note_cleaner: remove duplicate titles inside notes as added by repeated nvpy/resoph/nvalt migrations

note_cleaner.py

from pathlib import Path
import re, sys, os


def clean_note_lines(title: str, lines: list[str]):
    """
    Remove any lines matching title
    """
    nn = re.compile(r"[^a-zA-Z0-9]")
    def norm(s):
        return nn.sub('', s).lower()

    s0 = norm(title)
    for line in lines:
        if not norm(line)==s0:
            yield line


def clone_mtime(src: Path, dst: Path):
    s = src.stat()
    os.utime(dst, (s.st_ctime, s.st_mtime))


def clean_note(p: Path, po: Path):
    print(f"Cleaning: {p} --> {po}")
    with p.open(encoding='utf8') as f, po.open(mode='w',encoding='utf8', newline='\n') as fo:
        fo.writelines(clean_note_lines(p.stem, f))
    clone_mtime(p, po)


def clean_name(s):
    return re.sub('_', ' ', s)


def main():
    p0 = Path(sys.argv[1])
    p1 = Path(sys.argv[2])
    for p in p0.glob('*.md'):
        clean_note(p, p1 / clean_name(str(p.relative_to(p0))))

if __name__ == '__main__':
    main()