apache-airflow-providers-google 18.0.0__py3-none-any.whl → 18.1.0rc1__py3-none-any.whl

airflow/providers/google/cloud/hooks/vertex_ai/generative_model.py

@@ -90,6 +90,11 @@ class GenerativeModelHook(GoogleBaseHook):
         cached_context_model = preview_generative_model.GenerativeModel.from_cached_content(cached_content)
         return cached_context_model
 
+    @deprecated(
+        planned_removal_date="January 3, 2026",
+        use_instead="airflow.providers.google.cloud.hooks.gen_ai.generative_model.GenAIGenerativeModelHook.embed_content",
+        category=AirflowProviderDeprecationWarning,
+    )
     @GoogleBaseHook.fallback_to_default_project_id
     def text_embedding_model_get_embeddings(
         self,
@@ -114,6 +119,11 @@ class GenerativeModelHook(GoogleBaseHook):
 
         return response.values
 
+    @deprecated(
+        planned_removal_date="January 3, 2026",
+        use_instead="airflow.providers.google.cloud.hooks.gen_ai.generative_model.GenAIGenerativeModelHook.generate_content",
+        category=AirflowProviderDeprecationWarning,
+    )
     @GoogleBaseHook.fallback_to_default_project_id
     def generative_model_generate_content(
         self,
@@ -156,6 +166,11 @@ class GenerativeModelHook(GoogleBaseHook):
 
         return response.text
 
+    @deprecated(
+        planned_removal_date="January 3, 2026",
+        use_instead="airflow.providers.google.cloud.hooks.gen_ai.generative_model.GenAIGenerativeModelHook.supervised_fine_tuning_train",
+        category=AirflowProviderDeprecationWarning,
+    )
     @GoogleBaseHook.fallback_to_default_project_id
     def supervised_fine_tuning_train(
         self,
@@ -209,6 +224,11 @@ class GenerativeModelHook(GoogleBaseHook):
 
         return sft_tuning_job
 
+    @deprecated(
+        planned_removal_date="January 3, 2026",
+        use_instead="airflow.providers.google.cloud.hooks.gen_ai.generative_model.GenAIGenerativeModelHook.count_tokens",
+        category=AirflowProviderDeprecationWarning,
+    )
     @GoogleBaseHook.fallback_to_default_project_id
     def count_tokens(
         self,
@@ -296,6 +316,11 @@ class GenerativeModelHook(GoogleBaseHook):
 
         return eval_result
 
+    @deprecated(
+        planned_removal_date="January 3, 2026",
+        use_instead="airflow.providers.google.cloud.hooks.gen_ai.generative_model.GenAIGenerativeModelHook.create_cached_content",
+        category=AirflowProviderDeprecationWarning,
+    )
     def create_cached_content(
         self,
         model_name: str,
@@ -330,6 +355,11 @@ class GenerativeModelHook(GoogleBaseHook):
 
         return response.name
 
+    @deprecated(
+        planned_removal_date="January 3, 2026",
+        use_instead="airflow.providers.google.cloud.hooks.gen_ai.generative_model.GenAIGenerativeModelHook.generate_content",
+        category=AirflowProviderDeprecationWarning,
+    )
     def generate_from_cached_content(
         self,
         location: str,

airflow/providers/google/cloud/links/base.py

@@ -18,18 +18,10 @@
 from __future__ import annotations
 
 from typing import TYPE_CHECKING, ClassVar
+from urllib.parse import urlparse
 
-from airflow.providers.google.version_compat import (
-    AIRFLOW_V_3_0_PLUS,
-    BaseOperator,
-    BaseOperatorLink,
-    BaseSensorOperator,
-)
-
-if AIRFLOW_V_3_0_PLUS:
-    from airflow.sdk.execution_time.xcom import XCom
-else:
-    from airflow.models.xcom import XCom  # type: ignore[no-redef]
+from airflow.providers.common.compat.sdk import BaseOperatorLink, BaseSensorOperator, XCom
+from airflow.providers.google.version_compat import AIRFLOW_V_3_0_PLUS, BaseOperator
 
 if TYPE_CHECKING:
     from airflow.models.taskinstancekey import TaskInstanceKey
@@ -109,6 +101,14 @@ class BaseGoogleLink(BaseOperatorLink):
         if TYPE_CHECKING:
             assert isinstance(operator, (GoogleCloudBaseOperator, BaseSensorOperator))
 
+        # In cases when worker passes execution to trigger, the value that is put to XCom
+        # already contains link to the object in string format. In this case we don't want to execute
+        # get_config() again. Instead we can leave this value without any changes
+        link_value = XCom.get_value(key=self.key, ti_key=ti_key)
+        if link_value and isinstance(link_value, str):
+            if urlparse(link_value).scheme in ("http", "https"):
+                return link_value
+
         conf = self.get_config(operator, ti_key)
         if not conf:
             return ""

airflow/providers/google/cloud/links/dataproc.py

@@ -25,22 +25,14 @@ from typing import TYPE_CHECKING, Any
 import attr
 
 from airflow.exceptions import AirflowProviderDeprecationWarning
+from airflow.providers.common.compat.sdk import BaseOperatorLink, XCom
 from airflow.providers.google.cloud.links.base import BASE_LINK, BaseGoogleLink
-from airflow.providers.google.version_compat import (
-    AIRFLOW_V_3_0_PLUS,
-    BaseOperator,
-    BaseOperatorLink,
-)
 
 if TYPE_CHECKING:
     from airflow.models.taskinstancekey import TaskInstanceKey
+    from airflow.providers.google.version_compat import BaseOperator
     from airflow.utils.context import Context
 
-if AIRFLOW_V_3_0_PLUS:
-    from airflow.sdk.execution_time.xcom import XCom
-else:
-    from airflow.models.xcom import XCom  # type: ignore[no-redef]
-
 
 def __getattr__(name: str) -> Any:
     # PEP-562: deprecate module-level variable

airflow/providers/google/cloud/openlineage/CloudStorageTransferJobFacet.json

@@ -0,0 +1,68 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$defs": {
+    "CloudStorageTransferJobFacet": {
+      "allOf": [
+        {
+          "$ref": "https://openlineage.io/spec/2-0-2/OpenLineage.json#/$defs/JobFacet"
+        },
+        {
+          "type": "object",
+          "properties": {
+            "jobName": {
+              "type": "string",
+              "description": "Transfer job name assigned by GCP Storage Transfer Service."
+            },
+            "projectId": {
+              "type": "string",
+              "description": "GCP project ID."
+            },
+            "description": {
+              "type": "string",
+              "description": "Optional description of the transfer job."
+            },
+            "status": {
+              "type": "string",
+              "description": "Status of the transfer job (ENABLED, DISABLED)."
+            },
+            "sourceBucket": {
+              "type": "string",
+              "description": "Source AWS S3 bucket."
+            },
+            "sourcePath": {
+              "type": "string",
+              "description": "Prefix path inside the source bucket."
+            },
+            "targetBucket": {
+              "type": "string",
+              "description": "Target GCS bucket."
+            },
+            "targetPath": {
+              "type": "string",
+              "description": "Prefix path inside the target bucket."
+            },
+            "objectConditions": {
+              "type": "object",
+              "description": "Filtering conditions for objects transferred."
+            },
+            "transferOptions": {
+              "type": "object",
+              "description": "Transfer options such as overwrite or delete."
+            },
+            "schedule": {
+              "type": "object",
+              "description": "Transfer schedule details."
+            }
+          }
+        }
+      ],
+      "type": "object"
+    }
+  },
+  "type": "object",
+  "properties": {
+    "cloudStorageTransferJob": {
+      "$ref": "#/$defs/CloudStorageTransferJobFacet"
+    }
+  }
+}

airflow/providers/google/cloud/openlineage/CloudStorageTransferRunFacet.json

@@ -0,0 +1,60 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$defs": {
+    "CloudStorageTransferRunFacet": {
+      "allOf": [
+        {
+          "$ref": "https://openlineage.io/spec/2-0-2/OpenLineage.json#/$defs/RunFacet"
+        },
+        {
+          "type": "object",
+          "properties": {
+            "jobName": {
+              "type": "string",
+              "description": "Transfer job name associated with this run."
+            },
+            "operationName": {
+              "type": "string",
+              "description": "Transfer operation name if available."
+            },
+            "status": {
+              "type": "string",
+              "description": "Run status if available."
+            },
+            "startTime": {
+              "type": "string",
+              "description": "Start time of the transfer operation."
+            },
+            "endTime": {
+              "type": "string",
+              "description": "End time of the transfer operation."
+            },
+            "wait": {
+              "type": "boolean",
+              "description": "Whether the operator waited for completion."
+            },
+            "timeout": {
+              "type": ["number", "null"],
+              "description": "Timeout in seconds."
+            },
+            "deferrable": {
+              "type": "boolean",
+              "description": "Whether the operator used deferrable mode."
+            },
+            "deleteJobAfterCompletion": {
+              "type": "boolean",
+              "description": "Whether the transfer job was deleted after completion."
+            }
+          }
+        }
+      ],
+      "type": "object"
+    }
+  },
+  "type": "object",
+  "properties": {
+    "cloudStorageTransferRun": {
+      "$ref": "#/$defs/CloudStorageTransferRunFacet"
+    }
+  }
+}

airflow/providers/google/cloud/openlineage/DataFusionRunFacet.json

@@ -0,0 +1,32 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$defs": {
+    "DataFusionRunFacet": {
+      "allOf": [
+        {
+          "$ref": "https://openlineage.io/spec/2-0-2/OpenLineage.json#/$defs/RunFacet"
+        },
+        {
+          "type": "object",
+          "properties": {
+            "runId": {
+              "type": "string",
+              "description": "Pipeline run ID assigned by Cloud Data Fusion."
+            },
+            "runtimeArgs": {
+              "type": "object",
+              "description": "Runtime arguments provided when starting the pipeline."
+            }
+          }
+        }
+      ],
+      "type": "object"
+    }
+  },
+  "type": "object",
+  "properties": {
+    "dataFusionRun": {
+      "$ref": "#/$defs/DataFusionRunFacet"
+    }
+  }
+}

airflow/providers/google/cloud/openlineage/facets.py

@@ -24,13 +24,17 @@ from attr import define, field
 from airflow.providers.google import __version__ as provider_version
 
 if TYPE_CHECKING:
-    from openlineage.client.generated.base import RunFacet
+    from openlineage.client.generated.base import JobFacet, RunFacet
 
 try:
     try:
         from openlineage.client.generated.base import RunFacet
     except ImportError:  # Old OpenLineage client is used
         from openlineage.client.facet import BaseFacet as RunFacet  # type: ignore[assignment]
+    try:
+        from openlineage.client.generated.base import JobFacet
+    except ImportError:  # Old OpenLineage client is used
+        from openlineage.client.facet import BaseFacet as JobFacet  # type: ignore[assignment]
 
     @define
     class BigQueryJobRunFacet(RunFacet):
@@ -53,6 +57,100 @@ try:
                 f"providers-google/{provider_version}/airflow/providers/google/"
                 "openlineage/BigQueryJobRunFacet.json"
             )
+
+    @define
+    class CloudStorageTransferJobFacet(JobFacet):
+        """
+        Facet representing a Cloud Storage Transfer Service job configuration.
+
+        :param jobName: Unique name of the transfer job.
+        :param projectId: GCP project where the transfer job is defined.
+        :param description: User-provided description of the transfer job.
+        :param status: Current status of the transfer job (e.g. "ENABLED", "DISABLED").
+        :param sourceBucket: Name of the source bucket (e.g. AWS S3).
+        :param sourcePath: Prefix/path inside the source bucket.
+        :param targetBucket: Name of the destination bucket (e.g. GCS).
+        :param targetPath: Prefix/path inside the destination bucket.
+        :param objectConditions: Object selection rules (e.g. include/exclude prefixes).
+        :param transferOptions: Transfer options, such as overwrite behavior or whether to delete objects
+        from the source after transfer.
+        :param schedule: Schedule for the transfer job (if recurring).
+        """
+
+        jobName: str | None = field(default=None)
+        projectId: str | None = field(default=None)
+        description: str | None = field(default=None)
+        status: str | None = field(default=None)
+        sourceBucket: str | None = field(default=None)
+        sourcePath: str | None = field(default=None)
+        targetBucket: str | None = field(default=None)
+        targetPath: str | None = field(default=None)
+        objectConditions: dict | None = field(default=None)
+        transferOptions: dict | None = field(default=None)
+        schedule: dict | None = field(default=None)
+
+        @staticmethod
+        def _get_schema() -> str:
+            return (
+                "https://raw.githubusercontent.com/apache/airflow/"
+                f"providers-google/{provider_version}/airflow/providers/google/"
+                "openlineage/CloudStorageTransferJobFacet.json"
+            )
+
+    @define
+    class CloudStorageTransferRunFacet(RunFacet):
+        """
+        Facet representing a Cloud Storage Transfer Service job execution run.
+
+        :param jobName: Name of the transfer job being executed.
+        :param operationName: Name of the specific transfer operation instance.
+        :param status: Current status of the operation (e.g. "IN_PROGRESS", "SUCCESS", "FAILED").
+        :param startTime: Time when the transfer job execution started (ISO 8601 format).
+        :param endTime: Time when the transfer job execution finished (ISO 8601 format).
+        :param wait: Whether the operator waits for the job to complete before finishing.
+        :param timeout: Timeout (in seconds) for the transfer run to complete.
+        :param deferrable: Whether the operator defers execution until job completion.
+        :param deleteJobAfterCompletion: Whether the operator deletes the transfer job after the run completes.
+        """
+
+        jobName: str | None = field(default=None)
+        operationName: str | None = field(default=None)
+        status: str | None = field(default=None)
+        startTime: str | None = field(default=None)
+        endTime: str | None = field(default=None)
+        wait: bool = field(default=True)
+        timeout: float | None = field(default=None)
+        deferrable: bool = field(default=False)
+        deleteJobAfterCompletion: bool = field(default=False)
+
+        @staticmethod
+        def _get_schema() -> str:
+            return (
+                "https://raw.githubusercontent.com/apache/airflow/"
+                f"providers-google/{provider_version}/airflow/providers/google/"
+                "openlineage/CloudStorageTransferRunFacet.json"
+            )
+
+    @define
+    class DataFusionRunFacet(RunFacet):
+        """
+        Facet that represents relevant details of a Cloud Data Fusion pipeline run.
+
+        :param runId: The pipeline execution id.
+        :param runtimeArgs: Runtime arguments passed to the pipeline.
+        """
+
+        runId: str | None = field(default=None)
+        runtimeArgs: dict[str, str] | None = field(default=None)
+
+        @staticmethod
+        def _get_schema() -> str:
+            return (
+                "https://raw.githubusercontent.com/apache/airflow/"
+                f"providers-google/{provider_version}/airflow/providers/google/"
+                "openlineage/DataFusionRunFacet.json"
+            )
+
 except ImportError:  # OpenLineage is not available
 
     def create_no_op(*_, **__) -> None:
@@ -65,3 +163,6 @@ except ImportError:  # OpenLineage is not available
         return None
 
     BigQueryJobRunFacet = create_no_op  # type: ignore[misc, assignment]
+    CloudStorageTransferJobFacet = create_no_op  # type: ignore[misc, assignment]
+    CloudStorageTransferRunFacet = create_no_op  # type: ignore[misc, assignment]
+    DataFusionRunFacet = create_no_op  # type: ignore[misc, assignment]

airflow/providers/google/cloud/openlineage/mixins.py

@@ -97,7 +97,9 @@ class _BigQueryInsertJobOperatorOpenLineageMixin:
         run_facets: dict[str, RunFacet] = {
             "externalQuery": ExternalQueryRunFacet(externalQueryId=self.job_id, source="bigquery")
         }
-        self._client = self.hook.get_client(project_id=self.hook.project_id, location=self.location)
+        self._client = self.hook.get_client(
+            project_id=self.project_id or self.hook.project_id, location=self.location
+        )
         try:
             job_properties = self._client.get_job(job_id=self.job_id)._properties
 

airflow/providers/google/cloud/operators/bigquery.py

@@ -2370,20 +2370,13 @@ class BigQueryInsertJobOperator(GoogleCloudBaseOperator, _BigQueryInsertJobOpera
         if self.project_id is None:
             self.project_id = hook.project_id
 
-        # Handle missing logical_date. Example: asset-triggered DAGs (Airflow 3)
-        logical_date = context.get("logical_date")
-        if logical_date is None:
-            # Use dag_run.run_after as fallback when logical_date is not available
-            dag_run = context.get("dag_run")
-            if dag_run and hasattr(dag_run, "run_after"):
-                logical_date = dag_run.run_after
-
         self.job_id = hook.generate_job_id(
             job_id=self.job_id,
             dag_id=self.dag_id,
             task_id=self.task_id,
-            logical_date=logical_date,
+            logical_date=None,
             configuration=self.configuration,
+            run_after=hook.get_run_after_or_logical_date(context),
             force_rerun=self.force_rerun,
         )
 

airflow/providers/google/cloud/operators/cloud_run.py

@@ -441,9 +441,10 @@ class CloudRunCreateServiceOperator(GoogleCloudBaseOperator):
                 self.service_name,
                 self.region,
             )
-            return hook.get_service(
+            service = hook.get_service(
                 service_name=self.service_name, region=self.region, project_id=self.project_id
             )
+            return Service.to_dict(service)
         except google.cloud.exceptions.GoogleCloudError as e:
             self.log.error("An error occurred. Exiting.")
             raise e

airflow/providers/google/cloud/operators/cloud_sql.py

@@ -28,6 +28,7 @@ from googleapiclient.errors import HttpError
 
 from airflow.configuration import conf
 from airflow.exceptions import AirflowException
+from airflow.providers.common.compat.sdk import BaseHook
 from airflow.providers.google.cloud.hooks.cloud_sql import CloudSQLDatabaseHook, CloudSQLHook
 from airflow.providers.google.cloud.links.cloud_sql import CloudSQLInstanceDatabaseLink, CloudSQLInstanceLink
 from airflow.providers.google.cloud.operators.cloud_base import GoogleCloudBaseOperator
@@ -35,7 +36,6 @@ from airflow.providers.google.cloud.triggers.cloud_sql import CloudSQLExportTrig
 from airflow.providers.google.cloud.utils.field_validator import GcpBodyFieldValidator
 from airflow.providers.google.common.hooks.base_google import PROVIDE_PROJECT_ID, get_field
 from airflow.providers.google.common.links.storage import FileDetailsLink
-from airflow.providers.google.version_compat import BaseHook
 
 if TYPE_CHECKING:
     from airflow.models import Connection

airflow/providers/google/cloud/operators/cloud_storage_transfer_service.py

@@ -65,12 +65,14 @@ from airflow.providers.google.cloud.links.cloud_storage_transfer import (
 )
 from airflow.providers.google.cloud.operators.cloud_base import GoogleCloudBaseOperator
 from airflow.providers.google.cloud.triggers.cloud_storage_transfer_service import (
+    CloudDataTransferServiceRunJobTrigger,
     CloudStorageTransferServiceCheckJobStatusTrigger,
 )
 from airflow.providers.google.cloud.utils.helpers import normalize_directory_path
 from airflow.providers.google.common.hooks.base_google import PROVIDE_PROJECT_ID
 
 if TYPE_CHECKING:
+    from airflow.providers.openlineage.extractors import OperatorLineage
     from airflow.utils.context import Context
 
 
@@ -468,6 +470,8 @@ class CloudDataTransferServiceRunJobOperator(GoogleCloudBaseOperator):
         If set as a sequence, the identities from the list must grant
         Service Account Token Creator IAM role to the directly preceding identity, with first
         account from the list granting this role to the originating account (templated).
+    :param timeout: Time to wait for the operation to end in seconds. Defaults to 60 seconds if not specified.
+    :param deferrable: Run operator in the deferrable mode.
     """
 
     # [START gcp_transfer_job_run_template_fields]
@@ -489,6 +493,8 @@ class CloudDataTransferServiceRunJobOperator(GoogleCloudBaseOperator):
         api_version: str = "v1",
         project_id: str = PROVIDE_PROJECT_ID,
         google_impersonation_chain: str | Sequence[str] | None = None,
+        timeout: float | None = None,
+        deferrable: bool = conf.getboolean("operators", "default_deferrable", fallback=False),
         **kwargs,
     ) -> None:
         super().__init__(**kwargs)
@@ -497,6 +503,8 @@ class CloudDataTransferServiceRunJobOperator(GoogleCloudBaseOperator):
         self.gcp_conn_id = gcp_conn_id
         self.api_version = api_version
         self.google_impersonation_chain = google_impersonation_chain
+        self.timeout = timeout
+        self.deferrable = deferrable
 
     def _validate_inputs(self) -> None:
         if not self.job_name:
@@ -518,8 +526,32 @@ class CloudDataTransferServiceRunJobOperator(GoogleCloudBaseOperator):
                 job_name=self.job_name,
             )
 
+        if self.deferrable:
+            self.defer(
+                timeout=timedelta(seconds=self.timeout or 60),
+                trigger=CloudDataTransferServiceRunJobTrigger(
+                    job_name=self.job_name,
+                    project_id=project_id,
+                    gcp_conn_id=self.gcp_conn_id,
+                    impersonation_chain=self.google_impersonation_chain,
+                ),
+                method_name="execute_complete",
+            )
+
         return hook.run_transfer_job(job_name=self.job_name, project_id=project_id)
 
+    def execute_complete(self, context: Context, event: dict[str, Any]) -> Any:
+        """
+        Act as a callback for when the trigger fires.
+
+        This returns immediately. It relies on trigger to throw an exception,
+        otherwise it assumes execution was successful.
+        """
+        if event["status"] == "error":
+            raise AirflowException(event["message"])
+
+        return event["job_result"]
+
 
 class CloudDataTransferServiceGetOperationOperator(GoogleCloudBaseOperator):
     """
@@ -964,6 +996,7 @@ class CloudDataTransferServiceS3ToGCSOperator(GoogleCloudBaseOperator):
         self.aws_role_arn = aws_role_arn
         self.deferrable = deferrable
         self._validate_inputs()
+        self._transfer_job: dict[str, Any] | None = None
 
     def _validate_inputs(self) -> None:
         if self.delete_job_after_completion and not self.wait:
@@ -978,19 +1011,18 @@ class CloudDataTransferServiceS3ToGCSOperator(GoogleCloudBaseOperator):
 
         TransferJobPreprocessor(body=body, aws_conn_id=self.aws_conn_id, default_schedule=True).process_body()
 
-        job = hook.create_transfer_job(body=body)
-
+        self._transfer_job = hook.create_transfer_job(body=body)
         if self.wait:
             if not self.deferrable:
-                hook.wait_for_transfer_job(job, timeout=self.timeout)
+                hook.wait_for_transfer_job(self._transfer_job, timeout=self.timeout)
                 if self.delete_job_after_completion:
-                    hook.delete_transfer_job(job_name=job[NAME], project_id=self.project_id)
+                    hook.delete_transfer_job(job_name=self._transfer_job[NAME], project_id=self.project_id)
             else:
                 self.defer(
                     timeout=timedelta(seconds=self.timeout or 60),
                     trigger=CloudStorageTransferServiceCheckJobStatusTrigger(
-                        job_name=job[NAME],
-                        project_id=job[PROJECT_ID],
+                        job_name=self._transfer_job[NAME],
+                        project_id=self._transfer_job[PROJECT_ID],
                         gcp_conn_id=self.gcp_conn_id,
                         impersonation_chain=self.google_impersonation_chain,
                     ),
@@ -1040,6 +1072,57 @@ class CloudDataTransferServiceS3ToGCSOperator(GoogleCloudBaseOperator):
 
         return body
 
+    def get_openlineage_facets_on_complete(self, task_instance) -> OperatorLineage | None:
+        """Provide OpenLineage OperatorLineage for the S3->GCS transfer."""
+        from airflow.providers.common.compat.openlineage.facet import Dataset
+        from airflow.providers.google.cloud.openlineage.facets import (
+            CloudStorageTransferJobFacet,
+            CloudStorageTransferRunFacet,
+        )
+        from airflow.providers.openlineage.extractors import OperatorLineage
+
+        input_ds = Dataset(
+            namespace=f"s3://{self.s3_bucket}",
+            name=normalize_directory_path(self.s3_path) or "",
+        )
+
+        output_ds = Dataset(
+            namespace=f"gs://{self.gcs_bucket}",
+            name=normalize_directory_path(self.gcs_path) or "",
+        )
+
+        job = self._transfer_job or {}
+        job_facet = CloudStorageTransferJobFacet(
+            jobName=job.get(NAME),
+            projectId=job.get(PROJECT_ID, self.project_id),
+            description=job.get(DESCRIPTION, self.description),
+            status=job.get(STATUS),
+            sourceBucket=job.get(TRANSFER_SPEC, {})
+            .get(AWS_S3_DATA_SOURCE, {})
+            .get(BUCKET_NAME, self.s3_bucket),
+            sourcePath=job.get(TRANSFER_SPEC, {}).get(AWS_S3_DATA_SOURCE, {}).get(PATH, self.s3_path),
+            targetBucket=job.get(TRANSFER_SPEC, {}).get(GCS_DATA_SINK, {}).get(BUCKET_NAME, self.gcs_bucket),
+            targetPath=job.get(TRANSFER_SPEC, {}).get(GCS_DATA_SINK, {}).get(PATH, self.gcs_path),
+            objectConditions=job.get(TRANSFER_SPEC, {}).get("objectConditions", self.object_conditions),
+            transferOptions=job.get(TRANSFER_SPEC, {}).get("transferOptions", self.transfer_options),
+            schedule=job.get(SCHEDULE, self.schedule),
+        )
+
+        run_facet = CloudStorageTransferRunFacet(
+            jobName=job.get(NAME),
+            wait=self.wait,
+            timeout=self.timeout,
+            deferrable=self.deferrable,
+            deleteJobAfterCompletion=self.delete_job_after_completion,
+        )
+
+        return OperatorLineage(
+            inputs=[input_ds],
+            outputs=[output_ds],
+            job_facets={"cloudStorageTransferJob": job_facet},
+            run_facets={"cloudStorageTransferRun": run_facet},
+        )
+
 
 class CloudDataTransferServiceGCSToGCSOperator(GoogleCloudBaseOperator):
     """