def test_model(self, model, expected_score):
# assemblies
objectome = load_assembly()
probabilities = Path(__file__).parent / f'{model}-probabilities.nc'
probabilities = BehavioralAssembly(xr.load_dataarray(probabilities))
# metric
i2n = I2n()
score = i2n(probabilities, objectome)
score = score.sel(aggregation='center')
assert score == approx(expected_score, abs=0.005), f"expected {expected_score}, but got {score}"
def test_model(self, model, expected_score):
# assemblies
objectome = load_assembly()
probabilities = pd.read_pickle(
os.path.join(os.path.dirname(__file__),
f'{model}-probabilities.pkl'))['data']
probabilities = BehavioralAssembly(probabilities)
# metric
i2n = I2n()
score = i2n(probabilities, objectome)
score = score.sel(aggregation='center')
assert score == approx(
expected_score,
abs=0.005), f"expected {expected_score}, but got {score}"
def test_ceiling(self):
objectome = load_assembly()
i2n = I2n()
ceiling = i2n.ceiling(objectome)
assert ceiling.sel(aggregation='center') == approx(.4786, abs=.0064)
assert ceiling.sel(aggregation='error') == approx(.00537, abs=.0015)