Update tools/hf_backup.py
Browse files- tools/hf_backup.py +67 -66
tools/hf_backup.py
CHANGED
|
@@ -1,85 +1,86 @@
|
|
| 1 |
#!/usr/bin/env python3
|
| 2 |
# -*- coding: utf-8 -*-
|
| 3 |
-
import os, sys, time, tarfile, tempfile, shutil
|
| 4 |
-
from huggingface_hub import HfApi
|
| 5 |
-
|
| 6 |
-
def log(*a): print(time.strftime("[%Y-%m-%d %H:%M:%S]"), *a, flush=True)
|
| 7 |
-
|
| 8 |
-
token = os.environ.get("HF_TOKEN")
|
| 9 |
-
repo_id = os.environ.get("DATASET_ID")
|
| 10 |
-
max_backups = int(os.environ.get("MAX_BACKUPS", "10"))
|
| 11 |
-
backup_paths = [p.strip() for p in os.environ.get("BACKUP_PATHS", "/app/data,/app/logs").split(",") if p.strip()]
|
| 12 |
-
prefix = os.environ.get("BACKUP_PREFIX", "crs_backup_")
|
| 13 |
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
|
| 20 |
-
def list_backups():
|
| 21 |
files = api.list_repo_files(repo_id=repo_id, repo_type="dataset")
|
| 22 |
backs = [f for f in files if f.startswith(prefix) and f.endswith(".tar.gz")]
|
| 23 |
backs.sort()
|
| 24 |
return backs
|
| 25 |
|
| 26 |
-
def
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
|
|
|
|
|
|
| 31 |
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
log("add to archive:", d, "as", arc)
|
| 40 |
-
tar.add(d, arcname=arc, recursive=True)
|
| 41 |
-
return path, name, tmpdir
|
| 42 |
|
| 43 |
-
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
|
| 53 |
-
|
| 54 |
|
| 55 |
-
def
|
| 56 |
-
|
|
|
|
| 57 |
if not backs:
|
| 58 |
-
|
| 59 |
-
|
| 60 |
latest = backs[-1]
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
|
| 66 |
-
|
|
|
|
|
|
|
|
|
|
| 67 |
|
| 68 |
def main():
|
| 69 |
-
|
| 70 |
-
|
| 71 |
-
|
| 72 |
-
|
| 73 |
-
|
| 74 |
-
|
| 75 |
-
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
|
| 80 |
-
|
| 81 |
-
|
| 82 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 83 |
|
| 84 |
if __name__ == "__main__":
|
| 85 |
main()
|
|
|
|
| 1 |
#!/usr/bin/env python3
|
| 2 |
# -*- coding: utf-8 -*-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
|
| 4 |
+
import argparse
|
| 5 |
+
import os
|
| 6 |
+
import sys
|
| 7 |
+
import tempfile
|
| 8 |
+
from huggingface_hub import HfApi
|
| 9 |
|
| 10 |
+
def list_backups(api: HfApi, repo_id: str, prefix: str):
|
| 11 |
files = api.list_repo_files(repo_id=repo_id, repo_type="dataset")
|
| 12 |
backs = [f for f in files if f.startswith(prefix) and f.endswith(".tar.gz")]
|
| 13 |
backs.sort()
|
| 14 |
return backs
|
| 15 |
|
| 16 |
+
def upload(args):
|
| 17 |
+
api = HfApi(token=args.token)
|
| 18 |
+
# 可选:确认数据集存在
|
| 19 |
+
try:
|
| 20 |
+
api.dataset_info(repo_id=args.repo)
|
| 21 |
+
except Exception:
|
| 22 |
+
pass
|
| 23 |
|
| 24 |
+
# 上传
|
| 25 |
+
api.upload_file(
|
| 26 |
+
path_or_fileobj=args.file,
|
| 27 |
+
path_in_repo=os.path.basename(args.file),
|
| 28 |
+
repo_id=args.repo,
|
| 29 |
+
repo_type="dataset",
|
| 30 |
+
)
|
|
|
|
|
|
|
|
|
|
| 31 |
|
| 32 |
+
# 保留最新 N 份
|
| 33 |
+
if args.max and args.max > 0:
|
| 34 |
+
backs = list_backups(api, args.repo, args.prefix)
|
| 35 |
+
if len(backs) > args.max:
|
| 36 |
+
to_del = backs[: len(backs) - args.max]
|
| 37 |
+
for f in to_del:
|
| 38 |
+
try:
|
| 39 |
+
api.delete_file(path_in_repo=f, repo_id=args.repo, repo_type="dataset")
|
| 40 |
+
except Exception:
|
| 41 |
+
# 删除失败不致命
|
| 42 |
+
pass
|
| 43 |
|
| 44 |
+
def restore(args):
|
| 45 |
+
api = HfApi(token=args.token)
|
| 46 |
+
backs = list_backups(api, args.repo, args.prefix)
|
| 47 |
if not backs:
|
| 48 |
+
return # 没有备份则安静退出,由 shell 脚本处理
|
| 49 |
+
|
| 50 |
latest = backs[-1]
|
| 51 |
+
with tempfile.TemporaryDirectory() as d:
|
| 52 |
+
path = api.hf_hub_download(
|
| 53 |
+
repo_id=args.repo,
|
| 54 |
+
filename=latest,
|
| 55 |
+
repo_type="dataset",
|
| 56 |
+
local_dir=d
|
| 57 |
+
)
|
| 58 |
+
# 将本地文件路径直接输出给 shell
|
| 59 |
+
print(path)
|
| 60 |
|
| 61 |
def main():
|
| 62 |
+
p = argparse.ArgumentParser()
|
| 63 |
+
sub = p.add_subparsers(dest="cmd", required=True)
|
| 64 |
+
|
| 65 |
+
up = sub.add_parser("upload")
|
| 66 |
+
up.add_argument("--token", required=True)
|
| 67 |
+
up.add_argument("--repo", required=True)
|
| 68 |
+
up.add_argument("--file", required=True)
|
| 69 |
+
up.add_argument("--prefix", required=True)
|
| 70 |
+
up.add_argument("--max", type=int, default=10)
|
| 71 |
+
up.set_defaults(func=upload)
|
| 72 |
+
|
| 73 |
+
rs = sub.add_parser("restore")
|
| 74 |
+
rs.add_argument("--token", required=True)
|
| 75 |
+
rs.add_argument("--repo", required=True)
|
| 76 |
+
rs.add_argument("--prefix", required=True)
|
| 77 |
+
rs.set_defaults(func=restore)
|
| 78 |
+
|
| 79 |
+
args = p.parse_args()
|
| 80 |
+
try:
|
| 81 |
+
args.func(args)
|
| 82 |
+
except KeyboardInterrupt:
|
| 83 |
+
sys.exit(130)
|
| 84 |
|
| 85 |
if __name__ == "__main__":
|
| 86 |
main()
|