Reverse-engineering a ballot set from a specific table of pairwise voting results listed in a 2019 reddit post

reverse_cumulative_table.py

#!/usr/bin/python3
# Script for generating an answer to this:
# <https://www.reddit.com/r/Voting/comments/aqsxft/condorcet_winner_from_counts_of_ranks_possible/>

import json
import itertools
from collections import Counter

innerperm = [x for x in itertools.permutations(['P1', 'P2', 'P3'], 3)]

ballotsets = set()

for result in itertools.product(innerperm, repeat=10):
    topcount = Counter([x[0] for x in result])
    if(topcount['P1'] == 3 and topcount['P2'] == 2 and topcount['P3'] == 5):
        middlecount = Counter([x[1] for x in result])
        if(middlecount['P1'] == 2 and middlecount['P2'] == 6 and middlecount['P3'] == 2):
            bottomcount = Counter([x[2] for x in result])
            if(bottomcount['P1'] == 5 and bottomcount['P2'] == 2 and bottomcount['P3'] == 3):
                ballotcount = Counter(['>'.join(ballot) for ballot in result])
                ballotset = '\n'.join(
                    ["{}:{}".format(ballotcount[b], b) for b in sorted(ballotcount.keys())])
                ballotsets.add(ballotset)


for i, s in enumerate(ballotsets):
    print("Ballot set #{}:".format(i + 1))
    print(s)
    print()