Update tests for nappy, museum_victoria, provider_data_ingester

catalog/dags/providers/provider_api_scripts/nappy.py

@@ -26,7 +26,7 @@ class NappyDataIngester(ProviderDataIngester):
     endpoint = "https://api.nappy.co/v1/openverse/images"
     headers = {"User-Agent": prov.UA_STRING, "Accept": "application/json"}
 
-    # Hardoded to CC0, the only license Nappy.co uses
+    # Hardcoded to CC0, the only license Nappy.co uses
     license_info = get_license_info(
         "https://creativecommons.org/publicdomain/zero/1.0/"
     )

catalog/tests/dags/providers/provider_api_scripts/test_museum_victoria.py

@@ -119,6 +119,18 @@ def test_no_duplicate_records():
     assert actual_image_data is None
 
 
+@pytest.mark.parametrize(
+    "falsy_parameter",
+    ["id", "media"],
+)
+def test_get_record_data_returns_none_with_falsy_param(falsy_parameter):
+    media = _get_resource_json("record_data.json")
+    media[falsy_parameter] = ""
+    actual_image_data = mv.get_record_data(media)
+
+    assert actual_image_data is None
+
+
 def test_get_images_success():
     media = _get_resource_json("media_data_success.json")
     actual_image_data = mv._get_images([media])
@@ -144,52 +156,29 @@ def test_get_media_info_failure():
     assert actual_image_data is None
 
 
-def test_get_image_data_large():
-    image_data = _get_resource_json("large_image_data.json")
-
-    actual_image_url, actual_height, actual_width, actual_filesize = mv._get_image_data(
-        image_data
-    )
-
-    assert actual_image_url == (
-        "https://collections.museumsvictoria.com.au/content/media/45/"
-        "329745-large.jpg"
-    )
-    assert actual_height == 2581
-    assert actual_width == 2785
-    assert actual_filesize == 890933
-
-
-def test_get_image_data_medium():
-    image_data = _get_resource_json("medium_image_data.json")
-
-    actual_image_url, actual_height, actual_width, actual_filesize = mv._get_image_data(
-        image_data
-    )
-
-    assert actual_image_url == (
-        "https://collections.museumsvictoria.com.au/content/media/45/"
-        "329745-medium.jpg"
+@pytest.mark.parametrize(
+    "image_size, expected_height, expected_width, expected_filesize",
+    [
+        pytest.param("large", 2581, 2785, 890933, id="large"),
+        pytest.param("medium", 1390, 1500, 170943, id="medium"),
+        pytest.param("small", 500, 540, 20109, id="small"),
+    ],
+)
+def test_get_image_data(image_size, expected_height, expected_width, expected_filesize):
+    image_data = _get_resource_json(f"{image_size}_image_data.json")
+    expected_url = (
+        f"https://collections.museumsvictoria.com.au/content/media/45/"
+        f"329745-{image_size}.jpg"
     )
-    assert actual_height == 1390
-    assert actual_width == 1500
-    assert actual_filesize == 170943
-
-
-def test_get_image_data_small():
-    image_data = _get_resource_json("small_image_data.json")
 
     actual_image_url, actual_height, actual_width, actual_filesize = mv._get_image_data(
         image_data
     )
 
-    assert actual_image_url == (
-        "https://collections.museumsvictoria.com.au/content/media/45/"
-        "329745-small.jpg"
-    )
-    assert actual_height == 500
-    assert actual_width == 540
-    assert actual_filesize == 20109
+    assert actual_image_url == expected_url
+    assert actual_height == expected_height
+    assert actual_width == expected_width
+    assert actual_filesize == expected_filesize
 
 
 def test_get_image_data_none():

catalog/tests/dags/providers/provider_api_scripts/test_nappy.py

@@ -89,6 +89,17 @@ def test_get_should_continue(response_json, expected_result):
     [
         pytest.param({}, None, id="empty_dict"),
         pytest.param(FULL_BATCH_RESPONSE, None, id="no_urls"),
+        pytest.param(
+            {**SINGLE_ITEM, "foreign_landing_url": ""},
+            None,
+            id="falsy_foreign_landing_url",
+        ),
+        pytest.param(
+            {**SINGLE_ITEM, "foreign_identifier": ""},
+            None,
+            id="falsy_foreign_identifier",
+        ),
+        pytest.param({**SINGLE_ITEM, "url": ""}, None, id="falsy_url"),
         pytest.param(
             SINGLE_ITEM,
             {

catalog/tests/dags/providers/provider_api_scripts/test_provider_data_ingester.py

@@ -177,6 +177,23 @@ def test_process_batch_handles_list_of_records():
         assert image_store_mock.call_count == 1
 
 
+def test_process_batch_handles_empty_dictionary_of_records_from_get_record_data():
+    with (
+        patch.object(audio_store, "add_item") as audio_store_mock,
+        patch.object(image_store, "add_item") as image_store_mock,
+        patch.object(ingester, "get_record_data") as get_record_data_mock,
+    ):
+        # Mock `get_record_data` to return an empty list of records
+        get_record_data_mock.return_value = []
+
+        record_count = ingester.process_batch([{}])
+
+        # Both records are added, and to the appropriate stores
+        assert record_count == 0
+        assert audio_store_mock.call_count == 0
+        assert image_store_mock.call_count == 0
+
+
 def test_process_batch_halts_processing_after_reaching_ingestion_limit():
     # Set up an ingester with an ingestion limit of 1
     ingester = MockProviderDataIngester()