def test_defaults():
dummy_set = App('dummy_set', '', '1.0')
@dummy_set.metric(description='',
input_fields=[IN1, [IN2, IN3]],
output_fields=[OUT1])
def test_defaults(val, df=None):
sums = val.sum()
if df is not None:
sums += df.sum().sum()
return pd.Series([sums])
data_map = {'input': {}, 'output': pd.DataFrame()}
data_map['input']['data'] = DUMMY_DATAFRAME[[IN1.key[-1], IN2.key[-1]]]
report = dummy_set.run(data_map)
assert len(report['existing_results_skipped']) == 0
assert len(report['unneeded_metrics']) == 0
assert len(report['metrics_missing_input']) == 0
assert len(report['run_results']) == 1
assert report['run_results']['test_defaults']['result'] == "Success"
assert OUT1.get_by_path(data_map)[0] == 6
data_map['input']['data'] = DUMMY_DATAFRAME
report = dummy_set.run(data_map)
assert len(report['existing_results_skipped']) == 0
assert len(report['unneeded_metrics']) == 0
assert len(report['metrics_missing_input']) == 0
assert len(report['run_results']) == 1
assert report['run_results']['test_defaults']['result'] == "Success"
assert OUT1.get_by_path(data_map)[0] == 45
def test_subfield_missmatch():
with pytest.raises(AssertionError):
dummy_set = App('dummy_set', '', '1.0')
@dummy_set.metric(description='',
input_fields=[[IN1, IN_CONF]],
output_fields=[OUT1]) # pylint: disable=unused-variable
def test_subfield_missmatch1(df):
return df
dummy_set = App('dummy_set', '', '1.0')
@dummy_set.metric(description='',
input_fields=[[IN1, IN2]],
output_fields=[OUT1]) # pylint: disable=unused-variable
def test_subfield_missmatch(df):
return df
data_map = {'input': {}, 'output': pd.DataFrame()}
data_map['input']['data'] = {'in1': [1, 2, 3], 'in2': [4, 5, 6]}
report = dummy_set.run(data_map)
assert len(report['existing_results_skipped']) == 0
assert len(report['unneeded_metrics']) == 0
assert len(report['metrics_missing_input']) == 0
assert report['run_results']['test_subfield_missmatch']['result'].strip(
).endswith("is not a DataFrame and can't return a column set")
def test_output_mismatch():
def base_dummy(dummy):
ret = 'foo'
return tuple( ret for i in range(dummy) )
dummy_set2 = App('dummy_set', '', '1.0')
with pytest.raises(AssertionError):
@dummy_set2.metric(
description='',
input_fields=[FIELD_INPUT_DUMMY1],
output_fields=[]
)
def dummy_out0(dummy):
return base_dummy(dummy)
@dummy_set2.metric(
description='',
input_fields=[FIELD_INPUT_DUMMY1],
output_fields=[FIELD_RESULT_DUMMY_BAD_OUT]
)
def dummy_out1(dummy):
return base_dummy(dummy)
@dummy_set2.metric(
description='',
input_fields=[FIELD_INPUT_DUMMY1],
output_fields=[FIELD_RESULT_DUMMY_BAD_OUT2, FIELD_RESULT_DUMMY_BAD_OUT3]
)
def dummy_out2(dummy):
return base_dummy(dummy)
data_map = {'input': {}, 'output': {}}
for out_count in range(4):
FIELD_INPUT_DUMMY1.set_by_path(data_map, out_count)
report = dummy_set2.run(data_map)
assert len(report['existing_results_skipped']) == 0
assert len(report['unneeded_metrics']) == 0
assert len(report['metrics_missing_input']) == 0
assert len(report['run_results']) == 2
assert report['run_results']['dummy_out1']['result'] == "Success"
if out_count == 2:
assert report['run_results']['dummy_out2']['result'] == "Success"
else:
assert report['run_results']['dummy_out2']['result'].strip().endswith(
"Metric didn't produce expected number of outputs")
def test_mix():
dummy_set = App('dummy_set', '', '1.0')
val = 'foo'
@dummy_set.metric(description='',
input_fields=[[IN1, IN2], IN_CONF],
output_fields=[OUT1])
def test_mix(df, conf):
return pd.Series([conf + str(df.sum().sum())])
data_map = {'input': {}, 'output': pd.DataFrame()}
data_map['input']['data'] = DUMMY_DATAFRAME
data_map['input']['config'] = 'foo'
report = dummy_set.run(data_map)
assert len(report['existing_results_skipped']) == 0
assert len(report['unneeded_metrics']) == 0
assert len(report['metrics_missing_input']) == 0
assert report['run_results']['test_mix']['result'] == "Success"
test_sum = DUMMY_DATAFRAME[[IN1.key[-1], IN2.key[-1]]].sum().sum()
assert OUT1.get_by_path(data_map)[0] == val + str(test_sum)
def test_missing_one():
dummy_set = App('dummy_set', '', '1.0')
@dummy_set.metric(description='',
input_fields=[[IN1, IN5]],
output_fields=[OUT1])
def test_missing_one(df):
return pd.Series([df.sum().sum()])
data_map = {'input': {}, 'output': pd.DataFrame()}
data_map['input']['data'] = DUMMY_DATAFRAME
report = dummy_set.run(data_map)
assert len(report['existing_results_skipped']) == 0
assert len(
report['unneeded_metrics']) == len(dummy_set.graph.get_metrics()) - 1
assert report['metrics_missing_input'] == {
'test_missing_one': {
'bad_field': 'input/data/in5',
'reason': 'Missing input',
'has_default': False
}
}
assert len(report['run_results']) == 0