tests.py

Go to the documentation of this file.

#
# LSST Data Management System
#
# Copyright 2008-2017  AURA/LSST.
#
# This product includes software developed by the
# LSST Project (http://www.lsst.org/).
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the LSST License Statement and
# the GNU General Public License along with this program.  If not,
# see <https://www.lsstcorp.org/LegalNotices/>.
#
"""Support code for running unit tests"""
 
import contextlib
import gc
import inspect
import os
import subprocess
import sys
import unittest
import warnings
import numpy
import functools
import tempfile
 
__all__ = ["init", "MemoryTestCase", "ExecutablesTestCase", "getTempFilePath",
           "TestCase", "assertFloatsAlmostEqual", "assertFloatsNotEqual", "assertFloatsEqual",
           "debugger", "classParameters", "methodParameters"]
 
# File descriptor leak test will be skipped if psutil can not be imported
try:
    import psutil
except ImportError:
    psutil = None
 
# Initialize the list of open files to an empty set
open_files = set()
 
 
def _get_open_files():
    """Return a set containing the list of files currently open in this
    process.
 
    Returns
    -------
    open_files : `set`
        Set containing the list of open files.
    """
    if psutil is None:
        return set()
    return set(p.path for p in psutil.Process().open_files())
 
 
def init():
    """Initialize the memory tester and file descriptor leak tester."""
    global open_files
    # Reset the list of open files
    open_files = _get_open_files()
 
 
def sort_tests(tests):
    """Sort supplied test suites such that MemoryTestCases are at the end.
 
    `lsst.utils.tests.MemoryTestCase` tests should always run after any other
    tests in the module.
 
    Parameters
    ----------
    tests : sequence
        Sequence of test suites.
 
    Returns
    -------
    suite : `unittest.TestSuite`
        A combined `~unittest.TestSuite` with
        `~lsst.utils.tests.MemoryTestCase` at the end.
    """
 
    suite = unittest.TestSuite()
    memtests = []
    for test_suite in tests:
        try:
            # Just test the first test method in the suite for MemoryTestCase
            # Use loop rather than next as it is possible for a test class
            # to not have any test methods and the Python community prefers
            # for loops over catching a StopIteration exception.
            bases = None
            for method in test_suite:
                bases = inspect.getmro(method.__class__)
                break
            if bases is not None and MemoryTestCase in bases:
                memtests.append(test_suite)
            else:
                suite.addTests(test_suite)
        except TypeError:
            if isinstance(test_suite, MemoryTestCase):
                memtests.append(test_suite)
            else:
                suite.addTest(test_suite)
    suite.addTests(memtests)
    return suite
 
 
def suiteClassWrapper(tests):
    return unittest.TestSuite(sort_tests(tests))
 
 
# Replace the suiteClass callable in the defaultTestLoader
# so that we can reorder the test ordering. This will have
# no effect if no memory test cases are found.
unittest.defaultTestLoader.suiteClass = suiteClassWrapper
 
 
class MemoryTestCase(unittest.TestCase):
    """Check for resource leaks."""
 
    @classmethod
    def tearDownClass(cls):
        """Reset the leak counter when the tests have been completed"""
        init()
 
    def testFileDescriptorLeaks(self):
        """Check if any file descriptors are open since init() called."""
        if psutil is None:
            self.skipTest("Unable to test file descriptor leaks. psutil unavailable.")
        gc.collect()
        global open_files
        now_open = _get_open_files()
 
        # Some files are opened out of the control of the stack.
        now_open = set(f for f in now_open if not f.endswith(".car") and
                       not f.startswith("/proc/") and
                       not f.endswith(".ttf") and
                       not (f.startswith("/var/lib/") and f.endswith("/passwd")) and
                       not f.endswith("astropy.log"))
 
        diff = now_open.difference(open_files)
        if diff:
            for f in diff:
                print("File open: %s" % f)
            self.fail("Failed to close %d file%s" % (len(diff), "s" if len(diff) != 1 else ""))
 
 
class ExecutablesTestCase(unittest.TestCase):
    """Test that executables can be run and return good status.
 
    The test methods are dynamically created. Callers
    must subclass this class in their own test file and invoke
    the create_executable_tests() class method to register the tests.
    """
    TESTS_DISCOVERED = -1
 
    @classmethod
    def setUpClass(cls):
        """Abort testing if automated test creation was enabled and
        no tests were found."""
 
        if cls.TESTS_DISCOVERED == 0:
            raise Exception("No executables discovered.")
 
    def testSanity(self):
        """This test exists to ensure that there is at least one test to be
        executed. This allows the test runner to trigger the class set up
        machinery to test whether there are some executables to test."""
        pass
 
    def assertExecutable(self, executable, root_dir=None, args=None, msg=None):
        """Check an executable runs and returns good status.
 
        Prints output to standard out. On bad exit status the test
        fails. If the executable can not be located the test is skipped.
 
        Parameters
        ----------
        executable : `str`
            Path to an executable. ``root_dir`` is not used if this is an
            absolute path.
        root_dir : `str`, optional
            Directory containing executable. Ignored if `None`.
        args : `list` or `tuple`, optional
            Arguments to be provided to the executable.
        msg : `str`, optional
            Message to use when the test fails. Can be `None` for default
            message.
 
        Raises
        ------
        AssertionError
            The executable did not return 0 exit status.
        """
 
        if root_dir is not None and not os.path.isabs(executable):
            executable = os.path.join(root_dir, executable)
 
        # Form the argument list for subprocess
        sp_args = [executable]
        argstr = "no arguments"
        if args is not None:
            sp_args.extend(args)
            argstr = 'arguments "' + " ".join(args) + '"'
 
        print("Running executable '{}' with {}...".format(executable, argstr))
        if not os.path.exists(executable):
            self.skipTest("Executable {} is unexpectedly missing".format(executable))
        failmsg = None
        try:
            output = subprocess.check_output(sp_args)
        except subprocess.CalledProcessError as e:
            output = e.output
            failmsg = "Bad exit status from '{}': {}".format(executable, e.returncode)
        print(output.decode('utf-8'))
        if failmsg:
            if msg is None:
                msg = failmsg
            self.fail(msg)
 
    @classmethod
    def _build_test_method(cls, executable, root_dir):
        """Build a test method and attach to class.
 
        A test method is created for the supplied excutable located
        in the supplied root directory. This method is attached to the class
        so that the test runner will discover the test and run it.
 
        Parameters
        ----------
        cls : `object`
            The class in which to create the tests.
        executable : `str`
            Name of executable. Can be absolute path.
        root_dir : `str`
            Path to executable. Not used if executable path is absolute.
        """
        if not os.path.isabs(executable):
            executable = os.path.abspath(os.path.join(root_dir, executable))
 
        # Create the test name from the executable path.
        test_name = "test_exe_" + executable.replace("/", "_")
 
        # This is the function that will become the test method
        def test_executable_runs(*args):
            self = args[0]
            self.assertExecutable(executable)
 
        # Give it a name and attach it to the class
        test_executable_runs.__name__ = test_name
        setattr(cls, test_name, test_executable_runs)
 
    @classmethod
    def create_executable_tests(cls, ref_file, executables=None):
        """Discover executables to test and create corresponding test methods.
 
        Scans the directory containing the supplied reference file
        (usually ``__file__`` supplied from the test class) to look for
        executables. If executables are found a test method is created
        for each one. That test method will run the executable and
        check the returned value.
 
        Executable scripts with a ``.py`` extension and shared libraries
        are ignored by the scanner.
 
        This class method must be called before test discovery.
 
        Parameters
        ----------
        ref_file : `str`
            Path to a file within the directory to be searched.
            If the files are in the same location as the test file, then
            ``__file__`` can be used.
        executables : `list` or `tuple`, optional
            Sequence of executables that can override the automated
            detection. If an executable mentioned here is not found, a
            skipped test will be created for it, rather than a failed
            test.
 
        Examples
        --------
        >>> cls.create_executable_tests(__file__)
        """
 
        # Get the search directory from the reference file
        ref_dir = os.path.abspath(os.path.dirname(ref_file))
 
        if executables is None:
            # Look for executables to test by walking the tree
            executables = []
            for root, dirs, files in os.walk(ref_dir):
                for f in files:
                    # Skip Python files. Shared libraries are executable.
                    if not f.endswith(".py") and not f.endswith(".so"):
                        full_path = os.path.join(root, f)
                        if os.access(full_path, os.X_OK):
                            executables.append(full_path)
 
        # Store the number of tests found for later assessment.
        # Do not raise an exception if we have no executables as this would
        # cause the testing to abort before the test runner could properly
        # integrate it into the failure report.
        cls.TESTS_DISCOVERED = len(executables)
 
        # Create the test functions and attach them to the class
        for e in executables:
            cls._build_test_method(e, ref_dir)
 
 
@contextlib.contextmanager
def getTempFilePath(ext, expectOutput=True):
    """Return a path suitable for a temporary file and try to delete the
    file on success
 
    If the with block completes successfully then the file is deleted,
    if possible; failure results in a printed warning.
    If a file is remains when it should not, a RuntimeError exception is
    raised. This exception is also raised if a file is not present on context
    manager exit when one is expected to exist.
    If the block exits with an exception the file if left on disk so it can be
    examined. The file name has a random component such that nested context
    managers can be used with the same file suffix.
 
    Parameters
    ----------
 
    ext : `str`
        File name extension, e.g. ``.fits``.
    expectOutput : `bool`, optional
        If `True`, a file should be created within the context manager.
        If `False`, a file should not be present when the context manager
        exits.
 
    Returns
    -------
    `str`
        Path for a temporary file. The path is a combination of the caller's
        file path and the name of the top-level function
 
    Notes
    -----
    ::
 
        # file tests/testFoo.py
        import unittest
        import lsst.utils.tests
        class FooTestCase(unittest.TestCase):
            def testBasics(self):
                self.runTest()
 
            def runTest(self):
                with lsst.utils.tests.getTempFilePath(".fits") as tmpFile:
                    # if tests/.tests exists then
                    # tmpFile = "tests/.tests/testFoo_testBasics.fits"
                    # otherwise tmpFile = "testFoo_testBasics.fits"
                    ...
                    # at the end of this "with" block the path tmpFile will be
                    # deleted, but only if the file exists and the "with"
                    # block terminated normally (rather than with an exception)
        ...
    """
    stack = inspect.stack()
    # get name of first function in the file
    for i in range(2, len(stack)):
        frameInfo = inspect.getframeinfo(stack[i][0])
        if i == 2:
            callerFilePath = frameInfo.filename
            callerFuncName = frameInfo.function
        elif callerFilePath == frameInfo.filename:
            # this function called the previous function
            callerFuncName = frameInfo.function
        else:
            break
 
    callerDir, callerFileNameWithExt = os.path.split(callerFilePath)
    callerFileName = os.path.splitext(callerFileNameWithExt)[0]
    outDir = os.path.join(callerDir, ".tests")
    if not os.path.isdir(outDir):
        outDir = ""
    prefix = "%s_%s-" % (callerFileName, callerFuncName)
    outPath = tempfile.mktemp(dir=outDir, suffix=ext, prefix=prefix)
    if os.path.exists(outPath):
        # There should not be a file there given the randomizer. Warn and remove.
        # Use stacklevel 3 so that the warning is reported from the end of the with block
        warnings.warn("Unexpectedly found pre-existing tempfile named %r" % (outPath,),
                      stacklevel=3)
        try:
            os.remove(outPath)
        except OSError:
            pass
 
    yield outPath
 
    fileExists = os.path.exists(outPath)
    if expectOutput:
        if not fileExists:
            raise RuntimeError("Temp file expected named {} but none found".format(outPath))
    else:
        if fileExists:
            raise RuntimeError("Unexpectedly discovered temp file named {}".format(outPath))
    # Try to clean up the file regardless
    if fileExists:
        try:
            os.remove(outPath)
        except OSError as e:
            # Use stacklevel 3 so that the warning is reported from the end of the with block
            warnings.warn("Warning: could not remove file %r: %s" % (outPath, e), stacklevel=3)
 
 
class TestCase(unittest.TestCase):
    """Subclass of unittest.TestCase that adds some custom assertions for
    convenience.
    """
 
 
def inTestCase(func):
    """A decorator to add a free function to our custom TestCase class, while also
    making it available as a free function.
    """
    setattr(TestCase, func.__name__, func)
    return func
 
 
def debugger(*exceptions):
    """Decorator to enter the debugger when there's an uncaught exception
 
    To use, just slap a ``@debugger()`` on your function.
 
    You may provide specific exception classes to catch as arguments to
    the decorator function, e.g.,
    ``@debugger(RuntimeError, NotImplementedError)``.
    This defaults to just `AssertionError`, for use on `unittest.TestCase`
    methods.
 
    Code provided by "Rosh Oxymoron" on StackOverflow:
    http://stackoverflow.com/questions/4398967/python-unit-testing-automatically-running-the-debugger-when-a-test-fails
 
    Notes
    -----
    Consider using ``pytest --pdb`` instead of this decorator.
    """
    if not exceptions:
        exceptions = (Exception, )
 
    def decorator(f):
        @functools.wraps(f)
        def wrapper(*args, **kwargs):
            try:
                return f(*args, **kwargs)
            except exceptions:
                import sys
                import pdb
                pdb.post_mortem(sys.exc_info()[2])
        return wrapper
    return decorator
 
 
def plotImageDiff(lhs, rhs, bad=None, diff=None, plotFileName=None):
    """Plot the comparison of two 2-d NumPy arrays.
 
    Parameters
    ----------
    lhs : `numpy.ndarray`
        LHS values to compare; a 2-d NumPy array
    rhs : `numpy.ndarray`
        RHS values to compare; a 2-d NumPy array
    bad : `numpy.ndarray`
        A 2-d boolean NumPy array of values to emphasize in the plots
    diff : `numpy.ndarray`
        difference array; a 2-d NumPy array, or None to show lhs-rhs
    plotFileName : `str`
        Filename to save the plot to.  If None, the plot will be displayed in
        a window.
 
    Notes
    -----
    This method uses `matplotlib` and imports it internally; it should be
    wrapped in a try/except block within packages that do not depend on
    `matplotlib` (including `~lsst.utils`).
    """
    from matplotlib import pyplot
    if diff is None:
        diff = lhs - rhs
    pyplot.figure()
    if bad is not None:
        # make an rgba image that's red and transparent where not bad
        badImage = numpy.zeros(bad.shape + (4,), dtype=numpy.uint8)
        badImage[:, :, 0] = 255
        badImage[:, :, 1] = 0
        badImage[:, :, 2] = 0
        badImage[:, :, 3] = 255*bad
    vmin1 = numpy.minimum(numpy.min(lhs), numpy.min(rhs))
    vmax1 = numpy.maximum(numpy.max(lhs), numpy.max(rhs))
    vmin2 = numpy.min(diff)
    vmax2 = numpy.max(diff)
    for n, (image, title) in enumerate([(lhs, "lhs"), (rhs, "rhs"), (diff, "diff")]):
        pyplot.subplot(2, 3, n + 1)
        im1 = pyplot.imshow(image, cmap=pyplot.cm.gray, interpolation='nearest', origin='lower',
                            vmin=vmin1, vmax=vmax1)
        if bad is not None:
            pyplot.imshow(badImage, alpha=0.2, interpolation='nearest', origin='lower')
        pyplot.axis("off")
        pyplot.title(title)
        pyplot.subplot(2, 3, n + 4)
        im2 = pyplot.imshow(image, cmap=pyplot.cm.gray, interpolation='nearest', origin='lower',
                            vmin=vmin2, vmax=vmax2)
        if bad is not None:
            pyplot.imshow(badImage, alpha=0.2, interpolation='nearest', origin='lower')
        pyplot.axis("off")
        pyplot.title(title)
    pyplot.subplots_adjust(left=0.05, bottom=0.05, top=0.92, right=0.75, wspace=0.05, hspace=0.05)
    cax1 = pyplot.axes([0.8, 0.55, 0.05, 0.4])
    pyplot.colorbar(im1, cax=cax1)
    cax2 = pyplot.axes([0.8, 0.05, 0.05, 0.4])
    pyplot.colorbar(im2, cax=cax2)
    if plotFileName:
        pyplot.savefig(plotFileName)
    else:
        pyplot.show()
 
 
@inTestCase
def assertFloatsAlmostEqual(testCase, lhs, rhs, rtol=sys.float_info.epsilon,
                            atol=sys.float_info.epsilon, relTo=None,
                            printFailures=True, plotOnFailure=False,
                            plotFileName=None, invert=False, msg=None):
    """Highly-configurable floating point comparisons for scalars and arrays.
 
    The test assertion will fail if all elements ``lhs`` and ``rhs`` are not
    equal to within the tolerances specified by ``rtol`` and ``atol``.
    More precisely, the comparison is:
 
    ``abs(lhs - rhs) <= relTo*rtol OR abs(lhs - rhs) <= atol``
 
    If ``rtol`` or ``atol`` is `None`, that term in the comparison is not
    performed at all.
 
    When not specified, ``relTo`` is the elementwise maximum of the absolute
    values of ``lhs`` and ``rhs``.  If set manually, it should usually be set
    to either ``lhs`` or ``rhs``, or a scalar value typical of what is
    expected.
 
    Parameters
    ----------
    testCase : `unittest.TestCase`
        Instance the test is part of.
    lhs : scalar or array-like
        LHS value(s) to compare; may be a scalar or array-like of any
        dimension.
    rhs : scalar or array-like
        RHS value(s) to compare; may be a scalar or array-like of any
        dimension.
    rtol : `float`, optional
        Relative tolerance for comparison; defaults to double-precision
        epsilon.
    atol : `float`, optional
        Absolute tolerance for comparison; defaults to double-precision
        epsilon.
    relTo : `float`, optional
        Value to which comparison with rtol is relative.
    printFailures : `bool`, optional
        Upon failure, print all inequal elements as part of the message.
    plotOnFailure : `bool`, optional
        Upon failure, plot the originals and their residual with matplotlib.
        Only 2-d arrays are supported.
    plotFileName : `str`, optional
        Filename to save the plot to.  If `None`, the plot will be displayed in
        a window.
    invert : `bool`, optional
        If `True`, invert the comparison and fail only if any elements *are*
        equal. Used to implement `~lsst.utils.tests.assertFloatsNotEqual`,
        which should generally be used instead for clarity.
    msg : `str`, optional
        String to append to the error message when assert fails.
 
    Raises
    ------
    AssertionError
        The values are not almost equal.
    """
    if not numpy.isfinite(lhs).all():
        testCase.fail("Non-finite values in lhs")
    if not numpy.isfinite(rhs).all():
        testCase.fail("Non-finite values in rhs")
    diff = lhs - rhs
    absDiff = numpy.abs(lhs - rhs)
    if rtol is not None:
        if relTo is None:
            relTo = numpy.maximum(numpy.abs(lhs), numpy.abs(rhs))
        else:
            relTo = numpy.abs(relTo)
        bad = absDiff > rtol*relTo
        if atol is not None:
            bad = numpy.logical_and(bad, absDiff > atol)
    else:
        if atol is None:
            raise ValueError("rtol and atol cannot both be None")
        bad = absDiff > atol
    failed = numpy.any(bad)
    if invert:
        failed = not failed
        bad = numpy.logical_not(bad)
        cmpStr = "=="
        failStr = "are the same"
    else:
        cmpStr = "!="
        failStr = "differ"
    errMsg = []
    if failed:
        if numpy.isscalar(bad):
            if rtol is None:
                errMsg = ["%s %s %s; diff=%s with atol=%s"
                          % (lhs, cmpStr, rhs, absDiff, atol)]
            elif atol is None:
                errMsg = ["%s %s %s; diff=%s/%s=%s with rtol=%s"
                          % (lhs, cmpStr, rhs, absDiff, relTo, absDiff/relTo, rtol)]
            else:
                errMsg = ["%s %s %s; diff=%s/%s=%s with rtol=%s, atol=%s"
                          % (lhs, cmpStr, rhs, absDiff, relTo, absDiff/relTo, rtol, atol)]
        else:
            errMsg = ["%d/%d elements %s with rtol=%s, atol=%s"
                      % (bad.sum(), bad.size, failStr, rtol, atol)]
            if plotOnFailure:
                if len(lhs.shape) != 2 or len(rhs.shape) != 2:
                    raise ValueError("plotOnFailure is only valid for 2-d arrays")
                try:
                    plotImageDiff(lhs, rhs, bad, diff=diff, plotFileName=plotFileName)
                except ImportError:
                    errMsg.append("Failure plot requested but matplotlib could not be imported.")
            if printFailures:
                # Make sure everything is an array if any of them are, so we can treat
                # them the same (diff and absDiff are arrays if either rhs or lhs is),
                # and we don't get here if neither is.
                if numpy.isscalar(relTo):
                    relTo = numpy.ones(bad.shape, dtype=float) * relTo
                if numpy.isscalar(lhs):
                    lhs = numpy.ones(bad.shape, dtype=float) * lhs
                if numpy.isscalar(rhs):
                    rhs = numpy.ones(bad.shape, dtype=float) * rhs
                if rtol is None:
                    for a, b, diff in zip(lhs[bad], rhs[bad], absDiff[bad]):
                        errMsg.append("%s %s %s (diff=%s)" % (a, cmpStr, b, diff))
                else:
                    for a, b, diff, rel in zip(lhs[bad], rhs[bad], absDiff[bad], relTo[bad]):
                        errMsg.append("%s %s %s (diff=%s/%s=%s)" % (a, cmpStr, b, diff, rel, diff/rel))
 
    if msg is not None:
        errMsg.append(msg)
    testCase.assertFalse(failed, msg="\n".join(errMsg))
 
 
@inTestCase
def assertFloatsNotEqual(testCase, lhs, rhs, **kwds):
    """Fail a test if the given floating point values are equal to within the
    given tolerances.
 
    See `~lsst.utils.tests.assertFloatsAlmostEqual` (called with
    ``rtol=atol=0``) for more information.
 
    Parameters
    ----------
    testCase : `unittest.TestCase`
        Instance the test is part of.
    lhs : scalar or array-like
        LHS value(s) to compare; may be a scalar or array-like of any
        dimension.
    rhs : scalar or array-like
        RHS value(s) to compare; may be a scalar or array-like of any
        dimension.
 
    Raises
    ------
    AssertionError
        The values are almost equal.
    """
    return assertFloatsAlmostEqual(testCase, lhs, rhs, invert=True, **kwds)
 
 
@inTestCase
def assertFloatsEqual(testCase, lhs, rhs, **kwargs):
    """
    Assert that lhs == rhs (both numeric types, whether scalar or array).
 
    See `~lsst.utils.tests.assertFloatsAlmostEqual` (called with
    ``rtol=atol=0``) for more information.
 
    Parameters
    ----------
    testCase : `unittest.TestCase`
        Instance the test is part of.
    lhs : scalar or array-like
        LHS value(s) to compare; may be a scalar or array-like of any
        dimension.
    rhs : scalar or array-like
        RHS value(s) to compare; may be a scalar or array-like of any
        dimension.
 
    Raises
    ------
    AssertionError
        The values are not equal.
    """
    return assertFloatsAlmostEqual(testCase, lhs, rhs, rtol=0, atol=0, **kwargs)
 
 
def _settingsIterator(settings):
    """Return an iterator for the provided test settings
 
    Parameters
    ----------
    settings : `dict` (`str`: iterable)
        Lists of test parameters. Each should be an iterable of the same length.
        If a string is provided as an iterable, it will be converted to a list
        of a single string.
 
    Raises
    ------
    AssertionError
        If the ``settings`` are not of the same length.
 
    Yields
    ------
    parameters : `dict` (`str`: anything)
        Set of parameters.
    """
    for name, values in settings.items():
        if isinstance(values, str):
            # Probably meant as a single-element string, rather than an iterable of chars
            settings[name] = [values]
    num = len(next(iter(settings.values())))  # Number of settings
    for name, values in settings.items():
        assert len(values) == num, f"Length mismatch for setting {name}: {len(values)} vs {num}"
    for ii in range(num):
        values = [settings[kk][ii] for kk in settings]
        yield dict(zip(settings.keys(), values))
 
 
def classParameters(**settings):
    """Class decorator for generating unit tests
 
    This decorator generates classes with class variables according to the
    supplied ``settings``.
 
    Parameters
    ----------
    **settings : `dict` (`str`: iterable)
        The lists of test parameters to set as class variables in turn. Each
        should be an iterable of the same length.
 
    Examples
    --------
    ::
 
        @classParameters(foo=[1, 2], bar=[3, 4])
        class MyTestCase(unittest.TestCase):
            ...
 
    will generate two classes, as if you wrote::
 
        class MyTestCase_1_3(unittest.TestCase):
            foo = 1
            bar = 3
            ...
 
        class MyTestCase_2_4(unittest.TestCase):
            foo = 2
            bar = 4
            ...
 
    Note that the values are embedded in the class name.
    """
    def decorator(cls):
        module = sys.modules[cls.__module__].__dict__
        for params in _settingsIterator(settings):
            name = f"{cls.__name__}_{'_'.join(str(vv) for vv in params.values())}"
            bindings = dict(cls.__dict__)
            bindings.update(params)
            module[name] = type(name, (cls,), bindings)
    return decorator
 
 
def methodParameters(**settings):
    """Method decorator for unit tests
 
    This decorator iterates over the supplied settings, using
    ``TestCase.subTest`` to communicate the values in the event of a failure.
 
    Parameters
    ----------
    **settings : `dict` (`str`: iterable)
        The lists of test parameters. Each should be an iterable of the same
        length.
 
    Examples
    --------
    ::
 
        @methodParameters(foo=[1, 2], bar=[3, 4])
        def testSomething(self, foo, bar):
            ...
 
    will run::
 
        testSomething(foo=1, bar=3)
        testSomething(foo=2, bar=4)
    """
    def decorator(func):
        @functools.wraps(func)
        def wrapper(self, *args, **kwargs):
            for params in _settingsIterator(settings):
                kwargs.update(params)
                with self.subTest(**params):
                    func(self, *args, **kwargs)
        return wrapper
    return decorator