Alternative (not really!) to itertools.permutations. Sweet recursive indulgence. However, due to lack of in-built support for TRE in Python, this is a slow implementation. So please do not try this at home.

p&c

# inspired by the algorithm to find all possible strings from a given
# set of alphabets (as in the function permute here)
# the execution time is compared with itertools.permutations

def permute(lst):
    if len(lst) < 2:
        yield lst
    else:
        for perm in permute(lst[1:]):
            for i in range(len(lst)):
                yield(perm[:i] + lst[0] + perm[i:])


def combine(lst, r):
    if r == 1:
        for i in lst:
            yield i
    else:
        for i in range(len(lst)):
            for comb in combine(lst[i + 1:], r - 1):
                yield(lst[i] + comb)


def permutate(lst, r):
    for item in combine(lst, r):
        for out in permute(item):
            yield out


if __name__ == '__main__':
    for item in permutate(['a', 'b', 'c', 'd'], 2):
        print item
    from timeit import timeit
    print(timeit("permutate(['a', 'b', 'c', 'd'], 2)",
                 setup="from __main__ import permutate", number=1000))
    print(timeit("permutations(['a', 'b', 'c', 'd'], 2)",
                 setup="from itertools import permutations", number=1000))