Add app\core\hf_sync.py
Browse files- app//core//hf_sync.py +76 -0
app//core//hf_sync.py
ADDED
|
@@ -0,0 +1,76 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Romeo AI Research Assistant - High-Stability Sync Service
|
| 2 |
+
# Version: 2026.03.15
|
| 3 |
+
|
| 4 |
+
import os
|
| 5 |
+
import fcntl
|
| 6 |
+
import logging
|
| 7 |
+
from datetime import datetime
|
| 8 |
+
from huggingface_hub import hf_hub_download, HfApi
|
| 9 |
+
from apscheduler.schedulers.background import BackgroundScheduler
|
| 10 |
+
from app.core.config import settings
|
| 11 |
+
|
| 12 |
+
logger = logging.getLogger("romeo_sync")
|
| 13 |
+
api = HfApi()
|
| 14 |
+
scheduler = BackgroundScheduler()
|
| 15 |
+
|
| 16 |
+
# Configuration
|
| 17 |
+
HF_TOKEN = settings.HF_TOKEN
|
| 18 |
+
REPO_ID = settings.HF_DATASET_REPO
|
| 19 |
+
DB_NAME = "romeo_research.db"
|
| 20 |
+
LOCAL_DATA_DIR = "./data"
|
| 21 |
+
LOCAL_PATH = os.path.join(LOCAL_DATA_DIR, DB_NAME)
|
| 22 |
+
|
| 23 |
+
def download_db_from_hf():
|
| 24 |
+
"""Startup: Syncs DB with local directory creation."""
|
| 25 |
+
os.makedirs(LOCAL_DATA_DIR, exist_ok=True)
|
| 26 |
+
|
| 27 |
+
if not REPO_ID or not HF_TOKEN:
|
| 28 |
+
logger.info("Running in local-only mode (no HF sync variables found)")
|
| 29 |
+
return
|
| 30 |
+
|
| 31 |
+
try:
|
| 32 |
+
logger.info(f"Downloading {DB_NAME} from {REPO_ID}...")
|
| 33 |
+
hf_hub_download(
|
| 34 |
+
repo_id=REPO_ID,
|
| 35 |
+
filename=DB_NAME,
|
| 36 |
+
repo_type="dataset",
|
| 37 |
+
token=HF_TOKEN,
|
| 38 |
+
local_dir=LOCAL_DATA_DIR
|
| 39 |
+
)
|
| 40 |
+
logger.info("Database successfully synchronized.")
|
| 41 |
+
except Exception as e:
|
| 42 |
+
logger.warning(f"No existing DB found on HF (First Run): {e}")
|
| 43 |
+
|
| 44 |
+
def backup_db_to_hf():
|
| 45 |
+
"""Uploads with file locking to prevent corruption during active writes."""
|
| 46 |
+
if not REPO_ID or not HF_TOKEN or not os.path.exists(LOCAL_PATH):
|
| 47 |
+
return
|
| 48 |
+
|
| 49 |
+
try:
|
| 50 |
+
# Lock file during read/upload to prevent SQLite 'Database Disk Image is Malformed' errors
|
| 51 |
+
with open(LOCAL_PATH, 'rb') as f:
|
| 52 |
+
fcntl.flock(f, fcntl.LOCK_SH) # Shared lock for reading
|
| 53 |
+
api.upload_file(
|
| 54 |
+
path_or_fileobj=LOCAL_PATH,
|
| 55 |
+
path_in_repo=DB_NAME,
|
| 56 |
+
repo_id=REPO_ID,
|
| 57 |
+
repo_type="dataset",
|
| 58 |
+
token=HF_TOKEN,
|
| 59 |
+
commit_message=f"Romeo AI Backup: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}"
|
| 60 |
+
)
|
| 61 |
+
fcntl.flock(f, fcntl.LOCK_UN) # Unlock
|
| 62 |
+
logger.info("HF Backup completed successfully.")
|
| 63 |
+
except Exception as e:
|
| 64 |
+
logger.error(f"Backup failed: {e}")
|
| 65 |
+
|
| 66 |
+
def start_backup_scheduler():
|
| 67 |
+
"""Initialize the 5-minute interval backup."""
|
| 68 |
+
if HF_TOKEN and REPO_ID:
|
| 69 |
+
scheduler.add_job(backup_db_to_hf, 'interval', minutes=5)
|
| 70 |
+
scheduler.start()
|
| 71 |
+
logger.info("HF backup scheduler started (5min interval)")
|
| 72 |
+
|
| 73 |
+
def stop_backup_scheduler():
|
| 74 |
+
"""Graceful shutdown for the scheduler."""
|
| 75 |
+
if scheduler.running:
|
| 76 |
+
scheduler.shutdown()
|