datalabeling: ensure all tests use test endpoint (#2918)

* datalabeling: ensure all tests use test endpoint

* requires an input csv for text input, slight print statement cleanup

Co-authored-by: Leah E. Cole <[email protected]>

Author: nnegrey

datalabeling/create_instruction.py

@@ -53,7 +53,7 @@ def create_instruction(project_id, data_type, instruction_gcs_uri):
 
     # The format of the resource name:
     # project_id/{project_id}/instruction/{instruction_id}
-    print('The instruction resource name: {}\n'.format(result.name))
+    print('The instruction resource name: {}'.format(result.name))
     print('Display name: {}'.format(result.display_name))
     print('Description: {}'.format(result.description))
     print('Create time:')
@@ -62,7 +62,7 @@ def create_instruction(project_id, data_type, instruction_gcs_uri):
     print('Data type: {}'.format(
         datalabeling.enums.DataType(result.data_type).name))
     print('Pdf instruction:')
-    print('\tGcs file uri: {}'.format(
+    print('\tGcs file uri: {}\n'.format(
         result.pdf_instruction.gcs_file_uri))
 
     return result

datalabeling/label_text.py

@@ -47,9 +47,9 @@ def label_text(dataset_resource_name, instruction_resource_name,
         annotation_spec_set=annotation_spec_set_resource_name)
 
     response = client.label_text(
-        dataset_resource_name,
-        basic_config,
-        feature,
+        parent=dataset_resource_name,
+        basic_config=basic_config,
+        feature=feature,
         text_entity_extraction_config=config
     )
 

datalabeling/label_text_test.py

@@ -26,7 +26,7 @@
 import pytest
 
 PROJECT_ID = os.getenv('GCLOUD_PROJECT')
-INPUT_GCS_URI = 'gs://cloud-samples-data/datalabeling/text/text_dataset.csv'
+INPUT_GCS_URI = 'gs://cloud-samples-data/datalabeling/text/input.csv'
 
 
 @pytest.fixture(scope='function')

datalabeling/manage_dataset.py

@@ -35,20 +35,20 @@ def create_dataset(project_id):
     formatted_project_name = client.project_path(project_id)
 
     dataset = datalabeling.types.Dataset(
-        display_name='YOUR_ANNOTATION_SPEC_SET_DISPLAY_NAME',
+        display_name='YOUR_DATASET_SET_DISPLAY_NAME',
         description='YOUR_DESCRIPTION'
     )
 
     response = client.create_dataset(formatted_project_name, dataset)
 
     # The format of resource name:
     # project_id/{project_id}/datasets/{dataset_id}
-    print('The dataset resource name: {}\n'.format(response.name))
+    print('The dataset resource name: {}'.format(response.name))
     print('Display name: {}'.format(response.display_name))
     print('Description: {}'.format(response.description))
     print('Create time:')
     print('\tseconds: {}'.format(response.create_time.seconds))
-    print('\tnanos: {}'.format(response.create_time.nanos))
+    print('\tnanos: {}\n'.format(response.create_time.nanos))
 
     return response
 # [END datalabeling_create_dataset_beta]
@@ -59,6 +59,13 @@ def list_datasets(project_id):
     """Lists datasets for the given Google Cloud project."""
     from google.cloud import datalabeling_v1beta1 as datalabeling
     client = datalabeling.DataLabelingServiceClient()
+    # [END datalabeling_list_datasets_beta]
+    # If provided, use a provided test endpoint - this will prevent tests on
+    # this snippet from triggering any action by a real human
+    if 'DATALABELING_ENDPOINT' in os.environ:
+        opts = ClientOptions(api_endpoint=os.getenv('DATALABELING_ENDPOINT'))
+        client = datalabeling.DataLabelingServiceClient(client_options=opts)
+    # [START datalabeling_list_datasets_beta]
 
     formatted_project_name = client.project_path(project_id)
 
@@ -80,6 +87,13 @@ def get_dataset(dataset_resource_name):
     """Gets a dataset for the given Google Cloud project."""
     from google.cloud import datalabeling_v1beta1 as datalabeling
     client = datalabeling.DataLabelingServiceClient()
+    # [END datalabeling_get_dataset_beta]
+    # If provided, use a provided test endpoint - this will prevent tests on
+    # this snippet from triggering any action by a real human
+    if 'DATALABELING_ENDPOINT' in os.environ:
+        opts = ClientOptions(api_endpoint=os.getenv('DATALABELING_ENDPOINT'))
+        client = datalabeling.DataLabelingServiceClient(client_options=opts)
+    # [START datalabeling_get_dataset_beta]
 
     response = client.get_dataset(dataset_resource_name)
 
@@ -97,6 +111,13 @@ def delete_dataset(dataset_resource_name):
     """Deletes a dataset for the given Google Cloud project."""
     from google.cloud import datalabeling_v1beta1 as datalabeling
     client = datalabeling.DataLabelingServiceClient()
+    # [END datalabeling_delete_dataset_beta]
+    # If provided, use a provided test endpoint - this will prevent tests on
+    # this snippet from triggering any action by a real human
+    if 'DATALABELING_ENDPOINT' in os.environ:
+        opts = ClientOptions(api_endpoint=os.getenv('DATALABELING_ENDPOINT'))
+        client = datalabeling.DataLabelingServiceClient(client_options=opts)
+    # [START datalabeling_delete_dataset_beta]
 
     response = client.delete_dataset(dataset_resource_name)