copy_split.py

import tqdm
import argparse
from pathlib import Path
import shutil

def copy_files_preserve_structure(filelist_path, root_folder, target_root):
    filelist_path = Path(filelist_path)
    root_folder = Path(root_folder).resolve()
    target_root = Path(target_root).resolve()

    if not filelist_path.exists():
        raise FileNotFoundError(f"Filelist not found: {filelist_path}")

    with open(filelist_path, 'r', encoding='utf-8') as f:
        lines = f.readlines()

    copied = 0
    for line in tqdm.tqdm(lines):
        line = line.strip()
        if not line:
            continue

        try:
            filepath_str, _, _, _ = line.split("|", 3)
        except ValueError:
            print(f"Skipping malformed line: {line}")
            continue

        full_src = Path(filepath_str).resolve()
        if not full_src.exists():
            print(f"❌ Source file not found: {full_src}")
            continue

        try:
            rel_path = full_src.relative_to(root_folder)
        except ValueError:
            print(f"❌ File {full_src} is not under the root folder {root_folder}")
            continue

        dst = target_root / rel_path
        dst.parent.mkdir(parents=True, exist_ok=True)

        shutil.copy2(full_src, dst)
        copied += 1

    print(f"\n✅ Completed copying {copied} entries to: {target_root}")

if __name__ == "__main__":
    parser = argparse.ArgumentParser(description="Copy files from filelist to a new folder, preserving directory structure.")
    parser.add_argument("--filelist_path", type=str, required=True, help="Path to the filelist.txt")
    parser.add_argument("--root_folder", type=str, required=True, help="Root folder of original data")
    parser.add_argument("--target_folder", type=str, required=True, help="Root directory to copy files into")
    args = parser.parse_args()

    copy_files_preserve_structure(args.filelist_path, args.root_folder, args.target_folder)