Source code for tests.system.providers.google.cloud.vertex_ai.example_vertex_ai_custom_job

#
# 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.

# mypy ignore arg types (for templated fields)
# type: ignore[arg-type]

"""
Example Airflow DAG for Google Vertex AI service testing Custom Jobs operations.
"""
from __future__ import annotations

import os
from datetime import datetime
from pathlib import Path

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
from airflow.providers.google.cloud.operators.vertex_ai.custom_job import (
    CreateCustomTrainingJobOperator,
    DeleteCustomTrainingJobOperator,
)
from airflow.providers.google.cloud.operators.vertex_ai.dataset import (
    CreateDatasetOperator,
    DeleteDatasetOperator,
)
from airflow.providers.google.cloud.transfers.local_to_gcs import LocalFilesystemToGCSOperator
from airflow.utils.trigger_rule import TriggerRule

[docs]ENV_ID = os.environ.get("SYSTEM_TESTS_ENV_ID")
[docs]DAG_ID = "vertex_ai_custom_job_operations"
[docs]PROJECT_ID = os.environ.get("SYSTEM_TESTS_GCP_PROJECT", "default")
[docs]REGION = "us-central1"
[docs]CUSTOM_DISPLAY_NAME = f"train-housing-custom-{ENV_ID}"
[docs]MODEL_DISPLAY_NAME = f"custom-housing-model-{ENV_ID}"
[docs]CUSTOM_GCS_BUCKET_NAME = f"bucket_{DAG_ID}_{ENV_ID}"
[docs]DATA_SAMPLE_GCS_OBJECT_NAME = "vertex-ai/california_housing_train.csv"
[docs]CSV_ZIP_FILE_LOCAL_PATH = str(Path(__file__).parent / "resources" / "California-housing.zip")
[docs]CSV_FILE_LOCAL_PATH = "/custom-job/california_housing_train.csv"
[docs]TABULAR_DATASET = lambda bucket_name: { "display_name": f"tabular-dataset-{ENV_ID}", "metadata_schema_uri": schema.dataset.metadata.tabular, "metadata": ParseDict( {"input_config": {"gcs_source": {"uri": [f"gs://{bucket_name}/{DATA_SAMPLE_GCS_OBJECT_NAME}"]}}}, Value(),
), }
[docs]CONTAINER_URI = "gcr.io/cloud-aiplatform/training/tf-cpu.2-2:latest"
[docs]MODEL_SERVING_CONTAINER_URI = "gcr.io/cloud-aiplatform/prediction/tf2-cpu.2-2:latest"
[docs]REPLICA_COUNT = 1
[docs]LOCAL_TRAINING_SCRIPT_PATH = "/custom-job/california_housing_training_script.py"
with models.DAG( f"{DAG_ID}_custom", schedule="@once", start_date=datetime(2021, 1, 1), catchup=False, tags=["example", "vertex_ai", "custom_job"], ) as dag:
[docs] create_bucket = GCSCreateBucketOperator( task_id="create_bucket", bucket_name=CUSTOM_GCS_BUCKET_NAME, storage_class="REGIONAL", location=REGION,
) unzip_file = BashOperator( task_id="unzip_csv_data_file", bash_command=f"mkdir -p /custom-job && unzip {CSV_ZIP_FILE_LOCAL_PATH} -d /custom-job/", ) upload_files = LocalFilesystemToGCSOperator( task_id="upload_file_to_bucket", src=CSV_FILE_LOCAL_PATH, dst=DATA_SAMPLE_GCS_OBJECT_NAME, bucket=CUSTOM_GCS_BUCKET_NAME, ) create_tabular_dataset = CreateDatasetOperator( task_id="tabular_dataset", dataset=TABULAR_DATASET(CUSTOM_GCS_BUCKET_NAME), region=REGION, project_id=PROJECT_ID, ) tabular_dataset_id = create_tabular_dataset.output["dataset_id"] # [START how_to_cloud_vertex_ai_create_custom_training_job_operator] create_custom_training_job = CreateCustomTrainingJobOperator( task_id="custom_task", staging_bucket=f"gs://{CUSTOM_GCS_BUCKET_NAME}", display_name=CUSTOM_DISPLAY_NAME, script_path=LOCAL_TRAINING_SCRIPT_PATH, container_uri=CONTAINER_URI, requirements=["gcsfs==0.7.1"], model_serving_container_image_uri=MODEL_SERVING_CONTAINER_URI, # run params dataset_id=tabular_dataset_id, replica_count=REPLICA_COUNT, model_display_name=MODEL_DISPLAY_NAME, sync=False, region=REGION, project_id=PROJECT_ID, ) # [END how_to_cloud_vertex_ai_create_custom_training_job_operator] # [START how_to_cloud_vertex_ai_delete_custom_training_job_operator] delete_custom_training_job = DeleteCustomTrainingJobOperator( task_id="delete_custom_training_job", training_pipeline_id=create_custom_training_job.output["training_id"], custom_job_id=create_custom_training_job.output["custom_job_id"], region=REGION, project_id=PROJECT_ID, trigger_rule=TriggerRule.ALL_DONE, ) # [END how_to_cloud_vertex_ai_delete_custom_training_job_operator] 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_GCS_BUCKET_NAME, trigger_rule=TriggerRule.ALL_DONE, ) clear_folder = BashOperator( task_id="clear_folder", bash_command="rm -r /custom-job/*", ) ( # TEST SETUP create_bucket >> unzip_file >> upload_files >> create_tabular_dataset # TEST BODY >> create_custom_training_job # TEST TEARDOWN >> delete_custom_training_job >> delete_tabular_dataset >> delete_bucket >> clear_folder ) 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)

Was this entry helpful?