fix: component paths

notebooks/official/pipelines/rapid_prototyping_bqml_automl.ipynb

@@ -247,7 +247,7 @@
       "outputs": [],
       "source": [
         "# Install Python package dependencies.\n",
-        "! pip3 install --quiet 'google-cloud-pipeline-components==1.0.20' 'kfp<2'\n",
+        "! pip3 install --quiet google-cloud-pipeline-components kfp\n",
         "! pip3 install --quiet --upgrade google-cloud-aiplatform google-cloud-bigquery"
       ]
     },
@@ -457,7 +457,7 @@
       },
       "outputs": [],
       "source": [
-        "! gsutil mb -l $REGION -p $PROJECT_ID $BUCKET_URI"
+        "! gsutil mb -l {REGION} -p {PROJECT_ID} {BUCKET_URI}"
       ]
     },
     {
@@ -556,13 +556,15 @@
         "from typing import NamedTuple\n",
         "\n",
         "from google.cloud import aiplatform as vertex\n",
-        "from google_cloud_pipeline_components import \\\n",
-        "    aiplatform as vertex_pipeline_components\n",
-        "from google_cloud_pipeline_components.experimental import \\\n",
-        "    bigquery as bq_components\n",
-        "from kfp import dsl\n",
-        "from kfp.v2 import compiler\n",
-        "from kfp.v2.dsl import Artifact, Input, Metrics, Output, component"
+        "from google_cloud_pipeline_components.v1 import bigquery as bq_components\n",
+        "from google_cloud_pipeline_components.v1.automl.training_job import \\\n",
+        "    AutoMLTabularTrainingJobRunOp\n",
+        "from google_cloud_pipeline_components.v1.dataset import TabularDatasetCreateOp\n",
+        "from google_cloud_pipeline_components.v1.endpoint import (EndpointCreateOp,\n",
+        "                                                          ModelDeployOp)\n",
+        "from google_cloud_pipeline_components.v1.model import ModelUploadOp\n",
+        "from kfp import compiler, dsl\n",
+        "from kfp.dsl import Artifact, Input, Metrics, Output, component"
       ]
     },
     {
@@ -595,7 +597,7 @@
       },
       "outputs": [],
       "source": [
-        "PIPELINE_JSON_PKG_PATH = \"rapid_prototyping.json\"\n",
+        "PIPELINE_YAML_PKG_PATH = \"rapid_prototyping.yaml\"\n",
         "PIPELINE_ROOT = f\"{BUCKET_URI}/pipeline_root\"\n",
         "DATA_FOLDER = f\"{BUCKET_URI[5:]}/data\"\n",
         "\n",
@@ -1317,18 +1319,26 @@
         "    endpoint_display_name: str,\n",
         "    thresholds_dict_str: str,\n",
         "):\n",
+        "    from google_cloud_pipeline_components.types import artifact_types\n",
+        "    from kfp.dsl import importer_node\n",
+        "\n",
         "    # Imports data to BigQuery using a custom component.\n",
         "    import_data_to_bigquery_op = import_data_to_bigquery(\n",
-        "        project, bq_location, bq_dataset, gcs_input_file_uri\n",
+        "        project=project,\n",
+        "        bq_location=bq_location,\n",
+        "        bq_dataset=bq_dataset,\n",
+        "        gcs_data_uri=gcs_input_file_uri,\n",
         "    )\n",
         "    raw_dataset = import_data_to_bigquery_op.outputs[\"raw_dataset\"]\n",
         "\n",
         "    # Splits the BQ dataset using a custom component.\n",
-        "    split_datasets_op = split_datasets(raw_dataset, bq_location=bq_location)\n",
+        "    split_datasets_op = split_datasets(raw_dataset=raw_dataset, bq_location=bq_location)\n",
         "\n",
         "    # Generates the query to create a BQML using a static function.\n",
         "    create_model_query = _query_create_model(\n",
-        "        project, bq_dataset, split_datasets_op.outputs[\"dataset_uri\"]\n",
+        "        project_id=project,\n",
+        "        bq_dataset=bq_dataset,\n",
+        "        training_data_uri=split_datasets_op.outputs[\"dataset_uri\"],\n",
         "    )\n",
         "\n",
         "    # Builds BQML model using pre-built-component.\n",
@@ -1358,18 +1368,27 @@
         "    ).after(bqml_evaluate_op)\n",
         "    bqml_exported_gcs_path = bqml_export_op.outputs[\"exported_model_path\"]\n",
         "\n",
+        "    unmanaged_model_importer = importer_node.importer(\n",
+        "        artifact_uri=bqml_exported_gcs_path,\n",
+        "        artifact_class=artifact_types.UnmanagedContainerModel,\n",
+        "        metadata={\n",
+        "            \"containerSpec\": {\n",
+        "                \"imageUri\": \"us-docker.pkg.dev/cloud-aiplatform/prediction/tf2-cpu.2-3:latest\"\n",
+        "            }\n",
+        "        },\n",
+        "    )\n",
+        "\n",
         "    # Uploads the recently exported the BQML model from GCS into Vertex AI using a pre-built-component.\n",
-        "    bqml_model_upload_op = vertex_pipeline_components.ModelUploadOp(\n",
+        "    bqml_model_upload_op = ModelUploadOp(\n",
         "        project=project,\n",
         "        location=region,\n",
         "        display_name=DISPLAY_NAME + \"_bqml\",\n",
-        "        artifact_uri=bqml_exported_gcs_path,\n",
-        "        serving_container_image_uri=bqml_serving_container_image_uri,\n",
+        "        unmanaged_container_model=unmanaged_model_importer.outputs[\"artifact\"],\n",
         "    )\n",
         "    bqml_vertex_model = bqml_model_upload_op.outputs[\"model\"]\n",
         "\n",
         "    # Creates a Vertex AI Tabular dataset using a pre-built-component.\n",
-        "    dataset_create_op = vertex_pipeline_components.TabularDatasetCreateOp(\n",
+        "    dataset_create_op = TabularDatasetCreateOp(\n",
         "        project=project,\n",
         "        location=region,\n",
         "        display_name=DISPLAY_NAME,\n",
@@ -1377,7 +1396,7 @@
         "    )\n",
         "\n",
         "    # Trains an AutoML Tables model using a pre-built-component.\n",
-        "    automl_training_op = vertex_pipeline_components.AutoMLTabularTrainingJobRunOp(\n",
+        "    automl_training_op = AutoMLTabularTrainingJobRunOp(\n",
         "        project=project,\n",
         "        location=region,\n",
         "        display_name=f\"{DISPLAY_NAME}_automl\",\n",
@@ -1420,7 +1439,7 @@
         "        name=\"deploy_decision\",\n",
         "    ):\n",
         "        # Creates a Vertex AI endpoint using a pre-built-component.\n",
-        "        endpoint_create_op = vertex_pipeline_components.EndpointCreateOp(\n",
+        "        endpoint_create_op = EndpointCreateOp(\n",
         "            project=project,\n",
         "            location=region,\n",
         "            display_name=endpoint_display_name,\n",
@@ -1433,19 +1452,17 @@
         "            name=\"deploy_bqml\",\n",
         "        ):\n",
         "            # Deploys the BQML model (now on Vertex AI) to the recently created endpoint using a pre-built component.\n",
-        "            model_deploy_bqml_op = (\n",
-        "                vertex_pipeline_components.ModelDeployOp(  # noqa: F841\n",
-        "                    endpoint=endpoint_create_op.outputs[\"endpoint\"],\n",
-        "                    model=bqml_vertex_model,\n",
-        "                    deployed_model_display_name=DISPLAY_NAME + \"_best_bqml\",\n",
-        "                    dedicated_resources_machine_type=\"n1-standard-2\",\n",
-        "                    dedicated_resources_min_replica_count=2,\n",
-        "                    dedicated_resources_max_replica_count=2,\n",
-        "                    traffic_split={\n",
-        "                        \"0\": 100\n",
-        "                    },  # newly deployed model gets 100% of the traffic\n",
-        "                ).set_caching_options(False)\n",
-        "            )\n",
+        "            model_deploy_bqml_op = ModelDeployOp(  # noqa: F841\n",
+        "                endpoint=endpoint_create_op.outputs[\"endpoint\"],\n",
+        "                model=bqml_vertex_model,\n",
+        "                deployed_model_display_name=DISPLAY_NAME + \"_best_bqml\",\n",
+        "                dedicated_resources_machine_type=\"n1-standard-2\",\n",
+        "                dedicated_resources_min_replica_count=2,\n",
+        "                dedicated_resources_max_replica_count=2,\n",
+        "                traffic_split={\n",
+        "                    \"0\": 100\n",
+        "                },  # newly deployed model gets 100% of the traffic\n",
+        "            ).set_caching_options(False)\n",
         "\n",
         "            # Sends an online prediction request to the recently deployed model using a custom component.\n",
         "            validate_infrastructure(\n",
@@ -1458,19 +1475,17 @@
         "            name=\"deploy_automl\",\n",
         "        ):\n",
         "            # Deploys the AutoML model to the recently created endpoint using a pre-built component.\n",
-        "            model_deploy_automl_op = (\n",
-        "                vertex_pipeline_components.ModelDeployOp(  # noqa: F841\n",
-        "                    endpoint=endpoint_create_op.outputs[\"endpoint\"],\n",
-        "                    model=automl_model,\n",
-        "                    deployed_model_display_name=DISPLAY_NAME + \"_best_automl\",\n",
-        "                    dedicated_resources_machine_type=\"n1-standard-2\",\n",
-        "                    dedicated_resources_min_replica_count=2,\n",
-        "                    dedicated_resources_max_replica_count=2,\n",
-        "                    traffic_split={\n",
-        "                        \"0\": 100\n",
-        "                    },  # newly deployed model gets 100% of the traffic\n",
-        "                ).set_caching_options(False)\n",
-        "            )\n",
+        "            model_deploy_automl_op = ModelDeployOp(  # noqa: F841\n",
+        "                endpoint=endpoint_create_op.outputs[\"endpoint\"],\n",
+        "                model=automl_model,\n",
+        "                deployed_model_display_name=DISPLAY_NAME + \"_best_automl\",\n",
+        "                dedicated_resources_machine_type=\"n1-standard-2\",\n",
+        "                dedicated_resources_min_replica_count=2,\n",
+        "                dedicated_resources_max_replica_count=2,\n",
+        "                traffic_split={\n",
+        "                    \"0\": 100\n",
+        "                },  # newly deployed model gets 100% of the traffic\n",
+        "            ).set_caching_options(False)\n",
         "\n",
         "            # Sends an online prediction request to the recently deployed model using a custom component.\n",
         "            validate_infrastructure(\n",
@@ -1497,7 +1512,7 @@
       "source": [
         "compiler.Compiler().compile(\n",
         "    pipeline_func=train_pipeline,\n",
-        "    package_path=PIPELINE_JSON_PKG_PATH,\n",
+        "    package_path=PIPELINE_YAML_PKG_PATH,\n",
         ")\n",
         "\n",
         "\n",
@@ -1505,7 +1520,7 @@
         "\n",
         "pipeline_job = vertex.PipelineJob(\n",
         "    display_name=DISPLAY_NAME,\n",
-        "    template_path=PIPELINE_JSON_PKG_PATH,\n",
+        "    template_path=PIPELINE_YAML_PKG_PATH,\n",
         "    pipeline_root=PIPELINE_ROOT,\n",
         "    parameter_values=pipeline_params,\n",
         "    enable_caching=False,\n",