From 3b68e3a3dde1aee5c0b00e27beabe83921f426ac Mon Sep 17 00:00:00 2001 From: Matthew Roberson Date: Tue, 28 Apr 2026 10:31:04 -0500 Subject: [PATCH] [PLT-0] update test assets --- libs/labelbox/tests/conftest.py | 7 ++-- .../tests/data/annotation_import/conftest.py | 39 +++++++------------ libs/labelbox/tests/integration/conftest.py | 4 +- .../tests/integration/test_data_rows.py | 4 +- .../integration/test_data_rows_upsert.py | 6 +-- libs/lbox-alignerr/tests/conftest.py | 7 ++-- 6 files changed, 27 insertions(+), 40 deletions(-) diff --git a/libs/labelbox/tests/conftest.py b/libs/labelbox/tests/conftest.py index a2ffdd49d..42bb5e5e8 100644 --- a/libs/labelbox/tests/conftest.py +++ b/libs/labelbox/tests/conftest.py @@ -34,7 +34,6 @@ from labelbox.schema.quality_mode import QualityMode IMG_URL = "https://picsum.photos/200/300.jpg" -MASKABLE_IMG_URL = "https://storage.googleapis.com/labelbox-datasets/image_sample_data/2560px-Kitano_Street_Kobe01s5s4110.jpeg" 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 = 3 @@ -75,7 +74,7 @@ class Environ(Enum): @pytest.fixture def image_url() -> str: - return MASKABLE_IMG_URL + return IMG_URL @pytest.fixture @@ -880,8 +879,8 @@ def video_data(client, rand_gen, video_data_row, wait_for_data_row_processing): def create_video_data_row(rand_gen): return { - "row_data": "https://storage.googleapis.com/labelbox-datasets/video-sample-data/sample-video-1.mp4", - "global_key": f"https://storage.googleapis.com/labelbox-datasets/video-sample-data/sample-video-1.mp4-{rand_gen(str)}", + "row_data": "https://storage.googleapis.com/lb-test-data/cataflow/media/test_video_500kb.mp4", + "global_key": f"https://storage.googleapis.com/lb-test-data/cataflow/media/test_video_500kb.mp4-{rand_gen(str)}", "media_type": "VIDEO", } diff --git a/libs/labelbox/tests/data/annotation_import/conftest.py b/libs/labelbox/tests/data/annotation_import/conftest.py index f43151f2b..c3f574b2c 100644 --- a/libs/labelbox/tests/data/annotation_import/conftest.py +++ b/libs/labelbox/tests/data/annotation_import/conftest.py @@ -26,8 +26,8 @@ def video_data_row_factory(): def video_data_row(global_key): return { - "row_data": "https://storage.googleapis.com/labelbox-datasets/video-sample-data/sample-video-1.mp4", - "global_key": f"https://storage.googleapis.com/labelbox-datasets/video-sample-data/sample-video-1.mp4-{global_key}", + "row_data": "https://storage.googleapis.com/lb-test-data/cataflow/media/test_video_500kb.mp4", + "global_key": f"https://storage.googleapis.com/lb-test-data/cataflow/media/test_video_500kb.mp4-{global_key}", "media_type": "VIDEO", } @@ -38,8 +38,8 @@ def video_data_row(global_key): def audio_data_row_factory(): def audio_data_row(global_key): return { - "row_data": "https://storage.googleapis.com/labelbox-datasets/audio-sample-data/sample-audio-1.mp3", - "global_key": f"https://storage.googleapis.com/labelbox-datasets/audio-sample-data/sample-audio-1.mp3-{global_key}", + "row_data": "https://storage.googleapis.com/lb-test-data/cataflow/media/tomorrow-114848.mp3", + "global_key": f"https://storage.googleapis.com/lb-test-data/cataflow/media/tomorrow-114848.mp3-{global_key}", "media_type": "AUDIO", } @@ -82,8 +82,8 @@ def geospatial_data_row(global_key): def html_data_row_factory(): def html_data_row(global_key): return { - "row_data": "https://storage.googleapis.com/labelbox-datasets/html_sample_data/sample_html_1.html", - "global_key": f"https://storage.googleapis.com/labelbox-datasets/html_sample_data/sample_html_1.html-{global_key}", + "row_data": "https://storage.googleapis.com/lb-test-data/cataflow/media/about.html", + "global_key": f"https://storage.googleapis.com/lb-test-data/cataflow/media/about.html-{global_key}", } return html_data_row @@ -106,10 +106,10 @@ def document_data_row_factory(): def document_data_row(global_key): return { "row_data": { - "pdf_url": "https://storage.googleapis.com/labelbox-datasets/arxiv-pdf/data/99-word-token-pdfs/0801.3483.pdf", - "text_layer_url": "https://storage.googleapis.com/labelbox-datasets/arxiv-pdf/data/99-word-token-pdfs/0801.3483-lb-textlayer.json", + "pdf_url": "https://storage.googleapis.com/lb-test-data/cataflow/media/test-pdf.pdf", + "text_layer_url": "https://storage.googleapis.com/lb-test-data/cataflow/media/sample-pdf-textlayer.json", }, - "global_key": f"https://storage.googleapis.com/labelbox-datasets/arxiv-pdf/data/99-word-token-pdfs/0801.3483.pdf-{global_key}", + "global_key": f"https://storage.googleapis.com/lb-test-data/cataflow/media/test-pdf.pdf-{global_key}", "media_type": "PDF", } @@ -120,28 +120,17 @@ def document_data_row(global_key): def text_data_row_factory(): def text_data_row(global_key): return { - "row_data": "https://storage.googleapis.com/lb-artifacts-testing-public/sdk_integration_test/sample-text-2.txt", - "global_key": f"https://storage.googleapis.com/lb-artifacts-testing-public/sdk_integration_test/sample-text-2.txt-{global_key}", + "row_data": "https://storage.googleapis.com/lb-test-data/cataflow/media/sample-text.txt", + "global_key": f"https://storage.googleapis.com/lb-test-data/cataflow/media/sample-text.txt-{global_key}", "media_type": "TEXT", } return text_data_row -@pytest.fixture(scope="module", autouse=True) -def llm_human_preference_data_row_factory(): - def llm_human_preference_data_row(global_key): - return { - "row_data": "https://storage.googleapis.com/labelbox-datasets/sdk_test/llm_prompt_response_conv.json", - "global_key": global_key, - } - - return llm_human_preference_data_row - - @pytest.fixture(scope="module") def mmc_data_row_url(): - return "https://storage.googleapis.com/labelbox-datasets/conversational_model_evaluation_sample/offline-model-chat-evaluation.json" + return "https://storage.googleapis.com/lb-test-data/cataflow/media/sample-conversational-v2-4.json" @pytest.fixture(scope="module", autouse=True) @@ -700,7 +689,7 @@ def llm_prompt_response_creation_dataset_with_data_row( global_key = str(uuid.uuid4()) convo_data = { - "row_data": "https://storage.googleapis.com/labelbox-datasets/conversational-sample-data/pairwise_shopping_2.json", + "row_data": "https://storage.googleapis.com/lb-test-data/cataflow/media/sample-conversational-v2-4.json", "global_key": global_key, } @@ -1321,7 +1310,7 @@ def segmentation_inference(prediction_id_mapping): segmentation.update( { "mask": { - "instanceURI": "https://storage.googleapis.com/labelbox-datasets/image_sample_data/raster_seg.png", + "instanceURI": "https://storage.googleapis.com/lb-test-data/cataflow/media/dog.png", "colorRGB": (255, 255, 255), } } diff --git a/libs/labelbox/tests/integration/conftest.py b/libs/labelbox/tests/integration/conftest.py index 58d94aa43..c0d4ef8cc 100644 --- a/libs/labelbox/tests/integration/conftest.py +++ b/libs/labelbox/tests/integration/conftest.py @@ -245,8 +245,8 @@ def video_data(client, rand_gen, video_data_row, wait_for_data_row_processing): def create_video_data_row(rand_gen): return { - "row_data": "https://storage.googleapis.com/labelbox-datasets/video-sample-data/sample-video-1.mp4", - "global_key": f"https://storage.googleapis.com/labelbox-datasets/video-sample-data/sample-video-1.mp4-{rand_gen(str)}", + "row_data": "https://storage.googleapis.com/lb-test-data/cataflow/media/test_video_500kb.mp4", + "global_key": f"https://storage.googleapis.com/lb-test-data/cataflow/media/test_video_500kb.mp4-{rand_gen(str)}", "media_type": "VIDEO", } diff --git a/libs/labelbox/tests/integration/test_data_rows.py b/libs/labelbox/tests/integration/test_data_rows.py index fc668c05f..ba3c108d5 100644 --- a/libs/labelbox/tests/integration/test_data_rows.py +++ b/libs/labelbox/tests/integration/test_data_rows.py @@ -689,8 +689,8 @@ def test_data_row_update( assert data_row.row_data == image_url # tileLayer becomes a media attribute - pdf_url = "https://storage.googleapis.com/labelbox-datasets/arxiv-pdf/data/99-word-token-pdfs/0801.3483.pdf" - tileLayerUrl = "https://storage.googleapis.com/labelbox-datasets/arxiv-pdf/data/99-word-token-pdfs/0801.3483-lb-textlayer.json" + pdf_url = "https://storage.googleapis.com/lb-test-data/cataflow/media/test-pdf.pdf" + tileLayerUrl = "https://storage.googleapis.com/lb-test-data/cataflow/media/sample-pdf-textlayer.json" data_row.update(row_data={"pdfUrl": pdf_url, "tileLayerUrl": tileLayerUrl}) def custom_check(data_row): diff --git a/libs/labelbox/tests/integration/test_data_rows_upsert.py b/libs/labelbox/tests/integration/test_data_rows_upsert.py index 2ba7a9df9..b20eccb86 100644 --- a/libs/labelbox/tests/integration/test_data_rows_upsert.py +++ b/libs/labelbox/tests/integration/test_data_rows_upsert.py @@ -29,7 +29,7 @@ def all_inclusive_data_row(self, dataset, image_url): { "type": "PDF_URL", "name": "att3", - "value": "https://storage.googleapis.com/labelbox-datasets/arxiv-pdf/data/99-word-token-pdfs/0801.3483.pdf", + "value": "https://storage.googleapis.com/lb-test-data/cataflow/media/test-pdf.pdf", }, ], ) @@ -63,7 +63,7 @@ def test_create_data_row_with_upsert(self, client, dataset, image_url): { "type": AttachmentType.PDF_URL, "name": "att3", - "value": "https://storage.googleapis.com/labelbox-datasets/arxiv-pdf/data/99-word-token-pdfs/0801.3483.pdf", + "value": "https://storage.googleapis.com/lb-test-data/cataflow/media/test-pdf.pdf", }, ], "metadata": [ @@ -99,7 +99,7 @@ def test_create_data_row_with_upsert(self, client, dataset, image_url): assert attachments[2].attachment_type == AttachmentType.PDF_URL assert ( attachments[2].attachment_value - == "https://storage.googleapis.com/labelbox-datasets/arxiv-pdf/data/99-word-token-pdfs/0801.3483.pdf" + == "https://storage.googleapis.com/lb-test-data/cataflow/media/test-pdf.pdf" ) assert len(dr.metadata_fields) == 2 diff --git a/libs/lbox-alignerr/tests/conftest.py b/libs/lbox-alignerr/tests/conftest.py index 27e1140e1..5368b164e 100644 --- a/libs/lbox-alignerr/tests/conftest.py +++ b/libs/lbox-alignerr/tests/conftest.py @@ -34,7 +34,6 @@ from labelbox.schema.quality_mode import QualityMode IMG_URL = "https://picsum.photos/200/300.jpg" -MASKABLE_IMG_URL = "https://storage.googleapis.com/labelbox-datasets/image_sample_data/2560px-Kitano_Street_Kobe01s5s4110.jpeg" 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 = 3 @@ -72,7 +71,7 @@ class Environ(Enum): @pytest.fixture def image_url() -> str: - return MASKABLE_IMG_URL + return IMG_URL @pytest.fixture @@ -852,8 +851,8 @@ def video_data(client, rand_gen, video_data_row, wait_for_data_row_processing): def create_video_data_row(rand_gen): return { - "row_data": "https://storage.googleapis.com/labelbox-datasets/video-sample-data/sample-video-1.mp4", - "global_key": f"https://storage.googleapis.com/labelbox-datasets/video-sample-data/sample-video-1.mp4-{rand_gen(str)}", + "row_data": "https://storage.googleapis.com/lb-test-data/cataflow/media/test_video_500kb.mp4", + "global_key": f"https://storage.googleapis.com/lb-test-data/cataflow/media/test_video_500kb.mp4-{rand_gen(str)}", "media_type": "VIDEO", }