def test_aggregate_handles_axis(axis):
with pytest.raises(ValueError):
aggregate(
pd.DataFrame(index=pd.DatetimeIndex(['2017-01'])),
pd.DataFrame(index=pd.DatetimeIndex(['2017-01'])),
axis=axis,
)
def test_aggregate_output_type_weight_dataframe(
indices_3years,
weights_3years,
):
"""The aggregate function should return a pandas Series."""
# GIVEN indices and weights pandas DataFrames
# WHEN they are aggregated together
# THEN returned aggregated is of type pandas Series
aggregated = aggregate(indices_3years, weights_3years)
assert isinstance(aggregated, pd.core.series.Series)
def test_aggregate_output_type_weight_series(
indices_6months,
weights_6months,
):
"""The aggregate function should return a pandas Series."""
# GIVEN indices DataFrame and weights Series
# WHEN they are aggregated together
# THEN returned aggregated is of type pandas Series
aggregated = aggregate(indices_6months, weights_6months)
assert isinstance(aggregated, pd.Series)
def test_aggregate(aggregate_combinator):
"""Functional test to test for expected output.
Tests the following scenarios:
* 3 years, 3 indices, 3 weights
* 1 year, 3 indices, 3 weights
* 6 months, 3 indices, 3 weights
* 3 years, 3 indices (transposed), 3 weights (transposed)
* 3 years, 3 indices (missing values), 3 weights
* 3 years, 3 indices (missing values + transposed),
3 weights (transposed)
"""
# GIVEN indices and weights
# AND the outcome
# WHEN indices and weights are aggregated together
# THEN they should equal the outcome
indices, weights, outcome, axis = aggregate_combinator
aggregated = aggregate(indices, weights, axis=axis)
assert_series_equal(aggregated, outcome, check_names=False)
(Timestamp('2013-04-01 00:00:00'), 103.9939693, 103.707301,
102.8423593),
(Timestamp('2013-05-01 00:00:00'), 107.55573159999999, np.nan,
106.4030102),
(Timestamp('2013-06-01 00:00:00'), 102.15950079999999, 102.1210296,
103.6548477),
(Timestamp('2013-07-01 00:00:00'), 106.83231370000001,
105.71647240000001, 106.23764979999999),
], ).set_index(0, drop=True)
weights2 = pd.DataFrame.from_records([
(Timestamp('2013-01-01 00:00:00'), 13.1223919, 7.90254844, 3.20531341),
]).set_index(0, drop=True).squeeze()
aggregated2 = aggregate(indices2.T,
weights2.T,
axis=0,
ignore_na_indices=True)
aggregated3 = aggregate(indices2, weights2, axis=1, ignore_na_indices=True)
def aggregate_indices_missing(aggregate_indices_3years):
""" """
aggregate_indices_missing = aggregate_indices_3years.copy()
change_to_nans = [
('2012-06', 2),
('2012-12', 3),
('2013-10', 2),
('2014-07', 1),
]
for sl in change_to_nans: