
+        test_queue.push_tasks(tasks)
+
+        # Pull tasks
+        pulled_tasks = test_queue.pull_tasks(2)
+        assert len(pulled_tasks) == 2
+
+        # Verify deserialized tasks match original (FIFO order)
+        for i, pulled_task in enumerate(pulled_tasks):
+            original_task = tasks[i]  # FIFO order
+            assert isinstance(pulled_task, dict)
+            assert pulled_task["task_id"] == original_task["task_id"]
+            assert pulled_task["tenant_id"] == original_task["tenant_id"]
+            assert pulled_task["data"] == original_task["data"]
+            assert pulled_task["metadata"] == original_task["metadata"]
+
+    def test_mixed_task_types(self, test_queue, fake):
+        """Test pushing and pulling mixed string and object tasks."""
+        string_task = "simple_string_task"
+        object_task = {
+            "task_id": str(uuid4()),
+            "dataset_id": str(uuid4()),
+            "document_ids": [str(uuid4()) for _ in range(3)],
+        }
+
+        tasks = [string_task, object_task, "another_string"]
+
+        # Push mixed tasks
+        test_queue.push_tasks(tasks)
+
+        # Pull all tasks
+        pulled_tasks = test_queue.pull_tasks(3)
+        assert len(pulled_tasks) == 3
+
+        # Verify types and content
+        assert pulled_tasks[0] == string_task
+        assert isinstance(pulled_tasks[1], dict)
+        assert pulled_tasks[1] == object_task
+        assert pulled_tasks[2] == "another_string"
+
+    def test_empty_queue_operations(self, test_queue):
+        """Test operations on empty queue."""
+        # Pull from empty queue
+        tasks = test_queue.pull_tasks(5)
+        assert tasks == []
+
+        # Pull zero or negative count
+        assert test_queue.pull_tasks(0) == []
+        assert test_queue.pull_tasks(-1) == []
+
+    def test_task_ttl_expiration(self, test_queue):
+        """Test task key TTL expiration."""
+        # Set task with short TTL
+        short_ttl = 2
+        test_queue.set_task_waiting_time(short_ttl)
+
+        # Verify task key exists
+        assert test_queue.get_task_key() == b"1" or test_queue.get_task_key() == "1"
+
+        # Wait for TTL to expire
+        time.sleep(short_ttl + 1)
+
+        # Verify task key has expired
+        assert test_queue.get_task_key() is None
+
+    def test_large_task_batch(self, test_queue, fake):
+        """Test handling large batches of tasks."""
+        # Create large batch of tasks
+        large_batch = []
+        for i in range(100):
+            task = {
+                "task_id": str(uuid4()),
+                "index": i,
+                "data": fake.text(max_nb_chars=100),
+                "metadata": {"batch_id": str(uuid4())},
+            }
+            large_batch.append(task)
+
+        # Push large batch
+        test_queue.push_tasks(large_batch)
+
+        # Pull all tasks
+        pulled_tasks = test_queue.pull_tasks(100)
+        assert len(pulled_tasks) == 100
+
+        # Verify all tasks were retrieved correctly (FIFO order)
+        for i, task in enumerate(pulled_tasks):
+            assert isinstance(task, dict)
+            assert task["index"] == i  # FIFO order
+
+    def test_queue_operations_isolation(self, test_tenant_and_account, fake):
+        """Test concurrent operations on different queues."""
+        tenant, _ = test_tenant_and_account
+
+        # Create multiple queues for the same tenant
+        queue1 = TenantIsolatedTaskQueue(tenant.id, "queue1")
+        queue2 = TenantIsolatedTaskQueue(tenant.id, "queue2")
+
+        # Push tasks to different queues
+        queue1.push_tasks(["task1_queue1", "task2_queue1"])
+        queue2.push_tasks(["task1_queue2", "task2_queue2"])
+
+        # Verify queues are isolated
+        tasks1 = queue1.pull_tasks(2)
+        tasks2 = queue2.pull_tasks(2)
+
+        assert tasks1 == ["task1_queue1", "task2_queue1"]
+        assert tasks2 == ["task1_queue2", "task2_queue2"]
+        assert tasks1 != tasks2
+
+    def test_task_wrapper_serialization_roundtrip(self, test_queue, fake):
+        """Test TaskWrapper serialization and deserialization roundtrip."""
+        # Create complex nested data
+        complex_data = {
+            "id": str(uuid4()),
+            "nested": {"deep": {"value": "test", "numbers": [1, 2, 3, 4, 5], "unicode": "测试中文", "emoji": "🚀"}},
+            "metadata": {"created_at": fake.iso8601(), "tags": ["tag1", "tag2", "tag3"]},
+        }
+
+        # Create wrapper and serialize
+        wrapper = TaskWrapper(data=complex_data)
+        serialized = wrapper.serialize()
+
+        # Verify serialization
+        assert isinstance(serialized, str)
+        assert "测试中文" in serialized
+        assert "🚀" in serialized
+
+        # Deserialize and verify
+        deserialized_wrapper = TaskWrapper.deserialize(serialized)
+        assert deserialized_wrapper.data == complex_data
+
+    def test_error_handling_invalid_json(self, test_queue):
+        """Test error handling for invalid JSON in wrapped tasks."""
+        # Manually create invalid JSON task (not a valid TaskWrapper JSON)
+        invalid_json_task = "invalid json data"
+
+        # Push invalid task directly to Redis
+        redis_client.lpush(test_queue._queue, invalid_json_task)
+
+        # Pull task - should fall back to string since it's not valid JSON
+        task = test_queue.pull_tasks(1)
+        assert task[0] == invalid_json_task
+
+    def test_real_world_batch_processing_scenario(self, test_queue, fake):
+        """Test realistic batch processing scenario."""
+        # Simulate batch processing tasks
+        batch_tasks = []
+        for i in range(3):
+            task = {
+                "file_id": str(uuid4()),
+                "tenant_id": test_queue._tenant_id,
+                "user_id": str(uuid4()),
+                "processing_config": {
+                    "model": fake.random_element(["model_a", "model_b", "model_c"]),
+                    "temperature": fake.random.uniform(0.1, 1.0),
+                    "max_tokens": fake.random_int(1000, 4000),
+                },
+                "metadata": {
+                    "source": fake.random_element(["upload", "api", "webhook"]),
+                    "priority": fake.random_element(["low", "normal", "high"]),
+                },
+            }
+            batch_tasks.append(task)
+
+        # Push tasks
+        test_queue.push_tasks(batch_tasks)
+
+        # Process tasks in batches
+        batch_size = 2
+        processed_tasks = []
+
+        while True:
+            batch = test_queue.pull_tasks(batch_size)
+            if not batch:
+                break
+
+            processed_tasks.extend(batch)
+
+        # Verify all tasks were processed
+        assert len(processed_tasks) == 3
+
+        # Verify task structure
+        for task in processed_tasks:
+            assert isinstance(task, dict)
+            assert "file_id" in task
+            assert "tenant_id" in task
+            assert "processing_config" in task
+            assert "metadata" in task
+            assert task["tenant_id"] == test_queue._tenant_id
+
+
+class TestTenantIsolatedTaskQueueCompatibility:
+    """Compatibility tests for migrating from legacy string-only queues."""
+
+    @pytest.fixture
+    def fake(self):
+        """Faker instance for generating test data."""
+        return Faker()
+
+    @pytest.fixture
+    def test_tenant_and_account(self, db_session_with_containers, fake):
+        """Create test tenant and account for testing."""
+        # Create account
+        account = Account(
+            email=fake.email(),
+            name=fake.name(),
+            interface_language="en-US",
+            status="active",
+        )
+        db_session_with_containers.add(account)
+        db_session_with_containers.commit()
+
+        # Create tenant
+        tenant = Tenant(
+            name=fake.company(),
+            status="normal",
+        )
+        db_session_with_containers.add(tenant)
+        db_session_with_containers.commit()
+
+        # Create tenant-account join
+        join = TenantAccountJoin(
+            tenant_id=tenant.id,
+            account_id=account.id,
+            role=TenantAccountRole.OWNER,
+            current=True,
+        )
+        db_session_with_containers.add(join)
+        db_session_with_containers.commit()
+
+        return tenant, account
+
+    def test_legacy_string_queue_compatibility(self, test_tenant_and_account, fake):
+        """
+        Test compatibility with legacy queues containing only string data.
+
+        This simulates the scenario where Redis queues already contain string data
+        from the old architecture, and we need to ensure the new code can read them.
+        """
+        tenant, _ = test_tenant_and_account
+        queue = TenantIsolatedTaskQueue(tenant.id, "legacy_queue")
+
+        # Simulate legacy string data in Redis queue (using old format)
+        legacy_strings = ["legacy_task_1", "legacy_task_2", "legacy_task_3", "legacy_task_4", "legacy_task_5"]
+
+        # Manually push legacy strings directly to Redis (simulating old system)
+        for legacy_string in legacy_strings:
+            redis_client.lpush(queue._queue, legacy_string)
+
+        # Verify new code can read legacy string data
+        pulled_tasks = queue.pull_tasks(5)
+        assert len(pulled_tasks) == 5
+
+        # Verify all tasks are strings (not wrapped)
+        for task in pulled_tasks:
+            assert isinstance(task, str)
+            assert task.startswith("legacy_task_")
+
+        # Verify order (FIFO from Redis list)
+        expected_order = ["legacy_task_1", "legacy_task_2", "legacy_task_3", "legacy_task_4", "legacy_task_5"]
+        assert pulled_tasks == expected_order
+
+    def test_legacy_queue_migration_scenario(self, test_tenant_and_account, fake):
+        """
+        Test complete migration scenario from legacy to new system.
+
+        This simulates the real-world scenario where:
+        1. Legacy system has string data in Redis
+        2. New system starts processing the same queue
+        3. Both legacy and new tasks coexist during migration
+        4. New system can handle both formats seamlessly
+        """
+        tenant, _ = test_tenant_and_account
+        queue = TenantIsolatedTaskQueue(tenant.id, "migration_queue")
+
+        # Phase 1: Legacy system has data
+        legacy_tasks = [f"legacy_resource_{i}" for i in range(1, 6)]
+        redis_client.lpush(queue._queue, *legacy_tasks)
+
+        # Phase 2: New system starts processing legacy data
+        processed_legacy = []
+        while True:
+            tasks = queue.pull_tasks(1)
+            if not tasks:
+                break
+            processed_legacy.extend(tasks)
+
+        # Verify legacy data was processed correctly
+        assert len(processed_legacy) == 5
+        for task in processed_legacy:
+            assert isinstance(task, str)
+            assert task.startswith("legacy_resource_")
+
+        # Phase 3: New system adds new tasks (mixed types)
+        new_string_tasks = ["new_resource_1", "new_resource_2"]
+        new_object_tasks = [
+            {
+                "resource_id": str(uuid4()),
+                "tenant_id": tenant.id,
+                "processing_type": "new_system",
+                "metadata": {"version": "2.0", "features": ["ai", "ml"]},
+            },
+            {
+                "resource_id": str(uuid4()),
+                "tenant_id": tenant.id,
+                "processing_type": "new_system",
+                "metadata": {"version": "2.0", "features": ["ai", "ml"]},
+            },
+        ]
+
+        # Push new tasks using new system
+        queue.push_tasks(new_string_tasks)
+        queue.push_tasks(new_object_tasks)
+
+        # Phase 4: Process all new tasks
+        processed_new = []
+        while True:
+            tasks = queue.pull_tasks(1)
+            if not tasks:
+                break
+            processed_new.extend(tasks)
+
+        # Verify new tasks were processed correctly
+        assert len(processed_new) == 4
+
+        string_tasks = [task for task in processed_new if isinstance(task, str)]
+        object_tasks = [task for task in processed_new if isinstance(task, dict)]
+
+        assert len(string_tasks) == 2
+        assert len(object_tasks) == 2
+
+        # Verify string tasks
+        for task in string_tasks:
+            assert task.startswith("new_resource_")
+
+        # Verify object tasks
+        for task in object_tasks:
+            assert isinstance(task, dict)
+            assert "resource_id" in task
+            assert "tenant_id" in task
+            assert task["tenant_id"] == tenant.id
+            assert task["processing_type"] == "new_system"
+
+    def test_legacy_queue_error_recovery(self, test_tenant_and_account, fake):
+        """
+        Test error recovery when legacy queue contains malformed data.
+
+        This ensures the new system can gracefully handle corrupted or
+        malformed legacy data without crashing.
+        """
+        tenant, _ = test_tenant_and_account
+        queue = TenantIsolatedTaskQueue(tenant.id, "error_recovery_queue")
+
+        # Create mix of valid and malformed legacy data
+        mixed_legacy_data = [
+            "valid_legacy_task_1",
+            "valid_legacy_task_2",
+            "malformed_data_string",  # This should be treated as string
+            "valid_legacy_task_3",
+            "invalid_json_not_taskwrapper_format",  # This should fall back to string (not valid TaskWrapper JSON)
+            "valid_legacy_task_4",
+        ]
+
+        # Manually push mixed data directly to Redis
+        redis_client.lpush(queue._queue, *mixed_legacy_data)
+
+        # Process all tasks
+        processed_tasks = []
+        while True:
+            tasks = queue.pull_tasks(1)
+            if not tasks:
+                break
+            processed_tasks.extend(tasks)
+
+        # Verify all tasks were processed (no crashes)
+        assert len(processed_tasks) == 6
+
+        # Verify all tasks are strings (malformed data falls back to string)
+        for task in processed_tasks:
+            assert isinstance(task, str)
+
+        # Verify valid tasks are preserved
+        valid_tasks = [task for task in processed_tasks if task.startswith("valid_legacy_task_")]
+        assert len(valid_tasks) == 4
+
+        # Verify malformed data is handled gracefully
+        malformed_tasks = [task for task in processed_tasks if not task.startswith("valid_legacy_task_")]
+        assert len(malformed_tasks) == 2
+        assert "malformed_data_string" in malformed_tasks
+        assert "invalid_json_not_taskwrapper_format" in malformed_tasks

api/tests/test_containers_integration_tests/services/test_app_generate_service.py

@@ -5,6 +5,7 @@ import pytest
 from faker import Faker
 
 from core.app.entities.app_invoke_entities import InvokeFrom
+from enums.cloud_plan import CloudPlan
 from models.model import EndUser
 from models.workflow import Workflow
 from services.app_generate_service import AppGenerateService
@@ -32,7 +33,7 @@ class TestAppGenerateService:
             patch("services.app_generate_service.dify_config") as mock_dify_config,
         ):
             # Setup default mock returns for billing service
-            mock_billing_service.get_info.return_value = {"subscription": {"plan": "sandbox"}}
+            mock_billing_service.get_info.return_value = {"subscription": {"plan": CloudPlan.SANDBOX}}
 
             # Setup default mock returns for workflow service
             mock_workflow_service_instance = mock_workflow_service.return_value
@@ -430,7 +431,7 @@ class TestAppGenerateService:
 
         # Setup billing service mock for sandbox plan
         mock_external_service_dependencies["billing_service"].get_info.return_value = {
-            "subscription": {"plan": "sandbox"}
+            "subscription": {"plan": CloudPlan.SANDBOX}
         }
 
         # Set BILLING_ENABLED to True for this test
@@ -461,7 +462,7 @@ class TestAppGenerateService:
 
         # Setup billing service mock for sandbox plan
         mock_external_service_dependencies["billing_service"].get_info.return_value = {
-            "subscription": {"plan": "sandbox"}
+            "subscription": {"plan": CloudPlan.SANDBOX}
         }
 
         # Set BILLING_ENABLED to True for this test

api/tests/test_containers_integration_tests/services/test_feature_service.py

@@ -3,6 +3,7 @@ from unittest.mock import patch
 import pytest
 from faker import Faker
 
+from enums.cloud_plan import CloudPlan
 from services.feature_service import FeatureModel, FeatureService, KnowledgeRateLimitModel, SystemFeatureModel
 
 
@@ -173,7 +174,7 @@ class TestFeatureService:
             # Set mock return value inside the patch context
             mock_external_service_dependencies["billing_service"].get_info.return_value = {
                 "enabled": True,
-                "subscription": {"plan": "sandbox", "interval": "monthly", "education": False},
+                "subscription": {"plan": CloudPlan.SANDBOX, "interval": "monthly", "education": False},
                 "members": {"size": 1, "limit": 3},
                 "apps": {"size": 1, "limit": 5},
                 "vector_space": {"size": 1, "limit": 2},
@@ -189,7 +190,7 @@ class TestFeatureService:
             result = FeatureService.get_features(tenant_id)
 
         # Assert: Verify sandbox-specific limitations
-        assert result.billing.subscription.plan == "sandbox"
+        assert result.billing.subscription.plan == CloudPlan.SANDBOX
         assert result.education.activated is False
 
         # Verify sandbox limitations

api/tests/test_containers_integration_tests/services/test_file_service.py

@@ -11,7 +11,7 @@ from configs import dify_config
 from models import Account, Tenant
 from models.enums import CreatorUserRole
 from models.model import EndUser, UploadFile
-from services.errors.file import FileTooLargeError, UnsupportedFileTypeError
+from services.errors.file import BlockedFileExtensionError, FileTooLargeError, UnsupportedFileTypeError
 from services.file_service import FileService
 
 
@@ -943,3 +943,150 @@ class TestFileService:
 
         # Should have the signed URL when source_url is empty
         assert upload_file2.source_url == "https://example.com/signed-url"
+
+    # Test file extension blacklist
+    def test_upload_file_blocked_extension(
+        self, db_session_with_containers, engine, mock_external_service_dependencies
+    ):
+        """
+        Test file upload with blocked extension.
+        """
+        fake = Faker()
+        account = self._create_test_account(db_session_with_containers, mock_external_service_dependencies)
+
+        # Mock blacklist configuration by patching the inner field
+        with patch.object(dify_config, "inner_UPLOAD_FILE_EXTENSION_BLACKLIST", "exe,bat,sh"):
+            filename = "malware.exe"
+            content = b"test content"
+            mimetype = "application/x-msdownload"
+
+            with pytest.raises(BlockedFileExtensionError):
+                FileService(engine).upload_file(
+                    filename=filename,
+                    content=content,
+                    mimetype=mimetype,
+                    user=account,
+                )
+
+    def test_upload_file_blocked_extension_case_insensitive(
+        self, db_session_with_containers, engine, mock_external_service_dependencies
+    ):
+        """
+        Test file upload with blocked extension (case insensitive).
+        """
+        fake = Faker()
+        account = self._create_test_account(db_session_with_containers, mock_external_service_dependencies)
+
+        # Mock blacklist configuration by patching the inner field
+        with patch.object(dify_config, "inner_UPLOAD_FILE_EXTENSION_BLACKLIST", "exe,bat"):
+            # Test with uppercase extension
+            filename = "malware.EXE"
+            content = b"test content"
+            mimetype = "application/x-msdownload"
+
+            with pytest.raises(BlockedFileExtensionError):
+                FileService(engine).upload_file(
+                    filename=filename,
+                    content=content,
+                    mimetype=mimetype,
+                    user=account,
+                )
+
+    def test_upload_file_not_in_blacklist(self, db_session_with_containers, engine, mock_external_service_dependencies):
+        """
+        Test file upload with extension not in blacklist.
+        """
+        fake = Faker()
+        account = self._create_test_account(db_session_with_containers, mock_external_service_dependencies)
+
+        # Mock blacklist configuration by patching the inner field
+        with patch.object(dify_config, "inner_UPLOAD_FILE_EXTENSION_BLACKLIST", "exe,bat,sh"):
+            filename = "document.pdf"
+            content = b"test content"
+            mimetype = "application/pdf"
+
+            upload_file = FileService(engine).upload_file(
+                filename=filename,
+                content=content,
+                mimetype=mimetype,
+                user=account,
+            )
+
+            assert upload_file is not None
+            assert upload_file.name == filename
+            assert upload_file.extension == "pdf"
+
+    def test_upload_file_empty_blacklist(self, db_session_with_containers, engine, mock_external_service_dependencies):
+        """
+        Test file upload with empty blacklist (default behavior).
+        """
+        fake = Faker()
+        account = self._create_test_account(db_session_with_containers, mock_external_service_dependencies)
+
+        # Mock empty blacklist configuration by patching the inner field
+        with patch.object(dify_config, "inner_UPLOAD_FILE_EXTENSION_BLACKLIST", ""):
+            # Should allow all file types when blacklist is empty
+            filename = "script.sh"
+            content = b"#!/bin/bash\necho test"
+            mimetype = "application/x-sh"
+
+            upload_file = FileService(engine).upload_file(
+                filename=filename,
+                content=content,
+                mimetype=mimetype,
+                user=account,
+            )
+
+            assert upload_file is not None
+            assert upload_file.extension == "sh"
+
+    def test_upload_file_multiple_blocked_extensions(
+        self, db_session_with_containers, engine, mock_external_service_dependencies
+    ):
+        """
+        Test file upload with multiple blocked extensions.
+        """
+        fake = Faker()
+        account = self._create_test_account(db_session_with_containers, mock_external_service_dependencies)
+
+        # Mock blacklist with multiple extensions by patching the inner field
+        blacklist_str = "exe,bat,cmd,com,scr,vbs,ps1,msi,dll"
+        with patch.object(dify_config, "inner_UPLOAD_FILE_EXTENSION_BLACKLIST", blacklist_str):
+            for ext in blacklist_str.split(","):
+                filename = f"malware.{ext}"
+                content = b"test content"
+                mimetype = "application/octet-stream"
+
+                with pytest.raises(BlockedFileExtensionError):
+                    FileService(engine).upload_file(
+                        filename=filename,
+                        content=content,
+                        mimetype=mimetype,
+                        user=account,
+                    )
+
+    def test_upload_file_no_extension_with_blacklist(
+        self, db_session_with_containers, engine, mock_external_service_dependencies
+    ):
+        """
+        Test file upload with no extension when blacklist is configured.
+        """
+        fake = Faker()
+        account = self._create_test_account(db_session_with_containers, mock_external_service_dependencies)
+
+        # Mock blacklist configuration by patching the inner field
+        with patch.object(dify_config, "inner_UPLOAD_FILE_EXTENSION_BLACKLIST", "exe,bat"):
+            # Files with no extension should not be blocked
+            filename = "README"
+            content = b"test content"
+            mimetype = "text/plain"
+
+            upload_file = FileService(engine).upload_file(
+                filename=filename,
+                content=content,
+                mimetype=mimetype,
+                user=account,
+            )
+
+            assert upload_file is not None
+            assert upload_file.extension == ""

api/tests/test_containers_integration_tests/tasks/test_document_indexing_task.py

@@ -1,16 +1,33 @@
+from dataclasses import asdict
 from unittest.mock import MagicMock, patch
 
 import pytest
 from faker import Faker
 
+from core.entities.document_task import DocumentTask
+from enums.cloud_plan import CloudPlan
 from extensions.ext_database import db
 from models import Account, Tenant, TenantAccountJoin, TenantAccountRole
 from models.dataset import Dataset, Document
-from tasks.document_indexing_task import document_indexing_task
+from tasks.document_indexing_task import (
+    _document_indexing,  # Core function
+    _document_indexing_with_tenant_queue,  # Tenant queue wrapper function
+    document_indexing_task,  # Deprecated old interface
+    normal_document_indexing_task,  # New normal task
+    priority_document_indexing_task,  # New priority task
+)
 
 
-class TestDocumentIndexingTask:
-    """Integration tests for document_indexing_task using testcontainers."""
+class TestDocumentIndexingTasks:
+    """Integration tests for document indexing tasks using testcontainers.
+
+    This test class covers:
+    - Core _document_indexing function
+    - Deprecated document_indexing_task function
+    - New normal_document_indexing_task function
+    - New priority_document_indexing_task function
+    - Tenant queue wrapper _document_indexing_with_tenant_queue function
+    """
 
     @pytest.fixture
     def mock_external_service_dependencies(self):
@@ -197,7 +214,7 @@ class TestDocumentIndexingTask:
         # Configure billing features
         mock_external_service_dependencies["features"].billing.enabled = billing_enabled
         if billing_enabled:
-            mock_external_service_dependencies["features"].billing.subscription.plan = "sandbox"
+            mock_external_service_dependencies["features"].billing.subscription.plan = CloudPlan.SANDBOX
             mock_external_service_dependencies["features"].vector_space.limit = 100
             mock_external_service_dependencies["features"].vector_space.size = 50
 
@@ -223,7 +240,7 @@ class TestDocumentIndexingTask:
         document_ids = [doc.id for doc in documents]
 
         # Act: Execute the task
-        document_indexing_task(dataset.id, document_ids)
+        _document_indexing(dataset.id, document_ids)
 
         # Assert: Verify the expected outcomes
         # Verify indexing runner was called correctly
@@ -231,10 +248,11 @@ class TestDocumentIndexingTask:
         mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
 
         # Verify documents were updated to parsing status
-        for document in documents:
-            db.session.refresh(document)
-            assert document.indexing_status == "parsing"
-            assert document.processing_started_at is not None
+        # Re-query documents from database since _document_indexing uses a different session
+        for doc_id in document_ids:
+            updated_document = db.session.query(Document).where(Document.id == doc_id).first()
+            assert updated_document.indexing_status == "parsing"
+            assert updated_document.processing_started_at is not None
 
         # Verify the run method was called with correct documents
         call_args = mock_external_service_dependencies["indexing_runner_instance"].run.call_args
@@ -260,7 +278,7 @@ class TestDocumentIndexingTask:
         document_ids = [fake.uuid4() for _ in range(3)]
 
         # Act: Execute the task with non-existent dataset
-        document_indexing_task(non_existent_dataset_id, document_ids)
+        _document_indexing(non_existent_dataset_id, document_ids)
 
         # Assert: Verify no processing occurred
         mock_external_service_dependencies["indexing_runner"].assert_not_called()
@@ -290,17 +308,18 @@ class TestDocumentIndexingTask:
         all_document_ids = existing_document_ids + non_existent_document_ids
 
         # Act: Execute the task with mixed document IDs
-        document_indexing_task(dataset.id, all_document_ids)
+        _document_indexing(dataset.id, all_document_ids)
 
         # Assert: Verify only existing documents were processed
         mock_external_service_dependencies["indexing_runner"].assert_called_once()
         mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
 
         # Verify only existing documents were updated
-        for document in documents:
-            db.session.refresh(document)
-            assert document.indexing_status == "parsing"
-            assert document.processing_started_at is not None
+        # Re-query documents from database since _document_indexing uses a different session
+        for doc_id in existing_document_ids:
+            updated_document = db.session.query(Document).where(Document.id == doc_id).first()
+            assert updated_document.indexing_status == "parsing"
+            assert updated_document.processing_started_at is not None
 
         # Verify the run method was called with only existing documents
         call_args = mock_external_service_dependencies["indexing_runner_instance"].run.call_args
@@ -332,7 +351,7 @@ class TestDocumentIndexingTask:
         )
 
         # Act: Execute the task
-        document_indexing_task(dataset.id, document_ids)
+        _document_indexing(dataset.id, document_ids)
 
         # Assert: Verify exception was handled gracefully
         # The task should complete without raising exceptions
@@ -340,10 +359,11 @@ class TestDocumentIndexingTask:
         mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
 
         # Verify documents were still updated to parsing status before the exception
-        for document in documents:
-            db.session.refresh(document)
-            assert document.indexing_status == "parsing"
-            assert document.processing_started_at is not None
+        # Re-query documents from database since _document_indexing close the session
+        for doc_id in document_ids:
+            updated_document = db.session.query(Document).where(Document.id == doc_id).first()
+            assert updated_document.indexing_status == "parsing"
+            assert updated_document.processing_started_at is not None
 
     def test_document_indexing_task_mixed_document_states(
         self, db_session_with_containers, mock_external_service_dependencies
@@ -406,17 +426,18 @@ class TestDocumentIndexingTask:
         document_ids = [doc.id for doc in all_documents]
 
         # Act: Execute the task with mixed document states
-        document_indexing_task(dataset.id, document_ids)
+        _document_indexing(dataset.id, document_ids)
 
         # Assert: Verify processing
         mock_external_service_dependencies["indexing_runner"].assert_called_once()
         mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
 
         # Verify all documents were updated to parsing status
-        for document in all_documents:
-            db.session.refresh(document)
-            assert document.indexing_status == "parsing"
-            assert document.processing_started_at is not None
+        # Re-query documents from database since _document_indexing uses a different session
+        for doc_id in document_ids:
+            updated_document = db.session.query(Document).where(Document.id == doc_id).first()
+            assert updated_document.indexing_status == "parsing"
+            assert updated_document.processing_started_at is not None
 
         # Verify the run method was called with all documents
         call_args = mock_external_service_dependencies["indexing_runner_instance"].run.call_args
@@ -442,7 +463,7 @@ class TestDocumentIndexingTask:
         )
 
         # Configure sandbox plan with batch limit
-        mock_external_service_dependencies["features"].billing.subscription.plan = "sandbox"
+        mock_external_service_dependencies["features"].billing.subscription.plan = CloudPlan.SANDBOX
 
         # Create more documents than sandbox plan allows (limit is 1)
         fake = Faker()
@@ -469,15 +490,16 @@ class TestDocumentIndexingTask:
         document_ids = [doc.id for doc in all_documents]
 
         # Act: Execute the task with too many documents for sandbox plan
-        document_indexing_task(dataset.id, document_ids)
+        _document_indexing(dataset.id, document_ids)
 
         # Assert: Verify error handling
-        for document in all_documents:
-            db.session.refresh(document)
-            assert document.indexing_status == "error"
-            assert document.error is not None
-            assert "batch upload" in document.error
-            assert document.stopped_at is not None
+        # Re-query documents from database since _document_indexing uses a different session
+        for doc_id in document_ids:
+            updated_document = db.session.query(Document).where(Document.id == doc_id).first()
+            assert updated_document.indexing_status == "error"
+            assert updated_document.error is not None
+            assert "batch upload" in updated_document.error
+            assert updated_document.stopped_at is not None
 
         # Verify no indexing runner was called
         mock_external_service_dependencies["indexing_runner"].assert_not_called()
@@ -502,17 +524,18 @@ class TestDocumentIndexingTask:
         document_ids = [doc.id for doc in documents]
 
         # Act: Execute the task with billing disabled
-        document_indexing_task(dataset.id, document_ids)
+        _document_indexing(dataset.id, document_ids)
 
         # Assert: Verify successful processing
         mock_external_service_dependencies["indexing_runner"].assert_called_once()
         mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
 
         # Verify documents were updated to parsing status
-        for document in documents:
-            db.session.refresh(document)
-            assert document.indexing_status == "parsing"
-            assert document.processing_started_at is not None
+        # Re-query documents from database since _document_indexing uses a different session
+        for doc_id in document_ids:
+            updated_document = db.session.query(Document).where(Document.id == doc_id).first()
+            assert updated_document.indexing_status == "parsing"
+            assert updated_document.processing_started_at is not None
 
     def test_document_indexing_task_document_is_paused_error(
         self, db_session_with_containers, mock_external_service_dependencies
@@ -540,7 +563,7 @@ class TestDocumentIndexingTask:
         )
 
         # Act: Execute the task
-        document_indexing_task(dataset.id, document_ids)
+        _document_indexing(dataset.id, document_ids)
 
         # Assert: Verify exception was handled gracefully
         # The task should complete without raising exceptions
@@ -548,7 +571,317 @@ class TestDocumentIndexingTask:
         mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
 
         # Verify documents were still updated to parsing status before the exception
-        for document in documents:
-            db.session.refresh(document)
-            assert document.indexing_status == "parsing"
-            assert document.processing_started_at is not None
+        # Re-query documents from database since _document_indexing uses a different session
+        for doc_id in document_ids:
+            updated_document = db.session.query(Document).where(Document.id == doc_id).first()
+            assert updated_document.indexing_status == "parsing"
+            assert updated_document.processing_started_at is not None
+
+    # ==================== NEW TESTS FOR REFACTORED FUNCTIONS ====================
+    def test_old_document_indexing_task_success(self, db_session_with_containers, mock_external_service_dependencies):
+        """
+        Test document_indexing_task basic functionality.
+
+        This test verifies:
+        - Task function calls the wrapper correctly
+        - Basic parameter passing works
+        """
+        # Arrange: Create test data
+        dataset, documents = self._create_test_dataset_and_documents(
+            db_session_with_containers, mock_external_service_dependencies, document_count=1
+        )
+        document_ids = [doc.id for doc in documents]
+
+        # Act: Execute the deprecated task (it only takes 2 parameters)
+        document_indexing_task(dataset.id, document_ids)
+
+        # Assert: Verify processing occurred (core logic is tested in _document_indexing tests)
+        mock_external_service_dependencies["indexing_runner"].assert_called_once()
+        mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
+
+    def test_normal_document_indexing_task_success(
+        self, db_session_with_containers, mock_external_service_dependencies
+    ):
+        """
+        Test normal_document_indexing_task basic functionality.
+
+        This test verifies:
+        - Task function calls the wrapper correctly
+        - Basic parameter passing works
+        """
+        # Arrange: Create test data
+        dataset, documents = self._create_test_dataset_and_documents(
+            db_session_with_containers, mock_external_service_dependencies, document_count=1
+        )
+        document_ids = [doc.id for doc in documents]
+        tenant_id = dataset.tenant_id
+
+        # Act: Execute the new normal task
+        normal_document_indexing_task(tenant_id, dataset.id, document_ids)
+
+        # Assert: Verify processing occurred (core logic is tested in _document_indexing tests)
+        mock_external_service_dependencies["indexing_runner"].assert_called_once()
+        mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
+
+    def test_priority_document_indexing_task_success(
+        self, db_session_with_containers, mock_external_service_dependencies
+    ):
+        """
+        Test priority_document_indexing_task basic functionality.
+
+        This test verifies:
+        - Task function calls the wrapper correctly
+        - Basic parameter passing works
+        """
+        # Arrange: Create test data
+        dataset, documents = self._create_test_dataset_and_documents(
+            db_session_with_containers, mock_external_service_dependencies, document_count=1
+        )
+        document_ids = [doc.id for doc in documents]
+        tenant_id = dataset.tenant_id
+
+        # Act: Execute the new priority task
+        priority_document_indexing_task(tenant_id, dataset.id, document_ids)
+
+        # Assert: Verify processing occurred (core logic is tested in _document_indexing tests)
+        mock_external_service_dependencies["indexing_runner"].assert_called_once()
+        mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
+
+    def test_document_indexing_with_tenant_queue_success(
+        self, db_session_with_containers, mock_external_service_dependencies
+    ):
+        """
+        Test _document_indexing_with_tenant_queue function with no waiting tasks.
+
+        This test verifies:
+        - Core indexing logic execution (same as _document_indexing)
+        - Tenant queue cleanup when no waiting tasks
+        - Task function parameter passing
+        - Queue management after processing
+        """
+        # Arrange: Create test data
+        dataset, documents = self._create_test_dataset_and_documents(
+            db_session_with_containers, mock_external_service_dependencies, document_count=2
+        )
+        document_ids = [doc.id for doc in documents]
+        tenant_id = dataset.tenant_id
+
+        # Mock the task function
+        from unittest.mock import MagicMock
+
+        mock_task_func = MagicMock()
+
+        # Act: Execute the wrapper function
+        _document_indexing_with_tenant_queue(tenant_id, dataset.id, document_ids, mock_task_func)
+
+        # Assert: Verify core processing occurred (same as _document_indexing)
+        mock_external_service_dependencies["indexing_runner"].assert_called_once()
+        mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
+
+        # Verify documents were updated (same as _document_indexing)
+        # Re-query documents from database since _document_indexing uses a different session
+        for doc_id in document_ids:
+            updated_document = db.session.query(Document).where(Document.id == doc_id).first()
+            assert updated_document.indexing_status == "parsing"
+            assert updated_document.processing_started_at is not None
+
+        # Verify the run method was called with correct documents
+        call_args = mock_external_service_dependencies["indexing_runner_instance"].run.call_args
+        assert call_args is not None
+        processed_documents = call_args[0][0]
+        assert len(processed_documents) == 2
+
+        # Verify task function was not called (no waiting tasks)
+        mock_task_func.delay.assert_not_called()
+
+    def test_document_indexing_with_tenant_queue_with_waiting_tasks(
+        self, db_session_with_containers, mock_external_service_dependencies
+    ):
+        """
+        Test _document_indexing_with_tenant_queue function with waiting tasks in queue using real Redis.
+
+        This test verifies:
+        - Core indexing logic execution
+        - Real Redis-based tenant queue processing of waiting tasks
+        - Task function calls for waiting tasks
+        - Queue management with multiple tasks using actual Redis operations
+        """
+        # Arrange: Create test data
+        dataset, documents = self._create_test_dataset_and_documents(
+            db_session_with_containers, mock_external_service_dependencies, document_count=1
+        )
+        document_ids = [doc.id for doc in documents]
+        tenant_id = dataset.tenant_id
+        dataset_id = dataset.id
+
+        # Mock the task function
+        from unittest.mock import MagicMock
+
+        mock_task_func = MagicMock()
+
+        # Use real Redis for TenantIsolatedTaskQueue
+        from core.rag.pipeline.queue import TenantIsolatedTaskQueue
+
+        # Create real queue instance
+        queue = TenantIsolatedTaskQueue(tenant_id, "document_indexing")
+
+        # Add waiting tasks to the real Redis queue
+        waiting_tasks = [
+            DocumentTask(tenant_id=tenant_id, dataset_id=dataset.id, document_ids=["waiting-doc-1"]),
+            DocumentTask(tenant_id=tenant_id, dataset_id=dataset.id, document_ids=["waiting-doc-2"]),
+        ]
+        # Convert DocumentTask objects to dictionaries for serialization
+        waiting_task_dicts = [asdict(task) for task in waiting_tasks]
+        queue.push_tasks(waiting_task_dicts)
+
+        # Act: Execute the wrapper function
+        _document_indexing_with_tenant_queue(tenant_id, dataset.id, document_ids, mock_task_func)
+
+        # Assert: Verify core processing occurred
+        mock_external_service_dependencies["indexing_runner"].assert_called_once()
+        mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
+
+        # Verify task function was called for each waiting task
+        assert mock_task_func.delay.call_count == 1
+
+        # Verify correct parameters for each call
+        calls = mock_task_func.delay.call_args_list
+        assert calls[0][1] == {"tenant_id": tenant_id, "dataset_id": dataset_id, "document_ids": ["waiting-doc-1"]}
+
+        # Verify queue is empty after processing (tasks were pulled)
+        remaining_tasks = queue.pull_tasks(count=10)  # Pull more than we added
+        assert len(remaining_tasks) == 1
+
+    def test_document_indexing_with_tenant_queue_error_handling(
+        self, db_session_with_containers, mock_external_service_dependencies
+    ):
+        """
+        Test error handling in _document_indexing_with_tenant_queue using real Redis.
+
+        This test verifies:
+        - Exception handling during core processing
+        - Tenant queue cleanup even on errors using real Redis
+        - Proper error logging
+        - Function completes without raising exceptions
+        - Queue management continues despite core processing errors
+        """
+        # Arrange: Create test data
+        dataset, documents = self._create_test_dataset_and_documents(
+            db_session_with_containers, mock_external_service_dependencies, document_count=1
+        )
+        document_ids = [doc.id for doc in documents]
+        tenant_id = dataset.tenant_id
+        dataset_id = dataset.id
+
+        # Mock IndexingRunner to raise an exception
+        mock_external_service_dependencies["indexing_runner_instance"].run.side_effect = Exception("Test error")
+
+        # Mock the task function
+        from unittest.mock import MagicMock
+
+        mock_task_func = MagicMock()
+
+        # Use real Redis for TenantIsolatedTaskQueue
+        from core.rag.pipeline.queue import TenantIsolatedTaskQueue
+
+        # Create real queue instance
+        queue = TenantIsolatedTaskQueue(tenant_id, "document_indexing")
+
+        # Add waiting task to the real Redis queue
+        waiting_task = DocumentTask(tenant_id=tenant_id, dataset_id=dataset.id, document_ids=["waiting-doc-1"])
+        queue.push_tasks([asdict(waiting_task)])
+
+        # Act: Execute the wrapper function
+        _document_indexing_with_tenant_queue(tenant_id, dataset.id, document_ids, mock_task_func)
+
+        # Assert: Verify error was handled gracefully
+        # The function should not raise exceptions
+        mock_external_service_dependencies["indexing_runner"].assert_called_once()
+        mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
+
+        # Verify documents were still updated to parsing status before the exception
+        # Re-query documents from database since _document_indexing uses a different session
+        for doc_id in document_ids:
+            updated_document = db.session.query(Document).where(Document.id == doc_id).first()
+            assert updated_document.indexing_status == "parsing"
+            assert updated_document.processing_started_at is not None
+
+        # Verify waiting task was still processed despite core processing error
+        mock_task_func.delay.assert_called_once()
+
+        # Verify correct parameters for the call
+        call = mock_task_func.delay.call_args
+        assert call[1] == {"tenant_id": tenant_id, "dataset_id": dataset_id, "document_ids": ["waiting-doc-1"]}
+
+        # Verify queue is empty after processing (task was pulled)
+        remaining_tasks = queue.pull_tasks(count=10)
+        assert len(remaining_tasks) == 0
+
+    def test_document_indexing_with_tenant_queue_tenant_isolation(
+        self, db_session_with_containers, mock_external_service_dependencies
+    ):
+        """
+        Test tenant isolation in _document_indexing_with_tenant_queue using real Redis.
+
+        This test verifies:
+        - Different tenants have isolated queues
+        - Tasks from one tenant don't affect another tenant's queue
+        - Queue operations are properly scoped to tenant
+        """
+        # Arrange: Create test data for two different tenants
+        dataset1, documents1 = self._create_test_dataset_and_documents(
+            db_session_with_containers, mock_external_service_dependencies, document_count=1
+        )
+        dataset2, documents2 = self._create_test_dataset_and_documents(
+            db_session_with_containers, mock_external_service_dependencies, document_count=1
+        )
+
+        tenant1_id = dataset1.tenant_id
+        tenant2_id = dataset2.tenant_id
+        dataset1_id = dataset1.id
+        dataset2_id = dataset2.id
+        document_ids1 = [doc.id for doc in documents1]
+        document_ids2 = [doc.id for doc in documents2]
+
+        # Mock the task function
+        from unittest.mock import MagicMock
+
+        mock_task_func = MagicMock()
+
+        # Use real Redis for TenantIsolatedTaskQueue
+        from core.rag.pipeline.queue import TenantIsolatedTaskQueue
+
+        # Create queue instances for both tenants
+        queue1 = TenantIsolatedTaskQueue(tenant1_id, "document_indexing")
+        queue2 = TenantIsolatedTaskQueue(tenant2_id, "document_indexing")
+
+        # Add waiting tasks to both queues
+        waiting_task1 = DocumentTask(tenant_id=tenant1_id, dataset_id=dataset1.id, document_ids=["tenant1-doc-1"])
+        waiting_task2 = DocumentTask(tenant_id=tenant2_id, dataset_id=dataset2.id, document_ids=["tenant2-doc-1"])
+
+        queue1.push_tasks([asdict(waiting_task1)])
+        queue2.push_tasks([asdict(waiting_task2)])
+
+        # Act: Execute the wrapper function for tenant1 only
+        _document_indexing_with_tenant_queue(tenant1_id, dataset1.id, document_ids1, mock_task_func)
+
+        # Assert: Verify core processing occurred for tenant1
+        mock_external_service_dependencies["indexing_runner"].assert_called_once()
+        mock_external_service_dependencies["indexing_runner_instance"].run.assert_called_once()
+
+        # Verify only tenant1's waiting task was processed
+        mock_task_func.delay.assert_called_once()
+        call = mock_task_func.delay.call_args
+        assert call[1] == {"tenant_id": tenant1_id, "dataset_id": dataset1_id, "document_ids": ["tenant1-doc-1"]}
+
+        # Verify tenant1's queue is empty
+        remaining_tasks1 = queue1.pull_tasks(count=10)
+        assert len(remaining_tasks1) == 0
+
+        # Verify tenant2's queue still has its task (isolation)
+        remaining_tasks2 = queue2.pull_tasks(count=10)
+        assert len(remaining_tasks2) == 1
+
+        # Verify queue keys are different
+        assert queue1._queue != queue2._queue
+        assert queue1._task_key != queue2._task_key

api/tests/test_containers_integration_tests/tasks/test_rag_pipeline_run_tasks.py

@@ -0,0 +1,936 @@
+import json
+import uuid
+from unittest.mock import patch
+
+import pytest
+from faker import Faker
+
+from core.app.entities.app_invoke_entities import InvokeFrom, RagPipelineGenerateEntity
+from core.app.entities.rag_pipeline_invoke_entities import RagPipelineInvokeEntity
+from core.rag.pipeline.queue import TenantIsolatedTaskQueue
+from extensions.ext_database import db
+from models import Account, Tenant, TenantAccountJoin, TenantAccountRole
+from models.dataset import Pipeline
+from models.workflow import Workflow
+from tasks.rag_pipeline.priority_rag_pipeline_run_task import (
+    priority_rag_pipeline_run_task,
+    run_single_rag_pipeline_task,
+)
+from tasks.rag_pipeline.rag_pipeline_run_task import rag_pipeline_run_task
+
+
+class TestRagPipelineRunTasks:
+    """Integration tests for RAG pipeline run tasks using testcontainers.
+
+    This test class covers:
+    - priority_rag_pipeline_run_task function
+    - rag_pipeline_run_task function
+    - run_single_rag_pipeline_task function
+    - Real Redis-based TenantIsolatedTaskQueue operations
+    - PipelineGenerator._generate method mocking and parameter validation
+    - File operations and cleanup
+    - Error handling and queue management
+    """
+
+    @pytest.fixture
+    def mock_pipeline_generator(self):
+        """Mock PipelineGenerator._generate method."""
+        with patch("core.app.apps.pipeline.pipeline_generator.PipelineGenerator._generate") as mock_generate:
+            # Mock the _generate method to return a simple response
+            mock_generate.return_value = {"answer": "Test response", "metadata": {"test": "data"}}
+            yield mock_generate
+
+    @pytest.fixture
+    def mock_file_service(self):
+        """Mock FileService for file operations."""
+        with (
+            patch("services.file_service.FileService.get_file_content") as mock_get_content,
+            patch("services.file_service.FileService.delete_file") as mock_delete_file,
+        ):
+            yield {
+                "get_content": mock_get_content,
+                "delete_file": mock_delete_file,
+            }
+
+    def _create_test_pipeline_and_workflow(self, db_session_with_containers):
+        """
+        Helper method to create test pipeline and workflow for testing.
+
+        Args:
+            db_session_with_containers: Database session from testcontainers infrastructure
+
+        Returns:
+            tuple: (account, tenant, pipeline, workflow) - Created entities
+        """
+        fake = Faker()
+
+        # Create account and tenant
+        account = Account(
+            email=fake.email(),
+            name=fake.name(),
+            interface_language="en-US",
+            status="active",
+        )
+        db.session.add(account)
+        db.session.commit()
+
+        tenant = Tenant(
+            name=fake.company(),
+            status="normal",
+        )
+        db.session.add(tenant)
+        db.session.commit()
+
+        # Create tenant-account join
+        join = TenantAccountJoin(
+            tenant_id=tenant.id,
+            account_id=account.id,
+            role=TenantAccountRole.OWNER,
+            current=True,
+        )
+        db.session.add(join)
+        db.session.commit()
+
+        # Create workflow
+        workflow = Workflow(
+            id=str(uuid.uuid4()),
+            tenant_id=tenant.id,
+            app_id=str(uuid.uuid4()),
+            type="workflow",
+            version="draft",
+            graph="{}",
+            features="{}",
+            marked_name=fake.company(),
+            marked_comment=fake.text(max_nb_chars=100),
+            created_by=account.id,
+            environment_variables=[],
+            conversation_variables=[],
+            rag_pipeline_variables=[],
+        )
+        db.session.add(workflow)
+        db.session.commit()
+
+        # Create pipeline
+        pipeline = Pipeline(
+            id=str(uuid.uuid4()),
+            tenant_id=tenant.id,
+            workflow_id=workflow.id,
+            name=fake.company(),
+            description=fake.text(max_nb_chars=100),
+            created_by=account.id,
+        )
+        db.session.add(pipeline)
+        db.session.commit()
+
+        # Refresh entities to ensure they're properly loaded
+        db.session.refresh(account)
+        db.session.refresh(tenant)
+        db.session.refresh(workflow)
+        db.session.refresh(pipeline)
+
+        return account, tenant, pipeline, workflow
+
+    def _create_rag_pipeline_invoke_entities(self, account, tenant, pipeline, workflow, count=2):
+        """
+        Helper method to create RAG pipeline invoke entities for testing.
+
+        Args:
+            account: Account instance
+            tenant: Tenant instance
+            pipeline: Pipeline instance
+            workflow: Workflow instance
+            count: Number of entities to create
+
+        Returns:
+            list: List of RagPipelineInvokeEntity instances
+        """
+        fake = Faker()
+        entities = []
+
+        for i in range(count):
+            # Create application generate entity
+            app_config = {
+                "app_id": str(uuid.uuid4()),
+                "app_name": fake.company(),
+                "mode": "workflow",
+                "workflow_id": workflow.id,
+                "tenant_id": tenant.id,
+                "app_mode": "workflow",
+            }
+
+            application_generate_entity = {
+                "task_id": str(uuid.uuid4()),
+                "app_config": app_config,
+                "inputs": {"query": f"Test query {i}"},
+                "files": [],
+                "user_id": account.id,
+                "stream": False,
+                "invoke_from": "published",
+                "workflow_execution_id": str(uuid.uuid4()),
+                "pipeline_config": {
+                    "app_id": str(uuid.uuid4()),
+                    "app_name": fake.company(),
+                    "mode": "workflow",
+                    "workflow_id": workflow.id,
+                    "tenant_id": tenant.id,
+                    "app_mode": "workflow",
+                },
+                "datasource_type": "upload_file",
+                "datasource_info": {},
+                "dataset_id": str(uuid.uuid4()),
+                "batch": "test_batch",
+            }
+
+            entity = RagPipelineInvokeEntity(
+                pipeline_id=pipeline.id,
+                application_generate_entity=application_generate_entity,
+                user_id=account.id,
+                tenant_id=tenant.id,
+                workflow_id=workflow.id,
+                streaming=False,
+                workflow_execution_id=str(uuid.uuid4()),
+                workflow_thread_pool_id=str(uuid.uuid4()),
+            )
+            entities.append(entity)
+
+        return entities
+
+    def _create_file_content_for_entities(self, entities):
+        """
+        Helper method to create file content for RAG pipeline invoke entities.
+
+        Args:
+            entities: List of RagPipelineInvokeEntity instances
+
+        Returns:
+            str: JSON string containing serialized entities
+        """
+        entities_data = [entity.model_dump() for entity in entities]
+        return json.dumps(entities_data)
+
+    def test_priority_rag_pipeline_run_task_success(
+        self, db_session_with_containers, mock_pipeline_generator, mock_file_service
+    ):
+        """
+        Test successful priority RAG pipeline run task execution.
+
+        This test verifies:
+        - Task execution with multiple RAG pipeline invoke entities
+        - File content retrieval and parsing
+        - PipelineGenerator._generate method calls with correct parameters
+        - Thread pool execution
+        - File cleanup after execution
+        - Queue management with no waiting tasks
+        """
+        # Arrange: Create test data
+        account, tenant, pipeline, workflow = self._create_test_pipeline_and_workflow(db_session_with_containers)
+        entities = self._create_rag_pipeline_invoke_entities(account, tenant, pipeline, workflow, count=2)
+        file_content = self._create_file_content_for_entities(entities)
+
+        # Mock file service
+        file_id = str(uuid.uuid4())
+        mock_file_service["get_content"].return_value = file_content
+
+        # Act: Execute the priority task
+        priority_rag_pipeline_run_task(file_id, tenant.id)
+
+        # Assert: Verify expected outcomes
+        # Verify file operations
+        mock_file_service["get_content"].assert_called_once_with(file_id)
+        mock_file_service["delete_file"].assert_called_once_with(file_id)
+
+        # Verify PipelineGenerator._generate was called for each entity
+        assert mock_pipeline_generator.call_count == 2
+
+        # Verify call parameters for each entity
+        calls = mock_pipeline_generator.call_args_list
+        for call in calls:
+            call_kwargs = call[1]  # Get keyword arguments
+            assert call_kwargs["pipeline"].id == pipeline.id
+            assert call_kwargs["workflow_id"] == workflow.id
+            assert call_kwargs["user"].id == account.id
+            assert call_kwargs["invoke_from"] == InvokeFrom.PUBLISHED
+            assert call_kwargs["streaming"] == False
+            assert isinstance(call_kwargs["application_generate_entity"], RagPipelineGenerateEntity)
+
+    def test_rag_pipeline_run_task_success(
+        self, db_session_with_containers, mock_pipeline_generator, mock_file_service
+    ):
+        """
+        Test successful regular RAG pipeline run task execution.
+
+        This test verifies:
+        - Task execution with multiple RAG pipeline invoke entities
+        - File content retrieval and parsing
+        - PipelineGenerator._generate method calls with correct parameters
+        - Thread pool execution
+        - File cleanup after execution
+        - Queue management with no waiting tasks
+        """
+        # Arrange: Create test data
+        account, tenant, pipeline, workflow = self._create_test_pipeline_and_workflow(db_session_with_containers)
+        entities = self._create_rag_pipeline_invoke_entities(account, tenant, pipeline, workflow, count=3)
+        file_content = self._create_file_content_for_entities(entities)
+
+        # Mock file service
+        file_id = str(uuid.uuid4())
+        mock_file_service["get_content"].return_value = file_content
+
+        # Act: Execute the regular task
+        rag_pipeline_run_task(file_id, tenant.id)
+
+        # Assert: Verify expected outcomes
+        # Verify file operations
+        mock_file_service["get_content"].assert_called_once_with(file_id)
+        mock_file_service["delete_file"].assert_called_once_with(file_id)
+
+        # Verify PipelineGenerator._generate was called for each entity
+        assert mock_pipeline_generator.call_count == 3
+
+        # Verify call parameters for each entity
+        calls = mock_pipeline_generator.call_args_list
+        for call in calls:
+            call_kwargs = call[1]  # Get keyword arguments
+            assert call_kwargs["pipeline"].id == pipeline.id
+            assert call_kwargs["workflow_id"] == workflow.id
+            assert call_kwargs["user"].id == account.id
+            assert call_kwargs["invoke_from"] == InvokeFrom.PUBLISHED
+            assert call_kwargs["streaming"] == False
+            assert isinstance(call_kwargs["application_generate_entity"], RagPipelineGenerateEntity)
+
+    def test_priority_rag_pipeline_run_task_with_waiting_tasks(
+        self, db_session_with_containers, mock_pipeline_generator, mock_file_service
+    ):
+        """
+        Test priority RAG pipeline run task with waiting tasks in queue using real Redis.
+
+        This test verifies:
+        - Core task execution
+        - Real Redis-based tenant queue processing of waiting tasks
+        - Task function calls for waiting tasks
+        - Queue management with multiple tasks using actual Redis operations
+        """
+        # Arrange: Create test data
+        account, tenant, pipeline, workflow = self._create_test_pipeline_and_workflow(db_session_with_containers)
+        entities = self._create_rag_pipeline_invoke_entities(account, tenant, pipeline, workflow, count=1)
+        file_content = self._create_file_content_for_entities(entities)
+
+        # Mock file service
+        file_id = str(uuid.uuid4())
+        mock_file_service["get_content"].return_value = file_content
+
+        # Use real Redis for TenantIsolatedTaskQueue
+        queue = TenantIsolatedTaskQueue(tenant.id, "pipeline")
+
+        # Add waiting tasks to the real Redis queue
+        waiting_file_ids = [str(uuid.uuid4()) for _ in range(2)]
+        queue.push_tasks(waiting_file_ids)
+
+        # Mock the task function calls
+        with patch(
+            "tasks.rag_pipeline.priority_rag_pipeline_run_task.priority_rag_pipeline_run_task.delay"
+        ) as mock_delay:
+            # Act: Execute the priority task
+            priority_rag_pipeline_run_task(file_id, tenant.id)
+
+            # Assert: Verify core processing occurred
+            mock_file_service["get_content"].assert_called_once_with(file_id)
+            mock_file_service["delete_file"].assert_called_once_with(file_id)
+            assert mock_pipeline_generator.call_count == 1
+
+            # Verify waiting tasks were processed, pull 1 task a time by default
+            assert mock_delay.call_count == 1
+
+            # Verify correct parameters for the call
+            call_kwargs = mock_delay.call_args[1] if mock_delay.call_args else {}
+            assert call_kwargs.get("rag_pipeline_invoke_entities_file_id") == waiting_file_ids[0]
+            assert call_kwargs.get("tenant_id") == tenant.id
+
+            # Verify queue still has remaining tasks (only 1 was pulled)
+            remaining_tasks = queue.pull_tasks(count=10)
+            assert len(remaining_tasks) == 1  # 2 original - 1 pulled = 1 remaining
+
+    def test_rag_pipeline_run_task_legacy_compatibility(
+        self, db_session_with_containers, mock_pipeline_generator, mock_file_service
+    ):
+        """
+        Test regular RAG pipeline run task with legacy Redis queue format for backward compatibility.
+
+        This test simulates the scenario where:
+        - Old code writes file IDs directly to Redis list using lpush
+        - New worker processes these legacy queue entries
+        - Ensures backward compatibility during deployment transition
+
+        Legacy format: redis_client.lpush(tenant_self_pipeline_task_queue, upload_file.id)
+        New format: TenantIsolatedTaskQueue.push_tasks([file_id])
+        """
+        # Arrange: Create test data
+        account, tenant, pipeline, workflow = self._create_test_pipeline_and_workflow(db_session_with_containers)
+        entities = self._create_rag_pipeline_invoke_entities(account, tenant, pipeline, workflow, count=1)
+        file_content = self._create_file_content_for_entities(entities)
+
+        # Mock file service
+        file_id = str(uuid.uuid4())
+        mock_file_service["get_content"].return_value = file_content
+
+        # Simulate legacy Redis queue format - direct file IDs in Redis list
+        from extensions.ext_redis import redis_client
+
+        # Legacy queue key format (old code)
+        legacy_queue_key = f"tenant_self_pipeline_task_queue:{tenant.id}"
+        legacy_task_key = f"tenant_pipeline_task:{tenant.id}"
+
+        # Add legacy format data to Redis (simulating old code behavior)
+        legacy_file_ids = [str(uuid.uuid4()) for _ in range(3)]
+        for file_id_legacy in legacy_file_ids:
+            redis_client.lpush(legacy_queue_key, file_id_legacy)
+
+        # Set the task key to indicate there are waiting tasks (legacy behavior)
+        redis_client.set(legacy_task_key, 1, ex=60 * 60)
+
+        # Mock the task function calls
+        with patch("tasks.rag_pipeline.rag_pipeline_run_task.rag_pipeline_run_task.delay") as mock_delay:
+            # Act: Execute the priority task with new code but legacy queue data
+            rag_pipeline_run_task(file_id, tenant.id)
+
+            # Assert: Verify core processing occurred
+            mock_file_service["get_content"].assert_called_once_with(file_id)
+            mock_file_service["delete_file"].assert_called_once_with(file_id)
+            assert mock_pipeline_generator.call_count == 1
+
+            # Verify waiting tasks were processed, pull 1 task a time by default
+            assert mock_delay.call_count == 1
+
+            # Verify correct parameters for the call
+            call_kwargs = mock_delay.call_args[1] if mock_delay.call_args else {}
+            assert call_kwargs.get("rag_pipeline_invoke_entities_file_id") == legacy_file_ids[0]
+            assert call_kwargs.get("tenant_id") == tenant.id
+
+            # Verify that new code can process legacy queue entries
+            # The new TenantIsolatedTaskQueue should be able to read from the legacy format
+            queue = TenantIsolatedTaskQueue(tenant.id, "pipeline")
+
+            # Verify queue still has remaining tasks (only 1 was pulled)
+            remaining_tasks = queue.pull_tasks(count=10)
+            assert len(remaining_tasks) == 2  # 3 original - 1 pulled = 2 remaining
+
+        # Cleanup: Remove legacy test data
+        redis_client.delete(legacy_queue_key)
+        redis_client.delete(legacy_task_key)
+
+    def test_rag_pipeline_run_task_with_waiting_tasks(
+        self, db_session_with_containers, mock_pipeline_generator, mock_file_service
+    ):
+        """
+        Test regular RAG pipeline run task with waiting tasks in queue using real Redis.
+
+        This test verifies:
+        - Core task execution
+        - Real Redis-based tenant queue processing of waiting tasks
+        - Task function calls for waiting tasks
+        - Queue management with multiple tasks using actual Redis operations
+        """
+        # Arrange: Create test data
+        account, tenant, pipeline, workflow = self._create_test_pipeline_and_workflow(db_session_with_containers)
+        entities = self._create_rag_pipeline_invoke_entities(account, tenant, pipeline, workflow, count=1)
+        file_content = self._create_file_content_for_entities(entities)
+
+        # Mock file service
+        file_id = str(uuid.uuid4())
+        mock_file_service["get_content"].return_value = file_content
+
+        # Use real Redis for TenantIsolatedTaskQueue
+        queue = TenantIsolatedTaskQueue(tenant.id, "pipeline")
+
+        # Add waiting tasks to the real Redis queue
+        waiting_file_ids = [str(uuid.uuid4()) for _ in range(3)]
+        queue.push_tasks(waiting_file_ids)
+
+        # Mock the task function calls
+        with patch("tasks.rag_pipeline.rag_pipeline_run_task.rag_pipeline_run_task.delay") as mock_delay:
+            # Act: Execute the regular task
+            rag_pipeline_run_task(file_id, tenant.id)
+
+            # Assert: Verify core processing occurred
+            mock_file_service["get_content"].assert_called_once_with(file_id)
+            mock_file_service["delete_file"].assert_called_once_with(file_id)
+            assert mock_pipeline_generator.call_count == 1
+
+            # Verify waiting tasks were processed, pull 1 task a time by default
+            assert mock_delay.call_count == 1
+
+            # Verify correct parameters for the call
+            call_kwargs = mock_delay.call_args[1] if mock_delay.call_args else {}
+            assert call_kwargs.get("rag_pipeline_invoke_entities_file_id") == waiting_file_ids[0]
+            assert call_kwargs.get("tenant_id") == tenant.id
+
+            # Verify queue still has remaining tasks (only 1 was pulled)
+            remaining_tasks = queue.pull_tasks(count=10)
+            assert len(remaining_tasks) == 2  # 3 original - 1 pulled = 2 remaining
+
+    def test_priority_rag_pipeline_run_task_error_handling(
+        self, db_session_with_containers, mock_pipeline_generator, mock_file_service
+    ):
+        """
+        Test error handling in priority RAG pipeline run task using real Redis.
+
+        This test verifies:
+        - Exception handling during core processing
+        - Tenant queue cleanup even on errors using real Redis
+        - Proper error logging
+        - Function completes without raising exceptions
+        - Queue management continues despite core processing errors
+        """
+        # Arrange: Create test data
+        account, tenant, pipeline, workflow = self._create_test_pipeline_and_workflow(db_session_with_containers)
+        entities = self._create_rag_pipeline_invoke_entities(account, tenant, pipeline, workflow, count=1)
+        file_content = self._create_file_content_for_entities(entities)
+
+        # Mock file service
+        file_id = str(uuid.uuid4())
+        mock_file_service["get_content"].return_value = file_content
+
+        # Mock PipelineGenerator to raise an exception
+        mock_pipeline_generator.side_effect = Exception("Pipeline generation failed")
+
+        # Use real Redis for TenantIsolatedTaskQueue
+        queue = TenantIsolatedTaskQueue(tenant.id, "pipeline")
+
+        # Add waiting task to the real Redis queue
+        waiting_file_id = str(uuid.uuid4())
+        queue.push_tasks([waiting_file_id])
+
+        # Mock the task function calls
+        with patch(
+            "tasks.rag_pipeline.priority_rag_pipeline_run_task.priority_rag_pipeline_run_task.delay"
+        ) as mock_delay:
+            # Act: Execute the priority task (should not raise exception)
+            priority_rag_pipeline_run_task(file_id, tenant.id)
+
+            # Assert: Verify error was handled gracefully
+            # The function should not raise exceptions
+            mock_file_service["get_content"].assert_called_once_with(file_id)
+            mock_file_service["delete_file"].assert_called_once_with(file_id)
+            assert mock_pipeline_generator.call_count == 1
+
+            # Verify waiting task was still processed despite core processing error
+            mock_delay.assert_called_once()
+
+            # Verify correct parameters for the call
+            call_kwargs = mock_delay.call_args[1] if mock_delay.call_args else {}
+            assert call_kwargs.get("rag_pipeline_invoke_entities_file_id") == waiting_file_id
+            assert call_kwargs.get("tenant_id") == tenant.id
+
+            # Verify queue is empty after processing (task was pulled)
+            remaining_tasks = queue.pull_tasks(count=10)
+            assert len(remaining_tasks) == 0
+
+    def test_rag_pipeline_run_task_error_handling(
+        self, db_session_with_containers, mock_pipeline_generator, mock_file_service
+    ):
+        """
+        Test error handling in regular RAG pipeline run task using real Redis.
+
+        This test verifies:
+        - Exception handling during core processing
+        - Tenant queue cleanup even on errors using real Redis
+        - Proper error logging
+        - Function completes without raising exceptions
+        - Queue management continues despite core processing errors
+        """
+        # Arrange: Create test data
+        account, tenant, pipeline, workflow = self._create_test_pipeline_and_workflow(db_session_with_containers)
+        entities = self._create_rag_pipeline_invoke_entities(account, tenant, pipeline, workflow, count=1)
+        file_content = self._create_file_content_for_entities(entities)
+
+        # Mock file service
+        file_id = str(uuid.uuid4())
+        mock_file_service["get_content"].return_value = file_content
+
+        # Mock PipelineGenerator to raise an exception
+        mock_pipeline_generator.side_effect = Exception("Pipeline generation failed")
+
+        # Use real Redis for TenantIsolatedTaskQueue
+        queue = TenantIsolatedTaskQueue(tenant.id, "pipeline")
+
+        # Add waiting task to the real Redis queue
+        waiting_file_id = str(uuid.uuid4())
+        queue.push_tasks([waiting_file_id])
+
+        # Mock the task function calls
+        with patch("tasks.rag_pipeline.rag_pipeline_run_task.rag_pipeline_run_task.delay") as mock_delay:
+            # Act: Execute the regular task (should not raise exception)
+            rag_pipeline_run_task(file_id, tenant.id)
+
+            # Assert: Verify error was handled gracefully
+            # The function should not raise exceptions
+            mock_file_service["get_content"].assert_called_once_with(file_id)
+            mock_file_service["delete_file"].assert_called_once_with(file_id)
+            assert mock_pipeline_generator.call_count == 1
+
+            # Verify waiting task was still processed despite core processing error
+            mock_delay.assert_called_once()
+
+            # Verify correct parameters for the call
+            call_kwargs = mock_delay.call_args[1] if mock_delay.call_args else {}
+            assert call_kwargs.get("rag_pipeline_invoke_entities_file_id") == waiting_file_id
+            assert call_kwargs.get("tenant_id") == tenant.id
+
+            # Verify queue is empty after processing (task was pulled)
+            remaining_tasks = queue.pull_tasks(count=10)
+            assert len(remaining_tasks) == 0
+
+    def test_priority_rag_pipeline_run_task_tenant_isolation(
+        self, db_session_with_containers, mock_pipeline_generator, mock_file_service
+    ):
+        """
+        Test tenant isolation in priority RAG pipeline run task using real Redis.
+
+        This test verifies:
+        - Different tenants have isolated queues
+        - Tasks from one tenant don't affect another tenant's queue
+        - Queue operations are properly scoped to tenant
+        """
+        # Arrange: Create test data for two different tenants
+        account1, tenant1, pipeline1, workflow1 = self._create_test_pipeline_and_workflow(db_session_with_containers)
+        account2, tenant2, pipeline2, workflow2 = self._create_test_pipeline_and_workflow(db_session_with_containers)
+
+        entities1 = self._create_rag_pipeline_invoke_entities(account1, tenant1, pipeline1, workflow1, count=1)
+        entities2 = self._create_rag_pipeline_invoke_entities(account2, tenant2, pipeline2, workflow2, count=1)
+
+        file_content1 = self._create_file_content_for_entities(entities1)
+        file_content2 = self._create_file_content_for_entities(entities2)
+
+        # Mock file service
+        file_id1 = str(uuid.uuid4())
+        file_id2 = str(uuid.uuid4())
+        mock_file_service["get_content"].side_effect = [file_content1, file_content2]
+
+        # Use real Redis for TenantIsolatedTaskQueue
+        queue1 = TenantIsolatedTaskQueue(tenant1.id, "pipeline")
+        queue2 = TenantIsolatedTaskQueue(tenant2.id, "pipeline")
+
+        # Add waiting tasks to both queues
+        waiting_file_id1 = str(uuid.uuid4())
+        waiting_file_id2 = str(uuid.uuid4())
+
+        queue1.push_tasks([waiting_file_id1])
+        queue2.push_tasks([waiting_file_id2])
+
+        # Mock the task function calls
+        with patch(
+            "tasks.rag_pipeline.priority_rag_pipeline_run_task.priority_rag_pipeline_run_task.delay"
+        ) as mock_delay:
+            # Act: Execute the priority task for tenant1 only
+            priority_rag_pipeline_run_task(file_id1, tenant1.id)
+
+            # Assert: Verify core processing occurred for tenant1
+            assert mock_file_service["get_content"].call_count == 1
+            assert mock_file_service["delete_file"].call_count == 1
+            assert mock_pipeline_generator.call_count == 1
+
+            # Verify only tenant1's waiting task was processed
+            mock_delay.assert_called_once()
+            call_kwargs = mock_delay.call_args[1] if mock_delay.call_args else {}
+            assert call_kwargs.get("rag_pipeline_invoke_entities_file_id") == waiting_file_id1
+            assert call_kwargs.get("tenant_id") == tenant1.id
+
+            # Verify tenant1's queue is empty
+            remaining_tasks1 = queue1.pull_tasks(count=10)
+            assert len(remaining_tasks1) == 0
+
+            # Verify tenant2's queue still has its task (isolation)
+            remaining_tasks2 = queue2.pull_tasks(count=10)
+            assert len(remaining_tasks2) == 1
+
+            # Verify queue keys are different
+            assert queue1._queue != queue2._queue
+            assert queue1._task_key != queue2._task_key
+
+    def test_rag_pipeline_run_task_tenant_isolation(
+        self, db_session_with_containers, mock_pipeline_generator, mock_file_service
+    ):
+        """
+        Test tenant isolation in regular RAG pipeline run task using real Redis.
+
+        This test verifies:
+        - Different tenants have isolated queues
+        - Tasks from one tenant don't affect another tenant's queue
+        - Queue operations are properly scoped to tenant
+        """
+        # Arrange: Create test data for two different tenants
+        account1, tenant1, pipeline1, workflow1 = self._create_test_pipeline_and_workflow(db_session_with_containers)
+        account2, tenant2, pipeline2, workflow2 = self._create_test_pipeline_and_workflow(db_session_with_containers)
+
+        entities1 = self._create_rag_pipeline_invoke_entities(account1, tenant1, pipeline1, workflow1, count=1)
+        entities2 = self._create_rag_pipeline_invoke_entities(account2, tenant2, pipeline2, workflow2, count=1)
+
+        file_content1 = self._create_file_content_for_entities(entities1)
+        file_content2 = self._create_file_content_for_entities(entities2)
+
+        # Mock file service
+        file_id1 = str(uuid.uuid4())
+        file_id2 = str(uuid.uuid4())
+        mock_file_service["get_content"].side_effect = [file_content1, file_content2]
+
+        # Use real Redis for TenantIsolatedTaskQueue
+        queue1 = TenantIsolatedTaskQueue(tenant1.id, "pipeline")
+        queue2 = TenantIsolatedTaskQueue(tenant2.id, "pipeline")
+
+        # Add waiting tasks to both queues
+        waiting_file_id1 = str(uuid.uuid4())
+        waiting_file_id2 = str(uuid.uuid4())
+
+        queue1.push_tasks([waiting_file_id1])
+        queue2.push_tasks([waiting_file_id2])
+
+        # Mock the task function calls
+        with patch("tasks.rag_pipeline.rag_pipeline_run_task.rag_pipeline_run_task.delay") as mock_delay:
+            # Act: Execute the regular task for tenant1 only
+            rag_pipeline_run_task(file_id1, tenant1.id)
+
+            # Assert: Verify core processing occurred for tenant1
+            assert mock_file_service["get_content"].call_count == 1
+            assert mock_file_service["delete_file"].call_count == 1
+            assert mock_pipeline_generator.call_count == 1
+
+            # Verify only tenant1's waiting task was processed
+            mock_delay.assert_called_once()
+            call_kwargs = mock_delay.call_args[1] if mock_delay.call_args else {}
+            assert call_kwargs.get("rag_pipeline_invoke_entities_file_id") == waiting_file_id1
+            assert call_kwargs.get("tenant_id") == tenant1.id
+
+            # Verify tenant1's queue is empty
+            remaining_tasks1 = queue1.pull_tasks(count=10)
+            assert len(remaining_tasks1) == 0
+
+            # Verify tenant2's queue still has its task (isolation)
+            remaining_tasks2 = queue2.pull_tasks(count=10)
+            assert len(remaining_tasks2) == 1
+
+            # Verify queue keys are different
+            assert queue1._queue != queue2._queue
+            assert queue1._task_key != queue2._task_key
+
+    def test_run_single_rag_pipeline_task_success(
+        self, db_session_with_containers, mock_pipeline_generator, flask_app_with_containers
+    ):
+        """
+        Test successful run_single_rag_pipeline_task execution.
+
+        This test verifies:
+        - Single RAG pipeline task execution within Flask app context
+        - Entity validation and database queries
+        - PipelineGenerator._generate method call with correct parameters
+        - Proper Flask context handling
+        """
+        # Arrange: Create test data
+        account, tenant, pipeline, workflow = self._create_test_pipeline_and_workflow(db_session_with_containers)
+        entities = self._create_rag_pipeline_invoke_entities(account, tenant, pipeline, workflow, count=1)
+        entity_data = entities[0].model_dump()
+
+        # Act: Execute the single task
+        with flask_app_with_containers.app_context():
+            run_single_rag_pipeline_task(entity_data, flask_app_with_containers)
+
+        # Assert: Verify expected outcomes
+        # Verify PipelineGenerator._generate was called
+        assert mock_pipeline_generator.call_count == 1
+
+        # Verify call parameters
+        call = mock_pipeline_generator.call_args
+        call_kwargs = call[1]  # Get keyword arguments
+        assert call_kwargs["pipeline"].id == pipeline.id
+        assert call_kwargs["workflow_id"] == workflow.id
+        assert call_kwargs["user"].id == account.id
+        assert call_kwargs["invoke_from"] == InvokeFrom.PUBLISHED
+        assert call_kwargs["streaming"] == False
+        assert isinstance(call_kwargs["application_generate_entity"], RagPipelineGenerateEntity)
+
+    def test_run_single_rag_pipeline_task_entity_validation_error(
+        self, db_session_with_containers, mock_pipeline_generator, flask_app_with_containers
+    ):
+        """
+        Test run_single_rag_pipeline_task with invalid entity data.
+
+        This test verifies:
+        - Proper error handling for invalid entity data
+        - Exception logging
+        - Function raises ValueError for missing entities
+        """
+        # Arrange: Create entity data with valid UUIDs but non-existent entities
+        fake = Faker()
+        invalid_entity_data = {
+            "pipeline_id": str(uuid.uuid4()),
+            "application_generate_entity": {
+                "app_config": {
+                    "app_id": str(uuid.uuid4()),
+                    "app_name": "Test App",
+                    "mode": "workflow",
+                    "workflow_id": str(uuid.uuid4()),
+                },
+                "inputs": {"query": "Test query"},
+                "query": "Test query",
+                "response_mode": "blocking",
+                "user": str(uuid.uuid4()),
+                "files": [],
+                "conversation_id": str(uuid.uuid4()),
+            },
+            "user_id": str(uuid.uuid4()),
+            "tenant_id": str(uuid.uuid4()),
+            "workflow_id": str(uuid.uuid4()),
+            "streaming": False,
+            "workflow_execution_id": str(uuid.uuid4()),
+            "workflow_thread_pool_id": str(uuid.uuid4()),
+        }
+
+        # Act & Assert: Execute the single task with non-existent entities (should raise ValueError)
+        with flask_app_with_containers.app_context():
+            with pytest.raises(ValueError, match="Account .* not found"):
+                run_single_rag_pipeline_task(invalid_entity_data, flask_app_with_containers)
+
+        # Assert: Pipeline generator should not be called
+        mock_pipeline_generator.assert_not_called()
+
+    def test_run_single_rag_pipeline_task_database_entity_not_found(
+        self, db_session_with_containers, mock_pipeline_generator, flask_app_with_containers
+    ):
+        """
+        Test run_single_rag_pipeline_task with non-existent database entities.
+
+        This test verifies:
+        - Proper error handling for missing database entities
+        - Exception logging
+        - Function raises ValueError for missing entities
+        """
+        # Arrange: Create test data with non-existent IDs
+        fake = Faker()
+        entity_data = {
+            "pipeline_id": str(uuid.uuid4()),
+            "application_generate_entity": {
+                "app_config": {
+                    "app_id": str(uuid.uuid4()),
+                    "app_name": "Test App",
+                    "mode": "workflow",
+                    "workflow_id": str(uuid.uuid4()),
+                },
+                "inputs": {"query": "Test query"},
+                "query": "Test query",
+                "response_mode": "blocking",
+                "user": str(uuid.uuid4()),
+                "files": [],
+                "conversation_id": str(uuid.uuid4()),
+            },
+            "user_id": str(uuid.uuid4()),
+            "tenant_id": str(uuid.uuid4()),
+            "workflow_id": str(uuid.uuid4()),
+            "streaming": False,
+            "workflow_execution_id": str(uuid.uuid4()),
+            "workflow_thread_pool_id": str(uuid.uuid4()),
+        }
+
+        # Act & Assert: Execute the single task with non-existent entities (should raise ValueError)
+        with flask_app_with_containers.app_context():
+            with pytest.raises(ValueError, match="Account .* not found"):
+                run_single_rag_pipeline_task(entity_data, flask_app_with_containers)
+
+        # Assert: Pipeline generator should not be called
+        mock_pipeline_generator.assert_not_called()
+
+    def test_priority_rag_pipeline_run_task_file_not_found(
+        self, db_session_with_containers, mock_pipeline_generator, mock_file_service
+    ):
+        """
+        Test priority RAG pipeline run task with non-existent file.
+
+        This test verifies:
+        - Proper error handling for missing files
+        - Exception logging
+        - Function raises Exception for file errors
+        - Queue management continues despite file errors
+        """
+        # Arrange: Create test data
+        account, tenant, pipeline, workflow = self._create_test_pipeline_and_workflow(db_session_with_containers)
+
+        # Mock file service to raise exception
+        file_id = str(uuid.uuid4())
+        mock_file_service["get_content"].side_effect = Exception("File not found")
+
+        # Use real Redis for TenantIsolatedTaskQueue
+        queue = TenantIsolatedTaskQueue(tenant.id, "pipeline")
+
+        # Add waiting task to the real Redis queue
+        waiting_file_id = str(uuid.uuid4())
+        queue.push_tasks([waiting_file_id])
+
+        # Mock the task function calls
+        with patch(
+            "tasks.rag_pipeline.priority_rag_pipeline_run_task.priority_rag_pipeline_run_task.delay"
+        ) as mock_delay:
+            # Act & Assert: Execute the priority task (should raise Exception)
+            with pytest.raises(Exception, match="File not found"):
+                priority_rag_pipeline_run_task(file_id, tenant.id)
+
+            # Assert: Verify error was handled gracefully
+            mock_file_service["get_content"].assert_called_once_with(file_id)
+            mock_pipeline_generator.assert_not_called()
+
+            # Verify waiting task was still processed despite file error
+            mock_delay.assert_called_once()
+
+            # Verify correct parameters for the call
+            call_kwargs = mock_delay.call_args[1] if mock_delay.call_args else {}
+            assert call_kwargs.get("rag_pipeline_invoke_entities_file_id") == waiting_file_id
+            assert call_kwargs.get("tenant_id") == tenant.id
+
+            # Verify queue is empty after processing (task was pulled)
+            remaining_tasks = queue.pull_tasks(count=10)
+            assert len(remaining_tasks) == 0
+
+    def test_rag_pipeline_run_task_file_not_found(
+        self, db_session_with_containers, mock_pipeline_generator, mock_file_service
+    ):
+        """
+        Test regular RAG pipeline run task with non-existent file.
+
+        This test verifies:
+        - Proper error handling for missing files
+        - Exception logging
+        - Function raises Exception for file errors
+        - Queue management continues despite file errors
+        """
+        # Arrange: Create test data
+        account, tenant, pipeline, workflow = self._create_test_pipeline_and_workflow(db_session_with_containers)
+
+        # Mock file service to raise exception
+        file_id = str(uuid.uuid4())
+        mock_file_service["get_content"].side_effect = Exception("File not found")
+
+        # Use real Redis for TenantIsolatedTaskQueue
+        queue = TenantIsolatedTaskQueue(tenant.id, "pipeline")
+
+        # Add waiting task to the real Redis queue
+        waiting_file_id = str(uuid.uuid4())
+        queue.push_tasks([waiting_file_id])
+
+        # Mock the task function calls
+        with patch("tasks.rag_pipeline.rag_pipeline_run_task.rag_pipeline_run_task.delay") as mock_delay:
+            # Act & Assert: Execute the regular task (should raise Exception)
+            with pytest.raises(Exception, match="File not found"):
+                rag_pipeline_run_task(file_id, tenant.id)
+
+            # Assert: Verify error was handled gracefully
+            mock_file_service["get_content"].assert_called_once_with(file_id)
+            mock_pipeline_generator.assert_not_called()
+
+            # Verify waiting task was still processed despite file error
+            mock_delay.assert_called_once()
+
+            # Verify correct parameters for the call
+            call_kwargs = mock_delay.call_args[1] if mock_delay.call_args else {}
+            assert call_kwargs.get("rag_pipeline_invoke_entities_file_id") == waiting_file_id
+            assert call_kwargs.get("tenant_id") == tenant.id
+
+            # Verify queue is empty after processing (task was pulled)
+            remaining_tasks = queue.pull_tasks(count=10)
+            assert len(remaining_tasks) == 0

api/tests/unit_tests/core/file/test_models.py

@@ -23,3 +23,32 @@ def test_file():
     assert file.extension == ".png"
     assert file.mime_type == "image/png"
     assert file.size == 67
+
+
+def test_file_model_validate_with_legacy_fields():
+    """Test `File` model can handle data containing compatibility fields."""
+    data = {
+        "id": "test-file",
+        "tenant_id": "test-tenant-id",
+        "type": "image",
+        "transfer_method": "tool_file",
+        "related_id": "test-related-id",
+        "filename": "image.png",
+        "extension": ".png",
+        "mime_type": "image/png",
+        "size": 67,
+        "storage_key": "test-storage-key",
+        "url": "https://example.com/image.png",
+        # Extra legacy fields
+        "tool_file_id": "tool-file-123",
+        "upload_file_id": "upload-file-456",
+        "datasource_file_id": "datasource-file-789",
+    }
+
+    # Should be able to create `File` object without raising an exception
+    file = File.model_validate(data)
+
+    # The File object does not have tool_file_id, upload_file_id, or datasource_file_id as attributes.
+    # Instead, check it does not expose unrecognized legacy fields (should raise on getattr).
+    for legacy_field in ("tool_file_id", "upload_file_id", "datasource_file_id"):
+        assert not hasattr(file, legacy_field)

api/tests/unit_tests/core/helper/code_executor/javascript/test_javascript_transformer.py

@@ -0,0 +1,12 @@
+from core.helper.code_executor.javascript.javascript_code_provider import JavascriptCodeProvider
+from core.helper.code_executor.javascript.javascript_transformer import NodeJsTemplateTransformer
+
+
+def test_get_runner_script():
+    code = JavascriptCodeProvider.get_default_code()
+    inputs = {"arg1": "hello, ", "arg2": "world!"}
+    script = NodeJsTemplateTransformer.assemble_runner_script(code, inputs)
+    script_lines = script.splitlines()
+    code_lines = code.splitlines()
+    # Check that the first lines of script are exactly the same as code
+    assert script_lines[: len(code_lines)] == code_lines

api/tests/unit_tests/core/helper/code_executor/python3/test_python3_transformer.py

@@ -0,0 +1,12 @@
+from core.helper.code_executor.python3.python3_code_provider import Python3CodeProvider
+from core.helper.code_executor.python3.python3_transformer import Python3TemplateTransformer
+
+
+def test_get_runner_script():
+    code = Python3CodeProvider.get_default_code()
+    inputs = {"arg1": "hello, ", "arg2": "world!"}
+    script = Python3TemplateTransformer.assemble_runner_script(code, inputs)
+    script_lines = script.splitlines()
+    code_lines = code.splitlines()
+    # Check that the first lines of script are exactly the same as code
+    assert script_lines[: len(code_lines)] == code_lines

api/tests/unit_tests/core/rag/pipeline/test_queue.py

@@ -0,0 +1,301 @@
+"""
+Unit tests for TenantIsolatedTaskQueue.
+
+These tests verify the Redis-based task queue functionality for tenant-specific
+task management with proper serialization and deserialization.
+"""
+
+import json
+from unittest.mock import MagicMock, patch
+from uuid import uuid4
+
+import pytest
+from pydantic import ValidationError
+
+from core.rag.pipeline.queue import TaskWrapper, TenantIsolatedTaskQueue
+
+
+class TestTaskWrapper:
+    """Test cases for TaskWrapper serialization/deserialization."""
+
+    def test_serialize_simple_data(self):
+        """Test serialization of simple data types."""
+        data = {"key": "value", "number": 42, "list": [1, 2, 3]}
+        wrapper = TaskWrapper(data=data)
+
+        serialized = wrapper.serialize()
+        assert isinstance(serialized, str)
+
+        # Verify it's valid JSON
+        parsed = json.loads(serialized)
+        assert parsed["data"] == data
+
+    def test_serialize_complex_data(self):
+        """Test serialization of complex nested data."""
+        data = {
+            "nested": {"deep": {"value": "test", "numbers": [1, 2, 3, 4, 5]}},
+            "unicode": "测试中文",
+            "special_chars": "!@#$%^&*()",
+        }
+        wrapper = TaskWrapper(data=data)
+
+        serialized = wrapper.serialize()
+        parsed = json.loads(serialized)
+        assert parsed["data"] == data
+
+    def test_deserialize_valid_data(self):
+        """Test deserialization of valid JSON data."""
+        original_data = {"key": "value", "number": 42}
+        # Serialize using TaskWrapper to get the correct format
+        wrapper = TaskWrapper(data=original_data)
+        serialized = wrapper.serialize()
+
+        wrapper = TaskWrapper.deserialize(serialized)
+        assert wrapper.data == original_data
+
+    def test_deserialize_invalid_json(self):
+        """Test deserialization handles invalid JSON gracefully."""
+        invalid_json = "{invalid json}"
+
+        # Pydantic will raise ValidationError for invalid JSON
+        with pytest.raises(ValidationError):
+            TaskWrapper.deserialize(invalid_json)
+
+    def test_serialize_ensure_ascii_false(self):
+        """Test that serialization preserves Unicode characters."""
+        data = {"chinese": "中文测试", "emoji": "🚀"}
+        wrapper = TaskWrapper(data=data)
+
+        serialized = wrapper.serialize()
+        assert "中文测试" in serialized
+        assert "🚀" in serialized
+
+
+class TestTenantIsolatedTaskQueue:
+    """Test cases for TenantIsolatedTaskQueue functionality."""
+
+    @pytest.fixture
+    def mock_redis_client(self):
+        """Mock Redis client for testing."""
+        mock_redis = MagicMock()
+        return mock_redis
+
+    @pytest.fixture
+    def sample_queue(self, mock_redis_client):
+        """Create a sample TenantIsolatedTaskQueue instance."""
+        return TenantIsolatedTaskQueue("tenant-123", "test-key")
+
+    def test_initialization(self, sample_queue):
+        """Test queue initialization with correct key generation."""
+        assert sample_queue._tenant_id == "tenant-123"
+        assert sample_queue._unique_key == "test-key"
+        assert sample_queue._queue == "tenant_self_test-key_task_queue:tenant-123"
+        assert sample_queue._task_key == "tenant_test-key_task:tenant-123"
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_get_task_key_exists(self, mock_redis, sample_queue):
+        """Test getting task key when it exists."""
+        mock_redis.get.return_value = "1"
+
+        result = sample_queue.get_task_key()
+
+        assert result == "1"
+        mock_redis.get.assert_called_once_with("tenant_test-key_task:tenant-123")
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_get_task_key_not_exists(self, mock_redis, sample_queue):
+        """Test getting task key when it doesn't exist."""
+        mock_redis.get.return_value = None
+
+        result = sample_queue.get_task_key()
+
+        assert result is None
+        mock_redis.get.assert_called_once_with("tenant_test-key_task:tenant-123")
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_set_task_waiting_time_default_ttl(self, mock_redis, sample_queue):
+        """Test setting task waiting flag with default TTL."""
+        sample_queue.set_task_waiting_time()
+
+        mock_redis.setex.assert_called_once_with(
+            "tenant_test-key_task:tenant-123",
+            3600,  # DEFAULT_TASK_TTL
+            1,
+        )
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_set_task_waiting_time_custom_ttl(self, mock_redis, sample_queue):
+        """Test setting task waiting flag with custom TTL."""
+        custom_ttl = 1800
+        sample_queue.set_task_waiting_time(custom_ttl)
+
+        mock_redis.setex.assert_called_once_with("tenant_test-key_task:tenant-123", custom_ttl, 1)
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_delete_task_key(self, mock_redis, sample_queue):
+        """Test deleting task key."""
+        sample_queue.delete_task_key()
+
+        mock_redis.delete.assert_called_once_with("tenant_test-key_task:tenant-123")
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_push_tasks_string_list(self, mock_redis, sample_queue):
+        """Test pushing string tasks directly."""
+        tasks = ["task1", "task2", "task3"]
+
+        sample_queue.push_tasks(tasks)
+
+        mock_redis.lpush.assert_called_once_with(
+            "tenant_self_test-key_task_queue:tenant-123", "task1", "task2", "task3"
+        )
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_push_tasks_mixed_types(self, mock_redis, sample_queue):
+        """Test pushing mixed string and object tasks."""
+        tasks = ["string_task", {"object_task": "data", "id": 123}, "another_string"]
+
+        sample_queue.push_tasks(tasks)
+
+        # Verify lpush was called
+        mock_redis.lpush.assert_called_once()
+        call_args = mock_redis.lpush.call_args
+
+        # Check queue name
+        assert call_args[0][0] == "tenant_self_test-key_task_queue:tenant-123"
+
+        # Check serialized tasks
+        serialized_tasks = call_args[0][1:]
+        assert len(serialized_tasks) == 3
+        assert serialized_tasks[0] == "string_task"
+        assert serialized_tasks[2] == "another_string"
+
+        # Check object task is serialized as TaskWrapper JSON (without prefix)
+        # It should be a valid JSON string that can be deserialized by TaskWrapper
+        wrapper = TaskWrapper.deserialize(serialized_tasks[1])
+        assert wrapper.data == {"object_task": "data", "id": 123}
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_push_tasks_empty_list(self, mock_redis, sample_queue):
+        """Test pushing empty task list."""
+        sample_queue.push_tasks([])
+
+        mock_redis.lpush.assert_called_once_with("tenant_self_test-key_task_queue:tenant-123")
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_pull_tasks_default_count(self, mock_redis, sample_queue):
+        """Test pulling tasks with default count (1)."""
+        mock_redis.rpop.side_effect = ["task1", None]
+
+        result = sample_queue.pull_tasks()
+
+        assert result == ["task1"]
+        assert mock_redis.rpop.call_count == 1
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_pull_tasks_custom_count(self, mock_redis, sample_queue):
+        """Test pulling tasks with custom count."""
+        # First test: pull 3 tasks
+        mock_redis.rpop.side_effect = ["task1", "task2", "task3", None]
+
+        result = sample_queue.pull_tasks(3)
+
+        assert result == ["task1", "task2", "task3"]
+        assert mock_redis.rpop.call_count == 3
+
+        # Reset mock for second test
+        mock_redis.reset_mock()
+        mock_redis.rpop.side_effect = ["task1", "task2", None]
+
+        result = sample_queue.pull_tasks(3)
+
+        assert result == ["task1", "task2"]
+        assert mock_redis.rpop.call_count == 3
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_pull_tasks_zero_count(self, mock_redis, sample_queue):
+        """Test pulling tasks with zero count returns empty list."""
+        result = sample_queue.pull_tasks(0)
+
+        assert result == []
+        mock_redis.rpop.assert_not_called()
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_pull_tasks_negative_count(self, mock_redis, sample_queue):
+        """Test pulling tasks with negative count returns empty list."""
+        result = sample_queue.pull_tasks(-1)
+
+        assert result == []
+        mock_redis.rpop.assert_not_called()
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_pull_tasks_with_wrapped_objects(self, mock_redis, sample_queue):
+        """Test pulling tasks that include wrapped objects."""
+        # Create a wrapped task
+        task_data = {"task_id": 123, "data": "test"}
+        wrapper = TaskWrapper(data=task_data)
+        wrapped_task = wrapper.serialize()
+
+        mock_redis.rpop.side_effect = [
+            "string_task",
+            wrapped_task.encode("utf-8"),  # Simulate bytes from Redis
+            None,
+        ]
+
+        result = sample_queue.pull_tasks(2)
+
+        assert len(result) == 2
+        assert result[0] == "string_task"
+        assert result[1] == {"task_id": 123, "data": "test"}
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_pull_tasks_with_invalid_wrapped_data(self, mock_redis, sample_queue):
+        """Test pulling tasks with invalid JSON falls back to string."""
+        # Invalid JSON string that cannot be deserialized
+        invalid_json = "invalid json data"
+        mock_redis.rpop.side_effect = [invalid_json, None]
+
+        result = sample_queue.pull_tasks(1)
+
+        assert result == [invalid_json]
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_pull_tasks_bytes_decoding(self, mock_redis, sample_queue):
+        """Test pulling tasks handles bytes from Redis correctly."""
+        mock_redis.rpop.side_effect = [
+            b"task1",  # bytes
+            "task2",  # string
+            None,
+        ]
+
+        result = sample_queue.pull_tasks(2)
+
+        assert result == ["task1", "task2"]
+
+    @patch("core.rag.pipeline.queue.redis_client")
+    def test_complex_object_serialization_roundtrip(self, mock_redis, sample_queue):
+        """Test complex object serialization and deserialization roundtrip."""
+        complex_task = {
+            "id": uuid4().hex,
+            "data": {"nested": {"deep": [1, 2, 3], "unicode": "测试中文", "special": "!@#$%^&*()"}},
+            "metadata": {"created_at": "2024-01-01T00:00:00Z", "tags": ["tag1", "tag2", "tag3"]},
+        }
+
+        # Push the complex task
+        sample_queue.push_tasks([complex_task])
+
+        # Verify it was serialized as TaskWrapper JSON
+        call_args = mock_redis.lpush.call_args
+        wrapped_task = call_args[0][1]
+        # Verify it's a valid TaskWrapper JSON (starts with {"data":)
+        assert wrapped_task.startswith('{"data":')
+
+        # Verify it can be deserialized
+        wrapper = TaskWrapper.deserialize(wrapped_task)
+        assert wrapper.data == complex_task
+
+        # Simulate pulling it back
+        mock_redis.rpop.return_value = wrapped_task
+        result = sample_queue.pull_tasks(1)
+
+        assert len(result) == 1
+        assert result[0] == complex_task

api/tests/unit_tests/core/workflow/entities/test_variable_pool.py

@@ -111,3 +111,26 @@ class TestVariablePoolGetAndNestedAttribute:
         assert segment_false is not None
         assert isinstance(segment_false, BooleanSegment)
         assert segment_false.value is False
+
+
+class TestVariablePoolGetNotModifyVariableDictionary:
+    _NODE_ID = "start"
+    _VAR_NAME = "name"
+
+    def test_convert_to_template_should_not_introduce_extra_keys(self):
+        pool = VariablePool.empty()
+        pool.add([self._NODE_ID, self._VAR_NAME], 0)
+        pool.convert_template("The start.name is {{#start.name#}}")
+        assert "The start" not in pool.variable_dictionary
+
+    def test_get_should_not_modify_variable_dictionary(self):
+        pool = VariablePool.empty()
+        pool.get([self._NODE_ID, self._VAR_NAME])
+        assert len(pool.variable_dictionary) == 1  # only contains `sys` node id
+        assert "start" not in pool.variable_dictionary
+
+        pool = VariablePool.empty()
+        pool.add([self._NODE_ID, self._VAR_NAME], "Joe")
+        pool.get([self._NODE_ID, "count"])
+        start_subdict = pool.variable_dictionary[self._NODE_ID]
+        assert "count" not in start_subdict

api/tests/unit_tests/libs/test_datetime_utils.py

@@ -1,8 +1,10 @@
 import datetime
+from unittest.mock import patch
 
 import pytest
+import pytz
 
-from libs.datetime_utils import naive_utc_now
+from libs.datetime_utils import naive_utc_now, parse_time_range
 
 
 def test_naive_utc_now(monkeypatch: pytest.MonkeyPatch):
@@ -20,3 +22,247 @@ def test_naive_utc_now(monkeypatch: pytest.MonkeyPatch):
     naive_time = naive_datetime.time()
     utc_time = tz_aware_utc_now.time()
     assert naive_time == utc_time
+
+
+class TestParseTimeRange:
+    """Test cases for parse_time_range function."""
+
+    def test_parse_time_range_basic(self):
+        """Test basic time range parsing."""
+        start, end = parse_time_range("2024-01-01 10:00", "2024-01-01 18:00", "UTC")
+
+        assert start is not None
+        assert end is not None
+        assert start < end
+        assert start.tzinfo == pytz.UTC
+        assert end.tzinfo == pytz.UTC
+
+    def test_parse_time_range_start_only(self):
+        """Test parsing with only start time."""
+        start, end = parse_time_range("2024-01-01 10:00", None, "UTC")
+
+        assert start is not None
+        assert end is None
+        assert start.tzinfo == pytz.UTC
+
+    def test_parse_time_range_end_only(self):
+        """Test parsing with only end time."""
+        start, end = parse_time_range(None, "2024-01-01 18:00", "UTC")
+
+        assert start is None
+        assert end is not None
+        assert end.tzinfo == pytz.UTC
+
+    def test_parse_time_range_both_none(self):
+        """Test parsing with both times None."""
+        start, end = parse_time_range(None, None, "UTC")
+
+        assert start is None
+        assert end is None
+
+    def test_parse_time_range_different_timezones(self):
+        """Test parsing with different timezones."""
+        # Test with US/Eastern timezone
+        start, end = parse_time_range("2024-01-01 10:00", "2024-01-01 18:00", "US/Eastern")
+
+        assert start is not None
+        assert end is not None
+        assert start.tzinfo == pytz.UTC
+        assert end.tzinfo == pytz.UTC
+        # Verify the times are correctly converted to UTC
+        assert start.hour == 15  # 10 AM EST = 3 PM UTC (in January)
+        assert end.hour == 23  # 6 PM EST = 11 PM UTC (in January)
+
+    def test_parse_time_range_invalid_start_format(self):
+        """Test parsing with invalid start time format."""
+        with pytest.raises(ValueError, match="time data.*does not match format"):
+            parse_time_range("invalid-date", "2024-01-01 18:00", "UTC")
+
+    def test_parse_time_range_invalid_end_format(self):
+        """Test parsing with invalid end time format."""
+        with pytest.raises(ValueError, match="time data.*does not match format"):
+            parse_time_range("2024-01-01 10:00", "invalid-date", "UTC")
+
+    def test_parse_time_range_invalid_timezone(self):
+        """Test parsing with invalid timezone."""
+        with pytest.raises(pytz.exceptions.UnknownTimeZoneError):
+            parse_time_range("2024-01-01 10:00", "2024-01-01 18:00", "Invalid/Timezone")
+
+    def test_parse_time_range_start_after_end(self):
+        """Test parsing with start time after end time."""
+        with pytest.raises(ValueError, match="start must be earlier than or equal to end"):
+            parse_time_range("2024-01-01 18:00", "2024-01-01 10:00", "UTC")
+
+    def test_parse_time_range_start_equals_end(self):
+        """Test parsing with start time equal to end time."""
+        start, end = parse_time_range("2024-01-01 10:00", "2024-01-01 10:00", "UTC")
+
+        assert start is not None
+        assert end is not None
+        assert start == end
+
+    def test_parse_time_range_dst_ambiguous_time(self):
+        """Test parsing during DST ambiguous time (fall back)."""
+        # This test simulates DST fall back where 2:30 AM occurs twice
+        with patch("pytz.timezone") as mock_timezone:
+            # Mock timezone that raises AmbiguousTimeError
+            mock_tz = mock_timezone.return_value
+
+            # Create a mock datetime object for the return value
+            mock_dt = datetime.datetime(2024, 1, 1, 10, 0, 0)
+            mock_utc_dt = mock_dt.replace(tzinfo=pytz.UTC)
+
+            # Create a proper mock for the localized datetime
+            from unittest.mock import MagicMock
+
+            mock_localized_dt = MagicMock()
+            mock_localized_dt.astimezone.return_value = mock_utc_dt
+
+            # Set up side effects: first call raises exception, second call succeeds
+            mock_tz.localize.side_effect = [
+                pytz.AmbiguousTimeError("Ambiguous time"),  # First call for start
+                mock_localized_dt,  # Second call for start (with is_dst=False)
+                pytz.AmbiguousTimeError("Ambiguous time"),  # First call for end
+                mock_localized_dt,  # Second call for end (with is_dst=False)
+            ]
+
+            start, end = parse_time_range("2024-01-01 10:00", "2024-01-01 18:00", "US/Eastern")
+
+            # Should use is_dst=False for ambiguous times
+            assert mock_tz.localize.call_count == 4  # 2 calls per time (first fails, second succeeds)
+            assert start is not None
+            assert end is not None
+
+    def test_parse_time_range_dst_nonexistent_time(self):
+        """Test parsing during DST nonexistent time (spring forward)."""
+        with patch("pytz.timezone") as mock_timezone:
+            # Mock timezone that raises NonExistentTimeError
+            mock_tz = mock_timezone.return_value
+
+            # Create a mock datetime object for the return value
+            mock_dt = datetime.datetime(2024, 1, 1, 10, 0, 0)
+            mock_utc_dt = mock_dt.replace(tzinfo=pytz.UTC)
+
+            # Create a proper mock for the localized datetime
+            from unittest.mock import MagicMock
+
+            mock_localized_dt = MagicMock()
+            mock_localized_dt.astimezone.return_value = mock_utc_dt
+
+            # Set up side effects: first call raises exception, second call succeeds
+            mock_tz.localize.side_effect = [
+                pytz.NonExistentTimeError("Non-existent time"),  # First call for start
+                mock_localized_dt,  # Second call for start (with adjusted time)
+                pytz.NonExistentTimeError("Non-existent time"),  # First call for end
+                mock_localized_dt,  # Second call for end (with adjusted time)
+            ]
+
+            start, end = parse_time_range("2024-01-01 10:00", "2024-01-01 18:00", "US/Eastern")
+
+            # Should adjust time forward by 1 hour for nonexistent times
+            assert mock_tz.localize.call_count == 4  # 2 calls per time (first fails, second succeeds)
+            assert start is not None
+            assert end is not None
+
+    def test_parse_time_range_edge_cases(self):
+        """Test edge cases for time parsing."""
+        # Test with midnight times
+        start, end = parse_time_range("2024-01-01 00:00", "2024-01-01 23:59", "UTC")
+        assert start is not None
+        assert end is not None
+        assert start.hour == 0
+        assert start.minute == 0
+        assert end.hour == 23
+        assert end.minute == 59
+
+    def test_parse_time_range_different_dates(self):
+        """Test parsing with different dates."""
+        start, end = parse_time_range("2024-01-01 10:00", "2024-01-02 10:00", "UTC")
+        assert start is not None
+        assert end is not None
+        assert start.date() != end.date()
+        assert (end - start).days == 1
+
+    def test_parse_time_range_seconds_handling(self):
+        """Test that seconds are properly set to 0."""
+        start, end = parse_time_range("2024-01-01 10:30", "2024-01-01 18:45", "UTC")
+        assert start is not None
+        assert end is not None
+        assert start.second == 0
+        assert end.second == 0
+
+    def test_parse_time_range_timezone_conversion_accuracy(self):
+        """Test accurate timezone conversion."""
+        # Test with a known timezone conversion
+        start, end = parse_time_range("2024-01-01 12:00", "2024-01-01 12:00", "Asia/Tokyo")
+
+        assert start is not None
+        assert end is not None
+        assert start.tzinfo == pytz.UTC
+        assert end.tzinfo == pytz.UTC
+        # Tokyo is UTC+9, so 12:00 JST = 03:00 UTC
+        assert start.hour == 3
+        assert end.hour == 3
+
+    def test_parse_time_range_summer_time(self):
+        """Test parsing during summer time (DST)."""
+        # Test with US/Eastern during summer (EDT = UTC-4)
+        start, end = parse_time_range("2024-07-01 12:00", "2024-07-01 12:00", "US/Eastern")
+
+        assert start is not None
+        assert end is not None
+        assert start.tzinfo == pytz.UTC
+        assert end.tzinfo == pytz.UTC
+        # 12:00 EDT = 16:00 UTC
+        assert start.hour == 16
+        assert end.hour == 16
+
+    def test_parse_time_range_winter_time(self):
+        """Test parsing during winter time (standard time)."""
+        # Test with US/Eastern during winter (EST = UTC-5)
+        start, end = parse_time_range("2024-01-01 12:00", "2024-01-01 12:00", "US/Eastern")
+
+        assert start is not None
+        assert end is not None
+        assert start.tzinfo == pytz.UTC
+        assert end.tzinfo == pytz.UTC
+        # 12:00 EST = 17:00 UTC
+        assert start.hour == 17
+        assert end.hour == 17
+
+    def test_parse_time_range_empty_strings(self):
+        """Test parsing with empty strings."""
+        # Empty strings are treated as None, so they should not raise errors
+        start, end = parse_time_range("", "2024-01-01 18:00", "UTC")
+        assert start is None
+        assert end is not None
+
+        start, end = parse_time_range("2024-01-01 10:00", "", "UTC")
+        assert start is not None
+        assert end is None
+
+    def test_parse_time_range_malformed_datetime(self):
+        """Test parsing with malformed datetime strings."""
+        with pytest.raises(ValueError, match="time data.*does not match format"):
+            parse_time_range("2024-13-01 10:00", "2024-01-01 18:00", "UTC")
+
+        with pytest.raises(ValueError, match="time data.*does not match format"):
+            parse_time_range("2024-01-01 10:00", "2024-01-32 18:00", "UTC")
+
+    def test_parse_time_range_very_long_time_range(self):
+        """Test parsing with very long time range."""
+        start, end = parse_time_range("2020-01-01 00:00", "2030-12-31 23:59", "UTC")
+
+        assert start is not None
+        assert end is not None
+        assert start < end
+        assert (end - start).days > 3000  # More than 8 years
+
+    def test_parse_time_range_negative_timezone(self):
+        """Test parsing with negative timezone offset."""
+        start, end = parse_time_range("2024-01-01 12:00", "2024-01-01 12:00", "America/New_York")
+
+        assert start is not None
+        assert end is not None
+        assert start.tzinfo == pytz.UTC
+        assert end.tzinfo == pytz.UTC

api/tests/unit_tests/services/test_document_indexing_task_proxy.py

@@ -0,0 +1,317 @@
+from unittest.mock import Mock, patch
+
+from core.entities.document_task import DocumentTask
+from core.rag.pipeline.queue import TenantIsolatedTaskQueue
+from enums.cloud_plan import CloudPlan
+from services.document_indexing_task_proxy import DocumentIndexingTaskProxy
+
+
+class DocumentIndexingTaskProxyTestDataFactory:
+    """Factory class for creating test data and mock objects for DocumentIndexingTaskProxy tests."""
+
+    @staticmethod
+    def create_mock_features(billing_enabled: bool = False, plan: CloudPlan = CloudPlan.SANDBOX) -> Mock:
+        """Create mock features with billing configuration."""
+        features = Mock()
+        features.billing = Mock()
+        features.billing.enabled = billing_enabled
+        features.billing.subscription = Mock()
+        features.billing.subscription.plan = plan
+        return features
+
+    @staticmethod
+    def create_mock_tenant_queue(has_task_key: bool = False) -> Mock:
+        """Create mock TenantIsolatedTaskQueue."""
+        queue = Mock(spec=TenantIsolatedTaskQueue)
+        queue.get_task_key.return_value = "task_key" if has_task_key else None
+        queue.push_tasks = Mock()
+        queue.set_task_waiting_time = Mock()
+        return queue
+
+    @staticmethod
+    def create_document_task_proxy(
+        tenant_id: str = "tenant-123", dataset_id: str = "dataset-456", document_ids: list[str] | None = None
+    ) -> DocumentIndexingTaskProxy:
+        """Create DocumentIndexingTaskProxy instance for testing."""
+        if document_ids is None:
+            document_ids = ["doc-1", "doc-2", "doc-3"]
+        return DocumentIndexingTaskProxy(tenant_id, dataset_id, document_ids)
+
+
+class TestDocumentIndexingTaskProxy:
+    """Test cases for DocumentIndexingTaskProxy class."""
+
+    def test_initialization(self):
+        """Test DocumentIndexingTaskProxy initialization."""
+        # Arrange
+        tenant_id = "tenant-123"
+        dataset_id = "dataset-456"
+        document_ids = ["doc-1", "doc-2", "doc-3"]
+
+        # Act
+        proxy = DocumentIndexingTaskProxy(tenant_id, dataset_id, document_ids)
+
+        # Assert
+        assert proxy._tenant_id == tenant_id
+        assert proxy._dataset_id == dataset_id
+        assert proxy._document_ids == document_ids
+        assert isinstance(proxy._tenant_isolated_task_queue, TenantIsolatedTaskQueue)
+        assert proxy._tenant_isolated_task_queue._tenant_id == tenant_id
+        assert proxy._tenant_isolated_task_queue._unique_key == "document_indexing"
+
+    @patch("services.document_indexing_task_proxy.FeatureService")
+    def test_features_property(self, mock_feature_service):
+        """Test cached_property features."""
+        # Arrange
+        mock_features = DocumentIndexingTaskProxyTestDataFactory.create_mock_features()
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+
+        # Act
+        features1 = proxy.features
+        features2 = proxy.features  # Second call should use cached property
+
+        # Assert
+        assert features1 == mock_features
+        assert features2 == mock_features
+        assert features1 is features2  # Should be the same instance due to caching
+        mock_feature_service.get_features.assert_called_once_with("tenant-123")
+
+    @patch("services.document_indexing_task_proxy.normal_document_indexing_task")
+    def test_send_to_direct_queue(self, mock_task):
+        """Test _send_to_direct_queue method."""
+        # Arrange
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+        mock_task.delay = Mock()
+
+        # Act
+        proxy._send_to_direct_queue(mock_task)
+
+        # Assert
+        mock_task.delay.assert_called_once_with(
+            tenant_id="tenant-123", dataset_id="dataset-456", document_ids=["doc-1", "doc-2", "doc-3"]
+        )
+
+    @patch("services.document_indexing_task_proxy.normal_document_indexing_task")
+    def test_send_to_tenant_queue_with_existing_task_key(self, mock_task):
+        """Test _send_to_tenant_queue when task key exists."""
+        # Arrange
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+        proxy._tenant_isolated_task_queue = DocumentIndexingTaskProxyTestDataFactory.create_mock_tenant_queue(
+            has_task_key=True
+        )
+        mock_task.delay = Mock()
+
+        # Act
+        proxy._send_to_tenant_queue(mock_task)
+
+        # Assert
+        proxy._tenant_isolated_task_queue.push_tasks.assert_called_once()
+        pushed_tasks = proxy._tenant_isolated_task_queue.push_tasks.call_args[0][0]
+        assert len(pushed_tasks) == 1
+        assert isinstance(DocumentTask(**pushed_tasks[0]), DocumentTask)
+        assert pushed_tasks[0]["tenant_id"] == "tenant-123"
+        assert pushed_tasks[0]["dataset_id"] == "dataset-456"
+        assert pushed_tasks[0]["document_ids"] == ["doc-1", "doc-2", "doc-3"]
+        mock_task.delay.assert_not_called()
+
+    @patch("services.document_indexing_task_proxy.normal_document_indexing_task")
+    def test_send_to_tenant_queue_without_task_key(self, mock_task):
+        """Test _send_to_tenant_queue when no task key exists."""
+        # Arrange
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+        proxy._tenant_isolated_task_queue = DocumentIndexingTaskProxyTestDataFactory.create_mock_tenant_queue(
+            has_task_key=False
+        )
+        mock_task.delay = Mock()
+
+        # Act
+        proxy._send_to_tenant_queue(mock_task)
+
+        # Assert
+        proxy._tenant_isolated_task_queue.set_task_waiting_time.assert_called_once()
+        mock_task.delay.assert_called_once_with(
+            tenant_id="tenant-123", dataset_id="dataset-456", document_ids=["doc-1", "doc-2", "doc-3"]
+        )
+        proxy._tenant_isolated_task_queue.push_tasks.assert_not_called()
+
+    @patch("services.document_indexing_task_proxy.normal_document_indexing_task")
+    def test_send_to_default_tenant_queue(self, mock_task):
+        """Test _send_to_default_tenant_queue method."""
+        # Arrange
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+        proxy._send_to_tenant_queue = Mock()
+
+        # Act
+        proxy._send_to_default_tenant_queue()
+
+        # Assert
+        proxy._send_to_tenant_queue.assert_called_once_with(mock_task)
+
+    @patch("services.document_indexing_task_proxy.priority_document_indexing_task")
+    def test_send_to_priority_tenant_queue(self, mock_task):
+        """Test _send_to_priority_tenant_queue method."""
+        # Arrange
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+        proxy._send_to_tenant_queue = Mock()
+
+        # Act
+        proxy._send_to_priority_tenant_queue()
+
+        # Assert
+        proxy._send_to_tenant_queue.assert_called_once_with(mock_task)
+
+    @patch("services.document_indexing_task_proxy.priority_document_indexing_task")
+    def test_send_to_priority_direct_queue(self, mock_task):
+        """Test _send_to_priority_direct_queue method."""
+        # Arrange
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+        proxy._send_to_direct_queue = Mock()
+
+        # Act
+        proxy._send_to_priority_direct_queue()
+
+        # Assert
+        proxy._send_to_direct_queue.assert_called_once_with(mock_task)
+
+    @patch("services.document_indexing_task_proxy.FeatureService")
+    def test_dispatch_with_billing_enabled_sandbox_plan(self, mock_feature_service):
+        """Test _dispatch method when billing is enabled with sandbox plan."""
+        # Arrange
+        mock_features = DocumentIndexingTaskProxyTestDataFactory.create_mock_features(
+            billing_enabled=True, plan=CloudPlan.SANDBOX
+        )
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+        proxy._send_to_default_tenant_queue = Mock()
+
+        # Act
+        proxy._dispatch()
+
+        # Assert
+        proxy._send_to_default_tenant_queue.assert_called_once()
+
+    @patch("services.document_indexing_task_proxy.FeatureService")
+    def test_dispatch_with_billing_enabled_non_sandbox_plan(self, mock_feature_service):
+        """Test _dispatch method when billing is enabled with non-sandbox plan."""
+        # Arrange
+        mock_features = DocumentIndexingTaskProxyTestDataFactory.create_mock_features(
+            billing_enabled=True, plan=CloudPlan.TEAM
+        )
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+        proxy._send_to_priority_tenant_queue = Mock()
+
+        # Act
+        proxy._dispatch()
+
+        # If billing enabled with non sandbox plan, should send to priority tenant queue
+        proxy._send_to_priority_tenant_queue.assert_called_once()
+
+    @patch("services.document_indexing_task_proxy.FeatureService")
+    def test_dispatch_with_billing_disabled(self, mock_feature_service):
+        """Test _dispatch method when billing is disabled."""
+        # Arrange
+        mock_features = DocumentIndexingTaskProxyTestDataFactory.create_mock_features(billing_enabled=False)
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+        proxy._send_to_priority_direct_queue = Mock()
+
+        # Act
+        proxy._dispatch()
+
+        # If billing disabled, for example: self-hosted or enterprise, should send to priority direct queue
+        proxy._send_to_priority_direct_queue.assert_called_once()
+
+    @patch("services.document_indexing_task_proxy.FeatureService")
+    def test_delay_method(self, mock_feature_service):
+        """Test delay method integration."""
+        # Arrange
+        mock_features = DocumentIndexingTaskProxyTestDataFactory.create_mock_features(
+            billing_enabled=True, plan=CloudPlan.SANDBOX
+        )
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+        proxy._send_to_default_tenant_queue = Mock()
+
+        # Act
+        proxy.delay()
+
+        # Assert
+        # If billing enabled with sandbox plan, should send to default tenant queue
+        proxy._send_to_default_tenant_queue.assert_called_once()
+
+    def test_document_task_dataclass(self):
+        """Test DocumentTask dataclass."""
+        # Arrange
+        tenant_id = "tenant-123"
+        dataset_id = "dataset-456"
+        document_ids = ["doc-1", "doc-2"]
+
+        # Act
+        task = DocumentTask(tenant_id=tenant_id, dataset_id=dataset_id, document_ids=document_ids)
+
+        # Assert
+        assert task.tenant_id == tenant_id
+        assert task.dataset_id == dataset_id
+        assert task.document_ids == document_ids
+
+    @patch("services.document_indexing_task_proxy.FeatureService")
+    def test_dispatch_edge_case_empty_plan(self, mock_feature_service):
+        """Test _dispatch method with empty plan string."""
+        # Arrange
+        mock_features = DocumentIndexingTaskProxyTestDataFactory.create_mock_features(billing_enabled=True, plan="")
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+        proxy._send_to_priority_tenant_queue = Mock()
+
+        # Act
+        proxy._dispatch()
+
+        # Assert
+        proxy._send_to_priority_tenant_queue.assert_called_once()
+
+    @patch("services.document_indexing_task_proxy.FeatureService")
+    def test_dispatch_edge_case_none_plan(self, mock_feature_service):
+        """Test _dispatch method with None plan."""
+        # Arrange
+        mock_features = DocumentIndexingTaskProxyTestDataFactory.create_mock_features(billing_enabled=True, plan=None)
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = DocumentIndexingTaskProxyTestDataFactory.create_document_task_proxy()
+        proxy._send_to_priority_tenant_queue = Mock()
+
+        # Act
+        proxy._dispatch()
+
+        # Assert
+        proxy._send_to_priority_tenant_queue.assert_called_once()
+
+    def test_initialization_with_empty_document_ids(self):
+        """Test initialization with empty document_ids list."""
+        # Arrange
+        tenant_id = "tenant-123"
+        dataset_id = "dataset-456"
+        document_ids = []
+
+        # Act
+        proxy = DocumentIndexingTaskProxy(tenant_id, dataset_id, document_ids)
+
+        # Assert
+        assert proxy._tenant_id == tenant_id
+        assert proxy._dataset_id == dataset_id
+        assert proxy._document_ids == document_ids
+
+    def test_initialization_with_single_document_id(self):
+        """Test initialization with single document_id."""
+        # Arrange
+        tenant_id = "tenant-123"
+        dataset_id = "dataset-456"
+        document_ids = ["doc-1"]
+
+        # Act
+        proxy = DocumentIndexingTaskProxy(tenant_id, dataset_id, document_ids)
+
+        # Assert
+        assert proxy._tenant_id == tenant_id
+        assert proxy._dataset_id == dataset_id
+        assert proxy._document_ids == document_ids

api/tests/unit_tests/services/test_rag_pipeline_task_proxy.py

@@ -0,0 +1,483 @@
+import json
+from unittest.mock import Mock, patch
+
+import pytest
+
+from core.app.entities.rag_pipeline_invoke_entities import RagPipelineInvokeEntity
+from core.rag.pipeline.queue import TenantIsolatedTaskQueue
+from enums.cloud_plan import CloudPlan
+from services.rag_pipeline.rag_pipeline_task_proxy import RagPipelineTaskProxy
+
+
+class RagPipelineTaskProxyTestDataFactory:
+    """Factory class for creating test data and mock objects for RagPipelineTaskProxy tests."""
+
+    @staticmethod
+    def create_mock_features(billing_enabled: bool = False, plan: CloudPlan = CloudPlan.SANDBOX) -> Mock:
+        """Create mock features with billing configuration."""
+        features = Mock()
+        features.billing = Mock()
+        features.billing.enabled = billing_enabled
+        features.billing.subscription = Mock()
+        features.billing.subscription.plan = plan
+        return features
+
+    @staticmethod
+    def create_mock_tenant_queue(has_task_key: bool = False) -> Mock:
+        """Create mock TenantIsolatedTaskQueue."""
+        queue = Mock(spec=TenantIsolatedTaskQueue)
+        queue.get_task_key.return_value = "task_key" if has_task_key else None
+        queue.push_tasks = Mock()
+        queue.set_task_waiting_time = Mock()
+        return queue
+
+    @staticmethod
+    def create_rag_pipeline_invoke_entity(
+        pipeline_id: str = "pipeline-123",
+        user_id: str = "user-456",
+        tenant_id: str = "tenant-789",
+        workflow_id: str = "workflow-101",
+        streaming: bool = True,
+        workflow_execution_id: str | None = None,
+        workflow_thread_pool_id: str | None = None,
+    ) -> RagPipelineInvokeEntity:
+        """Create RagPipelineInvokeEntity instance for testing."""
+        return RagPipelineInvokeEntity(
+            pipeline_id=pipeline_id,
+            application_generate_entity={"key": "value"},
+            user_id=user_id,
+            tenant_id=tenant_id,
+            workflow_id=workflow_id,
+            streaming=streaming,
+            workflow_execution_id=workflow_execution_id,
+            workflow_thread_pool_id=workflow_thread_pool_id,
+        )
+
+    @staticmethod
+    def create_rag_pipeline_task_proxy(
+        dataset_tenant_id: str = "tenant-123",
+        user_id: str = "user-456",
+        rag_pipeline_invoke_entities: list[RagPipelineInvokeEntity] | None = None,
+    ) -> RagPipelineTaskProxy:
+        """Create RagPipelineTaskProxy instance for testing."""
+        if rag_pipeline_invoke_entities is None:
+            rag_pipeline_invoke_entities = [RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_invoke_entity()]
+        return RagPipelineTaskProxy(dataset_tenant_id, user_id, rag_pipeline_invoke_entities)
+
+    @staticmethod
+    def create_mock_upload_file(file_id: str = "file-123") -> Mock:
+        """Create mock upload file."""
+        upload_file = Mock()
+        upload_file.id = file_id
+        return upload_file
+
+
+class TestRagPipelineTaskProxy:
+    """Test cases for RagPipelineTaskProxy class."""
+
+    def test_initialization(self):
+        """Test RagPipelineTaskProxy initialization."""
+        # Arrange
+        dataset_tenant_id = "tenant-123"
+        user_id = "user-456"
+        rag_pipeline_invoke_entities = [RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_invoke_entity()]
+
+        # Act
+        proxy = RagPipelineTaskProxy(dataset_tenant_id, user_id, rag_pipeline_invoke_entities)
+
+        # Assert
+        assert proxy._dataset_tenant_id == dataset_tenant_id
+        assert proxy._user_id == user_id
+        assert proxy._rag_pipeline_invoke_entities == rag_pipeline_invoke_entities
+        assert isinstance(proxy._tenant_isolated_task_queue, TenantIsolatedTaskQueue)
+        assert proxy._tenant_isolated_task_queue._tenant_id == dataset_tenant_id
+        assert proxy._tenant_isolated_task_queue._unique_key == "pipeline"
+
+    def test_initialization_with_empty_entities(self):
+        """Test initialization with empty rag_pipeline_invoke_entities."""
+        # Arrange
+        dataset_tenant_id = "tenant-123"
+        user_id = "user-456"
+        rag_pipeline_invoke_entities = []
+
+        # Act
+        proxy = RagPipelineTaskProxy(dataset_tenant_id, user_id, rag_pipeline_invoke_entities)
+
+        # Assert
+        assert proxy._dataset_tenant_id == dataset_tenant_id
+        assert proxy._user_id == user_id
+        assert proxy._rag_pipeline_invoke_entities == []
+
+    def test_initialization_with_multiple_entities(self):
+        """Test initialization with multiple rag_pipeline_invoke_entities."""
+        # Arrange
+        dataset_tenant_id = "tenant-123"
+        user_id = "user-456"
+        rag_pipeline_invoke_entities = [
+            RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_invoke_entity(pipeline_id="pipeline-1"),
+            RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_invoke_entity(pipeline_id="pipeline-2"),
+            RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_invoke_entity(pipeline_id="pipeline-3"),
+        ]
+
+        # Act
+        proxy = RagPipelineTaskProxy(dataset_tenant_id, user_id, rag_pipeline_invoke_entities)
+
+        # Assert
+        assert len(proxy._rag_pipeline_invoke_entities) == 3
+        assert proxy._rag_pipeline_invoke_entities[0].pipeline_id == "pipeline-1"
+        assert proxy._rag_pipeline_invoke_entities[1].pipeline_id == "pipeline-2"
+        assert proxy._rag_pipeline_invoke_entities[2].pipeline_id == "pipeline-3"
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FeatureService")
+    def test_features_property(self, mock_feature_service):
+        """Test cached_property features."""
+        # Arrange
+        mock_features = RagPipelineTaskProxyTestDataFactory.create_mock_features()
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+
+        # Act
+        features1 = proxy.features
+        features2 = proxy.features  # Second call should use cached property
+
+        # Assert
+        assert features1 == mock_features
+        assert features2 == mock_features
+        assert features1 is features2  # Should be the same instance due to caching
+        mock_feature_service.get_features.assert_called_once_with("tenant-123")
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FileService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.db")
+    def test_upload_invoke_entities(self, mock_db, mock_file_service_class):
+        """Test _upload_invoke_entities method."""
+        # Arrange
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        mock_file_service = Mock()
+        mock_file_service_class.return_value = mock_file_service
+        mock_upload_file = RagPipelineTaskProxyTestDataFactory.create_mock_upload_file("file-123")
+        mock_file_service.upload_text.return_value = mock_upload_file
+
+        # Act
+        result = proxy._upload_invoke_entities()
+
+        # Assert
+        assert result == "file-123"
+        mock_file_service_class.assert_called_once_with(mock_db.engine)
+
+        # Verify upload_text was called with correct parameters
+        mock_file_service.upload_text.assert_called_once()
+        call_args = mock_file_service.upload_text.call_args
+        json_text, name, user_id, tenant_id = call_args[0]
+
+        assert name == "rag_pipeline_invoke_entities.json"
+        assert user_id == "user-456"
+        assert tenant_id == "tenant-123"
+
+        # Verify JSON content
+        parsed_json = json.loads(json_text)
+        assert len(parsed_json) == 1
+        assert parsed_json[0]["pipeline_id"] == "pipeline-123"
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FileService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.db")
+    def test_upload_invoke_entities_with_multiple_entities(self, mock_db, mock_file_service_class):
+        """Test _upload_invoke_entities method with multiple entities."""
+        # Arrange
+        entities = [
+            RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_invoke_entity(pipeline_id="pipeline-1"),
+            RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_invoke_entity(pipeline_id="pipeline-2"),
+        ]
+        proxy = RagPipelineTaskProxy("tenant-123", "user-456", entities)
+        mock_file_service = Mock()
+        mock_file_service_class.return_value = mock_file_service
+        mock_upload_file = RagPipelineTaskProxyTestDataFactory.create_mock_upload_file("file-456")
+        mock_file_service.upload_text.return_value = mock_upload_file
+
+        # Act
+        result = proxy._upload_invoke_entities()
+
+        # Assert
+        assert result == "file-456"
+
+        # Verify JSON content contains both entities
+        call_args = mock_file_service.upload_text.call_args
+        json_text = call_args[0][0]
+        parsed_json = json.loads(json_text)
+        assert len(parsed_json) == 2
+        assert parsed_json[0]["pipeline_id"] == "pipeline-1"
+        assert parsed_json[1]["pipeline_id"] == "pipeline-2"
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.rag_pipeline_run_task")
+    def test_send_to_direct_queue(self, mock_task):
+        """Test _send_to_direct_queue method."""
+        # Arrange
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        proxy._tenant_isolated_task_queue = RagPipelineTaskProxyTestDataFactory.create_mock_tenant_queue()
+        upload_file_id = "file-123"
+        mock_task.delay = Mock()
+
+        # Act
+        proxy._send_to_direct_queue(upload_file_id, mock_task)
+
+        # If sent to direct queue, tenant_isolated_task_queue should not be called
+        proxy._tenant_isolated_task_queue.push_tasks.assert_not_called()
+
+        # Celery should be called directly
+        mock_task.delay.assert_called_once_with(
+            rag_pipeline_invoke_entities_file_id=upload_file_id, tenant_id="tenant-123"
+        )
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.rag_pipeline_run_task")
+    def test_send_to_tenant_queue_with_existing_task_key(self, mock_task):
+        """Test _send_to_tenant_queue when task key exists."""
+        # Arrange
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        proxy._tenant_isolated_task_queue = RagPipelineTaskProxyTestDataFactory.create_mock_tenant_queue(
+            has_task_key=True
+        )
+        upload_file_id = "file-123"
+        mock_task.delay = Mock()
+
+        # Act
+        proxy._send_to_tenant_queue(upload_file_id, mock_task)
+
+        # If task key exists, should push tasks to the queue
+        proxy._tenant_isolated_task_queue.push_tasks.assert_called_once_with([upload_file_id])
+        # Celery should not be called directly
+        mock_task.delay.assert_not_called()
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.rag_pipeline_run_task")
+    def test_send_to_tenant_queue_without_task_key(self, mock_task):
+        """Test _send_to_tenant_queue when no task key exists."""
+        # Arrange
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        proxy._tenant_isolated_task_queue = RagPipelineTaskProxyTestDataFactory.create_mock_tenant_queue(
+            has_task_key=False
+        )
+        upload_file_id = "file-123"
+        mock_task.delay = Mock()
+
+        # Act
+        proxy._send_to_tenant_queue(upload_file_id, mock_task)
+
+        # If no task key, should set task waiting time key first
+        proxy._tenant_isolated_task_queue.set_task_waiting_time.assert_called_once()
+        mock_task.delay.assert_called_once_with(
+            rag_pipeline_invoke_entities_file_id=upload_file_id, tenant_id="tenant-123"
+        )
+
+        # The first task should be sent to celery directly, so push tasks should not be called
+        proxy._tenant_isolated_task_queue.push_tasks.assert_not_called()
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.rag_pipeline_run_task")
+    def test_send_to_default_tenant_queue(self, mock_task):
+        """Test _send_to_default_tenant_queue method."""
+        # Arrange
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        proxy._send_to_tenant_queue = Mock()
+        upload_file_id = "file-123"
+
+        # Act
+        proxy._send_to_default_tenant_queue(upload_file_id)
+
+        # Assert
+        proxy._send_to_tenant_queue.assert_called_once_with(upload_file_id, mock_task)
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.priority_rag_pipeline_run_task")
+    def test_send_to_priority_tenant_queue(self, mock_task):
+        """Test _send_to_priority_tenant_queue method."""
+        # Arrange
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        proxy._send_to_tenant_queue = Mock()
+        upload_file_id = "file-123"
+
+        # Act
+        proxy._send_to_priority_tenant_queue(upload_file_id)
+
+        # Assert
+        proxy._send_to_tenant_queue.assert_called_once_with(upload_file_id, mock_task)
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.priority_rag_pipeline_run_task")
+    def test_send_to_priority_direct_queue(self, mock_task):
+        """Test _send_to_priority_direct_queue method."""
+        # Arrange
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        proxy._send_to_direct_queue = Mock()
+        upload_file_id = "file-123"
+
+        # Act
+        proxy._send_to_priority_direct_queue(upload_file_id)
+
+        # Assert
+        proxy._send_to_direct_queue.assert_called_once_with(upload_file_id, mock_task)
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FeatureService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FileService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.db")
+    def test_dispatch_with_billing_enabled_sandbox_plan(self, mock_db, mock_file_service_class, mock_feature_service):
+        """Test _dispatch method when billing is enabled with sandbox plan."""
+        # Arrange
+        mock_features = RagPipelineTaskProxyTestDataFactory.create_mock_features(
+            billing_enabled=True, plan=CloudPlan.SANDBOX
+        )
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        proxy._send_to_default_tenant_queue = Mock()
+
+        mock_file_service = Mock()
+        mock_file_service_class.return_value = mock_file_service
+        mock_upload_file = RagPipelineTaskProxyTestDataFactory.create_mock_upload_file("file-123")
+        mock_file_service.upload_text.return_value = mock_upload_file
+
+        # Act
+        proxy._dispatch()
+
+        # If billing is enabled with sandbox plan, should send to default tenant queue
+        proxy._send_to_default_tenant_queue.assert_called_once_with("file-123")
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FeatureService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FileService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.db")
+    def test_dispatch_with_billing_enabled_non_sandbox_plan(
+        self, mock_db, mock_file_service_class, mock_feature_service
+    ):
+        """Test _dispatch method when billing is enabled with non-sandbox plan."""
+        # Arrange
+        mock_features = RagPipelineTaskProxyTestDataFactory.create_mock_features(
+            billing_enabled=True, plan=CloudPlan.TEAM
+        )
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        proxy._send_to_priority_tenant_queue = Mock()
+
+        mock_file_service = Mock()
+        mock_file_service_class.return_value = mock_file_service
+        mock_upload_file = RagPipelineTaskProxyTestDataFactory.create_mock_upload_file("file-123")
+        mock_file_service.upload_text.return_value = mock_upload_file
+
+        # Act
+        proxy._dispatch()
+
+        # If billing is enabled with non-sandbox plan, should send to priority tenant queue
+        proxy._send_to_priority_tenant_queue.assert_called_once_with("file-123")
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FeatureService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FileService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.db")
+    def test_dispatch_with_billing_disabled(self, mock_db, mock_file_service_class, mock_feature_service):
+        """Test _dispatch method when billing is disabled."""
+        # Arrange
+        mock_features = RagPipelineTaskProxyTestDataFactory.create_mock_features(billing_enabled=False)
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        proxy._send_to_priority_direct_queue = Mock()
+
+        mock_file_service = Mock()
+        mock_file_service_class.return_value = mock_file_service
+        mock_upload_file = RagPipelineTaskProxyTestDataFactory.create_mock_upload_file("file-123")
+        mock_file_service.upload_text.return_value = mock_upload_file
+
+        # Act
+        proxy._dispatch()
+
+        # If billing is disabled, for example: self-hosted or enterprise, should send to priority direct queue
+        proxy._send_to_priority_direct_queue.assert_called_once_with("file-123")
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FileService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.db")
+    def test_dispatch_with_empty_upload_file_id(self, mock_db, mock_file_service_class):
+        """Test _dispatch method when upload_file_id is empty."""
+        # Arrange
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+
+        mock_file_service = Mock()
+        mock_file_service_class.return_value = mock_file_service
+        mock_upload_file = Mock()
+        mock_upload_file.id = ""  # Empty file ID
+        mock_file_service.upload_text.return_value = mock_upload_file
+
+        # Act & Assert
+        with pytest.raises(ValueError, match="upload_file_id is empty"):
+            proxy._dispatch()
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FeatureService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FileService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.db")
+    def test_dispatch_edge_case_empty_plan(self, mock_db, mock_file_service_class, mock_feature_service):
+        """Test _dispatch method with empty plan string."""
+        # Arrange
+        mock_features = RagPipelineTaskProxyTestDataFactory.create_mock_features(billing_enabled=True, plan="")
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        proxy._send_to_priority_tenant_queue = Mock()
+
+        mock_file_service = Mock()
+        mock_file_service_class.return_value = mock_file_service
+        mock_upload_file = RagPipelineTaskProxyTestDataFactory.create_mock_upload_file("file-123")
+        mock_file_service.upload_text.return_value = mock_upload_file
+
+        # Act
+        proxy._dispatch()
+
+        # Assert
+        proxy._send_to_priority_tenant_queue.assert_called_once_with("file-123")
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FeatureService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FileService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.db")
+    def test_dispatch_edge_case_none_plan(self, mock_db, mock_file_service_class, mock_feature_service):
+        """Test _dispatch method with None plan."""
+        # Arrange
+        mock_features = RagPipelineTaskProxyTestDataFactory.create_mock_features(billing_enabled=True, plan=None)
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        proxy._send_to_priority_tenant_queue = Mock()
+
+        mock_file_service = Mock()
+        mock_file_service_class.return_value = mock_file_service
+        mock_upload_file = RagPipelineTaskProxyTestDataFactory.create_mock_upload_file("file-123")
+        mock_file_service.upload_text.return_value = mock_upload_file
+
+        # Act
+        proxy._dispatch()
+
+        # Assert
+        proxy._send_to_priority_tenant_queue.assert_called_once_with("file-123")
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FeatureService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.FileService")
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.db")
+    def test_delay_method(self, mock_db, mock_file_service_class, mock_feature_service):
+        """Test delay method integration."""
+        # Arrange
+        mock_features = RagPipelineTaskProxyTestDataFactory.create_mock_features(
+            billing_enabled=True, plan=CloudPlan.SANDBOX
+        )
+        mock_feature_service.get_features.return_value = mock_features
+        proxy = RagPipelineTaskProxyTestDataFactory.create_rag_pipeline_task_proxy()
+        proxy._dispatch = Mock()
+
+        mock_file_service = Mock()
+        mock_file_service_class.return_value = mock_file_service
+        mock_upload_file = RagPipelineTaskProxyTestDataFactory.create_mock_upload_file("file-123")
+        mock_file_service.upload_text.return_value = mock_upload_file
+
+        # Act
+        proxy.delay()
+
+        # Assert
+        proxy._dispatch.assert_called_once()
+
+    @patch("services.rag_pipeline.rag_pipeline_task_proxy.logger")
+    def test_delay_method_with_empty_entities(self, mock_logger):
+        """Test delay method with empty rag_pipeline_invoke_entities."""
+        # Arrange
+        proxy = RagPipelineTaskProxy("tenant-123", "user-456", [])
+
+        # Act
+        proxy.delay()
+
+        # Assert
+        mock_logger.warning.assert_called_once_with(
+            "Received empty rag pipeline invoke entities, no tasks delivered: %s %s", "tenant-123", "user-456"
+        )

dev/start-worker

@@ -7,4 +7,4 @@ cd "$SCRIPT_DIR/.."
 
 uv --directory api run \
 	celery -A app.celery worker \
-	-P gevent -c 1 --loglevel INFO -Q dataset,mail,ops_trace,app_deletion,plugin,workflow_storage,conversation,priority_pipeline,pipeline
+	-P gevent -c 1 --loglevel INFO -Q dataset,priority_dataset,generation,mail,ops_trace,app_deletion,plugin,workflow_storage,conversation,priority_pipeline,pipeline

docker/.env.example

@@ -492,6 +492,7 @@ VECTOR_INDEX_NAME_PREFIX=Vector_index
 # The Weaviate endpoint URL. Only available when VECTOR_STORE is `weaviate`.
 WEAVIATE_ENDPOINT=http://weaviate:8080
 WEAVIATE_API_KEY=WVF5YThaHlkYwhGUSmCRgsX3tD5ngdN8pkih
+WEAVIATE_GRPC_ENDPOINT=grpc://weaviate:50051
 
 # The Qdrant endpoint URL. Only available when VECTOR_STORE is `qdrant`.
 QDRANT_URL=http://qdrant:6333
@@ -762,6 +763,12 @@ UPLOAD_FILE_SIZE_LIMIT=15
 # The maximum number of files that can be uploaded at a time, default 5.
 UPLOAD_FILE_BATCH_LIMIT=5
 
+# Comma-separated list of file extensions blocked from upload for security reasons.
+# Extensions should be lowercase without dots (e.g., exe,bat,sh,dll).
+# Empty by default to allow all file types.
+# Recommended: exe,bat,cmd,com,scr,vbs,ps1,msi,dll
+UPLOAD_FILE_EXTENSION_BLACKLIST=
+
 # ETL type, support: `dify`, `Unstructured`
 # `dify` Dify's proprietary file extraction scheme
 # `Unstructured` Unstructured.io file extraction scheme
@@ -1351,6 +1358,9 @@ SWAGGER_UI_PATH=/swagger-ui.html
 # Set to false to export dataset IDs as plain text for easier cross-environment import
 DSL_EXPORT_ENCRYPT_DATASET_ID=true
 
+# Maximum number of segments for dataset segments API (0 for unlimited)
+DATASET_MAX_SEGMENTS_PER_REQUEST=0
+
 # Celery schedule tasks configuration
 ENABLE_CLEAN_EMBEDDING_CACHE_TASK=false
 ENABLE_CLEAN_UNUSED_DATASETS_TASK=false
@@ -1360,3 +1370,6 @@ ENABLE_CLEAN_MESSAGES=false
 ENABLE_MAIL_CLEAN_DOCUMENT_NOTIFY_TASK=false
 ENABLE_DATASETS_QUEUE_MONITOR=false
 ENABLE_CHECK_UPGRADABLE_PLUGIN_TASK=true
+
+# Tenant isolated task queue configuration
+TENANT_ISOLATED_TASK_CONCURRENCY=1

docker/docker-compose-template.yaml

@@ -17,7 +17,7 @@ services:
       PLUGIN_MAX_PACKAGE_SIZE: ${PLUGIN_MAX_PACKAGE_SIZE:-52428800}
       INNER_API_KEY_FOR_PLUGIN: ${PLUGIN_DIFY_INNER_API_KEY:-QaHbTe77CtuXmsfyhR7+vRjI/+XbV1AaFy691iy+kGDv2Jvy0/eAh8Y1}
     depends_on:
-      db:
+      db_postgres:
         condition: service_healthy
       redis:
         condition: service_started
@@ -44,7 +44,7 @@ services:
       PLUGIN_MAX_PACKAGE_SIZE: ${PLUGIN_MAX_PACKAGE_SIZE:-52428800}
       INNER_API_KEY_FOR_PLUGIN: ${PLUGIN_DIFY_INNER_API_KEY:-QaHbTe77CtuXmsfyhR7+vRjI/+XbV1AaFy691iy+kGDv2Jvy0/eAh8Y1}
     depends_on:
-      db:
+      db_postgres:
         condition: service_healthy
       redis:
         condition: service_started
@@ -66,7 +66,7 @@ services:
       # Startup mode, 'worker_beat' starts the Celery beat for scheduling periodic tasks.
       MODE: beat
     depends_on:
-      db:
+      db_postgres:
         condition: service_healthy
       redis:
         condition: service_started
@@ -101,8 +101,9 @@ services:
       ENABLE_WEBSITE_JINAREADER: ${ENABLE_WEBSITE_JINAREADER:-true}
       ENABLE_WEBSITE_FIRECRAWL: ${ENABLE_WEBSITE_FIRECRAWL:-true}
       ENABLE_WEBSITE_WATERCRAWL: ${ENABLE_WEBSITE_WATERCRAWL:-true}
-  # The postgres database.
-  db:
+
+  # The PostgreSQL database.
+  db_postgres:
     image: postgres:15-alpine
     restart: always
     environment:
@@ -126,7 +127,7 @@ services:
           "CMD",
           "pg_isready",
           "-h",
-          "db",
+          "db_postgres",
           "-U",
           "${PGUSER:-postgres}",
           "-d",
@@ -236,7 +237,7 @@ services:
     volumes:
       - ./volumes/plugin_daemon:/app/storage
     depends_on:
-      db:
+      db_postgres:
         condition: service_healthy
 
   # ssrf_proxy server
@@ -353,10 +354,67 @@ services:
       AUTHORIZATION_ADMINLIST_ENABLED: ${WEAVIATE_AUTHORIZATION_ADMINLIST_ENABLED:-true}
       AUTHORIZATION_ADMINLIST_USERS: ${WEAVIATE_AUTHORIZATION_ADMINLIST_USERS:-hello@dify.ai}
 
+  # OceanBase vector database
+  oceanbase:
+    image: oceanbase/oceanbase-ce:4.3.5-lts
+    container_name: oceanbase
+    profiles:
+      - oceanbase
+    restart: always
+    volumes:
+      - ./volumes/oceanbase/data:/root/ob
+      - ./volumes/oceanbase/conf:/root/.obd/cluster
+      - ./volumes/oceanbase/init.d:/root/boot/init.d
+    environment:
+      OB_MEMORY_LIMIT: ${OCEANBASE_MEMORY_LIMIT:-6G}
+      OB_SYS_PASSWORD: ${OCEANBASE_VECTOR_PASSWORD:-difyai123456}
+      OB_TENANT_PASSWORD: ${OCEANBASE_VECTOR_PASSWORD:-difyai123456}
+      OB_CLUSTER_NAME: ${OCEANBASE_CLUSTER_NAME:-difyai}
+      OB_SERVER_IP: 127.0.0.1
+      MODE: mini
+      LANG: en_US.UTF-8
+    ports:
+      - "${OCEANBASE_VECTOR_PORT:-2881}:2881"
+    healthcheck:
+      test:
+        [
+          "CMD-SHELL",
+          'obclient -h127.0.0.1 -P2881 -uroot@test -p${OCEANBASE_VECTOR_PASSWORD:-difyai123456} -e "SELECT 1;"',
+        ]
+      interval: 10s
+      retries: 30
+      start_period: 30s
+      timeout: 10s
+
+  # seekdb vector database
+  seekdb:
+    image: oceanbase/seekdb:latest
+    container_name: seekdb
+    profiles:
+      - seekdb
+    restart: always
+    volumes:
+      - ./volumes/seekdb:/var/lib/oceanbase
+    environment:
+      ROOT_PASSWORD: ${OCEANBASE_VECTOR_PASSWORD:-difyai123456}
+      MEMORY_LIMIT: ${SEEKDB_MEMORY_LIMIT:-2G}
+      REPORTER: dify-ai-seekdb
+    ports:
+      - "${OCEANBASE_VECTOR_PORT:-2881}:2881"
+    healthcheck:
+      test:
+        [
+          "CMD-SHELL",
+          'mysql -h127.0.0.1 -P2881 -uroot -p${OCEANBASE_VECTOR_PASSWORD:-difyai123456} -e "SELECT 1;"',
+        ]
+      interval: 5s
+      retries: 60
+      timeout: 5s
+
   # Qdrant vector store.
   # (if used, you need to set VECTOR_STORE to qdrant in the api & worker service.)
   qdrant:
-    image: langgenius/qdrant:v1.7.3
+    image: langgenius/qdrant:v1.8.3
     profiles:
       - qdrant
     restart: always
@@ -488,38 +546,6 @@ services:
       CHROMA_SERVER_AUTHN_PROVIDER: ${CHROMA_SERVER_AUTHN_PROVIDER:-chromadb.auth.token_authn.TokenAuthenticationServerProvider}
       IS_PERSISTENT: ${CHROMA_IS_PERSISTENT:-TRUE}
 
-  # OceanBase vector database
-  oceanbase:
-    image: oceanbase/oceanbase-ce:4.3.5-lts
-    container_name: oceanbase
-    profiles:
-      - oceanbase
-    restart: always
-    volumes:
-      - ./volumes/oceanbase/data:/root/ob
-      - ./volumes/oceanbase/conf:/root/.obd/cluster
-      - ./volumes/oceanbase/init.d:/root/boot/init.d
-    environment:
-      OB_MEMORY_LIMIT: ${OCEANBASE_MEMORY_LIMIT:-6G}
-      OB_SYS_PASSWORD: ${OCEANBASE_VECTOR_PASSWORD:-difyai123456}
-      OB_TENANT_PASSWORD: ${OCEANBASE_VECTOR_PASSWORD:-difyai123456}
-      OB_CLUSTER_NAME: ${OCEANBASE_CLUSTER_NAME:-difyai}
-      OB_SERVER_IP: 127.0.0.1
-      MODE: mini
-      LANG: en_US.UTF-8
-    ports:
-      - "${OCEANBASE_VECTOR_PORT:-2881}:2881"
-    healthcheck:
-      test:
-        [
-          "CMD-SHELL",
-          'obclient -h127.0.0.1 -P2881 -uroot@test -p$${OB_TENANT_PASSWORD} -e "SELECT 1;"',
-        ]
-      interval: 10s
-      retries: 30
-      start_period: 30s
-      timeout: 10s
-
   # Oracle vector database
   oracle:
     image: container-registry.oracle.com/database/free:latest