def testGetModelAndOutputNamesEmptyPredictions(self):
eval_config = config_pb2.EvalConfig(
model_specs=[config_pb2.ModelSpec()])
self.assertEmpty(
util.StandardExtracts({
constants.PREDICTIONS_KEY: {}
}).get_model_and_output_names(eval_config))
def testSetLabelsModelNameOutputNameOverwrite(self):
extracts = util.StandardExtracts({
'labels': {
'baseline': {
'output1': 3
},
'candidate': {
'output1': 3
}
}
})
extracts.set_labels(7, model_name='candidate', output_name='output1')
self.assertEqual(
7,
extracts.get_labels(model_name='candidate', output_name='output1'))
self.assertEqual(
{
'labels': {
'baseline': {
'output1': 3
},
'candidate': {
'output1': 7
}
}
}, dict(extracts))
def testGetLabelsNone(self):
self.assertIsNone(
util.StandardExtracts({
'labels': {
'candidate': None,
'baseline': None
},
}).get_labels('candidate'))
def testGetLabelsMultiOutput(self):
self.assertEqual(
7,
util.StandardExtracts({
'labels': {
'output1': 7
},
}).get_labels(output_name='output1'))
def testGetLabelsMultiModel(self):
self.assertEqual(
7,
util.StandardExtracts({
'labels': {
'candidate': 7,
'baseline': 8
},
}).get_labels('candidate'))
def testSetLabelsNonMappingRaisesException(self):
with self.assertRaisesRegex(
RuntimeError, 'set_by_keys failed with arguments:') as ctx:
util.StandardExtracts({
'labels': np.array([7]),
}).set_labels(7, model_name='candidate')
self.assertIsInstance(ctx.exception.__cause__, ValueError)
self.assertRegex(
str(ctx.exception.__cause__),
r'Cannot set keys \(\[\'candidate\'\]\) on a non-mapping root.*')
def testGetModelAndOutputNamesMultiOutput(self):
eval_config = config_pb2.EvalConfig(
model_specs=[config_pb2.ModelSpec()])
self.assertEqual([(None, 'output1'), (None, 'output2')],
util.StandardExtracts({
constants.PREDICTIONS_KEY: {
'output1': np.array([]),
'output2': np.array([])
}
}).get_model_and_output_names(eval_config))
def testGetLabelsMultiModelMultiOutput(self):
self.assertEqual(
7,
util.StandardExtracts({
'labels': {
'candidate': {
'output1': 7
},
'baseline': {
'output1': 8
},
},
}).get_labels('candidate', 'output1'))
def testGetModelAndOutputNamesMultiModel(self):
eval_config = config_pb2.EvalConfig(model_specs=[
config_pb2.ModelSpec(name=constants.BASELINE_KEY),
config_pb2.ModelSpec(name=constants.CANDIDATE_KEY)
])
self.assertEqual([(constants.BASELINE_KEY, None),
(constants.CANDIDATE_KEY, None)],
util.StandardExtracts({
constants.PREDICTIONS_KEY: {
constants.BASELINE_KEY: np.array([]),
constants.CANDIDATE_KEY: np.array([])
}
}).get_model_and_output_names(eval_config))
def testSetLabelsEmptyLabels(self):
extracts = util.StandardExtracts({'labels': {}})
extracts.set_labels(7, model_name='candidate')
self.assertEqual(7, extracts.get_labels(model_name='candidate'))
def testSetLabelsModelNameOutputNameEmptyExtracts(self):
extracts = util.StandardExtracts({})
extracts.set_labels(7, model_name='candidate', output_name='output1')
self.assertEqual(
7,
extracts.get_labels(model_name='candidate', output_name='output1'))
def testGetLabels(self):
self.assertEqual(7, util.StandardExtracts({'labels': 7}).get_labels())
def testGetModelAndOutputNamesEmptyExtracts(self):
eval_config = config_pb2.EvalConfig(
model_specs=[config_pb2.ModelSpec()])
self.assertEmpty(
util.StandardExtracts({}).get_model_and_output_names(eval_config))