[GitHub] [airflow] josh-fell commented on a change in pull request #20077: Create CustomJob and Datasets operators for Vertex AI service

[GitBox <gi...@apache.org>]

josh-fell commented on a change in pull request #20077:
URL: https://github.com/apache/airflow/pull/20077#discussion_r763591527



##########
File path: airflow/providers/google/cloud/example_dags/example_vertex_ai.py
##########
@@ -0,0 +1,306 @@
+#
+# 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 that demonstrates operators for the Google Vertex AI service in the Google
+Cloud Platform.
+
+This DAG relies on the following OS environment variables:
+
+* GCP_BUCKET_NAME - Google Cloud Storage bucket where the file exists.

Review comment:
       Should this be "GCP_VERTEX_AI_BUCKET" instead of "GCP_BUCKET_NAME"?

##########
File path: airflow/providers/google/cloud/example_dags/example_vertex_ai.py
##########
@@ -0,0 +1,306 @@
+#
+# 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 that demonstrates operators for the Google Vertex AI service in the Google
+Cloud Platform.
+
+This DAG relies on the following OS environment variables:
+
+* GCP_BUCKET_NAME - Google Cloud Storage bucket where the file exists.
+"""
+import os
+from uuid import uuid4
+
+from airflow import models
+from airflow.providers.google.cloud.operators.vertex_ai.custom_job import (
+    CreateCustomContainerTrainingJobOperator,
+    CreateCustomPythonPackageTrainingJobOperator,
+    CreateCustomTrainingJobOperator,
+    DeleteCustomTrainingJobOperator,
+    ListCustomTrainingJobOperator,
+)
+from airflow.providers.google.cloud.operators.vertex_ai.dataset import (
+    CreateDatasetOperator,
+    DeleteDatasetOperator,
+    ExportDataOperator,
+    GetDatasetOperator,
+    ImportDataOperator,
+    ListDatasetsOperator,
+    UpdateDatasetOperator,
+)
+from airflow.utils.dates import days_ago
+
+PROJECT_ID = os.environ.get("GCP_PROJECT_ID", "an-id")
+REGION = os.environ.get("GCP_LOCATION", "us-central1")
+BUCKET = os.environ.get("GCP_VERTEX_AI_BUCKET", "vertex-ai-system-tests")
+
+STAGING_BUCKET = f"gs://{BUCKET}"
+DISPLAY_NAME = str(uuid4())  # Create random display name
+CONTAINER_URI = "gcr.io/cloud-aiplatform/training/tf-cpu.2-2:latest"
+CUSTOM_CONTAINER_URI = os.environ.get("CUSTOM_CONTAINER_URI", "path_to_container_with_model")
+MODEL_SERVING_CONTAINER_URI = "gcr.io/cloud-aiplatform/prediction/tf2-cpu.2-2:latest"
+REPLICA_COUNT = 1
+MACHINE_TYPE = "n1-standard-4"
+ACCELERATOR_TYPE = "ACCELERATOR_TYPE_UNSPECIFIED"
+ACCELERATOR_COUNT = 0
+TRAINING_FRACTION_SPLIT = 0.7
+TEST_FRACTION_SPLIT = 0.15
+VALIDATION_FRACTION_SPLIT = 0.15
+
+PYTHON_PACKAGE_GCS_URI = os.environ.get("PYTHON_PACKAGE_GSC_URI", "path_to_test_model_in_arch")
+PYTHON_MODULE_NAME = "aiplatform_custom_trainer_script.task"
+
+LOCAL_TRAINING_SCRIPT_PATH = os.environ.get("LOCAL_TRAINING_SCRIPT_PATH", "path_to_training_script")
+
+TRAINING_PIPELINE_ID = "test-training-pipeline-id"
+CUSTOM_JOB_ID = "test-custom-job-id"
+
+IMAGE_DATASET = {
+    "display_name": str(uuid4()),
+    "metadata_schema_uri": "gs://google-cloud-aiplatform/schema/dataset/metadata/image_1.0.0.yaml",
+    "metadata": "test-image-dataset",
+}
+TABULAR_DATASET = {
+    "display_name": str(uuid4()),
+    "metadata_schema_uri": "gs://google-cloud-aiplatform/schema/dataset/metadata/tabular_1.0.0.yaml",
+    "metadata": "test-tabular-dataset",
+}
+TEXT_DATASET = {
+    "display_name": str(uuid4()),
+    "metadata_schema_uri": "gs://google-cloud-aiplatform/schema/dataset/metadata/text_1.0.0.yaml",
+    "metadata": "test-text-dataset",
+}
+VIDEO_DATASET = {
+    "display_name": str(uuid4()),
+    "metadata_schema_uri": "gs://google-cloud-aiplatform/schema/dataset/metadata/video_1.0.0.yaml",
+    "metadata": "test-video-dataset",
+}
+TIME_SERIES_DATASET = {
+    "display_name": str(uuid4()),
+    "metadata_schema_uri": "gs://google-cloud-aiplatform/schema/dataset/metadata/time_series_1.0.0.yaml",
+    "metadata": "test-video-dataset",
+}
+DATASET_ID = os.environ.get("DATASET_ID", "test-dataset-id")
+TEST_EXPORT_CONFIG = {"gcs_destination": {"output_uri_prefix": "gs://test-vertex-ai-bucket/exports"}}
+TEST_IMPORT_CONFIG = [
+    {
+        "data_item_labels": {
+            "test-labels-name": "test-labels-value",
+        },
+        "import_schema_uri": (
+            "gs://google-cloud-aiplatform/schema/dataset/ioformat/image_bounding_box_io_format_1.0.0.yaml"
+        ),
+        "gcs_source": {
+            "uris": ["gs://ucaip-test-us-central1/dataset/salads_oid_ml_use_public_unassigned.jsonl"]
+        },
+    },
+]
+DATASET_TO_UPDATE = {"display_name": "test-name"}
+TEST_UPDATE_MASK = {"paths": ["displayName"]}
+
+with models.DAG(
+    "example_gcp_vertex_ai_custom_jobs",
+    start_date=days_ago(1),

Review comment:
       Applicable to both DAGs in this file:
   
   - Example DAGs should use a static `start_date` value instead of a dynamic one as best practice. The value doesn't matter.
   
   - Also, we've been adding `catchup=False` to initially ward off new-user headaches when these example DAGs are copied and `start_date` is modified to fit needs. 

##########
File path: airflow/providers/google/cloud/example_dags/example_vertex_ai.py
##########
@@ -0,0 +1,306 @@
+#
+# 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 that demonstrates operators for the Google Vertex AI service in the Google
+Cloud Platform.
+
+This DAG relies on the following OS environment variables:
+
+* GCP_BUCKET_NAME - Google Cloud Storage bucket where the file exists.
+"""
+import os
+from uuid import uuid4
+
+from airflow import models
+from airflow.providers.google.cloud.operators.vertex_ai.custom_job import (
+    CreateCustomContainerTrainingJobOperator,
+    CreateCustomPythonPackageTrainingJobOperator,
+    CreateCustomTrainingJobOperator,
+    DeleteCustomTrainingJobOperator,
+    ListCustomTrainingJobOperator,
+)
+from airflow.providers.google.cloud.operators.vertex_ai.dataset import (
+    CreateDatasetOperator,
+    DeleteDatasetOperator,
+    ExportDataOperator,
+    GetDatasetOperator,
+    ImportDataOperator,
+    ListDatasetsOperator,
+    UpdateDatasetOperator,
+)
+from airflow.utils.dates import days_ago
+
+PROJECT_ID = os.environ.get("GCP_PROJECT_ID", "an-id")
+REGION = os.environ.get("GCP_LOCATION", "us-central1")
+BUCKET = os.environ.get("GCP_VERTEX_AI_BUCKET", "vertex-ai-system-tests")
+
+STAGING_BUCKET = f"gs://{BUCKET}"
+DISPLAY_NAME = str(uuid4())  # Create random display name
+CONTAINER_URI = "gcr.io/cloud-aiplatform/training/tf-cpu.2-2:latest"
+CUSTOM_CONTAINER_URI = os.environ.get("CUSTOM_CONTAINER_URI", "path_to_container_with_model")
+MODEL_SERVING_CONTAINER_URI = "gcr.io/cloud-aiplatform/prediction/tf2-cpu.2-2:latest"
+REPLICA_COUNT = 1
+MACHINE_TYPE = "n1-standard-4"
+ACCELERATOR_TYPE = "ACCELERATOR_TYPE_UNSPECIFIED"
+ACCELERATOR_COUNT = 0
+TRAINING_FRACTION_SPLIT = 0.7
+TEST_FRACTION_SPLIT = 0.15
+VALIDATION_FRACTION_SPLIT = 0.15
+
+PYTHON_PACKAGE_GCS_URI = os.environ.get("PYTHON_PACKAGE_GSC_URI", "path_to_test_model_in_arch")
+PYTHON_MODULE_NAME = "aiplatform_custom_trainer_script.task"
+
+LOCAL_TRAINING_SCRIPT_PATH = os.environ.get("LOCAL_TRAINING_SCRIPT_PATH", "path_to_training_script")
+
+TRAINING_PIPELINE_ID = "test-training-pipeline-id"
+CUSTOM_JOB_ID = "test-custom-job-id"
+
+IMAGE_DATASET = {
+    "display_name": str(uuid4()),
+    "metadata_schema_uri": "gs://google-cloud-aiplatform/schema/dataset/metadata/image_1.0.0.yaml",
+    "metadata": "test-image-dataset",
+}
+TABULAR_DATASET = {
+    "display_name": str(uuid4()),
+    "metadata_schema_uri": "gs://google-cloud-aiplatform/schema/dataset/metadata/tabular_1.0.0.yaml",
+    "metadata": "test-tabular-dataset",
+}
+TEXT_DATASET = {
+    "display_name": str(uuid4()),
+    "metadata_schema_uri": "gs://google-cloud-aiplatform/schema/dataset/metadata/text_1.0.0.yaml",
+    "metadata": "test-text-dataset",
+}
+VIDEO_DATASET = {
+    "display_name": str(uuid4()),
+    "metadata_schema_uri": "gs://google-cloud-aiplatform/schema/dataset/metadata/video_1.0.0.yaml",
+    "metadata": "test-video-dataset",
+}
+TIME_SERIES_DATASET = {
+    "display_name": str(uuid4()),
+    "metadata_schema_uri": "gs://google-cloud-aiplatform/schema/dataset/metadata/time_series_1.0.0.yaml",
+    "metadata": "test-video-dataset",
+}
+DATASET_ID = os.environ.get("DATASET_ID", "test-dataset-id")
+TEST_EXPORT_CONFIG = {"gcs_destination": {"output_uri_prefix": "gs://test-vertex-ai-bucket/exports"}}
+TEST_IMPORT_CONFIG = [
+    {
+        "data_item_labels": {
+            "test-labels-name": "test-labels-value",
+        },
+        "import_schema_uri": (
+            "gs://google-cloud-aiplatform/schema/dataset/ioformat/image_bounding_box_io_format_1.0.0.yaml"
+        ),
+        "gcs_source": {
+            "uris": ["gs://ucaip-test-us-central1/dataset/salads_oid_ml_use_public_unassigned.jsonl"]
+        },
+    },
+]
+DATASET_TO_UPDATE = {"display_name": "test-name"}
+TEST_UPDATE_MASK = {"paths": ["displayName"]}
+
+with models.DAG(
+    "example_gcp_vertex_ai_custom_jobs",

Review comment:
       Right now all of these task will execute in parallel and have no dependencies on each other. If there is a use case in which some of these tasks _could_ flow together in a pipeline it would be nice to demonstrate here.




-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@airflow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org