Spaces:
Running
Running
Create update_music.py
Browse files- update_music.py +99 -0
update_music.py
ADDED
@@ -0,0 +1,99 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
#!/usr/bin/env python3
|
2 |
+
from huggingface_hub import snapshot_download, HfApi
|
3 |
+
import os
|
4 |
+
import sys
|
5 |
+
import time
|
6 |
+
import json
|
7 |
+
import hashlib
|
8 |
+
from datetime import datetime
|
9 |
+
|
10 |
+
def get_dataset_info(repo_id, token):
|
11 |
+
"""获取数据集的最新信息,用于检测更新"""
|
12 |
+
try:
|
13 |
+
api = HfApi(token=token)
|
14 |
+
info = api.repo_info(repo_id=repo_id, repo_type="dataset")
|
15 |
+
return {
|
16 |
+
"sha": info.sha,
|
17 |
+
"last_modified": info.last_modified.isoformat() if info.last_modified else None
|
18 |
+
}
|
19 |
+
except Exception as e:
|
20 |
+
print(f"获取数据集信息出错: {str(e)}")
|
21 |
+
return None
|
22 |
+
|
23 |
+
def save_dataset_info(info, music_dir):
|
24 |
+
"""保存数据集信息到本地文件"""
|
25 |
+
info_file = os.path.join(music_dir, ".dataset_info.json")
|
26 |
+
try:
|
27 |
+
with open(info_file, "w") as f:
|
28 |
+
json.dump(info, f)
|
29 |
+
except Exception as e:
|
30 |
+
print(f"保存数据集信息出错: {str(e)}")
|
31 |
+
|
32 |
+
def load_dataset_info(music_dir):
|
33 |
+
"""从本地文件加载数据集信息"""
|
34 |
+
info_file = os.path.join(music_dir, ".dataset_info.json")
|
35 |
+
if not os.path.exists(info_file):
|
36 |
+
return None
|
37 |
+
|
38 |
+
try:
|
39 |
+
with open(info_file, "r") as f:
|
40 |
+
return json.load(f)
|
41 |
+
except Exception as e:
|
42 |
+
print(f"加载数据集信息出错: {str(e)}")
|
43 |
+
return None
|
44 |
+
|
45 |
+
def update_music(dataset_name, token, music_dir, force=False):
|
46 |
+
"""更新音乐文件,只在有变化时更新"""
|
47 |
+
print(f"[{datetime.now()}] 检查音乐数据集更新...")
|
48 |
+
|
49 |
+
# 获取远程数据集信息
|
50 |
+
remote_info = get_dataset_info(dataset_name, token)
|
51 |
+
if not remote_info:
|
52 |
+
print("无法获取远程数据集信息,跳过更新")
|
53 |
+
return False
|
54 |
+
|
55 |
+
# 获取本地数据集信息
|
56 |
+
local_info = load_dataset_info(music_dir)
|
57 |
+
|
58 |
+
# 检查是否需要更新
|
59 |
+
if not force and local_info and local_info.get("sha") == remote_info.get("sha"):
|
60 |
+
print("音乐数据集没有变化,无需更新")
|
61 |
+
return False
|
62 |
+
|
63 |
+
print(f"检测到音乐数据集有更新,开始下载...")
|
64 |
+
|
65 |
+
try:
|
66 |
+
# 下载数据集
|
67 |
+
snapshot_download(
|
68 |
+
repo_id=dataset_name,
|
69 |
+
repo_type="dataset",
|
70 |
+
local_dir=music_dir,
|
71 |
+
token=token
|
72 |
+
)
|
73 |
+
|
74 |
+
# 保存新的数据集信息
|
75 |
+
save_dataset_info(remote_info, music_dir)
|
76 |
+
|
77 |
+
print(f"[{datetime.now()}] 音乐数据集更新成功!")
|
78 |
+
return True
|
79 |
+
except Exception as e:
|
80 |
+
print(f"更新音乐数据集出错: {str(e)}")
|
81 |
+
return False
|
82 |
+
|
83 |
+
if __name__ == "__main__":
|
84 |
+
# 命令行参数: dataset_name token music_dir [interval] [force]
|
85 |
+
dataset_name = sys.argv[1]
|
86 |
+
token = sys.argv[2]
|
87 |
+
music_dir = sys.argv[3]
|
88 |
+
|
89 |
+
# 可选参数
|
90 |
+
interval = int(sys.argv[4]) if len(sys.argv) > 4 else 3600 # 默认1小时
|
91 |
+
force = sys.argv[5].lower() == "true" if len(sys.argv) > 5 else False
|
92 |
+
|
93 |
+
# 第一次运行时强制更新
|
94 |
+
update_music(dataset_name, token, music_dir, force=True)
|
95 |
+
|
96 |
+
# 定期检查更新
|
97 |
+
while True:
|
98 |
+
time.sleep(interval)
|
99 |
+
update_music(dataset_name, token, music_dir, force=force)
|