Merge branch 'main' into fix/chore-fix

api/.env.example

@@ -42,6 +42,11 @@ REDIS_SENTINEL_USERNAME=
 REDIS_SENTINEL_PASSWORD=
 REDIS_SENTINEL_SOCKET_TIMEOUT=0.1
 
+# redis Cluster configuration.
+REDIS_USE_CLUSTERS=false
+REDIS_CLUSTERS=
+REDIS_CLUSTERS_PASSWORD=
+
 # PostgreSQL database configuration
 DB_USERNAME=postgres
 DB_PASSWORD=difyai123456
@@ -234,6 +239,10 @@ ANALYTICDB_ACCOUNT=testaccount
 ANALYTICDB_PASSWORD=testpassword
 ANALYTICDB_NAMESPACE=dify
 ANALYTICDB_NAMESPACE_PASSWORD=difypassword
+ANALYTICDB_HOST=gp-test.aliyuncs.com
+ANALYTICDB_PORT=5432
+ANALYTICDB_MIN_CONNECTION=1
+ANALYTICDB_MAX_CONNECTION=5
 
 # OpenSearch configuration
 OPENSEARCH_HOST=127.0.0.1

api/README.md

@@ -18,12 +18,17 @@
    ```
 
 2. Copy `.env.example` to `.env`
+
+   ```cli
+   cp .env.example .env 
+   ```
 3. Generate a `SECRET_KEY` in the `.env` file.
 
+   bash for Linux
    ```bash for Linux
    sed -i "/^SECRET_KEY=/c\SECRET_KEY=$(openssl rand -base64 42)" .env
    ```
-
+   bash for Mac
    ```bash for Mac
    secret_key=$(openssl rand -base64 42)
    sed -i '' "/^SECRET_KEY=/c\\
@@ -41,14 +46,6 @@
    poetry install
    ```
 
-   In case of contributors missing to update dependencies for `pyproject.toml`, you can perform the following shell instead.
-
-   ```bash
-   poetry shell                                               # activate current environment
-   poetry add $(cat requirements.txt)           # install dependencies of production and update pyproject.toml
-   poetry add $(cat requirements-dev.txt) --group dev    # install dependencies of development and update pyproject.toml
-   ```
-
 6. Run migrate
 
    Before the first launch, migrate the database to the latest version.

api/commands.py

@@ -590,7 +590,7 @@ def upgrade_db():
             click.echo(click.style("Database migration successful!", fg="green"))
 
         except Exception as e:
-            logging.exception(f"Database migration failed: {e}")
+            logging.exception("Failed to execute database migration")
         finally:
             lock.release()
     else:
@@ -634,7 +634,7 @@ where sites.id is null limit 1000"""
                 except Exception as e:
                     failed_app_ids.append(app_id)
                     click.echo(click.style("Failed to fix missing site for app {}".format(app_id), fg="red"))
-                    logging.exception(f"Fix app related site missing issue failed, error: {e}")
+                    logging.exception(f"Failed to fix app related site missing issue, app_id: {app_id}")
                     continue
 
             if not processed_count:

api/configs/feature/__init__.py

@@ -674,6 +674,11 @@ class DataSetConfig(BaseSettings):
         default=False,
     )
 
+    PLAN_SANDBOX_CLEAN_MESSAGE_DAY_SETTING: PositiveInt = Field(
+        description="Interval in days for message cleanup operations - plan: sandbox",
+        default=30,
+    )
+
 
 class WorkspaceConfig(BaseSettings):
     """

api/configs/middleware/cache/redis_config.py

@@ -68,3 +68,18 @@ class RedisConfig(BaseSettings):
         description="Socket timeout in seconds for Redis Sentinel connections",
         default=0.1,
     )
+
+    REDIS_USE_CLUSTERS: bool = Field(
+        description="Enable Redis Clusters mode for high availability",
+        default=False,
+    )
+
+    REDIS_CLUSTERS: Optional[str] = Field(
+        description="Comma-separated list of Redis Clusters nodes (host:port)",
+        default=None,
+    )
+
+    REDIS_CLUSTERS_PASSWORD: Optional[str] = Field(
+        description="Password for Redis Clusters authentication (if required)",
+        default=None,
+    )

api/configs/middleware/vdb/analyticdb_config.py

@@ -1,6 +1,6 @@
 from typing import Optional
 
-from pydantic import BaseModel, Field
+from pydantic import BaseModel, Field, PositiveInt
 
 
 class AnalyticdbConfig(BaseModel):
@@ -40,3 +40,11 @@ class AnalyticdbConfig(BaseModel):
         description="The password for accessing the specified namespace within the AnalyticDB instance"
         " (if namespace feature is enabled).",
     )
+    ANALYTICDB_HOST: Optional[str] = Field(
+        default=None, description="The host of the AnalyticDB instance you want to connect to."
+    )
+    ANALYTICDB_PORT: PositiveInt = Field(
+        default=5432, description="The port of the AnalyticDB instance you want to connect to."
+    )
+    ANALYTICDB_MIN_CONNECTION: PositiveInt = Field(default=1, description="Min connection of the AnalyticDB database.")
+    ANALYTICDB_MAX_CONNECTION: PositiveInt = Field(default=5, description="Max connection of the AnalyticDB database.")

api/configs/packaging/__init__.py

@@ -9,7 +9,7 @@ class PackagingInfo(BaseSettings):
 
     CURRENT_VERSION: str = Field(
         description="Dify version",
-        default="0.11.1",
+        default="0.11.2",
     )
 
     COMMIT_SHA: str = Field(

api/controllers/console/app/app.py

@@ -9,6 +9,7 @@ from controllers.console.app.wraps import get_app_model
 from controllers.console.wraps import (
     account_initialization_required,
     cloud_edition_billing_resource_check,
+    enterprise_license_required,
     setup_required,
 )
 from core.model_runtime.utils.encoders import jsonable_encoder
@@ -29,6 +30,7 @@ class AppListApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
+    @enterprise_license_required
     def get(self):
         """Get app list"""
 
@@ -188,6 +190,7 @@ class AppApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
+    @enterprise_license_required
     @get_app_model
     @marshal_with(app_detail_fields_with_site)
     def get(self, app_model):

api/controllers/console/app/audio.py

@@ -70,7 +70,7 @@ class ChatMessageAudioApi(Resource):
         except ValueError as e:
             raise e
         except Exception as e:
-            logging.exception(f"internal server error, {str(e)}.")
+            logging.exception("Failed to handle post request to ChatMessageAudioApi")
             raise InternalServerError()
 
 
@@ -128,7 +128,7 @@ class ChatMessageTextApi(Resource):
         except ValueError as e:
             raise e
         except Exception as e:
-            logging.exception(f"internal server error, {str(e)}.")
+            logging.exception("Failed to handle post request to ChatMessageTextApi")
             raise InternalServerError()
 
 
@@ -170,7 +170,7 @@ class TextModesApi(Resource):
         except ValueError as e:
             raise e
         except Exception as e:
-            logging.exception(f"internal server error, {str(e)}.")
+            logging.exception("Failed to handle get request to TextModesApi")
             raise InternalServerError()
 
 

api/controllers/console/auth/forgot_password.py

@@ -14,7 +14,7 @@ from controllers.console.auth.error import (
     InvalidTokenError,
     PasswordMismatchError,
 )
-from controllers.console.error import EmailSendIpLimitError, NotAllowedRegister
+from controllers.console.error import AccountNotFound, EmailSendIpLimitError
 from controllers.console.wraps import setup_required
 from events.tenant_event import tenant_was_created
 from extensions.ext_database import db
@@ -51,7 +51,7 @@ class ForgotPasswordSendEmailApi(Resource):
                 token = AccountService.send_reset_password_email(email=args["email"], language=language)
                 return {"result": "fail", "data": token, "code": "account_not_found"}
             else:
-                raise NotAllowedRegister()
+                raise AccountNotFound()
         else:
             token = AccountService.send_reset_password_email(account=account, email=args["email"], language=language)
 

api/controllers/console/auth/login.py

@@ -16,9 +16,9 @@ from controllers.console.auth.error import (
 )
 from controllers.console.error import (
     AccountBannedError,
+    AccountNotFound,
     EmailSendIpLimitError,
     NotAllowedCreateWorkspace,
-    NotAllowedRegister,
 )
 from controllers.console.wraps import setup_required
 from events.tenant_event import tenant_was_created
@@ -76,7 +76,7 @@ class LoginApi(Resource):
                 token = AccountService.send_reset_password_email(email=args["email"], language=language)
                 return {"result": "fail", "data": token, "code": "account_not_found"}
             else:
-                raise NotAllowedRegister()
+                raise AccountNotFound()
         # SELF_HOSTED only have one workspace
         tenants = TenantService.get_join_tenants(account)
         if len(tenants) == 0:
@@ -119,7 +119,7 @@ class ResetPasswordSendEmailApi(Resource):
             if FeatureService.get_system_features().is_allow_register:
                 token = AccountService.send_reset_password_email(email=args["email"], language=language)
             else:
-                raise NotAllowedRegister()
+                raise AccountNotFound()
         else:
             token = AccountService.send_reset_password_email(account=account, language=language)
 
@@ -148,7 +148,7 @@ class EmailCodeLoginSendEmailApi(Resource):
             if FeatureService.get_system_features().is_allow_register:
                 token = AccountService.send_email_code_login_email(email=args["email"], language=language)
             else:
-                raise NotAllowedRegister()
+                raise AccountNotFound()
         else:
             token = AccountService.send_email_code_login_email(account=account, language=language)
 

api/controllers/console/datasets/datasets.py

@@ -10,7 +10,7 @@ from controllers.console import api
 from controllers.console.apikey import api_key_fields, api_key_list
 from controllers.console.app.error import ProviderNotInitializeError
 from controllers.console.datasets.error import DatasetInUseError, DatasetNameDuplicateError, IndexingEstimateError
-from controllers.console.wraps import account_initialization_required, setup_required
+from controllers.console.wraps import account_initialization_required, enterprise_license_required, setup_required
 from core.errors.error import LLMBadRequestError, ProviderTokenNotInitError
 from core.indexing_runner import IndexingRunner
 from core.model_runtime.entities.model_entities import ModelType
@@ -44,6 +44,7 @@ class DatasetListApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
+    @enterprise_license_required
     def get(self):
         page = request.args.get("page", default=1, type=int)
         limit = request.args.get("limit", default=20, type=int)

api/controllers/console/datasets/datasets_document.py

@@ -959,7 +959,7 @@ class DocumentRetryApi(DocumentResource):
                     raise DocumentAlreadyFinishedError()
                 retry_documents.append(document)
             except Exception as e:
-                logging.exception(f"Document {document_id} retry failed: {str(e)}")
+                logging.exception(f"Failed to retry document, document id: {document_id}")
                 continue
         # retry document
         DocumentService.retry_document(dataset_id, retry_documents)

api/controllers/console/error.py

@@ -52,8 +52,8 @@ class AccountBannedError(BaseHTTPException):
     code = 400
 
 
-class NotAllowedRegister(BaseHTTPException):
-    error_code = "unauthorized"
+class AccountNotFound(BaseHTTPException):
+    error_code = "account_not_found"
     description = "Account not found."
     code = 400
 
@@ -86,3 +86,9 @@ class NoFileUploadedError(BaseHTTPException):
     error_code = "no_file_uploaded"
     description = "Please upload your file."
     code = 400
+
+
+class UnauthorizedAndForceLogout(BaseHTTPException):
+    error_code = "unauthorized_and_force_logout"
+    description = "Unauthorized and force logout."
+    code = 401

api/controllers/console/remote_files.py

@@ -45,7 +45,7 @@ class RemoteFileUploadApi(Resource):
 
         resp = ssrf_proxy.head(url=url)
         if resp.status_code != httpx.codes.OK:
-            resp = ssrf_proxy.get(url=url, timeout=3)
+            resp = ssrf_proxy.get(url=url, timeout=3, follow_redirects=True)
         resp.raise_for_status()
 
         file_info = helpers.guess_file_info_from_response(resp)

api/controllers/console/workspace/account.py

@@ -14,7 +14,7 @@ from controllers.console.workspace.error import (
     InvalidInvitationCodeError,
     RepeatPasswordNotMatchError,
 )
-from controllers.console.wraps import account_initialization_required, setup_required
+from controllers.console.wraps import account_initialization_required, enterprise_license_required, setup_required
 from extensions.ext_database import db
 from fields.member_fields import account_fields
 from libs.helper import TimestampField, timezone
@@ -79,6 +79,7 @@ class AccountProfileApi(Resource):
     @login_required
     @account_initialization_required
     @marshal_with(account_fields)
+    @enterprise_license_required
     def get(self):
         return current_user
 

api/controllers/console/workspace/members.py

@@ -1,3 +1,5 @@
+from urllib import parse
+
 from flask_login import current_user
 from flask_restful import Resource, abort, marshal_with, reqparse
 
@@ -57,11 +59,12 @@ class MemberInviteEmailApi(Resource):
                 token = RegisterService.invite_new_member(
                     inviter.current_tenant, invitee_email, interface_language, role=invitee_role, inviter=inviter
                 )
+                encoded_invitee_email = parse.quote(invitee_email)
                 invitation_results.append(
                     {
                         "status": "success",
                         "email": invitee_email,
-                        "url": f"{console_web_url}/activate?email={invitee_email}&token={token}",
+                        "url": f"{console_web_url}/activate?email={encoded_invitee_email}&token={token}",
                     }
                 )
             except AccountAlreadyInTenantError:

api/controllers/console/workspace/models.py

@@ -72,7 +72,10 @@ class DefaultModelApi(Resource):
                     model=model_setting["model"],
                 )
             except Exception as ex:
-                logging.exception(f"{model_setting['model_type']} save error: {ex}")
+                logging.exception(
+                    f"Failed to update default model, model type: {model_setting['model_type']},"
+                    f" model:{model_setting.get('model')}"
+                )
                 raise ex
 
         return {"result": "success"}
@@ -156,7 +159,10 @@ class ModelProviderModelApi(Resource):
                         credentials=args["credentials"],
                     )
                 except CredentialsValidateFailedError as ex:
-                    logging.exception(f"save model credentials error: {ex}")
+                    logging.exception(
+                        f"Failed to save model credentials, tenant_id: {tenant_id},"
+                        f" model: {args.get('model')}, model_type: {args.get('model_type')}"
+                    )
                     raise ValueError(str(ex))
 
         return {"result": "success"}, 200

api/controllers/console/workspace/tool_providers.py

@@ -7,7 +7,7 @@ from werkzeug.exceptions import Forbidden
 
 from configs import dify_config
 from controllers.console import api
-from controllers.console.wraps import account_initialization_required, setup_required
+from controllers.console.wraps import account_initialization_required, enterprise_license_required, setup_required
 from core.model_runtime.utils.encoders import jsonable_encoder
 from libs.helper import alphanumeric, uuid_value
 from libs.login import login_required
@@ -608,6 +608,7 @@ class ToolLabelsApi(Resource):
     @setup_required
     @login_required
     @account_initialization_required
+    @enterprise_license_required
     def get(self):
         return jsonable_encoder(ToolLabelsService.list_tool_labels())
 

api/controllers/console/wraps.py

@@ -9,10 +9,10 @@ from configs import dify_config
 from controllers.console.workspace.error import AccountNotInitializedError
 from extensions.ext_database import db
 from models.model import DifySetup
-from services.feature_service import FeatureService
+from services.feature_service import FeatureService, LicenseStatus
 from services.operation_service import OperationService
 
-from .error import NotInitValidateError, NotSetupError
+from .error import NotInitValidateError, NotSetupError, UnauthorizedAndForceLogout
 
 
 def account_initialization_required(view):
@@ -147,3 +147,15 @@ def setup_required(view):
         return view(*args, **kwargs)
 
     return decorated
+
+
+def enterprise_license_required(view):
+    @wraps(view)
+    def decorated(*args, **kwargs):
+        settings = FeatureService.get_system_features()
+        if settings.license.status in [LicenseStatus.INACTIVE, LicenseStatus.EXPIRED, LicenseStatus.LOST]:
+            raise UnauthorizedAndForceLogout("Your license is invalid. Please contact your administrator.")
+
+        return view(*args, **kwargs)
+
+    return decorated

api/controllers/web/audio.py

@@ -59,7 +59,7 @@ class AudioApi(WebApiResource):
         except ValueError as e:
             raise e
         except Exception as e:
-            logging.exception(f"internal server error: {str(e)}")
+            logging.exception("Failed to handle post request to AudioApi")
             raise InternalServerError()
 
 
@@ -117,7 +117,7 @@ class TextApi(WebApiResource):
         except ValueError as e:
             raise e
         except Exception as e:
-            logging.exception(f"internal server error: {str(e)}")
+            logging.exception("Failed to handle post request to TextApi")
             raise InternalServerError()
 
 

api/core/agent/base_agent_runner.py

@@ -106,16 +106,9 @@ class BaseAgentRunner(AppRunner):
         # check if model supports stream tool call
         llm_model = cast(LargeLanguageModel, model_instance.model_type_instance)
         model_schema = llm_model.get_model_schema(model_instance.model, model_instance.credentials)
-        if model_schema and ModelFeature.STREAM_TOOL_CALL in (model_schema.features or []):
-            self.stream_tool_call = True
-        else:
-            self.stream_tool_call = False
-
-        # check if model supports vision
-        if model_schema and ModelFeature.VISION in (model_schema.features or []):
-            self.files = application_generate_entity.files
-        else:
-            self.files = []
+        features = model_schema.features if model_schema and model_schema.features else []
+        self.stream_tool_call = ModelFeature.STREAM_TOOL_CALL in features
+        self.files = application_generate_entity.files if ModelFeature.VISION in features else []
         self.query = None
         self._current_thoughts: list[PromptMessage] = []
 
@@ -243,7 +236,7 @@ class BaseAgentRunner(AppRunner):
         update prompt message tool
         """
         # try to get tool runtime parameters
-        tool_runtime_parameters = tool.get_runtime_parameters() or []
+        tool_runtime_parameters = tool.get_runtime_parameters()
 
         for parameter in tool_runtime_parameters:
             if parameter.form != ToolParameter.ToolParameterForm.LLM:

api/core/app/app_config/features/file_upload/manager.py

@@ -16,9 +16,7 @@ class FileUploadConfigManager:
         file_upload_dict = config.get("file_upload")
         if file_upload_dict:
             if file_upload_dict.get("enabled"):
-                transform_methods = file_upload_dict.get("allowed_file_upload_methods") or file_upload_dict.get(
-                    "allowed_upload_methods", []
-                )
+                transform_methods = file_upload_dict.get("allowed_file_upload_methods", [])
                 data = {
                     "image_config": {
                         "number_limits": file_upload_dict["number_limits"],

api/core/app/apps/advanced_chat/app_generator.py

@@ -373,5 +373,5 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
             if e.args[0] == "I/O operation on closed file.":  # ignore this error
                 raise GenerateTaskStoppedError()
             else:
-                logger.exception(e)
+                logger.exception(f"Failed to process generate task pipeline, conversation_id: {conversation.id}")
                 raise e

api/core/app/apps/advanced_chat/generate_task_pipeline.py

@@ -242,7 +242,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
                     start_listener_time = time.time()
                     yield MessageAudioStreamResponse(audio=audio_trunk.audio, task_id=task_id)
             except Exception as e:
-                logger.exception(e)
+                logger.exception(f"Failed to listen audio message, task_id: {task_id}")
                 break
         if tts_publisher:
             yield MessageAudioEndStreamResponse(audio="", task_id=task_id)

api/core/app/apps/base_app_generator.py

@@ -34,8 +34,8 @@ class BaseAppGenerator:
                 tenant_id=app_config.tenant_id,
                 config=FileUploadConfig(
                     allowed_file_types=entity_dictionary[k].allowed_file_types,
-                    allowed_extensions=entity_dictionary[k].allowed_file_extensions,
-                    allowed_upload_methods=entity_dictionary[k].allowed_file_upload_methods,
+                    allowed_file_extensions=entity_dictionary[k].allowed_file_extensions,
+                    allowed_file_upload_methods=entity_dictionary[k].allowed_file_upload_methods,
                 ),
             )
             for k, v in user_inputs.items()
@@ -48,8 +48,8 @@ class BaseAppGenerator:
                 tenant_id=app_config.tenant_id,
                 config=FileUploadConfig(
                     allowed_file_types=entity_dictionary[k].allowed_file_types,
-                    allowed_extensions=entity_dictionary[k].allowed_file_extensions,
-                    allowed_upload_methods=entity_dictionary[k].allowed_file_upload_methods,
+                    allowed_file_extensions=entity_dictionary[k].allowed_file_extensions,
+                    allowed_file_upload_methods=entity_dictionary[k].allowed_file_upload_methods,
                 ),
             )
             for k, v in user_inputs.items()
@@ -92,6 +92,9 @@ class BaseAppGenerator:
             )
 
         if variable_entity.type == VariableEntityType.NUMBER and isinstance(value, str):
+            # handle empty string case
+            if not value.strip():
+                return None
             # may raise ValueError if user_input_value is not a valid number
             try:
                 if "." in value:

api/core/app/apps/message_based_app_generator.py

@@ -80,7 +80,7 @@ class MessageBasedAppGenerator(BaseAppGenerator):
             if e.args[0] == "I/O operation on closed file.":  # ignore this error
                 raise GenerateTaskStoppedError()
             else:
-                logger.exception(e)
+                logger.exception(f"Failed to handle response, conversation_id: {conversation.id}")
                 raise e
 
     def _get_conversation_by_user(

api/core/app/apps/workflow/app_generator.py

@@ -310,5 +310,7 @@ class WorkflowAppGenerator(BaseAppGenerator):
             if e.args[0] == "I/O operation on closed file.":  # ignore this error
                 raise GenerateTaskStoppedError()
             else:
-                logger.exception(e)
+                logger.exception(
+                    f"Fails to process generate task pipeline, task_id: {application_generate_entity.task_id}"
+                )
                 raise e

api/core/app/apps/workflow/generate_task_pipeline.py

@@ -216,7 +216,7 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
                 else:
                     yield MessageAudioStreamResponse(audio=audio_trunk.audio, task_id=task_id)
             except Exception as e:
-                logger.exception(e)
+                logger.exception(f"Fails to get audio trunk, task_id: {task_id}")
                 break
         if tts_publisher:
             yield MessageAudioEndStreamResponse(audio="", task_id=task_id)

api/core/app/task_pipeline/message_cycle_manage.py

@@ -86,7 +86,7 @@ class MessageCycleManage:
                     conversation.name = name
                 except Exception as e:
                     if dify_config.DEBUG:
-                        logging.exception(f"generate conversation name failed: {e}")
+                        logging.exception(f"generate conversation name failed, conversation_id: {conversation_id}")
                     pass
 
                 db.session.merge(conversation)

api/core/app/task_pipeline/workflow_cycle_manage.py

@@ -381,7 +381,7 @@ class WorkflowCycleManage:
                 id=workflow_run.id,
                 workflow_id=workflow_run.workflow_id,
                 sequence_number=workflow_run.sequence_number,
-                inputs=workflow_run.inputs_dict or {},
+                inputs=workflow_run.inputs_dict,
                 created_at=int(workflow_run.created_at.timestamp()),
             ),
         )
@@ -428,7 +428,7 @@ class WorkflowCycleManage:
                 created_by=created_by,
                 created_at=int(workflow_run.created_at.timestamp()),
                 finished_at=int(workflow_run.finished_at.timestamp()),
-                files=self._fetch_files_from_node_outputs(workflow_run.outputs_dict or {}),
+                files=self._fetch_files_from_node_outputs(workflow_run.outputs_dict),
             ),
         )
 

api/core/file/models.py

@@ -28,8 +28,8 @@ class FileUploadConfig(BaseModel):
 
     image_config: Optional[ImageConfig] = None
     allowed_file_types: Sequence[FileType] = Field(default_factory=list)
-    allowed_extensions: Sequence[str] = Field(default_factory=list)
-    allowed_upload_methods: Sequence[FileTransferMethod] = Field(default_factory=list)
+    allowed_file_extensions: Sequence[str] = Field(default_factory=list)
+    allowed_file_upload_methods: Sequence[FileTransferMethod] = Field(default_factory=list)
     number_limits: int = 0
 
 

api/core/helper/moderation.py

@@ -55,7 +55,7 @@ def check_moderation(tenant_id: str, model_config: ModelConfigWithCredentialsEnt
                 if moderation_result is True:
                     return True
             except Exception as ex:
-                logger.exception(ex)
+                logger.exception(f"Fails to check moderation, provider_name: {provider_name}")
                 raise InvokeBadRequestError("Rate limit exceeded, please try again later.")
 
     return False

api/core/helper/module_import_helper.py

@@ -29,7 +29,7 @@ def import_module_from_source(*, module_name: str, py_file_path: AnyStr, use_laz
         spec.loader.exec_module(module)
         return module
     except Exception as e:
-        logging.exception(f"Failed to load module {module_name} from {py_file_path}: {str(e)}")
+        logging.exception(f"Failed to load module {module_name} from script file '{py_file_path}'")
         raise e
 
 

api/core/helper/ssrf_proxy.py

@@ -39,6 +39,7 @@ def make_request(method, url, max_retries=SSRF_DEFAULT_MAX_RETRIES, **kwargs):
         )
 
     retries = 0
+    stream = kwargs.pop("stream", False)
     while retries <= max_retries:
         try:
             if dify_config.SSRF_PROXY_ALL_URL:
@@ -52,6 +53,8 @@ def make_request(method, url, max_retries=SSRF_DEFAULT_MAX_RETRIES, **kwargs):
                     response = client.request(method=method, url=url, **kwargs)
 
             if response.status_code not in STATUS_FORCELIST:
+                if stream:
+                    return response.iter_bytes()
                 return response
             else:
                 logging.warning(f"Received status code {response.status_code} for URL {url} which is in the force list")

api/core/indexing_runner.py

@@ -29,6 +29,8 @@ from core.rag.splitter.fixed_text_splitter import (
     FixedRecursiveCharacterTextSplitter,
 )
 from core.rag.splitter.text_splitter import TextSplitter
+from core.tools.utils.text_processing_utils import remove_leading_symbols
+from core.tools.utils.web_reader_tool import get_image_upload_file_ids
 from extensions.ext_database import db
 from extensions.ext_redis import redis_client
 from extensions.ext_storage import storage
@@ -278,6 +280,19 @@ class IndexingRunner:
                 if len(preview_texts) < 5:
                     preview_texts.append(document.page_content)
 
+                # delete image files and related db records
+                image_upload_file_ids = get_image_upload_file_ids(document.page_content)
+                for upload_file_id in image_upload_file_ids:
+                    image_file = db.session.query(UploadFile).filter(UploadFile.id == upload_file_id).first()
+                    try:
+                        storage.delete(image_file.key)
+                    except Exception:
+                        logging.exception(
+                            "Delete image_files failed while indexing_estimate, \
+                                          image_upload_file_is: {}".format(upload_file_id)
+                        )
+                    db.session.delete(image_file)
+
         if doc_form and doc_form == "qa_model":
             if len(preview_texts) > 0:
                 # qa model document
@@ -500,11 +515,7 @@ class IndexingRunner:
                     document_node.metadata["doc_hash"] = hash
                     # delete Splitter character
                     page_content = document_node.page_content
-                    if page_content.startswith(".") or page_content.startswith("。"):
-                        page_content = page_content[1:]
-                    else:
-                        page_content = page_content
-                    document_node.page_content = page_content
+                    document_node.page_content = remove_leading_symbols(page_content)
 
                     if document_node.page_content:
                         split_documents.append(document_node)
@@ -554,7 +565,7 @@ class IndexingRunner:
                     qa_documents.append(qa_document)
                 format_documents.extend(qa_documents)
             except Exception as e:
-                logging.exception(e)
+                logging.exception("Failed to format qa document")
 
             all_qa_documents.extend(format_documents)
 

api/core/llm_generator/llm_generator.py

@@ -102,7 +102,7 @@ class LLMGenerator:
         except InvokeError:
             questions = []
         except Exception as e:
-            logging.exception(e)
+            logging.exception("Failed to generate suggested questions after answer")
             questions = []
 
         return questions
@@ -148,7 +148,7 @@ class LLMGenerator:
                 error = str(e)
                 error_step = "generate rule config"
             except Exception as e:
-                logging.exception(e)
+                logging.exception(f"Failed to generate rule config, model: {model_config.get('name')}")
                 rule_config["error"] = str(e)
 
             rule_config["error"] = f"Failed to {error_step}. Error: {error}" if error else ""
@@ -234,7 +234,7 @@ class LLMGenerator:
                 error_step = "generate conversation opener"
 
         except Exception as e:
-            logging.exception(e)
+            logging.exception(f"Failed to generate rule config, model: {model_config.get('name')}")
             rule_config["error"] = str(e)
 
         rule_config["error"] = f"Failed to {error_step}. Error: {error}" if error else ""
@@ -286,7 +286,9 @@ class LLMGenerator:
             error = str(e)
             return {"code": "", "language": code_language, "error": f"Failed to generate code. Error: {error}"}
         except Exception as e:
-            logging.exception(e)
+            logging.exception(
+                f"Failed to invoke LLM model, model: {model_config.get('name')}, language: {code_language}"
+            )
             return {"code": "", "language": code_language, "error": f"An unexpected error occurred: {str(e)}"}
 
     @classmethod

api/core/model_runtime/model_providers/gitee_ai/_common.py

@@ -1,47 +0,0 @@
-from dashscope.common.error import (
-    AuthenticationError,
-    InvalidParameter,
-    RequestFailure,
-    ServiceUnavailableError,
-    UnsupportedHTTPMethod,
-    UnsupportedModel,
-)
-
-from core.model_runtime.errors.invoke import (
-    InvokeAuthorizationError,
-    InvokeBadRequestError,
-    InvokeConnectionError,
-    InvokeError,
-    InvokeRateLimitError,
-    InvokeServerUnavailableError,
-)
-
-
-class _CommonGiteeAI:
-    @property
-    def _invoke_error_mapping(self) -> dict[type[InvokeError], list[type[Exception]]]:
-        """
-        Map model invoke error to unified error
-        The key is the error type thrown to the caller
-        The value is the error type thrown by the model,
-        which needs to be converted into a unified error type for the caller.
-
-        :return: Invoke error mapping
-        """
-        return {
-            InvokeConnectionError: [
-                RequestFailure,
-            ],
-            InvokeServerUnavailableError: [
-                ServiceUnavailableError,
-            ],
-            InvokeRateLimitError: [],
-            InvokeAuthorizationError: [
-                AuthenticationError,
-            ],
-            InvokeBadRequestError: [
-                InvalidParameter,
-                UnsupportedModel,
-                UnsupportedHTTPMethod,
-            ],
-        }

api/core/model_runtime/model_providers/gitee_ai/gitee_ai.py

@@ -1,36 +0,0 @@
-import logging
-
-import requests
-
-from core.model_runtime.errors.validate import CredentialsValidateFailedError
-from core.model_runtime.model_providers.__base.model_provider import ModelProvider
-
-logger = logging.getLogger(__name__)
-
-
-class GiteeAIProvider(ModelProvider):
-    def validate_provider_credentials(self, credentials: dict) -> None:
-        """
-        Validate provider credentials
-        if validate failed, raise exception
-
-        :param credentials: provider credentials, credentials form defined in `provider_credential_schema`.
-        """
-        try:
-            api_key = credentials.get("api_key")
-            if not api_key:
-                raise CredentialsValidateFailedError("Credentials validation failed: api_key not given")
-
-            # send a get request to validate the credentials
-            headers = {"Authorization": f"Bearer {api_key}"}
-            response = requests.get("https://ai.gitee.com/api/base/account/me", headers=headers, timeout=(10, 300))
-
-            if response.status_code != 200:
-                raise CredentialsValidateFailedError(
-                    f"Credentials validation failed with status code {response.status_code}"
-                )
-        except CredentialsValidateFailedError as ex:
-            raise ex
-        except Exception as ex:
-            logger.exception(f"{self.get_provider_schema().provider} credentials validate failed")
-            raise ex

api/core/model_runtime/model_providers/gitee_ai/gitee_ai.yaml

@@ -1,35 +0,0 @@
-provider: gitee_ai
-label:
-  en_US: Gitee AI
-  zh_Hans: Gitee AI
-description:
-  en_US: 快速体验大模型，领先探索 AI 开源世界
-  zh_Hans: 快速体验大模型，领先探索 AI 开源世界
-icon_small:
-  en_US: Gitee-AI-Logo.svg
-icon_large:
-  en_US: Gitee-AI-Logo-full.svg
-help:
-  title:
-    en_US: Get your token from Gitee AI
-    zh_Hans: 从 Gitee AI 获取 token
-  url:
-    en_US: https://ai.gitee.com/dashboard/settings/tokens
-supported_model_types:
-  - llm
-  - text-embedding
-  - rerank
-  - speech2text
-  - tts
-configurate_methods:
-  - predefined-model
-provider_credential_schema:
-  credential_form_schemas:
-    - variable: api_key
-      label:
-        en_US: API Key
-      type: secret-input
-      required: true
-      placeholder:
-        zh_Hans: 在此输入您的 API Key
-        en_US: Enter your API Key

api/core/model_runtime/model_providers/gitee_ai/llm/Qwen2-72B-Instruct.yaml

@@ -1,105 +0,0 @@
-model: Qwen2-72B-Instruct
-label:
-  zh_Hans: Qwen2-72B-Instruct
-  en_US: Qwen2-72B-Instruct
-model_type: llm
-features:
-  - agent-thought
-model_properties:
-  mode: chat
-  context_size: 6400
-parameter_rules:
-  - name: stream
-    use_template: boolean
-    label:
-      en_US: "Stream"
-      zh_Hans: "流式"
-    type: boolean
-    default: true
-    required: true
-    help:
-      en_US: "Whether to return the results in batches through streaming. If set to true, the generated text will be pushed to the user in real time during the generation process."
-      zh_Hans: "是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。"
-
-  - name: max_tokens
-    use_template: max_tokens
-    label:
-      en_US: "Max Tokens"
-      zh_Hans: "最大Token数"
-    type: int
-    default: 512
-    min: 1
-    required: true
-    help:
-      en_US: "The maximum number of tokens that can be generated by the model varies depending on the model."
-      zh_Hans: "模型可生成的最大 token 个数，不同模型上限不同。"
-
-  - name: temperature
-    use_template: temperature
-    label:
-      en_US: "Temperature"
-      zh_Hans: "采样温度"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The randomness of the sampling temperature control output. The temperature value is within the range of [0.0, 1.0]. The higher the value, the more random and creative the output; the lower the value, the more stable it is. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_p
-    use_template: top_p
-    label:
-      en_US: "Top P"
-      zh_Hans: "Top P"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The value range of the sampling method is [0.0, 1.0]. The top_p value determines that the model selects tokens from the top p% of candidate words with the highest probability; when top_p is 0, this parameter is invalid. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_k
-    use_template: top_k
-    label:
-      en_US: "Top K"
-      zh_Hans: "Top K"
-    type: int
-    default: 50
-    min: 0
-    max: 100
-    required: true
-    help:
-      en_US: "The value range is [0,100], which limits the model to only select from the top k words with the highest probability when choosing the next word at each step. The larger the value, the more diverse text generation will be."
-      zh_Hans: "取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。"
-
-  - name: frequency_penalty
-    use_template: frequency_penalty
-    label:
-      en_US: "Frequency Penalty"
-      zh_Hans: "频率惩罚"
-    type: float
-    default: 0
-    min: -1.0
-    max: 1.0
-    precision: 1
-    required: false
-    help:
-      en_US: "Used to adjust the frequency of repeated content in automatically generated text. Positive numbers reduce repetition, while negative numbers increase repetition. After setting this parameter, if a word has already appeared in the text, the model will decrease the probability of choosing that word for subsequent generation."
-      zh_Hans: "用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低。"
-
-  - name: user
-    use_template: text
-    label:
-      en_US: "User"
-      zh_Hans: "用户"
-    type: string
-    required: false
-    help:
-      en_US: "Used to track and differentiate conversation requests from different users."
-      zh_Hans: "用于追踪和区分不同用户的对话请求。"

api/core/model_runtime/model_providers/gitee_ai/llm/Qwen2-7B-Instruct.yaml

@@ -1,105 +0,0 @@
-model: Qwen2-7B-Instruct
-label:
-  zh_Hans: Qwen2-7B-Instruct
-  en_US: Qwen2-7B-Instruct
-model_type: llm
-features:
-  - agent-thought
-model_properties:
-  mode: chat
-  context_size: 32768
-parameter_rules:
-  - name: stream
-    use_template: boolean
-    label:
-      en_US: "Stream"
-      zh_Hans: "流式"
-    type: boolean
-    default: true
-    required: true
-    help:
-      en_US: "Whether to return the results in batches through streaming. If set to true, the generated text will be pushed to the user in real time during the generation process."
-      zh_Hans: "是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。"
-
-  - name: max_tokens
-    use_template: max_tokens
-    label:
-      en_US: "Max Tokens"
-      zh_Hans: "最大Token数"
-    type: int
-    default: 512
-    min: 1
-    required: true
-    help:
-      en_US: "The maximum number of tokens that can be generated by the model varies depending on the model."
-      zh_Hans: "模型可生成的最大 token 个数，不同模型上限不同。"
-
-  - name: temperature
-    use_template: temperature
-    label:
-      en_US: "Temperature"
-      zh_Hans: "采样温度"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The randomness of the sampling temperature control output. The temperature value is within the range of [0.0, 1.0]. The higher the value, the more random and creative the output; the lower the value, the more stable it is. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_p
-    use_template: top_p
-    label:
-      en_US: "Top P"
-      zh_Hans: "Top P"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The value range of the sampling method is [0.0, 1.0]. The top_p value determines that the model selects tokens from the top p% of candidate words with the highest probability; when top_p is 0, this parameter is invalid. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_k
-    use_template: top_k
-    label:
-      en_US: "Top K"
-      zh_Hans: "Top K"
-    type: int
-    default: 50
-    min: 0
-    max: 100
-    required: true
-    help:
-      en_US: "The value range is [0,100], which limits the model to only select from the top k words with the highest probability when choosing the next word at each step. The larger the value, the more diverse text generation will be."
-      zh_Hans: "取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。"
-
-  - name: frequency_penalty
-    use_template: frequency_penalty
-    label:
-      en_US: "Frequency Penalty"
-      zh_Hans: "频率惩罚"
-    type: float
-    default: 0
-    min: -1.0
-    max: 1.0
-    precision: 1
-    required: false
-    help:
-      en_US: "Used to adjust the frequency of repeated content in automatically generated text. Positive numbers reduce repetition, while negative numbers increase repetition. After setting this parameter, if a word has already appeared in the text, the model will decrease the probability of choosing that word for subsequent generation."
-      zh_Hans: "用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低。"
-
-  - name: user
-    use_template: text
-    label:
-      en_US: "User"
-      zh_Hans: "用户"
-    type: string
-    required: false
-    help:
-      en_US: "Used to track and differentiate conversation requests from different users."
-      zh_Hans: "用于追踪和区分不同用户的对话请求。"

api/core/model_runtime/model_providers/gitee_ai/llm/Qwen2.5-72B-Instruct.yaml

@@ -1,95 +0,0 @@
-model: Qwen2.5-72B-Instruct
-label:
-  zh_Hans: Qwen2.5-72B-Instruct
-  en_US: Qwen2.5-72B-Instruct
-model_type: llm
-features:
-  - agent-thought
-  - tool-call
-  - stream-tool-call
-model_properties:
-  mode: chat
-  context_size: 32768
-parameter_rules:
-  - name: max_tokens
-    use_template: max_tokens
-    label:
-      en_US: "Max Tokens"
-      zh_Hans: "最大Token数"
-    type: int
-    default: 512
-    min: 1
-    required: true
-    help:
-      en_US: "The maximum number of tokens that can be generated by the model varies depending on the model."
-      zh_Hans: "模型可生成的最大 token 个数，不同模型上限不同。"
-
-  - name: temperature
-    use_template: temperature
-    label:
-      en_US: "Temperature"
-      zh_Hans: "采样温度"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The randomness of the sampling temperature control output. The temperature value is within the range of [0.0, 1.0]. The higher the value, the more random and creative the output; the lower the value, the more stable it is. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_p
-    use_template: top_p
-    label:
-      en_US: "Top P"
-      zh_Hans: "Top P"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The value range of the sampling method is [0.0, 1.0]. The top_p value determines that the model selects tokens from the top p% of candidate words with the highest probability; when top_p is 0, this parameter is invalid. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_k
-    use_template: top_k
-    label:
-      en_US: "Top K"
-      zh_Hans: "Top K"
-    type: int
-    default: 50
-    min: 0
-    max: 100
-    required: true
-    help:
-      en_US: "The value range is [0,100], which limits the model to only select from the top k words with the highest probability when choosing the next word at each step. The larger the value, the more diverse text generation will be."
-      zh_Hans: "取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。"
-
-  - name: frequency_penalty
-    use_template: frequency_penalty
-    label:
-      en_US: "Frequency Penalty"
-      zh_Hans: "频率惩罚"
-    type: float
-    default: 0
-    min: -1.0
-    max: 1.0
-    precision: 1
-    required: false
-    help:
-      en_US: "Used to adjust the frequency of repeated content in automatically generated text. Positive numbers reduce repetition, while negative numbers increase repetition. After setting this parameter, if a word has already appeared in the text, the model will decrease the probability of choosing that word for subsequent generation."
-      zh_Hans: "用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低。"
-
-  - name: user
-    use_template: text
-    label:
-      en_US: "User"
-      zh_Hans: "用户"
-    type: string
-    required: false
-    help:
-      en_US: "Used to track and differentiate conversation requests from different users."
-      zh_Hans: "用于追踪和区分不同用户的对话请求。"

api/core/model_runtime/model_providers/gitee_ai/llm/Yi-1.5-34B-Chat.yaml

@@ -1,105 +0,0 @@
-model: Yi-1.5-34B-Chat
-label:
-  zh_Hans: Yi-1.5-34B-Chat
-  en_US: Yi-1.5-34B-Chat
-model_type: llm
-features:
-  - agent-thought
-model_properties:
-  mode: chat
-  context_size: 4096
-parameter_rules:
-  - name: stream
-    use_template: boolean
-    label:
-      en_US: "Stream"
-      zh_Hans: "流式"
-    type: boolean
-    default: true
-    required: true
-    help:
-      en_US: "Whether to return the results in batches through streaming. If set to true, the generated text will be pushed to the user in real time during the generation process."
-      zh_Hans: "是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。"
-
-  - name: max_tokens
-    use_template: max_tokens
-    label:
-      en_US: "Max Tokens"
-      zh_Hans: "最大Token数"
-    type: int
-    default: 512
-    min: 1
-    required: true
-    help:
-      en_US: "The maximum number of tokens that can be generated by the model varies depending on the model."
-      zh_Hans: "模型可生成的最大 token 个数，不同模型上限不同。"
-
-  - name: temperature
-    use_template: temperature
-    label:
-      en_US: "Temperature"
-      zh_Hans: "采样温度"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The randomness of the sampling temperature control output. The temperature value is within the range of [0.0, 1.0]. The higher the value, the more random and creative the output; the lower the value, the more stable it is. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_p
-    use_template: top_p
-    label:
-      en_US: "Top P"
-      zh_Hans: "Top P"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The value range of the sampling method is [0.0, 1.0]. The top_p value determines that the model selects tokens from the top p% of candidate words with the highest probability; when top_p is 0, this parameter is invalid. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_k
-    use_template: top_k
-    label:
-      en_US: "Top K"
-      zh_Hans: "Top K"
-    type: int
-    default: 50
-    min: 0
-    max: 100
-    required: true
-    help:
-      en_US: "The value range is [0,100], which limits the model to only select from the top k words with the highest probability when choosing the next word at each step. The larger the value, the more diverse text generation will be."
-      zh_Hans: "取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。"
-
-  - name: frequency_penalty
-    use_template: frequency_penalty
-    label:
-      en_US: "Frequency Penalty"
-      zh_Hans: "频率惩罚"
-    type: float
-    default: 0
-    min: -1.0
-    max: 1.0
-    precision: 1
-    required: false
-    help:
-      en_US: "Used to adjust the frequency of repeated content in automatically generated text. Positive numbers reduce repetition, while negative numbers increase repetition. After setting this parameter, if a word has already appeared in the text, the model will decrease the probability of choosing that word for subsequent generation."
-      zh_Hans: "用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低。"
-
-  - name: user
-    use_template: text
-    label:
-      en_US: "User"
-      zh_Hans: "用户"
-    type: string
-    required: false
-    help:
-      en_US: "Used to track and differentiate conversation requests from different users."
-      zh_Hans: "用于追踪和区分不同用户的对话请求。"

api/core/model_runtime/model_providers/gitee_ai/llm/_position.yaml

@@ -1,8 +0,0 @@
-- Qwen2.5-72B-Instruct
-- Qwen2-7B-Instruct
-- Qwen2-72B-Instruct
-- Yi-1.5-34B-Chat
-- glm-4-9b-chat
-- deepseek-coder-33B-instruct-chat
-- deepseek-coder-33B-instruct-completions
-- codegeex4-all-9b

api/core/model_runtime/model_providers/gitee_ai/llm/codegeex4-all-9b.yaml

@@ -1,105 +0,0 @@
-model: codegeex4-all-9b
-label:
-  zh_Hans: codegeex4-all-9b
-  en_US: codegeex4-all-9b
-model_type: llm
-features:
-  - agent-thought
-model_properties:
-  mode: chat
-  context_size: 40960
-parameter_rules:
-  - name: stream
-    use_template: boolean
-    label:
-      en_US: "Stream"
-      zh_Hans: "流式"
-    type: boolean
-    default: true
-    required: true
-    help:
-      en_US: "Whether to return the results in batches through streaming. If set to true, the generated text will be pushed to the user in real time during the generation process."
-      zh_Hans: "是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。"
-
-  - name: max_tokens
-    use_template: max_tokens
-    label:
-      en_US: "Max Tokens"
-      zh_Hans: "最大Token数"
-    type: int
-    default: 512
-    min: 1
-    required: true
-    help:
-      en_US: "The maximum number of tokens that can be generated by the model varies depending on the model."
-      zh_Hans: "模型可生成的最大 token 个数，不同模型上限不同。"
-
-  - name: temperature
-    use_template: temperature
-    label:
-      en_US: "Temperature"
-      zh_Hans: "采样温度"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The randomness of the sampling temperature control output. The temperature value is within the range of [0.0, 1.0]. The higher the value, the more random and creative the output; the lower the value, the more stable it is. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_p
-    use_template: top_p
-    label:
-      en_US: "Top P"
-      zh_Hans: "Top P"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The value range of the sampling method is [0.0, 1.0]. The top_p value determines that the model selects tokens from the top p% of candidate words with the highest probability; when top_p is 0, this parameter is invalid. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_k
-    use_template: top_k
-    label:
-      en_US: "Top K"
-      zh_Hans: "Top K"
-    type: int
-    default: 50
-    min: 0
-    max: 100
-    required: true
-    help:
-      en_US: "The value range is [0,100], which limits the model to only select from the top k words with the highest probability when choosing the next word at each step. The larger the value, the more diverse text generation will be."
-      zh_Hans: "取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。"
-
-  - name: frequency_penalty
-    use_template: frequency_penalty
-    label:
-      en_US: "Frequency Penalty"
-      zh_Hans: "频率惩罚"
-    type: float
-    default: 0
-    min: -1.0
-    max: 1.0
-    precision: 1
-    required: false
-    help:
-      en_US: "Used to adjust the frequency of repeated content in automatically generated text. Positive numbers reduce repetition, while negative numbers increase repetition. After setting this parameter, if a word has already appeared in the text, the model will decrease the probability of choosing that word for subsequent generation."
-      zh_Hans: "用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低。"
-
-  - name: user
-    use_template: text
-    label:
-      en_US: "User"
-      zh_Hans: "用户"
-    type: string
-    required: false
-    help:
-      en_US: "Used to track and differentiate conversation requests from different users."
-      zh_Hans: "用于追踪和区分不同用户的对话请求。"

api/core/model_runtime/model_providers/gitee_ai/llm/deepseek-coder-33B-instruct-chat.yaml

@@ -1,105 +0,0 @@
-model: deepseek-coder-33B-instruct-chat
-label:
-  zh_Hans: deepseek-coder-33B-instruct-chat
-  en_US: deepseek-coder-33B-instruct-chat
-model_type: llm
-features:
-  - agent-thought
-model_properties:
-  mode: chat
-  context_size: 9000
-parameter_rules:
-  - name: stream
-    use_template: boolean
-    label:
-      en_US: "Stream"
-      zh_Hans: "流式"
-    type: boolean
-    default: true
-    required: true
-    help:
-      en_US: "Whether to return the results in batches through streaming. If set to true, the generated text will be pushed to the user in real time during the generation process."
-      zh_Hans: "是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。"
-
-  - name: max_tokens
-    use_template: max_tokens
-    label:
-      en_US: "Max Tokens"
-      zh_Hans: "最大Token数"
-    type: int
-    default: 512
-    min: 1
-    required: true
-    help:
-      en_US: "The maximum number of tokens that can be generated by the model varies depending on the model."
-      zh_Hans: "模型可生成的最大 token 个数，不同模型上限不同。"
-
-  - name: temperature
-    use_template: temperature
-    label:
-      en_US: "Temperature"
-      zh_Hans: "采样温度"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The randomness of the sampling temperature control output. The temperature value is within the range of [0.0, 1.0]. The higher the value, the more random and creative the output; the lower the value, the more stable it is. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_p
-    use_template: top_p
-    label:
-      en_US: "Top P"
-      zh_Hans: "Top P"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The value range of the sampling method is [0.0, 1.0]. The top_p value determines that the model selects tokens from the top p% of candidate words with the highest probability; when top_p is 0, this parameter is invalid. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_k
-    use_template: top_k
-    label:
-      en_US: "Top K"
-      zh_Hans: "Top K"
-    type: int
-    default: 50
-    min: 0
-    max: 100
-    required: true
-    help:
-      en_US: "The value range is [0,100], which limits the model to only select from the top k words with the highest probability when choosing the next word at each step. The larger the value, the more diverse text generation will be."
-      zh_Hans: "取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。"
-
-  - name: frequency_penalty
-    use_template: frequency_penalty
-    label:
-      en_US: "Frequency Penalty"
-      zh_Hans: "频率惩罚"
-    type: float
-    default: 0
-    min: -1.0
-    max: 1.0
-    precision: 1
-    required: false
-    help:
-      en_US: "Used to adjust the frequency of repeated content in automatically generated text. Positive numbers reduce repetition, while negative numbers increase repetition. After setting this parameter, if a word has already appeared in the text, the model will decrease the probability of choosing that word for subsequent generation."
-      zh_Hans: "用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低。"
-
-  - name: user
-    use_template: text
-    label:
-      en_US: "User"
-      zh_Hans: "用户"
-    type: string
-    required: false
-    help:
-      en_US: "Used to track and differentiate conversation requests from different users."
-      zh_Hans: "用于追踪和区分不同用户的对话请求。"

api/core/model_runtime/model_providers/gitee_ai/llm/deepseek-coder-33B-instruct-completions.yaml

@@ -1,91 +0,0 @@
-model: deepseek-coder-33B-instruct-completions
-label:
-  zh_Hans: deepseek-coder-33B-instruct-completions
-  en_US: deepseek-coder-33B-instruct-completions
-model_type: llm
-features:
-  - agent-thought
-model_properties:
-  mode: completion
-  context_size: 9000
-parameter_rules:
-  - name: stream
-    use_template: boolean
-    label:
-      en_US: "Stream"
-      zh_Hans: "流式"
-    type: boolean
-    default: true
-    required: true
-    help:
-      en_US: "Whether to return the results in batches through streaming. If set to true, the generated text will be pushed to the user in real time during the generation process."
-      zh_Hans: "是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。"
-
-  - name: max_tokens
-    use_template: max_tokens
-    label:
-      en_US: "Max Tokens"
-      zh_Hans: "最大Token数"
-    type: int
-    default: 512
-    min: 1
-    required: true
-    help:
-      en_US: "The maximum number of tokens that can be generated by the model varies depending on the model."
-      zh_Hans: "模型可生成的最大 token 个数，不同模型上限不同。"
-
-  - name: temperature
-    use_template: temperature
-    label:
-      en_US: "Temperature"
-      zh_Hans: "采样温度"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The randomness of the sampling temperature control output. The temperature value is within the range of [0.0, 1.0]. The higher the value, the more random and creative the output; the lower the value, the more stable it is. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_p
-    use_template: top_p
-    label:
-      en_US: "Top P"
-      zh_Hans: "Top P"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The value range of the sampling method is [0.0, 1.0]. The top_p value determines that the model selects tokens from the top p% of candidate words with the highest probability; when top_p is 0, this parameter is invalid. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: frequency_penalty
-    use_template: frequency_penalty
-    label:
-      en_US: "Frequency Penalty"
-      zh_Hans: "频率惩罚"
-    type: float
-    default: 0
-    min: -1.0
-    max: 1.0
-    precision: 1
-    required: false
-    help:
-      en_US: "Used to adjust the frequency of repeated content in automatically generated text. Positive numbers reduce repetition, while negative numbers increase repetition. After setting this parameter, if a word has already appeared in the text, the model will decrease the probability of choosing that word for subsequent generation."
-      zh_Hans: "用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低。"
-
-  - name: user
-    use_template: text
-    label:
-      en_US: "User"
-      zh_Hans: "用户"
-    type: string
-    required: false
-    help:
-      en_US: "Used to track and differentiate conversation requests from different users."
-      zh_Hans: "用于追踪和区分不同用户的对话请求。"

api/core/model_runtime/model_providers/gitee_ai/llm/glm-4-9b-chat.yaml

@@ -1,105 +0,0 @@
-model: glm-4-9b-chat
-label:
-  zh_Hans: glm-4-9b-chat
-  en_US: glm-4-9b-chat
-model_type: llm
-features:
-  - agent-thought
-model_properties:
-  mode: chat
-  context_size: 32768
-parameter_rules:
-  - name: stream
-    use_template: boolean
-    label:
-      en_US: "Stream"
-      zh_Hans: "流式"
-    type: boolean
-    default: true
-    required: true
-    help:
-      en_US: "Whether to return the results in batches through streaming. If set to true, the generated text will be pushed to the user in real time during the generation process."
-      zh_Hans: "是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。"
-
-  - name: max_tokens
-    use_template: max_tokens
-    label:
-      en_US: "Max Tokens"
-      zh_Hans: "最大Token数"
-    type: int
-    default: 512
-    min: 1
-    required: true
-    help:
-      en_US: "The maximum number of tokens that can be generated by the model varies depending on the model."
-      zh_Hans: "模型可生成的最大 token 个数，不同模型上限不同。"
-
-  - name: temperature
-    use_template: temperature
-    label:
-      en_US: "Temperature"
-      zh_Hans: "采样温度"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The randomness of the sampling temperature control output. The temperature value is within the range of [0.0, 1.0]. The higher the value, the more random and creative the output; the lower the value, the more stable it is. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_p
-    use_template: top_p
-    label:
-      en_US: "Top P"
-      zh_Hans: "Top P"
-    type: float
-    default: 0.7
-    min: 0.0
-    max: 1.0
-    precision: 1
-    required: true
-    help:
-      en_US: "The value range of the sampling method is [0.0, 1.0]. The top_p value determines that the model selects tokens from the top p% of candidate words with the highest probability; when top_p is 0, this parameter is invalid. It is recommended to adjust either top_p or temperature parameters according to your needs to avoid adjusting both at the same time."
-      zh_Hans: "采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。"
-
-  - name: top_k
-    use_template: top_k
-    label:
-      en_US: "Top K"
-      zh_Hans: "Top K"
-    type: int
-    default: 50
-    min: 0
-    max: 100
-    required: true
-    help:
-      en_US: "The value range is [0,100], which limits the model to only select from the top k words with the highest probability when choosing the next word at each step. The larger the value, the more diverse text generation will be."
-      zh_Hans: "取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。"
-
-  - name: frequency_penalty
-    use_template: frequency_penalty
-    label:
-      en_US: "Frequency Penalty"
-      zh_Hans: "频率惩罚"
-    type: float
-    default: 0
-    min: -1.0
-    max: 1.0
-    precision: 1
-    required: false
-    help:
-      en_US: "Used to adjust the frequency of repeated content in automatically generated text. Positive numbers reduce repetition, while negative numbers increase repetition. After setting this parameter, if a word has already appeared in the text, the model will decrease the probability of choosing that word for subsequent generation."
-      zh_Hans: "用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低。"
-
-  - name: user
-    use_template: text
-    label:
-      en_US: "User"
-      zh_Hans: "用户"
-    type: string
-    required: false
-    help:
-      en_US: "Used to track and differentiate conversation requests from different users."
-      zh_Hans: "用于追踪和区分不同用户的对话请求。"

api/core/model_runtime/model_providers/gitee_ai/llm/llm.py

@@ -1,51 +0,0 @@
-from collections.abc import Generator
-from typing import Optional, Union
-
-from core.model_runtime.entities.llm_entities import LLMMode, LLMResult
-from core.model_runtime.entities.message_entities import (
-    PromptMessage,
-    PromptMessageTool,
-)
-from core.model_runtime.entities.model_entities import ModelFeature
-from core.model_runtime.model_providers.openai_api_compatible.llm.llm import OAIAPICompatLargeLanguageModel
-
-
-class GiteeAILargeLanguageModel(OAIAPICompatLargeLanguageModel):
-    MODEL_TO_IDENTITY: dict[str, str] = {
-        "Yi-1.5-34B-Chat": "Yi-34B-Chat",
-        "deepseek-coder-33B-instruct-completions": "deepseek-coder-33B-instruct",
-        "deepseek-coder-33B-instruct-chat": "deepseek-coder-33B-instruct",
-    }
-
-    def _invoke(
-        self,
-        model: str,
-        credentials: dict,
-        prompt_messages: list[PromptMessage],
-        model_parameters: dict,
-        tools: Optional[list[PromptMessageTool]] = None,
-        stop: Optional[list[str]] = None,
-        stream: bool = True,
-        user: Optional[str] = None,
-    ) -> Union[LLMResult, Generator]:
-        self._add_custom_parameters(credentials, model, model_parameters)
-        return super()._invoke(model, credentials, prompt_messages, model_parameters, tools, stop, stream, user)
-
-    def validate_credentials(self, model: str, credentials: dict) -> None:
-        self._add_custom_parameters(credentials, model, None)
-        super().validate_credentials(model, credentials)
-
-    def _add_custom_parameters(self, credentials: dict, model: str, model_parameters: dict) -> None:
-        if model is None:
-            model = "bge-large-zh-v1.5"
-
-        model_identity = GiteeAILargeLanguageModel.MODEL_TO_IDENTITY.get(model, model)
-        credentials["endpoint_url"] = f"https://ai.gitee.com/api/serverless/{model_identity}/"
-        if model.endswith("completions"):
-            credentials["mode"] = LLMMode.COMPLETION.value
-        else:
-            credentials["mode"] = LLMMode.CHAT.value
-
-        schema = self.get_model_schema(model, credentials)
-        if ModelFeature.TOOL_CALL in schema.features or ModelFeature.MULTI_TOOL_CALL in schema.features:
-            credentials["function_calling_type"] = "tool_call"

api/core/model_runtime/model_providers/gitee_ai/rerank/_position.yaml

@@ -1 +0,0 @@
-- bge-reranker-v2-m3

api/core/model_runtime/model_providers/gitee_ai/rerank/bge-reranker-v2-m3.yaml

@@ -1,4 +0,0 @@
-model: bge-reranker-v2-m3
-model_type: rerank
-model_properties:
-  context_size: 1024

api/core/model_runtime/model_providers/gitee_ai/rerank/rerank.py

@@ -1,128 +0,0 @@
-from typing import Optional
-
-import httpx
-
-from core.model_runtime.entities.common_entities import I18nObject
-from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType
-from core.model_runtime.entities.rerank_entities import RerankDocument, RerankResult
-from core.model_runtime.errors.invoke import (
-    InvokeAuthorizationError,
-    InvokeBadRequestError,
-    InvokeConnectionError,
-    InvokeError,
-    InvokeRateLimitError,
-    InvokeServerUnavailableError,
-)
-from core.model_runtime.errors.validate import CredentialsValidateFailedError
-from core.model_runtime.model_providers.__base.rerank_model import RerankModel
-
-
-class GiteeAIRerankModel(RerankModel):
-    """
-    Model class for rerank model.
-    """
-
-    def _invoke(
-        self,
-        model: str,
-        credentials: dict,
-        query: str,
-        docs: list[str],
-        score_threshold: Optional[float] = None,
-        top_n: Optional[int] = None,
-        user: Optional[str] = None,
-    ) -> RerankResult:
-        """
-        Invoke rerank model
-
-        :param model: model name
-        :param credentials: model credentials
-        :param query: search query
-        :param docs: docs for reranking
-        :param score_threshold: score threshold
-        :param top_n: top n documents to return
-        :param user: unique user id
-        :return: rerank result
-        """
-        if len(docs) == 0:
-            return RerankResult(model=model, docs=[])
-
-        base_url = credentials.get("base_url", "https://ai.gitee.com/api/serverless")
-        base_url = base_url.removesuffix("/")
-
-        try:
-            body = {"model": model, "query": query, "documents": docs}
-            if top_n is not None:
-                body["top_n"] = top_n
-            response = httpx.post(
-                f"{base_url}/{model}/rerank",
-                json=body,
-                headers={"Authorization": f"Bearer {credentials.get('api_key')}"},
-            )
-
-            response.raise_for_status()
-            results = response.json()
-
-            rerank_documents = []
-            for result in results["results"]:
-                rerank_document = RerankDocument(
-                    index=result["index"],
-                    text=result["document"]["text"],
-                    score=result["relevance_score"],
-                )
-                if score_threshold is None or result["relevance_score"] >= score_threshold:
-                    rerank_documents.append(rerank_document)
-            return RerankResult(model=model, docs=rerank_documents)
-        except httpx.HTTPStatusError as e:
-            raise InvokeServerUnavailableError(str(e))
-
-    def validate_credentials(self, model: str, credentials: dict) -> None:
-        """
-        Validate model credentials
-
-        :param model: model name
-        :param credentials: model credentials
-        :return:
-        """
-        try:
-            self._invoke(
-                model=model,
-                credentials=credentials,
-                query="What is the capital of the United States?",
-                docs=[
-                    "Carson City is the capital city of the American state of Nevada. At the 2010 United States "
-                    "Census, Carson City had a population of 55,274.",
-                    "The Commonwealth of the Northern Mariana Islands is a group of islands in the Pacific Ocean that "
-                    "are a political division controlled by the United States. Its capital is Saipan.",
-                ],
-                score_threshold=0.01,
-            )
-        except Exception as ex:
-            raise CredentialsValidateFailedError(str(ex))
-
-    @property
-    def _invoke_error_mapping(self) -> dict[type[InvokeError], list[type[Exception]]]:
-        """
-        Map model invoke error to unified error
-        """
-        return {
-            InvokeConnectionError: [httpx.ConnectError],
-            InvokeServerUnavailableError: [httpx.RemoteProtocolError],
-            InvokeRateLimitError: [],
-            InvokeAuthorizationError: [httpx.HTTPStatusError],
-            InvokeBadRequestError: [httpx.RequestError],
-        }
-
-    def get_customizable_model_schema(self, model: str, credentials: dict) -> AIModelEntity:
-        """
-        generate custom model entities from credentials
-        """
-        entity = AIModelEntity(
-            model=model,
-            label=I18nObject(en_US=model),
-            model_type=ModelType.RERANK,
-            fetch_from=FetchFrom.CUSTOMIZABLE_MODEL,
-            model_properties={ModelPropertyKey.CONTEXT_SIZE: int(credentials.get("context_size"))},
-        )
-
-        return entity

api/core/model_runtime/model_providers/gitee_ai/speech2text/_position.yaml

@@ -1,2 +0,0 @@
-- whisper-base
-- whisper-large

api/core/model_runtime/model_providers/gitee_ai/speech2text/speech2text.py

@@ -1,53 +0,0 @@
-import os
-from typing import IO, Optional
-
-import requests
-
-from core.model_runtime.errors.invoke import InvokeBadRequestError
-from core.model_runtime.errors.validate import CredentialsValidateFailedError
-from core.model_runtime.model_providers.__base.speech2text_model import Speech2TextModel
-from core.model_runtime.model_providers.gitee_ai._common import _CommonGiteeAI
-
-
-class GiteeAISpeech2TextModel(_CommonGiteeAI, Speech2TextModel):
-    """
-    Model class for OpenAI Compatible Speech to text model.
-    """
-
-    def _invoke(self, model: str, credentials: dict, file: IO[bytes], user: Optional[str] = None) -> str:
-        """
-        Invoke speech2text model
-
-        :param model: model name
-        :param credentials: model credentials
-        :param file: audio file
-        :param user: unique user id
-        :return: text for given audio file
-        """
-        # doc: https://ai.gitee.com/docs/openapi/serverless#tag/serverless/POST/{service}/speech-to-text
-
-        endpoint_url = f"https://ai.gitee.com/api/serverless/{model}/speech-to-text"
-        files = [("file", file)]
-        _, file_ext = os.path.splitext(file.name)
-        headers = {"Content-Type": f"audio/{file_ext}", "Authorization": f"Bearer {credentials.get('api_key')}"}
-        response = requests.post(endpoint_url, headers=headers, files=files)
-        if response.status_code != 200:
-            raise InvokeBadRequestError(response.text)
-        response_data = response.json()
-        return response_data["text"]
-
-    def validate_credentials(self, model: str, credentials: dict) -> None:
-        """
-        Validate model credentials
-
-        :param model: model name
-        :param credentials: model credentials
-        :return:
-        """
-        try:
-            audio_file_path = self._get_demo_file_path()
-
-            with open(audio_file_path, "rb") as audio_file:
-                self._invoke(model, credentials, audio_file)
-        except Exception as ex:
-            raise CredentialsValidateFailedError(str(ex))

api/core/model_runtime/model_providers/gitee_ai/speech2text/whisper-base.yaml

@@ -1,5 +0,0 @@
-model: whisper-base
-model_type: speech2text
-model_properties:
-  file_upload_limit: 1
-  supported_file_extensions: flac,mp3,mp4,mpeg,mpga,m4a,ogg,wav,webm

api/core/model_runtime/model_providers/gitee_ai/speech2text/whisper-large.yaml

@@ -1,5 +0,0 @@
-model: whisper-large
-model_type: speech2text
-model_properties:
-  file_upload_limit: 1
-  supported_file_extensions: flac,mp3,mp4,mpeg,mpga,m4a,ogg,wav,webm

api/core/model_runtime/model_providers/gitee_ai/text_embedding/_position.yaml

@@ -1,3 +0,0 @@
-- bge-large-zh-v1.5
-- bge-small-zh-v1.5
-- bge-m3

api/core/model_runtime/model_providers/gitee_ai/text_embedding/bge-large-zh-v1.5.yaml

@@ -1,8 +0,0 @@
-model: bge-large-zh-v1.5
-label:
-  zh_Hans: bge-large-zh-v1.5
-  en_US: bge-large-zh-v1.5
-model_type: text-embedding
-model_properties:
-  context_size: 200000
-  max_chunks: 20

api/core/model_runtime/model_providers/gitee_ai/text_embedding/bge-m3.yaml

@@ -1,8 +0,0 @@
-model: bge-m3
-label:
-  zh_Hans: bge-m3
-  en_US: bge-m3
-model_type: text-embedding
-model_properties:
-  context_size: 200000
-  max_chunks: 20

api/core/model_runtime/model_providers/gitee_ai/text_embedding/bge-small-zh-v1.5.yaml

@@ -1,8 +0,0 @@
-model: bge-small-zh-v1.5
-label:
-  zh_Hans: bge-small-zh-v1.5
-  en_US: bge-small-zh-v1.5
-model_type: text-embedding
-model_properties:
-  context_size: 200000
-  max_chunks: 20

api/core/model_runtime/model_providers/gitee_ai/text_embedding/text_embedding.py

@@ -1,31 +0,0 @@
-from typing import Optional
-
-from core.entities.embedding_type import EmbeddingInputType
-from core.model_runtime.entities.text_embedding_entities import TextEmbeddingResult
-from core.model_runtime.model_providers.openai_api_compatible.text_embedding.text_embedding import (
-    OAICompatEmbeddingModel,
-)
-
-
-class GiteeAIEmbeddingModel(OAICompatEmbeddingModel):
-    def _invoke(
-        self,
-        model: str,
-        credentials: dict,
-        texts: list[str],
-        user: Optional[str] = None,
-        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
-    ) -> TextEmbeddingResult:
-        self._add_custom_parameters(credentials, model)
-        return super()._invoke(model, credentials, texts, user, input_type)
-
-    def validate_credentials(self, model: str, credentials: dict) -> None:
-        self._add_custom_parameters(credentials, None)
-        super().validate_credentials(model, credentials)
-
-    @staticmethod
-    def _add_custom_parameters(credentials: dict, model: str) -> None:
-        if model is None:
-            model = "bge-m3"
-
-        credentials["endpoint_url"] = f"https://ai.gitee.com/api/serverless/{model}/v1/"

api/core/model_runtime/model_providers/gitee_ai/tts/ChatTTS.yaml

@@ -1,11 +0,0 @@
-model: ChatTTS
-model_type: tts
-model_properties:
-  default_voice: 'default'
-  voices:
-    - mode: 'default'
-      name: 'Default'
-      language: [ 'zh-Hans', 'en-US', 'de-DE', 'fr-FR', 'es-ES', 'it-IT', 'th-TH', 'id-ID' ]
-  word_limit: 3500
-  audio_type: 'mp3'
-  max_workers: 5

api/core/model_runtime/model_providers/gitee_ai/tts/FunAudioLLM-CosyVoice-300M.yaml

@@ -1,11 +0,0 @@
-model: FunAudioLLM-CosyVoice-300M
-model_type: tts
-model_properties:
-  default_voice: 'default'
-  voices:
-    - mode: 'default'
-      name: 'Default'
-      language: [ 'zh-Hans', 'en-US', 'de-DE', 'fr-FR', 'es-ES', 'it-IT', 'th-TH', 'id-ID' ]
-  word_limit: 3500
-  audio_type: 'mp3'
-  max_workers: 5

api/core/model_runtime/model_providers/gitee_ai/tts/_position.yaml

@@ -1,4 +0,0 @@
-- speecht5_tts
-- ChatTTS
-- fish-speech-1.2-sft
-- FunAudioLLM-CosyVoice-300M

api/core/model_runtime/model_providers/gitee_ai/tts/fish-speech-1.2-sft.yaml

@@ -1,11 +0,0 @@
-model: fish-speech-1.2-sft
-model_type: tts
-model_properties:
-  default_voice: 'default'
-  voices:
-    - mode: 'default'
-      name: 'Default'
-      language: [ 'zh-Hans', 'en-US', 'de-DE', 'fr-FR', 'es-ES', 'it-IT', 'th-TH', 'id-ID' ]
-  word_limit: 3500
-  audio_type: 'mp3'
-  max_workers: 5

api/core/model_runtime/model_providers/gitee_ai/tts/speecht5_tts.yaml

@@ -1,11 +0,0 @@
-model: speecht5_tts
-model_type: tts
-model_properties:
-  default_voice: 'default'
-  voices:
-    - mode: 'default'
-      name: 'Default'
-      language: [ 'zh-Hans', 'en-US', 'de-DE', 'fr-FR', 'es-ES', 'it-IT', 'th-TH', 'id-ID' ]
-  word_limit: 3500
-  audio_type: 'mp3'
-  max_workers: 5

api/core/model_runtime/model_providers/gitee_ai/tts/tts.py

@@ -1,79 +0,0 @@
-from typing import Optional
-
-import requests
-
-from core.model_runtime.errors.invoke import InvokeBadRequestError
-from core.model_runtime.errors.validate import CredentialsValidateFailedError
-from core.model_runtime.model_providers.__base.tts_model import TTSModel
-from core.model_runtime.model_providers.gitee_ai._common import _CommonGiteeAI
-
-
-class GiteeAIText2SpeechModel(_CommonGiteeAI, TTSModel):
-    """
-    Model class for OpenAI Speech to text model.
-    """
-
-    def _invoke(
-        self, model: str, tenant_id: str, credentials: dict, content_text: str, voice: str, user: Optional[str] = None
-    ) -> any:
-        """
-        _invoke text2speech model
-
-        :param model: model name
-        :param tenant_id: user tenant id
-        :param credentials: model credentials
-        :param content_text: text content to be translated
-        :param voice: model timbre
-        :param user: unique user id
-        :return: text translated to audio file
-        """
-        return self._tts_invoke_streaming(model=model, credentials=credentials, content_text=content_text, voice=voice)
-
-    def validate_credentials(self, model: str, credentials: dict) -> None:
-        """
-        validate credentials text2speech model
-
-        :param model: model name
-        :param credentials: model credentials
-        :return: text translated to audio file
-        """
-        try:
-            self._tts_invoke_streaming(
-                model=model,
-                credentials=credentials,
-                content_text="Hello Dify!",
-                voice=self._get_model_default_voice(model, credentials),
-            )
-        except Exception as ex:
-            raise CredentialsValidateFailedError(str(ex))
-
-    def _tts_invoke_streaming(self, model: str, credentials: dict, content_text: str, voice: str) -> any:
-        """
-        _tts_invoke_streaming text2speech model
-        :param model: model name
-        :param credentials: model credentials
-        :param content_text: text content to be translated
-        :param voice: model timbre
-        :return: text translated to audio file
-        """
-        try:
-            # doc: https://ai.gitee.com/docs/openapi/serverless#tag/serverless/POST/{service}/text-to-speech
-            endpoint_url = "https://ai.gitee.com/api/serverless/" + model + "/text-to-speech"
-
-            headers = {"Content-Type": "application/json"}
-            api_key = credentials.get("api_key")
-            if api_key:
-                headers["Authorization"] = f"Bearer {api_key}"
-
-            payload = {"inputs": content_text}
-            response = requests.post(endpoint_url, headers=headers, json=payload)
-
-            if response.status_code != 200:
-                raise InvokeBadRequestError(response.text)
-
-            data = response.content
-
-            for i in range(0, len(data), 1024):
-                yield data[i : i + 1024]
-        except Exception as ex:
-            raise InvokeBadRequestError(str(ex))

api/core/moderation/keywords/keywords.py

@@ -1,3 +1,6 @@
+from collections.abc import Sequence
+from typing import Any
+
 from core.moderation.base import Moderation, ModerationAction, ModerationInputsResult, ModerationOutputsResult
 
 
@@ -62,5 +65,5 @@ class KeywordsModeration(Moderation):
     def _is_violated(self, inputs: dict, keywords_list: list) -> bool:
         return any(self._check_keywords_in_value(keywords_list, value) for value in inputs.values())
 
-    def _check_keywords_in_value(self, keywords_list, value) -> bool:
-        return any(keyword.lower() in value.lower() for keyword in keywords_list)
+    def _check_keywords_in_value(self, keywords_list: Sequence[str], value: Any) -> bool:
+        return any(keyword.lower() in str(value).lower() for keyword in keywords_list)

api/core/moderation/output_moderation.py

@@ -126,6 +126,6 @@ class OutputModeration(BaseModel):
             result: ModerationOutputsResult = moderation_factory.moderation_for_outputs(moderation_buffer)
             return result
         except Exception as e:
-            logger.exception("Moderation Output error: %s", e)
+            logger.exception(f"Moderation Output error, app_id: {app_id}")
 
         return None

api/core/ops/langsmith_trace/entities/langsmith_trace_entity.py

@@ -49,6 +49,7 @@ class LangSmithRunModel(LangSmithTokenUsage, LangSmithMultiModel):
     reference_example_id: Optional[str] = Field(None, description="Reference example ID associated with the run")
     input_attachments: Optional[dict[str, Any]] = Field(None, description="Input attachments of the run")
     output_attachments: Optional[dict[str, Any]] = Field(None, description="Output attachments of the run")
+    dotted_order: Optional[str] = Field(None, description="Dotted order of the run")
 
     @field_validator("inputs", "outputs")
     @classmethod

api/core/ops/langsmith_trace/langsmith_trace.py

@@ -25,7 +25,7 @@ from core.ops.langsmith_trace.entities.langsmith_trace_entity import (
     LangSmithRunType,
     LangSmithRunUpdateModel,
 )
-from core.ops.utils import filter_none_values
+from core.ops.utils import filter_none_values, generate_dotted_order
 from extensions.ext_database import db
 from models.model import EndUser, MessageFile
 from models.workflow import WorkflowNodeExecution
@@ -62,6 +62,16 @@ class LangSmithDataTrace(BaseTraceInstance):
             self.generate_name_trace(trace_info)
 
     def workflow_trace(self, trace_info: WorkflowTraceInfo):
+        trace_id = trace_info.message_id or trace_info.workflow_app_log_id or trace_info.workflow_run_id
+        message_dotted_order = (
+            generate_dotted_order(trace_info.message_id, trace_info.start_time) if trace_info.message_id else None
+        )
+        workflow_dotted_order = generate_dotted_order(
+            trace_info.workflow_app_log_id or trace_info.workflow_run_id,
+            trace_info.workflow_data.created_at,
+            message_dotted_order,
+        )
+
         if trace_info.message_id:
             message_run = LangSmithRunModel(
                 id=trace_info.message_id,
@@ -76,6 +86,8 @@ class LangSmithDataTrace(BaseTraceInstance):
                 },
                 tags=["message", "workflow"],
                 error=trace_info.error,
+                trace_id=trace_id,
+                dotted_order=message_dotted_order,
             )
             self.add_run(message_run)
 
@@ -95,6 +107,8 @@ class LangSmithDataTrace(BaseTraceInstance):
             error=trace_info.error,
             tags=["workflow"],
             parent_run_id=trace_info.message_id or None,
+            trace_id=trace_id,
+            dotted_order=workflow_dotted_order,
         )
 
         self.add_run(langsmith_run)
@@ -177,6 +191,7 @@ class LangSmithDataTrace(BaseTraceInstance):
             else:
                 run_type = LangSmithRunType.tool
 
+            node_dotted_order = generate_dotted_order(node_execution_id, created_at, workflow_dotted_order)
             langsmith_run = LangSmithRunModel(
                 total_tokens=node_total_tokens,
                 name=node_type,
@@ -191,6 +206,9 @@ class LangSmithDataTrace(BaseTraceInstance):
                 },
                 parent_run_id=trace_info.workflow_app_log_id or trace_info.workflow_run_id,
                 tags=["node_execution"],
+                id=node_execution_id,
+                trace_id=trace_id,
+                dotted_order=node_dotted_order,
             )
 
             self.add_run(langsmith_run)

api/core/ops/ops_trace_manager.py

@@ -711,7 +711,7 @@ class TraceQueueManager:
                 trace_task.app_id = self.app_id
                 trace_manager_queue.put(trace_task)
         except Exception as e:
-            logging.exception(f"Error adding trace task: {e}")
+            logging.exception(f"Error adding trace task, trace_type {trace_task.trace_type}")
         finally:
             self.start_timer()
 
@@ -730,7 +730,7 @@ class TraceQueueManager:
             if tasks:
                 self.send_to_celery(tasks)
         except Exception as e:
-            logging.exception(f"Error processing trace tasks: {e}")
+            logging.exception("Error processing trace tasks")
 
     def start_timer(self):
         global trace_manager_timer

api/core/ops/utils.py

@@ -1,5 +1,6 @@
 from contextlib import contextmanager
 from datetime import datetime
+from typing import Optional, Union
 
 from extensions.ext_database import db
 from models.model import Message
@@ -43,3 +44,19 @@ def replace_text_with_content(data):
         return [replace_text_with_content(item) for item in data]
     else:
         return data
+
+
+def generate_dotted_order(
+    run_id: str, start_time: Union[str, datetime], parent_dotted_order: Optional[str] = None
+) -> str:
+    """
+    generate dotted_order for langsmith
+    """
+    start_time = datetime.fromisoformat(start_time) if isinstance(start_time, str) else start_time
+    timestamp = start_time.strftime("%Y%m%dT%H%M%S%f")[:-3] + "Z"
+    current_segment = f"{timestamp}{run_id}"
+
+    if parent_dotted_order is None:
+        return current_segment
+
+    return f"{parent_dotted_order}.{current_segment}"

api/core/rag/datasource/vdb/analyticdb/analyticdb_vector.py

@@ -1,310 +1,62 @@
 import json
 from typing import Any
 
-from pydantic import BaseModel
-
-_import_err_msg = (
-    "`alibabacloud_gpdb20160503` and `alibabacloud_tea_openapi` packages not found, "
-    "please run `pip install alibabacloud_gpdb20160503 alibabacloud_tea_openapi`"
-)
-
 from configs import dify_config
+from core.rag.datasource.vdb.analyticdb.analyticdb_vector_openapi import (
+    AnalyticdbVectorOpenAPI,
+    AnalyticdbVectorOpenAPIConfig,
+)
+from core.rag.datasource.vdb.analyticdb.analyticdb_vector_sql import AnalyticdbVectorBySql, AnalyticdbVectorBySqlConfig
 from core.rag.datasource.vdb.vector_base import BaseVector
 from core.rag.datasource.vdb.vector_factory import AbstractVectorFactory
 from core.rag.datasource.vdb.vector_type import VectorType
 from core.rag.embedding.embedding_base import Embeddings
 from core.rag.models.document import Document
-from extensions.ext_redis import redis_client
 from models.dataset import Dataset
 
 
-class AnalyticdbConfig(BaseModel):
-    access_key_id: str
-    access_key_secret: str
-    region_id: str
-    instance_id: str
-    account: str
-    account_password: str
-    namespace: str = ("dify",)
-    namespace_password: str = (None,)
-    metrics: str = ("cosine",)
-    read_timeout: int = 60000
-
-    def to_analyticdb_client_params(self):
-        return {
-            "access_key_id": self.access_key_id,
-            "access_key_secret": self.access_key_secret,
-            "region_id": self.region_id,
-            "read_timeout": self.read_timeout,
-        }
-
-
 class AnalyticdbVector(BaseVector):
-    def __init__(self, collection_name: str, config: AnalyticdbConfig):
-        self._collection_name = collection_name.lower()
-        try:
-            from alibabacloud_gpdb20160503.client import Client
-            from alibabacloud_tea_openapi import models as open_api_models
-        except:
-            raise ImportError(_import_err_msg)
-        self.config = config
-        self._client_config = open_api_models.Config(user_agent="dify", **config.to_analyticdb_client_params())
-        self._client = Client(self._client_config)
-        self._initialize()
-
-    def _initialize(self) -> None:
-        cache_key = f"vector_indexing_{self.config.instance_id}"
-        lock_name = f"{cache_key}_lock"
-        with redis_client.lock(lock_name, timeout=20):
-            collection_exist_cache_key = f"vector_indexing_{self.config.instance_id}"
-            if redis_client.get(collection_exist_cache_key):
-                return
-            self._initialize_vector_database()
-            self._create_namespace_if_not_exists()
-            redis_client.set(collection_exist_cache_key, 1, ex=3600)
-
-    def _initialize_vector_database(self) -> None:
-        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
-
-        request = gpdb_20160503_models.InitVectorDatabaseRequest(
-            dbinstance_id=self.config.instance_id,
-            region_id=self.config.region_id,
-            manager_account=self.config.account,
-            manager_account_password=self.config.account_password,
-        )
-        self._client.init_vector_database(request)
-
-    def _create_namespace_if_not_exists(self) -> None:
-        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
-        from Tea.exceptions import TeaException
-
-        try:
-            request = gpdb_20160503_models.DescribeNamespaceRequest(
-                dbinstance_id=self.config.instance_id,
-                region_id=self.config.region_id,
-                namespace=self.config.namespace,
-                manager_account=self.config.account,
-                manager_account_password=self.config.account_password,
-            )
-            self._client.describe_namespace(request)
-        except TeaException as e:
-            if e.statusCode == 404:
-                request = gpdb_20160503_models.CreateNamespaceRequest(
-                    dbinstance_id=self.config.instance_id,
-                    region_id=self.config.region_id,
-                    manager_account=self.config.account,
-                    manager_account_password=self.config.account_password,
-                    namespace=self.config.namespace,
-                    namespace_password=self.config.namespace_password,
-                )
-                self._client.create_namespace(request)
-            else:
-                raise ValueError(f"failed to create namespace {self.config.namespace}: {e}")
-
-    def _create_collection_if_not_exists(self, embedding_dimension: int):
-        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
-        from Tea.exceptions import TeaException
-
-        cache_key = f"vector_indexing_{self._collection_name}"
-        lock_name = f"{cache_key}_lock"
-        with redis_client.lock(lock_name, timeout=20):
-            collection_exist_cache_key = f"vector_indexing_{self._collection_name}"
-            if redis_client.get(collection_exist_cache_key):
-                return
-            try:
-                request = gpdb_20160503_models.DescribeCollectionRequest(
-                    dbinstance_id=self.config.instance_id,
-                    region_id=self.config.region_id,
-                    namespace=self.config.namespace,
-                    namespace_password=self.config.namespace_password,
-                    collection=self._collection_name,
-                )
-                self._client.describe_collection(request)
-            except TeaException as e:
-                if e.statusCode == 404:
-                    metadata = '{"ref_doc_id":"text","page_content":"text","metadata_":"jsonb"}'
-                    full_text_retrieval_fields = "page_content"
-                    request = gpdb_20160503_models.CreateCollectionRequest(
-                        dbinstance_id=self.config.instance_id,
-                        region_id=self.config.region_id,
-                        manager_account=self.config.account,
-                        manager_account_password=self.config.account_password,
-                        namespace=self.config.namespace,
-                        collection=self._collection_name,
-                        dimension=embedding_dimension,
-                        metrics=self.config.metrics,
-                        metadata=metadata,
-                        full_text_retrieval_fields=full_text_retrieval_fields,
-                    )
-                    self._client.create_collection(request)
-                else:
-                    raise ValueError(f"failed to create collection {self._collection_name}: {e}")
-            redis_client.set(collection_exist_cache_key, 1, ex=3600)
+    def __init__(
+        self, collection_name: str, api_config: AnalyticdbVectorOpenAPIConfig, sql_config: AnalyticdbVectorBySqlConfig
+    ):
+        super().__init__(collection_name)
+        if api_config is not None:
+            self.analyticdb_vector = AnalyticdbVectorOpenAPI(collection_name, api_config)
+        else:
+            self.analyticdb_vector = AnalyticdbVectorBySql(collection_name, sql_config)
 
     def get_type(self) -> str:
         return VectorType.ANALYTICDB
 
     def create(self, texts: list[Document], embeddings: list[list[float]], **kwargs):
         dimension = len(embeddings[0])
-        self._create_collection_if_not_exists(dimension)
-        self.add_texts(texts, embeddings)
+        self.analyticdb_vector._create_collection_if_not_exists(dimension)
+        self.analyticdb_vector.add_texts(texts, embeddings)
 
-    def add_texts(self, documents: list[Document], embeddings: list[list[float]], **kwargs):
-        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
-
-        rows: list[gpdb_20160503_models.UpsertCollectionDataRequestRows] = []
-        for doc, embedding in zip(documents, embeddings, strict=True):
-            metadata = {
-                "ref_doc_id": doc.metadata["doc_id"],
-                "page_content": doc.page_content,
-                "metadata_": json.dumps(doc.metadata),
-            }
-            rows.append(
-                gpdb_20160503_models.UpsertCollectionDataRequestRows(
-                    vector=embedding,
-                    metadata=metadata,
-                )
-            )
-        request = gpdb_20160503_models.UpsertCollectionDataRequest(
-            dbinstance_id=self.config.instance_id,
-            region_id=self.config.region_id,
-            namespace=self.config.namespace,
-            namespace_password=self.config.namespace_password,
-            collection=self._collection_name,
-            rows=rows,
-        )
-        self._client.upsert_collection_data(request)
+    def add_texts(self, texts: list[Document], embeddings: list[list[float]], **kwargs):
+        self.analyticdb_vector.add_texts(texts, embeddings)
 
     def text_exists(self, id: str) -> bool:
-        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
-
-        request = gpdb_20160503_models.QueryCollectionDataRequest(
-            dbinstance_id=self.config.instance_id,
-            region_id=self.config.region_id,
-            namespace=self.config.namespace,
-            namespace_password=self.config.namespace_password,
-            collection=self._collection_name,
-            metrics=self.config.metrics,
-            include_values=True,
-            vector=None,
-            content=None,
-            top_k=1,
-            filter=f"ref_doc_id='{id}'",
-        )
-        response = self._client.query_collection_data(request)
-        return len(response.body.matches.match) > 0
+        return self.analyticdb_vector.text_exists(id)
 
     def delete_by_ids(self, ids: list[str]) -> None:
-        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
-
-        ids_str = ",".join(f"'{id}'" for id in ids)
-        ids_str = f"({ids_str})"
-        request = gpdb_20160503_models.DeleteCollectionDataRequest(
-            dbinstance_id=self.config.instance_id,
-            region_id=self.config.region_id,
-            namespace=self.config.namespace,
-            namespace_password=self.config.namespace_password,
-            collection=self._collection_name,
-            collection_data=None,
-            collection_data_filter=f"ref_doc_id IN {ids_str}",
-        )
-        self._client.delete_collection_data(request)
+        self.analyticdb_vector.delete_by_ids(ids)
 
     def delete_by_metadata_field(self, key: str, value: str) -> None:
-        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
-
-        request = gpdb_20160503_models.DeleteCollectionDataRequest(
-            dbinstance_id=self.config.instance_id,
-            region_id=self.config.region_id,
-            namespace=self.config.namespace,
-            namespace_password=self.config.namespace_password,
-            collection=self._collection_name,
-            collection_data=None,
-            collection_data_filter=f"metadata_ ->> '{key}' = '{value}'",
-        )
-        self._client.delete_collection_data(request)
+        self.analyticdb_vector.delete_by_metadata_field(key, value)
 
     def search_by_vector(self, query_vector: list[float], **kwargs: Any) -> list[Document]:
-        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
-
-        score_threshold = kwargs.get("score_threshold") or 0.0
-        request = gpdb_20160503_models.QueryCollectionDataRequest(
-            dbinstance_id=self.config.instance_id,
-            region_id=self.config.region_id,
-            namespace=self.config.namespace,
-            namespace_password=self.config.namespace_password,
-            collection=self._collection_name,
-            include_values=kwargs.pop("include_values", True),
-            metrics=self.config.metrics,
-            vector=query_vector,
-            content=None,
-            top_k=kwargs.get("top_k", 4),
-            filter=None,
-        )
-        response = self._client.query_collection_data(request)
-        documents = []
-        for match in response.body.matches.match:
-            if match.score > score_threshold:
-                metadata = json.loads(match.metadata.get("metadata_"))
-                metadata["score"] = match.score
-                doc = Document(
-                    page_content=match.metadata.get("page_content"),
-                    metadata=metadata,
-                )
-                documents.append(doc)
-        documents = sorted(documents, key=lambda x: x.metadata["score"], reverse=True)
-        return documents
+        return self.analyticdb_vector.search_by_vector(query_vector)
 
     def search_by_full_text(self, query: str, **kwargs: Any) -> list[Document]:
-        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
-
-        score_threshold = float(kwargs.get("score_threshold") or 0.0)
-        request = gpdb_20160503_models.QueryCollectionDataRequest(
-            dbinstance_id=self.config.instance_id,
-            region_id=self.config.region_id,
-            namespace=self.config.namespace,
-            namespace_password=self.config.namespace_password,
-            collection=self._collection_name,
-            include_values=kwargs.pop("include_values", True),
-            metrics=self.config.metrics,
-            vector=None,
-            content=query,
-            top_k=kwargs.get("top_k", 4),
-            filter=None,
-        )
-        response = self._client.query_collection_data(request)
-        documents = []
-        for match in response.body.matches.match:
-            if match.score > score_threshold:
-                metadata = json.loads(match.metadata.get("metadata_"))
-                metadata["score"] = match.score
-                doc = Document(
-                    page_content=match.metadata.get("page_content"),
-                    vector=match.metadata.get("vector"),
-                    metadata=metadata,
-                )
-                documents.append(doc)
-        documents = sorted(documents, key=lambda x: x.metadata["score"], reverse=True)
-        return documents
+        return self.analyticdb_vector.search_by_full_text(query, **kwargs)
 
     def delete(self) -> None:
-        try:
-            from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
-
-            request = gpdb_20160503_models.DeleteCollectionRequest(
-                collection=self._collection_name,
-                dbinstance_id=self.config.instance_id,
-                namespace=self.config.namespace,
-                namespace_password=self.config.namespace_password,
-                region_id=self.config.region_id,
-            )
-            self._client.delete_collection(request)
-        except Exception as e:
-            raise e
+        self.analyticdb_vector.delete()
 
 
 class AnalyticdbVectorFactory(AbstractVectorFactory):
-    def init_vector(self, dataset: Dataset, attributes: list, embeddings: Embeddings):
+    def init_vector(self, dataset: Dataset, attributes: list, embeddings: Embeddings) -> AnalyticdbVector:
         if dataset.index_struct_dict:
             class_prefix: str = dataset.index_struct_dict["vector_store"]["class_prefix"]
             collection_name = class_prefix.lower()
@@ -313,26 +65,9 @@ class AnalyticdbVectorFactory(AbstractVectorFactory):
             collection_name = Dataset.gen_collection_name_by_id(dataset_id).lower()
             dataset.index_struct = json.dumps(self.gen_index_struct_dict(VectorType.ANALYTICDB, collection_name))
 
-        # handle optional params
-        if dify_config.ANALYTICDB_KEY_ID is None:
-            raise ValueError("ANALYTICDB_KEY_ID should not be None")
-        if dify_config.ANALYTICDB_KEY_SECRET is None:
-            raise ValueError("ANALYTICDB_KEY_SECRET should not be None")
-        if dify_config.ANALYTICDB_REGION_ID is None:
-            raise ValueError("ANALYTICDB_REGION_ID should not be None")
-        if dify_config.ANALYTICDB_INSTANCE_ID is None:
-            raise ValueError("ANALYTICDB_INSTANCE_ID should not be None")
-        if dify_config.ANALYTICDB_ACCOUNT is None:
-            raise ValueError("ANALYTICDB_ACCOUNT should not be None")
-        if dify_config.ANALYTICDB_PASSWORD is None:
-            raise ValueError("ANALYTICDB_PASSWORD should not be None")
-        if dify_config.ANALYTICDB_NAMESPACE is None:
-            raise ValueError("ANALYTICDB_NAMESPACE should not be None")
-        if dify_config.ANALYTICDB_NAMESPACE_PASSWORD is None:
-            raise ValueError("ANALYTICDB_NAMESPACE_PASSWORD should not be None")
-        return AnalyticdbVector(
-            collection_name,
-            AnalyticdbConfig(
+        if dify_config.ANALYTICDB_HOST is None:
+            # implemented through OpenAPI
+            apiConfig = AnalyticdbVectorOpenAPIConfig(
                 access_key_id=dify_config.ANALYTICDB_KEY_ID,
                 access_key_secret=dify_config.ANALYTICDB_KEY_SECRET,
                 region_id=dify_config.ANALYTICDB_REGION_ID,
@@ -341,5 +76,22 @@ class AnalyticdbVectorFactory(AbstractVectorFactory):
                 account_password=dify_config.ANALYTICDB_PASSWORD,
                 namespace=dify_config.ANALYTICDB_NAMESPACE,
                 namespace_password=dify_config.ANALYTICDB_NAMESPACE_PASSWORD,
-            ),
+            )
+            sqlConfig = None
+        else:
+            # implemented through sql
+            sqlConfig = AnalyticdbVectorBySqlConfig(
+                host=dify_config.ANALYTICDB_HOST,
+                port=dify_config.ANALYTICDB_PORT,
+                account=dify_config.ANALYTICDB_ACCOUNT,
+                account_password=dify_config.ANALYTICDB_PASSWORD,
+                min_connection=dify_config.ANALYTICDB_MIN_CONNECTION,
+                max_connection=dify_config.ANALYTICDB_MAX_CONNECTION,
+                namespace=dify_config.ANALYTICDB_NAMESPACE,
+            )
+            apiConfig = None
+        return AnalyticdbVector(
+            collection_name,
+            apiConfig,
+            sqlConfig,
         )

api/core/rag/datasource/vdb/analyticdb/analyticdb_vector_openapi.py

@@ -0,0 +1,309 @@
+import json
+from typing import Any
+
+from pydantic import BaseModel, model_validator
+
+_import_err_msg = (
+    "`alibabacloud_gpdb20160503` and `alibabacloud_tea_openapi` packages not found, "
+    "please run `pip install alibabacloud_gpdb20160503 alibabacloud_tea_openapi`"
+)
+
+from core.rag.models.document import Document
+from extensions.ext_redis import redis_client
+
+
+class AnalyticdbVectorOpenAPIConfig(BaseModel):
+    access_key_id: str
+    access_key_secret: str
+    region_id: str
+    instance_id: str
+    account: str
+    account_password: str
+    namespace: str = "dify"
+    namespace_password: str = (None,)
+    metrics: str = "cosine"
+    read_timeout: int = 60000
+
+    @model_validator(mode="before")
+    @classmethod
+    def validate_config(cls, values: dict) -> dict:
+        if not values["access_key_id"]:
+            raise ValueError("config ANALYTICDB_KEY_ID is required")
+        if not values["access_key_secret"]:
+            raise ValueError("config ANALYTICDB_KEY_SECRET is required")
+        if not values["region_id"]:
+            raise ValueError("config ANALYTICDB_REGION_ID is required")
+        if not values["instance_id"]:
+            raise ValueError("config ANALYTICDB_INSTANCE_ID is required")
+        if not values["account"]:
+            raise ValueError("config ANALYTICDB_ACCOUNT is required")
+        if not values["account_password"]:
+            raise ValueError("config ANALYTICDB_PASSWORD is required")
+        if not values["namespace_password"]:
+            raise ValueError("config ANALYTICDB_NAMESPACE_PASSWORD is required")
+        return values
+
+    def to_analyticdb_client_params(self):
+        return {
+            "access_key_id": self.access_key_id,
+            "access_key_secret": self.access_key_secret,
+            "region_id": self.region_id,
+            "read_timeout": self.read_timeout,
+        }
+
+
+class AnalyticdbVectorOpenAPI:
+    def __init__(self, collection_name: str, config: AnalyticdbVectorOpenAPIConfig):
+        try:
+            from alibabacloud_gpdb20160503.client import Client
+            from alibabacloud_tea_openapi import models as open_api_models
+        except:
+            raise ImportError(_import_err_msg)
+        self._collection_name = collection_name.lower()
+        self.config = config
+        self._client_config = open_api_models.Config(user_agent="dify", **config.to_analyticdb_client_params())
+        self._client = Client(self._client_config)
+        self._initialize()
+
+    def _initialize(self) -> None:
+        cache_key = f"vector_initialize_{self.config.instance_id}"
+        lock_name = f"{cache_key}_lock"
+        with redis_client.lock(lock_name, timeout=20):
+            database_exist_cache_key = f"vector_initialize_{self.config.instance_id}"
+            if redis_client.get(database_exist_cache_key):
+                return
+            self._initialize_vector_database()
+            self._create_namespace_if_not_exists()
+            redis_client.set(database_exist_cache_key, 1, ex=3600)
+
+    def _initialize_vector_database(self) -> None:
+        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
+
+        request = gpdb_20160503_models.InitVectorDatabaseRequest(
+            dbinstance_id=self.config.instance_id,
+            region_id=self.config.region_id,
+            manager_account=self.config.account,
+            manager_account_password=self.config.account_password,
+        )
+        self._client.init_vector_database(request)
+
+    def _create_namespace_if_not_exists(self) -> None:
+        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
+        from Tea.exceptions import TeaException
+
+        try:
+            request = gpdb_20160503_models.DescribeNamespaceRequest(
+                dbinstance_id=self.config.instance_id,
+                region_id=self.config.region_id,
+                namespace=self.config.namespace,
+                manager_account=self.config.account,
+                manager_account_password=self.config.account_password,
+            )
+            self._client.describe_namespace(request)
+        except TeaException as e:
+            if e.statusCode == 404:
+                request = gpdb_20160503_models.CreateNamespaceRequest(
+                    dbinstance_id=self.config.instance_id,
+                    region_id=self.config.region_id,
+                    manager_account=self.config.account,
+                    manager_account_password=self.config.account_password,
+                    namespace=self.config.namespace,
+                    namespace_password=self.config.namespace_password,
+                )
+                self._client.create_namespace(request)
+            else:
+                raise ValueError(f"failed to create namespace {self.config.namespace}: {e}")
+
+    def _create_collection_if_not_exists(self, embedding_dimension: int):
+        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
+        from Tea.exceptions import TeaException
+
+        cache_key = f"vector_indexing_{self._collection_name}"
+        lock_name = f"{cache_key}_lock"
+        with redis_client.lock(lock_name, timeout=20):
+            collection_exist_cache_key = f"vector_indexing_{self._collection_name}"
+            if redis_client.get(collection_exist_cache_key):
+                return
+            try:
+                request = gpdb_20160503_models.DescribeCollectionRequest(
+                    dbinstance_id=self.config.instance_id,
+                    region_id=self.config.region_id,
+                    namespace=self.config.namespace,
+                    namespace_password=self.config.namespace_password,
+                    collection=self._collection_name,
+                )
+                self._client.describe_collection(request)
+            except TeaException as e:
+                if e.statusCode == 404:
+                    metadata = '{"ref_doc_id":"text","page_content":"text","metadata_":"jsonb"}'
+                    full_text_retrieval_fields = "page_content"
+                    request = gpdb_20160503_models.CreateCollectionRequest(
+                        dbinstance_id=self.config.instance_id,
+                        region_id=self.config.region_id,
+                        manager_account=self.config.account,
+                        manager_account_password=self.config.account_password,
+                        namespace=self.config.namespace,
+                        collection=self._collection_name,
+                        dimension=embedding_dimension,
+                        metrics=self.config.metrics,
+                        metadata=metadata,
+                        full_text_retrieval_fields=full_text_retrieval_fields,
+                    )
+                    self._client.create_collection(request)
+                else:
+                    raise ValueError(f"failed to create collection {self._collection_name}: {e}")
+            redis_client.set(collection_exist_cache_key, 1, ex=3600)
+
+    def add_texts(self, documents: list[Document], embeddings: list[list[float]], **kwargs):
+        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
+
+        rows: list[gpdb_20160503_models.UpsertCollectionDataRequestRows] = []
+        for doc, embedding in zip(documents, embeddings, strict=True):
+            metadata = {
+                "ref_doc_id": doc.metadata["doc_id"],
+                "page_content": doc.page_content,
+                "metadata_": json.dumps(doc.metadata),
+            }
+            rows.append(
+                gpdb_20160503_models.UpsertCollectionDataRequestRows(
+                    vector=embedding,
+                    metadata=metadata,
+                )
+            )
+        request = gpdb_20160503_models.UpsertCollectionDataRequest(
+            dbinstance_id=self.config.instance_id,
+            region_id=self.config.region_id,
+            namespace=self.config.namespace,
+            namespace_password=self.config.namespace_password,
+            collection=self._collection_name,
+            rows=rows,
+        )
+        self._client.upsert_collection_data(request)
+
+    def text_exists(self, id: str) -> bool:
+        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
+
+        request = gpdb_20160503_models.QueryCollectionDataRequest(
+            dbinstance_id=self.config.instance_id,
+            region_id=self.config.region_id,
+            namespace=self.config.namespace,
+            namespace_password=self.config.namespace_password,
+            collection=self._collection_name,
+            metrics=self.config.metrics,
+            include_values=True,
+            vector=None,
+            content=None,
+            top_k=1,
+            filter=f"ref_doc_id='{id}'",
+        )
+        response = self._client.query_collection_data(request)
+        return len(response.body.matches.match) > 0
+
+    def delete_by_ids(self, ids: list[str]) -> None:
+        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
+
+        ids_str = ",".join(f"'{id}'" for id in ids)
+        ids_str = f"({ids_str})"
+        request = gpdb_20160503_models.DeleteCollectionDataRequest(
+            dbinstance_id=self.config.instance_id,
+            region_id=self.config.region_id,
+            namespace=self.config.namespace,
+            namespace_password=self.config.namespace_password,
+            collection=self._collection_name,
+            collection_data=None,
+            collection_data_filter=f"ref_doc_id IN {ids_str}",
+        )
+        self._client.delete_collection_data(request)
+
+    def delete_by_metadata_field(self, key: str, value: str) -> None:
+        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
+
+        request = gpdb_20160503_models.DeleteCollectionDataRequest(
+            dbinstance_id=self.config.instance_id,
+            region_id=self.config.region_id,
+            namespace=self.config.namespace,
+            namespace_password=self.config.namespace_password,
+            collection=self._collection_name,
+            collection_data=None,
+            collection_data_filter=f"metadata_ ->> '{key}' = '{value}'",
+        )
+        self._client.delete_collection_data(request)
+
+    def search_by_vector(self, query_vector: list[float], **kwargs: Any) -> list[Document]:
+        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
+
+        score_threshold = kwargs.get("score_threshold") or 0.0
+        request = gpdb_20160503_models.QueryCollectionDataRequest(
+            dbinstance_id=self.config.instance_id,
+            region_id=self.config.region_id,
+            namespace=self.config.namespace,
+            namespace_password=self.config.namespace_password,
+            collection=self._collection_name,
+            include_values=kwargs.pop("include_values", True),
+            metrics=self.config.metrics,
+            vector=query_vector,
+            content=None,
+            top_k=kwargs.get("top_k", 4),
+            filter=None,
+        )
+        response = self._client.query_collection_data(request)
+        documents = []
+        for match in response.body.matches.match:
+            if match.score > score_threshold:
+                metadata = json.loads(match.metadata.get("metadata_"))
+                metadata["score"] = match.score
+                doc = Document(
+                    page_content=match.metadata.get("page_content"),
+                    vector=match.values.value,
+                    metadata=metadata,
+                )
+                documents.append(doc)
+        documents = sorted(documents, key=lambda x: x.metadata["score"], reverse=True)
+        return documents
+
+    def search_by_full_text(self, query: str, **kwargs: Any) -> list[Document]:
+        from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
+
+        score_threshold = float(kwargs.get("score_threshold") or 0.0)
+        request = gpdb_20160503_models.QueryCollectionDataRequest(
+            dbinstance_id=self.config.instance_id,
+            region_id=self.config.region_id,
+            namespace=self.config.namespace,
+            namespace_password=self.config.namespace_password,
+            collection=self._collection_name,
+            include_values=kwargs.pop("include_values", True),
+            metrics=self.config.metrics,
+            vector=None,
+            content=query,
+            top_k=kwargs.get("top_k", 4),
+            filter=None,
+        )
+        response = self._client.query_collection_data(request)
+        documents = []
+        for match in response.body.matches.match:
+            if match.score > score_threshold:
+                metadata = json.loads(match.metadata.get("metadata_"))
+                metadata["score"] = match.score
+                doc = Document(
+                    page_content=match.metadata.get("page_content"),
+                    vector=match.values.value,
+                    metadata=metadata,
+                )
+                documents.append(doc)
+        documents = sorted(documents, key=lambda x: x.metadata["score"], reverse=True)
+        return documents
+
+    def delete(self) -> None:
+        try:
+            from alibabacloud_gpdb20160503 import models as gpdb_20160503_models
+
+            request = gpdb_20160503_models.DeleteCollectionRequest(
+                collection=self._collection_name,
+                dbinstance_id=self.config.instance_id,
+                namespace=self.config.namespace,
+                namespace_password=self.config.namespace_password,
+                region_id=self.config.region_id,
+            )
+            self._client.delete_collection(request)
+        except Exception as e:
+            raise e

api/core/rag/datasource/vdb/analyticdb/analyticdb_vector_sql.py

@@ -0,0 +1,245 @@
+import json
+import uuid
+from contextlib import contextmanager
+from typing import Any
+
+import psycopg2.extras
+import psycopg2.pool
+from pydantic import BaseModel, model_validator
+
+from core.rag.models.document import Document
+from extensions.ext_redis import redis_client
+
+
+class AnalyticdbVectorBySqlConfig(BaseModel):
+    host: str
+    port: int
+    account: str
+    account_password: str
+    min_connection: int
+    max_connection: int
+    namespace: str = "dify"
+    metrics: str = "cosine"
+
+    @model_validator(mode="before")
+    @classmethod
+    def validate_config(cls, values: dict) -> dict:
+        if not values["host"]:
+            raise ValueError("config ANALYTICDB_HOST is required")
+        if not values["port"]:
+            raise ValueError("config ANALYTICDB_PORT is required")
+        if not values["account"]:
+            raise ValueError("config ANALYTICDB_ACCOUNT is required")
+        if not values["account_password"]:
+            raise ValueError("config ANALYTICDB_PASSWORD is required")
+        if not values["min_connection"]:
+            raise ValueError("config ANALYTICDB_MIN_CONNECTION is required")
+        if not values["max_connection"]:
+            raise ValueError("config ANALYTICDB_MAX_CONNECTION is required")
+        if values["min_connection"] > values["max_connection"]:
+            raise ValueError("config ANALYTICDB_MIN_CONNECTION should less than ANALYTICDB_MAX_CONNECTION")
+        return values
+
+
+class AnalyticdbVectorBySql:
+    def __init__(self, collection_name: str, config: AnalyticdbVectorBySqlConfig):
+        self._collection_name = collection_name.lower()
+        self.databaseName = "knowledgebase"
+        self.config = config
+        self.table_name = f"{self.config.namespace}.{self._collection_name}"
+        self.pool = None
+        self._initialize()
+        if not self.pool:
+            self.pool = self._create_connection_pool()
+
+    def _initialize(self) -> None:
+        cache_key = f"vector_initialize_{self.config.host}"
+        lock_name = f"{cache_key}_lock"
+        with redis_client.lock(lock_name, timeout=20):
+            database_exist_cache_key = f"vector_initialize_{self.config.host}"
+            if redis_client.get(database_exist_cache_key):
+                return
+            self._initialize_vector_database()
+            redis_client.set(database_exist_cache_key, 1, ex=3600)
+
+    def _create_connection_pool(self):
+        return psycopg2.pool.SimpleConnectionPool(
+            self.config.min_connection,
+            self.config.max_connection,
+            host=self.config.host,
+            port=self.config.port,
+            user=self.config.account,
+            password=self.config.account_password,
+            database=self.databaseName,
+        )
+
+    @contextmanager
+    def _get_cursor(self):
+        conn = self.pool.getconn()
+        cur = conn.cursor()
+        try:
+            yield cur
+        finally:
+            cur.close()
+            conn.commit()
+            self.pool.putconn(conn)
+
+    def _initialize_vector_database(self) -> None:
+        conn = psycopg2.connect(
+            host=self.config.host,
+            port=self.config.port,
+            user=self.config.account,
+            password=self.config.account_password,
+            database="postgres",
+        )
+        conn.autocommit = True
+        cur = conn.cursor()
+        try:
+            cur.execute(f"CREATE DATABASE {self.databaseName}")
+        except Exception as e:
+            if "already exists" in str(e):
+                return
+            raise e
+        finally:
+            cur.close()
+            conn.close()
+        self.pool = self._create_connection_pool()
+        with self._get_cursor() as cur:
+            try:
+                cur.execute("CREATE TEXT SEARCH CONFIGURATION zh_cn (PARSER = zhparser)")
+                cur.execute("ALTER TEXT SEARCH CONFIGURATION zh_cn ADD MAPPING FOR n,v,a,i,e,l,x WITH simple")
+            except Exception as e:
+                if "already exists" not in str(e):
+                    raise e
+            cur.execute(
+                "CREATE OR REPLACE FUNCTION "
+                "public.to_tsquery_from_text(txt text, lang regconfig DEFAULT 'english'::regconfig) "
+                "RETURNS tsquery LANGUAGE sql IMMUTABLE STRICT AS $function$ "
+                "SELECT to_tsquery(lang, COALESCE(string_agg(split_part(word, ':', 1), ' | '), '')) "
+                "FROM (SELECT unnest(string_to_array(to_tsvector(lang, txt)::text, ' ')) AS word) "
+                "AS words_only;$function$"
+            )
+            cur.execute(f"CREATE SCHEMA IF NOT EXISTS {self.config.namespace}")
+
+    def _create_collection_if_not_exists(self, embedding_dimension: int):
+        cache_key = f"vector_indexing_{self._collection_name}"
+        lock_name = f"{cache_key}_lock"
+        with redis_client.lock(lock_name, timeout=20):
+            collection_exist_cache_key = f"vector_indexing_{self._collection_name}"
+            if redis_client.get(collection_exist_cache_key):
+                return
+            with self._get_cursor() as cur:
+                cur.execute(
+                    f"CREATE TABLE IF NOT EXISTS {self.table_name}("
+                    f"id text PRIMARY KEY,"
+                    f"vector real[], ref_doc_id text, page_content text, metadata_ jsonb, "
+                    f"to_tsvector TSVECTOR"
+                    f") WITH (fillfactor=70) DISTRIBUTED BY (id);"
+                )
+                if embedding_dimension is not None:
+                    index_name = f"{self._collection_name}_embedding_idx"
+                    cur.execute(f"ALTER TABLE {self.table_name} ALTER COLUMN vector SET STORAGE PLAIN")
+                    cur.execute(
+                        f"CREATE INDEX {index_name} ON {self.table_name} USING ann(vector) "
+                        f"WITH(dim='{embedding_dimension}', distancemeasure='{self.config.metrics}', "
+                        f"pq_enable=0, external_storage=0)"
+                    )
+                    cur.execute(f"CREATE INDEX ON {self.table_name} USING gin(to_tsvector)")
+            redis_client.set(collection_exist_cache_key, 1, ex=3600)
+
+    def add_texts(self, documents: list[Document], embeddings: list[list[float]], **kwargs):
+        values = []
+        id_prefix = str(uuid.uuid4()) + "_"
+        sql = f"""
+                INSERT INTO {self.table_name} 
+                (id, ref_doc_id, vector, page_content, metadata_, to_tsvector) 
+                VALUES (%s, %s, %s, %s, %s, to_tsvector('zh_cn',  %s));
+            """
+        for i, doc in enumerate(documents):
+            values.append(
+                (
+                    id_prefix + str(i),
+                    doc.metadata.get("doc_id", str(uuid.uuid4())),
+                    embeddings[i],
+                    doc.page_content,
+                    json.dumps(doc.metadata),
+                    doc.page_content,
+                )
+            )
+        with self._get_cursor() as cur:
+            psycopg2.extras.execute_batch(cur, sql, values)
+
+    def text_exists(self, id: str) -> bool:
+        with self._get_cursor() as cur:
+            cur.execute(f"SELECT id FROM {self.table_name} WHERE ref_doc_id = %s", (id,))
+            return cur.fetchone() is not None
+
+    def delete_by_ids(self, ids: list[str]) -> None:
+        with self._get_cursor() as cur:
+            try:
+                cur.execute(f"DELETE FROM {self.table_name} WHERE ref_doc_id IN %s", (tuple(ids),))
+            except Exception as e:
+                if "does not exist" not in str(e):
+                    raise e
+
+    def delete_by_metadata_field(self, key: str, value: str) -> None:
+        with self._get_cursor() as cur:
+            try:
+                cur.execute(f"DELETE FROM {self.table_name} WHERE metadata_->>%s = %s", (key, value))
+            except Exception as e:
+                if "does not exist" not in str(e):
+                    raise e
+
+    def search_by_vector(self, query_vector: list[float], **kwargs: Any) -> list[Document]:
+        top_k = kwargs.get("top_k", 4)
+        score_threshold = float(kwargs.get("score_threshold") or 0.0)
+        with self._get_cursor() as cur:
+            query_vector_str = json.dumps(query_vector)
+            query_vector_str = "{" + query_vector_str[1:-1] + "}"
+            cur.execute(
+                f"SELECT t.id AS id, t.vector AS vector, (1.0 - t.score) AS score, "
+                f"t.page_content as page_content, t.metadata_ AS metadata_ "
+                f"FROM (SELECT id, vector, page_content, metadata_, vector <=> %s AS score "
+                f"FROM {self.table_name} ORDER BY score LIMIT {top_k} ) t",
+                (query_vector_str,),
+            )
+            documents = []
+            for record in cur:
+                id, vector, score, page_content, metadata = record
+                if score > score_threshold:
+                    metadata["score"] = score
+                    doc = Document(
+                        page_content=page_content,
+                        vector=vector,
+                        metadata=metadata,
+                    )
+                    documents.append(doc)
+        return documents
+
+    def search_by_full_text(self, query: str, **kwargs: Any) -> list[Document]:
+        top_k = kwargs.get("top_k", 4)
+        with self._get_cursor() as cur:
+            cur.execute(
+                f"""SELECT id, vector, page_content, metadata_, 
+                ts_rank(to_tsvector, to_tsquery_from_text(%s, 'zh_cn'), 32) AS score
+                FROM {self.table_name}
+                WHERE to_tsvector@@to_tsquery_from_text(%s, 'zh_cn')
+                ORDER BY score DESC
+                LIMIT {top_k}""",
+                (f"'{query}'", f"'{query}'"),
+            )
+            documents = []
+            for record in cur:
+                id, vector, page_content, metadata, score = record
+                metadata["score"] = score
+                doc = Document(
+                    page_content=page_content,
+                    vector=vector,
+                    metadata=metadata,
+                )
+                documents.append(doc)
+        return documents
+
+    def delete(self) -> None:
+        with self._get_cursor() as cur:
+            cur.execute(f"DROP TABLE IF EXISTS {self.table_name}")

api/core/rag/datasource/vdb/couchbase/couchbase_vector.py

@@ -242,7 +242,7 @@ class CouchbaseVector(BaseVector):
         try:
             self._cluster.query(query, named_parameters={"doc_ids": ids}).execute()
         except Exception as e:
-            logger.exception(e)
+            logger.exception(f"Failed to delete documents, ids: {ids}")
 
     def delete_by_document_id(self, document_id: str):
         query = f"""

api/core/rag/datasource/vdb/lindorm/lindorm_vector.py

@@ -79,7 +79,7 @@ class LindormVectorStore(BaseVector):
                 existing_docs = self._client.mget(index=self._collection_name, body={"ids": batch_ids}, _source=False)
                 return {doc["_id"] for doc in existing_docs["docs"] if doc["found"]}
             except Exception as e:
-                logger.exception(f"Error fetching batch {batch_ids}: {e}")
+                logger.exception(f"Error fetching batch {batch_ids}")
                 return set()
 
         @retry(stop=stop_after_attempt(3), wait=wait_fixed(60))
@@ -96,7 +96,7 @@ class LindormVectorStore(BaseVector):
                 )
                 return {doc["_id"] for doc in existing_docs["docs"] if doc["found"]}
             except Exception as e:
-                logger.exception(f"Error fetching batch {batch_ids}: {e}")
+                logger.exception(f"Error fetching batch ids: {batch_ids}")
                 return set()
 
         if ids is None:
@@ -177,7 +177,7 @@ class LindormVectorStore(BaseVector):
             else:
                 logger.warning(f"Index '{self._collection_name}' does not exist. No deletion performed.")
         except Exception as e:
-            logger.exception(f"Error occurred while deleting the index: {e}")
+            logger.exception(f"Error occurred while deleting the index: {self._collection_name}")
             raise e
 
     def text_exists(self, id: str) -> bool:
@@ -201,7 +201,7 @@ class LindormVectorStore(BaseVector):
         try:
             response = self._client.search(index=self._collection_name, body=query)
         except Exception as e:
-            logger.exception(f"Error executing search: {e}")
+            logger.exception(f"Error executing vector search, query: {query}")
             raise
 
         docs_and_scores = []

api/core/rag/datasource/vdb/myscale/myscale_vector.py

@@ -142,7 +142,7 @@ class MyScaleVector(BaseVector):
                 for r in self._client.query(sql).named_results()
             ]
         except Exception as e:
-            logging.exception(f"\033[91m\033[1m{type(e)}\033[0m \033[95m{str(e)}\033[0m")
+            logging.exception(f"\033[91m\033[1m{type(e)}\033[0m \033[95m{str(e)}\033[0m")  # noqa:TRY401
             return []
 
     def delete(self) -> None:

api/core/rag/datasource/vdb/opensearch/opensearch_vector.py

@@ -158,7 +158,7 @@ class OpenSearchVector(BaseVector):
         try:
             response = self._client.search(index=self._collection_name.lower(), body=query)
         except Exception as e:
-            logger.exception(f"Error executing search: {e}")
+            logger.exception(f"Error executing vector search, query: {query}")
             raise
 
         docs = []

api/core/rag/embedding/cached_embedding.py

@@ -69,7 +69,7 @@ class CacheEmbedding(Embeddings):
                         except IntegrityError:
                             db.session.rollback()
                         except Exception as e:
-                            logging.exception("Failed transform embedding: %s", e)
+                            logging.exception("Failed transform embedding")
                 cache_embeddings = []
                 try:
                     for i, embedding in zip(embedding_queue_indices, embedding_queue_embeddings):
@@ -89,7 +89,7 @@ class CacheEmbedding(Embeddings):
                     db.session.rollback()
             except Exception as ex:
                 db.session.rollback()
-                logger.exception("Failed to embed documents: %s", ex)
+                logger.exception("Failed to embed documents: %s")
                 raise ex
 
         return text_embeddings
@@ -112,7 +112,7 @@ class CacheEmbedding(Embeddings):
             embedding_results = (embedding_results / np.linalg.norm(embedding_results)).tolist()
         except Exception as ex:
             if dify_config.DEBUG:
-                logging.exception(f"Failed to embed query text: {ex}")
+                logging.exception(f"Failed to embed query text '{text[:10]}...({len(text)} chars)'")
             raise ex
 
         try:
@@ -126,7 +126,7 @@ class CacheEmbedding(Embeddings):
             redis_client.setex(embedding_cache_key, 600, encoded_str)
         except Exception as ex:
             if dify_config.DEBUG:
-                logging.exception("Failed to add embedding to redis %s", ex)
+                logging.exception(f"Failed to add embedding to redis for the text '{text[:10]}...({len(text)} chars)'")
             raise ex
 
         return embedding_results

api/core/rag/extractor/word_extractor.py

@@ -229,7 +229,7 @@ class WordExtractor(BaseExtractor):
                                 for i in url_pattern.findall(x.text):
                                     hyperlinks_url = str(i)
                     except Exception as e:
-                        logger.exception(e)
+                        logger.exception("Failed to parse HYPERLINK xml")
 
         def parse_paragraph(paragraph):
             paragraph_content = []

api/core/rag/index_processor/processor/paragraph_index_processor.py

@@ -11,6 +11,7 @@ from core.rag.extractor.entity.extract_setting import ExtractSetting
 from core.rag.extractor.extract_processor import ExtractProcessor
 from core.rag.index_processor.index_processor_base import BaseIndexProcessor
 from core.rag.models.document import Document
+from core.tools.utils.text_processing_utils import remove_leading_symbols
 from libs import helper
 from models.dataset import Dataset
 
@@ -43,11 +44,7 @@ class ParagraphIndexProcessor(BaseIndexProcessor):
                     document_node.metadata["doc_id"] = doc_id
                     document_node.metadata["doc_hash"] = hash
                     # delete Splitter character
-                    page_content = document_node.page_content
-                    if page_content.startswith(".") or page_content.startswith("。"):
-                        page_content = page_content[1:].strip()
-                    else:
-                        page_content = page_content
+                    page_content = remove_leading_symbols(document_node.page_content).strip()
                     if len(page_content) > 0:
                         document_node.page_content = page_content
                         split_documents.append(document_node)

api/core/rag/index_processor/processor/qa_index_processor.py

@@ -18,6 +18,7 @@ from core.rag.extractor.entity.extract_setting import ExtractSetting
 from core.rag.extractor.extract_processor import ExtractProcessor
 from core.rag.index_processor.index_processor_base import BaseIndexProcessor
 from core.rag.models.document import Document
+from core.tools.utils.text_processing_utils import remove_leading_symbols
 from libs import helper
 from models.dataset import Dataset
 
@@ -53,11 +54,7 @@ class QAIndexProcessor(BaseIndexProcessor):
                     document_node.metadata["doc_hash"] = hash
                     # delete Splitter character
                     page_content = document_node.page_content
-                    if page_content.startswith(".") or page_content.startswith("。"):
-                        page_content = page_content[1:]
-                    else:
-                        page_content = page_content
-                    document_node.page_content = page_content
+                    document_node.page_content = remove_leading_symbols(page_content)
                     split_documents.append(document_node)
             all_documents.extend(split_documents)
         for i in range(0, len(all_documents), 10):
@@ -159,7 +156,7 @@ class QAIndexProcessor(BaseIndexProcessor):
                     qa_documents.append(qa_document)
                 format_documents.extend(qa_documents)
             except Exception as e:
-                logging.exception(e)
+                logging.exception("Failed to format qa document")
 
             all_qa_documents.extend(format_documents)
 

api/core/rag/rerank/weight_rerank.py

@@ -36,23 +36,21 @@ class WeightRerankRunner(BaseRerankRunner):
 
         :return:
         """
-        docs = []
-        doc_id = []
         unique_documents = []
+        doc_id = set()
         for document in documents:
-            if document.metadata["doc_id"] not in doc_id:
-                doc_id.append(document.metadata["doc_id"])
-                docs.append(document.page_content)
+            doc_id = document.metadata.get("doc_id")
+            if doc_id not in doc_id:
+                doc_id.add(doc_id)
                 unique_documents.append(document)
 
         documents = unique_documents
 
-        rerank_documents = []
         query_scores = self._calculate_keyword_score(query, documents)
-
         query_vector_scores = self._calculate_cosine(self.tenant_id, query, documents, self.weights.vector_setting)
+
+        rerank_documents = []
         for document, query_score, query_vector_score in zip(documents, query_scores, query_vector_scores):
-            # format document
             score = (
                 self.weights.vector_setting.vector_weight * query_vector_score
                 + self.weights.keyword_setting.keyword_weight * query_score
@@ -61,7 +59,8 @@ class WeightRerankRunner(BaseRerankRunner):
                 continue
             document.metadata["score"] = score
             rerank_documents.append(document)
-        rerank_documents = sorted(rerank_documents, key=lambda x: x.metadata["score"], reverse=True)
+
+        rerank_documents.sort(key=lambda x: x.metadata["score"], reverse=True)
         return rerank_documents[:top_n] if top_n else rerank_documents
 
     def _calculate_keyword_score(self, query: str, documents: list[Document]) -> list[float]:

api/core/tools/custom_tool/tool.py

@@ -6,6 +6,7 @@ from urllib.parse import urlencode
 
 import httpx
 
+from core.file.file_manager import download
 from core.helper import ssrf_proxy
 from core.tools.__base.tool import Tool
 from core.tools.__base.tool_runtime import ToolRuntime
@@ -145,6 +146,7 @@ class ApiTool(Tool):
         path_params = {}
         body = {}
         cookies = {}
+        files = []
 
         # check parameters
         for parameter in self.api_bundle.openapi.get("parameters", []):
@@ -173,8 +175,12 @@ class ApiTool(Tool):
                     properties = body_schema.get("properties", {})
                     for name, property in properties.items():
                         if name in parameters:
-                            # convert type
-                            body[name] = self._convert_body_property_type(property, parameters[name])
+                            if property.get("format") == "binary":
+                                f = parameters[name]
+                                files.append((name, (f.filename, download(f), f.mime_type)))
+                            else:
+                                # convert type
+                                body[name] = self._convert_body_property_type(property, parameters[name])
                         elif name in required:
                             raise ToolParameterValidationError(
                                 f"Missing required parameter {name} in operation {self.api_bundle.operation_id}"
@@ -189,7 +195,7 @@ class ApiTool(Tool):
         for name, value in path_params.items():
             url = url.replace(f"{{{name}}}", f"{value}")
 
-        # parse http body data if needed, for GET/HEAD/OPTIONS/TRACE, the body is ignored
+        # parse http body data if needed
         if "Content-Type" in headers:
             if headers["Content-Type"] == "application/json":
                 body = json.dumps(body)
@@ -205,6 +211,7 @@ class ApiTool(Tool):
                 headers=headers,
                 cookies=cookies,
                 data=body,
+                files=files,
                 timeout=API_TOOL_DEFAULT_TIMEOUT,
                 follow_redirects=True,
             )

api/core/tools/tool_engine.py

@@ -58,7 +58,7 @@ class ToolEngine:
             # check if this tool has only one parameter
             parameters = [
                 parameter
-                for parameter in tool.get_runtime_parameters() or []
+                for parameter in tool.get_runtime_parameters()
                 if parameter.form == ToolParameter.ToolParameterForm.LLM
             ]
             if parameters and len(parameters) == 1:

api/core/tools/tool_file_manager.py

@@ -98,7 +98,7 @@ class ToolFileManager:
             response.raise_for_status()
             blob = response.content
         except Exception as e:
-            logger.exception(f"Failed to download file from {file_url}: {e}")
+            logger.exception(f"Failed to download file from {file_url}")
             raise
 
         mimetype = guess_type(file_url)[0] or "octet/stream"

api/core/tools/tool_manager.py

@@ -526,7 +526,7 @@ class ToolManager:
                     yield provider
 
                 except Exception as e:
-                    logger.exception(f"load builtin provider {provider} error: {e}")
+                    logger.exception(f"load builtin provider {provider}")
                     continue
         # set builtin providers loaded
         cls._builtin_providers_loaded = True

api/core/tools/utils/configuration.py

@@ -145,7 +145,7 @@ class ToolParameterConfigurationManager:
         # get tool parameters
         tool_parameters = self.tool_runtime.entity.parameters or []
         # get tool runtime parameters
-        runtime_parameters = self.tool_runtime.get_runtime_parameters() or []
+        runtime_parameters = self.tool_runtime.get_runtime_parameters()
         # override parameters
         current_parameters = tool_parameters.copy()
         for runtime_parameter in runtime_parameters:

api/core/tools/utils/message_transformer.py

@@ -49,7 +49,7 @@ class ToolFileMessageTransformer:
                         meta=message.meta.copy() if message.meta is not None else {},
                     )
                 except Exception as e:
-                    logger.exception(e)
+                    logger.exception(f"Failed to download image from {url}")
                     yield ToolInvokeMessage(
                         type=ToolInvokeMessage.MessageType.TEXT,
                         message=ToolInvokeMessage.TextMessage(

api/core/tools/utils/parser.py

@@ -160,6 +160,9 @@ class ApiBasedToolSchemaParser:
     def _get_tool_parameter_type(parameter: dict) -> ToolParameter.ToolParameterType:
         parameter = parameter or {}
         typ = None
+        if parameter.get("format") == "binary":
+            return ToolParameter.ToolParameterType.FILE
+
         if "type" in parameter:
             typ = parameter["type"]
         elif "schema" in parameter and "type" in parameter["schema"]:

api/core/tools/utils/text_processing_utils.py

@@ -0,0 +1,16 @@
+import re
+
+
+def remove_leading_symbols(text: str) -> str:
+    """
+    Remove leading punctuation or symbols from the given text.
+
+    Args:
+        text (str): The input text to process.
+
+    Returns:
+        str: The text with leading punctuation or symbols removed.
+    """
+    # Match Unicode ranges for punctuation and symbols
+    pattern = r"^[\u2000-\u206F\u2E00-\u2E7F\u3000-\u303F!\"#$%&'()*+,\-./:;<=>?@\[\]^_`{|}~]+"
+    return re.sub(pattern, "", text)