#
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you 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
#
# http://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.
"""
Example Airflow DAG for Google ML Engine service.
"""
from __future__ import annotations
import os
from datetime import datetime
from google.cloud.aiplatform import schema
from google.protobuf.json_format import ParseDict
from google.protobuf.struct_pb2 import Value
from airflow import models
from airflow.operators.bash import BashOperator
from airflow.providers.google.cloud.operators.gcs import (
GCSCreateBucketOperator,
GCSDeleteBucketOperator,
GCSSynchronizeBucketsOperator,
)
from airflow.providers.google.cloud.operators.vertex_ai.batch_prediction_job import (
CreateBatchPredictionJobOperator,
DeleteBatchPredictionJobOperator,
)
from airflow.providers.google.cloud.operators.vertex_ai.custom_job import (
CreateCustomPythonPackageTrainingJobOperator,
)
from airflow.providers.google.cloud.operators.vertex_ai.dataset import (
CreateDatasetOperator,
DeleteDatasetOperator,
)
from airflow.providers.google.cloud.operators.vertex_ai.model_service import (
DeleteModelOperator,
DeleteModelVersionOperator,
GetModelOperator,
ListModelVersionsOperator,
SetDefaultVersionOnModelOperator,
)
from airflow.utils.trigger_rule import TriggerRule
[docs]PROJECT_ID = os.environ.get("SYSTEM_TESTS_GCP_PROJECT", "default")
[docs]ENV_ID = os.environ.get("SYSTEM_TESTS_ENV_ID", "default")
[docs]DAG_ID = "example_gcp_mlengine"
[docs]PACKAGE_DISPLAY_NAME = f"package-{DAG_ID}-{ENV_ID}".replace("_", "-")
[docs]MODEL_DISPLAY_NAME = f"model-{DAG_ID}-{ENV_ID}".replace("_", "-")
[docs]JOB_DISPLAY_NAME = f"batch_job_{DAG_ID}_{ENV_ID}".replace("-", "_")
[docs]RESOURCE_DATA_BUCKET = "airflow-system-tests-resources"
[docs]CUSTOM_PYTHON_GCS_BUCKET_NAME = f"bucket_python_{DAG_ID}_{ENV_ID}".replace("_", "-")
[docs]BQ_SOURCE = "bq://bigquery-public-data.ml_datasets.penguins"
[docs]TABULAR_DATASET = {
"display_name": f"tabular-dataset-{ENV_ID}",
"metadata_schema_uri": schema.dataset.metadata.tabular,
"metadata": ParseDict(
{"input_config": {"bigquery_source": {"uri": BQ_SOURCE}}},
Value(),
),
}
[docs]MACHINE_TYPE = "n1-standard-4"
[docs]ACCELERATOR_TYPE = "ACCELERATOR_TYPE_UNSPECIFIED"
[docs]TRAINING_FRACTION_SPLIT = 0.7
[docs]TEST_FRACTION_SPLIT = 0.15
[docs]VALIDATION_FRACTION_SPLIT = 0.15
[docs]PYTHON_PACKAGE_GCS_URI = f"gs://{CUSTOM_PYTHON_GCS_BUCKET_NAME}/vertex-ai/penguins_trainer_script-0.1.zip"
[docs]PYTHON_MODULE_NAME = "penguins_trainer_script.task"
[docs]TRAIN_IMAGE = "us-docker.pkg.dev/vertex-ai/training/tf-cpu.2-8:latest"
[docs]DEPLOY_IMAGE = "us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest"
with models.DAG(
dag_id=DAG_ID,
schedule="@once",
start_date=datetime(2021, 1, 1),
catchup=False,
tags=["example", "ml_engine"],
) as dag:
[docs] create_bucket = GCSCreateBucketOperator(
task_id="create_bucket",
bucket_name=CUSTOM_PYTHON_GCS_BUCKET_NAME,
storage_class="REGIONAL",
location=REGION,
)
move_data_files = GCSSynchronizeBucketsOperator(
task_id="move_files_to_bucket",
source_bucket=RESOURCE_DATA_BUCKET,
source_object="vertex-ai/penguins-data",
destination_bucket=CUSTOM_PYTHON_GCS_BUCKET_NAME,
destination_object="vertex-ai",
recursive=True,
)
create_tabular_dataset = CreateDatasetOperator(
task_id="tabular_dataset",
dataset=TABULAR_DATASET,
region=REGION,
project_id=PROJECT_ID,
)
tabular_dataset_id = create_tabular_dataset.output["dataset_id"]
# [START howto_operator_create_custom_python_training_job_v1]
create_custom_python_package_training_job = CreateCustomPythonPackageTrainingJobOperator(
task_id="create_custom_python_package_training_job",
staging_bucket=f"gs://{CUSTOM_PYTHON_GCS_BUCKET_NAME}",
display_name=PACKAGE_DISPLAY_NAME,
python_package_gcs_uri=PYTHON_PACKAGE_GCS_URI,
python_module_name=PYTHON_MODULE_NAME,
container_uri=TRAIN_IMAGE,
model_serving_container_image_uri=DEPLOY_IMAGE,
bigquery_destination=f"bq://{PROJECT_ID}",
# run params
dataset_id=tabular_dataset_id,
model_display_name=MODEL_DISPLAY_NAME,
replica_count=REPLICA_COUNT,
machine_type=MACHINE_TYPE,
accelerator_type=ACCELERATOR_TYPE,
accelerator_count=ACCELERATOR_COUNT,
training_fraction_split=TRAINING_FRACTION_SPLIT,
validation_fraction_split=VALIDATION_FRACTION_SPLIT,
test_fraction_split=TEST_FRACTION_SPLIT,
region=REGION,
project_id=PROJECT_ID,
)
# [END howto_operator_create_custom_python_training_job_v1]
model_id_v1 = create_custom_python_package_training_job.output["model_id"]
# [START howto_operator_gcp_mlengine_get_model]
get_model = GetModelOperator(
task_id="get_model", region=REGION, project_id=PROJECT_ID, model_id=model_id_v1
)
# [END howto_operator_gcp_mlengine_get_model]
# [START howto_operator_gcp_mlengine_print_model]
get_model_result = BashOperator(
bash_command=f"echo {get_model.output}",
task_id="get_model_result",
)
# [END howto_operator_gcp_mlengine_print_model]
# [START howto_operator_create_custom_python_training_job_v2]
create_custom_python_package_training_job_v2 = CreateCustomPythonPackageTrainingJobOperator(
task_id="create_custom_python_package_training_job_v2",
staging_bucket=f"gs://{CUSTOM_PYTHON_GCS_BUCKET_NAME}",
display_name=PACKAGE_DISPLAY_NAME,
python_package_gcs_uri=PYTHON_PACKAGE_GCS_URI,
python_module_name=PYTHON_MODULE_NAME,
container_uri=TRAIN_IMAGE,
model_serving_container_image_uri=DEPLOY_IMAGE,
bigquery_destination=f"bq://{PROJECT_ID}",
parent_model=model_id_v1,
# run params
dataset_id=tabular_dataset_id,
model_display_name=MODEL_DISPLAY_NAME,
replica_count=REPLICA_COUNT,
machine_type=MACHINE_TYPE,
accelerator_type=ACCELERATOR_TYPE,
accelerator_count=ACCELERATOR_COUNT,
training_fraction_split=TRAINING_FRACTION_SPLIT,
validation_fraction_split=VALIDATION_FRACTION_SPLIT,
test_fraction_split=TEST_FRACTION_SPLIT,
region=REGION,
project_id=PROJECT_ID,
)
# [END howto_operator_create_custom_python_training_job_v2]
model_id_v2 = create_custom_python_package_training_job_v2.output["model_id"]
# [START howto_operator_gcp_mlengine_default_version]
set_default_version = SetDefaultVersionOnModelOperator(
task_id="set_default_version",
project_id=PROJECT_ID,
region=REGION,
model_id=model_id_v2,
)
# [END howto_operator_gcp_mlengine_default_version]
# [START howto_operator_gcp_mlengine_list_versions]
list_model_versions = ListModelVersionsOperator(
task_id="list_model_versions", region=REGION, project_id=PROJECT_ID, model_id=model_id_v2
)
# [END howto_operator_gcp_mlengine_list_versions]
# [START howto_operator_start_batch_prediction]
create_batch_prediction_job = CreateBatchPredictionJobOperator(
task_id="create_batch_prediction_job",
job_display_name=JOB_DISPLAY_NAME,
model_name=model_id_v2,
predictions_format="bigquery",
bigquery_source=BQ_SOURCE,
bigquery_destination_prefix=f"bq://{PROJECT_ID}",
region=REGION,
project_id=PROJECT_ID,
machine_type=MACHINE_TYPE,
)
# [END howto_operator_start_batch_prediction]
# [START howto_operator_gcp_mlengine_delete_version]
delete_model_version_1 = DeleteModelVersionOperator(
task_id="delete_model_version_1",
project_id=PROJECT_ID,
region=REGION,
model_id=model_id_v2,
trigger_rule=TriggerRule.ALL_DONE,
)
# [END howto_operator_gcp_mlengine_delete_version]
# [START howto_operator_gcp_mlengine_delete_model]
delete_model = DeleteModelOperator(
task_id="delete_model",
project_id=PROJECT_ID,
region=REGION,
model_id=model_id_v1,
trigger_rule=TriggerRule.ALL_DONE,
)
# [END howto_operator_gcp_mlengine_delete_model]
delete_batch_prediction_job = DeleteBatchPredictionJobOperator(
task_id="delete_batch_prediction_job",
batch_prediction_job_id=create_batch_prediction_job.output["batch_prediction_job_id"],
region=REGION,
project_id=PROJECT_ID,
trigger_rule=TriggerRule.ALL_DONE,
)
delete_tabular_dataset = DeleteDatasetOperator(
task_id="delete_tabular_dataset",
dataset_id=tabular_dataset_id,
region=REGION,
project_id=PROJECT_ID,
trigger_rule=TriggerRule.ALL_DONE,
)
delete_bucket = GCSDeleteBucketOperator(
task_id="delete_bucket",
bucket_name=CUSTOM_PYTHON_GCS_BUCKET_NAME,
trigger_rule=TriggerRule.ALL_DONE,
)
(
# TEST SETUP
create_bucket
>> move_data_files
>> create_tabular_dataset
# TEST BODY
>> create_custom_python_package_training_job
>> create_custom_python_package_training_job_v2
>> create_batch_prediction_job
>> get_model
>> get_model_result
>> list_model_versions
>> set_default_version
# TEST TEARDOWN
>> delete_model_version_1
>> delete_model
>> delete_batch_prediction_job
>> delete_tabular_dataset
>> delete_bucket
)
from tests.system.utils.watcher import watcher
# This test needs watcher in order to properly mark success/failure
# when "tearDown" task with trigger rule is part of the DAG
list(dag.tasks) >> watcher()
from tests.system.utils import get_test_run # noqa: E402
# Needed to run the example DAG with pytest (see: tests/system/README.md#run_via_pytest)
[docs]test_run = get_test_run(dag)