def main(project_id,
zone,
cluster_name,
bucket_name,
pyspark_file=None,
create_new_cluster=True,
global_region=True):
# [START dataproc_get_client]
if global_region:
region = 'global'
# Use the default gRPC global endpoints.
dataproc_cluster_client = dataproc_v1.ClusterControllerClient()
dataproc_job_client = dataproc_v1.JobControllerClient()
else:
region = get_region_from_zone(zone)
# Use a regional gRPC endpoint. See:
# https://cloud.google.com/dataproc/docs/concepts/regional-endpoints
client_transport = (
cluster_controller_grpc_transport.ClusterControllerGrpcTransport(
address='{}-dataproc.googleapis.com:443'.format(region)))
job_transport = (
job_controller_grpc_transport.JobControllerGrpcTransport(
address='{}-dataproc.googleapis.com:443'.format(region)))
dataproc_cluster_client = dataproc_v1.ClusterControllerClient(
client_transport)
dataproc_job_client = dataproc_v1.JobControllerClient(job_transport)
# [END dataproc_get_client]
try:
spark_file, spark_filename = get_pyspark_file(pyspark_file)
if create_new_cluster:
create_cluster(dataproc_cluster_client, project_id, zone, region,
cluster_name)
wait_for_cluster_creation()
upload_pyspark_file(project_id, bucket_name, spark_filename,
spark_file)
list_clusters_with_details(dataproc_cluster_client, project_id, region)
(cluster_id,
output_bucket) = (get_cluster_id_by_name(dataproc_cluster_client,
project_id, region,
cluster_name))
# [START dataproc_call_submit_pyspark_job]
job_id = submit_pyspark_job(dataproc_job_client, project_id, region,
cluster_name, bucket_name, spark_filename)
# [END dataproc_call_submit_pyspark_job]
wait_for_job(dataproc_job_client, project_id, region, job_id)
output = download_output(project_id, cluster_id, output_bucket, job_id)
print('Received job output {}'.format(output))
return output
finally:
if create_new_cluster:
delete_cluster(dataproc_cluster_client, project_id, region,
cluster_name)
spark_file.close()
def main(project_id, region):
if region == "global":
# Use the default gRPC global endpoints.
dataproc_cluster_client = dataproc_v1.ClusterControllerClient()
else:
# Use a regional gRPC endpoint. See:
# https://cloud.google.com/dataproc/docs/concepts/regional-endpoints
client_transport = cluster_controller_grpc_transport.ClusterControllerGrpcTransport(
address="{}-dataproc.googleapis.com:443".format(region)
)
dataproc_cluster_client = dataproc_v1.ClusterControllerClient(client_transport)
list_clusters(dataproc_cluster_client, project_id, region)
def test_delete_cluster(self):
# Setup Expected Response
expected_response = {}
expected_response = empty_pb2.Empty(**expected_response)
operation = operations_pb2.Operation(
name='operations/test_delete_cluster', done=True)
operation.response.Pack(expected_response)
# Mock the API response
channel = ChannelStub(responses=[operation])
client = dataproc_v1.ClusterControllerClient(channel=channel)
# Setup Request
project_id = 'projectId-1969970175'
region = 'region-934795532'
cluster_name = 'clusterName-1018081872'
response = client.delete_cluster(project_id, region, cluster_name)
result = response.result()
assert expected_response == result
assert len(channel.requests) == 1
expected_request = clusters_pb2.DeleteClusterRequest(
project_id=project_id, region=region, cluster_name=cluster_name)
actual_request = channel.requests[0][1]
assert expected_request == actual_request
def post(self, request):
print("List of clusters initiated ......")
os.environ[
"GOOGLE_APPLICATION_CREDENTIALS"] = "C:\\Users\\t\\keys.json"
project_id = "deepak-cloud-trail"
zone = request.POST["zone"]
region = get_region_from_zone(zone)
zone_uri = \
'https://www.googleapis.com/compute/v1/projects/{}/zones/{}'.format(
project_id, zone)
client_transport = (
cluster_controller_grpc_transport.ClusterControllerGrpcTransport(
address='{}-dataproc.googleapis.com:443'.format(region)))
dataproc_client = dataproc_v1.ClusterControllerClient(client_transport)
cluster_name = request.POST["cluster_name"]
cluster_data = {
'project_id': project_id,
'cluster_name': cluster_name,
'config': {
'gce_cluster_config': {
'zone_uri': zone_uri
},
'master_config': {
'num_instances': 1,
'machine_type_uri': 'n1-standard-1'
},
'worker_config': {
'num_instances': 2,
'machine_type_uri': 'n1-standard-1'
}
}
}
cluster = dataproc_client.create_cluster(project_id, region,
cluster_data)
def __init__(self, cluster_metadata: MasterURLIdentifier) -> None:
"""Initializes the DataprocClusterManager with properties required
to interface with the Dataproc ClusterControllerClient.
"""
self.cluster_metadata = cluster_metadata
if self.cluster_metadata.region == 'global':
# The global region is unsupported as it will be eventually deprecated.
raise ValueError('Clusters in the global region are not supported.')
elif not self.cluster_metadata.region:
_LOGGER.warning(
'No region information was detected, defaulting Dataproc cluster '
'region to: us-central1.')
self.cluster_metadata.region = 'us-central1'
if not self.cluster_metadata.cluster_name:
self.cluster_metadata.cluster_name = ie.current_env(
).clusters.default_cluster_name
self._cluster_client = dataproc_v1.ClusterControllerClient(
client_options={
'api_endpoint': \
f'{self.cluster_metadata.region}-dataproc.googleapis.com:443'
})
if self.cluster_metadata in ie.current_env().clusters.master_urls.inverse:
self.master_url = ie.current_env().clusters.master_urls.inverse[
self.cluster_metadata]
self.dashboard = ie.current_env().clusters.master_urls_to_dashboards[
self.master_url]
else:
self.master_url = None
self.dashboard = None
self._fs = gcsfilesystem.GCSFileSystem(PipelineOptions())
self._staging_directory = None
def create_cluster(project_id, region, cluster_name, create_buckets=None):
# Create a client with the endpoint set to the desired cluster region.
cluster_client = dataproc.ClusterControllerClient(
client_options={
"api_endpoint": f"{region}-dataproc.googleapis.com:443"
})
staging_bucket = None
tmp_bucket = None
if create_buckets:
staging_bucket = create_bucket_if_not_exists(
project_id, region, f'{cluster_name}-staging')
tmp_bucket = create_bucket_if_not_exists(project_id, region,
f'{cluster_name}-tmp')
# Create the cluster config.
cluster = create_dataproc_config(project_id, cluster_name, region,
staging_bucket, tmp_bucket)
# Create the cluster.
operation = cluster_client.create_cluster(request={
"project_id": project_id,
"region": region,
"cluster": cluster
})
result = operation.result()
# Output a success message.
print(f"Cluster created successfully: {result.cluster_name}")
def post(request):
cluster_client = dataproc.ClusterControllerClient(
client_options={
'api_endpoint': 'europe-west2-dataproc.googleapis.com:443'
})
create_cluster(cluster_client, "bootcamp-bdmlv", "europe-west2-a",
"europe-west2", "hive")
def test_list_clusters(self):
project_id = os.environ["PROJECT_ID"]
client = dataproc_v1.ClusterControllerClient()
project_id_2 = project_id
region = "global"
response = client.list_clusters(project_id_2, region)
def setup_teardown():
storage_client = storage.Client()
bucket = storage_client.create_bucket(STAGING_BUCKET)
blob = bucket.blob(JOB_FILE_NAME)
blob.upload_from_string(SORT_CODE)
yield
cluster_client = dataproc.ClusterControllerClient(
client_options={"api_endpoint": "{}-dataproc.googleapis.com:443".format(REGION)}
)
# The quickstart sample deletes the cluster, but if the test fails
# before cluster deletion occurs, it can be manually deleted here.
clusters = cluster_client.list_clusters(
request={"project_id": PROJECT_ID, "region": REGION}
)
for cluster in clusters:
if cluster.cluster_name == CLUSTER_NAME:
cluster_client.delete_cluster(
request={
"project_id": PROJECT_ID,
"region": REGION,
"cluster_name": CLUSTER_NAME,
}
)
blob.delete()
bucket.delete()
def setup_and_teardown_cluster():
try:
# Create cluster using cluster client
cluster_client = dataproc.ClusterControllerClient(
client_options={
"api_endpoint": f"{CLUSTER_REGION}-dataproc.googleapis.com:443"
})
operation = cluster_client.create_cluster(project_id=PROJECT_ID,
region=CLUSTER_REGION,
cluster=CLUSTER_CONFIG)
# Wait for cluster to provision
operation.result()
yield
finally:
try:
# Delete cluster
operation = cluster_client.delete_cluster(
project_id=PROJECT_ID,
region=CLUSTER_REGION,
cluster_name=DATAPROC_CLUSTER)
operation.result()
except NotFound:
print("Cluster already deleted")
def test_create_cluster(self):
# Setup Expected Response
project_id_2 = 'projectId2939242356'
cluster_name = 'clusterName-1018081872'
cluster_uuid = 'clusterUuid-1017854240'
expected_response = {
'project_id': project_id_2,
'cluster_name': cluster_name,
'cluster_uuid': cluster_uuid
}
expected_response = clusters_pb2.Cluster(**expected_response)
operation = operations_pb2.Operation(
name='operations/test_create_cluster', done=True)
operation.response.Pack(expected_response)
# Mock the API response
channel = ChannelStub(responses=[operation])
client = dataproc_v1.ClusterControllerClient(channel=channel)
# Setup Request
project_id = 'projectId-1969970175'
region = 'region-934795532'
cluster = {}
response = client.create_cluster(project_id, region, cluster)
result = response.result()
assert expected_response == result
assert len(channel.requests) == 1
expected_request = clusters_pb2.CreateClusterRequest(
project_id=project_id, region=region, cluster=cluster)
actual_request = channel.requests[0][1]
assert expected_request == actual_request
def test_get_cluster(self):
# Setup Expected Response
project_id_2 = 'projectId2939242356'
cluster_name_2 = 'clusterName2875867491'
cluster_uuid = 'clusterUuid-1017854240'
expected_response = {
'project_id': project_id_2,
'cluster_name': cluster_name_2,
'cluster_uuid': cluster_uuid
}
expected_response = clusters_pb2.Cluster(**expected_response)
# Mock the API response
channel = ChannelStub(responses=[expected_response])
patch = mock.patch('google.api_core.grpc_helpers.create_channel')
with patch as create_channel:
create_channel.return_value = channel
client = dataproc_v1.ClusterControllerClient()
# Setup Request
project_id = 'projectId-1969970175'
region = 'region-934795532'
cluster_name = 'clusterName-1018081872'
response = client.get_cluster(project_id, region, cluster_name)
assert expected_response == response
assert len(channel.requests) == 1
expected_request = clusters_pb2.GetClusterRequest(
project_id=project_id, region=region, cluster_name=cluster_name)
actual_request = channel.requests[0][1]
assert expected_request == actual_request
def test_diagnose_cluster(self):
# Setup Expected Response
expected_response = {}
expected_response = empty_pb2.Empty(**expected_response)
operation = operations_pb2.Operation(
name='operations/test_diagnose_cluster', done=True)
operation.response.Pack(expected_response)
# Mock the API response
channel = ChannelStub(responses=[operation])
patch = mock.patch('google.api_core.grpc_helpers.create_channel')
with patch as create_channel:
create_channel.return_value = channel
client = dataproc_v1.ClusterControllerClient()
# Setup Request
project_id = 'projectId-1969970175'
region = 'region-934795532'
cluster_name = 'clusterName-1018081872'
response = client.diagnose_cluster(project_id, region, cluster_name)
result = response.result()
assert expected_response == result
assert len(channel.requests) == 1
expected_request = clusters_pb2.DiagnoseClusterRequest(
project_id=project_id, region=region, cluster_name=cluster_name)
actual_request = channel.requests[0][1]
assert expected_request == actual_request
def test_list_clusters(self):
# Setup Expected Response
next_page_token = ''
clusters_element = {}
clusters = [clusters_element]
expected_response = {
'next_page_token': next_page_token,
'clusters': clusters
}
expected_response = clusters_pb2.ListClustersResponse(
**expected_response)
# Mock the API response
channel = ChannelStub(responses=[expected_response])
patch = mock.patch('google.api_core.grpc_helpers.create_channel')
with patch as create_channel:
create_channel.return_value = channel
client = dataproc_v1.ClusterControllerClient()
# Setup Request
project_id = 'projectId-1969970175'
region = 'region-934795532'
paged_list_response = client.list_clusters(project_id, region)
resources = list(paged_list_response)
assert len(resources) == 1
assert expected_response.clusters[0] == resources[0]
assert len(channel.requests) == 1
expected_request = clusters_pb2.ListClustersRequest(
project_id=project_id, region=region)
actual_request = channel.requests[0][1]
assert expected_request == actual_request
def test_update_cluster_exception(self):
# Setup Response
error = status_pb2.Status()
operation = operations_pb2.Operation(
name='operations/test_update_cluster_exception', done=True)
operation.error.CopyFrom(error)
# Mock the API response
channel = ChannelStub(responses=[operation])
patch = mock.patch('google.api_core.grpc_helpers.create_channel')
with patch as create_channel:
create_channel.return_value = channel
client = dataproc_v1.ClusterControllerClient()
# Setup Request
project_id = 'projectId-1969970175'
region = 'region-934795532'
cluster_name = 'clusterName-1018081872'
cluster = {}
update_mask = {}
response = client.update_cluster(project_id, region, cluster_name,
cluster, update_mask)
exception = response.exception()
assert exception.errors[0] == error
def create_cluster(project_id, region, cluster_name):
"""This sample walks a user through creating a Cloud Dataproc cluster
using the Python client library.
Args:
project_id (string): Project to use for creating resources.
region (string): Region where the resources should live.
cluster_name (string): Name to use for creating a cluster.
"""
# Create a client with the endpoint set to the desired cluster region.
cluster_client = dataproc.ClusterControllerClient(
client_options={"api_endpoint": f"{region}-dataproc.googleapis.com:443"}
)
# Create the cluster config.
cluster = {
"project_id": project_id,
"cluster_name": cluster_name,
"config": {
"master_config": {"num_instances": 1, "machine_type_uri": "n1-standard-1"},
"worker_config": {"num_instances": 2, "machine_type_uri": "n1-standard-1"},
},
}
# Create the cluster.
operation = cluster_client.create_cluster(
request={"project_id": project_id, "region": region, "cluster": cluster}
)
result = operation.result()
# Output a success message.
print(f"Cluster created successfully: {result.cluster_name}")
def teardown():
yield
client = dataproc.ClusterControllerClient(
client_options={
'api_endpoint': '{}-dataproc.googleapis.com:443'.format(REGION)
})
# Client library function
client.delete_cluster(PROJECT_ID, REGION, CLUSTER_NAME)
def post(request):
cluster_client = dataproc.ClusterControllerClient(
client_options={
'api_endpoint': 'europe-west2-dataproc.googleapis.com:443'
})
operation = cluster_client.delete_cluster("bootcamp-bdmlv", "europe-west2",
"hive")
return operation.result()
def dataproc_cluster_client(zone):
"""
Lazily create a Dataproc ClusterController client to setup or
tear down dataproc clusters
"""
region = get_region_from_zone(zone)
client_transport = cluster_controller_grpc_transport.ClusterControllerGrpcTransport(
address="{}-dataproc.googleapis.com:443".format(region))
return dataproc_v1.ClusterControllerClient(client_transport)
def test_list_clusters_exception(self):
channel = ChannelStub(responses=[CustomException()])
client = dataproc_v1.ClusterControllerClient(channel=channel)
# Setup request
project_id = 'projectId-1969970175'
region = 'region-934795532'
paged_list_response = client.list_clusters(project_id, region)
with pytest.raises(CustomException):
list(paged_list_response)
def hello_pubsub(event, context):
cluster_transport = cluster_controller_grpc_transport.ClusterControllerGrpcTransport(
address='us-central1-dataproc.googleapis.com:443')
dataproc_cluster_client = dataproc_v1.ClusterControllerClient(
cluster_transport)
project_id = ''
region = 'us-central1'
pubsub_message = base64.b64decode(event['data']).decode('utf-8')
data = json.loads(pubsub_message)
cluster_name = data['resource']['labels']['cluster_name']
print(cluster_name + " dataproc cluster created")
def teardown():
yield
cluster_client = dataproc.ClusterControllerClient(
client_options={
'api_endpoint': f'{REGION}-dataproc.googleapis.com:443'
})
# Client library function
operation = cluster_client.delete_cluster(PROJECT_ID, REGION, CLUSTER_NAME)
# Wait for cluster to delete
operation.result()
def test_get_cluster_exception(self):
# Mock the API response
channel = ChannelStub(responses=[CustomException()])
client = dataproc_v1.ClusterControllerClient(channel=channel)
# Setup request
project_id = 'projectId-1969970175'
region = 'region-934795532'
cluster_name = 'clusterName-1018081872'
with pytest.raises(CustomException):
client.get_cluster(project_id, region, cluster_name)
def dataproc_cluster_client(self):
"""
Lazily create a Dataproc ClusterController client to setup or
tear down dataproc clusters
"""
if self._dataproc_cluster_client is None:
client_transport = cluster_controller_grpc_transport.ClusterControllerGrpcTransport(
address="{}-dataproc.googleapis.com:443".format(self._region))
self._dataproc_cluster_client = dataproc_v1.ClusterControllerClient(
client_transport)
return self._dataproc_cluster_client
def main(event, context):
'''
Triggered by a change to a Cloud Storage bucket.
:param event:
:param context:
:return:
'''
# Variables
PROJECT_ID = 'sas-ivnard'
CLUSTER_NAME = 'score-spark-demo'
BUCKET_NAME = 'network-spark-migrate'
REGION = 'europe-west6'
ZONE = 'europe-west6-b'
PIP_PACKAGES = "PyYAML==5.3.1 numpy==1.19.4 pandas==1.1.4 pyspark==3.0.1"
JOB_ID = 'Batch_Model_Score'
logging.basicConfig(
format='%(asctime)s %(name)-12s %(levelname)-8s %(message)s',
datefmt='%m/%d/%Y %I:%M:%S %p',
level=logging.INFO)
# Create a client with the endpoint set to the desired cluster region.
cluster_client = dataproc.ClusterControllerClient(
client_options={
"api_endpoint": f"{REGION}-dataproc.googleapis.com:443"
})
# Because we upload several files in a loop, the function will be trigger n times where
# n is the number of files uploaded. We need to create a filter.
# Notice: I have to investigate more. For demo purposes, it's enough.
if event[
'name'] == 'data/ML-MATT-CompetitionQT1920_val_processed.parquet/_SUCCESS':
logging.info("A new scoring process is starting...")
logging.info(f"Creating cluster {CLUSTER_NAME}...")
cluster = create_cluster(PROJECT_ID, CLUSTER_NAME, BUCKET_NAME, REGION,
ZONE, PIP_PACKAGES)
logging.info(f"Submitting job {JOB_ID}...")
cluster.add_done_callback(lambda _: submit_score_job(
PROJECT_ID, CLUSTER_NAME, REGION, JOB_ID))
while check_job_state(PROJECT_ID, REGION, JOB_ID) != 'state.done':
logging.info(f"Job {JOB_ID} is running...")
time.sleep(5)
logging.info(f"Job {JOB_ID} is done!")
logging.info(f"Deleting cluster {CLUSTER_NAME}...")
delete_cluster(PROJECT_ID, CLUSTER_NAME, REGION)
while check_if_cluster(PROJECT_ID, CLUSTER_NAME, REGION) is not None:
logging.info(f"Deleting {CLUSTER_NAME}...")
time.sleep(2)
logging.info(f"Cluster {CLUSTER_NAME} deleted!")
def __init__(self, cluster_metadata: ClusterMetadata) -> None:
"""Initializes the DataprocClusterManager with properties required
to interface with the Dataproc ClusterControllerClient.
"""
self.cluster_metadata = cluster_metadata
# Pipelines whose jobs are executed on the cluster.
self.pipelines = set()
self._cluster_client = dataproc_v1.ClusterControllerClient(
client_options={
'api_endpoint': \
f'{self.cluster_metadata.region}-dataproc.googleapis.com:443'
})
self._fs = gcsfilesystem.GCSFileSystem(PipelineOptions())
self._staging_directory = None
def test_list_clusters_exception(self):
channel = ChannelStub(responses=[CustomException()])
patch = mock.patch('google.api_core.grpc_helpers.create_channel')
with patch as create_channel:
create_channel.return_value = channel
client = dataproc_v1.ClusterControllerClient()
# Setup request
project_id = 'projectId-1969970175'
region = 'region-934795532'
paged_list_response = client.list_clusters(project_id, region)
with pytest.raises(CustomException):
list(paged_list_response)
def delete_cluster():
# Create a client with the endpoint set to the desired cluster region.
cluster_client = dataproc.ClusterControllerClient(client_options={
'api_endpoint':
'{}-dataproc.googleapis.com:443'.format('europe-west1')
})
# Create the cluster.
operation = cluster_client.delete_cluster('big-data-architecture-ricardo',
'europe-west1', 'dataproc-bda')
result = operation.result()
# Output a success message.
return 'Cluster deleted successfully'
def teardown():
yield
cluster_client = dataproc.ClusterControllerClient(
client_options={
"api_endpoint": f"{REGION}-dataproc.googleapis.com:443"
})
# Client library function
operation = cluster_client.delete_cluster(request={
"project_id": PROJECT_ID,
"region": REGION,
"cluster_name": CLUSTER_NAME,
})
# Wait for cluster to delete
operation.result()
def test_get_cluster_exception(self):
# Mock the API response
channel = ChannelStub(responses=[CustomException()])
patch = mock.patch('google.api_core.grpc_helpers.create_channel')
with patch as create_channel:
create_channel.return_value = channel
client = dataproc_v1.ClusterControllerClient()
# Setup request
project_id = 'projectId-1969970175'
region = 'region-934795532'
cluster_name = 'clusterName-1018081872'
with pytest.raises(CustomException):
client.get_cluster(project_id, region, cluster_name)
