fix(components): Add staging and temp locations to prophet trainer component

Signed-off-by: Michael Hu <humichael@google.com>
PiperOrigin-RevId: 631958163

Author: TheMichaelHu

components/google-cloud/RELEASE.md

@@ -1,6 +1,7 @@
 ## Upcoming release
 
 ## Release 2.15.0
+* Add staging and temp location parameters to prophet trainer component.
 * Add input parameter `autorater_prompt_parameters` to `_implementation.llm.online_evaluation_pairwise` component.
 * Mitigate bug in `v1.model_evaluation.autosxs_pipeline` where batch prediction would fail the first time it is run in a project by retrying.
 

components/google-cloud/google_cloud_pipeline_components/v1/automl/forecasting/prophet_predict_pipeline.yaml

@@ -1461,7 +1461,7 @@ deploymentSpec:
           \ = client.create_dataset(dataset=dataset, exists_ok=exists_ok)\n  return\
           \ collections.namedtuple('Outputs', ['project_id', 'dataset_id'])(\n   \
           \   ref.project, ref.dataset_id)\n\n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
     exec-bigquery-delete-dataset-with-prefix:
       container:
         args:
@@ -1495,7 +1495,7 @@ deploymentSpec:
           \    if dataset.dataset_id.startswith(dataset_prefix):\n      client.delete_dataset(\n\
           \          dataset=dataset.dataset_id,\n          delete_contents=delete_contents)\n\
           \n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
     exec-bigquery-query-job:
       container:
         args:
@@ -1583,7 +1583,7 @@ deploymentSpec:
           \        'datasetId': dataset_id,\n        'tableId': table_id,\n    }\n\
           \  if write_disposition:\n    config['write_disposition'] = write_disposition\n\
           \  return config\n\n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
     exec-build-job-configuration-query-2:
       container:
         args:
@@ -1617,7 +1617,7 @@ deploymentSpec:
           \        'datasetId': dataset_id,\n        'tableId': table_id,\n    }\n\
           \  if write_disposition:\n    config['write_disposition'] = write_disposition\n\
           \  return config\n\n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
     exec-get-first-valid:
       container:
         args:
@@ -1641,7 +1641,7 @@ deploymentSpec:
           \  import json\n  # pylint: enable=g-import-not-at-top,import-outside-toplevel,redefined-outer-name,reimported\n\
           \n  for value in json.loads(values):\n    if value:\n      return value\n\
           \  raise ValueError('No valid values.')\n\n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
     exec-get-table-location:
       container:
         args:
@@ -1677,7 +1677,7 @@ deploymentSpec:
           \  if table.startswith('bq://'):\n    table = table[len('bq://'):]\n  elif\
           \ table.startswith('bigquery://'):\n    table = table[len('bigquery://'):]\n\
           \  return client.get_table(table).location\n\n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
     exec-get-table-location-2:
       container:
         args:
@@ -1713,7 +1713,7 @@ deploymentSpec:
           \  if table.startswith('bq://'):\n    table = table[len('bq://'):]\n  elif\
           \ table.startswith('bigquery://'):\n    table = table[len('bigquery://'):]\n\
           \  return client.get_table(table).location\n\n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
     exec-load-table-from-uri:
       container:
         args:
@@ -1754,7 +1754,7 @@ deploymentSpec:
           \ source_format=source_format)\n  client.load_table_from_uri(\n      source_uris=csv_list,\n\
           \      destination=destination,\n      project=project,\n      location=location,\n\
           \      job_config=job_config).result()\n  return destination\n\n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
     exec-make-vertex-model-artifact:
       container:
         args:
@@ -1778,7 +1778,7 @@ deploymentSpec:
           Creates a google.VertexModel artifact.\"\"\"\n  vertex_model.metadata =\
           \ {'resourceName': model_resource_name}\n  vertex_model.uri = (f'https://{location}-aiplatform.googleapis.com'\n\
           \                      f'/v1/{model_resource_name}')\n\n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
     exec-maybe-replace-with-default:
       container:
         args:
@@ -1800,7 +1800,7 @@ deploymentSpec:
           \ *\n\ndef maybe_replace_with_default(value: str, default: str = '') ->\
           \ str:\n  \"\"\"Replaces string with another value if it is a dash.\"\"\"\
           \n  return default if not value else value\n\n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
     exec-model-batch-predict:
       container:
         args:
@@ -1879,7 +1879,7 @@ deploymentSpec:
           \ if use_bq_prefix:\n    bq_uri = 'bq://' + bq_uri\n  outputs.append(bq_uri)\n\
           \  return collections.namedtuple(\n      'Outputs',\n      ['project_id',\
           \ 'dataset_id', 'table_id', 'uri'],\n  )(*outputs)\n\n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
     exec-table-to-uri-2:
       container:
         args:
@@ -1909,7 +1909,7 @@ deploymentSpec:
           \ if use_bq_prefix:\n    bq_uri = 'bq://' + bq_uri\n  outputs.append(bq_uri)\n\
           \  return collections.namedtuple(\n      'Outputs',\n      ['project_id',\
           \ 'dataset_id', 'table_id', 'uri'],\n  )(*outputs)\n\n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
     exec-validate-inputs:
       container:
         args:
@@ -2005,7 +2005,7 @@ deploymentSpec:
           \   raise ValueError(\n          'Granularity unit should be one of the\
           \ following: '\n          f'{valid_data_granularity_units}, got: {data_granularity_unit}.')\n\
           \n"
-        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240214_1325
+        image: us-docker.pkg.dev/vertex-ai/automl-tabular/kfp-v2-base:20240419_0625
 pipelineInfo:
   description: Creates a batch prediction using a Prophet model.
   name: prophet-predict

components/google-cloud/google_cloud_pipeline_components/v1/automl/forecasting/prophet_trainer.py

@@ -180,6 +180,16 @@ def prophet_trainer(
                   '--dataflow_use_public_ips=',
                   dataflow_use_public_ips,
                   '", "',
+                  '--dataflow_staging_dir=',
+                  root_dir,
+                  (
+                      f'/{dsl.PIPELINE_JOB_ID_PLACEHOLDER}/{dsl.PIPELINE_TASK_ID_PLACEHOLDER}/dataflow_staging", "'
+                  ),
+                  '--dataflow_tmp_dir=',
+                  root_dir,
+                  (
+                      f'/{dsl.PIPELINE_JOB_ID_PLACEHOLDER}/{dsl.PIPELINE_TASK_ID_PLACEHOLDER}/dataflow_tmp", "'
+                  ),
                   '--gcp_resources_path=',
                   gcp_resources,
                   '", "',

components/google-cloud/google_cloud_pipeline_components/v1/automl/forecasting/prophet_trainer_pipeline.yaml

@@ -2418,7 +2418,10 @@ deploymentSpec:
           "\", \"", "--dataflow_service_account=", "{{$.inputs.parameters[''dataflow_service_account'']}}",
           "\", \"", "--dataflow_subnetwork=", "{{$.inputs.parameters[''dataflow_subnetwork'']}}",
           "\", \"", "--dataflow_use_public_ips=", "{{$.inputs.parameters[''dataflow_use_public_ips'']}}",
-          "\", \"", "--gcp_resources_path=", "{{$.outputs.parameters[''gcp_resources''].output_file}}",
+          "\", \"", "--dataflow_staging_dir=", "{{$.inputs.parameters[''root_dir'']}}",
+          "/{{$.pipeline_job_uuid}}/{{$.pipeline_task_uuid}}/dataflow_staging\", \"",
+          "--dataflow_tmp_dir=", "{{$.inputs.parameters[''root_dir'']}}", "/{{$.pipeline_job_uuid}}/{{$.pipeline_task_uuid}}/dataflow_tmp\",
+          \"", "--gcp_resources_path=", "{{$.outputs.parameters[''gcp_resources''].output_file}}",
           "\", \"", "--executor_input={{$.json_escape[1]}}\"]}}]}}"]}'
         command:
         - python3