Self organized maps on mixed dataset (categoricals / numerics features )

I have a dataset of mixed variables and i want to apply self organized maps on it how can i extend som to mixed dataset? can i use the gower distance instead of euclidienne distance in order to calculate the distance between input and prototype vectors ? and for the update rule i dont know to adjust it in that case ! is there any package or an implementation that do that in python or R… please help !