Add config for fixture profiling

Author: Val Brodsky

Makefile

@@ -13,6 +13,7 @@ test-local: build-image
 		-e LABELBOX_TEST_ENVIRON="local" \
 		-e DA_GCP_LABELBOX_API_KEY=${DA_GCP_LABELBOX_API_KEY} \
 		-e LABELBOX_TEST_API_KEY_LOCAL=${LABELBOX_TEST_API_KEY_LOCAL} \
+		-e FIXTURE_PROFILE=true \
 		local/labelbox-python:test pytest $(PATH_TO_TEST)
 
 test-staging: build-image

tests/integration/annotation_import/conftest.py

@@ -498,7 +498,6 @@ def get_data_row_id(indx=0):
 
 @pytest.fixture
 def configured_project(client, initial_dataset, ontology, rand_gen, image_url):
-    start_time = time.time()
     dataset = initial_dataset
     project = client.create_project(name=rand_gen(str),
                                     queue_mode=QueueMode.Batch)
@@ -515,15 +514,12 @@ def configured_project(client, initial_dataset, ontology, rand_gen, image_url):
         num_rows += 1
     project._wait_until_data_rows_are_processed(data_row_ids=data_row_ids,
                                                 sleep_interval=3)
-    if pytest.data_row_report:
-        pytest.data_row_report['times'] += time.time() - start_time
-        pytest.data_row_report['num_rows'] += num_rows
+
     project.create_batch(
         rand_gen(str),
         data_row_ids,  # sample of data row objects
         5  # priority between 1(Highest) - 5(lowest)
     )
-    print("After creating batch ", time.time() - start_time)
     project.data_row_ids = data_row_ids
 
     yield project
@@ -579,8 +575,6 @@ def dataset_conversation_entity(client, rand_gen, conversation_entity_data_row,
 @pytest.fixture
 def configured_project_with_one_data_row(client, ontology, rand_gen,
                                          initial_dataset, image_url):
-    start_time = time.time()
-
     project = client.create_project(name=rand_gen(str),
                                     description=rand_gen(str),
                                     queue_mode=QueueMode.Batch)
@@ -594,9 +588,6 @@ def configured_project_with_one_data_row(client, ontology, rand_gen,
     project._wait_until_data_rows_are_processed(data_row_ids=data_row_ids,
                                                 sleep_interval=3)
 
-    if pytest.data_row_report:
-        pytest.data_row_report['times'] += time.time() - start_time
-        pytest.data_row_report['num_rows'] += 1
     batch = project.create_batch(
         rand_gen(str),
         data_row_ids,  # sample of data row objects

tests/integration/annotation_import/test_data_types.py

@@ -135,10 +135,15 @@ def create_data_row_for_project(project, dataset, data_row_ndjson, batch_name):
     AudioData, ConversationData, DicomData, DocumentData, HTMLData, ImageData,
     TextData
 ])
-def test_import_data_types(client, configured_project, initial_dataset,
-                           rand_gen, data_row_json_by_data_type,
-                           annotations_by_data_type, data_type_class,
-                           one_datarow):
+def test_import_data_types(
+    client,
+    configured_project,
+    initial_dataset,
+    rand_gen,
+    data_row_json_by_data_type,
+    annotations_by_data_type,
+    data_type_class,
+):
 
     project = configured_project
     project_id = project.uid

tests/integration/conftest.py

@@ -27,8 +27,6 @@
 
 IMG_URL = "https://picsum.photos/200/300.jpg"
 SMALL_DATASET_URL = "https://storage.googleapis.com/lb-artifacts-testing-public/sdk_integration_test/potato.jpeg"
-DATA_ROW_PROCESSING_WAIT_TIMEOUT_SECONDS = 30
-DATA_ROW_PROCESSING_WAIT_SLEEP_INTERNAL_SECONDS = 5
 
 
 class Environ(Enum):
@@ -458,10 +456,8 @@ def configured_batch_project_with_label(project, dataset, data_row,
     One label is already created and yielded when using fixture
     """
     data_rows = [dr.uid for dr in list(dataset.data_rows())]
-    project._wait_until_data_rows_are_processed(
-        data_row_ids=data_rows,
-        wait_processing_max_seconds=DATA_ROW_PROCESSING_WAIT_TIMEOUT_SECONDS,
-        sleep_interval=DATA_ROW_PROCESSING_WAIT_SLEEP_INTERNAL_SECONDS)
+    project._wait_until_data_rows_are_processed(data_row_ids=data_rows,
+                                                sleep_interval=3)
     project.create_batch("test-batch", data_rows)
     project.data_row_ids = data_rows
 
@@ -604,7 +600,6 @@ def configured_project_with_complex_ontology(client, initial_dataset, rand_gen,
     project.setup(editor, ontology.asdict())
 
     yield [project, data_row]
-    dataset.delete()
     project.delete()
 
 
@@ -825,35 +820,31 @@ def upload_invalid_data_rows_for_dataset(dataset: Dataset):
     task.wait_till_done()
 
 
-@pytest.mark.skipif("FIXTURE_PROFILE" not in os.environ)
 def pytest_configure():
     pytest.report = defaultdict(int)
-    pytest.data_row_report = {'times': 0, 'num_rows': 0}
 
 
-@pytest.mark.skipif("FIXTURE_PROFILE" not in os.environ)
 @pytest.hookimpl(hookwrapper=True)
-def pytest_fixture_setup(fixturedef, request):
+def pytest_fixture_setup(fixturedef):
     start = time.time()
     yield
-
     end = time.time()
 
     exec_time = end - start
-    pytest.report[fixturedef.argname] += exec_time
+    if "FIXTURE_PROFILE" in os.environ:
+        pytest.report[fixturedef.argname] += exec_time
 
 
-@pytest.mark.skipif("FIXTURE_PROFILE" not in os.environ)
 @pytest.fixture(scope='session', autouse=True)
 def print_perf_summary():
     yield
 
-    sorted_dict = dict(
-        sorted(pytest.report.items(), key=lambda item: item[1], reverse=True))
-    num_of_entries = 10 if len(sorted_dict) >= 10 else len(sorted_dict)
-    slowest_fixtures = [
-        (aaa, sorted_dict[aaa]) for aaa in islice(sorted_dict, num_of_entries)
-    ]
-    print("\nTop slowest fixtures:\n", slowest_fixtures, file=sys.stderr)
-    print("Data row report:\n", pytest.data_row_report, file=sys.stderr)
-    # assert False
+    if "FIXTURE_PROFILE" in os.environ:
+        sorted_dict = dict(
+            sorted(pytest.report.items(),
+                   key=lambda item: item[1],
+                   reverse=True))
+        num_of_entries = 10 if len(sorted_dict) >= 10 else len(sorted_dict)
+        slowest_fixtures = [(aaa, sorted_dict[aaa])
+                            for aaa in islice(sorted_dict, num_of_entries)]
+        print("\nTop slowest fixtures:\n", slowest_fixtures, file=sys.stderr)