433 lines
16 KiB
Python
433 lines
16 KiB
Python
# Copyright 2023 Google LLC
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# https://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
# --------------------------------------------------------------------------------
|
|
# Load The Dependencies
|
|
# --------------------------------------------------------------------------------
|
|
|
|
import datetime
|
|
import time
|
|
|
|
from airflow import models
|
|
from airflow.models.variable import Variable
|
|
from airflow.operators import empty
|
|
from airflow.providers.google.cloud.operators.dataflow import \
|
|
DataflowStartFlexTemplateOperator
|
|
from airflow.providers.google.cloud.operators.bigquery import \
|
|
BigQueryInsertJobOperator, BigQueryUpsertTableOperator, \
|
|
BigQueryUpdateTableSchemaOperator
|
|
from airflow.utils.task_group import TaskGroup
|
|
|
|
# --------------------------------------------------------------------------------
|
|
# Set variables - Needed for the DEMO
|
|
# --------------------------------------------------------------------------------
|
|
BQ_LOCATION = Variable.get("BQ_LOCATION")
|
|
DATA_CAT_TAGS = Variable.get("DATA_CAT_TAGS", deserialize_json=True)
|
|
DWH_LAND_PRJ = Variable.get("DWH_LAND_PRJ")
|
|
DWH_LAND_BQ_DATASET = Variable.get("DWH_LAND_BQ_DATASET")
|
|
DWH_LAND_GCS = Variable.get("DWH_LAND_GCS")
|
|
DWH_CURATED_PRJ = Variable.get("DWH_CURATED_PRJ")
|
|
DWH_CURATED_BQ_DATASET = Variable.get("DWH_CURATED_BQ_DATASET")
|
|
DWH_CURATED_GCS = Variable.get("DWH_CURATED_GCS")
|
|
DWH_CONFIDENTIAL_PRJ = Variable.get("DWH_CONFIDENTIAL_PRJ")
|
|
DWH_CONFIDENTIAL_BQ_DATASET = Variable.get("DWH_CONFIDENTIAL_BQ_DATASET")
|
|
DWH_CONFIDENTIAL_GCS = Variable.get("DWH_CONFIDENTIAL_GCS")
|
|
GCP_REGION = Variable.get("GCP_REGION")
|
|
DRP_PRJ = Variable.get("DRP_PRJ")
|
|
DRP_BQ = Variable.get("DRP_BQ")
|
|
DRP_GCS = Variable.get("DRP_GCS")
|
|
DRP_PS = Variable.get("DRP_PS")
|
|
LOD_PRJ = Variable.get("LOD_PRJ")
|
|
LOD_GCS_STAGING = Variable.get("LOD_GCS_STAGING")
|
|
LOD_NET_VPC = Variable.get("LOD_NET_VPC")
|
|
LOD_NET_SUBNET = Variable.get("LOD_NET_SUBNET")
|
|
LOD_SA_DF = Variable.get("LOD_SA_DF")
|
|
ORC_PRJ = Variable.get("ORC_PRJ")
|
|
ORC_GCS = Variable.get("ORC_GCS")
|
|
ORC_GCS_TMP_DF = Variable.get("ORC_GCS_TMP_DF")
|
|
TRF_PRJ = Variable.get("TRF_PRJ")
|
|
TRF_GCS_STAGING = Variable.get("TRF_GCS_STAGING")
|
|
TRF_NET_VPC = Variable.get("TRF_NET_VPC")
|
|
TRF_NET_SUBNET = Variable.get("TRF_NET_SUBNET")
|
|
TRF_SA_DF = Variable.get("TRF_SA_DF")
|
|
TRF_SA_BQ = Variable.get("TRF_SA_BQ")
|
|
DF_KMS_KEY = Variable.get("DF_KMS_KEY", "")
|
|
DF_REGION = Variable.get("GCP_REGION")
|
|
DF_ZONE = Variable.get("GCP_REGION") + "-b"
|
|
|
|
# --------------------------------------------------------------------------------
|
|
# 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.datetime.now() - datetime.timedelta(days=1)
|
|
|
|
default_args = {
|
|
'owner': 'airflow',
|
|
'start_date': yesterday,
|
|
'depends_on_past': False,
|
|
'email': [''],
|
|
'email_on_failure': False,
|
|
'email_on_retry': False,
|
|
'retries': 1,
|
|
'retry_delay': datetime.timedelta(minutes=5),
|
|
}
|
|
|
|
dataflow_environment = {
|
|
'serviceAccountEmail': LOD_SA_DF,
|
|
'workerZone': DF_ZONE,
|
|
'stagingLocation': f'{LOD_GCS_STAGING}/staging',
|
|
'tempLocation': f'{LOD_GCS_STAGING}/tmp',
|
|
'subnetwork': LOD_NET_SUBNET,
|
|
'kmsKeyName': DF_KMS_KEY,
|
|
'ipConfiguration': 'WORKER_IP_PRIVATE'
|
|
}
|
|
|
|
# --------------------------------------------------------------------------------
|
|
# Main DAG
|
|
# --------------------------------------------------------------------------------
|
|
|
|
with models.DAG('data_pipeline_dc_tags_dag_flex', default_args=default_args,
|
|
schedule_interval=None) as dag:
|
|
start = empty.EmptyOperator(task_id='start', trigger_rule='all_success')
|
|
|
|
end = empty.EmptyOperator(task_id='end', trigger_rule='all_success')
|
|
|
|
# Bigquery Tables created here for demo porpuse.
|
|
# Consider a dedicated pipeline or tool for a real life scenario.
|
|
with TaskGroup('upsert_table') as upsert_table:
|
|
upsert_table_customers = BigQueryUpsertTableOperator(
|
|
task_id="upsert_table_customers",
|
|
project_id=DWH_LAND_PRJ,
|
|
dataset_id=DWH_LAND_BQ_DATASET,
|
|
impersonation_chain=[TRF_SA_DF],
|
|
table_resource={
|
|
"tableReference": {
|
|
"tableId": "customers"
|
|
},
|
|
},
|
|
)
|
|
|
|
upsert_table_purchases = BigQueryUpsertTableOperator(
|
|
task_id="upsert_table_purchases",
|
|
project_id=DWH_LAND_PRJ,
|
|
dataset_id=DWH_LAND_BQ_DATASET,
|
|
impersonation_chain=[TRF_SA_BQ],
|
|
table_resource={"tableReference": {
|
|
"tableId": "purchases"
|
|
}},
|
|
)
|
|
|
|
upsert_table_customer_purchase_curated = BigQueryUpsertTableOperator(
|
|
task_id="upsert_table_customer_purchase_curated",
|
|
project_id=DWH_CURATED_PRJ,
|
|
dataset_id=DWH_CURATED_BQ_DATASET,
|
|
impersonation_chain=[TRF_SA_BQ],
|
|
table_resource={"tableReference": {
|
|
"tableId": "customer_purchase"
|
|
}},
|
|
)
|
|
|
|
upsert_table_customer_purchase_confidential = BigQueryUpsertTableOperator(
|
|
task_id="upsert_table_customer_purchase_confidential",
|
|
project_id=DWH_CONFIDENTIAL_PRJ,
|
|
dataset_id=DWH_CONFIDENTIAL_BQ_DATASET,
|
|
impersonation_chain=[TRF_SA_BQ],
|
|
table_resource={"tableReference": {
|
|
"tableId": "customer_purchase"
|
|
}},
|
|
)
|
|
|
|
# Bigquery Tables schema defined here for demo porpuse.
|
|
# Consider a dedicated pipeline or tool for a real life scenario.
|
|
with TaskGroup('update_schema_table') as update_schema_table:
|
|
update_table_schema_customers = BigQueryUpdateTableSchemaOperator(
|
|
task_id="update_table_schema_customers", project_id=DWH_LAND_PRJ,
|
|
dataset_id=DWH_LAND_BQ_DATASET, table_id="customers",
|
|
impersonation_chain=[TRF_SA_BQ], include_policy_tags=True,
|
|
schema_fields_updates=[{
|
|
"mode": "REQUIRED",
|
|
"name": "id",
|
|
"type": "INTEGER",
|
|
"description": "ID"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "name",
|
|
"type": "STRING",
|
|
"description": "Name",
|
|
"policyTags": {
|
|
"names": [DATA_CAT_TAGS.get('2_Private', None)]
|
|
}
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "surname",
|
|
"type": "STRING",
|
|
"description": "Surname",
|
|
"policyTags": {
|
|
"names": [DATA_CAT_TAGS.get('2_Private', None)]
|
|
}
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "timestamp",
|
|
"type": "TIMESTAMP",
|
|
"description": "Timestamp"
|
|
}])
|
|
|
|
update_table_schema_purchases = BigQueryUpdateTableSchemaOperator(
|
|
task_id="update_table_schema_purchases", project_id=DWH_LAND_PRJ,
|
|
dataset_id=DWH_LAND_BQ_DATASET, table_id="purchases",
|
|
impersonation_chain=[TRF_SA_BQ], include_policy_tags=True,
|
|
schema_fields_updates=[{
|
|
"mode": "REQUIRED",
|
|
"name": "id",
|
|
"type": "INTEGER",
|
|
"description": "ID"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "customer_id",
|
|
"type": "INTEGER",
|
|
"description": "ID"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "item",
|
|
"type": "STRING",
|
|
"description": "Item Name"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "price",
|
|
"type": "FLOAT",
|
|
"description": "Item Price"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "timestamp",
|
|
"type": "TIMESTAMP",
|
|
"description": "Timestamp"
|
|
}])
|
|
|
|
update_table_schema_customer_purchase_curated = BigQueryUpdateTableSchemaOperator(
|
|
task_id="update_table_schema_customer_purchase_curated",
|
|
project_id=DWH_CURATED_PRJ, dataset_id=DWH_CURATED_BQ_DATASET,
|
|
table_id="customer_purchase", impersonation_chain=[TRF_SA_BQ],
|
|
include_policy_tags=True, schema_fields_updates=[{
|
|
"mode": "REQUIRED",
|
|
"name": "customer_id",
|
|
"type": "INTEGER",
|
|
"description": "ID"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "purchase_id",
|
|
"type": "INTEGER",
|
|
"description": "ID"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "name",
|
|
"type": "STRING",
|
|
"description": "Name",
|
|
"policyTags": {
|
|
"names": [DATA_CAT_TAGS.get('2_Private', None)]
|
|
}
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "surname",
|
|
"type": "STRING",
|
|
"description": "Surname",
|
|
"policyTags": {
|
|
"names": [DATA_CAT_TAGS.get('2_Private', None)]
|
|
}
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "item",
|
|
"type": "STRING",
|
|
"description": "Item Name"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "price",
|
|
"type": "FLOAT",
|
|
"description": "Item Price"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "timestamp",
|
|
"type": "TIMESTAMP",
|
|
"description": "Timestamp"
|
|
}])
|
|
|
|
update_table_schema_customer_purchase_confidential = BigQueryUpdateTableSchemaOperator(
|
|
task_id="update_table_schema_customer_purchase_confidential",
|
|
project_id=DWH_CONFIDENTIAL_PRJ, dataset_id=DWH_CONFIDENTIAL_BQ_DATASET,
|
|
table_id="customer_purchase", impersonation_chain=[TRF_SA_BQ],
|
|
include_policy_tags=True, schema_fields_updates=[{
|
|
"mode": "REQUIRED",
|
|
"name": "customer_id",
|
|
"type": "INTEGER",
|
|
"description": "ID"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "purchase_id",
|
|
"type": "INTEGER",
|
|
"description": "ID"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "name",
|
|
"type": "STRING",
|
|
"description": "Name",
|
|
"policyTags": {
|
|
"names": [DATA_CAT_TAGS.get('2_Private', None)]
|
|
}
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "surname",
|
|
"type": "STRING",
|
|
"description": "Surname",
|
|
"policyTags": {
|
|
"names": [DATA_CAT_TAGS.get('2_Private', None)]
|
|
}
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "item",
|
|
"type": "STRING",
|
|
"description": "Item Name"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "price",
|
|
"type": "FLOAT",
|
|
"description": "Item Price"
|
|
}, {
|
|
"mode": "REQUIRED",
|
|
"name": "timestamp",
|
|
"type": "TIMESTAMP",
|
|
"description": "Timestamp"
|
|
}])
|
|
|
|
customers_import = DataflowStartFlexTemplateOperator(
|
|
task_id='dataflow_customers_import', project_id=LOD_PRJ,
|
|
location=DF_REGION, body={
|
|
'launchParameter': {
|
|
'jobName': f'dataflow-customers-import-{round(time.time())}',
|
|
'containerSpecGcsPath': f'{ORC_GCS_TMP_DF}/csv2bq.json',
|
|
'environment': {
|
|
'serviceAccountEmail': LOD_SA_DF,
|
|
'workerZone': DF_ZONE,
|
|
'stagingLocation': f'{LOD_GCS_STAGING}/staging',
|
|
'tempLocation': f'{LOD_GCS_STAGING}/tmp',
|
|
'subnetwork': LOD_NET_SUBNET,
|
|
'kmsKeyName': DF_KMS_KEY,
|
|
'ipConfiguration': 'WORKER_IP_PRIVATE'
|
|
},
|
|
'parameters': {
|
|
'csv_file':
|
|
f'{DRP_GCS}/customers.csv',
|
|
'json_schema':
|
|
f'{ORC_GCS}/customers_schema.json',
|
|
'output_table':
|
|
f'{DWH_LAND_PRJ}:{DWH_LAND_BQ_DATASET}.customers',
|
|
}
|
|
}
|
|
})
|
|
|
|
purchases_import = DataflowStartFlexTemplateOperator(
|
|
task_id='dataflow_purchases_import', project_id=LOD_PRJ,
|
|
location=DF_REGION, body={
|
|
'launchParameter': {
|
|
'jobName': f'dataflow-purchases-import-{round(time.time())}',
|
|
'containerSpecGcsPath': f'{ORC_GCS_TMP_DF}/csv2bq.json',
|
|
'environment': {
|
|
'serviceAccountEmail': LOD_SA_DF,
|
|
'workerZone': DF_ZONE,
|
|
'stagingLocation': f'{LOD_GCS_STAGING}/staging',
|
|
'tempLocation': f'{LOD_GCS_STAGING}/tmp',
|
|
'subnetwork': LOD_NET_SUBNET,
|
|
'kmsKeyName': DF_KMS_KEY,
|
|
'ipConfiguration': 'WORKER_IP_PRIVATE'
|
|
},
|
|
'parameters': {
|
|
'csv_file':
|
|
f'{DRP_GCS}/purchases.csv',
|
|
'json_schema':
|
|
f'{ORC_GCS}/purchases_schema.json',
|
|
'output_table':
|
|
f'{DWH_LAND_PRJ}:{DWH_LAND_BQ_DATASET}.purchases',
|
|
}
|
|
}
|
|
})
|
|
|
|
join_customer_purchase = BigQueryInsertJobOperator(
|
|
task_id='bq_join_customer_purchase', gcp_conn_id='bigquery_default',
|
|
project_id=TRF_PRJ, location=BQ_LOCATION, configuration={
|
|
'jobType': 'QUERY',
|
|
'query': {
|
|
'query':
|
|
"""SELECT
|
|
c.id as customer_id,
|
|
p.id as purchase_id,
|
|
c.name as name,
|
|
c.surname as surname,
|
|
p.item as item,
|
|
p.price as price,
|
|
p.timestamp as timestamp
|
|
FROM `{dwh_0_prj}.{dwh_0_dataset}.customers` c
|
|
JOIN `{dwh_0_prj}.{dwh_0_dataset}.purchases` p ON c.id = p.customer_id
|
|
""".format(
|
|
dwh_0_prj=DWH_LAND_PRJ,
|
|
dwh_0_dataset=DWH_LAND_BQ_DATASET,
|
|
),
|
|
'destinationTable': {
|
|
'projectId': DWH_CURATED_PRJ,
|
|
'datasetId': DWH_CURATED_BQ_DATASET,
|
|
'tableId': 'customer_purchase'
|
|
},
|
|
'writeDisposition':
|
|
'WRITE_APPEND',
|
|
"useLegacySql":
|
|
False
|
|
}
|
|
}, impersonation_chain=[TRF_SA_BQ])
|
|
|
|
confidential_customer_purchase = BigQueryInsertJobOperator(
|
|
task_id='bq_confidential_customer_purchase',
|
|
gcp_conn_id='bigquery_default', project_id=TRF_PRJ, location=BQ_LOCATION,
|
|
configuration={
|
|
'jobType': 'QUERY',
|
|
'query': {
|
|
'query':
|
|
"""SELECT
|
|
customer_id,
|
|
purchase_id,
|
|
name,
|
|
surname,
|
|
item,
|
|
price,
|
|
timestamp
|
|
FROM `{dwh_cur_prj}.{dwh_cur_dataset}.customer_purchase`
|
|
""".format(
|
|
dwh_cur_prj=DWH_CURATED_PRJ,
|
|
dwh_cur_dataset=DWH_CURATED_BQ_DATASET,
|
|
),
|
|
'destinationTable': {
|
|
'projectId': DWH_CONFIDENTIAL_PRJ,
|
|
'datasetId': DWH_CONFIDENTIAL_BQ_DATASET,
|
|
'tableId': 'customer_purchase'
|
|
},
|
|
'writeDisposition':
|
|
'WRITE_APPEND',
|
|
"useLegacySql":
|
|
False
|
|
}
|
|
}, impersonation_chain=[TRF_SA_BQ])
|
|
|
|
start >> upsert_table >> update_schema_table >> [
|
|
customers_import, purchases_import
|
|
] >> join_customer_purchase >> confidential_customer_purchase >> end
|