From 3741c3668cce90a1e7f9a989d8ca542776a321dc Mon Sep 17 00:00:00 2001
From: Thomas Kluyver <thomas@kluyver.me.uk>
Date: Wed, 22 Nov 2023 08:32:27 +0000
Subject: [PATCH] Remove old code
---
.gitlab-ci.yml | 2 +-
tests/test_reference_runs/conftest.py | 14 +-
.../test_pre_deployment.py | 234 +-----------------
3 files changed, 3 insertions(+), 247 deletions(-)
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index 01a2d7ac0..5fe082129 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -68,7 +68,7 @@ automated_test:
- python3 -m pip install ".[automated_test]"
- echo "Running automated test. This can take sometime to finish depending on the test data."
- echo "Given variables are REFERENCE=$REFERENCE, OUTPUT=$OUTPUT, DETECTORS=$DETECTORS, CALIBRATION=$CALIBRATION"
- - python3 -m pytest ./tests/test_reference_runs --color yes --verbose --release-test --reference-folder /gpfs/exfel/data/scratch/xcaltst/test/$REFERENCE --out-folder /gpfs/exfel/data/scratch/xcaltst/test/$OUTPUT --detectors $DETECTORS --calibration $CALIBRATION --find-difference
+ - python3 -m pytest ./tests/test_reference_runs --color yes --verbose --release-test --reference-folder /gpfs/exfel/data/scratch/xcaltst/test/$REFERENCE --out-folder /gpfs/exfel/data/scratch/xcaltst/test/$OUTPUT --detectors $DETECTORS --calibration $CALIBRATION
timeout: 24 hours
cython-editable-install-test:
diff --git a/tests/test_reference_runs/conftest.py b/tests/test_reference_runs/conftest.py
index ce3b97458..c2c50991e 100644
--- a/tests/test_reference_runs/conftest.py
+++ b/tests/test_reference_runs/conftest.py
@@ -38,15 +38,6 @@ def pytest_addoption(parser):
"apply validation test on numerical data only."),
)
- parser.addoption(
- "--find-difference",
- action="store_true",
- default=False,
- help=(
- "In case of non numerical validation of h5file. "
- "Find the different attribute and fail on the first one."),
- )
-
parser.addoption(
"--use-slurm",
action="store_true",
@@ -80,8 +71,6 @@ def release_test_config(request):
"--no-numerical-validation")
validate_only = request.config.getoption(
"--validation-only")
- find_difference = request.config.getoption(
- "--find-difference")
use_slurm = request.config.getoption(
"--use-slurm")
picked_test = request.config.getoption("--picked-test")
@@ -90,8 +79,7 @@ def release_test_config(request):
return (
detectors, calibration, picked_test,
skip_numerical_validation, validate_only,
- find_difference, use_slurm, reference_folder,
- out_folder,
+ use_slurm, reference_folder, out_folder,
)
diff --git a/tests/test_reference_runs/test_pre_deployment.py b/tests/test_reference_runs/test_pre_deployment.py
index b1f99ce02..30ec362b2 100644
--- a/tests/test_reference_runs/test_pre_deployment.py
+++ b/tests/test_reference_runs/test_pre_deployment.py
@@ -23,236 +23,6 @@ from .callab_tests import automated_test_config
LOGGER = logging.getLogger(__name__)
-def file_md5(
- tested_file: str,
- block_size: int = 2 ** 20,
-) -> bytes:
- """Generating MD5 checksum for a file.
-
- Args:
- tested_file: File to be tested.
- block_size (_type_, optional): Block size for reading the file.
- Defaults to 2**20.
- """
- f = open(tested_file, "rb")
- md5 = hashlib.md5()
- while True:
- data = f.read(block_size)
- if not data:
- break
- md5.update(data)
- f.close()
- return md5.digest()
-
-
-def collect_attrs(groups, datasets, objects, exclude_attrs, name, node):
- """Collect h5 attrs in groups, datasets, and objects lists."""
- if node.name not in exclude_attrs:
- if isinstance(node, h5py.Group):
- groups.append(name)
- elif isinstance(node, h5py.Dataset):
- if node.dtype == 'object':
- objects.append(name)
- else:
- datasets.append(name)
-
-
-def compare_datasets(
- file1,
- file2,
- datasets: list
-):
- """Compare the values of datasets in two h5 files."""
- h5_diff = []
- for d in datasets:
- try:
- if not np.allclose(file1[d][()], file2[d][()], equal_nan=True):
- h5_diff.append(d)
- except ValueError as e:
- LOGGER.error(f"ValueError: {e}, {d}")
- h5_diff.append(d)
- except AttributeError as e:
- LOGGER.error(f"AttributeError: {e}, {d}")
- h5_diff.append(d)
- return h5_diff
-
-
-def compare_objects(
- file1,
- file2,
- objects: list
-):
- """Compare the objects in two h5 files."""
- h5_diff = []
- for d in objects:
- try:
- if isinstance(file1[d][()], bytes):
- if (
- file1[d][()].decode('utf-8') != file2[d][()].decode('utf-8') # noqa
- ):
- h5_diff.append(d)
- elif (
- file1[d][()].dtype != file1[d][()].dtype and
- not file1[d][()] != file2[d][()]
- ): # pnccd files has only list of bytes
- h5_diff.append(d)
- except ValueError as e:
- LOGGER.error(f"ValueError: {e}, {d}")
- h5_diff.append(d)
- except AttributeError as e:
- LOGGER.error(f"AttributeError: {e}, {d}, "
- f"{file1[d][()].decode('utf-8')}")
- h5_diff.append(d)
- return h5_diff
-
-
-def find_differences(
- test_file,
- reference_file,
- exclude_attrs,
-):
- """
- Find difference in groups, datasets, and objects between two h5files.
- Args:
- file1: first h5 file.
- file2: second h5 file.
- """
-
- groups_f1 = []
- datasets_f1 = []
- objects_f1 = []
-
- groups_f2 = []
- datasets_f2 = []
- objects_f2 = []
-
- with h5py.File(test_file, 'r') as file1, h5py.File(reference_file, 'r') as file2: # noqa
-
- # Fill groups, datasets, and objects list
- # to compare both h5files' attrs.
- file1.visititems(
- partial(
- collect_attrs,
- groups_f1,
- datasets_f1,
- objects_f1,
- exclude_attrs,
- ))
- file2.visititems(
- partial(
- collect_attrs,
- groups_f2,
- datasets_f2,
- objects_f2,
- exclude_attrs,
- ))
-
- start_time = time.perf_counter()
- # Compare groups, datasets, and objects to have the same content.
- assert set(groups_f1) == set(groups_f2), f"{test_file} and {reference_file} consists of different groups." # noqa
- assert set(datasets_f1) == set(datasets_f2), f"{test_file} and {reference_file} consists of different datasets." # noqa
- assert set(objects_f1) == set(objects_f2), f"{test_file} and {reference_file} consists of different datasets." # noqa
- duration = time.perf_counter() - start_time
- LOGGER.debug("Elapsed time comparing groups, "
- f"datasets, and objects: {duration} seconds")
- LOGGER.debug("Groups, datasets, and objects have the same content.")
-
- # Compare datasets and objects.
- start_time = time.perf_counter()
- h5_diff_datasets = compare_datasets(file1, file2, datasets_f1)
- duration = time.perf_counter() - start_time
- LOGGER.debug(f"Elapsed time comparing datasets: {duration} seconds")
- start_time = time.perf_counter()
- h5_diff_objects = compare_objects(file1, file2, objects_f1)
- LOGGER.debug(f"Elapsed time comparing objects: {duration} seconds")
-
- assert not h5_diff_datasets, f"{[d for d in h5_diff_datasets]} datasets contain different values for {test_file} and {reference_file}" # noqa
- LOGGER.debug("Datasets are validated.")
- assert not h5_diff_objects, f"{[d for d in h5_diff_objects]} objects contain different values for {test_file} and {reference_file}" # noqa
- LOGGER.debug("Objects are validated.")
-
-
-def validate_files(
- ref_folder: pathlib.PosixPath,
- out_folder: pathlib.PosixPath,
- exclude_attrs: list,
- test_file: pathlib.PosixPath,
-) -> Tuple[bool, pathlib.PosixPath]:
- """Validate file similarities. Create temporary files to exclude
- h5 attributes known to be different. e.g `report` for constants.
- If both files are not identical, the function is able to loop over
- both files and find and fail on the difference.
-
- Args:
- ref_folder: The reference folder for validating the files
- out_folder: The output folder for the test constant files.
- test_file: The output file to be validated.
- exclude_attrs: A list of datasets, groups to exclude
- from validated files.
- Returns:
- result: validation result for metadata.
- test_file: The validated file.
- """
- import h5py
- start_validating = time.perf_counter()
-
- def exclude_sources(source_file, dest, excluded_sources):
- # Open the source file in read-only mode
- with h5py.File(source_file, 'r') as source:
-
- # Recursively visit all objects in the source file
- def visit_func(name, obj):
- # Check if the object should be excluded
- if name in excluded_sources:
- return
-
- # Check if the object is a dataset
- if isinstance(obj, h5py.Dataset):
- # Create a new dataset in the destination
- # file and copy the data
- dest.create_dataset(name, data=obj[()])
-
- # Visit all objects in the source file and
- # copy them to the destination file
- source.visititems(visit_func)
-
- with tempfile.NamedTemporaryFile(
- dir=out_folder,
- suffix=".tmp",
- prefix="cal_",
- delete=True,
- ) as out_tf, tempfile.NamedTemporaryFile(
- dir=out_folder,
- suffix=".tmp",
- prefix="cal_",
- delete=True,
- ) as ref_tf:
-
- # Create in-memory HDF5 files for validation
- with h5py.File(out_tf.name, 'a') as hp1, h5py.File(ref_tf.name, 'a') as hp2: # noqa
-
- start_time = time.perf_counter()
- # Copy h5 files for validation and exclude selected attrs.
- exclude_sources(test_file, hp1, exclude_attrs)
-
- duration = time.perf_counter() - start_time
- LOGGER.debug(f"Elapsed time copying {test_file}: "
- f"{duration} seconds")
-
- start_time = time.perf_counter()
- exclude_sources(ref_folder / test_file.name, hp2, exclude_attrs)
-
- duration = time.perf_counter() - start_time
- LOGGER.debug(f"Elapsed time copying {ref_folder / test_file.name}: "
- f"{duration} seconds")
-
- start_time = time.perf_counter()
- result = file_md5(out_tf.name) == file_md5(ref_tf.name)
- LOGGER.debug(f"MD5 validation for {test_file}: {duration} seconds")
- duration = time.perf_counter() - start_validating
- return result, test_file
-
-
@dataclass
class ComparisonResult:
filename: str
@@ -475,7 +245,7 @@ def test_xfel_calibrate(
(
detectors, calibration, picked_test,
- skip_numerical_validation, only_validate, find_difference,
+ skip_numerical_validation, only_validate,
use_slurm, reference_dir_base, out_dir_base,
) = release_test_config
@@ -520,7 +290,6 @@ def test_xfel_calibrate(
out_folder,
reference_folder,
cal_type,
- find_difference,
)
return
@@ -558,5 +327,4 @@ def test_xfel_calibrate(
out_folder,
reference_folder,
cal_type,
- find_difference,
)
--
GitLab