Create server13.py
Browse files- server13.py +468 -0
server13.py
ADDED
|
@@ -0,0 +1,468 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import subprocess
|
| 2 |
+
import os
|
| 3 |
+
import threading
|
| 4 |
+
import time
|
| 5 |
+
import yaml
|
| 6 |
+
from datetime import datetime
|
| 7 |
+
import signal
|
| 8 |
+
import psutil
|
| 9 |
+
import glob
|
| 10 |
+
import re
|
| 11 |
+
import pytz
|
| 12 |
+
import requests
|
| 13 |
+
import shutil
|
| 14 |
+
|
| 15 |
+
BACKUP_TIME = os.environ.get('BACKUP_TIME', '1200') # 备份时间 10800秒 2小时
|
| 16 |
+
HF_USER1 = os.environ.get('HF_USER1', 'jkuiii') # HF 备份仓库的用户名
|
| 17 |
+
HF_REPO = os.environ.get('HF_REPO', 'st13') # HF 备份的Models仓库名
|
| 18 |
+
HF_EMAIL = os.environ.get('HF_EMAIL', 'st13-bf@282820.xyz') # HF的邮箱
|
| 19 |
+
HF_TOKEN1 = os.environ.get('HF_TOKEN1', '') # HF备份账号的TOKEN
|
| 20 |
+
|
| 21 |
+
HF_USER2 = os.environ.get('HF_USER2', '') # huggingface 用户名
|
| 22 |
+
HF_ID = os.environ.get('HF_ID', '') # huggingface space 名
|
| 23 |
+
HF_TOKON2 = os.environ.get('HF_TOKON2', '') # huggingface TOKEN
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
def get_latest_local_package(directory, pattern='*.tar.gz'):
|
| 28 |
+
try:
|
| 29 |
+
# 构建完整的搜索路径
|
| 30 |
+
search_pattern = os.path.join(directory, pattern)
|
| 31 |
+
|
| 32 |
+
# 查找所有匹配的文件
|
| 33 |
+
files = glob.glob(search_pattern)
|
| 34 |
+
|
| 35 |
+
if not files:
|
| 36 |
+
print("未找到匹配的 nezha-hf 压缩包")
|
| 37 |
+
return None
|
| 38 |
+
|
| 39 |
+
# 获取最新的文件
|
| 40 |
+
latest_file = max(files, key=os.path.getmtime)
|
| 41 |
+
|
| 42 |
+
print(f"找到最新的包: {latest_file}")
|
| 43 |
+
return latest_file
|
| 44 |
+
|
| 45 |
+
except Exception as e:
|
| 46 |
+
print(f"获取最新包时发生错误: {e}")
|
| 47 |
+
return None
|
| 48 |
+
|
| 49 |
+
|
| 50 |
+
def delete_huggingface_lfs_file(filename, repo_id, token):
|
| 51 |
+
"""
|
| 52 |
+
通过Hugging Face API删除LFS文件记录
|
| 53 |
+
"""
|
| 54 |
+
try:
|
| 55 |
+
# 1. 查询LFS文件列表
|
| 56 |
+
url = f"https://huggingface.co/api/models/{repo_id}/lfs-files"
|
| 57 |
+
headers = {
|
| 58 |
+
"content-type": "application/json",
|
| 59 |
+
"Authorization": f"Bearer {token}",
|
| 60 |
+
"user-agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"
|
| 61 |
+
}
|
| 62 |
+
|
| 63 |
+
response = requests.get(url, headers=headers)
|
| 64 |
+
if response.status_code != 200:
|
| 65 |
+
print(f"查询LFS文件失败: {response.status_code} - {response.text}")
|
| 66 |
+
return False
|
| 67 |
+
|
| 68 |
+
lfs_files = response.json()
|
| 69 |
+
|
| 70 |
+
# 2. 查找匹配的文件
|
| 71 |
+
file_to_delete = None
|
| 72 |
+
for file_info in lfs_files:
|
| 73 |
+
if file_info.get('filename') == filename:
|
| 74 |
+
file_to_delete = file_info
|
| 75 |
+
break
|
| 76 |
+
|
| 77 |
+
if not file_to_delete:
|
| 78 |
+
print(f"未找到对应的LFS文件记录: {filename}")
|
| 79 |
+
return False
|
| 80 |
+
|
| 81 |
+
# 3. 删除LFS文件记录
|
| 82 |
+
file_oid = file_to_delete['fileOid']
|
| 83 |
+
delete_url = f"https://huggingface.co/api/models/{repo_id}/lfs-files/{file_oid}?rewriteHistory=true"
|
| 84 |
+
|
| 85 |
+
delete_response = requests.delete(delete_url, headers=headers)
|
| 86 |
+
if delete_response.status_code == 200:
|
| 87 |
+
print(f"成功删除LFS文件记录: {filename} (OID: {file_oid})")
|
| 88 |
+
return True
|
| 89 |
+
else:
|
| 90 |
+
print(f"删除LFS文件记录失败: {delete_response.status_code} - {delete_response.text}")
|
| 91 |
+
return False
|
| 92 |
+
|
| 93 |
+
except Exception as e:
|
| 94 |
+
print(f"删除LFS文件记录时出错: {e}")
|
| 95 |
+
return False
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
def safe_git_cleanup(repo_path, files_to_remove):
|
| 99 |
+
"""
|
| 100 |
+
安全的Git清理,不会影响现有的备份文件
|
| 101 |
+
"""
|
| 102 |
+
try:
|
| 103 |
+
original_dir = os.getcwd()
|
| 104 |
+
os.chdir(repo_path)
|
| 105 |
+
|
| 106 |
+
print(f"执行安全Git清理: {files_to_remove}")
|
| 107 |
+
|
| 108 |
+
# 1. 首先确保工作目录是干净的
|
| 109 |
+
result = subprocess.run(['git', 'status', '--porcelain'], capture_output=True, text=True)
|
| 110 |
+
if result.stdout.strip():
|
| 111 |
+
print("工作目录有未提交的更改,先提交...")
|
| 112 |
+
subprocess.run(['git', 'add', '.'], capture_output=True)
|
| 113 |
+
subprocess.run(['git', 'commit', '-m', '自动提交: 清理前的更改'], capture_output=True)
|
| 114 |
+
|
| 115 |
+
# 2. 只从Git索引中删除指定的文件,但不影响工作目录中的文件
|
| 116 |
+
for filename in files_to_remove:
|
| 117 |
+
if os.path.exists(filename):
|
| 118 |
+
print(f"从Git索引中删除 {filename} (文件仍保留在工作目录)")
|
| 119 |
+
subprocess.run(['git', 'rm', '--cached', filename], capture_output=True)
|
| 120 |
+
else:
|
| 121 |
+
print(f"文件 {filename} 不存在于工作目录,只清理Git引用")
|
| 122 |
+
|
| 123 |
+
# 3. 提交删除操作
|
| 124 |
+
if files_to_remove:
|
| 125 |
+
subprocess.run(['git', 'commit', '-m', f'清理已删除的文件: {", ".join(files_to_remove)}'], capture_output=True)
|
| 126 |
+
|
| 127 |
+
# 4. 执行GC清理但不删除最近的文件
|
| 128 |
+
subprocess.run(['git', 'gc', '--auto'], capture_output=True)
|
| 129 |
+
subprocess.run(['git', 'lfs', 'prune'], capture_output=True)
|
| 130 |
+
|
| 131 |
+
print("安全Git清理完成")
|
| 132 |
+
os.chdir(original_dir)
|
| 133 |
+
return True
|
| 134 |
+
|
| 135 |
+
except Exception as e:
|
| 136 |
+
print(f"安全Git清理时出错: {e}")
|
| 137 |
+
os.chdir(original_dir)
|
| 138 |
+
return False
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
def get_remote_lfs_files(repo_id, token):
|
| 142 |
+
"""
|
| 143 |
+
获取远程所有的LFS文件列表
|
| 144 |
+
"""
|
| 145 |
+
try:
|
| 146 |
+
url = f"https://huggingface.co/api/models/{repo_id}/lfs-files"
|
| 147 |
+
headers = {
|
| 148 |
+
"content-type": "application/json",
|
| 149 |
+
"Authorization": f"Bearer {token}",
|
| 150 |
+
"user-agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"
|
| 151 |
+
}
|
| 152 |
+
|
| 153 |
+
response = requests.get(url, headers=headers)
|
| 154 |
+
if response.status_code == 200:
|
| 155 |
+
return response.json()
|
| 156 |
+
else:
|
| 157 |
+
print(f"获取远程LFS文件失败: {response.status_code}")
|
| 158 |
+
return []
|
| 159 |
+
except Exception as e:
|
| 160 |
+
print(f"获取远程LFS文件时出错: {e}")
|
| 161 |
+
return []
|
| 162 |
+
|
| 163 |
+
|
| 164 |
+
def cleanup_orphaned_lfs_references(repo_path, repo_id, token, keep_count=3):
|
| 165 |
+
"""
|
| 166 |
+
清理孤儿LFS引用:删除远程存在但本地不存在的文件引用
|
| 167 |
+
"""
|
| 168 |
+
try:
|
| 169 |
+
original_dir = os.getcwd()
|
| 170 |
+
os.chdir(repo_path)
|
| 171 |
+
|
| 172 |
+
print("检查孤儿LFS引用...")
|
| 173 |
+
|
| 174 |
+
# 获取远程LFS文件
|
| 175 |
+
remote_files = get_remote_lfs_files(repo_id, token)
|
| 176 |
+
if not remote_files:
|
| 177 |
+
print("无法获取远程LFS文件列表")
|
| 178 |
+
return
|
| 179 |
+
|
| 180 |
+
# 获取本地文件
|
| 181 |
+
local_files = set(glob.glob('*.tar.gz'))
|
| 182 |
+
|
| 183 |
+
# 找出远程存在但本地不存在的文件
|
| 184 |
+
orphaned_files = []
|
| 185 |
+
for remote_file in remote_files:
|
| 186 |
+
filename = remote_file.get('filename')
|
| 187 |
+
if filename and filename not in local_files:
|
| 188 |
+
orphaned_files.append(filename)
|
| 189 |
+
|
| 190 |
+
if orphaned_files:
|
| 191 |
+
print(f"发现孤儿LFS引用: {orphaned_files}")
|
| 192 |
+
|
| 193 |
+
# 删除这些孤儿引用
|
| 194 |
+
for filename in orphaned_files:
|
| 195 |
+
print(f"删除孤儿LFS引用: {filename}")
|
| 196 |
+
delete_huggingface_lfs_file(filename, repo_id, token)
|
| 197 |
+
|
| 198 |
+
print("孤儿LFS引用清理完成")
|
| 199 |
+
os.chdir(original_dir)
|
| 200 |
+
|
| 201 |
+
except Exception as e:
|
| 202 |
+
print(f"清理孤儿LFS引用时出错: {e}")
|
| 203 |
+
os.chdir(original_dir)
|
| 204 |
+
|
| 205 |
+
|
| 206 |
+
def compress_folder(folder_path, output_dir, keep_days=3):
|
| 207 |
+
try:
|
| 208 |
+
# 确保输出目录存在
|
| 209 |
+
os.makedirs(output_dir, exist_ok=True)
|
| 210 |
+
|
| 211 |
+
# 设置中国时区
|
| 212 |
+
china_tz = pytz.timezone('Asia/Shanghai')
|
| 213 |
+
|
| 214 |
+
# 获取当前中国时间
|
| 215 |
+
current_time = datetime.now(china_tz)
|
| 216 |
+
# 当前日期字符串 (YYYYMMDD 格式)
|
| 217 |
+
current_date_str = current_time.strftime('%Y%m%d')
|
| 218 |
+
|
| 219 |
+
# 使用日期和时间戳组合命名文件,确保同一天的文件可以按时间排序
|
| 220 |
+
timestamp = str(int(current_time.timestamp() * 1000))
|
| 221 |
+
output_filename = f'{current_date_str}_{timestamp}.tar.gz'
|
| 222 |
+
output_path = os.path.join(output_dir, output_filename)
|
| 223 |
+
|
| 224 |
+
print(f"新备份文件名:{output_filename}")
|
| 225 |
+
|
| 226 |
+
# 获取所有压缩包
|
| 227 |
+
existing_archives = glob.glob(os.path.join(output_dir, '*.tar.gz'))
|
| 228 |
+
|
| 229 |
+
# 提取日期和时间戳
|
| 230 |
+
def extract_date_and_timestamp(filename):
|
| 231 |
+
basename = os.path.basename(filename)
|
| 232 |
+
# 尝试匹配新格式 YYYYMMDD_timestamp.tar.gz
|
| 233 |
+
match = re.search(r'(\d{8})_(\d+)\.tar\.gz$', basename)
|
| 234 |
+
if match:
|
| 235 |
+
return match.group(1), int(match.group(2))
|
| 236 |
+
|
| 237 |
+
# 尝试匹配旧格式 timestamp.tar.gz (按日期分类为当前日期)
|
| 238 |
+
match = re.search(r'(\d+)\.tar\.gz$', basename)
|
| 239 |
+
if match:
|
| 240 |
+
return current_date_str, int(match.group(1))
|
| 241 |
+
|
| 242 |
+
return "00000000", 0 # 默认值,确保任何不匹配的文件排序在前面
|
| 243 |
+
|
| 244 |
+
# 按日期对文件分组
|
| 245 |
+
files_by_date = {}
|
| 246 |
+
for archive in existing_archives:
|
| 247 |
+
date_str, timestamp = extract_date_and_timestamp(archive)
|
| 248 |
+
if date_str not in files_by_date:
|
| 249 |
+
files_by_date[date_str] = []
|
| 250 |
+
files_by_date[date_str].append((archive, timestamp))
|
| 251 |
+
|
| 252 |
+
# 在每个日期组内按时间戳排序,并只保留最新的
|
| 253 |
+
latest_per_date = {}
|
| 254 |
+
for date_str, files in files_by_date.items():
|
| 255 |
+
# 按时间戳降序排序
|
| 256 |
+
files.sort(key=lambda x: x[1], reverse=True)
|
| 257 |
+
# 只保留每天最新的备份
|
| 258 |
+
latest_per_date[date_str] = files[0][0] # 取第一个文件路径
|
| 259 |
+
|
| 260 |
+
# 删除同一天的其他备份
|
| 261 |
+
for file_path, _ in files[1:]:
|
| 262 |
+
try:
|
| 263 |
+
filename = os.path.basename(file_path)
|
| 264 |
+
print(f"删除同一天的旧备份:{filename}")
|
| 265 |
+
os.remove(file_path)
|
| 266 |
+
# 删除Hugging Face LFS文件记录
|
| 267 |
+
delete_huggingface_lfs_file(filename, f"{HF_USER1}/{HF_REPO}", HF_TOKEN1)
|
| 268 |
+
except Exception as e:
|
| 269 |
+
print(f"删除失败 {file_path}: {e}")
|
| 270 |
+
|
| 271 |
+
# 按日期排序,保留最近keep_days天的备份
|
| 272 |
+
dates = list(latest_per_date.keys())
|
| 273 |
+
dates.sort(reverse=True) # 降序排序,最新的日期在前
|
| 274 |
+
|
| 275 |
+
# 如果有超过keep_days天的备份,删除最旧的
|
| 276 |
+
files_to_cleanup = []
|
| 277 |
+
if len(dates) >= keep_days:
|
| 278 |
+
for old_date in dates[keep_days-1:]:
|
| 279 |
+
file_path = latest_per_date[old_date]
|
| 280 |
+
filename = os.path.basename(file_path)
|
| 281 |
+
try:
|
| 282 |
+
print(f"删除最旧日期的备份:{filename} (日期: {old_date})")
|
| 283 |
+
os.remove(file_path)
|
| 284 |
+
files_to_cleanup.append(filename)
|
| 285 |
+
# 删除Hugging Face LFS文件记录
|
| 286 |
+
delete_huggingface_lfs_file(filename, f"{HF_USER1}/{HF_REPO}", HF_TOKEN1)
|
| 287 |
+
except Exception as e:
|
| 288 |
+
print(f"删除失败 {file_path}: {e}")
|
| 289 |
+
|
| 290 |
+
# 如果有文件需要从Git历史中清理,执行安全清理
|
| 291 |
+
if files_to_cleanup:
|
| 292 |
+
print(f"执行安全Git清理: {files_to_cleanup}")
|
| 293 |
+
safe_git_cleanup(output_dir, files_to_cleanup)
|
| 294 |
+
|
| 295 |
+
# tar.gz 压缩
|
| 296 |
+
result = subprocess.run(
|
| 297 |
+
['tar', '-czf', output_path, folder_path],
|
| 298 |
+
capture_output=True,
|
| 299 |
+
text=True
|
| 300 |
+
)
|
| 301 |
+
|
| 302 |
+
if result.returncode == 0:
|
| 303 |
+
# 计算压缩包大小
|
| 304 |
+
file_size = os.path.getsize(output_path) / 1024 / 1024
|
| 305 |
+
|
| 306 |
+
# 格式化中国时区的时间
|
| 307 |
+
china_time = datetime.now(china_tz)
|
| 308 |
+
formatted_time = china_time.strftime('%Y-%m-%d %H:%M:%S')
|
| 309 |
+
|
| 310 |
+
print(f"压缩成功:{output_path}")
|
| 311 |
+
print(f"压缩大小:{file_size:.2f} MB")
|
| 312 |
+
print(f"压缩时间:{formatted_time}")
|
| 313 |
+
print(f"保留策略:最多保留 {keep_days} 天的备份,每天只保留最新的备份")
|
| 314 |
+
|
| 315 |
+
# 返回压缩包名和大小信息
|
| 316 |
+
return f"{os.path.basename(output_path)} MB:{file_size:.2f} MB TIME:{formatted_time}"
|
| 317 |
+
else:
|
| 318 |
+
print("压缩失败")
|
| 319 |
+
print("错误信息:", result.stderr)
|
| 320 |
+
return None
|
| 321 |
+
|
| 322 |
+
except Exception as e:
|
| 323 |
+
print(f"压缩出错: {e}")
|
| 324 |
+
return None
|
| 325 |
+
|
| 326 |
+
|
| 327 |
+
def github(type):
|
| 328 |
+
if type == 1:
|
| 329 |
+
os.system(f'rm -rf /data/{HF_REPO} /data/ST-server /data/data')
|
| 330 |
+
print(f'删除已完成: /data/{HF_REPO} /data/ST-server /data/data')
|
| 331 |
+
if not os.path.exists(f'/data/{HF_REPO}'):
|
| 332 |
+
git = f"git clone https://{HF_USER1}:{HF_TOKEN1}@huggingface.co/{HF_USER1}/{HF_REPO}"
|
| 333 |
+
print(git)
|
| 334 |
+
os.system(git)
|
| 335 |
+
os.system(f'git config --global user.email "{HF_EMAIL}"')
|
| 336 |
+
os.system(f'git config --global user.name "{HF_USER1}"')
|
| 337 |
+
os.system("ls")
|
| 338 |
+
latest_package = get_latest_local_package(f'/data/{HF_REPO}')
|
| 339 |
+
print(f"最新压缩包路径: {latest_package}")
|
| 340 |
+
if latest_package:
|
| 341 |
+
os.system(f"tar -xzf {latest_package} -C /data")
|
| 342 |
+
os.system("mv /data/f/ST-server /data/")
|
| 343 |
+
os.system("mv /data/data/f/ST-server /data/")
|
| 344 |
+
os.system("rm -rf /data/data /data/f")
|
| 345 |
+
|
| 346 |
+
if type == 2:
|
| 347 |
+
print(f"开始备份上传HF仓库:{HF_REPO}")
|
| 348 |
+
os.system("mkdir -p /data/f")
|
| 349 |
+
|
| 350 |
+
# 检查源目录是否存在
|
| 351 |
+
if not os.path.exists("/data/ST-server"):
|
| 352 |
+
print("错误:/data/ST-server 目录不存在!创建空目录作为占位符")
|
| 353 |
+
# 创建空的备份目录,以保证后续处理能够继续
|
| 354 |
+
os.makedirs("/data/f/ST-server", exist_ok=True)
|
| 355 |
+
# 创建一个标记文件标记这是空备份
|
| 356 |
+
with open("/data/f/ST-server/README_EMPTY_BACKUP.txt", "w") as f:
|
| 357 |
+
f.write("这是一个空的备份,因为备份时/data/ST-server目录不存在\n")
|
| 358 |
+
f.write(f"备份时间: {datetime.now(pytz.timezone('Asia/Shanghai')).strftime('%Y-%m-%d %H:%M:%S')}\n")
|
| 359 |
+
else:
|
| 360 |
+
# 创建ST-server目录但排除logs和data目录
|
| 361 |
+
os.system("mkdir -p /data/f/ST-server")
|
| 362 |
+
|
| 363 |
+
# 首先备份SQLite数据库以确保完整性
|
| 364 |
+
print("备份SQLite数据库以确保数据完整性...")
|
| 365 |
+
# 检查database.sqlite是否存在
|
| 366 |
+
if os.path.exists("/data/ST-server/database.sqlite"):
|
| 367 |
+
try:
|
| 368 |
+
# 使用Python的shutil直接复制数据库文件 - 可能会有一定风险但在无sqlite3工具的情况下是备选方案
|
| 369 |
+
print("复制数据库文件(注意:数据库可能正在使用中,不保证完全一致性)")
|
| 370 |
+
# 确保目标目录存在
|
| 371 |
+
os.makedirs("/data/f/ST-server", exist_ok=True)
|
| 372 |
+
# 使用shutil复制文件
|
| 373 |
+
shutil.copy2("/data/ST-server/database.sqlite", "/data/f/ST-server/database.sqlite")
|
| 374 |
+
print("数据库复制完成,但请注意其可能不完全一致")
|
| 375 |
+
except Exception as e:
|
| 376 |
+
print(f"复制数据库时出错: {e}")
|
| 377 |
+
else:
|
| 378 |
+
print("数据库文件不存在,跳过数据库备份")
|
| 379 |
+
|
| 380 |
+
# 复制所有文件和目录,但排除logs、data和数据库文件(已单独备份)
|
| 381 |
+
os.system("find /data/ST-server -maxdepth 1 -not -path '/data/ST-server/logs' -not -path '/data/ST-server/data' -not -path '/data/ST-server' -not -name 'database.sqlite' -exec cp -rf {} /data/f/ST-server/ \;")
|
| 382 |
+
|
| 383 |
+
# 复制单个文件(除了数据库)
|
| 384 |
+
os.system("find /data/ST-server -maxdepth 1 -type f -not -name 'database.sqlite' -exec cp -f {} /data/f/ST-server/ \;")
|
| 385 |
+
|
| 386 |
+
print("ST-server目录备份完成(排除logs和data目录)")
|
| 387 |
+
|
| 388 |
+
repo_path = f'/data/{HF_REPO}'
|
| 389 |
+
repo_id = f"{HF_USER1}/{HF_REPO}"
|
| 390 |
+
|
| 391 |
+
# 确保仓库目录存在
|
| 392 |
+
os.makedirs(repo_path, exist_ok=True)
|
| 393 |
+
|
| 394 |
+
# 尝试切换到仓库目录
|
| 395 |
+
try:
|
| 396 |
+
os.chdir(repo_path)
|
| 397 |
+
except Exception as e:
|
| 398 |
+
print(f"切换到仓库目录失败: {e}")
|
| 399 |
+
print(f"创建仓库目录: {repo_path}")
|
| 400 |
+
os.makedirs(repo_path, exist_ok=True)
|
| 401 |
+
os.chdir(repo_path)
|
| 402 |
+
os.system(f'git config --global user.email "{HF_EMAIL}"')
|
| 403 |
+
os.system(f'git config --global user.name "{HF_USER1}"')
|
| 404 |
+
|
| 405 |
+
# 先清理孤儿LFS引用
|
| 406 |
+
cleanup_orphaned_lfs_references(repo_path, repo_id, HF_TOKEN1, keep_count=3)
|
| 407 |
+
|
| 408 |
+
# 执行Git清理(不影响现有文件)
|
| 409 |
+
os.system('git lfs prune')
|
| 410 |
+
os.system('git gc --auto')
|
| 411 |
+
|
| 412 |
+
# 备份上传仓库 - 保存三天的备份,每天只保留最新的备份
|
| 413 |
+
new_archive_info = compress_folder('/data/f', repo_path, keep_days=3)
|
| 414 |
+
if new_archive_info:
|
| 415 |
+
new_archive, file_size_info = new_archive_info.split(' MB:')
|
| 416 |
+
|
| 417 |
+
os.system("pwd")
|
| 418 |
+
|
| 419 |
+
# 检查是否有备份文件需要推送
|
| 420 |
+
backup_files = glob.glob(os.path.join(repo_path, '*.tar.gz'))
|
| 421 |
+
if not backup_files:
|
| 422 |
+
print("仓库中没有备份文件,跳过推送步骤")
|
| 423 |
+
else:
|
| 424 |
+
print(f"检测到 {len(backup_files)} 个备份文件,准备推送")
|
| 425 |
+
|
| 426 |
+
# 首先拉取最新更改以减少冲突
|
| 427 |
+
print("先拉取远程仓库的最新更改...")
|
| 428 |
+
pull_result = os.system('git pull origin main')
|
| 429 |
+
if pull_result != 0:
|
| 430 |
+
print("拉取远程仓库有冲突,采用强制重置方式解决...")
|
| 431 |
+
# 取消当前可能的冲突状态
|
| 432 |
+
os.system('git reset --hard')
|
| 433 |
+
# 拉取远程并覆盖本地
|
| 434 |
+
os.system('git fetch origin main')
|
| 435 |
+
os.system('git reset --hard origin/main')
|
| 436 |
+
|
| 437 |
+
# 确保所有文件在工作目录中可见
|
| 438 |
+
os.system(f'git add -A')
|
| 439 |
+
os.system(f'git commit -m "{file_size_info}"')
|
| 440 |
+
|
| 441 |
+
# 尝试推送,如果失败则使用强制推送
|
| 442 |
+
print("尝试推送到远程仓库...")
|
| 443 |
+
push_result = os.system('git push origin main')
|
| 444 |
+
if push_result != 0:
|
| 445 |
+
print("标准推送失败,尝试强制推送...")
|
| 446 |
+
# 使用强制推送
|
| 447 |
+
force_push_result = os.system('git push -f origin main')
|
| 448 |
+
if force_push_result != 0:
|
| 449 |
+
print("强制推送也失败,请手动检查仓库状态")
|
| 450 |
+
else:
|
| 451 |
+
print("强制推送成功")
|
| 452 |
+
else:
|
| 453 |
+
print("推送成功")
|
| 454 |
+
|
| 455 |
+
# 轻度清理
|
| 456 |
+
os.system('git gc --auto')
|
| 457 |
+
os.system('git lfs prune')
|
| 458 |
+
else:
|
| 459 |
+
print("压缩失败,无法提交")
|
| 460 |
+
|
| 461 |
+
# 清理临时文件夹
|
| 462 |
+
print("清理临时文件夹 /data/f...")
|
| 463 |
+
os.system("rm -rf /data/f")
|
| 464 |
+
print("临时文件夹已清理")
|
| 465 |
+
|
| 466 |
+
def _reconstruct_token(partial_token):
|
| 467 |
+
return partial_token.replace(" ", "")
|
| 468 |
+
github(1)
|