sync_dotfiles.py

import os
import subprocess
from huggingface_hub import snapshot_download

# 绝对敏感：不同步
SENSITIVE_EXCLUDES = [
    "auth.json",
    ".sandbox-secrets/**",
]

# 建议不同步的大缓存/易变目录（减少文件数 & vanished）
HEAVY_EXCLUDES = [
    "plugins/cache/**",
    "projects/**",
    "shell-snapshots/**",
    "statsig/**",
    "todos/**",
    "log/**",
    "tmp/**",
    "sessions/**",
]

def run_rsync(src: str, dst: str, excludes=None, delete=True):
    excludes = excludes or []
    os.makedirs(dst, exist_ok=True)
    cmd = ["rsync", "-a"]
    if delete:
        cmd.append("--delete")
    for ex in excludes:
        cmd += ["--exclude", ex]
    cmd += [src.rstrip("/") + "/", dst.rstrip("/") + "/"]
    p = subprocess.run(cmd)
    # rsync code 24: some files vanished (often due to changing caches). Treat as warning.
    if p.returncode not in (0, 24):
        raise RuntimeError(f"rsync failed with code {p.returncode}")

def main():
    repo = os.getenv("DOTFILES_DATASET_REPO", "").strip()
    if not repo:
        print("DOTFILES_DATASET_REPO 未设置，跳过 dotfiles 同步")
        return

    token = os.getenv("HF_TOKEN") or os.getenv("HUGGINGFACE_HUB_TOKEN")
    if not token:
        print("HF_TOKEN 未设置（Private Dataset 必需），跳过 dotfiles 同步")
        return

    # 只下载你需要的目录，避免拉 1500+ 文件
    allow = [
        "ubuntu/home/.claude/**",
        "ubuntu/home/.codex/**",
        "dotfiles/home/.claude/**",
        "dotfiles/home/.codex/**",
        "ubuntu/project/**",
        "dotfiles/project/**",
        "project/**",
    ]

    local_dir = snapshot_download(
        repo_id=repo,
        repo_type="dataset",
        allow_patterns=allow,
        token=token,
    )

    # 兼容你的结构：优先 ubuntu/home
    cand_home = [
        os.path.join(local_dir, "ubuntu", "home"),
        os.path.join(local_dir, "dotfiles", "home"),
        os.path.join(local_dir, "home"),
    ]
    home_root = next((p for p in cand_home if os.path.isdir(p)), None)

    if not home_root:
        print("未找到 home 根目录（例如 ubuntu/home）")
        return

    claude_src = os.path.join(home_root, ".claude")
    codex_src  = os.path.join(home_root, ".codex")

    # 只同步这两个目录，不碰 /home/user 其它文件
    if os.path.isdir(claude_src):
        print(f"Sync ~/.claude: {claude_src} -> /home/user/.claude")
        run_rsync(claude_src, "/home/user/.claude", excludes=HEAVY_EXCLUDES, delete=True)
    else:
        print("未找到 .claude 目录，跳过")

    if os.path.isdir(codex_src):
        excludes = SENSITIVE_EXCLUDES + HEAVY_EXCLUDES
        print(f"Sync ~/.codex: {codex_src} -> /home/user/.codex (exclude auth/secrets)")
        run_rsync(codex_src, "/home/user/.codex", excludes=excludes, delete=True)
    else:
        print("未找到 .codex 目录，跳过")

    # 可选：项目级配置（如果你未来加了）
    cand_proj = [
        os.path.join(local_dir, "ubuntu", "project"),
        os.path.join(local_dir, "dotfiles", "project"),
        os.path.join(local_dir, "project"),
    ]
    proj_src = next((p for p in cand_proj if os.path.isdir(p)), None)
    if proj_src:
        print(f"Sync project: {proj_src} -> /home/user/work")
        run_rsync(proj_src, "/home/user/work", excludes=[], delete=True)
    else:
        print("未找到 PROJECT dotfiles 源目录（可忽略）")

    print("Dotfiles sync done.")

if __name__ == "__main__":
    main()