feat: Add comprehensive Python testing infrastructure with Poetry

.gitignore

@@ -167,3 +167,9 @@ group_config/
 # local dev
 local/
 *legacy*
+
+# Claude settings
+.claude/
+
+# Poetry - DO NOT ignore poetry.lock
+# poetry.lock is intentionally not ignored to ensure consistent dependencies

pyproject.toml

@@ -0,0 +1,104 @@
+[tool.poetry]
+name = "openood"
+version = "1.5"
+description = "This package provides a unified test platform for Out-of-Distribution detection."
+authors = ["openood dev team <[email protected]>"]
+readme = "README.md"
+homepage = "https://github.com/Jingkang50/OpenOOD"
+repository = "https://github.com/Jingkang50/OpenOOD"
+classifiers = [
+    "Programming Language :: Python :: 3",
+    "License :: OSI Approved :: Apache Software License",
+    "Operating System :: OS Independent",
+]
+packages = [{include = "openood"}]
+
+[tool.poetry.dependencies]
+python = "^3.8"
+torch = ">=1.13.1"
+torchvision = ">=0.13"
+scikit-learn = "*"
+json5 = "*"
+matplotlib = "*"
+scipy = "*"
+tqdm = "*"
+pyyaml = ">=5.4.1"
+pre-commit = "*"
+opencv-python = ">=4.4.0.46"
+imgaug = ">=0.4.0"
+pandas = "*"
+diffdist = ">=0.1"
+Cython = ">=0.29.30"
+faiss-cpu = ">=1.7.2"
+gdown = ">=4.7.1"
+
+[tool.poetry.group.dev.dependencies]
+pytest = "^7.4.0"
+pytest-cov = "^4.1.0"
+pytest-mock = "^3.11.1"
+
+[tool.poetry.scripts]
+test = "pytest:main"
+tests = "pytest:main"
+
+[tool.pytest.ini_options]
+minversion = "7.0"
+testpaths = ["tests"]
+python_files = ["test_*.py", "*_test.py"]
+python_classes = ["Test*", "*Test"]
+python_functions = ["test_*"]
+addopts = [
+    "-v",
+    "--strict-markers",
+    "--cov=openood",
+    "--cov-report=term-missing",
+    "--cov-report=html",
+    "--cov-report=xml",
+    "--cov-fail-under=0",  # TODO: Set to 80 once tests are written
+]
+markers = [
+    "unit: Unit tests",
+    "integration: Integration tests",
+    "slow: Slow running tests",
+]
+filterwarnings = [
+    "ignore::DeprecationWarning",
+    "ignore::PendingDeprecationWarning",
+]
+
+[tool.coverage.run]
+source = ["openood"]
+omit = [
+    "*/tests/*",
+    "*/test_*.py",
+    "*/__init__.py",
+    "*/setup.py",
+    "*/venv/*",
+    "*/.venv/*",
+]
+
+[tool.coverage.report]
+exclude_lines = [
+    "pragma: no cover",
+    "def __repr__",
+    "def __str__",
+    "raise AssertionError",
+    "raise NotImplementedError",
+    "if __name__ == .__main__.:",
+    "if TYPE_CHECKING:",
+    "class .*\\bProtocol\\):",
+    "@(abc\\.)?abstractmethod",
+]
+precision = 2
+show_missing = true
+skip_covered = false
+
+[tool.coverage.html]
+directory = "htmlcov"
+
+[tool.coverage.xml]
+output = "coverage.xml"
+
+[build-system]
+requires = ["poetry-core>=1.0.0"]
+build-backend = "poetry.core.masonry.api"

tests/conftest.py

@@ -0,0 +1,260 @@
+"""Shared pytest fixtures and configuration for OpenOOD tests."""
+
+import tempfile
+import shutil
+from pathlib import Path
+from typing import Dict, Any, List, Generator
+import pytest
+import torch
+import numpy as np
+import yaml
+import json
+
+
+@pytest.fixture
+def temp_dir() -> Generator[Path, None, None]:
+    """Create a temporary directory for test files."""
+    temp_path = tempfile.mkdtemp()
+    yield Path(temp_path)
+    shutil.rmtree(temp_path)
+
+
+@pytest.fixture
+def mock_config() -> Dict[str, Any]:
+    """Provide a mock configuration dictionary for testing."""
+    return {
+        "dataset": {
+            "name": "cifar10",
+            "data_dir": "/tmp/data",
+            "batch_size": 32,
+            "num_workers": 2,
+        },
+        "network": {
+            "name": "resnet18",
+            "pretrained": False,
+            "num_classes": 10,
+        },
+        "optimizer": {
+            "name": "sgd",
+            "lr": 0.1,
+            "momentum": 0.9,
+            "weight_decay": 5e-4,
+        },
+        "scheduler": {
+            "name": "cosine",
+            "num_epochs": 100,
+        },
+        "trainer": {
+            "name": "base",
+            "save_dir": "/tmp/checkpoints",
+            "save_epoch": 10,
+        },
+        "evaluator": {
+            "name": "ood",
+            "save_output": True,
+        },
+    }
+
+
+@pytest.fixture
+def sample_tensor() -> torch.Tensor:
+    """Provide a sample tensor for testing."""
+    return torch.randn(4, 3, 32, 32)
+
+
+@pytest.fixture
+def sample_labels() -> torch.Tensor:
+    """Provide sample labels for testing."""
+    return torch.randint(0, 10, (4,))
+
+
+@pytest.fixture
+def sample_features() -> torch.Tensor:
+    """Provide sample feature vectors for testing."""
+    return torch.randn(4, 512)
+
+
+@pytest.fixture
+def mock_model():
+    """Provide a mock neural network model."""
+    class MockModel(torch.nn.Module):
+        def __init__(self, num_classes=10):
+            super().__init__()
+            self.conv = torch.nn.Conv2d(3, 64, 3, padding=1)
+            self.pool = torch.nn.AdaptiveAvgPool2d(1)
+            self.fc = torch.nn.Linear(64, num_classes)
+
+        def forward(self, x):
+            x = self.conv(x)
+            x = torch.nn.functional.relu(x)
+            x = self.pool(x)
+            x = x.view(x.size(0), -1)
+            return self.fc(x)
+
+    return MockModel()
+
+
+@pytest.fixture
+def sample_dataset_config(temp_dir: Path) -> Dict[str, Any]:
+    """Create a sample dataset configuration."""
+    data_dir = temp_dir / "data"
+    data_dir.mkdir(exist_ok=True)
+
+    # Create mock image list files
+    train_list = data_dir / "train.txt"
+    val_list = data_dir / "val.txt"
+    test_list = data_dir / "test.txt"
+
+    # Write sample image paths
+    for file_path, num_samples in [(train_list, 100), (val_list, 20), (test_list, 20)]:
+        with open(file_path, 'w') as f:
+            for i in range(num_samples):
+                f.write(f"image_{i}.jpg {i % 10}\n")
+
+    return {
+        "data_dir": str(data_dir),
+        "train_list": str(train_list),
+        "val_list": str(val_list),
+        "test_list": str(test_list),
+        "num_classes": 10,
+    }
+
+
+@pytest.fixture
+def yaml_config_file(temp_dir: Path, mock_config: Dict[str, Any]) -> Path:
+    """Create a temporary YAML configuration file."""
+    config_path = temp_dir / "config.yml"
+    with open(config_path, 'w') as f:
+        yaml.dump(mock_config, f)
+    return config_path
+
+
+@pytest.fixture
+def json_config_file(temp_dir: Path, mock_config: Dict[str, Any]) -> Path:
+    """Create a temporary JSON configuration file."""
+    config_path = temp_dir / "config.json"
+    with open(config_path, 'w') as f:
+        json.dump(mock_config, f)
+    return config_path
+
+
+@pytest.fixture
+def mock_checkpoint(temp_dir: Path, mock_model) -> Path:
+    """Create a mock model checkpoint file."""
+    checkpoint_path = temp_dir / "checkpoint.pth"
+    torch.save({
+        'epoch': 50,
+        'model_state_dict': mock_model.state_dict(),
+        'optimizer_state_dict': {},
+        'loss': 0.123,
+    }, checkpoint_path)
+    return checkpoint_path
+
+
+@pytest.fixture
+def sample_metrics() -> Dict[str, float]:
+    """Provide sample evaluation metrics."""
+    return {
+        "accuracy": 0.95,
+        "auroc": 0.89,
+        "aupr_in": 0.91,
+        "aupr_out": 0.87,
+        "fpr95": 0.15,
+    }
+
+
+@pytest.fixture
+def mock_dataloader(sample_tensor: torch.Tensor, sample_labels: torch.Tensor):
+    """Create a mock dataloader for testing."""
+    from torch.utils.data import TensorDataset, DataLoader
+
+    dataset = TensorDataset(sample_tensor, sample_labels)
+    return DataLoader(dataset, batch_size=2, shuffle=False)
+
+
+@pytest.fixture
+def numpy_rng():
+    """Provide a seeded numpy random number generator."""
+    return np.random.RandomState(42)
+
+
+@pytest.fixture
+def torch_device():
+    """Provide the appropriate torch device (CPU/GPU)."""
+    return torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+
+@pytest.fixture(autouse=True)
+def reset_torch_seed():
+    """Reset torch random seed before each test."""
+    torch.manual_seed(42)
+    if torch.cuda.is_available():
+        torch.cuda.manual_seed_all(42)
+
+
+@pytest.fixture
+def sample_ood_scores() -> Dict[str, np.ndarray]:
+    """Provide sample OOD detection scores."""
+    num_id = 100
+    num_ood = 100
+
+    # Generate scores where ID samples have lower scores than OOD
+    id_scores = np.random.normal(0, 1, num_id)
+    ood_scores = np.random.normal(2, 1, num_ood)
+
+    return {
+        "id_scores": id_scores,
+        "ood_scores": ood_scores,
+        "id_labels": np.zeros(num_id),
+        "ood_labels": np.ones(num_ood),
+    }
+
+
+@pytest.fixture
+def mock_logger(temp_dir: Path):
+    """Create a mock logger for testing."""
+    import logging
+
+    logger = logging.getLogger("test_logger")
+    logger.setLevel(logging.DEBUG)
+
+    # Add file handler
+    log_file = temp_dir / "test.log"
+    handler = logging.FileHandler(log_file)
+    handler.setLevel(logging.DEBUG)
+
+    formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s')
+    handler.setFormatter(formatter)
+    logger.addHandler(handler)
+
+    return logger
+
+
+@pytest.fixture
+def sample_transform():
+    """Provide a sample image transformation pipeline."""
+    from torchvision import transforms
+
+    return transforms.Compose([
+        transforms.Resize((32, 32)),
+        transforms.ToTensor(),
+        transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
+    ])
+
+
+@pytest.fixture
+def mock_args():
+    """Provide mock command-line arguments."""
+    class MockArgs:
+        config = "configs/test_config.yml"
+        dataset = "cifar10"
+        network = "resnet18"
+        batch_size = 32
+        num_workers = 2
+        gpu = "0"
+        seed = 42
+        output_dir = "/tmp/output"
+        checkpoint = None
+        verbose = False
+
+    return MockArgs()