筆記一鍵修正¶

# --- 設定（請依需要修改）---
TARGET_PATHS = ["./mkdocs"]  # 檔案或資料夾路徑，可多個
DRY_RUN = False  # True=僅預覽；False=寫入檔案
VERBOSE = False   # 顯示每個檔案和變更數

# --- 轉換工具（自包含）---
import os, glob, re
from pathlib import Path
from typing import Optional, Tuple, List

# 中文數字對照
CN_DIGIT = {
    '零': 0, '〇': 0, '○': 0, '◯': 0,
    '一': 1, '二': 2, '兩': 2, '三': 3, '四': 4, '五': 5,
    '六': 6, '七': 7, '八': 8, '九': 9,
}
CN_UNIT = {'十': 10, '百': 100, '千': 1000}

def cn_to_int(text: str) -> Optional[int]:
    """將中文數字（含 十/百/千/萬）或阿拉伯數字轉為整數。遇到未知字元則回傳 None。
    """
    if re.fullmatch(r'\d+', text):
        return int(text)
    total = 0
    section = 0
    number = 0
    for ch in text:
        if ch in CN_DIGIT:
            number = CN_DIGIT[ch]
        elif ch in CN_UNIT:
            unit = CN_UNIT[ch]
            if number == 0:
                number = 1  # 例如：十=10（前面的 1 省略）
            section += number * unit
            number = 0
        elif ch == '萬':
            part = section + number
            if part == 0:
                part = 1  # 單獨的「萬」視為 1 萬
            total += part * 10000
            section = 0
            number = 0
        else:
            return None
    return total + section + number

# 僅匹配緊貼形式的 第…條 / 第…項
PATTERN = re.compile(r'第([零〇○◯兩二一三四五六七八九十百千萬0-9]+)(條|項)')

def convert_ordinals(s: str) -> Tuple[str, int]:
    count = 0
    def repl(m):
        nonlocal count
        numtxt = m.group(1)
        unit = m.group(2)
        val = cn_to_int(numtxt)
        if val is None:
            return m.group(0)
        if str(val) == numtxt:
            return m.group(0)  # 已是阿拉伯數字
        count += 1
        return f'第{val}{unit}'
    out = PATTERN.sub(repl, s)
    return out, count

# 清單首項前自動補空行（不處理程式碼區塊內文）
LIST_ITEM_RE = re.compile(r'^([\t ]*)(?:> ?)*([\t ]*)(?:[-*+]|\d+[.)])\s+')
FENCE_RE = re.compile(r'^(?:[\t ]*)(```|~~~)')

def ensure_blank_line_before_lists(s: str) -> Tuple[str, int]:
    lines = s.splitlines()
    out: List[str] = []
    fixes = 0
    in_code = False
    fence_seq = None
    for i, line in enumerate(lines):
        # 追蹤程式碼區塊的開始/結束
        m_f = FENCE_RE.match(line)
        if m_f:
            tick = m_f.group(1)
            if not in_code:
                in_code = True
                fence_seq = tick
            elif tick == fence_seq:
                in_code = False
                fence_seq = None
        # 非程式碼區塊時，檢查是否為清單首項，需要補空行
        if not in_code and LIST_ITEM_RE.match(line):
            prev = out[-1] if out else ''
            if prev.strip() != '' and not LIST_ITEM_RE.match(prev):
                out.append('')
                fixes += 1
        out.append(line)
    return '\n'.join(out), fixes

def iter_target_files(paths: List[Path]) -> List[Path]:
    # Traverse every level under each target and collect files.
    # - Recursively walks all subdirectories (follows symlinks).
    # - Supports wildcard patterns in targets (e.g., **/*.md).
    # - Filters to .md/.markdown/.txt (case-insensitive).

    exts = {'.md', '.markdown', '.txt'}
    files: List[Path] = []

    def maybe_add(fp: Path):
        try:
            if fp.is_file() and fp.suffix.lower() in exts:
                files.append(fp)
        except Exception:
            pass

    for p in paths:
        pat = str(p)
        if any(ch in pat for ch in '*?['):
            for g in glob.glob(pat, recursive=True):
                gp = Path(g)
                if gp.is_dir():
                    for root, dirs, names in os.walk(gp, followlinks=True):
                        for name in names:
                            maybe_add(Path(root) / name)
                else:
                    maybe_add(gp)
        elif Path(p).is_dir():
            for root, dirs, names in os.walk(p, followlinks=True):
                for name in names:
                    maybe_add(Path(root) / name)
        elif Path(p).exists():
            maybe_add(Path(p))
    return files