Skip to content

chore: clean up legacy unit tests #6903

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
copybara-service merged 1 commit into from
Jun 19, 2026
+0 −311
Merged

chore: clean up legacy unit tests #6903

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
311 changes: 0 additions & 311 deletions tests/unit/aiplatform/test_language_models.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4883,317 +4883,6 @@ def teardown_method(self):
self._job_wait_patcher.stop()
self._log_wait_patcher.stop()

@pytest.mark.usefixtures(
"get_model_with_tuned_version_label_mock",
"get_endpoint_with_models_mock",
)
@pytest.mark.parametrize(
"job_spec",
[_TEST_EVAL_PIPELINE_SPEC_JSON, _TEST_EVAL_PIPELINE_JOB],
)
@pytest.mark.parametrize(
"mock_request_urlopen_eval",
["https://us-kfp.pkg.dev/proj/repo/pack/latest"],
indirect=True,
)
def test_model_evaluation_text_generation_task_with_gcs_input(
self,
job_spec,
mock_pipeline_service_create_eval,
mock_pipeline_job_get_eval,
mock_successfully_completed_eval_job,
mock_pipeline_bucket_exists,
mock_load_yaml_and_json,
mock_request_urlopen_eval,
):
aiplatform.init(project=_TEST_PROJECT, location=_TEST_LOCATION)

with mock.patch.object(
target=model_garden_service_client.ModelGardenServiceClient,
attribute="get_publisher_model",
return_value=gca_publisher_model.PublisherModel(
_TEXT_BISON_PUBLISHER_MODEL_DICT
),
):
my_model = preview_language_models.TextGenerationModel.get_tuned_model(
test_constants.ModelConstants._TEST_MODEL_RESOURCE_NAME
)

eval_metrics = my_model.evaluate(
task_spec=preview_language_models.EvaluationTextGenerationSpec(
ground_truth_data="gs://my-bucket/ground-truth.jsonl",
),
)

assert isinstance(eval_metrics, preview_language_models.EvaluationMetric)
assert eval_metrics.bleu == _TEST_TEXT_GENERATION_METRICS["bleu"]

@pytest.mark.usefixtures(
"get_model_with_tuned_version_label_mock",
"get_endpoint_with_models_mock",
)
@pytest.mark.parametrize(
"job_spec",
[_TEST_EVAL_PIPELINE_SPEC_JSON, _TEST_EVAL_PIPELINE_JOB],
)
def test_populate_eval_template_params(
self,
job_spec,
mock_pipeline_service_create,
mock_model_evaluate,
mock_pipeline_job_get,
mock_successfully_completed_eval_job,
mock_pipeline_bucket_exists,
mock_load_yaml_and_json,
):
aiplatform.init(project=_TEST_PROJECT, location=_TEST_LOCATION)

with mock.patch.object(
target=model_garden_service_client.ModelGardenServiceClient,
attribute="get_publisher_model",
return_value=gca_publisher_model.PublisherModel(
_TEXT_BISON_PUBLISHER_MODEL_DICT
),
):
my_model = preview_language_models.TextGenerationModel.get_tuned_model(
test_constants.ModelConstants._TEST_MODEL_RESOURCE_NAME
)

task_spec = preview_language_models.EvaluationTextGenerationSpec(
ground_truth_data="gs://my-bucket/ground-truth.jsonl",
)

formatted_template_params = (
_evaluatable_language_models._populate_eval_template_params(
task_spec=task_spec, model_name=my_model._model_resource_name
)
)

assert (
"batch_predict_gcs_destination_output_uri" in formatted_template_params
)
assert "model_name" in formatted_template_params
assert "evaluation_task" in formatted_template_params

# This should only be in the classification task pipeline template
assert "evaluation_class_labels" not in formatted_template_params
assert "target_column_name" not in formatted_template_params

@pytest.mark.usefixtures(
"get_model_with_tuned_version_label_mock",
"get_endpoint_with_models_mock",
)
@pytest.mark.parametrize(
"job_spec",
[_TEST_EVAL_PIPELINE_SPEC_JSON, _TEST_EVAL_PIPELINE_JOB],
)
def test_populate_template_params_for_classification_task(
self,
job_spec,
mock_pipeline_service_create,
mock_model_evaluate,
mock_pipeline_job_get,
mock_successfully_completed_eval_job,
mock_pipeline_bucket_exists,
mock_load_yaml_and_json,
):
aiplatform.init(project=_TEST_PROJECT, location=_TEST_LOCATION)

with mock.patch.object(
target=model_garden_service_client.ModelGardenServiceClient,
attribute="get_publisher_model",
return_value=gca_publisher_model.PublisherModel(
_TEXT_BISON_PUBLISHER_MODEL_DICT
),
):
my_model = preview_language_models.TextGenerationModel.get_tuned_model(
test_constants.ModelConstants._TEST_MODEL_RESOURCE_NAME
)

task_spec = preview_language_models.EvaluationTextClassificationSpec(
ground_truth_data="gs://my-bucket/ground-truth.jsonl",
target_column_name="test_targ_name",
class_names=["test_class_name_1", "test_class_name_2"],
)

formatted_template_params = (
_evaluatable_language_models._populate_eval_template_params(
task_spec=task_spec, model_name=my_model._model_resource_name
)
)

assert "evaluation_class_labels" in formatted_template_params
assert "target_field_name" in formatted_template_params

@pytest.mark.usefixtures(
"get_model_with_tuned_version_label_mock",
"get_endpoint_with_models_mock",
"mock_storage_blob_upload_from_filename",
)
@pytest.mark.parametrize(
"job_spec",
[_TEST_EVAL_PIPELINE_SPEC_JSON, _TEST_EVAL_PIPELINE_JOB],
)
def test_populate_template_params_with_dataframe_input(
self,
job_spec,
mock_pipeline_service_create,
mock_pipeline_job_get,
mock_successfully_completed_eval_job,
mock_pipeline_bucket_exists,
mock_load_yaml_and_json,
):
aiplatform.init(project=_TEST_PROJECT, location=_TEST_LOCATION)

with mock.patch.object(
target=model_garden_service_client.ModelGardenServiceClient,
attribute="get_publisher_model",
return_value=gca_publisher_model.PublisherModel(
_TEXT_BISON_PUBLISHER_MODEL_DICT
),
):
my_model = preview_language_models.TextGenerationModel.get_tuned_model(
test_constants.ModelConstants._TEST_MODEL_RESOURCE_NAME
)

task_spec = preview_language_models.EvaluationTextGenerationSpec(
ground_truth_data=_TEST_EVAL_DATA_DF,
)

formatted_template_params = (
_evaluatable_language_models._populate_eval_template_params(
task_spec=task_spec, model_name=my_model._model_resource_name
)
)

# The utility method should not modify task_spec
assert isinstance(task_spec.ground_truth_data, pd.DataFrame)

assert (
"batch_predict_gcs_destination_output_uri" in formatted_template_params
)
assert "model_name" in formatted_template_params
assert "evaluation_task" in formatted_template_params

# This should only be in the classification task pipeline template
assert "evaluation_class_labels" not in formatted_template_params
assert "target_column_name" not in formatted_template_params

def test_evaluate_raises_on_ga_language_model(
self,
):
aiplatform.init(project=_TEST_PROJECT, location=_TEST_LOCATION)

with mock.patch.object(
target=model_garden_service_client.ModelGardenServiceClient,
attribute="get_publisher_model",
return_value=gca_publisher_model.PublisherModel(
_TEXT_BISON_PUBLISHER_MODEL_DICT
),
):
model = language_models.TextGenerationModel.from_pretrained(
"text-bison@001"
)

with pytest.raises(AttributeError):
model.evaluate()

@pytest.mark.usefixtures(
"get_endpoint_with_models_mock",
)
@pytest.mark.parametrize(
"job_spec",
[_TEST_EVAL_PIPELINE_SPEC_JSON, _TEST_EVAL_PIPELINE_JOB],
)
@pytest.mark.parametrize(
"mock_request_urlopen_eval",
["https://us-kfp.pkg.dev/proj/repo/pack/latest"],
indirect=True,
)
def test_model_evaluation_text_generation_task_on_base_model(
self,
job_spec,
mock_pipeline_service_create_eval,
mock_pipeline_job_get_eval,
mock_successfully_completed_eval_job,
mock_pipeline_bucket_exists,
mock_load_yaml_and_json,
mock_request_urlopen_eval,
):
aiplatform.init(project=_TEST_PROJECT, location=_TEST_LOCATION)

with mock.patch.object(
target=model_garden_service_client.ModelGardenServiceClient,
attribute="get_publisher_model",
return_value=gca_publisher_model.PublisherModel(
_TEXT_BISON_PUBLISHER_MODEL_DICT
),
):
my_model = preview_language_models.TextGenerationModel.from_pretrained(
"text-bison@001"
)

eval_metrics = my_model.evaluate(
task_spec=preview_language_models.EvaluationTextGenerationSpec(
ground_truth_data="gs://my-bucket/ground-truth.jsonl",
),
)

assert isinstance(eval_metrics, preview_language_models.EvaluationMetric)

@pytest.mark.usefixtures(
"get_endpoint_with_models_mock",
)
@pytest.mark.parametrize(
"job_spec",
[
_TEST_EVAL_CLASSIFICATION_PIPELINE_SPEC_JSON,
_TEST_EVAL_CLASSIFICATION_PIPELINE_JOB,
],
)
@pytest.mark.parametrize(
"mock_request_urlopen_eval_classification",
["https://us-central1-kfp.pkg.dev/proj/repo/pack/latest"],
indirect=True,
)
def test_model_evaluation_text_classification_base_model_only_summary_metrics(
self,
job_spec,
mock_pipeline_service_create_eval_classification,
mock_pipeline_job_get_eval_classification,
mock_successfully_completed_eval_classification_job,
mock_pipeline_bucket_exists,
mock_load_yaml_and_json,
mock_request_urlopen_eval_classification,
):
aiplatform.init(project=_TEST_PROJECT, location=_TEST_LOCATION)

with mock.patch.object(
target=model_garden_service_client.ModelGardenServiceClient,
attribute="get_publisher_model",
return_value=gca_publisher_model.PublisherModel(
_TEXT_BISON_PUBLISHER_MODEL_DICT
),
):
my_model = preview_language_models.TextGenerationModel.from_pretrained(
"text-bison@001"
)

eval_metrics = my_model.evaluate(
task_spec=preview_language_models.EvaluationTextClassificationSpec(
ground_truth_data="gs://my-bucket/ground-truth.jsonl",
target_column_name="test_targ_name",
class_names=["test_class_name_1", "test_class_name_2"],
)
)

assert isinstance(
eval_metrics,
preview_language_models.EvaluationClassificationMetric,
)
assert eval_metrics.confidenceMetrics is None
assert eval_metrics.auPrc == _TEST_TEXT_CLASSIFICATION_METRICS["auPrc"]

@pytest.mark.parametrize(
"job_spec",
[
Expand Down
Loading