Your reasoning is correct. There is a bug. This is because we calculate the SD and SEM directly
using Pandas tools instead of considering propagation
We need to fix this. Thanks for reporting it.
You can create an issue to follow upI don't have time to work on this at the moment, so any PR is welcome.
In future releases, we're considering major changes to this API to enhance data processing efficiency, improve the responsiveness and speed of the GUI, and optimize data storage.