def aggregate_loss(self, groupby=None, kwargs=None):
"""
Aggregate data by the `groupby` attribute, using the `kwargs` to
perform any arithmetic aggregation on fields (e.g. summation,
mean, etc.)
:param str groupby: A column in the `DataFrame` that corresponds to
regions by which to aggregate data
:param dict kwargs: A `dict` with keys of valid column names (from the
`DataFrame`) and values being lists of aggregation functions to apply
to the columns.
For example::
kwargs = {'REPLACEMENT_VALUE': ['mean', 'sum'],
'structural_loss_ratio': ['mean', 'std']}
See
https://pandas.pydata.org/pandas-docs/stable/user_guide/groupby.html#aggregation
for more guidance on using aggregation with `DataFrames`
"""
LOGGER.info(f"Aggregating loss using {groupby} attribute")
a1 = self.prov.activity(":AggregateLoss",
datetime.now().strftime(DATEFMT), None, {
"prov:type": "Aggregation",
"void:aggregator": repr(groupby)
})
self.prov.wasInformedBy(a1, self.provlabel)
self.exposure_agg = aggregate.aggregate_loss_atts(
self.exposure_att, groupby, kwargs)
def test_aggregate_loss_atts(self):
data_frame = pd.DataFrame({'A': ['X', 'Y', 'Y'], 'B': [1, 2, 3]})
aggregated = aggregate_loss_atts(data_frame, 'A', {'B': 'sum'})
assert_frame_equal(pd.DataFrame({
'A': ['X', 'Y'],
'B': [1, 5]
}), aggregated)
def test_aggregate_loss_atts_invalid_field(self):
with self.assertRaises(SystemExit) as context:
data_frame = pd.DataFrame({'A': ['X', 'Y', 'Y'], 'B': [1, 2, 3]})
aggregate_loss_atts(data_frame, 'Z')
self.assertEqual(context.exception.code, 1)