google-cloud-pipeline-components 2.16.0__py3-none-any.whl → 2.17.0__py3-none-any.whl

google_cloud_pipeline_components/_implementation/llm/generated/refined_image_versions.py

@@ -17,4 +17,4 @@
 DO NOT EDIT - This file is generated, manual changes will be overridden.
 """
 
-IMAGE_TAG = '20240623_1707'
+IMAGE_TAG = '20240818_1707'

google_cloud_pipeline_components/_implementation/model_evaluation/import_evaluation/__init__.py

@@ -0,0 +1,14 @@
+# Copyright 2023 The Kubeflow Authors. All Rights Reserved.
+#
+# Licensed 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.
+"""Google Cloud Pipeline Evaluation Import Evaluation Component."""

google_cloud_pipeline_components/_implementation/model_evaluation/import_evaluation/component.py

@@ -0,0 +1,208 @@
+# Copyright 2023 The Kubeflow Authors. All Rights Reserved.
+#
+# Licensed 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.
+
+
+from typing import List, Optional
+
+from google_cloud_pipeline_components import _image
+from google_cloud_pipeline_components.types.artifact_types import ClassificationMetrics
+from google_cloud_pipeline_components.types.artifact_types import ForecastingMetrics
+from google_cloud_pipeline_components.types.artifact_types import RegressionMetrics
+from google_cloud_pipeline_components.types.artifact_types import VertexModel
+from kfp import dsl
+from kfp.dsl import Input
+from kfp.dsl import Metrics
+
+
+@dsl.container_component
+def model_evaluation_import(
+    model: Input[VertexModel],
+    gcp_resources: dsl.OutputPath(str),
+    evaluation_resource_name: dsl.OutputPath(str),
+    metrics: Optional[Input[Metrics]] = None,
+    row_based_metrics: Optional[Input[Metrics]] = None,
+    problem_type: Optional[str] = None,
+    classification_metrics: Optional[Input[ClassificationMetrics]] = None,
+    forecasting_metrics: Optional[Input[ForecastingMetrics]] = None,
+    regression_metrics: Optional[Input[RegressionMetrics]] = None,
+    text_generation_metrics: Optional[Input[Metrics]] = None,
+    question_answering_metrics: Optional[Input[Metrics]] = None,
+    summarization_metrics: Optional[Input[Metrics]] = None,
+    explanation: Optional[Input[Metrics]] = None,
+    feature_attributions: Optional[Input[Metrics]] = None,
+    embedding_metrics: Optional[Input[Metrics]] = None,
+    display_name: str = "",
+    dataset_path: str = "",
+    dataset_paths: List[str] = [],
+    dataset_type: str = "",
+):
+  # fmt: off
+  """Imports a model evaluation artifact to an existing Vertex model with
+  ModelService.ImportModelEvaluation.
+
+  For more details, see
+  https://cloud.google.com/vertex-ai/docs/reference/rest/v1/projects.locations.models.evaluations
+  One of the four metrics inputs must be provided, metrics & problem_type,
+  classification_metrics, regression_metrics, or forecasting_metrics.
+
+  Args:
+    model: Vertex model resource that will be the parent resource of the
+      uploaded evaluation.
+    metrics: Path of metrics generated from an evaluation component.
+    row_based_metrics:
+      Path of row_based_metrics generated from an evaluation component.
+    problem_type: The problem type of the metrics being imported to the
+      VertexModel. `classification`, `regression`, `forecasting`,
+      `text-generation`, `question-answering`, and `summarization` are the
+      currently supported problem types. Must be provided when `metrics` is
+      provided.
+    classification_metrics: google.ClassificationMetrics artifact generated from
+      the ModelEvaluationClassificationOp component.
+    forecasting_metrics: google.ForecastingMetrics artifact generated from
+      the ModelEvaluationForecastingOp component.
+    regression_metrics: google.ClassificationMetrics artifact generated from
+      the ModelEvaluationRegressionOp component.
+    text_generation_metrics: system.Metrics artifact generated from
+      the LLMEvaluationTextGenerationOp component. Subject to change to
+      google.TextGenerationMetrics.
+    question_answering_metrics: system.Metrics artifact generated from
+      the LLMEvaluationTextGenerationOp component. Subject to change to
+      google.QuestionAnsweringMetrics.
+    summarization_metrics: system.Metrics artifact generated from
+      the LLMEvaluationTextGenerationOp component. Subject to change to
+      google.SummarizationMetrics.
+    explanation: Path for model explanation metrics generated from an evaluation
+      component.
+    feature_attributions: The feature attributions metrics artifact generated
+      from the feature attribution component.
+    embedding_metrics: The embedding metrics artifact generated from the
+      embedding retrieval metrics component.
+    display_name: The display name for the uploaded model evaluation resource.
+  """
+  # fmt: on
+  return dsl.ContainerSpec(
+      image=_image.GCPC_IMAGE_TAG,
+      command=[
+          "python3",
+          "-u",
+          "-m",
+          "google_cloud_pipeline_components.container._implementation.model_evaluation.import_model_evaluation",
+      ],
+      args=[
+          dsl.IfPresentPlaceholder(
+              input_name="metrics",
+              then=[
+                  "--metrics",
+                  metrics.uri,
+                  "--metrics_explanation",
+                  metrics.metadata["explanation_gcs_path"],
+              ],
+          ),
+          dsl.IfPresentPlaceholder(
+              input_name="row_based_metrics",
+              then=[
+                  "--row_based_metrics",
+                  row_based_metrics.uri,
+              ],
+          ),
+          dsl.IfPresentPlaceholder(
+              input_name="explanation",
+              then=[
+                  "--explanation",
+                  explanation.metadata["explanation_gcs_path"],
+              ],
+          ),
+          dsl.IfPresentPlaceholder(
+              input_name="classification_metrics",
+              then=[
+                  "--classification_metrics",
+                  classification_metrics.uri,
+              ],
+          ),
+          dsl.IfPresentPlaceholder(
+              input_name="forecasting_metrics",
+              then=[
+                  "--forecasting_metrics",
+                  forecasting_metrics.uri,
+              ],
+          ),
+          dsl.IfPresentPlaceholder(
+              input_name="regression_metrics",
+              then=[
+                  "--regression_metrics",
+                  regression_metrics.uri,
+              ],
+          ),
+          dsl.IfPresentPlaceholder(
+              input_name="text_generation_metrics",
+              then=[
+                  "--text_generation_metrics",
+                  text_generation_metrics.uri,
+              ],
+          ),
+          dsl.IfPresentPlaceholder(
+              input_name="question_answering_metrics",
+              then=[
+                  "--question_answering_metrics",
+                  question_answering_metrics.uri,
+              ],
+          ),
+          dsl.IfPresentPlaceholder(
+              input_name="summarization_metrics",
+              then=[
+                  "--summarization_metrics",
+                  summarization_metrics.uri,
+              ],
+          ),
+          dsl.IfPresentPlaceholder(
+              input_name="feature_attributions",
+              then=[
+                  "--feature_attributions",
+                  feature_attributions.uri,
+              ],
+          ),
+          dsl.IfPresentPlaceholder(
+              input_name="embedding_metrics",
+              then=[
+                  "--embedding_metrics",
+                  embedding_metrics.uri,
+              ],
+          ),
+          dsl.IfPresentPlaceholder(
+              input_name="problem_type",
+              then=[
+                  "--problem_type",
+                  problem_type,
+              ],
+          ),
+          "--display_name",
+          display_name,
+          "--dataset_path",
+          dataset_path,
+          "--dataset_paths",
+          dataset_paths,
+          "--dataset_type",
+          dataset_type,
+          "--pipeline_job_id",
+          dsl.PIPELINE_JOB_ID_PLACEHOLDER,
+          "--pipeline_job_resource_name",
+          dsl.PIPELINE_JOB_RESOURCE_NAME_PLACEHOLDER,
+          "--model_name",
+          model.metadata["resourceName"],
+          "--gcp_resources",
+          gcp_resources,
+          "--evaluation_resource_name",
+          evaluation_resource_name,
+      ],
+  )

google_cloud_pipeline_components/_implementation/model_evaluation/llm_evaluation/component.py

@@ -32,6 +32,7 @@ def model_evaluation_text_generation(
     row_based_metrics: Output[Metrics],
     project: str,
     location: str,
+    model_name: str,
     evaluation_task: str = 'text-generation',
     target_field_name: str = 'instance.output_text',
     prediction_field_name: str = 'predictions.content',
@@ -55,6 +56,7 @@ def model_evaluation_text_generation(
   Args:
     project: The GCP project that runs the pipeline component.
     location: The GCP region that runs the pipeline component.
+    model_name: The name of the model to be evaluated.
     evaluation_task: The task that the large language model will be evaluated
       on. The evaluation component computes a set of metrics relevant to that
       specific task. Currently supported tasks are: `summarization`,
@@ -124,6 +126,7 @@ def model_evaluation_text_generation(
           machine_type=machine_type,
           image_uri=version.LLM_EVAL_IMAGE_TAG,
           args=[
+              f'--model_name={model_name}',
               f'--evaluation_task={evaluation_task}',
               f'--target_field_name={target_field_name}',
               f'--prediction_field_name={prediction_field_name}',

google_cloud_pipeline_components/_implementation/model_evaluation/llm_evaluation_preprocessor/component.py

@@ -101,10 +101,8 @@ def evaluation_dataset_preprocessor_internal(
               f'--gcs_source_uris={gcs_source_uris}',
               f'--input_field_name={input_field_name}',
               f'--role_field_name={role_field_name}',
-              (
-                  f'--target_field_name={target_field_name}'
-                  f'--model_name={model_name}'
-              ),
+              f'--target_field_name={target_field_name}',
+              f'--model_name={model_name}',
               f'--output_dirs={output_dirs}',
               '--executor_input={{$.json_escape[1]}}',
           ],

google_cloud_pipeline_components/_implementation/model_evaluation/version.py

@@ -14,7 +14,7 @@
 """Version constants for model evaluation components."""
 
 _EVAL_VERSION = 'v0.9.4'
-_LLM_EVAL_VERSION = 'v0.6'
+_LLM_EVAL_VERSION = 'v0.7'
 
 _EVAL_IMAGE_NAME = 'gcr.io/ml-pipeline/model-evaluation'
 _LLM_EVAL_IMAGE_NAME = 'gcr.io/ml-pipeline/llm-model-evaluation'

google_cloud_pipeline_components/container/preview/custom_job/remote_runner.py

@@ -32,23 +32,32 @@ def insert_system_labels_into_payload(payload):
   return json.dumps(job_spec)
 
 
-def cast_accelerator_count_to_int(payload):
-  """Casts accelerator_count from string to an int."""
+def is_json(test_string: str) -> bool:
+  try:
+    json.loads(test_string)
+  except ValueError:
+    return False
+  return True
+
+
+def parse_nested_json_strings(payload):
+  """Parse nested json strings in the payload."""
 
   job_spec = json.loads(payload)
-  # TODO(b/353577594): accelerator_count placeholder is not resolved to int.
-  # Need to typecast to int to avoid type mismatch error. Can remove when fix
-  # placeholder resolution.
-  if (
-      'accelerator_count'
-      in job_spec['job_spec']['worker_pool_specs'][0]['machine_spec']
+  # TODO(b/353577594): Nested placeholder fields inside worker_pool_specs are
+  # not parsed correctly in backend. Can remove when fix backend logic.
+  worker_pool_spec = job_spec['job_spec']['worker_pool_specs'][0]
+  if is_json(
+      worker_pool_spec.get('machine_spec', {}).get('accelerator_count', '')
+  ):
+    worker_pool_spec['machine_spec']['accelerator_count'] = json.loads(
+        worker_pool_spec['machine_spec']['accelerator_count']
+    )
+  if is_json(
+      worker_pool_spec.get('disk_spec', {}).get('boot_disk_size_gb', '')
   ):
-    job_spec['job_spec']['worker_pool_specs'][0]['machine_spec'][
-        'accelerator_count'
-    ] = int(
-        job_spec['job_spec']['worker_pool_specs'][0]['machine_spec'][
-            'accelerator_count'
-        ]
+    worker_pool_spec['disk_spec']['boot_disk_size_gb'] = json.loads(
+        worker_pool_spec['disk_spec']['boot_disk_size_gb']
     )
   return json.dumps(job_spec)
 
@@ -107,7 +116,7 @@ def create_custom_job(
     # Create custom job if it does not exist
     job_name = remote_runner.check_if_job_exists()
     if job_name is None:
-      payload = cast_accelerator_count_to_int(payload)
+      payload = parse_nested_json_strings(payload)
       job_name = remote_runner.create_job(
           create_custom_job_with_client,
           insert_system_labels_into_payload(payload),

google_cloud_pipeline_components/preview/automl/forecasting/forecasting_ensemble.py

@@ -72,7 +72,7 @@ def automl_forecasting_ensemble(
   # fmt: on
   job_id = dsl.PIPELINE_JOB_ID_PLACEHOLDER
   task_id = dsl.PIPELINE_TASK_ID_PLACEHOLDER
-  image_uri = 'us-docker.pkg.dev/vertex-ai-restricted/automl-tabular/forecasting-training:20240710_0625'
+  image_uri = 'us-docker.pkg.dev/vertex-ai-restricted/automl-tabular/forecasting-training:20240808_0625'
   display_name = f'automl-forecasting-ensemble-{job_id}-{task_id}'
 
   error_file_path = f'{root_dir}/{job_id}/{task_id}/error.pb'

google_cloud_pipeline_components/preview/automl/forecasting/forecasting_stage_1_tuner.py

@@ -99,14 +99,14 @@ def automl_forecasting_stage_1_tuner(
                       ' 1, "machine_spec": {"machine_type": "n1-standard-8"},'
                       ' "container_spec": {"image_uri":"'
                   ),
-                  'us-docker.pkg.dev/vertex-ai-restricted/automl-tabular/forecasting-training:20240710_0625',
+                  'us-docker.pkg.dev/vertex-ai-restricted/automl-tabular/forecasting-training:20240808_0625',
                   '", "args": ["forecasting_mp_l2l_stage_1_tuner',
                   '", "--region=',
                   location,
                   '", "--transform_output_path=',
                   transform_output.uri,
                   '", "--training_docker_uri=',
-                  'us-docker.pkg.dev/vertex-ai-restricted/automl-tabular/forecasting-training:20240710_0625',
+                  'us-docker.pkg.dev/vertex-ai-restricted/automl-tabular/forecasting-training:20240808_0625',
                   '", "--reduce_search_space_mode=',
                   reduce_search_space_mode,
                   f'", "--component_id={dsl.PIPELINE_TASK_ID_PLACEHOLDER}',

google_cloud_pipeline_components/preview/automl/forecasting/forecasting_stage_2_tuner.py

@@ -97,14 +97,14 @@ def automl_forecasting_stage_2_tuner(
                       ' 1, "machine_spec": {"machine_type": "n1-standard-8"},'
                       ' "container_spec": {"image_uri":"'
                   ),
-                  'us-docker.pkg.dev/vertex-ai-restricted/automl-tabular/forecasting-training:20240710_0625',
+                  'us-docker.pkg.dev/vertex-ai-restricted/automl-tabular/forecasting-training:20240808_0625',
                   '", "args": ["forecasting_mp_l2l_stage_2_tuner',
                   '", "--region=',
                   location,
                   '", "--transform_output_path=',
                   transform_output.uri,
                   '", "--training_docker_uri=',
-                  'us-docker.pkg.dev/vertex-ai-restricted/automl-tabular/forecasting-training:20240710_0625',
+                  'us-docker.pkg.dev/vertex-ai-restricted/automl-tabular/forecasting-training:20240808_0625',
                   f'", "--component_id={dsl.PIPELINE_TASK_ID_PLACEHOLDER}',
                   '", "--training_base_dir=',
                   root_dir,