A Python solution to the same problem as my 2.clj gist.

2.py

#!/usr/bin/env python3

# Given a list of "passwords" for which e.g. "abc" and "cba" or "cat" and "tac"
# are not considered unique but "abc" and "bac" are considered unique, count the
# unique words in the list.
#
# Usage: python3 2.py < words.txt

import sys
from collections.abc import Set

class PassSet(Set):
    def __init__(self, *args):
        self.data = set(self.normalize_value(arg) for arg in args)

    def __contains__(self, value):
        return self.normalize_value(value) in self.data

    def __iter__(self):
        return iter(self.data)

    def __len__(self):
        return len(self.data)

    def add(self, value):
        self.data.add(self.normalize_value(value))

    def normalize_value(self, value):
        """Note that value must be a string."""

        first, last = value[0], value[-1]
        if first < last:
            return value
        else:
            return ''.join(reversed(value))

def main():
    pd = PassSet()
    for line in sys.stdin: # note `line` ends with \n
        pd.add(line.rstrip())
    print(len(pd))

if __name__ == '__main__':
    main()