dgCMatrix Support for get.exp.stat function

[Cosmic1Neko]

Hello,
Thank you for this amazing package！
The current implementation of the get.exp.stat function can be inefficient when dealing with sparse matrices (dgCMatrix). Specifically, the line:

BayesPrism/BayesPrism/R/process_input.R

Line 185 in eeb6a9c

dat.tmp <- log2(dat.tmp + pseudo.count) - log2(pseudo.count)

If modified to the following code, it can avoid converting the dgCMatrix to a dense matrix when the input is a dgCMatrix:

dat.tmp@x <- log2(dat.tmp@x + pseudo.count) - log2(pseudo.count) 

Thank you for considering this suggestion.