def test_median_votes_groupped_by(self):
data = pd.core.frame.DataFrame([
{'name': 'Joao', 'party': 'PT', 'state': 'PB', 'poll1': 1},
{'name': 'Joana', 'party': 'PT', 'state': 'PB', 'poll1': 1},
{'name': 'Marcio', 'party': 'PT', 'state': 'PB', 'poll1': 0},
{'name': 'Pedro', 'party': 'PSOL', 'state': 'PE', 'poll1': 0},
])
expected_result = {
'poll1': {
'PT': 1,
'PSOL': 0,
}
}
groupby = 'party'
rollcall = Rollcall(data)
result = rollcall.median_votes_groupped_by(groupby)
self.assertEqual(result.to_dict(), expected_result)
def main(self, csv_path,
majority_percentual=None, groupby=None,
metric_method=RiceIndex().calculate_adjusted, **filters):
"""Calculates the adjusted Rice Index polls contained in a CSV
Args:
csv_path (string): Path to a CSV file with polls as columns and
people/groups' votes as rows. The votes should be either 1 for
YES, 2 for NO, or empty for Not Voted. Example:
poll1,poll2,poll3
0,,1
1,1,1
majority_percentual (float): Removes votes where the majority was
greater than this percentual. Defaults to None.
groupby (string): Column on the metadata to group the votes by.
This is useful when you want to compare a larger party with a
smaller one. You would set groupby = "party", so this method
will get each party's most common vote to calculate the
cohesion. Defaults to None.
metric_method (function): Method that receives a list of votes and
returns a score. For an example, check RiceIndex().calculate.
Defaults to RiceIndex().calculate_adjusted.
filters (kwargs): dict of filters to limit which votes we consider
when calculating the metric. Defaults to None.
Returns:
list(OrderedDict): A list of dicts with each poll name in the keys
and the resulting metric score in the values. If there's no
`metric_method`, it'll simply remove unanimous votes, apply
groups and filters.
"""
votes = Rollcall.from_csv(csv_path)\
.remove_unanimous_votes(majority_percentual)\
.filter(filters)\
.median_votes_groupped_by(groupby)
if metric_method:
metrics = self.calculate_metric(votes, metric_method)
return [collections.OrderedDict(zip(votes.columns, metrics))]
if groupby:
votes.insert(0, groupby, votes.index)
rows = []
columns = votes.columns.tolist()
replace_nan_with_none = lambda df: df.where(pd.notnull(df), None)
for row in replace_nan_with_none(votes).itertuples(False):
row = [self._convert_to_int_if_possible(r) for r in row]
rows.append(collections.OrderedDict(zip(columns, row)))
return rows
