too complex

Small nitpick, but I believe space complexity is still O(1) in practice. It will depend on the number of unique characters, not the total count. The number should remain relatively low (<256 for ASCII), although it can technically be as high as n for variable-length encodings (e.g., UTF-8).

I give points for "clever" use of acollections.

But not best practice. Python has better solutions that are more concise. Not a fan of using 2 for loops here. My time trials show that a translation table is better, but also you are correct that using a dict is better than count.

Nice solution 👏👏👏

Brilliant

Allocation for new string is expensive.

Very informative solution!

That doesn't make the complexity exponential, that would just make it O(n + n), which is still just O(n). O(n^2) would be if there was a loop that went through n and a nested loop that always went through n, or something similar. Maybe you mean that .join would be faster than +=.

TIL, thanks.

Is there a way to use defaultdict as dict comprehension?

You can round expression $p0 = $p0 + ($p0 * ($percent / 100.0) + $aug); to $p0 += ($p0 * ($percent / 100) + $aug);

@nUG48YpaiHBG you mean in terms of performace?

@felyxjet thank you for this comment, I knew strings were immutable but didn't put together that it meant creating a new string every time.
Does this mean that using the str.join() method is like using the list.append() method?
TIA!

The version you suggest will be slower (because that thing is implemented in pure python, while the present one is done with cython). But yes, the present use isn't mandatory.