# -*- coding: utf-8 -*- # Copyright 2024 Google LLC # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # import json import logging as std_logging import pickle from typing import Callable, Dict, Optional, Sequence, Tuple, Union import warnings from google.api_core import gapic_v1, grpc_helpers, operations_v1 import google.auth # type: ignore from google.auth import credentials as ga_credentials # type: ignore from google.auth.transport.grpc import SslCredentials # type: ignore from google.longrunning import operations_pb2 # type: ignore from google.protobuf import empty_pb2 # type: ignore from google.protobuf.json_format import MessageToJson import google.protobuf.message import grpc # type: ignore import proto # type: ignore from google.ai.generativelanguage_v1alpha.types import tuned_model as gag_tuned_model from google.ai.generativelanguage_v1alpha.types import model, model_service from google.ai.generativelanguage_v1alpha.types import tuned_model from .base import DEFAULT_CLIENT_INFO, ModelServiceTransport try: from google.api_core import client_logging # type: ignore CLIENT_LOGGING_SUPPORTED = True # pragma: NO COVER except ImportError: # pragma: NO COVER CLIENT_LOGGING_SUPPORTED = False _LOGGER = std_logging.getLogger(__name__) class _LoggingClientInterceptor(grpc.UnaryUnaryClientInterceptor): # pragma: NO COVER def intercept_unary_unary(self, continuation, client_call_details, request): logging_enabled = CLIENT_LOGGING_SUPPORTED and _LOGGER.isEnabledFor( std_logging.DEBUG ) if logging_enabled: # pragma: NO COVER request_metadata = client_call_details.metadata if isinstance(request, proto.Message): request_payload = type(request).to_json(request) elif isinstance(request, google.protobuf.message.Message): request_payload = MessageToJson(request) else: request_payload = f"{type(request).__name__}: {pickle.dumps(request)}" request_metadata = { key: value.decode("utf-8") if isinstance(value, bytes) else value for key, value in request_metadata } grpc_request = { "payload": request_payload, "requestMethod": "grpc", "metadata": dict(request_metadata), } _LOGGER.debug( f"Sending request for {client_call_details.method}", extra={ "serviceName": "google.ai.generativelanguage.v1alpha.ModelService", "rpcName": client_call_details.method, "request": grpc_request, "metadata": grpc_request["metadata"], }, ) response = continuation(client_call_details, request) if logging_enabled: # pragma: NO COVER response_metadata = response.trailing_metadata() # Convert gRPC metadata `` to list of tuples metadata = ( dict([(k, str(v)) for k, v in response_metadata]) if response_metadata else None ) result = response.result() if isinstance(result, proto.Message): response_payload = type(result).to_json(result) elif isinstance(result, google.protobuf.message.Message): response_payload = MessageToJson(result) else: response_payload = f"{type(result).__name__}: {pickle.dumps(result)}" grpc_response = { "payload": response_payload, "metadata": metadata, "status": "OK", } _LOGGER.debug( f"Received response for {client_call_details.method}.", extra={ "serviceName": "google.ai.generativelanguage.v1alpha.ModelService", "rpcName": client_call_details.method, "response": grpc_response, "metadata": grpc_response["metadata"], }, ) return response class ModelServiceGrpcTransport(ModelServiceTransport): """gRPC backend transport for ModelService. Provides methods for getting metadata information about Generative Models. This class defines the same methods as the primary client, so the primary client can load the underlying transport implementation and call it. It sends protocol buffers over the wire using gRPC (which is built on top of HTTP/2); the ``grpcio`` package must be installed. """ _stubs: Dict[str, Callable] def __init__( self, *, host: str = "generativelanguage.googleapis.com", credentials: Optional[ga_credentials.Credentials] = None, credentials_file: Optional[str] = None, scopes: Optional[Sequence[str]] = None, channel: Optional[Union[grpc.Channel, Callable[..., grpc.Channel]]] = None, api_mtls_endpoint: Optional[str] = None, client_cert_source: Optional[Callable[[], Tuple[bytes, bytes]]] = None, ssl_channel_credentials: Optional[grpc.ChannelCredentials] = None, client_cert_source_for_mtls: Optional[Callable[[], Tuple[bytes, bytes]]] = None, quota_project_id: Optional[str] = None, client_info: gapic_v1.client_info.ClientInfo = DEFAULT_CLIENT_INFO, always_use_jwt_access: Optional[bool] = False, api_audience: Optional[str] = None, ) -> None: """Instantiate the transport. Args: host (Optional[str]): The hostname to connect to (default: 'generativelanguage.googleapis.com'). credentials (Optional[google.auth.credentials.Credentials]): The authorization credentials to attach to requests. These credentials identify the application to the service; if none are specified, the client will attempt to ascertain the credentials from the environment. This argument is ignored if a ``channel`` instance is provided. credentials_file (Optional[str]): A file with credentials that can be loaded with :func:`google.auth.load_credentials_from_file`. This argument is ignored if a ``channel`` instance is provided. scopes (Optional(Sequence[str])): A list of scopes. This argument is ignored if a ``channel`` instance is provided. channel (Optional[Union[grpc.Channel, Callable[..., grpc.Channel]]]): A ``Channel`` instance through which to make calls, or a Callable that constructs and returns one. If set to None, ``self.create_channel`` is used to create the channel. If a Callable is given, it will be called with the same arguments as used in ``self.create_channel``. api_mtls_endpoint (Optional[str]): Deprecated. The mutual TLS endpoint. If provided, it overrides the ``host`` argument and tries to create a mutual TLS channel with client SSL credentials from ``client_cert_source`` or application default SSL credentials. client_cert_source (Optional[Callable[[], Tuple[bytes, bytes]]]): Deprecated. A callback to provide client SSL certificate bytes and private key bytes, both in PEM format. It is ignored if ``api_mtls_endpoint`` is None. ssl_channel_credentials (grpc.ChannelCredentials): SSL credentials for the grpc channel. It is ignored if a ``channel`` instance is provided. client_cert_source_for_mtls (Optional[Callable[[], Tuple[bytes, bytes]]]): A callback to provide client certificate bytes and private key bytes, both in PEM format. It is used to configure a mutual TLS channel. It is ignored if a ``channel`` instance or ``ssl_channel_credentials`` is provided. quota_project_id (Optional[str]): An optional project to use for billing and quota. client_info (google.api_core.gapic_v1.client_info.ClientInfo): The client info used to send a user-agent string along with API requests. If ``None``, then default info will be used. Generally, you only need to set this if you're developing your own client library. always_use_jwt_access (Optional[bool]): Whether self signed JWT should be used for service account credentials. Raises: google.auth.exceptions.MutualTLSChannelError: If mutual TLS transport creation failed for any reason. google.api_core.exceptions.DuplicateCredentialArgs: If both ``credentials`` and ``credentials_file`` are passed. """ self._grpc_channel = None self._ssl_channel_credentials = ssl_channel_credentials self._stubs: Dict[str, Callable] = {} self._operations_client: Optional[operations_v1.OperationsClient] = None if api_mtls_endpoint: warnings.warn("api_mtls_endpoint is deprecated", DeprecationWarning) if client_cert_source: warnings.warn("client_cert_source is deprecated", DeprecationWarning) if isinstance(channel, grpc.Channel): # Ignore credentials if a channel was passed. credentials = None self._ignore_credentials = True # If a channel was explicitly provided, set it. self._grpc_channel = channel self._ssl_channel_credentials = None else: if api_mtls_endpoint: host = api_mtls_endpoint # Create SSL credentials with client_cert_source or application # default SSL credentials. if client_cert_source: cert, key = client_cert_source() self._ssl_channel_credentials = grpc.ssl_channel_credentials( certificate_chain=cert, private_key=key ) else: self._ssl_channel_credentials = SslCredentials().ssl_credentials else: if client_cert_source_for_mtls and not ssl_channel_credentials: cert, key = client_cert_source_for_mtls() self._ssl_channel_credentials = grpc.ssl_channel_credentials( certificate_chain=cert, private_key=key ) # The base transport sets the host, credentials and scopes super().__init__( host=host, credentials=credentials, credentials_file=credentials_file, scopes=scopes, quota_project_id=quota_project_id, client_info=client_info, always_use_jwt_access=always_use_jwt_access, api_audience=api_audience, ) if not self._grpc_channel: # initialize with the provided callable or the default channel channel_init = channel or type(self).create_channel self._grpc_channel = channel_init( self._host, # use the credentials which are saved credentials=self._credentials, # Set ``credentials_file`` to ``None`` here as # the credentials that we saved earlier should be used. credentials_file=None, scopes=self._scopes, ssl_credentials=self._ssl_channel_credentials, quota_project_id=quota_project_id, options=[ ("grpc.max_send_message_length", -1), ("grpc.max_receive_message_length", -1), ], ) self._interceptor = _LoggingClientInterceptor() self._logged_channel = grpc.intercept_channel( self._grpc_channel, self._interceptor ) # Wrap messages. This must be done after self._logged_channel exists self._prep_wrapped_messages(client_info) @classmethod def create_channel( cls, host: str = "generativelanguage.googleapis.com", credentials: Optional[ga_credentials.Credentials] = None, credentials_file: Optional[str] = None, scopes: Optional[Sequence[str]] = None, quota_project_id: Optional[str] = None, **kwargs, ) -> grpc.Channel: """Create and return a gRPC channel object. Args: host (Optional[str]): The host for the channel to use. credentials (Optional[~.Credentials]): The authorization credentials to attach to requests. These credentials identify this application to the service. If none are specified, the client will attempt to ascertain the credentials from the environment. credentials_file (Optional[str]): A file with credentials that can be loaded with :func:`google.auth.load_credentials_from_file`. This argument is mutually exclusive with credentials. scopes (Optional[Sequence[str]]): A optional list of scopes needed for this service. These are only used when credentials are not specified and are passed to :func:`google.auth.default`. quota_project_id (Optional[str]): An optional project to use for billing and quota. kwargs (Optional[dict]): Keyword arguments, which are passed to the channel creation. Returns: grpc.Channel: A gRPC channel object. Raises: google.api_core.exceptions.DuplicateCredentialArgs: If both ``credentials`` and ``credentials_file`` are passed. """ return grpc_helpers.create_channel( host, credentials=credentials, credentials_file=credentials_file, quota_project_id=quota_project_id, default_scopes=cls.AUTH_SCOPES, scopes=scopes, default_host=cls.DEFAULT_HOST, **kwargs, ) @property def grpc_channel(self) -> grpc.Channel: """Return the channel designed to connect to this service.""" return self._grpc_channel @property def operations_client(self) -> operations_v1.OperationsClient: """Create the client designed to process long-running operations. This property caches on the instance; repeated calls return the same client. """ # Quick check: Only create a new client if we do not already have one. if self._operations_client is None: self._operations_client = operations_v1.OperationsClient( self._logged_channel ) # Return the client from cache. return self._operations_client @property def get_model(self) -> Callable[[model_service.GetModelRequest], model.Model]: r"""Return a callable for the get model method over gRPC. Gets information about a specific ``Model`` such as its version number, token limits, `parameters `__ and other metadata. Refer to the `Gemini models guide `__ for detailed model information. Returns: Callable[[~.GetModelRequest], ~.Model]: A function that, when called, will call the underlying RPC on the server. """ # Generate a "stub function" on-the-fly which will actually make # the request. # gRPC handles serialization and deserialization, so we just need # to pass in the functions for each. if "get_model" not in self._stubs: self._stubs["get_model"] = self._logged_channel.unary_unary( "/google.ai.generativelanguage.v1alpha.ModelService/GetModel", request_serializer=model_service.GetModelRequest.serialize, response_deserializer=model.Model.deserialize, ) return self._stubs["get_model"] @property def list_models( self, ) -> Callable[[model_service.ListModelsRequest], model_service.ListModelsResponse]: r"""Return a callable for the list models method over gRPC. Lists the ```Model``\ s `__ available through the Gemini API. Returns: Callable[[~.ListModelsRequest], ~.ListModelsResponse]: A function that, when called, will call the underlying RPC on the server. """ # Generate a "stub function" on-the-fly which will actually make # the request. # gRPC handles serialization and deserialization, so we just need # to pass in the functions for each. if "list_models" not in self._stubs: self._stubs["list_models"] = self._logged_channel.unary_unary( "/google.ai.generativelanguage.v1alpha.ModelService/ListModels", request_serializer=model_service.ListModelsRequest.serialize, response_deserializer=model_service.ListModelsResponse.deserialize, ) return self._stubs["list_models"] @property def get_tuned_model( self, ) -> Callable[[model_service.GetTunedModelRequest], tuned_model.TunedModel]: r"""Return a callable for the get tuned model method over gRPC. Gets information about a specific TunedModel. Returns: Callable[[~.GetTunedModelRequest], ~.TunedModel]: A function that, when called, will call the underlying RPC on the server. """ # Generate a "stub function" on-the-fly which will actually make # the request. # gRPC handles serialization and deserialization, so we just need # to pass in the functions for each. if "get_tuned_model" not in self._stubs: self._stubs["get_tuned_model"] = self._logged_channel.unary_unary( "/google.ai.generativelanguage.v1alpha.ModelService/GetTunedModel", request_serializer=model_service.GetTunedModelRequest.serialize, response_deserializer=tuned_model.TunedModel.deserialize, ) return self._stubs["get_tuned_model"] @property def list_tuned_models( self, ) -> Callable[ [model_service.ListTunedModelsRequest], model_service.ListTunedModelsResponse ]: r"""Return a callable for the list tuned models method over gRPC. Lists created tuned models. Returns: Callable[[~.ListTunedModelsRequest], ~.ListTunedModelsResponse]: A function that, when called, will call the underlying RPC on the server. """ # Generate a "stub function" on-the-fly which will actually make # the request. # gRPC handles serialization and deserialization, so we just need # to pass in the functions for each. if "list_tuned_models" not in self._stubs: self._stubs["list_tuned_models"] = self._logged_channel.unary_unary( "/google.ai.generativelanguage.v1alpha.ModelService/ListTunedModels", request_serializer=model_service.ListTunedModelsRequest.serialize, response_deserializer=model_service.ListTunedModelsResponse.deserialize, ) return self._stubs["list_tuned_models"] @property def create_tuned_model( self, ) -> Callable[[model_service.CreateTunedModelRequest], operations_pb2.Operation]: r"""Return a callable for the create tuned model method over gRPC. Creates a tuned model. Check intermediate tuning progress (if any) through the [google.longrunning.Operations] service. Access status and results through the Operations service. Example: GET /v1/tunedModels/az2mb0bpw6i/operations/000-111-222 Returns: Callable[[~.CreateTunedModelRequest], ~.Operation]: A function that, when called, will call the underlying RPC on the server. """ # Generate a "stub function" on-the-fly which will actually make # the request. # gRPC handles serialization and deserialization, so we just need # to pass in the functions for each. if "create_tuned_model" not in self._stubs: self._stubs["create_tuned_model"] = self._logged_channel.unary_unary( "/google.ai.generativelanguage.v1alpha.ModelService/CreateTunedModel", request_serializer=model_service.CreateTunedModelRequest.serialize, response_deserializer=operations_pb2.Operation.FromString, ) return self._stubs["create_tuned_model"] @property def update_tuned_model( self, ) -> Callable[[model_service.UpdateTunedModelRequest], gag_tuned_model.TunedModel]: r"""Return a callable for the update tuned model method over gRPC. Updates a tuned model. Returns: Callable[[~.UpdateTunedModelRequest], ~.TunedModel]: A function that, when called, will call the underlying RPC on the server. """ # Generate a "stub function" on-the-fly which will actually make # the request. # gRPC handles serialization and deserialization, so we just need # to pass in the functions for each. if "update_tuned_model" not in self._stubs: self._stubs["update_tuned_model"] = self._logged_channel.unary_unary( "/google.ai.generativelanguage.v1alpha.ModelService/UpdateTunedModel", request_serializer=model_service.UpdateTunedModelRequest.serialize, response_deserializer=gag_tuned_model.TunedModel.deserialize, ) return self._stubs["update_tuned_model"] @property def delete_tuned_model( self, ) -> Callable[[model_service.DeleteTunedModelRequest], empty_pb2.Empty]: r"""Return a callable for the delete tuned model method over gRPC. Deletes a tuned model. Returns: Callable[[~.DeleteTunedModelRequest], ~.Empty]: A function that, when called, will call the underlying RPC on the server. """ # Generate a "stub function" on-the-fly which will actually make # the request. # gRPC handles serialization and deserialization, so we just need # to pass in the functions for each. if "delete_tuned_model" not in self._stubs: self._stubs["delete_tuned_model"] = self._logged_channel.unary_unary( "/google.ai.generativelanguage.v1alpha.ModelService/DeleteTunedModel", request_serializer=model_service.DeleteTunedModelRequest.serialize, response_deserializer=empty_pb2.Empty.FromString, ) return self._stubs["delete_tuned_model"] def close(self): self._logged_channel.close() @property def get_operation( self, ) -> Callable[[operations_pb2.GetOperationRequest], operations_pb2.Operation]: r"""Return a callable for the get_operation method over gRPC.""" # Generate a "stub function" on-the-fly which will actually make # the request. # gRPC handles serialization and deserialization, so we just need # to pass in the functions for each. if "get_operation" not in self._stubs: self._stubs["get_operation"] = self._logged_channel.unary_unary( "/google.longrunning.Operations/GetOperation", request_serializer=operations_pb2.GetOperationRequest.SerializeToString, response_deserializer=operations_pb2.Operation.FromString, ) return self._stubs["get_operation"] @property def list_operations( self, ) -> Callable[ [operations_pb2.ListOperationsRequest], operations_pb2.ListOperationsResponse ]: r"""Return a callable for the list_operations method over gRPC.""" # Generate a "stub function" on-the-fly which will actually make # the request. # gRPC handles serialization and deserialization, so we just need # to pass in the functions for each. if "list_operations" not in self._stubs: self._stubs["list_operations"] = self._logged_channel.unary_unary( "/google.longrunning.Operations/ListOperations", request_serializer=operations_pb2.ListOperationsRequest.SerializeToString, response_deserializer=operations_pb2.ListOperationsResponse.FromString, ) return self._stubs["list_operations"] @property def kind(self) -> str: return "grpc" __all__ = ("ModelServiceGrpcTransport",)