samples: NLP import and predict samples for SDK (#276)

* samples: NLP import and predict samples for SDK

* fix: mismatched region tags

* samples: add more import data samples

* samples: adds prediction NLP samples

* fix: lint

* samples: prediction tests

* samples: adds endpoint fixtures

* fix: lint

* Update all MBSDK Dataset sample test mocks, fix NLP sample tests

* Remove reference to ipdb

Co-authored-by: Vinny Senthil <vinnysenthil@gmail.com>

samples/model-builder/conftest.py

@@ -25,48 +25,117 @@ def mock_sdk_init():
         yield mock
 
 
-# ----------------------------------------------------------------------------
-# Dataset Fixtures
-# ----------------------------------------------------------------------------
+"""
+----------------------------------------------------------------------------
+Dataset Fixtures
+----------------------------------------------------------------------------
+"""
+
+"""Dataset objects returned by SomeDataset(), create(), import_data(), etc. """
 
 
 @pytest.fixture
-def mock_dataset():
-    mock = MagicMock(aiplatform.datasets.Dataset)
+def mock_image_dataset():
+    mock = MagicMock(aiplatform.datasets.ImageDataset)
     yield mock
 
 
 @pytest.fixture
-def mock_new_dataset(mock_dataset):
-    with patch.object(aiplatform.datasets.Dataset, "__new__") as mock_new_dataset:
-        mock_new_dataset.return_value = mock_dataset
-        yield mock_new_dataset
+def mock_tabular_dataset():
+    mock = MagicMock(aiplatform.datasets.TabularDataset)
+    yield mock
 
 
 @pytest.fixture
-def mock_init_dataset(mock_new_dataset):
-    with patch.object(aiplatform.datasets.Dataset, "__init__") as mock_init_dataset:
-        mock_init_dataset.return_value = None
-        yield mock_init_dataset
+def mock_text_dataset():
+    mock = MagicMock(aiplatform.datasets.TextDataset)
+    yield mock
 
 
 @pytest.fixture
-def mock_create_dataset():
-    with patch.object(aiplatform.datasets.Dataset, "create") as mock:
-        mock.return_value = MagicMock(aiplatform.Dataset)
-        yield mock
+def mock_video_dataset():
+    mock = MagicMock(aiplatform.datasets.VideoDataset)
+    yield mock
+
+
+"""Mocks for getting an existing Dataset, i.e. ds = aiplatform.ImageDataset(...) """
+
+
+@pytest.fixture
+def mock_get_image_dataset(mock_image_dataset):
+    with patch.object(aiplatform, "ImageDataset") as mock_get_image_dataset:
+        mock_get_image_dataset.return_value = mock_image_dataset
+        yield mock_get_image_dataset
 
 
 @pytest.fixture
-def mock_create_image_dataset():
-    with patch.object(aiplatform.datasets.ImageDataset, "create") as mock:
-        mock.return_value = MagicMock(aiplatform.Dataset)
+def mock_get_tabular_dataset(mock_tabular_dataset):
+    with patch.object(aiplatform, "TabularDataset") as mock_get_tabular_dataset:
+        mock_get_tabular_dataset.return_value = mock_tabular_dataset
+        yield mock_get_tabular_dataset
+
+
+@pytest.fixture
+def mock_get_text_dataset(mock_text_dataset):
+    with patch.object(aiplatform, "TextDataset") as mock_get_text_dataset:
+        mock_get_text_dataset.return_value = mock_text_dataset
+        yield mock_get_text_dataset
+
+
+@pytest.fixture
+def mock_get_video_dataset(mock_video_dataset):
+    with patch.object(aiplatform, "VideoDataset") as mock_get_video_dataset:
+        mock_get_video_dataset.return_value = mock_video_dataset
+        yield mock_get_video_dataset
+
+
+"""Mocks for creating a new Dataset, i.e. aiplatform.ImageDataset.create(...) """
+
+
+@pytest.fixture
+def mock_create_image_dataset(mock_image_dataset):
+    with patch.object(aiplatform.ImageDataset, "create") as mock_create_image_dataset:
+        mock_create_image_dataset.return_value = mock_image_dataset
+        yield mock_create_image_dataset
+
+
+@pytest.fixture
+def mock_create_tabular_dataset(mock_tabular_dataset):
+    with patch.object(
+        aiplatform.TabularDataset, "create"
+    ) as mock_create_tabular_dataset:
+        mock_create_tabular_dataset.return_value = mock_tabular_dataset
+        yield mock_create_tabular_dataset
+
+
+@pytest.fixture
+def mock_create_text_dataset(mock_text_dataset):
+    with patch.object(aiplatform.TextDataset, "create") as mock_create_text_dataset:
+        mock_create_text_dataset.return_value = mock_text_dataset
+        yield mock_create_text_dataset
+
+
+@pytest.fixture
+def mock_create_video_dataset(mock_video_dataset):
+    with patch.object(aiplatform.VideoDataset, "create") as mock_create_video_dataset:
+        mock_create_video_dataset.return_value = mock_video_dataset
+        yield mock_create_video_dataset
+
+
+"""Mocks for SomeDataset.import_data() """
+
+
+@pytest.fixture
+def mock_import_text_dataset(mock_text_dataset):
+    with patch.object(mock_text_dataset, "import_data") as mock:
         yield mock
 
 
-# ----------------------------------------------------------------------------
-# TrainingJob Fixtures
-# ----------------------------------------------------------------------------
+"""
+----------------------------------------------------------------------------
+TrainingJob Fixtures
+----------------------------------------------------------------------------
+"""
 
 
 @pytest.fixture
@@ -84,9 +153,11 @@ def mock_run_automl_image_training_job():
         yield mock
 
 
-# ----------------------------------------------------------------------------
-# Model Fixtures
-# ----------------------------------------------------------------------------
+"""
+----------------------------------------------------------------------------
+Model Fixtures
+----------------------------------------------------------------------------
+"""
 
 
 @pytest.fixture
@@ -102,12 +173,34 @@ def mock_batch_predict_model():
         yield mock
 
 
-# ----------------------------------------------------------------------------
-# Job Fixtures
-# ----------------------------------------------------------------------------
+"""
+----------------------------------------------------------------------------
+Job Fixtures
+----------------------------------------------------------------------------
+"""
 
 
 @pytest.fixture
 def mock_create_batch_prediction_job():
     with patch.object(aiplatform.jobs.BatchPredictionJob, "create") as mock:
         yield mock
+
+
+"""
+----------------------------------------------------------------------------
+Endpoint Fixtures
+----------------------------------------------------------------------------
+"""
+
+
+@pytest.fixture
+def mock_endpoint():
+    mock = MagicMock(aiplatform.models.Endpoint)
+    yield mock
+
+
+@pytest.fixture
+def mock_get_endpoint(mock_endpoint):
+    with patch.object(aiplatform, "Endpoint") as mock_get_endpoint:
+        mock_get_endpoint.return_value = mock_endpoint
+        yield mock_get_endpoint

samples/model-builder/create_and_import_dataset_text_sample.py

@@ -0,0 +1,43 @@
+# Copyright 2021 Google LLC
+#
+# 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
+#
+#     https://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, Union
+
+from google.cloud import aiplatform
+
+#  [START aiplatform_sdk_create_and_import_dataset_text_sample]
+def create_and_import_dataset_text_sample(
+    project: str,
+    location: str,
+    display_name: str,
+    src_uris: Union[str, List[str]],
+    sync: bool = True,
+):
+    aiplatform.init(project=project, location=location)
+
+    ds = aiplatform.TextDataset.create(
+        display_name=display_name,
+        gcs_source=src_uris,
+        import_schema_uri=aiplatform.schema.dataset.ioformat.text.single_label_classification,
+        sync=sync,
+    )
+
+    ds.wait()
+
+    print(ds.display_name)
+    print(ds.resource_name)
+    return ds
+
+
+#  [END aiplatform_sdk_create_and_import_dataset_text_sample]

samples/model-builder/create_and_import_dataset_text_sample_test.py

@@ -0,0 +1,39 @@
+# Copyright 2021 Google LLC
+#
+# 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
+#
+#     https://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.
+
+
+import test_constants as constants
+import create_and_import_dataset_text_sample
+
+from google.cloud.aiplatform import schema
+
+
+def test_create_and_import_dataset_text_sample(mock_sdk_init, mock_create_text_dataset):
+
+    create_and_import_dataset_text_sample.create_and_import_dataset_text_sample(
+        project=constants.PROJECT,
+        location=constants.LOCATION,
+        src_uris=constants.GCS_SOURCES,
+        display_name=constants.DISPLAY_NAME,
+    )
+
+    mock_sdk_init.assert_called_once_with(
+        project=constants.PROJECT, location=constants.LOCATION
+    )
+    mock_create_text_dataset.assert_called_once_with(
+        display_name=constants.DISPLAY_NAME,
+        gcs_source=constants.GCS_SOURCES,
+        import_schema_uri=schema.dataset.ioformat.text.single_label_classification,
+        sync=True,
+    )

samples/model-builder/create_training_pipeline_image_classification_sample.py

@@ -32,7 +32,7 @@ def create_training_pipeline_image_classification_sample(
 
     job = aiplatform.AutoMLImageTrainingJob(display_name=display_name)
 
-    my_image_ds = aiplatform.Dataset(dataset_id)
+    my_image_ds = aiplatform.ImageDataset(dataset_id)
 
     model = job.run(
         dataset=my_image_ds,

samples/model-builder/create_training_pipeline_image_classification_sample_test.py

@@ -18,12 +18,12 @@
 import create_training_pipeline_image_classification_sample
 
 
-@pytest.mark.usefixtures("mock_init_dataset")
 def test_create_training_pipeline_image_classification_sample(
     mock_sdk_init,
+    mock_image_dataset,
     mock_init_automl_image_training_job,
-    mock_dataset,
     mock_run_automl_image_training_job,
+    mock_get_image_dataset,
 ):
 
     create_training_pipeline_image_classification_sample.create_training_pipeline_image_classification_sample(
@@ -38,14 +38,16 @@ def test_create_training_pipeline_image_classification_sample(
         disable_early_stopping=False,
     )
 
+    mock_get_image_dataset.assert_called_once_with(constants.RESOURCE_ID)
+
     mock_sdk_init.assert_called_once_with(
         project=constants.PROJECT, location=constants.LOCATION
     )
     mock_init_automl_image_training_job.assert_called_once_with(
         display_name=constants.DISPLAY_NAME
     )
     mock_run_automl_image_training_job.assert_called_once_with(
-        dataset=mock_dataset,
+        dataset=mock_image_dataset,
         model_display_name=constants.DISPLAY_NAME_2,
         training_fraction_split=constants.TRAINING_FRACTION_SPLIT,
         validation_fraction_split=constants.VALIDATION_FRACTION_SPLIT,

samples/model-builder/import_data_text_classification_single_label_sample.py

@@ -0,0 +1,43 @@
+# Copyright 2021 Google LLC
+#
+# 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
+#
+#     https://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, Union
+
+from google.cloud import aiplatform
+
+#  [START aiplatform_sdk_import_data_text_classification_single_label_sample]
+def import_data_text_classification_single_label(
+    project: str,
+    location: str,
+    dataset: str,
+    src_uris: Union[str, List[str]],
+    sync: bool = True,
+):
+    aiplatform.init(project=project, location=location)
+
+    ds = aiplatform.TextDataset(dataset)
+    ds.import_data(
+        gcs_source=src_uris,
+        import_schema_uri=aiplatform.schema.dataset.ioformat.text.single_label_classification,
+        sync=sync,
+    )
+
+    ds.wait()
+
+    print(ds.display_name)
+    print(ds.resource_name)
+    return ds
+
+
+#  [END aiplatform_sdk_import_data_text_classification_single_label_sample]