Alra not using Scaled data

[TomLuijts]

I noticed that Alra doesn't use the scaled data for it's imputation. Is this by design?
I'm not using the default SCTransform function to normalize my data but the 'old' NormalizeData %>% FindVariableFeatures %>% ScaleData pipeline. I noticed that I get the same result from Alra regardless of whether I perform the ScaleData step.

code to reproduce:

library(Seurat)
library(SeuratData)
InstallData("stxBrain")
set.seed(123)
brain <- LoadData("stxBrain", type = "anterior1")
brain <- NormalizeData(brain , verbose = F)
brain <- FindVariableFeatures(brain , verbose = F)
set.seed(123)
brain_unscaled <- RunALRA(brain)
brain <- ScaleData(brain , verbose = F)
set.seed(123)
brain_scaled <- RunALRA(brain)

all(brain_unscaled@assays[["alra"]]@data == brain_scaled@assays[["alra"]]@data)

It seems Alra is using the data values and not the values stored in scaled.data. If this is by design, I guess I need to run ScaleData after Alra to obtain scaled expression values? If this is the case, it seems a bit misleading to me that one would think the data is scaled/centered when it isn't after Alra has run.