makeCappedVolumeBox: Utilities to make capped volume boxes

In DelayedArray: A unified framework for working transparently with on-disk and in-memory array-like datasets

Description

makeCappedVolumeBox returns the dimensions of the biggest multidimensional box (a.k.a. hyperrectangle) that satisfies 3 constraints: (1) its volume is capped, (2) it fits in the constraining box, (3) it has the specified shape.

makeRegularArrayGridOfCappedLengthViewports makes a RegularArrayGrid object with grid elements that are capped volume boxes with the specified constraints.

These are low-level utilities used internally to support defaultAutoGrid and family.

Usage

makeCappedVolumeBox(maxvol, maxdim, shape=c("hypercube",
                                            "scale",
                                            "first-dim-grows-first",
                                            "last-dim-grows-first"))

makeRegularArrayGridOfCappedLengthViewports(refdim,
                           viewport_len,
                           viewport_shape=c("hypercube",
                                            "scale",
                                            "first-dim-grows-first",
                                            "last-dim-grows-first"))

Arguments

maxvol	The maximum volume of the box to return.
maxdim	The dimensions of the constraining box.
shape	The shape of the box to return.
refdim	The dimensions of the reference array of the grid to return.
viewport_len	The maximum length of the elements (a.k.a. viewports) of the grid to return.
viewport_shape	The shape of the elements (a.k.a. viewports) of the grid to return.

Details

makeCappedVolumeBox returns the dimensions of a box that satisfies the following constraints:

1. The volume of the box is as close as possibe to (but no bigger than) maxvol.

2. The box fits in the constraining box i.e. in the box whose dimensions are specified via maxdim.

3. The box has a non-zero volume if the constraining box has a non-zero volume.

4. The shape of the box is as close as possible to the requested shape.

The supported shapes are:

• hypercube: The box should be as close as possible to an hypercube (a.k.a. n-cube), that is, the ratio between its biggest and smallest dimensions should be as close as possible to 1.

• scale: The box should have the same proportions as the constraining box.

• first-dim-grows-first: The box will be grown along its 1st dimension first, then along its 2nd dimension, etc...

• last-dim-grows-first: Like first-dim-grows-first but starting along the last dimension.

See Also

• defaultAutoGrid and family to generate automatic grids to use for block processing of array-like objects.

• ArrayGrid for the formal representation of grids and viewports.

Examples

## ---------------------------------------------------------------------
## makeCappedVolumeBox()
## ---------------------------------------------------------------------

maxdim <- c(50, 12)  # dimensions of the "constraining box"

## "hypercube" shape:
makeCappedVolumeBox(40, maxdim)
makeCappedVolumeBox(120, maxdim)
makeCappedVolumeBox(125, maxdim)
makeCappedVolumeBox(200, maxdim)

## "scale" shape:
makeCappedVolumeBox(40, maxdim, shape="scale")
makeCappedVolumeBox(160, maxdim, shape="scale")

## "first-dim-grows-first" and "last-dim-grows-first" shapes:
makeCappedVolumeBox(120, maxdim, shape="first-dim-grows-first")
makeCappedVolumeBox(149, maxdim, shape="first-dim-grows-first")
makeCappedVolumeBox(150, maxdim, shape="first-dim-grows-first")

makeCappedVolumeBox(40, maxdim, shape="last-dim-grows-first")
makeCappedVolumeBox(59, maxdim, shape="last-dim-grows-first")
makeCappedVolumeBox(60, maxdim, shape="last-dim-grows-first")

## ---------------------------------------------------------------------
## makeRegularArrayGridOfCappedLengthViewports()
## ---------------------------------------------------------------------

grid1a <- makeRegularArrayGridOfCappedLengthViewports(maxdim, 40)
grid1a
as.list(grid1a)  # turn the grid into a list of ArrayViewport objects
table(lengths(grid1a))
stopifnot(maxlength(grid1a) <= 40)  # sanity check

grid1b <- makeRegularArrayGridOfCappedLengthViewports(maxdim, 40,
                                            "first-dim-grows-first")
grid1b
as.list(grid1b)  # turn the grid into a list of ArrayViewport objects
table(lengths(grid1b))
stopifnot(maxlength(grid1b) <= 40)  # sanity check

grid2a <- makeRegularArrayGridOfCappedLengthViewports(maxdim, 120)
grid2a
as.list(grid2a)  # turn the grid into a list of ArrayViewport objects
table(lengths(grid2a))
stopifnot(maxlength(grid2a) <= 120)  # sanity check

grid2b <- makeRegularArrayGridOfCappedLengthViewports(maxdim, 120,
                                            "first-dim-grows-first")
grid2b
as.list(grid2b)  # turn the grid into a list of ArrayViewport objects
table(lengths(grid2b))
stopifnot(maxlength(grid2b) <= 120)  # sanity check

grid3a <- makeRegularArrayGridOfCappedLengthViewports(maxdim, 200)
grid3a
as.list(grid3a)  # turn the grid into a list of ArrayViewport objects
table(lengths(grid3a))
stopifnot(maxlength(grid3a) <= 200)  # sanity check

grid3b <- makeRegularArrayGridOfCappedLengthViewports(maxdim, 200,
                                            "first-dim-grows-first")
grid3b
as.list(grid3b)  # turn the grid into a list of ArrayViewport objects
table(lengths(grid3b))
stopifnot(maxlength(grid3b) <= 200)  # sanity check

DelayedArray documentation built on March 25, 2021, 6:01 p.m.