feat: update create_training_pipeline samples (#142)

dizcology · web-flow · commit 624a08d65c20 · 2020-12-22T11:17:08.000-08:00
* feat: update create_training_pipeline samples.  Use the schema types.
diff --git a/.sample_configs/process_configs.yaml b/.sample_configs/process_configs.yaml
@@ -69,48 +69,33 @@ create_hyperparameter_tuning_job_sample: {}
 create_specialist_pool_sample: {}
 create_training_pipeline_custom_job_sample: {}
 create_training_pipeline_custom_training_managed_dataset_sample: {}
-create_training_pipeline_entity_extraction_sample: {}
-create_training_pipeline_image_classification_sample: {}
-create_training_pipeline_image_object_detection_sample: {}
+create_training_pipeline_image_classification_sample:
+  schema_types:
+    training_task_inputs_dict: trainingjob.definition.AutoMlImageClassificationInputs
+create_training_pipeline_image_object_detection_sample:
+  schema_types:
+    training_task_inputs_dict: trainingjob.definition.AutoMlImageObjectDetectionInputs
 create_training_pipeline_sample: {}
 create_training_pipeline_tabular_classification_sample: {}
 create_training_pipeline_tabular_regression_sample: {}
-create_training_pipeline_text_classification_sample: {}
+create_training_pipeline_text_classification_sample:
+  schema_types:
+    training_task_inputs_dict: trainingjob.definition.AutoMlTextClassificationInputs
 create_training_pipeline_text_entity_extraction_sample:
-  skip:
-  - predict_schemata
-  - supported_export_formats
-  - container_spec
-  - deployed_models
-  - explanation_spec
+  schema_types:
+    training_task_inputs_dict: trainingjob.definition.AutoMlTextExtractionInputs
 create_training_pipeline_text_sentiment_analysis_sample:
-  skip:
-  - predict_schemata
-  - supported_export_formats
-  - container_spec
-  - deployed_models
-  - explanation_spec
+  schema_types:
+    training_task_inputs_dict: trainingjob.definition.AutoMlTextSentimentInputs
 create_training_pipeline_video_action_recognition_sample:
-  skip:
-  - predict_schemata
-  - supported_export_formats
-  - container_spec
-  - deployed_models
-  - explanation_spec
+  schema_types:
+    training_task_inputs_dict: trainingjob.definition.AutoMlVideoActionRecognitionInputs
 create_training_pipeline_video_classification_sample:
-  skip:
-  - predict_schemata
-  - supported_export_formats
-  - container_spec
-  - deployed_models
-  - explanation_spec
+  schema_types:
+    training_task_inputs_dict: trainingjob.definition.AutoMlVideoClassificationInputs
 create_training_pipeline_video_object_tracking_sample:
-  skip:
-  - predict_schemata
-  - supported_export_formats
-  - container_spec
-  - deployed_models
-  - explanation_spec
+  schema_types:
+    training_task_inputs_dict: trainingjob.definition.AutoMlVideoObjectTrackingInputs
 delete_batch_prediction_job_sample: {}
 delete_custom_job_sample: {}
 delete_data_labeling_job_sample: {}
diff --git a/samples/snippets/create_training_pipeline_image_classification_sample.py b/samples/snippets/create_training_pipeline_image_classification_sample.py
@@ -30,14 +30,13 @@ def create_training_pipeline_image_classification_sample(
     # Initialize client that will be used to create and send requests.
     # This client only needs to be created once, and can be reused for multiple requests.
     client = aiplatform.gapic.PipelineServiceClient(client_options=client_options)
-
-    icn_training_inputs = trainingjob.definition.AutoMlImageClassificationInputs(
+    training_task_inputs_object = trainingjob.definition.AutoMlImageClassificationInputs(
         multi_label=True,
-        model_type=trainingjob.definition.AutoMlImageClassificationInputs.ModelType.CLOUD,
+        model_type="CLOUD",
         budget_milli_node_hours=8000,
-        disable_early_stopping=False
+        disable_early_stopping=False,
     )
-    training_task_inputs = icn_training_inputs.to_value()
+    training_task_inputs = training_task_inputs_object.to_value()
 
     training_pipeline = {
         "display_name": display_name,
diff --git a/samples/snippets/create_training_pipeline_image_object_detection_sample.py b/samples/snippets/create_training_pipeline_image_object_detection_sample.py
@@ -14,8 +14,7 @@
 
 # [START aiplatform_create_training_pipeline_image_object_detection_sample]
 from google.cloud import aiplatform
-from google.protobuf import json_format
-from google.protobuf.struct_pb2 import Value
+from google.cloud.aiplatform.schema import trainingjob
 
 
 def create_training_pipeline_image_object_detection_sample(
@@ -31,12 +30,12 @@ def create_training_pipeline_image_object_detection_sample(
     # Initialize client that will be used to create and send requests.
     # This client only needs to be created once, and can be reused for multiple requests.
     client = aiplatform.gapic.PipelineServiceClient(client_options=client_options)
-    training_task_inputs_dict = {
-        "modelType": "CLOUD_HIGH_ACCURACY_1",
-        "budgetMilliNodeHours": 20000,
-        "disableEarlyStopping": False,
-    }
-    training_task_inputs = json_format.ParseDict(training_task_inputs_dict, Value())
+    training_task_inputs_object = trainingjob.definition.AutoMlImageObjectDetectionInputs(
+        model_type="CLOUD_HIGH_ACCURACY_1",
+        budget_milli_node_hours=20000,
+        disable_early_stopping=False,
+    )
+    training_task_inputs = training_task_inputs_object.to_value()
 
     training_pipeline = {
         "display_name": display_name,
diff --git a/samples/snippets/create_training_pipeline_text_classification_sample.py b/samples/snippets/create_training_pipeline_text_classification_sample.py
@@ -14,8 +14,7 @@
 
 # [START aiplatform_create_training_pipeline_text_classification_sample]
 from google.cloud import aiplatform
-from google.protobuf import json_format
-from google.protobuf.struct_pb2 import Value
+from google.cloud.aiplatform.schema import trainingjob
 
 
 def create_training_pipeline_text_classification_sample(
@@ -31,8 +30,10 @@ def create_training_pipeline_text_classification_sample(
     # Initialize client that will be used to create and send requests.
     # This client only needs to be created once, and can be reused for multiple requests.
     client = aiplatform.gapic.PipelineServiceClient(client_options=client_options)
-    training_task_inputs_dict = {}
-    training_task_inputs = json_format.ParseDict(training_task_inputs_dict, Value())
+    training_task_inputs_object = (
+        trainingjob.definition.AutoMlTextClassificationInputs()
+    )
+    training_task_inputs = training_task_inputs_object.to_value()
 
     training_pipeline = {
         "display_name": display_name,
diff --git a/samples/snippets/create_training_pipeline_text_entity_extraction_sample.py b/samples/snippets/create_training_pipeline_text_entity_extraction_sample.py
@@ -14,8 +14,7 @@
 
 # [START aiplatform_create_training_pipeline_text_entity_extraction_sample]
 from google.cloud import aiplatform
-from google.protobuf import json_format
-from google.protobuf.struct_pb2 import Value
+from google.cloud.aiplatform.schema import trainingjob
 
 
 def create_training_pipeline_text_entity_extraction_sample(
@@ -31,8 +30,8 @@ def create_training_pipeline_text_entity_extraction_sample(
     # Initialize client that will be used to create and send requests.
     # This client only needs to be created once, and can be reused for multiple requests.
     client = aiplatform.gapic.PipelineServiceClient(client_options=client_options)
-    training_task_inputs_dict = {}
-    training_task_inputs = json_format.ParseDict(training_task_inputs_dict, Value())
+    training_task_inputs_object = trainingjob.definition.AutoMlTextExtractionInputs()
+    training_task_inputs = training_task_inputs_object.to_value()
 
     training_pipeline = {
         "display_name": display_name,
diff --git a/samples/snippets/create_training_pipeline_text_sentiment_analysis_sample.py b/samples/snippets/create_training_pipeline_text_sentiment_analysis_sample.py
@@ -14,8 +14,7 @@
 
 # [START aiplatform_create_training_pipeline_text_sentiment_analysis_sample]
 from google.cloud import aiplatform
-from google.protobuf import json_format
-from google.protobuf.struct_pb2 import Value
+from google.cloud.aiplatform.schema import trainingjob
 
 
 def create_training_pipeline_text_sentiment_analysis_sample(
@@ -32,8 +31,10 @@ def create_training_pipeline_text_sentiment_analysis_sample(
     # This client only needs to be created once, and can be reused for multiple requests.
     client = aiplatform.gapic.PipelineServiceClient(client_options=client_options)
     # Use sentiment_max of 4
-    training_task_inputs_dict = {"sentiment_max": 4}
-    training_task_inputs = json_format.ParseDict(training_task_inputs_dict, Value())
+    training_task_inputs_object = trainingjob.definition.AutoMlTextSentimentInputs(
+        sentiment_max=4
+    )
+    training_task_inputs = training_task_inputs_object.to_value()
 
     training_pipeline = {
         "display_name": display_name,
diff --git a/samples/snippets/create_training_pipeline_video_action_recognition_sample.py b/samples/snippets/create_training_pipeline_video_action_recognition_sample.py
@@ -14,8 +14,7 @@
 
 # [START aiplatform_create_training_pipeline_video_action_recognition_sample]
 from google.cloud import aiplatform
-from google.protobuf import json_format
-from google.protobuf.struct_pb2 import Value
+from google.cloud.aiplatform.schema import trainingjob
 
 
 def create_training_pipeline_video_action_recognition_sample(
@@ -32,11 +31,11 @@ def create_training_pipeline_video_action_recognition_sample(
     # Initialize client that will be used to create and send requests.
     # This client only needs to be created once, and can be reused for multiple requests.
     client = aiplatform.gapic.PipelineServiceClient(client_options=client_options)
-    training_task_inputs_dict = {
-        # modelType can be either 'CLOUD' or 'MOBILE_VERSATILE_1'
-        "modelType": model_type
-    }
-    training_task_inputs = json_format.ParseDict(training_task_inputs_dict, Value())
+    # modelType can be either 'CLOUD' or 'MOBILE_VERSATILE_1'
+    training_task_inputs_object = trainingjob.definition.AutoMlVideoActionRecognitionInputs(
+        model_type=model_type
+    )
+    training_task_inputs = training_task_inputs_object.to_value()
 
     training_pipeline = {
         "display_name": display_name,
diff --git a/samples/snippets/create_training_pipeline_video_classification_sample.py b/samples/snippets/create_training_pipeline_video_classification_sample.py
@@ -14,8 +14,7 @@
 
 # [START aiplatform_create_training_pipeline_video_classification_sample]
 from google.cloud import aiplatform
-from google.protobuf import json_format
-from google.protobuf.struct_pb2 import Value
+from google.cloud.aiplatform.schema import trainingjob
 
 
 def create_training_pipeline_video_classification_sample(
@@ -31,8 +30,10 @@ def create_training_pipeline_video_classification_sample(
     # Initialize client that will be used to create and send requests.
     # This client only needs to be created once, and can be reused for multiple requests.
     client = aiplatform.gapic.PipelineServiceClient(client_options=client_options)
-    training_task_inputs_dict = {}
-    training_task_inputs = json_format.ParseDict(training_task_inputs_dict, Value())
+    training_task_inputs_object = (
+        trainingjob.definition.AutoMlVideoClassificationInputs()
+    )
+    training_task_inputs = training_task_inputs_object.to_value()
 
     training_pipeline = {
         "display_name": display_name,
diff --git a/samples/snippets/create_training_pipeline_video_object_tracking_sample.py b/samples/snippets/create_training_pipeline_video_object_tracking_sample.py
@@ -14,8 +14,7 @@
 
 # [START aiplatform_create_training_pipeline_video_object_tracking_sample]
 from google.cloud import aiplatform
-from google.protobuf import json_format
-from google.protobuf.struct_pb2 import Value
+from google.cloud.aiplatform.schema import trainingjob
 
 
 def create_training_pipeline_video_object_tracking_sample(
@@ -31,8 +30,10 @@ def create_training_pipeline_video_object_tracking_sample(
     # Initialize client that will be used to create and send requests.
     # This client only needs to be created once, and can be reused for multiple requests.
     client = aiplatform.gapic.PipelineServiceClient(client_options=client_options)
-    training_task_inputs_dict = {"modelType": "CLOUD"}
-    training_task_inputs = json_format.ParseDict(training_task_inputs_dict, Value())
+    training_task_inputs_object = trainingjob.definition.AutoMlVideoObjectTrackingInputs(
+        model_type="CLOUD"
+    )
+    training_task_inputs = training_task_inputs_object.to_value()
 
     training_pipeline = {
         "display_name": display_name,
