def test_remove_duplicate_marks(param_dict):
b = BallotMarks(param_dict['input'])
computed = BallotMarks.remove_duplicate_candidate_marks(b)
assert computed.marks == param_dict['expected']['marks']
assert computed.unique_marks == param_dict['expected']['unique_marks']
assert computed.unique_candidates == param_dict['expected'][
'unique_candidates']
def _compute_cvr_stat_table(self) -> None:
cvr = self.get_cvr_dict()
candidates = self.get_candidates()
df = pd.DataFrame()
df['weight'] = cvr['weight']
df['valid_ranks_used'] = [len(b.unique_candidates) for b in cvr['ballot_marks']]
df['ranks_used_times_weight'] = df['valid_ranks_used'] * df['weight']
df['used_last_rank'] = [True if b.marks[-1] != BallotMarks.SKIPPED else False for b in cvr['ballot_marks']]
df['undervote'] = [b.unique_marks == {BallotMarks.SKIPPED} for b in cvr['ballot_marks']]
df['ranked_single'] = df['valid_ranks_used'] == 1
df['ranked_multiple'] = df['valid_ranks_used'] > 1
df['ranked_3_or_more'] = df['valid_ranks_used'] > 2
ballot_marks_no_skipped = [BallotMarks.remove_mark(b, [BallotMarks.SKIPPED]) for b in cvr['ballot_marks']]
first_round = [b.marks[0] if b.marks else 'NA' for b in ballot_marks_no_skipped]
df['first_round'] = pd.Series(first_round, dtype='category')
df['first_round_overvote'] = df['first_round'].eq(BallotMarks.OVERVOTE)
df['contains_overvote'] = [BallotMarks.OVERVOTE in b.unique_marks for b in cvr['ballot_marks']]
# contains_skipped
# {SKIPVOTE} & {x} - {y}
# this checks that x == SKIPVOTE and that y then != SKIPVOTE
# (the y check is important to know whether or not the ballot contains marks
# following the skipped rank)
df['contains_skip'] = [any({BallotMarks.SKIPPED} & {x} - {y} for x, y in zip(b.marks, b.marks[1:]))
for b in cvr['ballot_marks']]
# contains_duplicate
# remove overvotes and undervotes
dup_check = [BallotMarks.remove_mark(b, [BallotMarks.SKIPPED, BallotMarks.OVERVOTE]) for b in cvr['ballot_marks']]
# count all ranks for candidates
counters = [collections.Counter(b.marks) for b in dup_check]
# check if any candidates were ranked more than once
df['contains_duplicate'] = [max(counter.values()) > 1 if counter else False for counter in counters]
irregular_condtions = ['contains_overvote', 'contains_skip', 'contains_duplicate']
df['irregular'] = df[irregular_condtions].any(axis='columns')
# fully_ranked
candidates_combined_writeins = BallotMarks.combine_writein_marks(candidates)
candidates_excluded_writeins = BallotMarks.remove_mark(candidates_combined_writeins, [BallotMarks.WRITEIN])
candidate_set = candidates_excluded_writeins.unique_candidates
ballot_marks_cleaned = [BallotMarks.remove_mark(b, [BallotMarks.OVERVOTE, BallotMarks.SKIPPED])
for b in cvr['ballot_marks']]
ballot_marks_cleaned = [BallotMarks.remove_duplicate_candidate_marks(b) for b in ballot_marks_cleaned]
fully_ranked = [(set(b.marks) & candidate_set) == candidate_set or
# voters ranked every possible candidate
len(a.marks) == len(b.marks)
# or did not, had no skipped ranks, overvotes, or duplicates
for a, b in zip(cvr['ballot_marks'], ballot_marks_cleaned)]
df['fully_ranked'] = fully_ranked
self._cvr_stat_table = df
def test_remove_duplicate_marks_errors(error_type, input):
with pytest.raises(error_type):
BallotMarks.remove_duplicate_candidate_marks(input)