@pytest.mark.twelve def test_csv_3899_eb2_small__procedure_hifa_calimage__regression(): """PIPE-2245: Run small ALMA cal+image regression to cover various heuristics Recipe name: procedure_hifa_calimage Dataset: CSV-3899-EB2-small """ ref_directory = 'pl-regressiontest/CSV-3899-EB2-small' pt = PipelineTester( visname=['uid___A002_X1181695_X1c6a4_8ant.ms'], recipe='procedure_hifa_calimage.xml', input_dir=ref_directory, expectedoutput_dir=ref_directory, ) > pt.run(omp_num_threads=1) tests/regression/fast/alma_if_fast_test.py:181: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ tests/testing_utils.py:381: in run self.__compare_results(new_file, default_relative_tolerance) _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ self = <tests.testing_utils.PipelineTester object at 0x7f68d8fc4800> new_file = 'uid___A002_X1181695_X1c6a4_8ant.ms.NEW.results.txt' relative_tolerance = 1e-07 def __compare_results(self, new_file: str, relative_tolerance: float) -> None: """ Compare results between new one loaded from file and old one. Args: new_file : file path of new results relative_tolerance : relative tolerance of output value """ with open(self.expectedoutput_file) as expected_fd, open(new_file) as new_fd: expected_results = expected_fd.readlines() new_results = new_fd.readlines() errors = [] worst_diff = (0, 0) worst_percent_diff = (0, 0) for old, new in zip(expected_results, new_results): try: oldkey, oldval, tol = self.__sanitize_results_string(old) newkey, newval, _ = self.__sanitize_results_string(new) except ValueError as e: errorstr = "The results: {0} could not be parsed. Error: {1}".format(new, str(e)) errors.append(errorstr) continue assert oldkey == newkey, f"Expected key {oldkey} does not match new key {newkey}." tolerance = tol if tol else relative_tolerance if newval is not None: LOG.info('Comparing %s to %s with a rel. tolerance of %s', oldval, newval, tolerance) if oldval != pytest.approx(newval, rel=tolerance): diff = oldval-newval percent_diff = (oldval-newval)/oldval * 100 if oldval != 0 else 100 if abs(diff) > abs(worst_diff[0]): worst_diff = diff, oldkey if abs(percent_diff) > abs(worst_percent_diff[0]): worst_percent_diff = percent_diff, oldkey errorstr = f"{oldkey}\n\tvalues differ by > a relative difference of {tolerance}\n\texpected: {oldval}\n\tnew: {newval}\n\tdiff: {diff}\n\tpercent_diff: {percent_diff}%" errors.append(errorstr) elif oldval is not None: # If only the new value is None, fail errorstr = f"{oldkey}\n\tvalue is None\n\texpected: {oldval}\n\tnew: {newval}" errors.append(errorstr) else: # If old and new values are both None, this is expected, so pass LOG.info('Comparing %s and %s... both values are None.', oldval, newval) [LOG.warning(x) for x in errors] n_errors = len(errors) if n_errors > 0: summary_str = f"Worst absolute diff, {worst_diff[1]}: {worst_diff[0]}\nWorst percentage diff, {worst_percent_diff[1]}: {worst_percent_diff[0]}%" errors.append(summary_str) > pytest.fail("Failed to match {0} result value{1} within tolerance{1} :\n{2}".format( n_errors, '' if n_errors == 1 else 's', '\n'.join(errors)), pytrace=True) E Failed: Failed to match 2 result values within tolerances : E s16.hifa_gfluxscaleflag.uid___A002_X1181695_X1c6a4_8ant.qa.metric.CorrectedampflagQAHandler E values differ by > a relative difference of 1e-07 E expected: 0.0053688072299937365 E new: 0.0 E diff: 0.0053688072299937365 E percent_diff: 100.0% E s16.hifa_gfluxscaleflag.uid___A002_X1181695_X1c6a4_8ant.qa.score.CorrectedampflagQAHandler E values differ by > a relative difference of 1e-07 E expected: 0.9946311927700062 E new: 1.0 E diff: -0.005368807229993777 E percent_diff: -0.539778690736802% E Worst absolute diff, s16.hifa_gfluxscaleflag.uid___A002_X1181695_X1c6a4_8ant.qa.score.CorrectedampflagQAHandler: -0.005368807229993777 E Worst percentage diff, s16.hifa_gfluxscaleflag.uid___A002_X1181695_X1c6a4_8ant.qa.metric.CorrectedampflagQAHandler: 100.0% tests/testing_utils.py:440: Failed