liftCellChat alters weight values dependent on group.new vector order

[BornToBeJan]

Hi,

I noticed a problem in liftCellchat(). When I use the function to lift a cellchat object by using the rownames() vector of its own weight matrix for group.new, it does not change anything, which is as expected.

However, if I change the order of values in the rownames vector and use it for group.new, I get completely different weight values, even the values in the group.new vector are still the same, only in different order.

Can someone tell me if this is intended or is this a bug?

[BornToBeJan]

I think it is a bug, because the new weight value of 0.744... is also part of the original matrix. it seems like the function only swaps the matrix indices and keeps the values at the same place, which totally screws up the data.