Plots: 'Plot' wrapper intended for use in a SpaDES module

View source: R/Plots.R

PlotsR Documentation

Plot wrapper intended for use in a SpaDES module

Description

This is a single function call that allows a user to change which format in which the plots will occur. Specifically, the two common formats would be to "screen" or to disk as an image file, such as "png". This has currently been tested with ggplot2, RasterLayer, and tmap objects. The default (or change with e.g., ⁠fn = "print", usePlot = FALSE⁠) uses Plot internally, so individual plots may be rearranged. When saved to disk (e.g., via type = 'png'), then Plot will not be used and the single object that is the result of this Plots call will be saved to disk. This function requires at least 2 things: a plotting function and arguments passed to that function (which could include data, but commonly would simply be named arguments required by fn). See below and examples.

Usage

Plots(
  data,
  fn,
  filename,
  types = quote(params(sim)[[currentModule(sim)]]$.plots),
  path = quote(figurePath(sim)),
  .plotInitialTime = quote(params(sim)[[currentModule(sim)]]$.plotInitialTime),
  ggsaveArgs = list(),
  usePlot = getOption("spades.PlotsUsePlot", FALSE),
  deviceArgs = list(),
  ...
)

Arguments

data

An (optional) arbitrary data object. If supplied, it will be passed as the first argument to Plot function, and should contain all the data required for the inner plotting. If passing a RasterLayer, it may be a good idea to set names(RasterLayer) so that multiple layers can be plotted without overlapping each other. When a custom fn is used and all arguments for fn are supplied and named, then this can be omitted. See examples.

fn

An arbitrary plotting function. If not provided, defaults to using quickPlot::Plot

filename

A name that will be the base for the files that will be saved, i.e, do not supply the file extension, as this will be determined based on types. If a user provides this as an absolute path, it will override the path argument.

types

Character vector, zero or more of types. If used within a module, this will be deduced from the P(sim)$type and can be omitted. See below.

path

Currently a single path for the saved objects on disk. If filename is supplied as an absolute path, path will be set to dirname(filename), overriding this argument value.

.plotInitialTime

A numeric. If NA then no visual on screen. Anything else will have visuals plotted to screen device. This is here for backwards compatibility. A developer should set in the module to the intended initial plot time and leave it, i.e., not NA.

ggsaveArgs

An optional list of arguments passed to ggplot2::ggsave

usePlot

Logical. If TRUE, the default, then the plot will occur with quickPlot::Plot, so it will be arranged with previously existing plots.

deviceArgs

An optional list of arguments passed to one of png, pdf, tiff, bmp, or jgeg. This is useful when the plotting function is not creating a ggplot object, e.g., plotting a RasterLayer.

...

Anything needed by fn, all named.

Details

  • type

    • "screen" – Will plot to the current device, normally a plot window

    • "object" – Will save the plot object, e.g., ggplot object

    • "raw" – Will save the raw data prior to plotting, e.g., the data argument

    • "png" – or any other type save-able with ggsave

Value

Called for its side effect of plot creation.

Recording of files saved

In cases where files are saved, and where Plots is used within a SpaDES module, the file(s) that is/are saved will be appended to the outputs slot of the simList of the module. This will, therefore, keep a record of figures saved within the simList

Note

This is still experimental and could change in the next release.

Plots now has experimental support for "just a Plot call", but with types specified. See examples. The devices to save on disk will have some different behaviours to the screen representation, since "wiping" an individual plot on a device doesn't exist for a file device.

This offers up to 4 different actions for a given plot:

  • To screen device

  • To disk as raw data (limited testing)

  • To disk as a saved plot object (limited testing)

  • To disk as a ‘.png’ or other image file, e.g., ‘.pdf

To turn off plotting both to screen and disk, set both .plotInititalTime = NA and .plots = NA or any other value that will not trigger a TRUE with a grepl with the types argument (e.g., "" will omit all saving).

Examples


  # Note: if this is used inside a SpaDES module, do not define this
  #  function inside another function. Put it outside in a normal
  #  module script. Otherwise, it will cause a memory leak.
  if (requireNamespace("ggplot2")) {
    fn <- function(d)
      ggplot2::ggplot(d, ggplot2::aes(a)) +
      ggplot2::geom_histogram()
    sim <- simInit()
    sim$something <- data.frame(a = sample(1:10, replace = TRUE))

    Plots(data = sim$something, fn = fn,
          types = c("png"),
          path = file.path("figures"),
          filename = tempfile(),
          .plotInitialTime = 1
          )

    # plot to active device and to png
    Plots(data = sim$something, fn = fn,
          types = c("png", "screen"),
          path = file.path("figures"),
          filename = tempfile(),
          .plotInitialTime = 1
          )

    # Can also be used like quickPlot::Plot, but with control over output type
    r <- terra::rast(terra::ext(0,10,0,10), vals = sample(1:3, size = 100, replace = TRUE))
    Plots(r, types = c("screen", "png"), deviceArgs = list(width = 700, height = 500),
          usePlot = TRUE)

    # with ggplotify, Plots can also be used to plot/save
    # non-ggplot objects:


    if (require("ggplotify")) {
      if (!require("lattice")) stop("please install lattice")

      plotFile <- tempfile()

      p1 <- densityplot(~mpg|cyl, data=mtcars)
      Plots(data = p1, fn = as.ggplot, filename = plotFile,
            ggsaveArgs = list(width = 5, height = 4, dpi = 300, bg = "white", units = "in"),
            types = c("screen", "png"), .plotInitialTime = 1)
    }
  } # end ggplot
 # end of dontrun

SpaDES.core documentation built on June 22, 2024, 10:32 a.m.