R/group_rates.R

In rafalab/smallcount: Methods for Small Counts Stored in a Sparse Matrix

#' Row-wise Rates for Groups
#'
#' @param y Sparse matrix (can be a matrix, dgCMatrix, or SparseMatrix)
#' @param g Factor specifying the group for each column
#' 
#' @return Row-wise rates for each group
#'
#' @importFrom SparseArray nzvals nzwhich
#' @export
groupRates <- function(y, g) {
    y <- .convertToSparse(y)

    if (!is.factor(g)) {
        warning("Coercing g into a factor")
        g <- as.factor(g)
    }

    # Compute row sums for each group
    nz_ind <- nzwhich(y)
    nz_rows <- .nzrows(y, nz_ind)
    nz_cols <- .nzcols(y, nz_ind)
    group_sums <- tapply(nzvals(y),
        list(
            factor(nz_rows, levels = seq_len(nrow(y))),
            g[nz_cols]
        ),
        sum,
        default = 0
    )

    # Standardize column sums to 1
    rates <- sweep(group_sums, 2, colSums(group_sums), FUN = .safeDivide)
    colnames(rates) <- levels(g)
    rownames(rates) <- rownames(y)
    return(rates)
}