偏旁过滤改md库驱动: bot只查Unicode列码点, md缺失才回退硬编码范围, 维护只改md不改代码

monitor.py

@@ -41,18 +41,52 @@ import difflib
 import os
 
 # ⭐ 偏旁部首 Unicode 库：用于 OCR 碎片过滤
-# CJK Radicals Supplement: U+2E80-U+2EFF
-# CJK Radicals Extended: U+2F00-U+2FDF
-# 这些字符在正常中文聊天中不会出现，OCR 拆字碎片才会有
-_RADICAL_RANGES = (
+# 单一数据源 = unicode_radicals.md（人和机器共用）：
+#   - 人用肉眼看「偏旁」列维护
+#   - bot 启动时解析「Unicode」列（U+XXXX）构建码点集合，比对只看码点
+# 维护方式：编辑 md 添加/删除行即可，无需改代码。
+# 仅当 md 缺失/解析失败时才回退到硬编码码位范围（保底，正常不触发）。
+_RADICALS_MD = os.path.join(os.path.dirname(os.path.abspath(__file__)), "unicode_radicals.md")
+# 保底范围：CJK Radicals Supplement U+2E80-U+2EFF + Extended-A U+2F00-U+2FDF
+_RADICAL_RANGES_FALLBACK = (
     (0x2E80, 0x2EFF),  # CJK Radicals Supplement
     (0x2F00, 0x2FDF),  # CJK Radicals Extended-A
 )
 
+def _load_radical_codepoints() -> frozenset:
+    """从 unicode_radicals.md 解析「Unicode」列，返回码点集合。
+
+    行格式: | 亻 | U+2E80 | 单人旁 | 件→亻牛 | 件 |
+    Unicode 列必须是 U+XXXX / U+XXXXX 十六进制；解析失败的行跳过。
+    """
+    cps = set()
+    try:
+        with open(_RADICALS_MD, "r", encoding="utf-8") as f:
+            for line in f:
+                m = re.search(r"U\+([0-9A-Fa-f]{4,5})\b", line)
+                if m:
+                    cps.add(int(m.group(1), 16))
+    except OSError as e:
+        logger.warning(f"[ocr] 偏旁库 {_RADICALS_MD} 读取失败({e})，回退码位范围")
+        return frozenset()
+    if not cps:
+        logger.warning("[ocr] 偏旁库为空，回退码位范围")
+        return frozenset()
+    logger.debug(f"[ocr] 偏旁库加载 {len(cps)} 个码点 <- {_RADICALS_MD}")
+    return frozenset(cps)
+
+_RADICAL_CODEPOINTS = _load_radical_codepoints()
+
 def _is_radical_char(ch: str) -> bool:
-    """检查字符是否为 CJK 偏旁字符（U+2E80-U+2FDF）。"""
+    """检查字符是否为 CJK 偏旁字符。
+
+    优先查 md 库的 Unicode 码点（维护只改 md）；
+    库未加载成功时回退到硬编码码位范围（保底）。
+    """
     cp = ord(ch)
-    return any(lo <= cp <= hi for lo, hi in _RADICAL_RANGES)
+    if _RADICAL_CODEPOINTS:
+        return cp in _RADICAL_CODEPOINTS
+    return any(lo <= cp <= hi for lo, hi in _RADICAL_RANGES_FALLBACK)
 
 import subprocess
 import json