apache-airflow-providers-google 10.14.0rc1__py3-none-any.whl → 10.15.0rc1__py3-none-any.whl

airflow/providers/google/cloud/hooks/gcs.py

@@ -74,7 +74,9 @@ def _fallback_object_url_to_object_name_and_bucket_name(
     object_name_keyword_arg_name="object_name",
 ) -> Callable[[T], T]:
     """
-    Decorator factory that convert object URL parameter to object name and bucket name parameter.
+    Convert object URL parameter to object name and bucket name parameter.
+
+    This method is a Decorator factory.
 
     :param object_url_keyword_arg_name: Name of the object URL parameter
     :param bucket_name_keyword_arg_name: Name of the bucket name parameter
@@ -160,7 +162,7 @@ class GCSHook(GoogleBaseHook):
         )
 
     def get_conn(self) -> storage.Client:
-        """Returns a Google Cloud Storage service object."""
+        """Return a Google Cloud Storage service object."""
         if not self._conn:
             self._conn = storage.Client(
                 credentials=self.get_credentials(), client_info=CLIENT_INFO, project=self.project_id
@@ -176,7 +178,7 @@ class GCSHook(GoogleBaseHook):
         destination_object: str | None = None,
     ) -> None:
         """
-        Copies an object from a bucket to another, with renaming if requested.
+        Copy an object from a bucket to another, with renaming if requested.
 
         destination_bucket or destination_object can be omitted, in which case
         source bucket/object is used, but not both.
@@ -304,7 +306,7 @@ class GCSHook(GoogleBaseHook):
         user_project: str | None = None,
     ) -> str | bytes:
         """
-        Downloads a file from Google Cloud Storage.
+        Download a file from Google Cloud Storage.
 
         When no filename is supplied, the operator loads the file into memory and returns its
         content. When a filename is supplied, it writes the file to the specified location and
@@ -366,7 +368,7 @@ class GCSHook(GoogleBaseHook):
         num_max_attempts: int | None = 1,
     ) -> bytes:
         """
-        Downloads a file from Google Cloud Storage.
+        Download a file from Google Cloud Storage.
 
         When no filename is supplied, the operator loads the file into memory and returns its
         content. When a filename is supplied, it writes the file to the specified location and
@@ -399,7 +401,7 @@ class GCSHook(GoogleBaseHook):
         user_project: str | None = None,
     ) -> Generator[IO[bytes], None, None]:
         """
-        Downloads the file to a temporary directory and returns a file handle.
+        Download the file to a temporary directory and returns a file handle.
 
         You can use this method by passing the bucket_name and object_name parameters
         or just object_url parameter.
@@ -435,7 +437,7 @@ class GCSHook(GoogleBaseHook):
         user_project: str | None = None,
     ) -> Generator[IO[bytes], None, None]:
         """
-        Creates temporary file, returns a file handle and uploads the files content on close.
+        Create temporary file, returns a file handle and uploads the files content on close.
 
         You can use this method by passing the bucket_name and object_name parameters
         or just object_url parameter.
@@ -478,7 +480,7 @@ class GCSHook(GoogleBaseHook):
         user_project: str | None = None,
     ) -> None:
         """
-        Uploads a local file or file data as string or bytes to Google Cloud Storage.
+        Upload a local file or file data as string or bytes to Google Cloud Storage.
 
         :param bucket_name: The bucket to upload to.
         :param object_name: The object name to set when uploading the file.
@@ -498,7 +500,7 @@ class GCSHook(GoogleBaseHook):
 
         def _call_with_retry(f: Callable[[], None]) -> None:
             """
-            Helper functions to upload a file or a string with a retry mechanism and exponential back-off.
+            Upload a file or a string with a retry mechanism and exponential back-off.
 
             :param f: Callable that should be retried.
             """
@@ -572,7 +574,7 @@ class GCSHook(GoogleBaseHook):
 
     def exists(self, bucket_name: str, object_name: str, retry: Retry = DEFAULT_RETRY) -> bool:
         """
-        Checks for the existence of a file in Google Cloud Storage.
+        Check for the existence of a file in Google Cloud Storage.
 
         :param bucket_name: The Google Cloud Storage bucket where the object is.
         :param object_name: The name of the blob_name to check in the Google cloud
@@ -601,7 +603,7 @@ class GCSHook(GoogleBaseHook):
 
     def is_updated_after(self, bucket_name: str, object_name: str, ts: datetime) -> bool:
         """
-        Checks if an blob_name is updated in Google Cloud Storage.
+        Check if an blob_name is updated in Google Cloud Storage.
 
         :param bucket_name: The Google Cloud Storage bucket where the object is.
         :param object_name: The name of the object to check in the Google cloud
@@ -621,7 +623,7 @@ class GCSHook(GoogleBaseHook):
         self, bucket_name: str, object_name: str, min_ts: datetime, max_ts: datetime
     ) -> bool:
         """
-        Checks if an blob_name is updated in Google Cloud Storage.
+        Check if an blob_name is updated in Google Cloud Storage.
 
         :param bucket_name: The Google Cloud Storage bucket where the object is.
         :param object_name: The name of the object to check in the Google cloud
@@ -642,7 +644,7 @@ class GCSHook(GoogleBaseHook):
 
     def is_updated_before(self, bucket_name: str, object_name: str, ts: datetime) -> bool:
         """
-        Checks if an blob_name is updated before given time in Google Cloud Storage.
+        Check if an blob_name is updated before given time in Google Cloud Storage.
 
         :param bucket_name: The Google Cloud Storage bucket where the object is.
         :param object_name: The name of the object to check in the Google cloud
@@ -680,7 +682,7 @@ class GCSHook(GoogleBaseHook):
 
     def delete(self, bucket_name: str, object_name: str) -> None:
         """
-        Deletes an object from the bucket.
+        Delete an object from the bucket.
 
         :param bucket_name: name of the bucket, where the object resides
         :param object_name: name of the object to delete
@@ -954,7 +956,7 @@ class GCSHook(GoogleBaseHook):
 
     def get_size(self, bucket_name: str, object_name: str) -> int:
         """
-        Gets the size of a file in Google Cloud Storage.
+        Get the size of a file in Google Cloud Storage.
 
         :param bucket_name: The Google Cloud Storage bucket where the blob_name is.
         :param object_name: The name of the object to check in the Google
@@ -971,7 +973,7 @@ class GCSHook(GoogleBaseHook):
 
     def get_crc32c(self, bucket_name: str, object_name: str):
         """
-        Gets the CRC32c checksum of an object in Google Cloud Storage.
+        Get the CRC32c checksum of an object in Google Cloud Storage.
 
         :param bucket_name: The Google Cloud Storage bucket where the blob_name is.
         :param object_name: The name of the object to check in the Google cloud
@@ -991,7 +993,7 @@ class GCSHook(GoogleBaseHook):
 
     def get_md5hash(self, bucket_name: str, object_name: str) -> str:
         """
-        Gets the MD5 hash of an object in Google Cloud Storage.
+        Get the MD5 hash of an object in Google Cloud Storage.
 
         :param bucket_name: The Google Cloud Storage bucket where the blob_name is.
         :param object_name: The name of the object to check in the Google cloud
@@ -1016,7 +1018,7 @@ class GCSHook(GoogleBaseHook):
         labels: dict | None = None,
     ) -> str:
         """
-        Creates a new bucket.
+        Create a new bucket.
 
         Google Cloud Storage uses a flat namespace, so you can't
         create a bucket with a name that is already in use.
@@ -1076,7 +1078,7 @@ class GCSHook(GoogleBaseHook):
         self, bucket_name: str, entity: str, role: str, user_project: str | None = None
     ) -> None:
         """
-        Creates a new ACL entry on the specified bucket_name.
+        Create a new ACL entry on the specified bucket_name.
 
         See: https://cloud.google.com/storage/docs/json_api/v1/bucketAccessControls/insert
 
@@ -1111,7 +1113,7 @@ class GCSHook(GoogleBaseHook):
         user_project: str | None = None,
     ) -> None:
         """
-        Creates a new ACL entry on the specified object.
+        Create a new ACL entry on the specified object.
 
         See: https://cloud.google.com/storage/docs/json_api/v1/objectAccessControls/insert
 
@@ -1184,7 +1186,7 @@ class GCSHook(GoogleBaseHook):
         delete_extra_files: bool = False,
     ) -> None:
         """
-        Synchronizes the contents of the buckets.
+        Synchronize the contents of the buckets.
 
         Parameters ``source_object`` and ``destination_object`` describe the root sync directories. If they
         are not passed, the entire bucket will be synchronized. If they are passed, they should point
@@ -1336,15 +1338,22 @@ def gcs_object_is_directory(bucket: str) -> bool:
     return len(blob) == 0 or blob.endswith("/")
 
 
-def parse_json_from_gcs(gcp_conn_id: str, file_uri: str) -> Any:
+def parse_json_from_gcs(
+    gcp_conn_id: str,
+    file_uri: str,
+    impersonation_chain: str | Sequence[str] | None = None,
+) -> Any:
     """
-    Downloads and parses json file from Google cloud Storage.
+    Download and parses json file from Google cloud Storage.
 
     :param gcp_conn_id: Airflow Google Cloud connection ID.
     :param file_uri: full path to json file
         example: ``gs://test-bucket/dir1/dir2/file``
     """
-    gcs_hook = GCSHook(gcp_conn_id=gcp_conn_id)
+    gcs_hook = GCSHook(
+        gcp_conn_id=gcp_conn_id,
+        impersonation_chain=impersonation_chain,
+    )
     bucket, blob = _parse_gcs_url(file_uri)
     with NamedTemporaryFile(mode="w+b") as file:
         try:
@@ -1390,6 +1399,9 @@ class GCSAsyncHook(GoogleBaseAsyncHook):
     sync_hook_class = GCSHook
 
     async def get_storage_client(self, session: ClientSession) -> Storage:
-        """Returns a Google Cloud Storage service object."""
-        with await self.service_file_as_context() as file:
-            return Storage(service_file=file, session=cast(Session, session))
+        """Return a Google Cloud Storage service object."""
+        token = await self.get_token(session=session)
+        return Storage(
+            token=token,
+            session=cast(Session, session),
+        )

airflow/providers/google/cloud/hooks/gdm.py

@@ -49,7 +49,7 @@ class GoogleDeploymentManagerHook(GoogleBaseHook):
         )
 
     def get_conn(self) -> Resource:
-        """Returns a Google Deployment Manager service object."""
+        """Return a Google Deployment Manager service object."""
         http_authorized = self._authorize()
         return build("deploymentmanager", "v2", http=http_authorized, cache_discovery=False)
 
@@ -61,7 +61,7 @@ class GoogleDeploymentManagerHook(GoogleBaseHook):
         order_by: str | None = None,
     ) -> list[dict[str, Any]]:
         """
-        Lists deployments in a google cloud project.
+        List deployments in a google cloud project.
 
         :param project_id: The project ID for this request.
         :param deployment_filter: A filter expression which limits resources returned in the response.
@@ -84,7 +84,7 @@ class GoogleDeploymentManagerHook(GoogleBaseHook):
         self, project_id: str | None, deployment: str | None = None, delete_policy: str | None = None
     ) -> None:
         """
-        Deletes a deployment and all associated resources in a google cloud project.
+        Delete a deployment and all associated resources in a google cloud project.
 
         :param project_id: The project ID for this request.
         :param deployment: The name of the deployment for this request.

airflow/providers/google/cloud/hooks/kms.py

@@ -32,12 +32,12 @@ if TYPE_CHECKING:
 
 
 def _b64encode(s: bytes) -> str:
-    """Base 64 encodes a bytes object to a string."""
+    """Encode a Base64 bytes object to a string."""
     return base64.b64encode(s).decode("ascii")
 
 
 def _b64decode(s: str) -> bytes:
-    """Base 64 decodes a string to bytes."""
+    """Decode a Base64 string to bytes."""
     return base64.b64decode(s.encode("utf-8"))
 
 
@@ -75,7 +75,7 @@ class CloudKMSHook(GoogleBaseHook):
 
     def get_conn(self) -> KeyManagementServiceClient:
         """
-        Retrieves connection to Cloud Key Management service.
+        Retrieve connection to Cloud Key Management service.
 
         :return: Cloud Key Management service object
         """

airflow/providers/google/cloud/hooks/kubernetes_engine.py

@@ -28,10 +28,10 @@ from __future__ import annotations
 import contextlib
 import json
 import time
-import warnings
 from functools import cached_property
 from typing import TYPE_CHECKING, Sequence
 
+from deprecated import deprecated
 from gcloud.aio.auth import Token
 from google.api_core.exceptions import NotFound
 from google.api_core.gapic_v1.method import DEFAULT, _MethodDefault
@@ -48,6 +48,7 @@ from urllib3.exceptions import HTTPError
 
 from airflow import version
 from airflow.exceptions import AirflowException, AirflowProviderDeprecationWarning
+from airflow.providers.cncf.kubernetes.kube_client import _enable_tcp_keepalive
 from airflow.providers.cncf.kubernetes.utils.pod_manager import PodOperatorHookProtocol
 from airflow.providers.google.common.consts import CLIENT_INFO
 from airflow.providers.google.common.hooks.base_google import (
@@ -98,22 +99,23 @@ class GKEHook(GoogleBaseHook):
 
     # To preserve backward compatibility
     # TODO: remove one day
+    @deprecated(
+        reason=(
+            "The get_conn method has been deprecated. "
+            "You should use the get_cluster_manager_client method."
+        ),
+        category=AirflowProviderDeprecationWarning,
+    )
     def get_conn(self) -> container_v1.ClusterManagerClient:
-        warnings.warn(
-            "The get_conn method has been deprecated. You should use the get_cluster_manager_client method.",
-            AirflowProviderDeprecationWarning,
-            stacklevel=2,
-        )
         return self.get_cluster_manager_client()
 
     # To preserve backward compatibility
     # TODO: remove one day
+    @deprecated(
+        reason="The get_client method has been deprecated. You should use the get_conn method.",
+        category=AirflowProviderDeprecationWarning,
+    )
     def get_client(self) -> ClusterManagerClient:
-        warnings.warn(
-            "The get_client method has been deprecated. You should use the get_conn method.",
-            AirflowProviderDeprecationWarning,
-            stacklevel=2,
-        )
         return self.get_conn()
 
     def wait_for_operation(self, operation: Operation, project_id: str | None = None) -> Operation:
@@ -176,7 +178,7 @@ class GKEHook(GoogleBaseHook):
         retry: Retry | _MethodDefault = DEFAULT,
         timeout: float | None = None,
     ) -> Operation | None:
-        """Deletes the cluster, the Kubernetes endpoint, and all worker nodes.
+        """Delete the cluster, the Kubernetes endpoint, and all worker nodes.
 
         Firewalls and routes that were configured during cluster creation are
         also deleted. Other Google Compute Engine resources that might be in use
@@ -352,6 +354,7 @@ class GKEPodHook(GoogleBaseHook, PodOperatorHookProtocol):
         self,
         cluster_url: str,
         ssl_ca_cert: str,
+        disable_tcp_keepalive: bool | None = None,
         gcp_conn_id: str = "google_cloud_default",
         impersonation_chain: str | Sequence[str] | None = None,
         **kwargs,
@@ -363,6 +366,7 @@ class GKEPodHook(GoogleBaseHook, PodOperatorHookProtocol):
         )
         self._cluster_url = cluster_url
         self._ssl_ca_cert = ssl_ca_cert
+        self.disable_tcp_keepalive = disable_tcp_keepalive
 
     @cached_property
     def api_client(self) -> client.ApiClient:
@@ -397,6 +401,10 @@ class GKEPodHook(GoogleBaseHook, PodOperatorHookProtocol):
     def get_conn(self) -> client.ApiClient:
         configuration = self._get_config()
         configuration.refresh_api_key_hook = self._refresh_api_key_hook
+
+        if self.disable_tcp_keepalive is not True:
+            _enable_tcp_keepalive()
+
         return client.ApiClient(configuration)
 
     def _refresh_api_key_hook(self, configuration: client.configuration.Configuration):
@@ -500,14 +508,15 @@ class GKEPodAsyncHook(GoogleBaseAsyncHook):
         :param name: Name of the pod.
         :param namespace: Name of the pod's namespace.
         """
-        async with Token(scopes=self.scopes) as token:
-            async with self.get_conn(token) as connection:
-                v1_api = async_client.CoreV1Api(connection)
-                pod: V1Pod = await v1_api.read_namespaced_pod(
-                    name=name,
-                    namespace=namespace,
-                )
-            return pod
+        with await self.service_file_as_context() as service_file:  # type: ignore[attr-defined]
+            async with Token(scopes=self.scopes, service_file=service_file) as token:
+                async with self.get_conn(token) as connection:
+                    v1_api = async_client.CoreV1Api(connection)
+                    pod: V1Pod = await v1_api.read_namespaced_pod(
+                        name=name,
+                        namespace=namespace,
+                    )
+                return pod
 
     async def delete_pod(self, name: str, namespace: str):
         """Delete a pod.
@@ -515,18 +524,21 @@ class GKEPodAsyncHook(GoogleBaseAsyncHook):
         :param name: Name of the pod.
         :param namespace: Name of the pod's namespace.
         """
-        async with Token(scopes=self.scopes) as token, self.get_conn(token) as connection:
-            try:
-                v1_api = async_client.CoreV1Api(connection)
-                await v1_api.delete_namespaced_pod(
-                    name=name,
-                    namespace=namespace,
-                    body=client.V1DeleteOptions(),
-                )
-            except async_client.ApiException as e:
-                # If the pod is already deleted
-                if e.status != 404:
-                    raise
+        with await self.service_file_as_context() as service_file:  # type: ignore[attr-defined]
+            async with Token(scopes=self.scopes, service_file=service_file) as token, self.get_conn(
+                token
+            ) as connection:
+                try:
+                    v1_api = async_client.CoreV1Api(connection)
+                    await v1_api.delete_namespaced_pod(
+                        name=name,
+                        namespace=namespace,
+                        body=client.V1DeleteOptions(),
+                    )
+                except async_client.ApiException as e:
+                    # If the pod is already deleted
+                    if e.status != 404:
+                        raise
 
     async def read_logs(self, name: str, namespace: str):
         """Read logs inside the pod while starting containers inside.
@@ -539,19 +551,22 @@ class GKEPodAsyncHook(GoogleBaseAsyncHook):
         :param name: Name of the pod.
         :param namespace: Name of the pod's namespace.
         """
-        async with Token(scopes=self.scopes) as token, self.get_conn(token) as connection:
-            try:
-                v1_api = async_client.CoreV1Api(connection)
-                logs = await v1_api.read_namespaced_pod_log(
-                    name=name,
-                    namespace=namespace,
-                    follow=False,
-                    timestamps=True,
-                )
-                logs = logs.splitlines()
-                for line in logs:
-                    self.log.info("Container logs from %s", line)
-                return logs
-            except HTTPError:
-                self.log.exception("There was an error reading the kubernetes API.")
-                raise
+        with await self.service_file_as_context() as service_file:  # type: ignore[attr-defined]
+            async with Token(scopes=self.scopes, service_file=service_file) as token, self.get_conn(
+                token
+            ) as connection:
+                try:
+                    v1_api = async_client.CoreV1Api(connection)
+                    logs = await v1_api.read_namespaced_pod_log(
+                        name=name,
+                        namespace=namespace,
+                        follow=False,
+                        timestamps=True,
+                    )
+                    logs = logs.splitlines()
+                    for line in logs:
+                        self.log.info("Container logs from %s", line)
+                    return logs
+                except HTTPError:
+                    self.log.exception("There was an error reading the kubernetes API.")
+                    raise

airflow/providers/google/cloud/hooks/life_sciences.py

@@ -19,10 +19,10 @@
 from __future__ import annotations
 
 import time
-import warnings
 from typing import Sequence
 
 import google.api_core.path_template
+from deprecated import deprecated
 from googleapiclient.discovery import build
 
 from airflow.exceptions import AirflowException, AirflowProviderDeprecationWarning
@@ -32,6 +32,15 @@ from airflow.providers.google.common.hooks.base_google import GoogleBaseHook
 TIME_TO_SLEEP_IN_SECONDS = 5
 
 
+@deprecated(
+    reason=(
+        "This hook is deprecated. Consider using "
+        "Google Cloud Batch Operators' hook instead. "
+        "The Life Sciences API (beta) will be discontinued "
+        "on July 8, 2025 in favor of Google Cloud Batch."
+    ),
+    category=AirflowProviderDeprecationWarning,
+)
 class LifeSciencesHook(GoogleBaseHook):
     """
     Hook for the Google Cloud Life Sciences APIs.
@@ -76,17 +85,9 @@ class LifeSciencesHook(GoogleBaseHook):
         )
         self.api_version = api_version
 
-        warnings.warn(
-            """This hook is deprecated. Consider using Google Cloud Batch Operators' hook instead.
-            The Life Sciences API (beta) will be discontinued on July 8, 2025 in favor
-            of Google Cloud Batch.""",
-            AirflowProviderDeprecationWarning,
-            stacklevel=3,
-        )
-
     def get_conn(self) -> build:
         """
-        Retrieves the connection to Cloud Life Sciences.
+        Retrieve the connection to Cloud Life Sciences.
 
         :return: Google Cloud Life Sciences service object.
         """
@@ -98,7 +99,7 @@ class LifeSciencesHook(GoogleBaseHook):
     @GoogleBaseHook.fallback_to_default_project_id
     def run_pipeline(self, body: dict, location: str, project_id: str) -> dict:
         """
-        Runs a pipeline.
+        Run a pipeline.
 
         :param body: The request body.
         :param location: The location of the project. For example: "us-east1".
@@ -136,7 +137,7 @@ class LifeSciencesHook(GoogleBaseHook):
 
     def _wait_for_operation_to_complete(self, operation_name: str) -> None:
         """
-        Waits for the named operation to complete - checks status of the asynchronous call.
+        Wait for the named operation to complete - checks status of the asynchronous call.
 
         :param operation_name: The name of the operation.
         :return: The response returned by the operation.

airflow/providers/google/cloud/hooks/looker.py

@@ -41,9 +41,8 @@ class LookerHook(BaseHook):
     def __init__(
         self,
         looker_conn_id: str,
-        **kwargs,
     ) -> None:
-        super().__init__(**kwargs)
+        super().__init__()
         self.looker_conn_id = looker_conn_id
         # source is used to track origin of the requests
         self.source = f"airflow:{version}"
@@ -55,7 +54,7 @@ class LookerHook(BaseHook):
         query_params: dict | None = None,
     ):
         """
-        Submits a PDT materialization job to Looker.
+        Submit a PDT materialization job to Looker.
 
         :param model: Required. The model of the PDT to start building.
         :param view: Required. The view of the PDT to start building.
@@ -84,7 +83,7 @@ class LookerHook(BaseHook):
         materialization_id: str,
     ):
         """
-        Gets the PDT materialization job status from Looker.
+        Get the PDT materialization job status from Looker.
 
         :param materialization_id: Required. The materialization id to check status for.
         """
@@ -101,7 +100,7 @@ class LookerHook(BaseHook):
         materialization_id: str,
     ) -> dict:
         """
-        Gets the PDT materialization job status.
+        Get the PDT materialization job status.
 
         :param materialization_id: Required. The materialization id to check status for.
         """
@@ -121,7 +120,7 @@ class LookerHook(BaseHook):
         materialization_id: str,
     ):
         """
-        Starts a PDT materialization job cancellation request.
+        Start a PDT materialization job cancellation request.
 
         :param materialization_id: Required. The materialization id to stop.
         """
@@ -141,7 +140,7 @@ class LookerHook(BaseHook):
         timeout: int | None = None,
     ) -> None:
         """
-        Helper method which polls a PDT materialization job to check if it finishes.
+        Poll a PDT materialization job to check if it finishes.
 
         :param materialization_id: Required. The materialization id to wait for.
         :param wait_time: Optional. Number of seconds between checks.
@@ -186,7 +185,7 @@ class LookerHook(BaseHook):
         self.log.info("PDT materialization job completed successfully. Job id: %s.", materialization_id)
 
     def get_looker_sdk(self):
-        """Returns Looker SDK client for Looker API 4.0."""
+        """Return Looker SDK client for Looker API 4.0."""
         conn = self.get_connection(self.looker_conn_id)
         settings = LookerApiSettings(conn)
 
@@ -212,7 +211,7 @@ class LookerApiSettings(api_settings.ApiSettings):
 
     def read_config(self):
         """
-        Fetches the connection settings from Airflow's connection object.
+        Fetch the connection settings from Airflow's connection object.
 
         Overrides the default logic of getting connection settings.
         """