| from __future__ import annotations | |
| def normalize_text(text: str) -> str: | |
| if not text: | |
| return "" | |
| text = text.replace("\\n", "").replace("\\\n", "") | |
| return text.replace(" ", "").replace("\u3000", "").replace("\t", "").replace("\r", "").replace("\n", "") | |
| _PUNCT_CHARS = ( | |
| r"""!"#$%&'()*+,-./:;<=>?@[\]^_`{|}~""" | |
| "，。、；：「」『』（）【】〔〕〈〉《》" | |
| "！？…—–·．“”‘’「」『』〝〞" | |
| "￥％＃＆＊＠／＼｜＋－＝＿" | |
| "～｀＾" | |
| ) | |
| _PUNCT_TRANS = str.maketrans("", "", _PUNCT_CHARS) | |
| def normalize_for_parsing(text: str) -> str: | |
| """字符提取 / 1-NED 评分专用：基础清洗 + 去标点。""" | |
| return normalize_text(text).translate(_PUNCT_TRANS) | |