Introduction

All of loon's displays have plot states. Plot states specify what is displayed, how it is displayed and if and how the plot is linked with other loon plots. loon's plot states are derived from Tk's configurable options. A large part of loon's framework revolves around modifying states, tracking state changes and the synchronization of plot states between plots.

For example, the plot states of the scatterplot display include x, y, color, size, selected, xlabel, ylabel, zoomX, zoomY, panX, panY, showScales and showGuides. The scatterplot display has more than 30 states.

To get a complete list of the plot states for a particular loon widget use the l_info_states function

p <- l_plot(iris[,1:2])

istates <- l_info_states(p)
names(istates)
# or together as
names(p)

l_info_states(p, 'x')
# or
istates$x

l_info_states(p, c('x', 'y', 'xTemp'))

When possible then the data structure for each state is either a scalar or a flat vector. One exception is the data state which contains a data.frame.

Query and Modify

To query a state, say showScales, of the plot p use either the accessor method [ or the l_cget function

p['showScales']
l_cget(p, 'showScales')

To modify a single state, say showLabels, use one of

p['showLabels'] <- FALSE
l_configure(p, showLabels=FALSE)

to modify multiple states, say showLabels and showScales, use

l_configure(p, showScales=FALSE, showLabels=FALSE)

Note that you should never use the tkconfigure function (defined in the tcltk package) instead of the l_configure function! One reason is that the l_configure function is customized for loon plots and takes care of some R to Tcl data structure conversions that are otherwise not supported (e.g. data.frame and nested lists).

State Dimension

The dimension of a state is either explicit, i.e. a number, or abstract, i.e. a letter.

• Abstract dimensions take on a value at plot creation time.

• The value of abstract dimensions can be changed when changing the dominant states for a dimension together. The dominant states for the displays are: x for the histogram, x and y for the scatterplot, data for the serialaxes display, and for the graph display nodes is dominant for n dimensional states and from and to are dominant for the p dimensional states.

    p <- l_plot(iris, color=iris$Species)
    p['n']
    data(olive)
    with(olive, l_configure(p, x=oleic, y=stearic, color=Area))
    l_scaleto_world(p)
    p['n']

• Changing an abstract dimension will assign default values to all non-dominant states that have that dimension.

• When assigning a single value to a state that has an abstract dimension then that value gets repeated accordingly. For example, for the n dimensional state selected

      p <- l_plot(iris)
      p['selected'] <- TRUE

  will repeat TRUE n=150 times.

• It is possible to modify a subset of a state with abstract dimension using the corresponding which argument

      p <- l_plot(iris)
      l_configure(p, size=15, which_n=as.character(iris$Species) == "setosa")

  where which_* accepts logical and index sub-setting. In R indexing starts at 1. which_* also accepts a state name that is of type Boolean and has the same abstract dimension. l_configure(p, active=FALSE, which_n='selected')

  Note that which_n is equivalent to which.

• Note that when switching between R and Tcl: in R vector indexing starts at 1 and in Tcl at 0. loon will honor the indexing origin in R and Tcl depending on where you use it.