Package 'cmrutils'

Helper Routine

Description

Creates empty copy of object.

Usage

empty(x)

Arguments

x	An object to create copy from.

Value

Returns empty (filled with NAs) object having the same structure and attributes as x.

Examples

empty(0)
  empty(1:10)
  empty(matrix(1:4, 2, 2))
  empty(array(1:24, dim = c(2, 3, 4)))

  empty(list(a = 1, b = 2))
  empty(data.frame(a = 1:2, b = 3:4))

  empty(ts(1:10, frequency = 5))

---

Helper Plotting Routine

Description

Sets up graphical device and plots figure. For PDF device, also takes care of embedding fonts (see Details).

Usage

makefig(plotfun, ..., device = c('default', 'pdf', 'svg'),
        width = 8, height = 6, scale = pointsize/12, pointsize = 12, 
        filename = 'Rplot', family = 'Helvetica')

Arguments

plotfun	A function which does the actual plotting.
...	Additional arguments passed to plotfun.
device	The type of graphical device.
width, height	The unscaled width and height of the graphics region in inches. Actual size is c(width, height)*scale.
scale	The scaling factor. Is 1 for the default pointsize.
pointsize	The base font size in points.
filename	The filename for PDF or SVG device. If "standard" extension is missing, it is added automagically.
family	The font family for PDF device (see Details).

Details

For PDF device, function does font embedding via call to embedFonts. The package provides additional font families "CMRoman", "CMSans" which work nicely with cyrillics. These families are based on CM-Super font package by Vladimir Volovich and CMSYASE symbol font by Paul Murrell.

Both metric (‘*.afm’) and outline (‘*.pfb’) files for the "CMRoman", "CMSans" fonts are provided with the package. This allows to embed these fonts even if they are not installed in the system.

Value

Returns the result value of call to plotfun invisibly.

See Also

dev.new, pdf, svg,

Paul Murrell (2005?) Using Computer Modern Fonts in R Graphics. http://www.stat.auckland.ac.nz/~paul/R/CM/CMR.html

Examples

x <- rnorm(10)

  myplot <- function(x, legend, ...)
  {
    plot(x, ..., pch = 1)
    legend('bottomright', bg = 'white',
           legend = legend, pch = 1)
  }

  makefig(myplot, x = x, legend = 'Simulation')
  makefig(myplot, x = x, legend = 'Simulation', pointsize = 8)


  makefig(myplot, x = x, legend = 'Simulation',
          device = 'pdf', filename = 'embedding.pdf', family = 'CMSans')

  mapply(makefig, device = c('window', 'pdf', 'svg'),
         MoreArgs = list(plotfun = myplot, x = x, legend = 'Simulation',
           filename = 'test', family = 'CMSans'))

---

Helper Routine

Description

Replaces non-finite values with NAs in array-like objects.

Usage

nan2na(x)

Arguments

x	An array-like object.

Value

Returns x with non-finite values replaced with NAs.

See Also

is.finite, NA.

Examples

nan2na(Inf)
  nan2na(rep(c(0, -Inf, Inf), 3))
  nan2na(matrix(c(0, Inf, -Inf, 0), 2, 2))
  nan2na(array(c(0, -Inf, Inf, 1, NaN), dim = c(2, 3, 4)))
  nan2na(ts(rep(c(0, -Inf, Inf), 2), frequency = 5))

---

Helper Routine

Description

Slices array by specified dimension.

Usage

slice(x, MARGIN, n)

Arguments

x	An array-like object.
MARGIN	An integer giving the dimension to slice by.
n	Number of slice.

Value

Returns array of dimensions dim(x)[-MARGIN]. The corresponding dimnames are preserved.

See Also

slice.index.

Examples

x <- 1:10
  slice(x, 1, 1)

  x <- matrix(1:4, 2, 2)
  rownames(x) <- c("Row 1", "Row 2")
  colnames(x) <- c("Col 1", "Col 2")
  x
  slice(x, 1, 1)
  slice(x, 1, 2)
  slice(x, 2, 1)
  slice(x, 2, 2)

  x <- array(1:24, dim = c(2, 3, 4))
  dimnames(x) <- list(letters[1:2], letters[3:5], letters[6:9])
  x
  slice(x, 1, 1) 
  slice(x, 2, 1)
  slice(x, 3, 1)

---

Helper Time-Series Routines

Description

Converts beetween Date objects and sampling times of daily time series with the natural time period of 1 year.

Usage

time365toDate(x)
time365fromDate(x)

Arguments

x	A vector of sampling times (time365toDate) or dates (time365fromDate).

Details

A simple approach to modelling daily time series with the natural time period of 1 year is to use ts object with $\code{frequency} = 365$, that is to assume no leap days.

time365fromDate returns NA for the leap day February 29.

Value

A vector of dates (time365fromDate) or sampling times (time365toDate).

See Also

time.

Examples

## non-leap year
  x <- ts(1:4, start = c(1991, 58), frequency = 365)
  time365toDate(time(x))
  d <- seq.Date(as.Date('1991-02-27'), as.Date('1991-03-02'), 1)
  time365fromDate(d)

  ## leap year
  x <- ts(1:4, start = c(1992, 58), frequency = 365)
  time365toDate(time(x))
  d <- seq.Date(as.Date('1992-02-27'), as.Date('1992-03-02'), 1)
  time365fromDate(d)

  x <- seq(1900, 2100, by = 1/365)
  stopifnot(all(abs(x - time365fromDate(time365toDate(x))) <
                getOption("ts.eps")))

---