Common.meta: Common meta data structure in cranvas
In ggobi/cranvas: Interactive statistical graphics based on Qt

Description Usage Format Details Author(s) Examples

A meta data is created with reference classes (setRefClass) in the methods and signaling fields in the objectProperties package (properties). It is used to make several plotting routines more modularized (e.g. the axes and background layers).

1	Common.meta

List of 26
 $ alpha          : chr "numeric"
 $ main           : chr "character"
 $ xlim           : chr "numeric"
 $ ylim           : chr "numeric"
 $ xat            : chr "numeric"
 $ yat            : chr "numeric"
 $ xlab           : chr "character"
 $ ylab           : chr "character"
 $ xlabels        : chr "character"
 $ ylabels        : chr "character"
 $ limits         : chr "matrix"
 $ color          : chr "character"
 $ border         : chr "character"
 $ size           : chr "numeric"
 $ start          : chr "numeric"
 $ pos            : chr "numeric"
 $ active         : chr "logical"
 $ brush.move     : chr "logical"
 $ brush.size     : chr "numeric"
 $ brush.adj      : chr "numeric"
 $ manual.brush   : chr "function"
 $ minor          : chr "character"
 $ keys           : chr "character"
 $ identified     : chr "integer"
 $ identify.labels: chr "character"
 $ handlers       : chr "list"

This object only provides a template for the common meta data, we have to call properties in order to really create a meta object.

We can use Meta.template$new() to generate an object, which is essentially an environment. This environment object is dynamic in the sense that no matter where we copy it to or how we modify the components in it, all its copies can sync with each other. This free us from considering lexical scopes frequently when we do interactive graphics, which often involves with global or local changes of variables.

Another advantage is that we can attach events on this object. When a component is changed, an event can be triggered (see examples below).

The common meta structure contains the following components:

alpha: the alpha transparency
main: the main title
xlim, ylim: the real x- and y-axis limits calculated from original data
xat, yat: the tickmarks locations on the x and y axis
xlab, ylab: x and y axis titles
xlabels, ylabels: the tickmarks labels
minor: possible values are '', 'x', 'y' and 'xy', indicating which axis to draw minor grid lines
limits: the limits of the main plot layer (it is often the case that some other layers need to sync with the main plot layer in limits, so this component is very important)
color, border, size: the real color, border and size vectors used to draw the graphical elements, which can be different from the corresponding columns in the original data
start, pos: the starting position and the current position of the cursor
brush.move: indicating if the brush is in the move mode (if not, the brush will be resized when mouse moves
brush.size: the size of the brush rectangle (a numeric vector of length 2)
brush.adj: the amount of x and y adjustment for the brush rectangle to capture a graphical element without having to cover its center
manual.brush: a function to manually brush the plot given the mouse position
identified, identify.labels: the identified indices and the text label to draw in the plot for the identified cases
handlers: a list of event handlers (functions of two arguments: layer and event)
active: logical: if the current plot window is active

Yihui Xie <http://yihui.name>

library(objectSignals)
My.meta <- setRefClass("My_meta", fields = objectProperties::properties(c(Common.meta, 
    list(horizontal = "logical", xleft = "numeric"))))
meta <- My.meta$new(alpha = 1)
meta$alpha
meta$xat <- 1:5
meta$pos <- c(1, 4)
meta$horizontal <- FALSE

meta2 <- meta
meta2$xat <- 2:6
meta$xat  # meta is changed too

meta$posChanged$connect(function() {
    message("the mouse position is (", paste(meta$pos, collapse = ","), ") now")
})
meta$pos <- c(2, 0)

ls(meta)  # list all the objects in it