SrSSKJr SSKJr SSKJr SSKJr SSKJr SSKJ r SSKJ r SS K J r SS KJr SS KJr SS KJr SS KJr SSKJr SSKJr SSKJr SSKJr SSKJr SSKJr SrSr SSjr!Sr"Sr#Sr$"SS\%5r&g)z5Utilities for dealing with AI Platform endpoints API.)absolute_import)division)unicode_literals)encoding) exceptions) extra_types) list_pager)utilclient)apis) constants)errors)flags) properties) resources)requests) http_clientc[RRUU[RR R RS.SS9$)z/Parses a model ID into a model resource object.) locationsId projectsIdz$aiplatform.projects.locations.models)params collection)rREGISTRYParserVALUEScoreproject GetOrFail)model_id location_ids 1lib/googlecloudsdk/api_lib/ai/endpoints/client.py _ParseModelr#&sK    ! !$"))..66@@8 " cZUVs/sHn[R"X5PM sn$s snfN)rPyValueToMessage) message_typevaluesvs r"_ConvertPyListToMessageListr+2s%>D Ef( # #L 4f EE Es (Nc[R"US9RU5n[R"U5nUSnUb:SU;a3[ R "SRUR555egURS5 US$![Ra( [ R "SUR5-5ef=f![a USSs$f=f)aGets the deployment resource type of a model. Args: model_ref: a model resource object. client: an apis.GetClientInstance object. shared_resources_ref: str, the shared deployment resource pool the model should use, formatted as the full URI Returns: A string which value must be 'DEDICATED_RESOURCES', 'AUTOMATIC_RESOURCES' or 'SHARED_RESOURCES' Raises: ArgumentError: if the model resource object is not found. r z\There is an error while getting the model information. Please make sure the model %r exists.!supportedDeploymentResourcesTypesSHARED_RESOURCESz,Shared resources not supported for model {}.r) model_client ModelsClientGetapitools_exceptions HttpErrorr ArgumentError RelativeNamerMessageToPyValueformatremove ValueError) model_refr shared_resources_ref model_msgmodel_resource$supported_deployment_resources_typess r"_GetModelDeploymentResourceTyper?6s $))8<q AABs#B>CO KK,Lo"":y/C/CD & H&15=5H5H /1E1E6 *5&11#Y%9%9 46 6h2 -- I I((*(0 J C ;; 3 3 : :3 ??r$c Sn U(aURRUS9n Sn U(aURRUS9n SnU (aURRU S9n[R "S[ R5"UUUUU U US9nU b`[R "S[ R5"SU (aU OS[R "S [ R5"U S 9S 9UlURRUR5UUS 9nURRRU5$) aCreates a new endpoint using v1beta1 API. Args: location_ref: Resource, the parsed location to create an endpoint. display_name: str, the display name of the new endpoint. labels: list, the labels to organize the new endpoint. description: str or None, the description of the new endpoint. network: str, the full name of the Google Compute Engine network. endpoint_id: str or None, the id of the new endpoint. encryption_kms_key_name: str or None, the Cloud KMS resource identifier of the customer managed encryption key used to protect a resource. gdce_zone: str or None, the name of the GDCE zone. gdc_zone: str or None, the name of the GDC zone. request_response_logging_table: str or None, the BigQuery table uri for request-response logging. request_response_logging_rate: float or None, the sampling rate for request-response logging. Returns: A long-running operation for Create. Nrh)rUrj)rkrlrmrnro gdceConfigrTrpTrqrrrsru)ryrz$googleCloudAiplatformV1beta1Endpoint)rc*GoogleCloudAiplatformV1beta1EncryptionSpec&GoogleCloudAiplatformV1beta1GdceConfig%GoogleCloudAiplatformV1beta1GdcConfigr}r~r BETA_VERSIONrrr5r rr)rVrrrmrlrnrr gdce_zonegdc_zonerrr gdce_config gdc_configrXrs r" CreateBetaEndpointsClient.CreateBetasVFO -- B B0 C  KMMHHIkJ==FFGj"":y/E/EF &H&15=5H5H /1G1G6 *5&11#Y%;%; 46 6h2 -- I I((*-5 J C ;; 3 3 : :3 ??r$cURRUR5S9nURRR U5$)zDeletes an existing endpoint.name)rc1AiplatformProjectsLocationsEndpointsDeleteRequestr5r rDeleterVrWrs r"rEndpointsClient.Delete!sE -- I I  & & ( J C ;; 3 3 : :3 ??r$cURRUR5S9nURRR U5$)zGets details about an endpoint.r)rc.AiplatformProjectsLocationsEndpointsGetRequestr5r rr1rs r"r1EndpointsClient.Get(sE -- F F  & & ( G C ;; 3 3 7 7 <>r$c URR5n /n UR(a"URU lU R S5 UbX:lU R S5 Ubu/n [ UR55H1upU R U R5RXS95 M3 U RU S9U l U R S5 U(aSU l U R S5 UbXJl U R S5 UcUbURU5RnU(d*[R"S[ R"5"5nS UlUb.[R"S [ R"5"US 9UlUbUUlUU l U R S 5 U (aAURU5RnU(aS UlUU l U R S 5 U (d[*R,"S5eURR/UR15U SR3U 5S9nUR4R6R9U5$)aUpdates an endpoint using v1beta1 API. Args: endpoint_ref: Resource, the parsed endpoint to be updated. labels_update: UpdateResult, the result of applying the label diff constructed from args. display_name: str or None, the new display name of the endpoint. description: str or None, the new description of the endpoint. traffic_split: dict or None, the new traffic split of the endpoint. clear_traffic_split: bool, whether or not clear traffic split of the endpoint. request_response_logging_table: str or None, the BigQuery table uri for request-response logging. request_response_logging_rate: float or None, the sampling rate for request-response logging. disable_request_response_logging: bool, whether or not disable request-response logging of the endpoint. Returns: The response message of Patch. Raises: NoFieldsSpecifiedError: An error if no updates requested. rmNrrrrrlrpTrrrsrFrr)rrr)rc$GoogleCloudAiplatformV1beta1Endpointrrmrrkrrrrrrlr1rr}r~rrrvrxrwrrrr5rr rrrs r" PatchBetaEndpointsClient.PatchBetasrH}}AACHK!!%,,ho")(  }2245*#$$  & & ( ; ; <  6 '8849h)"h)(' '2 ( 4(, )++&-*2*=*= 193I3I+ + '15%- ' 3    5y7M7M N8  (; ' 2 ) (4 *2BC'(, )++& )27'/ )2BC   ) )*A BB -- H H  & & (-588K( I C ;; 3 3 9 9# >>r$cjURR[[RUS5S9nSU;a-[ R "[RUS5UlURRUR5US9nURRRU5$)z+#> JO ~%#+#<#<    !=$o  -- J J**,.= K C ;; 3 3 ; ;C @@r$cjURR[[RUS5S9nSU;a-[ R "[RUS5UlURRUR5US9nURRRU5$)zASends online prediction request to an endpoint using v1beta1 API.rrr)rX*googleCloudAiplatformV1beta1PredictRequest) rc*GoogleCloudAiplatformV1beta1PredictRequestr+rrrr'rrr5r rrrs r" PredictBetaEndpointsClient.PredictBeta7smmNN-  ! !>+#> OO ~%#+#<#<    !=$o  -- J J**,3B K C ;; 3 3 ; ;C @@r$c.SRURR[URS5UR 55n[ XBU5upVnU[ R:wa'[R"SUR5-5eXg4$)3Sends online raw prediction request to an endpoint.z{}{}/{}:rawPredict_VERSIONzHTTP request failed. Response: ) r7r rIgetattrr5rLrOKcore_exceptionsErrordecode)rVrWrDrFrIstatusresponse_headersrKs r" RawPredictEndpointsClient.RawPredictIs  % %  Z(!!# C *5S7)K&Fh   ! ! ,x/@ @   %%r$c## SRURR[URS5UR 55n[ XBU5HnUv M g7f)rz{}{}/{}:streamRawPredictrN)r7r rIrr5rR)rVrWrDrFrIrPs r"StreamRawPredict EndpointsClient.StreamRawPredictYsR $ + +  Z(!!# C "#8 j9sA"A$cURR[URRUS5S9nSU;a3[R "URRUS5UlURRUR5US9nURRRU5$)zCSends online direct prediction request to an endpoint using v1 API.inputsrr)rX+googleCloudAiplatformV1DirectPredictRequest) rc+GoogleCloudAiplatformV1DirectPredictRequestr+GoogleCloudAiplatformV1Tensorrr'r8AiplatformProjectsLocationsEndpointsDirectPredictRequestr5r r DirectPredictrVrW inputs_jsondirect_predict_requestrs r"rEndpointsClient.DirectPredictds AA. ;;H% B {"*2*C*C -- 5 5{<7P+' NN!..08N O  ;; 3 3 A A# FFr$cURR[URRUS5S9nSU;a3[R "URRUS5UlURRUR5US9nURRRU5$)zHSends online direct prediction request to an endpoint using v1beta1 API.rrr)rX0googleCloudAiplatformV1beta1DirectPredictRequest) rc0GoogleCloudAiplatformV1beta1DirectPredictRequestr+"GoogleCloudAiplatformV1beta1Tensorrr'rrr5r rrrs r"DirectPredictBeta!EndpointsClient.DirectPredictBeta{s FF. @@H% G {"*2*C*C -- : : l #+' -- P P**,9O Q C ;; 3 3 A A# FFr$c (URR[USS5URSURS55S9nURR UR 5US9nUR RRU5$)zGSends online direct raw prediction request to an endpoint using v1 API.inpututf-8 methodName method_namerr)rX.googleCloudAiplatformV1DirectRawPredictRequest) rc.GoogleCloudAiplatformV1DirectRawPredictRequestbytesget;AiplatformProjectsLocationsEndpointsDirectRawPredictRequestr5r rDirectRawPredictrVrW input_jsondirect_raw_predict_requestrs r"r EndpointsClient.DirectRawPredicts!%!]!]Jw'1>>, }0MN"^" -- S S**,7Q T C ;; 3 3 D DS IIr$c (URR[USS5URSURS55S9nURR UR 5US9nUR RRU5$)zLSends online direct raw prediction request to an endpoint using v1beta1 API.rrrr r )rX3googleCloudAiplatformV1beta1DirectRawPredictRequest) rc3GoogleCloudAiplatformV1beta1DirectRawPredictRequestr rrr5r rrrs r"DirectRawPredictBeta$EndpointsClient.DirectRawPredictBetas!%!b!bJw'1>>, }0MN"c" -- S S**,+#> JO ~%#+#<#<    !=$o  )(,(>(>o% -- J J**,.= K C ;; 3 3 ; ;C @@r$cURR[[RUS5S9nSU;a-[ R "[RUS5UlSU;a3[ R "URRUS5Ul URbURUl URRUR5US9nURRR!U5$)rrrrexplanation_spec_override)rX*googleCloudAiplatformV1beta1ExplainRequest)rc*GoogleCloudAiplatformV1beta1ExplainRequestr+rrrr'r3GoogleCloudAiplatformV1beta1ExplanationSpecOverrideexplanationSpecOverriderrr r5r rr!r"s r" ExplainBetaEndpointsClient.ExplainBetasmmNN-  ! !>+#> OO ~%#+#<#<    !=$o #n4080I0I -- K K 4 51o- )(,(>(>o% -- J J**,3B K C ;; 3 3 ; ;C @@r$c[X#5n[UUR5nUS:XGaURR 5nUbUUlUbUUlUbUUl[R"U[R5nUb"URUl URUl U b*[R"U [R5UlURR!UUS9nU =(d SUlU bU UlU bU UlU bf/n[)U R+55H@unnUR-URR/[R0UUS95 MB UUlURR5UUUR75S9nOWURR95nU bU UlU bU UlURR5UUUR75S9nUUlUUlUbUUlUbUUl URRCUS9nUbf/n[)UR+55H3un n!UR-URE5RGU U!S95 M5 UREUS 9Ul$URRKUR75US 9n"URRLROU"5$) a#Deploys a model to an existing endpoint using v1 API. Args: endpoint_ref: Resource, the parsed endpoint that the model is deployed to. model: str, Id of the uploaded model to be deployed. region: str, the location of the endpoint and the model. display_name: str, the display name of the new deployed model. machine_type: str or None, the type of the machine to serve the model. tpu_topology: str or None, the topology of the TPU to serve the model. multihost_gpu_node_count: int or None, the number of nodes per replica for multihost GPU deployments. accelerator_dict: dict or None, the accelerator attached to the deployed model from args. min_replica_count: int or None, the minimum number of replicas the deployed model will be always deployed on. max_replica_count: int or None, the maximum number of replicas the deployed model may be deployed on. required_replica_count: int or None, the required number of replicas the deployed model will be considered successfully deployed. reservation_affinity: dict or None, the reservation affinity of the deployed model which specifies which reservations the deployed model can use. autoscaling_metric_specs: dict or None, the metric specification that defines the target resource utilization for calculating the desired replica count. spot: bool, whether or not deploy the model on spot resources. enable_access_logging: bool, whether or not enable access logs. disable_container_logging: bool, whether or not disable container logging. service_account: str or None, the service account that the deployed model runs as. traffic_split: dict or None, the new traffic split of the endpoint. deployed_model_id: str or None, id of the deployed model. Returns: A long-running operation for DeployModel. DEDICATED_RESOURCES machineSpecspot metricNametargetdedicatedResourcesrkmodelautomaticResourcesrkr9 deployedModelrr)rX)googleCloudAiplatformV1DeployModelRequest)(r#r?r rc"GoogleCloudAiplatformV1MachineSpec machineType tpuTopologymultihostGpuNodeCountrParseAcceleratorFlagrracceleratorTypeacceleratorCountParseReservationAffinityFlagreservationAffinity)GoogleCloudAiplatformV1DedicatedResourcesminReplicaCountmaxReplicaCountrequiredReplicaCountrrr,GoogleCloudAiplatformV1AutoscalingMetricSpec!OP_AUTOSCALING_METRIC_NAME_MAPPERautoscalingMetricSpecs$GoogleCloudAiplatformV1DeployedModelr5)GoogleCloudAiplatformV1AutomaticResourcesenableAccessLoggingdisableContainerLoggingserviceAccountid)GoogleCloudAiplatformV1DeployModelRequestrrr6AiplatformProjectsLocationsEndpointsDeployModelRequestr DeployModel)#rVrWr9regionr machine_type tpu_topologymultihost_gpu_node_countaccelerator_dictmin_replica_countmax_replica_countrequired_replica_countreservation_affinityautoscaling_metric_specsr2enable_access_loggingdisable_container_loggingservice_accountrrr: resource_type machine_spec accelerator dedicatedautoscaling_metric_specs_listrr6deployed_model automaticdeployed_model_reqrrrrs# r"rWEndpointsClient.DeployModelstE*I3It{{KM--]]EEGl  !#/   !#/  ! --E *.. I00k  '2'B'B $(3(D(D %  )+0+M+M )"6"6, (--II"Ji #4"8qi  &$5 !  +)? & ! -(*%"#;#A#A#CDLD& ' . .mmHH&HHNI E,I (}}II&"&&(Jn--IIKi  &$5 !  &$5 !}}II&"&&(Jn *?N&-FN*"&5n#$+n ??( @    }2245*#u$$  0 0 2 E Eu F  6 );(L(L4)M)% -- N N**,2D O C ;; 3 3 ? ? DDr$c[X5nU(a{URRSURRRRSS9nURR USSS9nURR UUUS9nGO[X#5n[UURU5nUS:XGa&URR5nUbUUl UbUUl UbUUl [R"U[R 5nUb"UR"UlUR$UlU b*[R&"U [R 5UlUbUUlURR UUS9nU cSOU UlU cUR,S:XaSUlO U bU UlU bU UlUcUR,S:XaSUlO UbUUlU bf/n [5U R755H@un!n"U R9URR;[R<U!U"S 95 MB U UlURRA5n#S n$UbS RCU5U#l"S n$UbS RCU5U#l#S n$U$(aU#Ul$URR UUURK5S 9nOUS:XaXURRM5n%U bU U%lU bU U%lURR U%UURK5S9nO7URR UURK5URK5S9nUUl'UUl(UbUUl)UbUUl*URRWUS9n&Ubf/n'[5UR755H3un(n)U'R9U&RY5R[U(U)S95 M5 U&RYU'S9U&l.URR_URK5U&S9n*URR`RcU*5$)a Deploys a model to an existing endpoint using v1beta1 API. Args: endpoint_ref: Resource, the parsed endpoint that the model is deployed to. model: str, Id of the uploaded model to be deployed. region: str, the location of the endpoint and the model. display_name: str, the display name of the new deployed model. machine_type: str or None, the type of the machine to serve the model. tpu_topology: str or None, the topology of the TPU to serve the model. multihost_gpu_node_count: int or None, the number of nodes per replica for multihost GPU deployments. accelerator_dict: dict or None, the accelerator attached to the deployed model from args. min_replica_count: int or None, the minimum number of replicas the deployed model will be always deployed on. max_replica_count: int or None, the maximum number of replicas the deployed model may be deployed on. required_replica_count: int or None, the required number of replicas the deployed model will be considered successfully deployed. reservation_affinity: dict or None, the reservation affinity of the deployed model which specifies which reservations the deployed model can use. autoscaling_metric_specs: dict or None, the metric specification that defines the target resource utilization for calculating the desired replica count. spot: bool, whether or not deploy the model on spot resources. enable_access_logging: bool, whether or not enable access logs. enable_container_logging: bool, whether or not enable container logging. service_account: str or None, the service account that the deployed model runs as. traffic_split: dict or None, the new traffic split of the endpoint. deployed_model_id: str or None, id of the deployed model. shared_resources_ref: str or None, the shared deployment resource pool the model should use min_scaleup_period: str or None, the minimum duration (in seconds) that a deployment will be scaled up before traffic is evaluated for potential scale-down. Defaults to 1 hour if min replica count is 0. idle_scaledown_period: str or None, the duration after which the deployment is scaled down if no traffic is received. This only applies to deployments enrolled in scale-to-zero. initial_replica_count: int or None, the initial number of replicas the deployment will be scaled up to. This only applies to deployments enrolled in scale-to-zero. gpu_partition_size: str or None, the partition size of the GPU accelerator. Returns: A long-running operation for DeployModel. z n1-standard-2r3)r@rDrE)r1rIrJ)r8rkgdcConnectedModelr/r0rr4Fz{}sTr7AUTOMATIC_RESOURCESr:)rkr9sharedResourcesr<rr)rX.googleCloudAiplatformV1beta1DeployModelRequest)2rZrc'GoogleCloudAiplatformV1beta1MachineSpecAcceleratorTypeValueValuesEnumNVIDIA_TESLA_T4.GoogleCloudAiplatformV1beta1DedicatedResources)GoogleCloudAiplatformV1beta1DeployedModelr#r?r r@rArBrrCrrrDrErFrGgpuPartitionSizerIrJrKinitialReplicaCountrrr1GoogleCloudAiplatformV1beta1AutoscalingMetricSpecrMrN=GoogleCloudAiplatformV1beta1DedicatedResourcesScaleToZeroSpecr7minScaleupPeriodidleScaledownPeriodscaleToZeroSpecr5.GoogleCloudAiplatformV1beta1AutomaticResourcesrQenableContainerLoggingrSrT.GoogleCloudAiplatformV1beta1DeployModelRequestrrrrVrrW)+rVrWr9rXrrYrZr[r\r]r^r_r`rar2rbenable_container_loggingrdrrr;min_scaleup_periodidle_scaledown_periodinitial_replica_countgpu_partition_sizeis_gdc_ggs_modelrfrhrjr:rergrirr6stz_specstz_spec_modifiedrkrlrrrrs+ r"DeployModelBetaEndpointsClient.DeployModelBetausX+4>]]JJ%--OOnn~~Kl --NN"AqOi}}NN&"!On e,i5 T[["6m / /}}LLN  #%1, "  #%1, " # //G, ,00 i44   ")4)D)D, &*5*F*F, '  +-2-O-O"I$:$:. , *  )*<, ' MM H H(t I  #*A0A !  $)B)Ba)G&') #  *(9I % ! -+A) ( ! (Y-F-F!-K*+) ' " .,AI ) # /*, '$%=%C%C%EFldF ) 0 0 OO(JJ " P G.K) * MM W W Y "  )&+ll3E&F( #"  ,).6K)L( &"  &.) # MM C C#,(,,. D   1 1 MM H H J   (&7) #  (&7) # MM C C#,(,,. D   MM C C(,,. 4 A A C D  *?N&,DN)"&5n#$+n DD( E    }2245*#u$$  0 0 2 E Eu F  6 );(L(L4)M)% -- N N**,7I O C ;; 3 3 ? ? DDr$cURRUS9nUbd/n[UR55H1upgUR UR 5R XgS95 M3 UR US9UlURRUR5US9nURRRU5$)aPUndeploys a model from an endpoint using v1 API. Args: endpoint_ref: Resource, the parsed endpoint that the model is undeployed from. deployed_model_id: str, Id of the deployed model to be undeployed. traffic_split: dict or None, the new traffic split of the endpoint. Returns: A long-running operation for UndeployModel. rrr)rX+googleCloudAiplatformV1UndeployModelRequest) rc+GoogleCloudAiplatformV1UndeployModelRequestrrrrrr8AiplatformProjectsLocationsEndpointsUndeployModelRequestr5r r UndeployModel rVrWrrundeployed_model_reqrrrrs r"rEndpointsClient.UndeployModelfs AA- B    }2245*#$$ 2 2 4 G G H  6  0 0#8 1 ' NN!..08L O  ;; 3 3 A A# FFr$cURRUS9nUbd/n[UR55H1upgUR UR 5R XgS95 M3 UR US9UlURRUR5US9nURRRU5$)aUUndeploys a model from an endpoint using v1beta1 API. Args: endpoint_ref: Resource, the parsed endpoint that the model is undeployed from. deployed_model_id: str, Id of the deployed model to be undeployed. traffic_split: dict or None, the new traffic split of the endpoint. Returns: A long-running operation for UndeployModel. rrr)rX0googleCloudAiplatformV1beta1UndeployModelRequest) rc0GoogleCloudAiplatformV1beta1UndeployModelRequestrrrrrrrr5r rrrs r"UndeployModelBeta!EndpointsClient.UndeployModelBetas FF- G    }2245*#$$ 2 2 4 G G H  6  0 0#8 1 ' -- P P**,9M Q C ;; 3 3 A A# FFr$)r rc)NNN)NNNNNN)NNNNNNNN)NN)NNNFNNF)NNNNNNNNNFFFNNN)NNNNNNNNNFFFNNNNNNNNr&)__name__ __module__ __qualname____firstlineno____doc__rerrrr1rrrrrrrrrrrr!r,rWrrr__static_attributes__r$r"r\r\s]@<"%)$(=@H"%)$(Q@f@= $%)$(',r?p%)$(',r?hA$A$& G.G, J JA(A>#!# ! %)YEB#!# !$  3oEb&GR<@&Gr$r\r&)'r __future__rrrapitools.base.pyrrr2rr googlecloudsdk.api_lib.air r} googlecloudsdk.api_lib.ai.modelsr r/googlecloudsdk.api_lib.utilr googlecloudsdk.command_lib.airrrgooglecloudsdk.corerrrgooglecloudsdk.core.credentialsr six.movesrr#r+r?rLrRrZobjectr\rr$r"rsw<&'%>('6C,30/=*)4! F -1/BdB oGfoGr$