composer-dags/demo_pipeline_composer.py [34:95]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    DataformCreateCompilationResultOperator,
    DataformCreateWorkflowInvocationOperator,
)
from airflow.providers.google.cloud.operators.dataproc import (
    DataprocCreateBatchOperator,
    DataprocGetBatchOperator
)
from airflow.providers.google.cloud.sensors.dataproc import DataprocBatchSensor
from google.cloud.dataform_v1beta1 import WorkflowInvocation
from google.cloud import storage
from datetime import datetime, timedelta

# --------------------------------------------------------------------------------
# Read variables from GCS parameters file for the job
# --------------------------------------------------------------------------------
storage_client = storage.Client()
jobs_bucket = Variable.get("DATA_TRANSFORMATION_GCS_BUCKET")

batch_id = f"aef-{str(uuid.uuid4())}"

def extract_job_params(job_name, function_name, encoding='utf-8'):
    """Extracts parameters from a JSON job file.

    Args:
        bucket_name: Bucket containing the JSON parameters file .

    Returns:
        A dictionary containing the extracted parameters.
    """

    json_file_path = f'gs://{jobs_bucket}/{function_name}/{job_name}.json'

    parts = json_file_path.replace("gs://", "").split("/")
    bucket_name = parts[0]
    object_name = "/".join(parts[1:])
    bucket = storage_client.bucket(bucket_name)
    blob = bucket.blob(object_name)

    json_data = blob.download_as_bytes()
    params = json.loads(json_data.decode(encoding))
    return params


# --------------------------------------------------------------------------------
# Set default arguments
# --------------------------------------------------------------------------------

# If you are running Airflow in more than one time zone
# see https://airflow.apache.org/docs/apache-airflow/stable/timezone.html
# for best practices
yesterday = datetime.now() - timedelta(days=1)

default_args = {
    'owner': 'airflow',
    'start_date': yesterday,
    'depends_on_past': False,
    'email': [''],
    'email_on_failure': False,
    'email_on_retry': False,
    'retries': 0,
    'retry_delay': timedelta(minutes=5)
}
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



workflows-generator/composer-templates/workflow.py [34:95]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    DataformCreateCompilationResultOperator,
    DataformCreateWorkflowInvocationOperator,
)
from airflow.providers.google.cloud.operators.dataproc import (
    DataprocCreateBatchOperator,
    DataprocGetBatchOperator
)
from airflow.providers.google.cloud.sensors.dataproc import DataprocBatchSensor
from google.cloud.dataform_v1beta1 import WorkflowInvocation
from google.cloud import storage
from datetime import datetime, timedelta

# --------------------------------------------------------------------------------
# Read variables from GCS parameters file for the job
# --------------------------------------------------------------------------------
storage_client = storage.Client()
jobs_bucket = Variable.get("DATA_TRANSFORMATION_GCS_BUCKET")

batch_id = f"aef-{str(uuid.uuid4())}"

def extract_job_params(job_name, function_name, encoding='utf-8'):
    """Extracts parameters from a JSON job file.

    Args:
        bucket_name: Bucket containing the JSON parameters file .

    Returns:
        A dictionary containing the extracted parameters.
    """

    json_file_path = f'gs://{jobs_bucket}/{function_name}/{job_name}.json'

    parts = json_file_path.replace("gs://", "").split("/")
    bucket_name = parts[0]
    object_name = "/".join(parts[1:])
    bucket = storage_client.bucket(bucket_name)
    blob = bucket.blob(object_name)

    json_data = blob.download_as_bytes()
    params = json.loads(json_data.decode(encoding))
    return params


# --------------------------------------------------------------------------------
# Set default arguments
# --------------------------------------------------------------------------------

# If you are running Airflow in more than one time zone
# see https://airflow.apache.org/docs/apache-airflow/stable/timezone.html
# for best practices
yesterday = datetime.now() - timedelta(days=1)

default_args = {
    'owner': 'airflow',
    'start_date': yesterday,
    'depends_on_past': False,
    'email': [''],
    'email_on_failure': False,
    'email_on_retry': False,
    'retries': 0,
    'retry_delay': timedelta(minutes=5)
}
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -