Skip to content

Commit

Permalink
SCHEMACODE: Add data loader class (#1506)
Browse files Browse the repository at this point in the history
* ENH: Add data loader to bst.data and bst.tests.data

* RF: Use data loaders in place of __file__

* DOC: Install pyparsing, mock pytest
  • Loading branch information
effigies committed Jun 25, 2023
1 parent bda5ffe commit 8ef9920
Show file tree
Hide file tree
Showing 7 changed files with 127 additions and 33 deletions.
99 changes: 99 additions & 0 deletions tools/schemacode/bidsschematools/data/__init__.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,99 @@
"""Module containing data files, including the schema source files
.. autofunction:: load_resource
"""
import atexit
import os
from contextlib import ExitStack
from functools import cached_property
from pathlib import Path
from types import ModuleType
from typing import Union

try:
from functools import cache
except ImportError: # PY38
from functools import lru_cache as cache

try: # Prefer backport to leave consistency to dependency spec
from importlib_resources import as_file, files
except ImportError:
from importlib.resources import as_file, files # type: ignore

__all__ = ["load_resource"]


class Loader:
"""A loader for package files relative to a module
This class wraps :mod:`importlib.resources` to provide a getter
function with an interpreter-lifetime scope. For typical packages
it simply passes through filesystem paths as :class:`~pathlib.Path`
objects. For zipped distributions, it will unpack the files into
a temporary directory that is cleaned up on interpreter exit.
This loader accepts a fully-qualified module name or a module
object.
Expected usage::
'''Data package
.. autofunction:: load_data
'''
from bidsschematools.data import Loader
load_data = Loader(__package__)
:class:`~Loader` objects implement the :func:`callable` interface
and generate a docstring, and are intended to be treated and documented
as functions.
"""

def __init__(self, anchor: Union[str, ModuleType]):
self._anchor = anchor
self.files = files(anchor)
self.exit_stack = ExitStack()
atexit.register(self.exit_stack.close)
# Allow class to have a different docstring from instances
self.__doc__ = self._doc

@cached_property
def _doc(self):
"""Construct docstring for instances
Lists the public top-level paths inside the location, where
non-public means has a `.` or `_` prefix or is a 'tests'
directory.
"""
top_level = sorted(
os.path.relpath(p, self.files) + "/"[: p.is_dir()]
for p in self.files.iterdir()
if p.name[0] not in (".", "_") and p.name != "tests"
)
doclines = [
f"Load package files relative to ``{self._anchor}``.",
"",
"This package contains the following (top-level) files/directories:",
"",
*(f"* ``{path}``" for path in top_level),
]

return "\n".join(doclines)

@cache
def __call__(self, *segments) -> Path:
"""Ensure data is available as a :class:`~pathlib.Path`.
Any temporary files that are created remain available throughout
the duration of the program, and are deleted when Python exits.
Results are cached so that multiple calls do not unpack the same
data multiple times, but the cache is sensitive to the specific
argument(s) passed.
"""
return self.exit_stack.enter_context(as_file(self.files.joinpath(*segments)))


load_resource = Loader(__package__)
9 changes: 9 additions & 0 deletions tools/schemacode/bidsschematools/tests/data/__init__.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,9 @@
"""Test data module
.. autofunction:: load_test_data
"""
from ...data import Loader

__all__ = ("load_test_data",)

load_test_data = Loader(__package__)
9 changes: 3 additions & 6 deletions tools/schemacode/bidsschematools/tests/test_schema.py
Original file line number Diff line number Diff line change
Expand Up @@ -6,15 +6,12 @@

from bidsschematools import __bids_version__, schema, types

from ..data import load_resource


def test__get_bids_version(tmp_path):
# Is the version being read in correctly?
schema_path = os.path.join(
os.path.abspath(os.path.dirname(__file__)),
os.pardir,
"data",
"schema",
)
schema_path = str(load_resource("schema"))
bids_version = schema._get_bids_version(schema_path)
assert bids_version == __bids_version__

Expand Down
34 changes: 10 additions & 24 deletions tools/schemacode/bidsschematools/tests/test_validator.py
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,9 @@
from bidsschematools.conftest import BIDS_ERROR_SELECTION, BIDS_SELECTION
from bidsschematools.validator import select_schema_path, validate_bids

from ..data import load_resource
from .data import load_test_data


def test_inheritance_examples():
correct_inheritance = [
Expand Down Expand Up @@ -68,20 +71,11 @@ def test_write_report(tmp_path):
"rawdata/sub-EXC022/anat/sub-EXC022_ses-MRI_flip-1_VFA.nii.gz"
]

report_path = os.path.join(
tmp_path,
"output_bids_validator_xs_write.log",
)
expected_report_path = os.path.join(
os.path.abspath(os.path.dirname(__file__)),
"data/expected_bids_validator_xs_write.log",
)
write_report(validation_result, report_path=report_path)
with open(report_path, "r") as f:
report_text = f.read()
with open(expected_report_path, "r") as f:
expected_report_text = f.read()
assert report_text == expected_report_text
report_path = tmp_path / "output_bids_validator_xs_write.log"
write_report(validation_result, report_path=str(report_path))

expected_report_path = load_test_data("expected_bids_validator_xs_write.log")
assert report_path.read_text() == expected_report_path.read_text()


@pytest.mark.skipif(
Expand Down Expand Up @@ -125,12 +119,7 @@ def test_validate_bids(bids_examples, tmp_path):
assert len(result["path_tracking"]) == 0

# Is the schema version recorded correctly?
schema_path = os.path.join(
os.path.abspath(os.path.dirname(__file__)),
os.pardir,
"data",
"schema",
)
schema_path = load_resource("schema")
with open(os.path.join(schema_path, "BIDS_VERSION")) as f:
expected_version = f.readline().rstrip()
assert result["bids_version"] == expected_version
Expand All @@ -148,10 +137,7 @@ def test_broken_json_dataset(bids_examples, tmp_path):
dataset_path = os.path.join(bids_examples, dataset)
dataset_json = os.path.join(dataset_path, "dataset_description.json")

broken_json = os.path.join(
os.path.abspath(os.path.dirname(__file__)),
"data/broken_dataset_description.json",
)
broken_json = load_test_data("broken_dataset_description.json")
shutil.copyfile(broken_json, dataset_json)

# No assert, will simply raise JSON reader error if not catching it properly.
Expand Down
5 changes: 3 additions & 2 deletions tools/schemacode/bidsschematools/utils.py
Original file line number Diff line number Diff line change
@@ -1,6 +1,7 @@
"""Utility functions for the bids-specification schema."""
import logging
import os.path as op

from . import data


def get_bundled_schema_path():
Expand All @@ -11,7 +12,7 @@ def get_bundled_schema_path():
str
Absolute path to the directory containing schema-related files.
"""
return op.abspath(op.join(op.dirname(__file__), "data", "schema"))
return str(data.load_resource("schema"))


def get_logger(name=None):
Expand Down
3 changes: 2 additions & 1 deletion tools/schemacode/docs/conf.py
Original file line number Diff line number Diff line change
Expand Up @@ -51,8 +51,9 @@
exclude_patterns = ["_build", "tests"]


# Mock internal modules to avoid building docs
autosummary_mock_imports = [
"pytest",
# Mock internal modules to avoid building docs
"bidsschematools.conftest",
]

Expand Down
1 change: 1 addition & 0 deletions tools/schemacode/docs/requirements.txt
Original file line number Diff line number Diff line change
@@ -1,3 +1,4 @@
sphinx
furo
myst_parser
pyparsing

0 comments on commit 8ef9920

Please sign in to comment.