pytc.vmc.test.test_adaptive_burn_in.TestAdaptiveBurnIn

class pytc.vmc.test.test_adaptive_burn_in.TestAdaptiveBurnIn(methodName='runTest')

Bases: TestCase

Task #12 tier-2: burn_in() should terminate on ensemble-stability evidence rather than a fixed, easy-to-under-provision step count.

Methods

__call__	Call self as a function.
__delattr__	Implement delattr(self, name).
__dir__	Default dir() implementation.
__eq__	Return self==value.
__format__	Default object formatter.
__ge__	Return self>=value.
__getattribute__	Return getattr(self, name).
__getstate__	Helper for pickle.
__gt__	Return self>value.
__hash__	Return hash(self).
__init__	Create an instance of the class that will use the named test method when executed.
__init_subclass__	This method is called when a class is subclassed.
__le__	Return self<=value.
__lt__	Return self<value.
__ne__	Return self!=value.
__new__
__reduce__	Helper for pickle.
__reduce_ex__	Helper for pickle.
__repr__	Return repr(self).
__setattr__	Implement setattr(self, name, value).
__sizeof__	Size of object in memory, in bytes.
__str__	Return str(self).
__subclasshook__	Abstract classes can override this to customize issubclass().
_addDuration
_addExpectedFailure
_addUnexpectedSuccess
_baseAssertEqual	The default assertEqual implementation, not type specific.
_callCleanup
_callSetUp
_callTearDown
_callTestMethod
_formatMessage	Honour the longMessage attribute when generating failure messages.
_fresh_walkers
_getAssertEqualityFunc	Get a detailed comparison function for the types of the two args.
_truncateMessage
addClassCleanup	Same as addCleanup, except the cleanup items are called even if setUpClass fails (unlike tearDownClass).
addCleanup	Add a function, with arguments, to be called when the test is completed.
addTypeEqualityFunc	Add a type specific assertEqual style function to compare a type.
assertAlmostEqual	Fail if the two objects are unequal as determined by their difference rounded to the given number of decimal places (default 7) and comparing to zero, or by comparing that the difference between the two objects is more than the given delta.
assertCountEqual	Asserts that two iterables have the same elements, the same number of times, without regard to order.
assertDictEqual
assertEqual	Fail if the two objects are unequal as determined by the '==' operator.
assertFalse	Check that the expression is false.
assertGreater	Just like self.assertTrue(a > b), but with a nicer default message.
assertGreaterEqual	Just like self.assertTrue(a >= b), but with a nicer default message.
assertIn	Just like self.assertTrue(a in b), but with a nicer default message.
assertIs	Just like self.assertTrue(a is b), but with a nicer default message.
assertIsInstance	Same as self.assertTrue(isinstance(obj, cls)), with a nicer default message.
assertIsNone	Same as self.assertTrue(obj is None), with a nicer default message.
assertIsNot	Just like self.assertTrue(a is not b), but with a nicer default message.
assertIsNotNone	Included for symmetry with assertIsNone.
assertLess	Just like self.assertTrue(a < b), but with a nicer default message.
assertLessEqual	Just like self.assertTrue(a <= b), but with a nicer default message.
assertListEqual	A list-specific equality assertion.
assertLogs	Fail unless a log message of level level or higher is emitted on logger_name or its children.
assertMultiLineEqual	Assert that two multi-line strings are equal.
assertNoLogs	Fail unless no log messages of level level or higher are emitted on logger_name or its children.
assertNotAlmostEqual	Fail if the two objects are equal as determined by their difference rounded to the given number of decimal places (default 7) and comparing to zero, or by comparing that the difference between the two objects is less than the given delta.
assertNotEqual	Fail if the two objects are equal as determined by the '!=' operator.
assertNotIn	Just like self.assertTrue(a not in b), but with a nicer default message.
assertNotIsInstance	Included for symmetry with assertIsInstance.
assertNotRegex	Fail the test if the text matches the regular expression.
assertRaises	Fail unless an exception of class expected_exception is raised by the callable when invoked with specified positional and keyword arguments.
assertRaisesRegex	Asserts that the message in a raised exception matches a regex.
assertRegex	Fail the test unless the text matches the regular expression.
assertSequenceEqual	An equality assertion for ordered sequences (like lists and tuples).
assertSetEqual	A set-specific equality assertion.
assertTrue	Check that the expression is true.
assertTupleEqual	A tuple-specific equality assertion.
assertWarns	Fail unless a warning of class warnClass is triggered by the callable when invoked with specified positional and keyword arguments.
assertWarnsRegex	Asserts that the message in a triggered warning matches a regexp.
countTestCases
debug	Run the test without collecting errors in a TestResult
defaultTestResult
doClassCleanups	Execute all class cleanup functions.
doCleanups	Execute all cleanup functions.
enterClassContext	Same as enterContext, but class-wide.
enterContext	Enters the supplied context manager.
fail	Fail immediately, with the given message.
id
run
setUp	Hook method for setting up the test fixture before exercising it.
setUpClass	Hook method for setting up class fixture before running tests in the class.
shortDescription	Returns a one-line description of the test, or None if no description has been provided.
skipTest	Skip this test.
subTest	Return a context manager that will return the enclosed block of code in a subtest identified by the optional message and keyword parameters.
tearDown	Hook method for deconstructing the test fixture after testing it.
tearDownClass	Hook method for deconstructing the class fixture after running all tests in the class.
test_acceptance_target_zero_rejected	acceptance_target=0 would divide by zero in the step-size controller; it must raise ValueError instead.
test_burn_in_honours_adapt_step_size_false	burn_in with adapt_step_size=False must return the initial step_size unchanged (controller owned by the caller).
test_first_complete_interval_uses_all_interval_samples	After exactly one full interval, the returned step_size must equal initial * clip(mean(all interval acceptances) / 0.5).
test_hits_hard_cap_without_hanging	An impossible stability tolerance must still terminate at max_steps rather than loop forever.
test_no_adaptation_before_first_complete_interval	With n_steps < report_interval, no adaptation may fire at all -- in particular not at step 0 off a single acceptance sample.
test_nonpositive_chunk_and_max_steps_rejected	chunk_size<=0 never advances total_steps (infinite loop) and max_steps<=0 is meaningless; both must raise ValueError.
test_oversized_step_never_zeros_step_size	Regression: with a wildly oversized initial step (acceptance ~0) the old unclipped single-step controller could drive step_size to exactly 0, killing the chain permanently.
test_pregate_skips_evaluation_when_acceptance_out_of_band	A near-zero acceptance tolerance on freshly-initialized (highly accepted) walkers should never pass the pre-gate, so every chunk's mean_energy/variance stay None.
test_step_size_adapts_between_chunks_on_chunk_mean	The between-chunk controller must pull step_size toward the acceptance target from both sides: starting far apart, two runs with the same seed must end closer together than they started.
test_terminates_early_when_stable	With loose tolerances, should stop well short of max_steps and report the E/Var that triggered termination.

Attributes

__annotations__
__dict__
__doc__
__module__
__weakref__	list of weak references to the object
_classSetupFailed
_class_cleanups
_diffThreshold
longMessage
maxDiff

classmethod setUpClass()

Hook method for setting up class fixture before running tests in the class.

test_terminates_early_when_stable()

With loose tolerances, should stop well short of max_steps and report the E/Var that triggered termination.

test_hits_hard_cap_without_hanging()

An impossible stability tolerance must still terminate at max_steps rather than loop forever.

test_pregate_skips_evaluation_when_acceptance_out_of_band()

A near-zero acceptance tolerance on freshly-initialized (highly accepted) walkers should never pass the pre-gate, so every chunk’s mean_energy/variance stay None.

test_step_size_adapts_between_chunks_on_chunk_mean()

The between-chunk controller must pull step_size toward the acceptance target from both sides: starting far apart, two runs with the same seed must end closer together than they started.

test_oversized_step_never_zeros_step_size()

Regression: with a wildly oversized initial step (acceptance ~0) the old unclipped single-step controller could drive step_size to exactly 0, killing the chain permanently. The clipped chunk-mean controller must keep it positive and shrinking.

test_no_adaptation_before_first_complete_interval()

With n_steps < report_interval, no adaptation may fire at all – in particular not at step 0 off a single acceptance sample.

test_first_complete_interval_uses_all_interval_samples()

After exactly one full interval, the returned step_size must equal initial * clip(mean(all interval acceptances) / 0.5).

test_acceptance_target_zero_rejected()

acceptance_target=0 would divide by zero in the step-size controller; it must raise ValueError instead.

test_nonpositive_chunk_and_max_steps_rejected()

chunk_size<=0 never advances total_steps (infinite loop) and max_steps<=0 is meaningless; both must raise ValueError.

test_burn_in_honours_adapt_step_size_false()

burn_in with adapt_step_size=False must return the initial step_size unchanged (controller owned by the caller).