pdf_title_image/scan_dir.py

import os
from typing import List


def scan_dir(path, suffix: str = None):
    results = []
    files = os.listdir(path)

    for file in files:
        file_d = os.path.join(path, file)
        if os.path.isdir(file_d):
            results.extend(scan_dir(file_d, suffix))
        else:
            if (suffix and file.endswith(suffix)) or (not suffix):
                results.append(file_d)
    return results


def batch_ln(files: List[str], target: str):
    cmd = 'ln -s {} {}'
    for f in files:
        if os.path.isfile(f):
            os.system(cmd.format(f, os.path.join(target, os.path.basename(f))))


if __name__ == '__main__':
    fs = scan_dir('/home/zzh/ocr/pdf', 'pdf')
    batch_ln(fs, './all_pdf')