tests: allow for parallel test execution

.gitignore

@@ -229,3 +229,4 @@ data/**
 
 test_data
 
+*.lock

config/coverage.toml

@@ -3,6 +3,7 @@ omit = [
   "src/imgtools/logging/**/*.py",
   "src/imgtools/cli/**/*.py",
   "src/imgtools/dicom/index/**/*.py",
+  "tests/**/*.py",
 ]
 
-[tool.coverage.report]
+[tool.coverage.report]

config/pytest.ini

@@ -14,7 +14,7 @@ addopts =
     --showlocals
     # Generate coverage report
     # Tracks code coverage during test execution
-    --cov=imgtools
+    --cov
     # Output coverage report in terminal
     # Provides immediate feedback on coverage
     --cov-report=term-missing
@@ -27,6 +27,17 @@ addopts =
     # Point to coverage config file
     # Allows customization of coverage report generation
     --cov-config=config/coverage.toml
+    # Append coverage data from previous runs
+    # Ensures coverage data is appended
+    --cov-append
+    # numprocessors to use for xdist plugin
+    # Sets number of processors to use for parallel test execution
+    --numprocesses=auto
+    # max processes
+    # Sets maximum number of processes to use for parallel test execution
+    --maxprocesses=8
+    # group xdist
+    --dist=loadgroup
 
 # Patterns for test discovery
 # Defines which files are considered test files

pixi.toml

@@ -49,6 +49,7 @@ pytest-cov = "*"
 pytest-xdist = "*"
 pytest-mock = ">=3.14.0,<4"
 sqlalchemy-stubs = ">=0.4,<0.5"
+filelock = ">=3.16.1,<4"
 
 [feature.test.pypi-dependencies]
 med-imagetools = { path = ".", editable = true }
@@ -64,6 +65,10 @@ inputs = ["coverage-report/coverage.xml", "config/coverage.toml"]
 depends-on = ["test"]
 description = "Run pytest and generate coverage report"
 
+[feature.test.tasks.clean_tests]
+cmd = "rm -rf .pytest_cache ./data ./tests/temp"
+description = "Clean up the test cache and data"
+
 ############################################## DOCS ###############################################
 [feature.docs.dependencies]
 mkdocs = "*"

src/imgtools/ops/__init__.py

@@ -1 +1,41 @@
-from .ops import *
+from .ops import (
+    BaseInput,
+    BaseOp,
+    BaseOutput,
+    CentreCrop,
+    ClipIntensity,
+    Crop,
+    HDF5Output,
+    ImageAutoInput,
+    ImageAutoOutput,
+    ImageStatistics,
+    MinMaxScale,
+    NumpyOutput,
+    Resample,
+    Resize,
+    StandardScale,
+    StructureSetToSegmentation,
+    WindowIntensity,
+    Zoom,
+)
+
+__all__ = [
+    "ImageAutoInput",
+    "ImageAutoOutput",
+    "StructureSetToSegmentation",
+    "BaseOp",
+    "BaseInput",
+    "BaseOutput",
+    "CentreCrop",
+    "ClipIntensity",
+    "Crop",
+    "HDF5Output",
+    "ImageStatistics",
+    "MinMaxScale",
+    "NumpyOutput",
+    "Resample",
+    "Resize",
+    "StandardScale",
+    "WindowIntensity",
+    "Zoom",
+]

tests/conftest.py

@@ -4,68 +4,67 @@
 from zipfile import ZipFile
 
 import pytest
+from filelock import FileLock
 
 from imgtools.logging import logger
 
 
-@pytest.fixture(scope='session')
+@pytest.fixture(scope="package")
 def curr_path():
-    return pathlib.Path(__file__).parent.parent.resolve().as_posix()
+    return pathlib.Path().cwd().resolve().absolute()
 
 
-@pytest.fixture(scope='session')
-def dataset_path(curr_path):
-    quebec_path = pathlib.Path(curr_path, 'data', 'Head-Neck-PET-CT')
+@pytest.fixture(scope="session")
+def prepare_dataset():
+    """Prepares the dataset if not already downloaded."""
+    curr_path = pathlib.Path().cwd().resolve().absolute()
+    quebec_path = pathlib.Path(curr_path, "data", "Head-Neck-PET-CT").absolute()
 
-    if not (quebec_path.exists() and len(list(quebec_path.glob('*'))) == 2):
-        quebec_path.mkdir(parents=True, exist_ok=True)
+    # when running xdist, use lockfile to prevent all processors from trying to download the dataset
+    lock_path = quebec_path / ".dataset.lock"
 
-        # Download QC dataset
-        logger.info('Downloading the test dataset...')
-        quebec_data_url = (
-            'https://github.com/bhklab/tcia_samples/blob/main/Head-Neck-PET-CT.zip?raw=true'
+    with FileLock(lock_path):
+        logger.info(
+            "Checking if the test dataset is downloaded...",
+            curr_path=curr_path,
+            quebec_path=quebec_path,
         )
-        quebec_zip_path = pathlib.Path(quebec_path, 'Head-Neck-PET-CT.zip').as_posix()
-        request.urlretrieve(quebec_data_url, quebec_zip_path)
-        with ZipFile(quebec_zip_path, 'r') as zipfile:
-            zipfile.extractall(quebec_path)
-        os.remove(quebec_zip_path)
-    else:
-        logger.info('Data already downloaded...')
+        if not (quebec_path.exists() and len(list(quebec_path.glob("*"))) == 2):
+            quebec_path.mkdir(parents=True, exist_ok=True)
 
-    output_path = pathlib.Path(curr_path, 'tests', 'temp').as_posix()
-    quebec_path = quebec_path.as_posix()
+            # Download QC dataset
+            logger.info("Downloading the test dataset...")
+            quebec_data_url = "https://github.com/bhklab/tcia_samples/blob/main/Head-Neck-PET-CT.zip?raw=true"
+            quebec_zip_path = pathlib.Path(
+                quebec_path, "Head-Neck-PET-CT.zip"
+            ).as_posix()
+            request.urlretrieve(quebec_data_url, quebec_zip_path)
+            with ZipFile(quebec_zip_path, "r") as zipfile:
+                zipfile.extractall(quebec_path)
+            os.remove(quebec_zip_path)
+        else:
+            logger.info("Data already downloaded...")
 
-    # Dataset name
-    dataset_name = os.path.basename(quebec_path)
-    imgtools_path = pathlib.Path(os.path.dirname(quebec_path), '.imgtools')
-
-    # Defining paths for autopipeline and dataset component
-    crawl_path = pathlib.Path(imgtools_path, f'imgtools_{dataset_name}.csv').as_posix()
-    edge_path = pathlib.Path(imgtools_path, f'imgtools_{dataset_name}_edges.csv').as_posix()
-    # json_path =  pathlib.Path(imgtools_path, f"imgtools_{dataset_name}.json").as_posix()  # noqa: F841
+    yield quebec_path
 
-    yield quebec_path, output_path, crawl_path, edge_path
+    # Delete the lock file
+    if lock_path.exists():
+        lock_path.unlink()
 
 
-@pytest.fixture(scope='session')
-def modalities_path(curr_path):
-    qc_path = pathlib.Path(curr_path, 'data', 'Head-Neck-PET-CT', 'HN-CHUS-052')
-    assert qc_path.exists(), 'Dataset not found'
+@pytest.fixture(scope="package")
+def dataset_path(prepare_dataset):
+    """Provides paths related to the dataset for tests."""
+    curr_path = pathlib.Path().cwd().resolve().absolute()
+    output_path = pathlib.Path(curr_path, "tests", "temp").as_posix()
 
-    path = {}
-    path['CT'] = pathlib.Path(
-        qc_path, '08-27-1885-CA ORL FDG TEP POS TX-94629/3.000000-Merged-06362'
-    ).as_posix()
-    path['RTSTRUCT'] = pathlib.Path(
-        qc_path,
-        '08-27-1885-OrophCB.0OrophCBTRTID derived StudyInstanceUID.-94629/Pinnacle POI-41418',
-    ).as_posix()
-    path['RTDOSE'] = pathlib.Path(
-        qc_path,
-        '08-27-1885-OrophCB.0OrophCBTRTID derived StudyInstanceUID.-94629/11376',
-    ).as_posix()
-    path['PT'] = pathlib.Path(
-        qc_path, '08-27-1885-CA ORL FDG TEP POS TX-94629/532790.000000-LOR-RAMLA-44600'
+    # Paths
+    quebec_path = prepare_dataset.as_posix()
+    dataset_name = os.path.basename(quebec_path)
+    imgtools_path = pathlib.Path(os.path.dirname(quebec_path), ".imgtools")
+    crawl_path = pathlib.Path(imgtools_path, f"imgtools_{dataset_name}.csv").as_posix()
+    edge_path = pathlib.Path(
+        imgtools_path, f"imgtools_{dataset_name}_edges.csv"
     ).as_posix()
-    return path
+
+    yield quebec_path, output_path, crawl_path, edge_path