def test_simple(self):
class C(datasets.DataSet):
def __init__(self):
pass
def __len__(self):
return 5
def __getitem__(self, item):
if item == 2:
raise ValueError()
return datasets.PredictionItem(item, "H", 0)
r = ds.DatasetTrial(C())
self.assertEqual(len(r.errors), 1)
pass
def test_binary(self):
class C(datasets.DataSet):
def __init__(self):
pass
def __len__(self):
return 5
def __getitem__(self, item):
return datasets.PredictionItem(item, "H", [item % 2])
r = ds.DatasetTrial(C())
self.assertEqual(len(r.errors), 0)
self.assertEqual(r.outputTrials[0].numFirst, 3)
self.assertEqual(r.get_inputs_count(), 1)
self.assertEqual(r.get_outputs_count(), 1)
pass
def test_multiclass_binary3(self):
class C(datasets.DataSet):
def __init__(self):
pass
def __len__(self):
return 5
def __getitem__(self, item):
return datasets.PredictionItem(item, "H", (np.array(
[float(item % 2), float(item % 3 == 0)])))
r = ds.DatasetTrial(C())
self.assertEqual(len(r.errors), 0)
self.assertEqual(r.get_inputs_count(), 1)
self.assertEqual(r.get_outputs_count(), 1)
self.assertEqual(r.outputTrials[0].counts[0][0], 3)
pass
def test_multiclass(self):
class C(datasets.DataSet):
def __init__(self):
pass
def __len__(self):
return 5
def __getitem__(self, item):
return datasets.PredictionItem(item, "H",
(np.array([(item % 3)])))
r = ds.DatasetTrial(C())
self.assertEqual(len(r.errors), 0)
self.assertEqual(r.get_inputs_count(), 1)
self.assertEqual(r.get_outputs_count(), 1)
self.assertEqual(len(r.outputTrials), 1)
for i in r.outputTrials:
self.assertEqual(i.isMulti, False)
pass
def test_multiclass_binary4(self):
class C(datasets.DataSet):
def __init__(self):
pass
def __len__(self):
return 5
def __getitem__(self, item):
return datasets.PredictionItem(item, "H", (np.array(
[float(item % 2), float(item % 3)])))
r = ds.DatasetTrial(C())
self.assertEqual(len(r.errors), 0)
self.assertEqual(r.get_inputs_count(), 1)
self.assertEqual(r.get_outputs_count(), 1)
for i in r.outputTrials:
if isinstance(i, ds.MultiClassification):
self.assertTrue(False)
pass
def test_multiclass2(self):
class C(datasets.DataSet):
def __init__(self):
pass
def __len__(self):
return 5
def __getitem__(self, item):
return datasets.PredictionItem(item, "H",
(np.array([(item % 3), 6, 7])))
r = ds.DatasetTrial(C())
self.assertEqual(len(r.errors), 0)
self.assertEqual(r.get_inputs_count(), 1)
self.assertEqual(r.get_outputs_count(), 1)
self.assertEqual(len(r.outputTrials), 2)
for i in r.outputTrials:
if isinstance(i, ds.MultiSetClassification):
self.assertEqual(i.isMulti, True)
pass