def test_distinct_missing_input_implies_no_code_success():
arguments = {
'parameters': {},
'named_inputs': {},
'named_outputs': {
'output data': 'out'
}
}
instance = DistinctOperation(**arguments)
assert instance.generate_code() is None
def test_remove_duplicated_by_attributes_success():
params = {'attributes': ['name']}
n_in = {'input data': 'df1'}
n_out = {'output data': 'out'}
instance = DistinctOperation(params,
named_inputs=n_in,
named_outputs=n_out)
code = instance.generate_code()
expected_code = "{} = {}.drop_duplicates(subset={}, keep='first')"\
.format(n_out['output data'], n_in['input data'], params['attributes'])
result, msg = compare_ast(ast.parse(code), ast.parse(expected_code))
assert result, msg + format_code_comparison(code, expected_code)
def test_remove_duplicated_minimal_params_success():
params = {}
n_in = {'input data': 'df1'}
n_out = {'output data': 'out'}
instance = DistinctOperation(params,
named_inputs=n_in,
named_outputs=n_out)
code = instance.generate_code()
expected_code = "{out} = {in1}.drop_duplicates(subset=None, keep='first')"\
.format(out=n_out['output data'], in1=n_in['input data'])
result, msg = compare_ast(ast.parse(code), ast.parse(expected_code))
assert result, msg + format_code_comparison(code, expected_code)
def test_distinct_missing_attributes_param_success():
df = util.iris(['sepallength'], size=10)
df.loc[0:3, 'sepallength'] = 'test'
df.loc[6:9, 'sepallength'] = 'distinct'
test_df = df.copy()
arguments = {
'parameters': {},
'named_inputs': {
'input data': 'df',
},
'named_outputs': {
'output data': 'out'
}
}
instance = DistinctOperation(**arguments)
result = util.execute(instance.generate_code(), {'df': df})
assert result['out'].equals(test_df.drop(index=[1, 2, 3, 7, 8, 9]))
def test_distinct_invalid_attribute_param_fail():
df = util.iris(['petallength'], 10)
df.loc[0:3, 'petallength'] = 10
df.loc[6:9, 'petallength'] = 10
arguments = {
'parameters': {
'attributes': 'invalid'
},
'named_inputs': {
'input data': 'df',
},
'named_outputs': {
'output data': 'out'
}
}
instance = DistinctOperation(**arguments)
with pytest.raises(NameError) as nam_err:
util.execute(instance.generate_code(), {'df': df})
assert "name 'invalid' is not defined" in str(nam_err.value)
def test_distinct_success():
slice_size = 10
df = [
'df',
util.iris(['sepallength', 'sepalwidth', 'petalwidth', 'petallength'],
slice_size)
]
arguments = {
'parameters': {},
'named_inputs': {
'input data': df[0],
},
'named_outputs': {
'output data': 'out'
}
}
instance = DistinctOperation(**arguments)
result = util.execute(instance.generate_code(), dict([df]))
assert result['out'].equals(util.iris(size=slice_size))