def mk_evolution_step(tp, tn, fp, fn):
step = EvolutionStepEstimator()
step._true_positive = tp
step._true_negative = tn
step._false_positive = fp
step._false_negative = fn
return step
class TestEvolutionStepEstimator(unittest.TestCase):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.sut = EvolutionStepEstimator()
def mk_cyk_result(self, belongs_to_grammar, is_positive):
result = CykResult()
result.belongs_to_grammar = belongs_to_grammar
result.is_positive = is_positive
return result
def assert_estimation(self, tp, tn, fp, fn, total, positives, negatives, fitness):
assert_that(self.sut.true_positive, is_(equal_to(tp)))
assert_that(self.sut.true_negative, is_(equal_to(tn)))
assert_that(self.sut.false_positive, is_(equal_to(fp)))
assert_that(self.sut.false_negative, is_(equal_to(fn)))
assert_that(self.sut.total, is_(equal_to(total)))
assert_that(self.sut.positives_that_has_occurred, is_(equal_to(positives)))
assert_that(self.sut.negatives_that_has_occurred, is_(equal_to(negatives)))
assert_nearly_equal_or_both_nan(self.sut.fitness, fitness)
def test_step_estimation(self):
result_tp = self.mk_cyk_result(True, True)
result_fp = self.mk_cyk_result(True, False)
result_tn = self.mk_cyk_result(False, False)
result_fn = self.mk_cyk_result(False, True)
self.assert_estimation(tp=0, tn=0, fp=0, fn=0, total=0, positives=0, negatives=0,
fitness=float('nan'))
self.sut.append_result(result_tp)
self.assert_estimation(tp=1, tn=0, fp=0, fn=0, total=1, positives=1, negatives=0,
fitness=1)
self.sut.append_result(result_tp)
self.assert_estimation(tp=2, tn=0, fp=0, fn=0, total=2, positives=2, negatives=0,
fitness=1)
self.sut.append_result(result_fp)
self.assert_estimation(tp=2, tn=0, fp=1, fn=0, total=3, positives=2, negatives=1,
fitness=0.67)
self.sut.append_result(result_tn)
self.assert_estimation(tp=2, tn=1, fp=1, fn=0, total=4, positives=2, negatives=2,
fitness=0.75)
self.sut.append_result(result_fn)
self.assert_estimation(tp=2, tn=1, fp=1, fn=1, total=5, positives=3, negatives=2,
fitness=0.6)
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.sut = EvolutionStepEstimator()