StratigrapheR (version 1.3.1)

neatPick: Interactive user modification of the arguments of a repeated function

Description

This opens a shiny app that will allow to manipulate the arguments of a function interactively, with different conditions that the user can provide a priori and modify at will

Usage

neatPick(
  fun,
  n,
  args = list(),
  class.args = list(),
  pick = NA,
  fix = NA,
  buttonswidth = 2,
  text = "output",
  textwidth = 4,
  plotwidth = 800,
  plotheight = 600,
  args.only = F,
  width = 10,
  height = 10,
  name = "fig",
  dir = tempdir(),
  gfile = "onePDF",
  openfile = TRUE,
  folder = "Rfig",
  gfun = "jpeg",
  ext = ".jpeg",
  gargs = list(units = "in", res = 300),
  pargs = list(ps = 12, cex = 1.5)
)

Arguments

fun

the function to be applied n times.

n

number of runs.

args

the arguments to be supplied to fun. Should be a list of each argument to be supplied to fun, having n elements stored indiscriminately in list or in vector form.

class.args

the class of the arguments, in a list. This is useful when the starting arguments are NA

pick

which arguments to be able to adapt interactively

fix

which arguments that cannot be chosen interactively (if pick is NA)

buttonswidth

the width of the buttons panel (integer from 1 to 12)

text

which information to send to the text panel. The default is the output of the current element (ni); "output". Can be the whole dataset of arguments; "all". Otherwise the panel does not show.

textwidth

the width of the text panel (integer from 1 to 12)

plotwidth

the width of the plot panel (arbitrary units)

plotheight

the height of the plot panel (arbitrary units)

args.only

whether to be only allowed to download and return the arguments (this simplifies things and makes the workflow more efficient)

width, height, name, dir, gfile, openfile, folder, gfun, ext, gargs

arguments to be supplied to neatPicked, the equivalent of neatPick without interactivity: it runs the function for each ni and saves the output (normal and graphical). In neatPick this happens when the button 'Run and Download Output' is clicked. See ?neatPicked function help page for details.

pargs

the arguments to transmit to par(), in neatPick and neatPicked

Details

This is a complicated function. A few basics:

neatPick works using the formals() function. It evaluates the arguments and their default values of any function that you provide without parentheses, like this for instance: formals(multigons).

neatPick is capable of providing interaction with arguments of class integer or numeric (e.g. 10, or 13,58745), character (e.g. "BlipBlapBLoup") and logical (T or F), as long as for each iteration (n) the length of the argument is one (you cannot use arguments like xlim = c(0,1), however you can use xmin = 0 and xmax = 1 for instance). But you can provide a different value for each iteration n (if n = 3, you can provide col = c("red", "blue", "green") in the args list of arguments)

You can chose which arguments are interactive or not using the 'pick' and 'fix' arguments.

To return the arguments or the output, you have to click on 'End & Return Arguments' or 'End & Return Output', respectively.

You can also save the obtained output and arguments via the download buttons: you geta .RData file were the arguments are in the object saved.arguments and the output is in the saved.output object. The arguments can also be found at saved.output$args. The arguments can be provided to the args argument of the same neatPick function to rework the changes you made.

Examples

if (FALSE) {
# You create a simple function. The one below creates sinusoidal waves between
# x0 = 0 and x1 = 1. You want to personalise the amplitude (delta), the y
# offset (pos, see ?sinpoint for more details), the phase (phase, expressed
# in multiples of pi), the number of waves between x0 and x1, and the number
# of intervals between each discrete point (nint).
# So you set all these as arguments of the function. This function can also
# have a graphical output of one plot (which can be subdivided if necessary
# using par(mfrow)). And the function can return output.

fun <- function(delta = 1, pos = 1, phase = 1.5, nwave = 1, nint = 50)
{

  res <- sinpoint(1, 0, delta = delta, pos = pos, phase = phase,
                  nwave = nwave, nint = nint)

  plot(res$x,res$y)

  return(res)

}

# Once this simple function is coded, it can be integrated to neatPick(). The
# argument n defines to number of different realisations of the function.

# WHEN YOU ARE HAPPY WITH THE OUTPUTS, click on 'END & RETURN ARGUMENTS'

a <- neatPick(fun, n = 10, args.only = TRUE)

# If you have clicked right (on the 'END & RETURN ARGUMENTS' button), the
# arguments will be returned and stored in a;

a

# These arguments can then serve for a more efficient function:

seg <- sinpoint(1, 0, delta = a$delta, pos = a$pos, phase = a$phase,
                nwave = a$nwave, nint = a$nint)

# Basically neatPick applies a for loop to fun, but if you work on a large
# dataset, you can also create a function that can handle the arguments more
# efficiently. This is what sinpoint does here

# Now you can see the results imported in R and do whatever you want with:

plot(seg$x, seg$y, type = "n")

multilines(seg$i, seg$x, seg$y)

# You can even rework your initial changes:

b <- neatPick(fun, n = 10, args.only = TRUE, args = a)}