The R Project SVN R-packages

Rev

Rev 4444 | Blame | Compare with Previous | Last modification | View Log | Download | RSS feed

### Define Methods that can be inherited for all subclasses


setAs("dMatrix", "matrix",
      function(from) as(as(from, "dgeMatrix"), "matrix"))

##-> "dMatrix" <--> "lMatrix"   ---> ./lMatrix.R

## these two are parallel to "n <-> l" in the above :
setAs("nMatrix", "dMatrix",
      function(from) {
      cl <- class(from)
      nCl <- sub("^n", "d", cl)
      r <- new(nCl)# default => no validity check; and copy slots:
      ## result is "same", for sparse just with an 'x' slot
      if(extends(cl, "sparseMatrix"))
          slot(r, "x") <- rep.int(1., nnzero(from))
      for(nm in slotNames(cl))
          slot(r, nm) <- slot(from, nm)
      r
      })

setAs("dMatrix", "nMatrix",
      function(from) {
      if(any(is.na(from@x)))
          stop("\"dMatrix\" object with NAs cannot be coerced to \"nMatrix\"")
      ## i.e. from@x are only TRUE (or FALSE in dense case)
      cl <- class(from)
      if(extends(cl, "diagonalMatrix")) # have no "ndi*" etc class
          cl <- class(from <- as(from, "sparseMatrix"))
      nCl <- sub("^d", "n", cl)
      sNams <- slotNames(if(extends(cl, "sparseMatrix")) .sp.class(cl) else cl)
      r <- new(nCl)# default => no validity check; and copy slots:
      for(nm in sNams)
          slot(r, nm) <- slot(from, nm)
      r
      })


## Methods for operations where one argument is integer
## No longer made use of (and confusing hence) since R version 2.1.0
## where "integer" goes as part of "numeric"

## Note: Use as.matrix() {not directly array()} :
##  1) to ensure consistency with "numeric" (non-matrix)
##  2) names -> dimnames {potentially}
## setMethod("%*%", signature(x = "dMatrix", y = "integer"),
##           function(x, y) callGeneric(x, as.numeric(y)))

## setMethod("%*%", signature(x = "integer", y = "dMatrix"),
##           function(x, y) callGeneric(as.numeric(x), y))

## setMethod("crossprod", signature(x = "dMatrix", y = "integer"),
##           function(x, y = NULL) callGeneric(x, as.numeric(y)))

## setMethod("crossprod", signature(x = "integer", y = "dMatrix"),
##           function(x, y = NULL) callGeneric(as.numeric(x), y))

## setMethod("solve", signature(a = "dMatrix", b = "integer"),
##           function(a, b, ...) callGeneric(a, as.numeric(b)))

setMethod("expm", signature(x = "dMatrix"),
          function(x) callGeneric(as(x, "dgeMatrix")))


## Group Methods, see ?Arith (e.g.)
## -----
## >>> More specific methods for sub-classes (sparse), use these as "catch-all":

## the non-Ops ones :
setMethod("Math2",
          ## Assume that  Generic(u, k) |--> u for u in {0,1}
          ## which is true for round(), signif() ==> all structure maintained
          signature(x = "dMatrix", digits = "numeric"),
      function(x, digits) {
              x@x <- callGeneric(x@x, digits = digits)
              x
          })

## round(x) == round(x, 0)  etc
setMethod("Math2",
      signature(x = "dMatrix", digits = "missing"),
      function(x, digits) callGeneric(x, digits = 0))

## This needs extra work in ./AllGeneric.R :
setMethod("Summary", signature(x = "dMatrix", na.rm = "ANY"),
          function(x, ..., na.rm) callGeneric(x@x, ..., na.rm = na.rm))

## "Ops" ("Arith", "Compare", "Logic") --> ./Ops.R

## -- end{group generics} -----------------------




## Methods for single-argument transformations

setMethod("zapsmall", signature = list(x = "dMatrix"),
          function(x, digits = getOption("digits")) {
              x@x <- zapsmall(x@x, digits)
              x
          })

## -- end(single-argument transformations) ------