Map reference array positions to grid positions and vice-versa
Use mapToGrid() to map a set of reference array positions to
grid positions.
Use mapToRef() for the reverse mapping.
mapToGrid(Mindex, grid, linear=FALSE) mapToRef(major, minor, grid, linear=FALSE)
Mindex |
An M-index containing absolute positions, that is,
positions with respect to the underlying array (i.e. to the reference
array of For convenience, Note that no bounds checking is performed, that is, values in the j-th
column of |
grid |
An ArrayGrid object. |
linear |
By default (i.e. when |
major, minor |
The |
For mapToGrid(): A list with 2 components, major
and minor.
Each row in input matrix Mindex is an n-uplet that contains
the coordinates of an absolute position.
By default (i.e. when linear is FALSE), the 2
components of the returned list are integer matrices of the same
dimensions as the input matrix. A row in the major (or
minor) matrix is called a "major n-uplet" (or "minor n-uplet").
So for each "input position" (i.e. for each row in the input matrix),
2 n-uplets are returned: the "major n-uplet" and the "minor n-uplet".
The "major n-uplet" contains the coordinates of the "input position"
in the grid coordinate system, that is, the coordinates of the
grid element where the position falls in.
The "minor n-uplet" represents where exactly the "input position"
falls inside the grid element reported by the "major n-uplet".
The coordinates in the "minor n-uplet" are relative to this
grid element.
When linear is TRUE, the major and minor
components are returned as linear indices. In this case, both are
integer vectors containing 1 linear index per "input position".
For mapToRef(): A numeric matrix like one returned
by base::arrayInd describing positions
relative to the reference array of grid.
ArrayGrid for the formal representation of grids and viewports.
Lindex2Mindex and Mindex2Lindex in
this package (DelayedArray) for converting back and forth
between linear indices and matrix indices.
array objects in base R.
## Create an arbitrary-spaced grid on top of a 15 x 9 matrix:
grid2 <- ArbitraryArrayGrid(list(c(2L, 7:10, 13L, 15L), c(5:6, 6L, 9L)))
## Create a set of reference array positions:
Mindex <- rbind(c( 2, 5), # bottom right corner of 1st grid element
c( 3, 1), # top left corner of 2nd grid element
c(14, 9), # top right corner of last grid element
c(15, 7), # bottom left corner of last grid element
c(15, 9)) # bottom right corner of last grid element
## Map them to grid positions:
majmin <- mapToGrid(Mindex, grid2)
majmin
## Reverse mapping:
Mindex2 <- mapToRef(majmin$major, majmin$minor, grid2)
stopifnot(all.equal(Mindex2, Mindex))
majmin <- mapToGrid(Mindex, grid2, linear=TRUE)
majmin
Mindex2 <- mapToRef(majmin$major, majmin$minor, grid2, linear=TRUE)
stopifnot(all.equal(Mindex2, Mindex))
## Map all the valid positions:
all_positions <- seq_len(prod(refdim(grid2)))
Mindex <- arrayInd(all_positions, refdim(grid2))
majmin <- data.frame(mapToGrid(Mindex, grid2, linear=TRUE))
majmin
## Sanity checks:
min_by_maj <- split(majmin$minor,
factor(majmin$major, levels=seq_along(grid2)))
stopifnot(identical(lengths(min_by_maj, use.names=FALSE), lengths(grid2)))
stopifnot(all(mapply(isSequence, min_by_maj, lengths(min_by_maj))))
Mindex2 <- mapToRef(majmin$major, majmin$minor, grid2, linear=TRUE)
stopifnot(identical(Mindex2, Mindex))
## More mapping:
grid4 <- RegularArrayGrid(c(50, 20), spacings=c(15L, 9L))
Mindex <- rbind(c( 1, 1),
c( 2, 1),
c( 3, 1),
c(16, 1),
c(16, 2),
c(16, 10),
c(27, 18))
mapToGrid(Mindex, grid4)
mapToGrid(Mindex, grid4, linear=TRUE)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.