src/regx/normalize.py

import re, unicodedata

WS_RE = re.compile(r"[ \t]+")
MULTI_NL_RE = re.compile(r"\n{3,}")

def normalize_text(s: str) -> str:
    s = unicodedata.normalize("NFKC", s)
    s = s.replace("\r\n", "\n").replace("\r", "\n")
    s = WS_RE.sub(" ", s)
    s = MULTI_NL_RE.sub("\n\n", s)
    return s.strip()

def breadcrumb(path: list[str]) -> str:
    return " > ".join(path)