Add benchmark

[jakewilliami]

Look into benchmarking to make it easy to see if changes make it faster to run. We currently have an efficient first case for this purpose but I would still be interested in performance in future.

[jakewilliami]

Compare to other alternatives:

import unicodedata
import regex

def ispunct_unicodedata(c: str) -> bool:
    # May be more efficient than:
    #     UTF8PROC_CATEGORY_PC <= category_code(c) <= UTF8PROC_CATEGORY_PO
    # Although category_code may be more accurate/correct
    return unicodedata.category(c).startswith("P")

def ispunct_regex(c: str) -> bool:
    return bool(regex.match(r"\p{P}", c))