Improve scaling function

[njohnso6]

See gitlab commit entitled "Improved Scaling Function". Essentially, not only can the scaling function be vectorized, but the heatmap function can be passed any function for normalizing.
This is the function that should be used to normalize:
normalize <- function(x) {
x <- as.matrix(x)
minAttr=apply(x, 1, min)
maxAttr=apply(x, 1, max)
x <- sweep(x, 1, minAttr, FUN="-")
x=sweep(x, 1, maxAttr-minAttr, "/")

return (x)
}
Taken from Lui Kimman (https://stats.stackexchange.com/users/77829/lui-kimman), How to normalize data to 0-1 range?, URL (version: 2015-05-22): https://stats.stackexchange.com/q/153544
Modified to go by row instead of column.