diff --git a/arkindex_worker/worker/training.py b/arkindex_worker/worker/training.py index c53cffee4f2633307a94b9131ca3146415e5c283..ffe52d6f9632882743dbc7a5340d4fc5b1c5ef66 100644 --- a/arkindex_worker/worker/training.py +++ b/arkindex_worker/worker/training.py @@ -50,12 +50,10 @@ def create_archive(path: DirPath) -> Tuple[Path, Hash, FileSize, Hash]: # Create an uncompressed tar archive with all the needed files # Files hierarchy ifs kept in the archive. - file_list = [] + with tarfile.open(path_to_tar_archive, "w") as tar: - for p in path.glob("**/*"): - x = p.relative_to(path) - tar.add(p, arcname=x, recursive=False) - file_list.append(p) + tar.add(path) + file_list = [member for member in tar.getnames() if os.path.isfile(member)] # Sort by path file_list.sort() diff --git a/tests/conftest.py b/tests/conftest.py index d058dc878d05de7160d06ce8d0380008d5711a14..8b6f16ef834712c769badc8308143c81e58a3a88 100644 --- a/tests/conftest.py +++ b/tests/conftest.py @@ -26,7 +26,7 @@ from arkindex_worker.worker import BaseWorker, ElementsWorker from arkindex_worker.worker.transcription import TextOrientation FIXTURES_DIR = Path(__file__).resolve().parent / "data" -SAMPLES_DIR = Path(__file__).resolve().parent / "samples" +SAMPLES_DIR = Path("tests") / "samples" __yaml_cache = {} @@ -280,6 +280,11 @@ def model_file_dir(): return SAMPLES_DIR / "model_files" +@pytest.fixture +def model_file_dir_with_subfolder(): + return SAMPLES_DIR / "model_files_with_subfolder" + + @pytest.fixture def fake_dummy_worker(): api_client = MockApiClient() diff --git a/tests/samples/model_files_with_subfolder/model-best/model_file.pth b/tests/samples/model_files_with_subfolder/model-best/model_file.pth new file mode 100644 index 0000000000000000000000000000000000000000..cc78ba3026c620f5d0e8c5b65071ae8ae2dfe157 --- /dev/null +++ b/tests/samples/model_files_with_subfolder/model-best/model_file.pth @@ -0,0 +1 @@ +Wow this is actually the data of the best model ever created on Arkindex \ No newline at end of file diff --git a/tests/samples/model_files_with_subfolder/model-last/model_file.pth b/tests/samples/model_files_with_subfolder/model-last/model_file.pth new file mode 100644 index 0000000000000000000000000000000000000000..cc78ba3026c620f5d0e8c5b65071ae8ae2dfe157 --- /dev/null +++ b/tests/samples/model_files_with_subfolder/model-last/model_file.pth @@ -0,0 +1 @@ +Wow this is actually the data of the best model ever created on Arkindex \ No newline at end of file diff --git a/tests/test_elements_worker/test_training.py b/tests/test_elements_worker/test_training.py index e8f419621bbf964a4313c93c130ffe14d049af7f..4448554da951fcaaa87ccde715af8c31865a92fd 100644 --- a/tests/test_elements_worker/test_training.py +++ b/tests/test_elements_worker/test_training.py @@ -45,6 +45,24 @@ def test_create_archive(model_file_dir): assert not os.path.exists(zst_archive_path), "Auto removal failed" +def test_create_archive_with_subfolder(model_file_dir_with_subfolder): + """Create an archive when the model's file is in a folder containing a subfolder""" + + with create_archive(path=model_file_dir_with_subfolder) as ( + zst_archive_path, + hash, + size, + archive_hash, + ): + assert os.path.exists(zst_archive_path), "The archive was not created" + assert ( + hash == "e2fa86cefc33b24502ad4151a638dd29" + ), "Hash was not properly computed" + assert 300 < size < 1500 + + assert not os.path.exists(zst_archive_path), "Auto removal failed" + + @pytest.mark.parametrize( "tag, description", [