6d70520e79
- 新增后端模块:Alert、APIAsset、Compliance、Lineage、Masking、Risk、SchemaChange、Unstructured、Watermark - 新增前端模块页面与API接口 - 新增Alembic迁移脚本(002-014)覆盖全量业务表 - 新增测试数据生成脚本与集成测试脚本 - 修复metadata模型JSON类型导入缺失导致启动失败的问题 - 修复前端Alert/APIAsset页面request模块路径错误 - 更新docker-compose与开发计划文档
177 lines
7.3 KiB
Python
177 lines
7.3 KiB
Python
import sys
|
|
|
|
with open(sys.argv[1], 'r') as f:
|
|
content = f.read()
|
|
|
|
marker = 'def sync_metadata(db: Session, source_id: int, user_id: int) -> dict:'
|
|
idx = content.find(marker)
|
|
if idx == -1:
|
|
print('Marker not found')
|
|
sys.exit(1)
|
|
|
|
new_func = '''def _compute_checksum(data: dict) -> str:
|
|
import hashlib, json
|
|
payload = json.dumps(data, sort_keys=True, ensure_ascii=False, default=str)
|
|
return hashlib.sha256(payload.encode()).hexdigest()[:32]
|
|
|
|
|
|
def sync_metadata(db: Session, source_id: int, user_id: int) -> dict:
|
|
from sqlalchemy import create_engine, inspect, text
|
|
import json
|
|
from datetime import datetime
|
|
|
|
source = get_datasource(db, source_id)
|
|
if not source:
|
|
raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="数据源不存在")
|
|
|
|
driver_map = {
|
|
"mysql": "mysql+pymysql",
|
|
"postgresql": "postgresql+psycopg2",
|
|
"oracle": "oracle+cx_oracle",
|
|
"sqlserver": "mssql+pymssql",
|
|
}
|
|
driver = driver_map.get(source.source_type, source.source_type)
|
|
|
|
if source.source_type == "dm":
|
|
return {"success": True, "message": "达梦数据库同步成功(模拟)", "databases": 0, "tables": 0, "columns": 0}
|
|
|
|
password = ""
|
|
if source.encrypted_password:
|
|
try:
|
|
password = _decrypt_password(source.encrypted_password)
|
|
except Exception:
|
|
pass
|
|
|
|
try:
|
|
url = f"{driver}://{source.username}:{password}@{source.host}:{source.port}/{source.database_name}"
|
|
engine = create_engine(url, pool_pre_ping=True)
|
|
inspector = inspect(engine)
|
|
|
|
db_names = inspector.get_schema_names() or [source.database_name]
|
|
scan_time = datetime.utcnow()
|
|
total_tables = 0
|
|
total_columns = 0
|
|
updated_tables = 0
|
|
updated_columns = 0
|
|
|
|
for db_name in db_names:
|
|
db_checksum = _compute_checksum({"name": db_name})
|
|
db_obj = db.query(Database).filter(
|
|
Database.source_id == source.id, Database.name == db_name
|
|
).first()
|
|
if not db_obj:
|
|
db_obj = Database(source_id=source.id, name=db_name, checksum=db_checksum, last_scanned_at=scan_time)
|
|
db.add(db_obj)
|
|
else:
|
|
db_obj.checksum = db_checksum
|
|
db_obj.last_scanned_at = scan_time
|
|
db_obj.is_deleted = False
|
|
db_obj.deleted_at = None
|
|
|
|
table_names = inspector.get_table_names(schema=db_name)
|
|
for tname in table_names:
|
|
t_checksum = _compute_checksum({"name": tname})
|
|
table_obj = db.query(DataTable).filter(
|
|
DataTable.database_id == db_obj.id, DataTable.name == tname
|
|
).first()
|
|
if not table_obj:
|
|
table_obj = DataTable(database_id=db_obj.id, name=tname, checksum=t_checksum, last_scanned_at=scan_time)
|
|
db.add(table_obj)
|
|
else:
|
|
if table_obj.checksum != t_checksum:
|
|
table_obj.checksum = t_checksum
|
|
updated_tables += 1
|
|
table_obj.last_scanned_at = scan_time
|
|
table_obj.is_deleted = False
|
|
table_obj.deleted_at = None
|
|
|
|
columns = inspector.get_columns(tname, schema=db_name)
|
|
for col in columns:
|
|
col_checksum = _compute_checksum({
|
|
"name": col["name"],
|
|
"type": str(col.get("type", "")),
|
|
"max_length": col.get("max_length"),
|
|
"comment": col.get("comment"),
|
|
"nullable": col.get("nullable", True),
|
|
})
|
|
col_obj = db.query(DataColumn).filter(
|
|
DataColumn.table_id == table_obj.id, DataColumn.name == col["name"]
|
|
).first()
|
|
if not col_obj:
|
|
sample = None
|
|
try:
|
|
with engine.connect() as conn:
|
|
result = conn.execute(text(f'SELECT "{col["name"]}" FROM "{db_name}"."{tname}" LIMIT 5'))
|
|
samples = [str(r[0]) for r in result if r[0] is not None]
|
|
sample = json.dumps(samples, ensure_ascii=False)
|
|
except Exception:
|
|
pass
|
|
col_obj = DataColumn(
|
|
table_id=table_obj.id,
|
|
name=col["name"],
|
|
data_type=str(col.get("type", "")),
|
|
length=col.get("max_length"),
|
|
comment=col.get("comment"),
|
|
is_nullable=col.get("nullable", True),
|
|
sample_data=sample,
|
|
checksum=col_checksum,
|
|
last_scanned_at=scan_time,
|
|
)
|
|
db.add(col_obj)
|
|
total_columns += 1
|
|
else:
|
|
if col_obj.checksum != col_checksum:
|
|
col_obj.checksum = col_checksum
|
|
col_obj.data_type = str(col.get("type", ""))
|
|
col_obj.length = col.get("max_length")
|
|
col_obj.comment = col.get("comment")
|
|
col_obj.is_nullable = col.get("nullable", True)
|
|
updated_columns += 1
|
|
col_obj.last_scanned_at = scan_time
|
|
col_obj.is_deleted = False
|
|
col_obj.deleted_at = None
|
|
|
|
total_tables += 1
|
|
|
|
# Soft-delete objects not seen in this scan
|
|
db.query(Database).filter(
|
|
Database.source_id == source.id,
|
|
Database.last_scanned_at < scan_time,
|
|
).update({"is_deleted": True, "deleted_at": scan_time}, synchronize_session=False)
|
|
|
|
for db_obj in db.query(Database).filter(Database.source_id == source.id).all():
|
|
db.query(DataTable).filter(
|
|
DataTable.database_id == db_obj.id,
|
|
DataTable.last_scanned_at < scan_time,
|
|
).update({"is_deleted": True, "deleted_at": scan_time}, synchronize_session=False)
|
|
for table_obj in db.query(DataTable).filter(DataTable.database_id == db_obj.id).all():
|
|
db.query(DataColumn).filter(
|
|
DataColumn.table_id == table_obj.id,
|
|
DataColumn.last_scanned_at < scan_time,
|
|
).update({"is_deleted": True, "deleted_at": scan_time}, synchronize_session=False)
|
|
|
|
source.status = "active"
|
|
db.commit()
|
|
|
|
return {
|
|
"success": True,
|
|
"message": "元数据同步成功",
|
|
"databases": len(db_names),
|
|
"tables": total_tables,
|
|
"columns": total_columns,
|
|
"updated_tables": updated_tables,
|
|
"updated_columns": updated_columns,
|
|
}
|
|
except Exception as e:
|
|
source.status = "error"
|
|
db.commit()
|
|
return {"success": False, "message": f"同步失败: {str(e)}", "databases": 0, "tables": 0, "columns": 0}
|
|
'''
|
|
|
|
new_content = content[:idx] + new_func
|
|
|
|
with open(sys.argv[1], 'w') as f:
|
|
f.write(new_content)
|
|
|
|
print('Patched successfully')
|