Getting Levenshtein distance from `diifflib.ndiff`

Question

The levenshtein distance algorithm in Python is too slow as I am comparing many strings.

So I want to use difflib.ndiff to do it.

I tried parsing the output by interpreting "+", "-", " " from the result of ndiff, but failed.

Here is what I tried:

import difflib
edit_dist = sum(op[0] == "-" for op in difflib.ndiff("split", "sitting"))

But the returned result is incorrect.

I want to use solutions from stdlib. Any advice / solutions?

P.S. I need to get the nunber edit distance, not ratio so SequenceMatcher.ratio doesn't work