我采取了以下步骤:

1. 将存储过程转储结果制作到临时表中，以便将结果集展平为单个结果集.这摆脱了multi=True
的需要
2. 此外，我确保IP地址的用户可以访问数据库本身中的调用存储过程.

3. 最后，我继续研究 callproc 函数.最终，另一个站点上的某个人提出了以下有效的代码:

1. Made my stored procedure dump results into a temporary table so as to flatten the result set to a single result set. This got rid of the need for multi=True
2. In addition, I made sure the user at my IP address had access to call stored procedures in the database itself.

3. Finally, I continued to research the callproc function. Eventually someone on another site suggested the following code, which worked:

cur = connections["SomeDB"].cursor()
cur.callproc("spGetLocationPath", [id, someval])
res = next(cur.stored_results()).fetchall()
cur.close()