Update sync_data.sh
Browse files- sync_data.sh +22 -15
sync_data.sh
CHANGED
|
@@ -1,13 +1,5 @@
|
|
| 1 |
#!/bin/bash
|
| 2 |
|
| 3 |
-
# 检查Tunnel
|
| 4 |
-
check_tunnel() {
|
| 5 |
-
while ! curl -s http://localhost:8090 >/dev/null; do
|
| 6 |
-
echo "Waiting for Cloudflare Tunnel to be ready..."
|
| 7 |
-
sleep 5
|
| 8 |
-
done
|
| 9 |
-
}
|
| 10 |
-
|
| 11 |
# 检查环境变量
|
| 12 |
if [[ -z "$HF_TOKEN" ]] || [[ -z "$DATASET_ID" ]]; then
|
| 13 |
echo "Starting without backup functionality - missing HF_TOKEN or DATASET_ID"
|
|
@@ -20,17 +12,30 @@ source /opt/venv/bin/activate
|
|
| 20 |
|
| 21 |
# Python 函数: 上传备份
|
| 22 |
upload_backup() {
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
|
| 28 |
-
|
| 29 |
from huggingface_hub import HfApi
|
| 30 |
import sys
|
| 31 |
import os
|
| 32 |
-
|
| 33 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
api = HfApi(token='$token')
|
| 35 |
try:
|
| 36 |
api.upload_file(
|
|
@@ -40,6 +45,8 @@ try:
|
|
| 40 |
repo_type='dataset'
|
| 41 |
)
|
| 42 |
print(f'Successfully uploaded $file_name')
|
|
|
|
|
|
|
| 43 |
except Exception as e:
|
| 44 |
print(f'Error uploading file: {str(e)}')
|
| 45 |
"
|
|
|
|
| 1 |
#!/bin/bash
|
| 2 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
# 检查环境变量
|
| 4 |
if [[ -z "$HF_TOKEN" ]] || [[ -z "$DATASET_ID" ]]; then
|
| 5 |
echo "Starting without backup functionality - missing HF_TOKEN or DATASET_ID"
|
|
|
|
| 12 |
|
| 13 |
# Python 函数: 上传备份
|
| 14 |
upload_backup() {
|
| 15 |
+
file_path="$1"
|
| 16 |
+
file_name="$2"
|
| 17 |
+
token="$HF_TOKEN"
|
| 18 |
+
repo_id="$DATASET_ID"
|
| 19 |
|
| 20 |
+
python3 -c "
|
| 21 |
from huggingface_hub import HfApi
|
| 22 |
import sys
|
| 23 |
import os
|
| 24 |
+
|
| 25 |
+
def manage_backups(api, repo_id, max_files=50):
|
| 26 |
+
files = api.list_repo_files(repo_id=repo_id, repo_type='dataset')
|
| 27 |
+
backup_files = [f for f in files if f.startswith('halo_backup_') and f.endswith('.tar.gz')]
|
| 28 |
+
backup_files.sort()
|
| 29 |
+
|
| 30 |
+
if len(backup_files) >= max_files:
|
| 31 |
+
files_to_delete = backup_files[:(len(backup_files) - max_files + 1)]
|
| 32 |
+
for file_to_delete in files_to_delete:
|
| 33 |
+
try:
|
| 34 |
+
api.delete_file(path_in_repo=file_to_delete, repo_id=repo_id, repo_type='dataset')
|
| 35 |
+
print(f'Deleted old backup: {file_to_delete}')
|
| 36 |
+
except Exception as e:
|
| 37 |
+
print(f'Error deleting {file_to_delete}: {str(e)}')
|
| 38 |
+
|
| 39 |
api = HfApi(token='$token')
|
| 40 |
try:
|
| 41 |
api.upload_file(
|
|
|
|
| 45 |
repo_type='dataset'
|
| 46 |
)
|
| 47 |
print(f'Successfully uploaded $file_name')
|
| 48 |
+
|
| 49 |
+
manage_backups(api, '$repo_id')
|
| 50 |
except Exception as e:
|
| 51 |
print(f'Error uploading file: {str(e)}')
|
| 52 |
"
|