Files
geMoldInsight/scripts/migrations/backfill_mold_cavity_confidence.py
T

82 lines
2.6 KiB
Python
Raw Normal View History

2026-05-02 00:39:04 +08:00
#!/usr/bin/env python3
"""
回填 mold_cavity_data 的多方案可信化字段。
回填字段:
- best_scheme_id
- confidence_score
- is_fallback
- fallback_reason
"""
import asyncio
import json
import sys
from pathlib import Path
project_root = Path(__file__).resolve().parents[2]
src_root = project_root / "src"
sys.path.insert(0, str(project_root))
sys.path.insert(0, str(src_root))
from sqlalchemy import select
from database.database import db_manager
from moldinsight.models import MoldCavityData
2026-05-02 00:39:04 +08:00
from storage.rustfs_storage import rustfs_manager
from config.settings import settings
from services.storage_integration_rustfs import StorageIntegrationService
from utils.logger import get_logger
logger = get_logger(__name__)
async def backfill():
await db_manager.connect()
await rustfs_manager.connect(
endpoint=settings.RUSTFS_ENDPOINT,
access_key=settings.RUSTFS_ACCESS_KEY,
secret_key=settings.RUSTFS_SECRET_KEY,
timeout=settings.RUSTFS_TIMEOUT,
)
summary = {"total": 0, "updated": 0, "failed": 0}
try:
async with db_manager.session() as session:
result = await session.execute(select(MoldCavityData))
rows = result.scalars().all()
summary["total"] = len(rows)
for row in rows:
try:
cavity_bytes = await rustfs_manager.download_file(
file_type="mold_cavities",
object_key=row.detailed_object_key,
)
cavity_json = json.loads(cavity_bytes.decode("utf-8"))
payload = StorageIntegrationService._resolve_best_scheme_payload(cavity_json)
best_scheme = payload.get("best_scheme") or {}
row.best_scheme_id = payload.get("best_scheme_id")
row.confidence_score = best_scheme.get("confidence_score")
row.is_fallback = best_scheme.get("is_fallback")
row.fallback_reason = best_scheme.get("fallback_reason")
summary["updated"] += 1
except Exception as exc:
summary["failed"] += 1
logger.warning(f"回填失败 id={row.id}: {exc}")
await session.commit()
finally:
await rustfs_manager.close()
await db_manager.disconnect()
print("=== Backfill 完成 ===")
print(f"总记录: {summary['total']}")
print(f"成功更新: {summary['updated']}")
print(f"失败: {summary['failed']}")
if __name__ == "__main__":
asyncio.run(backfill())